University of Southern California Dissertations and Theses

Two-coordinate coinage metal complexes for solar fuels and organic LED chemistry

(USC Thesis Other)

Two-coordinate coinage metal complexes for solar fuels and organic LED chemistry

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2023 Collin Muñiz TWO-COORDINATE COINAGE METAL COMPLEXES FOR SOLAR FUELS AND ORGANIC LED CHEMISTRY by Collin Muñiz A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (CHEMISTRY) August 2023 ii Dedication This dissertation is dedicated to my parents: Sarah Prince, Thiago Valle, and Juan & Marlene Muñiz, and to my wonderful partner, Mahta Barekatain. iii Epigraph “For what shall it profit a man, if he shall gain the whole world, and lose his own soul?” Mark 8:36 iv Acknowledgements My journey to achieve a doctorate degree in Chemistry has been part of a long-term journey to understand nature. First, I would like to acknowledge my parents: Sarah Prince, Thiago Valle, Juan Muñiz, and Marlene Muñiz, who have fully supported me in pursuing my scientific curiosity. Their love and support have allowed me to accomplish all of my achievements thus far. I want to acknowledge Mahta Barekatain for the love and support that she has given me. She has been there to celebrate all of the big and little achievements during my PhD, and she has taught me many valuable life lessons that will serve me beyond my time at USC. I must thank my best friend of 20 years, Nykell Hunter, who has served as a continuous positive influence in my upbringing and adulthood. I want to acknowledge my siblings, Gabriel Muñiz and Giselle Muñiz for their love and support; I could not be more proud of them. My journey in science started in Beaverton High School where I was first inspired by math, physics, and chemistry. Thus, I must acknowledge some important people from the early portion of my journey. Thank you, Jason Sarmiento, for offering precalculus; my first difficult math course where it was made clear that my success will be a product of my personal effort in learning the material. Thank you to Patrick Kreipe for performing the first chemistry demonstration which resulted in a major paradigm shift in my life. My curiosity was awakened after we performed the copper cycle lab in your sophomore chemistry class. Thank you to Sharon Cooper for offering AP Physics, and pulling me aside to tell me that you saw a bright path ahead of me in science. Thank you to Rishika Krishna for your endearing friendship over the last 13 years. After high school, I decided to pursue a degree in Chemistry and Physics at Oregon State University; I would like to acknowledge several important people that I met at OSU. Thank you, professor May Nyman, for giving me my first opportunity to pursue professional science as an v undergraduate researcher. Your mentorship solidified my path in chemistry and physics. Thank you for allowing me to spearhead my own project as a young undergraduate student, which resulted in my first publication and opened many doors for my career. Your continued support is deeply appreciated. Thank you to Dr. Corrine Manogue, Dr. Elizabeth Gire, Dr. Matthew Graham, Dr. David McIntyre, Dr. Ethan Minot, Dr. Oksana Ostroverkhova, Dr. Kenneth Walsch, and Dr. Janet Tate for your instruction in Physics. All of you brought passion to every class that was deeply appreciated. Thank you to Professor Chong Fang for your physical chemistry lab, to Doug Keszler for allowing me to do research at the CSMC, to Paula Kristie and Luanne Johnson for their administrative support. Thank you, Yousif Almulla, for being an inspiration in science and a dear friend. Thank you, Gabriel Nowak, for your friendship, and for your comradery through our physics courses. Thank you to Noah Hoffman and Jacob Van De Lindt for taking on leadership of IE; you both continue to impress me to this day. Thank you, Yitong Qi, for your friendship. Thank you, Zach Konkel, for being a kind soul and for the beautiful memories of jamming, going around the world, and for your continued brothership. Finally, I would like to acknowledge many people who were influential to me at the University of Southern California. Thank you to Professor Richard Brutchey for recruiting me into the program. Thank you to Professor Michael Inkpen for the tea exchange sessions, and for letting me bother Joe and Christina in our first year at USC. Thank you to Professor Smaranda Marinescu for giving me the opportunity to do research in your lab. Thank you, Professor Anupam Madhukar, for enlightening me on what it truly means to understand a scientific concept. Your MASC 501 class changed my world and inspired me to challenge myself on new levels. Thank you, Lucas Jordao and Qi Huang, for your comradery in science; you both inspire me to push myself. Thank vi you, Alexander Schmitt, for always being there for me. Thank you, Joseph Parr, for having my back. Thank you, Cindy Tseng, for your wonderful friendship and support in grad school. Lastly, I want to thank the members of the Mark Thompson lab. Thank you to Professor Mark Thompson who welcomed me into his lab after making a difficult decision to switch groups. My time in Mark’s lab has dramatically improved my capabilities as a professional scientist. Thank you to Peter Djurovich for teaching me many valuable lessons in synthetic and analytical chemistry. I will carry these lessons forward in my career. Thank you to Judy Fong for supporting me (and all students in the group) and being a consistently reliable administrator. I had the pleasure of meeting and befriending many talented scientists at USC. Thank you to Narcisse Ukwitegetse, my inkoramutima, for welcoming me into Mark Thompson’s lab and inspiring me to be a better human and scientist. Thank you to the senior members Savannah Kapper, Daniel Sylvinson, Moon Chul Jung, Abegail Tadle, and Muazzam Idris for helping me get started in Mark’s lab. Thank you, Jie Ma, for showing me how to purify two-coordinate complexes and for your friendship. Thank you, Austin Mencke, for your shenanigans that brought character to the MET group, and your comradery as we cross the finish line together. Thank you to Dr. Anton Razgoniaev for the late nights in lab together, for rides home, and for fruitful scientific collaboration and life advice. Thank you, Dr. Sunil Kandappa, for being a great resource for learning lab techniques and for being a great friend and person. Thank you, Eric McClure, for introducing me to Illimat, and for being a great resource in lab. Thank you to my cohort Mahsa Rezaiyan (original hoodie), Konstantin Mallon, and Jonas Schaab for being a great scientific collaborator and friend. I am especially grateful for Jonas pushing me to jump into a bio- luminescent algae infested ocean at 2am on Catalina Island. Thank you to Allen Shariaty for the late nights wings, spicy food adventures, brilliant chess games, and scientific collaboration. Thank vii you for showing me Thank You Scientist, Nina B. Reyes, and for the late nights talking about mesmerizing science concepts, for the stickers, and for your friendship. Thank you to Gemma Goh for your friendship and the boxes you gave me when I needed to move. Thank you Megan Cassingham for the delicious baked goods and for teaching me about carcinization. Thank you, Marsel Shafikov, for the many fruitful discussions over molecular photophysics, and for inspiring me to be a better scientist. Thank you, Junru Su, for the late-night comradery. Thank you, James Fortwengler, for your friendship and scientific collaboration. Thank you to Frances Yau for exhilarating games of speed and for being such a polite person. Thank you, Mattia & Michela Di Niro, for making me feel like family was just across the street. Thank you, Claire Archer, for being a wonderful undergraduate mentee. I see a very bright future ahead of you and am so proud of all that you have accomplished. Thank you, Kelly Biv, for your contagious, unwavering passion for science, and warm friendship (and the GDR). The East coast isn’t so bad after all. I am very grateful for my time at USC, and I feel lucky to have made so many great friends and colleagues. viii Table of Contents Dedication ....................................................................................................................................... ii Epigraph ......................................................................................................................................... iii Acknowledgements ........................................................................................................................ iv List of Tables ................................................................................................................................ xii List of Figures .............................................................................................................................. xiv Abstract ...................................................................................................................................... xxiv 1 Chapter 1: An Introduction to Organic Light Emitting Diodes and Solar Fuels ..................... 1 1.1 Solar Fuels: An Overview ................................................................................................ 1 1.2 Advantages of Solar Fuels................................................................................................ 2 1.3 Challenges in Solar Fuels ................................................................................................. 3 1.4 Organic LEDs: An Overview ........................................................................................... 4 1.5 Modern Challenges of OLED Technology ...................................................................... 5 1.6 Chapter 1 References ....................................................................................................... 7 2 Chapter 2: Introductory Molecular Photophysics .................................................................... 9 2.1 Absorption and Emission ................................................................................................. 9 2.2 Oscillator strength & Extinction Coefficient ................................................................. 14 2.3 Excited State Lifetime, Quantum Yield, and Triplets .................................................... 18 2.4 General Computational Chemistry ................................................................................. 31 2.5 Geometry Optimization .................................................................................................. 33 ix 2.6 Molecular Orbital Energies and Plotting Wavefunctions .............................................. 36 2.7 Calculating Optical Transitions ...................................................................................... 40 2.8 Chapter 2 References ..................................................................................................... 44 3 Chapter 3: π-Extended Ligands in Two-Coordinate Coinage Metal Complexes for Organic LED Applications .................................................................................................... 46 3.1 A Closer Look at the Blue Problem ............................................................................... 47 3.2 Introduction .................................................................................................................... 52 3.3 Synthesis of all Complexes ............................................................................................ 56 3.4 Crystal Structures of the Complexes .............................................................................. 57 3.5 Electrochemical Methods ............................................................................................... 63 3.6 Computational Methods ................................................................................................. 63 3.7 Photophysical Measurements ......................................................................................... 68 3.8 Structural Results ........................................................................................................... 69 3.9 Electrochemical & Absorption Results .......................................................................... 72 3.10 Computational Results ................................................................................................ 75 3.11 Photophysical Properties at Room Temperature ........................................................ 80 3.12 Temperature Dependent Photophysics ....................................................................... 87 3.13 Discussion ................................................................................................................... 90 3.14 Future Work ................................................................................................................ 95 3.15 Synthesis Of All Complexes & Precursors ............................................................... 101 3.16 NMR & Mass Spec of All Reported Complexes in Ch 3 ......................................... 121 x 3.17 Chapter 3 References ................................................................................................ 158 4 Chapter 4: Two-Coordinate Coinage Metal Complexes as Photosensitizers for Solar Fuels .................................................................................................................................... 165 4.1 Solar Photosensitizer Design Principles ....................................................................... 166 4.2 Existing Photosensitizers: Case Studies ....................................................................... 176 4.3 Exploring Two-Coordinate Coinage Metal Complexes as Photosensitizers ............... 178 4.4 Introducing Two-Coordinates ...................................................................................... 178 4.5 Experimental: Synthesis ............................................................................................... 182 4.6 Computational Parameters ........................................................................................... 183 4.7 Electrochemical Methods ............................................................................................. 183 4.8 Photophysical Measurements ....................................................................................... 183 4.9 Stern-Volmer Quenching Experiments ........................................................................ 184 4.10 Photocatalysis Methods ............................................................................................ 185 4.11 Computational Results .............................................................................................. 187 4.12 Photophysical Properties .......................................................................................... 189 4.13 Electrochemical properties ....................................................................................... 195 4.14 Redox Properties of the Excited State ...................................................................... 197 4.15 Generating Solar Fuels: Photoreduction of Water .................................................... 207 4.16 Conclusion ................................................................................................................ 213 4.17 Stern-Volmer Quenching Data ................................................................................. 216 4.18 Synthesis & Characterization ................................................................................... 221 xi 4.19 Future Work for Solar Fuels with cMa Complexes .................................................. 228 4.20 Chapter 4 References ................................................................................................ 234 xii List of Tables Table 1.1 Gravimetric and volumetric energy densities of predominant fuel sources. .................. 3 Table 3.1 Crystallographic parameters for BZAC-Au-Cz, PAC-Cu-BCz, BZAC-Au-bim and CAAC-Au-bim. ............................................................................................................................. 59 Table 3.2: Crystallographic parameters for MAC-Au-bim, PAC-Au-bim and BZI-Au-Mbim. ....................................................................................................................................................... 61 Table 3.3 Selected bond lengths and angles for the(carbene)M(amide) complexes. .................... 72 Table 3.4 Cyclic voltammetry & absorption results: measurements were carried out in DMF solution with 0.1 M NBu4PF6 electrolyte, and the potentials are listed relative to a ferrocene internal reference. The absorption edge is taken as the point where ICT absorbance has dropped to 10% of the peak absorbance for a toluene solution of the cMa. ................................. 73 Table 3.5 Ground and excited state dipole calculations for all cMa complexes ........................... 76 Table 3.6 Computational Results of all cMa complexes. ............................................................. 78 Table 3.7 Photophysical parameters for cMa complexes in toluene (tol) solution and polystyrene (PS) thin film (1% by weight). .................................................................................. 84 Table 3.8 Photophysical parameters for cMa complexes in toluene (tol) solution and polystyrene (PS) thin film (1% by weight) extracted from temperature dependent measurements. ............................................................................................................................... 89 Table 3.9 Hole/electron separation distances for (carbene)Au(amide) complexes. ..................... 94 Table 4.1 Table of electron transfer rates and corresponding lifetimes at various concentrations. ............................................................................................................................ 175 Table 4.2 Photosensitizer figures of merit for reported organometallic complexes. Entries with (--) were not reported in the literature source. .................................................................... 177 Table 4.3 Computational results of each photosensitizer. Here, f corresponds to the oscillator strength of the transition, and h+,e- is the overlap between the hole and electron wavefunctions. ............................................................................................................................ 188 Table 4.4 Photophysical parameters of all reported complexes in toluene, THF, CH2Cl2, DMF, and MeCN. ....................................................................................................................... 194 xiii Table 4.5 Electrochemical potentials of cMa complexes in various solvents. Potentials are in Volts vs. Fc +/0 couple. See the manuscript for the full CV and DPV for all complexes. ...... 196 Table 4.6 Ground and excited state redox potentials for the cMa complexes in THF. .............. 197 Table 4.7 Stern-Volmer quenching rate constants (kq) for the cMa complexes measured in THF. Reduction potentials of the quenching molecules were measured using differential pulse voltammetry and are reported vs. Fc +/0 .............................................................................. 202 Table 4.8 Parameters from Rehm-Weller fit of kq vs. E 0/- of the quenching molecule. ............. 202 Table 4.9 Stern-Volmer results in toluene. ................................................................................. 206 Table 4.10 Stern-Volmer results in MeCN. ................................................................................ 207 Table 4.11 Crystallographic Parameters for 𝐶𝑢 𝐵𝐶 𝑧 𝑀𝐴𝐶 , 𝐴 𝑢 𝐵𝐶𝑧 𝑀𝐴𝐶 and 𝐶𝑢 𝑃 ℎ𝐶𝑧 𝑀𝐴𝐶 .................................. 224 xiv List of Figures Figure 1.1 A simple schematic of generating solar fuels using the photovoltaic (PV) approach (left) and molecular approach (right). ............................................................................. 2 Figure 1.2 The difference in material requirements for LCD/LED and OLED displays. This image was reprinted from https://www.concept-phones.com/news/oled-vs-lcd-whats-the- difference/. 13 .................................................................................................................................... 4 Figure 1.3 Flexible OLED display by Samsung. 16 ......................................................................... 5 Figure 2.1 The general chemistry model of molecular absorption and emission of a photon (left). The expected absorption/emission spectra corresponding to this simple model (right). ....................................................................................................................................................... 10 Figure 2.2 The absorption and emission spectra of anthracene in solution. This figure was modified from a literature source. 1 ............................................................................................... 10 Figure 2.3 A Jablonski representation of transitions between the S0 to the S1 state of a molecule. Note that Jablonski diagrams are much more complex than what is shown here, but more detail will be added later as more concepts are discussed. ............................................ 13 Figure 2.4 A Jablonski diagram showing excitation from S0→S3 followed by internal conversion to the S1, and subsequent emission. ............................................................................ 