1 the Excited State Properties of Dirhodium (II,II) Complexes

Home , Phenazine

The Excited State Properties of Dirhodium (II,II) Complexes: Application for Solar

Energy Conversion

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

Congcong Xue

Graduate Program in Chemistry

The Ohio State University

2019

Dissertation Committee

Dr. Claudia Turro, Advisor

Dr. Terry Gustafson

Dr. Yiying Wu

Copyrighted by

Congcong Xue

2019

Abstract

Transition metal complexes have been widely explored for solar energy conversion. The absorption of high energy ultraviolet and blue photons typically results in more energetic excited states, which renders them better reductants and/or oxidants as compared to their ground states. The excited states of dirhodium(II,II) complexes were recently discovered, which may serve as dyes in dye-sensitized solar cell applications owing to their strong, panchromatic absorption throughout the visible and near-infrared spectral regions. These "black dyes" are important to maximize the absorption of the photons from sunlight that reach the earth.

T In the present work, the energy of the triplet excited states, E00 , and the redox potentials of a series of dirhodium complexes with the type cis-[Rh2(μ-

DTolF)2(L)2][BF4]2, where L = np (1; 1,8-naphthyridine), npCOOH (2; 1,8- naphthyridine-2-carboxylic acid), phen (3; 1,10-phenanthroline), dpq (4; dipyrido[3,2- f:2’,3’-h]quinoxaline), dppz (5; dipyrido[3,2-a:2’,3’-c]phenazine), and dppn (6; benzo[i]dipyrido[3,2-a:2’,3’-h]quinoxaline); DTolF = N,N’-di-p-tolylformamidinate), were measured. Complexes 1 – 6 are not emissive, such that their triplet state energies were determined from energy transfer quenching experiments with a series of organic

3 T sensitizers with known * excited state energies, resulting in E00 values estimated to be ~1.1 eV for 1 and 2, ~1.5 eV for 3, and ~1.2 eV for 4 – 6 (vs Ag/AgCl). ii

The reductive excited states of the dirhodium complexes make them potential dyes for n-type dye-sensitized solar cell (DSSC) applications. For this purpose, two new

Rh2(II,II) dyes with methyl ester anchoring groups, [Rh2(DTolF)2(menp)2][BF4]2 (7; menp = 4-carbomethoxy-1,8-naphthyridine) and [Rh2(DTolF)2(dmeb)2][BF4]2 (8; dmeb =

4,4ʹ-carbomethoxy-2,2ʹ-bipyridine) were synthesized and their photophysical properties were characterized and compared. Femtosecond transient absorption (fsTA) and time- resolved infrared (fsTRIR) spectroscopies reveal that the initially populated 1ML-LCT excited states of 7 and 8 decay to the corresponding 3ML-LCT excited states with time constants of 4 ps and 2.5 ps, respectively. The 3ML-LCT excited states of 7 and 8 repopulate the ground state with lifetimes of 460 ps and 56 ps, respectively. The shorter

1ML-LCT and 3ML-LCT lifetimes of 8 as compared to those of 7 are attributed to the longer Rh-Rh bond in the former, which provides a fast deactivation pathway through a metal-centered (MC) state that involves population of the Rh2(*) molecular orbital.

Photoinitiated electron injection into the semiconductor TiO2 upon low energy light irradiation was achieved through the excitation of Rh2(II,II) dyes associate to TiO2 nanoparticles through the methyl-ester substituent (7@TiO2 and 8@TiO2) and ultrafast electron injection was observed by fsTRIR with low energy excitation, 600 nm for 7 and

520 nm for 8.

2+ Complex 11, Rh2(DPhF)2(bncn)2 (form= diphenylformamidinate, bncn = benzo[c]cinnoline), with a shorter bridging bncn ligand was synthesized and characterized. The crystal structure indicates that 11 has a shorter Rh-Rh bond length compared to 1 – 8, which leads to a longer triplet excited state lifetime, T ~ 19 ns. Bulk

iii electrolysis of 11 with trifluoromethanesulfonic acid shows H2 production with 98%

Faradaic efficiency. The highly oxidative and nanosecond long triplet excited state of 11, together with its catalytic active bimetallic core, result in photocatalytic activity.

Irradiation at 670 nm in acidic solutions with a sacrificial electron donor, 1-benzyl-1,4- dihydronicotinamide (BNAH), results in the catalytic production of hydrogen exceeding

170 turnovers (TON) in 24 hours with an initial rate of 28 TONs per hour. The catalysis proceeds through two stepwise excited state redox events, a feature previously unknown in homogeneous photocatalysis, which permits the storage of two redox equivalents on the dirhodium catalyst using low energy light with high efficiency. Lastly, the important intermediate, the one-electron reduced complex [11]1−, exhibits a triplet lifetime T~ 0.5 ns and Ered* ~ +0.66 eV, which makes the second electron transfer event favorable with

BNAH as an electron donor to support the proposed mechanism.

Dedication

For my parents. Thanks for all your love and supports.

Acknowledgments

First and foremost, I’d like to thank my advisor, Professor Claudia Turro for her guidance and enlightenment. I’m so grateful that she always helps me see the positive part when I encounter any difficulties and encourages me when I doubt myself. I’d like to thank her for shaping me into a more confident version of myself. Claudia gave me a lot of opportunities to collaborate with amazing people and work on interesting projects. She also allowed me to travel around the country to share my research results, talk to fantastic people and get inspired. I genuinely appreciate all my incredible experience in the past five years as a Turro group member.

I also want to acknowledge the amazing group. Dr. Tyler Whittemore, you’re like a mentor and role model to me. Working with you is such an incredible experience, you’re so brilliant yet hardworking and your positivity keeps influencing me. You taught me all the laser techniques, answered so many of my questions and guided me through a lot of tough situations. Dr. Hannah Sayre and Dr. Suzanne Witt, you’re like big sisters to me in the group and taught me everything about catalysis and electrochemistry. Hannah influences me not only as a scientist but also as a human being who cares about the world around us, environment, women’s rights and so much more fun topics you brought up with. Suzanne, thank you for laying such a good foundation on the energy side of the group and the talk with you about future career. Dr. William Kender, you’re so smart and vi just like a walking Wikipedia and thank you for all the enlightenment when I lost the direction. Dr. Travis and Jessica White, you’re the first mentors I had in group. I also appreciate the opportunity to collaborate with Travis when he’s a professor at OU.

Lauren, you’ve always been super helpful, thoughtful and sweet. Jie, thanks for taking your own time and synthesizing all the complexes for me. To my current and past coworkers, Regina, TJ, Malik, Katy, Camila, Hemanthi, Austin, Sean, Jessica, Matt,

Massiel, Shaoyang and Allen. It’s been such a pleasure to work with all of you guys.

Thanks for my dearest friends Juan, Fen, Celia and Jiahui, without you girls grad school wouldn’t be so enjoyable. To my Mon and Dad, thanks for encouraging me to go aboard and pursue my PhD. I cannot go through these alone without your support and video calls every weekend.

vii

Vita

June 2014 ...... B.S. Chemistry, Beijing University of

Chemical Technology, Beijing, China

2014 to present ...... Graduate Research and Teaching Associate,

Department of Chemistry, The Ohio State

University

Publications

C. Xue, H. J. Sayre, C. Turro, Chem. Comm., 2019, DOI: 10.1039/c9cc04677a

S. Saeedi, C. Xue, B. McCullough, S. Roe, B. Neyhouse, T. White, ACS Appl. Energy

Mater., 2019

T. N. Rohrabaugh Jr., K. A. Collins, C. Xue, J. K. White, J. J. Kodanko, C. Turro, Dalton

Trans. 2018, 47, 11851

T. J. Whittemore, A. Millet, H. J. Sayre, C. Xue, B. S. Dolinar, E. G. White, K. R.

Dunbar, C. Turro, J. Am. Chem. Soc., 2018, 140, 5161

T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci, C. Turro, J. Am.

Chem. Soc., 2017, 139, 14724

S. Chen,† W. Zhou,† Y. Cao, C. Xue, and C. Lu, J. Phys. Chem. C, 2014, 118, 2851 viii

Fields of Study

Major Field: Chemistry

Table of Contents

Abstract ...... ii Dedication ...... v Acknowledgments...... vi Vita ...... viii Table of Contents ...... x List of Tables ...... xiii List of Figures ...... xiv Chapter 1. Introduction and Background ...... 1 1.1 Solar Energy Need ...... 1 1.2 Dye Sensitized Solar Cell ...... 4 1.2.1 Operational Principle ...... 4 1.2.2 Dyes or Sensitizers ...... 6 1.2.3 Electron Transfer Dynamics ...... 9 1.3 Hydrogen Fuel ...... 12 1.3.1 Hydrogen Production and Storage ...... 12 1.3.2 Biological Hydrogen Production ...... 15 1.4 Metal-metal bonds ...... 19 1.4.1 History of Metal-metal Bonding ...... 19 1.4.2 Dirhodium Complexes ...... 22 Bibliography ...... 26 Chapter 2. Experimental Methods ...... 31 2.1 Materials ...... 31 2.2 Synthesis and sample preparation ...... 31

2.2.1 Synthesis of [Rh2(DTolF)2(-NN)2][BF4]2 (NN= menp (7) and dmeb (8)) ..... 31 2.3 Methods and Instrumentation ...... 35 x

2.3.1 General ...... 35 2.3.2 Electrochemistry ...... 36 2.3.3 Time-resolved spectroscopies ...... 38 2.3.4 Photolysis ...... 41 Bibliography ...... 42 Chapter 3. Dirhodium Complexes: Excited State Properties and Electron Transfer .... 43 3.1 Background ...... 43 3.2 Results and Discussion ...... 46 3.2.1 Excited State Quenching of Organic Sensitizers ...... 46 3.2.2 Quenching Mechanism ...... 50 3.2.3 Excited State Electron Transfer ...... 53 3.3 Conclusions ...... 56 Bibliography ...... 57

Chapter 4. Panchromatic Dirhodium Photosensitizers and Electron Injection to TiO2 59 4.1 Background ...... 59 4.2 Results and Discussion ...... 60 4.2.1 Electronic Absorption and Electrochemistry ...... 60 4.2.2 Excited State Properties of the Dirhodium Dyes ...... 63

4.2.3 Electron Injection to TiO2 ...... 70 4.3 Conclusions ...... 75 Bibliography ...... 75

Chapter 5. Single Molecule Dirhodium Photocatalyst for H2 Generation Using Low Energy Light ...... 78 5.1 Background ...... 78 5.1.1 Multicomponent Photocatalytic System ...... 78 5.1.2 Single Molecule/Chromophore Photocatalysts ...... 82 5.2 Results and Discussion ...... 85 5.2.1 Characterization ...... 85 5.2.2 Electrocatalytic Proton Reduction ...... 94 5.2.3 Photocatalytic Proton Reduction...... 99 5.2.4 Mechanism Studies ...... 101 5.2.5 Photophysical and photochemical properties of [11]1− ...... 107 5.3 Conclusions ...... 115 xi

Bibliography ...... 116 Chapter 6. Concluding Remarks and Future Work ...... 121 Bibliography ...... 125 Bibliography ...... 126

xii

List of Tables

Table 1.1 Electronic absorption maxima, abs, extinction coefficient, , and reduction potentials for 4-6 in CH3CN...... 24

Table 3.1 Selected organic sensitizers and E(3ππ*)...... 47

Table 3.2 Quenching constants for 1 and 2 with organic EnDs ...... 50

Table 3.3 Quenching constants for 3 - 4 with organic EnDs ...... 50

Table 3.4 Reduction potential, E(EnD0/−), excited state reduction potential, E(EnD*/−) and driving force for reductive quenching ΔGRQ with 1, 2 and 4...... 52

Table 3.5 Oxidation potential, E(EnD+/0), excited state oxidation potential, E(EnD+/*) and driving force for oxidative quenching ΔGOQ with 1, 2 and 4...... 52

Table 4.1 Electronic Absorption Maxima, Extinction Coefficients, and Reduction Potentials of 1, 7, 8 and 9 in CH3CN...... 62

Table 5.1 Electronic Absorption and Electrochemical Potentials of 11 in DMF ...... 87

Table 5.2 Control experiments for each component and hydrogen production...... 101

Table 5.3 Energy transfer donors and their 3* excited state energy (E(3*)), lifetime (0), excitation energy (exc) and energy transfer quenching constant (KEnT) ...... 110

Table 5.4 Electron donor, peak anodic potential (Epa) and 5% maximum current onset oxidation (Eonset) vs. Ag/AgCl in 0.1 M TBAPF6 ...... 114

Table 6.1 Complex 1, 2, 8 and 11, lowest absorption maxima, extinction coefficient, diimine ligands binding mode, Rh-Rh bond length, axial coordination, singlet and triplet excited state lifetimes...... 124

xiii

List of Figures

Figure 1.1 Solar flux as a function of wavelength from the UV to the near-IR and integrated photon flux (dotted line).6 ...... 3

Figure 1.2 Mechanism of photosensitization in dye-sensitized solar cells: (a) n-type DSSCs (b) p-type DSSCs...... 5

Figure 1.3 The components of a typical metal complex photosensitizer...... 7

Figure 1.4 N3 dye and corresponding electronic absorption spectrum...... 8

Figure 1.5 State diagram representation of the kinetics of n-type DSSC function with − − I /I3 as the redox mediator. Recombination processes are indicated by orange arrows. . 10

Figure 1.6 Hydrogen generation pathways and the near/mid/long-term goals in the future. CCS: CO2 capture and storage. PEC: photoelectrosynthesis cell. STCH: solar thermochemical hydrogen production...... 13

Figure 1.7 Direct biophotolysis. (H2ase: hydrogenase) ...... 16

Figure 1.8 Indirect biophotolysis...... 16

Figure 1.9 Photofermentations by photosynthetic bacteria (N2ase: nitrogenase)...... 17

Figure 1.10 Diagram of the overlaps of d-orbitals and the resulting energy levels as they are involved in the formation of M-M multiple bonds in a X4M-MX4 structure. In practice, the ordering of the orbitals, especially those having antibonding nature might differ...... 20

Figure 1.11 M-M bond order change by removal of d-electrons or addition of antibonding electrons. Orbital ordering may change for antibonding orbitals...... 21

Figure 1.12 Structure of the Rh2(O2CCH3)4(L)2 complexes (L = H2O, py, PPh3; py = pyridine, PPh3 = triphenylphosphine)...... 22

2+ Figure 1.13 Structure of cis-[Rh2(-DTolF)2(NN)2] , DTolF = N,N’-di-p- tolylformamidinate, NN = chelating diimine ligands, dpq (4), dppz (5), dppn (6)...... 24

xiv

1 Figure 2.1 H NMR of 7 in DMSO-d6 ...... 32

1 Figure 2.2 H NMR of 8 in CH3CN-d3 ...... 33

Figure 2.3 Electronic absorption spectra collected before (black) and after (red) anchoring of (a) 7 and (b) 8 to TiO2 nanoparticles...... 34

Figure 2.4 Bulk electrolysis cell ...... 38

Figure 2.5 Four of the 670 nm LEDs on an acrylic plate ...... 42

Figure 3.1 Structures of cis-[Rh2(μ-DTolF)2(L)2][BF4]2, L= bridging diimine ligand, np (1) and npCOO (2), or L = chelating diimine ligand phen (3), dpq (4), dppz (5), dppn (6)...... 45

Figure 3.2 Energy transfer mechanism of the organic sensitizers (S) and dirhodium complexes (Rh2)...... 46

Figure 3.3 Kinetic traces of the decay of the 3* state of anthracene monitored at 420 nm as a function of [1]...... 49

Figure 3.4 Stern-Volmer plot of anthracene with 1...... 49

III Figure 3.5 Structure of Co (dmgH)2(py)(Cl)...... 53

Figure 3.6 Transient absorption spectra comparison of 0.25 mM 2 (red) and 0.14 mM 2 III with 5 mM Co (dmgH)2(py)(Cl) (blue) in CH3CN at 24 ns after laser pulse (exc= 532 nm, 5 mJ/pulse) ...... 55

III Figure 3.7 Transient absorption spectrum of 0.13 mM 2 and 0.5 mM Co in CH3CN at 86 II ns after excitation (exc= 532 nm, 5 mJ/pulse), which is assigned to Co ...... 55

III Figure 3.8 Kinetic trace at 650 nm of 0.13 mM 2 and 0.5 mM Co in CH3CN (exc= 532 nm, 5 mJ/pulse)...... 56

Figure 4.1 Schematic representation of Rh2(II,II) complexes ...... 60

Figure 4.2 Electronic absorption spectra of 7 (solid) and 8 (dashed) in CH3CN ...... 62

Figure 4.3 fsTRIR spectra of (a) 7 and (b) 8 in CD3CN (λexc = 600 nm, 2µJ) and corresponding ground state IR spectra (dashed lines)...... 65

Figure 4.4 FsTA spectra of (a) 7 (λex = 600 nm, 2µJ) and (b) 8 (λexc = 520 nm, 2µJ) in CH3CN ...... 66

Figure 4.5 Spectroelectrochemistry of 8 at applied potential of +0.137 V (blue spectrum) and −0.570 V (red spectrum) in 0.1 M Bu4NPF6 CH3CN ...... 68

Figure 4.6 Modified Latimer Diagram for 7 (a) and 8 (b) ...... 70

Figure 4.7 fsTRIR of (a) 7@TiO2 (2 J, exc = 600 nm) and (b) 8@TiO2 in CD3CN (2 J, exc = 520 nm) ...... 71

-1 Figure 4.8 Kinetic trace of 7@TiO2 (a) at 1787 cm within 5 ps after excitation (b) at 1610 cm-1 with long decay components...... 74

Figure 5.1 Traditional multicomponent photocatalysis system for proton reduction shown in oxidative quenching pathway...... 79

Figure 5.2 Structure of 11, where solvent in the axial positions are omitted...... 86

Figure 5.3 Electronic absorption spectrum of 11 in CH3CN...... 87

Figure 5.4 Emission and excitation spectra at 77K of 11 in 10:1 H2O:EtOH glass (black) and CH3CN (red) where the excitations (dotted lines) are monitored at 750 nm and the emissions (solid lines and done irradiating with 515 nm...... 88

Figure 5.5 Cyclic voltammogram of 11 in DMF vs Ag/AgCl (calibrated with ferrocene) scan rate 200 mV/s...... 89

Figure 5.6 (A)Femtosecond transient absorption of complex 11 (650 μM) in acetonitrile (630 nm excitation; 2.5 μJ) (B) Kinetic trace at 440 nm (hollow circles) and 650 nm (solid circles) ...... 91

Figure 5.7 A.) Nanosecond transient absorption of 1 in DMF when excited with 630 nm light (5 mJ/pulse) and B.) kinetic trace taken at 420 nm...... 92

Figure 5.8 Nanosecond transient absorption of (A) 11 in 15 mM PTZ acetonitrile solution taken at 30 μs (B) and kinetic trace taken at 420 nm ...... 93

Figure 5.9 LSV upon addition of dimethylformamidinium triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11 ...... 96

Figure 5.10 LSV upon addition of p-toluenesulfonic acid triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11 ...... 96

Figure 5.11 LSV upon addition of trifluoroacetic acid in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11 ...... 97

Figure 5.12 LSV upon addition of acetic acid in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11 ...... 97 xvi

Figure 5.13 CV upon addition of dimethylformamidinium triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11 ...... 98

Figure 5.14 CV upon addition of dimethylformamidinium triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of benzo[c]cinnoline ligand ...... 98

Figure 5.15 Turnover number (TON) versus time graph of 11 (75 μM) irradiated with 670 nm LEDs for 24 hours in the presence of BNAH (0.03 M) and TsOH (0.1 M) produces over 170 equivalents of H2...... 99

Figure 5.16 Turnover number (TON) versus time graph of 11 (75 μM) irradiated with 735 nm LEDs for 24 hours in the presence of BNAH (0.03 M) and TsOH (0.1 M) produces over 70 equivalents of H2...... 101

Figure 5.17 Reduced 11 ([11]1−) and doubly reduced 11 ([11]2−) chemically generated through the addition of 1 and 2 equivalents of cobaltocene in DMF ...... 102

Figure 5.18 Difference spectrum associated with A.) reduced 11 ([11]1−) and B.) doubly 2− reduced 11 (([11] )) generated via electrochemically in 0.1 Bu4NPF6 in dry DMF ..... 103

Figure 5.19 Difference in electronic absorption spectrum of catalytic mixture (0.1 mM 11, 0.1 M TsOH, 0.03 M BNAH in DMF) at A.) early times and B.) late times (bottom) ...... 104

Figure 5.20 Conversion of 1-electron species ([11]1−) to 2-electron species [11]2− upon irradiation with 670 nm light ...... 105

Figure 5.21 Addition of TsOH (0.1 M) to photochemically generated 2-electron species (65 μM 11, 0.2 M BNAH in DMF) (blue) yields ground state complex (black) ...... 106

Figure 5.22 Proposed mechanism of the photocatalytic H2 production by 11 ...... 106

1− Figure 5.23 fsTA (λex = 650 nm, irf = 85 fs) of [11] after reduction with one equivalent of cobaltocene showing (A) various traces (smoothed) and (B) kinetic trace centered at 428 ± 3 nm showing biexponential decay...... 108

Figure 5.24 A. Kinetic traces at 460 nm (exc = 470 nm, 5 mJ/pulse). B. Transient absorption spectra at 1s for tetracene (red) and tetracene with 50 mM [11]− added (blue) in CH3CN...... 110

− Figure 5.25 Photolysis of [11] with 30 mM A. DPA and B. TPA in CH3CN (irr = 670 nm) at various irradiation times...... 112

Figure 5.26 UV-Vis monitored photolysis of [11] in the presence of A. TEA (30 mM) and B. TEOA (30 mM) in CH3CN at various irradiation times...... 113

xvii

Figure 5.27 LSV of electron donor 0.1 M TBAPF6 in CH3CN vs Ag/AgCl...... 114

xviii

Chapter 1. Introduction and Background

1.1 Solar Energy Need

The increasing demand of energy and the limitation of fossil fuels urge us to find renewable, sustainable, efficient and clean energy sources. In 2018, the United State primary energy consumption reached the highest record of 101.3 quadrillion British thermal unit (Btu; 1 Btu = 1.055 × 103 joules), increased by 4% compared to 2017 and up

0.3% from the previous record in 2007.1 80% of the U.S. energy consumption was from fossil fuels: petroleum, natural gas, and coal. Coal consumption declined by 4% from

2017. However, it was offset by the 10% increase in natural gas consumption, together with smaller increases in the petroleum, renewable energy and nuclear electric power.

