INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quaiity of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
ProQuest Information and Learning 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 800-521-0600 UMI
NOTE TO USERS
Page(s) not included in the original manuscript and are unavailable from the author or university. The manuscript was microfilmed as received.
121
This reproduction is the best copy availabie.
UMI
PHOTOCHEMICAL STUDIES OF ZEOLITE-BASED SYSTEMS
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
Anand S. Vaidyalingam. M.S.
*****
The Ohio State University 2001
Dissertation Committee: Approved by Professor Prabir K. Dutta. Adviser
Professor Richard L. McCreery Adviser Professor Patrick M. Woodward Department of Chemistry UMI Number; 3031278
UMI’
UMI Microform 3031278 Copyright 2002 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 ABSTRACT
This work is aimed at building a zeolite-based hydrogen evolving artificial photosynthetic system. Toward this goal, we have adopted the well-known sacrificial D-
S-A model. EDTA-Ru(bpy) 3 -'-bipyridinium-catalyst scheme. The stability of the
sensitizer. RuCbpylj-' was evaluated under photolytic conditions by developing a chromatographic method to separate, quantitate and identify the decomposition products.
Effect of various photolytic parameters like zeolite encapsulation. pH. photolysis time,
quencher concentration, nature and concentration of buffers were studied to find the
optimum conditions where the decomposition could be minimized. It was found that the
extent of decomposition is dependent on the nature and concentration of the buffer anion
and decreases with increasing quenching efficiency in the presence of quenchers.
It is well known that the electron acceptor employed in this system, bipyridinium
is susceptible to decomposition under hydrogen evolving photolytic conditions, limiting
the practical applicability. The catalyst involved in the electron transfer from the relay is
widely believed to be responsible for this limiting condition. Our research group has
earlier developed a zeolite-based RuGi catalyst for the photooxidation of water to
oxygen. This catalyst was evaluated for the photoreduction of water to hydrogen using
the above-mentioned system and was found to be effective. Performance of this catalyst was compared with other known catalysts as for as rate of hydrogen evolution, stability and the rate of bipyridinium decomposition. Attempts were made to selectively poison the sites responsible for the bipyridinium reduction, while keeping the sites involved in the hydrogen evolution unperturbed. The sensitizer was also covalently linked to the zeolitic surface that also contains the catalyst as a first step toward a proposed integrated solar water splitting assembly.
Intrazeolitic electron transfer was studied with zeolite encapsulated Ru(bpy) 3 -* and bipyridinium ions. Molecular modeling of guest species in constrained zeolitic cages involved in photoelectron transfer allowed to gain insight into their restricted rotational and diffiisional mobility. A systems dynamics modeling approach was employed to simulate the intrazeolitic electron transfer processes and the developed model allowed the extraction of various kinetic parameters. We propose that the zeolite architecture plays a crucial role in aiding long-lived charge separation.
Ill Dedicated to my mother
Smt. V. Mangaiyarkarasi
IV ACKNOWLEDGMENTS
I am grateful to my adviser. Prof. Prabir Dutta. for his guidance, support and encouragement, and especially for his patience without which this thesis would not have been possible. His work ethic and dedication are truly inspirational.
1 thank Samar Das and Norma Castagnola for all their help, ideas and suggestions.
1 wish to thank Marcello Vitale and Nancy Ortins-Savage for obtaining time-resolved diffuse reflectance spectra. I also wish to thank all the past and present group members.
Mario Castagnola. Michael Coûtant. Amitava Das. John Doolittle. Estelle Each. Marla
Frank. Arwa Ginwalla. Astrid Guglielmi. Brian Hogg. Yanghee Kim. Bob Kristovich.
Tracy Krueger. Ty Le. Hyunjung Lee. Zhaohui Lei. Vincent Maloney. Kefa Onchoke.
Pramatha Payra. Ramachandra Rao. Ramsharan Singh. Nick Szabo and Joe Trimboli for their help, support and for putting up with me during my stay in the Dutta research group.
I am grateful to Dr. Gordon Renkes for training me on various analytical instruments and to Tim Henthome for his help with excellent glass blowing.
I would like to thank David Amsbary. my roommate for four years, for his friendship and support. I would also like to thank my family for their love, support and patience. VITA
March 29. 1972 ...... Bom - Pudukottai. India.
1992 ...... B.Sc.. Chemistry. University of Madras.
1995 ...... M.Sc., Organic Chemistry Indian Institute of Technology. Bombay.
1995 -present ...... Graduate Teaching and Research Assistant. The Ohio State University.
PUBLICATIONS
A. S. Vaidyalingam, M. A. Coûtant, P. K. Dutta, “Electron-transfer Processes in Zeolites and Related Microheterogeneous Media", in Electron Transfer in Chemistry. V. Balzani (Ed); WILEY-VCH, Weinheim. Germany. 2001. Vol 4. 412.
2. A. Vaidyalingam. P. K. Dutta, “Analysis of the photodecomposition products of
Ru(bpy)3 ' ‘ in various buffers and upon zeolite encapsulation"..-Inai. Chem. 2000. 72. 5219.
3. M. Vitale. N. B. Castagnola. N. J. Ortins. J. A. Brooke, A. Vaidyalingam. P. K. Dutta. “Intrazeolitic Photochemical Charge Separation for
Ru(bpy)3 -*-Bipyridinium System: Role of the Zeolite Structure". J. Phys. Chem. B 1999. 103, 2408.
VI 4. B. D. Hogg, P. K. Dutta, J. F. Long, A. Vaidyalingam, “Carcinogenicity of mineral erionite fibers: measurements and hypothesis of activity”, in Proc. Int. Zeolite Conf, 12th Meeting Date 1998. Treacy, M. M. J (Ed); Materials Research Society, Warrendale, Pa. 1999. Vol 4, 2927.
5. Z. Lei, A. Vaidyalingam, P. K. Dutta, “Photochemistry of azobenzene in microporous aluminophosphate AlPOj-S", J. Phys. Chem. B 1998, 102. 8557.
FIELDS OF STUDY
Major Field: Chemistry
Minor Field: Analytical Chemistry
Vll TABLE OF CONTENTS
Page
Abstract ...... n
Dedication ...... iv
Acknowledgments ...... v
V ita...... vi
List of Tables...... xü
List of Figures ...... xiü
List of Schemes ...... xix
Chapters:
1. Introduction ...... 1
Alternative Energy Sources...... 3
Fuel Cells and Hydrogen ...... 6 Water Splitting ...... 10 Natural Photosynthesis ...... II Light-initiated Electron-transfer Reactions ...... 13 Kinetics of Electron-transfer Reactions...... 15 Requirements for Artificial Photosynthesis ...... 19 Artificial Photosynthetic System s ...... 21 Sacrificial model systems ...... 22 Artificial homogeneous PS 11 ...... 22 Liposome-based artificial photosynthetic membrane ...... 24 Polymer molecular assembly...... 25 Lamellar Assemblies ...... 26 Heterogeneous semiconductor systems .. ; ...... 26 Self-assembled monolayer based photoelectrochemical cell ...... 27
viii Zeolite based systems ...... 29 Photochemistry of Tris(2,2'-bipyridyi)ruthenium(II) ...... 30 Ruthenium Dioxide ...... 31 Zeolites...... 33 Scope of this Work ...... 35 References ...... 37
Photostability of Ru(bpy) 3 -' in Solution and upon Zeolite Encapsulation...... 41
Introduction...... 41 Experimental Section ...... 45 Synthesis ...... 45 Photolysis ...... 46 Instrumentation ...... 47 Separation ...... 48 Results and Discussion ...... 49 Chromatographic methodology ...... 49 Aquation chemistry of RuIbpyl^CU ...... 58
Photoaquation chemistry of Ru(bpy) 3 *‘ ...... 64
Photolysis of Ru(bpy) 3 -‘ in water ...... 67
Photolysis of Ru(bpy) 3 "' in buffers ...... 70 Acetate buffer (pH 5 ) ...... 72 Phosphate buffer (pH 5 ) ...... 89 Phthalate buffer (pH 5) ...... 91 Effect of p H ...... 95
Photolysis of Zeolite encapsulated Ru(bpy) 3 -' ...... 98
Photolysis of Ru(bpy) 3 '‘ in the presence of a quencher 104
Photostability of Ru(bpy>3 ^ '...... 119 C onclusion...... 122 References ...... 124
Catalyst Development for Photochemical Hydrogen Production ...... 128
Introduction ...... 128 Photochemical hydrogen production ...... 128 Ruthenium dioxide ...... 131 Structure...... 132 Preparation ...... 132 Surface derivatization techniques ...... 133 Aluminosilicate surface immobilization of metal complexes .. 134 Metal oxide surface derivatization ...... 135 Ruthenium complexes with silyl-substituted bipyridyl ligands . 135 Scope and organization ...... 136
ix Experimental Section ...... 138 Catalyst preparation ...... 138 Zeolite ba^ed RuO; ...... 138 Catalyst poisoning ...... 139 Platiniun catalyst ...... 139 Zeolite surface derivatization...... 140 Synthesis of silyl-substituted biptyridyl ligands ...... 140 Anchoring to zeolite surface ...... 143 Photolysis ...... 145 Chromatography ...... 145 HPLC ...... 145 GC ...... 146 Instrumental methods ...... 146 Results and Discussion ...... 147 Catalysts synthesis and characterization ...... 147 RuO; synthesis from RujCCO),,...... 148 Catalyst characterization ...... 149 XRD ...... 149 Diffuse Reflectance ...... 153 XPS ...... 155 TEM ...... 159 Photolysis experiments ...... 166 Chromatography ...... 166 HPLC ...... 166 GC ...... 168 Hydrogen evolution experiments with RuOn-Y catalysts ...... 171 Comparison of RuO? with other catalytic system ...... 183 MV-‘ decomposition during photolysis ...... 188 Photochemical stability of MV radical ...... 194 Strategies for reducing MV * destruction ...... 197 Covalently modified zeolites ...... 202 Synthesis ...... 203 Photochemistry ...... 215 Conclusion...... 220 References ...... 221
4 Kinetic Analysis and Modeling of Intrazeolitic Photoinduced Charge Separation ...... 227
Introduction...... 227 Experimental Section ...... 230 Results and Discussion ...... 232
Ru(bpy)3 -'-zeolite Y ...... 232 Ru(bpy)3 -*’ quenching with bipyridinium electron acceptors ...... 238 Back electron transfer from bipyridinium radical ion to
Ru(bpy)3 '” in the zeolite ...... 244 Low viologen loadings ...... 244 High loading levels ...... 249 Stella M odel ...... 251 Loading effects ...... 271 Magnitude of the back electron transfer reaction rate constants 272 Relationship between back electron transfer and bipyridinium ion reduction potentials .. 277 Longer lived charge separation in high loaded bipyridinium samples . 278 Quantum Yield...... 282 Cage Escape Yield ...... 283 Conclusion...... 283 References ...... 284
Bibliography ...... 288
XI LIST OF TABLES
Table Pagg
1.1 Components employed in the early sacrificial model systems ...... 23
2.1 Calculated chromatographic parameters for Rulbpylj-' and M V -' ...... 56
2 . 2 Ru(bpy)3 -‘ decomposition after eight hours of photolysis ...... 80
2.3 Retention time and spectral properties of the chromatographic peaks ...... 94
2.4 Emission intensity and lifetime quenching of Rulbpylj-' by MV- ...... 112
2.5 Correlation of the decrease in Ru(bpy);-' photodecomposition and quenching efficiency by MV*' ...... *...... 117
3.1 Comparison of hydrogen evolution efficiency of RuOi-Y catalysts ...... 177
3.2 Structural parameters of various surfaces of single crystal RuO^ ...... 179
3.3 Comparison of hydrogen evolution efficiency of RuOyYiQo. Pt. and RuO? ... 187
3.4 Performance of alkali metal nitrate modified powder RuO, catalysts ...... 199
4.1 2.2'- and 4,4'-bipyridinium ions used as electron acceptors ...... 239
4.2 Bipyridinium loading levels in the Bipyridinium - Rufbpy);-' - Zeolite Y system s...... 242
4.3 Back electron transfer rate constants in the low loading Bipyridinium - Rufbpy);-' - Zeolite Y systems ...... 249
4.4 Rate Constants in the high loading Bipyridinium - Rufbpylj-' - Zeolite Y svstem s...... 271
Xll LIST OF FIGURES
Figure EafiÊ
1.1 Actual and projected world energy consumption ...... 2
1.2 Schematic representation of a fuel cell with a proton conducting electrolyte ------7
1 .3 Prototype miniature fuel cells, (a) Medis DMFC, and (b) Fraunhofer-ISE Hydrogen fuel c e l l ...... 9
1.4 a) The Z-scheme. aerobic photosynthesis mechanism, b) Kok cycle, oxygen
evolution mechanism ...... 1 2
1.5 Schematic representation of (a) adiabatic and non-adiabatic surfaces, (b) surfaces involved in photoinduced electron transfer reactions ...... 16
1. 6 a) Schematic representation of surfaces under different driving force (AGs^). b) Comparison of Marcus and Rehm-Weller models ...... 18
1.7 Sacrificial model system for photoevolution of hydrogen ...... 23
1.8 Liposome-based artificial photosynthetic membrane ...... 25
1.9 Photovoltaic-photoelectrochemical system for water reduction, (a) Schematic representation, and (b) Idealized energy level diagram ...... 28
1.10 Schematic representation of a zeolite based permanent photoinduced charge separation system ...... 30 ■V 1.11 Rutile Structure of RuOi...... 33
1.12 Schematic representation of various zeolite structures with sodalite as the building block ...... 34
2.1 Chromatographic separation method and background chromatograms observed at various relevant wavelengths ...... 52
xiii 2.2 Chromatographic separation of Ru(bpy) 3 '" and MV-' ...... 54
2.3 Ru(bpy)3 -' calibration...... 57
2.4 Photolysis of aqueous Ru(bpy);Cl2 (a) Chromatographic separation of photoproducts, (b) Absorption spectra of the photoproducts ...... 60
2.5 Absorption spectra of (a) the 11.8 min peak in the chromatogram of photolyzed
aqueous Ru(bpy)2 Cl2 . b) the 1 2 . 6 min peak in the chromatogram of methanolic
Ru(bpy)2 Cl2 ...... 62
2.6 Photolysis of Ru(bpy) 3 -* in distilled water (a) Solution absorption spectra, (b)
Chromatographic separation of photoproducts ...... 6 8
2.7 Photolysis of Ru(bpy) 3 -' in distilled water (a) Chromatographic separation of photoproducts, (b) Absorption spectra of the photoproducts ...... 69
2.8 Photolysis of Ru(bpy) 3 -' in pH 5, 2 M acetate buffer (a) Solution absorption spectra, (b) Chromatographic separation of photoproducts ...... 73
2.9 Photolysis of Ru(bpy) 3 -' in pH 5,2 M acetate buffer (a) Chromatographic separation of photoproducts (b) .Absorption spectra of the 12.3 min peak ...... 75
2.10 Absorption spectra of the peaks observed in the chromatograms of photolyzed
Ru(bpy)3 -' in pH 5,2 M acetate buffer (a) 13.3 min peak, and (b) 14.1 min peak76
2.11 Chromatographic separation of photolyzed solution of Ru(bpy) 3 -‘ in pH 5 acetate buffers ...... 78
2.12 Absorption spectra of the peaks observed in the chromatograms of photolyzed
Ru(bpy)3 -‘ in pH 5 acetate buffers (a) 12.3 min peak, (b) 13.3 min peak, and (c) 14.1 min peaks ...... 79
2.13 Chromatographic separation of photolyzed solution of Ru(bpy) 3 -' in pH 5, 2 M acetate buffer in the presence of added N aCl ...... 84
2.14 Comparison of chromatograms monitored at 450 nm and 282 nm in the
separation of photolyzed solution of Ru(bpy) 3 -' in pH 5,2 M acetate buffer in the presence of added N aC l ...... 85
2.15 Absorption spectra of the peaks observed in the chromatograms of the
photolyzed solution of Ru(bpy) 3 ‘* in pH 5. 2 M acetate buffer in the presence
of added NaCl ...... 8 6
xiv 2.16 2,2'-bipyridine a) Chromatogram, (b) Absorption spectrum ...... 8 8
2.17 Photolysis of Ru(bpy)f' in pH 5, 0.025 M phosphate buffer a) Chromatographic separation of photoproducts b) Absorption spectra of photoproducts ...... 90
2.18 Photolysis of Ru(bpy)]-' in pH 5, 0.025 M phthalate buffer a) Solution absorption spectra (b) Chromatographic separation of photoproducts ...... 92
2.19 Photolysis of Rufbpy))-' in pH 5,0.025 M phthalate buffer a) Chromatographic separation of photoproducts (b) Absorption spectra of the photoproducts ...... 93
2.20 Photolysis of Ru(bpy)]-' in pH 4,2 M acetate buffer a) Chromatographic separation of photoproducts (b) Absorption spectra of the photoproducts ...... 96
2.21 Photolysis of Ru(bpy)/' in pH 3. 0.025 M phthalate buffer a) Solution absorption spectra (b) Chromatographic separation of photoproducts ...... 97
2.22 Diffuse reflectance spectra of Rufbpy);-' encapsulated in nanocrystalline zeolite Y ...... 99
-I -> Chromatographic separation of the citric acid extract of photolyzed Ru(bpy);,-* encapsulated in nanocrystalline zeolite Y ...... 101
2.24 Absorption spectra of the peaks observed in the chromatograms of the citric acid extract of photolyzed Rufbpy);-' encapsulated in nanocrystalline 102
2.25 Emission quenching of Rufbpy)/' by MV-' (a) Intensity quenching b) Lifetime quenching ...... 109
2.26 Stem-Volmer plot of the emission intensity and lifetime quenching of Ru(bpy);-' by M V -'...... I ll
2.27 Chromatographic separation of photolyzed solution of Rufbpy)/' in pH 5. 2 M acetate buffer in the presence of MV- ...... 114
2.28 Chromatographic separation (monitored at 256 nm) of photolyzed solution of Rufbpy)]-' in pH 5,2 M acetate buffer in the presence of M V -' ...... 115
2.29 Ru(bpy) 3 -' peak area decrease during the photolysis of Rufbpy);-' in pH 5,2 M acetate buffer in the presence of MV- ...... 116
2.30 Correlation of the photodecomposition quenching to the fluorescence quenching of Ru(bpy)]-' by MV- ...... 118
XV 2.31 Photolysis of Ru(bpy) 3 *^ in pH 4, acetate buffer in the presence of
[Co(NH3 );Cl]-" a) Chromatographic separation of photoproducts (b) Absorption spectra of the photoproducts ...... 121
3.1 X-ray diffraction patterns of zeolite Y and that of RuOi-Y catalysts ...... 150
3.2 X-ray diffraction patterns of zeolite Y and that of RuO,-Y catalysts with peak deconvolution of the most intense peak of RuOi at 35.1 ° for Ru02-Y^oo 151
3.3 Diffuse reflectance spectra of RuO^-YiQo and RuO^-Yj^...... 154
3.4 X-ray photoelectron spectrum survey scans of RuO.-Y catalysts...... 157
3.5 X-ray photoelectron spectra of RuO^-Y catalysts showing the Ru hd region .. 158
3.6 Transmission electron micrograph o fR u Oi-Y r x ...... 160
3.7 Transmission electron micrograph of RuOi-Y^oo ...... 161
3.8 Transmission electron micrograph of RuO^-Yjoo ...... 162
3.9 Transmission electron micrograph of a particle of RuOyY , 0 0 ...... 163
3.10 Transmission electron micrograph of RuOi-Yjoo ...... 164
3.11 HPLC calibration for methylviologen ...... 169
3.12 GC calibration for hydrogen ...... 170
3.13 Photogeneration of hydrogen with RuOi-Yioq catalyst (a) Gas chromatograms of headspace during the photolysis (b) Comparison of hydrogen evolution under standard conditions with that in the absence of catalyst ...... 173
3.14 Comparison of the hydrogen evolution performance of RuO^-Y^oo catalyst (a) at different EOTA concentrations and b) at different pH ...... 175
3.15 Comparison of hydrogen evolution performance of the RuOi-Y catalysts .... 176
3.16 Long term stability of the RuOi-Y^oo catalyst ...... 184
3.17 Comparison of hydrogen evolution performance of RuO^-Yioo, Pt and RuOi.. 185
XVI 3.18 Comparison of hydrogen evolution performance of RuOi-Yigo, Pt and RuOi (a) peak area (b) total yield ...... 186
3.19 Chromatographic estimation of MV-" decomposition after hydrogen evolution photolysis with RuOj-Y iqo catalyst ...... 190
3.20 Chromatographic estimation of MV- decomposition after hydrogen evolution photolysis with RuOi catalyst ...... 191
3.21 Comparison of MV’* decomposition for RuO^-Y^ oü. Pt. and R u O ...... 193
3.22 Photostability of methylviologen radical a) Absorption spectra of Ru(bpy);-' - MV-' - EOTA solution during continuous photolysis b) Chromatographic estimation of MV-' decomposition after photolysis ...... 195
3.23 Effect of RuO. catalyst mass on a) H, evolution rate and MV’ decomposition during photolysis, b) Hydrogen yield and MV-' effective life cycle ...... 201
3.24 Schematic representation of the photoinitiated electron transfer processes with RuO.-Y surface anchored Rufbpy);-' ...... 204
3.25 NMR spectrum of 4-methyl.4'-triethoxysilylmethyl-2,2'-bipyridine ...... 207
3.26 Diffuse reflectance spectra of zeolite surface derivatized with a) bipyridine and (b) Ru(bpy)j’' ...... 210
3.27 a) Diffuse reflectance spectra of surface derivatized RuO.-Y. qo (1%). b) Fluorescence emission spectra of surface derivatized zeolite Y and RuO.-Y:oo(l%) ...... 214
3.28 a) Solution absorption spectra during the photolysis with surface derivatized zeolite Y. b) Hydrogen evolution during the photolysis with surface derivatized RuO.-Y^oo ...... 217
3.29 Schematic representation of the proposed integrated photosystem ...... 219
4.1 Molecular model of Rufbpy)]’* encapsulated in a supercage of zeolite Y .... 233
4.2 Molecular model of Rufbpy),’* encapsulated in a hypercage of EMT ...... 235
4.3 a) Normalized TRDR difference spectra of intrazeolitic Rufbpy)/' 100 ns after the laser excitation, b) The transient TRDR decay data ...... 237
xvii 4.4 Molecular models of (a) Ru(bpy)3 -" - zeolite Y with (b) low loading of MV-* and (c) high loading of MV-" ...... 241
4.5 Comparison of the normalized TRDR difference spectra obtained 100 ns after
the laser excitation of Ru(bpy) 3 '"-zeolite V with that of low loading MV- -
Ru(bpy)3 -'-zeolite V ...... 245
4.6 The transient TRDR decay data for low loading bipyridinium - Ru(bpy);-' - zeolite V for (a) MV-*, and (b) 2DQ- ...... 247
4.7 The transient TRDR decay data for low loading bipyridinium - Ru(bpy) 3 -' - zeolite Y for (a) 3DQ- . and (b) 4DQ- ...... 248
4.8 Comparison of the TRDR difference spectra 100 ns after the laser excitation
of Ru(bpy)3 -*-zeolite Y with that of high loading MV-'-Ru(bpy) 3 -*-zeolite Y . 250
4.9 a) Normalized TRDR spectra of the high loading MV-'-Ru(bpy) 3 - -zeolite at various time delays, b) The transient TRDR decay d a ta ...... 252
4.10 a) Schematic representation of a compartmental model of drug absorption and excretion, b) The equivalent STELLA m odel ...... 254
4.11 Simulation of the STELLA model for drug absorption and excretion ...... 257
4.12 Preliminary STELLA model for the photoinduced charge transfer processes
involved in the MV-'- Ru(bpy) 3 -*-zeolite system ...... 260
4.13 Integrated STELLA model for the charge transfer processes involved in the
MV-*- Ru(bpy) 3 -‘-zeolite system ...... 262
4.14 STELLA model for the charge transfer processes originating from the
methylviologen radical involved in the MV-'- Ru(bpy) 3 -*-zeolite system .... 264
4.15 Schematic representation of the arrangement of supercages in zeolite Y ...... 266
4.16 The transient TRDR decay data for high loading bipyridinium-Rufbpy);-'- zeolite and STELLA model simulated decay for (a) MV-*, (b) 2D Q -* ...... 269
4.17 The transient TRDR decay data for high loading bipyridinium-Rufbpy),-'- zeolite and STELLA model simulated decay for (a) 3DQ-*, (b) 4DQ-* ...... 270
4.18 Plot of the rate constants for the back electron transfer under high bipyridinium loading (kb'*‘)versus the driving force (A E°) ...... 279
xviii LIST OF SCHEMES
Scheme Page
2.1 Schematic representation of the energy levels and the excited state decay pathways for Ru(bpy)/' ...... 42
Schematic representation of the mechanism proposed in the literature for
ligand photosubstitution reactions of Rulbpy),-' ...... 6 6
2.3 Proposed mechanism for the photolysis of Ru(bpy)/' in distilled water and in acetate buffers ...... 71
XIX CHAPTER I
INTRODUCTION
"Energy is the very fuel of society, and societies without competitive energy suffer", according to Mr. Lee Raymond, the chairman of ExxonMobil, the largest energy company in the world [1]. This is true for living organisms in any ecosystem, with nature striking a delicate energy balance between consumption and conversion. Human need for energy is growing at an ever increasing rate. During the 20'*’ century alone, the human population quadrupled and energy use grew 16-fold [2]. The world energy consumption
is projected to increase by 60% in the next 2 0 years as the developing countries' need for energy is expected to increase at an alarming pace (Figure 1.1) [3]. The insatiable appetite for energy has fueled the growth of the energy industry to become the largest business in the world. A conservative estimate puts global energy turnover for the year
2000 at US $1.7-2 trillion and in the same period the energy brokerage e-commerce website of Enron handled some 550 million transactions valued at US $ 345 billion, making it the most successful internet effort to date [1]. These figures highlight the importance of the energy industry that is about to imdergo a massive change. 700 Total Projections «aasjtei EE + FSU 600 - Industrialized Developing
500 -
400 -
OQ 'C 300 -
1985 1990 1995 1999 2005 2010 2015 2020 Year
Figure 1.1 Actual and projected world energy consumption (adapted form ref. 3). BTU - British Thermal Units; EE - Eastern Europe; FSU - Former Soviet Union; Throughout the history of modem society, industrial revolutions like the steam engine and internal combustion engine have been closely related to the human ability to decarbonize the fuel. In the beginning it was biomass with high carbon, then came coal with moderate carbon and during the past century it was oil with its lower carbon content.
Extrapolating this global transition in energy systems would lead ultimately to clean-
burning gases with no or very low carbon content. Considerable effort has already been
invested in exploiting natural gas with very low carbon content but for futuristic gaseous
fuel with no carbon content we have to look beyond traditional fossil fuel sources.
ALTERNATIVE ENERGY SOURCES
“The stone age did not end because the world ran out of stones, and the oil age
will not end because we run out of oil", according to Mr. Don Huberts, head of Shell
Hydrogen, a division of Royal Dutch/Shell [4]. Though experts disagree when and why
the oil age will end. it is a fact that one day it will certainly end. Fossil fuels by nature
are limited in quantities and by one estimate we would have used up half of all available
oil sometime between 2015 and 2030 [I]. Volatile prices, growing environmental
concerns, governmental deregulations and innovations in science and technology are all
fueling a rapid transition. In a demand-supply dictated free market economy, scarcity,
whether artificial or real, leads to an increase in price. In the past two years crude oil
prices have been very volatile, fluctuating from US$ 10 to US$ 35 per barrel. Burning
fossil fuels release gases like CO], leading to global warming due to greenhouse effect.
SO]. NO,^ leading to acid rain and particulates leading to urban smog. Environmental
activism has lead to measures such as the Kyoto Protocol, a 2000 United Nations treaty
3 among industrial nations to curb global warming and California Air Resources Board's
zero-emission requirement for 1 0 % of the cars sold in that state by big manufacturers by
2004. These reasons, coupled with modest governmental subsidies and impressive
technological advances have made alternative fuel sources economically viable.
Ancient windmills in The Netherlands, water powered rafts and mills around the
world and solar powered cooking in the tropical coimtries are all testaments to the
ingenuity with which man has been able to harness various abundant and clean energy
sources. Modem day technology evolved out of these ancient practices and has made
alternative energy (non-fossil fueled) sources practical and attractive. Wind powered
energy producing units are getting ever bigger and provide as much as 25% of the power
supply to large parts of Denmark, Spain and Germany [ 1 ]. Huge hydrothermal dams are
producing several gigawatts of power around the world. Though tidal wave energy is
estimated to be one of the largest untapped reserve, it is also technologically difficult at
this point. Solar powered turbines have been around for a while and semiconductor based
solar panels provide power to remote areas as well as to various small applications. In
places like Iceland, geothermal energy has been tapped as a clean alternative. Though not
renewable in the strict sense, in the past few decades, nuclear energy has provided and
continues to provide a substantial portion of the energy needs in several countries.
Some of these alternative energy sources like tidal and geothermal are restricted to
certain geographic areas of the world. On the other hand, though cleaner than fossil fuels,
hydrothermal and nuclear power plants have their own environmental problems. Issues
with dams include CO; emission firom substantial areas of submerged rotting vegetation,
4 irreversible ecological changes like almost extinct salmon population in some rivers, increased seismological activities around the dams, high initial cost and limited lifetime due to silting. Nuclear power plants pose a dangerous operational safety problem as is evident from the Three Mile Island and Chernobyl accidents; spent fuel with its thousands of years of radiative half life poses an unacceptable storage and disposal risk.
Solar energy on the other hand suffers from none of these issues, is available in most parts of the world (though in differing quantities) and should be so for the
foreseeable future. Solar energy is also clean and poses no safety risk now or in the
future. According to one estimate every day the earth's surface receives solar energy equivalent to thirty years of total human consumption [5]. The cost and efficiency remain as issues with common forms of tapping solar energy, either into electricity or heat.
Semiconductor photovoltaic (PV) cells that can form electron-hole pair need very expensive pure crystalline silicon and the solar heaters suffer from poor efficiency.
Ruthenium polypyridyl dye sensitized nanocrystalline semiconductor TiOi solar cells
present a cheaper alternative and have been shown to have an overall energy conversion
efficiency as high as 10.4% [ 6 ]. Even if the shortcomings are overcome, the storage and
transmission of heat and low-power electricity thus generated would be cumbersome. An
attractive alternative is to convert solar energy to storable chemical energy. Among other
ways to attain this conversion, the splitting of water into oxygen and hydrogen has
attracted considerable attention. The explosive limit of H, in air (4.00%) is higher than
that of butane (1.8 6 %). With the development of storage technologies employing metal
hydrides or carbon nanotubes, transportation and storage of would not be difficult.
5 This route to solar energy conversion is very attractive, as with the introduction of fuel cell based power generators and prototype cars, hydrogen is poised to replace oil as the
fuel of the future.
FUEL CELLS AND HYDROGEN
Hydrogen is the energy currency' of the future [2]. When used in a fuel cell,
hydrogen would be a true zero emission’ power source as water is the only byproduct.
Typical internal combustion engines convert less than 20% of the energy content of their
fuel into work, whereas for fuel cells it is as high as 30%. making them 50% more
efficient. At the high power end. fuel cells even outperform the combined-cycle gas
turbine, the present gold standard in power generation. The fuel cell technology is older
than that of the gasoline internal combustion engine itself. It was serendipitously
discovered during water electrolysis experiments in 1839 by Sir William Grove, and the
first modem fuel cell was designed in 1889 [7].
Fuel cells are electrochemical cells with an ion conducting electrolyte and work
by the principle of reverse electrolysis using a fuel and an oxidant. Fuel cells differ in
fuel (hydrogen, methane, methanol), catalyst (gold, platinum), electrode (ceramic, metal),
electrolyte (alkali like potassium hydroxide, molten carbonate, solid mixed oxide like
yttria and zirconia, phosphoric acid, proton exchange membrane often referred to as
PEM) and operational temperatures (~50°C to ~1000°C) [ 8 ]. In the common form,
hydrogen is oxidized at the anode, electrons flow through the external circuit providing
energy and protons move through electrolyte towards the cathode where they reduce
oxygen to water (Figure 1.2).
6 For over a century fuel cells remained an exquisite but expensive technology to commercialize, as the electrodes are usually coated with noble metal catalysts like platinum that are veiy expensive. Among others, the platinum dispersing technology of
Johnson Matthey coupled with the development PEM technology by Ballard Power
Systems is making a practical fuel cell a reality. Automakers consider fuel cell as their savior to comply with the Californian emission requirement. Fuel cell powered commercial buses by Daimler-Benz and prototypes like NECAR from Daimler-Benz and
Hypercar from Rocky Mountain Institute attest to this new found status.
Oxidant
ii-K-A
Fuel
Anode Cathode Water
Figure 1.2 Schematic representation of a fuel cell with a proton conducting electrolyte. Low or no emission, higher efficiency and modular scalability of fuel cells have attracted the attention of the power industry for distributed generation that would cut the
massive transmission losses. Power generation companies like General Public Utilities
(USA). Toshiba (Japan) and equipment makers like General Electric, ABB, Siemens are
showing increasing interest in fuel cells. Mass produced fuel cell power generators like
PC25 by a collaboration between Toshiba and International Fuel Cells are expected to
generate 200 MW in Japan alone by the year 2010.
On the lower end of the power spectrum, micro fuel cells are being touted as the
energy source of the future for portable electronic devices. Various successful prototypes
of direct methanol fuel cells (DMFC) reveal a glimpse of the future of micro fuel cell
technology [9]. MicroFuel Cell™ invented by Mr. Robert Hockaday at Manhattan
Scientifics. Inc. (New York) is one of the first to demonstrate the viability of this novel
application [10]. The Hockaday cell employing a PEM has been shown to produce 300
Watt Hour/Kg of fuel, three times as much as a Lithium-ion cell phone battery.
Motorola's energy technology lab in collaboration with Los Alamos National Laboratory
and Department of Energy recently demonstrated a ceramic fuel cell membrane electrode
assembly, measuring about 2" x 2" x capable of producing 100 mW continuously.
Medis Fuel cell (Medis Technologies. Yehud. Israel) employs a liquid electrolyte (H 3PO4
- NaH^PO^ - Na^HPO^) instead of PEM and platinum containing polyaniline particles as
electrocatalyst (Figure 1.3a) [11]. Medis Technologies has been tapped by Sagem SA. a
French telecommimication group, to build a pilot plant capable of producing 50 million
fuel cells annually [9].
8 Though these prototypes employ diluted methanol or ethanol, efforts are well underway to develop hydrogen based miniature fuel cells also. At the 2001 Hannover
Fair. Fraunhofer Institute for Solar Energy Systems (Freiburg, Germany) demonstrated a camcorder powered by the first ever fully operational and integrated fuel cell system
about the size of a matchbox producing 10 W at 8 V [12]. The hydrogen based fuel cell system had a replaceable metal hydride reservoir that served as a hydrogen tank. The prototype cell shown in Figure 1.3b developed by Fraunhofer ISE has a power density of
1 W/cm\ The multi-billion dollar collaborative effort between the car makers, power firms, portable electronic device manufacturers, even oil companies and fuel cell developers marks the maturation of the experimental technology into a commercial one.
a) b)
Figure 1.3 Prototype miniature fuel cells, (a) Medis DMFC (from ref. 13), (b) Fraunhofer-ISE Hydrogen fuel cell (from ref. 9).
- "X, - * WATER SPLITTING
Solar powered water splitting is a “Holy Grail", a much coveted but difficult to attain renewable energy source to produce hydrogen [14]. Hydrogen by itself is only a means of storing energy and not a primary source of energy. There are several routes envisioned to produce Hi. including reforming methanol or hydrocarbons, and none would be as clean as water splitting. Earlier ideas to split water into hydrogen and oxygen invariably relied on energy intensive electrolysis. This would involve generating electricity by other means and at every step on the way wasting energy, thereby ending up as an inefficient process. Solar powered water splitting, if made practical, would combine two of the most abundant resources in this world, namely water and sunlight.
There are several potential reactions that involve water to produce energy rich fuel like ammonia, methane, methanol and hydrogen. Of these hydrogen production as shown
below involves the fewest electrons, highest energy stored per gram of fuel and for our
purposes is of greatest importance.
H .O — > H. + '/SO. AG' = 237.2 kJ mol ' E’ = 1.23 V (1.1)
From the free energy of this reaction it is obvious that the energy content of Hi is
118.6 kJ g ', which is three times as high as that of oil [15]. Water does not absorb
visible light and there are various other scientific hurdles to overcome. Several research
groups around the world have been working since the early 1970s, on this area of
research that is commonly referred to as artificial photosynthesis’ [14]. In our attempts
to emulate photosynthesis, it would be prudent to look at how nature does it.
10 NATURAL PHOTOSYNTHESIS
“O Surya (sun)...The radiant God...Thou by whose lustre all the world of life comes forth...” - Rig Veda Book X; Hymn XXXVII; Verse 7,8,9; [16]. Sun via photosynthesis is the source o f energy, either directly or indirectly, for all life on earth.
There are two kinds of organisms capable of photosynthesis. Anaerobic bacteria can photochemically reduce CO? to carbohydrate using reduced carbon and sulfur compounds as electron source. These bacteria are not capable of using water as an electron source and hence are not of great interest for our present purpose. Photosynthesis in aerobic cyanobacteria, algae and plants use water as an electron source, employ various pigments
for light absorption, evolve oxygen and is a trans-membrane process [17]. Chlorophyll
multimers (P^o) in photosystem 11 (PSIl) can be excited either by direct light absorption or sensitization via the light collecting complex. The photoexcited P^go reduces the electron acceptor, pheophytin (Ph a) which in turn reduces quinones (PQ \and PQg) that act as electron relays. Water oxidation in the oxygen evolving complex (OEC) is
catalyzed by an oxo-bridged tetra manganese unit. Electrons obtained in this process are
utilized to reduce the reaction center" of the PSII via a tyrosine residue (Z) that acts as an
electron relay. The transmembrane electron propagation aided by the second light
absorbing chlorophyll imit (P 700) in the photosystem I (PSI) ultimately results in the
reduction of nicotinamide adenine dinucleotide phosphate, NADP* to NADPH on the
other side of the membrane. This process can be schematically described by the now
familiar "Z-scheme" (Figure 1.4a).
11 PSI a) -1.5—1 P700'
- 10- PSII
> - 0 .5 -
N A D P e Fd-NADP'
0 —
Oxygen- Cytochrome b f • 'lasiocyanin evolving Complex a: +0.5 — P700 complex Light Proton Gradient
+ 10 — P680
Light
+ 15 — Proton Gradient
b) O,
(HCO3)
2 H, 0 -----> O, + 4 H+ + 4 e-
Figure 1.4 a) The Z-scheme. aerobic photosynthesis mechanism (adapted from ref.
2 0 ). b) Kok cycle, oxygen evolution mechanism (adapted from ref. 2 1 ).
12 After 3 billion years of evolution all aerobic photosynthetic species are believed to possess qualitatively the same OEC. Recent reports of x-ray diffraction at 3.8 to 4.2 Â
resolution [18] and electron diffraction at 8 A resolution [19] of the OEC complex shed some light into the structure previously inferred by spectroscopic and sequence studies, but the exact structure is yet to be solved. The inorganic cofactor of OEC contains an
oxo-bridged tetra manganese unit and can be represented as Mn^Ca,Cl,. 2 0 ., [21]. Though the molecular mechanism of oxygen evolution is still not proven, the Kok cycle summarizes the current understanding (Figure 1.4b). Four sequential one electron
oxidation reactions result in the formation of five intermediates, Sq to S 4 , the subscripts
indicating their total oxidation state with S 4 being the most oxidized state. The Sq and S,
are dark stable. Si and S 3 are metastable and S 4 is an unstable state returning to S„ accompanied by the oxidation of water to oxygen. Calcium is proposed to offer structural stability and act as a gatekeeper in restricting access of water whereas chloride is proposed to block premature water oxidation [21]. Various attempts to mimic the OEC with synthetic manganese complexes are also providing insights into the structure and mechanism of this complicated system.
LIGHT-INITIATED ELECTRON-TRANSFER REACTIONS
Natural photosynthesis is ideally compartmentalized to keep the charge separated species spatially isolated. The three important components of the reaction center are chlorophyll, the light absorbing sensitizer, pheophytin. the electron acceptor and tyrosine, the electron relay. As water itself does not absorb solar radiation except in the infrared region, we need a sensitizer molecule S, that would absorb light photons of appropriate
13 energy. Sensitizer molecules that are in an electronically excited state S* are often better electron donors and/or acceptors than in the ground state. Ideally in the presence of a suitable catalysts. S* should reduce water and the oxidized S' should oxidize water. But practical and kinetic conditions prevent such a system from being operational, requiring additional components.
The excited sensitizer S* could radiatively or non-radiatively relax to the ground state. When the sensitizers are better donors S* could also undergo oxidative quenching in the presence of an electron acceptor A. This light driven electron transfer reaction that would produce the oxidized sensitizer S' and a reduced acceptor A . marks the beginning of the conversion of light energy into chemical energy. Often the charge separated species. S' and A . undergo energy wasting recombination. This could be prevented by clever manipulation of the system or as nature does it by employing an electron donor D.
In the presence of appropriate catalysts, the reduced acceptor A reduces water, and either the oxidized sensitizer S' or the oxidized donor D' oxidizes water to oxygen. These reactions are summarized as shown below [14.15]:
hv S s* (1.2)
kd S* — ------> s S (1.3)
kf S * + A S' +A- (1.4)
S' + A — S + A (1.5)
S' + D S + D' (1.6)
14 2A + 2H.0 > 2A + H, + 20H- (1.7)
4S + H ,0 > 4S + 0 i + 4H (1.8)
4D +H.0 > 4D + Oi + 4H' (1.9)
KINETICS OF ELECTRON-TRANSFER REACTIONS
The bimolecular electron-transfer reaction as shown in equation 1.4 represents a simplified picture of the actual complicated process. In the first step of this process, an
encounter complex’ (EC) between the excited sensitizer S* and the electron acceptor A. is formed by diffusional collision, surrounded by a solvent cage. Electron transfer reaction with a rate constant ofk* is one of the possibilities for S* and A in the EC, and a fraction of these emerge out of the solvent cage as charge separated species as shown in
equation 1 . 1 0 :
kjif S* + A ^ [S* • • • A] ^ [ S '• • • A ] S + A (1.10) EC
The kinetics of the electron transfer step could be treated depending on whether the potential energy surfaces of S* and that of A interact (adiabatic) or intersect (non- adiabatic) either by the classical model or by the quantum mechanical model, respectively
(Figure 1.5a) [22]. The electron transfer rate constant according to the classical model,
similar to the absolute reaction-rate theory, is given by equation 1 . 1 1 , where k", is the
frequency factor and is the reorganization free energy (neglecting entropy
contribution) required for the electron transfer between the sensitizer and the acceptor to occur obeying Franck-Condon principle (Figure 1.5b) [23].
15 a) Adiabatic Non-adiabatic
V B ÙÜ
Nuclear Motion Nuclear Motion
b) Transition S* + A ^ State
'SA
S + A
Nuclear Motion
Figure 1.5 Schematic representation of (a) interacting (adiabatic) and intersecting (non- adiabatic) surfaces, (b) surfaces involved in photoinduced electron transfer reactions shown in equation 1.2 and 1.4.
16 kç, = k%exp(-AGsA*/RT) (111)
k% - kvn 1 ( 1 2 )
AGsA^ = AG„b* + AG„r (1.13)
The frequency factor can be expressed as in equation 1.12. where k is the
transmission coefficient ( 1 . 0 for adiabatic systems), wherein the potential energy surface of S* and A interact appreciably to form two new surfaces and is the effective frequency for nuclear motion. The reorganization free energy (AG^,J) Is determined by the structural distortions and solvation changes necessary to obey the Franck-Condon principle during the electron transfer process. AGsa* can be separated into vibrational and solvent reorganizational components as shown in equation 1.13. According to the now
famous Marcus theory. AG^^' can be calculated using equation 1.14 from AG^(0)sa> the
intrinsic reorganizational parameter (when AG = 0) and AGja- the free energy change for the electron transfer process.
A G \, = AG^(0) sa {l+[AGsA/4AΠ(0W }: (1.14)
where 4AG*(0)sa = k. the total reorganization energy. This expression on substitution for
AG^sa in equation 1 . 1 1 . describes the relationship between the rate constant, k^,. and the
driving force AG^^ According to the Marcus theory, the rate constant increases with the
driving force up to the point when AGsa = -X. (normal region), beyond which the rate
decreases with the driving force (inverted region) as shown in Figure 1. 6 . Not all
reactions involving electron transfer behave as predicted by the Marcus theory when
17 a)
V s Normal Region Inverted Region U Ë Ùm
Nuclear Motion
b)
en
Figure 1.6 a) Schematic representation of surfaces involved in electron transfer reactions under different driving force conditions (AG sa ). (b) Comparison of Marcus (solid curve, a) and Rehm-Weller (dashed curve, b) models.
18 AG sa < and these situations can be explained either by the modified Marcus model
(Marcus-Agmon-Levine) or by the Rehm-Weller model (equation 1.15).
AG% a = (AG sa/2) + {(AG sa/2)' + [AG*(0)sA]-}' (115)
According to the quantum mechanical approach when the potential energy surfaces of S* and A intersect and not interact, electron transfer occurs by tunneling that can be explained by the perturbation theory and is beyond the scope of the present discussion. From the Marcus theory, it is clear that for an efficient artificial photosynthetic system it would be advantageous to have the forward electron transfer between S* and A (equation 1.4) in the normal region' and the back electron transfer
(equation 1.5) in the inverted region, thereby maximizing the charge separation (Figure
1.6 ). Also the choice of solvent would influence the solvent reorganizational energy and hence the electron transfer rate constant.
REQUIREMENTS FOR ARTIFICIAL PHOTOSYNTHESIS
Any practical artificial photosynthesis system must satisfy some theoretical and practical requirements. As the reaction potential of water splitting is 1.23 V. theoretically photons with wavelength less than 1008 nm should be able to effect it [15]. Light of higher wavelength is not useful and that of lower wavelength is partially lost to various
wasteful processes like internal conversion, limiting the efficiency of the system. The energy efficiency of a solar energy system is also inherently limited by thermodynamic considerations. The maximum theoretically achievable efficiency, rip, for any solar energy system as shown below, is a function of the threshold excitation wavelength
[24].
19 _ _ '^g^^ex^com % " ^ (1.16)
\n in
where is the photon flux with X < X^, Ap^x is the chemical potential of the excited state relative to the ground state, „nv is the quantum yield of the conversion process and E, is the total incident irradiance density (W m ’). In the expression for (equation 1.16). is the minimum wavelength with significant output (for terrestrial solar spectrum it is usually 300 nm), E,(>.) is irradiance in W m - nm ', h is the Planck's constant and c is the velocity of light. Theoretical analysis of the Equation 1.16 suggests that for a likely system with a single kind of sensitizer that absorbs two photons for every molecule of
hydrogen, the ideal and practical efficiency would be about 31% and 1 0 %, respectively
[24]. As in natural photosynthesis, it is also possible to use multiple sensitizers to generate a combined potential greater than 1.23 eV that would allow the use of higher wavelengths and better efficiency. For a photosystem with two kinds of sensitizers that
absorb a total of four photons per molecule of hydrogen the ideal and practical efficiency
would be about 41% and 16%, respectively [24]. However, in natural photosynthesis eight to ten photons are absorbed per molecule of CO] fixed (or O, evolved) and the overall efficiency is about 4-6%.
As the solar radiation is rather diffuse, for an economically viable practical system
the minimum necessary efficiency is about 1 0 % for the cost of hydrogen thus produced to
20 be competitive with other forms of fiiel [24]. The sensitizer must have appropriate optical and electrochemical properties. The relaxation process of S* must be slower than the oxidative or reductive quenching. The sensitizer and the relay should be photochemically stable. The catalyst must be efficient in coupling the single electron light reactions to multi electron dark reactions and should not catalyze any unwanted side reactions. Any practical system should be reasonably long lived with all the active components lasting for at least several years and millions of photons.
ARTIFICIAL PHOTOSYNTHETIC SYSTEMS
From our present understanding, emulating nature is possible as far as light collection, oxygen evolution, electron propagation and the general architecture. But for
Hi production we have to rely on our ingenuity as plants instead produce NADPH. For an artificial photosynthetic system we need a light collecting and energy propagating unit, a redox unit, electron propagation unit and catalysts with the components arranged in appropriate spatial orientation. There are various diverse strategies reported in the literature in order to realize a functional artificial photosynthetic system. They can be distinguished into sacrificial and cyclic, biological and synthetic, homogeneous and heterogeneous, self-assembling multicomponent microheterogeneous and covalent supramolecular, photocatalytic and photoelectrocatalytic systems. The sacrificial model systems almost always involve actual gas evolution studies, whereas most other systems with a few notable exceptions study only the charge separation reaction, concentrating on the forward and back electron transfer rates, charge separation efficiency and the lifetime of the charge separated species. Several model systems are discussed below.
21 Sacrificial model systems
Most of the early attempts fail into this category, with the different components simply dissolved or dispersed in a solution [15]. These model systems function by employing sacrificial components that decompose irreversibly, thus altering the thermodynamics of the reaction from inherently endothermie to exothermic conversion
[14]. Though not of any great practical value, they do serve as a valuable tool to evaluate the feasibility and to compare the performance of new components. The sacrificial model systems almost always involve only one half of the water splitting reaction, either
hydrogen or oxygen evolution. For example in equation 1. 6 , if D is an unstable species that undergoes rapid decomposition then the system would produce only hydrogen
(Figure 1.7). Similarly when A is unstable the system would produce only oxygen. But simple combinations of proven and O, evolving systems often do not work in conjunction to split water. Unless inexpensive, industrial waste chemicals that are
produced in large quantities could be used as the sacrificial agents, these systems pose no commercial potential. The sensitizers, acceptors, donors, and catalysts employed in the early systems for hydrogen production are shown in Table 1.1.
Artificial homogeneous PS II
Donor-sensitizer-acceptor (D-S-A) units made of a manganese(II) complex as
electron donor covalently linked to the sensitizer, Rufbpyjj-', in the presence of
methylviologen as an electron acceptor in acetonitrile has been shown to produce
photoinitiated electron transfer [25]. The photoexcited Ru-‘ transfers an electron to
methylviologen in solution and in turn is reduced by the Mn(II) complex. It has also been Ox Pdt
jCat
Figure 1.7 Sacrificial model system for hydrogen evolution. S - sensitizer; A - acceptor: D - donor; Cat - catalyst.
Sensitizer Acceptor Donor Catalyst Ref
Acridine Yellow Eu^* / Cysteine PtO: [26]
Rufbpylj-* Rh(bpy) 3-' Triethanolamine K: PtCl, [27]
Rufbpy),-* MV-- EDTA Colloidal Pt [28]
Ru(bpy);-' MV-* Triethanolamine PtO: [29]
Table 1.1 Components employed in the early sacrificial model systems.
23 shown that the same sensitizer with tyrosine as the donor and Co(NH 3)jCl-* or sodium persulfate as acceptor in an aqueous solution on irradiation produced tyrosine radical cation in a manner similar to above. When dipicolylamine is covalently linked to the tyrosine moiety, the intramolecular electron transfer rate constant in this process increased by 2 orders of magnitude. The system is supposed to mimic the Mn cluster.