14 Figure 2.5 A schematic of the interaction between the oscillating electric field of a photon and the electron and ion cores of the atom at times t1 and t2. ....................................................... 15 Figure 2.6 A representation of the difference in overlap of the occupied 1s and unoccupied 2p orbital with and without an induced dipole.............................................................................. 15 Figure 2.7 The absorption spectrum of a tetrahedral zinc compound. This figure was modified from Ukwitegetse N et al. Inorganic Chemistry 2021, 60 (2), 866-871 with permission. 5 ................................................................................................................................... 17 Figure 2.8 A simplified energy diagram and kinetic scheme for a molecule “M” that absorbs light to form M*, and decays back to the ground state through radiative or non-radiative relaxation with rate constants kr and knr respectively. ................................................................... 19 Figure 2.9 A state energy diagram involving S 0, S1, and T1. ISC denotes intersystem crossing. Only radiative decay is explicitly drawn here, but the S1 and T1 can both decay non-radiatively back to S0. ............................................................................................................ 22 xv Figure 2.10 Molecular orbital diagram of formaldehyde. Here the HOMO-1 is a  bonding MO, the HOMO is approximated as a non-bonding px orbital, and the LUMO is a * MO (a). The conversion of the S1 to the T1 via orbital rotation (b). .................................................... 25 Figure 2.11 A spatial diagram of two circular HOMO (HO) and LUMO (LU) wavefunctions with one electron each. Case 1 is where the wavefunctions overlap. Case 2 is where there is no overlap of the wavefunctions. J represents this exchange energy. .......................................... 28 Figure 2.12 A sketch of the energy of the H2 molecule as a function of ion core separation |R1-R2|. .......................................................................................................................................... 33 Figure 2.13 A sketch of an energy diagram of a molecule with many atoms as a function of nuclear coordinates. Several molecular geometries of interest are labeled RA-RE. ...................... 34 Figure 2.14 A geometry optimization of methane using the universal force field method. The top starting geometry was drawn with highly distorted bonds in a square planar conformation. The bottom geometry was drawn with distorted bonds at random angles. The two starting geometries resulted in different optimized geometries, but the square planar conformation is 2.3 eV higher in energy than the tetrahedral geometry. ..................................... 36 Figure 2.15 The frontier orbitals of formaldehyde calculated using the Hartree-Fock method and 6-31G** functional. The MOs are plotted with iso-value = 0.1 Å -3/2 . Hydrogen atoms are silver, carbon is dark grey, and the oxygen atom is red. ......................................................... 37 Figure 2.16 The frontier orbitals of formaldehyde calculated using the Hartree-Fock method and 6-31G** functional plotted side-by-side with iso-value = 0.1 and 0.5. ................................. 39 Figure 2.17 The TDDFT results of a zinc porphyrin complex using 6-31G** basis and the B3LYP exchange interaction. The 8th singlet transition is highlighted here which has an excitation energy of 3.854 eV, an oscillator strength of 0.051, and corresponds to partial HOMO – 3 to LUMO and partial HOMO - 3 to LUMO + 1 character (relative contributions of the orbital pairs are related to the transparency of the purple arrow). Orbital occupancy is denoted by a green dot and the alpha and beta labels denote spin up or spin down. The absorption spectrum is shown with delta function and Lorentzian representation. The width of the Lorentzian peaks were artificially chosen. ......................................................................... 40 Figure 2.18 The S8 excited state represented by MOs (top) and NTOs (bottom). ........................ 43 Figure 3.1 A simplified schematic of electrochemical generation of the excited state in OLED devices. A real device contains a hole and electron transport layer, and a host material which help efficiently transport holes and electrons onto the dopant molecule. .......................... 47 xvi Figure 3.2 Efficient electroluminescence from fast phosphorescence (left) and Thermally Assisted Delayed Fluorescence (right). This simple schematic assumes that the radiative and ISC rates are much greater than any non-radiative decay out of the S1 or T1. Not shown is prompt fluorescence that is competitive with ISC, but represents a minor component of the emitted photons. ............................................................................................................................ 49 Figure 3.3 The kinetic scheme for emission via TADF , where 𝑘𝑟 S1 and 𝑘𝑟 TADF are radiative decay rates of S1 state and TADF process, Keq indicates the equilibrium constant between T1 and S1 states. .............................................................................................................. 52 Figure 3.4 Compounds considered in this work (Ar = 2,6-diisopropylphenyl). 𝑀𝑋𝑃𝐴𝐶 : M = Cu, Ag, Au; X = Cz, BCz and 𝑀𝑋𝐵𝑍𝐴𝐶 : M = Cu, Au; X = Cz, BCz 𝐴𝑢𝑎𝑚𝑐𝑎𝑟𝑏𝑒𝑛𝑒 : carbene = IPr, BZI, PZI, CAAC, MAC, PAC, BZAC; am = Cz, bim, Mbim, Obim ................................ 56 Figure 3.5 Synthetic chart for all materials prepared in this study. M’ = Cu(I), Ag(I), or Au(I). M’’ = Cu(I) or Au(I) .......................................................................................................... 57 Figure 3.6 The S0→S1 transition of 𝐴𝑢𝐶𝑧𝑃𝐴𝐶 involves an electron moving from the HOMO to the LUMO and HOMO to LUMO + 1. The hole and electron NTOs (Ψ h+ and Ψe-) are appropriate visualizations of the hole and electron wavefunctions corresponding to this transition. One can see that the electron NTO is made up of a linear combination of the ΨLUMO and ΨLUMO+1. The isopropyl groups are omitted from the diisopropylphenyl groups for clarity. ...................................................................................................................................... 64 Figure 3.7 The S0→S1 transition of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 involves an electron moving from the HOMO to LUMO and HOMO to LUMO + 1. The hole and electron NTOs (Ψh+ and Ψe-) are appropriate visualizations of the hole and electron wavefunctions corresponding to this transition. One can see that the electron NTO is made up of a linear combination of the ΨLUMO and ΨLUMO+1. The isopropyl groups are omitted from the diisopropylphenyl groups for clarity. ...................................................................................................................................... 65 Figure 3.8 The NTO (left) and center of charge representation (right) for the hole and electron wavefunctions of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 . The isopropyl groups from the diisopropylphenyl substituents have been deleted for clarity. The iso value was set to 0.1 and the phase information omitted for the wavefunction visualization. The red density represents the hole wavefunction and the blue density represents the electron wavefunction respectively. A red dummy atom was placed at the rh+ coordinate and a blue dummy atom was placed at the re- coordinate. ..................................................................................................................................... 68 Figure 3.9 Thermal ellipsoid drawings of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 , 𝐴𝑢𝐵𝐶𝑧𝑃𝐴𝐶 , 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐴𝐶 and 𝐴𝑢𝐵𝑖𝑚𝐶𝐴𝐴𝐶 . ............................................................................................................................... 70 Figure 3.10 Thermal ellipsoid drawings of 𝐴𝑢𝐵𝑖𝑚𝑀𝐴𝐶 , 𝐴𝑢𝐵𝑖𝑚𝑃𝐴𝐶 and 𝐴𝑢𝑀𝐵𝑖𝑚𝐵𝑍𝐼 ........... 70 xvii Figure 3.11 Electrochemical redox potentials and transition energies for the 1 ICT state. The energy of the 1 ICT state (in toluene) was estimated from the onset of the absorption band where the intensity was 0.10 the value at max. ............................................................................. 75 Figure 3.12 Molecular orbitals (MOs) of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 (left) and 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐴𝐶 (right). The HOMO is displayed with red and blue phases and the LUMO is displayed with turquoise and cream phases (isovalue = 0.1). A magnified perspective is presented to highlight contribution of the d orbital to the HOMO and LUMO. The 2,6-isopropyl groups have been removed for clarity. ....................................................................................................................... 76 Figure 3.13 BZAC/PAC extinction (a) and emission (b) in toluene and polystyrene. Inset shows the spectra of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 in toluene and polystyrene (1 wt %), normalized at the Cz absorbance. Extinction spectra in toluene (c) and emission spectra in polystyrene (d) of 𝐴𝑢𝑏𝑖𝑚𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes. ......................................................................................................... 81 Figure 3.14𝐴𝑢𝐶𝑧𝐶𝑎𝑟𝑏𝑒𝑛𝑒 absorption spectra in toluene. ............................................................ 81 Figure 3.15 Absorption and emission spectra for (BZI)Au(amide), amide = bim, Mbim and Obim. Extinction spectra are recorded in toluene solution and emission spectra in 1% doped polystyrene thin films. .................................................................................................................. 83 Figure 3.16 The absorption spectrum of bim precursors in toluene (a). The excitation and emission spectra of bim in MeTHF at 77K(b). ............................................................................. 85 Figure 3.17 Solvent dependent emission of 𝑀𝐴𝑚𝑖𝑑𝑒𝐶𝑎𝑟𝑏𝑒𝑛𝑒 in DCM, MeTHF, Tol, PS film, and MeCyclohexane ............................................................................................................. 86 Figure 3.18 (a) The ICT transition and the nature of the interaction between the N lone pairs and carbene p-orbital of the N-heterocycle carbene (NHC) for the cMa complexes are illustrated. In this representation the plane of the NHC ligand is perpendicular to the page. (b) LUMO orbitals are shown for 𝑀𝑏𝑖𝑚𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes. The dipp groups are only shown for the IPr and CAAC complexes. In the other complexes the dipp groups are not involved in the LUMO. The LUMO energies (in eV) from DFT calculations are given below the acronym of each ligand. .......................................................................................................... 91 Figure 3.19 (a) The centers of negative charge (blue spheres) and positive charge (red spheres) are shown on the molecular frame for the ICT excited states. Ar = 2,6-diisopropylphenyl and Ad = adamantyl. (b) The rate of TADF emission (𝑘𝑟𝑇𝐴𝐷𝐹 ) at room temperature for a doped polystyrene film is plotted as a function of the hole/electron separation distance. ....................................................................................................................... 93 Figure 3.20 Hole-electron separation calculations for a series of substituted BZAC-Au-Cz complexes. The electron is visualized on the chemdraw based on the calculated <re-> xviii coordinates from QChem. The computations were performed at the same level of theory as the preceding hole-electron calculations. ...................................................................................... 96 Figure 3.21 The HOMO of 𝐴𝑢𝐶𝑧𝐵𝑍𝐼 (left) and 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐼 (right) shown with the iso-value set to 0.35. ..................................................................................................................................... 97 Figure 3.22 State analysis of 𝐴𝑢𝐶𝑧𝑃𝑍𝐼 and 𝐴𝑢𝐶𝑧𝐶𝑁 2𝑃𝑍𝐼 . ........................................................ 98 Figure 3.23 PL spectra of 𝐶𝑢𝐶𝑧𝐶𝑁 2𝑃𝐴𝐶 and 𝐴𝑢𝐶𝑧𝑃𝐴𝐶 in toluene. .......................................... 99 Figure 3.24 Synthesis of PZI-HCl. ............................................................................................. 103 Figure 3.25 1 H NMR of PAC OTf in acetone-d6 ....................................................................... 123 Figure 3.26 1 H NMR of BZAC OTf (400 MHz in acetone-d6).................................................. 123 Figure 3.27 1 H NMR of 𝐶𝑢𝐶𝑙𝑃𝐴𝐶 (400 MHz in acetone-d6). .................................................. 124 Figure 3.28 1 H NMR of 𝐴𝑔𝐵𝐹 4𝑃𝐴𝐶 (400 MHz in acetone-d6). .............................................. 124 Figure 3.29 1 H NMR of 𝐴𝑢 𝐶𝑙𝑃𝐴𝐶 (400 MHz in acetone-d6). .................................................. 125 Figure 3.30 1 H NMR of 𝐶𝑢𝐶𝑙𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). ............................................... 125 Figure 3.31 1 H NMR of 𝐴𝑢𝐶𝑙𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). ............................................... 126 Figure 3.32 1 H and 13 C{ 1 H} NMR of 𝐴𝑢𝐶𝑧𝑃𝐴𝐶 (400 MHz in acetone-d6). ............................ 127 Figure 3.33 Mass spectrometry (MALDI) of 𝐴𝑢𝐶𝑧𝑃𝐴𝐶 vs Bruker Peptide Standard ............... 128 Figure 3.34 1 H and 13 C{ 1 H} NMR of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). ......................... 129 Figure 3.35 Mass spectrometry (MALDI) of 𝐴𝑢𝐶𝑧𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard............. 130 Figure 3.36 1 H and 13 C{ 1 H} NMR of 𝐶𝑢𝐵𝐶𝑧𝑃𝐴𝐶 (400 MHz in acetone-d6). ......................... 131 Figure 3.37 Mass spectrometry (MALDI) of 𝐶𝑢𝐵𝐶𝑧𝑃𝐴𝐶 vs Bruker Peptide Standard ............. 132 Figure 3.38 1 H and 13 C{ 1 H} NMR of 𝐴𝑔𝐵𝐶𝑧𝑃𝐴𝐶 (400 MHz in acetone-d6). ......................... 133 xix Figure 3.39 Mass spectrometry (MALDI) of 𝐴𝑔𝐵𝐶𝑧𝑃𝐴𝐶 vs Bruker Peptide Standard ............ 134 Figure 3.40 1 H and 13 C{ 1 H} NMR of 𝐴𝑢𝐵𝐶𝑧𝑃𝐴𝐶 (400 MHz in acetone-d6). ......................... 135 Figure 3.41 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝐶𝑧𝑃𝐴𝐶 vs Bruker Peptide Standard ............ 136 Figure 3.42 1 H and 13 C{ 1 H} NMR of 𝐴𝑢𝐵𝐶𝑧𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). ....................... 137 Figure 3.43 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝐶𝑧𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard .......... 138 Figure 3.44 1 H and 13 C{ 1 H} NMR of 𝐶𝑢𝐵𝐶𝑧𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). ....................... 139 Figure 3.45 Mass spectrometry (MALDI) of 𝐶𝑢𝐵𝐶𝑧𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard .......... 140 Figure 3.46 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝐵𝑖𝑚𝑃𝑍𝐼 in acetone-d6 ........................................... 141 Figure 3.47 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝐵𝑖𝑚𝑃𝐴𝐶 in CDCl3 ................................................. 142 Figure 3.48 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝑖𝑚𝑃𝐴𝐶 vs Bruker Peptide Standard ............ 143 Figure 3.49 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝐵𝑖𝑚𝑀𝐴𝐶 in CDCl3 ................................................ 144 Figure 3.50 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝑖𝑚𝑀𝐴𝐶 vs Bruker Peptide Standard ........... 145 Figure 3.51 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝐵𝑖𝑚𝐶𝐴𝐴𝐶 in CDCl3 ............................................... 146 Figure 3.52 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝑖𝑚𝐶𝐴𝐴𝐶 vs Bruker Peptide Standard .......... 147 Figure 3.53 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐴𝐶 in CDCl3 ............................................... 148 Figure 3.54 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard .......... 148 Figure 3.55 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐼 in CDCl3 .................................................. 149 Figure 3.56 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝑖𝑚𝐵𝑍𝐼 vs Bruker Peptide Standard ............. 150 Figure 3.57 1 H and 13 C{ 1 H} spectra for 𝐴𝑢𝑀𝑒𝐵𝑖𝑚𝐵𝑍𝐼 in CDCl3 ............................................. 151 Figure 3.58 Mass spectrometry (MALDI) of 𝐴𝑢𝑀𝑒𝐵𝑖𝑚𝐵𝑍𝐼 vs Bruker Peptide Standard ........ 152 xx Figure 3.59 1 H and 13 C{1H} spectra for 𝐴𝑢𝑀𝑒𝑂𝐵𝑖𝑚𝐵𝑍𝐼 in CDCl3 ......................................... 153 Figure 3.60 Mass spectrometry (MALDI) of 𝐴𝑢𝑀𝑒𝑂𝐵𝑖𝑚𝐵 𝑍 𝐼 vs Bruker Peptide Standard ..... 153 Figure 3.61 1 H and 13 C{1H} spectra for 𝐴𝑢𝐵𝑖𝑚𝑖𝑃𝑟 in CDCl3 .................................................. 154 Figure 3.62 Mass spectrometry (MALDI) of 𝐴𝑢𝐵𝑖𝑚𝑖𝑃𝑟 vs Bruker Peptide Standard .............. 155 Figure 3.63 1 H and 13 C{1H} spectra for 𝐴𝑢𝐵𝑖𝑚𝑃𝑍𝐼 in acetone-d6 ........................................... 156 Figure 3.64 1 H and 13 C{1H} spectra for 𝐴𝑢𝐶𝑧𝑃𝑍𝐼 in acetone-d6 .............................................. 157 Figure 4.1 The solar emission spectrum. 7 ................................................................................... 167 Figure 4.2 A 0 th order approximation of the energetic driving force of photo-reduction (a) and photo-oxidation (b). HO represents HOMO and LU represents LUMO. Note that the HO and LU of PS* are no longer accurate labels in the excited state, but they are used here for simplicity..................................................................................................................................... 168 Figure 4.3 A Latimer diagram showing the relationship between the ground state and excited state redox potentials. E00 represents the S0→S1 transition energy. ........................................... 170 Figure 4.4 An accurate diagram of energetically favorable (a) and unfavorable (b) photoredox between the excited state photosensitizer (PS*) and the quenching molecule (Q). The axis here represent potential vs. Fc +/0 . ................................................................................. 171 Figure 4.5 A schematic for reversible (a) and irreversible (b) oxidation in cyclic voltammetry. ............................................................................................................................... 173 Figure 4.6 Kinetic scheme of the deactivation of the excited state photosensitizer (PS*) via normal radiative/non-radiative decay (kr + knr) or photoreducing the electrocatalyst (Q) with rate constant keT. ......................................................................................................................... 174 Figure 4.7 Structure of the three photosensitizers investigated in this work: MRCzMAC [M = Cu or Au, R = tert-butyl (B) or phenyl (Ph)] (iPr = isopropyl). MCzMAC were reported previously and were used for comparison in this study. 43 .......................................................... 181 Figure 4.8 Experimental Setup for photocatalysis ...................................................................... 185 Figure 4.9 GC H2 calibration curve. ........................................................................................... 186 xxi Figure 4.10 The HOMO [red/blue] and LUMO [cyan/beige] wavefunctions (left), and hole [red] and electron [blue] NTO wavefunctions (right) of CuBCzMAC displayed with isovalue = 0.1. Hydrogen atoms and the 3,6-diisopropylphenyl groups were omitted for clarity. The HOMO/LUMO densities of CuPhCzMAC and AuBCzMAC are qualitatively the same. ............ 188 Figure 4.11 a) Molar absorptivity and (b) the emission of all compounds in THF. ................... 190 Figure 4.12 Molar absorptivity spectra in THF for all reported photosensitizers. ..................... 191 Figure 4.13 Solvent dependent PL of CuBCzMAC : (a) absorption, (b) emission. ..................... 192 Figure 4.14 Solvent dependent PL of AuBCzMAC : (a) absorption, (b) emission. ...................... 192 Figure 4.15 Solvent dependent PL of CuPhCzMAC : (a) absorption, (b) emission. ................... 192 Figure 4.16 Cyclic voltammetry of all complexes in CH2Cl2 (a) and THF (b). The reduction event was outside of the solvent window for CH2Cl2 therefore, only the oxidative sweeps are shown here. ........................................................................................................................... 197 Figure 4.17 Kinetic scheme for the oxidative quenching of a photosensitizer. .......................... 199 Figure 4.18 Stern-Volmer quenching if of AuBCzMAC with NMeP in THF. ............................ 200 Figure 4.19 Transient absorption of AuBCzMAC and NMeP in THF excited at 355nm. The blue dashed line is the absorption spectrum of AuBCzMAC and the dot-dashed line is the AuBCzMAC cation spectra obtained from bulk elecrolysis (a). The cationic and anion absorption spectra of AuBCzMAC (b). ........................................................................................ 201 Figure 4.20 Transient absorption of AuBCzMAC in THF with ~0.1 mM NMeP. 73 ................... 203 Figure 4.21 Rehm-Weller analysis of the cMa complexes in THF. ........................................... 205 Figure 4.22 Transient absorption of AuBCzMAC with NMeP in toluene excited at 355nm. The absorption spectrum of the ground state complex is corresponds to the dashed blue line, and the navy blue dot-dashed line is the absorption of the cation determined by bulk electrolysis. 