Driven by the tremendous efforts in improving the efficiency of renewable energy, its consumption reached a record high of 11.5 quadrillion Btu in 2018, increased by 3% from 2017. Biomass consumption accounted for 45% of all renewable energy consumption, mostly in the form of transportation fuels including ethanol and biodiesel.

The primary increments result from wind electricity and solar power, rose by 8% and

22%, respectively.2 To weaken the energy reliance on fossil fuels, the continuous transition to renewable energy source is imperative.

Fortunately, there is an enormous supply of energy from the sun to the earth, about 10,000-fold greater than the entire current global energy consumption in a year.3 1

Solar energy application has benefited fields such as dye-sensitized solar cells (DSSCs), dye-sensitized photoelectrosynthesis cells (DSPECs), and artificial photosynthesis.

Among various dyes or photosensitizer, transition metal complexes have received extensive attention for solar energy conversion due to their synthetic tunability, rich redox chemistry and long-term stability. The absorption of high energy ultraviolet and blue photons typically results in more energetic excited states, which renders them better reductants and/or oxidants as compared to their ground states. However, efficient collection of sunlight remains a challenge, in part because not all incident photons are harvested by current photosensitizers, such that much of the sunlight remain unused.4 The energy required to split water under standard condition is 1.23 eV (~1,000 nm).5

Although the solar spectrum that can potentially be used by sensitizers extends broadly from UV to the near-IR, traditional dyes typically do not absorb low energy photons in the red and near-IR regions, which limits the potential incident photon to current efficiency (IPCE) of a given solar cell. The area shaded in red in Figure 1.1 shows the range where the traditional dyes absorb light, while the area shaded in gray is the result of the extension of the absorption from 650 to 900 nm. The integration of both red and gray areas represents a 2-fold increase as compared to that of the red area alone in the number of photons available, and thus the theoretical IPCE that can be reached. panchromatic dyes are often limited by their low excited state oxidation and reduction potentials, making charge injection unfavorable. Therefore, a need remains for stable dye molecules with a broad absorption range and favorable excited state redox properties to inject electrons or holes to n- or p-type semiconductors, respectively.

Figure 1.1 Solar flux as a function of wavelength from the UV to the near-IR and integrated photon flux (dotted line).6

Herein, the goals of the present work are to develop new dyes that absorbs broadly from UV to infrared region for solar cell applications and new photocatalysts for the cathodic half reaction of water splitting (equation 1.1). Dirhodium(II,II) complexes with energetic excited states were recently discovered, which may serve as dyes in solar cell applications owing to their strong, panchromatic absorption throughout the visible and near-infrared spectral regions.6-9 These "black dyes" are important to maximize the absorption of the photons from sunlight that reach the earth. Furthermore, these complexes possess rich redox chemistry and catalytic active metal centers, which were recently explored as electrocatalysts.10,11 Together with their energetic excited states,

3 synthetic modification may result in new single-molecule/single-chromophore photocatalysts for proton reduction without additional photosensitizer.

+ − 2 H + 2 hv + 2 e → H2 (1.1)

1.2 Dye Sensitized Solar Cell

1.2.1 Operational Principle

Solar light has been widely applied in the fields of dye-sensitized solar cells

(DSSCs) and dye-sensitized photoelectrosynthesis cells (DSPECs) to convert solar energy to electrical charge and to store energy in chemical bonds.6,12-28 The Grätzel-type

DSSCs utilize a wide-band gap semiconductor modified with an inorganic dye

(sensitizer) to absorb photons.14

The operation mechanisms of n-type and p-type DSSCs are shown in Figure 1.2, where S, D and A represent sensitizer, electron donor and acceptor, respectively. Ru- complexes have been studied extensively as sensitizers since the early days. Since then, other inorganic dyes such as osmium and rhenium complexes, and organometallic complexes including porphyrins and phthalocyanines have been explored. The electron donor and acceptor system which is also known as the redox mediator, such as iodide and

− − − − − − triiodide couple (I /I3 ), SCN /(SCN)3 and SeCN /(SeCN)3 , is used to regenerate the sensitizer after electron/hole injection.29-31

Figure 1.2 Mechanism of photosensitization in dye-sensitized solar cells: (a) n-type DSSCs (b) p-type DSSCs.

In n-type DSSCs, photoexcited dyes (S*) inject electrons into the conduction band

(CB) of the n-type semiconductor, resulting in the generation of current on the photoanode. The oxidized dye (S+) is subsequently reduced by the redox mediator in the organic electrolyte. The regeneration of the sensitizer also prevents the electron recapture by the oxidized dye, S+, from the CB.2 Similarly, in p-type DSSCs, photoexcited dyes that are good oxidizing agents may inject holes to the valance band (VB) of p-type semiconductors on the photocathode. The restoration of the reduced dye (S−) is facilitated by the electron acceptor of the redox mediate couple in the electrolyte, which also intercepts the electron back transfer from the reduced dye, S−, to the VB.3

The relative positions of the different energy levels for semiconductor, dye and the redox mediator are essential to the function of DSSCs.4 The energy of the excited state oxidation potential (Eox*) of the dye (or sensitizer, S) should be higher than the CB 5 edge of n-type semiconductor, so that the efficient electron transfer between the excited dye (S*) and CB is thermodynamically favorable. For the dye regeneration, the reduction potential of oxidized dye (E S•+/0) should be more positive than the oxidation potential of electron donor (E D•+/0) in the redox relay. In contrast, for p-DSSCs, the reduction potential of the excited dye (Ered*) should be at more positive potential than the VB level of p-type semiconductor. The reduction potential of the electron acceptor (E A0/•−) should be more positive compared to the oxidation potential of the reduced dye (E S0/•−).

1.2.2 Dyes or Sensitizers

The sensitizer represents a crucial component of DSSCs and should possesses the following characteristics for efficient function: (1) absorbs broadly from the ultraviolet

(UV), visible, and into the near-IR spectral ranges, (2) strong anchoring group for attachment to p-/n-type semiconductor for hole/charge injection (e.g. -COOH, -H2PO3, -

SO3H), (3) suitable excited state redox potential for hole/charge injection into the valance or conduction band of the p- or n-type semiconductor, (4) redox potential should be

− − compatible for regeneration by the redox couple in solution (e.g. I /I3 ), (5) minimal charge recombination, (6) no aggregation on the surface of the semiconductor, and (7) thermal and electro-/photochemical stability over long times for commercial applications.

Figure 1.3 The components of a typical metal complex photosensitizer.

Among various dyes, transition metal complexes have been extensively applied as sensitizers in DSSCs due to their relative broad absorption spectra and favorable excited state photophysical and redox properties. Generally, metal complex photosensitizers are composed of metal center(s) with ancillary ligands and at least one anchoring group to bind to semiconductor (Figure 1.3).9 The ancillary ligands can be synthetically modified to tune the energy levels, excited state energy and lifetimes, and redox properties of the sensitizer, while the anchoring group can facilitate the electron transfer between the dye and the semiconductor.

Thiocyanate-based Ru(II) polypyridyl complexes have been investigated extensively since their introduction in 1993 owing to their relatively broad absorption range compared with other Ru(II) polypyridyl dyes, as well as their suitable redox and excited state properties (Figure 1.4).32 The molecular structure of the representational N3 dye, Ru(dcbpyH)2(NCS)2 (dcbpyH = 4,4ʹ-carboxy-2,2ʹ-bipyridine, NCS = isothiocyanate), and its absorption spectrum are shown in Figure 1.4. The intense visible

7 absorption bands have been assigned to singlet metal-to-ligand charge-transfer (1MLCT) transitions from the t2g d orbitals localized on Ru metal center to * orbitals on the carboxylated bipyridyl ligand. Thus, the excitation process promotes electrons from the metal center to the carboxylated bipyridyl ligand, which generates a formally oxidized

Ru(III) and a bipyridine radical anion.33 A solar-to-electricity efficiency of 10% was first obtained for the N3 dye sensitized system.34 The doubly protonated form of N3 dye,

(Bu4N)2[Ru(dcbpyH)2(NCS)2] (N719), exhibited an improved performance on power conversion. The N3 dye is considered as the reference sensitizer in the field and used as a guide to design other Ru-based dyes. However, the instability of the NCS– donor ligands and the limited ability to harvest photons beyond 600 nm reduce their performance.35

Figure 1.4 N3 dye and corresponding electronic absorption spectrum.

Porphyrin and phthalocyanine systems are known for their intense spectral response in the near-IR region, but they suffer from unfavorable aggregation on the

8 semiconductor surface.36 Organic dyes, such as coumarin and indoline,37,38 are cost- efficient alternatives to ruthenium dyes, but their narrow absorption features and cumbersome synthetic routes for modification represent drawbacks. Recently, two- photon absorption materials have gained significant attention. By absorbing two low energy photons simultaneously, the ground state molecule can be excited. Nonetheless, this feature requires intense light irradiation which is not suitable for harvesting ambient solar light.

1.2.3 Electron Transfer Dynamics

Understanding the electron transfer dynamics at the sensitizer-semiconductor nanoparticle interface is crucial to improve the solar energy conversion as well as to aid in the development of new systems. This includes the charge injection to the semiconductor, excited state decay, migration within the semiconductor, regeneration of the ground-state dye, and back electron transfer or charge recombination (Figure 1.5).6, 39

The forward pathways are shown in blue arrows while the competing loss pathways are shown in orange.

Figure 1.5 State diagram representation of the kinetics of n-type DSSC function with − − I /I3 as the redox mediator. Recombination processes are indicated by orange arrows.

Various spectroscopic measurements can provide insights into these processes.

Transient absorption (TA) spectroscopy probing visible and near-IR region follows the electronic transitions of various states of the dye adsorbate (ground and excited state, reduced and oxidized form), and the signals arising from the electrons/holes injected into the semiconductor. However, the board electronic absorption features originating from various states of the dye adsorbate can make it difficult to discern each species and monitor their dynamics independently.

Transient absorption spectroscopy probing the mid-IR region detects the vibrational transitions of various forms of the dye adsorbate. To be noted, free carriers generate broad absorption in mid-IR which makes it easier to study the dynamics of the electron signal without the superposition of the various forms of the dye adsorbate.40-42

For example, from the prior studies using Ru-based N3-sensitized solar cells,7,8 the

10 kinetics of the mid-IR absorption band reveal biphasic electron injection process with an ultrafast electron injection (< 100 fs) from the unrelaxed, locally-excited singlet excited state and a slower component on 1-100 ps from the triplet excited state.

For a DSSC to be efficient, it is not important that the electron injection is ultrafast, but it is crucial that the quantum yield of the injection (휑inj) is high, expressed in eq. 1.1, where kinj and ko represent the rate constants of electron injection into the semiconductor and excited state decay, respectively.13

푘푖푛푗 휑푖푛푗 = (1.1) 푘푖푛푗+ 푘표

Efficient dye regeneration requires the rate for reduction of the oxidized dye with the redox mediator in electrolyte to be faster than the recombination with the injected electron in the CB. This recombination is shown to be dependent on the electron density in the CB,43 spatial separation of the highest occupied molecular orbital (HOMO) of the dye cation from the semiconductor surface,44 electrolyte concentration, and viscosity and dye structure.45 Furthermore, efficient charge collection requires the charge transport to the electrodes to be faster than the charge recombination processes mentioned above. The typical time constants for these steps are shown in Figure 1.5.39

1.3 Hydrogen Fuel

1.3.1 Hydrogen Production and Storage

H2 is a of desirable energy source because it possesses the highest energy content of any common fuel by weight.46-50 It can be used in fuel cells to generate power through chemical reactions, producing only water and heat as products, thus making it a clean fuel that does not release molecules harmful to the environment upon combustion, such as

CO2. Fuel cells function like batteries; however, they can operate as long as supplied with fuel and oxidants (usually oxygen from air) without recharging process. Typically, fuel cells are composed of anode, cathode and electrolyte sandwiched in between them.51

Based on the type of electrolytes, fuel cells are defined as different types, such as proton exchange membrane fuel cells (PEMFCs), phosphoric acid fuel cell (PAFC), solid acid fuel cell (SAFC) and alkaline fuel cells (AFC). At the anode, a catalyst, usually fine platinum powder, oxidizes hydrogen, separating it into protons and electrons, which take different pathways to the cathode. Electrons are transported through an external circuit, creating electricity, while protons travel through the electrolyte. At the cathode, protons, electrons and oxygen combine to produce water. Fuel cells can be used to power electric cars, appliances in homes, and to provide electricity to a number of other applications.

Currently, most of the commercial hydrogen is produced from a process known as reforming, which separates hydrogen from hydrocarbons, including gasoline, natural gas and propane, through heat. Hydrogen can also be generated by electrolysis, a process in which an electrical current is applied to water to separate it into hydrogen and oxygen.

Some algae and bacteria can photobiologically produce hydrogen by only using solar 12 light under certain conditions. These different hydrogen generation pathways are shown in Figure 1.6.52

Figure 1.6 Hydrogen generation pathways and the near/mid/long-term goals in the future. CCS: CO2 capture and storage. PEC: photoelectrosynthesis cell. STCH: solar thermochemical hydrogen production.

The current challenge to scalable hydrogen production is the cost of current methods. As such, the Department of Energy aims at the development of technologies that can produce H2 at a cost of less than $4/gge (gge = gasoline gallon equivalent; the amount of energy equivalent to a gallon of gasoline). In the short term, this goal can be met by reforming from natural gas through existing facilities, however, natural gas is not a renewable source. In order to produce H2 economically and in a manner that is environmentally friendly and sustainable, renewable technologies are required. Moving forward, hydrogen production from biomass and wind-generated electricity is expected to be viable, but in long run, the utilization of free, abundant solar energy to split water with 13 zero carbon emissions are needed to meet the goals of cost, clean energy, and sustainable sources. Solar technologies currently being explored include solar thermochemical hydrogen (STCH), dye-sensitized photoelectrosynthesis cells (DSPECs), electrolysis, and photobiological pathways. Among the various hydrogen production methods available, the production of hydrogen using solar energy is of special interest to the present work.

We focus on the cathode half reaction of water splitting, equation 1.2.

+ − 2 H + 2 hv + 2 e → H2 (1.2)

Hydrogen storage is also a crucial factor to advance H2 as a fuel in transportation, as well as in stationary and portable power. The low ambient temperature density of hydrogen results in low energy per unit volume, although each molecule has highest energy per mass of any fuel. By mass, H2 provides three times more energy as compared to gasoline, 120 MJ/kg vs 44 MJ/kg, respectively. However, on a volume basis, the situation is reversed. Liquid hydrogen (achieved by high pressure and low temperature) has 9 MJ/L of energy, while gasoline has an energy density of 32 MJ/L.46,47

Hydrogen storage can be realized by physical-based methods, such as compressed gas or cryogenically cooled liquid H2. One of the drawbacks of these methods is the high energy compensation for storage because up to 20% and 40% of the energy content is required to compress or liquify hydrogen, respectively. Therefore, technologies to advance hydrogen storage methods to increase its energy density at standard temperature and pressure is of great interest. In search of major technology breakthrough, material- 14 based methods including adsorbent, complex hydride, and chemical hydrogen have been developed.48 An interesting alternative method is to produce synthetic hydrocarbons from hydrogen and CO2 extracted from the atmosphere. The storage of these fuels is much like

47 gasoline and are CO2 neutral.

1.3.2 Biological Hydrogen Production

Biological hydrogen production processes allow the conversion of water or biomass to hydrogen. More importantly, they are less energy demanding, more environment friendly comparing to thermo/electro-chemical processes. Biological hydrogen production usually involves photosynthetic or fermentative organisms. The main pathways include direct and indirect biophotolysis, photofermantation, dark fermentation, or the combination of these processes.53-57

Direct biophotolysis is similar to photosynthesis found in plants and algal but differs in the generation of hydrogen instead of biomass, shown in equation 1.3 and

Figure 1.7. In green plants, photosynthesis involves light absorption by photosystem I and II, where two photons are used to split each electron in H2O and then used in CO2 reduction to store energy in biomass. Due to the lack of hydrogen generation enzyme, hydrogenases, only CO2 reduction takes place and no hydrogen is formed. In contrast, microalgae with hydrogenases (such as green algae) can produce hydrogen under specific conditions. However, hydrogen production efficiency is lower than CO2 reduction arising

54 from the small amount of O2 that inhibits the activity of the hydrogenase.

2 H2O + ‘light energy’ → 2 H2 + O2 (1.3)

Figure 1.7 Direct biophotolysis. (H2ase: hydrogenase)

Figure 1.8 Indirect biophotolysis.

In indirect biophotolysis shown in Figure 1.8, the oxygen sensitivity of the hydrogen evolution process is circumvented by separating oxygen evolution and hydrogen evolution processes temporarily and/or spatially. For example, cyanobacteria

16 cell can use CO2 in the air as a carbon source to produce cell material in stage 1 (equation

1.4), which are subsequently utilized in the anaerobic stage 2 for hydrogen production

(equation 1.5). Some bacteria can also consume organic substrates derivable from water waste, which makes them good candidates for wastewater treatment.

12 H2O + 6 CO2 + ‘light energy’ → C6H12O6 + 6 O2 (1.4)

C6H12O6 + 12 H2O+ ‘light energy’ → 12 H2 + 6 CO2 (1.5)

Figure 1.9 Photofermentations by photosynthetic bacteria (N2ase: nitrogenase).

Photofermentation with photosynthetic bacteria have long been studied for their hydrogen production ability through nitrogenase system (Figure 1.9). H2 production by 17 purple non-sulfur bacteria uses solar light and organic acids (such as acetic acid) under nitrogen-deficient conditions (equation 1.6). In these organisms, the lack of Photosystem

II, which is responsible for O2 evolution, eliminates the inhibition of H2 production.

Photosynthetic bacteria can use a wide range of organic or inorganic acid to convert light energy to hydrogen in batch processes.

CH3COOH +2 H2O + ‘light energy’ → 4 H2 + 2 CO2 (1.6)

In contrast from the processes mentioned above, dark hydrogen fermentation does not need light energy and usually takes place under anoxic conditions. Bacteria grow on organic substrates, which are degraded to form building blocks for the organism through oxidation. The electrons generated from the oxidation process need to be combined with other compounds to maintain electrical neutrality, such as protons to form hydrogen.

In summary, biological hydrogen production can serve dual purposes: bioremediation and clean energy generation. However, the scale up of such processes remains a huge challenge due to lower H2 yield and rate of H2 production. The search or biomodification of suitable microbial cultures and the use of cheaper raw materials are essential to handle waste efficiently.

1.4 Metal-metal Bonds

1.4.1 History of Metal-metal Bonding

Prior to the 1960s, transition metal chemistry was based on the notions developed by Alfred Werner, with the concept of a single metal ion surrounded by a set of ligands.

Progress was made that expanded upon Werner’s one-center coordination chemistry theme when more sophisticated characterization methods became available, such as x-ray crystallography. These techniques led to the discovery of multi-metallic chemistry, with compounds involving direct metal-metal bonds. This realm of transition metal chemistry was not recognized and explored until 1963.

The first non-Wernerian coordination compounds were brought into attention at

− about the same time by two independent research groups. Re3Cl12 contains a triangular

Re3 trimetallic core, where it was found that the Re-Re distances were much shorter than those in metallic rhenium, 2.47 Å vs 2.75 Å, and were formulated as double bonds.58-60

One of the innovative aspect of the development of the multicenter transition metal chemistry is the recognition of chemical bonds of an order higher than three, resulting in the discovery of the quadruple bond. The first quadruple bond complexes were

2− 61 discovered in 1964 in [Re2Cl8] with a shorter Re-Re distance of only 2.22 Å. The fact

2− that [Re2Cl8] possesses an eclipsed configuration, rather than staggered, gives a  bond, which further justifies the existence of the quadruple bond.62 Since that time, thousands of quadruply-bonded bimetallic complexes have been synthesized and characterized.

According to Huckel theory, atomic orbitals (AOs) combine to generate molecular orbitals (MOs), and the resulting energy splitting of these MOs is proportional 19 to the overlap integrals of the AOs (Figure 1.10). These overlaps increase in the order of

 <<  <<  thus the orbitals in a bimetallic complex are expected to be ordered in the energy as follows:63

 <         

Many ground state electron configurations are possible in the energy diagrams shown in Figure 1.11Figure 1.1163 According to the conventional MO theory definition of bond order:

bond order = 푛푏−푛푎 2

where nb and na are the numbers of electrons occupying bonding and antibonding orbitals, respectively. Bond orders may vary in steps of 0.5 by removing d-electrons or addition of antibonding electrons, from 0.5 to 4.