Chlorophyll. Tyr^ and His 190 moieties of PSII. Though oxygen evolution is yet to be realized, this system is a credible model for the early reactions of water oxidation.
Liposome-based artificial photosynthetic membrane
D-S-A units made of carotenoid-porphyrin-quinone (C-P-Q) triads inserted into the lipid bilayer of liposome vesicles span the width of the bilayer with the hydrophobic carotenoid on the inside and the hydrophilic carboxylic containing quinone on the outside of the vesicle as shown in Figure 1.8 [30]. Photoinduced charge separation leads to C" -
P-Q*', followed by reduction of a free Q, by Q" near the outer surface to a semiquinone anion. The basic semiquinone anion upon protonation from the external aqueous medium becomes neutral and diffuses through the lipid bilayer which upon encountering a C" gets oxidized, releasing the proton inside the liposome. This results in a net proton translocation across the membrane without any counterion, thereby creating a transmembrane potential. By incorporating membrane spanning spinach ATP synthase enzymes onto the liposomes, the transmembrane proton gradient was utilized to synthesize ATP from ADP on the external aqueous medium. The quantum yield of proton translocation and ATP synthesis were found to be 0.5-0.6% and 0.15%. respectively.
24 ATP
Figure 1.8 Liposome-based artificial photosynthetic membrane (from ref. 30a).
Polymer molecular assembly
D-S-A units were made of amine frmctionalized polystyrene [co-PS-CH^NH;] matrix derivatized with ruthenium polypyridyl complexes. [Ru(R'-bpy-R-)(RLbpy-
R'*)(Me-bpy-COOH)]-\ where R'-R'* = H (antenna sites, AS. 2.13 eV) or R'. R^ = CHj and R- = electron donor (phenothiazine, PTZ) and R"* = acceptor (methylviologen. MV-*) covalently linked via CH? linkers (reaction center, RC, 2.04 eV) [31]. The RC alone when irradiated at 460 nm formed an intramolecular redox-separated (RS, lifetime
25 = 160 ns), [Ru(Me-bpy-PTZ"')(Me-bpy-MV"')(Me-bpy-COOH)]‘‘* with an efficiency of tIrs = 0.35 ± 0.05%. In a styrene assembly that contained 17:3 AS:RC, irradiation led to the formation of RS state which was estimated approximately 50% due to direct excitation and ET at the RC and the other 50% due to sensitization, excitation of the antenna element (favored by 0.09 eV), followed by energy transfer and ET. The net efficiency of the AS sensitized RS fonnation was found to be irradiance dependant (11.5-
17.5%) decreasing with increasing intensity due to self quenching.
Lamellar Assemblies
Layer by layer self-assembly of organic/inorganic films results in a ‘ photoredox onion' [32]. Polymers containing the electron acceptor pendant group MV-' adsorbed onto functionalized silica are spatially separated from polymers containing sensitizers like
Ru(bpy)j-' (polycations) or coumarin-fluorescein-porphyrin (poly-ions) by layered semiconductor materials like niobates or phosphates. In the presence of a sacrificial donor. PET across the semiconductor layer was observed and the combined energy/electron transfer process was found to have a quantum efficiency as high as 61%.
Heterogeneous semiconductor systems
When a semiconductor material is irradiated with photons of energy matching the bandgap, electron-hole separation can be effected. When the energetics are favorable, the charge separated electron and hole can be used for water reduction and oxidation respectively, in the presence of appropriate catalysts or in a photoelectrochemical cell y (PEC). The main issues with this approach are the bandgap that is often too high for visible light, efficiency and corrosion. Various LTV photocatalysts like niobates and
26 titanates have been reported to split water with co-catalysts [33]. Water oxidation with visible light in the presence of a sacrificial electron acceptor, AgNOj. has been reported with RbPbiNbOj, BiVO^ and IniO^CZnO)). Water oxidation by photoelectrocatalysis with AgCl coated electrodes in the presence of Ag' in the solution has also been reported
[34]. A recent report on a monolithic PV-PEC tandem cell involving a GaAs bottom cell and a GalnPi top cell, with two photon excitation capable of reducing water to H, with an efficiency of 12.4% has evoked a lot of interest (Figure 1.9) [35].
Self-assembled monolayer based photoelectrochemical cell
Fullerene thiol containing mixed self-assembled monolayers on gold electrodes are used to realize vectorial multistep electron transfer where Pyrene / Boron-Dipyrrin act as light harvesting unit and Ferrocene-Porphyrin-Fullerene (Fc-P-Cjo) triad as the reaction center [36]. Photoinduced electron transfer (PET) from P to Cf,„ followed by electron donation by Fc results in Fc'-P-C^o. Electron donation by C^,' to relays like MV-' and by the Au electrode to Fc' complete the circuit to generate a cathodic photocurrent. The small reorganizational energy of accelerates the forward ET while retarding the back
ET. Pyrene / Boron-Dipyrrin pigments that have overlapping emission with the absorption of P enhance the useful spectral range of absorption by undergoing energy
transfer from ('B*) to P. High quantum yields based on absorbed photons (-50%) and
incident photon to current efficiencies were reported ( 1 . 6 and 0.6% at 430 and 510 nm
respectively). This system similar to the semiconductor photovoltaic system produces an
electrical potential upon irradiation and hence different from our present approach.
27 a )
/ / / 7 Z
1 (/} 0 . c hv è 1 u 6 . i 1 o
1/
1.23V
ph2
4-0
Transparent Ohmic Semiconductor Metal interconnect contact Electrolyte interlace anode
Figure 1.9 Photovoltaic-photoelectrochemical system for water reduction. (a) Schematic representation (PV-PEC). and (b) Idealized energy level diagram (adapted from ref. 35).
28 Zeolite based systems
Early attempts to exploit zeolites for artificial photosynthesis involving tris(bipyridyl) complexes of ruthenium showed considerable promise for suppressing energy wasting back electron transfer rate and prolonging the lifetime of the charge separated species. Mallouk and co-workers investigated zeolite surface exchanged ruthenium tris(bipyridyl) complexes covalently linked to bipyridinium acceptors and found that the lifetime of the charge separated species could be five orders longer than in solution [37]. Recently Dutta and co-workers studied the totally intrazeolitic Rulbpy)/'- bipyridinium ion system using zeolite Y and found that up to 30% of the charge separated species formed during the 15 ns laser irradiation was present after I ms [38]. Efforts to convert the long-lived charge separation into permanent charge separation led to ternary systems using a second electron acceptor.
Dutta and co-workers surrounded the encapsulated Ru(bpy)/' in zeolite Y with a
2.2'-bipyridinium ion (4DQ-‘, E° = -0.65 V vs NHE) and suspended the zeolite particles
in a solution of propylviologen sulfate (PVS, E° = -0.4 V) as shown in Figure 1.10 [39].
Steady state photolysis of the deaerated solution turned the solution blue due to the PVS
radical anion, marking the permanent charge separation albeit with a very low efficiency
* (5 X 10"*). This system has been improved recently by Kincaid and co-workers by
employing an intrazeolitic donor molecule, Ru(mmb)]-' in a cage adjacent to the one
containing the sensitizer, Ru(bpy)ibpr^ [40]. The system employing both the ruthenium
polypyridyl complexes showed four-fold increase in radical production over the systems
employing only one of them. Though this approach results in permanent charge
29 Zeolite Solution
PVS O H D i
E ° = -0.65 V I SO3 = -0.41 V
Figure 1.10 Schematic representation of a zeolite based permanent photoinduced charge separation system (adapted from ref. 39).
separation, the present inability to access the oxidized component restricts the use of these systems for any practical applications.
PHOTOCHEMISTRY OF TRIS(2.2'-BIPYRIDYL)RUTHENIUM(II)
Photochemical and photophysical properties of ruthenium (II) polypyridyl complexes have been studied in extensive detail as they are finding extensive applications
in diverse fields [41]. Ru(bpy))-' is a popular choice of sensitizer for photochemical
water splitting systems and is finding increasing use as a building block for the
preparation of photoactive supramolecular assemblies as discussed before. Upon visible
excitation of Rufbpy)]-'. metal-to-ligand-charge transfer (MLCT) from the metal dit^
30 orbital to that of the k* of the bipyridine ligand occurs, resulting in the formation of a dTT^Ti*' excited state [42]. This initially formed excited state is largely singlet in character
('MLCT) and undergoes rapid intersystem crossing to a long lived triplet state (^MLCT) with a quantum efficiency close to unity. Besides the excited MLCT state, these molecules are often characterized by the presence of thermally accessible excited dd states (dTrMa*') that lead to photodecomposition [43]. The ^MLCT state is the one usually involved in the electron transfer, emission or quenching chemistry. This 'MLCT
state has favorable redox properties as shown below making Ru(bpy) 3-' a very good choice for water oxidation as well as reduction.
Ru(bpy)3^' — —— ------> *Ru(bpy)3-' Î.2. eV
Ru(bpy)3' ^ ------> Ru(bpy)3-
RUTHENIUM DIOXIDE
Ruthenium dioxide (RuOi) is a blue-black, tetragonal, metallic oxide with rutile
structure (Figure l.l 1) [44]. The optical property of RuOi in the UV-Visible region is
determined by the interband transitions (p-d) of the metallic oxide [45]. The room
temperature resistivity of RuOi has been reported to be in the range 3.5-5 x lO - Q cm
[44], making it more conductive than mercury (9.6 x 10'^ Q cm) but less than that of
copper (1.7 x 10 ’ cm) [46]. Good conductivity and catalytic activity, coupled with
low overpotential for O, and CL evolution has made RuO; attractive to diverse
technologically important fields. RuO? has found various interesting applications in
31 catalysis of organic transformations, electrocatalysis, photocatalysis and optoelectronics
[44].
Different forms of RuOj employed for studying the basic properties and potential applications include single crystals, polycrystalline films and powder, hydrous RuO^
(RuOi.nHiO) and ruthenium oxychlorides containing small amounts of chloride in the lattice. Single crystals and films are almost always prepared by thermal decomposition of polycrystalline RuOi and/or Ru metal powder at high temperatures in an oxygen stream
[47]. Polycrystalline powders are prepared either by precipitation in an alkaline medium or by the thermal decomposition of precursors like ruthenium chlorides and nitrates [46].
Recently low temperature decomposition of organometallic complexes of ruthenium has been explored as an easier alternative route to RuOi [45].
Application of RuO; in the field of water splitting has been studied by various
research groups. Taking a cue from the electrochemical studies of chlorine and oxygen evolution on RuO, electrodes, earlier efforts concentrated on the application of RuO, for
water oxidation [48]. Amouyal and co-workers were the first to study the catalytic use of
commercially available powder RuOn for light-induced Hi generation from water using
the Rufbpy)]-' - MV-' - EOTA system [49]. Using the same system, Kleijn and co
workers have studied the catalytic properties of colloidal RuOi prepared by the thermal
decomposition of RuClj [50]. Dutta and co-workers have prepared RuOi held to the
surface of zeolite Y by thermal decomposition of RujfCO),! and found such catalysts to
outperform various known catalysts for photochemical oxidation of water to oxygen.
32 # au
c O 0 I I i K < 4
Figure 1.11 Rutile Structure of RuOj (from ref. 44).
ZEOLITES
Zeolites are cry stalline aluminosilicates made up of oxygen comer sharing SiOj
and AIO 4 tetrahedra. The comer sharing tetrahedra form various simple substructures
like D4R and D6 R and complex building blocks like sodalite and pentasil. These building blocks along with substructures form a variety of complex frameworks that are often characterized by the presence of well defined channels, cages and windows (Figure
1.12). The presence of aluminum renders the framework negatively charged and is
balanced by alkali or alkaline earth metal cations that are ion-exchangeable. When the
ion-exchangeable cations are replaced by protons by chemical treatment, zeolites behave
like solid superacid. Because of tliese properties, traditionally zeolites have found
extensive applications in petroleum refining and catalysis.
33 Sodalite Cage
Stnicutre 6
Figure 1.12 Schematic diagram of various zeolite structures with sodalite as the building block.
34 The well defined void space inside the crystalline zeolite frameworks coupled with their ion-exchangeability offer intriguing possibilities in the present context [51].
Zeolites with their highly porous structure when synthesized as very small particles could act as an ideal host for artificial photosynthetic applications [52]. The presence of cations in close proximity would render zeolites as an active host with tunable electrostatic properties. The void volume could be controlled by the size of the counterion or by synthetically grafting spacer units. It is possible to build photoactive supramolecular structures with precise control, either inside the cages of zeolite by the ship in a bottle’ type synthesis or outside utilizing the surface silanol group.
SCOPE OF THIS WORK
The work described in this thesis was aimed at building a zeolite based hydrogen evolving artificial photosynthetic system. Toward this goal, we have adopted the well-
known sacrificial D- S-A model. EDTA-Rufbpyl^-'-bipyridinium-catalyst scheme. The
stability of the sensitizer. Rufbpy),-' was evaluated under photolytic conditions by
developing a chromatographic method to separate, quantitate and identify the
decomposition products. Effect of various photolytic parameters like zeolite
encapsulation. pH. photolysis time, quencher concentration, nature and concentration of
buffers were studied to find the optimum conditions where the decomposition could be
minimized.
It is well known that the electron acceptor employed in this system, bipyridinium
is susceptible to decomposition under H. evolving photolytic conditions, limiting the
practical applicability. The catalyst involved in the electron transfer from the relay is
35 widely believed to be responsible for this limiting condition. Our research group has earlier developed a zeolite-based RuOi catalyst for the photooxidation of water to Oi.
This catalyst was evaluated for the photoreduction of water to Hi using the above- mentioned system and was found to be effective. Performance of this catalyst was compared with other known catalysts as for as rate of hydrogen evolution, stability and the rate of bipyridinium decomposition. Attempts were made to selectively poison the sites responsible for the bipyridinium reduction, while keeping the sites involved in the
H; evolution unperturbed. The sensitizer was also covalently linked to the zeolitic
surface that also contains the catalyst as a first step toward a proposed integrated solar water splitting assembly.
Intrazeolitic electron transfer was studied with zeolite encapsulated sensitizer and
acceptor. Molecular modeling of guest species in constrained zeolitic cages involved in
photoelectron transfer allowed insight into their restricted rotational and diffusional
mobility. A compartmental systems dynamics modeling approach was employed to
simulate the intrazeolitic electron transfer processes and the developed model allowed the
extraction of various kinetic parameters.
36 REFERENCES
1. The Economist 2001, Feb 10. A survey of energy : page 6 .
2. Kordesch, K. V.; Simader, G. R. Chem. Rev. 1995, 95. 191.
3. "The International Energy Outlook - 2001," Energy Information Administration, The U. S. DOE, 2001.
4. The Economist 1999, July 22. Fuel cells meet big business.
5. McCusker, J. K.Science 2001,293. 1599.
6 . a) Hagfeldt, A.; Gratzel, M. Acc. Chem. Res. 2000, 3 3 . 269.
b) Nazeeruddin, M. K.; Pechy, P.; Renouard, T.; Zakeeruddin, S. M.; Humphry- Baker, R.; Comte, P.; Liska, P.; Cevey, L.; Costa, E.; Shklover, V.; Spiccia, L.; Deacon, G. S.; Bignozzi, C. A.; Gratzel, M. J. Am. Chem. Soc. 2001,123. 1613.
7. Appleby. A. J. J. Power Sources 1990. 29. 3.
8 . Larminie, J.; Dicks, “Fuel cell systems explained". John Wiley. Chichester, 2000.
9. Ashley, S. Scientific American 2001, July.
10. Hockaday, R. G.; Turner, P. S.; Maslow, M.; Cooper, M. Micro-fuel cell power devices. International Patent. WO 00/35032. 2000.
11. Finkelshtain, G.; Katzman, Y.; Khidekel, M.; Borover. G. .4 new class of electrocatalysts and a gas diffusion electrode based thereon. International Patent. WO 01/15253,2001.
12. “Camcorder fueled with Hydrogen”. April 09 Press Release, Fraunhofer Institute for Solar Energy Systems, Freiburg, 2001. http://w^ww.lse.fhg.de/english/press/pi_2001/pdf/lSE-Pl_Camcorder_HMI_e.pdf
13. http://www.medisel.com/img/battery 2 .gif
14. Bard. A. J.; Fox, M. A. Acc. Chem. Res. 1995, 28. 141.
15. Amouyal. E. Sol. Energy Mater. Sol. Cells 1995. 38 .249.
37 16. Griffith, R. T. (Translated with a popular commentary) “Hymns of the Rgveda”. J. L. Shastri (Ed), Motilal Banarsidass, Delhi, 1973. Vol. 1, pp 538.
17 a) Yagi, M.; Kaneko, M. Chem. Rev. 2001,101, 21. b) Ruettinger, W.; Dismukes, G. C.Chem. Rev. 1997.97, 1.
18. Zouni. A.; Witt, H.-T.; Kern, J.; Fromme, P.; Krauss, N.; Saenger, W.; Orth. P. Nature 2001. 409, 739.
19. Rhee. K.-H.; Morris. E. P.; Barber, J.; Kuhlbrandt, W. Nature 1998. 396, 283.
20. Moran, L. A.; Scrimgeour, K. G.; Horton. H. R.; Ochs, R. S.; Rawn, J. D. “Biochemisty”. Prentice-Hall. Englewood Cliffs, 1994.
21. Dismukes. G. C.Science 2001, 292, 447.
22. a) “Photoinduced Electron Transfer”, M. A. Fox and M. Chanon (Eds), Elsevier. Amsterdam. 1988. b) Kavamos, G. J. “Fundamentals of Photoinduced Electron Transfer”. VCH. New York. 1993.
23. “Energy Resources through Photochemistry and Catalysis”. M. Gratzel (Ed). Academic Press, New York. 1983.
24. Bolton. J. R. Sol. Energy 1996. 57, 37.
25. Sun. L.; Hammarstrom, L.; Akermark, B.; Styring, S.Chem. Soc. Rev. 2001. 30, 36.
26. Koryakin. B. V.; Dzhabiev. T. S.; Shilov. A. E. Dokl. Akad. Nauk. S.S.S.R. 1977.233, 620.
27. Lehn. J. M.; Sauvage. J. P. Nouv. J. Chim. 1977. /. 449.
28. Moradpour. A.; Amouyal, E.; Keller. P.; Kagan, H. Nouv. J. Chim. 1978. 2, 547.
29. Kalyanasundaram. K.; Kiwi, J.; Gratzel, M. Helv. Chim. Acta. 1978. 61,2720.
30. a) Steinberg-Yfrach, G.; Rigaud. J.-L.; Durantini. E. N.; Moore. A. L.; Gust. D.: Moore, T. A. Nature 1998.392.479. b) Gust, D.; Moore. T. A.; Moore. A. L. Acc. Chem. Res. 2001. 34. 40.
31. Sykora, M.; Maxwell, K. A.; DeSimone, J. M.; Meyer. T. J. Proc. Natl. Acad. Sci. U. S. A 2000, 97, 7687.
38 32. Kaschak, D. M.; Johnson, S. A.; Waraksa, C. C.; Pogue, J.; Mallouk, T. E.Coord. Chem. Rev 1999, 185-186,403.
33. Takata, T.; Tanaka, A.; Hara, M.; fCondo, J. N.; Domen, K.Catal. Today 1998, 44, 17.
34. Lanz, M.; Schurch, D.; Calzaferri, G. J. Photochem. PhotobioL, A 1999, 120, 105.
35. Khaselev, O.; Turner, J. A. Science 1998, 280, 425.
36. Imahori, H.; Norieda. H.; Yamada, H.; Nishimura, Y.; Yamazaki. 1.; Sakata. Y.; Fukuzumi.S.J. Am. Chem. Soc. 2001,123, 100.
37. Yonemoto, E. H.; Kim, Y. 1.; Schmehl, R. H.; Wallin. J. O.; Shoulders, B. A.; Richardson, B. R.; Haw. J. P.; Mallouk, T. E. J. Am. Chem. Soc. 1994, 116, 10557.
38. Vitale. M.; Castagnola. N. B.; Ortins, N. J.; Brooke. J. A.; Vaidyalingam, A.; Dutta. P. K. J. Phys. Chem. B 1999, 103, 2408.
39. Boija. M.; Dutta. P. K. Nature 1993. 362, 43.
40. Sykora, M.; Kincaid. J. Nature 1997. 387. 162.
41. a) Balzani, V.; Bolletta. P.; Gandolfl. M. T.; Maestri. M. Topp. Curr. Chem. 1978,
ô*. 1 . b) Balzani. V.; Moggi, L.; Manfrin. M. P.; Bolletta, P.; LawTence, G. S. Coord. Chem. Rev. 1975, 15, 321. c) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159.
42. Juris. A.; Balzani, V.; Barigelletti, P.; Campagna, S.; Belser. P.; Von Zelewsky, A. Coord. Chem. Rev. 1988. 84, 85.
43. a) Van Houten, J.; Watts. R. J. J. Am. Chem. Soc. 1976, 98. 4853. b) Durham. B.; Caspar, J. V.; Nagle. J. K.; Meyer, T. J. J. Am. Chem. Soc. 1982. 104.4803.
44. Seddon. E. A.; Seddon, K. R. “The Chemistry of Ruthenium”, Elsevier, Amsterdam. 1984.
45. Barreca, D.; Buchberger, A.; Daolio, S.; Depero, L. E.; Pabrizio, M.; Morandini, P.; Rizzi. G. A.; Sangaletti, L.; Tondello, E. Langmuir 1999, 15.4537.
46. Rard, J. A. Chem. Rev. 1985, 8 5 ,1.
39 47. Huang, Y. S.; Park, H. L.; Poliak, F. H.Mat. Res. Bull. 1982, /7, 1305.
48. a) Kiwi, J.; Gratzel. M. Angew. Chem., Int. Ed. Engl. 1979, 18, 624. b) Lehn, J. M.; Sauvage, J. P.; Ziessel, R. Nouv. J. Chim. 1979, 3, 423.
49. a) Amouyal, E.; Keller, P.; Moradpour, A. J. C. S. Chem. Comm. 1980, 1019. b) Keller. P.; Moradpour, A.; Amouyal. E. J. Chem. Soc., Faraday Trans. I 1982. 7,9, 3331.
50. a) Kleijn. J. M.; Rouwendal. E.; Van Leeuwen, H. P.; Lyklema. J.J. Photochem. PhotobioL, A 1988, 44, 29. b) Kleijn. J. M.; Boschloo. G. K. J. Electroanal. Chem. 1991. 300, 595.
51. Vaidyalingam, A. S.; Coûtant, M. A.; Dutta, P. K. Electron-transfer Processes in Zeolites and Related Microheterogeneous Media in “Electron Transfer in Chemistry”. V. Balzani (Ed), WILEY-VCH, Weinheim, 2001. Vol. 4, pp 412.
52. Castagnola. N. B.; Dutta. P. K. J. Phys. Chem. B 1998. 102, 1696.
40 CHAPTER2
PHOTOSTABILITY OF Ru(bpy)j*^ IN SOLUTION AND UPON ZEOLITE
ENCAPSULATION
INTRODUCTION
Novel photochemical properties of ruthenium polypyridyl complexes are being extensively exploited in many diverse applications. In analytical chemistry. tris(2,2'- bipyridyl)ruthenium(II) (Ru(bpy]^’*) is being used as a chemiluminescence label for quantitating various analytes [I]. Polypyridyl complexes of ruthenium are also finding increasing applications as a luminescent probe in oxygen sensors [2]. Ru(bpy)j‘' is a popular choice as photosensitizer for photochemical water splitting systems [3]. In addition. Ru(bpy)j‘* type molecules are also finding increasing use as a building block for the preparation of photoactive supramolecular assemblies [4a]. Photochemical and photophysical properties of ruthenium (II) polypyridyl complexes have been studied in extensive detail [4]. Upon visible excitation, a long-lived metal-to-ligand-charge transfer
(MLCT) state is formed, which is usually the state that is involved in the electron transfer, emission or quenching chemistry [4b]. Besides the excited MLCT state, these molecules are often characterized by the presence of thermally accessible excited dd states (dtt^do*') as sho'vn in Scheme 2.1 [5].
41 'MLCT—I d-d AE
I Photoproducts tS io
Scheme 2.1 Schematic representation of the energy levels and the excited state decay pathways for Ru(bpy)) A problem with the use of ruthenium polypyridyl complexes in photochemical applications is the photo-initiated decomposition arising from the population of the low lying dd states [5]. There are four possible strategies discussed in the literature to minimize the photodecomposition [4b,5b]. First is the preclusion of solvents of low dielectric constants and strongly ligating anions in the system. The photodissociation
Involves an intermediate in which the bpy ligand is partially de-ligated and coordinating anions can complex to the Ru promoting complete deligation [5b.6]. The use of aqueous solutions alleviates this problem to some degree, but often the aqueous solutions require the use of buffers where the buffer anions can exhibit significant coordinating ability.
Second is the use of lower temperatures which via a Boltzmann effect diminishes the population of the low-lying dd states [5], thereby decreasing photodissociation. However, in analytical and photochemical applications, experiments at lower temperature may not be practical. Third, for reactions in which a quencher molecule reacts with the MLCT state of Ru(bpy);-'. increased concentrations of the quencher should diminish the photodissociation chemistry [4b.5b]. Fourth, the photodecomposition can be alleviated if the dd ligand field state can be destabilized, thus lowering the probability of populating it.
It has been proposed that encapsulation in rigid media like cellulose acetate [7], cation
exchange polymer resins like Sephadex [ 8 ] or zeolites [9] should reduce the photolabilit)'.
The origin of the reduced photolability can arise from destabilization of the dd state as
well as the increased probability of bpy ring enclosure due to encapsulation.
Reports on photochemical water splitting applications of Ru(bpy)/' have often
relied on absorption spectroscopic study of the photolyzed solutions to estimate the
43 photodegradation [10]. In the case of heterogenous constrained media like zeolites, the conclusions regarding ligand loss were inferred from spectroscopic data such as emission lifetimes, and temperature-dependent lifetimes via the extent of destabilization of the ligand field state [9]. Even when photochemical studies were attempted in rigid media like poly(methyl methacrylate) (PMMA) there has been no attempt to separate the photoproducts [11]. Spectroscopic methods on unseparated mixtures containing small amounts of decomposition products makes it difficult to identify the photoproducts and also the extent of decomposition, especially when the products have spectroscopic properties similar to that of the parent Rufbpylj''. Also, to the best of our knowledge, there has been no study reported in the literature aimed at analyzing the photodegradation products of Rufbpy);-' entrapped in a heterogenous constrained media.
High-performance liquid chromatography (HPLC). especially if employed with diode-array detection is an ideal tool to study potentially complicated systems. Reverse-
phase HPLC has been used for separation and identification of a long lived intermediate
with a monodentate ligand in the photosubstitution ofRu(bpy),(dmbpy)'' [12]. Other
relevant applications of reverse-phase HPLC include the separation of 1) hydrolysis
products of tris(2.2'-bipyridyl)ruthenium(II) derivatives [13], 2) water reduction products
of Ru(bpy)/' [14], and 3) spontaneous reduction products of unstable Ru(IV) oxo
complexes in basic medium [15]. Use of reverse-phase HPLC to study the
photodecomposition of ruthenium(Il) polypyridyl complexes in dichloromethane has also
been reported, however the details of the chromatographic method has not been included
[16]. Ru(bpy)3 ‘‘ has also been studied chromatographically as a probe to evaluate
44 important electrostatic properties like the surface charge of various stationary phases employed in reverse phase HPLC involving silanol groups [17]. Cation exchange HPLC
has also been used for the synthesis and photochemical studies of bis( 2 ,2 '-bipyridyl) ruthenium(II) complexes [18].
In spite of the above mentioned applications of HPLC for the separation and identification of Ru(II) polypyridyl complexes, there has been no systematic study of the fate of Ru(bpy))-' upon photolysis in different environments. We have developed a reverse phase ion-pair HPLC method to estimate the extent of photodecomposition and to separate and identify the different decomposition products of Ru(bpy);-'. We have investigated the photolysis of Ru(bpy);-' in solution media with a set of commonly used buffers, in the presence of an electron transfer quenching agent and encapsulated in the supercages of zeolite Y.
EXPERIMENTAL SECTION
Synthesis: The nanocrystalline zeolite- Y was synthesized by Dr. Norma Castagnola. following a patent literature procedure, from aqueous silicate (27% SiO, in 14% NaOH.
Sigma) and aluminate (Na^ALO^.SHnO. Chem Service) solutions with a molar ratio of
SiOVALOj ~ 6 [19]. Intrazeolitic synthesis of tris(2.2'-bipyridyl)ruthenium(ll) inside the supercages of nanocrystalline zeolite-Y was carried out according to literature method
and was found to contain 33.6 pmol of Ru(bpy)/' per gram of zeolite, or - 1 complex per
15 supercages [20]. Silylation of the zeolitic surface of Rufbpy),*' containing nano crystals was accomplished using OTS (n-octadecyltrichlorosilane, 95% Aldrich) in toluene (AR. Mallinckrodt) [21].
45 Photolysis: A model AlOIO xenon-arc lamp equipped with a LPS 250 power supply
(Photon Technology International) was used for photolysis experiments. A 3" long water filter and a 420 nm glass cutoff filter were used to remove IR and UV radiation respectively, and the radiation was reflected off a mirror that reflects light in the range
420-650 nm and focused to a 1 cm spot. The power of the incident radiation as measured using a Coherent 210 power meter was found to be about 300 mW/cm’. Sodium phosphate monobasic hydrate (ACS, Fischer), sodium phosphate dibasic dodecahydrate
(Baker analyzed), sodium acetate trihydrate (Jenneile), glacial acetic acid (Mallinckrodt). potassium hydrogen phthalate (Baker analyzed), sodium hydroxide (AR. Mallinckrodt) and hydrochloric acid (ACS, Fisher) were used in the preparation of buffers in nanopure water (18 MQ cm ’).
Photolysis reactions were carried out either in buffer solutions of given pH and buffer concentration or in distilled water containing 2 * 10“* M Rufbpy),*' (Strem). In studies involving /V.A^-dimethyl-4,4'-bipyridine (Aldrich), weighed amounts of the reagent was added to the pre-made Ru(bpy);-' solutions. In the case of zeolite entrapped
Ru(bpy)3''. the hydrophobic silylated particles (33-35 mg) were suspended in toluene (5 mL) to get an optically clear 2 x 10^ M in Ru(bpy))“' solution. In all cases, 5 mL of the
solution in a modified spectrophotometric quartz cell ( 1 cm pathlength) was purged with
nitrogen gas for 45 min and was photolyzed for 8 hours, under stirring. Solution absorption spectra were recorded and HPLC separation were carried out, before and immediately after photolysis. For analysis during the time course of the photolysis,
absorption spectra were recorded at two hour intervals and 1 0 0 of the photolyzed
46 sample was pre-elevated using a syringe through the septa and analyzed immediately while the photolysis was continued. For chromatographic analysis of the zeolite sample, the toluene suspension was centrifuged and the separated solid zeolite particles were treated with citric acid (Jenneile Enterprises) to extract the zeolite-entrapped complexes
[22]. A known amount c/j-dichlorobis(2,2'-bipyridine)ruthenium(Il) dihydrate (Strem) to give 2 X 10"* M was suspended in water and stirred for 2 hours under N, atmosphere
shielded from ambient light. The insoluble Ru(bpy)iCl 2 remaining after two hours of
stirring was removed by centrifugation and the clear solution obtained was used for
further analysis.
Instrumentation: Absorbance measurements were carried out using a Shimadzu model
UV-265 spectrophotometer and diffuse reflectance spectra were recorded using a Harrick
diffuse reflectance accessory. Fluorescence spectra were recorded using a Spex Fluorolog
model FLl 12 spectrofluorometer equipped with a 450-W Xenon lamp and a Hamamatsu
R928 photomultiplier tube. Fluorescence lifetimes were measured by excitation with the
second harmonic (532 nm) of a Quantel Nd:YAG Q-switched laser ( I Hz), and by
monitoring the emission at 620 nm perpendicular to the excitation source. The emission
was monitored with Acton Research Corp SpectraPro 275 single monochromator,
Hamamatsu R928 PMT detector and Tektronix DSA 601 Digitizing Signal Analyzer,
with a 11A52 Two Channel Amplifier. Coming model 125 pH meter with a general
purpose combination electrode was used for all pH measurements.
A Shimadzu high performance liquid chromatograph equipped with a DGU-14A
on-line degasser, FCV-IOALv/? low pressure gradient flow controller, LClO-ATvp pump,
47 SPD-MlOAvp diode array absorbance detector (standard 10 mm optical pathlength cell), and 30 cm x 3.9 mm i.d. pBondapak octadecylsiiane (ODS) column (Waters Associates) was used for the separation studies. A static type solvent mixer was employed between the pump and the injector to ensure the homogeneity of the mobile phase in gradient analysis, with the mixer volume set at 1.7 mL for a mobile phase flow rate of 1 mL/min.
Samples were injected through a Rheodyne 7725i injector equipped with a 10 pi sample
loop. The mobile phases were degassed by ultrasonication, filtered through Whatman
grade 1 filter paper prior to use in a solvent reservoir equipped with an on-line filter.
Separation: Separations were carried out with a binary gradient using pH 5,0.025 M
phosphate buffer (A) and methanol (B) (Fisher Optima), both containing 5 x 10 " M 1-
heptanesulfonic acid sodium salt (Puriss, Fluka) as ion pairing agent with a flow rate of
1.0 mL/min. Optimum separation was achieved with a 12 min linear gradient at 5%
B/min from 20% B to 80 % B and held for 8 min. Equilibration time between runs was
10 minutes. The separated components were detected by their absorption spectrum (190
to 550 or 600 nm) and the chromatogram were recorded at given wavelengths.
Calibration was carried out with six standards of Ru(bpy)/' of known concentration in
the range of 1 x I O'" M to I x 10"’ M Three independent injections of the standards were
carried out and the average was employed. The peak area as a function of concentration
was found to be linear (r^ = 0.999) and this calibration curve was used to estimate the
extent of Rufbpy);-' decomposition.
48 RESULTS AND DISCUSSION
Chromatographic methodology:
The use of reversed-phase high performance liquid chromatography (RP-HPLC)
for the analysis of metal complexes is rapidly gaining in popularity [23]. Among the
various methods of RP-HPLC, the ion-pair (IP) technique is often used to control
retention and selectivity, to achieve the simultaneous separation of samples containing
neutral and ionized molecules. Several theoretical models such as the ion-pair model, the dynamic ion-exchange model and the electrostatic model have been proposed to describe
the retention behavior in IPRP-HPLC. In the simplest ion-pair model it has been
proposed that ion-pairing reagent forms a complex with the analyte in the mobile phase
whereas in more complex models, the IP agent has been proposed to adsorb reversibly to
the stationary phase by hydrophobic interaction prior to complexation with the analyte.
Separation in IPRP- HPLC is controlled by a number of experimental variables:
ion-pair agent (type, size and concentration), organic modifier (type and concentration),
pH, temperature and stationary phase. For benzalkonium salts, strong and weak bases
and inorganic ions, the commonly employed ion-pairing agents are alkyl- or
arylsulfonates and alkylsulfates. For the separation of ruthenium polypyridyl complexes
alkyl sulfonates are commonly employed and we chose 1 -heptanesulfonic acid sodium
salt [13,24] at a commonly employed IPA concentration of 5 mM [13-15], while others
have also used octanesulfonic acid sodium salt [14,15]. For chemically bonded stationary
phases, the working range is about pH 2-8. As our photochemical studies were mostly
carried out at pH 5, the aqueous mobile phase was also maintained at pH 5 using
49 phosphate buffer (25 mM), a commonly employed chromatographic buffer with little coordinating ability and high optical transparency at a concentration to prevent salting-out of other ionic components in the presence of an organic phase.
Since our HPLC instrument was not equipped with a column oven all separations were carried out at room temperature. We chose to work with a C,g bonded stationary phase as it is one of the commonly employed general purpose column and it has also been used in the literature for the separation of ruthenium polypyridyl complexes [13-15,25].
This left the type and concentration of the organic modifier present in the mobile phase to be manipulated to achieve the desired separation. In IPRP-HPLC the retention decreases with increasing solvent strength i.e., with increasing concentration of the organic modifier. The commonly employed organic modifiers in the separation of ruthenium polypyridyl complexes are tetrahydrofuran [13-15] and methanol [15,25]. We chose to work with methanol and decided to study the effect of mobile phase composition consisting of methanol and pH 5 phosphate buffer (25 mM), both containing 1- heptanesulfonic acid sodium salt (5 mM) on the separation efficiency using a C,g bonded stationary phase (pBondapak ODS column). As the aim of the present work is to study the photostability of Rufbpy),** under various conditions including in the presence of an electron accepting, oxidative quencher. iV,A’-dimethyl-4.4'-bipyridine (methylviologen,
MV-'), the chromatographic method employed should be capable of separating
Ru(bpy)/' from the decomposition products as well as from MV-'. Hence the choice of initial separation conditions were also influenced by the reported chromatographic conditions for MV*' (see chapter 3 for a detailed discussion) [26].
50 Initial attempts with fixed mobile phase composition showed high retention times at low methanol compositions and at high methanol compositions the separation of
Ru(bpy)3'* from MV** was poor. Modification to a binary gradient method, starting with a relatively ‘weak’ solvent composition of 80 % phosphate buffer (A) : 20% methanol (B) and applying a continuous gradient of (5 % B/min) to a final strong’ composition of 20%
A ; 80% B that was held, gave good separation and low peak broadening within
reasonable time. The hold time at (20:80, A:B) was 8 min at the end of which the composition was changed to the starting composition. The elution was monitored using the diode array detector 190 - 550 nm (or 190 - 600 nm) the entire time, including the equilibration time of 10 min resulting in 30 min long runs.
Figure 2.1 shows the programmed composition change and the resulting observed absorbance (chromatogram) for methanol at 198 nm and at other relevant wavelengths
(256.450 nm). Though the gradient program with increasing methanol composition
started at 0 min. the time it took to observe this by the increase in absorbance at 198 nm is about 5 min. The dead volume of the column presently employed has been reported to be
2.4-1 . 8 mL [13.26a] and we have also evaluated it to be about 2.4 mL. Hence the remainder must be due to the dead volume associated with the plumbing including the solvent mixer with a capacity of 1.7 mL, employed to ensure the homogeneity of the mobile phase in gradient analyses. The increasing absorbance of the mobile phase with
increasing methanol content can cause anomalously high absorbance below 2 2 0 nm for
the eluted components, restricting the useful wavelengths to > 2 2 0 nm, however, this did
not affect the separation or identification of the components in any manner.
51 100 Method - 0.6 198 nm 256 nm 450 nm i 0.4 ■S
- 0.2 X)1 <
- 0.0
0 5 10 15 20 25 30 Time (min)
Figure 2.1 Chromatographic separation method (gradient program) showing the mobile phase methanol composition and background chromatograms observed at various relevant wavelengths. ® rest - pH 5 phosphate buffer (25 x 10'^ M), both components containing 1-heptanesulfonic acid sodium salt (5 x 10^ M).
52 The chromatographic method was evaluated for the separation of Ru(bpy) 3 *' from
MV‘* and Figure 2.2 shows the chromatogram monitored at 256 nm and 450 nm for a
sample that contained 2 x 10"' M Ru(bpy) 3 -" and I x 10'^ M MV-*. The chromatogram showed peaks with retention times (RT) of 2.4 min. 12.34 min, and 14.18 min.
Comparison with the chromatograms of pure sample and the UV-Vis absorption spectra obtained from the diode array detector indicated that the peak at 12.34 min (M) is that of
MV-% the peak at 14.18 min (R) is that of Ru(bpy) 3‘* while the peak at 2.4 min is due to unretained anionic salts like acetate and EDTA. Separation of polar and ionic compounds like cationic organometallic complexes and organic bases on an ODS column by (RP-
HPLC) is often problematic due to the surface charges present on the chromatographic stationary phase. Electrostatic attraction between the analyte and the silanol groups as well as acidic sites due to metallic impurities often give rise to zone tailing and increased retention [17]. By manipulating the mobile phase composition and by employing 1- heptanesulfonic acid sodium salt as the ion-pairing agent, we were able to reduce the peak broadening significantly.
The peak asymmetry factor, calculated as the ratio of the trailing to leading half width at 10% of the peak height [17], for the chromatogram shown in Figure 2.2. was 2.7 for MV-' and 3.6 for Rulbpylj’*. Peak asymmetry factor greater than one due to deviation from Gaussian peak shape (tailing) is a reflection of the surface charges, but is still adequate for the separation of quarteraary nitrogen containing organic bases and ionic compounds on ODS column. Peak asymmetry factors greater than 4 has been reported for Rufbpy);-" on a monomeric C,g stationary phase bonded to type A silica at pH 5 [17].