73 ............................................................................................................................... 206 Figure 4.23 (a) TON versus irradiation time for AuBCzMAC (160 M), Co(dmgH)2pyCl (290 M), and BIH (200 mM) in 5 mL of the THF/water mixture (5% v/v). (b) Photo-HER in the presence of a mercury droplet for AuBCzMAC and CuPhCzMAC . ................................... 208 xxii Figure 4.24 Irradiation of the sample with (left) and without (right) the photosensitizer as a control experiment. This proves that the photosensitizer is necessary to generate H2. .............. 210 Figure 4.25 Control experiment of photo driven HER with and without water after one hour of irradiation in a 355nm photoreactor chamber......................................................................... 210 Figure 4.26 Photocatalysis without Co(dmgH)2pyCl. Samples were prepared with 160 mM AuBCzMAC (left), and 63 mM CuPhCzMAC (right) in wet THF (5% water by vol), 200mM BIH, and irradiated with a 470nm LED. ..................................................................................... 211 Figure 4.27 Photostability of AuBCzMAC in air free water/THF (5% v/v) irradiated with a 470nm LED at ~486 mW intensity. The absorption spectrum (a) and the concentration of AuBCzMAC calculated using Beer’s law (b). .............................................................................. 212 Figure 4.28 Mercury poisoning tests after ~23 hrs of irradiation without Co(dmgH)2pyCl for 𝐶𝑢𝑃 ℎ𝐶𝑧𝑀𝐴𝐶 (left), and 𝐴𝑢𝐵𝐶𝑧𝑀𝐴𝐶 (right). The concentrations of the photosensitizer these experiments were ~ 300 mM for the copper experiments, and ~150 mM for the gold experiments. ................................................................................................................................ 213 Figure 4.29 Stern-Volmer study of AuBCzMAC quenched by Co(dmgH)2pyCl in THF. The kq is near diffusion limited as predicted by the Rehm-Weller analysis. ..................................... 216 Figure 4.30 Stern-Volmer plots measured in THF. The corresponding photosensitizer and electron accepting quencher are displayed on the plots. ............................................................. 219 Figure 4.31 Stern-Volmer plots measured in Toluene. The corresponding photosensitizer and electron accepting quencher are displayed on the plots ....................................................... 220 Figure 4.32 Stern-Volmer plots measured in MeCN. The corresponding photosensitizer and electron accepting quencher are displayed on the plots. ............................................................. 221 Figure 4.33 Synthetic scheme of the new cMa complexes CuBCzMAC , AuBCzMAC and CuPhCzMAC ................................................................................................................................ 221 Figure 4.34 Single crystal structures of all new cMa complexes. .............................................. 224 Figure 4.35 1 H and 13 C NMR of 𝐶𝑢𝐵𝐶𝑧𝑀𝐴𝐶 in acetone-d6. The peaks at 2.05ppm and 2.8ppm in the 1 H NMR correspond to acetone and water respectively. ..................................... 225 Figure 4.36 1 H and 13 C NMR of 𝐶𝑢𝑃 ℎ𝐶𝑧𝑀𝐴𝐶 in acetone-d6. The peaks at 5.2ppm, 2.8ppm, and 2.05ppm in the 1 H NMR correspond to DCM, water, and acetone-d6 respectively. ................................................................................................................................ 226 xxiii Figure 4.37 1 H and 13 C NMR of AuBCzMAC in acetone-d6. The peaks at 2.05ppm and 2.8ppm in the 1 H NMR correspond to acetone and water respectively. ..................................... 227 Figure 4.38 Two macrocycle targets for cMa complexes with enhanced stability. MCzBZI }2 (left) and MBCzBZI }1 (right). ..................................................................................................... 229 Figure 4.39 Retrosynthesis of MBCzBZI }1. ................................................................................ 231 Figure 4.40 Retrosynthesis of MCzBZI }2. .................................................................................. 232 Figure 4.41Schematic of photocatalysis in aqueous solution with cMa complexes hosted in Savie micelles. cMa* represents the excited cMa complex, ECat represents an electrocatalyst molecule, and BIH is the sacrificial reductant. ........................................................................... 233 xxiv Abstract Generating renewable fuel from sunlight and designing molecules for organic-LED (OLED) displays are two prominent fields of modern research. While these technologies are very different in what they aim to accomplish, they are united by their fundamental starting point: namely, the excited state of a molecule. This dissertation highlights accomplishments in both fields. Chapter 1 provides a top-level overview of solar fuels and OLED technology. The technical details of the molecular photophysics that govern these technologies are presented in Chapter 2. Chapter 3 highlights work that was published in J. Am. Chem. Soc on the discovery of blue emissive molecules that have record fast radiative rates involving triplet states. An explanation of the structural design that enables fast photon emission is given, and future work to further improve the radiative rate is proposed. Chapter 4 highlights a new class of molecular photosensitizers for solar fuels that feature abundant metal complexes that was published in J. Am. Chem. Soc. We demonstrate figures of merit that are comparable to scarce metal complexes such as ruthenium and iridium. The photosensitzers are paired with a cobalt electrocatalyst and sacrificial reductant to convert water to hydrogen upon irradiation with a blue LED. 1 1 Chapter 1: An Introduction to Organic Light Emitting Diodes and Solar Fuels My time at USC was spent investigating two major research topics: solar fuels and organic light emitting diode (OLED) chemistry. While these fields are drastically different in what they aim to accomplish, their chemistry and underlying physics are closely related. Each topic will have its own introductory section to provide the reader with important details for understanding the research accomplishments that will follow. 1.1 Solar Fuels: An Overview The main objective of solar fuels is to use energy emitted by the sun to directly convert abundant molecules into chemical fuel. This is related to photovoltaic (PV) technology (solar panels) except photovoltaics use sunlight to generate electricity that must be immediately stored in batteries. This is distinctly different from solar fuels technology because the generated fuel does not need to be collected immediately, and the fuel can be stored in tanks rather than batteries. One popular example of solar fuels is using sunlight to split water into hydrogen and oxygen (2 H2O → 2 H2 + O2). Hydrogen and oxygen can be readily stored as fuel which can be burned to produce thermal energy and water vapor (with no greenhouse gas emission) or used to power a fuel cell to generate electricity. Solar panels can also be used to perform electrolysis of water to directly generate fuel. In this case, the electricity harvested by the solar panels is directed to electrodes which can split water into hydrogen and oxygen. Solar fuels can also be generated using a molecular approach in which molecules are dissolved in water that become active towards water splitting upon absorption of sunlight (Figure 1.1). This process is known as “molecular photocatalysis” because the molecules involved are not consumed during water splitting, and the process is driven by light absorption. 2 My research focused on the molecular approach; thus, I will dedicate significant discussion to the molecular route in Chapter 4. Figure 1.1 A simple schematic of generating solar fuels using the photovoltaic (PV) approach (left) and molecular approach (right). 1.2 Advantages of Solar Fuels One of the major advantages of storing solar energy in chemical fuels is the energy density associated with liquid fuel. This can be directly compared to using solar panels to store energy in batteries. The volumetric and mass energy densities of gasoline, rechargeable batteries, methanol, and hydrogen are shown in Table 1.1. The gravimetric energy density of hydrogen far exceeds lithium-ion batteries by ~3 orders of magnitude, ~6x higher than methanol, and ~2x higher than gasoline. However, gasoline stands out as the most volume efficient for energy source with an energy density of ~35 MJ/L. The large volumetric density of gasoline (and fossil fuels in general) is a primary reason why it is still the dominant energy source for automotive vehicles and in the energy grid. Renewable fuels such as hydrogen and methanol have a larger volumetric energy density than lithium-ion batteries which makes them more efficient for storage on the basis of 3 weight and volume. Methanol is a much more volume-efficient fuel source than hydrogen because it is a liquid at room temperature. Hydrogen is a gas at room temperature; therefore, it must be stored at significantly high pressures to reach the liquid phase to have a reasonable volumetric energy density. Table 1.1 Gravimetric and volumetric energy densities of predominant fuel sources. 1-5 Gasoline Lithium-Ion Batteries Methanol Hydrogen (700 bar) Gravimetric Energy Density (MJ/kg) 42.9 0.9 21 120 Volumetric Energy Density (MJ/L) 35.3 2.3 17 5.6 While lithium-ion batteries have been commercialized and are currently implemented in renewable energy storage, solar fuels provide significant improvements to energy storage on the basis of weight and volume. Another consideration between methanol in hydrogen is their combustion biproducts. Burning methanol generates CO2, while burning hydrogen only produces water vapor. Thus, burning hydrogen is more sustainable with respect to greenhouse gas emission. 1.3 Challenges in Solar Fuels Molecular solar fuels generation is accomplished by designing two classes of molecules that work in concert to convert an abundant molecule into fuel upon absorption of sunlight. The molecular design considerations and technical details of the science are covered extensively in Chapter 4. The current state-of-the-art photocatalytic molecular systems include scarce metal complexes based on iridium, ruthenium, and platinum. 6-12 These ultra-rare metal complexes are among the most efficient reported molecules for facilitating photocatalysis, but their scarcity 4 restricts their ability to be scaled up for significant contributions to the energy grid. Thus, a predominant challenge of this technology is to design molecules based on abundant metals to drive photocatalysis. Chapter 4 highlights a project where we designed copper complexes that have competitive figures of merit to ruthenium and iridium complexes. These complexes were even successfully used to convert water into hydrogen. 1.4 Organic LEDs: An Overview Organic LEDs (OLEDs) are an emerging display technology that offer better picture quality compared to LED/LCD displays. The improvement in the display is due to the underlying mechanism of how the two types of displays generate light. A traditional LED/LCD display makes use of an LED backlight to generate the image. This light is controlled using several additional layers of material including lenses, prisms, and diffusers to produce an image (Figure 1.2). The LED backlight is always on in an LCD display, and the backlight can scatter off the other optical materials which lowers the quality of the display. For example, areas that should be completely black with high contrast will have some light due to scattered backlight. Figure 1.2 The difference in material requirements for LCD/LED and OLED displays. This image was reprinted from https://www.concept-phones.com/news/oled-vs-lcd-whats-the-difference/. 13 5 However, OLEDs are designed such that each individual pixel is electronically connected to the device. OLED displays are made up of three primary pixel colors (red, green, and blue) which can be used to generate any color in the visible spectrum. Emission from the pixels is activated by passing current to the pixels which contain an electroluminescent organic molecule. This allows individual pixels to be turned on and off, which eliminates the need for extra materials to control the display. The marked control over each pixel enables high contrast images, and “true black” can be achieved by simply not supplying electricity to the pixels. Another added benefit of OLEDs is that they require a minimal amount of material to function; the pixels are only ~ 20-100 m thick, which is about the diameter of a human hair. 14 As a result, the OLED pixels do not have structural rigidity which allows them to be used in flexible displays as shown in Figure 1.3. 15 Figure 1.3 Flexible OLED display by Samsung. 16 1.5 Modern Challenges of OLED Technology While OLED displays offer enhanced picture quality in comparison to LED/LCD displays, they suffer from pixel degradation. Failure of the pixels arises from the mechanism of light 6 generation. The technical details of this mechanism are provided in Chapter 3 (section 3.1). In summary, the organic molecule in the OLED pixel is charged by the display’s power source (battery, outlet, etc.), which effectively stores energy on the molecule. The charged molecule can release the energy by emitting a photon which generates the display. However, there is a period of time between the initial charging of the molecule and the release of the energy as a photon. This time is known as the molecule’s “excited state lifetime”. Two organic molecules in the excited state can combine their energy which generates an ultra-excited state that can destroy the bonds in the molecule, resulting in permanent failure of the pixel. Interestingly, this is a major problem for blue pixels, but modern green and red pixels last for more than 20,000 hours of operation. However, blue pixels fail within 1,000 hours of operation. 17 If the blue pixels fail, then the display is compromised because red and green pixels alone cannot generate all colors on the visible spectrum. The relatively rapid failure of blue OLED pixels is known as the “blue problem”. As aforementioned, the mechanism of pixel failure has to do with excited state molecules combining their energy. The contemporary solution to the blue problem is to make molecules that release their energy as quickly as possible upon charging. This strategy minimizes the amount of time that the emissive molecule spends in the excited state, which prevents multiple molecules in the excited state from combining their energy and decomposing. Chapter 3 details a project where we discovered blue emitting molecules have record fast photon emission capabilities. We synthesized several analogous molecules to understand the structure-function relationship for increasing the rate of photon emission. Finally, we provide a simple theoretical picture for making molecules with fast radiative rates, and future work is proposed. 7 1.6 Chapter 1 References (1) Møller, K. T.; Jensen, T. R.; Akiba, E.; Li, H.-w. Hydrogen - A sustainable energy carrier. Progress in Natural Science: Materials International 2017, 27 (1), 34-40. DOI: https://doi.org/10.1016/j.pnsc.2016.12.014. (2) Ioannis, G.; Dimitrios, M.; Theodoros, G.; Panagiotis, X.; Dimitrios, K.; Zisis, T. Performance and Emission Characteristics of Spark Ignition Engine Fuelled with Ethanol and Methanol Gasoline Blended Fuels. In Alternative Fuel, Maximino, M. Ed.; IntechOpen, 2011; p Ch. 7. (3) Li, S.-h.; Wen, Z.; Hou, J.; Xi, S.; Fang, P.; Guo, X.; Li, Y.; Wang, Z.; Li, S. Effects of Ethanol and Methanol on the Combustion Characteristics of Gasoline with the Revised Variation Disturbance Method. ACS Omega 2022, 7 (21), 17797-17810. DOI: 10.1021/acsomega.2c00991. (4) Yoshida, K.; Tanaka, S.; Tomonari, S.; Satoh, D.; Esashi, M. High-energy density miniature thermoelectric generator using catalytic combustion. Journal of Microelectromechanical Systems 2006, 15 (1), 195-203. DOI: 10.1109/JMEMS.2005.859202. (5) Manthiram, A. An Outlook on Lithium Ion Battery Technology. ACS Central Science 2017, 3 (10), 1063-1069. DOI: 10.1021/acscentsci.7b00288. (6) Curtin, P. N.; Tinker, L. L.; Burgess, C. M.; Cline, E. D.; Bernhard, S. Structure−Activity Correlations Among Iridium(III) Photosensitizers in a Robust Water-Reducing System. Inorganic Chemistry 2009, 48 (22), 10498-10506. DOI: 10.1021/ic9007763. (7) Bevernaegie, R.; Wehlin, S. A. M.; Piechota, E. J.; Abraham, M.; Philouze, C.; Meyer, G. J.; Elias, B.; Troian-Gautier, L. Improved Visible Light Absorption of Potent Iridium(III) Photo- oxidants for Excited-State Electron Transfer Chemistry. Journal of the American Chemical Society 2020, 142 (6), 2732-2737. DOI: 10.1021/jacs.9b12108. (8) Goldsmith, J. I.; Hudson, W. R.; Lowry, M. S.; Anderson, T. H.; Bernhard, S. Discovery and High-Throughput Screening of Heteroleptic Iridium Complexes for Photoinduced Hydrogen Production. Journal of the American Chemical Society 2005, 127 (20), 7502-7510. DOI: 10.1021/ja0427101. (9) Lu, Z.; Gao, Y.; Chen, H.; Liu, Z.; Sun, L. Water oxidation catalyzed by a charge-neutral mononuclear ruthenium(iii) complex. Dalton Transactions 2017, 46 (4), 1304-1310, 10.1039/C6DT04160D. DOI: 10.1039/C6DT04160D. (10) Gholamkhass, B.; Mametsuka, H.; Koike, K.; Tanabe, T.; Furue, M.; Ishitani, O. Architecture of Supramolecular Metal Complexes for Photocatalytic CO2 Reduction: Ruthenium−Rhenium Bi- and Tetranuclear Complexes. Inorganic Chemistry 2005, 44 (7), 2326-2336. DOI: 10.1021/ic048779r. (11) Wang, X.; Goeb, S.; Ji, Z.; Pogulaichenko, N. A.; Castellano, F. N. Homogeneous photocatalytic hydrogen production using π-conjugated platinum (II) arylacetylide sensitizers. Inorganic chemistry 2011, 50 (3), 705-707. 8 (12) Harriman, A.; Porter, G.; Richoux, M.-C. Colloidal platinum catalysts for the reduction of water to hydrogen, photosensitised by reductive quenching of zinc porphyrins. Journal of the Chemical Society, Faraday Transactions 2: Molecular and Chemical Physics 1981, 77 (10), 1939- 1948, 10.1039/F29817701939. DOI: 10.1039/F29817701939. (13) Kelvin. OLED vs. LCD: What’s the Difference? https://www.concept- phones.com/news/oled-vs-lcd-whats-the-difference/. (14) Lee, W.-K.; Chen, Y.-T.; Wen, S.-W.; Liao, P.-H.; Lee, M.-C.; Hsu, T.-S.; Chen, Y.-J.; Su, G.-D.; Lin, H. Y.; Chen, C.-C.; et al. Three-dimensional pixel configurations for optical outcoupling of OLED displays—optical simulation. Journal of the Society for Information Display 2019, 27 (5), 273-284, https://doi.org/10.1002/jsid.769. DOI: https://doi.org/10.1002/jsid.769 (acccessed 2023/04/05). (15) Sugimoto, A.; Ochi, H.; Fujimura, S.; Yoshida, A.; Miyadera, T.; Tsuchida, M. Flexible OLED displays using plastic substrates. IEEE Journal of Selected Topics in Quantum Electronics 2004, 10 (1), 107-114. DOI: 10.1109/JSTQE.2004.824112. (16) Lee, J.-c.; Lee, H.-y. Flexible OLED Q3 sale 1.5 times flat OLED, Samsung Elec supplying 96.5%: IHS. https://pulsenews.co.kr/view.php?year=2017&no=852666. (17) Geffroy, B.; le Roy, P.; Prat, C. Organic light-emitting diode (OLED) technology: materials, devices and display technologies. Polymer International 2006, 55 (6), 572-582, https://doi.org/10.1002/pi.1974. DOI: https://doi.org/10.1002/pi.1974 (acccessed 2023/04/05). 9 2 Chapter 2: Introductory Molecular Photophysics Both solar fuels and OLEDs are technologies that operate based on the excited state of a material. Understanding the scientific principles that govern the quantitative behavior of the excited state is crucial for successfully design materials to push the boundaries of these fields. The rest of the chapter is dedicated to explaining concepts of photophysics that are used abundantly in my graduate research. Understanding these concepts are necessary to understanding the research accomplishments demonstrated in this thesis. 2.1 Absorption and Emission A crude model for the absorption and emission properties of a molecule are first demonstrated in general chemistry as electronic transitions between the molecular orbitals (MOs). For example, the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) as shown in Figure 2.1. In this model, a photon with equal energy to the HOMO/LUMO energy gap can be absorbed by the molecule. The mechanism of absorption involves promoting the electron from the HOMO to the LUMO. A notable shortcoming of this simple model is that the energies of absorption and emission are identical, and the energies of the molecular orbitals remain constant between the ground and excited state. If this were a complete picture, the absorption and emission spectra would be superimposed delta functions at various energies corresponding to differences between pairs of occupied and unoccupied molecular orbitals. 10 Figure 2.1 The general chemistry model of molecular absorption and emission of a photon (left). The expected absorption/emission spectra corresponding to this simple model (right). The general chemistry model is a nice introduction to the concept of absorption and emission of a molecule, but it fails to explain many important phenomena observed in nature. Figure 2.2 shows the absorption and emission spectra of a real molecule using anthracene as an example. Figure 2.2 The absorption and emission spectra of anthracene in solution. This figure was modified from a literature source. 