Figure 1.11 M-M bond order change by removal of d-electrons or addition of antibonding electrons. Orbital ordering may change for antibonding orbitals.

1.4.2 Dirhodium Complexes

Since 1965, the range of metallic complexes investigated, including their number and variety of the metal and coordination sphere, have grown drastically. There are over

4000 bimetallic complexes involving elements from groups 5 to 10.63 The specific element and structures of interest in the present work are those composed of two Rh(II) centers in a paddlewheel and half-lantern ligation arrangement. Complexes with four bridging ligands, as in the case of the dirhodium tetraacetate Rh2(O2CCH3)4(L)2 systems in Figure 1.12,64 are called paddlewheel compounds. The RhII-RhII core with 14 electrons adopts a ground state electronic configuration of ()2()4()2()2()4()0, resulting in

II,II a Rh-Rh single bond. The Rh2 (O2CCH3)4 complex is known to bind duplex DNA and to inhibit DNA replication.65

Figure 1.12 Structure of the Rh2(O2CCH3)4(L)2 complexes (L = H2O, py, PPh3; py = pyridine, PPh3 = triphenylphosphine).

The related dirhodium(II) tetraformamidinate complexes, Rh2(R-form)4 (R = p-

CF3, p-Cl, p-OCH3, m-OCH3, form = N,N’-di-p-tolylformamidinate), bear more electron 22 donating bridging ligands due to the lower electron affinity of the nitrogen atom comparing to the tetraacetate complexes. It should be noted that the formamidinate ligands also impact the metal orbitals that make up the HOMO.66-70 While typical d7d7 metal paddlewheels, considering only d orbitals, have a (σ)2(π)4(δ)2(δ*)2(π*)4(σ*)0 electronic configuration,62,71 formamidinate ligands (π*) orbitals have the correct symmetry to interact with Rh2(δ*) MO, thus raising its energy above the Rh2(π*) orbitals in some cases. This switch is observed in tetraformamidinate complexes of rhodium,66 as

67-69 well as those of other metals, where the formamidinate π* orbitals of b1u symmetry interact with the δ* orbitals of the same symmetry to impart mixed DTolF(π*)/Rh2(δ*)

HOMO character.

Rh2(R-form)4 complexes possess multiple redox states and have accessible redox- active excited states to photochemically reduce alkyl halides (RX) through an outer

II,III sphere electron transfer mechanism to generate Rh2 (R-form)4X and radical R•; the

72 latter combine to form R2. However, one of the drawbacks of these dirhodium tetraformamidinate systems is their poor solubility in polar solvents, which limits their application in photocatalysis, solar cells, and medicine.

To solve this issue, the substitution of two anionic formamidinate bridging ligands with two neutral chelating ligands produces cationic complexes, cis-[Rh2(-

2+ DTolF)2(NN)2] ( 4-6; DTolF = N,N’-di-p-tolylformamidinate, NN = chelating diimine ligands, Figure 1.13), which are called half-lantern complexes.7,10 They show enhanced solubility in water and other polar solvents while maintaining rich electrochemistry and redox-active excited states.

2+ Figure 1.13 Structure of cis-[Rh2(-DTolF)2(NN)2] , DTolF = N,N’-di-p- tolylformamidinate, NN = chelating diimine ligands, dpq (4), dppz (5), dppn (6).

Table 1.1 Electronic absorption maxima, abs, extinction coefficient, , and reduction potentials for 4-6 in CH3CN.

a −1 b c vs. Ag/AgCl; 0.1 M TBAPF6; 0.2 V s . Quasi-reversible. Irreversible.

The electronic absorption and reduction potentials for 4-6 are summarized in

Table 1.1.7 Density functional theory (DFT) calculation reveals that the highest occupied

24 molecular orbital (HOMO) for these complexes is Rh2 (*) and DTolF (*) mixed while the lowest unoccupied molecular orbital (LUMO) is localized on the diimine (*). These complexes feature modest absorption band in the visible region in CH3CN with extinction coefficients ~ 1,000 M−1 cm−1, which is assigned to singlet metal-ligand to ligand charge transfer (ML-LCT) transition. The extended conjugation system on 5 and 6 does not affect the electronic transition features significantly, which is due to the limited electronic communication between the metal center and the distal part of the diimine ligand, which has also been observed with Ru(II) systems.73,74

Electrochemistry reveals that the first oxidation and reduction for 4-6 are ~ + 1.0

V and ~ 0.4 V vs Ag/AgCl in CH3CN, respectively. On the reduction side, pyridine titration of these complexes does not affect the second and third reduction peaks while the first reduction shifts with the addition of pyridine. This observation indicates that

LUMO is metal-based in character, due to the pyridine interaction on the complexes axial position which strongly disturb the metal-based orbitals. To be noted, the electrochemical

LUMO assignment is in conflict with the calculated results, which predicts a diimine ligand based LUMO.

To further extend the absorption profile of the dirhodium complexes into the red and near-IR region, we have now developed the next generation of these systems and that the excited state and redox properties of these complexes, including photoinduced charge injection and photocatalysis, are the focus of the work presented hereby.

Bibliography

1. U. S. Energy Information Administration, Annual Energy Outlook 2019, 2019.

2. U. S. Energy Information Administration, Monthly Energy Review, June, 2019.

3. T. W. Brown, T. Bischof-Niemz, K. Blok, C. Breyer, H. Lund and B. V. Mathiesen, Renew. Sust. Energ. Rev., 2018, 92, 834-847.

4. S. Mozaffari, M. R. Nateghi and M. B. Zarandi, Renew. Sust. Energ. Rev., 2017, 71, 675-686.

5. A. J. Bard and M. A. Fox, Acc. Chem. Res., 1995, 28, 141-145.

6. T. J. Whittemore, A. Millet, H. J. Sayre, C. Xue, B. S. Dolinar, E. G. White, K. R. Dunbar and C. Turro, J. Am. Chem. Soc., 2018, 140, 5161-5170.

7. Z. Li, N. A. Leed, N. M. Dickson-Karn, K. R. Dunbar and C. Turro, Chem. Sci., 2014, 5, 727-737.

8. T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

9. H. J. Sayre, A. Millet, K. R. Dunbar and C. Turro, Chem. Commun., 2018, 54, 8332-8334.

10. T. A. White, S. E. Witt, Z. Li, K. R. Dunbar and C. Turro, Inorg. Chem., 2015, 54, 10042-10048.

11. S. E. Witt, T. A. White, Z. Li, K. R. Dunbar and C. Turro, Chem. Commun., 2016, 52, 12175-12178.

12. M. K. Nazeeruddin, E. Baranoff and M. Grätzel, Sol. Energy, 2011, 85, 1172- 1178.

13. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595-6663.

14. M. Freitag, J. Teuscher, Y. Saygili, X. Zhang, F. Giordano, P. Liska, J. Hua, S. M. Zakeeruddin, J.-E. Moser, M. Grätzel and A. Hagfeldt, Nat. Photonics, 2017, 11, 372.

15. W. J. Youngblood, S.-H. A. Lee, Y. Kobayashi, E. A. Hernandez-Pagan, P. G. Hoertz, T. A. Moore, A. L. Moore, D. Gust and T. E. Mallouk, J. Am. Chem. Soc., 2009, 131, 926-927.

16. J. Li and N. Wu, Catal. Sci. Technol., 2015, 5, 1360-1384.

17. D. Wang, B. D. Sherman, B. H. Farnum, M. V. Sheridan, S. L. Marquard, M. S. Eberhart, C. J. Dares and T. J. Meyer, Proc. Natl. Acad. Sci., 2017, 114, 9809.

18. P. Xu, T. Huang, J. Huang, Y. Yan and T. E. Mallouk, Proc. Natl. Acad. Sci., 2018, 115, 6946.

19. P. Xu, C. L. Gray, L. Xiao and T. E. Mallouk, J. Am. Chem. Soc., 2018, 140, 11647-11654.

20. M. Urbani, M.-E. Ragoussi, M. K. Nazeeruddin and T. Torres, Coord. Chem. Rev., 2019, 381, 1-64.

21. D. Wang, S. L. Marquard, L. Troian-Gautier, M. V. Sheridan, B. D. Sherman, Y. Wang, M. S. Eberhart, B. H. Farnum, C. J. Dares and T. J. Meyer, J. Am. Chem. Soc., 2018, 140, 719-726.

22. M. K. Brennaman, R. J. Dillon, L. Alibabaei, M. K. Gish, C. J. Dares, D. L. Ashford, R. L. House, G. J. Meyer, J. M. Papanikolas and T. J. Meyer, J. Am. Chem. Soc., 2016, 138, 13085-13102.

23. R. J. Kamire, K. L. Materna, W. L. Hoffeditz, B. T. Phelan, J. M. Thomsen, O. K. Farha, J. T. Hupp, G. W. Brudvig and M. R. Wasielewski, J. Phys. Chem. C, 2017, 121, 3752-3764.

24. J. B. Asbury, E. Hao, Y. Wang and T. Lian, J. Phys. Chem. B, 2000, 104, 11957- 11964.

25. Y. Xu and M. A. A. Schoonen, Am. Mineral., 2000, 85, 543-556.

26. I. Purnama, Salmahaminati, M. Abe, M. Hada, Y. Kubo and J. Y. Mulyana, Dalton Trans., 2019, 48, 688-695.

27. J. Sun, Y. Yu, A. E. Curtze, X. Liang and Y. Wu, Chem. Sci., 2019, 10, 5519- 5527.

28. V. Nikolaou, A. Charisiadis, G. Charalambidis, A. G. Coutsolelos and F. Odobel, J. Mater. Chem. A, 2017, 5, 21077-21113.

29. P. Wang, S. M. Zakeeruddin, J.-E. Moser, R. Humphry-Baker and M. Grätzel, J. Am. Chem. Soc., 2004, 126, 7164-7165.

30. B. V. Bergeron, A. Marton, G. Oskam and G. J. Meyer, J. Phys. Chem. B, 2005, 109, 937-943.

31. G. Oskam, B. V. Bergeron, G. J. Meyer and P. C. Searson, J. Phys. Chem. B, 2001, 105, 6867-6873.

32. M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Müller, P. Liska, N. Vlachopoulos and M. J. J. o. t. A. C. S. Grätzel, 1993, 115, 6382-6390.

33. L. C. T. Shoute and G. R. Loppnow, J. Am. Chem. Soc., 2003, 125, 15636-15646.

34. M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Mueller, P. Liska, N. Vlachopoulos and M. Graetzel, J. Am. Chem. Soc., 1993, 115, 6382-6390.

35. G. C. Vougioukalakis, A. I. Philippopoulos, T. Stergiopoulos and P. Falaras, Coord. Chem. Rev., 2011, 255, 2602-2621.

36. M.-E. Ragoussi, M. Ince and T. Torres, Eur. J. Org. Chem., 2013, 2013, 6475- 6489.

37. K. Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh, H. Sugihara, Y. Dan-oh, C. Kasada, A. Shinpo and S. Suga, J. Phys. Chem. B, 2005, 109, 15476-15482.

38. S. Ito, H. Miura, S. Uchida, M. Takata, K. Sumioka, P. Liska, P. Comte, P. Péchy and M. Grätzel, Chem. Commun., 2008, DOI: 10.1039/B809093A, 5194-5196.

39. A. Listorti, B. O’Regan and J. R. Durrant, Chem. Mater., 2011, 23, 3381-3399.

40. J. Sá, P. Friedli, R. Geiger, P. Lerch, M. H. Rittmann-Frank, C. J. Milne, J. Szlachetko, F. G. Santomauro, J. A. van Bokhoven, M. Chergui, M. J. Rossi and H. Sigg, Analyst, 2013, 138, 1966-1970.

41. M. Abdellah, A. M. El-Zohry, L. J. Antila, C. D. Windle, E. Reisner and L. Hammarström, J. Am. Chem. Soc., 2017, 139, 1226-1232.

42. M. Juozapavicius, M. Kaucikas, J. J. van Thor and B. C. O’Regan, J. Phys. Chem. C, 2013, 117, 116-123.

43. B. C. O'Regan, K. Bakker, J. Kroeze, H. Smit, P. Sommeling and J. R. Durrant, J. Phys. Chem. B, 2006, 110, 17155-17160.

44. S. A. Haque, S. Handa, K. Peter, E. Palomares, M. Thelakkat and J. R. Durrant, Angew. Chem. Int. Ed., 2005, 44, 5740-5744.

45. A. N. M. Green, E. Palomares, S. A. Haque, J. M. Kroon and J. R. Durrant, J. Phys. Chem. B, 2005, 109, 12525-12533.

46. P. P. Edwards, V. L. Kuznetsov, W. I. F. David and N. P. Brandon, Energy Policy, 2008, 36, 4356-4362.

47. A. Züttel, A. Remhof, A. Borgschulte and O. Friedrichs, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2010, 368, 3329-3342.

48. Department of Energy, https://www.energy.gov/eere/fuelcells/hydrogen-storage, (accessed 07/09, 2019).

49. F. Zhang, P. Zhao, M. Niu and J. Maddy, Int. J. Hydrogen Energy, 2016, 41, 14535-14552.

50. J. Qi, W. Zhang and R. Cao, Advanced Energy Materials, 2018, 8, 1701620.

51. A. Kirubakaran, S. Jain and R. K. Nema, Renew. Sust. Energ. Rev., 2009, 13, 2430-2440.

52. Department of Energy, Hydrogen Production Pathways, https://www.energy.gov/eere/fuelcells/hydrogen-production-pathways, (accessed Jul 05, 2019).

53. P. C. Hallenbeck and Y. Liu, Int. J. Hydrogen Energy, 2016, 41, 4446-4454.

54. D. Das and T. N. Veziroǧlu, Int. J. Hydrogen Energy, 2001, 26, 13-28.

55. D. Das and T. N. Veziroglu, Int. J. Hydrogen Energy, 2008, 33, 6046-6057.

56. S. Manish and R. Banerjee, Int. J. Hydrogen Energy, 2008, 33, 279-286.

57. P. C. Hallenbeck and J. R. Benemann, Int. J. Hydrogen Energy, 2002, 27, 1185- 1193.

58. J. A. Bertrand, F. A. Cotton and W. A. Dollase, J. Am. Chem. Soc., 1963, 85, 1349-1350.

59. J. A. Bertrand, F. A. Cotton and W. A. Dollase, Inorg. Chem., 1963, 2, 1166- 1171. 29

60. Proceedings of the Chemical Society, 1963, DOI: 10.1039/PS9630000101, 101- 124.

61. V. G. Kuznetsov and P. A. Koz'min, J. Struct. Chem., 1963, 4, 49-55.

62. F. A. Cotton, Inorg. Chem., 1965, 4, 334-336.

63. C. A. M. F. Albert Cotton, Richard A. Walton, in Multiple Bonds Between Metal Atoms, ed. C. A. M. F. Albert Cotton, Richard A. Walton, Springer, Boston, MA, ch. 1, pp. 1-21.

64. P. K. L. Fu, P. M. Bradley and C. Turro, Inorg. Chem., 2001, 40, 2476-2477.

65. R. A. Howard, A. P. Kimball and J. L. Bear, Cancer Res., 1979, 39, 2568.

66. G. A. Rizzi, M. Casarin, E. Tondello, P. Piraino and G. Granozzi, Inorg. Chem., 1987, 26, 3406-3409.

67. D. L. Lichtenberger, M. A. Lynn and M. H. Chisholm, J. Am. Chem. Soc., 1999, 121, 12167-12176.

68. F. A. Cotton and T. Ren, Inorg. Chem., 1991, 30, 3675-3679.

69. F. A. Cotton, T. Ren and J. L. Eglin, Inorg. Chem., 1991, 30, 2559-2563.

70. P. M. Bradley, L. T. Smith, J. L. Eglin and C. Turro, Inorg. Chem., 2003, 42, 7360-7362.

71. H. T. Chifotides and K. R. Dunbar, in Multiple Bonds Between Metal Atoms, eds. F. A. Cotton, C. A. Murillo and R. A. Walton, Springer US, Boston, MA, 2005, DOI: 10.1007/0-387-25829-9_12, pp. 465-589.

72. D. A. Lutterman, N. N. Degtyareva, D. H. Johnston, J. C. Gallucci, J. L. Eglin and C. Turro, Inorg. Chem., 2005, 44, 5388-5396.

73. W. Chen, C. Turro, L. A. Friedman, J. K. Barton and N. J. Turro, J. Phys. Chem. B, 1997, 101, 6995-7000.

74. T. J. Whittemore, T. A. White and C. Turro, J. Am. Chem. Soc., 2018, 140, 229- 234.

Chapter 2. Experimental Methods

2.1 Materials

All materials were used as received unless stated otherwise. Reaction grade solvents were purchased from Fisher Scientific. Acetonitrile was distilled over calcium hydride and molecular sieves or purified by a Braun MB-SPS solvent purification system.

Spectrophotometric grade dimethylformamide (DMF) were purchased from Fisher

Scientific and Sigma Aldrich. The metal salts RhCl3•xH2O were obtained from Pressure

Chemical Company. 1-benzyl-1.4-dihydronicotinamide (BNAH) were purchased from

TCI Chemical. Phenazine, anthracene, perylene, tetracene, rubrene, -carotene, trifluoromethanesulfonic acid, p-Toluenesulfonic acid monohydrate, trifluoroacetic acid, glacial acetic acid, triethylamine, triethanolamine, triphenylamine, diphenylamine were purchased from Sigma Aldrich and used without further purification. Cobaltocene were purchased from Sigma Aldrich and stored in a nitrogen-filled Braun glovebox.

2.2 Synthesis and sample preparation

2.2.1 Synthesis of [Rh2(DTolF)2(-NN)2][BF4]2 (NN= menp (7) and dmeb (8))

Synthesis of the ligand 4-methoxycarbonyl-1.8-napthyridine was performed by dissolving 25 mg (0.14 mmol) 1,8-Naphthyridine-4-carboxylic acid in 2.0 mL MeOH 31 with 200 L H2SO4. After 1 week at room temperature, methanol was removed by evaporation and the acid was neutralized with NaHCO3. ~ 5 mL water was added and the ligand was extracted with CH2Cl2. The ligand 2,2ʹ-bipyridine-4,4ʹ-methoxycarbonyl was synthesized following similar procedures. The complex [Rh2(DTolF)2(-menp)2][BF4]2

(7) and [Rh2(DTolF)2(-dmeb)2][BF4]2 (8) was synthesized by the modified procedure for

1 2 1 and [Rh2(DTolF)2(bpy)2][BF4]2 (9).

1 7 H NMR (250 MHz, DMSO-d6): δ 2.21 (12 H, s, tolyl CH3), 3.99 (6 H, s,

OCH3), 6.55 (8 H, m, tolyl), 7.01 (8 H, m, tolyl), 7.57 (2 H, dd, NCHN), 8.02 (2 H, ddd,

6), 8.33 (2 H, dd, 3), 9.29 (2 H, dd, 5), 9.42 (2 H, dd, 7), 9.54 (2 H, dd, 2), shown in

Figure 2.1.

1 8 H NMR (250 MHz, CD3CN-d3): δ 2.25 (12 H, s, tolyl CH3), 4.02 (12 H, s,

OCH3), 7.00 (16 H, m, tolyl), 7.79 (2 H, d, 5), 8.13 (2 H, t, NCHN), 8.32 (2 H, s, 3), 8.66

(2 H, d, 6), shown in Figure 2.2.

1 Figure 2.1 H NMR of 7 in DMSO-d6 32

1 Figure 2.2 H NMR of 8 in CH3CN-d3

TiO2 sensitization

7@TiO2 7 (~6 mg) was added to 1 mL CH3CN and the electronic absorption spectrum was taken in a 1 mm pathlength cuvette. P25 TiO2 nanoparticles (~50mg) was added to the solution subsequently to prepare sensitized TiO2 (7@TiO2). Upon mixing,

TiO2 quickly took on a blue color, indicating anchoring of the dye to TiO2. The nanoparticles were centrifuged, and the supernatant were removed and examined by electronic absorption spectroscopy. 8@TiO2 was prepared follow similar procedures. The electronic absorption spectra before and after anchoring is shown in Figure 2.3.

Figure 2.3 Electronic absorption spectra collected before (black) and after (red) anchoring of (a) 7 and (b) 8 to TiO2 nanoparticles.

cis-[Rh2(DPhF)2(bncn)2](BF4)2 (11). Samples of 11 were prepared from cis-

[Rh2(DPhF)2(CH3CN)6](BF4)2 (58 mg, 0.057 mmol) and benzo[c]cinnoline (23 mg, 0.125 mmol) in 20 mL of dry CH3CN and refluxed for 16 h under a positive pressure of Ar or

N2 in an oil bath to give a blue solution. After 16 h, the reaction was cooled to room temperature and the solution was concentrated under reduced pressure to ~2 mL. The solution was added dropwise to rapidly stirring diethyl ether and subsequently filtered through a medium frit under vacuum. The filtrate was washed with ether and 100 mL of hexanes to get rid of excess benzo[c]cinnoline until the washings were no longer yellow 34

(65 % yield). X-ray quality crystals were obtained by the slow diffusion of ether into concentrated solution of acetonitrile at 0°C. 1H NMR δH (400 MHz; CD3CN-d3): 9.69

(4H, d, J = 8.2 Hz, bncn), 8.83 (4H, d, J = 8.2 Hz, J = 1.4 Hz, bncn), 8.24 (8H, m, bncn),

7.24 (12H, m, J = 8.2 Hz, Phenyl), 7.01 (4H, t, J = 6.4 Hz, Phenyl), 7.02 (2H, t, J = 4.3

Hz, NCN), 6.95 (8H, dd, J = 7.5 Hz, J = 1.2 Hz, Phenyl). ESI-MS in CH3CN: m/z: 982.2

+ for [Rh2(DPhF)2(bncn)2][CN] .