53 a) 0.25 256 nm 450 nm 0.20 E e
- 0.15 e3
O.IO
- 0.05 S)I < Jl 0.00
10 15 20 Time (min)
2.0 0.25 b) 0.20 £ I e VO o 0.10 ■eo - 0.05 1 < 0.0 - 0.00 12 13 1514 Time (min) Figure 2J. Chromatographic separation of Ru(bpy)g^* (2 x 10^ M), (1 x 10 ^ M), and EDTA (0.2 M) in pH 5 buffer, a) Chromatograms monitored at 256 nm and 450 nm. b) Expanded chromatographic region of interest showing peak asymmetry. 54 Other calculated chromatographic parameters are given in Table 2.1 and shows good separation (Resolution factor, a = 7.75) and high efficiency (height equivalent to a theoretical plate, HETP < 5 x 10"* cm). Performance of the column over the period (> year) of this investigation deteriorated with continuous use, indicated by an increase in retention time of MV*' and Ru(bpy);-' to about 13.5 min and 16 min, respectively. Also the peak width increased leading to the decrease in the calculated column efficiency (HETP > 5 x 1 0 "^ cm), along with an increase in the peak asymmetry factor. Similar observation has been reported in the separation of 1,10-phenanthroline complexes of Fe(ll). Ni(ll), and Ru(Il), with the retention time of Ru(phen);-" increasing from 10.24 min on a new column to 19.6 min and 24.51 min on the same column after six and ten months of use. respectively, along with an increase in the degree of tailing [27]. This behavior has been explained on the basis of a continuous loss of the organic phase along with the strong adsorption of the ion-pairing sulfonates on the silanol surface making the stationary phase more "polar” [27] and a similar situation could exist under the present circumstances. Figure 2.3 shows the results of Ru(bpy) 3*' calibration monitored at 450 nm with a set of six standards in the concentration range of 1 x 1 0 - M to 1 x 10'^ M. With a sample volume of lÛAiL, this corresponds to a range of I x io '“ to 1 x 10 * mol of Rulbpy);"' with a linear detector response over the entire range. With the higher molar absorptivity of Ru(bpy).-' at 280 nm coupled with the excellent detector S/N parameters, the detectabilit}' limit of ruthenium polypyridyl complexes would be at least two orders of magnitude less than 1 x 10'"* mol. With a Ru(bpy) 3 ‘* concentration of 2 x 10"* M 55 Retention Volume* MV- 12.34 Vr (mL) Ru(bpy)3 -' 14.18 MV-- 0 . 1 2 FWHM" (mL) Ru(bpy)3 -' 0.16 MV-- 2.7 Asymmetry Factof, t| Ru(bpy)3 ‘* 3.6 MV-* 4.00 Capacity factor^, k' Ru(bpy)3-' 4.74 Separation factor*, a 1.19 Resolution'. R, 7.75 MV-* 5.12 X 10 * HETP« (cm) Ru(bpy)3-' 6.89 X 10-* Theoretical Plate MV- 58500 Number". N Ru(bpy)3-' 43500 Mobile phase flow rate = 1 mL/min; FWl- M is peak width at 50% peak intensity; ‘ Asymmetry factor is the ratio of the trailing to leading half width at 10% of the peak maximum intensity; Capacity factor = (Vg-VJ/Vg; ' Separation factor = (VR2 -VJ/(VR,-Vg); ' Resolution =1.18 (Vr,-Vr|)/(W, + WJ; * HETP, Height equivalent to a theoretical plate = L / [5.54(Vr, / W,)*]; Theoretical Plate Number = L/HETP; where V^, and W, are the retention volume and FWHM of MV-\ and W, are the retention volume and FWHM of Rufbpy)^-'. respectively; V„, void volume of the column = 2.34 min; L. height of the column = 30 cm; Table 2.1 Calculated chromatographic parameters for the chromatographic separation of Ru(bpy)3‘' and MV-'. 56 0.25 J a) ------1 * 1 0 "M I 0.20 - o 0.15 ® 5*10"* M 8 0.10 0.05 - I A ------2 * 1 0 "* M < P \ i* io 1 m 0.00 ------5*10 :M 0 10 15 20 25 30 Time (min) 12 b) 10 8 I 6 < 4 I 2 0 0 10 [Ru(bpy),'l (10-^ M) Figure 2 3 Ru(bpy)g * Calibration, a) Chromatograms as a function of concentration (three injections were made for each concentration and the chromatograms are shown), b) Calibration plot showing the peak area as a linear function of concentration (r^ = 0.999). 57 employed, the estimated lower detection limit would then allow monitoring decomposition as low as 0.05%. Though the standard deviation of Ru(bpy)]-' peak area is of the order of 1 %, the detection of new decomposition products at very small quantities would be possible with the estimated detection limit. The retention time of RuCbpy),*' peak in Figure 2.3a changes from 16.23 min at 1 X 10^ M to about 17.68 min at 1 x 10'^ M with 16.48, 16.61. 16.75. 17.05 min at intermediate concentrations and such behavior is indicative of the amount of analyte exceeding the linear range for the column that we could not avoid. However this concentration dependant change in retention time did not adversely affect the separation process, as all the components of interest elute before that of Ru(bpy) 3‘' (see below). Strong concentration dependence of capacity factor {k') for RuCbpy),’* at sub-millimolar concentrations has been observed with monomeric C,, stationary phase on type A as well type B silica, with the leading end of the zone eluting earlier, as the concentration exceeds the linear region [17]. However, this non-linearity in the chromatographic efficiency should not be confused with the linearity of the detector response. Aquation chemistry of Ru(bpy)zCl2 Any chromatographic separation method used in the study of photodecomposition of Ru(bpy)3 ‘‘ should be capable of separating complexes that differ in their net charge, and more importantly, species that have the same charge with slightly different ligands that may occur as the photoproducts. As some of the primary products may be photoactive that can undergo cis-tram photoisomerization, the method to be employed should also be capable of separating geometrical isomers. In addition to these 58 requirements the mobile phase composition employed in the chromatographic analysis should not promote any unwanted secondary changes such as ligand exchange during the separation process. In order to evaluate the suitability of the current method it was decided to study the well known aquation of Ru(bpy)iCli and the photochemistry of the resulting aquo products. Ru(bpy),Ch - Ru(bpy),(S)Cr - Ru(bpy),(S)f' (2.1) Ru(bpy)iCli was sparingly soluble in water as the solubility arises from the replacement of Cl ligands by HiO leading to the formation of cationic complexes. Ru(bpy)nCl2 is known to be substitutionally labile, with both the Cl' ligands capable of being replaced by the solvent molecules (S) yielding Ru(bpy),(S) 2 ' ‘ as shown in equation 2.1 [28]. It has been observed from conductance measurements that Ru(bpy)iCh in aqueous solutions behaves like a2 : 1 electrolyte confirming the complete aquation leading to Ru(bpy)2 (OH 2 ) f * [28]. In fact this procedure is a common synthetic route for Ru(bpy)nXY"^ type molecules [29,30]. Upon chromatographic separation of an aqueous Ru(bpy)iCl2 solution, peaks were noted with retention times of 11.0 (peak *), 12.4 (peak T) and 13.3 min (peak C), with the first peak being very small and the last peak being the most intense. The intensity of the 11.0 min peak decreased with the time of stirring before analysis. Ru(bpy);(0 Hi)2 ‘* is known to undergo cis-tram photoisomerization in acidic aqueous solutions yielding a tram rich photostationary state with thermal conversion of tram to cis (equation 2.2) [31 ]. Photolysis of the aqueous solution of Ru(bpy)iCl2 resulted in increase of the intensity of peak (T) with concomitant decrease of peak (C) and the resulting chromatograms are shown in Figure 2.4a. The absorption 59 a) 12 Before o 1 hour 2 hour 8 I 6 4 2 I 0 < 2 10 12 14 16 18 20 Time (min) 4 b) 3 ■o 2 350 400 450 500 5:0 1 0 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.4 Photolysis of aqueous Ru(bpy)2 Cl2 a) Chromatographic separation of the photoproducts, b) Absorption spectra of the peaks marked T and C in the chromatograms, as obtained horn the diode array detector. 60 spectra of species T exhibited bands at 245, 295,353 and 501 nm, whereas species C showed bands at 243, 291, 340 and 486 nm (Figure 2.4b). The absorption spectra of the species (*) that elutes with a retention time of 11.0 min is shown in Figure 2.5a and has bands at 246,295, 356 and 512 nm. hv cis-Ru(bpy)2(OH2)2‘* ^ trans-Ru(bpy)2(OH2)2‘' (2.2) Comparison of the absorption spectra of species T and C with literature suggest that the peaks labeled (T) and (C) are due to that of trans- and c/j-Ru(bpy),(OH 2 )2 ‘’, respectively [31]. Though a large number of Ru(bpy),XY"* type complexes show similar spectra, the lowest energy band due to the Ru(II) metal centered orbitals (Ru 4dJ to bpy 7t* metal-to-ligand-charge-transfer (MLCT) band shows significant changes depending on the nature of the ligands X and Y. The MLCT band for the cis isomers is known to occur at higher energy than that for the trans isomer and is consistent with our assignment. This property has been attributed to the d, levels of the cis isomer being more stable than that of the trans isomer due to the enhanced it stabilization imparted by having the strong bpy a acids trans to the weak ancillary it acids (Cl', H 2O etc.) rather than trans to each other [24]. When the non-bpy ligands (X or Y) are 7t-donating in nature, Ru 4d, get destabilized relative to the tr-accepting bpy ligand centered orbitals and as a result the metal to ligand charge transfer (MLCT) gets red shifted. The small peak (*) at 11.0 min showed a MLCT band at 512 nm, red shifted compared to the bisaquo complexes. Compared to the ligand, HjO, present in the final aquation product. Cl , present in the starting complex is a better Tt-donating ligand. The observed red shift in addition to the fact that it disappears with 61 a) 5 4 3 y 2 I 1 < 0 250 300 350 400 450 500 550 Wavelength (nm) 10 b) 8 6 4 190 490 900 5 0 i 2 0 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.S Absorption spectra of a) the il.0 min peak in the chromatogram of photolyzed aqueous RufbpyljClj. b) 12.6 min peak (the only component) in the chromatogram of methanolic RufbpyljC^. 62 time suggests that this species could be cw-Ru(bpy)i( 0 H2 )C r, the first (incomplete) aquation product that leads finally to the significant product observed, cis- Ru(bpy)i(0H2)2-\ The appearance of small amount of species attributed to the trans-isomer in the "before photolysis” sample could be due to the small extent of photoisomerization during the handling of the solution in the ambient environment. Also, it should be noted that the reported molar absorptivity at 495 nm for the trans isomer (11,800 M 'cm ') is higher than that of cis (7,300 M 'cm ') [31], making the small trans component seem larger, when monitored at 496 nm. The ratio of the intensity of the bands due to trans. cis increased with photolysis (Figure 2.4a) and is consistent with the reported photoisomerization of Ru(bpy)2 (0 H,)i-' favoring the trans isomer [31 ]. However no attempt was made to estimate the ratio of [trans\l[cis] quantitatively in the present study as our aim was only to use this system to evaluate the capability of the chromatographic method. This photoisomerization is thought to proceed via the photodeligation of one H,0 molecule followed by the conversion of the pentacoordinate intermediate from a square pyramidal to trigonal bipyramidal structure followed by religation [31 ] similar to the photoracemization of Ru(bpy)/' [32], The observed area ratio of the peaks assigned to trans'xis obtained after photolysis could be reversed by heat treatment, without any changes in the absorption spectra. It has been reported that due to steric crowding of the bulky bpy ligands in the trans configuration, the cis isomer is thermally more stable [31 ] and our observation is consistent with this report. However we could not obtain complete recovery of the peak intensity due to the starting cis form. The chromatographic 63 observations could be summarized as stepwise aquation followed by cis-irans photoisomerization, with thermal conversion of the trans to cis isomer. It has been reported that during the photochemical studies of RuCbpyl^lviz),'' (where viz = N-vinylimidazole) by cation exchange HPLC, the photoproduct [Ru(bpy)2 (viz)HiO]-' was stable dunng the elution with acetonitrile containing mobile phase whereas [RuCbpyljHnOCl]* rapidly exchanged water for acetonitrile [18]. In the HPLC analysis of the reduction products of polypyridyl complexes containing Rd^=0. it has been suggested that substitution of aquo ligands by acetate anion present in the mobile phase is possible [15]. In order to eliminate the possibility of ligand exchange during the chromatographic separation on the column, we also analyzed a methanolic solution of Ru(bpy)iCL. Ru(bpy)iCL is much more readily soluble in MeOH compared to water and the chromatographic analysis showed only one primary peak with a retention time of 12.6 min. The absorption spectra of this species is shown in Figure 2.5b and we tentatively assign this species to m-Ru(bpy);(MeOH) 2‘'. As the retention time and the absorption spectra of this species is different from the ones observed for Ru(bpy)nCL in water (cis- and trans- RulbpyljlOH,)!"'). we conclude that there could be no significant ligand exchange processes occurring during the chromatographic separation involving methanol/phosphate buffer in the mobile phase. Photoaquation chemistry of Rufbpy)]^^ Ru(bpy)3 ‘" was originally considered to be a photochemically inert sensitizer towards deligation of bpy and ligand substitution, however increasing evidence suggest the contrary [4b]. Van Houten and Watts were the first to report the photosubstitution of 64 Ru(bpy)3 ** and found that the complex that was photoinert at 25 °C in 0.1 M HCl underwent deligation at 95°C when irradiated with 436 nm light [33]. These authors, in a subsequent detailed investigation, with the help of temperature dependent studies have shown the dd state that lies - 3500 cm ' above the ^CT to be responsible for the photoactivity [5a]. The quantum yield of Ru(bpy) 3‘* disappearance was found to be temperature and pH dependant in the range of 10'’ - 10'\ however the exact nature of the products were not clear [5a]. Other workers have subsequently reported similar photodecomposition studies in organic media such as dimethylformamide [ 6 b], dichloromethane [5b,6a,34] and acetonitrile [35] in the presence of coordinating anions like chloride or thiocyanate and quantum yields were in the range of 10"' - 10'\ Quantum yield in these studies were determined from the unseparated photolysed solution either from absorption [5b,34] or emission [5a] measurements. Meyer and co-workers’ study of the photochemistry of Ru(bpy)3 ’‘ in dichloromethane essentially agreed with the earlier proposals by Van Houten and Watts [5a] and concluded that the apparent photochemical stability in water is a ramification of the importance of ring closure and not due to inherent low photochemical activity and the proposed mechanism is shown in Scheme 2.2 [5b]. The mechanism of the photodecomposition is thought to proceed by a dissociative mechanism through population of the thermally accessible ^dd which leads to a Ru-N bond lengthening resulting in dechelation and photosubstitution by the ion-paired counteranion. The intermediate in this process is thought to involve a five-coordinate complex with one of the bpy ligands attached via only one of the nitrogens. The 65 d-d =bpy X = Coordinating anion or solvent Q = Quencher +X -bpy Scheme 2.2 Schematic representation o f the mechanism proposed in the literature for ligand photosubstitution reactions o f Ru(bpy))-^ (adapted from ref. 5b). previously observed photoracemization of optically pure Ru(bpy)j‘"^ in aqueous solutions at neutral pH [32] has also been explained by this mechanism wherein the pentacoordinated square pyramidal intermediate could rearrange to a trigonal bipyramidal product which can lead back to either of the isomer [5b]. Photolysis o f Ru(bpy)/* in water Photolysis of Ru(bpy)j’" in distilled water (2 x 10"* M) for 8 hours showed very little change in the solution absorption spectra, with slight broadening of the lowest energy absorption at 450 nm (Figure 2.6a). The solution after 8 hours of photolysis was analyzed by HPLC and the chromatogram is shown in Figure 2.6b. The peak marked (R) in the chromatogram and eluting at 16.3 min exhibited a spectrum identical to Ru(bpy)/ and eluted at about the same retention time as the unphotolyzed sample. The other two peaks (T) and (C) in the chromatogram after photolysis were similar in retention time (Figure 2.7a) and absorption spectra (Figure 2.7b) to those obtained in the case of Ru(bpy)2 CL aquation. The area of the Ru(bpy);'' (peak R) decreased upon photolysis as compared to the unphotolyzed sample and using the calibration curve it was estimated that the extent of decomposition after 8 hours of photolysis was about 8 %. A determination of the photochemical quantum yield from our present data is not possible as light with a wide range of wavelengths (420-650 nm) was employed for the irradiation. However the results can be used to compare the various systems we have employed when the experimental conditions like lamp power, tjme of photolysis and concentration of Ru(bpy).-' were maintained the same. The reported low quantum yield in water for Ru(bpy)3 ‘* photodecomposition at 343 K (<10‘^) [5a] and that of 67 a ) 4 Before After 3 2 1 1 0 200 300 400 SOO 600 700 800 900 Wavelength (nm) 5 b) Before After 4 E e 0 3 2 1 1 < 0 10 12 14 16 18 20 Time (min) Figure 2.6 Photolysis of Ru(bpy)^^* in distilled water (2 x 10^ M) for 8 hours a) Solution absorption spectra, b) Chromatographic separation of the photoproducts. 68 a ) 2 Before I d After S c 1 0 0 1 1 2 13 14 15 Time (min) 20 b) 15 10 4 0 0 490 MO 5 0 1 5 < 0 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.7 Photolysis of Ru(bpy)^^* in distilled water (2 x 10“* M) for 8 hours a) Chromatographic separation of the photoproducts, b) Absorption spectra of the peaks marked T and C in the chromatogram. 69 photoracemization at 276 K (5.2 x 10'’) could explain the observed low extent of photodecomposition [32]. At the sub-millimolar concentrations of Ru(bpy) 3CU examined in this study, photolysis of an aqueous solution led to the formation of two decomposition products, which on the basis of retention times and absorption spectra were identified as irans- and CM-Ru(bpy ) 2 (OHi)?- . We have no experimental evidence as to whether the irans- species is a primary photoproduct Ru(bpy);-' arising from the conversion of the pentacoordinate square pyramidal to trigonal bipyramidal conversion or a secondary product from the cis-trans photoisomerization of c/j-Ru(bpy )2 (OHi);-'. Considering the low concentration of the chloride ions (-4 x 10"* M) in the medium, the absence of any other peak, for example the one previously attributed to c/ 5 -Ru(bpy)i(OH2 )C r is hardly surprising. As shown in Scheme 2.3, the photochemistry occurring in aqueous solutions is therefore photoaquation with the bpy ligand replacement occurring directly via water attack (reactions 1.2, 3 in Scheme 2.3), resulting in the formation of c/ 5 -Ru(bpy), (OH,),'* which can undergo photoisomerization (reaction 4 in Scheme 2.3) to irans- Ru(bpy),(OH,),-*. Photolysis o f Ru(bpy)j‘* in buffers As mentioned before there have been several studies aimed at understanding the photochemistry of Ru(bpy)j'* in organic media [5b,6,35] and a few in aqueous media [5a,32], but none of these studies have addressed the issue of photodecomposition under practical conditions. Most of the numerous reports aimed at water splitting involving Ru(bpy)3'* with very few exceptions have also ignored the fact about photodecomposition. However, Golin and Getoffhave reported that Ru(bpy) 3'* is not 70 ® ® H ,0 hv HjO 12.3 min 13.3 min © h v | | ® hv 1 H ,0 14.1 min 12.3 mm Scheme 2.3 Proposed mechanism for the photolysis of Ru(hpy)/* in distilled water and in acetate bufïers. photochemically stable during prolonged irradiation ( 6 hrs) under hydrogen evolving conditions (5 x 10 ' M RuCbpy),’*, 4 x 10'^ M MV-‘ and 4 x lO'^ M EDTA with PtO,), though no detailed study has been made [36]. Most photocatalytic water splitting studies involving Ru(bpy)]-' require buffers of pH in the range of 4-5. From the literature it is clear that acetate buffer is the medium of choice [10,26c,37] and buffer concentrations up to 2 M have been employed [37bcj. Another effective buffer in this pH range involves potassium hydrogen phthalate (KHP/NaOH). Though phosphate buffers are not good choices due to the poor buffer capacity in this pH region. Rufbpylj’" photodecomposition was studied in this medium also, as phosphate anions are known to be poorly coordinating. We have examined the extent and decomposition pattern of Ru(bpy)r' in various relevant buffers: a) pH 5 acetate (0.025 M, O.l M and 2 M). b) pH 4 acetate (2 M), c) pH 5 phosphate (0.025 M. 2 M) d) pH 5 KHP/NaOH (0.025 M). and e) pH 3 KHP/HCl (0.025 M). Acetate buffer (pH 5) The color of the Ru(bpy);-' solution changed during photolysis and was easily noticeable in the case of 2 M acetate changing from initial yellow-orange to orange and then finally to red-orange, and this change was much less pronounced with decreasing buffer concentrations. The Rufbpylj** solution absorption spectra before and during photolysis at various time intervals in the case of 2 M acetate buffer are compared in Figure 2.8a, and show sharply decreased absorbance in the 450 nm region, the appearance and growth of a shoulder at 500 nm along with three noticeable isosbestic points at 475. 381 and 335 nm in the visible region. This is remarkably similar to the isosbestic points 72 a ) 4 Before 2 Hour 4 Hour 3 6 Hour 8 Hour 2 x>I < 1 0 200 300 400 500 600 700 Wavelength (nm) 5 b) Before 2 hour 4 4 hour S e 6 hour 0 3 8 hour 2 1 1JD < 0 1 0 12 14 16 18 20 Time (min) Figure 2.8 Photolysis of Ru(bpy)^^* (2 x 10^ M) in pH 5 acetate buffer (2 M) for 8 hours a) Solution absorption spectra, b) Chromatographic separation of the photoproducts. 7 3 (480, 393, and 338 nm) observed for Ru(bpy) 3 ‘* in DMF in the presence of added thiocyanate [ 6 b]. This resemblance, though the medium and counterions were different, is probably due to the fact that the different photoproducts have similar absorption spectra. The chromatograms obtained with the photolyzed solution (2 M acetate buffer) at various time intervals of photolysis are shown in Figure 2.8b and Figure 2.9a. At all photolysis time intervals, along with the Ru(bpy)j‘' peak, three new peaks with RT of 12.3±0.1, 13.3±O.I and 14.1±0.1 min were observed, and their absorption spectra are shown in Figure 2.9b and Figure 2.10. The peak due to Rufbpy),- became smaller with photolysis time indicating substantial photodecomposition. The retention times of the first two peaks were similar to the previously observed peaks due to trans- and cis- RufbpyljlOHi);*'. After the first two hours of photolysis, the spectrum of the 12.3 min component matched that of /ra/tr-Ru(bpy) 2 (0 H2 )i-' (Figure 2.9b). However, with continuous photolysis the spectrum showed subtle changes with the lowest energy band shifting from 500 nm to higher wavelengths and after 8 hours of photolysis shifting to about -515 nm (Figure 2.9b). Close inspection of the contour plot showed that this peak was indeed due to two different components that were not sufficiently resolved (within 0.1 min). The early eluting component showed a spectrum (Figure 2.9b) that matched that of the trans- species, while the late eluting component had a peak at -520. The absorption spectra of the second peak at 13.3 min did not change with photolysis time (Figure 2.10a) and was found to be the same as that of the c/ 5 -Ru(bpy)2 (OH;)2 ‘*. The 74 a ) 12 Before 10 2 hour 4 hour S 8 s 6 hour 0 8 hour 6 4 2 1JO 0 < 2 12 13 14 15 Time (min) b) o 20 250 300 350 400 450 500 550 600 Wavelength (nm) Figure 2.9 Photolysis of Ru(bpy)g^^ (2 x 10^ M) in pH 5 acetate buffer (2 M) for 8 hours a) Chromatographic separation of the photoproducts, b) Absorption spectra of the 12.3 min peak, after two hours of photolysis (— ); and after eight hours of photolysis (—»—), a composite of early eluting component (—— ) and later eluting component (—#—). 75 a ) 3 2 y 3S0 400 SOO iSO I < 0 200 250 300 350 400 450 500 550 600 Wavelength (nm) b) 12 9 y 6 I MO S» * 0 < 3 0 250 300 350 400 450 500 550 600 Wavelength (nm) Figure 2.10 Absorption spectra of the peaks observed in the chromatograms of photolyzed Ru(bpy)j^^ solution (2 x 10“* M) in pH 5 acetate buffer (2 M) a) 13.3 min peak, after 2 hours (- -) and 8 hours of photolysis (—^ ) and b) 14.1 min peak, after 2 hours (- -) and 8 hours of photolysis (—#—). 76 prominent new peak at 14.1 min showed peaks at 292, 350 and 495 nm and did not change with photolysis time (Figure 2.10b) (discussed below). The chromatograms obtained from photolyzed solutions for eight hours with various acetate buffer concentrations are shown in Figure 2.11. The extent of Ru(bpy)3 *' peak disappearance decreased with decreasing acetate buffer concentrations and after 8 hours of photolysis showed 50, 18,12% of Ru(bpy);'' decomposition in 2 M, 0.1 M, and 0.025M pH 5 acetate buffers, respectively. Table 2.2 compares the extent of photodecomposition in various media. The enhanced photodecomposition with increasing buffer anion concentration suggests that acetate is playing a role in the primary photoprocess, as shown in Scheme 2.3 (reaction 5, page 71). The same three peaks observed at 12.3±0.1. 13.3±0.1 and 14.1 ±0.1 min in 2 M acetate were also observed in lower buffer concentrations and the absorption spectra corresponding to the peak positions are shown in Figure 2.12. The spectra of the component eluting at 13.3 min at all buffer concentrations and photolysis time intervals showed the same absorption spectra (Figure 2.12b) as that of the cfj-Ru(bpy)i( 0 Hi)2 -' and hence it can be concluded to be due to c/.y-Ru(bpy^^(OHi);' . It has been reported that deprotonation of Ru(bpy);(H; 0 )2 ' ’ to Ru(bpy)2 (H;0 )(0 H)* required a pH > 12 [25], while Meyer and co-workers have reported a pK, of 10.81 for a similar complex, Ru(bpy)i(Py)(H 2 0 )‘* [38]. Under the mildly acidic conditions employed here (pH 5), such deprotonation of the aquo ligand would not be possible. The appearance of new components during the photolysis of Ru(bpy);-' in acetate buffers as compared to that in distilled water could then arise from the acetate 77 a ) 6 Before 5 0.025 M After 0.1 M After I 4 2 M After I 3 2 1 I 0 10 12 14 16 18 20 Time (min) b) 10 a 8 e 0 6 4 2 0 1 < 2 12 13 14 15 Time (min) Figure 2.11 a) Chromatograms of the photolysis reaction of Ru(bpy)j^^ (2 x 10"* M), before and after 8 hours of photolysis in pH 5 acetate buffer of various concentrations, b) Same as (a) with the region of interest expanded. 78 0.03 0.02 0.01 - MO «00 MO S 0 0.00 - 0.03 - Ii 0 0 2 . ^ 0.01 - i 0 0.08 - 0.04 - 300MO S 0 0.00 4 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.12 Absorption spectra of the peaks observed in the chromatograms of 8 hours photolyzed Ru(bpy).^* solution (2 x 10~* M) inI inpH pH 5 acetate5 acetate buffer buffer of of various various concentration. concentn 0.025 M (-© -), 0.1 M( ), 2 M ( - A - ) . a) 12.3 min peak, b) 13.3 min peak, and c) 14.1 min peaks. 79 Medium Buffer pH Fraction Concentration Loss" (M) Acetate 2 . 0 4.0 0.60 Acetate 2 . 0 5.0 0.50 Acetate 0 . 1 0 5.0 0.18 Acetate 0.025 5.0 0 . 1 2 Phthalate 0.025 5.0 0.42 Phosphate 0.025 5.0 0.08 Phosphate 2 . 0 5.0 0.13 Water — — 0.06 Zeolite — — 0.06 Defined as loss of Ru(bpy) 3 ** upon photolysis relative to the original sample. Table 2.2 Rufbpy),'* decomposition after eight hours of photolysis under various conditions. 80 anions acting as a ligand. While the spectra of the component eluting at 14.1 min remained the same except for the change in intensities at different buffer concentrations (Figure 2.12c), the abundance increased with increasing acetate buffer concentrations, suggesting tliat this component contain acetate as a ligand. The shift of the lowest energy band from 486 nm for c«-Ru(bpy),(OH 2 )2‘" to 495 nm for the component eluting at 14.1 min can arise from replacing one of the HjO ligands by acetate, consistent with the better n donating character of the carboxylate ligand. In the chromatographic analysis of the reduction products of polypyridyl complexes containing Ru'^=0, it has been suggested that thermal substitution of aquo ligands in ruthenium complexes by acetate anion is possible [15]. If acetate participates only via thermal substitution of Ru(bpy),(OH2)2‘'. then the extent of RuCbpy),-' decomposition should be independent of acetate concentration and only the relative proportion of Ru(bpy) 2 (OH2)2 ‘* should decrease. However, the observed photodecomposition appears to be directly influenced by the acetate concentration. If the Ru(bpy)2 (CH 3 COO)(OH2 )* complex arises from the thermal replacement of Ru(bpy)2 (0 H2 )2 ' \ then the peaks due to cis- and /rawj-Ru(bpy) 2 (OH2 )2"‘ in the case of acetate buffer must be less than that of the same in distilled water. However, the observed decomposition patterns even at the highest acetate concentration, where thermal substitution if at all possible would be the most, indicate the presence of higher cis- and /ram-Ru(bpy)2 (OH 2 )2 '' products. This indicates that Ru(bpy) 2 (CH3 COO)(OH2 )' must be a primary photoproduct. This argument is also supported by the reported identification of Ru(bpy)2 (CH 3 COO)* among the unseparated photodecomposition products of Rufbpy),-' 81 in the presence of acetic acid by electrospray mass spectrometry (ES-MS), suggesting photosubstitution of bpy ligand by acetate [39]. If the origin of this component is photochemical as argued above, then it would be more likely that it is a cis- isomer (A^) as shown in Scheme 2.3 (page 71), formed via reactions 5 , 6 , and 7. If this new component is photoactive and capable of undergoing processes like cis-trans isomerization etc., then it would be accompanied by other species as well. Also, if this component is a trans- isomer (a secondary photoproduct), then it ought to be accompanied by the presence of the corresponding cis- isomer at same time. However, we could not see any other chromatographically observable peaks even after 8 hours of photolysis in low buffer concentrations as well as during the early photolysis period in high buffer concentration. The photochemical origin, absence of cis-trans isomerization, as well as the absorption spectra suggest that this component (A^) is cis- Ru(bpy)2 (CH 3COO)(OH 2)‘. It should be noted that if a diacetate complex is formed it would be neutral and hence insoluble in water and we could not detect any insoluble material at the low concentrations of Rulbpy);-' employed. The behavior of the component eluting at 12.3 min after 8 hours of photolysis was not as simple as that of the other two components. At low acetate concentrations (0.025M) the absorption spectrum was similar to that of the rra« 5 -Ru(bpy)2 (OH 3)2 ** (Figure 2.12a) with the peak position of the lowest energy absorption at 500 nm. At 0.1 M acetate, the peak slightly shifted to higher wavelengths (-505 nm) and at 2 M acetate the peak position had shifted to -520 nm. It is also not clear whether this component is a primary photoproduct or a secondary photoproduct or even a secondary thermal 82 substitution product. The fact that it appears only later during photolysis suggest that it could be a thermal product, however it could also be a secondary photoproduct with a very small quantum yield. If the photolysis of Ru(bpy]^'' in 2 M acetate buffer was stopped after the first two hours, when the first peak showed primarily the spectra corresponding to that of rra/ 7j-Ru(bpy)2(OH,)2 ‘" and the photolysed solution analyzed after standing overnight, the spectra of the 12.3 min eluting component completely converted to that of the new one while the spectra corresponding to that of the other peaks remained the same. This suggests the existence of a thermal route and the conversion of the /ram-RuCbpyliCOHi),'' to the new component, Aj. as shown in Scheme 2.3 (page 71) reaction 9. The concentration dépendance of this component and the shift of the lowest energy band to higher wavelengths and thermal conversion from rran 5 -Ru(bpy),(OH2 )2 ‘’. suggest that this new component (Aj) could be rrans-Ru(bpy) 2 (CH 3C 0 0 )(0 Hi) . The proposed (A^) rraw^-RuCbpyliCCHjCOOlCOH,)' showed a 520 nm band red shifted from 500 nm observed for the proposed (A^) cif-Ru(bpy) 2(CH;C 0 0 )(0 H2 )' internally consistent with the previous explanation for the lowest energy MLCT band of cis-trans isomers [24]. In order to confirm the assignments of the photoproducts and to eliminate the commonly suspected photoproducts containing chloride ligands, photolysis of Ru(bpy)/' in pH 5. 2 M acetate buffer was carried out in the presence of 0.5 M NaCl. Figure 2.13 shows the chromatogram obtained after photolysis in the presence of NaCl, which has two new peaks at 1 1 . 1 and 1 2 . 8 min (Cc, and 0 ^, marked by arrows) along with the previously observed peaks. Comparison of the chromatograms monitored at 450 nm with 83 a ) 7 Before o After G 5 e 0 4 (§) 3 2 1 1 < 0 1 0 1 2 14 16 18 20 Time (min) b) 6 O 5 E 4 e o 3 1 I 0 < -1 11 12 13 14 15 Time (min) Figure 2.13 a) Chromatograms of the photolyzed Ru(bpy)^^* solution (2 x 10"* M) in pH 5 acetate buffer (2 M) in the presence of added NaCl (0.5 M). b) Same as (a) with the region of interest expanded. 84 a ) 3 282 nm 450 nm •o s I 2 e 8 (S I I 1 < 0 0 I 1 0 1 2 14 16 18 20 Time (min) b) 5 s 4 - 4 e I 2 1 - 0 XiI 1 < < 0 1 1 12 13 14 15 Time (min) Figure 2.14 Comparison of chromatograms monitored at 450 nm and 282 nm of the photolyzed Ru(bpy)g^* solution (2 x 10"* M) in pH 5 acetate buffer (2 M) in the presence of added NaCl (0.5 M). b) Same as (a) with the region of interest expanded. 85 a ) 15 -, 1 2 - m ‘o 9 - i JO 6 - : < 3 - 0 - 300 350 400 450 550 Wavelength (nm) 8 b) 6 4 490 900 5 0 2 0 250 300 350 400 450 500 550 Wavelength (nm) Figure 2. IS Absorption spectra of the peaks observed in the chromatograms of the photolyzed Ru(bpy)g^^ solution (2 x 10^ M) in pH 5 acetate buffer (2 M) in the presence of added NaCl (0.5 M). a) II.2 min peak (C^.), 12.6 min peak (A.^) 12.8 min peak (C^.); b) 13.3 min peak (B), 13.4 min peak (C), 14.1 min peak (A^,). 86 that of the 282 nm (Figure 2.14), showed the presence of an unresolved new component (B) at about 13.3 min that was noticeable only in the 282 nm chromatogram. To avoid interference from the closely eluting cis bisaquo complex, the absorption spectra of the 13.3 min peak was monitored at a time when the 450 nm chromatogram showed low absorbance. The absorption spectra of this new component (Figure 2.15b) showed peaks in the UV region (232 and 281 nm). The absorption spectra of all the components are shown in Figtwe 2.15. The two components found in Figure 2.13. Q and C^. could be cis- and /raw 5 -Ru(bpy)2 (Cl)(OH 2 )'. respectively. Though the experimental evidence is inconclusive, the first peak at 1 1 .1 min can be compared to the small peak previously observed in the aquation studies of c/ 5 -Ru(bpy)2(Cl)2 that we assigned to cis- Ru(bpy);(Cl)(0 H2)'. The spectrum of the 13.3 min peak component resembled that of •free bipyridine' that would be a natural byproduct of photodeligation of Ru(bpy)^-'. In order to confirm this, chromatographic analysis of a bipyridine solution was carried out. and Figure 2.16 shows the results. The chromatogram showed a prominent peak at about 13.2 min (Figure 2.16a) and the spectrum showed peaks at 234 and 282 nm (Figure 2.16b). This confirmed the detection of free bpy' in the photolyzed solution. Unlike this instance the 'free bpy’ peak often completely overlapped with that of the c/s-bis aquo complex and hence the direct identification of free bpy was not possible. However, the presence of free bpy' in those instances was inferred from the anomalously high relative absorbance of the unresolved peak monitored at 282 nm compared to that of the same monitored at 450 nm. 87 a ) 0.4 I « 3 - S ? 0 .2 - I S 0 1 - < 0.0 10 12 13 14 1511 Time (min) b) 0.6 - Î 0.4 I ^ 0.2 0.0 200 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.16 a) Chromatogram of 2,2 -bipyndine (1 x 10 ^ M). b) Absorption spectrum of the 13.2 min peak in the chromatogram. 88 Phosphate buffer (pH 5) Chromatograms of 2 x 10"* M Ru(bpy) 3'" solution in 0.025 M. pH 5 phosphate before and after 8 hours of photolysis are shown in Figure 2.17a. Two photoproducts were observed at about 12.5 and 13.4 min (T. and C) similar to that of photolysis in distilled water and the extent of Ru(bpy) 3'* decomposition was about 8 %. The absorption spectra of the peaks are shown in Figure 2.17b and correspond to trans- and cis- Ru(bpy)2(0 H2)2' \ Similar results were obtained when 2 * 10"* M Ru(bpy) 3** solution in 2 M, pH 5 phosphate was photolyzed for 8 hours and the extent of decomposition was estimated to be about 13% and the modest increase from 8 % in 0.025M phosphate can be attributed to ionic strength [5a]. It can be concluded that phosphate ions play a negligible role in enhancing the photodecomposition due to their poor coordination ability and the observed decomposition can be explained by simple photoaquation. The relative peak areas of the photoproducts, cis- and tra/w-Ru(bpy) 2(OH2)2‘' appeared to have reversed, with the detection of relatively more cis- isomer when photolysis was carried out in 2 M phosphate. Though the effect of ionic strength in the cis-trans isomerization of Ru(bpy)2(OH 2)2‘* is not known, in the related photoracemization of optically active Ru(bpy)3-' in neutral aqueous solution, the quantum yield has been reported to increase with ionic strength and at this point we do not have an explanation for the observed photoproduct composition. However it is worth comparing photolysis of Rufbpy);-' in 2 M phosphate with that in 2 M acetate to notice the coordinating ability of acetate and to eliminate effects due to ionic strength. 89 a ) 5 Before After 4 S c 0 3 2 1 1 < 0 10 1 2 14 16 18 20 Time (min) 0.020 b) 0.015 - I 0.010 - s 3 0 0 4 0 0 4 8 0 9 0 0 9 0 < 0.005 - 0.000 - 200 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.17 Photolysis of Ru(bpy)j^^ (2 x 10"* M) in pH 5 phosphate buffer (0.025 M) (a) Chromatographic separation of photoproducts (b) Absorption spectra of the peaks observed in the chromatograms, 12.5 min peak (T), 13.4 min peak (C). 90 Phthalate buffer (pH 5) Photolysis of 2 x KT* M Ru(bpy) 3'* solution in 0.025 M phthalate (KHP/NaOH) buffer for 8 hours, showed appreciable solution absorption changes (Figure 2.18a) similar to that observed in 2 M acetate. The chromatogram showed two peaks. the first one at about 12.6 min (T) and a broad peak at about 13.1 miq (Figure 2.18b, Figure 2.19a). The absorption spectrum of the 12.3 min eluting component was similar to that of trans- Ru(bpy)i(OH3)2-" (Figure 2.19b). Cross-sections of the contour plot at various positions along the second peak at 13.1 min showed that it is in fact a mixture of two components. The component eluting at the tail end of the broad peak (C) was similar to that of the previously observed c/ 5 -Ru(bpy)2(OH 2);‘" (Figure 2.19b). The absorption spectrum of the new component that elutes at the beginning of the broad peak (P) was distinct from that of the cw-bis aquo complex, and had absorption bands at 294. 358 and 511 nm (Figure 2.19b). For ease of comparison, the absorption spectra of all three components were normalized to the 290 nm band and are shown in Figure 2.19b. The new component (P) can be assigned to cw-Ru(bpy)i(T]'-phthalate)( 0 H2)' similar to the acetate case, though there is no conclusive evidence to prove it. It should be noted that a bidentate phthalate in CIS- form or two monodentate KHP molecules bound either in cis- or in trans- form would make the complex neutral and precipitate out of the aqueous photolysis solutions. The extent of Ru(bpy) 3** decomposition in the KHP/NaOH buffer system as measured from the decrease in peak area was 42%. Comparison of this high degree of photodecomposition to that of 8 and 1 2 % in phosphate and acetate buffer, respectively, of similar pH and buffer concentration shows the greater coordinating ability of phthalate 91 a ) 4 Before After 3 2 1 Si < 0 200 300 400 500 600 700 800 900 Wavelength (nm) 5 b) Before After 4 E c o 3 I I 0 < 1 0 12 14 16 18 20 Time (min) Figure 2.18 Photolysis of Ru(bpy)]^* (2 x 10^ M) in pH 5 phthalate buffer (0.025 M) a) Solution absorption spectra b) Chromatographic separation of the photoproducts. 92 a ) 7 Before 6 After I 5 4 I 3 2 1 1 0 < -1 12 13 14 15 Time (min) b) « 0.25 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.19 Photolysis of Ru(bpy)j^* (2 x 10^ M) in pH 5 phthalate buffer (0.025 M) a) Chromatographic separation of the photoproducts b) Absorption spectra (intensities normalized to the -290 band) of the chromatographic peaks. 93 Peak Retention (nm) Species Time (min) R 16.1 244, 254 (sh). 286,450 Ru(bpy) 3-' C 13.4 243,291,340,486 cw-Ru(bpy)2(OH 3)f’* T 12.3 245,295,353,501 rraws-Ru(bpy)2(OH 2)2'' M 1 2 . 6 245,293, 348, 500 cw-Ru(bpy)2 (MeOH)2 '' Cc 1 1 .1 246. 295. 356,512 m-Ru(bpy)2(OH 2)C r Cr 1 2 . 8 296,363, 508 rmm-Ru(bpy)2 (Cl)(0 H2 )' (?) Ac 14.1 243,292, 350,495 c/5-Ru(bpy)2(CH3COO)(OH2)' At 12.3 296, 360,519 rra«5-Ru(bpy)2(CH3COO)(OH2)' P 13.3 -243.294,358,511 m-Ru(bpy)2 (C<,H,COOHCOO)(OH2 )' B 13.3 234,281 2 ,2 '-bipyridine M 13.4 256 MV-- Table 2.3 Retention time and spectral properties of the chromatographic peaks and proposed structural assignment. 94 and highlights the importance of careful of choice of buffer anion. The retention times and spectral properties of all the identified chromatographic peaks are shown in Table Effect o f pH The effect of pH on the photodecomposition of Ru(bpy);-' was studied by the photolysis of 2 x 10"* M Ru(bpy);-' solution in 2 M pH 4 acetate buffer and comparison to that of the same in pH 5 acetate buffer. The chromatogram of eight hour photolyzed sample showed the presence of three photoproducts (T. C. previously assigned to tram-, cis- bis aquo and c/ 5 -acetatoaquo, though the decomposition of Rufbpylj'* was enhanced to 60% (Figure 2.20a). The absorption spectra of the components are shown in Figure 2.20b. The enhanced decomposition can be explained by the higher extent of protonation of the monodentate bpy ligand in the intermediate, preventing rechelation, and thereby enhancing the observed photodecomposition. The relatively lesser extent of the acetate containing complex may be due to the reduced ionized acetate content at this lower pH (pH 5,2 M acetate = 1.4 M CHjCOO' + 0.6 M CHjCOOH; pH 4.2 M acetate = 0.4 M CHiCOO + 1. 6 M CHjCOOH). It should be noted that the low concentration of Ru(bpy)/' (2 X 10“*) would not alter the composition of the buffer media appreciably. Photolysis of 2 x 10“* M Rufbpy)/' solution in 0.025 M pH 3 KHP/HCl buffer brought about drastic changes in the solution absorption spectra, and at the end of eight hours of photolysis the solution was more or less colorless (Figure 2.21a). The chromatogram of four hour photolysed (Figure 2.21b) sample showed a small peak at about 1 1 min (Cc) at about the same that the same time cû-Ru(bpy) 2 (OH2)C r eluted in 95 a ) ------Before 0 —«— After C 4 B 1 : 3 ® 2 y Ac a \ 1 - c |\ 1 — < 0 I N V _ 10 1 2 14 16 18 20 Time (min) 0.15 b) 0.12 - 0.09 - 430 J 0.06 - < 0.03 0.00 -L 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.20 Photolysis of Ru(bpy)g^* (2 x 10^ M) in pH 4 acetate buffer (2 M) (a) Chromatographic separation of photoproducts (b) Abso^tion spectra of the peaks observed in the chromatograms, 12.5 min peak (T), 13.3 min peak (C), 14.1 min peak (A^,). 96 4 a) Before 4 hour 8 hour 3 2 ISi < 1 0 200 300 400 500 600 700 800 900 Wavelength (nm) b) 5 Before 4 hour 2 hour E 4 e 0 3 2 1 1 0 10 12 14 16 18 20 Time (min) Figure 2.21 Photolysis of Ru(bpy)g^* (2 x 10^ M) in pH 3 phthalate buffer (0.025 M) a) Solution absorption spectra (b) Chromatographic separation of photoproducts. 97 c r containing media and as the buffer was made up with HCi such a product may be possible. Peaks (T, P, C) at about the retention times previously assigned trans- Ru(bpy)2 (OH 2)2 ’", c/s-Ru(bpy)2 (Ti‘-phthalate)(OH 2 )" and cM-Ru(bpy) 2 (0 H2)i-' were also observed and the absorption spectra were comparable. In addition to these four components, an unresolved component also eluted at about 12.5 min (N) and the identity of this component could not be established from the present data. Upon further photolysis all the peaks were found to decrease in intensity greatly as indicated by the bleaching of the dye; after 8 hours > 95% of the starting Rufbpy),’* was found to have been photodecomposed. From the chromatographic analysis it can be concluded that for a given pH and concentration of buffer, decomposition follows the trend, phosphate < acetate « phthalate. Lowering the buffer pH to 4 and 3 enhances the decomposition of Rufbpy)]"' up to 60% and > 95%, respectively and is consistent with what is known in the literature [5a]. Photolysis o f Zeolite encapsulated Ru(bpy)/* Zeolite encapsulation affords the opportunity to examine the photodecomposition of Rufbpy),’* in a constrained medium. Moreover, since the charge neutralization of the Rufbpy)]-' occurs by the charge on the zeolite framework, there are no coordinating anions present. Use of rigid media like zeolite crystals normally would preclude the use of solution medium for any meaningful photochemical studies, often requiring a pressed pellet wherein only the surface could be sampled. By silylating the surface of the nanocrystalline zeolite-Y crystals containing Rufbpy),’* we were able to render these 98 3 D e 2 s S i I 1 tS 0 200 300 400 SOO 600 700 800 Wavelength (nm) Figure 2.22 Diffuse reflectance spectra of Ru(bpy)g^* encapsulated in nanocrystalline zeolite Y, after silylation (loading level 1 Ru(bpy)^^* per 15 supercages). 99 crystals highly hydrophobic. Figure 2.22 shows the diffuse reflectance spectra of Ru(bpy)3 ‘* encapsulated in nanocrystalline zeolite Y with a loading level of 1/15 after silylation. These hydrophobic crystals could be suspended in organic media like toluene and it was found that at least 80% of the encapsulated complexes could be spectroscopically sampled [20]. It is important to note that though the medium is organic, the immediate vicinity of Ru(bpy)3"* inside the supercages contains plenty of water molecules. Employing these relatively clear nanocrystallite suspensions allowed us to compare the photodecomposition of Ru(bpy) 3‘* in aqueous solution directly with that in the zeolite. Photolysis of the toluene suspension followed by the extraction of the encapsulated complexes into aqueous solution showed the same two peaks (Figure 2.23) as with aqueous Ru(bpy)3‘*, and their absorption spectra are shown in Figure 2.24. The after photolysis sample was compared to the Ru(bpy);-' extracted from identical mass of Rulbpy),'* -nano Y. without any photolysis. The before photolysis sample also showed a small broad peak with a retention time similar to the free bipyridine and the c/.y-bis aquo complex when monitored at 282 nm. However, the same sample did not show any peaks in the 450 nm chromatogram and hence it could be from the excess bipyridine employed during the intrazeolitic synthesis of RuCbpy),’' that was incompletely removed by Soxhlet extraction. The extent of decomposition compared to the before photolysis sample, taking into account the dilution factor due to the extraction procedure was of the order of 6 %. comparable to that in distilled water. 100 a) 12 Before 10 After I 8 0 6 4 2 0 1 2 10 12 14 16 18 20 Time (min) 1.0 2.0 b) CS After 282 nm 0.8 Before 282 nm After 450 nm 0.6 £ Before 450 nm c 0 8 0.4 (N 0.5 ( § ) 0.2 0.0 0.0 s -0.2 - -0.5 < -0.4 1< 12 13 14 15 Time (min) 2 + Figure 2.23 a) Chromatographic separation of the citric acid extract of Rufbpy), encapsulated in nanocystalline zeolite Y, before and after 8 hours of photolysis in toluene b) Same as (a) with the region of interest expanded and the chromatograms monitored at 450 nm and 282 nm are compared. 101 a) 0.004 - < 0.002 - 0.000 - 250 300 350 400 450 500 550 Wavelength (nm) 0.05 - b) 0.04 - 5 0.03 - jq 0.02 - 0.01 - 0.00 - 250 300 350 400 450 500 550 Wavelength (nm) Figure 2.24 Absorption spectra of the peaks observed in the chromatograms of the citric acid extract of 8 hours photolyzed Ru(bpy)g"^ encapsulated in zeolite Y. a) 12.4 min peak (T), 13.1 min peak (B), 13.3 min peak (T), and b) 16.1 min peak (R). 102 It can be expected that zeolite encapsulation of Rulbpy),"* should promote stability for the following reasons: 1 ) the zeolite encapsulation is shown to destabilize the Md level of ruthenium(ll) polypyridyl complexes [9,40], 2) the absence of photodechelation promoting, small, coordinating anions inside the zeolite and 3) the cage structure is expected to help in ligand rebinding [11]. The charge neutralization of Ru(bpy)3 ** in the zeolite is by the framework aluminate ions which have no mobility and considering that the Ru is at the center of the -13 A supercage, the aluminates cannot complex the ruthenium. Thus, bpy ligand replacement can only be mediated by intrazeolitic H^O molecules. The comparable photochemical decomposition between Ru(bpy)3 ‘* in distilled water and in zeolite suggests that the proposed destabilization of the Md level in the zeolite as well as the rigid cage that should promote the annealing of the intermediate monodentate bpy ligand do not appear to be diminishing the photodecomposition. This may have to do with the nature of intrazeolitic water. In the zeolite, there are 2-4 H^O molecules wedged between the bpy ligands and they are well positioned to promote the ligand replacement reaction. The analogy of the intrazeolitic environment can be made with accelerated photodecomposition of Rufbpy);"' in organic solvents where small molecules and ions such as H^O and Cl are wedged between the bpy ligands and available to bond to Ru once dechelation of bpy occurs [5b, 15]. It has also been observed that [Ru(bpy) 2 (/-Bu(py)),](PF6 ) 3 is photoactive in CHiCK even in the absence of added coordinating anions, and Ru(bpy)i(r-Bu(py))( 0 H2 )-" was found to be the major photoproduct [34]. The concentration of water in reagent grade CH 2 CI2 under the conditions of this study was found to be 10-fold or greater compared to the concentration 103 of the complex employed and was found to be responsible for the observed photochemistry [34]. It can be concluded that under zeolite encapsulation, the enhanced stability of Ru(bpy)j'"* due to the destabilization of Md state and cage effect is being negated by the close proximity and packing of H,0 molecules around the Ru center. Photolysis of Ru(bpy)f* in the presence of a quencher Since the early 1970s, one of the most extensively studied properties of Ru(bpy)]-' is the ability of its triplet MLCT excited state to photosensitize various quencher molecules. Fujita and Kobayashi were the first to report on the possibility of photosensitization with *Ru(bpy) 3 ‘" by observing the low temperature (77 K) energy- transfer sensitized emission o f [Crfox))]^' on charge-transfer excitation of Rufbpy);"' [41]. Adamson and co-workers showed that *Ru(bpy) 3'' can also sensitize substitution reactions of substrates like PtClj’" in room temperature solutions by the previously observed energy-transfer mechanism with a concomitant quenching of *Ru(bpy Ij'* emission [42]. The emission intensity and lifetime quenching of ^Rulbpy)-’' as a function of the substrate concentration (Stem-Volmer plot) was found to be adequately described by the following equations. 