1 There are several notable differences here: 11 1) The absorption and emission of the molecule are not delta functions, rather broadened bands of intensity. 2) The absorption and emission do not occur at the same energies. The emission occurs at lower energy than the absorption. 3) The absorption and emission mirror each other These are important features that can be accounted for by relaxing the assumptions made in the simple general chemistry model. First, it is important to understand that the MO energies in Figure 2.1 are governed by the molecular geometry. The MO energies are given by Eq. 2.1 𝐻 ̂ 𝛹 𝑗 ⃑⃑⃑ (𝑟 i ⃑ ⃑ ) = 𝐸 𝛹 𝑗 ⃑⃑⃑ (𝑟 i ⃑ ⃑ ) Eq.2.1 where Ĥ is the Hamiltonian which includes kinetic energy and Coulombic terms for the ion cores and electrons, Ψ j ⃑⃑⃑⃑ are the molecular orbitals, and r i ⃑ ⃑ represent the electron coordinates. 2 The Coulombic terms in the Hamiltonian are purely a function of the relative position of the ion cores and electrons (the molecular structure/conformation). The general chemistry model in Figure 2.1 assumes that all molecules in the sample exist in the same conformation. In a real sample, such as anthracene in Figure 2.2, the molecules exist in an ensemble of different conformations, which will lead to a distribution of energies centered around the MO energies corresponding to the lowest energy molecular conformation. This distribution of MO energies will lead to broadening of the observed absorption and emission spectra. The concept of MO energies being governed by the molecular geometry in Eq. 2.1 is also relevant to the second observation. The process of photon absorption results in a redistribution of electron density as an electron is promoted from an occupied MO to an unoccupied MO. The ion cores rearrange in order to respond to this change in electron density which will change the molecular geometry by definition. The change in molecular 12 geometry results in a change of molecular orbital energies; thus, one cannot expect the absorption and emission energy to be the same. However, this does not rigorously mean that the emission energy is lower than the absorption energy as depicted in Figure 2.2, nor does it explain why the absorption and emission spectra are mirror images. To further understand these details, we must move to a more complete model: The Jablonksi Model. Here we will start with a simple Jablonski diagram and add complexity to the model as we explore deeper concepts of photophysics. A Jablonski diagram considers the entire state of the molecule, rather than solely considering the molecular orbital energies. This is because absorption an emission are transitions between molecular states, not just electronic states. When the spins of the electrons in the molecule are paired, the net spin angular momentum is zero and the state is known as a “singlet” denoted by “Sn”. The name singlet comes from the fact that there is only a single measurable spin angular momentum for the molecule (namely 0). The ground state singlet (S0) consists of paired electrons up to the ground state HOMO, with no electrons in higher lying orbitals. The lowest energy excited singlet is known as the S1. In a simple picture, the S1 consists of one electron in the HOMO and one in the LUMO with opposite spins. Each molecular state has unique vibrational states i. Absorption of light can promote a molecule in the lowest vibrational state of the S0 (v1) to any vibrational state in the S1. This results in vibrational structure within the S0→S1 transition manifold. Vibrational structure is clearly observed in the anthracene example in Figure 2.2. However, higher lying vibrational states in the S1 manifold quickly relax to the lowest vibrational state 1 before the energy is re-emitted. Radiative relaxation from the S1 back to the S0 can also occur between the lowest vibrational state of 1, and any of the vibrational states of the S0 as shown in Figure 2.3. 13 Figure 2.3 A Jablonski representation of transitions between the S0 to the S1 state of a molecule. Note that Jablonski diagrams are much more complex than what is shown here, but more detail will be added later as more concepts are discussed. Figure 2.3 demonstrates that the S0→S1 transitions corresponding to absorption are greater than or equal in energy to the S1→S0 transitions corresponding to emission. Moreover, there is an apparent symmetry in the energy of S0→S1 and S1→S0 transitions which explains the mirror image between the absorption and emission, and the point of overlap in Figure 2.2. Photoexcitation into higher lying singlets (S0→Sn for n>1) is also possible, however emission from the Sn→S0 is not observed. This phenomenon is known as Kasha’s rule. 3 This is because the rate of non-radiative decay from the Sn state to a lower lying singlet state is much faster than the rate of Sn→S0. Thus, the Sn state will non-radiatively decay to lower lying singlet states until the S1 state is reached. The process of non-radiative decay between two singlet states Sn→Sm (for n>m) is known as “internal conversion” and is denoted “IC”. This process is shown in the Jablonski diagram below in Figure 2.4. 14 Figure 2.4 A Jablonski diagram showing excitation from S0→S3 followed by internal conversion to the S1, and subsequent emission. An observation that has not been addressed in this section is that the intensity of the different absorption transitions is variable. This section explained why the state transitions occur at different energies, but did not explain what governs the intensity. The quantification of transition intensity is the subject of “oscillator strength”, and the intensity of a given transition per molecule is known as the “extinction coefficient”. 2.2 Oscillator strength & Extinction Coefficient Optically exciting a molecule involves an instantaneous conversion of the ground state to the excited state under the influence of the electric field of the incoming photon. The incoming electric field with wavevector “k ⃑ ” interacts with the positively charged ion cores and the negatively charged electrons which polarizes the molecule. An example of this process is shown for a simple H atom in Figure 2.5, however this interaction also occurs for larger systems such as 15 molecules. The direction of the induced dipole oscillates because the electric field is oscillating in time and space. Figure 2.5 A schematic of the interaction between the oscillating electric field of a photon and the electron and ion cores of the atom at times t1 and t2. The induced oscillations result in two important phenomena. First, the resultant dipole coulombically interacts with the electrons and ion cores. Second, the wavefunction that describes the orbital distorts and can begin to resemble other orbitals. Figure 2.6 demonstrates how the resemblance (or overlap) of the 1s orbital and 2p orbital changes in the presence of interaction with a photon. Figure 2.6 A representation of the difference in overlap of the occupied 1s and unoccupied 2p orbital with and without an induced dipole. 16 The net overlap between the 1s and 2p orbital is rigorously zero in the absence of an electric field. In other words, there is equal overlap of the 1s orbital with the positive (blue) and negative (red) component of the p orbital, which sum to zero. However, the overlap between the 1s and 2p orbital becomes non-zero in the presence of the induced dipole. Note that a dipole is only induced on the 1s orbital and not the 2p orbital because the 2p orbital is unoccupied, so there are no electrons for the electric field to interact with. If the energy of the incoming photon matches the difference in energy between the occupied 1s orbital and the unoccupied 2p orbital, the photon can be absorbed and the electron will instantaneously move from the 1s to the 2p orbital. The efficiency of this transition is quantified by the oscillator strength “ f” in Eq. 2.2 𝑓 ∝ |⟨𝑆 0 |µ ̂|𝑆 1 ⟩| 2 Eq.2.2 where “μ ̂” represents the quantum mechanical dipole operator. 4 Note that Eq. 2.2 is the mathematical description of the concepts in the preceding paragraph. It is an explicit overlap integral of the S0 and S1 wavefunctions under the influence of the induced dipole . The higher the oscillator strength, the more allowed (or efficient) the transition is. Transitions between different molecular orbitals will each have different oscillator strengths, which is why an absorption spectrum has variable intensity for different transitions as shown in Figure 2.7. 17 Figure 2.7 The absorption spectrum of a tetrahedral zinc compound. This figure was modified from Ukwitegetse N et al. Inorganic Chemistry 2021, 60 (2), 866-871 with permission. 5 The example zinc complex in Figure 2.7 illustrates that the S0→S1 absorption is more efficient than the S0→S2 absorption feature. This suggests that the oscillator strength is higher for the S0→S 2 transition. The concept of oscillator strength is closely related to the concept of the extinction coefficient. Beer’s law states that the absorption of a molecule dissolved in a solution of length “l” is given by: 𝐴 𝜆 = 𝘀 𝜆 𝑙𝑐 Eq.2.3 where “A” is the absorption and “” is the extinction coefficient at wavelength “”, “l” is the path length, and “c” is the concentration of the molecule in solution. Beer’s law highlights that one can increase the amount of light absorbed by increasing the concentration of the sample, or by increasing the pathlength. However, the  is a concomitant property of the molecule that governs 18 the relative intensity of the transition (much like the oscillator strength). These terms are often used interchangeably and the explicit relationship between them is provided below in Eq. 2.4. 𝑓 ∝ ∫𝘀 𝑑 ν ̅ Eq.2.4 where “ν ̅” is in units of energy. 4 The oscillator strength of a transition is given by the integral of the extinction spectra as a function of energy. The higher the extinction coefficient, the higher the oscillator strength. Highly allowed transitions typically in the range of  ~ 10 5 M -1 cm -1 , and is as low as 10 -4 M -1 cm -1 for forbidden transitions (f ~ 0). Once the molecule has absorbed light and the excited state is formed, there is some time where the excited state exists before it is able to re-emit the light. Unsurprisingly, the time it takes to relax from S1→S0 is also related to the oscillator strength. The more allowed the transition is, the faster the radiative decay constant “kr” will be. However, spatial overlap of the wavefunctions under the influence of the induced oscillating dipole is not the only thing that governs the allowedness of the transition. One must also consider the effects of spin. 2.3 Excited State Lifetime, Quantum Yield, and Triplets In this section we will consider the amount of time that passes before molecules in their excited state return to the ground state, and the role of electron spin in this phenomenon. The simple kinetic scheme of a molecule absorbing a photon and decaying back to the ground state is shown in Figure 2.8. A photon with energy h is absorbed by molecule “M” which instantaneously generates the excited state “M*”. The excited state molecule then relaxes back to the ground state by emitting a lower energy photon h’ with rate constant “kr” or it relaxes non-radiatively with rate constant knr. There are many possible mechanisms for non-radiative decay; this includes any 19 process, that quenches the excited state, that does not involve emission of a photon. A few examples of non-radiative decay include bond vibration which generates heat, chemical decomposition which generates a new species, charge transfer which generates ions, and energy transfer to another molecule. For simplicity, Figure 2.8 considers loss of the excited state M* by radiative decay, or non-radiative decay via bond vibration. Figure 2.8 A simplified energy diagram and kinetic scheme for a molecule “M” that absorbs light to form M*, and decays back to the ground state through radiative or non-radiative relaxation with rate constants kr and knr respectively. The rate law that governs these processes is shown in Eq. 2.5 𝑑 [𝑀 ∗ ] 𝑑𝑡 = −(𝑘 𝑟 + 𝑘 𝑛𝑟 )[𝑀 ∗ ] Eq.2.5 The integrated form of this expression yields [𝑀 ∗ (𝑡 )] = [𝑀 ∗ ] 0 𝑒 −(𝑘 𝑟 +𝑘 𝑛𝑟 )𝑡 Eq.2.6 Equation 2.6 describes the loss of the excited state as an exponential decay with an effective rate constant kr + knr. The time constant for exponential decay is defined as the inverse of the rate of decay and is shown explicitly in Eq. 2.7 20 𝜏 = 1 𝑘 𝑟 + 𝑘 𝑛𝑟 Eq.2.7 The time constant  is known as the “excited state lifetime” and corresponds to the point at which the population of the excited state has decreased by a factor of 1/e which can be shown by evaluating Eq. 2.6 for t = . The radiative and non-radiative relaxation processes are in competition with one another the moment the excited states are generated. Some of the excited states may decay radiatively and produce a photon, and some may decay non-radiatively. The efficiency of radiative decay is known as the photoluminescence “quantum yield” and is commonly represented with the Greek letter PL and is defined by Eq. 2.8. Φ PL = 𝑘 𝑟 𝑘 𝑟 + 𝑘 𝑛𝑟 Eq.2.8 In the extreme case where the radiative rate is much faster than the non-radiative rate (kr>>knr), Eq. 2.8 simplifies to unity (100% PL efficiency). When the non-radiative rate dominates Eq. 2.8 converges to 0 which means all excited states relax non-radiatively. Any regime in between where kr and knr are comparable will yield PL between 0 and 1. The radiative and non-radiative rate are governed by completely different physics. The radiative decay of a molecule’s excited state is essentially the reverse process of photon absorption. Section 2.5 demonstrated the concept of oscillator strength as the primary metric for allowdness photon absorption. To a first order approximation, the same orbitals that are involved in excitation are also involved in emission (i.e. excitation is moving an electron from HOMO to LUMO, emission is moving an electron from LUMO to HOMO). Thus, transitions with a large oscillator strength will also have a fast radiative rate. This model breaks down for molecules with large 21 structural differences in the ground and excited state, because the HOMO and LUMO orbitals will change as a function of molecular geometry. A transition that may have had large f in the ground state may have a lower f in the excited state if the molecular geometry significantly rearranges. Assuming the ground and excited state geometries are approximately the same, the oscillator strength “ f” is explicitly related to kr through Eq. 2.9 𝑘 𝑟 ≅ 𝑣 ̅ 0 2 𝑓 Eq.2.9 where 𝑣 ̅ 0 represents the energy of the absorption maxima corresponding to the S0→S1 transition. 4 The higher the oscillator strength, the faster the radiative rate will be. Optimizing the oscillator strength of the S0→S1 transition is a valid way to increase the quantum yield of a molecule through molecular design. The factors that govern the non-radiative rate are much more complicated and include a broad scope of mechanisms and considerations. For example, oxygen can quench M* through electron or energy transfer which is a form of non-radiative decay. This rate of decay depends on the concentration of oxygen in the sample. It is common to observe an increase in emission brightness of a sample after sparging it with N2 for several minutes. This is directly due to suppressing the knr of oxygen quenching which results in an increase in PL. One non-bimolecular path for knr is via bond vibration and rotation. One can design a molecular structure to be more rigid which damps knr and results in an increase in PL. An example of this was shown in a highly cited paper by Rasha Hamze published in Science. 6 The net spin angular momentum of the electrons in the molecule also significantly affects the excited state lifetime. Thus far into chapter, we have only considered systems where the electron spins are paired. Consider a molecule in its ground state (S0: two spin paired electrons in 22 the HOMO and a completely vacant LUMO). An incoming photon with energy h is absorbed which drives the S0→S1 transition and promotes an electron from the HOMO to the LUMO. Since the excitation is instantaneous, the spins did not have time to flip and initially remain paired in the S1 excited state. This may decay back to the ground state radiatively or non-radiatively. However, there exists some probability for the spin of one of the electrons to flip as time passes. The process of spin flip is known as “intersystem crossing” (ISC). This generates a spin 1 system which has three measurable spin angular momentum values: -1 ħ, 0 , and 1 ħ and is known as a “triplet” state denoted by “Tn”. This is demonstrated below in Figure 2.9. Figure 2.9 A state energy diagram involving S0, S1, and T1. ISC denotes intersystem crossing. Only radiative decay is explicitly drawn here, but the S 1 and T1 can both decay non-radiatively back to S0. The electron configuration of the T1 state in Figure 2.9 shows one electron in the HOMO and LUMO that are both spin-up. It is clear that there is an issue with radiative decay from the T1 23 because the spin up electron in the LUMO cannot relax into the HOMO; that would result in two spin-up electrons in the HOMO which is Pauli forbidden. The T1 can radiatively decay to S0, but that requires a spontaneous spin flip which is theoretically forbidden. Mathematically, this means that the oscillator strength of T1→S0 is zero. Experimentally, it means that the process does still occur but at a very slow rate. Note that radiative decay out of the S 1 state does not require a spontaneous change of spin which makes it a completely allowed transition. Fully allowed radiative relaxation to the ground state occurs quickly and is known as fluorescence (with rate constant kfl). Since decay out of the S1 is spin-allowed, fluorescence is commonly associated with radiative decay out of the S1 state. Slow radiative relaxation to the ground state is known as phosphorescence and occurs with rate constant kphos. Phosphorescence is commonly associated with emission from the T1 state since it is spin-forbidden. The typical excited state lifetime of a triplet is much greater than that of a singlet (  S1) by Eq. 2.7. Singlet excited states radiatively decay on the order of hundreds of picoseconds to nanoseconds, whereas triplets of purely organic molecules take milliseconds to hours to radiatively decay. It should be noted that spin is a useful way to distinguish fluorescence (S1→S0) and phosphorescence (T1→S0), but spin is not what rigorously distinguishes fluorescence and phosphorescence. The original definitions were purely based on the timescale of observed emission; fast emission corresponds to fluorescence and slow emission corresponds to phosphorescence. 7 Nevertheless, the following work in this dissertation will abide by spin labels; emission out of a singlet is labeled fluorescence, and emission of a triplet is labeled phosphorescence. The radiative rate of a triplet state can be improved if there is some mechanism by which the spin can be perturbed. One way to think of this is finding a way to flip the spin back from spin unpaired to spin paired. Electron spin only interacts with magnetic fields, thus a local magnetic 24 field must be installed in the molecule to interact with electron spin. One way to go about this is to include a heavy atom such as late transition metals (i.e. Ru, Re, Pd, Os, Ir, Pt, Au). These heavy metals have a concomitantly high nuclear charge which accelerates the “orbit” of the atom’s electrons. Charged particles moving in circular motion generate a magnetic field proportional to the square of the speed of the particle. Thus, the electrons orbiting late transition metals generate a magnetic field significant enough to appreciably interact with electron spin. The interaction between this magnetic field and the spin of the electron is known as “spin orbit coupling” denoted by “SOC”. The ability of an atom to observe SOC with its electrons is quantified as the “spin orbit coupling constant”. For a given group of the periodic table, the SOC constant increases quadratically from left to right. 8 Notable accomplishments in organic LEDs were made in Mark Thompson’s research group by taking advantage of the large SOC constant of iridium to speed up the rate of emission out of the triplet. 9-11 Coordinating iridium(III) to organic molecules lead to an increase of kr which resulted in an increase in photoluminescence efficiency by Eq. 2.8. The phosphorescence lifetime (1/kphos) of such compounds is in the 1-10 s regime. 12-14 The SOC factor can be further enhanced by designing complexes with multiple heavy metal centers. This is reasonable since more heavy atoms in a molecule yields more local magnetic field strength and thus enhances SOC. Shafikov et al. demonstrated a dinuclear iridium complex with a phosphorescence lifetime as low as 440ns which is approaching the lifetime of slow emission from a singlet state. 