2.3 Methods and Instrumentation

2.3.1 General

Electronic spectroscopy was performed on a Hewlett-Packard 8453 diode array spectrometer; extinction coefficients were obtained by serial dilution of ~10 mg of complex using a Mettler Toledo XSE105 DualRange balance and values are agreed within 2% of one another. Samples were prepared in 1 mm or 1 cm quartz cuvette so that the absorbance was below the value of 1. Steady state emission and excitation spectra were obtained at 77 K in a homebuilt dewar in acetonitrile as well as an ethanol:methanol glass (10:1) on a Fluorolog spectrofluorimeter (Horiba).

Electrospray mass spectroscopy was performed on a Bruker MicroTOF spectrometer in positive ion mode by the serial dilution of the sample in acetonitrile.

All nuclear magnetic resonance (NMR) spectroscopy was conducted in deuterated solvent at 298 K on a Bruker 400 MHz spectrometer and chemical shifts were referenced to the residual peak of the deuterated solvent signal.

Gaseous products were quantified and qualified by a Shimadzu GC-2014 gas chromatograph (GC). Aliquots of headspace were removed using a Hamilton gastight syringe and injected into a GC (He carrier gas for H2 generation) with a ShinCarbon column (2 m long × 1/8 in. OD × 2.0 mm ID) and a Shimadzu TCD-2014 thermal conductivity detector. The GC conditions were as follows: injector temperature, 41.0 °C; column temperature, 30 °C; detector temperature, 150 °C; and gas flow, 25 mL/min. The calibration curve was constructed by injecting a series of known amount of 5% H2/N2 mixture in triplicate.

2.3.2 Electrochemistry

Cyclic voltammograms were recorded under an atmosphere of N2 with a BASi

CV-50W potentiostat (Bioanalytical Systems, Inc.; West Lafayette, IN, USA) using a three-electrode cell consisting of a glassy carbon disc working electrode, Pt wire counter electrode and a Ag/AgCl (3 M NaCl) reference electrode standardized to ferrocene (E1/2 =

+0.55 V and + 0.44 V vs Ag/AgCl in DMF and CH3CN, respectively) and a 200 mV/s scan rate. Approximately 0.5 mM of analyte in 0.1 M Bu4NPF6 acetonitrile or DMF electrolyte solution. Equivalents of trifluoromethanesulfonic acid, p-toluenesulfonic acid monohydrate, trifluoroacetic acid and glacial acetic acid were added to the electrolyte containing 0.5 mM analyte to determine catalytic proton reduction activities. The glassy carbon electrode was polished with 0.3 m alumina and rinsed with water after each scan to prevent any deposit on the electrode surface.

Bulk electrolysis experiments were performed in a customized, airtight, two- compartment cell separated by a glass frit, shown in Figure 2.4. A high-density extruded graphite rod (Graphtek LLC, Buffalo Grove, IL, USA) or glassy carbon rod working electrode, a Ag/AgCl reference electrode and a magnetic stir bar were placed in the working compartment while a Pt wire counter electrode was in the auxiliary compartment. The working electrode was polished before each measurement, this is particularly important in the presence of acid, where adsorption phenomena were observed when applying voltages more negative than -800 mV at high concentrations of strong acid. The electrolysis solution contained 0.5 mM of 1 and 5 mM dimethylformamidinium triflate ([DMFH][OTf]) in 0.1 M Bu4NPF6 DMF electrolyte. The electrolyte in the auxiliary compartment does not contain any substrate or catalyst. Both sides were purged with N2 for 20 min before the experiments. Aliquots of the headspace was injected to GC to determine H2 production.

Spectroelectrochemical measurements were performed in a homemade H-cell under inert gas where the Ag/AgCl reference and glassy carbon rod or graphite working electrode were kept in a working compartment and the silver coil wire counter electrode in the auxiliary compartment during the electrolysis of the 0.1 M Bu4NPF6 acetonitrile electrolyte solution at potentials ~50-200 mV past the oxidation/reduction.

Figure 2.4 Bulk electrolysis cell

2.3.3 Time-resolved spectroscopies

Nanosecond transient absorption experiments in Chapter 3 were carried out on a system that have been described previously,3 but upgraded to include a new laser source and upgraded software. The experiment was performed with a Lab-190 (Spectra Physics)

Nd:YAG laser. The fundamental output was frequency doubled, isolated using a set of dichroic mirrors, and the pulse (8 ns fwhm) was set to 5 mJ at the sample. The output of a

150 W Xe arc lamp and an LPS-220 was focused into a H-20 Monochromator (Horiba) onto a PMT. This output was read using a TDS-380 Oscilloscope and the experiment was designed using a home-built LabView program. Except for rubrene, which was excited at

532 nm (5 mJ/pulse), all other organic sensitizers were excited at 355 nm (5 mJ/pulse).

~0.5 OD solution in DMSO deaerated by N2 for 20 min was prepared for each organic sensitizer at excitation wavelength and 0 ~ 100 M dirhodium complexes were added to

38 obtain the lifetimes. The transient absorption spectra were examined before and after the addition of dirhodium complexes to ensure no redox product was formed. In these experiments, Stern-Volmer plots, 0/ vs [Rh2 complex], were constructed to obtain the quenching constants.

In later chapters, nanosecond transient absorption was performed on a setup previously reported.4 The pump beam (~5 mJ) was generated using a BasiScan OPO

(Spectra Physics) pumped with the third harmonic of an Nd:YAG laser (Spectra Physics,

INDI-40) at a rate of 10 Hz and power of 130 mJ. The output of a continuous 150 W xenon arc lamp gated using a Uniblitz shutter was used as the probe. The pump and probe pulse were overlapped at a 90° geometry at the sample. Time-resolved absorption spectra were obtained on an Edinburgh LP980 spectrometer with an ICCD-based broadband camera. The kinetic traces were collected with a PMT and an oscilloscope. All transient absorption samples were deoxygenated prior to study and UV-Vis spectra were collected before and after to ensure no sample degradation occurred upon irradiation. All samples were prepared to an optical density of ~0.5 OD at the pump wavelength in a 1 by 1 cm cell and all fits were performed as the sum of exponentials using Igor Pro (6.3) software.

Ultrafast transient absorption (fsTA) experiments were performed on a system previously described in detail.5 Briefly, the output of Ti:sapphire regenerative amplifier

(Astrella 1K-USP, Coherent) was split to generate the white light probe through rotating

CaF2 crystal and to pump an OPA (OPerA Solo, Coherent/Positive Light) to generate the pump pulse at 600 nm. A thermally cooled CCD camera and home-built software written in LabVIEW 2015 were used to collect and manipulate data. The sample solution was

39 prepared with ~ 0.5 OD at the excitation wavelength and ~5 mL solution containing the sample was flowed through a 1 mm path-length Harrick Scientific flow cell (1 mm thick

CaF2 windows) and excited with ~ 2.5 µJ at the pump wavelength. An instrument response of fwhm ~ 85 fs was measured using Kerr effect in cyclohexane. The polarization angle between the pump and probe was set to 54.7° to avoid the rotational diffusion effects.

Time-resolved infrared spectroscopy (TRIR) has been reported previously.6 TRIR was performed with a Ti:Sapphire regenerative amplifier (Coherent Legend; 1kHz; 300 fwhm) seeded by a short pulse oscillator (Coherent, Mantis). The pump beam was generated using an optical parametric amplifier (OPerA, Coherent) to create the desired wavelength, which was then filtered and focused onto an optical delay line. The probe was created from a DFG module in front of a second OPerA optical parametric amplifier, which produces an adjustable output of 2 to 10 μm radiation. This beam was split using a germanium beam splitter into a probe and reference that were both put through a 0.1 mm pathlength airtight Harrick sample cell (CaF2) that was continuously actuated perpendicular to pump beam to ensure lack of photoproduct or decomposition. The pump beam was directed from the optical delay line through a polarizer set at 54.7 degrees relative to the probe wavelength to overcome anisotropy effects, chopped at a rate of 500 Hz to create a pump on/pump off measurement, and spatially overlapped with the IR probe beam. The probe and reference were dispersed using a spectrograph (Triax

320) and imaged on a liquid nitrogen cooled HgCdTe CCD (32 by 2) with 4 cm-1 resolution for each window. The pulse stability was corrected to be <1% for

40 measurements. All TRIR measurements were performed under an N2 gas positive pressure environment and at room temperature. Solutions were made to an absorbance of

0.5−0.8 OD in the 0.1 mm cell path separated by a Teflon spacer sandwiched by two

CaF2 windows. Samples were checked before and after collection using ground state IR

(Perkin Elmer; Spectrum 65) to ensure that no decomposition occurred.

2.3.4 Photolysis

All irradiation studies were performed using light emitting diodes (Luxeon Star), where 670 nm irradiation was performed with 4 LEDs on a homebuilt irradiation apparatus, where the sum of all the LED output was found to be 800 mW and irradiation with the 735 nm LEDs was performed with two, lower output LEDs where the sum of all

LED output was found to be 350 mW. Acrylic plates were designed to hold a 1 cm cuvette in the center of the LEDs, each at a 1.6 cm distance from the photolysis cuvette,

1− shown in Figure 2.5. Samples of 11 or 11 were prepared ~0.5 OD in DMF or CH3CN in glovebox to minimize the oxygen concentration in the sample prior to photolysis.

Dark controls were performed by irradiating samples in an aluminum foil sleeve to ensure the effect was not simply temperature dependent. Quantification of hydrogen was performed by irradiation in septum capped vials or vacuum valve equipped 1 cm cuvettes. Volume measurements of cells were performed through weight when filled with deionized water, all measurements made in triplicate.

Figure 2.5 Four of the 670 nm LEDs on an acrylic plate

Bibliography

1. T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

2. H. T. Chifotides, K. V. Catalan and K. R. Dunbar, Inorg. Chem., 2003, 42, 8739- 8747.

3. P. M. Bradley, B. E. Bursten and C. Turro, Inorg. Chem., 2001, 40, 1376-1379.

4. T. N. Rohrabaugh, K. A. Collins, C. Xue, J. K. White, J. J. Kodanko and C. Turro, Dalton Trans., 2018, 47, 11851-11858.

5. L. M. Loftus, A. Li, K. L. Fillman, P. D. Martin, J. J. Kodanko and C. Turro, J. Am. Chem. Soc., 2017, 139, 18295-18306.

6. J. Wang, G. Burdzinski, J. Kubicki, T. L. Gustafson and M. S. Platz, J. Am. Chem. Soc., 2008, 130, 5418-5419.

Chapter 3. Dirhodium Complexes: Excited State Properties and Electron Transfer

3.1 Background

The excited state processes of transition metal complexes have been widely explored for solar energy conversion.1,2 Due to the absorbed photon energy, the excited states are better reductants and/or oxidants comparing to their ground state counterparts.

One challenge in harvesting the solar output is that not every photon is usable because many are scattered and the frequency of others is too low to drive useful reactions.3 But a large portion of the remaining incident light in the visible and near-infrared (near-IR) spectral regions are not absorbed by traditional photosensitizers, such that they remain unused in light harvesting and dye-sensitized applications. Panchromatic light absorbing dirhodium complexes cis-[Rh2(μ-DTolF)2(L)2][BF4]2 , where DTolF = N,N’-di-p- tolylformamidinate and L represents the bridging ligands np (1; 1,8-naphthyridine) and npCOOH (2; 1,8-naphthyridine-2-carboxylic acid) was synthesized and characterized.4,5

Femtosecond and nanosecond transient absorption (fsTA, nsTA) and time- resolved infrared (fsTRIR) spectroscopies reveal that the initially populated singlet metal- ligand to ligand charge transfer (1ML-LCT) excited states of 1 and 2 decay to the corresponding 3ML-LCT excited states with time constants of 13 ps and 7 ps, respectively. The 3ML-LCT excited states of 1 and 2 repopulate the ground state with 43 lifetimes of 640 ps and 25 ns, respectively. The shorter 1ML-LCT and 3ML-LCT lifetimes of 1 as compared to those of 2 are attributed to the blocked axial position in the later, which prevent the solvent assisted nonradiative deactivation pathway.4,6 Similarly, the related complexes with chelating diimine ligands, L, cis-[Rh2(μ-DTolF)2(L)2][BF4]2, where L = phen (3; 1,10-phenanthroline), dpq (4; dipyrido[3,2-f:2’,3’-h]quinoxaline), dppz (5; dipyrido[3,2-a:2’,3’-c]phenazine) and dppn (6; benzo[i]dipyrido[3,2-a:2’,3’- h]quinoxaline), also exhibit long-lived excited states.4,7 The molecular structures of 1 – 6 are shown in Figure 3.1 and their excited state properties make these compounds promising dyes for applications that involve dye-sensitized solar cells, DSSCs. In order to effectively inject holes or electrons into p-type or n-type semiconductors, it is important for dye molecules to possess proper excited state redox potentials relative to their valence or conduction band, respectively.8,9 In order to determine if the sensitized charge injection is favorable for a given molecule, the excited state reduction and/or oxidation potential must first be determined.

T T The triplet excited state reduction and oxidation potentials, E* red and E* ox, respectively, can be calculated from equations 3.1 and 3.2, where Ered and Eox are the

T ground state reduction and oxidation potentials, respectively, and E00 represents the energy difference between the lowest energy vibrational levels of the ground state and that of the lowest lying triplet excited state. For emissive complexes, this value can be obtained from the highest energy peak in the vibrational progression typically observed at

10 T low temperature. However, dirhodium complexes 1 – 6 are not emissive. Thus, E00

T cannot be readily obtained from the low temperature emission spectra. Herein, the E00 of

44 these dirhodium complexes were estimated through energy transfer quenching experiments with a series of organic sensitizers with varying triplet excited state energy,

E(3ππ*).

T T E* red = Ered + E00 (3.1)

T T E* ox = Eox − E00 (3.2)

Figure 3.1 Structures of cis-[Rh2(μ-DTolF)2(L)2][BF4]2, L= bridging diimine ligand, np (1) and npCOO (2), or L = chelating diimine ligand phen (3), dpq (4), dppz (5), dppn (6).

3.2 Results and Discussion

3.2.1 Excited State Quenching of Organic Sensitizers

Organic sensitizers were chosen with a variety of known triplet excited state energies, E(3ππ*), as listed in Table 3.1.11 In these experiments, the organic sensitizers were used as energy transfer donors (EnDs) from their long-lived 3ππ* excited states, as depicted in Figure 3.2. Upon excitation, the singlet excited state of the EnD is populated,

1S*, which intersystem crosses to the corresponding triplet excited state, 3S*.

Subsequently, the excited EnD can be quenched by a Rh2 complex via energy transfer if

3 the Rh2 has a lower excited state energy (Figure 3.2). However, if the value of E( ππ*) of

T the EnD is lower than E00 of the Rh2 complex, then energy transfer would be uphill, and quenching would not occur. Thus, the excited state energy of the Rh2 can be estimated from the 3* energies of different reference organic EnDs (Figure 3.2).

Figure 3.2 Energy transfer mechanism of the energy transfer donors (D) and dirhodium complexes (Rh2).

Table 3.1 Selected organic EnDs and E(3ππ*).

EnD E(3ππ*)/eV11 Phenazine 1.93 Anthracene 1.83 Perylene 1.53 Tetracene 1.29 Phthalocyanine 1.23 Rubrene 1.15 Pentacene 0.99 -carotene 0.78

The organic EnDS were excited with the third harmonic output of a Nd:YAG laser (fwhm ∼ 8 ns, 50 mW) at 355 nm, with the exception of rubrene, which was excited with the doubled output at 532 nm from the same laser. The experiments for the chelating

Rh2 complexes 3 – 6 were performed in dimethylformamide (DMF), whereas those with the bridging Rh2 complexes 1 and 2 were conducted in dimethyl sulfoxide (DMSO). The lifetime of the monoexponential decay of the transient absorption signal from each EnD,

o in the absence of quencher, was measured, resulting in shorter lifetimes, , as a function the concentration of Rh2 was increased. In this manner, Stern-Volmer plots, o /

 vs [Rh2], were generated for each EnD/Rh2 pair. The energy transfer quenching rate

T constant, kEN was obtained from the slope of each plot (Equation 3.3).

휏0 T = 푘EN × 휏 [푅ℎ ] + 1 (3.3) 휏 0 2

As an example, Figure 3.3 shows the kinetic traces at 420 nm for anthracene,

3 E( ππ*) = 1.83 eV and 0 = 44.4 s, as the EnD with varying concentrations of 1, showing the decrease in the anthracene 3* state lifetime as a function of quencher concentration. The resulting Stern-Volmer plot, 0/ vs [1], is shown in Figure 3.4,

T 9 -1 -1 resulting in a triplet-triplet energy transfer rate constant, kEN , of 2.15 × 10 M s . This value is within the order of magnitude for diffusion-controlled rate constant. Perylene and rubrene , with E(3ππ*) values of 1.52 and 1.15 eV, respectively, were also quenched by 1.

However, quenching was not observed for β-carotene, with E(3ππ*) = 0.78 eV. It should also be noted that a large decrease in the energy-transfer rate constant was observed for

T 8 -1 -1 T 9 -1 -1 rubrene (kEN ≈ 10 M s ) relative to that measured for perylene (kEN ≈ 10 M s ).

These results lead to the estimation that the triplet excited states of 1 lies ∼1.1 eV above the ground state. Similar results were obtained for 2 and are summarized in Table 3.2.

The quenching rate constants decrease with the decreasing value of E(3ππ*) of the EnD, which is indicative the relationship between quenching driving force and the excited state energy. Similar experiments were performed for 3 - 6 with anthracene, perylene, tetracene, rubrene and β-carotene and summarized in Table 3.3. 4 - 6 can quench

T tetracene but not rubrene which results in E00 ~1.2 eV. Nonetheless, 3 cannot be

T quenched by perylene and other lower energy EnDs. Therefore, the E00 of the 3 was estimated ~1.7 eV.

[1] 0.25 0 µM 0.20 6 µM 12 µM

0.15 18 µM

OD  0.10

0.05

0.00

-20 0 20 40 60 80 100 120 140  /µs

Figure 3.3 Kinetic traces of the decay of the 3* state of anthracene monitored at 420 nm as a function of [1].

Figure 3.4 Stern-Volmer plot of anthracene with 1.

Table 3.2 Quenching constants for 1 and 2 with organic EnDs

k T / M-1s-1 E(eV)11 EN

1 2 9 9 Anthracene 1.83 2.15×10 2.16×10 9 9 Perylene 1.53 1.35×10 2.01×10 8 8 Rubrene 1.15 6.47×10 2.03×10 -carotene 0.78 NQa NQ a. NQ stands for no quenching was observed

Table 3.3 Quenching constants for 3 - 4 with organic EnDs

-1 -1 k T / M s E(eV)11 EN 3 4 5 6 9 9 9 9 Phenazine 1.93 2.34×10 4.84×10 2.93×10 2.22×10 9 9 9 9 Anthracene 1.83 1.95×10 1.22×10 2.17×10 2.03×10 9 8 9 Perylene 1.53 NQ 1.10×10 6.60×10 1.87×10 8 8 8 Tetracene 1.27 NQ 6.33×10 5.02×10 3.26×10 Rubrene 1.15 NQ NQ NQ NQ β-carotene 0.78 NQ NQ NQ NQ

3.2.2 Quenching Mechanism

Quenching can take place via different mechanisms, such as energy transfer, reductive quenching and oxidative quenching, shown in Equation 3.4 - 3.6.12

*EnD + Rh2 → EnD + *Rh2 Energy Transfer (3.4)

− + *EnD + Rh2 → EnD + Rh2 Reductive Quenching (3.5)

+ − *EnD + Rh2 → EnD + Rh2 Oxidative Quenching (3.6)

The driving force for reductive quenching, ΔGRQ, depends on the reduction potential of the organic EnD in its excited state, E(EnD*/−), and the first oxidation

+/0 */− potential of the ground state of the Rh2 complex, E(Rh2 ). The value of E(EnD ) can be calculated from the ground state reduction potential of EnD, E(EnD0/−), and its triplet excited state energy, E(3ππ*), following equations 3.7 and 3.8.

+/0 */− ΔGRQ = E(Rh2 ) – E(EnD ) (3.7)

E(EnD*/−) = E(EnD0/−) + E(3ππ*) (3.8)

Similarly, the driving force for oxidative quenching, ΔGOQ, is given by equation

3.9 and the excited state oxidation potential of the EnD, E(EnD+/*), is shown in equation

3.10.

+/ 0/− ΔGOQ = E(EnD *) − E(Rh2 ) (3.9)

E(EnD+/*) = E(EnD+/0) – E(3ππ*) (3.10)

In Tables 3.4 and 3.5, 1, 2 and 4 are used as examples to calculate the driving force for reductive oxidative quenching. The driving forces for the reductive quenching by 1 and 2 and oxidative quenching for 1 – 6 are positive, which indicates these processes are not thermodynamically favorable. Interestingly, reductive quenching is thermodynamically possible for 4 with all the EnDs, however, there is no trend between the quenching 51 constant and the reductive quenching driving force. In addition, no charge transfer products were observed in transient absorption experiments with the EnDs in the presence of 4. As a result, the quenching mechanism is primarily energy transfer.