7 " ' (2.4, where. Iq and I are the emission intensity in the absence and in the presence of the quencher, respectively and ig and t are the luminescence lifetime in the absence and in the 104 presence of the quencher, respectively. In the above expressions, K^, the Stem-Volmer quenching constant, is the ratio of the bimolecular, second order quenching constant (k^) and the first order rate constant (k„) of decay of *Ru(bpyX-' in the absence o f quencher (K,^ = kq / ko), and [Q] is the concentration of the substrate quencher. The decay constant ko can be calculated from the lifetime data (ky = 1/Tq). The Stem-Volmer quenching constant. K,^, for the quenching of *Ru(bpy)^-' with PtCl/' was found to be 4.42 x 10^ M' ' [42]. Subsequently, Adamson and co-workers showed that in addition to energy- transfer, Ru^bpy);-' can also sensitize electron-transfer reduction of suitable substrates like [Co(NH3 );Br](NO ) ) 2 with a Stem-Volmer quenching constant (K,„) of -250 M'' [43]. These authors have also noticed that the emission from the excited ^CT was partially quenched by dissolved oxygen (K,^ = 2130 M ‘), with air saturated solutions exhibiting two-thirds of the emission intensity of those deaerated by Nn bubbling [43]. Meyer. Whitten and co-workers also observed quenching of *Ru(bpy)]-' with organic molecules like 1.1 -dimethyl-4,4'-bipyridine dications (MV-*) in acetonitrile solutions (K„ = 2120 M ‘) [44]. These authors have argued that since the triplet energy of MV’* (71.5 kcal/mol) is too high for efficient quenching of *Ru(bpy) 3’* = 49 kcal/mol). the quenching process must occur by electron-transfer mechanism (Equation 2.5) as confirmed by the transient absorbance of MV* at 603 and 393 nm [44]. *Ru(bpy) 3 '* + MV-* - Ru(bpy) 3^* + MV*' (2.5) This quenching reaction was found to be diffusion controlled with a second order rate constant (k^) of 2.4 x 10’ M*' s*‘ [44,45]. 105 The importance of the oxidative quenching of *Ru(bpy) 3 ' ’ with bipyridinium ions for the photochemical hydrogen production was soon realized by several groups as the redox potential of MV-'/MV'* (-0.44 V) in neutral aqueous solution is close to that of the water reduction (Eq' = -0.42 V). Moradpour and co-workers studied this reaction in aqueous buffered solutions and found that (5.5 x 10* M'' s ') was invariant at pH 3 and pH 6 [46]. Gratzel and co-workers observed similar k^ (5 x 10* M ' s ') and found that the first order rate constant (k„) of decay of *Ru(bpy))-' in the absence MV-' was 1 . 6 x 10* s ' yielding a Stem-Volmer quenching constant of -310 (K,^ = k,, / k^) [47]. With these rate constants, Gratzel and co-workers have reported a quenching efficiency (q^) of 0.76 with [Ru(bpy)3 - ] = 4 X 10-' M at [MV-'] = 1 x iQ'- M. + ^ 0 (2 .6 ) The quenching efficiency of 0.76 indicates that at this quencher concentration, about 24% of *Ru(bpy);-' is available for other processes like radiative and non-radiative decay as well as photochemistry. Sasse and co-workers have studied the oxidative quenching of *Ru(bpy) 3 "' with a series of bipyridinium ions in pH 5.0.2 M acetate buffer and have reported a of 615 M"' and calculated k^ (9.6 x 10* M'' s ') using a k^ of 1 . 6 x 10* s'' for methylviologen [48]. As discussed above if the origin of the photoanation of Ru(bpy) 3'' is the Md state that is thermally populated from the ^MLCT state, then efficient quenching of tlie latter should minimize the photodecomposition, so conditions leading to higher 106 values should lower the decomposition. As shown in the expression above, T]q can be evaluated from a knowledge of k^, k„ and the quencher concentration. Van Houten and Watts have noticed a modest decrease in the quantum yield of Ru(bpy)j*" decomposition when the solution was air saturated as compared to deaerated solutions [5a]. Contrarily, Hoggard and Porter have found the quantum yield of photodecomposition in the photolysis of Ru(bpy) 3'*(SCN ) 2 in dimethylformamide to increase with dissolved oxygen [ 6 b]. Jones and Cole-Hamilton have suggested that the suppression of photosubstitution of Ru(bpy)j‘* in acetonitrile by dissolved oxygen could be due to the quenching of the photochemically excited state [35]. Meyer and co-workers have shown the quantum yield for photochemical decomposition of Ru(bpy)]-' in dichloromethane to decrease in the presence of oxygen and have attributed it to the lowered efficiency for the ^CT to dd transition due to the efficient bimolecular quenching of ^CT by On [5b]. Porter and Sparks have studied the effect of electron transfer quenchers on the photoracemization of Rufbpy);"' in aqueous neutral solutions [32]. The quenchers employed were the oxidative quenchers. Cofphen);^^. Fe^* (aqueous), and Co(acac); as well as the reductive quencher, FefCN)^^ and it was found that none of the racemization quenching constants could be as high as the luminescence lifetime quenching, constants apparently due to the large ionic strength effects associated with the highly charged donor-acceptor pairs [32]. In the photochemical hydrogen production system involving Ru(bpy)^-‘ and Ti(IIl), 8-10% reduced Ru(bpy),*" decomposition has been reported in the presence of 1 x 10'- M Ti(IIi) that has been correlated to the fluorescence lifetime and intensity quenching 107 [49]. Since our primary interest in Ru(bpy)]-' stems from its potential use in the photochemical hydrogen production via the sensitized generation of MV"'. it would be prudent to study the effect of MV*" concentration on the photostability of Ru(bpy);-". Though the various reported values for the rate constant (k^) for *Ru(bpy)/' decay, in the absence of any quencher agree well, there is widespread discrepancies in the quenching constants (k^) reported for the oxidative quenching of *Ru(bpyh-' by MV-' (vide supra). Hence, we decided to evaluate the quenching constant, k, from the Stem-Volmer plot, under our experimental conditions for an accurate comparison of the quenching efficiency to the extent of decomposition. Emission intensity and lifetime studies were carried out under the conditions that showed the maximum photodecomposition, by employing a deaerated Ru(bpy),-' solution (2 X 10"* M) in pH 5 acetate buffer (2 M). Quenching studies were carried out in the presence of MV’* at concentrations up to 1 x 10 - M. Typical photochemical hydrogen evolution studies reported in the literature have employed MV- concentrations ranging from 3 X 1 0 '^ to 1 . 5 X 10^ M [10,37a]. Emission spectra of Rufbpy)/' were recorded in the presence of a given concentration of MV"', before and after lifetime studies. In order to eliminate artifacts due to the small day-to-day instrumental variations, for comparison purposes emission spectra of Ru(bpy)/' in the absence of MV-' were also recorded on the same day. Figure 2.25a shows the uncorrected emission spectra o f Rufbpy),-* in the absence and in the presence of MV-' (2 x 10‘^ M). As expected, quenching of *Ru(bpy);-' emission was observed in the presence of MV-'. The spectra recorded before 108 a) 6 0 M MV'*, befote lifetime 5 0 M MV''. after lifetime 2*10 ’ M MV**. before Ufetime 2'iO 'MMV\ after Ufetime 4 2 1 0 500 600 700 800 900 Wavelength (nm) b) ^ 1.0 - 0 M MV^* Data 0 M MV^* Fit 0 M MV^* Residuals Si 0.8 - 2*10'^MMV^*Data 2*I0'^MMV^*Fit 0.6 - 2* 10'^ MMV^* Residuals . 1 0.4 I 0.2 - Z 0 0 ^ 2 e- 6 4e-6 6 e - 6 8 e - 6 Time (s) Figure 2.25 a) Emission spectra (uncorrected) of deaerated Ru(bpy)^^* (2 x 10^ M) in pH 5 acetate buffer (2 M) in the absence and in the presence of (2 x 10 ^ M). b) Transient emission signal decay of Ru(bpy)g^^ observed at 610 nm after the 532 nm excitation, in the absence and in the presence of (2 x 10 ^ M). 109 and after lifetime measurement showed small variations, and the average intensity at 605 nm was employed for the Stem-Volmer plot. Figure 2.25b shows the transient emission signal decay of *Ru(bpy)^‘' observed at 610 nm after the 532 nm excitation, in the absence and in the presence of MV-' (2 x 10'^ M). Increasing concentrations of MV’* progressively shortened the emission lifetime. Figure 2.25b also shows the fitting of the transient signal to a single exponential decay along with the residuals. As expected for homogeneous solutions, the decay at all MV’* concentrations fitted well to a single exponential as observed from the residuals and the exponential term (1/t) increased with increasing MV’ concentrations. For each MV’* concentration employed the transient decay was recorded thrice and the average lifetime (±4 ns) is shown in Table 2.4. The emission lifetime of *Ru(bpy);’* (Tq) varied slightly more from day to day and was found to be 652±l 1 ns. Table 2.4 shows the emission intensity and lifetime quenching data for various MV’ concentrations. Figure 2.26 shows the Stem-Volmer plot of the emission intensity and lifetime quenching as a function of MV’* concentration (data from Table 2.4). The slope of the Stem-Volmer plot of the lifetime quenching (K,„ = 858 M*‘) and intensity quenching (K,„ = 777 M*') were slightly different due to the error of measurements. Using an average K,, of 820 M ' and an average kg of 1.53 x IQ® s ' (1/Tq) the calculated bimolecular. second order quenching constant (Iq, = K^*kg) was 1.26 x iQ’ M 's '. The quenching efficiency (t^,) calculated from these values using the above mentioned expression (equation 2 .6 ) for the highest studied MV’ concentration of 1 x 10*’ M was 0.89. no 10 8 6 4 2 G 0.000 0.002 0.004 0.006 0.008 0.010 Figure 2.26 Stem-Volmer plot of the emission intensity and lifetime quenching of deaerated Ru(bpy).^* (2 x 10^ M) by in pH 5 acetate buffer (2 M). Ill [MV:*], Intensity, lo/r Lifetime To/X^ (M) X 1 0 ^ ( 1 0 -' s) 0 5.10 1 . 0 0 652 1 . 0 0 4 X lO"* 4.08 1.23 472 1.37 2 X 1 0 ': 2.19 2.47 234 2.75 4 X 10': 1.04 4.06 147 4.45 6 X 1 0 ': 0.77 5.62 108 6 . 1 0 I X 10': 0.48 8.83 70 9.60 in the presence and absence olFMV: respectively. ’’ r, t lifetime in the presence and absence of MV**. respectively. Table 2.4 Emission intensity and lifetime quenching of Ru(bpy);'' by MV- in pH. 2 M acetate buffer. In order to correlate the calculated quenching efficiencies to the extent of photodecomposition, photolysis were carried out with deaerated solutions identical to those employed for the quenching studies. Solutions containing Ru(bpy)j-* (2 x 10"* M) and various concentrations of MV** up to 1 x 10'- M solution in pH 5 acetate buffer (2 M) were photolyzed for 8 hours. Photolysis of Ru(bpy);-' solution in the presence of I x IQ - M MV" showed no visually perceptible change and the spectra showed small decrease in the 450 nm region and small increase in the 500 nm shoulder that was obvious at a lower MV-' concentration of 2 x 10'^ M. Chromatographic analysis of the Ru(bpy),-* solution after photolysis in the presence of MV’* (Figure 2.27) showed the same three decomposition products as obtained in acetate buffer in the absence of MV’*. In order to 112 correlate the fluorescence quenching efficiency measured at the beginning of the photolysis to quenching efficiency of decomposition, the concentration of the quencher should remain the same during the period of hydrolysis. Figure 2.28 shows the chromatograms monitored at 256 nm during the photolysis of Ru(bpy)j‘" (2 x lO"* M) in the presence of MV'" (1 x 10 ' M) in 2 M pH 5 acetate buffer and after eight hours of photolysis the concentration of MV’" was found to remain the same within the experimental error range. In the absence of external electron donors like EDTA, MV' was found to be generally stable during photolysis with Ru(bpy)^**. Figure 2.29 shows the Rufbpylj'* peak area as a function of photolysis time under various MV'* concentrations. The extent of decrease in the Rufbpy),*' peak area after photolysis and the concomitant increase in the peak area of the decomposition products were found to be inversely proportional to the MV'* concentration. Table 2.5 shows the extent of decomposition of Ru(bpy)/' for various MV' concentrations and ranges from -50% in the absence of MV'* to 7% for the highest MV'* concentration employed (1 x iQ-- M). Figure 2.30 shows a plot of the percentage improvement in photodecomposition of Ru(bpy>3 '* along with the percentage quenching efficiency (r^* 100) calculated from the intensity and lifetime quenching data for various concentrations of MV'*. It is evident from Figure 2.30 and Table 2.5 that the magnitude and the trend in the photodecomposition of Rufbpy);'* in the presence of MV** is similar to that of the quenching efficiency. The observed close correlation between the photodecomposition of Rufbpy);'* and the quenching efficiency as a function of the MV'* concentration suggests that in reactions involving the photoexcited Rufbpy),'*. such as electron transfer 113 0.04 - Before 2 Hour 4 Hour 0.03 - 6 Hour 8 Hour 0.02 - s c 0.01 - 0 0.00 - 8 0.04 - 1Si 0.03 - < 0.02 - 0.01 - 0.00 - 14 16 Time (min) Figure 2.27 Chromatographic separation of the photolyzed solution of Ru(bpy) 2+ (2 X 10"* M) in pH S acetate buffer (2 M) a) in the presence of (2 x 10^ M) and b) in the presence of (1 x 10 ^ M). 114 Before 2 Hour 4 Hour 6 Hour 8 Hour I < 0.2 - 0.0 - 10 1214 16 18 20 Time (min) Figure 2.28 Chromatographic separation (monitored at 256 nm) of the photolyzed solution of Ru(bpy)g (2 X 10"* in pH 5 acetate buffer (2 M) a) in the presence o fM V,2 + " (l _X 10,n -2 ^ M). 115 A m = 0 .7 - 1 0.6 . ZI 0 .5 - 0.4 0 100 200 300 400 500 Time (min) Figure 2.29 Plot of Ru(bpy);"^ peak area decrease as a function of time during the photolysis of Rufbpy)^^^ (2 x 10^ M) in pH 5 acetate buffer (2 M) in the presence of various concentrations. 116 [MV-'] Fraction Decrease in Quenching (M) Loss' Photodecomposition'’ efficiency, (%) (%) 0 0.50 0 0 4 X 10“* 0.47 7 23 2 X 1 0 '^ 0.23 54 62 4 X 10'' 0 . 1 2 77 76 6 X 1 0 '' 0.07 85 83 1 X 10-- 0.07 8 6 89 ' Defined as loss of Ru(bpy);-' upon photolysis relative to the original sample. ’’ Defined as (D q-D) / Do, where Do is the fraction lost in the absence of MV-* and D is the fraction lost in the presence of given concentration of MV-' (pH 5.2 M acetate buffer). '■ = lOOx kJMV-']/(kJMV-*]-(-ko). L = 1.26 X 10’ M ' s ' and k„ = 1.53 x 10" s '. Table 2.5 Correlation of the decrease in Ru(bpy)/' photodecomposition and quenching efficiency by MV- (eight hours photolysis in pH 5. 2 M acetate buffer). 117 100 100 I 80 i II 60 I 40 20 Lifetime | . S Intensity 1 0 Decomposition 0.000 0.002 0.004 0.006 0.008 0.010 [MV^*] (M ) Figure 2 JO Correlation of the percentage improvement in photodecomposition of Ru(bpy),^* to the percentage quenching efficiency (q*100) calculated from the intensity and lifetime data, as a function of concentration. 118 quenching, the extent of Ru(bpy) 3** decomposition can be minimized by increasing the quencher. However, it is clear from Table 2.5, it takes a high concentration of MV- (1 >= 10 ’ M) to bring down the level of photodecomposition of RuCbpy),"' in 2 M acetate buffer to the order of 7%, comparable to that observed for photolysis in distilled water in the absence of MV’*. More typical photolysis conditions involve concentrations of MV’ an order of magnitude or two lower [10,37a], which would lead to significant photodecomposition of RuCbpy)]’* upon long term photolysis in buffers containing coordinating ligands. Hence it is preferable to employ a sufficiently high quencher concentration and avoid high buffer concentrations and buffers containing coordinating anions to ensure the long term stability of Ru(bpy) 3 ’\ Photostability of Ru(bpy)j^* In order to ensure that all the decomposition products that we are observing are indeed due to RuCbpy),’**, we examined the products formed upon photolysis of Ru(bpy)3 ^*. Ru(bpy) 3 ^* was generated in-situ by excited state electron transfer quenching of Ru(bpy)3 -'' by [Co(NH 3);Cl]’* in 2 M acetate buffer at pH 4.0. As in the case of quenching with MV’ , use of [Co(III)(NH 3 );Cl]’' as an oxidative quencher would eliminate the decomposition via Ru(bpy) 3 ’*' pathway. Unlike the case of MV’*. rapid back electron transfer from the reduced quencher [Co(H)(NH 3 );]’* to the oxidized sensitizer, Ru(bpy)3"*, is not possible as the Co(Il) is not hydrolytically stable and undergoes irreversible decomposition. In this context [Co(III)(NH 3 );Cl]’* acts as a sacrificial electron acceptor leading to the steady build-up of Ru(bpy)j'* with irradiation. Figure 2.3 la shows the chromatograms obtained for the before and after photolysis 119 samples. The peak marked as 'Co’ is due to that of [Co(NH 3 )sCl]'* and the peak due to Ru(bpy)3** (R) has almost disappeared after photolysis. Peak 1 is a [Co(NH 3 );Cl]'" decomposition product and the other major decomposition products (labeled as 2,3, and 4) were all characterized by strong MLCT charge transfer bands around 450 - 460 nm (Figure 2.31b). The use of [Co(T4 H 3 );Cl]-' in our studies as mentioned above led to the formation of Co(II) aquo species. Such Co(II) aquo species under alkaline conditions could undergo pH dependent hydrolysis to form Co(IV), which can act as a catalyst for the oxidation of water to dioxygen by Ru(bpy);^' thereby minimizing the decomposition due to Ru(bpy)3" ' as well as Ru(bpy)3 ’’. However under acidic conditions formation of such catalytic Co(lV) species is not possible and hence the Ru(bpy) 3^' decomposition cannot be prevented. It has been reported that Ru(III) complexes like RulbpyjtCL' and Ru(bpy)n(py)Cl-' are substitutionally inert in water and polar organic solvents [5b]. Also, it has been observed that Ru(bpy) 3"' underwent 15-20% destruction via ligand oxidation after 4 hours of irradiation when employed under oxygen evolving conditions involving [Co(IlI)(NH3 );Cl]-' and IrO, as a photocatalyst [50]. Sutin and co-workers have studied the thermal and light induced reduction of Ru(bpy)3 ^* in aqueous solution and found the process to involve nucleophilic addition of water or hydroxide to a bipyridine carbon with the associated decomposition of the complex [14]. About 0.1 mol of decomposition was observed for every mole of Ru(bpy)3 ^* that got reduced to Ru(bpy) 3 -\ under both thermal and photochemical conditions, leading to the formation of Ru(II) complex with modified bpy ligand (like 120 NOTE TO USERS Page(s) not included in the original manuscript and are unavailable from the author or university. The manuscript was microfilmed as received. 1 2 1 This reproduction is the best copy available. UMI Ru(bpy)i(bpyOH)^") and is thought to proceed via the intermediate, Ru(bpy)i(bpyHOH)', formed as shown below: OH Ru(bpy)3 " - [Ru'"(bpyMbpyHOH)]-' Ru(bpyMbpyHOH)-' (2.7) hv HiO Ru"'(bpy)3- -> [Ru“(bpy>,(bpy")]'' - Ru(bpy).(bpyHGH)-' (2.8) -H The hydroxyl group can be present at different positions of the same bpy ligand and/or on multiple bpy ligands leading to various slightly different modified products. Since the three major photoproducts observed in the chromatogram have all showed MLCT band similar to Rulbpy),^*, it can be concluded that the decomposition products were formed by modifications of the bpy ligand of Ru(bpy),''. Since none of the previously monitored decomposition product had MLCT band similar to that of the decomposition products observed here, the earlier decomposition must be independent of Ru(bpy) 3^' and as it has been shown to be quenched by MV-‘. it certainly goes through the Ru(bpy);-* state as had been postulated in the literature [5]. CONCLUSION Photo lability of Ru(bpy) 3-' in aqueous and heterogenous media was studied by reverse-phase ion-pair HPLC and separation of the photodecomposition products from the starting Ru(bpy) 3 ‘‘ was achieved. Comparison of these chromatographic profiles with that of aqueous Ru(bpy)2 CL showed that the common products are cis and trans- Ru(bpy)2 (0 Hi)2 -\ In the case of acetate and phthalate buffers, other species were found and based on the absorption spectra and concentration dépendance on the buffer anion. 1 2 2 were identified as cis~ and /ran^-Ru(bpy) 2 (L)(OH2 )", where L is acetate or phthalate. Additional photoproducts like cis- and rran.r-Ru(bpy)i(Cl)(OHi) ' were also formed when Ru(bpy)3 '' was photolysed in the presence of Cl' ions. It can be concluded that for a given pH and concentration of buffer, Ru(bpy)j'* decomposition follows the trend, phosphate < acetate « phthalate. Presence of the electron transfer quenching agent, N.N'-dimethyl-4,4'-bipyridinium ion in the medium led to a decrease of the photodecomposition, and closely followed the quenching efficiency as measured by intensity and lifetime quenching studies. But under typical conditions found in the literature involving MV-* concentrations of less than 1 x 10'- M, significant photodecomposition of Ru(bpy)--' will still occur if high concentrations of buffers containing coordinating anions are used. Photodecomposition pattern of the Ru(bpy) 3 '* encapsulated in nanocrystalline zeolite-Y crystals indicate that the enhanced stability of RuCbpy),-'* due to the destabilization of ^dd orbitals and cage effect is being negated by the close proximity and packing of H^O molecules, wedged between the bipyridine ligands around the Ru center. 123 REFERENCES 1. Gerardi, R. D.; Bamett, N. W.; Lewis, S. W. Anal. Chim. Acta 1999, 378, 1. 2. "Fiber Optic Chemical Sensors and Biosensors”, O. S. Wolfbeis, CRC Press, Boca Raton, 1991. 3. Amouyal, E. Sol. Energy Mater. Sol. Cells 1995, 38, 249. 4. a) Balzani, V.; Juris, A.; Venturi, M.; Campagna, S.; Serroni, S. Chem. Rev 1996. 96, 759 . b) Juris. A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser. P.; VonZelewsky, A. Coord. Chem. Rev. 1988, 84, 85. c) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159. d) Balzani, V.; Bolletta, F.; Gandolfi, M. T.; Maestri, M. Top. Curr. Chem. 1978, 75, 1 . 5. a) Van Houten, J.; Watts, R. J. Inorg Chem. 1 978, 17, 3381. b) Durham, B.; Caspar, J. V.; Nagle, J. K.; Meyer. T. J. J. Am. Chem. Soc. 1982, 104,4803. 6 . a) Gleria, M.; Minto. F.; Begglato, G.; Bortolus. P. J. Chem. Soc., Chem. Commun. 1978.285. b) Hoggard. P. E.; Porter, G. B. J. Am. Chem. Soc. 1978. 100, 1457. 7. Lumpkin, R. S.; Kober, E. M.; Worl, L. A.; Murtaza, Z.; Meyer. T. J.J. Phys. Chem. 1990, 94, 239. 8 . Masschelein. A.; Kirsch-De Mesmaeker, A.; Willsher, C. J.; Wilkinson, F. J. Chem. Soc. Faraday Trans. 1991. 87, 259. 9. a) Maruszewski. K.; Strommen. D. P.; Kincaid. J. R. J. Am. Chem. Soc. 1993,115, 8345. b) Maruszewski, K.; Kincaid, J. R.Inorg. Chem. 1995. 34, 2002. 10. Kleijn. J. M.; Rouwendal, E.; Van Leeuwen, H. P.; Lyklema, J.J. Photochem. PhotobioL A 1988. 4 429. , 11. Adelt, M.; Devenney. M.; Meyer, T. J.; Thompson. D. W.; Treadway. J. A. Inorg. C/iem. 1998.37.2616. 12. Tachiyashiki, S.; Ikezawa, H.; Mizumachi. K. Inorg. Chem. 1994, 3 3 , 623. 124 13. Valenty, S. J.; Behnken, P. E. Anal. Chem. 1978,50, 834. 14. Ghosh, P. K.; Brunschwig, B. S.; Chou, M.; Creutz, C.; Sutin, N. J. Am. Chem. Soc. 1984, 106,4772. 15. Roecker, L.; Kutner, W.; Gilbert, J. A.; Simmons, M.; Murray, R. W.; Meyer, T. J. Inorg. Chem. 1985,2 4 ,3784. 16. Jones, W. E., Jr.; Smith, R. A.; Abramo, M. T.; Williams, M. D.; Van Houten, J. Inorg. Chem. 1989,28, 2281. 17. Fairbank, R. W. P.; Wirth, M. J. Anal. Chem. 1997,69, 2258. 18. Buchanan, B. E.; McGovern, E.; Harkin, P.; Vos, J. G. Inorg. Chim. Acta 1988,154, 1 . 19. Ambs. W. J. Microcrystalline synthetic faujasite, U.S. Patent, US4372931, 1983. 20. Castagnola, N. B.; Dutta, P. K. J. Phys. Chem. B 1998.102, 1696. 21. Singh. R.; Dutta, P. K. Microporous Mesoporous Mater. 1999.32, 29. 22. Maruszewski, K.; Strommen, D. P.; Handrich, K.; Kincaid. J. R. Inorg. Chem. 1991. 30, 4579. 23. Wang. P.; Lee. H. K. J. Chromatogr., A 1997.789, 437. 24. Kroener. R.; Heeg, M. J.; Deutsch. E. Inorg. Chem. 1988.27, 558. 25. Moreira. I. S.; Santiago, M. O. Chromatographia 1996,43, 322. 26. a) Twitchett, P. J.; Moffat. A. C. J. Chromatogr. 1975, III, 149. b) Keller, P.; Moradpour, A.; Amouyal. E.; Kagan. H. B. Nouv. J. Chim. 1980. 4, 377. c) Johansen. O.; Launikonis, A.; Loder. J. W.; Mau, A. W.-H.: Sasse. W. H. P.; Swift, J. D.; Wells. D. Aust. J. Chem. 1981. 34, 981. d) Hodgeson, J. W.; Bashe, W. J.; Eichelberger, J. W. "Determination of diquat and paraquat in drinking water by liquid-solid extraction and high performance liquid chromatography with ultraviolet detection," U.S. Environmental Protection Agency, 1992. e) Sakai, K.; Kizaki, Y.; Tsubomura, T.; Matsumoto, K.J. Mol. Catal. 1993,79, 141. 125 27. O'Laughlin, J. W.; Hanson, R. S. Anal. Chem. 1980,5 2 ,2263. 28. Davies, N. R.; Mullins. T. L. Aust. J. Chem. 1967,20, 657. 29. Wallace, W. M.; Hoggard, P. E. Inorg. Chem. 1979,18, 2934. 30. Geraty, S. M.; Vos, J. 0 . J. Chem. Sac., Dalton Trans. 1987,3073. 31. Durham, B.; Wilson, S. R.; Hodgson, D. J.; Meyer, T. J. J. Am. Chem. Soc. 1980, 102, 600. 32. Porter. 0 . B.; Sparks, R. H. J. Photochem. 1980,15, 123. 33. Van Houten. J.; Watts, R. J. J. Am. Chem. Soc. 1976.98, 4853. 34. Durham, B.; Walsh. J. L.; Carter. C. L.; Meyer, T. J. Inorg. Chem. 1980,19, 860. 35. Jones, R. F.; Cole-Hamilton, D. J. Inorg. Chim. Acta 1981,53, L3. 36. Cohn. M.; GetofF. N. Z Naturforsch., A1979,54. 1135. 37. a) Amouyal. E.; Koffi, P. J. Photochem. 1985,2 9 ,227. b) Lehn, J. M.; Sauvage. J. P.; Ziessel, R. Nouv. J. Chim. 1981,5,291. c) Das, S. D.; Dutta, P. K. Microporous and Mesoporous Materials 1998.22, 475. 38. Moyer, B. A.; Meyer, T. J. J. Am. Chem. Soc. 1978.100, 3601. 39. Arakawa. R.; Lu, J.; Yoshimura, A.; Nozaki. K.; Ohno. T.; Doe. H.; Matsuo. T. Inorg. Chem. 1995, 54, 3874. 40. Bhuiyan, A. A.; Kincaid, J. R. Inorg. Chem. 1998,57, 2525. 41. Fujita, I.; Kobayashi, H. J. Chem. Phys. 1970,52,4904. 42. Demas. J. N.; Adamson, A. W. J. Amer. Chem. Soc. 1971, 95, 1800. 43. a) Gafhey. H. D.; Adamson, A. W. J. Amer. Chem. Soc. 1972. 94, 8238. b) Demas, J. N.; Adamson. A. W. J. Amer. Chem. Soc. 1973, 95, 5159. 44. Bock, C. R.; Meyer. T. J.; Whitten, D. G. J. Amer. Chem. Soc. 1974, 96, 4710. 45. Young, R. C.; Meyer, T. J.; Whitten, D. G. J. Am. Chem. Soc. 1976,98, 286. 126 46. Moradpour, A.; Amouyal, E.; Keller. P.; Kagan, H. Nouv. J. Chim. 1978, 2, 547. 47. Kalyanasundaram, K.; Kiwi, J.; Graetzel, M. Helv. Chim. Acta 1978, 61,2720. 48. Launikonis, A.; Loder, J. W.; Mau, A. W.-H.; Sasse, W. H. F.; Summers. L. A.; Wells, D. Aust. J. Chem. 1982, 35, 1341. 49. Giro, G.; Casalbore, G.; Di Marco, P. G. Chem. Phys. Lett. 1980, 71,476. 50. Werner, H. A. P.; Bauer. R. J. Mol. Catal. 1994, 88, 185. 127 CHAPTER3 CATALYST DEVELOPMENT FOR PHOTOCHEMICAL HYDROGEN PRODUCTION INTRODUCTION I. Photochemical hydrogen production: Direct conversion of solar energy into chemical energy to produce non-polluting fuels still remains a challenge at the dawn of the new century. Among the various interesting practical solar energy conversion processes, production of hydrogen from water, a clean-burning fuel from an abundant source has attracted considerable attention (Equation 3.1). As water does not absorb the visible light, photoproduction of hydrogen from water requires additional components that would absorb visible light, utilize the excitation energy to effect charge separation, and transfer the separated charges to water in a concerted, cyclic manner. There have been several strategies reported for the photo splitting of water into molecular hydrogen and oxygen. H ,0 ------> H, + V2O2 AG = 237.2 kJ mol ' E’ = 1.23 V (3.1 ) 1 2 8 One such popular multi-molecular system is the well-known sacrificial donor- sensitizer-acceptor model, EDTA-Ru(bpy) 3 -"-bip>Tidinium-catalyst scheme. The various reactions involved in this system that lead to the photoproduction of hydrogen can be summarized as below: hv Ru(bpy);-' *Ru(bpy) 3 -' (3.2) *Ru(bpy);-' kd Rulbpy);" (3.3) kf *Ru(bpy);-' + MV- Ru(bpy)3 "' + MV ' (3.4) Ru(bpy)j^' + MV ' kb RuCbpylj" + MV" (3.5) Ru(bpy)/' + E D I A kfcd Rulbpylj" + EDTA (3.6) 2 M V + 2H .0 2 MV" + H: + 20H (3.7) Cat EDTA„, Oxidized Products (3.8) EDTA'_ + MV" MV ' + H' + EDTA" (3.9) Visible light excitation of the sensitizer, Ru(bpy) 3 " , leads to the formation of metal to ligand charge-transfer (MLCT) excited state. *Ru(bpy) 3 -‘ (t = 600 ns, Equation 3.2). The sensitizer in the MLCT excited state is a very good reducing agent, thermodynamically capable of water reduction (E° 3 . .3 . = -0.87 V). However even in the presence of a catalyst, direct reduction of water to hydrogen is not known to occur, necessitating the use of an electron relay [1], A suitable electron relay should be capable of reduction by *Ru(bpy) 3 ‘* and the reduced relay should be energetic enough to reduce water. 129 Methylviologen (MA^-dimethyl-4,4'-bipyridine dication) is a commonly employed one electron redox indicator with a blue colored radical reduced form. Methylviologen has also been widely used as an electron relay as it meets the energetic requirement (E° 2 - .. = -0.44 V) for water reduction [2]. In the presence of sufficiently high concentrations of methylviologen, *Ru(bpy)/' undergoes rapid electron transfer to produce methylviologen radical (Equation 3.4, k, = 1 x 10^ M'‘s ') effectively eliminating the decay pathway (Equation 3.3, kj = 1 . 6 x 10* s '). But the high energy intermediates Ru(bpy);"' and MV*' undergo rapid energy wasting back electron transfer (Equation 3.5. kb = 2 X 10’ M 's ') requiring the use of a sacrificial electron donor to quench Ru(bpy) 3 ^*. Among the various sacrificial donors employed in the literature, Na^EDTA is very popular due to its relatively high reduction rate of Ru(bpy);,^' (Equation 3.6. ^ 10" M's') and solubility. Methylviologen radical in the presence of catalysts like colloidal platinum or metal oxides like IrOi is known to be capable of evolving hydrogen (Equation 3.7). The one electron reduction product of EDTA is a nitrogen radical cation and is unstable, rapidly undergoing decomposition (Equation 3.8) by proton elimination from the adjacent carbon, leading to the elimination of the back reaction [3]. The homogeneous multicomponent system described above suffers from various performance limiting drawbacks. The lifetime of the charge separated products is too low to be effectively utilized requiring the use of sacrificial electron donors to drive the reaction forward as mentioned above. Organization of the various components in a microheterogeneous media like zeolites has been shown to retard the back electron transfer prolonging the lifetime of the charge separated products. In homogeneous 130 systems, even in the presence of sacrificial donor, the efficiency of the hydrogen production process is limited by the diffiisional process by the end of which the methylviologen radical, MV*' is oxidized back to MV-* on the surface of a catalyst. The reactive methylviologen radical intermediate is susceptible to photodecomposition resulting in the irreversible loss of the electron relay. The photodecomposition of the electron relay arises from hydrogenation as well as from the poorly understood inherent photoinstability of the radical. The catalytic reduction of methylviologen under hydrogen evolving photo lytic conditions can be alleviated by employing selective catalysts like RuOi. efficient in hydrogen production but inefficient in hydrogenation. Inherent photodecomposition of the radical can be minimized by generating the radical close to the catalyst thereby minimizing the time required for the radical to encounter a catalytic site. This chapter reports our efforts toward an integrated photochemical hydrogen production system. The next two subsections review ruthenium dioxide and surface derivatization techniques and the last subsection details the scope and organization of the rest of the chapter. 11. Ruthenium dioxide The commonly employed catalysts for hydrogen evolution like colloidal platinum are also well known hydrogenation catalysts. In comparison, ruthenium dioxide has been reported to be poorer in its ability to hydrogenate the electron relay (methylviologen), while maintaining good hydrogen evolving ability [3]. Ruthenium dioxide (RuOi) has attracted a good deal of attention recently due to its technologically interesting applications in various fields. Traditionally RuOi has been employed as a heterogeneous 131 catalyst for various reactions including the hydrogenation of propanone and 2 -propen- 1 - ol and for the hydroformylation of propene [4]. Due to its low overpotential RuO; has been used in electrocatalysis as dimensionally stable electrodes for the production of chlorine as well as oxygen evolution [5]. RuOi is also being studied as an electrode material for oxide dielectric ultracapacitors in gigabit-scale dynamic random access memories [6 ], Importantly, RuO, either by itself [2,7] or supported on titanates [ 8 ] or niobates [9] has been extensively studied for photocatalytic water splitting. Structure: RuOn is a tetragonal metal oxide with two molecules per unit cell and has a rutile structure (space group PA-Jmnm, No. 136; a = 4.4911 ± 0.0008  and c = 3.1065 ± 0.0003 Â) [4]. It is made up of RuO^ comer sharing octahedra with two short Ru-0 bonds ( 1.942 Â) and four longer bonds (1.984 Â). The comer sharing octahedra form a chain along the c axis of the crystal with a neighboring Ru-Ru distance of 3.107 Â. Each such chain is also linked to four other neighboring chain through comer sharing 0 atoms with a longer neighboring Ru-Ru distance of 3.535 Â, resulting in a 3-d network. Preparation: RuOi is the stable oxide phase of Ru in an oxygen atmosphere below about 1850 K [ 1 0 ]. Single crystals of RuOi have been synthesized from polycrystalline RuOn and/or Ru metal powder by chemical vapor transport at temperatures greater than 1300 K in an oxygen stream [11]. Volatilization occurs by formation of higher oxides and these thermodynamically unstable oxides decompose upon cooling to form single crystals. When bulk or sintered Ru powder are exposed to air at room temperature to 473 K, a thin film (4-8 Â) of native oxide forms which 132 prevents further oxidation. Oxidation of bulk or powdered Ru at higher temperatures gives RuG,, but oxidation of Ru black can occur as low as 390 K [10]. Thermal decomposition of RuClj.nHiO in air yields a polycrystalline mixture of RuOi and ruthenium oxychlorides that are oxygen deficient [10]. When annealed in an oxygen atmosphere at high temperature, the oxygen content increases and the excess oxygen exists as RuGj probably as part of a gross defect structure. Precipitation from aqueous solution of a ruthenium salt often results in hydrated RuG? and requires annealing at high temperatures. Thermal decomposition of various volatile organometallic compounds of Ru has also been employed for the preparation of polycrystalline RuG? [12]. RuGi films have been prepared by various techniques like epitaxy [13], pulsed laser deposition [ 6 ], chemical vapor deposition, spray pyrolysis, and sputtering [14]. III. Surface derivatization techniques As mentioned before, the photodecomposition of the electron relay can be decreased and the efficiency of hydrogen production can be increased if the production of the reduced radical form of the relay take place close to a catalytic site. This can be achieved by increasing the effective concentration of the catalyst thereby reducing the diffiisional pathlength. However, the concentration of the commonly employed heterogeneous catalysts, metal or metal oxide particles sometimes supported on inert material, cannot be increased beyond a certain level without adversely affecting the efficiency due to inner-filter effect caused by the dark catalysts preventing a significant fraction of the light from getting absorbed by the sensitizer. Another way to generate the 133 "S""' reduced radical close to a catalytic site is by immobilizing the relay or the sensitizer, either on the catalyst surface or close to the catalyst on the surface of the zeolite matrix. Though techniques such as self-assembly or ion-exchange are easier to achieve, immobilization by covalent surface derivatization involving the zeolite surface silanol (Si-OH ) groups, offers a highly stable, almost permanent anchoring. Aluminosilicate surface immobilization o f metal complexes: Covalent surface attachment of transition metal complexes has been achieved on quartz plates pretreated with 3-bromopropyltrimethoxysilane as well as 3- bromopropyltrihalosilane [15]. Coupling of the silanized surfaces with nucleophilic carbanions (CH, ) generated from transition metal complexes containing methyl bipyridine led to the surface confinement of the chromophores [ 15a]. When simple methyl substituted pyridine or polypyridine carbanions were used, subsequent complexation with transition metals resulted in the covalent surface attachment of transition metal complexes [15b,c]. Wrighton and co-workers have functionalized high surface area silica with 4-[2-(trimethoxysilyl)ethyl]pyridine to give [Si 0 2 ]-SiEtpyr. which on complexation with [(CH 3 CN)Re(CO)3 (2 ,2 '-bipyridine)]* yielded [S1 0 3 ]-[(SiEtpyr)Re(C 0 )3 (2 .2 '-bipyridine)]* [16]. This group has also immobilized monolayer quantities of metal complexes like {L'Re(C0 )3 [(4-L.4'-L)-2.2'-bipyridine]} ' where L = C 0 0 -n-pr-SiCl 3 and L' = 4-ethylpyridine on modified Pt and indium tin oxide (ITO) electrodes [17]. In this case, about 50% of the surface anchored complex was lost after 30 min of repeated voltammetric scanning, probably due to the unstable ester linkage. Corma and co-workers have synthesized chiral metal complexes containing the 134 ligand, (S)-2-(3-triethoxysilyl)propylaminocarbonylpyrrolldine for catalytic studies and have derivatized silica and modified ultrastable zeolite Y surfaces. The ligands are proposed to be anchored to the surface of the zeolite that also contains supermicropores (12-30 Â) [18]. Metal oxide surface derivatization: Fox and co-workers have derivatized colloidal RuOi with a surfactant-like Zn- porphyrins through a coordinatively attached bipyridine unit [19]. The 2,2'-bipyridine unit was proposed to complex with Ru centers present on the surface of the colloidal RuOi during its formation from RuClj [19]. Dye sensitization of nanocrystalline TiOi with polypyridyl complexes of ruthenium and osmium is an active area of research [ 2 0 ]. Surface attachment often has been achieved by simple dipping of the TiOi film in a solution of the carboxylate group containing polypyridyl complexes. Spontaneous formation of a monolayer has been proposed to occur via the ionized carboxyl group either by the interaction with the exposed Ti ions or through hydrogen bonding with surface hydroxyl groups. Covalent attachment to the TiOi surface has also been reported for complexes containing suitable anchoring group like carboxylate. phosphonate. salicylate and acetylacetonate [ 2 1 ]. Ruthenium complexes with silyl-substituted bipyridyl ligands: As discussed above, derivatization involving surface silanol groups would need ligands modified with silyl groups, to create covalent linkages via stable Si-O-Si bonds. To accomplish this for our present purposes with Ru(bpy) 3 -‘ would require silyl- substituted bipyridine ligands. Hosseini and co-workers have synthesized 2,2'-bip>Tidine 135 ligands interconnected by silane spacers and their binuclear ruthenium complexes [ 2 2 ]. Mono- or dilithium-carbanions obtained by the treatment of 4,4’-dimethyl-2,2'-bipyridine with LDA. were treated with bis(chloromethyl)dimethylsilane to form the ligands which on re fluxing with Ru(bpy)iCl 2 gave the binuclear complexes [22]. Kira and co-workers have synthesized homoleptic and heteroleptic ruthenium complexes, RuL)-' or Ru(bpy)iL-\ where L = 5.5'-disilyl-2,2'-bipyridine that were prepared by Ni(0) catalyzed dehalocoupling of 2-bromo-5-silyl pyridines, which in turn were obtained from 2.5- dibromopyridine lithiation followed by treatment with alkylchlorosilanes [23]. IV. Scope and organization: In this chapter we describe the development of a catalyst for photochemical hydrogen production in our efforts toward an integrated solar water splitting assembly. Our research group has earlier developed a zeolite-based RuGj catalyst for the photooxidation of water to O?. This catalyst was evaluated for the photoreduction of water to Hi using the above-mentioned Rufbpy);-' - MV- - EDTA system and the RuOi- zeolite Y catalyst heat treated at 300°C was found to be the most effective. Performance of the RuOi-Y catalyst was compared with other known catalysts as for as rate of hydrogen evolution, stability and the rate of bipyridinium decomposition. Though the rate of hydrogen evolution was slower than that of colloidal platinum. RuOi-Y catalyst was found to be better as for as the stability and bipyridinium decomposition were concerned. Attempts were made to selectively poison the sites responsible for the bipyridinium reduction, while keeping the sites involved in the H, evolution unperturbed. The sensitizer was covalently linked to the zeolitic surface that also contained the catalyst 136 in our efforts to limit methylviologen decomposition and also as a first step towards a proposed integrated solar water splitting assembly. The rest of this chapter is divided into two sections. The experimental section is in turn divided into five subsections and in the first subsection we present an account of the preparation of the hydrogen evolving catalysts. The second subsection concerns surface derivatization including the synthesis of the silyl-bipyridine ligands and anchoring to the zeolite surface. The last three subsections detail the photolysis procedure for Hi evolution, chromatographic techniques to study MV- decomposition as well as Hi evolution and instrumental methods employed to characterize the as synthesized and surface derivatized catalysts. The section on results and discussion is divided into three subsections. In the first subsection we discuss the synthesis of RuOi-Y catalysts from Ru^fCO),] and the characterization of the catalyst by various analytical techniques. In the photolysis subsection we discuss the general chromatographic results, hydrogen evolution with RuOi-zeolite Y. comparison of RuOi-zeolite Y to other hydrogen evolving catalysts, photochemical stability of MV*' radical and strategies to reduce methylviologen decomposition. In the surface derivatization subsection we discuss the results of the synthetic approach and the photochemical experiments with the resulting surface derivatized RuOi-zeolite Y. 137 EXPERIMENTAL SECTION I. Catalyst preparation a. Zeolite based RuO; The catalysts were prepared following a procedure developed in our laboratory [7]. Two grams of commercial zeolite Y (Union Carbide, LZY-52. Lot No. 968083061080-5-10) was stirred overnight in 200 mL of 1 M NaCl solution in order to remove ion-exchangeable cations that may be present. After overnight stirring, the slurry was centrifuged, the supernatant decanted and washed thrice with distilled water. The NaCl ion-exchanged zeolite Y was calcined at 500°C overnight in a steady stream of oxygen to remove organic impurities. After calcination, the chamber containing the zeolite was cooled down at first under a flow of pure nitrogen and then under moist nitrogen to prevent the active zeolite from absorbing ambient impurities. The calcined zeolite Y (1.5 g) was dehydrated at 400°C in vacuum to s 1 x 10“* Torr for 12 hours and was transferred to a glove box in a sealed tube. Triruthenium dodecacarbonyl (Aldrich) crystals were ground to fine powder and a weighed amount of the powder was transferred to the glove box in a glass vial. The amount of Ru:(C 0 ) , 2 employed normally was 10% of the zeolite Y by weight (0.15 g) and 1% for catalysts to be surface derivatized. The dehydrated zeolite and powdered Ru^fCO),! were mixed together intimately in the glove box under inert nitrogen atmosphere to yield a uniformly yellow powder. The solid mixture was heat treated under vacuum (1 x 10"* Torr) at 70“C for one hour and then at 170°C for another five hours. During the heat treatment the pressure inside the vacuum line increased due to the thermal decomposition of RujfCO),,. After heat treatment the gray solid was transferred back to the glove box, thoroughly 138 ground and was stored in the inert nitrogen atmosphere before further treatment. This solid was divided into five portions and one portion was exposed to ambient laboratory atmosphere at room temperature and the rest were heat treated in air at 100°C. 200°C. 300°C and 400°C for 24 hours and are referred to as RuOi-Y rj, RuO^-Y joo , RuOj-Y iqo , RuO.-Yjoo and RuO^-Yjoo, respectively. b. Catalyst poisoning:Poisoning of commercial RuO^ powder was attempted by aqueous solution impregnation method. The poisoning agents tested were alkali metal nitrate salts (lithium, sodium, potassium and cesium), ammonium chloride and cesium chloride and the ratios tried were 10:1. 1:1,1:2 and 1:5 (Ru:poison). A weighed amoimt of the salt to achieve the require ratio was dissolved in 1 mL of distilled water and 100 mg of RuOi powder was added and stirred for two hours in a vacuum tube. After stirring the tube was transferred to a vacuum line, dried and heated to 400°C for 6 hours. In a different method the impregnated RuO, after drying was transferred to a ceramic boat and heated to 400 °C for 6 hours in a calcination chamber under flowing nitrogen. c. Platinum catalyst:Colloidal platinum catalyst stabilized with polyvinyl alcohol (PVA) was prepared following procedure originally described by Nord and Rampino [24], later modified by others [25]. A 0.5% Pt solution was prepared by the addition of 21.3 mg of dipotassium tetrachloroplatinate (Aldrich) to 2 mL of deionized water. A 2% PVA solution was prepared by the addition of 500 mg of polyvinyl alcohol (Aldrich) in small portions to 25 mL of stirred hot (80°C) deionized water. Stirring and heating was continued (two to three hours) until the solution 'vas clear. The solution was cooled to room temperature and was filtered through glass wool. To 12.5 mL of the clear 2% PVA 139 solution in a round bottom flask, 11 mL of water was added imder stirring and to this solution 1 mL of aqueous 0.5% Pt solution was added slowly in a dropwise manner under brisk stirring. This solution was neutralized by the dropwise addition of 0.4 mL of 4% NaOH solution and was heated to boiling for 5 min and then cooled to room temperature. To this brown cloudy solution, 4 mg of sodium borohydride (Aldrich) was added to reduce the dissolved platinum species to colloidal platinum (l^lO'^ M Ft) and was stirred overnight. II. Zeolite surface derivatization a. Synthesis of silyl-substituted bipyridyl ligands:Synthesis of silyl-bpy was carried out according to scheme 3.1. Tetrahydrofliran (THF) was freshly distilled over potassium prior to use. Diisopropylamine (redistilled. 99.5%) and n-butyllithium (w-BuLi. 2.0 M in cyclohexane) were used as obtained from Aldrich in Sure/Seal bottles. Triethylamine (Aldrich. 99.5%) was dried over KOH pellets and absolute ethanol (Pharmco. 200 proof) was dried over molecular sieves. All reactions were carried out under dry nitrogen atmosphere using standard Shlenk procedures using oven dried glassware and syringes. Reactions were monitored by TLC on silica plates, usually with 20% ethylacetate in hexane and UV detection. Silica plates were deactivated with 10% EtjN in hexanes prior to use. Lithium diisopropylamide (LDA) was generated by the dropwise addition of n- BuLi ( 1.3 mL. 2.8 mmol) to diisopropylamine (0.44 mL, 3.1 mmol) in dry THF (4 mL) at -78°C. The solution was stirred at -78°C for 20 min and was transferred dropwise using a carmula to a solution of 4,4'-dimethyl-2,2'-bipyridine (0.460 g, 2.5 mmol) in dry THF 140 H Li i. + Li -i, -irc THF, 20 min LDA LDA THF N — EtOH. Et;N SiCL N - = THF N — THF N — S i- C I S i - 0 . Scheme 3.1 (II mL ). the flask and the cannula were rinsed with dry THF (2 mL). The clear solution turned orange and finally dark brown and was stirred at -78°C for 20 min. The solution was warmed to -10°C for 25 min and then was cooled back to -78°C and was added dropwise using a cannula to SiC^ (0.3 mL. 2.5 mmol) in THE (10 mL) and the dark brown solution turned orange on stirring. The reaction was quenched by the addition of ethanol/triethylamine (EtOH/EtjN) and was either worked up or treated with zeolite as such before quenching. 141 For the standard aqueous workup, the cold solution after the addition of EtOH/EtjN was treated with saturated NaHCOj and was allowed to warm to room temperature. The reaction products were thrice extracted with ethylacetate (100 mL), the organic fractions were combined, washed with saturated NaCl solution (100 mL) and dried over NaiSO^. Filtration to remove the solid Na^SO^ and drying in a rotovap to remove the solvent yielded the product often contaminated with the starting material. Purification of the reaction product was found to be difficult. While monitoring the reaction with TLC on SiOi, product spot was found to be streaky indicating that the product might not be stable on silica precluding purification by column chromatography. Attempts at purification by distillation under reduced pressure were also unsuccessful due to thermal decomposition of the product. Hence it was decided to employ the unseparated reaction mixture (obtained before quenching with ethanol) for the zeolite surface derivatization as the unreacted 4.4'-dimethyl-2,2'-bipyridine would not interfere in the derivatization process. Product: 'H NMR Ô 1.24 (t, = 7 Hz. 9H, Si-O-CH.-C^)), 2.35 (s. 2H, bpy-C/f,-Si), 2.43 (s, 3H. C ^ 3 -bpy). 3.85 (q, 'J = 7 Hz, 6 H, Si-O-C^.-CH^), 6.95 (dd, U = 5.05 Hz, "J = 1.78 Hz. 1H, 5-H), 7.11 (dd, U = 4.88 Hz, "J = 0.96 Hz. 1H, 5 '^ . 8.08 (d, "J = 1.24 Hz. IH. 3 -^ , 8.19 (hr s. IH, 3'-W), 8.48 (hr d, V = 4.96 Hz, 2H, 6,6'-H) Reactant: 'H NMR 8 2.44 (s, 6 H, bpy-C^j), 7.13 (d = 5.1 Hz. 2H. 5S-ff), 8.22 (s, 2H. 2,y-H). 8.55 (d, U = 4.57 Hz, 2H, 6,6'-H) 142 b) Anchoring to zeolite surface: Zeolite surface derivatization with bipyridyl ligand was carried out according to scheme 3.2. Zeolite Y (100 mg) was dried under vacuum at 70°C and was dispersed in dry THF (10 mL). The cold reaction mixture obtained during the synthesis of silyl- substituted bipyridine after the addition of SiCl., was transferred to the zeolite-THF suspension. During this addition, EtjN (1 mL) was also added simultaneously and the reaction mixture was stirred (4 hrs) and was allowed to warm to room temperature. The -OH EtiN -OH -cAsr + -OH -o N— THF -OH -OH S i-C l Zeolite c { Cl surface Scheme 3.2 dark reaction mixture was filtered and washed with THF and then with EtOH. The washed derivatized zeolite was stirred overnight in 1 M NaCl solution (100 mL) to remove ion-exchangeable impurities and dried. The dry derivatized zeolite was Soxhlet extracted with EtOH to remove any unreacted bpy ligand and the solvent was monitored by UV-Vis. After a week of extraction there was no change in the absorption peak of the bpy ligand in solution and the solvent was removed. To ensure the complete removal of 143 the unreacted ligand, solvent extraction was continued for two more weeks with fresh solvent every week. After Soxhlet extraction the derivatized zeolite was washed with EtOH and dried in the vacuum oven. Complexation of Ru(bpy) 2 '* to the surface anchored bipyridine was carried out as shown in Scheme 3.3, by adapting a literature procedure for the homogeneous solution synthesis of heteroleptic ruthenium polypyridyl complexes [26]. C/j-Ru(bpy) 2 C E 2HiO (Strem. 