15 Purely organic molecules can also exhibit SOC if their electronic structure contains MOs that feature orthogonal p orbitals on the same atom. El-Sayed developed theory for SOC in purely organic molecules which will be summarized here. 16 El-Sayed’s rule is underpinned by the conservation of momentum rather than spin interaction with the local magnetic field of a heavy 25 atom. Consider the case of formaldehyde; the MO diagram is shown in Figure 2.10a. The frontier orbitals are highlighted in blue and drawn explicitly on the diagram. The S1 state is composed of an n-* transition. There is a lower lying T1 state corresponding to a - transition. Notice that the S1 state is identical to the T1 state upon a 90° rotation of the p orbitals on the oxygen and a spin flip of one of the electrons (Figure 2.10b). Rotating the orbital generates angular momentum “r” where  is the angular frequency of the rotation. However, Newton’s third law still applies; momentum must be conserved. The angular momentum required for orbital rotation can be offset by the momentum required to change spin. Figure 2.10 Molecular orbital diagram of formaldehyde. Here the HOMO-1 is a  bonding MO, the HOMO is approximated as a non-bonding px orbital, and the LUMO is a * MO (a). The conversion of the S1 to the T1 via orbital rotation (b). 26 This is a form of SOC that does not require a heavy atom and is achieved through orbital rotation. The mathematical description of this type of SOC is shown in Eq. 2.10 and gauges the degree of overlap between the S1 and Tn states upon an operator that performs a 90° rotation (Ĥ SOC). The metric for SOC here is “ESOC”: higher values of ESOC correspond to larger degrees of SOC. 𝐸 𝑆𝑂𝐶 = ⟨𝑆 1 |𝐻 ̂ 𝑆𝑂𝐶 |𝑇 𝑛 ⟩ 𝐸𝑞 .2.10 This effect is maximized when the two orbitals in question are on the same atom (px and pz on the oxygen atom in the case of formaldehyde), and when the orbitals are spatially orthogonal. Thus, this effect would not be observed between a - singlet and -* triplet in the same orientation, because a 90° rotation of two p orbitals that were initially in the same orientation yields two orthogonal orbitals with a spatial overlap of zero. In fact, if we look at any two states with the same orbital parentage (i.e. an n- singlet and n-* triplet), the ESOC will always be zero. Consider that the wavefunction describing the state will either be odd or even. However, the Ĥ SOC operator applies a 90° rotation to the function which is analogous to switching the function from odd to even or even to odd. Thus, the ESOC in Eq. 2.10 will always yield an overlap integral of an odd function if the singlet and triplet states have the same orbital parentage, which integrates to zero. As an example, consider |S1> and |T1> to be a wavefunctions described by cos(cS1x) and cos(cT1x) over x = - to  where cS1 and cT1 are constants. These are even functions. However, Ĥ SOC|T1> is a “90° rotation” of cos(cix) which yields sin(cix). The integral for ESOC yields an overlap integral of cos(cS1)sin(cT1x) from - to  which is zero. Additionally, - transitions of different orbital parentage (i.e. and S1 - state from px orbitals and a Tn - state from pz orbitals) also yields ESOC = 0 because it would require two p orbital rotations and that induce two spin flips. In that situation, there is no way to end up with a - triplet; the first spin flip is canceled out by the 27 second spin flip resulting in a singlet configuration. Thus, El-Sayed arrived at the conclusion that only n-* and -* states can effectively exhibit SOC in organic molecules. A general rule of thumb is that a molecule with lone pairs on an atom with double bonds (like carbonyls or imines) will have n-* and -* states and have efficient ISC. Another important detail in Figure 2.9 is that the T1 state is lower in energy than the S1 state. This is another ubiquitous feature of triplets compared to singlets of the same orbital parentage. Thus, not only are triplets slower to emit but they also emit lower energy light than the corresponding singlets. The difference in energy is a consequence of the Pauli Exclusion Principle which states that two fermions (in this case electrons) cannot occupy the same point in space with the same spin. The restriction of spatial overlap between electrons with the same spin means that the electrons in a triplet state will be farther apart from one another on average compared to the corresponding singlet state, which has considerable consequences on the energy of the molecular state since the energy is predominantly governed by Coulombic interactions. Equation 2.11 is the potential experienced by two electrons (with charge “e”) separated by distance |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ |. 𝑉 = 𝑒 |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ | Eq.2.11 The Coulombic potential is inversely proportional to the distance between the two electrons. Thus, the singlet state will experience more electron-electron repulsion than the triplet state since there is no restriction on 𝑟 1 ⃑⃑⃑ = 𝑟 2 ⃑⃑⃑ when the spins are paired. The triplet state is restricted such that 𝑟 1 ⃑⃑⃑ can never be 𝑟 2 ⃑⃑⃑ which means it will experience less electron-electron repulsion on average which results in a lower energy of the state with respect to the singlet. 28 Now that we have established that a Tn state is lower in energy than an Sn state, we move on to quantifying the energy difference. Consider the thought experiment shown below in Figure 2.11 to begin to understand the quantitative picture of the singlet triplet energy gap “EST”. Figure 2.11 A spatial diagram of two circular HOMO (HO) and LUMO (LU) wavefunctions with one electron each. Case 1 is where the wavefunctions overlap. Case 2 is where there is no overlap of the wavefunctions. J represents this exchange energy. Case 1 in Figure 2.11 has significant overlap between the HOMO and LUMO wavefunctions, which means that there are multiple opportunities for the two electrons to occupy the same point in space at the same point in time. This results in an increase of energy via electron-electron repulsion which is more significant for the singlet state than the triplet state following the logic demonstrated above. However, Case 2 in Figure 2.11 shows a scenario in which there is no overlap of the HOMO and LUMO wavefunctions. Thus, there is no probability of the electrons occupying the same point in space at the same point in time regardless of spin. It follows that there is no difference in electron-electron repulsion between the singlet and triplet states in Case 2 which renders them as degenerate states (EST = 0). It is evident that the EST depends on the overlap of the molecular orbitals involved; a larger degree of overlap will result in a larger EST. 29 The quantification of this energy gap is known as the “Exchange Energy” and is denoted here by “J”. As explained in the preceding paragraph, the exchange integral arises from the difference in electron-electron repulsion in the singlet and triplet states. Consider Figure 2.11 which has one electron occupying HO and one occupying LU. Mathematically, we cannot assign a specific electron to the HOMO or LUMO. Thus, we express the excited state in Eq. 2.12 as a superposition of electron 1 (with coordinate r1) and electron 2 (with spatial coordinate r2) occupying the HOMO (|𝛹 𝐻𝑂 ⟩) and LUMO (|𝛹 𝐿𝑈 ⟩). |𝛹 𝑒𝑙𝑒𝑐 ⟩ = (2 − 1 2 )(|𝛹 (𝑟 1 ⃑⃑⃑ ) 𝐻𝑂 ⟩|𝛹 (𝑟 2 ⃑⃑⃑ ) 𝐿𝑈 ⟩ ± |𝛹 (𝑟 2 ⃑⃑⃑ ) 𝐻𝑂 ⟩|𝛹 (𝑟 1 ⃑⃑⃑ ) 𝐿𝑈 ⟩) 𝐸𝑞 . 2.12 Where 2 -1/2 is a normalization constant. This expression ensure that neither of the electrons is formally labeled. In other words, we do not state that “electron 1” is in the HOMO and “electron 2” is in the LUMO. Rather, this expression states there is a 50% change of electron 1 being in the HOMO and electron 2 being in the LUMO, and there is a 50% chance of electron 2 being in the HOMO and electron 1 being in the LUMO. The relative spin of the electrons (paired or unpaired) is determined by the “±” sign between the MO products. The + and – linear combination correspond to the singlet and triplet, respectively. 17 The difference in electron-electron repulsion between the singlet and triplet can be directly evaluated with Eq. 2.12 and the coulombic repulsion between electron 1 and 2 (Eq. 2.13). ⟨𝛹 𝑒𝑙𝑒𝑐 | 1 |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ | |𝛹 𝑒𝑙𝑒𝑐 ⟩ = 𝐾 ± 𝐽 𝐸𝑞 .2.13 where J and K are the exchange and coulombic energy terms, respectively. The singlet state has an electron-electron repulsion energy of K + J, and the triplet state has a repulsion energy of K -J. 30 Thus, the difference in repulsion energy between the two states is 2J. The exchange energy is given in Eq. 2.14. 𝐽 = ⟨𝛹 (𝑟 2 ⃑⃑⃑ ) 𝐻𝑂 |⟨𝛹 (𝑟 1 ⃑⃑⃑ ) 𝐿𝑈 | 1 |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ | |𝛹 (𝑟 1 ⃑⃑⃑ ) 𝐻𝑂 ⟩|𝛹 (𝑟 2 ⃑⃑⃑ ) 𝐿𝑈 ⟩ 𝐸𝑞 . 2.14 and there is no easy way to simplify this integral without making assumptions. The exchange energy “J” appears to be related to the overlap of the HOMO and LUMO wavefunctions. However, the exchange integral is more complicated than simple orbital overlap since the coulomb potential and molecular orbital wavefunctions depend on the variable electron positions (𝑟 i ⃑ ⃑ ⃑ ). However, if we assume that the electron separation (|𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ |) is constant, then we can pull it out of the integral and the math simplifies drastically. In this assumption, the electron separation can be taken as the difference between the center of electron distribution in the two wavefunctions. The coulomb term can then be pulled out of the integral to yield Eq. 2.15 𝐽 ≈ 1 |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ | (⟨𝛹 (𝑟 1 ⃑⃑⃑ ) 𝐿𝑈 |𝛹 (𝑟 1 ⃑⃑⃑ ) 𝐻𝑂 ⟩)(⟨𝛹 (𝑟 2 ⃑⃑⃑ ) 𝐿𝑈 |𝛹 (𝑟 2 ⃑⃑⃑ ) 𝐻𝑂 ⟩) 𝐸𝑞 .2.15 and since there is no difference between the individual overlap integrals ⟨𝛹 (𝑟 1 ⃑⃑⃑⃑ ) 𝐿𝑈 |𝛹 (𝑟 1 ⃑⃑⃑⃑ ) 𝐻𝑂 ⟩ and ⟨𝛹 (𝑟 2 ⃑⃑⃑⃑ ) 𝐿𝑈 |𝛹 (𝑟 2 ⃑⃑⃑⃑ ) 𝐻𝑂 ⟩, Eq. 2.15 further simplifies to Eq. 2.16. 𝐽 ≈ 1 |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ | ⟨𝛹 𝐿𝑈 |𝛹 𝐻𝑂 ⟩ 2 𝐸𝑞 .2.16 and we see that a crude approximation yields an expression that directly depends on the overlap of the HOMO and LUMO wavefunctions, and the coulombic repulsion between the electrons in the HOMO and LUMO, respectively, evaluated at their center of probability distributions. This offers a guideline for molecular design. Certain applications such as organic-LEDs benefit from having emissive molecules with low singlet triplet gaps (low J). In that case, one can either design 31 molecules that have a spatially orthogonal HOMO and LUMO to minimize ⟨𝛹 𝐿𝑈 |𝛹 𝐻𝑂 ⟩, or design molecules with a large separation in the HOMO and LUMO centers of distribution (high |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ |). There are also applications that benefit from large singlet-triplet gaps such as singlet fission. In that case, it is better to design a molecule with a highly localized HOMO and LUMO that have significant wavefunction overlap. A version of the electron separation was directly utilized in the discoveries presented in Ch. 3, and will be covered more extensively there. 2.4 General Computational Chemistry Computational chemistry is extremely useful for quantitatively predicting the properties of a hypothetical molecule before it is ever synthesized in a laboratory. Calculating the mathematical wavefunction of a molecule allows one to predict any measurable property of the molecule by Eq. 2.17 𝑚 𝑖 = ⟨𝛹 𝑖 |𝑀 ̂ |𝛹 𝑖 ⟩ 𝐸𝑞 .2.17 where “mi” is the measurable property (i.e. energy, momentum, position, etc.) corresponding to the molecular state “|𝛹 𝑖 ⟩”, and “𝑀 ̂ ” is the corresponding operator. A specific example of this is Schrödinger’s equation which was shown earlier in Eq. 2.1 where “𝑀 ̂ ” was the Hamiltonian “Ĥ” and “mi” was the energy of the i th state “Ei”. The Hamiltonian is the operator for energy and consists of all the kinetic “KEi” and Coulombic “Vi,j” terms in the molecule as shown in Eq. 2.15 𝐻 ̂ = ∑𝐾𝐸 𝑖 (𝑟 i ⃑ ⃑ ) 𝑖 + ∑𝐾 𝐸 𝑗 (𝑅 j ⃑⃑⃑ ) 𝑖 + 1 2 ∑𝑉 𝑖 ,𝑗 (𝑟 i ⃑ ⃑ ,𝑅 j ⃑⃑⃑ ) 𝑖 ,𝑗 𝐸𝑞 .2.15 where “𝑟 i ⃑ ⃑ ” are the positions of the electrons and “𝑅 j ⃑⃑⃑ ” are the positions of the ion cores. 18 The Coulombic terms include electron-ion core attraction terms, and the Coulombic repulsion terms between the ion cores and the electrons respectively. The solutions to Eq. 2.15 for the Hamiltonian 32 operator are molecular states “|𝛹 𝑖 ⟩”. Since the Hamiltonian is a function of the ion core positions (𝑅 j ⃑⃑⃑ ) and electron positions (𝑟 i ⃑ ⃑ ), the resulting wavefunctions are a function of 𝑟 i ⃑ ⃑ and 𝑅 j ⃑⃑⃑ . This function of multiple variables can be further simplified under the Born-Oppenheimer approximation which recognizes that the ion cores are ~10 3 -10 4 heavier than the electrons. Thus, the motion of the electrons does not significantly impact the position of the ion cores on short timescales. Another effect of this mass difference is that the electrons are moving much faster than the ion-cores; the position of the ion-core is fixed from the perspective of the electron. Thus, the electron wavefunctions only have a parametric dependence on the equilibrium positions of the ion-cores; “𝑅 j ⃑⃑⃑ ” is effectively a constant. Thus, the Born-Oppenheimer approximation allows the wavefunction to be separable into a product of electronic (|𝛹 𝑒𝑙𝑒𝑐 ,𝑖 (𝑟 i ⃑ ⃑ ,{𝑅 j ⃑⃑⃑ })⟩) and ion-core (|𝛺 𝐼𝐶 ,𝑖 (𝑅 j ⃑⃑⃑ )⟩) functions (Eq. 2.18) |𝛹 𝑖 (𝑟 i ⃑ ⃑ ,𝑅 j ⃑⃑⃑ )⟩ = (|𝛹 𝑒𝑙𝑒𝑐 ,𝑖 (𝑟 i ⃑ ⃑ ,{𝑅 j ⃑⃑⃑ })⟩)(|𝛺 𝐼𝐶 ,𝑖 (𝑅 j ⃑⃑⃑ )⟩) 𝐸𝑞 .2.18 and the energy is the sum of the electronic and nuclear components by Eq. 2.17. 𝐸 𝑖 = 𝐸 𝑒𝑙𝑒𝑐 ,𝑖 + 𝐸 𝐼𝐶 ,𝑖 𝐸𝑞 .2.19 The electronic wavefunctions (|𝛹 𝑒𝑙𝑒𝑐 ,𝑖 (𝑟 i ⃑ ⃑ ,{𝑅 j ⃑⃑⃑ })⟩) are the molecular orbitals and they are approximated as a linear combination of atomic orbitals (Eq. 2.20). |𝛹 𝑖 ⟩ = ∑𝑐 𝑖 |𝜑 𝑖 ⟩ 𝑖 𝐸𝑞 .2.20 where “ci” are the scalar coefficients of the atomic wavefunctions that contribute to the molecular orbital. The solutions to the Schrödinger equation are found by optimizing the coordinates and coefficients to yield the lowest energy ground state Ψ0 which determines the optimized molecular 33 structure (𝑅 j ⃑⃑⃑ ). The most common method for solving Schrödinger’s equation is known as “Density Functional Theory”; the mathematical details are beyond the scope of this dissertation but are covered in detail in several other papers. 19 2.5 Geometry Optimization The first step in computational chemistry is to optimize the molecular geometry since the properties and molecular orbitals are dependent on the coordinates of the ion cores. A simple example is shown for the H2 molecule in Figure 2.12. Figure 2.12 A sketch of the energy of the H2 molecule as a function of ion core separation |R1-R2|. Starting from infinite separation of the two hydrogen atoms, the total energy is equal to the sum of the kinetic energies of the individual atoms. The sum of the kinetic energies is the reference energy (the zero point on the y axis). As the ion cores move towards each other, the positively charged ion cores begin to attract the electron on the other atom which induces a dipole. The long- range dipole-dipole interaction stabilizes the system. Once the atoms get close enough, they enter a new regime in which the stabilization from the Coulombic attraction of the ion cores and electrons competes with the destabilization from electron-electron and ion core-ion core repulsion. 34 A global minimum is reached which corresponds to the molecular geometry with the lowest energy. The ion core separation at the local minimum is also the bond length for the H 2 atom. Beyond the minimum at closer ion core separation, the repulsion terms dominate, and the system becomes infinitely unstable. Note that the molecular orbitals and properties would all be different depending on the |R1-R2|. Solving the Schrödinger equation using the minimum energy geometry will yield the molecular orbitals and properties of the molecule in the ground state. The model shown in Figure 2.12 can be extended to many atom molecules. However, the x-axis can no longer be |R1-R2| because now there are more than two atoms. There are not enough axis to separately account for the distance between all the atoms in a molecule with more than two atoms. Instead, the interpretation of the x-axis is changed. Now, each point on the x-axis represents a different arrangement of the ion core positions. We expect similar behavior; there will exist a lowest energy geometry and the energy will diverge to infinity as the ion-cores converge to the same position. However, the existence of multiple atoms can bring about local minima and maxima in the geometry curve. An example for a many-atom molecule is shown below in Figure 2.12. Figure 2.13 A sketch of an energy diagram of a molecule with many atoms as a function of nuclear coordinates. Several molecular geometries of interest are labeled RA-RE. 35 Geometry optimizations are calculated iteratively by taking a starting arrangement of the ion cores and computing the energy corresponding to that molecular geometry. Slight changes in the geometry are made and the energy is recalculated. Changing the molecular geometry corresponds to moving slightly to the left or right from a starting point on the x-axis of Figure 2.13. If the energy is lower than the initial geometry, then the change of molecular coordinates is accepted, and the process is repeated. This will continue until a local energy minimum is reached, at which point any slight change in the molecular geometry will be uphill in energy. This is the only computationally tractable way to perform a geometry minimization; producing a full energy diagram such as the one in Figure 2.13 is too computationally expensive. The drawback of this method is that one does not know if they have converged at the global minimum, or if they are stuck in a local minimum. For example, if the starting geometry is RA or RB in Figure 2.13, then the geometry optimization will converge to the left most local energy minimum. Starting at R C will yield the global minimum geometry. Starting at RD will optimize to a relatively unstable state. Finally, starting at RE will yield yet another final geometry. In practice, the starting geometry is the way the molecule is drawn in the computational software. The best one can do is to perform several different geometry optimizations with different starting geometries and see if they converge to the same geometry. If the calculations do not converge to the same geometry, then one can compare the total energy of each geometry (this value is calculated by the software package) which clarifies which minimum is lower in energy. It is common practice to input the crystal structure molecular geometry as a starting point if that has been experimentally obtained beforehand. The crystal structure geometry will likely not end up being the final optimized geometry because calculations correspond to the molecule in a vacuum, whereas the crystal 36 structure is the minimum energy of the molecular geometry in the lattice environment. A real example of a geometry optimization QChem v2.14.0 is shown below in Figure 2.14 Figure 2.14 A geometry optimization of methane using the universal force field method. The top starting geometry was drawn with highly distorted bonds in a square planar conformation. The bottom geometry was drawn with distorted bonds at random angles. The two starting geometries resulted in different optimized geometries, but the square planar conformation is 2.3 eV higher in energy than the tetrahedral geometry. 2.6 Molecular Orbital Energies and Plotting Wavefunctions Once the geometry is optimized, the Schrödinger equation is solved for that particular molecular geometry which produces the corresponding molecular orbitals (|𝛹 𝑖 ⟩”) and their energies (Eelec,i) from equations 2.17 and 2.18. The MOs can be plotted to provide a visual representation of the linear combination of atomic orbitals; this also allows one to determine the bonding/antibonding character of the MO. An example is shown below for the HOMO and LUMO of the formaldehyde molecule that was analyzed by hand earlier. 37 Figure 2.15 The frontier orbitals of formaldehyde calculated using the Hartree-Fock method and 6-31G** functional. The MOs are plotted with iso-value = 0.1 Å -3/2 . Hydrogen atoms are silver, carbon is dark grey, and the oxygen atom is red. Here we see the HOMO-1 is  bonding between the 2pz orbitals on the carbon and oxygen atoms as predicted by Figure 2.10. The HOMO is a  orbital between the s orbitals on the hydrogens and the oxygen px orbital. The HOMO in Figure 2.10 was approximated as a non-bonding orbital but the results from Figure 2.15 are much more rigorous. The LUMO is a * orbital between the pz atomic orbitals on the carbon and oxygen atoms which is consistent with the approximation in Figure 2.10. The orbital energies are highly relevant for predicting properties such as oxidation (removing an electron from the HOMO) and reduction (putting an electron in the LUMO) potentials. The HOMO/LUMO energy gap also typically trends with the S 1 absorption and emission energy; especially for transitions that are well described by an electron being promoted from the HOMO to the LUMO (Figure 2.3 and Figure 2.9). The visualization of the orbitals is helpful in determining the nature of the MO and which atoms contribute the most to the MO. For example, the LUMO has more density on the carbon atom than the oxygen atom. It is  in nature so one would expect a population of the LUMO through reduction or photoexcitation to weaken the bond order of the formaldehyde double bond, resulting in a bond elongation. 