Table 3.4 Reduction potential, E(EnD0/−), excited state reduction potential, E(EnD*/−) and driving force for reductive quenching ΔGRQ with 1, 2 and 4. ΔGRQ/V Donor E(EnD0/−)a/V E(EnD*/−)/V 1 2 4

Anthracene −1.71 +0.12 +0.94 +0.78 −0.32

Perylene −1.43 +0.10 +0.96 +0.70 −0.30

Tetracene −1.34 −0.05 +1.11 +0.85 −0.15

Rubrene −1.17 −0.02 +1.08 +0.82 −0.18

a. Potential vs. NHE in DMF11

Table 3.5 Oxidation potential, E(EnD+/0), excited state oxidation potential, E(EnD+/*) and driving force for oxidative quenching ΔGOQ with 1, 2 and 4.

ΔGOQ/V Donor E(EnD+/0)a/V E(EnD+/*)/V 1 2 4 Anthracene +1.33 −0.50 +0.25 +0.38 +0.26

Perylene +1.09 −0.44 +0.31 +0.44 +0.32

Tetracene +1.01 −0.28 +0.47 +0.60 +0.48

Rubrene > +1.1 > −0.05 > +0.70 > +0.83 > +0.71

a. Potential vs. NHE in DMF11

3.2.3 Excited State Electron Transfer

The ability to transfer electrons to other molecules is a prerequisite for dye sensitizer or a photosensitizer in photocatalytic system. As a photosensitizer, a long triplet excited state lifetime is important for bimolecular charge transfer. Complex 2, with axially blocking acetate group, possesses a triplet metal-ligand to ligand charge transfer

(3ML-LCT) excited state lifetime of 19 ns which is about 30 times increasement comparing to its unblocked analogue, 1 (T = 0.64 ps). Hereby, 2 as an electron donor,

III Co (dmgH)2(py)(Cl) (dmg = dimethylglyoxime, py = pyridine; shown in Figure 3.5), a widely applied hydrogen evolution catalyst was used as an electron acceptor to study the electron transfer event in a photocatalytic system. Nanosecond transient absorption was used to examine the excited state spectral change with the addition of the CoIII in 2.13

III Figure 3.5 Structure of Co (dmgH)2(py)(Cl).

Without the addition of CoIII, the excitation of 2 in acetonitrile at 532 nm generated the triplet excited state (3ML-LCT) with absorption peaks around 520 nm and

800 nm and ground state bleach (GSB) ~650 nm as shown in the transient absorption spectrum at 24 ns after the laser pulse in Figure 3.6 (blue trace). The kinetic trace at 480 nm were fitted monoexponentially with a triplet excited state lifetime T~18 ns as reported.4

In contrast, with 5 mM CoIII (Figure 3.6, red trace), upon the excitation of the mixture under the same condition, the 3ML-LCT of 2 decays faster due to the quenching by CoIII as shown in the equation below. In other words, upon excitation, [2]2+* donates an electron to CoIII to generate one electron reduced CoII and oxidized [2]3+. In the red trace of Figure 3.6, at 24 ns after the laser excitation, the GSB at 650 nm corresponding to 2 overlaps with the positive absorption feature of the one electron reduced CoII. A more pronounced spectra, without the complexation of [2]2+* at 86 ns, with a broad absorption from 450-800 ns and a peak at 470 nm is assigned to CoII (shown in Figure

3.7).16 The kinetic trace at 650 nm reveals that right after excitation, GSB of 2 is formed and followed by the rise of CoII in ~20 ns. CoII and [2]3+ recombines in ~88 ns with the rate constant of 1.14 × 107 s−1.

2+ III 3+ II [2] * + Co (dmgH)2(py)(Cl) → [2] + Co (dmgH)2(py)(Cl)

TA@24ns

10 OD

 0

-10 0.25 mM Rh2

-3 0.13 mM Rh2 + Co -20x10

400 500 600 700 800  / nm

Figure 3.6 Transient absorption spectra comparison of 0.25 mM 2 (red) and 0.14 mM 2 III with 5 mM Co (dmgH)2(py)(Cl) (blue) in CH3CN at 24 ns after laser pulse (exc= 532 nm, 5 mJ/pulse)

-3 12x10 TA@86ns with Co

OD  4

0 400 500 600 700 800  / nm

III Figure 3.7 Transient absorption spectrum of 0.13 mM 2 and 0.5 mM Co in CH3CN at 86 II ns after excitation (exc= 532 nm, 5 mJ/pulse), which is assigned to Co .

0 OD

 -10 Coefficient values ± one standard deviation y0 =0.00014518 ± 0.000358 -20 A =0.01744 ± 0.00073 tau =8.8321e-008 ± 8.41e-009 Constant: -3 -30x10 X0 =5.8e-008

-9 0 100 200 300 400x10 t / s

III Figure 3.8 Kinetic trace at 650 nm of 0.13 mM 2 and 0.5 mM Co in CH3CN (exc= 532 nm, 5 mJ/pulse).

3.3 Conclusions

The energy difference between the zero vibrational levels of the ground state and

T the lowest lying triplet excited state, E00 , for the non-emissive dirhodium formamidinate complexes 1 - 6 were determined by energy transfer quenching experiments with a series of organic compounds with known E(3*) energies. Stern-Volmer plots were constructed from the lifetime quenching of the 3* states of the organic energy transfer donors monitored using transient absorption spectroscopy to obtain the energy transfer

T T 3 rate constant, kEn . The values of kEn were found to increase as the magnitude of E( *)

T increases. Complexes 1 and 2 were able to quench rubrene such that their E00 values were estimated to be ~1.1 eV. In contrast, 4 – 6 are not able to undergo energy transfer

T with rubrene but quenched teracene, resulting in E00 ~ 1.2 eV. Oxidative and reductive quenching of the 3* states were ruled out based on the lack of charge transfer products in the transient absorption spectra. Moreover, for most sensitizer/Rh2 pairs, charge transfer was calculated to be thermodynamically unfavorable. In addition, the values of the quenching rate constants do not follow a trend with the ease of reduction or oxidation driving force. Finally, bimolecular charge transfer was achived from the triplet excited

III state of 2 to a known hydrogen evolution catalyst, Co (dmgH)2(py)Cl, resulting in the

II observation of the one-electron reduced cobalt complex, Co (dmgH)2(py)Cl. These results indicate that these complexes have the potential to be useful as sensitizers in

DSSCs and related solar energy conversion schemes, such as photocatalysis and dye- sensitized photoelectrosynthesis cells (DSPECs).

Bibliography

1. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737-740.

2. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595-6663.

3. J. Moan and A. Juzeniene, J. Photochem. Photobio. B, 2010, 101, 109-110.

4. T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

5. T. A. White, K. R. Dunbar, R. P. Thummel and C. Turro, Polyhedron, 2016, 103, 172-177.

6. H. J. Sayre, A. Millet, K. R. Dunbar and C. Turro, Chem. Commun., 2018, 54, 8332-8334.

7. Z. Li, N. A. Leed, N. M. Dickson-Karn, K. R. Dunbar and C. Turro, Chem. Sci., 2014, 5, 727-737.

8. M. Buchalska, J. Kuncewicz, E. Świętek, P. Łabuz, T. Baran, G. Stochel and W. Macyk, Coord. Chem. Rev., 2013, 257, 767-775.

9. S. Ardo and G. J. Meyer, Chem. Soc. Rev., 2009, 38, 115-164.

10. Y. Liu, R. Hammitt, D. A. Lutterman, R. P. Thummel and C. Turro, Inorg. Chem., 2007, 46, 6011-6021.

11. S. L. Murov, Handbook of Photochemistry, Marcel Dekker, New York, 1973.

12. P. C. Vincenzo Balzani, Alberto Juris, Photochemistry and Photophysics: Concepts, Research, Applications, Wiley, Weinheim, Germany, 2014.

13. M. Kumar, E. Natarajan and P. Neta, J. Phys. Chem., 1994, 98, 8024-8029.

Chapter 4. Panchromatic Dirhodium Photosensitizers and Electron Injection to TiO2

4.1 Background

Previous work has demonstrated that Rh2(II,II) formamidinate complexes of the type [Rh2(DTolF)2(L)2][BF4]2 (DTolF= p-ditolylformamidinate; L = neutral bidentate chelating or bridging diimine ligand) feature a low energy transition that corresponds to the movement of electron density from the Rh2(*)/formamidinate (*) highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) that is localized on an unoccupied * MO of the diimine ligand.1-3 Upon excitation, the metal/ligand-to-ligand charge transfer (1ML-LCT) excited state is populated and undergoes intersystem crossing to generate the corresponding 3ML-LCT state. In

2+ [Rh2(DTolF)2(np)2] (np = 1,8-naphthyridine; 1), the structure of which is shown in

Figure 4.1, the lifetimes of these singlet and triplet states were measured to be 13 and 640 ps, respectively, in CH3CN. Synthetic modification of 1 with ligands that are able to

coordinate to the axial positions, such as in Rh2(DTolF)2(npCOO)2 (2, shown in Figure

3.1), result in the extension of the 3ML-LCT lifetimes from 640 ps in 1 to ~25 ns in 2.3

Excited state charge transfer reactions with methyl viologen (MV2+) and p- phenylenediamine (PDA) have demonstrated that some Rh2(II,II) complexes can perform bimolecular charge transfer, acting as both excited state reductants and oxidants.

The broad absorption and excited state redox properties of Rh2(II,II) complexes have already resulted in their application as photosensitizers for H2 production upon 655 nm excitation.4 Herein, the charge injection by the panchromatic dyes 7 and 8 (Figure

4.1) into the n-type semiconductor, anatase TiO2. Complexes 7 and 8 were anchored to the TiO2 surface and the dynamics of electron injection and recombination were measured using ultrafast transient absorption and time-resolved IR (TRIR spectroscopies) upon low energy irradiation, 600 nm for 7 and 520 nm for 8.

Figure 4.1 Schematic representation of Rh2(II,II) complexes

4.2 Results and Discussion

4.2.1 Electronic Absorption and Electrochemistry

Complexes 7 and 8 were prepared according to published synthetic procedures for

2 the related complexes [Rh2(DTolF)2(np)2][BF4]2 (1) and [Rh2(DTolF)2(bpy)2][BF4]2 (9;

(bpy = 2,2'-bipyridine),5 which are described in detail in the Chapter 2. The bridging np

60 ligands shorten the Rh-Rh bond length to 2.4466(7) Å in 1 compared to 2.5821(5) Å in 9, which contains chelating bpy ligands.5,6 The shorter Rh-Rh bond length results in a destabilized Rh2(σ*) MO composed on the antibonding linear combination of the dz2 orbitals in each metal as explained in Chapter 1 and, consequently, lead to higher energy

3 6 metal centered ( MC) states. Based on comparisons with structurally related Rh2(II,II) complexes, it is expected that the Rh-Rh bond length of 7 will be similar to that for 1, and that for 8 close to the Rh-Rh distance reported for 9.1,7

The steady-state electronic absorption spectra of 7 and 8 in CH3CN are shown in

Figure 4.2 and summarized in Table 4.1. A broad, low-energy absorption band with maximum at 595 nm ( = 1,700 M–1cm–1) is observed for 8, which shifts to a lower energy in 7, 630 nm ( = 2,800 M–1cm–1). These transitions are consistent with those of

–1 –1 previously reported complexes, such as 1 (abs = 566 nm,  = 3600 M cm ) and

–1 –1 [Rh2(DTolF)2(dpq)2][BF4]2 (4, abs = 525 nm,  = 1300 M cm ). Based on these comparisons, these bands are assigned as 1ML-LCT in nature arising from a

Rh2(δ*)/DTolF(π*)→menp(π*) transition in 7 and Rh2(δ*)/DTolF(π*)→dmeb(π*) in 8.

Figure 4.2 Electronic absorption spectra of 7 (solid) and 8 (dashed) in CH3CN

Table 4.1 Electronic Absorption Maxima, Extinction Coefficients, and Reduction Potentials of 1 and 7 – 9 in CH3CN.

3 –1 –1 a Complex λabs / nm (ε / 10 M cm ) E1/2 / V 1 300 (24.9), 436 (1.8), 566 (3.6)b +0.87, −0.94, −1.12c 7 271 (27.9), 306 (22.7), 450 (1.7), 630 (2.8) +0.93, −0.61, −0.72, −1.33 8 241 (30.0), 312 (31.0), 387 (9.8), 595 (1.7) +1.08, −0.28, −0.81, −1.16 9 272 (28.3), 418 (2.3), 510 (0.6) +0.86, −0.45, −1.20, −1.74d a b [Complex] = 0.5 mM, 0.1 M Bu4NPF6 CH3CN under N2 atmosphere; vs Ag/AgCl. From ref. 6. cIrreversible, from ref. 8 dFrom this work.

The half-wave potentials measured for 7 and 8 in CH3CN and are listed in Table

4.1, and compared to those of 1 and 9. Complexes 7 and 8 exhibit a reversible one-

III,II/II,II electron oxidation corresponding to Rh2 couples at +0.93 V and +1.08 V vs

Ag/AgCl in CH3CN (0.1 M Bu4NPF6), respectively, which compare well to those measured for 1 and 9 at +0.87 V and +0.86 V vs Ag/AgCl, respectively, in the same solvent. The first and second cathodic couples of 7 are assigned to the sequential reduction on the two menp ligands and are observed at −0.61 V and −0.72 V vs Ag/AgCl.

These values compare well to those reported for the reduction of the np ligands in 1 at –

0.81V and –1.16 V vs Ag/AgCl. The anodic shifts observed of the reduction of menp ligands in 7 as compared to the np ligands in 1 is attributed to the electron withdrawing methyl ester group on the menp ligand.2 For 8, the first cathodic couple at −0.28 V is

II,II/II,I, assigned to the one-electron reduction of the bimetallic core, Rh2 The reduction of one dmeb ligand is observed in 8 at −0.81 V, followed by another reversible couple at –

1.16 V vs Ag/AgCl in CH3CN (Table 1). A similar anodic shift of the ligand-based reduction of dmeb in 8 relative to that of bpy in 9 arises from the presence of the electron donating methyl ester substituent.

4.2.2 Excited State Properties of the Dirhodium Dyes

The photophysical properties of 7 and 8 were examined in CH3CN and compared to those of the corresponding parent complexes, 1 and 9, respectively. The femtosecond time-resolved infrared spectra (fsTRIR) of 7 following 600 nm excitation (IRF = 85 ps)

(Figure 4.3) exhibit two ground state bleach features at 1512 cm–1 and 1574 cm–1 corresponding to two asymmetric (N=C−N) stretches of the DTolF ligand, as previously reported for related complex 7 at 1507 and 1577 cm–1.2 One (C=O) stretch bleach at 63

1732 cm-1 associated with the methyl ester functional group of 7 is also observed upon excitation, consistent with the ground state IR stretch of the complex and that of the free menp ligand at 1729 cm–1 (Figure 4.3a). The excited state (N=C−N) stretches shift to higher energy, 1536 cm-1 and 1600 cm-1, whereas the (C=O) vibration is observed at

1673 cm-1 in the excited state (Figure 4.3a). The shift of the DTolF vibrations to higher energy and ester vibration to lower energy in the excited state arise from the movement of electron density from the Rh2/DTolF MO to the menp ligand, which confirms the assignment of an ML-LCT excited state in this complex. The lifetimes of the positive signals can be fitted to biexponential functions with lifetimes of 2.6 ps and 429 ps associated with the 1ML-LCT and 3ML-LCT excited states, respectively. Similar excited state features were reported for 1 with 1ML-LCT and 3ML-LCT lifetimes of 14 ps and

2 640 ps in CH3CN.

Similar results were observed with 8, where the fsTRIR spectra exhibit three asymmetric (N=C−N) vibrations ground state bleaches at 1520 cm–1, 1580 cm–1 and

1620 cm–1, and one (C=O) bleach at 1740 cm-1 (Figure 4.3b). The lack of the positive excited state absorption signal associated with the (N=C−N) bleaches could due to a smaller extent of charge transfer excited state resulting from enhanced mixing with the lower-energy MC state in 8 as compared to 7, attributed to the longer Rh-Rh bond in the former. The excited state (C=O) stretch of 8 is observed at 1720 cm–1. The smaller C=O shift in 8,  = − 20 cm–1, as compared to that in 7,  = − 60 cm–1, is attributed to the presence of two symmetric methyl ester groups on the dmeb ligand relative to only one on menp. It is expected that there is a higher degree of polarization in the ML-LCT 64 excited state of 7 because the electron density will be localized asymmetrically on one side of the ligand. Similar effects on the magnitudes of (C) shifts in the excited states were previously reported for ruthenium polypyridyl complexes. [Ru(bpy)2(4,4'-

2+ (CO2Et)2bpy)] , with a symmetric ethyl ester substituted bpy ligand, shows a smaller shift when compared to the asymmetrically substituted complex, [Ru(bpy)2(4-CO2Et-4'-

2+ 9 CH3bpy)] .

Figure 4.3 fsTRIR spectra of (a) 7 and (b) 8 in CD3CN (λexc = 600 nm, 2µJ) and corresponding ground state IR spectra (dashed lines).

Figure 4.4 FsTA spectra of (a) 7 (λex = 600 nm, 2µJ) and (b) 8 (λexc = 520 nm, 2µJ) in CH3CN

The femtosecond transient absorption (fsTA) spectra of 7 and 8 are characterized by broad positive signal in the visible region, as shown in Figure 4.4. 7, corresponding to excited state absorption that in the 400 to 540 nm range, with two peaks at 435 and 510 nm. The spectra at wavelengths longer than 540 nm were not collected due to the scattered light from 600 nm pump. The lower intensity from 450−500 nm is attributed to the ground-state bleach band ~450 nm, which overlaps with the positive signal. The absorption changes at 440 nm as a function of time were fitted to a biexponential function, resulting in decays with 1= 4 ps (43%) and 2 = 460 ps (57%). Previously 66 reported fsTA spectra for 1 also show broad absorption from 350 to 550 nm with peaks at

~400 and ~500 nm, which superimpose well with the features from the oxidized

Rh2(III,II) measured using spectroelectrochemistry. These results are consistent with the fsTRIR data, where the short and long component of the excited states are assigned to

1ML-LCT and 3ML-LCT, respectively.

As shown in Figure 4.4b, the fsTA spectra of 8 features a broad absorption band from 450 to 650 nm with an apparent maximum at ~640 nm in CH3CN (exc = 520 nm,

IRF = 85 fs). The decay at 640 nm is fitted biexponentially to 2.5 ps (46%) and 56 ps

(54%). The spectroelectrochemistry of 8 collected at an applied potential, Eapp, of +1.37

V vs Ag/AgCl, expected to result in the formation of the one-electron oxidized Rh2(III,II) complex, which exhibits minimal spectral changes (Figure 4.5). In contrast, reduced 8, recorded at Eapp= −570 mV corresponding to the Rh2(II,I) complex, possesses characteristic peaks at 470 and 570 nm that are not present in the fsTA spectrum (Figure

4.4b). The observed broad absorption at  > 500 nm is consistent with * transitions of the reduced dmeb ligand, as previously reported for complexes with one-electron reduced

10-12 bpy and 4,4ʹ-(COOEt)2-2,2ʹ-bpy ligands.

0.5 Reduction at -570 mV 0.4 Oxidation at +1370mV

0.3

OD 0.2 

0.1

0.0

-0.1 400 500 600 700 800  / nm

Figure 4.5 Spectroelectrochemistry of 8 at applied potential of +0.137 V (blue spectrum) and −0.570 V (red spectrum) in 0.1 M Bu4NPF6 CH3CN

It is important to note that the electrochemistry shows that the lowest unoccupied molecular orbital (LUMO) of 8 is localized on the dirhodium core, resulting in a spectrum of the ground state one-electron reduced complex associated with the Rh2(II,I).

However, the excited state detected using fsTA resembles the spectral features of reduced dmeb ligand, demonstrating that the lowest excited state must not correspond to the

HOMO-LUMO transition. Together with the electron density shift observed in the fsTRIR of 8, the results point at singlet and triplet excited states that are ML-LCT in character. In the case of 7, the first reduction is attributed to placement of an electron on the menp ligand, since the bridging nature of the substituted np ligand raises the energy of the Rh2(*) MO. In this case, the results are consistent with a lowest excited state in which the LUMO localized on menp(*) is populated in the ML-LCT state.

The electron transfer from excited Ru(I) dyes to TiO2 has been shown to take place from both the 1MLCT and 3MLCT states, although the singlet state is short-lived,

<50 fs, such that extended lifetimes are not necessary for a dye to inject electrons into n- type semiconductors.13,14 Given the longer 1ML-LCT lifetimes of 7 and 8 as compared to

Ru(II) sensitizers, it is possible that these states would be able to be useful in similar charge injection into semiconductors. To this end, the singlet excited states oxidation

1 potentials, *Eox, of 7 and 8 were estimated to determine their driving forces for electron injection to TiO2 and the modified Latimer diagrams for 7 and 8 are shown in Figure 4.6.

1 S The energy of the ML-LCT state, E00 , for 7 is estimated to be ~1.7 eV from the tail of

1 its electronic absorption spectrum, resulting in *Eox ~ −0.8 V vs Ag/AgCl. The same

1 protocols were applied to 8 which yield *Eox ~ −0.7 V vs Ag/AgCl. As a result, the

1 higher *Eox compared to the lower limit of the TiO2 (ECB ~ −0.4 eV vs Ag/AgCl) conduction band makes electron injection thermodynamically favorable for both complexes. However, electron injection from the triplet excited state is not expected to be

3 thermodynamically favorable, with calculated values of *Eox of ~ –0.2 V and ~ –0.1 V vs Ag/AgCl for 7 and 8, respectively.