65 mg, 1.25 x 10"* moles) was dissolved in 50 mL of degassed EtOH and 100 mg of the derivatized zeolite Y was added. The solution was heated to reflux under nitrogen atmosphere for about 4 hrs. The reaction mixture was filtered, washed with EtOH and O—Si O - r S i N EtOH + Ru(bpy>2Cl2 —OH R eflux Scheme 3.3 then with CHiCL to remove excess Ru(bpy)iCl 2 and again with EtOH. The washed, dried, derivatized zeolite V was stirred overnight in 1 M NaCl solution (100 mL), Soxhlet extracted with CHiCL followed by Soxhlet extraction with EtOH until the solution showed no characteristic absorption. RuOi-Y catalysts were also derivatized similarly. 144 III. Photolysis The photolysis lamp setup was described in chapter 2. Photolysis reactions for Hi evolution were carried out either in a modified spectrophotometric quartz cell ( 1 cm pathlength, total volume 9 mL) or in a cylindrical glass cell (total volume 52 mL) sealed with a rubber septum. The buffered solutions of a given pH and concentration were first purged with nitrogen for 45 min and were photolyzed under stirring. During photolysis, the evolved hydrogen gas in the head space of the cell was analyzed using a 50 pL airtight syringe (Hamilton) and the solution was analyzed for methylviologen destruction. For sealed tube photolysis, 12 mg of surface derivatized zeolite Y was placed in an NMR tube along with a micro magnetic bar. To the NMR tube 0.5 mL of pH 5 phosphate buffer solution containing 5 x 10'^ M methylviologen and 0.2 M EDTA was added and the solution was degassed by repeating freeze-pump-thaw cycle three times. After degassing, the tube was sealed in vacuum, photolyzed under stirring and the growth MV radical monitored by absorption spectrophotometry. IV. Chromatography a) HPLC: The chromatographic instrument and the separation method are as described in chapter 2. Calibration of MV-' was carried out with five standards of known concentration in the range of 5 x 1 0 '^ M to 2 x 10'^ M. Three independent injections of the standards were carried out and the average was employed for constructing a calibration curve. The peak area as a function of concentration was found to be nonlinear above 1 x 10'^ M and a quadratic expression was used (r = 0.999) to estimate the extent of MV-' decomposition. 145 b) GC: Headspace analysis of photogenerated hydrogen was carried out using a Hewlett Packard model 5890 Series II gas chromatograph equipped with a thermal conductivity detector and a 3' x V* molecular sieve column (60/80 mesh, 5 A, Suppelco). GC analysis of hydrogen in the headspace of the photolysis cell was carried out with argon (prepurified) as the carrier gas. Good separation of hydrogen, oxygen and nitrogen was achieved under isothermal conditions while maintaining the injector port and the detector at 110°C and the column oven at 60°C. Calibration was carried out with five different volumes in the range of 10 pL to 50 pL of 5% hydrogen in nitrogen and the peak area as a function of moles of amount of hydrogen was found to be linear (r^ = 0.999). V. Instrumental methods X-ray diffraction patterns (XRD) were recorded using Ni-filtered Cu radiation (40 kV. 25 mA) with a Rigaku Geigerflex model D/Max 2-B diffractometer. X-ray photoelectron spectra (XPS) were collected with A1 radiation (14.5 kV, 20 mA) using a ESCALAB MARK II Surface Analysis system from VG Scientific Fisions Instruments. Pressed pellet (13 mm) samples were affixed to copper stubs using double sided sticky tapes and the stubs were mounted on to the quick insertion probe and were analyzed. Survey scans were recorded with a step size of 1 eV and selected individual peaks were recorded with a step size of 0.1 eV. Transmission electron micrographs (TEM) were obtained with Philips CM-300 instrument equipped with a high brightness field emission gun (PEG) with an accelerating voltage of 300 kV. Powder samples were suspended in ethanol, dispersed by sonication (5 min) and the suspension was transferred on to holey carbon polymer coated copper grids and air dried. 146 Electronic absorption spectra were obtained using a Shimadzu model UV-265 spectrophotometer. Diffuse reflectance UV-Vis spectra of pressed pellet (13 mm) were recorded on a Perkin-Elmer Lambda 900 spectrophotometer equipped with a diffuse reflectance accessory. 'H NMR spectra were obtained using either Bruker DPX-250 or DPX-400 spectrometers. Fluorescence spectra were recorded using a Spex Fluorolog model FLl 12 spectrofluorometer equipped with a 450-W Xenon lamp and a Hamamatsu R928 photomultiplier tube. A Coming model 125 pH meter with a general purpose combination electrode was used for all pH measurements. RESULTS AND DISCUSSION I. Catalysts synthesis and characterization The study of supported metal carbonyl clusters is an active area of research due to their inherently interesting catalytic properties and also due to their use as precursors for metal and metal oxides [27]. The use of volatile carbonyl clusters as precursors for supported metal and metal oxide catalysts is attracting interest over conventional methods due to the cleaner, milder preparation conditions and better dispersion. Zeolite Y supported RujfCO),; has been studied for its inherent catalytic properties and also as a precursor for zeolite Y dispersed metallic Ru catalyst [28]. We have synthesized RuOi catalyst particles supported on zeolite Y via the thermal decomposition of ruthenium carbonyl followed by air oxidation and the RuO? Y catalysts thus prepared are characterized by XRD, diffuse reflectance, XPS and TEM. 147 a) RuO; synthesis from RuifCO),;: RuOi-Y catalysts were prepared starting from RujlCO),! and dehydrated zeolite Y intimately mixed in an inert atmosphere and heated in vacuum at 70°C for one hour. At tliis temperature under vacuum, Ru3 (C 0 ),i is expected to undergo volatilization and may diffuse into the activated empty (dehydrated) zeolite cages. The size of the RujlCO),, cluster, 6 . 6  x 7.4  x 8.1 Â, is larger than the zeolite Y supercage window (~7 Â) and hence direct incorporation by thermal diffusion at low temperature into the zeolite would be difficult. However, it is possible that Ru 3 (CO ) , 2 could decompose into the metastable Ru(CO), (diameter 6.3 Â) and can reform Ru3 (C 0 ),, inside the zeolite cavities [28,29]. It is also possible for a partially decomposed ruthenium carbonyl to anchor to the zeolite surface with Ru-O-Si linkages. Species like [HRu3 (CO),o(p-OSi)] and [Ru(CO)„(p- OSi)i] where n = 2 and/or 3, have been proposed to form during the preparation of RuOi incorporated into porous vycor glass (PVG) starting from RujfCO),, [30]. After one hour at 70°C, heat treatment was continued at 170'C for 5 h. Complete thermal decomposition of the incorporated or surface held Rujf CO),, at 170°C lead to well dispersed Ru clusters like Ruj that could act as nucléation sites for further crystal growth. This decomposition is expected to follow the same stepwise mechanism observed for Ru^fCO),] on silica as shown in Scheme 3.4 [31 ]. Thermal decomposition Ru3(CO)|2 Ru3(C0)g —* Ru3(C0); —» Ru^ —* RuOt Scheme 3.4 148 was complete within the 5 h heating period as indicated by the return of the vacuum pressure to a stable I x 10"* Torr. Uytterhoeven and co-workers have observed that dispersed metallic Ru in zeolite Y is not stable and undergo oxidation upon exposure to oxygen at ambient temperature (296 K) [32]. This facile air oxidation of the zeolite supported ruthenium clusters was carried out either by exposing the sample to ambient air or at elevated temperatures, b) Catalyst characterization: XRD: Figure 3.1 shows the X-ray diffraction pattern obtained for the starting zeolite Y and that of the other RuO^-Y catalysts with peaks normalized to the most intense zeolite peak at 6.1 °. Similarly Figure 3.2 shows the diffraction pattern for the catalysts in a narrower 16 region of 25-60° with the peaks normalized to the zeolite peak at 31.5°. The catalysts obtained after exposure to air at room temperature and heated at 100°C, RuO?- Yrt and RuOi-Y,oo, respectively, showed no new peaks and except for reflections with somewhat increased intensity for the zeolite peaks in the region 40° to 45°. the diffraction patterns were similar to that of the starting zeolite Y. For the catalyst heated in air at 200°C three new peaks as marked by * on Figure 3.1 and 3.2 appear at 26 values of 28.0°. 35.1 ° and 54.4°. These peaks grew in intensity with increase in the temperature of the heat treatment. Metallic Ru has a Mg-type hexagonal close-packed structure (hep, space group P6~Jmmc, a = 2.7059, c = 4.2820) and show intense reflections for (100), (002), and (101) planes at 26 values of 38.41 ° (25%), 42.18° (25%) and 44.04° (100%), respectively. The reflections due to (002), and (101) planes of metallic Ru appear in the same region as that 149 M 5 LuLiL Figure 3.1 X-ray diffraction patterns of zeolite Y and that of RuO^-Y catalysts prepared by air oxidation under ambient conditions (RT), IOO“C, 200°C, 300°C, and 400‘*C, with peaks normalized to the most intense zeolite peak at 6 . 1 °. 150 Zeolite Y RuOj-Y™ RuO,-Y™ R u O j -Y jq j , R u O j-Y ^ Ul Figure 3J1 X-ray diffraction patterns of zeolite Y and that of RuO^ Y catalysts prepared by air oxidation under ambient conditions (RT), 100®C, 200°C, 300®C, and 400®C, with peaks normalized to the most intense zeolite peak at 31.5°. Inset shows peak deconvolution of the most intense peak of RuO^ at 35.1° for RuO^ Y ^ by curve fitting. of some of the zeolite reflections but that of ( 1 0 0 ) appears in a region devoid of any zeolite peak interference. Though RuOj-Yrj and RuO.-Yioo catalysts showed increased zeolite peak intensities in the 26 region 40° to 45°, one o f the important reflections at 38.41 ° for metallic Ru was absent. This could indicate either the absence of Ru metal or the presence of very small crystallites. In the earlier mentioned study by Uytterhoeven and co-workers also the small RuOi particles obtained from low temperature oxidation did not show aiiy reflections in the XRD but when heated to 200°C showed diffraction peaks [32]. The three new peaks observed for catalysts heat treated at 200 °C and higher compare well to that of (110). (101) and (211) crystal planes of the rutile phase of RuO. expected at 26 values of 28.04° (100%), 35.09° (80%). and 54.30° (63%) [33]. Reflections due to other crystal planes of RuO, are generally weak and are not observable for supported RuOi at low concentrations [6,12]. The absence of any RuOi reflections in the low temperature heat treated samples indicate that at this stage the material either is amorphous or the crystallites were too small for coherent X-ray diffraction or both. The size of the coherently diffracting RuO; crystallites can be calculated from the diffraction peak width using the Scherrer formula [34]. As all three observed peaks due to RuO, were overlapping with that of the zeolite peaks we had to resort to curve fitting to determine the peak width. The results of the curve fitting for the most intense peak of RuOi at about 35.1 ° for RuOi-Yjoo is shown in Figure 3.2 (inset) and the particle size calculated from the full- width at half-maximum (FWHM) of the peak is about 15 nm. Similarly particle sizes 152 calculated from FWHM obtained from curve fitting for the RuOi-Y^oo and RuOi-Yioo catalysts are 9 nm and 5 nm, respectively. These calculated values give a lower estimate of the particle size as various other factors can also contribute to the XRD peak broadening. Though these values appear to be smaller than the particle sizes observed from TEM (see below) the trend of increasing particle size with increase in the temperature of the heat treatment is evident. For RuOi rutile phase, the reflection due to (110) crystal plane is the most intense but as shown in Figure 3.2 the (101) reflection appears to be the well defined peak in our samples. This could indicate some preferential orientation of the RuOi crystallites but in the absence of comparison to diffraction patterns obtained with a fixed incident angle while sweeping the IQ axis (fixed-w configuration) it is difficult to conclude anything about the orientation. It would be of interest to note that RuOi thin films have been found to show preferential orientation depending on the preparation conditions [6.14]. Diffuse Reflectance: Figure 3.3 shows representative diffuse reflectance spectra of the greyish catalysts. RuOi-Yiqo and RuOi-Yjoo- The spectra of RuOi-Y catalysts showed decreasing absorbance with decreasing wavelength up to about 525 nm and showed peaks with maxima at about 375 nm, and 245 nm with a shoulder at 215 nm. The absence of a peak at 392 nm arising from the o-o* transition due to the Ru-Ru bond and peaks at 270 and 238 nm due to LMCT in Ru^fCO),! indicates the decomposition of the precursor. The opto-electronic properties of RuOi have been well studied theoretically [35] and experimentally [36]. It is known that the optical properties of RuOi in the IR region are determined by intraband transitions (d-d) and that above a threshold (1.5 - 1.8 eV) in the 153 1.00 0.75 1 1 U 0.25 0.00 200 300 400 500 600 700 800 Wavelength (nm) Figure 3 J Diffuse reflectance spectra of RuO,-Y,QQ and RuO,-Y300- 154 visible and UV region by interband transitions (p-d) [12,36d]. The spectra of RuOi-Y catalysts are similar to that observed for RuOi encapsulated in PVG that showed peaks at about 250 nm, 360 nm with a broad peak around 800 nm [30]. RuOi single crystals as well as films show decreasing absorbance with increasing energy with a minimum near 2.0 eV (~ 620 nm) with peaks around 3 eV (~ 410 nm) and 5 eV (~ 250 nm) [36c]. From the literature evidence, we assign the observed peaks in the visible-UV region to transitions from valence oxygen p levels to conduction ruthenium d levels. It is difficult to assign these peaks to specific transitions from our present understanding. The sloping baseline that shows increasing absorbance in the near IR region could be due to intraband d-d transitions and has been proposed to be indicative of the metallic nature of the RuOi particles [36b]. XPS: X-ray photoelectron spectroscopy is a good technique to characterize the RuOi-Y catalysts as it would exclusively sample the near surface RuOi species due to the small escape depth of the excited electrons (~ 40 Â) compared to the size of the crystallites (~ pm). Figure 3.4 shows the X-ray photoelectron spectrum survey scans with the binding energy (BE) uncorrected for charging effects. Though RuOi is metallic the zeolite substrate is not and hence shifts in the binding energies due to charging was observed. The carbon Is line with a BE of 284.8 eV with respect to the Fermi level is one of the commonly employed BE calibrant. However, for ruthenium containing samples C Is peak cannot be used as the internal standard for BE calibration as there is interference from the ruthenium 3dj,T line. Hence silicon 2p%i emission observed at 102.7 eV for various aluminosilicates was chosen as the internal standard as it has been used in the 155 literature for BE calibration of zeolitic material. Figure 3.5 shows the Ru 7>dcore level spectrum of the RuOi-Y catalysts with corrected binding energy. The intensity ratio of the two Ru 3d spin-orbit split components I( 3 d;/i)/I( 3 d], 2 ) is expected to be about 1.5. But Figure 3.5 shows a much lower value for this ratio and suggests severe interference of C Is emission with the Ru 3 d 3 ,T component indicating the presence of carbon on the surface. This carbon could be from the CO ligand but it is more likely from contamination, adventitious carbon species adsorbed on the surface during air exposure [37]. The Ru 3 d 3 ,i component appeared at about 285 eV and Ru 3d< i appeared at about 280 eV and are expected to be 4.1 eV apart for RuOi [38]. The binding energies are in the range reported by several recent XPS studies of RuOi [12,13,39]. The Ru 3d;,i BE for catalysts treated at RT, 100° C, 200° C, 300° C, and 400° C were 280.7, 280.6, 280.4, 280.0, and 280.3 eV, respectively. The catalysts show gradually lower BE with increasing heat treatment temperature, except for RuOi-Yjoo which shows the lowest BE. The observed Ru 3d< i BE, 280.7 eV, for RuOi-Yr^ is in the middle of the range of values reported for supported RuOi [12.13.39a,bj. Heat treatment of this catalyst in an oxygen atmosphere could only promote further oxidation and particle growth. But RuGi-Y^go, the catalyst heat treated to the highest temperature (400° C) shows a Ru 3d;/i BE less than that of RuGi-Yrx suggesting that the ruthenium in the latter case could not exist as metallic ruthenium. This indicates that the Ru clusters formed after the decarbonylation of the precursor is getting oxidized to RuGi simply upon exposure to air imder ambient conditions, as mentioned before. The ability to undergo oxidation at low temperature 156 O RT Is O K-Lj 3 L 2 3 100 200 300 400 O Ru Is Na K-L2 3L2 3 BsSat I u Si Si / 2 s 2 p ^ Al Al 1 0 0 0 900 600 500 400 Binding Energy (eV) Figure 3.4 X-ray photoelectron spectrum survey scans of RuO,-Y™ RuO,-Y,q„, RuO,-Y,„, RuO^ Y^^, and RuO^-Y ' 2 * 400' c Is + Ru 3d 3 / 2 RT 100 200 300 400 e a 290 288 286 284 282 280 278 276 Binding Energy (eV) Figure 3.5 X-ray photoelectron spectra of RuO^-Y catalysts prepared by air oxidation under ambient conditions (RT), 100°C, 200°C, 3(X)°C, and 400°C showing the Ru 3d region. 158 could be due to the low Ru-0 bond strength in RuO, compared to other transition metal oxides [40]. Through XPS studies, Lunsford and co-workers have concluded that Ru dispersed in zeolite Y gets oxidized to RuO? upon exposure to even small amounts of oxygen at room temperature (300 Torr, 25° C) [38]. The small decrease in BE upon heat treatment could either be due to the migration of the intrazeolitic clusters to the surface [38] and/or due to increasing particle size effect [13] and it is hard to distinguish between these two processes. The Ru 3d; i emission also shows slight narrowing with increasing heat treatment temperature. This could indicate that at low temperatures, the signal is due to both intrazeolitic and extrazeolitic clusters, whereas at high temperatures, it could be mostly due to the RuOi crystallites on the surface [38]. In the literature, the broadness of the Ru emission from RuO, has also been attributed to the strong satellites shifted to higher BE from the primary spin-orbit component [13] and to the final-state screening effects [ 1 2 ]. TEM: Figures 3.6 to 3.10 show the TEM micrograph for the various RuO^-Y catalysts. RuO^-Yrt shows no distinct features with very small (2-5 nm) well-dispersed particles (Figure 3.6). The RuO^-Yioo sample also showed very small particles but appears to have some degree of surface aggregation leading to irregular shaped particles. The sample heat treated to 200°C showed few small particles but most particles were well shaped and are about 5-10 nm wide and 20-30 nm long (Figure 3.7). Most particles had a characteristic angular V-shaped edges and showed lattice fiinge patterns indicating their crystalline nature. TEM of RuOj-Yjœ (Figure 3.8) showed well formed rod like 159 0 00 mn Figure 3.6 Transmission electron micrograph of RuOi-Yr^. 160 Figure 3.7 Transmission electron micrograph of RuOi-Yi,200- 161 Figure 3.8 Transmission electron micrograph of RuOi-Y300- 1 6 2 Figure 3.9 Transmission electron micrograph of a particle of RuOj-Y300- 163 Figure 3.10 Transmission electron micrograph of RuOi-Y.400- 164 crystallites with morphology similar to that of RuOi-Y^oo but are slightly longer. Figure 3.9 shows one of the widest particle and the crystal planes are clearly visible with an interplanar distance of about 3.2 Â. It is interesting to note that the d spacing of the ( 101 ) planes that give rise to the most intense XRD peak is 3.179 Â. The RuOyY^ sample appears to have bigger particles with their edges somewhat less defined (Figure 3.10). It is interesting to note that similar rod-like morphology has been observed in the literature for Ru metal on zeolite Y heat treated up to 300°C and the morphology changed from fiber-like to more circular on heat treatment to higher temperature (500°C) [40]. It should be noted that commercial powder RuO, showed poorly defined morphology with aggregates of 1 0 nm particles. The faceted edges of the RuOi particles could indicate that the growth front was in a solid-vapor equilibrium involving volatile higher oxides like RuO, and RuO^ [ 6 ]. However such faceted particles were observed even at a low heat treatment temperature of 200 °C. which may be too low to cause such volatilization on a gross scale. Many of the reported RuOi rod shaped single crystals also showed such faceted growing edges with (011), (111). (101) and (1-1 1 ) planes with (101) as the largest face [11]. Similar faceted RuOi particles have been observed by electron microscopy in the cross-sectional morphology of films grown at much higher temperatures (600°C) [ 6 ]. One common feature among the samples heat treated to temperatiu’es above 2 0 0 '’C is that the finger like particles that appear to be sticking out of the zeolite support. Also since most particles were seen to be anchored to the zeolite surface it is possible that the particles were growing from inside out. 165 In conclusion, thermal decomposition of RujCCO),, offers a convenient route to RuOj-Y catalysts. From XRD, XPS, and TEM analysis the following could be concluded: 1) Ru clusters formed upon decarbonylation of the Ru^fCO),? precursor undergoes oxidation upon exposure to air even at room temperature, 2 ) upon heat treatment the RuOi particles grow out of the zeolite to form fiber like crystallites on the zeolite surface and 3) the crystals appear to have well defined morphology which might be important for catalytic purposes. II. Photolysis experiments It is clear from XRD and XPS results the RuOi-Y catalysts prepared by heat treatment to various temperatures exhibit different extent of crystallinity and morphology and it can be expected that their catalytic performance would be different too. The performance of the RuOi-Y catalysts was evaluated and compared with other hydrogen catalysts like commercial powder RuOi and colloidal platinum with respect to the hydrogen production efficiency and methylviologen destruction under identical photolysis conditions by GC and HPLC. respectively. In this context we shall discuss the chromatographic methods of separation and calibration at first followed by the comparison of the catalytic performance. Finally we shall discuss the instability of the methylviologen radical and our attempts to minimize it. a) Chromatography: HPLC: Quantitative study of the extent of decomposition of the electron relay, methylviologen, is possible only after separation as the UV absorption bands of other components present in the system interfere with that of MV-*. Reverse-phase ion-paired 166 chromatographic separation of methylviologen from other nitrogen containing organic bases has been studied [41]. There have been a few previous attempts to study the photochemical decomposition of methylviologen during hydrogen evolution by employing RP-IP technique using an ODS column [25,42]. Moradpour and co-workers have adapted an earlier report [41a] for quantitative determination of methylviologen decomposition under hydrogen evolution conditions with colloidal platinum from the decrease in MV- peak area [25a,43]. From a non-photochemical reaction that did not employ chromatographic separation the decomposition product was reported as N,N'- dimethyl-4,4'-bipiperidine. Sasse and co-workers by employing an ODS column but with a different elution method have been able to separate MV-' from the hydrogenation product after hydrogen evolving photolysis with colloidal platinum and identified the latter as l-methyl-4- (pyridinyl-4'-yl)pyridinium cation [25b]. According this study, the hydrogenation product reported by Moradpour and co-workers was observed only under uncontrolled hydrogenation and not under photolytic conditions. However, our attempts to simply apply these methods were not very successful with MV- peak showing significant peak tailing and overlap with that ofRufbpy))-'. Unfortunately these reports did not include any chromatogram preventing comparison [25]. It may be interesting to note that Matsumoto and co-workers have failed to detect any hydrogenation products with chromatographic conditions similar to that reported by Sasse and co-workers [42]. Significant modification of the separation procedure from the reported isocratic methods to that of the mobile phase program presently employed offered good separation 167 of MV-" form Ru(bpy)3 ‘* and low peak broadening (see chapter 2 for a detailed discussion on the chromatographic method and separation). Figure 3.11a shows the HPLC chromatograms obtained for different concentrations of MV-' with a 10 /^L sample loop. Three injections were made for each concentration as shown in Figure 3.11a and the average peak area was employed for the calibration plot. The calibration plot (Figure 3.1 lb) shows significant deviation from linearity for MV-' concentrations above 1 x 10 '’ M which could be alleviated by employing a smaller sample size. However, manual attempts to inject smaller volumes using the sample loop resulted in reproducibility problems and also employing a 2 ^L sample loop would cause S/N problems at low concentrations. Hence it was decided to employ a quadratic expression for the calibration as shown in Figure 3.1 lb. Since the typical MV- concentration in photolysis experiments is of the order of 1 x 10'^ M this should not cause any problems. When higher concentrations of MV-' were employed the solutions were analyzed after quantitative dilution with known amount of distilled water. GC: Figure 3.12a shows the GC chromatograms obtained with different volumes of 5% hydrogen in nitrogen. The peak with RT of 0.85 min is that of hydrogen and the peak at 1.9 min is that of nitrogen and the very small broad feature at around 1.2 min is that of residual oxygen. It is interesting to note that the molecular sieve GC colunm employed does not elute CO, thus eliminating potential interference. Three injections were made for each standard as shown in Figure 3.12a and the average peak area was employed for the calibration plot. Calibration curve as shown in Figure 3.12b was linear (r^ = 0.999) and the calibration process was repeated periodically. 168 a ) 2 * i 6 ^ M I 2.0- 1*10^ M VO 1.5 - (§) I§ 1.0 4 5*10^ M I 0.5 4 2.5*10"^ M < 0.0 5*10'® M — 1— “T— - 1 — 11 13 14 Time (min) 4 b) Peak area = A*[MV^*]*[MV^*] + B*[MV^^ + C; 3 I 2 I 1 A = -6.42*10 B = 3.06*10“ C = -4.76*10' 0 0.0 0.5 1.0 1.5 2.0 [MV2+] (1Q‘^ M) Figure 3.11 HPLC calibration for a) Chromatograms of 10 of standards b) Calibration plot showing the peak area as a function of concentration (r^ = 0.999). 169 a ) 14 1 2 0 10 8 6 1c 4 Î 2 0 0 Time (min) a) m 0 X 1.2 ca % 0.9 1 0.6 £ 0 . 3 0.0 0 10 20 30 40 50 60 Volume of Standard (jiL) Figure 3.12 GC calibration for hydrogen a) Chromatograms of different volumes of hydrogen standard b) Calibration plot showing the peak area as a linear function of the volume of hydrogen standard (r^ = 0.999). 170 b) Hydrogen evolution experiments with RUO2-Y catalysts The photochemical water oxidation ability of the RuOi-Y catalysts has been well studied by our group and was found to outperform other traditional catalysts [7]. In order to effect a complete photochemical water splitting, it is necessary to have a catalyst that would also catalyze the reduction half reaction. To evaluate the photochemical catalytic performance of our RuOi-zeolite Y samples it was decided to employ the well studied sacrificial system with Ru(bpy)j-" as the sensitizer, methylviologen as the electron relay and EOTA as the sacrificial electron donor. Sacrificial systems have no practical use as discussed before, but can be used to evaluate the performance of individual components like our RuOi-Y catalysts with existing ones under actual photochemical conditions. Under constant irradiation of a standard solution containing Ru(bpy) 3-*, MV-'. EOTA and RuOi-Y catalyst, the formation of small gas bubbles is noticeable. The evolved gases that get accumulated could be easily analyzed by gas chromatography, and Figure 3.13a shows the chromatograms obtained at various time intervals during one such photolysis run with RuOi-Yiqo- The chromatogram before photolysis shows only an intense peak for nitrogen that was used to purge air from the photolysis cell and a very small broad peak for oxygen. This small broad feature due to oxygen remained constant after about 40 min of purging with inert gases and was present even after several hours of purging. This could be due to the small amount of residual oxygen present in the system even after purging or more likely is introduced via the syringe. Though the syringe employed to sample the headspace is an airtight one. it was not equipped with an airlock 171 and could have a very small amount of oxygen contamination, and as it remained constant it was deemed acceptable. Upon photolysis, the area under the hydrogen peak steadily increases and that of the nitrogen shows a small decrease indicating the changing composition in a closed system. With argon as the carrier gas, the sensitivity of the thermal conductivity detector for the lighter hydrogen gas is much higher than that of the heavier nitrogen making the compositional change of the former more apparent than the latter. The decrease in nitrogen peak area is apparent when the photolysis is continued for longer time period or a change in the composition of the photolysis solution that leads to the production of significant amounts of hydrogen. Under such conditions, to prevent the accumulation of hydrogen that could cause a pressure artifact, the head space needed to be vented and purged with nitrogen before continuing the photolysis. When the concentrations of the various components involved and the head space volume are the same, with the help of a calibration plot, the hydrogen peak area could be used for quantitative comparison of various catalysts. Figure 3.13b compares the hydrogen evolution with RuOi-Y^oo under standard conditions (0.2 M EDTA) with that in the absence of EDTA. As expected, there was no hydrogen evolution in the absence of EDTA due to the efficient back electron transfer. Similarly if Ru(bpy)j-" or MV*" was omitted during photolysis there was no hydrogen production either. However, in the absence of the catalyst the Rufbpy))-' - MV-‘ - EDTA system alone did produce very small amounts of hydrogen and could be due to trace amounts of impurities present in EDTA or due to inefficient cleaning of the photolysis 172 a ) B c I CO 0 1 2 3 4 Time (min) 35 b) 0.2 M 30 0.2 M, No Catalyst C 25 0.0 M % 20 2I" 10 I 5 0 I ▲ “I" 0 50 100 150 200 250 Time (min) Figure 3.13 Photogeneration of hydrogen with RuOj-Yj^o a) Gas chromatograms of the headspace (4 mL) obtained during the photolysis of 5 mL of solution (pH 5.1,0.1 M acetate buffer) containing Ru(bpy)^^* (2 x 10^ M), MV ^ ^ ( 1 x 10 ^ M), EDTA (0.2 M), and RuOj-Yjqo catalyst (3 pmol). b) Comparison of hydrogen evolution under standard conditions (0.2 M EDTA) with that in the absence of EDTA (0 M EDTA) and in the absence of catalyst (0.2 M EDTA, No catalyst). 173 cell. This observation has been made in the literature as well [25]. This small hydrogen evolution was almost always observed after relatively long periods of photolysis involving EDTA, and the origin of this effect is beyond the scope of this work. Figure 3.14a compares the performance of RuOj-Yiqo as a function of EDTA concentration. The rate of hydrogen evolution increases with increasing EDTA concentration, however above 0.2 M, solubility of EDTA becomes an issue. Figure 3.14b compares the performance of RuOj-Y.qo at different pH conditions (pH 4 and 5 acetate; pH 6 phosphate). It was found that the optimum pH was around 5 and is in accordance with what has been observed in the literature [3,44]. The effect of pH is due to two opposing effects, decreasing pH favoring of the formation of hydrogen from the pH dependence of the redox potential and decreasing efficiency of EDTA as a sacrificial donor. Figure 3.15 compares the water photoreduction performance of the various RuOi- Y catalysts obtained by heat treatment to different temperatures. The catalyst obtained by heat treatment and oxidation at 300°C showed the highest rate of hydrogen evolution whereas the room temperature oxidized catalyst showed the least. The catalyst performance with respect to the rate of hydrogen evolution followed the trend: RuOj-Yjoo > RuOi-Yioo > RUO2 -Y400 > RuOi-Y100 > RuGi-Yrt- Table 3.1 summarizes the rate and amount o f hydrogen produced with different catalysts. Catalysts obtained from the same batch showed small changes in the rate (< 5%) while those obtained from different batches varied as much as 10% but followed the same general trend. When care is taken to recover all of the used catalyst, it could be used again without any significant loss in activity. 174 a ) 0.20 M 30 - 0.02 M 0.00 M QI g 15 - î1 .0 . 0 50 100 150 200 250 Time (min) b) 30 - 5.9 3.96 o I 20 - I 15 - t 1 0 . X 0 50 100 150 200 250 300 Time (min) Figure 3.14 Comparison of the hydrogen evolution performance of RuO j-Y jqq a) at different EDTA concentrations and b) at different pH (pH 3.96 and 5.1 acetate; pH 5.9 phosphate), under conditions as described in F i^re 3.13. 175 i < I I 0 100 15050 200 250 300 Time (min) Figure 3.15 Comparison of hydrogen evolution performance of the various RuO^-Y catalysts obtained by heat treatment to different temperatures, during the photolysis of 5 mL of solution (pH 5.1,0.1 M acetate buffer) containing Ru(bpy)^ *(2 x 10^ M), ( 1 X 10 ^ M), Na^EDTA (0.2 M), and RuOj-Y catalyst (3 ^unol), headspace volume 4 mL. 176 Catalyst Hi Rate (/zmol/hr) Hi Yield (//mol) RuOi-Y rx 2.4 9.8 RuOi-Y100 3.9 15.8 RuOi-Y200 7.7 30.9 RuOi-Y300 10.3 41.3 RuOi-Y4 oo 5.9 23.5 5 mL of photolysis solution (pf 5.1,0.1 M acetate buffer) containing Ru(bpy))-' (2 x 10 M), MV-' (1 X 10'^ M), NajEDTA (0.2 M), Catalyst (3 yumol). Headspace volume 4 mL. Table 3.1 Comparison of hydrogen evolution efficiency of RuO^-Y catalysts. Though the mechanism of hydrogen evolution from MV" radicals on RuOi catalytic particles is yet to be established, we can draw some inferences from the proposed pathways for metal colloids. In the earliest and still widely accepted mechanism, analogous to the electrochemical reduction of water on bulk electrodes, the colloidal Ag, Au or Pt particles act as electron-accumulating microelectrodes [45]. The reduced M V radicals get oxidized on the catalyst surface with the concomitant reduction of the metal atoms. The multicharged particles then pick up protons from the solution at a rate determined by the overpotential, forming metal hydride type intermediates which ultimately produce molecular hydrogen. In another less popular model the methylviologen radical picks up a proton in solution forming MVH- ‘ which upon encountering a catalyst transfers the hydrogen atom [46]. In this model the catalyst acts 177 merely as a site for hydrogen atom deposition and recombination to produce molecular hydrogen. In either of these models the key steps that precede hydrogen evolution are the formation of intermediates of the type M-H, where M is a surface metal site and recombination of the surface held hydrogen atoms. If the formation of a Ru-H type intermediate is important then the number of covalently unsaturated ruthenium sites per unit area of the catalyst would have an important role in determining the efficiency. Also the distance between such covalently unsaturated ruthenium atoms might be important for the combination of the surface held hydrogen atoms to form molecular hydrogen. As shown in Table 3.2 the different surfaces of RuO^ crystals exhibit differences in both of these crucial compositional properties. Recently the catalytic reactivity of the RuO, (110) surface has been studied in detail with respect to oxidation of carbon monoxide and the covalently unsaturated ruthenium atoms in the atomic-scale structure are found to be the reaction center [47]. The electrochemical properties of the ( 110), (GDI ), ( 111 ), ( 101 ) and (100) exposed faces of single crystal RuO; electrodes are known to be very different [5.48]. The idealized (111) and (101) surfaces of RuO? has only Ru atoms on the surface (Table 3.2) and showed only one pair of cathodic/anodic peaks at about -0.3 V (SCE) and has been attributed to the reversible hydrogen adsorption/desorption on Ru sites. The (110) and (001 ) surfaces in addition to Ru also has surface O atoms and showed in addition to the 178 Surface Composition Ru-Ru distance Number of broken per unit cell (Â) Ru-0 bonds ( 1 1 0 ) 2 Ru, 2 0 3.11,3.53 2 (Ru'), 1 (Ru") (0 0 1 ) 1 Ru,2 0 4.49 2 ( 1 1 1 ) '/zRu 5.46,6.35 3 ( 1 0 1 ) 2 Ru 3.53,4.49, 5.46 3 ( 1 0 0 ) 1 Ru 3.11,4.49 3 Table 3.2 Structural parameters of various surfaces of single crystal RuOj (adapted from Ref. 5). above mentioned peaks another pair of cathodic/anodic peaks at about 0 V (SCE) and has been attributed to the hydrogen chemisorption on the O sites. The electrocatalytic properties of these surfaces followed the order ( 1 0 1 ) best > ( 1 1 1 ) good > ( 1 1 0 ). (0 0 1 ) poor, with hydrogen evolution starting at -500 mV, -520 mV and -620 mV (SCE), respectively. The observed differences have been attributed to the structural and electrochemical differences between the different surfaces [5.48]. As all the RuOi-Y catalysts are obtained from the same Ru 3(CO),i-zeolite Y precursor, the differences in their catalytic performance must relate to the structural differences in the generated RuO? particles. X-ray diffraction of the catalysts show increasing crystallinity with heat treatment temperature. If the catalytic performance is related to the crystallinity alone, then the RuOj-Y^oo catalyst should be the best, yet it consistently performed poorer than RuOi-Y^qo and RuOi-Yiqo sample. The less than 179 optimum performance of RuOi-Y^oo indicates that the activity depended not only on the extent of crystallinity but may also depend on the morphology and/or the size of the particles. The RuO, particle size as observed from TEM studies grows from 2 nm to >50 nm with heat treatment. As the particles grow, the effective surface area decreases and could lead to decreased catalytic activity. However, our observation is contrary to this expected behavior. Gratzel and co-workers have reported an exceptional water reduction activity for colloidal platinum with an optimum particle size of 110  [49]. However, the validity of this conclusion has been questioned by others discounting the particle size effect in hydrogen evolution [43], As all the RuOi-Y catalysts show oxidized ruthenium in X-ray photoelectron spectroscopy, the only other reason for the difference in the catalytic performance could be the particle morphology. Recently the importance of specific crystal faces, orientation, and morphology for photochemical processes has attracted considerable attention. For the photochemical reduction of Ag' to Ag metal using TiO, crystals, the {101} facets have been found to be far more reactive than the others [50]. Use of inert support to manipulate the exposed crystal faces, orientation and morphology is a novel way to control the involved photophysical properties and photochemical reactions. In a recent article Scaiano, Garcia and co-workers have reported the preparation of TiO^ clusters in zeolite Y, p, and mordenite and have noticed the photophysical properties of the intrazeolitic TiOi clusters to be dependent on the zeolite matrix [51]. Anpo and co-workers have recently reported that efficiency and selectivity of products in the photocatalytic decomposition of NO with TiOVzeolites depended on the method of preparation and the zeolite employed [52]. 180 Deposition of M 0 O 3 crystallites on silica surface has been reported to control its morphology [53]. The silica support apparently stabilizes the crystallites with predominating side faces and a smaller fraction of the basal (OkO) faces. Oxidation of allyl bromide to acrolein has been correlated to the basal (010) face of the M 0 O3 . From these reports, it is clear that morphology of the active catalyst could be manipulated by the 'inert' support and hence could affect the catalytic performance. The RuGj-Y photocatalysts employed in this study for water reduction have been previously shown in our laboratory to possess catalytic efficiency for water oxidation, and it was found that the RuGi-Yioo is the best catalyst [7]. The differences in activity have been attributed to the unique fibrous morphology of the RuGi crystals growing on the zeolite surface. In a previous study it has been shown that heat treatment of hydrated RuGi to 150°C produce the best catalysts for water oxidation that were partially dehydrated and corrosion resistant [54]. However Kleijn and co-workers have reported that colloidal RuGi prepared by the thermal decomposition of RUCI 3 at 300°C. 400°C and 500°C showed negligible difference in hydrogen production using the Ru(bpy) 3 -' - MV-' - EOTA system [55]. The authors have noted that though the specific surface area decreases with increasing decomposition temperatures, the aggregates were of the same size and the small changes in performance were attributed to the difference in the mass transport of MV * to the catalytic surface. As the catalysts used in our study contained RuG; well dispersed in the zeolite matrix with no aggregate formation, differences in mass transport could not explain the observed differences. Also, comparison of our results with that of Kleijn and co-workers underline the importance of the zeolite matrix 181 and not just heat treatment for explaining the observed differences. In the case of RuO, dispersed on barium titanates, it has been reported that the UV photocatalytic activity for water splitting depended on the RuO, preparation temperature [ 8 ]. In this report, the activity dependence of catalysts with similar RuO, particle size and dispersion on the preparation temperature has been proposed due to the temperature dependent creation of optimum states in RuO, for charge transfer. Though the mechanism in this case is different it is interesting to note that the RuOi/barium titanate catalyst prepared at 573 K outperformed catalysts prepared at other temperatures. Since the RuOi particles in RuO, -Y grow on the zeolite surface presumably from the inside it is not unreasonable to expect the particles to show specific exposed surfaces and morphologies. As the unsaturated ruthenium atom densities as well as Ru-Ru distances are different in the various faces (Table 3.2), the morphology can control the hydrogen evolution efficiency of the catalysts. The literature evidence discussed above suggest that the observed differences in the efficiency of RuO^-Y catalysts may arise from the differences in the hydrogen evolution efficiency of different crystal faces of the RuO; particles. The catalyst preparation temperature as well as the ability of the zeolite matrix to control the morphology and photophysical properties of the supported metal oxides may in turn determine the exposed crystal faces and hence the efficiency of the RuOyY catalysts. From our present understanding it is difficult to pinpoint the origin of oxidation temperature dependence of the hydrogen evolution efficiency and further research is needed to elucidate the mechanistic and structural reasons. It might be useful to prepare RuOi on zeolites with cage and channel structures different from that of 1 8 2 zeolite-Y and compare them with that of RuOi-Y catalysts in order to better our understanding of the role of zeolites in generating the RuO, catalysts. Figure 3.16 shows the long term stability of the RuOi-Y iqo catalyst. At the end of each run the evolved hydrogen was purged off the headspace and the photolysis continued. The catalyst did not show any significant change in its ability to produce hydrogen even after 40 hours of photolysis. This is in contrast to the hydrolytic instability of the popular colloidal platinum catalyst as discussed below, c) Comparison of RuOi with other catalytic system Figure 3.17 compares the performance of RuOi-Yi^ with that of colloidal platinum (henceforth referred to as platinum or simply as Pt) and commercial powder RuOi (the conditions are as in Table 3.3). Both Pt with lower effective catalyst content ( 1 jimol) and RuOi powder with higher effective catalyst concentration content ( 15 pmol) outperform RuOi-Yiqq ( 1 . 2 pmol) as for as the initial rate of hydrogen evolution is concerned. However when the photolysis is continued after venting the hydrogen and purging the head space with nitrogen, both Pt and RuOi powder show sharply reduced rates whereas that of R u O i -Y, qo remained almost the same. Figure 3.18a compares the rate of these three systems normalized for the amount of catalyst employed and shows that RuOi-Yioo is better than RuOi powder but is still slower than Pt in initial rate. Figure 3 .18b compares the total amount of hydrogen produced at the end of the second photolysis run (total photolysis time 16 h) normalized to the maximum in the case of Pt. The highest amount of Hi produced, in the case of colloidal platinum is less than 2% of the EDTA employed in terms of molar quantities. Hence the decrease in the rate of 183 Run 1 Run 2 Run 3 es < I 0 200 600 800400 1000 1200 Time (min) Figure 3.16 Long term catalytic stability of the RuOj-Y jqo catalyst under Hj evolving conditions in 30 mL of solution (pH 5.1,0.1 M acetate) containing Ru(bpy)g^* (2 x 10"* M), MV^*(2 X 10 ^ M) and EDTA (0.2 M), headspace volume 22 mL. 184 RuOg-Y^QQ Run I O RuOj-Yjoo Run 2 ■ RuOj Run 1 □ RuO^ Run 2 A Pt Run 1 A Pt Run 2 I # o I # o □ □ 00 A A L n A ID ▲ A -i------1------1------1------1------1 if-r- — I 1------1------1------r y ~T —I------1------1------1----- 0 100 200 300 400 500 0 100 200 300 400 500 0 100 200 300 400 500 Time (min) Figure 3.17 Comparison of the evolution performance of RuO^-Y^ with that of commercial powder RuO^ and colloidal platinum (Pt), under the conditions given in Table 3.3. a ) e RuOj-Yjoo Run 1 A A 0 RuOj-Yjf^ Run 2 ■ RuOj Run 1 A □ RuOj Run 2 A Pt Run 1 A I A PtRun2 I A A • o A # A ^ z A ^ ■ □ " b ■ Q a o ■ 0 100 200 300 400 500 600 T im e (m in) b) •a I I I % I 0 Figure 3.18 a) Comparison of the rate of Hj evolution for RuOj-Yjp^, powder RuO^^ and Pt, normalized for the amount of catalyst employed, b) Comparison of the total amount of hydrogen produced at the end of the second photolysis run (total photolysis time 16 h) norm ali^ to the maximum (Pt), under the conditions given in Table 3.3. 186 hydrogen evolution with photolysis time in the case of Pt and RuO^ powder could not be due to lack of EDTA. As mentioned in Chapter 2, the sensitizer is prone to photodecomposition. However the conditions employed in these three cases were identical, and under these conditions there should not be appreciable Ru(bpy)j-* decomposition which was confirmed by HPLC. as discussed below. After photolysis the pH of the buffered medium slightly increases by about 0.1 to 0.2 units. This small pH change which was almost the same among the three catalysts could not be the reason for the decrease in rate. The headspace was also purged between the two runs to prevent the accumulation of excess hydrogen in the headspace that could impede hydrogen evolution slowing down the rate. This leaves two other possible reasons, irreversible decomposition of the electron relay or deactivation of the catalyst. RuO^ powder is known to undergo oxidative Catalyst amount® H, Yield " (/^mol) (/^mol) Colloidal Pt 1 1 1 2 . 6 Powder RuO, 15 101.7 RuO,-Y200 1 . 2 93.2 ' Colloidal Pt -1 mL of 1*10'^ M Pt solution; RuOi - 2 mg; RuOi-Y^oo - 4 mg; ’’ Total amount of H? produced upon 16 hours of photolysis of 30 mL of solution (pH 5. 0.1 M acetate buffer) containing Ru(bpy)3 -* (2 x 10"* M), MV-* (2 x 10'^ M) and Na^EDTA (0.2 M), Headspace volume 22 mL. Table 3 J Comparison of hydrogen evolution efficiency of RuO^-Yioo, Pt, and RuO,. 187 corrosion under oxygen evolving conditions [56] and should be stable during hydrogen evolving reducing conditions. The electron relay methylviologen, as noted before in the literature, often undergoes hydrogenation during water photoreduction [3,44,25b]. In order to decouple these two possible reasons, after the initial photolysis in the case of RuOj powder, the catalyst was isolated with care and reused. No appreciable change in hydrogen evolution rate could be noticed suggesting stability of the RuOi powder catalyst and MV-* decomposition. However such isolation in the case of colloidal Pt was not possible and hence ftesh Pt catalyst was added to the solution after the initial photolysis, and on continued photolysis, the initial rate could not be restored back again suggesting MV-' destruction. However it may be interesting to note that with Pt catalyst, after the complete stoppage of hydrogen evolution, if fresh methylviologen was added and photolysed further, the originaKrate could not be restored suggesting that the PVA stabilized colloidal Pt is not very stable under long term photolysis. In order to evaluate the MV- destruction quantitatively it was decided to employ HPLC. d) decomposition during photolysis The chromatogram of the solution before and after four hours of photolysis with RuOi-Yjoo are shown in Figure 3.19. Absorption spectra of the components eluting at 12.3 and 14.2 min corresponded to that of MV-' and Rufbpy);-'. respectively. The chromatogram showed decreased MV-' concentration after photolysis and almost constant Rufbpy)/' with the spectra showing no change. However, it should be noted that the zeolite matrix of the RuOi-Y catalysts is microporous with significant ion- exchange capacity and since MV-' cations are smaller than the zeolite cage windows 1 8 8 MV-* could get exchanged in to the zeolite. Hence after photolysis, the catalyst particles were carefully recovered, washed thrice with distilled water and ion-exchanged with 5 mL, I M NaCl overnight. The chromatogram of this solution is also shown in Figure 3.19. The solution after NaCl exchange showed a sizeable MV- peak and no significant Rufbpy)]-' indicating that the washing was complete and that the MV-' was in fact from the zeolite cages. To quantitate the MV-' destruction, the difference between the amount present before photolysis and the sum of the MV-' present after photolysis and after NaCl exchange was determined using the calibration curve. Figure 3.20 shows the chromatograms before and after four hours of photolysis with RuO; powder. At the concentrations employed, methylviologen adsorption on RuO, particles is negligible. In this case (as well as in Pt) the difference between these two samples could then only be due to decomposition, as there is no ion-exchange involved and the chromatography directly yields the extent of decomposition. As can be seen from Figures 3.19 and 3.20 the only noticeable changes in the chromatograms after photolysis are the decrease in the methylviologen peak and the minor changes around the EDTA peak. Unfortunately, the chromatograms did not show any new significant peaks that would correspond to the methylviologen decomposition product(s). However, there were a few very minor peaks in the early part of the chromatogram that correspond to the Rufbpy);-' photoproducts described in Chapter 2. The shape of the EDTA peak changed after photolysis due to the EDTA oxidation product eluting together with EDTA. The inability of the chromatographic method employed to elute the decomposition product(s) of MV-' is disappointing as it precludes 189 a) •Ê I < 0.0 0 5 10 15 20 25 30 Time (min) b) Before photolysis a After photolysis s After NaCl exchange VO «N I •e 0.5 - x> < 0.0 12 13 14 15 Time (min) Figure 3.19 Chromatographic estimation of decomposition after 4 hours of hydrogen evolution photolysis with RuOj-Yjgg in 5 mL of photolysis solution (pH 5.1,0.1 M acetate buffer) containing Ru(bpy)^^* (2 x 10"* M), (1 x 10 ^ M), EDTA (0.2 M), and R u O ^-Y ^ (3 jimol). 