38 The MOs (Ψelec,i(x,y,z)) represent root probability densities of the electron in a given electronic state at a given point in space. MOs are four-dimensional functions because the root probability density varies over the three spatial coordinates x, y, and z. Plotting 4-dimensional functions requires picking an “iso-value” of the function and plotting an iso-surface that satisfy Ψelec,i(x,y,z) = C (where “C” is the iso-value). A mathematical example is the four-dimensional function f(x,y,z) = x 2 + y 2 + z 2 . This function is impossible to plot normally as one cannot draw 4- axis to represent f,x,y, and z. However, picking different iso-values (Ci) of f yields the equation Ci = x 2 + y 2 + z 2 which is the equation of a sphere of radius Ci 1/2 . Picking larger iso-values for this function yields spheres with larger radii. In the case of MOs, Ψelec,i(x,y,z) are built of a linear combination of atomic orbitals which ubiquitously undergo exponential decay. The decaying nature of Ψelec,i(x,y,z) as a function of distance results in smaller surfaces for larger iso-values and larger surfaces for smaller iso-values. Thus, plotting the MOs of a molecule with larger iso-values allows one to visually identify the atomic orbitals that have the highest contribution to the MO. Another way to think about this is that plotting MOs with high iso-values (typically > 0.3) reveals the 1 st order nature of the MO, and low iso-values (< 0.2) shows the higher order perturbations. These concepts are demonstrated in Figure 2.16. 39 Figure 2.16 The frontier orbitals of formaldehyde calculated using the Hartree-Fock method and 6-31G** functional plotted side-by-side with iso-value = 0.1 and 0.5. Plotting the MOs with a higher iso-value evinces that the HO-1  bonding state is weighted towards the oxygen pz orbital. The HOMO appears to be a non-bonding px orbital on the oxygen atom due to poor mixing with the hydrogen 1s orbitals. Both of the above observations are not clear when iso-value = 0.1 is chosen. Revealing that the HOMO is non-bonding in character is relevant to the concepts of El-Sayed’s rule discussed in previous sections; n-* states are required to efficiently couple to -* states to enhance ISC. In larger molecules, identifying the atoms with the largest contribution in the HOMO and LUMO can also be helpful for guiding molecular design. The energy of these states will be most effectively tuned by making structural modifications near the highest contributing atoms. 40 2.7 Calculating Optical Transitions The absorption spectra of a molecule can be calculated using “Time Dependent Density Functional Theory” or “TDDFT”. This technique is extremely useful for predicting optical transition energies and the corresponding oscillator strengths. Thus, TDDFT can be used to simulate the absorption spectrum of a molecule and provide peak assignments in the transitions. Another strength of TDDFT is that it does not make the assumption that only one pair of MOs is involved in photoexcitation (i.e. HOMO to LUMO, or HOMO to LUMO + 1). Multiple MOs can participate in photoexcitation as demonstrated in Figure 2.17. Figure 2.17 The TDDFT results of a zinc porphyrin complex using 6-31G** basis and the B3LYP exchange interaction. The 8th singlet transition is highlighted here which has an excitation energy of 3.854 eV, an oscillator strength of 0.051, and corresponds to partial HOMO – 3 to LUMO and partial HOMO - 3 to LUMO + 1 character (relative contributions of the orbital pairs are related to the transparency of the purple arrow). Orbital occupancy is denoted by a green dot and the alpha and beta labels denote spin up or spin down. The absorption spectrum is shown with delta function and Lorentzian representation. The width of the Lorentzian peaks were artificially chosen. 41 The S0→S8 transition in Figure 2.17 involves a transition from the HOMO - 3 to LUMO, and the HOMO - 3 to LUMO + 1. The contributions of these MO pairs are 40% and 60% respectively which is given in the output file of the calculation. The notion of more than two MOs being involved in a single transition may seem strange at first, but there is no reason why an incoming photon cannot induce a dipole between more than two MOs. As discussed in section 2.2, an induced dipole from a photon with the appropriate energy results in an electronic transition; in this case that transition would involve more than two MOs. Such a transition must be described by a linear combination of MOs as shown for the S8 state of zinc-porphyrin in Eq. 2.21. |𝑆 8 ′ ⟩ = 𝑐 1 |𝛹 𝐻 −3 ⟩ + 𝑐 2 |𝛹 𝐿 ⟩ + 𝑐 3 |𝛹 𝐿 +1 ⟩ 𝐸𝑞 .2.21 Note that the S8 state also consists of all the lower lying occupied orbitals but only the orbitals that are different than the MOs in S0 are shown here, which is why S8 is marked with a prime. The coefficients of each MO reflect their contribution to the excited state. In this case, one expects the values of c3 to be larger than the c2 because the HOMO - 3 to LUMO + 1 transition contributes 60% to the overall S0→S8 transition. The excited state can also be thought of as a partially oxidized and reduced state. This is because photoexcitation involves electrons moving from occupied orbitals to unoccupied orbitals. Removing electrons from the occupied orbitals is effectively a local oxidation on the molecule corresponding to the occupied MO density (i.e. HOMO, HOMO – 1, etc.). Adding electrons to unoccupied orbitals is effectively a local reduction on the molecule corresponding to the unoccupied MO density (i.e. LUMO, LUMO + 1, etc.). Once the photoexcitation occurs, these are no longer formally HOMO - n and LUMO + m orbitals anymore as they are now singly occupied 42 MOs. The previously occupied MOs involved in the photoexcitation now bear a positive charge which is referred to as a “hole”. The newly occupied MOs now bear a negative charge which is referred to as the “electron”. In the case of a simple transition that involves a single orbital pair (i.e. HOMO to LUMO), the hole wavefunction is simply the HOMO and the electron wavefunction is described by the LUMO. In the case of the S8 of zinc porphyrin corresponding to Eq. 2.21, the electron moves from the HOMO – 3 to two unoccupied orbitals, namely the LUMO and LUMO + 1. Thus, the hole is perfectly described by the HOMO – 3 but the electron is a linear combination of the LUMO and LUMO + 1 densities. Visualizing the electron in the MO basis is non-trivial because one must visually compare two densities and mentally add them and weight them by their coefficients. The solution to this is to change the basis to “Nature Transition Orbitals” or “NTOs”. Molecular orbitals are generated by solving Schrödinger’s equation, but NTOs are generated by a “Transition Operator” which acts on the ground state to generate the excited state. NTOs are the wavefunctions that come from the singular value decomposition of the transition operator as shown in Eq. 2.22 [𝑈 † 𝑇𝑉 ] 𝑖𝑗 = √𝜆 𝑖 𝛿 𝑖𝑗 𝐸𝑞 .2.22 where U represents a matrix of occupied MOs, V represents a matrix of unoccupied MOs, and T represents the transition matrix, and ij is the delta function. 20 The eigenvalues λi are the contribution of a given occupied and unoccupied MO pair which are arranged in a positive definite matrix. The orbitals in U and V are in a different basis (the NTO basis) than the MOs, but both describe the same orbital densities. However, NTOs often require less orbitals to describe a transition and make visualizing the hole and electron significantly easier. This is summarized below in Figure 2.18. 43 Figure 2.18 The S8 excited state represented by MOs (top) and NTOs (bottom). The phases in NTOs are not important which is why they are plotted with a single solid color in Figure 2.18. Since the HOMO – 3 is the only MO that is involved in the hole wavefunction, the HOMO – 3 and Hole NTO are visually and mathematically the same. The electron wavefunction is composed of a linear combination of the LUMO and LUMO + 1; the two wavefunctions in the MO basis are shown in Figure 2.18. The electron NTO is composed of a single function that describes 95% of the transition and is plotted below the electron MOs. The electron NTO is a linear combination of the electron MOs which is visually apparent when you add the LUMO to the LUMO + 1 and account for constructive and deconstructive interference. Thus, NTOs are a much more ideal basis for plotting the hole and electron wavefunctions that make up an excited state. It should be noted that the hole and electron are not always well described by a single pair of NTOs, but it generally takes less NTOs to plot the wavefunction than MOs. 44 2.8 Chapter 2 References (1) Drummer, M.; Singh, V.; Gupta, N.; Gesiorski, J.; Weerasooriya, R.; Glusac, K. Photophysics of nanographenes: from polycyclic aromatic hydrocarbons to graphene nanoribbons. Photosynthesis Research 2022, 151. DOI: 10.1007/s11120-021-00838-y. (2) Levine, I. N.; Busch, D. H.; Shull, H. Quantum chemistry; Pearson Prentice Hall Upper Saddle River, NJ, 2009. (3) Kasha, M. Characterization of electronic transitions in complex molecules. Discussions of the Faraday Society 1950, 9 (0), 14-19, 10.1039/DF9500900014. DOI: 10.1039/DF9500900014. (4) Turro, N. J.; Ramamurthy, V.; Scaiano, J. Modern Molecular Photochemistry of Organic Molecules; University Science Books, 2010. (5) Ukwitegetse, N.; Femia, D.; Muthiah Ravinson, D. S.; Djurovich, P. I.; Thompson, M. E. Synthesis and Characterization of Zinc(II) Complexes Bearing 4-Acridinol and 1-Phenazinol Ligands. Inorganic Chemistry 2021, 60 (2), 866-871. DOI: 10.1021/acs.inorgchem.0c02942. (6) Hamze, R.; Peltier, J. L.; Sylvinson, D.; Jung, M.; Cardenas, J.; Haiges, R.; Soleilhavoup, M.; Jazzar, R.; Djurovich, P. I.; Bertrand, G.; et al. Eliminating nonradiative decay in Cu(I) emitters: >99% quantum efficiency and microsecond lifetime. Science 2019, 363 (6427), 601. DOI: 10.1126/science.aav2865. (7) Ravinson, D. S. M.; Thompson, M. E. Thermally assisted delayed fluorescence (TADF): fluorescence delayed is fluorescence denied. Materials Horizons 2020, 7 (5), 1210-1217, 10.1039/D0MH00276C. DOI: 10.1039/D0MH00276C. (8) Koseki, S.; Matsunaga, N.; Asada, T.; Schmidt, M. W.; Gordon, M. S. Spin–Orbit Coupling Constants in Atoms and Ions of Transition Elements: Comparison of Effective Core Potentials, Model Core Potentials, and All-Electron Methods. The Journal of Physical Chemistry A 2019, 123 (12), 2325-2339. DOI: 10.1021/acs.jpca.8b09218. (9) Adachi, C.; Baldo, M. A.; Thompson, M. E.; Forrest, S. R. Nearly 100% internal phosphorescence efficiency in an organic light-emitting device. Journal of Applied Physics 2001, 90 (10), 5048-5051. DOI: 10.1063/1.1409582 (acccessed 2023/03/16). (10) Baldo, M. A.; O'Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.; Thompson, M. E.; Forrest, S. R. Highly efficient phosphorescent emission from organic electroluminescent devices. Nature 1998, 395 (6698), 151-154. DOI: 10.1038/25954. (11) Adachi, C.; Baldo, M. A.; Forrest, S. R.; Lamansky, S.; Thompson, M. E.; Kwong, R. C. High-efficiency red electrophosphorescence devices. Applied Physics Letters 2001, 78 (11), 1622- 1624. DOI: 10.1063/1.1355007 (acccessed 2023/03/16). (12) Suzuri, Y.; Oshiyama, T.; Ito, H.; Hiyama, K.; Kita, H. Phosphorescent cyclometalated complexes for efficient blue organic light-emitting diodes. Science and Technology of Advanced Materials 2014, 15 (5), 054202. DOI: 10.1088/1468-6996/15/5/054202. 45 (13) Holzer, W.; Penzkofer, A.; Tsuboi, T. Absorption and emission spectroscopic characterization of Ir(ppy)3. Chemical Physics 2005, 308 (1), 93-102. DOI: https://doi.org/10.1016/j.chemphys.2004.07.051. (14) Shafikov, M. Z.; Zaytsev, A. V.; Kozhevnikov, V. N.; Czerwieniec, R. Aligning π-Extended π-Deficient Ligands to Afford Submicrosecond Phosphorescence Radiative Decay Time of Mononuclear Ir(III) Complexes. Inorganic Chemistry 2023, 62 (2), 810-822. DOI: 10.1021/acs.inorgchem.2c03403. (15) Shafikov, M. Z.; Daniels, R.; Kozhevnikov, V. N. Unusually Fast Phosphorescence from Ir(III) Complexes via Dinuclear Molecular Design. The Journal of Physical Chemistry Letters 2019, 10 (22), 7015-7024. DOI: 10.1021/acs.jpclett.9b03002. (16) El‐Sayed, M. A. Spin—Orbit Coupling and the Radiationless Processes in Nitrogen Heterocyclics. The Journal of Chemical Physics 1963, 38 (12), 2834-2838. DOI: 10.1063/1.1733610 (acccessed 2023/03/17). (17) Voorhis, P. R. G. G. T. V. Two Electrons: Exited State. 2007. https://ocw.mit.edu/courses/5- 61-physical-chemistry-fall 2007/d93711b502f4deabc67454a94cd922e1_lecture26.pdf (accessed. (18) Epifanovsky, E.; Gilbert, A. T. B.; Feng, X.; Lee, J.; Mao, Y.; Mardirossian, N.; Pokhilko, P.; White, A. F.; Coons, M. P.; Dempwolff, A. L.; et al. Software for the frontiers of quantum chemistry: An overview of developments in the Q-Chem 5 package. The Journal of Chemical Physics 2021, 155 (8), 084801. DOI: 10.1063/5.0055522 (acccessed 2023/01/08). (19) Gill, P. M. W.; von Rague Schleyer, P. Density functional theory (DFT), Hartree-Fock (HF), and the self-consistent field. J. Chem. Phys 1994, 100, 5066-5075. (20) Martin, R. L. Natural transition orbitals. The Journal of Chemical Physics 2003, 118 (11), 4775-4777. DOI: 10.1063/1.1558471 (acccessed 2023/03/21). 46 3 Chapter 3: π-Extended Ligands in Two-Coordinate Coinage Metal Complexes for Organic LED Applications This chapter details published research on record fast blue photon emitting materials for OLED applications. Dr. Anton Razgoniaev discovered a two-coordinate complex with record fast blue photon emitting capabilities. Jonas Schaab and I worked hard to synthesize and characterize a family of analogous complexes to understand the underlying structure-function relationship that resulted in the marked fast rate of emission. I lead efforts in understanding and explaining the theoretical physics that resulted in the fast emission, while Jonas Schaab lead efforts in determining temperature-dependant photophysical properties. This work was published in J. Am. Chem Soc. 1 Two-coordinate carbene-M-amide (cMa, M = Cu(I), Ag(I), Au(I)) complexes have emerged as highly efficient luminescent materials for use in a variety of photonic applications due to their extremely fast radiative rates. In this work, a series of cMa derivatives was prepared to examine the variables that affect the radiative rate, with the goal of understanding the parameters that control the radiative rate in these materials. We find that blue-emissive complexes with high photoluminescence efficiencies (ΦPL > 0.95) and fast radiative rates (kr = 4 × 10 6 s –1 ) can be achieved by selectively extending the π-system of the carbene and amide ligands. Of note is the role played by the increased separation between the hole and electron density distributions in the ICT excited state. Analysis of temperature-dependent luminescence data and theoretical calculations indicate that the hole–electron separation exerts a primary effect on the energy gap between the lowest-energy singlet and triplet states (ΔEST) while keeping the radiative rate for the singlet state relatively unchanged. This interpretation provides guidelines for the design of new cMa derivatives with even faster radiative rates in addition to those with slower radiative rates and thus extended excited state lifetimes. 47 3.1 A Closer Look at the Blue Problem As mentioned in Chapter 1, OLEDs are an emerging display technology that offer better picture quality than traditional LED/LCD displays. OLED displays are made up of red, green, and blue pixels. The pixels contain the emissive molecule (the “dopant” molecule) that is contained within a stack of materials that facilitate electroluminescence of the dopant molecule. Unlike photoluminescence, the excited states here are generated by electrochemically oxidizing the HOMO and reducing the LUMO as shown in Figure 3.1. Figure 3.1 A simplified schematic of electrochemical generation of the excited state in OLED devices. A real device contains a hole and electron transport layer, and a host material which help efficiently transport holes and electrons onto the dopant molecule. Holes are transferred to the HOMO which oxidizes the dopant, and electrons are transferred to the LUMO which reduces the dopant. When both of these processes occur on the same molecule, one electron remains in the HOMO and one is placed in the LUMO. This is an alternate way to generate excited states compared to the photo-excitation mechanism that has been primarily discussed in earlier chapters. A modern challenge of this technology is that blue pixels stop working much faster than red and green pixels, which eliminates 1/3 of the display. 2-6 This is because an excited 48 state that emits blue light carries much more energy than a green or red emitting excited state. Two molecules that are simultaneously in the excited state can perform energy transfer and combine their energy by returning one of the molecules to the ground state and doubling the energy of the other molecule’s excited state (E + E = 2E). Doubling the excited state energy is a particularly deleterious process for blue dopant molecules, because doubling the energy of blue light leads to rupturing of the molecular bonds and decomposition of the dopant. The contemporary solution to the so called “blue problem” is to design blue dopants that rapidly decay to the ground state with high PL before energy transfer can occur. The most important photophysical parameter to accomplish this goal is maximizing kr. One is essentially racing the rate of energy transfer by increasing the radiative rate. Long lived excited states are an issue for blue-emitting dopants for OLEDs. One feature of the electroluminescence mechanism shown in Figure 3.1 is that there is no control over the net spin of the excited state. In photoexcitation of the S0 state, the spontaneous transition conserves spin such that only singlet states are generated (S0→Sn). However, electroluminescence offers no control over spin selection in the electron that is oxidized from the homo, nor the electron that is put into the LUMO. As mentioned in section 2.3, there is only one way to make a singlet state (one electron spin up, one electron spin down), but there are three unique ways to make a triplet. The 1:3 ratio of state preparation yields a 25% to 75% probability of generating singlets and triplets, respectively. This initially posed a large issue for OLED technology because triplets have slow kr due to the spin floridness of the T1→S0 transition. There is no restriction on the knr of a triplet state which results in poor PL. This leads to inefficient devices because a significant portion of the energy provided to the pixel is lost non-radiatively. Another issue for purely organic systems is that the triplet state can be very long lived (0.1 ms - hours) which is counterproductive to the blue 49 problem. A significant advancement of OLED technology was accomplished by the Thompson and Forrest groups by using organometallic complexes as the dopant molecules based on iridium. The inclusion of a heavy atom such as iridium induces strong SOC to facilitate radiative decay from the T1 to the S0 (see section 2.3). The kr in iridium complexes can be as high as 1-1.3 × 10 6 s -1 corresponding to excited state lifetimes in the 0.7-10 s regime with PL ~1.0. 7-10 The contemporary approach to increasing the radiative rate out of the triplet is to design molecules that have a small EST on the order of 25 to 100 meV. A small energy gap allows for thermal promotion from the T1 into the S1 which is referred to as “Reverse Inter-System Crossing” or “RISC”, despite the definition of ISC not being specific to S 1→T1 or T1→S1. Once the triplet has thermally populated to the singlet, it is no longer forbidden to relax to the ground state. The process of thermal population of the S1 from the T1, followed by emission is known as “Thermally Assisted Delayed Fluorescence” or “TADF” and is demonstrated in Figure 3.2. Figure 3.2 Efficient electroluminescence from fast phosphorescence (left) and Thermally Assisted Delayed Fluorescence (right). This simple schematic assumes that the radiative and ISC rates are much greater than any non-radiative decay out of the S1 or T1. Not shown is prompt fluorescence that is competitive with ISC, but represents a minor component of the emitted photons. 50 TADF is not as fast as regular fluorescence (ps to ns) because the excited state spends time in the T1 before it is able to reverse ISC into the S1 resulting in delayed fluorescence. However, TADF can be faster than phosphorescence from heavy metal complexes. The two principal factors that govern the TADF rate constant (kTADF) are the radiative rate of the singlet kr(S1→S0), and the EST. Smaller EST results in a less energetically uphill RISC process, which shifts the equilibrium between the S1 and T1 population more towards the S1 population. A higher S1 population results in faster emission. Once the S1 state is populated, the rate of radiative relaxation to the ground state is governed by kr(S1→S0) (or “kS1”). The full kinetic model for molecules with fast ISC and near unity PL is shown in Eq. 3.1. 