Figure 4.6 Modified Latimer Diagram for 7 (a) and 8 (b)

4.2.3 Electron Injection to TiO2

FsTRIR was used to examine the excited state electron injection from 7 into TiO2.

The Mid-IR spectral region is ideal to monitor electron injection due to the broad absorption resulting from free moving electrons in the conduction band (CB) of TiO2, without any spectral features of the dye in the ground or excited state or in its oxidized

15 form. Complex 7 was anchored to the surface of TiO2 nanoparticles and purified as described in Chapter 2, and the resulting 7@TiO2 nanoparticles were suspended in

CD3CN for the spectroscopic studies. Unlike the fsTRIR spectra collected for 7 (Figure

4.3a), those of 7@TiO2 show broad positive signal in the mid-IR region that decays over time (Figure 4.7a). This broad signal has been previously attributed to the free moving

1 14-18 electrons injected to the CB in TiO2 from the hot ML-LCT excited state.

Superimposed on this broad band absorption, a weak bleach at ~1570 cm–1 is observed

70 that corresponds to the (N=C−N) stretch of the anchored dye. In the absence of sensitizer, it should be noted that no broad electron signal was observed for TiO2 alone upon excitation under the similar experimental conditions. As expected from the absence of anchoring groups on the np ligand, control experiments with 1 and TiO2 did not result in charge injection or broad mid-IR signal, although the same sample preparation procedure was followed as in the case of 7.

Figure 4.7 fsTRIR of (a) 7@TiO2 (2 J, exc = 600 nm) and (b) 8@TiO2 in CD3CN (2 J, exc = 520 nm) 71

Upon excitation, the TiO2 CB mid-IR signal rises within the instrument response time ~130 fs, consistent with an ultrafast injection from the 1ML-LCT excited state of 7

(Figure 4.8a). The change in intensity of the signal at 1610 cm-1 as a function of time was fitted to a biexponential function, resulting in decay constants of 1 = 9 ps (59%) and 2 =

243 ps (28%) (Figure 4.8b). There is also a third long component that is beyond the range of the experiment, estimated 3 > 3 ns (13%). It has been previously reported that “hot” electrons injected from a higher energy level to the conduction band cool down, resulting in a decrease in the IR absorption cross-section to lower the signal.19 When slow injection from the triplet excited state was inhibited, a fast and slow electron decay component of

~90 ps and 10 ns, respectively, was observed by Lian for charge injection into TiO2 from

20 Re(CO)3Cl(dcbpy). Whereas, the oxidized Re signal showed negligible decay in 1 ns.

Therefore, the observed decay of the IR signal is attributed to the relaxation of the injected hot electrons. Similar assignments of ~50 ps decay was made by Chisholm for

21 cis-Mo2 paddlewheel complex anchored on TiO2. Herein, the ultrafast electron injection indicates that the injection occurs from an unrelaxed excited state and the fast 1 may be assigned to hot electron relaxation within the CB. However, we cannot rule out back electron transfer (BET) from the CB to the oxidized dye for this fast component, since it also leads to the decrease of the mid-IR signal.

BET has been reported to occur in the pico-, nano-, micro-, and millisecond timescales, depending on the specific electron donor and anchoring groups.15,20,21 It should also be pointed out that in nearly all experiments that measure the dynamics of charge recombination following electron injection into TiO2, two or three decay

72 components are observed when the changes are fit to multi-exponential functions. Given the heterogenous nature of the systems, many authors fit the decay data to stretch exponential function, but this method limits the physical interpretation of the system.

In the system studied by Gosh where Os(II)-polypyridyl dye adsorbed on TiO2, the decay rate of 12 ps (8%), 109 ps (7%) and >400 ps (85%) was observed. All three components are consistent with the dynamics of the dye recovery to the ground state; thus, they are assigned to BET. Similar BET was observed for the Ru(II) dye in this work with decay lifetimes of 120 ps (10%) and >400 ps (90%).22 Similar Ru(II) dye with resorcinol anchoring group studied by the same group, showed the recombination of 33

23 ps and >400 ps. In the system where dye fluorescein 27 adsorbed on a TiO2 surface studied by Sundstrom, multiexponential decay with charge recombination rate of ~10 ps

(~70%) and ns-s (~30%) was observed.24 The 10 ps component is assigned to germinate charge recombination to the same dye (same component observed in electrons on TiO2 and on the dye). However, they noted that this rate is 105 faster than expected from

Marcus theory. Multiexponential recombination from cresyl violet into SnO2 due to the semiconductor surface and trap states was reported by Kamat.25 The work from Hupp showed short and long recombination components for Ru(II) dyes on TiO2, where the short component is independent of pH and applied potential while the long component that attributes to the trap states recombination varies.26 The BET reported by Wachtyeitl, where the 1.5- 2.6 ps and 400 ps recombination arising from shallow trap states and deep trap states, respectively, shows pH dependence.27

Hereby, the longer component 2 and 3 were assigned to rapid BET. Similar results were observed for 8@TiO2 where upon 520 nm excitation, injected free moving electrons in TiO2 CB was observed (Figure 4.7b).

-1 Kinetic trace at 1787 cm

-6 a 500x10 Rh2@TiO2 400 fitted Rh2@TiO2 Coefficient values ± one standard deviation

300 OD y0 =0.0005178 ± 7.01e-006  A =-0.00047595 ± 2.94e-005 200 tau =0.13894 ± 0.0187 Constant: 100 X0 =-0.142239

0 -1 0 1 2 3 t/ps

-6 500x10 b Coefficient values ± one standard deviation y0 =6.6621e-005 ± 8.87e-006 400 A1 =0.00031498 ± 1.65e-005 tau1 =8.7444 ± 1.11 300 A2 =0.00015099 ± 1.58e-005 tau2 =242.77 ± 64.9

OD Constant:  200 X0 =0.642239

100

-100 0 500 1000 1500 2000 2500 Time / ps

-1 Figure 4.8 Kinetic trace of 7@TiO2 (a) at 1787 cm within 5 ps after excitation (b) at 1610 cm-1 with long decay components.

4.3 Conclusions

In conclusion, this work shows for the first time that Rh2 photosensitizers can be anchored to TiO2 nanoparticles and that they are able to inject electrons upon low energy visible light photosensitization for solar cell applications. Both 7 and 8 process picosecond-long 1ML-LCT excited state lifetimes and these singlet states decay into the corresponding 3ML-LCT excited state in each complex. The geometry of these complexes creates a charge separated excited state in which holes are localized on

Rh2(δ*)/DTolF(π*) MO and electrons are localized on menp(π*) or dmeb(π*) MO. The methyl ester group on the diimine ligand facilitates the binding to TiO2 and electron injection. The singlet and triplet excited state energy estimation predicted that 7 and 8 are energetic dyes for singlet injection but not thermodynamically favorable to inject electrons from their triplet excited states. As such, ultrafast electron injection is observed in these complexes without a slower component. The panchromatic dirhodium dyes can serve as to complement the known charge injection data of the well-known ruthenium polypyridyl dyes, as well as to harvest more lower energy photons to make full use of the solar spectrum. Like ruthenium dyes, synthetic modification on the bridging or chelating ligands are versatile to tune the energetics of the excited state to further develop energetic near-IR light absorbing dyes.

Bibliography

1. Z. Li, N. A. Leed, N. M. Dickson-Karn, K. R. Dunbar and C. Turro, Chem. Sci., 2014, 5, 727-737. 75

2. T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

3. T. J. Whittemore, A. Millet, H. J. Sayre, C. Xue, B. S. Dolinar, E. G. White, K. R. Dunbar and C. Turro, J. Am. Chem. Soc., 2018, 140, 5161-5170.

4. H. J. Sayre, A. Millet, K. R. Dunbar and C. Turro, Chem. Commun., 2018, 54, 8332-8334.

5. H. T. Chifotides, K. V. Catalan and K. R. Dunbar, Inorg. Chem., 2003, 42, 8739- 8747.

6. T. A. White, K. R. Dunbar, R. P. Thummel and C. Turro, Polyhedron, 2016, 103, 172-177.

7. H. T. Chifotides, B. Saha, N. J. Patmore, K. R. Dunbar and J. K. Bera, Group 9 Metal–Metal Bonds, 2015.

8. T. J. Whittemore, H. J. Sayre, C. C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

9. K. M. Omberg, G. D. Smith, D. A. Kavaliunas, P. Chen, J. A. Treadway, J. R. Schoonover, R. A. Palmer and T. J. Meyer, Inorg. Chem., 1999, 38, 951-956.

10. C. M. Elliott and E. J. Hershenhart, J. Am. Chem. Soc., 1982, 104, 7519-7526.

11. G. A. Heath, L. J. Yellowlees and P. S. Braterman, J. Chem. Soc., Chem. Commun., 1981, DOI: 10.1039/C39810000287, 287-289.

12. D. M. Peloquin, D. R. Dewitt, S. S. Patel, J. W. Merkert, B. T. Donovan-Merkert and T. A. Schmedake, Dalton Trans., 2015, 44, 18723-18726.

13. K. Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh, H. Sugihara, Y. Dan-oh, C. Kasada, A. Shinpo and S. Suga, J. Phys. Chem. B, 2005, 109, 15476-15482.

14. A. Listorti, B. O’Regan and J. R. Durrant, Chem. Mater., 2011, 23, 3381-3399.

15. N. A. Anderson and T. Lian, Annu. Rev. Phys. Chem., 2004, 56, 491-519.

16. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595-6663.

17. L. C. T. Shoute and G. R. Loppnow, J. Am. Chem. Soc., 2003, 125, 15636-15646. 76

18. Y. Tachibana, S. A. Haque, I. P. Mercer, J. R. Durrant and D. R. Klug, J. Phys. Chem. B, 2000, 104, 1198-1205.

19. J. B. Asbury, N. A. Anderson, E. Hao, X. Ai and T. Lian, J. Phys. Chem. B, 2003, 107, 7376-7386.

20. J. B. Asbury, E. Hao, Y. Wang and T. Lian, J. Phys. Chem. B, 2000, 104, 11957- 11964.

21. S. E. Brown-Xu, M. H. Chisholm, C. B. Durr, T. L. Gustafson and T. F. Spilker, J. Am. Chem. Soc., 2014, 136, 11428-11435.

22. T. Banerjee, A. K. Biswas, T. S. Sahu, B. Ganguly, A. Das and H. N. Ghosh, Dalton Trans., 2014, 43, 13601-13611.

23. T. Banerjee, S. Kaniyankandy, A. Das and H. N. Ghosh, Inorg. Chem., 2013, 52, 5366-5377.

24. M. Hilgendorff and V. Sundström, J. Phys. Chem. B, 1998, 102, 10505-10514.

25. D. Liu, R. W. Fessenden, G. L. Hug and P. V. Kamat, J. Phys. Chem. B, 1997, 101, 2583-2590.

26. S. G. Yan and J. T. Hupp, J. Phys. Chem., 1996, 100, 6867-6870.

27. V. V. Matylitsky, M. O. Lenz and J. Wachtveitl, J. Phys. Chem. B, 2006, 110, 8372-8379.

Chapter 5. Single Molecule Dirhodium Photocatalyst for H2 Generation Using Low

Energy Light

5.1 Background

5.1.1 Multicomponent Photocatalytic System

In natural photosynthesis, H2O is oxidized to O2 and electrons, which are used to generate biological reductant NAD(P)H from NAD(P)+.1 Eventually, NAD(P)H is used to reduce CO2 to form biomass and O2 as a byproduct of water oxidation. In other words, photosynthesis utilizes the energy provided by solar light to drive the endothermic reaction, reduction of CO2 to biomass to store solar energy in the form of chemical bonds. Traditional homogeneous artificial photocatalytic system mimics natural photosynthesis, which features separate chlorophyll derivative light absorbing units, photosystem I and photosystem II, and manganese-based catalyst center, as well as quinone based electron shuttle.2 In other words, multicomponent photocatalytic system usually involves electron donor (ED), light absorber (LA), electron relay (ER) and catalyst (CAT), shown in Figure 5.1. Upon excitation, in the case of oxidative quenching, light absorber is pumped to its energetic excited state which is oxidatively quenched by

ER and the electron is transferred to CAT. Or in some systems, direct oxidative quenching by CAT without an ER is also viable. The oxidized LA+ can be regenerated by

78 the ED in the system. In other cases, excited LA* maybe reductively quenched by ED to form LA− and followed by electron transfer to CAT to perform catalysis. Both cycles need to be repeated twice for the 2-electron process for proton reduction.

Figure 5.1 Traditional multicomponent photocatalysis system for proton reduction shown in oxidative quenching pathway.

To be an efficient LA, a broad-band absorption profile across the solar spectrum from ultraviolet to visible and extended to near IR region is desirable. A longer excited state lifetime (> ns) is also of interest for efficient diffusion controlled bimolecular quenching processes.3 The ground and excited state redox potentials are required to carefully match the energy levels of ED, ER and CAT for efficient electron transfer.

Basically, for oxidative quenching that involves LA* and ER, thermodynamically favorable electron transfer can occur when E(LA+/*) < E(ER0/−) and E(LA+/0) >

E(ED+•/0). In the case of reductive quenching mechanism, photoinduced electron transfer involves LA* and ED, thus E(LA*/−) > E(ED+•/0) while E(LA0/−) < E(ER0/−). Reductive

79 quenching mechanism is slightly more favorable than oxidative quenching due to LA− is a stronger reductant than LA*, rendering a wider choice of CAT and more driving force.

2+ The well-known Ru(bpy)3 is an energetic LA with excited state reduction and oxidation potential of +0.84 V and −0.86 V vs. NHE, respectively. It has an absorption maximum at 452 nm and tails down to 500 nm in water. The triplet excited state lifetime is about 650 ns in water. Other Ru (II) and Ir (III) d6 pseudo-octahedral complexes4-10 have been studied extensively due to their relative ease of synthesis, visible-light- absorbing character, long excited state lifetimes for bimolecular electron transfer and tunable photoredox properties. They have been utilized with a wide range of catalyst for hydrogen generation in multicomponent systems, such as Co4-6, Fe7-9 and Ni10 catalysts.

However, the Ru and Ir LAs suffer from poor absorption in the red to near infrared wavelengths region which limits their utilization of the full solar spectrum.

Recent efforts have spent to establish functioning LAs using more earth-abundant metals, such as Fe(II)11-14, Cr(0)15, Zr(IV)16,17 and Cu(I)18-21 to lower the cost and to enhance our understanding of their complicated excited state dynamics, allowing for synthetic modifications to control excited state properties such as emission quantum yield, lifetime, and redox potentials. Nonetheless, first row transition metal complexes also have limited absorption in the lower energy region of the solar spectrum and typically suffers from short-lived excited state and instability.

Some of the dirhodium complexes explored by our group has shown photocatalytic H2 generation in multicomponent systems using red light. Panchromatic

photosensitizer Rh2(DTolF)2(npCOO)2 (2, DTolF = N,N’-di-p-tolylformamidinate,

-1 -1 npCOO = 1,8-naphthyridine-2-carboxylic acid; max=640 nm,  =3500 M cm in

2+ CH3CN) and Rh2(DTolF)2(qnnp)2 (10, qnnp = 2-(quinolin-2-yl)-1,8-naphthyridine,

-1 -1 max=704 nm,  =3400 M cm in CH3CN) with nanosecond long lifetime, 25 ns and 7 ns, respectively, can photocatalytically produce H2 under 655 nm irradiation with methyl viologen (MV2+) as ER, Pt nanoparticles as CAT and 1-benzyl-1,4-dihydronicotinamide as sacrificial ED.22 The catalytic reaction with 2 processes through oxidative quenching mechanism while 10 operates by reductive quenching mechanism. This system shows

2+ comparable activity using Ru(bpy)3 as LA under 447 nm irradiation under similar experimental conditions which demonstrates the ability of dirhodium complexes to achieve photocatalysis with low energy red light.

In traditional multicomponent systems, LA and CAT are separated which requires photoinduced electron transfer from the LA to the CAT to propel the reaction. Catalysis is largely dependent on the efficient electron transfer step. The supramolecular approach, in which chromophore and catalyst are covalently linked together, eliminates diffusion controlled bimolecular electron transfer. Ideally, the supramolecule should contain a chromophore that absorbs broadly and facilitate the accumulation of multiple charge carriers allowing multi-electron reactions, such as two-electron reduction of proton to hydrogen and carbon dioxide to carbon monoxide. Bimetalic assembles including Ru-

Pt23, Ru-Rh24, Ir-Co25, zine porphyrin-cobaloximes26; trimetallic varietals Ru-Rh-Ru27-29;

30 tetrametallic Ru3-Pt complexes, and pigment–acceptor–catalyst triad having a

31 PtCl2(dpbpy) chromophore, [PtCl2(dpbpyMV4)], have been studied and showed promising photocatalytic H2 generation by Brewer, Sakai and other groups.

However, the eventual goal of scaling these systems to be fully heterogeneous requires LA-CAT interactions as specific distances and orientations, which remain difficult to control at the surfaces and in supramolecular systems. Therefore, the design of a convergent system which LA-CAT units are inseparable is crucial for the efficient photocatalytic reactions.

5.1.2 Single Molecule/Chromophore Photocatalysts

Sensitizer-free single molecule/chromophore photocatalyst, which can act as both

LA and CAT is extremely rare. In this type of system, light absorption and subsequent catalysis proceed within one molecular unit without electron transfer between LA and

CAT. Furthermore, the simplicity of this system could allow for lower cost for further improvement on heterogeneous surface for bulk production.

The pioneering work reported by Zeissel and coworkers in 1983 revealed

-1 -1 Re(CO)3(bpy)Cl (bpy= 2.2′-bipyridine, max = 392 nm,  = 2680 M cm ) can photocatalytically reduce CO2 to produce CO under UV-light irradiation without additional sensitization.32 Ishitani group reported iridium (III) complex, [Ir(tpy)(R‐ ppy)Cl] (R=H, Me, CF3, tpy = 2,2′:6′,2′′-terpyridine; ppy = 2-phenylpyridine, which can selectively reduce CO2 to CO with triethanolamine (TEOA) as electron donor in CH3CN under visible light (410-750 nm) irradiation with a TON of about 38 in 5 hours. This system is twice as efficient as the well-known Re(CO)3(bpy)Cl under the same

33 experimental condition. Fujita group synthesized the CO2 reduction key intermediate

+ -1 -1 for iridium catalyst, the hydride [Ir(tpy)(ppy)H] (max = 500 nm,  ~ 1300 M cm ),

82 which can produce CO under similar conditions with TON of about 50 over 5 h.34 More progress made by Masaoka group in 2018 has shown a phosphine-substituted Ru(II) polypyridyl complex, trans(P,MeCN)-[RuII(tpy)(pqn)-MeCN]]2+ (pqn = 8-

-1 -1 (diphenylphosphanyl)quinoline, max = 475 nm,  ~ 8500 M cm ), can photocatalytically produce CO with 1,3-dimethyl-2-phenyl-2,3-dihydro-1H-benzo[d]imidazole (BIH) under visible light irradiation (420~750 nm) with 353 TON in 24h.35.

A few photocatalysts for H2 generation have also been reported but with much

36 lower activity. Zampella group reported diiron hydride , [(μ-H)Fe2(pdt)(CO)4(dppv)]+

-1 -1 (dppv = cis-1,2-C2H2(PPh2)2, max = 500 nm,  ~ 1000 M cm ), can catalyze the formation H2 from triflic acid and ferrocene, however, it is oxygen sensitive and exhibit a maximum TON, of only 4. Recently, a (pyridyl)-N-heterocyclic carbene tungsten

37 -1 -1 tetracarbonyl complex , W(pyNHC)(CO)4 (max = 468 nm,  = 3700 M cm ), showed visible light driven H2 production in the presence of triﬂic acid and decamethylferrocene with a TON of 17 over 3h. In general, these H2 generation photocatalysts are inefficient.

Furthermore, poor light harvesting in the red and near-infrared (NIR) spectral regions by these systems is common because these photoredox processes require large driving forces for favorable multi-electron transformations. Moreover, the excited state lifetimes are limited by these red/near-IR light absorbing chromophores as predicted by the Energy

Gap law, thus reducing the time the molecule is spending in its energetic excited states.

Bimetallic complexes are some of the promising structures that can achieve this sensitizer free single molecule photocatalyst goal due to their ability to store multiple redox equivalents in one molecule and facile synthesis. In the early 1980s, the Gray

83 group reported noncatalytic hydrogen production by dirhodium(I) isocyanides complex from hydrohalic acids solution, which represents a seminal contribution to the field of solar fuels.38 The Nocera group reported mix-valance dirhodium complex,

II,0 Rh2 (dfpma)3X2(L) (X = Cl, Br; L = CO, PR3, CNR, bridged by dfpma = bis(difluorophosphino)methylamine), to produce hydrogen from hydrohalic acid.39 To date, the most efficient sensitizer-free dirhodium photocatalysts is the mix-valence system that can produce H2 with TON ~ 7 over 72 to 144 h with high energy ultraviolet light irradiation.40,41 Although the bimetallic complexes are of considerable interests due to the merits mentioned above, they suffer from instability in ambient environment and require high energy ultraviolet and visible light activation and limited robustness. Thus, better single molecular systems need to be explored to meet the objectives.