190 ------Before photolysis 1.5 - After photolysis cs so « 1.0 - 0.5 - I< l I 0.0 i. . 1------,------T r"" " —1----- “ T" 0 5 10 15 20 25 3 Time (min) Figure 3 JOChromatographic estimation of decomposition after 4 hours of hydrogen evolution photolysis with RuO^ in 5 mL of photolysis solution (pH 5.1,0.1 M acetate buffer) containing Ru(bpy)g"* (2 x 10“* M), (1 x 10^ M), EDTA (0.2 M), and powder RuO, (15 pmol). 191 the study of the possible pathways of decomposition and only the extent of decomposition can be determined. Figure 3.21 compares the methylviologen decomposition as a function of photolysis time for the three different catalysts after the production of similar quantities of hydrogen (Pt - 41.7 ^imol; RuOi - 39.9 /umol; RuG i-Y iqo - 38.3 umol). The Pt catalyzed nm showed the highest extent of decomposition and that of the RuOi-Y iqo the least. The MV-' appeared to be decreasing fairly linear in the cases of Pt and RuOi powder, with increasing time. In contrast, with RuOi-Y iqo there appeared to be an initial decrease in the concentration followed by a more gradual decrease. The difference in the profile could be explained by the ion-exchange process by which MV-' leaves the solution into the zeolite matrix as mentioned before. The effective life cycle of methylviologen calculated as the ratio of moles of hydrogen produced to the moles of methylviologen destroyed from different runs were about 10-12 for colloidal Pt, 20-25 for powder RuOi and 35-40 for RuOi-Y iqo - These effective life cycle numbers indicate that though the initial rate of hydrogen evolution was better with Pt, the destruction of methylviologen for comparable total hydrogen evolution was also higher. The effective life cycle numbers for methylviologen corresponding to the different RuOi-Y catalysts for which the hydrogen evolution rates were shown in Figure 3.15 are 28, 35, 38,24 and 19. for 400°C. 300 °C, 200 °C. 100 °C, and RT heat treated catalysts, respectively. If catalyzed hydrogenation is the only or predominant decomposition pathway with RuOi-Y catalysts then the methylviologen decomposition should increase with increasing hydrogen production, making the methylviologen effective life cycle (the ratio) invariant. 192 1.1 e RUO2-Y200 1.0 B Ru0 2 Powder A Colloidal Pt 0.9 -j ^ 0.8 - S "o 0.7 - I 0.6 - 0.5 0.4 4 0.3 - r 0 10 Time (hrs) Figures 3.21 Comparison of methylviologen decomposition as a function of photolysis time for RuO^-Y^oQ, Pt, and RuO^. 193 However the observed differences in the methylviologen effective life cycle indicate that the methylviologen destruction is not linearly dependent on hydrogen production alone. The methylviologen effective life cycle ntunbers rather indicate that for identical mass of RuOi Y catalysts the faster the hydrogen evolution, the slower is the decomposition resulting in higher methylviologen effective life cycles. This could mean that either the catalysts exhibit different ability to catalyze the hydrogen evolution as compared to methylviologen hydrogenation or the faster the MV" radical gets turned over to MV-' the lesser is the destruction. In order to evaluate the latter idea it was decided to evaluate the stability of MV " radical under photochemical conditions, e) Photochemical stability of MV radical: Figure 3.22a shows the absorption spectra of the nitrogen purged, sealed Rufbpy)]-' - MV-' - EDTA system upon continuous photolysis in the absence of any catalyst and Figure 3.22b shows the chromatograms of the solution obtained before and after such photolysis. Though the solution contained 1 x 10 - M MV- and 0.2 M EDTA. after about 15 min of photolysis, the solution attained the maximum [MV "] and on further photolysis, the [M V ] steadily decreased. The maximum concentration of the radical calculated with the absorption coefficient ( ejo:„m = 1 . 4 x 1 0 ^ M 'em ' ) was about 1.27 X 10"* M indicating that only about 12% of the dication has been converted into the radical form. From Figure 3.22a it is obvious that the MV" radical absorption in the 450 nm region is very small and hence cannot significantly interfere with the photoefficiency of the MLCT excited state production of the sensitizer via inner filter effect. In the presence of the small amoimts of MV" (<1.3 x 10"* M) the large excess of EDTA (0.2 M) 194 a) I I— I 1------1------1 0 100 200 Time (min) 200 400 600 800 Wavelength (nm) b) Before photolysis I After photolysis so «n CN @ l-O 8 I •2I 0.5 0.0 12 13 14 15 Time (min) Figure 3.22 Photostability of methylviologen radical a) Absorption spectra of the anaerobic Ru(bpy);^^ - - EDTA solution during continuous photolysis in the absence of catalyst b) Chromatographic estimation of decomposition after photolysis. 195 should effectively eliminate the BET pathway. The decreased solubility of the radical cation compared to the dication could not account for this observation at such low radical concentration. At this low concentration dimerization of MV" is reported not to take place to any significant extent [44]. The reduction of MV" radical to neutral dihydrobipyridyl is known to occur with powerful chemical agents such as zinc or sodium dithionite [57]. However with EDTA alone this reduction is not known to occur. In the absence of oxygen, the reason that only a small fraction of the available MV- is getting reduced to MV" could then only be due to the steady consumption of the radical due to other side reactions. It has been proposed in the literature that the oxidized product of EDTA can convert MV" back to MV-' leading to decreased yield [3.44]. Even if this is the case, then we should see the establishment of some kind of steady state MV * concentration as the EDTA radicals could only be generated with concomitant generation of MV" radical. However the observation that we have made has been that the MV" concentration steadily decreases with increasing photolysis time and the extent of MV-' disappearance estimated from HPLC correlates with the rate of decay of the 600 nm band of MV" radical. This indicates that the MV- is not getting regenerated by the oxidation of MV" by the oxidized products of EDTA but rather is getting irreversibly destroyed. Electrochemically generated MV" radical is known to be stable under acidic, basic and neutral pH in the absence of oxygen. It has been reported that addition of oxygen to the solution, containing MV" radical oxidizes the radical back to the dication form. However it has also been noted by various authors that the methylviologen radical 1 9 6 is photochemically unstable and decomposes even in the absence of any catalyst [3,25b,44], After four hours of photolysis the system containing Ru(bpy) 3 ** (2 x 10“* M), MV-* (1 X lQ-5 M) and Na^EDTA (0.2 M) showed about 70% MV-* decomposition (Figure 3.22b) as compared to a maximum of 30% in the case of photolysis with Pt under hydrogen evolving conditions. This observation is similar to the reported enhanced stability of methylviologen in the presence of hydrogen evolution catalysts [25a]. Addition of catalysts effectively prevents the accumulation of MV" radical thereby reduces the decomposition. This indicates that if the effective MV" radical concentration is kept at an optimum level then the photodecomposition could be eliminated or at least minimized. f) Strategies for reducing MV^ destruction: It has been reported in the literature that it is possible to modify the electron transfer rate from MV" radical to platinum hydrosols by the addition of cations that adsorb strongly on the catalytic surface [58]. In the presence of small amounts of Pb** cations (5 x 10“* M) electron transfer rate was found to increase by three orders of magnitude. MV" radicals did not react with Pb*’ in the absence of Pt hydrosols and the observed effect has been attributed to the modification of the catalytic properties of Pt by Pb’*. However, in this study it was observed that the hydrogen evolution was inhibited by the presence of Pb’* [58]. Ideally it would be preferable to increase the rate of electron transfer from the MV** radicals to the catalyst without inhibiting the catalytic hydrogen evolution. 197 Addition of alkali metal cations as promoters or as selective poisoning agents has been widely practiced in the field of heterogeneous catalysis. Hydrogenation of unsaturated hydrocarbons with zinc oxide has been found to be dependant on the alkali metal dopant [59]. The oxidative dehydrogenation of propane on V^Oj/TiO, (VTi) catalyst has been found to be affected by the alkali metal additives and the rate has been found to decrease in the order VTi>LiVTi>KVTi>RbVTi [60]. Activity of supported ruthenium catalysts for water gas shift reaction [61 ] as well as NHj synthesis [62] was found to be affected by the addition of alkali metal promoters. In our attempts to modify the catalytic properties of RuO, catalysts to enhance the rate of oxidation of MV" radicals and to minimize the hydrogenation without affecting the hydrogen evolution ability, we chose to study the effect of alkali metal cations and chloride salts. Modification of catalytic properties with doped poisoning agents often involve high temperature heat treatment of the involved catalysts. As mentioned above the efficiency of RuOj - zeolite Y catalysts is sensitive to the temperature of heat treatment and hence it was decided to study the commercial powder RuOi catalysts at first. By following literature methods employed for supported ruthenium catalysts for water gas shift reaction [61] and NHj synthesis [62], powder RuO, catalysts were modified by solution impregnation method and calcined at 400°C. Table 3.4 shows the results for one such series of modified catalysts with a poison to ruthenium molar ratio of 1:1. These catalysts showed small variations in the methylviologen life cycle numbers and imfortimately among the different poisons and ratio tried there was no significant effect or trend. 198 Catalyst H? Yield ' MV-" lost MV-* (/imol) (/imol) life cycle RuO, ' 52.73 2.06 25.6 LiNOj-RuO? 55.99 1.98 28.3 NaNGj-RuG? 54.99 1.83 30.1 KNGj-RuG? 50.94 2.13 23.9 CsNGj-RuG?. 49.79 2 . 2 1 22.5 “ 5 mL of photolysis solution (pH 5.1, 0.1 M acetate buffer) containing Ru(bpy)/' (2 x 10"* M), MV- (1 X 10'^ M), NaiEDTA (0.2 M), and catalyst (2 mg). Headspace volume 4 mL. *’ Life cycle number defined as the ratio of moles of hydrogen produced to the moles of methylviologen destroyed. RuO? heat treated to 400°C. Table 3.4 Performance of alkali metal nitrate modified powder RuO? catalysts (poison to ruthenium molar ratio of 1 : 1 ). Colloidal platinum has been reported to be capable of hydrogenating methylviologen [25], whereas with RuO? based catalysts literature reports indicate that hydrogenation does not take place [3.44]. Hence in the case of Pt, the observed MV-* destruction would be a combination of the photochemical destruction and that of the hydrogenation whereas with the RuO? based catalyst it would be mostly the catalyst independent decomposition of the MV“ radical as discussed above. If this is the case, then by optimizing the rate of MV" radical production to equal the rate of the radical oxidation by the catalyst, the radical accumulation can be minimized. This could be achieved by decreasing the methylviologen concentration, increasing the effective catalyst concentration or by producing the radical close to the catalyst surface eliminating the diffusional time delay for the radical to encounter a catalyst particle in order to get 199 oxidized. However reducing the methylviologen to produce smaller amounts of methylviologen radical is not an attractive option as it would reduce the extent of quenching of the MLCT excited state of the sensitizer leading to reduced yield and probably increased photodecomposition of the sensitizer itself. Effective catalyst concentration optimization with RuOj-Y catalysts would involve large changes in the total mass of the catalyst as the Ru content was very low (-3%). This could complicate the system not only due to the different light absorbing efficiencies due to different extent of light scattering but also due to the methylviologen ion-exchange into the zeolites which would decrease the effective solution MV- concentration to different extent. Hence, it was decided to employ RuO? powder for the optimization purpose as we could use the general observations from this catalyst system to others as well. Increasing the amount of RuO, powder added led to the gradual darkening of the photolysis solution and produced different amounts of hydrogen at varying rates as shown in Figure 3.23a. Increasing amounts of RuOi catalyst beyond an optimum concentration (1.5 mg / 5 mL) decreased the rate of hydrogen evolution probably due to the darkening of the solution and hence for a given period of photolysis, lower hydrogen yield. However, with increasing amounts of catalyst, the extent of decomposition steadily decreased and hence the methylviologen effective life cycle increased as shown in Figure 3.23b. It is obvious from Figure 3.23b that it is possible to more than double the methylviologen effective life cycle just by increasing the catalyst content. However as the amount of hydrogen produced decreases for the same light intensity, the efficiency suffers in this approach. 200 a) 1.1 50 1.0 - \ d^6 mg 4 - 40 « ■ □ 4m g % A 0.9 -. AA 2 mg è (A > 1.5 mg « s o ♦ O 1 mg e 0.8 - % *0 # 0 0.5 mg 20 1 0.7 - t ' 2 A L 0.6 J # e • r 4 6 b) -I # ■ 46 - - 36 ■ 43 - ■ - 30 • Hj produced _ A • I ■ # 40 - e ■ Lifecycle - 24 £ > 37 - - 18 ■ ■ • • 1 2 34 , 1------r ■ 1------1------1------r ' - 3 4 Catalyst (mg) Figure 3.23 Effect of RuO^ catalyst mass on a) evolution rate and methylviologen decomposition b) Total amount of hydrogen produced and effective life cycle (moles of produced / moles of lost) in 5 mL of photolysis solution (pH 5.1,0.1 M acetate buffer) containing Ru(bpy)^^* (2 x 10^ M), ( 1 x 10^ M), and EDTA (0.2 M). 201 Another way to minimize the photochemical destruction of the radical is to generate the radical close to the catalyst surface, and our first attempt involved exploiting the ion-exchange ability of zeolite matrix of the RuO^-Y catalysts. It has been shown that MV ' radical exhibits enhanced stability in constrained media like small pore silica gel [63], intercalation compound with titanates [64] and in organic host-guest type complexes with P-cyclodextrin [65]. Initial attempts with low levels of starting methylviologen loading (calculated loading levels: 0.3,0.7 MV-* per supercage) did not show any appreciable change in the methylviologen effective life cycle (35-40) with RuO^-Yjoo- It should be pointed out that the photolysis solutions also contain fairly high concentrations of exchange cations due to NaiEDTA and the buffering agent that could continuously ion-exchange the methylviologen out. However when RuOi-Yiqo was intentionally ion- exchanged with high levels of MV-* (1.4 per SC) the effective life cycle of methylviologen did in fact almost double (80-100). But when this MV- ion exchanged RuOi-Ynoo was employed in a solution of Ru(bpy)3 **, H^EDTA and tetrabutylammonium hydroxide unfortunately there was no measurable hydrogen evolution. This is probably due to the inabilit}' of the totally intrazeolitic methylviologen to efficiently quench the Ru(bpy))-' molecules present in the solution. III. Covalently modified zeolites From our study of the photochemical stability of MV*’ radical and attempts to minimize the methylviologen decomposition during hydrogen evolving photolysis, we realized the importance of generating the radical close to the zeolite surface containing the catalyst. This can be achieved by covalently anchoring either the sensitizer or the 202 electron relay on to the external zeolite surface and as discussed earlier there are plenty of strategies available in the literature for such derivatization. However, anchoring the relay to the surface would severely curtail its mobility making it difficult to reach a catalytic site. On the other hand, if the sensitizer is anchored to the surface of the zeolite, generation of the methylviologen relay could take place close to the catalytic sites as shown schematically in Figure 3.24 and could also access the catalytic sites easily. This approach was attempted with the aim of curtailing the methylviologen decomposition, a) Synthesis Chemical modifications of silicate surfaces involve the reactive surface silanol groups and are often carried out with silyl reagents containing reactive Si-X groups (where X = Cl, OMe or OEt). The surface density of the necessary silanol groups for this derivatization, often referred to as the silanol number, is estimated to be in the range of 1 - 5.5 (OH/nm-) for silica [6 6 ], 3 for ZSM-5 [67] and 2 for high silica faujasite [ 6 8 ]. The surface reaction is proposed to proceed either through hydrolysis of the Si-X bond followed by condensation or direct condensation catalyzed by a nucleophile (NH 3 , EtjN). Covalent linkage of the moiety of interest to the surface has been achieved either by direct reaction or by coupling with previously surface grafted organic moieties that also contain a reactive site (C-Cl, C-NH,). Coupling reactions often yield reduced surface coverage than the direct immobilization [15a,c,69]. Coupling of functional groups to previously anchored moieties often involve either amide and ester linkages that are hydrolytically unstable or ether and amine that are redox-sensitive. The commonly employed derivatizing reagents include trichlorosilanes, triethoxysilanes and 203 Figure 3.24 Schematic representation of the photoinitiated electron transfer processes with RuOi-Y surface anchored Ru(bpy);-\ 204 aminopropyltriethoxysilane (APTS). In the case of APTS, the silylating agent contains both the reactive group and the self catalyzing amino group which could also be used for further coupling. Several reviews have appeared in the recent past dealing with this growing field with applications in the fields of catalysis and separation [66,69]. Silyl reagents are commonly prepared by the nucleophilic substitution reaction of organolithium or Grignard reagents with compoimds containing reactive Si-X (where X = Cl, OR etc.) groups. Halosilanes are more reactive than silylalkoxides and when both groups are present, it is almost always the Si-Cl bond that reacts [70]. Ideally, for our present purpose it would be desirable to use (EtO) 3SiCl as the product (EtO) 3 SiR could be directly reacted with zeolite surface. But due to purity problems, (EtO) 3 SiCl was commercially unavailable at the time of this work and this led us to try other reagents including tetraethoxyorthosilane (TEGS) and silicon tetrachloride. The use of TEGS for the preparation of silyl pyridines [71] and various other alkyl and aryl silyl [72] reagents has been reported. However, Initial attempts with tetraethoxysilane were unsuccessful with very poor yields probably due to the sluggish leaving group (-GEt) and it was decided to try SiCl^ instead. There are a few literature reports detailing the use of the highly reactive SiCl^ for grafting. SiCl^ has been employed with 4 molar equivalents of R-CHi-Li to form (R- CH]-)^Si in THF or Et,G [73]. SiCl^ has also been employed in excess to react with dilithiated diimine to form a stable silylene in THF [74]. Equimiloar quantities of di- Grignard reagent and SiC^ in EtiG were reacted to yield a dichlorosilyl product [75]. 205 Conversion of Si-Cl bond to Si-OR has been accomplished with NaOR and ROH in EtiO [75], trfluoroethanol and EtjN [76] or EtOH and Et^N [77]. Synthesis of the silyl-substituted bipyridine ligand was carried out by a modified literature procedure. Fraser and co-workers have synthesized 4,4'- bis[(trimethylsilylmethyl]-2,2'-bipyridine in high yield (-99%) as an intermediate in their new synthetic route towards the synthesis of halomethyl and other bipyridine derivatives [78]. This trimethylsilyl intermediate was synthesized by deprotonating 4.4'-dimethyl- 2.2’-bipyridine with LDA base resulting in the formation of bpy(CH 2 Li) 2 dianion that was trapped with trimethylsilyl chloride (TMSCl). Though in some cases the product was contaminated with the starting material and/or a yellow impurity and over-silylated products, this reaction in general was found to be quantitative in yield. Because of this reason and the ease of this single step procedure, it was decided to adapt this procedure for our present purposes by employing SiCl., instead of TMSCl as shown in Scheme 3.1 (page 140). The NMR spectra (Figure 3.25) of 4-methyl.4'-triethoxysilylmethyl-2.2'- bipyridine obtained after aqueous work up and solvent removal showed a triplet ( 6 1.24) and a quartet (3.85) corresponding to the methyl and methylene protons of the S 1-0 -CH2 - CHj moiety, a singlet (2.35) due to the methylene protons of the bpy-CH 2-Si moiety and five peaks in the aromatic region confirming the presence of the expected product along with the peaks due to small amounts of starting compound and trace amounts of what appears to be the disubstituted product. Purification of silyl compoimds are often carried out by ordinary distillation under reduced pressure, Kugelrohr (bulb to bulb) distillation, 206 § I ,, 11 JL AJL i _ A L Figure 3.25 NMR spectrum of 4 -methy 1,4-thethoxysilyImethy12,2 -bipyridine recrystallization or chromatography. Attempts to purify the product by distillation were not successful and the products decomposed to yield the starting material 4,4'-dimethyl- 2.2'-bipyridine and TEOS, the hydrolyzed product of SiCl^. Problems associated with thermal decomposition [79] and hydrolysis [75] diuing distillation are known. Our limited experience with the reactive silyl-compounds prevented the purification of the product. It was decided to employ the unseparated reaction products for further zeolite derivatization as the presence of unsilylated starting compound would neither affect the anchoring process nor the complexation step as explained below. Figure 3.26a shows the diffuse reflectance spectra of zeolite surface derivatized bipyridine (CHj-bpy-CHi-Si-zeolite) as obtained and after extensive Soxhlet extraction with ethanol. Soxhlet extraction removed all the non-covalently held ligands and the spectra showed a corresponding decrease in absorbance in the UV region. After three weeks of Soxhlet extraction, the diffuse reflectance spectrum showed almost no further change and also the solvent showed no absorbance characteristic of the bpy ligand and at that point the extraction was assumed to be complete. Corma and co-workers have Soxhlet extracted silica and modified ultrastable zeolite Y after derivatization with chiral metal complexes for only 24 h [18a]. Though it is difficult to completely rule out the presence of non-covalently held bpy ligands, the surface concentration of such species after the three week Soxhlet extraction period compared to that of the covalently held species would be negligible. Any non-covalently held bpy ligands, if at all present in appreciable concentration, would be inside the cages of the zeolites and would not interfere with subsequent complexation step. 208 After Soxhlet extraction the spectra shows three somewhat broad peaks In the UV region with of 209, 242 and 283 nm and are slightly red shifted compared to 4,4'- dimethyl-2,2'-bipyridine in solution. 3,3'-Silyl-2,2'-bipyridine has been reported to show broadened zr-zt* absorption peaks that are red shifted to 294 nm and 303 nm when the silyl substitution were SiMej and SijMe^, respectively in comparison to unmodified 2 ,2 '- bipyridine [23]. The red shift has been attributed to the lowering of LUMO due to the electron withdrawing effect of the silyl substituents [23]. Red shift and broadening of peaks in the absorption spectra on surface immobilization has been observed before in the literature with the 4-Me-bpy-4'-(CHih-Si-Quartz showing an absorption peak at 287 nm [15a]. Such peak broadening and small red shifts upon surface immobilization has been observed for other ligands [15b,c]. Since the silyl substitution in our case is on the side chain rather than on the bipyridine ring, we ascribe the observed peak broadening and small red shift to the surface immobilization. Following literature procedure employed for the synthesis of homogeneous heteroleptic ruthenium polypyridyl complexes, zeolite surface derivation with Ru(bpy) 3 -' was carried out by refluxing RuCbpyl^CK in ethanol with CH^-bpy-CHi-Si-zeolite (Scheme 3.3) [26]. Since the size of the solvated ruthenium species in an ethanol solution. Ru(bpy)i(EtOH)-' (>10 Â), is larger than the zeolite pore diameter (7 Â) only the surface bpy ligands could complex to form the surface anchored Ru(bpy)iL-' (where L = CHj-bpy-CHj-Si-zeolite). Any remaining uncoordinated bpy ligands after Soxhlet extraction that were simply adsorbed inside the zeolite would not be able to react with Ru(bpy)2 (EtOH)2 -" to form intrazeolitic Ru(bpy)2 (4 ,4 '-Me-bpy)-*. Figure 3.26b shows the 209 a) bpy derivatized zeolite Y After Soxhlet extraction g 1.2 - 1 ^ 0.9- I ^ 0.6 - 5I 0.3- 0.0 . 200 300 500 600400 700 Wavelength (nm) b) Ru(bpy), surf. denv. zeolite Y a 0.8 - Ru(bpy), surf. ads. zeolite Y § e s Is 400 500 700 Wavelength (nm) Figure 3.26 Diffuse reflectance spectra of zeolite surface derivatized with a) bipyridine (CHg-bpy CHg-Si-zeolite) as obtained and after extensive Soxhlet extraction with ethanol b) Ru(bpy)jL^^ (L = CH^ bpy- CH^-Si-zeolite) along with surface adsorbed Ru(bpy)j^* for comparison. 210 diffuse reflectance spectra of zeolite surface anchored Ru(bpy)iL-" (where L = CH^-bpy- CHi-Si-zeolite) along with surface adsorbed Rufbpy)]-^ for comparison. The MLCT band of surface anchored Ru(bpy) 2 L‘* was red shifted and has a maximum of 457 nm compared to 450 nm for the surface adsorbed Rufbpy);-'. The intraligand k-k* peak at 289 nm was also slightly red shifted compared to 286 nm. The high energy peaks at 213 and 243 nm were of comparatively greater intensity. Since we have followed a strategy involving modified ligand derivatized support -► anchored complex, there might be some uncomplexed surface held ligand which could give rise to the ligand centered absorption peaks with higher intensity. Brunei and co-workers have derivatized micelle templated silica (MTS) with 3-chIoropropyltrimethoxysilane with a surface coverage of 14.3 X 1 Q-* mol/g. This functionalized silica on treatment with ephedrine gave a surface that contained both ephedrine capped propylsilyl groups ( 6 . 1 x 1 0 "* mol/g) as well as unreacted 3-chloropropylsilyl groups (8.3 x 10"* mol/g) [69]. Only when the silyl substituted ligand containing complexes are employed for derivatization (modified ligand organometallic compound anchored complex), would it be possible to completely eliminate the presence of uncomplexed surface held ligand [18b], but this could result in greatly reduced overall surface coverage. Gafney and co-workers have noted red shifted visible band for Ru( 4 -Xbpy) 3"* compared to Rufbpy)]-' when X was an electron withdrawing group [80]. When X = PEtj’ (n=5) or NO, (n=2 ) red shifted peaks were observed at 466 and 486 nm, respectively and has been explained by the strong electron withdrawing nature of the substituents, confirming the MLCT origin of this band. There have been two reports on 211 ruthenium polypyridine complexes containing silyl-substituted bipyridines. Hosseini and co-workers have synthesized 2 ,2 '-bipyridine ligands interconnected by silane spacers and their binuclear ruthenium complexes [22]. Unfortunately in this report the authors did not include any spectral data other than the fact that the crystals were slightly yellow or orange in color. Homoleptic and heteroleptic ruthenium complexes, RuL;-' or Ru(bpy)2 L-*. where L = 3.3'-silyl-2,2'-bipyridine with SiMej and SiiMe^ groups, were found to show red shifted MLCT and intra-ligand 7t-7t* absorption peaks because of the lowering of LUMO due to the electron withdrawing effect of silyl substituents [23]. Kincaid and co-workers have synthesized zeolite entrapped Ru(bpy) 2(4 ,4 '-Me-bpy) and found broadened absorption bands at 288 and 464 nm [81] compared to that of 287 and 456 nm in solution [82]. As described above the silylmethyl-bipyridine is expected to behave more like methyl-bipyridine than ring substituted silyl-bipyridine and correspondingly the red shifts are not due to the electron withdrawing effects and could only be explained as due to that of surface anchoring effect as evidenced in the literature [15]. RUO2 -Y surface derivation with Ru(bpy)/' was carried out similar to that of zeolite surface modification. Since the previously discussed RUO 2-Y that were prepared with 1 0 % Ru,(C0 ) , 2 by weight were very dark in color any surface derivatization would permit only a small ftaction of the light to be absorbed by the anchored sensitizer. Hence, it was decided to carry out the surface derivatization of a catalyst that was prepared with only 1% Ru3(CO),2 . Figure 3.27a shows the diftuse reflectance spectra of bare RUO 2-Y 200 ( 1 %), bpy surface derivatized RUO 2-Y 2 00 ( 1 %) and surface anchored Ru(bpy) 2 L-* (where 212 L = CHs-bpy-CHi-Si-zeolite-RuOi). The MLCT band was red shifted and has a maximum of 476 nm and the intraligand tc-tti* and were also slightly red shifted to 248 and 291 nm, respectively similar to the case of bare zeolite derivatization. The much larger MLCT red shift is likely due to unreacted Ru(bpy):CL contamination. The bare catalyst RuOi-Y^oo (1%) also has absorption peak in the UV region and gently rising baseline in the visible region as discussed before. Figure 3.27b shows the fluorescence emission spectra of surface derivatized RBY on zeolite Y and on RuO^-Yioq (1%) (note the change in Y axis). The uncorrected emission peak in both cases has a maximum at around 610 nm and is slightly blue shifted compared to Ru(bpy)]-' (618 nm). Fox and co-workers have derivatized colloidal RuOi with surfactant like Zn-porphyrins containing a bipyridine unit [19]. The fluorescence emission of the porphyrin did not show any change in wavelength but was quenched (> 90%) when the coordinating bpy moiety was present and was attributed to intramolecular electron transfer. Simple physical mixing of Zn-porphyrin-bpy with colloidal RuOi did not cause any fluorescence quenching. In our case since the RuOi was preformed on the zeolite surface before the covalent linking of the silyl-bipyridine. there would not be any complexation to the RuOi particles and hence there should not be any fluorescence quenching. However Figure 3.27b shows different emission intensities for the surface derivatized zeolite Y ( 8 x 10*) compared to that of (1%) RuOi-Yiqo (2.5 x 10*). Though there could be some difference in the surface coverage, this would not be large enough to account for the observed difference. However, the optical densities of the two matrices 213 a) 2.5 Ru02*Y2oq (1%) a 2.0 - bpy-RuOj-Yîoo (1%) e Ru(bpy)3^^-Ru02-Y2oo (1%) s 1.5 - E S 1.0 - I 0.5 - 0.0 - 400 500 Wavelength (nm) b) Ru(bpy))^* Zeolite Y I Ru(bpy)3^+-RuO2-Y20Q (1%) - 3 & S 0 0 500 550 600 650 700 750 800 Wavelength (nm) Figure 3.27 a) Diffuse reflectance spectra of bareR uOj-Y jqq (1%), bpy surface derivatized Ru Oj-Y jqo (I%) and surface anchored Rufbpy)^^^ (L = CH^ bpy-CHg-Si-zeolite-RuO^). b) Fluorescence emission spectra of surface anchored Ru(bpy);L^* on zeolite Y and on RuOj-Yjpo (1%) (L = CHj-bpy-CHj-Si-zeolite or CHj-bpy-CHj-Si-zeolite-RuOj). 214 are very different, zeolite is white whereas RuOi-Yjoo ( 1 %) is grey in color, making quantitative comparison of the solids difficult. b)Photocheiiiistry Wrighton and co-workers have found that the silica surface immobilized Re- bipyridyl complex on photoexcitation was capable of undergoing electron transfer quenching with MV-' to yield MV ' [16]. This group has also found that immobilization of Re-bipyridyl complexes on Pt causes the excited-state lifetime to decrease ten fold (35 ns) compared to the free complex in solution (339 ns) due to Pt quenching. Even though the photoelectron could reside in the only bpy ligand that is also covalently bound to the surface, it was still capable of photooxldizing a sacrificial electron donor, triethanolamine present in solution leading to a sustained anodic photocurrent. The photochemical viability of the surface anchored Ru(bpy);-' was tested by photolyzing Ru(bpy) 3 *‘ anchored to zeolite Y in the presence of MV-' and EOTA in solution. If the surface held Ru(bpy);-' is capable of undergoing photoelectron transfer, then we should be able to observe the accumulation of blue methylviologen radical in solution, which could be monitored by absorption spectroscopy. Figure 3.28a shows the UV-Vis absorption spectra obtained during such a photolysis over a period of time carried out in a sealed NMR tube after degassing by three ffeeze-pump-thaw cycles in a vacuum line. After about 15 min of photolysis the generation of MV** radical was evident from the appearance of peaks at 395 nm and 600 nm. The spectra showed no significant absorbance at 450 nm due to the MLCT band of Ru(bpy) 3 -* in solution suggesting that the MV** radical generation is in fact from the surface held Ru(bpy) 3 -*. 215 The ready formation of MV” radical indicate that the surface anchored species is in fact capable of undergoing photoelectron transfer and should be suitable for photogeneration of hydrogen in the presence of the RuOi catalyst. Attempts at using surface derivatized sensitizer or electron relay for photogeneration of hydrogen has been reported before. Wrighton and co-workers have studied electrode materials derivatized with N,N'-substituted-4,4' bipyridine [83]. In the case of derivatized p-Si, shining light with energy greater than the band gap of Si (1.1 eV. -630 nm) resulted in the reduction of surface viologen at potentials about 500 mV more positive than its 5%,^, demonstrating photoelectrocatalysis and the reduced viologen was found to be capable of mediating the reduction of solution redox species as well. When these bipyridine derivatized substrates were deposited with Pd or Pt metal, photoelectrochemical reduction of V-* to V” results in hydrogen evolution with a maximum efficiency of about 5% at pH 4. Photolysis experiment was carried out with Rulbpy);-' derivatized RuO^-Yjoo ( 1%). with MV- and EOTA in solution. During photolysis, formation of small gas bubbles observed for the earlier systems, was not visually noticeable. However, the gas chromatogram after 30 min of photolysis showed the hydrogen peak and the peak grew with photolysis time albeit at a much slower rate compared to the earlier systems. Figure 3.28b shows the plot of H, peak area as a function of time and after 8 hours of photolysis, the total amount o f hydrogen produced was less than a micromole ( - 0 . 8 x 1 0 '* mole). This was only about 2.5 times that of the system containing Rufbpy))- ', MV-* and EOTA in solution in the absence of the RuO^-Y catalyst, where the hydrogen evolution was 216 a) 4 ^ Before 30 min 60 min 3 - 120 min 330 min 390 min 1 2 - < I - 0 300 400 SOO 600 700 Wavelength (nm) b) V 100 200 300 500 Time (min) Figure 3.28 a) Absorption spectra obtained at various time intervals during the photolysis of 12 mg of surface derivatized zeolite Y in 0.5 mL of pH 5 buffer solution containing (5 X 10^ M), EDTA (0.2 M) in a sealed NMR tube under anaerobic conditions, b) Hj evolution during the photolysis of 12 mg of surface derivatized RuO^-Y^ (1%) in pH 5 buffer solution containing (5 x 10 ^ M), EDTA (0.2 M), headspace 1.2 mL. 217 earlier attributed to the impurities present in the EDTA sample. However, it should be pointed out that in this case if Ru(bpy) 3 -* was omitted, i.e., a solution containing only MV-* and EDTA upon photolysis produced no hydrogen. Hence the hydrogen formation observed in the case of surface derivatized RuOi-Yjoo in the absence of added Ru(bpy) 3 ** in solution must be due to the surface anchored complex sensitization. The production of very small amounts of hydrogen is probably due to the small quantities of surface anchored sensitizer as compared to that in the earlier homogeneous solution case. For an accurate comparison, we need a better characterized surface derivatized RuO^-Y sample as well as a suitable method to estimate the actual fraction of the excitation light absorbed by the surface held sensitizer. At present, due to the absence of significant hydrogen production and a suitable blank to compare with, we did not attempt to estimate the MV-' decomposition. As indicated by the earlier UV-Vis studies of the photolyzed solutions in the absence of catalyst as well as with increasing amounts of RuOi catalyst, the MV- decomposition does in fact decrease when the radical was shielded front tight or turned over faster. In light of this the observation of hydrogen evolution the surface anchored case is somewhat promising as it could lead to an integrated system like the one shown in Figure 3.29. In this schematic representation the sensitizer is covalently linked to an acceptor and is held on the surface of a zeolite membrane containing the RuO? that is also loaded with methylviologen. The photoinitiated charge separation from the sensitizer to the acceptor would be followed by electron transfer to the ion-exchanged MV-' which could 218 R uO j Zeolite Membrane (side view) Figure 3.29 Schematic representation of the proposed integrated photosystem. 219 efficiently undergo charge hopping (discussed in the next chapter) leading to charge separation across the membrane. Due to the tight packing of MV-* that is possible in the zeolite cages charge hopping would be very efficient and in this arrangement the viologen radical would be shielded fi’om the photodegradation. Ideally the oxidized sensitizer would oxidize water to oxygen in the presence of nanocrystalline zeolite particles containing RuO, on one side of the membrane and the migrated charge would result in the reduction of water to hydrogen on the other side of the RuO, containing membrane. CONCLUSION RuOi-zeolite Y catalysts, prepared by the thermal decomposition of Ru-lCO),, on zeolite Y, are efficient in catalyzing the reduction of water to hydrogen by photogenerated methylviologen radicals. RuOi-Y catalyst are better than commercial RuOi and colloidal Pt as far as the stability of the methylviologen is concerned. Attempts to selectively poison RuO; catalytic sites responsible for the viologen hydrogenation were not successful. The non-catalytic degradation due to the photoinstability of the methylviologen radical may be a significant decomposition pathway. The effective lifecycle of the electron relay can be increased with efficient radical turnover. Surface anchored Rufbpylj’* on RuOi-Y is capable of sensitizing hydrogen production. However further efforts are needed in order to realize an non-sacrificial. integrated system. 220 REFERENCES 1. Kalyanasundaram, K.; Kiwi, J.; Gratzel, M. Helv. Chim. Acta 1978, 6 1 ,2720. 2. Amouyal, E. Sol. Energy Mater. Sol. Cells 1995, 38, 249. 3. Keller, P.; Moradpour, A.; Amouyal, E. J. Chem. Soc., Faraday Trans. I 1982, 78, 3331. 4. Seddon, E. A.; Seddon, K. R. "The Chemistry of Ruthenium", Elsevier, Amsterdam, 1984. 5. Hepel, T.; Poliak, F. H.; O'Grady, W. E. J. Electrochem. Soc. 1984, 131, 2094. 6. Hiratani, M.; Matsui, Y.; Imagawa, K.; Kimura, S. Thin Solid Films 2000, 366, 102. 7. Das. S. D.; Dutta, P. K. Microporous and Mesoporous Mater. 1998, 22,475. 8. Kohno. M.; Kaneko, T.; Ogura, S.; Sato, K.; Inoue, Y. J. Chem. Soc.. Faraday Trans. 1998, 94, 89. 9. Takata, T.; Tanaka, A.; Hara. M.: Kondo, J. N.; Domen. K. Catal. Today 1998, 44. 17. 10. Rard, J. A. Chem. Rev. 1985, 85. 1. 11. Huang, Y. S.; Park. H. L.; Poliak, F. H.Mat. Res. Bull. 1982, 17. 1305. 12. Barreca, D.; Buchberger, A.; Daolio, S.; Depero, L. E.; Fabrizio, M.; Morandini, F.; Rizzi, G. A.: Sangaletti, L.; Tondello, E. Langmuir 1999, 15.4537. 13. Rizzi, G. A.; Magrin, A.; Granozzi, G. Surface Science 1999, 443. 277. 14. Meng, Li.-Jian.; Dos Santos, M. P. Mat. Res. Soc. Symp. Proc. 1999, 541, 141. 15. a) Paulson. S.; Motfis, K ; Sullivan, B. P.J. Chem. Soc., Chem. Commun. 1992, 1615. b) Liang, Y.; Schmehl, R. H. J. Chem. Soc., Chem. Commun. 1995, 1007. c) Zhou, M.; Laux, J. M.; Edwards, K. D.; Hemminger, J. C.; Hong, B. Chem. Commun. 1997,1977. 16. Mabrouk, P. A.; Wrighton, M. S. Spectrochim. Acta, Part A 1989, 45A, 17. 221 17. Christ, C. S., Jr.; Yu, J.; Zhao, X.; Palmore, G. T. R.; Wrighton, M. S. Inorg. Chem. 1992,3L 4439. 18. a) Corma, A.; Iglesias, M.; Del Pino, C.; Sanchez, F. J. Chem. Soc., Chem. Commun. 1991,18, 1253. b) Corma A.; Iglesias, M.; Del Pino, C.; Sanchez, F.V. Organomet. Chem. 1992, 431, 233. c) Sanchez, F.; Iglesias, M.; Corma A.; Del Pino, C. J. Mol. Catal. 1991, 70,369. 19. Resch, U.; Fox, M. A. J. Phys. Chem. 1991, 9 56316. , 20. Hagfeldt, A.; Gratzel, M. Acc. Chem. Res. 2000,3 3 , 269. 21. Bonhote, P.; Moser, J.-E.; Humphry-Baker, R.; Vlachopoulos, N.; Zakeeruddin. S. M.; Walder, L.; Gratzel, M. J. Am. Chem. Soc. 1999, 121, 1324. 22. Kaes, C.; Hosseini, M. W.; Cian, A. D.; Fischer, J. Chem. Commun. 1997, 2229. 23. Stange, A. F.; Tokura, S.: Kira, M.J. Organomet. Chem. 2000, 612. 117. 24. Rampino, L. D.; Nord, F. F. J. Am. Chem. Soc. 1941, 63, 2745. 25. a) Keller, P.; Moradpour, Amouyal, E.; Kagan, H. B. /Vouv. J. Chim. 1980, 4. 377. b) Johansen, O.; Launikonis, A.; Loder, J. W.; Mau, A. W.-H.; Sasse, W. H. F.; Swift, J. D.; Wells, D. Aust. J. Chem. 1981, 34. 981. 26. Sprintschnik, G.; Sprintschnik, H. W.; Kirsch, P. P.; Whitten, D. G. J. Am. Chem. Soc. 1977. 99, 4947. 27. Gates, B. C.: Guczi, L.; Knozinger, H. “Metal Clusters in Catalysis”, Elsevier, New York, 1986. 28. Lin, L.-H.; Chao, K.-J. Studies in Surface Science and Catalysis 1994, 84. 749. 29. Shen, G.-C.; Liu. .A.. M.; Ichikawa M. Inorg. Chem. 1998, 37, 5497. 30. Zarbin, A. J. G.; Vargas, M. D.; Alves, O. L. J. Mater. Chem. 1999, 9, 519. 31. Robertson, J.; Webb, G. Proc. R. Soc. London, Ser. A 1974, 341. 383. 32. Pearce, J. R.; Mortier, W. J.; Uytterhoeven, J. B. J. Chem. Soc., Faraday Trans. I 1979, 75, 1395. 222 33. a) Bolzan, A. A.; Fong, C.; Kennedy, B. J.; Howard, C. J. Ac/a Crystallogr., Sect. B: Struct. Sci. 1997,353, 373. b) CPDS Card # 3782, Institute of Experimental Mineralogy, Chemogolovka, 1997. http://database.iem.ac.ru/mincryst/s_carta.php37RUTlLE-t-3782 34. Cullity, B. D. “Elements of X-Ray Diffraction”, Addison-Wesley, Reading, 1978. 35. a) Zhao, G. L.; Bagayoko, D.; Williams, T. D. Phys. Rev. B 1999,60, 1563. b) Krasovska, O. V.: Krasovskii, E. E.; Antonov. V. N.Phys. Rev. B 1995, 52, 11825. c) Xu, J. H.; Jarlborg, T.; Freeman, A. J. Phys. Rev. B 1989,40, 7939. d) Mattheiss, L. F. Phys. Rev. B 1976,13, 2433. 36. a) Huang, Y. S.; Liao. P. C. Sol. Energy Mater. Sol. Cells 1998,55, 179. b) Belkind, A.; Orban, Z.; Vossen. J. L.; Woo I lam. J. A. Thin Solid Films 1992.207, 242. c) Canart-Martin, M. C.; Canivez. V.; Declercq, R.; Laude, L. Wautelet, M. Thin Solid Films 1992,92, 323. d) Goel, A. K.; Skorinko. G.; Poliak, F. H.Phys. Rev. B 1981,24, 7432. 37. Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. “Handbook of X-ray Photoelectron Spectroscopy", Physical Electronics Division. Perkin-Elmer. Eden Prairie. 1992. 38. Pedersen, L. A.; Lunsford, J. H. J. Catal. 1980. 61 ,39. 39. a) Blouin, M.; Guay, D. J. Electrochem. Soc. 1997,144, 573. b) Chan, H. V. H.; Takoudis, C. G.; Weaver, M. J. J. Catal. 1997,172, 336. c) Wagner, C. D.; Naumkin, A. V.; Kraut-Vass, A.; Allison, J. W.; Powell, C. J.; Rumble, J. R. J. “NIST X-ray Photoelectron Spectroscopy Database” 2000. http://srdata.nist.gov/xps/ 40. Verdonck, J. J.; Jacobs, P. A.; Genet, M.; Poncelet. G. J. Chem. Soc., Faraday Trans. I 1980, 76,403. 41. a) Twitchett, P. J.; Moffat,_A. C. J. Chromatogr. 1975, III. 149. b) Hodgeson, J. W.; Bashe, W. J.; Eichelberger, J. W. "Determination of diquat and paraquat in drinking water by liquid-solid extraction and high performance liquid chromatography with ultraviolet detection," U.S. Environmental Protection Agency, 1992. 42. Sakai, K.; Kizaki, Y.; Tsubomura, T.; Matsumoto, J.K. Mol. Catal. 1993,7 9 ,141. 223 43. Keller, P.; Moradpour, A. J. Am. Chem. Soc. 1980,102,7193. 44. Kleijn, J. M.; Rouwendal, E.; Van Leeuwen, H. P.; Lyklema, J.J. Photochem. Photobiol., A 1988,44, 29. 45. Amouyal, E. “Development of Catalysts for Water Photoreduction: Improvement, Poisoning and Catalytic Mechanism”, E. Pelizzetti, N. Serpone (Ed), D. Reidel Publishing, Dordrecht, 1986. 46. Bauer. R.; Werner. H. A. F. J. Mol. Catal. 1992.72, 67. 47. Over. H.; Kim. Y. D.; Seitsonen. A. P.; Wendt. S.; Lundgren, E.; Schmid, M.; Varga. P.; Morgante. A.; Ertl. G. Science 2000. 287, 1474. 48. Hepel. T.; Poliak, F. H.; O'Grady, W. E. Proc. - Electrochem. Soc. 1984. 84-12, 512. 49. Kiwi. J.: Gratzel. M. J. Am. Chem. Soc. 1979.lOl, 7214. 50. a) Hostenpiller. P. A. M.; Bolt, J. D.: Fameth. W. E.; Lowekamp. J. B.; Rohrer, G. S. J. Phys. Chem. B 1998.102. 3216. b) Lowekamp, J. B.: Rohrer. G. S.; Hostenpiller. P. A. M.; Bolt. J. D.; Fameth, W. E. J. Phys. Chem. B 1998.102, 7323. 51. Corrent. S.; Cosa, G.; Scaiano. J. C.; Galletero. M. S.; Alvaro. M.; Garcia, H. Chem. Mater. 2001. 13, 715. 52. Zhang, J.; Hu, Y.; Matsuoka, M.; Yamashita, H.; Minagawa, M.; Hidaka. H.: Anpo. M.y. Phys. Chem. B. 2001. 105, 8395. 53. Grabowski, R.; Machej. T.; Mazurkiewicz. A.; Slocynski. J.Bull. Pol. Acad. Sci. Chem. 1987,35, 141. 54. Mills, A.; Giddings, S.; Patel, I.; Lawrence, C. J. Chem. Soc.. Faraday Trans. I1987. 83, 2331. 55. Kleijn. J. M.: Boschloo. G. K. J. Electroanal. Chem. 1991.300, 595. 56. Mills. A.; Giddings. S.; Patel. I. J. Chem. Soc., Faraday Trans. I1987.83. 2317. 57. Carey. J. G.; Cairns, J. F.; Colchester, J. E. J. Chem. Soc. D 1969, 1280. 58. Nenadovic, M. T.; Micic, O. L; Adzic, R. R. J. Chem. Soc., Faraday Trans. I1982, 78, 1065. 224 59. Winiarek, P.; Kijenski, J. J. Chem. Soc. Faraday Trans. 1998, 94, 167. 60. Grabowski, R.; Grzybowska, B.; Kozlowska, A.; Sloczynski, J.; Wcislo, K.Top. Catal. 1996, i. 111. 61. Basinska, A.; Domka, F.Catal. Lett. 1997, 42, 59. 62. Aika, K; Kawahara, T.; Murata, S.; Onishi, T. Bull. Chem. Soc. Jpn. 1990, 6 3 ,1221. 63. Xiang, B.; Kevan, L.J. Phys. Chem. 1994, 98, 5120. 64. Miyata. H.; Sugahara, Y.; Kuroda. K.; Kate, C.J. Chem. Soc.. Faraday Trans. 1988, 84, 2677. 65. Petr, A.; Dunsch, L.; Koradecki, D.: Kutner, W.J. Electroanal. Chem. Interfacial Electrochem. \99\, 300, 129. 66. Impens, N. R. E. N.; Van der Voort, P.; Vansant, E. F. Microporous and Mesoporous Mater. 1999,25,217. 67. Vankelecom, 1. F. J.; Van den Broeck, S.; Merckx, E.; Geerts, H.; Grobet, P.; Uytterhoeven. J. B. J. Phys. Chem. 1996. 100, 3753. 68. Kawai, T.; Tsutsumi, K. J. Colloid Int. Sci. 1999, 2/2, 310. 69. Brunei, D.; Bellocq, N.; Sutra, P.; Cauvel, A.; Lasperas, M.; Moreau, P.; Di Renzo, F.; Galameau, A.; Fajula, F.Coord. Chem. Rev. 1998, 178-180, 1085. 70. Eabom, C. “Organosilicon Compounds", Academic Press, New York, 1960. 71. Riedmiller, F.; Jockisch, A.: Schmidbaur, H. Organometallics 1998. 17, 4444. 72. Tour, J. M.; John, J. A.; Stephens, E. B. J. Organomet. Chem. 1992, 429, 301. 73. Strohmarm, C.; Ludtke, S.; Ulbrich, O. Organometallics 2000, 19,4223. 74. Haaf. M.; Schmiedl, A.; Schmedake, T. A.; Powell, D. R.; Millevolte, A. J.; Denk, M.; West, R.7. Am. Chem. Soc. 1998, 120, 12714. 75. Cremer, S. E.; Blankenship, C. Organometallics 1986, 5, 1329. 76. Perrin, C. L.; Kim, Y.-J. Inorg. Chem. 2000,39, 3902. 225 77. Matsumoto, Y.; Ohno, A.; Hayashi, T. 1993, 12,4051. 78. Fraser, C. L.; Anastasi, N. R; Lamba, J. J. S. J. Org. Chem. 1997, <52,9314. 79. Rubinsztajn, S.; Zeldin, M.; Fife, W. K. Synth. React. Inorg. Met. -Org Chem. 1990, 20,495. 80. Basu, A.; Weiner, M. A.; Strekas, T. C.; Gafney, H. D. Inorg. Chem. 1982, 21. 1085. 81. a) Maruszewski, K.; Strommen, D. P.; Handrich, K.; Kincaid. J. R. Inorg. Chem. 1991, 30, 4579. b) Maruszewski, K.; Strommen, D. P.; Kincaid, J. R.J. Am. Chem. Soc. 1993, 115, 8345. 82. McClanahan, S. F.; Dallinger, R. F.; Holler, F. J.; Kincaid, J. R. J. Am. Chem. Soc. 1985, 107. 4853. 83. Bruce, J. A.; Wrighton, M. S. Isr. J. Chem. 1982, 22, 184. 