11 𝑑 (ℎ𝜈 ) 𝑑𝑡 = 𝑘 𝐼𝑆𝐶 𝑘 𝑆 1 𝑘 𝑅𝐼𝑆𝐶 [𝑇 1 ] 𝐸𝑞 .3.1 The equilibrium constant between the singlet and triplet states is simply K eq = kISC/kRISC, thus the radiative rate constant out of the T1 state is kS1 × Keq. Thus, the TADF lifetime “TADF” is given by Eq. 3.2 𝜏 𝑇𝐴𝐷𝐹 = 𝜏 𝑆 1 𝐾 𝑒𝑞 𝐸𝑞 .3.2 Thus, increasing kTADF (or equivalently, decreasing TADF) can be done by increasing kS1 or decreasing Keq. We established in section 2.3 that the radiative rate is related to the oscillator strength. The radiative rate is given by Eq. 3.3 𝑘 𝑟 (𝑆 1 → 𝑆 0 ) = 8𝜋 2 〈𝜈̃ 𝑓 〉 3 3ℎ 2 |⟨𝑆 1 |𝜇 ̂ |𝑆 0 ⟩| 2 𝐸𝑞 .3.3 where “h” is Planck’s constant, and “f” is the emission energy. The radiative rate generally benefits from a higher degree of overlap in the S0 and S1 NTOs. However, the equilibrium constant 51 is largely governed by the singlet-triplet energy gap which is given by the exchange integral discussed in section 2.3 𝐽 ≈ 1 |𝑟 1 ⃑⃑⃑ − 𝑟 2 ⃑⃑⃑ | ⟨𝛹 ℎ+ |𝛹 𝑒 + ⟩ 2 𝐸𝑞 .2.16 where h+ and e- represent the hole and electron wavefunctions (which are approximately equivalent to the HOMO and LUMO wavefunctions respectively). The exchange integral would be optimized if the hole and electron wavefunctions were completely separated which would result in an exchange energy of zero and would maximize Keq. However, spatially separated wavefunctions also drop kr to zero. Thus, there is a tradeoff in wavefunction design. It is very challenging to mathematically optimize the hole and electron wavefunctions such that kr × Keq is optimized. Even if a mathematical solution was found, there is no guarantee that a molecule could be designed with a wavefunction that matches the solution to the optimization problem. Instead, research groups have taken an empirical approach to finding optimal wavefunction design. One strategy that has been used is to design a series of molecules with a varying degree of hole/electron NTO overlap, and plot that against the experimental kTADF. 12 These studies find a fairly scattered distribution with an apparent maximum at around 28% overlap of the hole and electron NTOs. One of the issues with this analysis is that neither the exchange integral nor the oscillator strength are pure overlap integrals of the wavefunction. They are both modified by an operator and that changes the outcome of the integral. In the following work, we looked at plotting kTADF against the distance between the centers of distribution for the hole and electron wavefunction for two-coordinate Au(I) compounds. The idea being that spatial separation of the wavefunctions is also an explicit term in the exchange energy (and thus ST), and is expected to directly impact Keq and kTADF. 52 3.2 Introduction The luminescent properties of coinage metal complexes (M (I) = Cu, Ag and Au) were reported over fifty years ago, 13, 14 with the first report of emission from a two-coordinate d 10 coinage metal complex in 1987. 15 Several papers have highlighted emission in the solid state and in solution for M (I) L2 + and LM (I) X complexes (L = phosphine, carbene, X = halide, acetylide, aryl, amide). 16-30, 2, 31-40 Of particular interest here is the promise of (carbene)M (I) (amide) (cMa) complexes as efficient luminescent materials. 2, 31-40 The cMa complexes can have high photoluminescent quantum yields (PL), short luminescence decay lifetimes () in the s regime and shorter, and emission color tunable over the entire visible spectrum in solid, solution and doped films. 2, 29, 32-41 These luminophores have properties similar to transition metal phosphors that contain Ru, Os, Ir and Pt used in a range of applications including organic electronics and LEDs 2, 17, 31-40, 42-45 , photocatalysis 46-48 , chemo- and bio-sensing 49-51 and solar energy conversion. Unlike the noble metal phosphors which luminesce from triplet states, the majority of the reported cMa complexes emit via thermally assisted delayed fluorescence (TADF), Figure 3.3. 11 Figure 3.3 The kinetic scheme for emission via TADF , where 𝑘 𝑟 S 1 and 𝑘 𝑟 TADF are radiative decay rates of S1 state and TADF process, Keq indicates the equilibrium constant between T1 and S1 states. The carbene ligand serves as an electron acceptor (A) and amide ligand serves as an electron donor 53 The lowest energy (emitting) excited state is an interligand charge transfer (ICT) transition between the two ligands. The energy of the ICT state depends on the choice of ligands but is relatively insensitive to the identity of the metal atom. 35 The linear geometry of the cMa complexes leads to a large spatial separation between the donor and acceptor groups/ligands of ~4 Å. This spacing restricts the overlap between the -orbitals of the two ligands, consequently limiting the exchange energy and thus the energy gap between lowest singlet (S1) and triplet (T1) states (EST). A small EST enhances thermal population of the singlet state, which improves the luminescence efficiency for TADF by increasing the radiative rate for emission. 11 Organic TADF molecules have distinct lifetimes for prompt ( = 1-100 ns) and delayed ( = 1-1000 s) emission that are controlled by EST and the rate of intersystem crossing S1→T1 (typically kISC < 10 7 s -1 ). 52-57 In contrast, the intersystem crossing (ISC) rates in metal containing TADF complexes are fast enough (kISC ≥ 10 10 s -1 ) to outcompete the radiative rates for the S1 state, making delayed emission (TADF) independent of kISC. 11 The result is extremely fast prompt emission ( < 200 ps) and comparatively short TADF lifetimes,  TADF = 0.5-3 s, leading to high luminescence efficiency. 30, 32, 33, 35, 58 For compounds where the ISC rate exceeds 𝑘 𝑟 S 1 , a pre-equilibrium approximation can be made such that the equilibrium constant (𝐾 𝑒𝑞 ,𝑇 1 ⇄ 𝑆 1 ) becomes a principal factor in determining 𝑘 𝑟 TADF as shown in Eq. 3.4: 11 𝑘 𝑟 TADF = 𝑘 𝑟 S 1 ∙ 𝐾 𝑒𝑞 𝐸𝑞 .3.4 In this equation, 𝑘 𝑟 TADF is only dependent on 𝑘 𝑟 S 1 and 𝐾 𝑒𝑞 , with Keq tied to EST, Eq. 3.5. 59 𝐾 𝑒𝑞 (𝑇 1 ⇄ 𝑆 1 ) = 1 3 𝑒 − ∆𝐸 𝑆𝑇 𝑘 𝑏 𝑇 𝐸𝑞 .3.5 54 Therefore, predictions can be made regarding the TADF properties for TADF emitters with fast ISC (high spin orbit coupling) without prior knowledge of the ISC rates since only 𝑘 𝑟 S 1 and EST are needed to estimate 𝑘 𝑟 TADF . Boltzmann fits of temperature dependent luminescence data can be used to accurately derive EST and 𝑘 𝑟 S 1 values, and thus K eq (Eqn. 2). 11, 12, 35, 36 The values of 𝑘 𝑟 S 1 can also be estimated experimentally from absorption spectra using the Strickler-Berg analysis. 60 Being able to control the rate of TADF with EST and 𝑘 𝑟 S 1 is useful when designing chromophores for different applications. A small EST and large 𝑘 𝑟 S 1 leads to a fast 𝑘 𝑟 TADF , 11 which is important for applications where a short excited state lifetime is important, such as organic LEDs. To that end, this work describes cMa complexes with some of the shortest TADF lifetimes reported to date ( ~ 250 ns). Conversely, one can design molecules where the EST value is large, which will make 𝐾 𝑒𝑞 (𝑇 1 ⇄ 𝑆 1 ) very small and push (TADF) in the 10-100 s regime. These long-lived materials could be used as sensitizers for photoelectrochemical reactions where the diffusion of the excited species to an electrocatalysis in solution is a key step in the process. A EST value of 1000 cm -1 is sufficient to increase the TADF lifetime to the s regime, while only sacrificing ~100 meV in electrochemical potential for the excited state. Both 𝑘 𝑟 S 1 and EST parameters are related to the degree of overlap between the hole and electron wavefunctions that describe the excited state. This overlap can be evaluated using the natural transition orbitals (NTOs) and is referred to as 𝛬 𝑁𝑇𝑂 . 4, 12, 61 While 𝛬 𝑁𝑇𝑂 appears to be useful in predicting both 𝑘 𝑟 S 1 and Keq, it affects the two parameters in opposite ways. A large 𝛬 𝑁𝑇𝑂 leads to a high value of 𝑘 𝑟 S 1 , but it also increases the value for EST and thus lowers Keq. In a previous study we found that an 𝛬 𝑁𝑇𝑂 = 0.25-0.3 is an optimal range to give the fastest 𝑘 𝑟 TADF for 55 cMa complexes of Cu, Ag and Au with a variety of carbene and amide ligands. 12 The lowest 𝛬 𝑁𝑇𝑂 values were obtained for silver based cMa complexes. In this work, the synthesis and characterization of a family of new cMa materials is reported. The photophysical properties of these complexes was studied with an eye to further explore how ligands can affect the excited state properties. The compounds discussed here are illustrated in Figure 3.4. Neither the BZAC and PAC carbene ligands, nor the use of bim (Figure 3.4) as the amide, in a cMa complex has been reported previously. We find that the choice of carbene ligand affects the excited state energy but has a weaker influence on the photophysical properties and EST of the cMa complexes. Interestingly, shifting from a carbazolide to a bim ligand markedly increases 𝑘 𝑟 TADF even though the energy of the HOMO is effectively the same for both complexes. 56 Figure 3.4 Compounds considered in this work (Ar = 2,6-diisopropylphenyl). 𝑀 𝑋 𝑃𝐴𝐶 : M = Cu, Ag, Au; X = Cz, BCz and 𝑀 𝑋 𝐵𝑍𝐴𝐶 : M = Cu, Au; X = Cz, BCz 𝐴𝑢 𝑎𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 : carbene = IPr, BZI, PZI, CAAC, MAC, PAC, BZAC; am = Cz, bim, Mbim, Obim 3.3 Synthesis of all Complexes The synthetic scheme of all complexes in work is shown below in Figure 3.5. All complexes studied here are nearly indefinitely air stable in the solid state. The complexes are stable in solution for prolonged periods, with the exception of 𝐴𝑔 𝐵𝐶𝑧 𝑃𝐴𝐶 which decomposes over several days in all solvents. Dr. Anton O. Razgoniaev lead the synthesis of bim derivatives, and Jonas Schaab contributed significantly to synthesis and purification of the reported complexes. Full details of the synthesis and characterization are included in section 3.15. 57 Figure 3.5 Synthetic chart for all materials prepared in this study. M’ = Cu(I), Ag(I), or Au(I). M’’ = Cu(I) or Au(I) 3.4 Crystal Structures of the Complexes The single crystal structures were solved by Jonas Schaab. All crystals were grown by recrystallization. Vapor diffusion of hexanes or pentane into a solution of the compound in dichloromethane. A Cryo-Loop was used to mount the sample with Paratone oil. All single crystal structures were determined at 100K with Rigaku Xta LAB Synergy S, equipped with an HyPix- 58 600HE detector and an Oxford Cryostream 800 low Temperature unit, using Cu K  PhotonJet-S X-ray source. The frames were integrated using the SAINT algorithm to give the hkl files. Data were corrected for absorption effects using the multi-scan method (SADABS) with Rigaku CrysalisPro. The structures were solved by intrinsic phasing and refined with the SHELXTL Software Package. 62 If necessary, the disordered solvent treatment method BYPASS for co- crystalizing solvent molecules, was implemented and marked in the CCDC entry. Each structure was deposited in the Cambridge Crystallographic Data Centre with the following accension codes: 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 2168084, 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 2170320, 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 2182717, 𝐴𝑢 𝑏 𝑖𝑚 𝐶𝐴𝐴𝐶 2155241, 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 2167514, 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 2170000 and 𝐴𝑢 𝑀𝑏𝑖𝑚 𝐵𝑍𝐼 2168086. All cif files and report data including atom position, bond lengths and bond angle can be downloaded from the CCDC database, using the database number in the right column of the following table. Furthermore, the following table is giving the most important bond lengths, angles and the torsion angle around the metal center. The conformer ratio represents the ratio of two different disordered conformations of the molecule. Only 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 crystallized with two molecules in the asymmetric unit, therefore the parameters were given for both molecules in the asymmetric unit. All others crystalized with one molecule in the asymmetric unit. 59 Table 3.1 Crystallographic parameters for BZAC-Au-Cz, PAC-Cu-BCz, BZAC-Au-bim and CAAC-Au-bim. Identification code BZAC-Au-Cz PAC-Cu-BCz BZAC-Au-bim CAAC-Au-bim Empirical formula C176H192Au4N12 C104H124Cu2N6O2 C45H48AuN5 C40H47AuN4 Formula weight 191.96 1613.16 855.883 780.78 Temperature/K 293(2) 100.00(10) 100.01(10) 107(11) Crystal system triclinic triclinic orthorhombic monoclinic Space group P-1 P-1 Fdd2 I2/a a/Å 9.0083(2) 16.5805(2) 57.2043(10) 23.3536(3) b/Å 19.3190(3) 17.3257(2) 30.8094(5) 9.61920(10) c/Å 23.0688(2) 20.0203(2) 8.8252(1) 33.2792(4) α/° 82.1710(10) 68.7580(10) 90 90 β/° 80.8960(10) 72.6580(10) 90 100.5580(10) γ/° 83.406(2) 62.0990(10) 90 90 Volume/Å 3 3909.28(11) 4679.32(11) 15553.8(4) 7349.37(15) Z 2 2 16 8 ρcalcg/cm 3 1.386 1.145 1.462 1.411 μ/mm -1 7.306 0.934 7.389 7.752 F(000) 1648 1724 6870.6 3152 Crystal size/mm 3 0.549 × 0.058 × 0.029 0.25 × 0.195 × 0.135 0.118 × 0.025 × 0.015 0.138 × 0.098 × 0.029 Radiation Cu Kα (λ = 1.54184) Cu Kα (λ = 1.54184) Cu Kα (λ = 1.54184) Cu Kα (λ = 1.54184) 60 2Θ range for data collection/° 5.694 to 163.426 5.984 to 161.91 6.18 to 160.66 5.402 to 161.006 Index ranges -11 ≤ h ≤ 11, -24 ≤ k ≤ 24, -27 ≤ l ≤ 29 -21 ≤ h ≤ 21, -21 ≤ k ≤ 22, -25 ≤ l ≤ 25 -68 ≤ h ≤ 72, -39 ≤ k ≤ 37, -7 ≤ l ≤ 10 -29 ≤ h ≤ 29, -12 ≤ k ≤ 8, -42 ≤ l ≤ 42 Reflections collected 92189 301930 24795 60195 Independent reflections 16775[Rint = 0.1410, Rsigma = 0.0759] 20281[Rint = 0.0668, Rsigma = 0.0231] 6414[Rint = 0.0343, Rsigma = 0.0274] 7981[Rint = 0.0901, Rsigma = 0.0472] Data/restraints/parameters 16775/0/882 20281/0/1055 6414/1/468 7981/0/412 Goodness-of-fit on F 2 1.14 1.052 1.038 1.085 Final R indexes [I>=2σ (I)] R1 = 0.0791, wR2 = 0.2219 R1 = 0.0675, wR2 = 0.1654 R1 = 0.0359, wR2 = 0.0946 R1 = 0.0364, wR2 = 0.0915 Final R indexes [all data] R1 = 0.0911, wR2 = 0.2383 R1 = 0.0725, wR2 = 0.1689 R1 = 0.0377, wR2 = 0.0953 R1 = 0.0431, wR2 = 0.0960 Largest diff. peak/hole / e Å -3 4.79/-3.15 1.59/-1.48 2.12/-1.16 2.49/-1.83 #CCDC 2168084 2170320 2182717 2155241 61 Table 3.2: Crystallographic parameters for MAC-Au-bim, PAC-Au-bim and BZI-Au-Mbim. Identification code MAC-Au-bim PAC-Au-bim BZI-Au-Mbim Empirical formula C43H50AuN5O C45H46AuN5O C45H48AuN5 Formula weight 849.876 869.83 855.85 Temperature/K 123(30) 100.5(8) 100(90) Crystal system monoclinic monoclinic monoclinic Space group C2/c P21/n P21/c a/Å 15.9731(1) 13.61560(10) 8.91960(10) b/Å 22.6254(2) 18.11870(10) 29.5199(5) c/Å 21.5638(2) 16.34850(10) 15.2229(3) α/° 90 90 90 β/° 93.533(1) 109.8980(10) 99.193(2) γ/° 90 90 90 Volume/Å 3 7778.30(11) 3792.35(5) 3956.79(11) Z 8 4 4 ρcalcg/cm 3 1.451 1.523 1.437 μ/mm -1 7.4 7.609 7.263 F(000) 3419.4 1752 1728 Crystal size/mm 3 0.27 × 0.14 × 0.13 0.104 × 0.083 × 0.043 0.29 × 0.059 × 0.032 62 Radiation Cu Kα (λ = 1.54184) Cu Kα (λ = 1.54184) Cu Kα (λ = 1.54184) 2Θ range for data collection/° 6.78 to 161.6 7.33 to 160.548 5.988 to 176.412 Index ranges -20 ≤ h ≤ 20, -28 ≤ k ≤ 28, -24 ≤ l ≤ 27 -17 ≤ h ≤ 16, -23 ≤ k ≤ 23, -20 ≤ l ≤ 20 -9 ≤ h ≤ 11, -37 ≤ k ≤ 37, -19 ≤ l ≤ 19 Reflections collected 132598 128799 80685 Independent reflections 8489 [Rint = 0.0702, Rsigma = 0.0200] 8259 [Rint = 0.0596, Rsigma = 0.0195] 8698 [Rint = 0.1226, Rsigma = 0.0429] Data/restraints/parameters 8489/0/471 8259/0/477 8698/0/476 Goodness-of-fit on F 2 0.945 1.178 1.128 Final R indexes [I>=2σ (I)] R1 = 0.0634, wR2 = 0.1472 R1 = 0.0314, wR2 = 0.0852 R1 = 0.0622, wR2 = 0.1347 Final R indexes [all data] R1 = 0.0639, wR2 = 0.1474 R1 = 0.0334, wR2 = 0.0863 R1 = 0.0701, wR2 = 0.1549 Largest diff. peak/hole / e Å -3 2.20/-2.14 1.10/-1.51 2.41/-2.89 #CCDC 2167514 2170000 2168086 63 3.5 Electrochemical Methods Cyclic voltammetry and differential pulsed voltammetry were performed using a VersaSTAT potentiostat measured at 100 mV/s scan. Anhydrous dimethylformamide was used as the solvent, with 0.1 M tetra(n-butyl)ammonium hexafluorophosphate as the supporting electrolyte. The redox potentials are based on values measured from differential pulsed voltammetry and are reported relative to the ferrocenium/ferrocene (Cp2Fe + /Cp2Fe) redox couple using either ferrocene or decamethylferrocene as an internal reference. Electrochemical reversibility was determined using cyclic voltammetry. 3.6 Computational Methods The electronic properties of the complexes were modelled in Q-Chem using Density Functional Theory (DFT) and Time Dependent DFT (TDDFT). 63 The electronic properties of the complexes were modeled using Density Functional Theory (DFT) and Time Dependent DFT (TDDFT). First, a geometry optimization was performed using the B3LYP functional and LACVP basis set which accounts for molecules containing transition metals. TDDFT calculations were performed on the ground-state optimized structures to estimate the S0→S1 transition energies and corresponding oscillator strengths. The TDDFT calculations were performed using the LACVP basis, the CAM-B3LYP exchange, the fit-LACVP effective core potential, the random phase approximation, and the omega value set to 0.2 arbitrary units. Natural transition orbitals were generated by performing a singular value decomposition on the transition density matrix using the Q-Chem software package. The resultant eigenstates are a linear combination of molecular orbitals involved in the S0 → Sn transition which are separated into hole and electron pairs. The advantage of NTOs is that electronic transitions are often 64 composed of more than two molecular orbitals (i.e., HOMO to LUMO). Interestingly, the HOMO and LUMO of the compounds in this work make up >99.9% of the S 0→S1 transition, aside from the PAC complexes which mix the LUMO + 1 into the electron NTO. This is demonstrated with 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 and 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 as example compounds in Figure 3.6 and Figure 3.7 respectively. Figure 3.6 The S0→S1 transition of 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 involves an electron moving from the HOMO to the LUMO and HOMO to LUMO + 1. The hole and electron NTOs (Ψh+ and Ψe-) are appropriate visualizations of the hole and electron wavefunctions corresponding to this transition. One can see that the electron NTO is made up of a linear combination of the ΨLUMO and ΨLUMO+1. The isopropyl groups are omitted from the diisopropylphenyl groups for clarity. All PAC compounds give a mixed LUMO/LUMO+1 behavior in the electron NTO which is consistent with the electron NTO being consistent for a given carbene across different metals and donors. However, the Ipr, CAAC, BZAC, BZI, and MAC complexes all have S 1→S0 NTOs that are described purely by the HOMO and LUMO, localized on the amide and carbene ligands, 65 respectively. Figure 3.7 The S0→S1 transition of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 involves an electron moving from the HOMO to LUMO and HOMO to LUMO + 1. The hole and electron NTOs (Ψh+ and Ψe-) are appropriate visualizations of the hole and electron wavefunctions corresponding to this transition. One can see that the electron NTO is made up of a linear combination of the ΨLUMO and ΨLUMO+1. The isopropyl groups are omitted from the diisopropylphenyl groups for clarity. Further, all compounds in this study are described by a single hole and electron NTO pair. This simplifies overlap integral calculations to a single spatial integral between two NTOs. The spatial NTO overlap integrals 𝛬 𝑁 𝑇𝑂 were calculated for these compounds using a script that draws from the integration methods developed by G. Herman, et al., 64 and S. G. P. Castro, et al., 65 using a method reported previously. 4, 12 This makes use of the integral in Eq. 3.6 which computes a weighted average of spatial overlap of all NTO pairs that contribute to the excited state 𝛬 𝑁𝑇𝑂 ≈ ∑𝑐 𝑘 𝑘 ∰|𝜑 k,ℎ+ ||𝜑 k,𝑒 − |𝑑𝜏 ∑𝑐 𝑘 𝑘 𝐸𝑞 .3.6 66 As aforementioned, all compounds in the study only require one electron/hole NTO pair to describe the excited state which simplifies the integral. The overlap calculations reveal the drawbacks of relying on iso-value representations of the wavefunction. All NTO/MO visualizations with isovalue = 0.1 suggest that the hole and electron wavefunctions have very poor spatial overlap, and that most of the overlap comes from the metal d orbital. Certainly, one would not expect the calculated 𝛬 𝑁𝑇𝑂 values of 26-43% based on the wavefunction images in Figure 3.6 and Figure 3.7. Thus, iso-value representations must be viewed with caution because the viewer does not know the probability of finding an electron within that surface, but only the iso-surface corresponding to a probability per unit volume. In the case of cMa complexes, the volume within iso-value = 0.1 only contains 15% probability of finding an electron within the surface, which explains why the overlap integral can be large despite the iso-value = 0.1 visualization. As mentioned in section 3.1, we opted to inspect the empirical relationship between kTADF and the hole and electron density distributions. The center of charge for the hole and electron NTOs of a given transition are extracted from the position expectation values of the respective wavefunctions 〈𝑟 ℎ+ 〉 = ⟨𝛹 ℎ+ |𝑟 ̂ |𝛹 ℎ+ ⟩ 𝐸𝑞 .3.7 〈𝑟 𝑒 − 〉 = ⟨𝛹 𝑒 − |𝑟 ̂ |𝛹 𝑒 − ⟩ 𝐸𝑞 .3.8 where |𝛹 ℎ+ ⟩ and |𝛹 𝑒 − ⟩ represent the hole and electron NTO wavefunctions. This can be calculated component-wise to extract the vectors 𝑟⃗ ℎ+ = ∑⟨𝛹 ℎ+ |𝑟 ̂ 𝑗 |𝛹 ℎ+ ⟩ 𝑗 |𝑗 ⟩ 𝐸𝑞 .3.9 𝑟⃗ 𝑒 − = ∑⟨𝛹 𝑒 − |𝑟 ̂ 𝑗 |𝛹 𝑒 − ⟩ 𝑗 |𝑗 ⟩ 𝐸𝑞 .3.10 67 Where j represents a direction in a coordinate system. The hole-electron separation is the difference of Eq. 3.9 and Eq. 3.10 𝑑 (ℎ + ,𝑒 − ) = |𝑟⃗ ℎ+ − 𝑟⃗ 𝑒 − | 𝐸𝑞 .3.11 which describes the separation between the hole and electron charges that constitute the exciton. Eq. 3.9 and Eq. 3.10 are also useful for visualizing the charges because a pseudo-particle can be imposed on the corresponding molecule by letting |𝑗 ⟩ represent molecular coordinates as demonstrated in Figure 3.8. Again, the centers of charge are readily determined with the QChem software package. Center of charge descriptions are a much simpler parameter to understand compared to the wavefunction overlap 𝛬 𝑁𝑇𝑂 . Similar behavior is expected; as the +/- separation increases, the exchange integral and the radiative rate should decrease. The utility of the center of charge descriptor is that it simply describes where the wavefunctions are centered, rather than integrating their products. 