Last but not the least, in both multi- and single component photocatalytic systems, the electron donor (ED) is an important part. In natural photosynthesis, H2O is served as a sacrificial electron donor in the natural photosystem. However, in artificial photosynthesis, extracting electrons directly from H2O is particularly difficult and requires multiple processes including, light harvesting, charge accumulation, charge transportation and catalysis. To understand the overall photosynthesis, the oxidative half

(PSII) and reductive half (PSI) are investigated separately. While the oxidative artificial photosynthesis is crucial, in this work, the reductive half of photosynthesis is of particular interest. Instead of H2O, electron donors with milder oxidation potential were used to simplify the reductive half artificial photosynthesis. In a multicomponent system, except the thermodynamic requirements mentioned earlier, irreversible mono-electron oxidation

84 of ED to inert products to prevent back electron transfer is necessary. Aliphatic and aromatic amines, ascorbic acid, oxalate and NADP(H)+ derivatives are some of the commonly used sacrificial electron donor. The reaction solvent, pH value and the concentration of ED can strongly affect the overall efficiency of the photocatalytic systems.

5.2 Results and Discussion

5.2.1 Characterization

The Turro group has developed dirhodium(II,II) electrocatalysts,

2+ Rh2(DTolF)2(L)2 (L= chelating diimine ligands, phen (3; 1,10-phenanthroline), dppz (5; dipyrido[3,2-a:2’,3’-c]phenazine), dppn (6; benzo[i]dipyrido[3,2-a:2’,3’-h]quinoxaline)),

42 for H2 generation electrochemically in the presence of acetic acid in CH3CN and DMF.

The open axial coordination site is the key to substrate binding which facilitates the transfer of reducing equivalents. However, these complexes exhibit short triplet excited state lifetime which limits their application as LA. The chelating napthyridine ligands in

2+ complex 1, Rh2(DTolF)2(np)2 , maximize the symmetric overlap between the Rh2(*) and form (*) orbitals to ensure increased oscillator strength and shorten the Rh-Rh bond comparing to the bridging diimine analogues. The shorter Rh-Rh bond stabilizes the bonding Rh2() orbital and destabilizes the antibonding Rh2() orbital to raise the energy level of the deactivating metal-centered (MC) states, such that the lowest triplet state of np-containing complexes is charge transfer in character. Moreover, to extend the excited state lifetime, synthetic modification of dirhodium complexes with axial blocking 85 bridging ligand prevents solvent assisted nonradiative decay, 2 and 10 exhibit 25 and 7 ns long lifetime,43 while 1 is only about 640 ps.44 As described in 5.1.1, these panchromatic dirhodium(II,II) light absorbers have yielded application in multicomponent system.22

Nonetheless, the blocked axial position also inhibits substrate binding for catalysis.

Therefore, a dirhodium system with long excited state lifetime and open axial coordination site is required for the purpose of single molecule photocatalyst.

Figure 5.2 Structure of 11, where solvent in the axial positions are omitted.

2+ Rh2(form)2(bncn)2 (11, bncn = benzo[c]cinnoline, Figure 5.2) with the short N-

N bound ligand was synthesized and is stable to exposure and manipulation in air and water. Complex 11 absorbs broadly from the ultraviolet to visible and extended to near-

IR region, shown in Figure 5.3 and summarized in Table 5.1. Complex 11 exhibits ligand-centered * transitions at 332 nm ( = 12,000 M-1cm-1). A low-energy, broad absorption with maximum at 624 nm ( = 8,600 M-1cm-1) is also observed for 11 and 86 assigned as a Rh2(*)/form to bncn (*) ML-LCT transition, which tails down to ~ 800 nm. The short binding bncn ligand maximize the symmetric overlap between the Rh2(*) and form (*) orbitals to ensure increased oscillator strength and a larger extinction coefficient.

Figure 5.3 Electronic absorption spectrum of 11 in CH3CN.

Table 5.1 Electronic Absorption and Electrochemical Potentials of 11 in DMF

3 −1 −1 a Complex λabs / nm (ε / 10 M cm ) E1/2 / V

1 332(12.0), 401(5.6), 624(8.6) +1.28b, −0.38, −0.63

a n b Potentials taken vs Ag/AgCl in 0.1 M Bu4NPF6 in DMF, Taken in CH3CN

The crystal structure of 11 shows that the Rh-Rh bond length is 2.4049(5) Å, which is shorter comparing to the napthyridine analogue, 2.4466(7) Å for complex 1.

This shorter Rh-Rh bond contributes to a higher lying MC state which is responsible for nonradiative deactivation pathway. The reduced nonradiative deactivation in 11 is further demonstrated by the first observed 77 K emission for the dirhodium system, Figure 5.4.

The results from cyclic voltammetry measurements in DMF are shown in Figure 5.5 and summarized in Table 5.1, which indicates that 11 exhibits reversible RhIII,II/II,II oxidation at +1.28 V vs Ag/AgCl. The one electron reversible couples at −0.38 V and −0.63 V are assigned to the reduction on the two bncn ligands, respectively.

A 

Current / -5

-10

-15 0 -200 -400 -600 -800 -1000 -1200 -1400 Potential / mV (vs Ag/AgCl)

Figure 5.5 Cyclic voltammogram of 11 in DMF vs Ag/AgCl (calibrated with ferrocene) scan rate 200 mV/s.

Transient absorption was performed to study the photophysical properties of 11, shown in Figure 5.6 and Figure 5.7. Upon 630 nm excitation, a ground state bleach ~ 650 nm is observed. The spectra from 605 to 640 nm were omitted due to the scattering of the pump at 630 nm. An excited state absorption peak at 440 nm is observed and slightly blue-shifted to 430 nm over ~ 60 ps and lasted over the range of the ultrafast experiment.

1 The short component, 1, is fitted to 15 ps at 440 nm and assigned to the ML-LCT excited state, shown in Figure 5.6B. The picosecond long singlet excited state lifetime is comparable to other dirhodium complexes.43,44 While sub-picosecond intersystem rates are relatively common in transition metal complexes,45,46 some bimetallic complexes behave in a manner that similar to traditional organic molecules, including diiridium(I,I) pyrazolyl, diplatinum(II,II) pyrophosphito, dirhodium (I,I) isocyanide and a number of 89 quadruple bonded ditungsten and dimolybdenum carboxylates.46-49 These bimetallic complexes exhibit relatively long singlet lifetimes, ranging from 15 ps to 1.6 ns. The slower intersystem crossing rate in bimetallic complexes may be attributed to the rigidity of the paddlewheel architecture. To obtain the longer component, nanosecond transient absorption was performed, shown in Figure 5.7A. The excited state absorption at 430 nm and ground state bleach at 650 nm matches with the observation in the fsTA. An additional absorption peak at 380 nm was also observed in nsTA due to a wider response of the iCCD camera and more higher energy probe photons in the xenon arc lamp comparing to the white light probe generated by CaF2 crystal in fsTA. The longer component is fitted to ~ 19 ns at 420 nm which is assigned to 3ML-LCT excited state

(Figure 5.7B). The longer triplet lifetime comparing to the napthyridine analogue, 1, 640 ps is attributed to the higher lying MC excited state deactivation pathway due to the shorter Rh-Rh bond as explained earlier.44 Unlike 2 which requires axial blocking to achieve longer lifetimes, the long triplet lifetime of 11 is achieved without axial capping.43

Figure 5.6 (A)Femtosecond transient absorption of complex 11 (650 μM) in acetonitrile (630 nm excitation; 2.5 μJ) (B) Kinetic trace at 440 nm (hollow circles) and 650 nm (solid circles)

Figure 5.7 A.) Nanosecond transient absorption of 1 in DMF when excited with 630 nm light (5 mJ/pulse) and B.) kinetic trace taken at 420 nm.

The triplet excited state oxidation potential is calculated from equation 5.1, where the E00 is obtained from the emission maximum at 77 K, ~1.55 eV. The ability of the long-lived triplet excited state of 11 to serve as a strong oxidizing agent (E*red = +1.17 V

20 vs Ag/AgCl) was demonstrated using phenothiazine (PTZ, E1/2 = +0.59 V vs Ag/AgCl) as a reversible electron donor to yield singly reduced 11 and PTZ•+ radical cation upon

•+ excitation with red light, λex = 650 nm (Figure 5.8A). PTZ radical cation features the sharp peak at 420 nm and recombines with the reduced 11 over 19 s (Figure 5.8B).

ERED* = ERED − E00 (5.1)

Figure 5.8 Nanosecond transient absorption of (A) 11 in 15 mM PTZ acetonitrile solution taken at 30 μs (B) and kinetic trace taken at 420 nm

5.2.2 Electrocatalytic Proton Reduction

In electrochemistry, the electrons are provided by the working electrode from electricity without the complexation of electron transfer steps as in photocatalysis.

Investigating the electrocatalytical properties can facilitate the understanding of the role as a catalyst for the complex. The cyclic voltammogram of 11 has reversible 1-electron couples at −0.38 and −0.65 V vs Ag/AgCl in DMF (Figure 5.5). The reduction of 11 beyond its second reduction couple results in the electrocatalytic generation of dihydrogen in acidic solutions. Catalytic current enhancement is observed in voltammograms of 11 upon the addition of a number of acids, including dimethylformidinium triflate ([DMFH][OTf]), p-toluene sulfonic acid (TsOH), trifluoroacetic acid, and acetic acid (Figure 5.9 − Figure 5.12). It is estimated that the formation of the two-electron species may result in a reactive rhodium hydride intermediate, as has been previously reported in similar dirhodium electrocatalysts.42

Bulk electrolysis of 11 at −650 mV with 10 equivalents of trifluoromethanesulfonic acid demonstrates that 11.5 μmol of dihydrogen are formed over one hour, with a Faradaic

Efficiency of 98%.

In the presence of acid, the CV of 11 features a reversible pre-peak at ~ −200 mV that comprises a fraction of the overall current passed at the electrode. This feature is most apparent in the presence of the strong acid [DMFH][OTf] (pKa = 0) but is also present in lower proportion with weaker acids, including p-toluene sulfonic acid (pKa =

2.6) and trifluoroacetic acid (pKa = 6) (Figure 5.13). Similar to the free benzo[c]cinnoline ligand which shifts 122 mV/pH unit with strong acid, the first reduction couple of

94 complex 11 exhibits a positive shift of 120 mV/pH unit indicative of a two-proton, one- electron intermediate. This feature is reversible in 11, unlike the pseudo-irreversible (Epa-

Epc > 450 mV) behavior in the free ligand (Figure 5.14). Spectroelectrochemical measurements electrolyzing at −300 mV and at −180 mV vs Ag/AgCl yield no buildup of the protonated intermediate, which demonstrates that this species may not persist at long time scales. Electron transfer studies to isolate this protonated species using ferrocene as an electron donor were performed. In the presence of 13 mM ferrocene, complex 11 (50

µM) is readily reduced to form [11]1− upon excitation (λex = 630 nm). In ferrocene solutions containing [DMFH][OTf] or p-toluene sulfonic acid, [11]1− is also observed.

New spectral features are not present that could be associated with the protonated ligand

1− in [11] and the excited state lifetime of the reduced species in the presence of acid is identical to that measured for [11]1−, suggesting a difference in the photochemical and electrochemical behavior of this molecule attributed to the distinct conditions required in the two experiments. Similar studies using the experimental conditions of the photocatalytic solutions of 11 also yield the same results, where only [11]1− is observed.

Thus, the electrochemical intermediate is not anticipated to be involved in photocatalysis.

Figure 5.9 LSV upon addition of dimethylformamidinium triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11

Figure 5.10 LSV upon addition of p-toluenesulfonic acid triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11

Figure 5.11 LSV upon addition of trifluoroacetic acid in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11

Figure 5.12 LSV upon addition of acetic acid in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11

Figure 5.13 CV upon addition of dimethylformamidinium triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of 11

Figure 5.14 CV upon addition of dimethylformamidinium triflate in DMF (0.1 V/s) vs Ag/AgCl in 0.5 mM solution of benzo[c]cinnoline ligand

5.2.3 Photocatalytic Proton Reduction

The electrocatalytic production of dihydrogen upon the double reduction of 11, together with its excited state electron transfer properties, provide strong evidence that a photochemical approach may also generate H2. Irradiation of 11 (52 µM, 0.26 µmoles) in

DMF with red light (λex = 670 nm) in the presence of the sacrificial electron donor 1- benzyl-1,4-dihydronicotinamide (BNAH; 0.03 M) and p-toluene sulfonic acid (0.1 M) results in the production of hydrogen, with 84 turnovers in the first 3 hours and a maximum of 170 ± 5 TON (45 ± 3 µmoles) after 24 hours of irradiation (Figure 5.15)

Figure 5.15 Turnover number (TON) versus time graph of 11 (75 μM) irradiated with 670 nm LEDs for 24 hours in the presence of BNAH (0.03 M) and TsOH (0.1 M) produces over 170 equivalents of H2.

Because the majority of the BNAH is consumed during the catalysis (60% at 24 h), a rescue study was performed where the same initial amount of TsOH and BNAH were added to the catalytic mixture (530 and 140 µmoles, respectively), and the solution was irradiated for an additional 20 hours, which resulted in further H2 production (37

TON). Moreover, photolysis using near-infrared (NIR) light also led to hydrogen generation, resulting in 70 turnovers after irradiation with 735 nm light over 24 hours

(Figure 5.16). A longer induction period is observed with NIR light, attributed to the

−1 −1 lower extinction coefficient of the complex at this wavelength, 735 = 1300 M cm .

Similar results were recorded using [DMFH][OTf] and trifluroracetic acid. Control experiments demonstrate that each component is necessary to produce H2. No hydrogen was observed in the absence of p-toluene sulfonic acid, BNAH, complex 11, or light

(Table 5.2). In addition, the turnover numbers decreased upon the introduction of oxygen to the catalytic solution and the chemical oxidation of BNAH using AgBF4 in acidic solutions also inhibits H2 generation. Additional experiments using 2,

Rh2(DTolF)2(npCOO)2, a known red light-absorbing photosensitizer with a multi- nanosecond excited state lifetime,22 instead of 11 under the same 670 nm illumination, keeping the other solution components constant, does not yield any hydrogen over 24h.

100

20 70

60 H

2 15 ( Evolved

) 50 Cat

/n 40 H2 10



30 mol) TON (n 20 5 10

0 0 0 5 10 15 20 25 Time / h

Figure 5.16 Turnover number (TON) versus time graph of 11 (75 μM) irradiated with 735 nm LEDs for 24 hours in the presence of BNAH (0.03 M) and TsOH (0.1 M) produces over 70 equivalents of H2.

Table 5.2 Control experiments for each component and hydrogen production.

Entry Light 11 BNCN TsOH H2 1 x √ √ √ x

2 √ x √ √ x

3 √ √ x √ x

4 √ √ √ x x

5.2.4 Mechanism Studies

Reduced 11 is the key for catalysis, therefore, one-electron and two-electron reduced 11, [11]1− and [11]2−, were chemically prepared with the addition of one or two 101

1− equivalents of cobaltocene (EOX = −0.78 V vs Ag/AgCl in DMF, Figure 5.17). [11] evolves a new peak at 440 nm and the original lowest absorption maximum at 624 nm, which corresponds to 1ML-LCT, red shifted to ~680 nm. After adding another equivalent of cobaltocene, [11]2− was prepared and also shows the new peak at 440 nm but the lowest energy absorption band intensity decreases which is consistent with the loss of

ML-LCT transition from the Rh2(*)/form(*) to bncn (*) due to the reduction on both bncn ligands in [11]2−. The formation of these species was confirmed by comparing their electronic absorption spectra to those obtained using spectroelectrochemistry at applied potentials more negative than the first and the second reduction waves, respectively, under an inert N2 atmosphere in dry DMF in the absence of acid (Figure 5.18).

Figure 5.17 Reduced 11 ([11]1−) and doubly reduced 11 ([11]2−) chemically generated through the addition of 1 and 2 equivalents of cobaltocene in DMF

102

Figure 5.18 Difference spectrum associated with A.) reduced 11 ([11]1−) and B.) doubly 2− reduced 11 (([11] )) generated via electrochemically in 0.1 Bu4NPF6 in dry DMF

Irradiation of the solution used for catalysis (0.03 M BNAH, 0.1 M TsOH, 52 µM

11 in DMF) results in spectra consistent with [11]1− within a few minutes (Figure 5.19), and the spectral features of the solution continue to change over longer time scales to an intermediate with an absorption maximum at 600 nm, a feature inconsistent with the chemically generated two-electron reduced complex [11]2− (Figure 5.19). Continued

103 irradiation of this intermediate leads to further spectral changes and the production of hydrogen.

Figure 5.19 Difference in electronic absorption spectrum of catalytic mixture (0.1 mM 11, 0.1 M TsOH, 0.03 M BNAH in DMF) at A.) early times and B.) late times (bottom)

In order to probe the reactivity of [11]1−, the one-electron reduced species was chemically generated in the absence of acid (Figure 5.20, red curve) in DMF, and it was determined that its absorption spectrum did not change upon the addition of BNAH over 104 several hours in the dark. However, upon irradiation with 670 nm light, the spectrum of the solution containing [11]1− and BNAH changes over several minutes to hours to yield complete conversion to spectra consistent with the two-electron reduced species, [11]2−

(Figure 5.20, dashed line). The addition of TsOH to this solution results in the regeneration of the ground state of 1 and a quantitative amount of H2 (Figure 5.21, black solid line). This result clearly indicates that [11]2− reacts with acid to produce dihydrogen.

Figure 5.20 Conversion of 1-electron species ([11]1−) to 2-electron species [11]2− upon irradiation with 670 nm light

105

Figure 5.21 Addition of TsOH (0.1 M) to photochemically generated 2-electron species (65 μM 11, 0.2 M BNAH in DMF) (blue) yields ground state complex (black)

Figure 5.22 Proposed mechanism of the photocatalytic H2 production by 11 106

Taken together, the evidence presented herein supports the mechanism outlined in

Figure 5.22, where the catalysis proceeds through two stepwise redox events, where both

11 and [11]1− are capable of oxidizing BNAH from their respective excited states. In the absence of acid, the photolysis of [11]1− forms [11]2− through the irreversible BNAH oxidation. Electrochemical experiments show that [11]2− is catalytically active in the presence of acids such as [DMFH][OTf], p-toluene sulfonic acid, and trifluoroacetic acid

(Figure 5.9-Figure 5.11), but not acetic acid, which requires a third electron before catalytic current is observed. These findings parallel the photolysis results where hydrogen evolution is observed for [11]2− with the same stronger acid set, but not with acetic acid.

5.2.5 Photophysical and photochemical properties of [11]1−

The one electron reduced intermediate, [11]1−, can be irradiated and obtain an electron from BNAH to form [11]2− as shown in 5.2.4. Therefore, it is important to understand its photophysical and photochemical properties for the second electron transfer event. fsTA spectra of [11]1− is shown in Figure 5.23A. [11]1− is prepared by adding slightly over one equivalent of cobaltocene until seeing the absorption at 680 nm starts to decrease to ensure no ground state 11 was left in the solution to complicate the fsTA spectra. Upon excitation at 650 nm, an excited state absorption peak at 430 nm and a broader band at 520 nm are observed. The lower intensity in the 440 ~ 480 nm region is attributed to the superimposed ground state bleach signal with absorption maximum at

440 nm. The kinetic trace at 428 nm was fitted biexponentially with 1 = 0.6 ps and 2 =

107

450 ps. The sub-nanosecond excited state lifetime coincides with the long induction time in forming [11]2− after first generating [11]1− under irradiation in catalytic condition in a few minutes. However, with the excess amount of electron donor, BNAH, in solution, diffusion controlled bimolecular interaction is still viable.

108

T − The triplet excited state reduction potential, Ered ([11] ), is important for the second photoinduced electron transfer event. [11]− is not emissive at room temperature or

T 77 K, thus the triplet excited state energy (E00 ) cannot be readily obtained by luminescence. A series of organic compounds with known 3ππ* energies were used as

T − 44 energy-transfer donors (EnD) to estimate the E00 of [11] . Perylene, tetracene and rubrene were examined respectively with [11]−, results are summarized in Table 5.3.

Upon excitation of perylene and tetracene, the addition of [11]− resulted in the decrease in lifetime of the 3ππ* excited state comparing to the EnDs by themselves. The energy transfer quenching constant (KEnT) is calculated from the Stern-Volmer equation,

4.49×109 M-1s-1 and 5.38×108 M-1s-1, for perylene and rubrene, respectively. However, no quenching was observed in rubrene with the addition of [11]−. Reductive and oxidative quenching with the 3ππ* of EnDs were ruled out due to no redox products were observed.

Tetracene is used as an example and the kinetic traces are shown in Figure 5.24A. In

Figure 5.24B, the transient absorption spectra at 1 s is the same regardless of the addition of [11]− which indicates no redox product was formed. Since no quenching of

T – the rubrene triplet state was observed, the value of E00 of [11] must be greater than 1.15 eV. The fact that the quenching rate constant of tetracene is 1-2 orders of magnitude lower than the diffusion limit and an order of magnitude lower than that for perylene

T – points at a small driving force for energy transfer, thus placing E00 of [11] just below or

− T nearly equal to that of tetracene. Thus, for [11] , E00 is estimated ~ 1.29 eV. The excited

− −/2− state reduction potential Ered*(11 ) = E(11 ) + E00 = −0.63 V + 1.29 V = ~ +0.66 V vs

Ag/AgCl.

109

Table 5.3 Energy transfer donors and their 3* excited state energy (E(3*)), lifetime (0), excitation energy (exc) and energy transfer quenching constant (KEnT)

3 50 c -1 -1 EnD E( *)/eV 0/s exc/nm KEnT/M S Perylene 1.53 13.7a 450 4.48×109 Tetracene 1.29 13.1a 470 5.38×108 Rubrene 1.15 40.0b 525 No Quenching a. Measured in CH3CN b. Measured in DMF due to limited solubility in CH3CN c. laser power at sample 5 mJ/pulse

Figure 5.24 A. Kinetic traces at 460 nm (exc = 470 nm, 5 mJ/pulse). B. Transient absorption spectra at 1s for tetracene (red) and tetracene with 50 mM [11]1− added (blue) in CH3CN.