226 CHAPTER 4 KINETIC ANALYSIS AND MODELING OF INTRAZEOLITIC PHOTOINDUCED CHARGE SEPARATION INTRODUCTION Photoinduced charge separation in microheterogeneous systems has been extensively studied for potential long-lived charge separation [I]. Aluminosilicate zeolites with unique arrangement of cages and channels lend themselves well as hosts for such studies, and can be exploited via intrazeolitic assembly of photoactive systems [2]. Tris-bipyridine ruthenium (11) as a sensitizer and bipyridinium ions as electron acceptors have been extensively investigated in homogeneous solution [3]. In homogeneous solution, the energy wasting back electron transfer is an order of magnitude faster than that of forward electron transfer reaction resulting in no long term charge separation. This paradigmatic system can be used to explore the role of zeolitic hosts in photochemical charge separation. Previous studies on the steady-state photochemistry of intrazeolitic tris-bipyridine ruthenium (11) and viologens have shown the possibility of long-term charge separation [4]. There are three reports on the use of time-resolved diffuse reflectance (TRDR) 227 spectroscopy to study the kinetics of the photoelectron process in this system [5]. However, two of those studies by Mailouk and co-workers involved Ru(bpy)3 -" moiety only on the surface of the zeolite with intrazeolitic methylviologen [5a,b], In these studies, the back electron transfer rate constant on the zeolite surface was found to be considerably slower than for comparable systems in solution. Dutta and Turbeville studied the totally intrazeolitic Ru(bpy) 3 -*-methylviologen system by monitoring the transient viologen radical signal decay at 610 nm [5c]. We have investigated the effect of reaction driving force and the electron acceptor loading level on the intrazeolitic photoelectron transfer kinetics of tris-bipyridine ruthenium (II). For this purpose, electron acceptors methylviologen and a series of 2.2' bipyridinium ions with varying reduction potentials were employed at loading levels ranging from 1/15 to 1.7/1 (acceptor molecules/supercage). The various reactions involved in the intrazeolitic Ru(bpy) 3 --bipyridinium (BP) system are as shown below in its simplest form: Ru(bpy)3 -‘ —— > ♦Ru(bpy)3 *' (4.1) ♦Ru(bpy)3 ^* > Ru(bpy) 3 -' (4.2) *Ru(bpy) 3 -‘ + BP- (next cage) —^ ------> Ru(bpy)3 ' ‘ + BP" (next cage)(4.3) Ru(bpy)3 ^* + BP"(next cage) — — > Ru(bpy) 3 -* + BP- (next cage)(4.4) TRDR spectroscopy can be used to study the kinetic behavior of this system by exploiting the differences in the electronic spectra of groimd state and transient state of the species involved. In the absence of any bipyridinium quenchers, the difference in the 228 diffuse reflectance between the ground state Rufbpylj*" and transient excited state, *Ru(bpy) 3 ** can be exploited to study the kinetics of the excited state decay process and lifetime. The photoelectron transfer process can be studied by the introduction of bipyridinium ions (BP-") via ion-exchange into the Rufbpyl^-'-zeolite Y sample and monitoring the TRDR spectra of the transient BP*’ species. The back electron transfer (equation 4.4) in the zeolite from the bipyridinium radical to Rufbpy),^* was smaller by several orders of magnitude compared to the forward reaction (equation 4.3) suggesting a possible role for the zeolite structure. The dynamics of the electron transfer was found to be dependent on the loading levels of bipyridinium ions in the zeolite. The rate of back electron transfer at low loadings of the bipyridinium ions, decreases with increasing driving force, as expected for reactions occurring in the Marcus inverted region. In the case of high loading for all the electron acceptors studied, transient diffuse reflectance spectra recorded at wavelengths between 340 - 460 nm show up to 30% of the bipyridinium radical signal even after 1 ms. In the previous studies [5], there has been no attempt to kinetically model the long term charge separation and we have attempted to remedy this situation and this chapter describes our modeling effort. The long lived radical signal required a detailed analysis that led to a kinetic scheme more complicated than that has been previously proposed by Dutta and Turbeville [5c]. Migration of the photogenerated electron via hopping among adjacent bipyridinium molecules away from the ruthenium center has been proposed to explain the long-lived charge separation. A compartmental model was developed using STELLA, a commercial software package, to mimic the charge hopping process. 229 Simulation of the proposed model agreed closely with the experimental data enabling the extraction of the rate constants. Molecular modeling of this system offered an opportunity to visualize some of the intrazeolitic processes and was found to support the proposed model. Modeling of the kinetic data was done using STELLA (Structural Thinking Experimental Learning Laboratory with Animation) Research, a system dynamics modeling software. The use of STELLA as a simulation software has been critically reviewed and found to perform well with notable ease of use [6]. STELLA has been used to teach chemical kinetics, simulating the concentration profile of a species undergoing a simple second order reaction as well as model complex oscillatory systems like Belousov-Zhabotinskii reaction [7]. Numerical simulation using STELLA has been used to understand and predict diverse phenomena [8], from the reaction mechanism of the activity control of cytochrome c oxidase [8a] to the phrmacokinetics of dermally exposed volatile organic chemicals [8b]. STELLA has found applications in various fields including biochemistry [8c], pharmacology [8d] and environmental science [8e]. The extensive use of STELLA in ecological studies is apparent from the recent special issue in the journal. 'Ecological Modelling", on modeling ecological systems with STELLA [9]. The use of STELLA continues to draw attention in teaching kinetics and thermodynamics and is evident from a recent ACS conference [10]. EXPERIMENTAL SECTION Synthesis of intrazeolitic Ru(bpy) 3 ** and loading of bipyridinium ions by ion- exchange onto zeolite Y were carried out by Dr. Norma Castagnola [11]. Acquisition of 230 the time-resolved diffuse reflectance (TRDR) measurements were carried out by Dr. Nancy Ortins and Dr. Marcello Vitale using the second harmonic (532 nm) of a NdrYAG Q-switched laser (Quantel) excitation. The pulsed laser source had a temporal width of 15 ns with a pulse power of 21 -24 mW and was repeated at a frequency of 1 Hz. A pulsed xenon lamp (250 W. Applied Photophysics) was used as the probe and the diffusely reflected light was collected and focused into a spectrometer (150 gr/mm gratings. SPEX 1877 Triplemate) and detected with a photodiode array (EG&G Princeton Applied Research). Timing of the laser pulse, lamp pulse, time delay, and detection window ( 100 ns) was done with a EG&G high voltage pulse generator and a home made timing box for triggering the Q-switch and lamp pulse. TRDR spectra of the rigorously deaerated samples in a sealed NMR tube were acquired at 100 ns after sample excitation and at various time delays up to 10 ms. Modeling of the kinetic data was done using STELLA Research (v 4.0.2) from High Performance Systems. Hanover, NH. The finite difference equations were solved with a time interval of 3.2 x 10'^ s using fourth-order Runge-Kutta algorithm on a personal computer with windows operating system. STELLA models were simulated by manual iteration with a range of values using the "sensitivity specs' option of the run controller. The simulation results were compared with the experimental data visually for the best fits, since STELLA is not equipped with fitting capability. Molecular models were generated using previously published crystallographic data [12] for zeolite Y [12a]. Ru(bpy)/' [12b], and methylviologen [12c] on an IBM RISC System/6000 computer on AIX platform, with the program Cerius- (release 3), from Molecular Simulations Inc. 231 RESULTS AND DISCUSSION For the sake of clarity this section is divided into four subsections. The first subsection deals with the Ru(bpy) 3 -" alone encapsulated in zeolite Y, with a discussion on the molecular modeling and the *Ru(bpy) 3 -' excited state decay kinetic analysis. The second subsection deals with the *Ru(bpy) 3 -' quenching studies with bipyridinium electron acceptors with a discussion on the forward electron transfer rate. The third subsection deals with back electron transfer (BET) from bipyridinium radical ion to Ru(bpy)3 ^* in the zeolite leading to the development of a new kinetic model. This subsection also includes a discussion on factors like the loading level and redox potential of the bipyridinium ions that affect BET rate. The final subsection deals with the role of the zeolite structure in promoting charge separation in high loading bipyridinium samples and estimation of transient quantum yields and supercage escape yields. Ru(bpy)3^-zeolite Y Figure 4.1 shows a model of RuCbpylj-' encapsulated in a supercage based on crystallographic data for zeolite Y and Ru(bpy) 3 -* [12a,b]. When the interatomic distance between an atom of the encapsulated Rulbpy);-' and a zeolite framework atom is less than 90% of the sum of van der Waals radii, the energy of the interaction becomes unfavorable and can be designated as a bad' contact [13]. The modeling program provides the interatomic distances between the atoms of Ru(bpy) 3 ** molecule and the surrounding zeolite framework. A large number of orientations of Ru(bpy)3 ‘" in a supercage is obviously possible. For every orientation of Ru(bpy) 3 *' that was examined, there were several such bad contacts, usually between the H and C atoms of Ru(bpy) 3 ** 232 Figure 4.1 Molecular model of Ru(bpy)3 *^ encapsulated in a supercage of zeolite Y, dotted line indicates ‘bad contact". 233 and the O and in some cases Si atoms of the zeolite framework. For the orientation shown in Figure 4.1, the number of bad contacts is 14 (interatomic distances varying from 1.95 to 3.30 Â). Of course, the manual search that we perform does not exhaust all the possible orientations. The presence of these bad contacts suggests that though it is possible to synthesize Ru(bpy)]-' within a supercage, the rotational mobility of the Ru(bpy)/" will be restricted in the zeolite. For example, for each rotation of the Rufbpy),-* in Figure 4.1 by a 5° angle about the x, y or z axis, we find that the number of bad contacts remain around 15 or so, though from different parts of the molecule. What is unclear from the present analysis is the excess energy over that available at room temperature that will be required to rotate Ru(bpy);-' within the supercage and the dynamics of such rotation. In the absence of such energetic information, comparison of Ru(bpy);-' encapsulated in a zeolite Y supercage with that of other zeolites can provide some useful information. Structure 6 (EMT) is a zeolite framework structurally similar to zeolite Y but has two kinds of cavities that differ in their sizes. Figure 4.2 shows Ru(bpy)r' encapsulated in a hypercage, larger of the two kinds of cages in EMT with very few bad' contacts. Manual rotation of Rufbpy),-' in this case did not increase the number of ‘bad’ contacts and there were a few orientations with no bad' contacts. Presently effort is imderway in our laboratory to study the rotational mobility of encapsulated Rufbpy);-' in zeolitic cages by temperature dependent '^C CP-MAS NMR and preliminary results correlate well with the present prediction. 234 Figure 4.2 Molecular model of RuCbpy),-" encapsulated in a hypercage of EMT. 235 The loading level employed for this study is 1 RuCbpy),-* in 15 supercages. If the distribution of 1 Ru(bpy)3 ‘^ in 15 supercages is random as expected, then 1 in 4 Ru(bpy)3 ^^ has another Ru(bpy) 3 '’^ in a neighboring supercage and three empty supercages, whereas for the rest of the Ru(bpy) 3 '* molecules, all the neighboring supercages are empty [14]. At a loading level of 1 Ru(bpy) 3 *" in 15 supercages according to this probabilistic distribution there would not be any significant self-quenching while at the same time the signal would be sufficiently intense enough to be studied. The primary reaction of the photoactive assembly involves the laser irradiation (532 nm) of the zeolite encapsulated Ru(bpy) 3 '* with a 15 ns pulse resulting in the population of the metal to ligand charge-transfer (MLCT) excited state. *Ru(bpy) 3 -' (equation 4.1 ). The normalized transient diffuse reflectance difference spectra of intrazeolitic Ru(bpy)3 -' obtained 100 ns after the laser excitation shows intensity depletion at 450 nm and concomitant increase at 370 nm as shown in Figure 4.3a. In the TRDR spectra the band at 450 nm is due to the depletion of the MLCT band of the ground state Ru(bpy) 3 -' and the band at 370 nm that resembles that of the 2,2'-bipyridine radical anion is due to the ligand-centered transition of the freshly populated *Ru(bpy) 3 " . similar to that of *Ru(bpy) 3 -" in solution [15]. This excitation process is sufficiently fast and the signal maximizes within the time duration of the smallest detection window (100 ns). However, the decay of this signal can be monitored with ease by recording TRDR spectra at various time delays after the excitation pulse and would be a measure of the lifetime of the photoexcited state. The decay data at 370 nm can be fitted to a single exponential decay with a rate constant of 1.5 x 10* s ' as shown in Figure 4.3b indicating 2 3 6 a) 0 360 380 400 420 440 460 Wavelength (nm) b) 1.0 0.8 0.6 0.4 0.2 • • 0.0 0 le-6 2e-6 3e-6 4e-6 Time (sec) Figure 4.3 a) Normalized TRDR difference spectra of intrazeolitic Ru(bpy)^'^ obtained 100 ns after the laser excitation pulse, b) The transient TRDR decay data at 370 nm fitted to a single-exponential decay (rate constant = 1.5 x 10^ s '). 237 the homogeneous distribution of Ru(bpy)]-\ The corresponding excited state lifetime is 670 ns indicates the absence of any self-quenching and is comparable to that of 620 ns observed by Mailouk and coworkers [5a]. R u(bpy)/" quenching with bipyridinium electron acceptors The electron transfer quenching of Ru(bpy)]-' by several bipyridinium ions in solution has been examined by others, and the quenching rate constants are found to increase as the bipyridinium reduction potentials become less negative [16]. Similar results were also reported for covalently held Ru(bpy -viologen assemblies, as the viologen reduction potential was altered [17]. The bipyridinium electron acceptors examined in this study are shown in Table 4.1, along with their structures and the reduction potentials. Three of these acceptors are based on the 2.2' bipyridinium moiety and their reduction potentials are controlled by the length of the connecting methylene chain units. It has been established that as the (CH^ln chain bridging the N.N’ atoms increases in length, the two pyridinium rings are distorted from planarity [18]. In the reduced form of the bipyridinium ions, the rings are planar. Thus the increased distortion with increase in CHi units leads to increasingly negative reduction potentials. These 2,2’ bipyridinium ions belong to the diquat family and are abbreviated in this study as nDQ- . The fourth viologen is the 1,T dimethyl-4,4' bipyridinium (methylviologen, MV-*). The oxidative quenching of intrazeolitic photoexcited Ru(bpy)/' occurs by electron transfer to bipyridinium ions (reaction 4.3). As the bipyridinium ions are doubly charged, they can be ion-exchanged into the Ru(bpy)]-'-zeolite Y. It is obvious from the model of Ru(bpy)j-* in a supercage (Figure 4.1a) that there is not enough room to fit a 238 STRUCTURE ABBREV. E°,:(V)“ AE“ (V)» 2DQ-' -0.37 1.63 ^N—y MV" -0.44 1.70 CK) 3DQ- -0.55 1.81 ChO 4DQ-- -0.65 1.91 “Volts vs. NHE (ref. 33). *’ The value E°,,i (Ru“*/Ru*‘) = 1.26 V (vs. NHE) was used in this calculation. Table 4.1 2,2'- and 4,4'-bipyridinium ions used as electron acceptors. 239 bipyridinium ion in the same supercage, and these molecules must be in supercages not occupied by Ru(bpy) 3 -* molecules. Upon ion-exchanging bipyridinium ions into the zeolite, they fill the empty supercages as shown in Figure 4.4. At low loading levels of 1 bipyridinium ion in 10-15 supercages, the probability of more than one bipyridinium ion in a supercage is low (< 1%). Figure 4.4b shows a MV-' ion in a cage neighboring that of Ru(bpy)3 -' and it appears that there is considerable room for MV-' to move around in the supercage. Crystallographic data was used to generate the MV- structure [12cj. Figure 4.4c shows the packing of methylviologen under high loading conditions, which corresponds to little over three methylviologen molecules in two cages. This is obviously one of the many arrangements that are possible. The figure is meant to reflect the crowding of the MV- in the supercages and the close approach of the molecules upon high loading. In the case of the diquats, the loadings were slightly lower though comparable (Table 4.2). In the high loading case, the Ru(bpy) 3 -' is surrounded by the bipyridinium ions in all four neighboring supercages, and there is less mobility of the bipyridinium molecules as compared to the low loading samples. In the zeolite, the diffusional contribution to electron transfer quenching is expected to be small. The bipyridinium ions will need to diffuse through die 7  windows, and considering that the lateral width of the smallest of these molecules is 6.3 Â, intercage diffusion is likely to be slow. In the low loading samples, the bipyridinium ion has some mobility within its supercage, whereas, in the high loading samples, the bipyridinium ions are considerably more constrained, since the whole zeolite crystallite is packed with the bipyridinium ions. Because of these diffusional limitations, 240 a ) b) c) Figure 4.4 Molecular models of (a) Ru(bpy)j-‘ - zeolite Y with (b) low loading of MV-* and (c) high loading of MV*". 241 Bipyridinium Low Loading * High Loading ® 2DQ- 1/15 1.6/1 MV-* 1/15 1.7/1 3 DQ-* 1/10 1.4/1 4DQ-* 1/10 1.2/1 Loading indicates bipyridinium molecules per supercage. Table 4.2 Bipyridinium loading levels in the Bipyridinium - Rulbpy);-' - Zeolite Y Systems. the quenching of intrazeolitic Ru(bpy) 3 -'* is expected to be more of a static process, with photochemical electron transfer occurring between Rulbpy))- ' in a supercage with the bipyridinium molecules in only the neighboring supercages. Kim and Mailouk have reported that if the Ru(bpy) 3 -' is ion-exchanged on the zeolite surface, and MV- is in the zeolite, then the quenching process is primarily dynamic [5a]. Thus, constraining the Ru(bpy);-' inside the zeolite has a potential effect on the nature of the quenching process. For the high loading bipyridinium systems, the signal for the bipyridinium ion radicals maximized within the first 100 ns window of measurement (minimum time for which the detector gate is left open). Though the exact rate constants could not be determined due to this limited time resolution of the instrument, a lower limit 10^ s ' can be estimated for the forward electron transfer indicating efficient oxidative quenching. There are several reasons for the effective quenching. The cage structure of the zeolite and the packing of the bipyridinium ions limits the distance between the Ru center and the bipyridinium ions to a narrow range. Edge to edge distances between the bpy ligand 242 in a zeolite supercage and the bipyridinium molecule in a neighboring supercage can vary from as close as van der Waals contact (Figure 4.4) to about 6 A, as the bipyridinium ion moves towards the other edge of the supercage. For covalently linked Ru(bpy)3 ‘"-viologen systems, the forward electron transfer rate constant over a 6 A range varies from 5.9 x 10'“ to 6.5 x 10^ s ' [19]. Typically, if these covalently bound complexes are ion-exchanged onto the surface of zeolites or colloidal aluminosilicate particles, the forward electron transfer is slowed down [5b,20]. The slowing down of the electron transfer on adsorption has been proposed as due to restricted conformational arrangements [5b]. In the case of intrazeolitic quenching, the influence of conformational limitations is going to be less significant since each Ru(bpy)3 -* is surrounded by bipyridinium ions on four sides as compared to only one edge for covalently held complexes. The four supercages surrounding the entrapped Ru(bpy)3 -‘ on average contain between 4.8 - 6.8 molecules of bipyridinium ions. The intrazeolitic synthesis of Ru(bpy) 3 -' occurs in a stepwise manner with the final step being the ligation of the bis complex, Ru(bpy) 2 L2 "' with a third bpy molecule that must approach from a neighboring supercage. Once the tris complex, Rufbpylj-" is formed, it is almost immobilized (Figure 4.1a) and hence the last bpy ligand must face the window through which it migrated. In modeling rotational arrangements of Ru(bpy) 3 -' in the supercage, in fact two of the bpy ligands are always found to be positioned facing two of the four 7 A cage windows leading to an asymmetric environment around Ru(bpy) 3 ‘* in zeolite Y. It is generally agreed that in the MLCT state, the electron is localized on the bipyridine ligand [21], and it is likely that it would be localized on a bpy ligand facing the 243 window rather than on a ligand facing the negatively charged zeolite surface for electrostatic reasons. Based on these factors, the fast quenching process inside the zeolite is understandable as the forward electron transfer can occur from the bpy * to a bipyridinium ion readily across the 7  window. Back electron transfer from bipyridinium radical ion to Rufbpy)]^ in the zeolite The rate of back electron transfer increased with the loading level. We present the kinetic data analysis with two extreme loadings of bipyridinium ions, a low loading of I in 15 or 1 in 10 supercages, and a high loading of 1.2 -1.7 molecules per supercage. As the complexity of the bipyridinium radical decay kinetics increases with the loading, we shall first discuss the simpler low loading case and the high loading case later. a) Low viologen loadings Figure 4.5 compares the normalized transient diffuse reflectance difference spectra 100 ns after the laser pulse of Rufbpy),'-zeolite Y with the methylviologen loaded sample (1 in 15 supercages). This low loading results in only a fraction of the Ru(bpy)3 -* molecules in the zeolite being surrounded by bipyridinium ions. For those Rufbpy);-' molecules that have a bipyridinium ion in a neighboring supercage, oxidative quenching as shown in reaction 4.3 can occur resulting in MV ’. However, since not all of the Ru(bpy);-' molecules have a bipyridinium ion in the next supercage, the quenching of the Rufbpy))-'" is only partial. Accordingly Figure 4.5 shows a band at 370 nm due to the unquenched Rufbpy);-'" along with a shoulder around 389 nm due to the MV ' [22]. The signal due to MV ' significantly overlaps with that of the unquenched Rufbpy);-' and decays with time as a result of the back electron transfer to Ru(bpy) 3 ^* 244 0 0 360 380400 420 440 460 Wavelength (nm) Figure 4.5 Comparison of the normalized TRDR difference spectra obtained 100 ns after the laser pulse excitation of a) Ru(bpy)^'- zeolite Y with b) the low loading MV' - Ru(bpy)^'^- zeolite Y (1 MV'* in 15 supercages). 245 (reaction 4.4). In order to measure the rate of the back electron transfer, we fitted the decay at two wavelengths around the observed maximum of the absorption band of MV*’ (387,389 nm) to a sum of two exponentials. One of the two exponentials has the decay rate constant of 1.5 x 10* s ', characteristic of the Ru(bpy) 3 -*' molecule isolated from interaction with bipyridinium ions, and was held constant for the fit. The other exponential provided the measure of the rate of the back electron transfer reaction. Figure 4.6a shows the decay of the intensity at 389 nm and the corresponding fit by a double exponential function. The back electron transfer rate constant ky" (11-low loading) obtained for the two wavelengths was averaged and found to be 1.7 x iO'* s '. In a similar fashion k^" for 2DQ-* at a loading level of 1 in 15 supercages was determined to be 4.0 x 10* s ' by monitoring the decay at wavelengths of 374, 376 and 379 nm and Figure 4.6b shows the decay at 379 nm. For 3- and 4-DQ^. the loading levels of 1 in 15 supercages was insufficient to create enough bipyridinium radical to noticeably perturb the transient diffuse reflectance spectra and it resembled that of Ru(bpy) 3 -*’-zeolite. For these two bipyridinium ions, the loading level was increased to 1 in 10 supercages and Figure 4.7 shows the TRDR signal decay for 3- and 4-DQ-* at 383 nm and 377 nm. respectively. At this loading level, back electron transfer rate constants of 1.1 x 10"* and 7.3 x 10^ s*' were found for 3- and 4-DQ’* respectively, by monitoring wavelengths at 379,381 and 383 nm for 3DQ-* and 374, 376 and 379 nm for 4DQ-*. These values are listed in Table 4.3. 246 a) 0.8 0.6 • ## I 0.4 0.2 0.0 2e-6 4e-6 6e-6 8e-6 le-5 Time (sec) b) 0.8 0.6 I 0.4 0.2 0.0 2e-6 4e-6 6e-6 8e-6 le-5 Time (sec) Figure 4.6 The transient TRDR data showing the decay of the bipyridinium radical signal at a low loading level of 1 bipyridinium in 15 supercages and the corresponding fit by a double exponential function for a) at 389 nm, b) 2DQ^ at 379 nm. 247 a) 0.8 O 0.6 I 0.4 ## 0.2 0.0 0 2e-6 4e-6 8e-6 le-56e-6 Time (sec) b) 0.8 0.6 I 0.4 0.2 0.0 0 2e-6 4e-6 6e-6 8e-6 le-5 Time (sec) Figure 4.7 The transient TRDR data showing the decay of the bipyridinium radical signal at a low loading level of 1 bipyridinium in 10 supercages and the corresponding fit by a double exponential function for a) 3DQ‘* at 383 nm, b) 4DQ^^ at 377 nm. 248 Bipyridinium Loading ® kb" (sec') 2DQ-" 1/15 4.0 X 10" MV-* 1/15 1.7 X 10" 3 DQ:" 1/10 1.1 X 10" 4DQ- 1/10 7.3 X 10^ ' Loading indicates bipyridinium molecules per supercage, " low loading. Table 4.3 Back electron transfer rate constants in the low loading Bipyridinium - Ru(bpy);-' - Zeolite Y systems. b) High loading levels The highest loading levels obtained for 2DQ- . MV-'. 3DQ-' and 4DQ-' were 1.6. 1.7. 1.4 and 1.2 molecules per supercage. Transient diffuse reflectance data showed efficient quenching in all these cases. As an example. Figure 4.8b shows the normalized diffuse reflectance difference spectra 100 ns after the laser pulse for a high loading Ru(bpy)]-'-MV-'-zeolite Y. Comparison to Figure 4.8a shows that the peak maximum has changed from 370 nm to 390 nm. This arises because the Rufbpy))- ' is completely quenched by MV-' and the signal present is due to the MV* radical absorption band at 390 nm [22]. Similar results were also observed with the high loading nDQ-* samples. The dynamics of the back electron transfer reaction (reaction 4.2) could be readily obtained by monitoring the decay of the signal due to the bipyridinium radical as a function of time. Figure 4.9a shows some representative time-resolved diffuse reflectance (TRDR) spectra of the high loading Ru(bpy)]-'-MV-' (1.7 molecule per supercage)- zeolite in the wavelength range of 340 - 460 nm. The points in Figure 4.9b 249 0 0 360 380 400 420 440 460 Wavelength (nm) Figure 4.8 Comparison of the normalized TRDR difference spectra obtained 100 ns after the laser excitation pulse of a) Ru(bpy)^"^ - zeolite Y with b) the high loading MV‘^ - Ru(bpy),"’^ - zeolite Y (1.7 per supercage). 250 are the intensity of the MV ' signal at 389 nm as a function of time. In the high loading case, the bipyridinium radical signal was longer-lived by several orders of magnitude compared to the low loading samples, and does not follow a monoexponential decay. This observation was consistent for all four electron acceptors studied and required the development of a new model to explain the complex kinetics. Stella Model One common type of mathematical model that can be used to describe a chemical system is the compartmental model. In such a model, chemical species merely change state according to their concentration and the rate of the process. STELLA can be used to numerically simulate and predict time-dependent behavior of complex phenomena by constructing compartmental models [23]. The easiest way to describe the STELLA modeling process is to illustrate with a simple example of dermal absorption of a drug and its subsequent excretion, both assumed to be first order processes depending only on the drug concentration [6b]. In terms of compartmental modeling this reaction could be represented as shown in Figure 4.10a where the change can be represented as the drug molecule moving state from skin compartment (S) into the blood compartment (B) with a transport coefficient of k,. followed by the excretion with a transport coefficient of k,. Without a simulation software this system represented by equation 4.5 can be solved by writing the transport equations for each compartment. If the initial concentration of drug (X) in S and B are and Xg then a system of coupled differential equations results: 251 a ) 10 msec 2.0 msec 0.5 msec 0.1 msec 20psec 2.5 ^isec 0.5 iisec 0.1 usee T------r 420 440 460 Wavelength (nm) b) 1.0 0.8 0.6 0.4 0.2 0.0 0.0 5.0e-3 l.Oe-2 Time (sec) Figure 4.9 a) Representative TRDR spectra of the high loading MV^*-Ru(bpy) 2 ^*-zeolite (1.7 per supercage) at various time delays after the laser excitation, b) The transient TRDR data at 389 nm showing the decay of the methylviologen radical signal with time. 252 K K S -* B -* E (4.5) dXs/dt = - k.X^ (4.6) dXg/dt = - k^B (4.7) The STELLA model corresponding to this compartmental model is shown in Figure 4.1 Ob. In the modeling environment of STELLA, the rectangular block that represents a compartment is known as a ’stock’ and can be initiated to any starting concentration. The valve with pipes connecting the stocks represent the flux and is known as a ’flow’. The flows’ modulate the rate of change of state and is controlled by a ■flow rule’ employed that is equivalent to the rate expression in kinetics. The open circle in the model is a multipurpose component known as a converter’ and can be used for parameters that change from one simulation to another, for e.g., the rate constant. Another common application for the converter is to evaluate mathematical functions involving one or more variables in the model, for example to calculate the total amount of a species present in various compartments. Connection between components in a STELLA model are made with curved arrows called ‘connectors’. The cloud shaped icon in this model indicates an infinite sink and is used to drain species that are of no consequence to the model after reaching that state. In the model shown in Figure 4.10b, drug flows from the skin compartment into the blood compartment and finally gets excreted, according to the flow rule established by the connectors between the stock, converter and the flows. 253 a ) BloodSkin b) Drug Metabolism jm . Skin Blood Absorption Excretion Figure 4.10 a) Schematic representation of a compartmental model of drug absorption and excretion, b) the equivalent STELLA model. 254 The STELLA modeling environment employs the top down approach with three distinct layers, the high-level mapping layer, model construction layer and the equation layer. The high-level mapping layer shields the actual model complexity and can be used for model initialization as well as for displaying the simulation results. The model shown in Figure 4.10b resides in the model construction layer that can be toggled between mapping and modeling mode. The equation layer has a document window that lists all the finite difference equations tha& get generated automatically as shown below. Drug Metabolism I I Blood(t) = Blood(t - dt) + (Absorption - Excretion) * dt INIT Blood = 0 INFLOWS: Absorption = k_a*Skin OUTFLOWS: Excretion = k_e* Blood I I Skin(t) = Skin(t - dt) + (- Absorption) * dt INIT Skin = 1 OUTFLOWS: Absorption = k_a*Skin k_a = 0.5 O k_e = 0.1 The navigation icon can be used for easy movement between layers and the run controller icon is used to start the simulation as well as set various specifications. After initializing all the stocks, converters and establishing the flow rule, the model can be simulated for any length of time. The finite difference equations can then be solved with a given time 255 interval using one of the three available integration options, Euler, second-order Runge- Kutta or fourth-order Runge-Kutta methods. The program uses the input values and calculates the resultant concentrations of every species (compartments) in the model for each time interval. The results of the simulation of the model as shown in Figure 4.10b for 24 hours (dt = 0.001 hr) with an initial concentration of drug in S and B of I and 0. respectively, when k, and k, are 0.5 hr ' and 0.1 hr ', respectively, is shown in Figure 4.11. Though in this example STELLA acts only as an equation solving software, its ability to model unknown phenomena with ease and to predict observed changes is what attracted our attention. The use of STELLA to model the disposition of ibuprofen in human blood plasma leading to the identification of systemic bioinversion pathway is one such example [24]. In this approach, a preliminary STELLA model is constructed from a basic understanding of the physical system. The kinetic behavior of the model is simulated and is compared to the experimentally observed data. In the next step, the model is changed so that the simulated data closely predicts experimental results. The model can be simulated either with a new set of initial variables or changed with the addition or omission of components. This process is iteratively continued until the fit between the simulated and experimental data is acceptable. Simulation alone can never prove a model positively, but can help in excluding ill-fitting solutions and in identifying other possible solutions. The possible solutions can then be addressed either experimentally or theoretically from a basic understanding of the system under study. 256 s Figure 4.11 The results of the simulation of the model shown in Figure 4.10b for 24 hrs (dt = 0.(X)1 hr) using Runge-Kutta four point method. For this approach, STELLA offers the advantage of ease with which the model can changed but suffers from a disadvantage as it does not offer any fitting capabilities requiring an independent fitting procedure. Possible strategies identified in the literature include, a) downhill simplex with several random start vectors, b) a genetic algorithm where the population would be a parameter vector and the evaluation would be a complete set of simulations and c) simulated annealing. These approaches would involve extracting the finite difference equations from the STELLA model, embedding it in a program employing one of the above mentioned three strategies and would be computationally demanding. For our present modeling study, such an effort did not seem to be warranted, given that we were rather interested in unearthing the physical phenomenon underlying the observed data than in an extremely good fit. Hence we restricted ourselves to running a series of simulations varying the model components at first and then each parameter selecting their optimum value by visual comparison of simulated curve with experimental data points. This approach is similar to the approach employed in solving the mechanism of autoreceptor-mediated regulation of G ABA release [25]. Previous studies of the Ru(bpy)3 -'-bipyridinium system in homogeneous as well as heterogeneous medium indicate long-lived charge separation with complex recombination dynamics. In homogeneous solution the cage escape efficiency has been found to increase with increasing MV-* concentration [26]. Mallouk and coworkers have invoked the idea of self-exchange between viologens to explain the longer lived charge separation in zeolites [4a]. Dutta and Turbeville have also used the self-exchange 258 argument to help explain the steady state charge separation observed on photolysis of Ru(bpy)3 --methylviologen zeolite samples [5c]. However, the previous studies have not attempted any kinetic analysis of the long lived charge separation [5]. In order to model the complex BET process, we have included the following reactions at high loading levels in addition to that of reactions 4 .1-4.4 (using MV-" as the example) ^hop MV ' + MV: •« MV-' + MV" (4.8) Ru(bpy)/' + MV" ------► Ru(bpy)j-' + MV- (4.9) Electron hopping between bipyridinium ions (kh„p) provides a pathway (reaction 4.8) for the electron to move farther away from the Ru(bpy)/' center. Once the electron has hopped on to neighboring cages, it can “diffuse” around In the zeolite via charge hopping until it encounters a Ru(bpy) 3 "'. This recombination process is represented in reaction 4.9 and is controlled by ki. a second order rate constant. Our first attempt at kinetic modeling involved the construction of a preliminary STELLA model for the entire physical system of all six required processes (equations 4.1-4.4.4.8, and 4.9) resulting in the STELLA model shown in Figure 4.12. The cloud shaped icons in this model indicate infinite sinks and are used to drain species that are of no consequence to the model after reaching the sink. This component is employed to mimic the discrete nature of the laser pulse as we are modeling the decay of the signal due to a single pulse. Though comprehensive, this model was found to be unnecessarily complex involving the fast excitation of Ru(bpy) 3 ** (equation 4.1 ). The excitation of 259 Oorx» Rale of BET Ru2 Ru 2 e« Ru3 kb Rate ol Decay Rate ot FET Rate ot Excitation Recombination Rate 8 Acceptor Rate of recombination FET F&te MV tai MV rad BET Rate Rate ot BET total MV rad Figure 4.12 Preliminary STELLA model for the photoinduced charge transfer processes (equations 4.1-4.4,4.8, and 4.9) involved in the Ru(bpy);^*-MV^+- zeolite system. Ru(bpy)3 ‘" to the singlet MLCT state and the subsequent efficient intersystem crossing to the triplet MLCT state are known to be ultra-fast compared to the timescale of our experiments. Hence it was decided not to include equation 4.1 in the subsequent modeling attempts. On restricting the model to the five electron transfer processes necessary (equations 4.2- 4.4,4.5,4.6), and integrating the donor and acceptor components, the STELLA model simplified as shown in Figure 4.13. Though this model is simpler than the one considered before it still did not accurately predict the system that we are trying to simulate. In the TRDR experimental data, the signal originates with the population of Ru(bpy)3 ^* and bipyridinium radical. Within the first 100 ns window of measurement, the minimum time for which the detector gate was left open, the TRDR signal maximized indicating the forward electron transfer rate constant (kf, equation 4.3) is greater than 10' s ' for all the bipyridinium ions. Hence it was decided to eliminate equation 4.3 from the model as the lower estimate of the kf would be the same for all the cases. The TRDR data under high bipyridinium loading conditions that we are attempting to model had no contribution from *Ru(bpy)j-*. Since the forward electron transfer process is not to be included, in the absence of signal due to *Ru(bpy) 3 - \ the decay due to processes not Involving electron transfer (radiative and non-radiative) as shown in equation 4.2 need not be involved in the model. Even after these attempts the predicted behavior was very different from the observed data and led to the identification of the cyclical sub-system involving ‘Radicals in cage’ and ‘Supercage escaped’ resulting in spurious long-lived bipyridinium radicals. When the outflow from the ‘Supercage escaped’ state was 261 Integrated Model Supercage escaped Hop Recomtrined Rate ol BET Recombination g Rate ol FET K2 G total rad Ru2Ex Rate ol decay Figure 4.13 Integrated STELLA model for the charge transfer processes (equations 4 2-4.4,4.8, and 4.9) involved in the Ru(bpy)]^*-MV^*- zeolite system. decoupled from the "Radicals in cage’ and instead made to drain into an infinite sink the model improved considerably. This model was varied in a number of ways and led to the final model as shown in Figure 4.14 involving equations 4.4,4.8. and 4.9. The model shown in Figure 4.14 is elegantly simple and shows the state of the photoelectron instead of the individual chemical entities. The flows indicate the migration of the photoelectron among the different states instead of the conventional conversion of one chemical species into another. In this model each compartment represents three cages and their contents, for example the stock labeled *Ru3 MVrad MV2' represents Rufbpy);^" in a cage, MV" radical in an adjacent cage, and MV- in another cage. The model was initialized (populated) with an unit concentration of ‘Ru3 MVrad MV2' corresponding to the maximum of the normalized [(J-Jq) / Jq] plot observed with the smallest time delay of 100 ns (Figure 4.9b). Migration of the charge separated photoelectron either by back electron transfer (equation 4.4) or hopping (equation 4.8) results in the depletion of Ru3 Mvrad MV2' state. Both these processes are first order in nature. Hopping provides the pathway necessary for the longer-lived charge separation and is followed by the second order recombination process (equation 4.9). In the low loading case, hopping would not be able to compete effectively resulting in the reduction of the model to a single first order process. It is important to explore the physical system behind the photoinduced charge separation in zeolite in order to understand the limitations of this model. In zeolite Y, each supercage. 13  in diameter, is surrounded by four identical supercages in a tetrahedral manner and are intercotmected by 7  windows. This tetrahedral arrangement 263 Working Model TOTAL MV RAD Recombination Ru2 MV2 MV2 Ru3 MVrad MV2 Ru3 MV2 MVrad & BET Hop Figure 4.14 STELLA model for the charge transfer processes originating from the radical cation (equations 4.4,4.8, and 4.9) involved in the Ru(bpy)g^*-MV^*- zeolite system. can be represented schematically by an infinite diamond structure (a typical pm sized zeolite particle corresponds to 10’ -10'° supercages), a part of which is shown in Figure 4.15 with the spherical balls representing the supercage and the bonds representing the windows. The intrazeolitic distribution of Ru(bpy)]-' calculated from purely statistical considerations has been reported in the literature [14]. At a loading level of one Rulbpy);-' per 15 supercages (Figure 4.15, cage ‘a’) employed in this study, the probability of finding another Rufbpylj-' molecule in one of the four adjacent supercage (b) is about 25%. Similarly the probability of finding a Ru(bpy)/' in a supercage once removed (c) and twice removed (d) are about 44% and 18%, respectively and for those in cages farther away the probabilities are much less [14]. When the bipyridinium cations are exchanged into the zeolite, they fill the supercages not occupied by Rufbpyjj-*. Considering the Rufbpy);-' in the supercage marked 'a' in Figure 4.15. upon excitation, electron transfer can occur to the bipyridinium present in any of the adjacent cages. This photoseparated charge present on a bipyridinium radical cation can imdergo well defined first order processes: 1 ) recombination (Iq,) with the Ru(bpy)/' present in the adjacent cage, “a’, or 2) migration to the bipyridinium cations present in the neighboring cages (k^op). The fate of the cage escaped charge is ill-defined and could recombine with the original Rufbpy),^" in cage a' by back tracking its path or by following the six-membered ring of supercages or by a myriad of other pathways or could recombine with a different RufbpyJs"" present in any of the other supercages (b, c. d ...). In homogeneous solutions and in some heterogeneous media at any concentration of Rufbpy)]-' and bipyridinium cations a continuum of 265 Figure 4.15 Schematic representation of the arrangement of supercages in zeolite Y. 266 intermolecular distances is possible. Also the molecules could diffuse freely in solution resulting in short average migration pathway for the recombination of the separated charges. However the regular structure of zeolite Y permits only discrete restricted intermolecular distances and also well defined migration pathways for the photoseparated charges but with a range in the number of cages traveled by a charge before recombination (k,). In our modeling attempts we have represented this complex recombination process by a simple second order process (k?) akin to that of the diffusive process in solution. Also no distinction is made between the bipyridinium cations present in the same cage and only those present in different cages were distinguished. By employing an admittedly simplified representation, the actual complex process has been reduced to a pair of simultaneous reactions (4.4. 4.8) and a consecutive reaction (4.9) involving the bipyridinium radicals. The kinetic analysis of such a system is complex and integration of non-linear differential equations would be considerably difficult, especially when the rate constants were comparable [27]. In this respect, STELLA offers an advantage as it converts the coupled differential equations into a set of finite difference equations which can be simulated with ease. One of the important step in simulating the kinetic data using the STELLA model is choosing appropriate initial state variables describing the system. The following strategy was used to determine the initial values of the rate constants, ky and k,. First, Iq, was estimated by numerically fitting the bipyridinium radical ion decay for time intervals < 1 psec when radical disappearance due to recombination would not be significant. Similarly, k? was estimated by fitting the same decay at time intervals > 10 psec, when most of the surviving radicals 267 would have cage escaped and radical disappearance due to BET would not be significant. These estimated values were then used to initiate the kinetic model and the decay was simulated using STELLA. In order to eliminate systematic errors due to the integration method, the time interval was halved until there was absolutely no difference between two such successive halving. However for a simulation period of 0.01 s, the smallest time interval permitted by STELLA was 3.2 x 10'^ s. As this time interval is larger than the first three data points (1,2.3 x 10'^ s) the validity of the simulation results were tested by comparison with that of a simulation for shorter period with appropriately shorter time intervals and were found to be almost identical. The simulation results were compared with the experimental data visually and the rate constants were then manually iterated for the best fit and the corresponding kb**', k, and khop are reported in Table 4.4 The solid lines through the decay in Figure 4 .16 and 4.17 are the simulated fits using this model. There are three rate constants involved in the process: kb""' (hl-high loading), kbop and k,. In order to convert k, to concentration units, the absorptivity of the bipyridinium radical ion inside the zeolite and the penetration depth need to be known. Previous studies have considered the absorptivity of viologen in the zeolite to be the same as in solution and is a reasonable approximation [5b]. However, the penetration depth is not readily determined and so we are not reporting k, in more conventional units. The back electron transfer rates are higher in magnitude as compared to the low loading bipyridinium samples and follow the order : 2DQ’* > 3DQ-* > 4DQ-* > MV-“. 268 I 0.4 l.Oe-7 1.0 e-5 0 0.4 l.Oe-7 0 0.0 5.0e-3 l.Oe-2 Time (sec) Figure 4.16 The transient TRDR data showing the decay of the bipyridinium radical signal for high loading bipyridinium-Ru(bpy-zeolite (Inset shows decay at shorter times) for (a) MV'^ (1.7/1), (b) (1.6/1). Solid lines represent the decay simulated with the STELLA model shown in Figure 4.14 with the rate constants given in Table 4.4. 269 # 0.4 .Oe-7 1.0 e-5 0 0.4 .Oe-7 0 0.0 S.Oe-3 l.Oe-2 Tim e (sec) Figure 4.17 The transient TRDR data showing the decay of the bipyridinium radical signal for high loading bipyridinium-Ru(bpy)^"-zeolite (Inset shows decay at shorter times) for (a) 2DQ^^ (1.4/1), (b) 4DQ‘^ (1.2/1). Solid lines represent the decay simulated with the STELLA model shown in Figure 4.14 with the rate constants given in Table 4.4. 270 Bipyridinium Loading * kb"(sec') khop (sec') ki (cone' sec') 2DQ-* 1.6/1 2.5 X 10^ 2.5 X 105 6.0 X 10' MV-' 1.7/1 9.0 X lO-* 2.0 X 10' 3.0 X 10' 3 DQ-- 1.4/1 1.8 X 1Q5 2.3 X 10' 1.0 X 10' 4D Q -' 1.2/1 1.2 X 105 1.8 X 10' 1.5 X 10' Loading indicates bipyridinium molecules per supercage, high loading. Table 4.4 Rate Constants in the high loading Bipyridinium - RuCbpy);-' - Zeolite Y systems. There are several issues that are related to the back electron transfer rate constants. These include the dependence of the back electron transfer rate constants on the viologen loading levels, the decreased value of the back electron transfer rate constants in the zeolite as compared to the forward electron transfer and the dependence of the back electron transfer rate constant on the bipyridinium reduction potential. Loading effects There are two sets of back electron transfer rate constants reported in Table 4.3 (page 248) and Table 4.4 determined by the loading of the bipyridinium ions into the zeolite. The low loading samples have rate constants that are smaller by a factor of ten as compared to the high loading samples, for example in the case of 3DQ-* the ky" and kb*'’ were 1.1 x 10"* s ' and 1.8 x 10^ s'', respectively. In the low loading case (Figure 4.4b), a supercage that contains a bipyridinium ion has rest of the volume of the supercage occupied by water and hydrated Na* ions. This provides the bipyridinium radical ion the possibility of exploring the volume of the supercage leading to greater separation between 271 the Ru(bpy) 3 ^* and the radical ion. In the high loading case, as is evident from Figure 4.4c. the bipyridinium ions are closer to the Ru(bpy)j-* and their mobility is considerably restricted since the whole zeolite crystallite is filled with bipyridinium ions. Thus, the separation distance between the Ru(bpy))^' and the bipyridinium radical ion is smaller on average in the high loading case and should result in a faster back electron transfer reaction. Magnitude o f the back electron transfer reaction rate constants In the intrazeolitic Ru(bpy) 3 -'-bipyridinium samples, there appears to be at least a two order of magnitude decrease in the back versus forward electron transfer rate constants, for example in the case of high loading 2DQ- the kf and kb’’' were >1 x 10’ s ' and 2.5 x iQ'* s ', respectively. Considerable research has been done on the RuCbpy)^’'- viologen and Ru(bpy) 3 ’*-diquat systems in solution and on various matrices [1,28]. In the covalently held Ru(bpy) 3 ’ -viologen or Ru(bpy) 3 ’*-diquat complexes, the rate constants for the back electron transfer reaction are comparable or slightly faster than the forward electron transfer, and the orders of magnitude are 10'“ s ' [17]. In solution, the forward electron transfer is diffusion controlled, and the back electron transfer typical of the geminate recombination of Ru(bpy) 3 ^* and viologen radicals is also of the order of 10'“ s ' [26]. If the Ru(bpy) 3 ’* and viologen are held in micelles, then the rate constant for the back electron transfer was reported as 5.7 x 10“ s ', and the forward electron transfer rate constant was not reported [29]. If covalently held Ru(bpy) 3 ’ -viologen complexes are adsorbed on alumina-coated colloidal silica particles, nanosecond transient studies suggest that the back electron transfer is faster than the forward reaction [20]. In the case 272 of Ru(bpy)3 -"-viologen ion-exchanged onto surface zeolite sites via the charged viologen moiety, a significant increase in the lifetime of charge separated species was reported [5b]. It was proposed that lateral charge transfer between Ru and/or viologen centers was occurring on the zeolite surface. However, in the present study, Iq," was determined from low loading samples of bipyridinium ions, which essentially contain isolated Rufbpy);-'- viologen units separated spatially from each other and there is no opportunity for lateral charge transfer. Thus, the differences between the forward and back electron transfer rate constants for reactions within the zeolite must be due to the role of the framework. The rates of electron transfer are controlled by the exoergicity of the reaction, the reorganization energy, adequate spin-orbit coupling for triplet to singlet conversion in the back electron transfer process and appropriate orbital overlap between donor and acceptor. All of these factors can be influenced by zeolite encapsulation. Electrochemistry on zeolite Y modified electrodes show that at loadings of 1 viologen per 2 supercages, the redox potentials of different viologens are close to the values in solution [30]. Electrochemical studies on Rufbpy),-' ion-exchanged on the surface of zeolite Y show shifts in the redox potential to more positive values [31], decreasing the driving force for the forward electron transfer quenching with bipyridinium ions. Thus, it was found that the quenching of Rufbpy)/ ' emission on the surface of the zeolite by 3DQ- was poor [4a]. However, according to the observations made by Dr. Norma Castagnola in our research group, the decrease of intrazeolitic Rufbpy),-*' fluorescence emission upon 3DQ- quenching is greater than a factor of 60, and resembles that observed in solution [11]. Thus, even though the values of the reduction potentials inside the zeolite 273 are not known, the similarities that are observed for quenching reactions in the zeolite and in solution for comparable systems indicate that the reduction potentials inside the zeolite are not being altered dramatically. The influence of the solvent reorganization energy in zeolites on the electron transfer dynamics is difficult to evaluate since the intrazeolitic solvation environment is not well studied. In the high loading bipyridinium case, there could be a parallel to previous observations in rigid/frozen media. Solvent molecules cannot reorient in the rigid/frozen media and profound effects on electron transfer dynamics are observed [32]. For example, in the case of [(4,4'-(C02Et)2bpy)Re(C0)3(py-PTZ)]', the rate constant for the final back electron transfer after photoexcitation and intramolecular electron transfer is decreased by a factor of 42 in a rigid poly(methylmethacrylate) matrix as compared to fluid CHjCN [33]. This arises because the back electron transfer occurs in the inverted region, and immobilization in a rigid medium increases the energy gap. In a zeolite sample, there are water molecules coordinated to ions, as well as bulk water in the supercages [34]. As molecules like Rufbpy);-' and bipyridinium ions fill the supercages, they replace the bulk water molecules, leaving behind proportionately increasing levels of water molecules that are strongly coordinated to the sodium cations and the zeolite framework, and are less mobile. Others in our group have studied this “water-transfer effect” by replacing the sodium cations with inert bulky organic tetraethylammonium (TEA) cations in Rufbpy);-' encapsulated zeolite Y [35]. Exchange of TEA profoundly influenced the fluorescence properties of *Ru(bpy) 3 ' \ shifting the emission to higher energy, increasing the quantum 274 yield and lifetime. These observations have been explained by the destabilization of the dipolar MTCT state. Ru"‘(bpy) 2 (bpy)*’ due to the intrazeolitic bulk-like uncomplexed water being replaced by the TEA cations and the subsequent diminution in k„,. Similarly, in the present case when the bipyridinium cations capable of quenching the MLCT state were exchanged into Ru(bpy)j-" encapsulated zeolite Y, the bulk-like uncomplexed water would get replaced leaving behind only the water molecules tightly held by the cations and the framework. We hypothesize that upon intrazeolitic charge transfer, these bound water molecules will have difficulty redistributing in order to equilibrate to the new charge distribution. Even though the media is not as rigid as a frozen solution, it will appear so because the solvent molecules are already strongly coordinated and slower to reorient. More research needs to be done to evaluate the contributions of the intrazeolitic medium to solvent reorganization energy, including the role of the zeolite framework. The back electron transfer process is also accompanied by a spin inversion of the mostly triplet Ru(bpy) 3 ^'-bipyridinium radical ion which is formed immediately after quenching to the singlet ground Ru(bpy) 3 -‘-bipyridinium ion. The rotational motion of the Ru(III) complex promotes the triplet to singlet conversion [36]. Constraining the Rufbpy))-' in the zeolite supercage will disrupt the spin conversion pathway and can contribute to slowing of the back electron transfer reaction. Modification of the extent of orbital overlap between the acceptor and donor can occur because of entrapment. Unlike the forward electron transfer that occurs from the bpy ligand to the bipyridinium ions across the 7  window, the back electron transfer needs to occur from the bipyridinium radical ion to the dn (A,) orbital of ruthenium [37], 275 in tlie center of the neighboring supercage. The dtt (A,) orbital does not interact with the bpy ligand orbitals [38]. This will necessitate Ru(bpy) 3 -" reorienting inside the zeolite to allow for a favorable overlap of the metal and viologen orbitals. We propose that the likelihood of Ru(bpy) 3 -* rotating inside the supercage is small because of the steric constraints in the zeolite due to the number of bad contacts of the bpy ligand with the framework atoms (Figure 4.1 ). Any reorientation of the viologen molecule in the neighboring supercage would not get it closer to the Ru center, because the 7 A window is blocked by the bpy ligand (Figure 4.4). This zeolite imposed restricted orientation decreases electronic coupling between acceptor and donor and slows the back electron transfer. The influence of orientational effects on electron transfer rates is well recognized in the literature. For example, the rate constant for the back electron transfer within the association complex {[(bpy*‘)Re'(CO)3(4-Etpy)] 10-MePTZ**} is 3 orders of magnitude faster than in the covalently linked [(bpy*')Re‘(C0)3(py-PTZ**)]**. though the electron transfer is occurring from the same donor to acceptor and therefore, the driving force for both reactions is the same [39]. This is due to the fact that the -CH,- link in the covalently held complex is inhibiting the electronic coupling by imposing restrictions on the relative orientation between bpy" and PTZ**. Electron transfer within a bisporphyrin in a low temperature glass was found to be slower than in fluids and attributed partly to stereochemical effects [40]. In tr-bridged systems, rates of electron transfer can change by a factor of 20 as the angle between bridging k -electron clouds change by 45° [41]. Unlike these examples where chemical linkers or freezing are required, in the present study the motional constraints imposed by the zeolite architecture is responsible for 276 minimizing orbital overlap. Theoretical calculations of the electronic coupling between donor and acceptor sites for metallocene-metallocenium redox pairs have shown a striking dependence on the relative orientations [42]. Recently, an electron confinement concept for explaining altered chemical and photochemical reactivity in zeolites has been proposed [43] and is related to the orbital overlap argument just advanced. The negative charge of the zeolite is partially delocalized over the aluminosilicate framework that forms the walls of the cages. The electron density of the guest molecule orbitals that extend to the walls of the cages will be repelled by the electron cloud of the framework, leading to a spatial contraction of the orbitals. This effect will be pronounced for molecules that occupy the entire cage volume, such as Ru(bpy)/\ If this orbital contraction does occur, then the overlap of donor-acceptor orbitals, even for similar geometrical arrangements will be different in the zeolite as compared to other media and will influence the back electron transfer reaction in the Rufbpyj^-'-bipyridinium system more so than the forward reaction. Relationship between back electron transfer and bipyridinium ion reduction potentials We will primarily discuss the high loading bipyridinium ion back electron transfer rate constants, since the geometry of the donor acceptor pairs are better defined because of the restricted mobility of the bipyridinium ions. The back electron transfer rate constants between the different bipyridinium ions follow the order 2DQ-* > 3DQ-* > 4DQ-* > MV-'. Based on the reduction potentials of the bipyridinium ions in solution, the driving force for the back electron transfer reaction for this series is expected to be 2DQ-* < MV-" < 3DQ- < 4DQ-*. Figure 4.18 is a plot of the rate constants for the back 277 electron transfer versus the driving force, AE derived from the reduction potentials in solution. For the diquat series, the reactions follow the Marcus inverted behavior. Similar results have been reported for reaction of Ru(bpy) 3 ‘* with diquat ions in solution [16]. Covalently held Ru(bpy) 3 -'-viologen and diquat moieties also show behavior expected for reactions occurring within the inverted region [17c,44]. Thus, the intrazeolitic behavior is consistent with previous studies. However, MV-' is not following the trend, the electron transfer rate being considerably slower from what would be expected from its reduction potential. In the case of MV-'. both the quenching of Ru(bpy) 3 -'* and the back electron transfer from the viologen radical to Ru(bpy) 3 '' are not consistent with the prediction based on its reduction potential. As has been mentioned before, the geometry of the MV-' is quite distinct from the diquat ions. Its cylindrical shape allows for more efficient contact of the pyridinium part of the molecule with the bpy ligand of Ru(bpy) 3 -'. This will promote the quenching reaction. Also, as compared to the diquat series, where the rotation around the bipyridinium units is essentially controlled by the methylene units, there is no restraint on the twisting around the C-C bond in MV-', and can be promoted by high loading. Such geometric changes may influence the electron transfer dynamics. At present, we do not have all the experimental evidence to explain the anomalous results for MV-'. Longer lived charge separation in high loaded bipyridinium samples Despite the larger back electron transfer rate constants for the high-loading bipyridinium samples, long-lived charge-separated species are observed. Previous studies of Ru(bpy)j-' and MV^\ as well as the covalently linked Ru(bpy) 3 -'-viologen 278 2.8e+5 2DQ 1.2e+5 - MV* # 8.0e+4 - 1.6 1.7 1.8 1.9 2.0 A E" (V) Figure 4.18 Plot of the rate constants for the back electron transfer under high bipyridinium loading versus the driving force (AE°) derived from the reduction potentials in solution. 279 samples on zeolite surfaces have shown that the recombination dynamics are complex [4,5]. In the intra^iitic study described here, the complexity in the back electron transfer dynamics for the high loading samples arises from the competition between back # electron transfer and self-exchange processes. The model that we have used to explain long-lived charge separation is parallel to the process that occurs in solution [28a]. In solution, after the oxidative quenching, Ru(bpy) 3 ^* and bipyridinium radical ion are produced, similar to that in the zeolite. This is followed by a geminate recombination with rate constants of the order of 10'“ s ' in solution and 10’ s ' in the zeolite. Species that escape cage recombination in solution combine by a second order diffusive process. In the zeolite, the comparable process to cage escape in solution is electron hopping via packed viologens to supercages further from the Ru center. The hole (Rulbpyl^’' ) and electron (bipyridinium radical ion) recombine by a second order process, controlled by hopping of charge within the zeolite. Thus, the difference between the solution and the zeolite is in the time scale, with events being slowed down in the zeolite and the packing of the bipyridinium ions making it possible for charge migration. It is interesting to note that cage escape efficiencies in solution also increase with increasing MV-' concentration [26]. At high MV-' concentrations, aggregates are formed. It may be possible that, as in the zeolite, that the MV- aggregates promote self-exchange. The idea that self-exchange between viologens can lead to longer lived charge separation in zeolites was first noted by Mallouk and coworkers, based on transient spectroscopy [5a]. Dutta and Turbeville have used the self-exchange argument to help explain the steady state charge separation observed on photolysis of Ru(bpy) 3 *'- 280 methylviologen zeolite samples [5c]. It is of interest to compare the dynamics of electron transfer as one proceeds from the surface of the zeolite to the bulk. For covalently held Ru(bpy)3 -'-viologen systems ion-exchanged onto the surface of zeolites, it was estimated that the electron exchange rate is about 10 times slower than the back electron transfer reaction [5b]. On the external surface of the (111) face of zeolite Y, the spacing between centers of the supercages is of the order of 11 Â. In the completely intrazeolitic Ru(bpy)3 -'-bipyridinium systems, the electron hopping rate constants are comparable to the back electron transfer rate constant. The dense packing of the bipyridinium ions in the high loading case can lead to van der Waals contacts between the ions and thereby promote more efficient charge hopping as compared to the external surface of the zeolite. The zeolite Y topology also aids in the charge separation process. As the electron escapes from the Ru(bpy) 3 ^' - bipyridinium radical ion assembly in adjacent cages to a neighboring supercage, it can return back to the same supercage. However, at this stage, it has the option of recombining with Ru(bpy) 3 ^* in one of the cages, or moving to a bipyridinium ion in 3 of the other cages as described before (Figure 4.15). This would not occur in zeolites with two-dimensional channels such as zeolite L or mordenite. There have been several other instances where long-lived charge separation has been reported in zeolites, including charge transfer between arenes as electron donors and viologen as electron acceptors [45]. Unlike the case presented here, where donors and acceptors are in neighboring cages, the arene and the viologen occupy the same cage. The reason for the long-lived charge separation was proposed to be electrostatic in nature. 281 where the arene cation and viologen radical ion are pulled to opposite ends of the supercage, thereby slowing down the electron transfer reaction. Other examples of long-lived charge separation include cases where the zeolite framework is the electron donor. Thomas and coworkers have reported that pyrene anion radicals are formed by photochemical electron transfer from the zeolite [46]. The special intrazeolitic environment is required for this reaction. No evidence of pyrene radicals are observed on the exterior of the zeolite. Scaiano and coworkers have noted that laser flash photolysis of MV-‘-zeolite at 266/355 nm results in electron transfer from the framework and long-lived formation of the methylviologen radical cation [47]. The rate of the back electron transfer was controlled by the framework basicity. Quantum Yield: In the case of MV-'. we also estimated a transient quantum yield of formation of Ru(bpy);^' and M V ' 100 ns after the exciting pulse to be between 0.44-0.59 (multiple measurements). This was done by estimating the absolute (J-Jo)/Jo values from Figure 4.8a and 4.8b. Two assumptions were needed to get the quantiun yield. First, the molar absorptivities of the viologen radical at 389 run and the RuCbpy),-*’ at 370 nm in the zeolite were taken to be 42,100 L mole ' cm"' and 27,300 L mole'' cm ', respectively, the same as in solution [48]. Secondly, the penetration depth of the radiation is considered to be similar for both Ru(bpy) 3 -*-zeolite Y and Ru(bpy) 3 -'-bipyridinium-zeolite V. These assiunptions are similar to previous studies in the literature [5b]. Also, no correction factors were applied for differences in penetration depths at wavelengths 390 and 370 nm, considering that these wavelengths are close to each other [49]. Since the absorptivities 282 of the nDQ ' radical ions are not well-established, we did not estimate the quantum yields for the diquat systems. Cage Escape Yield: A more meaningful parameter from the charge utilization point of view is a zeolite supercage escape yield. We define this as the fraction of bipyridinium ion radicals that no longer have Rufbpy);^" in a neighboring cage and is khop/(khop + k j, which amounts to about 0.69 for the high loading Rufbpy);-' - MV-' - zeolite Y. Similar calculations for 2DQ-'. 3DQ-', and 4DQ-* result in zeolite supercage escape yields of 0.50, 0.56 and 0.60 respectively. For the low loading samples, such a cage escape yield would be zero since there is no charge propagation out of the supercage in which the bipyridinium radical ion was generated. CONCLUSION In conclusion, this study establishes that the zeolite framework has a significant role in influencing electron transfer rates by virtue of specific geometric constrainment of acceptor and donor molecules. Forward electron transfer is favorable due to good positioning of donor and acceptor, but the back electron transfer is restricted because of unfavorable overlap between metal center and viologen orbitals. Lowered mobility of water molecules in the zeolite may also favor a smaller back electron transfer rate constant. The rates of the back electron transfer decrease with increasing driving force for the reaction, indicating that the reactions are occurring in the Marcus inverted region. The favorable packing of the bipyridinium acceptor molecules via the ion-exchange properties of the zeolite also provides a pathway for the photogenerated electron to migrate via bipyridinium self-exchange, leading to long-term charge separation. 283 REFERENCES 1. a) Grâtzel, M.; Kalyanasundaram, K. “Kinetics and Catalysis in Microheterogeneous Systems", Marcel Dekker, New York, 1991. b) Ramamurthy, V. “Photochemistry in Organized and Constrained Media”. VCH, New York, 1991. 2. Vaidyalingam, A. S.; Coûtant, M. A.; Dutta, P. K. “Electron-transfer Processes in Zeolites and Related Microheterogeneous Media”, V. Balzani (Ed), Wiley-VCH, Weinheim, 2001; Vol. 4, pp 412. 3. Kalyanasundaram, K. “Photochemistiy of Polypyridine and Porphyrin Complexes” Academic Press, London, 1992. 4. a) Krueger, J. S.; Mayer, J. E.; Mallouk, T. E.J. Am. Chem. Soc. 1988, 110, 8232. b) Borja, M.; Dutta, P. K. Nature 1993, 362, 43. c) Sykora, M.; Kincaid, J. R. Nature 1997, 387, 162. 5. a) Kim, Y. I.; Mallouk, T. E.J. Phys. Chem. 1992. 96, 2879. b) Yonemoto, E. H.; Kim, Y. L; Schmehl, R. H.; Wallin, J. O.; Shoulders, B. A.; Richardson, B. R.; Haw, J. P.; Mallouk, T. E. J. Am. Chem. Soc. 1994, 116, 10557. c) Dutta, P. K.; Turbeville, W.J. Phys. Chem. 1992, 969410. . 6. a) Dunn, I. J.; Heinzle, E.; Dettwiler, B.; Saner, U.: Ryhiner, G.; Ingham, J. Chem. Eng. Sci. 1988, 4 3 ,1897. b) Bogen, D. K. Science 1989, 246, 138. 7. a) Melka, R. P.; Olsen, G.; Beavers, L.; Draeger, J. A. J. Chem. Educ. 1992, 69, 596. b) Steffen, L. K.; Holt, P. L. J. Chem. Educ. 1993, 70,991. 8. a) Nicholls, P.; Cooper, C. E.; Wrigglesworth, J. M. Biochem. Cell Biol. 1990, 68, 1128. b) Shatkin, J. A.; Brown, H. S. Environ. Res. 1991, 5 690. , c) Gakis, C.; Cappio-Borlino, A.; Pulina, G. Biochem. Mol. Biol. Int. 1998, 4 6487. , d) Sutinen, R.; Paronen, P.; Saano, V.; Urtti, A. Eur. J. Pharm. Sci. 2000, II, 25. e) Aelion, C. M.; Kirtland, B. C. Environ. Sci. Technol. 2000, 34, 3167. 9. Costanza, R.; Duplisea, D.; Kautsky, U. Ecol. Modell. 1998, 110. 10. a) DeKock. R. L. CHED-1025, Book of Abstracts, 219th ACS National Meeting, American Chemical Society, Washington, D C., 2000. 284 b) Metz, C. CHED-1021, Book of Abstracts, 219th ACS National Meeting, American Chemical Society, Washington, D C., 2000. c) Soltzberg, L. J. CHED-1020, Book of Abstracts, 219th ACS National Meeting, American Chemical Society, Washington, D C., 2000. 11. a) Vitale, M.; Castagnola, N. B.; Ortins, N. J.; Brooke, J. A.; Vaidyalingam, A.; Dutta, P. K. J. Phys. Chem. B 1999, 103, 2408 . b) Castagnola, N. B. Ph. D. Dissertation, The Ohio State University, Columbus, OH, 1999. 12. a) Olson, D. H. J. Phys. Chem. 1970, 74,2758. b) Rillema, D. P.; Jones. D. S.; Woods, C.; Levy. H. A. Inorg. Chem. 1992. 31. 2935. c) Russell. J. H.; Wallwork, S. C.Acta Crystallogr., Sect. B 1972. 28. 1527. 13. Derycke. I.; Vigneron. J. P.; Lambin. P.; Lucas. A. A.; Derouane. E.G.J. Chem. Phys. 1991. 94.4620. 14. Sykora, M.; Kincaid, J. R.; Dutta, P. K.; Castagnola, N. B. J. Phys. Chem. B 1999, 103. 309. 15. a) Lachish. U.; Infelta. P. P.; Gratzel, M. Chem. Phys. Lett. 1979. 62. 317. b) Creutz. C.; Chou. M.; Netzel, T. L.; Okumura, M.; Sutin. N.J. Am. Chem.Soc. 1980. 102, 1309. 16. Amouyal, E.; Zidler. B. Isr. J. Chem. 1982, 22, 117. 17. a) Yonemoto. E. H.; Riley, R. L.; Kim, V. L; Atherton, S. J.; Schmehl. R. H.; Mallouk. T. E.J. Am. Chem. Soc. 1992, 114, 8081. b) Cooley, L. F.; Headford, C. E. L.; Elliott, C. M.; Kelley. D. F. J. Am. Chem. Soc. 1988. 110, 6673. c) Kelly, L. A.; Rodgers. M. A. J.J. Phys. Chem. 1995, 99, 13132. 18. Willner, L; Ayalon. A.; Rabinovitz, M. NewJ. Chem. 1990 14.685. 19. Yonemoto. E. H.; Saupe. G. B.; Schmehl, R. H.; Hubig, S. M.; Riley, R. L.; Iverson, B. L.; Mallouk, T. E.J. Am. Chem. Soc. 1994. 116,4786. 20. Kelly. L. A.; Rodgers, M. A. J. J. Phys. Chem. 1994, 9 8 ,6386. 21. Bradley, P. G.; Kress, N.; Homberger, B. A.; Dallinger, R. F.; Woodruff, W. H. J. Am. Chem. Soc. 1981, 103, 7441. 285 22. a) Evans, A. G.; Dodson, N. K.; Rees, N. H. J. Chem. Soc., Perkin Trans. 1976,2, 859. b) Yoshimura, A.; Hoffman, M. Z.; Sun, H. J. Photochem. PhotobioL, A 1993, 70, 29. 23. “An Academic User's Guide to STELLA", High Perfromance Systems, Inc., Hanover, NH, 1996. 24. Romero, A. J.; Rackley, R. J.; Rhodes, C. T. Chirality 1991,5,418. 25. Christen, T.; Baumann, P. A.; Waldmeier, P. C. Naunyn-Schmiedeberg's Arch. Pharmacol. 1993, 618. 26. Hoffman, M. Z. J Phys. Chem. 1988, 92, 3458. 27. Chien, J.-Y. J. Am. Chem. Soc. 1948, 70, 2256. 28. a) Gratzel, M. “Energy Resources Through Photochemistry and Catalysis", Academic Press, New York, 1983. b) Sauvage, J. P.; Collin, J. P.; Chambron, J. C.; Guillerez, S.; Coudret, C.; Balzani. V.; Barigelletti, F.; De Cola, L.; Flamigni, L. Chem. Rev. 1994, 94, 993. c) Hoffman. M. Z.; Bolletta, F.; Moggi. L.; Hug. G. L. J. Phys. Chem. Ref. Data 1989,75,219. 29. Rodgers, M. A. J.; Becker, J. C. J. Phys. Chem. 1980.84, 2762. 30. a) Gemborys, H. A.; Shaw, B. R. J. Electroanal. Chem. Interfacial Electrochem. 1986,208, 95. b) Li. Z.; Wang, C. M.; Persaud. L.; Mallouk, T. E.J. Phys. Chem. 1988, 92, 2592. 31. Li, Z.; Mallouk, T. E.J. Phys. Chem. 1987,91, 643. 32. a) Gaines, G. L., Ill; O'Neil, M. P.; Svec, W. A.; Niemczyk, M. P.; Wasielewski, M. R.J. Am. Chem. Soc. 1991,113, 719. b) Marcus, R. A. J. Phys. Chem. 1990,94, 4963. c) Chen. P.; Meyer, T. J. Inorg. Chem. 1996,35, 5520. 33. Jones, W. E., Jr.; Chen, P.; Meyer, T. J. J. Am. Chem. Soc. 1992,114. 387. 34. Hanke, W.; Moeller. K. Zeolites 1984,4 ,244. 35. Coûtant, M. A.; Le, T.; Castagnola, N.; Dutta, P. K. J. Phys. Chem. B 2000.104, 10783. 286 36. Buerssner, D.; Steiner, U. E. Coord. Chem. Rev. 1994, 732, 51. 37. Sun. H.; Yoshimura, A.; Hoffinan, M. Z. J. Phys. Chem. 1994, 98, 5058. 38. Ceulemans. A.; Vanquickenbome, L. G.J. Am. Chem. Soc. 1981,103, 2238. 39. Chen, P.; Duesing, R.; Graff, D. K..; Meyer, T. J. J. Phys. Chem. 1991,95, 5850. 40. Harriman, A.; Heitz. V.; Ebersole, M.; van Willigen, H. J. Phys. Chem. 1994, 98, 4982. 41. Helms, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc. 1991.113, 4325. 42. Newton. M. D.; Ohta, K.; Zhong, E. J. Phys. Chem. 1991,95. 2317. 43. Zicovich-Wilson. C. M.; Corma, A.; Viruela, P. J. Phys. Chem. 1994, 98, 10863. 44. Larson. S. L.; Cooley. L. F.; Elliott. C. M.; Kelley. D. F. J. Am. Chem. Soc. 1992, 114. 9504. 45. Yoon. K. B.; Hubig, S. M.; Kochi, J. K. J. Phys. Chem. 1994. 98. 3865. 46. Liu. X.; lu. K.-K.; Thomas, J. K. J. Phys. Chem. 1994. 98. 7877. 47. Alvaro. M.; Garcia. H.; Garcia. S.; Marquez. F.; Scaiano. J. C. J. Phys. Chem. B 1997. 101. 3043. 48. a) Kalyanasundaram, K. Coord. Chem. Rev. 1982. 46. 159. b) Watanabe. T.; Honda. K. J. Phys. Chem. 1982, 86. 2617. 49. Kim, Y. L; Atherton, S. J.; Brigham, E. S.; Mallouk. T. E. J. Phys. Chem. 1993.97. 11802. 287 BIBLIOGRAPHY Chapter 1 1. The Economist 2001, Feb 10, A survey of energy : page 6. 2. Kordesch. K. V.; Simader, G. R. Chem. Rev. 1995, 95, 191. 3. "The International Energy Outlook - 2001," Energy Information Administration, The U. S. DOE, 2001. 4. The Economist 1999.July 22, Fuel cells meet big business. 5. McCusker, J. K.Science 2001, 293, 1599. 6. a) Hagfeldt, A.; Gratzel. M. Acc. Chem. Res. 2000, 33, 269. b) Nazeeruddin, M. K.; Pechy, P.; Renouard, T.; Zakeeruddin, S. M.; Humphry- Baker, R.; Comte, P.; Liska, P.; Cevey, L.; Costa, E.; Shklover, V.; Spiccia, L.; Deacon, G. B.; Bignozzi, C. A.; Gratzel, M. J Am. Chem. Soc. 2001,123, 1613. 7. Appleby, A. J. J. Power Sources 1990,29, 3. 8. Larminie, J.; Dicks, A. “Fuel cell systems explained", John Wiley. Chichester, 2000. 9. Ashley, S. Scientific American 2001, July. 10. Hockaday, R. G.; Turner, P. S.; Maslow, M.; Cooper, M. Micro-fuel cell power devices. International Patent, WO 00/35032, 2000. 11. Finkelshtain. G.; Katzman, Y.; Khidekel, M.; Borover, G. A new class of electrocatalysts and a gas diffusion electrode based thereon. International Patent, WO 01/15253, 2001. 12. “Camcorder fueled with Hydrogen”, April 09 Press Release, Fraunhofer Institute for Solar Energy Systems, Freiburg, 2001. http://www.ise.fhg.de/engIish/press/pi_200I/pdfi1SE-PI_Camcorder_HMI_e.pdf 288 13. http://www.medisel.com/img/battery2.gif 14. Bard, A. J.; Fox, M. A. Acc. Chem. Res. 1995,2 8 ,141. 15. Amouyal, E. Sol. Energy Mater. Sol. Cells 1995, 38, 249. 16. Griffith, R. T. (Translated with a popular commentary) “Hymns of the Rgveda”, J. L. Shastri (Ed), Motilal Banarsidass, Delhi, 1973. Vol. 1, pp 538. 17 a) Yagi, M.; Kaneko, M. Chem. Rev. 2001, lOl, 21. b) Ruettinger, W.; Dism'ikes, G. C. Chem. Rev. 1997, 97, 1. 18. Zouni, A.; Witt, H.-T.; Kern, J.; Fromme, P.; Krauss, N.; Saenger, W.; Orth. P. Nature 2001, 409, 739. 19. Rhee. K.-H.; Morris. E. P.; Barber, J.; Kuhlbrandt, W. Nature 1998,396, 283. 20. Moran, L. A.; Scrimgeour, K. G.; Horton, H. R.; Ochs, R. S.; Rawn, J. D. “Biochemisty", Prentice-Hall, Englewood Cliffs, 1994. 21. Dismukes, G. C.Science 2001,292,447. 22. a) “Photoinduced Electron Transfer". M. A. Fox and M. Chanon (Eds), Elsevier. Amsterdam. 1988. b) Kavamos, G. J. “Fundamentals of Photoinduced Electron Transfer". VCH. New York, 1993. 23. “Energy Resources through Photochemistry and Catalysis". M. Gratzel (Ed). Academic Press. New York. 1983. 24. Bolton. J. R. Sol. Energy 1996.57. 37. 25. Sun. L.; Hammarstrom, L.; Akermark, B.; Styring, S.Chem. Soc. Rev. 2001.30, 36. 26. Koryakin, B. V.; Dzhabiev, T. S.; Shilov, A. E. Dokl. Akad Nauk. S.S.S.R. 1977.233, 620. 27. Lehn. J. M.; Sauvage. J. P. Nouv. J. Chim. 1977,1 ,449. 28. Moradpour, A.; Amouyal, E.; Keller, P.; Kagan, H. Nouv. J. Chim. 1978,2, 547. 29. Kalyanasundaram, K.; Kiwi, J.; Gratzel, M. Helv. Chim. Acta. 1978,61, 2720. 2 8 9 30. a) Steinberg-Yfrach, G.; Rigaud, J.-L.; Durantini, E. N.; Moore, A. L.; Gust, D.; Moore, T. A. Nature 1998, 392,479. b) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2001,3 4 ,40. 31. Sykora, M.; Maxwell, K. A.; DeSimone, J. M.; Meyer, T. J. Proc. Natl. Acad. Sci. U. S. A 2000, 97, 7687. 32. Kaschak, D. M.; Johnson, S. A.; Waraksa, C. C.; Pogue, J.; Mallouk, T. E.Coord. Chem. Rev 1999, 185-186,403. 33. Takata, T.; Tanaka, A.; Hara, M.; Kondo, J. N.; Domen, K.Catal. Today 1998, 44, 17. 34. Lanz, M.; Schurch, D.; Calzaferri, G. J. Photochem. Photobiol., A 1999, 120, 105. 35. Khaselev, O.; Turner, J. A. Science 1998, 280,425. 36. Imahori, H.; Norieda, H.; Yamada, H.; Nishimura. Y.; Yamazaki, 1.; Sakata, Y.; Fukuzumi, S.J. Am. Chem. Soc. 2001,123, 100. 37. Yonemoto, E. H.; Kim, Y. I.; Schmehl, R. H.; Wallin, J. O.; Shoulders, B. A.; Richardson, B. R.; Haw, J. P.; Mallouk, T. E.J. Am. Chem. Soc. 1994. 116.. 10557. 38. Vitale, M.; Castagnola, N. B.; Ortins, N. J.; Brooke, J. A.; Vaidyalingam, A.; Dutta, P. K. J. Phys. Chem. B 1999, 103, 2408. 39. Boija, M.; Dutta. P. K. Nature 1993, 362,43. 40. Sykora, M.; Kincaid, J. Nature 1997, 387, 162. 41. a) Balzani, V.; Bolletta. P.; Gandolfi, M. T.; Maestri, M. Topp. Curr. Chem. 1978, 75, 1. b) Balzani, V.; Moggi, L.; Manfrin, M. P.; Bolletta, P.; Lawrence, G. S. Coord. Chem. Rev. 1975, 15, 321. c) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 4 6159. , 42. Juris, A.: Balzani, V.; Barigelletti, P.; Can^agna, S.; Belser, P.; Von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85. 43. a) Van Houten, J.; Watts, R. J. J. Am. Chem. Soc. 1976, 98,4853. b) Durham, B.; Caspar, J. V.; Nagle, J. K.; Meyer, T. J. J. Am. Chem. Soc.1982, 104,4803. 290 44. Seddon, E. A.; Seddon, K. R. “The Chemistry of Ruthenium”, Elsevier, Amsterdam. 1984. 45. Barreca, D.; Buchberger, A.; Oaolio, S.; Depero, L. E.: Fabrizio, M.; Morandini, P.; Rizzi, G. A.; Sangaletti, L.; Tondello, E. Langmuir 1999, 15,4537. 46. Rard. J. A. Chem. Rev. 1985, 85, 1. 47. Huang, Y. S.: Park, H. L.; Poliak, F. H.Mat. Res. Bull. 1982, 17, 1305. 48. a) Kiwi, J.; Gratzel, M. Angew. Chem., Int. Ed Engl. 1979. 18. 624. b) Lehn. J. M.; Sauvage, J. P.; Ziessel, R. Nouv. J. Chim. 1979. 3, 423. 49. a) Amouyal. E.; Keller, P.; Moradpour, A. J. C. S. Chem. Comm. 1980, 1019. b) Keller, P.; Moradpour, A.; Amouyal, E. J. Chem. Soc., Faraday Trans. 1982,1 7&3331. 50. a) Kleijn. J. M.; Rouwendal, E.; Van Leeuwen, H. P.; Lyklema. J.J. Photochem. Photobiol., A 1988, 44. 29. b) Kleijn. J. M.; Boschloo, G. K. J. Electroanal. Chem. 1991. 300. 595. 51. Vaidyalingam. A. S.; Coûtant, M. A.; Dutta. P. K. Electron-transfer Processes in Zeolites and Related Microheterogeneous Media in “Electron Transfer in Chemistry", V. Balzani (Ed), WILEY-VCH, Weinheim. 2001. Vol. 4, pp 412. 52. Castagnola. N. B.; Dutta. P. K. J. Phys. Chem. B 1998, 102, 1696. Chapter 2 1. Gerardi, R. D.; Barnett, N. W.; Lewis, S. W. Anal. Chim. Acta 1999, 378, 1. 2. 'Tiber Optic Chemical Sensors and Biosensors”. O. S. Wolfbeis, CRC Press, Boca Raton. 1991. 3. Amouyal. E. Sol. Energy Mater. Sol. Cells 1995. 38, 249. 4. a) Balzani, V.; Juris, A.; Venturi. M.; Campagna, S.; Serroni. S. Chem. Rev 1996, 96, 759. b) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85. c) Kalyanasundaram, K. Coord Chem. Rev. 1982, 46, 159. d) Balzani, V.; Bolletta, P.; Gandolfi, M. T.; Maestri, M. Top. Curr. Chem. 1978, 75, 1. 2 9 1 5. a) Van Houten, J.; Watts, R. J. Inorg. Chem. 1978, / 7, 3381. b) Durham, B.; Caspar, J. V.; Nagle, J. K.; Meyer, T. J. J. Am. Chem. Soc. 1982, 104,4803. 6. a) Gleria, M.; Minto, F.; Beggiato, G.; Bortolus, P. J. Chem. Soc., Chem. Commun. 1978,285. b) Hoggard, P. E.; Porter, G. B. J. Am. Chem. Soc. 1978, 100, 1457. 7. Lumpkin, R. S.; Kober, E. M.; Worl, L. A.; Murtaza, Z.; Meyer, T. J.J. Phys. Chem. 1990, 94, 239. 8. Masschelein, A.; Kirsch-De Mesmaeker, A.; Wiilsher, C. J.; Wilkinson. F. J. Chem. Soc. Faraday Trans. 1991. 57. 259. 9. a) Maruszewski. K.; Strommen, D. P.; Kincaid, J. K.J. Am. Chem. Soc. 1993. 115, 8345. b) Maruszewski. K.; Kincaid, J. R.Inorg. Chem. 1995, 34, 2002. 10. Kleijn, J. M.; Rouwendal, E.; Van Leeuwen, H. P.; Lyklema. J.J. Photochem. Photobiol.. A 1988. 44, 29. 11. Adelt. M.; Devenney, M.; Meyer. T. J.; Thompson. D. W.; Treadway. J. A. Inorg. Chem. 1998. 37, 2616. 12. Tachiyashiki, S.; Ikezawa. H.; Mizumachi. K. Inorg. Chem. 1994, 33, 623. 13. Valenty. S. J.; Behnken, P. E. Anal. Chem. 1978. 50, 834. 14. Ghosh. P. K.; Brunschwig, B. S.; Chou. M.; Creutz, C.; Sutin. N. J. Am. Chem. Soc. 1984. 106, 4772. 15. Roecker. L.; Kutner. W.; Gilbert. J. A.; Simmons. M.; Murray. R. W.; Meyer. T. J. Inorg. Chem. 1985, 24, 3784. 16. Jones. W. E.. Jr.; Smith. R. A.; Abramo, M. T.; Williams. M. D.; Van Houten, J. Inorg. Chem. 1989. 28, 2281. 17. Fairbank, R. W. P.; Wirth, M. J. Anal. Chem. 1997, 69, 2258. 18. Buchanan, B. E.; McGovern, E.; Harkin, P.; Vos. J. G. Inorg. Chim. Acta 1988, 154, 1 . 19. Ambs, W. J. Microcrystalline synthetic faujasite, U.S. Patent, US4372931, 1983. 292 20. Castagnola, N. B.; Dutta, P. K. J. Phys. Chem. B 1998, 102,1696. 21. Singh, R.; Dutta, P. K. Microporous Mesoporous Mater. 1999, 3 2 , 29. 22. Maruszewski, K.; Strommen, D. P.; Handrich, K.; Kincaid, J. R. Inorg. Chem. 1991. 3 0 , 4579. 23. Wang, P.; Lee, H. K. J. Chromatogr., A 1997, 789,437. 24. Kroener, R.; Heeg, M. J.; Deutsch, E. Inorg. Chem. 1988.27, 558. 25. Moreira, I. S.; Santiago, M. O. Chromatographia 1996, 43. 322. 26. a) Twitchett, P. J.; Moffat, A. C. J. Chromatogr. 1975, III, 149. b) Keller, P.; Moradpour, A.; Amouyal, E.; Kagan, H. B. A omv . J. Chim. 1980, 4. 377. c) Johansen. 0.; Launikonis, A.; Loder, J. W.; Mau. A. W.-H.; Sasse. W. H. F.; Swift. J. D.; Wells, D. Aust. J. Chem. 1981. 3 4 .981. d) Hodgeson, J. W.; Bashe, W. J.; Eichelberger, J. W. "Determination of diquat and paraquat in drinking water by liquid-solid extraction and high performance liquid chromatography with ultraviolet detection," U.S. Environmental Protection Agency. 1992. e) Sakai, K.; Kizaki, Y.; Tsubomura, T ; Matsumoto. J.K. Mol. Catal. 1993. 79. 141. 27. O'Laughlin. J. W.; Hanson, R. S. Anal. Chem. 1980. 52. 2263. 28. Davies, N. R.; Mullins, T. L. Aust. J. Chem. 1967, 2 0 .657. 29. Wallace, W. M.; Hoggard, P. E. Inorg. Chem. 1979. 18. 2934. 30. Geraty. S. M.; Vos. J. G. J. Chem. Soc., Dalton Trans. 1987. 3073. 31. Durham, B.; Wilson, S. R.; Hodgson, D. J.; Meyer. T. J. J. Am. Chem. Soc. 1980. 102. 600. 32. Porter. G. B.; Sparks. R. H. J. Photochem. 1980, 13. 123. 33. Van Houten, J.; Watts, R. J. J. Am. Chem. Soc. 1976. 98,4853. 34. Durham, B.; Walsh, J. L.; Carter, C. L.; Meyer, T. J. Inorg. Chem. 1980, /9. 860. 35. Jones, R. P.; Cole-Hamilton, D. J. Inorg. Chim. Acta 1981, 53, L3. 2 9 3 36. Gohn, M.; Getoff, N. Z Naturforsch., A1979,34, 1135. 37. a) Amouyal, E.; Koffi, P. J. Photochem. 1985,29,227. b) Lehn, J. M.; Sauvage, J. P.; Ziessel, R. Nouv. J. Chim. 1981, J, 291. c) Das, S. D.; Dutta, P. K.. .Microporous and Mesoporous Materials 1998.22,475. 38. Moyer, B. A.; Meyer. T. J. J. Am. Chem. Soc. 1978,100, 3601. 39. Arakawa. R.; Lu. J.; Yoshimura, A.; Nozaki, K.; Ohno, T.; Doe. H.; Matsuo. T. Inorg. Chem. 1995. 34, 3874. 40. Bhuiyan, A. A.; Kincaid, J. R. Inorg. Chem. 1998,37, 2525. 41. Fujita, L; Kobayashi. H. J. Chem. Phys. 1970. 52, 4904. 42. Demas, J. N.; Adamson. A. W. J. Amer. Chem. Soc. 1971, 93, 1800. 43. a) Gafney. H. D.; Adamson, A. W. J. Amer. Chem. Soc. 1972, 94, 8238. b) Demas, J. N.; Adamson, A. W. J. Amer. Chem. Soc. 1973. 95. 5159. 44. Bock. C. R.; Meyer. T. J.; Whitten, D. G. J. Amer. Chem. Soc. 1974, 96, 4710. 45. Young, R. C.; Meyer, T. J.; Whitten. D. G. J. Am. Chem. Soc. 1976. 98, 286. 46. Moradpour. A.; Amouyal. E.; Keller, P.; Kagan, H. Nouv. J. Chim. 1978.2. 547. 47. Kalyanasundaram, K.; Kiwi. J.; Graetzel. M. Helv. Chim. Acta 1978.61, 2720. 48. Launikonis, A.; Loder, J. W.; Mau. A. W.-H.; Sasse, W. H. F.; Summers. L. A.; Wells. D. Aust. J. Chem. 1982,35, 1341. 49. Giro, G.; Casalbore. G.; Di Marco, P. G. Chem. Phys. Lett. 1980. 71,476. 50. Wemer. H. A. F.; Bauer, R. J. Mol. Catal. 1994,88, 185. Chapter 3 1. Kalyanasundaram, K.; Kiwi, J.; Gratzel, M. Helv. Chim. Acta 1978,61, 2720. 2. Amouyal, E. Sol. Energy Mater. Sol. Cells 1995,3 8 , 249. 3. Keller, P.; Moradpour, A.; Amouyal, E. J. Chem. Soc., Faraday Trans. I1982, 78, 3331. 2 9 4 4. Seddon, E. A.; Seddon, K. R. “The Chemistry of Ruthenium", Elsevier, Amsterdam, 1984. 5. Hepel, T.; Poliak, F. H.; O'Grady, W. E. J. Electrochem. Soc. 1984, 131, 2094. 6. Hiratani, M.; Matsui, Y.; Imagawa, K.; Kimura, S. Thin Solid Films 2000, 366, 102. 7. Das. S. D.; Dutta, P. K. Microporous and Mesoporous Mater. 1998,22, 475. 8. Kohno, M.; Kaneko, T.; Ogura, S.; Sato, K.; Inoue. Y. J. Chem. Soc., Faraday Trans. 1998,94. 89. 9. Takata, T.; Tanaka, A.; Hara, M.; Kondo, J. N.; Domen, K. Catal. Today 1998,44, 17. 10. Rard. J. A. Chem. Rev. 1985, 85, 1. 11. Huang, Y. S.; Park, H. L.; Poliak, F. H.Mat. Res. Bull. 1982, 17. 1305. 12. Barreca, D.; Buchberger, A.; Daolio, S.; Depero, L. E.; Fabrizio, M.; Morandini, F.: Rizzi, G. A.; Sangaletti, L.; Tondello, E. Langmuir 1999,15, 4537. 13. Rizzi, G. A.; Magrin, A.; Granozzi, G. Surface Science 1999,443, 277. 14. Meng, Li.-Jian.; Dos Santos, M. P. Mat. Res. Soc. Symp. Proc. 1999,541, 141. 15. a) Paulson, S.; Morris, K.; Sullivan, B. P.J. Chem. Soc.. Chem. Commun. 1992, 1615. b) Liang, Y.; Schmehl, R. W.J. Chem. Soc., Chem. Commun. 1995, 1007. c) Zhou, M.; Laux, J. M.; Edwards, K. D.; Hemminger, J. C.; Hong, B. Chem. Commun. 1997,1977. 16. Mabrouk, P. A.; Wrighton, M. S. Spectrochim. Acta, Part A1989,45A, 17. 17. Christ, C. S., Jr.; Yu, J.; Zhao, X.; Palmore, G. T. R.; Wrighton, M. S. Inorg. Chem. 1992,i / , 4439. 18. a) Corma. A.; Iglesias, M.; Del Pino, C.: Sanchez, F. J. Chem. Soc., Chem. Commun. 1991,18, 1253. b) Corma. A.; Iglesias, M.; Del Pino, C.; Sanchez, F. J. Organomet. Chem. 1992, 431,222. c) Sanchez, F.; Iglesias, M.; Corma. A.; Del Pino, C. J. Mol. Catal. 1991,70, 369. 295 19. Resch, U.; Fox, M. A. J. Phys. Chem. 1991,9 56316. , 20. Hagfeldt, A.; Gratzel, M. Acc. Chem. Res. 2000, 5 3 ,269. 21. Bonhote, P.; Moser, J.-E.; Humphry-Baker, R.; Vlachopoulos, N.; Zakeeruddin, S. M.; Walder, L.; Gratzel, M. 7. Am. Chem. Soc. 1999,121, 1324. 22. Kaes, C.; Hosseini, M. W.; Cian, A. D.; Fischer, J. Chem. Commun. 1997,2229. 23. Stange, A. F.; Tokura, S.; Kira, M.J. Organomet. Chem. 2000, 612, 117. 24. Rampino, L. D.; Nord, F. F. J. Am. Chem. Soc. 1941, 63 ,2745. 25. a) Keller, P.; Moradpour, A.; Amouyal, E.; Kagan, H. B. Nouv. J. Chim. 1980, 4, 377. b) Johansen, O.; Launikonis, A.; Loder, J. W.; Mau, A. W.-H.; Sasse, W. H. F.; Swift, J. D.; Wells, D. Aust. J. Chem. 1981, 34, 981. 26. Sprintschnik, G.; Sprintschnik, H. W.; Kirsch. P. P.; Whitten, D. G. J. Am. Chem. Soc. 1977, 99, 4947. 27. Gates, B. C.; Guczi, L.; Knozinger, H. “Metal Clusters in Catalysis”, Elsevier, New York, 1986. 28. Lin. L.-H.; Chao, K.-J. Studies in Surface Science and Catalysis 1994, 84, 749. 29. Shen, G.-C.; Liu, A. M.; Ichikawa, M. Inorg. Chem. 1998.37, 5497. 30. Zarbin, A. J. G.; Vargas, M. D.; Alves, O. L. J. Mater. Chem. 1999,9, 519. 31. Robertson, J.; Webb. G. Proc. R. Soc. London. Ser. A 1974, 341, 383. 32. Pearce, J. R.; Mortier. W. J.; Uytterhoeven, J. B. J. Chem. Soc., Faraday Trans. 1 1979. 7J. 1395. 33. a) Bolzan, A. A.; Fong, C.; Kennedy, B. J.; Howard, C. J. Acta Crystallogr., Sect. B: Struct. Sci. 1997.853, 373. b) CPDS Card # 3782, Institute of Experimental Mineralogy, Chemogolovka, 1997. http://database.iem.ac.ru/mincryst/s_carta.php3?RUTILE+3782 34. Cullity, B. D. “Elements of X-Ray Diffraction”, Addison-Wesley, Reading, 1978. 35. a) Zhao, G. L.; Bagayoko, D.; Williams, T. D. Phys. Rev. B 1999,6 0 ,1563. 296 b) Krasovska, O. V.; Krasovskii, E. E.; Antonov, V. N.Phys. Rev. B 1995,52. 11825. c) Xu, J. H.; Jarlborg, T.; Freeman, A. J. Phys. Rev. B 1989,40. 7939. d) Mattheiss, L. F. Phys. Rev. B 1976, 13. 2433. 36. a) Huar g, Y. S.; Liao, P. C. Sol. Energy Mater. Sol. Cells 1998.5 5 .179. b) Beikind, A.; Orban, Z.; Vossen, J. L.; Woollam, J. A. Thin Solid Films 1992, 207, 242. c) Canart-Martin, M. C.; Canivez, Y.; Declercq, R.; Laude, L. D.; Wautelet, M. Thin Solid Films m 2 . 92, 323. d) Goel, A. K.; Skorinko, G.; Pollak, F. H.Phys. Rev. B 1981. 24, 7432. 37. Moulder. J. F.; Stickle, W. F.; Sobol, P. E.; Bomben. K. D. “Handbook of X-ray Photoelectron Spectroscopy". Physical Electronics Division. Perkin-Elmer. Eden Prairie. 1992. 38. Pedersen, L. A.; Lunsford, J. H. J. Catal. 1980, 61, 39. 39. a) Blouin. M.; Guay, D. J. Electrochem. Soc. 1997.144. 573. b) Chan, H. Y. H.; Takoudis, C. G.; Weaver, M. J. J. Catal. 1997,172. 336. c) Wagner, C. D.; Naumkin, A. V.; Kraut-Vass, A.; Allison, J. W.; Powell. C. J.; Rumble. J. R. J. “NIST X-ray Photoelectron Spectroscopy Database" 2000. http ://srdata. nist.gov/xps/ 40. Verdonck. J. J.; Jacobs. P. A.; Genet. M.: Poncelet. G. J. Chem. Soc., Faraday Trans. I 1980, 76.403. 41. a) Twitchett, P. J.; Moffat. A. C. J. Chromatogr. 1975, 111, 149. b) Hodgeson, J. W.; Bashe. W. J.; Eichelberger, J. W. "Determination of diquat and paraquat in drinking water by liquid-solid extraction and high performance liquid chromatography with ultraviolet detection." U.S. Environmental Protection Agency. 1992. 42. Sakai. K.; Kizaki. Y.; Tsubomura. T.; Matsumoto. K.J. Mol. Catal. 1993.79. 141. 43. Keller, P.; Moradpour, A. J. Am. Chem. Soc. 1980,102, 7193. 44. Kleijn, J. M.; Rouwendal, E.; Van Leeuwen, H. P.; Lyklema. J.J. Photochem. Photobiol., A 1988,4 429. , 45. Amouyal, E. “Development of Catalysts for Water Photoreduction: Improvement, Poisoning and Catalytic Mechanism", E. Pelizzetti, N. Serpone (Ed), D. Reidel Publishing, Dordrecht, 1986. 297 46. Bauer, R.; Wemer, H. A. F. J. Mol. Calai 1992,72,67. 47. Over, H.; Kim, Y. D.; Seitsonen, A. P.; Wendt, S.; Lundgren, E.; Schmid, M.; Varga, P.; Morgante, A.; Ertl, G. Science 2000,287, 1474. 48. Hepel, T.; Pollak, F. H.; O'Grady, W. E. Proc. - Electrochem. Soc. 1984,84-12, 512. 49. Kiwi, J.; Grâtzel, M. J. Am. Chem. Soc. 1979,lOl, 7214. 50. a) Hostenpiller, P. A. M.; Boit, J. D.; Fameth, W. E.; Lowekamp, J. B.; Rohrer, G. S. J. Phys. Chem. B 1998,102, 3216. b) Lowekamp, J. B.; Rohrer, G. S.; Hostenpiller, P. A. M.; Boit, J. D.; Fameth, W. E. J. Phys. Chem. B 1998,102, 7323. 51. Corrent, S.; Cosa, G ; Scaiano, J. C.; Galletero, M. S.; Alvaro, M.; Garcia, H. Chem. Mater. 2001, 13, 715. 52. Zhang, J.; Hu, Y.; Matsuoka, M.; Yamashita, H.; Minagawa. M.: Hidaka, H.; Anpo. M.y. Phys. Chem. B, 2001,105, 8395. 53. Grabowski, R.; Machej, T.; Mazurkiewicz, A.; Slocynski. J.Bull PoL Acad. Sci. Chem. 1987.35. 141. 54. Mills, A.; Giddings, S.: Patel. L; Lawrence, C. J. Chem. Soc.. Faraday Trans. 11987. 83. 2331. 55. Kleijn, J. M.; Boschloo, G. K. J. Electroanai Chem. 1991,300. 595. 56. Mills, A.; Giddings, S.; Patel, I. J. Chem. Soc., Faraday Trans. l1987.83. 2317. 57. Carey. J. G.; Caims, J. F.; Colchester, J. E. J. Chem. Soc. D 1969.1280. 58. Nenadovic, M. T.; Micic, O. L; Adzic, R. R. J. Chem. Soc., Faraday Trans. 11982, 78. 1065. 59. Winiarek, P.; Kijenski, J.J. Chem. Soc. Faraday Trans. 1998,94. 167. 60. Grabowski, R.; Grzybowska, B.; Kozlowska, A.; Sloczynski. J.; Wcislo, K.Top. Catai 1996,3 , 277. 61. Basinska, A.; Domka, F.Catai Lett. 1997,4 3 ,59. 62. Aika, K; Kawahara, T.; Murata, S.; Onishi, T. Buli Chem. Soc. Jpn. 1990,6 3 ,1221. 298 63. Xiang, B.; Kevan, L. J. Phys. Chem. 1994,98, 5120. 64. Miyata, H.; Sugahara, Y.; Kuroda, K.; Kato, C. J. Chem. Soc., Faraday Trans. 1988, 84, 2677. 65. Petr, A.; Dunsch, L.; Koradecki, D.; Kutner, W.J. Electroanal. Chem. Interfacial Electrochem. 1991,300, 129. 66. Impens, N. R. E. N.; Van der Voort, P.; Vansant, E. F. Microporous and Mesoporous Mater. 1999,28, 217. 67. Vankelecom, I. F. J.; Van den Broeck, S.; Merckx, E.; Geerts, H.; Grobet, P.; Uytterhoeven, J. B. J. Phys. Chem. 1996,100, 3753. 68. Kawai, T.; Tsutsumi, K. J. Colloid Int. Sci. 1999,212, 310. 69. Brunei, D.; Bellocq, N.; Sutra, P.; Cauvel, A.; Lasperas, M.; Moreau, P.; Di Renzo, F.; Galameau. A.; Fajula, F.Coord. Chem. Rev. 1998,178-180. 1085. 70. Eabom, C. “Organosilicon Compounds", Academic Press, New York. 1960. 71. Riedmiller. F.; Jockisch, A.; Schmidbaur, H. Organometallics 1998,17.4444. 72. Tour. J. M.; John, J. A.; Stephens, E. B. J. Organomet. Chem. 1992,429. 301. 73. Strohmann, C.; Ludtke, S.; Ulbrich, O. Organometallics 2000,19.4223. 74. Haaf, M.; Schmiedl, A.; Schmedake, T. A.; Powell, D. R.; Millevolte, A. J.; Denk, M.; West, R.J. Am. Chem. Soc. 1998.120, 12714. 75. Cremer, S. E.; Blankenship, C. Organometallics 1986,5, 1329. 76. Perrin, C. L.; Kim, Y.-J. Inorg. Chem. 2000, 39, 3902. 77. Matsumoto. Y.; Ohno, A.; Hayashi, T. 1993, 12,4051. 78. Fraser, C. L.; Anastasi, N. R; Lamba, J. J. S. J. Org. Chem. 1997,62, 9314. 79. Rubinsztajn, S.; Zeldin, M.; Fife, W. K. Synth. React. Inorg. Met. -Org Chem. 1990, 2 0 ,495. 80. Basu, A.; Weiner, M. A.; Strekas, T. C.; Gafiiey, H. D. Inorg. Chem. 1982, 21. 1085. 299 81. a) Maruszewski, K.; Strommen, D. P.; Handrich, K..; Kincaid, J. R. Inorg. Chem. 1991,3 0 , 4579. b) Maruszewski, K.; Strommen, D. P.; Kincaid, J. R.J. Am. Chem. Soc. 1993, 115, 8345. 82. McClanahan, S. F.; Dallinger, R. P.; Holler, P. J.; Kincaid, J. R. J. Am. Chem. Soc. 1985, 107, 4853. 83. Bruce, J. A.; Wrighton, M. S. Isr. J. Chem. 1982, 22, 184. Chapter 4 1. a) Gratzel, M.; Kalyanasundaram, K. “Kinetics and Catalysis in Microheterogeneous Systems", Marcel Dekker, New York, 1991. b) Ramamurthy, V. “Photochemistry in Organized and Constrained Media”, VCH, New York, 1991. 2. Vaidyalingam, A. S.; Coûtant, M. A.; Dutta, P. K. “Electron-transfer Processes in Zeolites and Related Microheterogeneous Media", V. Balzani (Ed), Wiley-VCH, Weinheim, 2001; Vol. 4, pp412. 3. Kalyanasundaram. K. “Photochemistry of Polypyridine and Porphyrin Complexes” Academic Press, London, 1992. 4. a) Krueger, J. S.; Mayer, J. E.; Mallouk, T. E.J. Am. Chem. Soc. 1988.110. 8232. b) Boija, M.; Dutta, P. K. Nature 1993,362,43. c) Sykora. M.; Kincaid. J. R. Nature 1997,3 8 7 ,162. 5. a) Kim, Y. 1.; Mallouk, T. E.J. Phys. Chem. 1992,9 6,2879. b) Yonemoto, E. H.; Kim, Y. 1.; Schmehl, R. H.; Wallin, J. O.; Shoulders. B. A.; Richardson, B. R.; Haw. J. P.; Mallouk, T. E. J. Am. Chem. Soc. 1994. 116. 10557. c) Dutta, P. K.; Turbeville, W.J. Phys. Chem. 1992,96, 9410. 6. a) Dunn, 1. J.; Heinzle, E.; Dettwiler, B.; Saner, U.; Ryhiner, G.; Ingham, J. Chem. Eng. Sci. 1988,43, 1897. b) Bogen, D. K. Science 1989,2 4 6138. , 7. a) Melka, R. P.; Olsen, G.; Beavers, L.; Draeger, J. A. J. Chem. Educ. 1992,69. 596. b) Steffen, L. K.; Holt, P. L. J. Chem. Educ. 1993,70, 991. 8. a) Nicholls, P.; Cooper. C. E.; Wrigglesworth, J. M. Biochem. Cell Biol. 1990,68, 1128. 300 b) Shatkin, J. A.; Brown, H. S. Environ. Res. 1991,56,90. c) Gakis, C.; Cappio-Borlino, A.; Pulina, G. Biochem. Mol. Biol. Int. 1998.46, 487. d) Sutinen. R.; Paronen,’?.; Saano, V.; Urtti, A. Eur. J. Pharm. Sci. 2000, II, 25. e) Aelion, C. M.; Kirtland, B. C. Environ. Sci. Technol. 2000. 34, 3167. 9. Costanza, R.; Duplisea, D.; Kautsky, U.Ecol. Modell. 1998,IlO. 10. a) DeKock, R. L. CHED-1025, Book of Abstracts, 219th ACS National Meeting, American Chemical Society, Washington, D C.. 2000. b) Metz. C. CHED-1021, Book of Abstracts, 219th ACS National Meeting. American Chemical Society, Washington, D C., 2000. c) Soltzberg, L. J. CHED-1020, Book of Abstracts, 219th ACS National Meeting, American Chemical Society, Washington, D C., 2000. 11. a) Vitale. M.; Castagnola, N. B.; Ortins, N. J.; Brooke, J. A.; Vaidyalingam. A.; Dutta, P. K. J. Phys. Chem. B 1999,103, 2408 . b) Castagnola. N. B. Ph. D. Dissertation. The Ohio State University, Columbus. OH. 1999. 12. a) Olson. D. H. J. Phys. Chem. 1970, 74, 2758. b) Rillema. D. P.; Jones, D. S.; Woods. C.; Levy. H. A. Inorg. Chem. 1992.31. 2935. c) Russell. J. H.; Wallwork. S. C.Acta Crystallogr.. Sect. B 1972. 28, 1527. 13. Derycke. 1.; Vigneron, J. P.; Lambin, P.; Lucas, A. A.; Derouane. E. G.J. Chem. Phys. 1991.9 4,4620. 14. Sykora. M.; Kincaid. J. R.; Dutta. P. K.; Castagnola, N. B. J. Phys. Chem. B 1999. 103, 309. 15. a) Lachish, U.; Infelta, P. P.; Gratzel. M. Chem. Phys. Lett. 1979.62, 317. b) Creutz, C.; Chou, M.; Netzel, T. L.; Okumura. M.; Sutin, N.J. Am. Chem. Soc. 1980.102, 1309. 16. Amouyal, E.; Zidlcr, B. Isr. J. Chem. 1982, 22, 117. 17. a) Yonemoto, E. H.; Riley, R. L.; Kim, V. L; Atherton, S. J.; Schmehl, R. H.; Mallouk, T. E.J. Am. Chem. Soc. 1992. 114, 8081. b) Cooley, L. F.; Headford, C. E. L.; Elliott, C. M.; Kelley, D. F. J. Am. Chem. Soc. 1988,7/0,6673. c) Kelly. L. A.; Rodgers, M. A. J. J. Phys. Chem. 1995, 99, 13132. 18. Willner, L; Ayalon, A.; Rabinovitz. M. NewJ. Chem. 1990,14,685. 301 19. Yonemoto, E. H.; Saupe, G. B.; Schmehl, R. H.; Hubig, S. M.; Riley, R. L.; Iverson, B. L.; Maliouk, T.E :^A m . Chem. Soc. 1994,116,4786. 20. Kelly, L. A.; Rodgers, M. A. J. J. Phys. Chem. 1994, 98,6386. 21. Bradley, P. G.; Kress, N.; Homberger, B. A.; Dallinger, R. F.; Woodruff, W. H. J. Am. Chem. Soc. 1981,103, 7441. 22. a) Evans, A. G.; Dodson, N. K.; Rees, N. H. J. Chem. Sac.. Perkin Trans. 1976,2, 859. b) Yoshimura, A.; Hoffman, M. Z.; Sun, H. J. Photochem. Photobioi. A 1993. 70, 29. 23. “An Academic User's Guide to STELLA”, High Perfromance Systems, Inc., Hanover, NH, 1996. 24. Romero, A. J.; Rackley, R. J.; Rhodes, C. T. Chirality 1991,i. 418. 25. Cliristen, T.; Baumann. P. A.; Waldmeier, P. C. Naunyn-Schmiedeberg's Arch. Pharmacol. 1993,348,618. 26. Hoffman, M. Z. J. Phys. Chem. 1988, 92, 3458. 27. Chien, J.-Y. J. Am. Chem. Soc. 1948, 70, 2256. 28. a) Gratzel, M. “Energy Resources Through Photochemistry and Catalysis”. Academic Press, New York, 1983. b) Sauvage, J. P.; Collin, J. P.; Chambron, J. C.; Guilierez. S.; Coudret, C.; Balzani. V.; Barigelletti, P.; De Cola, L.: Flamigni, L. Chem. Rev. 1994. 9 4993. , c) Hoffman, M. Z ; Bolletta, P.; Moggi, L.; Hug, G. L. J. Phys. Chem. Ref. Data 1989,78,219. 29. Rodgers, M. A. J.; Becker, J. C. J. Phys. Chem. 1980,84, 2762. 30. a) Gemborys, H. A.; Shaw, B. R. J. Electroanal. Chem. Interfacial Electrochem. 1986,208, 95. b) Li, Z.; Wang, C. M.; Persaud, L.; Maliouk, T. E.J. Phys. Chem. 1988,92, 2592. 31. Li, Z.; Maliouk, T.E.J. Phys. Chem. 1987,91, 643. 32. a) Gaines, G. L., Ill; O'Neil, M. P.; Svec, W. A.; Niemczyk, M. P.; Wasielewski, M. R. J. Am. Chem. Soc. 1991,113, 719. b) Marcus. R. A. J. Phys. Chem. 1990,9 44963. , 302 c) Chen, P.; Meyer, T. J. Inorg. Chem. 1996, 35, 5520. 33. Jones, W. E., Jr.; Chen, P.; Meyer, T. J. J. Am. Chem. Soc. 1992,114, 387. 34. Hanke, W.; Moeller, K. Zeolites 1984, 4 ,244. 35. Coûtant, M. A.; Le, T.; Castagnola, N.; Dutta, P. K. J. Phys. Chem. B 2000,104, 10783. 36. Buerssner, D.; Steiner, U. E. Coord. Chem. Rev. 1994,132, 51. 37. Sun, H.; Yoshimura, A.; HofFman, M. Z. J. Phys. Chem. 1994, 98, 5058. 38. Ceulemans, A.; Vanquickenbome, L.G.J. Am. Chem. Soc. 1981.103, 2238. 39. Chen, P.; Duesing, R.; Graff, D. K.; Meyer, T. J. J. Phys. Chem. 1991,95, 5850. 40. Harriman. A.; Heitz, V.; Ebersole, M.; van Willigen, H. J. Phys. Chem. 1994, 98, 4982. 41. Helms, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc. 1991,113. 4325. 42. Newton, M. D.; Ohta, K.; Zhong, E.J. Phys. Chem. 1991,9 52317. , 43. Zicovich-Wilson, C. M.; Corma, Viruela, P. J. Phys. Chem. 1994, 98, 10863. 44. Larson. S. L.; Cooley. L. F.; Elliott. C. M.; Kelley, D. F. J. Am. Chem. Soc. 1992, 114, 9504. 45. Yoon. K. B.; Hubig, S. M.; Kochi, J. K. J. Phys. Chem. 1994. 98, 3865. 46. Liu. X.; lu. K.-K.; Thomas, J. K. J. Phys. Chem. 1994, 98, 7877. 47. Alvaro, M.; Garcia, H.; Garcia, S.; Marquez, F.; Scaiano, J. C. J. Phys. Chem. B 1997. 101, 3043. 48. a) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 4 6159. , b) Watanabe, T.; Honda, K. J. Phys. Chem. 1982,8 6 ,2617. 49. Kim. Y. L; Atherton. S. J.; Brigham, E. S.; Mallouk, T. E. J. Phys. Chem. 1993,97, 11802. 303