110

Another indirect estimation method is the formation of the doubly reduced product, [11]2–, following steady-state irradiation in the presence of sacrificial electron donors (EDs) with various oxidation potentials. As discussed earlier in 5.2.4 (Figure

5.21), photolysis of chemically prepared [11]1− with BNAH results in [11]2−. The same protocols were followed by diphenylamine (DPA) and triphenylamine (TPA), however,

[11]2− was not observed after 7h irradiation (Figure 5.25). The photolysis of 11 with triethylamine (TEA) and triethanolamine (TEOA) generate [11]2− in extended irradiation time which indicates a very mild driving force for the second electron transfer event

(Figure 5.26). However, due to the irreversible nature of the one electron oxidation of the

+ electron donors (ED /ED), the anodic peak potential (EOX) can only be an estimated of the thermodynamics of the electron donating process. Therefore, various values of EOX are reported in the literature.51 Linear sweep voltammograms (Figure 5.27) were obtained and onset oxidation potential (Eonset vs Ag/AgCl) for the oxidation process were determined for BNAH, TEA and TEOA, DPA and TPA (Table 5.4). The onset potentials of DPA and TPA are ~0.2 V more positive than those of TEA and TEOA. Given the slow reaction observed with TEA and TEOA, it is likely that the driving force is minimal for the oxidation of these SDs by the excited state of [11]1–. These findings lead to the

1− estimation of Ered*(11 ) near the onset potentials of TEA and TEOA, ~+0.66 V vs

Ag/AgCl.

1− From the two methods, similar values of Ered*(11 ), +0.66 V vs Ag/AgCl was obtained. And this milder excited state reduction potential comparing to the ground state

111

11, 1.09 V, as well as the shorter triplet state lifetime, agree with the slow conversion from the 111− to 112− under catalytic condition when BNAH is used as an ED.

1− Figure 5.25 Photolysis of [11] with 30 mM A. DPA and B. TPA in CH3CN (irr = 670 nm) at various irradiation times.

112

Figure 5.26 UV-Vis monitored photolysis of [11] in the presence of A. TEA (30 mM) and B. TEOA (30 mM) in CH3CN at various irradiation times.

113

Figure 5.27 LSV of electron donor 0.1 M TBAPF6 in CH3CN vs Ag/AgCl.

Table 5.4 Electron donor, peak anodic potential (Epa) and 5% maximum current onset oxidation (Eonset) vs. Ag/AgCl in 0.1 M TBAPF6

Sacrificial +/0 b,c +/0 b,d e Donor (SD)a Ep(SD ) / V Eonset(SD ) / V Formation of [SD]; tirr [1]2–? BNAH +0.79 +0.47 YES, fast 30 mM; 1 h TEA +0.91 +0.66 YES, slow 30 mM; 8 h TEOA +0.95 +0.66 YES, slow 30 mM; 8 h DPA +1.00 +0.87 NO 30 mM; 7 h TPA +1.03 +0.88 NO 30 mM; 7 h aBNAH = ; TEA = triethylamine; TEOA = triethanolamine; DPA = diphenylamine; TPA = b triphenylamine. From linear sweep voltammograms in CH3CN, 0.1 M tBuNPF6, vs Ag/AgCl. cIrreversible peak anodic potential. dOnset oxidation determined from 5% of the maximum e – current. Irradiation of 1 or [1] with 670 nm in CH3CN.

114

5.3 Conclusions

A new Rh2(II,II) scaffold for the self-sensitized photocatalytic production of dihydrogen from protic solutions is presented. Importantly, the catalyst operates with unprecedented low energy light that allows it to better harness the solar output and its efficiency far surpasses those previously reported. This photocatalysis with red/NIR photons is facilitated through two stepwise excited state reduction reactions of 11 from an electron donor, resulting in the formation of the two-electron reduced catalytically reactive species, [11]2−. The reaction of the latter with two equivalents of protons in acidic solutions results in the production of H2 while regenerating the starting material,

11. The synthesis of 11 is facile and the complex is stable under ambient conditions.

Notably, the catalysis is triggered at all wavelengths absorbed by both the ground and one-electron reduced species, both of which have absorption cross sections that span the visible spectral range into the near infrared.

Moreover, the double reduction of Rh2(II,II) to begin catalysis in the present work is reminiscent of early work on Rh2(I,I) dimers by the Gray group,52,53where the thermal reaction of Rh2(I,I) diisocyanopropanes with hydrochloric acid results in half an equivalent of H2 and a tetranuclear [Rh2(diisocyanopropane)4Cl]2 complex. The latter evolves another half an equivalent of H2 upon irradiation of its metal-centered transitions.

In contrast, the current catalyst stores two electron equivalents via versatile benzo[c]cinnoline ligands that destabilize metal-centered transitions, allowing catalysis to operate through two excited state redox events to create the catalytically active species,

[11]2−. Finally, the evidence that complex 11, when reduced, is able to undergo further 115 photoredox chemistry to generate a doubly reduced species is remarkable and highlights the ability of the bimetallic complex to store two charge equivalents to effect a multi- redox transformation. Prior reports of the light-activated storage of multiple electron equivalents have not been demonstrated without the use of photosensitizers, high photon flux, or covalently tethered electron acceptors.23,25,31,54,55 The present findings open scientific avenues for the design of molecular systems that utilize low energy light via two unprecedented sequential reduction steps following light absorption to access the photocatalytic generation of clean fuels. In particular, the ability of the reduced catalyst to undergo a second photoinduced event to generate the catalytically active doubly-reduced species reported herein is in stark contrast to the manner in which the reducing equivalents are generated and stored in photosynthesis. The present work highlights that artificial photosynthesis may be realized in ways that do not mimic the photosensitizer- catalytic center found in Nature, but instead by creating single molecules that can undergo multi-redox reactions to form catalytically active species.

Bibliography

1. J. Barber and D. Tran Phong, J. R. Soc. Interface, 2013, 10, 20120984.

2. T. Mirkovic, E. E. Ostroumov, J. M. Anna, R. van Grondelle, Govindjee and G. D. Scholes, Chem. Rev., 2017, 117, 249-293.

3. D. Shoup, G. Lipari and A. Szabo, Biophys. J., 1981, 36, 697-714.

4. W. M. Singh, T. Baine, S. Kudo, S. Tian, X. A. N. Ma, H. Zhou, N. J. DeYonker, T. C. Pham, J. C. Bollinger, D. L. Baker, B. Yan, C. E. Webster and X. Zhao, 2012, 51, 5941-5944.

116

5. E. Deponti and M. Natali, Dalton Trans., 2016, 45, 9136-9147.

6. J. I. Goldsmith, W. R. Hudson, M. S. Lowry, T. H. Anderson and S. Bernhard, J. Am. Chem. Soc., 2005, 127, 7502-7510.

7. P. Zhang, M. Wang, Y. Na, X. Li, Y. Jiang and L. Sun, Dalton Trans., 2010, 39, 1204-1206.

8. Y. Na, M. Wang, J. Pan, P. Zhang, B. Åkermark and L. Sun, Inorg. Chem., 2008, 47, 2805-2810.

9. F. Gärtner, B. Sundararaju, A.-E. Surkus, A. Boddien, B. Loges, H. Junge, P. H. Dixneuf and M. Beller, 2009, 48, 9962-9965.

10. M. P. McLaughlin, T. M. McCormick, R. Eisenberg and P. L. Holland, Chem. Commun., 2011, 47, 7989-7991.

11. P. Chábera, K. S. Kjaer, O. Prakash, A. Honarfar, Y. Liu, L. A. Fredin, T. C. B. Harlang, S. Lidin, J. Uhlig, V. Sundström, R. Lomoth, P. Persson and K. Wärnmark, J. Phys. Chem. Lett., 2018, 9, 459-463.

12. L. Liu, T. Duchanois, T. Etienne, A. Monari, M. Beley, X. Assfeld, S. Haacke and P. C. Gros, Phys. Chem. Chem. Phys., 2016, 18, 12550-12556.

13. Y. Liu, P. Persson, V. Sundström and K. Wärnmark, Acc. Chem. Res., 2016, 49, 1477-1485.

14. S. M. Fatur, S. G. Shepard, R. F. Higgins, M. P. Shores and N. H. Damrauer, J. Am. Chem. Soc., 2017, 139, 4493-4505.

15. L. A. Büldt, X. Guo, R. Vogel, A. Prescimone and O. S. Wenger, J. Am. Chem. Soc., 2017, 139, 985-992.

16. Y. Zhang, J. L. Petersen and C. Milsmann, J. Am. Chem. Soc., 2016, 138, 13115- 13118.

17. Y. Zhang, T. S. Lee, J. L. Petersen and C. Milsmann, J. Am. Chem. Soc., 2018, 140, 5934-5947.

18. S. Saeedi, C. Xue, B. J. McCullough, S. E. Roe, B. J. Neyhouse and T. A. White, ACS Applied Energy Materials, 2019, 2, 131-143.

19. Y. Zhang, M. Schulz, M. Wächtler, M. Karnahl and B. Dietzek, Coord. Chem. Rev., 2018, 356, 127-146. 117

20. M. Sandroni, Y. Pellegrin and F. Odobel, C. R. Chim., 2016, 19, 79-93.

21. M. W. Mara, K. A. Fransted and L. X. Chen, Coord. Chem. Rev., 2015, 282-283, 2-18.

22. H. J. Sayre, A. Millet, K. R. Dunbar and C. Turro, Chem. Commun., 2018, 54, 8332-8334.

23. H. Ozawa, M. Kobayashi, B. Balan, S. Masaoka and K. Sakai, 2010, 5, 1860- 1869.

24. K. Peuntinger, T. D. Pilz, R. Staehle, M. Schaub, S. Kaufhold, L. Petermann, M. Wunderlin, H. Görls, F. W. Heinemann, J. Li, T. Drewello, J. G. Vos, D. M. Guldi and S. Rau, Dalton Trans., 2014, 43, 13683-13695.

25. S. Jasimuddin, T. Yamada, K. Fukuju, J. Otsuki and K. Sakai, Chem. Commun., 2010, 46, 8466-8468.

26. P. Zhang, M. Wang, C. Li, X. Li, J. Dong and L. Sun, Chem. Commun., 2010, 46, 8806-8808.

27. T. A. White, K. Rangan and K. J. Brewer, J. Photochem. Photobio. A, 2010, 209, 203-209.

28. T. A. White, S. L. H. Higgins, S. M. Arachchige and K. J. Brewer, 2011, 50, 12209-12213.

29. T. A. White, B. N. Whitaker and K. J. Brewer, J. Am. Chem. Soc., 2011, 133, 15332-15334.

30. J. D. Knoll, S. M. Arachchige and K. J. Brewer, 2011, 4, 252-261.

31. K. Kitamoto and K. Sakai, Chem. Commun., 2016, 52, 1385-1388.

32. J. Hawecker, J.-M. Lehn and R. Ziessel, J. Chem. Soc., Chem. Commun., 1983, DOI: 10.1039/C39830000536, 536-538.

33. S. Sato, T. Morikawa, T. Kajino and O. Ishitani, 2013, 52, 988-992.

34. K. Garg, Y. Matsubara, M. Z. Ertem, A. Lewandowska-Andralojc, S. Sato, D. J. Szalda, J. T. Muckerman and E. Fujita, 2015, 54, 14128-14132.

118

35. S. K. Lee, M. Kondo, M. Okamura, T. Enomoto, G. Nakamura and S. Masaoka, J. Am. Chem. Soc., 2018, 140, 16899-16903.

36. W. Wang, T. B. Rauchfuss, L. Bertini and G. Zampella, J. Am. Chem. Soc., 2012, 134, 4525-4528.

37. A. J. Huckaba, H. Shirley, R. W. Lamb, S. Guertin, S. Autry, H. Cheema, K. Talukdar, T. Jones, J. W. Jurss, A. Dass, N. I. Hammer, R. H. Schmehl, C. E. Webster and J. H. Delcamp, ACS Catalysis, 2018, 8, 4838-4847.

38. H. B. Gray and A. W. Maverick, Science, 1981, 214, 1201.

39. A. F. Heyduk and D. G. Nocera, Science, 2001, 293, 1639.

40. N. Elgrishi, T. S. Teets, M. B. Chambers and D. G. Nocera, Chem. Commun., 2012, 48, 9474-9476.

41. D. C. Powers, S. J. Hwang, S.-L. Zheng and D. G. Nocera, Inorg. Chem., 2014, 53, 9122-9128.

42. T. A. White, S. E. Witt, Z. Li, K. R. Dunbar and C. Turro, Inorg. Chem., 2015, 54, 10042-10048.

43. T. J. Whittemore, A. Millet, H. J. Sayre, C. Xue, B. S. Dolinar, E. G. White, K. R. Dunbar and C. Turro, J. Am. Chem. Soc., 2018, 140, 5161-5170.

44. T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

45. L. S. Forster, Coord. Chem. Rev., 2006, 250, 2023-2033.

46. A. C. Bhasikuttan, M. Suzuki, S. Nakashima and T. Okada, J. Am. Chem. Soc., 2002, 124, 8398-8405.

47. B. G. Alberding, M. H. Chisholm and T. L. Gustafson, Inorg. Chem., 2012, 51, 491-498.

48. V. M. Miskowski, S. F. Rice, H. B. Gray and S. J. Milder, J. Phys. Chem., 1993, 97, 4277-4283.

49. L. S. Fox, J. L. Marshall, H. B. Gray and J. R. Winkler, J. Am. Chem. Soc., 1987, 109, 6901-6902.

50. S. L. Murov, Handbook of Photochemistry, Marcel Dekker, New York, 1973. 119

51. Y. Pellegrin and F. Odobel, C. R. Chim., 2017, 20, 283-295.

52. K. R. Mann, N. S. Lewis, V. M. Miskowski, D. K. Erwin, G. S. Hammond and H. B. Gray, J. Am. Chem. Soc., 1977, 99, 5525-5526.

53. J. Nomrowski and O. S. Wenger, J. Am. Chem. Soc., 2018, 140, 5343-5346.

54. I. S. Sigal, K. R. Mann and H. B. Gray, J. Am. Chem. Soc., 1980, 102, 7252-7256.

55. M. P. Neil, M. P. Niemczyk, W. A. Svec, D. Gosztola, G. L. Gaines and M. R. Wasielewski, Science, 1992, 257, 63.

120

Chapter 6. Concluding Remarks and Future Work

The excited state processes of transition metal complexes have been widely explored for solar energy conversion including solar cells, photoelectrosynthesis cells and artificial photosynthesis.1,2 Due to the absorbed photon energy, the excited states are better reductants and/or oxidants compared to their ground state counterparts. In chapter

3, for the class of dirhodium complexes, cis-[Rh2(μ-DTolF)2(L)2][BF4]2 (L = np (1, 1,8- naphthyridine), npCOOH (2, 1,8-naphthyridine-2-carboxylic acid), phen (3, 1,10- phenanthroline), dpq (4, dipyrido[3,2-f:2’,3’-h]quinoxaline), dppz (5, dipyrido[3,2- a:2’,3’-c]phenazine) and dppn (6, benzo[i]dipyrido[3,2-a:2’,3’-h]quinoxaline); DTolF =

N,N’-di-p-tolylformamidinate), the excited state redox potentials were calculated from

T 3,4 T the zero-zero energy (E00 ) values estimated from this work. E00 is the energy difference between the zero vibrational levels of the ground state and the lowest lying

T triplet excited state. The E00 energies for the non-emissive triplet states of the dirhodium complexes cannot be readily obtained from luminescence and were determined by energy transfer quenching experiments with a series of organic sensitizers (energy transfer doners) with known 3* excited state energy.

The excited state redox potential estimation in chapter 3 reveals that the highly reductive excited states of these complexes, which make them promising dyes for n-type dye sensitized solar cell (DSSC) applications. In chapter 4, two new Rh2(II,II) dyes with 121 methyl ester anchoring groups, [Rh2(DTolF)2(menp)2][BF4]2 (7, menp = 4- carbomethoxy-1,8-naphthyridine) and [Rh2(DTolF)2(dmeb)2][BF4]2 (8, dmeb = 4,4ʹ- carbomethoxy-2,2ʹ-bipyridine) were synthesized and their photophysical properties were characterized and compared. Femtosecond transient absorption (fsTA) and time-resolved infrared (fsTRIR) spectroscopies reveal that the initially populated 1ML-LCT excited states of 7 and 8 decay to the corresponding 3ML-LCT excited states with time constants of 4 ps and 2.5 ps, respectively. The 3ML-LCT excited states of 7 and 8 repopulate the ground state with lifetimes of 460 ps and 56 ps, respectively. The shorter 1ML-LCT and

3ML-LCT lifetimes of 8 as compared to those of 7 are attributed to the longer Rh-Rh bond in the former, which provides a fast deactivation pathway through a metal-centered state that involves population of the Rh2(*) molecular orbital. Photoinitiated electron injection into the semiconductor TiO2 upon low energy light irradiation was achieved through the excitation of Rh2(II,II) dyes associate to TiO2 nanoparticles through the methyl-ester substituent (7@TiO2 and 8@TiO2) and ultrafast electron injection was observed by fsTRIR with low energy excitation, 600 nm for 7 and 520 nm for 8.

In chapter 5, synthetic modification of the dirhodium paddlewheel complex with a shorter bridging bncn (bncn = benzo[c]cinnoline) ligand, complex 11,

2+ Rh2(form)2(bncn)2 (form= formamidinate), was synthesized and characterized. The crystal structure reveals that 11 has a shorter Rh-Rh bond length as compared to previously mentioned dirhodium complexes, which results in a higher lying Rh2(*) molecular orbital that is responsible for metal-center (MC) nonradiative deactivation pathway. The less accessible MC state leads to a longer 3ML-LCT excited state lifetime

122 in the nanosecond timescale, T~19 ns without axial blocking geometry, like 2. The open axial coordination site with the rich electrochemistry make it electro and photocatalytically active in the presence of acid. Bulk electrolysis with trifluoromethanesulfonic acid shows H2 production with 98% Faradaic efficiency. The highly oxidative and nanosecond long triplet excited state, together with the catalytic active metal centers make 11 promising single-molecule/chromophore photocatalyst.

Irradiation at 670 nm in acidic solutions with a sacrificial electron donor, 1-benzyl-1,4- dihydronicotinamide (BNAH), results in the catalytic production of hydrogen exceeding

170 turnovers in 24 hours with an initial rate of 28 turnovers/hour. Mechanistic studies reveal that catalysis proceeds through two stepwise excited state redox events, a feature previously unknown in homogeneous photocatalysis, which permits the storage of multiple redox equivalents on a dirhodium catalyst using low energy light with high efficiency. Lastly, the important intermediate, one-electron reduced complex [11]1−,

T exhibits triplet lifetime  ~ 0.5 ns and ERED* ~ 0.66 eV, which makes the second electron transfer event favorable with BNAH as an electron donor to support the proposed mechanism.

Throughout this work, synthetic modification plays a crucial role in the excited state properties of the dirhodium complex, summarized in Table 6.1. The mode of binding of the diimine ligand, bridging or chelating, as well as the structure of the ligand significantly affect the Rh-Rh bond length. The shorter Rh-Rh bond results in a less accessible MC excited state, thus decreases the nonradiative deactivation pathway, resulting in a longer excited state lifetime, as in the comparison among 8, 1 and 11. Axial

123 coordination also affects the excited state lifetime by blocking solvent-assisted deactivation pathways, as seen in the case of 1 and 2.

Understanding the relationship between structure and function can help us design new molecules with specific properties geared toward various purposes. As a dye used in solar cell, favorable singlet excited state redox potential relative to the semiconductor conduction band or valence band is essential. To be a light absorber in multicomponent photocatalytic systems, besides suitable excited state redox potential, at least nanosecond triplet lifetime is crucial for diffusion controlled bimolecular electron transfer. As a single-molecule/single-chromophore photocatalyst, the open coordination site for catalytic turnover is also desirable in addition to the requirements mentioned for a light absorber.

Table 6.1 Complex 1, 2, 8 and 11, lowest absorption maxima, extinction coefficient, diimine ligands binding mode, Rh-Rh bond length, axial coordination, singlet and triplet excited state lifetimes.

abs / nm (ε / Binding Axial Complex Rh-Rh / Å τS / ps τT / ns 103 M–1cm–1) Mode Coordination? 13 566(3.6)  No 2.4466(7) 14 0.64

23 640(3.5)  Yes 2.4482(4) 7.4 25

8 595(1.7)  No 2.5821(5)a 2.5 0.056

11 642(8.6)  No 2.4049(5) 15 19 aRh-Rh bond length is obtained from structural related complex 9 5 (Rh2(DTolF)2(bpy)2][BF4]2, bpy = 2,2'-bipyridine)

124

Bibliography

1. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737-740.

2. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595-6663.

3. T. J. Whittemore, H. J. Sayre, C. Xue, T. A. White, J. C. Gallucci and C. Turro, J. Am. Chem. Soc., 2017, 139, 14724-14732.

4. T. J. Whittemore, A. Millet, H. J. Sayre, C. Xue, B. S. Dolinar, E. G. White, K. R. Dunbar and C. Turro, J. Am. Chem. Soc., 2018, 140, 5161-5170.

5. H. T. Chifotides, K. V. Catalan and K. R. Dunbar, Inorg. Chem., 2003, 42, 8739- 8747.

125

Bibliography

Chapter 1