68 Figure 3.8 The NTO (left) and center of charge representation (right) for the hole and electron wavefunctions of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 . The isopropyl groups from the diisopropylphenyl substituents have been deleted for clarity. The iso value was set to 0.1 and the phase information omitted for the wavefunction visualization. The red density represents the hole wavefunction and the blue density represents the electron wavefunction respectively. A red dummy atom was placed at the rh+ coordinate and a blue dummy atom was placed at the re- coordinate. 3.7 Photophysical Measurements Samples in fluid solution were both sparged and examined under N2. Doped polystyrene thin films were prepared from a solution of polystyrene in toluene, drop cast onto a quartz substrate and measured under N2. The UV-visible spectra were recorded on a Hewlett-Packard 4853 diode array spectrometer. Steady state excitation and emission spectra were obtained using a Photon Technology International QuantaMaster spectrofluorimeter. Photoluminescence quantum yields were recorded using a Hamamatsu C9920 integrating sphere equipped with a xenon lamp. Luminescence lifetimes were measured using Time-Correlated Single Photon Counting (TCSPC) on an IBH Fluorocube apparatus interfaced to a Horiba FluoroHub+ controller. Variable temperature photophysical measurements were carried out on a Janis SHI-4-2 (0.2 W 4K) optical cryocooler with a Lakeshore 335 Temperature controller and evacuated by an Drytel 31 69 Turbomolecular pump to 1.2x 10 -4 mTorr. The IBH Fluorocube was used as a detector for luminescence lifetimes and the Photon Technology International QuantaMaster spectrofluorimeter as a detector for steady state emission spectra with 375 nm LED (Thorlabs M375L4, 1270 mW) as excitation source. Doped polystyrene thin films were spin coated onto a round sapphire substrate that was used to insure good thermal conductivity at low temperatures. More information on this setup is provided in the published manuscript. 1 A Horibia Fluorohub+ with a Horiba Jobin Yvon detector with monochromator was used for excited state lifetime studies, with a a NanoLED 407N (405 nm) or IBH SpectraLED S-03 (372 nm) excitation source. Values for the measured lifetime were plotted against temperature. High temperature data (150 K – 300 K) was fit to an Arrhenius equation according to Hamze, et al. 35 EST and S1 were obtained from the slope and intercept of ln(kTADF) vs. 1/T respectfully. 3.8 Structural Results Single crystal X-ray diffraction was used to determine the molecular structures of 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 , 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 , 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 , 𝐴𝑢 𝑏𝑖𝑚 𝐶𝐴𝐴𝐶 , 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 , 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 and 𝐴𝑢 𝑀𝑏𝑖𝑚 𝐵𝑍𝐼 . Crystallographic data for the seven complexes are given in Table 3.1 and Table 3.2 in section 3.4. Representative structures of the compounds of 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 , 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 and 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 are shown in Figure 3.9 and Figure 3.10. 70 Figure 3.9 Thermal ellipsoid drawings of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 , 𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 , 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 and 𝐴𝑢 𝐵𝑖𝑚 𝐶𝐴𝐴𝐶 . Figure 3.10 Thermal ellipsoid drawings of 𝐴𝑢 𝐵𝑖𝑚 𝑀𝐴𝐶 , 𝐴𝑢 𝐵𝑖𝑚 𝑃𝐴𝐶 and 𝐴𝑢 𝑀𝐵𝑖𝑚 𝐵𝑍𝐼 All the complexes have a coplanar conformation of carbene and amide ligands (dihedral angles = 0.7-7.3°) and close to 180˚ angle around the metal center (C-Au-N = 174 - 179˚). The C-Au bond 71 lengths are in the range 1.97-2.01 Å and the Au-N bond is 2.01 – 2.03 Å (Table 3.3). Values for the C-Au and Au-N bond lengths are similar to analogous (carbene)M(Cz) complexes published previously. 31, 34, 35, 38, 66 The only Cu compound structurally characterized in this report, 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 , has C-Cu and Cu-N bond lengths of 1.89 Å and 1.85 Å, respectively, that are shorter than the Au based compounds and consistent with the difference in ionic radii of the metals. 35 Two different conformers are possible for 𝐴𝑢 𝐵𝑖 𝑚 𝑀𝐴𝐶 , one with the carbonyl anti to the phenylene (pictured in Figure 3) and the other where the carbonyl is syn to the phenylene. Both conformers were present in the crystal and the diffraction data was best fit by treating the crystal as disordered with a 10% “impurity” of the syn-conformer in a crystal of the anti-conformer. Only a single conformer for 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 , 𝐴𝑢 𝐵𝑖𝑚 𝑃𝐴𝐶 and 𝐴𝑢 𝐵𝑖𝑚 𝐶𝐴𝐴𝐶 was observed, even though a similar stereocenter to that of 𝐴𝑢 𝐵𝑖𝑚 𝑀𝐴𝐶 was present in these molecules. The crystals of 𝐴𝑢 𝑀𝑏𝑖𝑚 𝐵𝑍𝐼 contain two tautomers, with the major component (84%) having the methyl group closest to the Au-N bond. The ratio of tautomers in the single crystals is close to the 75:25 ratio observed using NMR spectroscopy. 72 Table 3.3 Selected bond lengths and angles for the(carbene)M(amide) complexes. compound C-M (A) M-N (A) C-M-N (˚) Torsion (˚) NC-M-NC Conformer ratio CCDC # 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 2.004(7) 1.984(8) 2.020(6) 1.999(7) 179.1(3) 1.79.5(3) 3.7 100/0 2168084 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 1.879(2) 1.854(2) 173.9(1) 6.1 100/0 2170320 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 2.005(1) 2.025(7) 176.1(4) 2.7 100/0 2182717 𝐴𝑢 𝑏𝑖𝑚 𝐶𝐴𝐴𝐶 1.974(4) 2.014(3) 178.1(2) 7.2 100/0 2155241 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 1.992(7) 2.023(6) 178.8(3) 4.0 90/10 2167514 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 1.995(3) 2.018(3) 178.6(2) 7.3 100/0 2170000 𝐴𝑢 𝑀𝑏𝑖𝑚 𝐵𝑍𝐼 2.005(7) 2.031(6) 174.1(3) 0.7 84/16 2168086 3.9 Electrochemical & Absorption Results The redox properties of the cMa complexes were examined by cyclic and differential pulse voltammetry in dimethylformamide (DMF), and the potentials relative to an internal ferrocene reference are listed in below in Table 3.4 and graphically presented in Figure 3.11. Oxidations are irreversible for all cMa complexes with Cz and bim amides. A common decomposition pathway for carbazoles upon oxidation is dimerization at the 3,6-positions. 73 Table 3.4 Cyclic voltammetry & absorption results: measurements were carried out in DMF solution with 0.1 M NBu4PF6 electrolyte, and the potentials are listed relative to a ferrocene internal reference. The absorption edge is taken as the point where ICT absorbance has dropped to 10% of the peak absorbance for a toluene solution of the cMa. Complex Eox (V) Ered (V) Eredox (V) Abs. edge (eV) 𝑪𝒖 𝑩𝑪𝒛 𝑩𝒁𝑨𝑪 0.06 -2.90 2.96 2.74 𝑨𝒖 𝑩𝑪𝒛 𝑩𝒁𝑨𝑪 0.15 -2.83 2.98 2.77 𝑨𝒖 𝑪𝒛 𝑩𝒁𝑨𝑪 0.26 -2.81 3.07 2.89 𝑪𝒖 𝑩𝑪𝒛 𝑷𝑨𝑪 0.12 -2.28 2.40 2.25 𝑨𝒈 𝑩𝑪𝒛 𝑷𝑨𝑪 0.02 -2.24 2.26 2.27 𝑨𝒖 𝑩𝑪𝒛 𝑷𝑨𝑪 0.21 -2.24 2.45 2.33 𝑨𝒖 𝑪𝒛 𝑷𝑨𝑪 0.29 -2.24 2.53 2.49 𝑨𝒖 𝑪𝒛 𝑷𝒁𝑰 0.29 -1.99 2.30 2.36 𝑨𝒖 𝒃𝒊𝒎 𝑷𝒁𝑰 0.36 -1.92 2.28 2.38 𝑨𝒖 𝒃𝒊𝒎 𝑷𝑨𝑪 0.32 -2.16 2.48 2.52 𝑨𝒖 𝒃𝒊𝒎 𝑴𝑨𝑪 0.33 -2.37 2.70 2.65 𝑨𝒖 𝒃𝒊𝒎 𝑪𝑨𝑨𝑪 0.32 -2.62 2.94 2.86 𝑨𝒖 𝒃𝒊𝒎 𝑩𝒁𝑨𝑪 0.31 -2.69 3.00 2.99 𝑨𝒖 𝒃𝒊𝒎 𝑩𝒁𝑰 0.33 -2.79 3.12 3.08 𝑨𝒖 𝑴𝒃𝒊𝒎 𝑩𝒁𝑰 0.26 -2.75 3.01 3.00 𝑨𝒖 𝑶𝒃𝒊𝒎 𝑩𝒁𝑰 0.18 -2.81 2.99 2.93 𝑨𝒖 𝒃𝒊𝒎 𝑰𝒑𝒓 0.33 * * 3.59 * The reduction potential of 𝑨𝒖 𝒃𝒊𝒎 𝑰𝒑𝒓 was outside of the solvent window. This is consistent with it having the largest optical LUMO of all compounds in this study. 74 Alkyl substitution at these sites inhibits this reaction pathway, 67 leading to reversible oxidation for the 𝐴𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 and 𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 complexes. The oxidation potentials for silver and gold derivatives of 𝑀 𝐵𝐶𝑧 𝑃𝐴𝐶 are similar; however, the copper analog is easier to oxidize. This is contrary to what was observed for the 𝑀 𝐶𝑧 𝑀𝐴𝐶 and 𝑀 𝐶𝑧 𝐶𝐴𝐴𝐶 complexes, where the oxidation and reduction potentials were unaffected by the choice of metal ion. 35 A similar trend is observed for 𝑀 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 where the oxidation potential is lower for the copper derivative than the gold analog. Addition of the t Bu groups to Cz destabilizes the oxidation potential for a given complex by ~100 mV, e.g. 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 /𝐴𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 and 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 /𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 . This effect parallels the shift observed for the energy for the longest wavelength absorption band in the two complexes, which also decreases by roughly 100 mV. The oxidation potentials for complexes with bim ligands are ca. 60 mV greater than those of the analogous Cz based complexes, leading to a similar shift in the absorption energies. The dependence of the absorption energy on the oxidation and reduction potentials of the complex is consistent with an interligand charge transfer (ICT) transition for these complexes, as seen for other cMa complexes. Reduction waves were observed for all complexes except 𝐴𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 , which falls outside the solvent window for DMF. All reductions are reversible, except for derivatives with BZAC ligands. The identity of the metal ion for the complex does not affect the reduction potential (see 𝑀 𝐵𝐶𝑧 𝑃𝐴𝐶 ). A gradual shift to more negative potentials, along with a concomitant increase in energy for the lowest absorption band, is found upon going from complexes with the PZI, PAC, MAC, CAAC, BZAC and BZI carbenes, which is again consistent with an ICT assignment for the transition. 75 Cu-BCz Au-BCz Au-Cz Cu-BCz Ag-BCz Au-BCz Au-Cz PZI PAC MAC CAAC BZAC BZI Au BZI Mbim Au BZI Obim 0.4 0.2 0.0 -2.0 -2.2 -2.4 -2.6 -2.8 -3.0 Potential vs. Fc/Fc+ 1.5 2.0 2.5 3.0 3.5 4.0 Absorption Energy (eV) Abs. Energy Reduction Oxidation Au carbene bim M PAC Cz, BCz M BZAC Cz, BCz Figure 3.11 Electrochemical redox potentials and transition energies for the 1 ICT state. The energy of the 1 ICT state (in toluene) was estimated from the onset of the absorption band where the intensity was 0.10 the value at max. 3.10 Computational Results The electronic properties of the complexes were modeled using Density Functional Theory (DFT) and Time Dependent DFT (TDDFT), details are given in section 3.6. The HOMO and LUMO energies from these calculations are listed in Table 3.6 and representative orbital iso- surfaces are shown in Figure 3.12. TDDFT methods were used to estimate the oscillator strengths for the singlet transitions, as well as vertical energies and dipole moments for the singlet and triplet excited states. In all cases but 𝐴𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 , the transitions from S0 to the S1 are ICT in nature and have > 99.5% HOMO (amide) → LUMO (carbene) character. 76 Figure 3.12 Molecular orbitals (MOs) of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 (left) and 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 (right). The HOMO is displayed with red and blue phases and the LUMO is displayed with turquoise and cream phases (isovalue = 0.1). A magnified perspective is presented to highlight contribution of the d orbital to the HOMO and LUMO. The 2,6-isopropyl groups have been removed for clarity. For the majority of the cMa complexes the T1 state is comprised of the same orbitals as the S1 state. The 𝐴𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 and 𝐴𝑢 𝐶𝑧 𝐼𝑝𝑟 68 complexes are a special case where the energy of the LUMO is sufficiently destabilized that the lowest triplet state is 3 bim and 3 Cz, respectively, rather than 3 ICT. The dipole moments for the ground, 1 ICT and 3 ICT states are large in magnitude ( = 10-20 D); however, the excited state dipoles are antiparallel to those of the ground state (Table 3.5). This feature is common in cMa complexes and is a consequence of the high degree of charge transfer in the ICT excited state (Figure 3.12). Table 3.5 Ground and excited state dipole calculations for all cMa complexes 77  - S0 (D)  - S1 (D)  - T1 (D)  (S0-S1) (D)  (S0-T1) (D) 𝐶𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 11.6 -12.4 -10.4 24 22 𝐶𝑢 𝐶𝑧 𝑀𝐴𝐶 10.1 -15.3 -13.6 25.4 23.7 𝐶𝑢 𝐶𝑧 𝑃𝐴𝐶 10.4 -16.8 -14.4 27.2 24.8 𝐴𝑔 𝐶𝑧 𝑃𝐴𝐶 11.9 -19.3 -17.7 31.2 29.6 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 10.0 -17.6 -15.2 27.6 25.2 𝐴𝑢 𝐶𝑧 𝑀𝐴𝐶 9.8 -15.9 -14.5 25.7 24.3 𝐴𝑢 𝐶𝑧 𝐶𝐴𝐴𝐶 15.3 -10.1 -8.84 25.4 24.14 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 11.4 -12.8 -10.7 24.2 22.1 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 11.2 -13.4 -10.7 24.6 21.9 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 8.2 -20.2 -17.8 28.4 26 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 10.3 -18.3 -16.8 28.6 27.1 𝐴𝑢 𝑏𝑖𝑚 𝐶𝐴𝐴𝐶 10.6 -12.7 -11.0 23.3 21.6 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 10.7 -15.0 -12.5 25.7 23.2 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 10.8 -15.8 -13.1 26.6 23.9 𝐴𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 11.0 -9.9 12.3* 20.9 -1.3 𝐻 𝐶𝑧 𝑀𝐴𝐶 11.6 -18.2 -15.8 29.8 NA 𝐿𝑖 𝐶𝑧 𝑀𝐴𝐶 11.5 -15.7 -15.4 27.2 NA The antiparallel alignment of dipoles gives rise to hypsochromic shifts in absorption and bathochromic shifts for emission in solvents with increasing polarity. The natural transition orbitals (NTOs) for the S1 state were calculated using the S0 optimized geometry and the overlap between the hole and electron wavefunctions (NTO) determined as described in the experimental section (Table 3.6). 78 Table 3.6 Computational Results of all cMa complexes. HO (eV) LU (eV) ELU-HO (eV) S1→S0 (eV/f) T1→S0 (eV) EST a (eV) NTO (S1) 𝐶𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 -4.04 -1.39 2.65 2.95/0.12 2.76 0.19 0.36 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 -4.18 -1.41 2.76 3.10/0.19 2.87 0.23 0.39 𝐶𝑢 𝐶𝑧 𝑃𝐴𝐶 -4.19 -2.20 1.99 2.43/0.13 2.21 0.22 0.36 𝐴𝑔 𝐶𝑧 𝑃𝐴𝐶 -4.07 -2.29 1.78 2.39/0.09 2.27 0.11 0.26 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 -4.35 -2.20 2.15 2.61/0.19 2.36 0.25 0.39 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 -4.38 -2.54 1.84 2.36/0.20 2.13 0.23 0.37 𝐶𝑢 𝐶𝑧 𝑀𝐴𝐶 -4.17 -1.99 2.17 2.48/0.11 2.27 0.21 0.36 𝐴𝑢 𝐶𝑧 𝑀𝐴𝐶 -4.32 -1.99 2.33 2.66/0.16 2.41 0.25 0.40 𝐴𝑢 𝐶𝑧 𝐶𝐴𝐴𝐶 -4.29 -1.60 2.68 2.86/0.16 2.56 0.30 0.43 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 -4.22 -1.46 2.76 3.09/0.20 2.87 0.22 0.40 𝐴𝑢 𝐶𝑧 𝐼𝑝𝑟 -4.16 -0.81 3.35 3.44 3.15 b 0.29 0.42 𝐴𝑢 𝑏𝑖𝑚 𝑃𝑍𝐼 -4.44 -2.58 1.85 2.45/0.16 2.29 0.16 0.32 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 -4.43 -2.21 2.22 2.75/0.17 2.56 0.19 0.35 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 -4.41 -2.0 2.41 2.81/0.14 2.63 0.18 0.35 𝐴𝑢 𝑏𝑖𝑚 𝐶𝐴𝐴𝐶 -4.39 -1.61 2.78 3.02/0.15 2.80 0.22 0.38 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 -4.29 -1.42 2.87 3.25/0.17 3.08 0.17 0.35 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 -4.30 -1.50 2.80 3.22/0.21 3.04 0.18 0.35 𝐴𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 -4.27 -0.82 3.46 3.62/0.16 3.32 b 0.30 0.36 𝐻 𝐶𝑧 𝑀𝐴𝐶 -3.99 -2.15 1.84 2.35/0.001 2.34 0.01 0.06 𝐿𝑖 𝐶𝑧 𝑀𝐴𝐶 -3.83 -2.08 1.75 2.17/0.03 2.11 0.06 0.18 The transition energies in the series of Group 11 metals in 𝑀 𝐶𝑧 𝑃𝐴𝐶 follow the same trends as reported previously for 𝑀 𝐶𝑧 𝐶𝐴𝐴𝐶 and 𝑀 𝐶𝑧 𝑀𝐴𝐶 . 35 The energies for the S1 and T1 states are independent of the metal center, whereas values for EST fall in the order Au > Cu > Ag, which mirror NTO values of 0.39, 0.36 and 0.26, respectively. The interligand carbene CꞏꞏꞏN distances for the three complexes 79 fall in the order Ag > Au > Cu. A long interligand distance might be expected to give rise to a small NTO, but the higher value for the Au complex suggests a greater participation of the metal ion in the excited state than for the Cu and Ag complexes. Examining the data for the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes it is apparent that the carbene ligand markedly affects the S1 and T1 energies, whereas the EST, NTO and oscillator strengths for the S1 state are only moderately altered by the nature of this ligand. Small differences in EST and NTO among the complexes with various carbene ligands suggests that the stabilization imparted on the carbene −system by addition of electron withdrawing carbonyl groups (MAC and PAC) or benzannulated arene rings (BZI, PZI, BZAC and PAC) does not markedly alter overlap between the electron and hole wavefunctions. In contrast, both EST and NTO decrease upon shifting from a Cz to a bim donor, despite there being only a minor increase in energy for the S1 and T1 states. The lowest energy excited state in these cMa complexes is best characterized as interligand, not metal-to-ligand, charge-transfer in character. In the cMa complexes the metal ion contributes equally to both the HOMO and LUMO in the ICT state, so there is no net charge transfer between the metal and the ligands. However, the metal ion is still an important participant in these transitions. This can be seen in comparing the modeling data for the complexes in Table 1 to data for 𝐻 𝐶𝑧 𝑀𝐴𝐶 and 𝐿 𝑖 𝐶𝑧 𝑀𝐴𝐶 . The separation between the central ion (neither of which have accessible d-orbitals available for bonding) and the ligands was kept at the same distance as for the copper ion in the geometry optimized structure of 𝐶𝑢 𝐶𝑧 𝑀𝐴𝐶 . Predictably, the values for EST and NTO decrease precipitously in these two complexes, illustrating the contribution of the metal ion to the valence orbitals of the cMa complexes. 80 3.11 Photophysical Properties at Room Temperature The UV-visible absorption spectra of the complexes were recorded in toluene solution. Spectra for the carbazole-based cMa complexes with BZAC and PAC are shown in Figure 3.13a. The absorption spectra of the 𝑀 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 and 𝑀 𝐵𝐶𝑧 𝑃𝐴𝐶 complexes display bands at high energy ( < 325 nm) that are assigned to transitions localized on the carbene and carbazolide ligands. Structured bands at 375 nm are assigned transitions on the carbazolide ligand, whereas bands assigned to absorption from the 1 ICT state are centered at 410 and 480 nm for the BZAC and PAC complexes, respectively. As observed for other cMa complexes, 35 values for the extinction coefficients of the ICT bands fall in the order Au > Cu > Ag. Changing the methylene moiety in BZAC to carbonyl in PAC stabilizes the LUMO and leads to a marked red shift for complexes with the PAC ligand. The absorption energies of the ICT state vary depending on the carbene ligands used here, which can best be seen by comparing spectra of the Au complexes using the bim donor (𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 ), Figure 3.13c because the LE band on bim is well separated from the ICT band. However, this trend is also observed in the Cz analogues which further supports that the carbene ligand can tune the ICT absorption independent of the donor ligand (Figure 3.14). In these complexes the absorption transitions localized on the amide (bim) ligand that appear at 310 nm ensures minimal overlap with the ICT band of the complexes. An exception is in 𝐴 𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 where the LUMO is destabilized sufficiently to raise the energy of the ICT transition to be comparable to that of the bim ligand. For the other complexes, the ICT bands are distinct and have energies that fall in the order BZI > BZAC > CAAC > MAC > PAC > PZI. 81 300 350 400 450 500 550 600 0 5,000 10,000 15,000 20,000 300 350 400 450 500 0.0 0.5 1.0 1.5 Au BZAC Cz PS Au BZAC Cz Toluene Extinction Coeff. (M -1 cm -1 ) Wavelength (nm) Au BZAC BCz Cu BZAC BCz Au PAC BCz Cu PAC BCz Ag PAC BCz (a) 400 500 600 700 800 0.0 0.5 1.0 1.5 2.0 Photoluminescnce (arb. units) Wavelength (nm) PAC-toluene BZAC-toluene PAC- PS BZAC PS 300K 77K (b) Au carbene BCz 300 350 400 450 500 550 0 5,000 10,000 15,000 20,000 25,000 30,000 Extinction Coeff. (M -1 cm -1 ) Wavelength (nm) Au PZI bim Au PAC bim Au MAC bim Au CAAC bim Au BZAC bim Au BZI bim Au Ipr bim (c) 400 500 600 700 0.0 0.5 1.0 1.5 2.0 Photoluminescence (arb. units) Wavelegth (nm) 300 K 77 K (d) Figure 3.13 BZAC/PAC extinction (a) and emission (b) in toluene and polystyrene. Inset shows the spectra of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 in toluene and polystyrene (1 wt %), normalized at the Cz absorbance. Extinction spectra in toluene (c) and emission spectra in polystyrene (d) of 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes. 350 400 450 500 550 0.0 0.5 1.0 Au BZI Cz Au BZAC Cz Au CAAC Cz Au MAC Cz Au PAC Cz Au PZI Cz Absorbance AU (AU) Wavelength (nm) (e) Figure 3.14𝐴𝑢 𝐶𝑧 𝐶𝑎𝑟𝑏𝑒𝑛𝑒 absorption spectra in toluene. 82 The absorption spectra of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 in toluene, overlaid with that in polystyrene at 1 wt% loading and normalized at 375 nm, is shown in the inset to Figure 3.13a. The close match in absorption profiles confirm that the two media have similar solvation properties. However, the polystyrene matrix is expected to hinder geometric rearrangement, particularly rotation around the metal-ligand bond axis. Luminescence spectra for all cMa complexes were thus recorded in toluene and polystyrene (Figure 3.13 (b, d). The spectra are broad and featureless at room temperature, indicative of emission from an ICT state. Spectra for complexes measured in toluene are slightly red shifted relative to the same compounds in polystyrene (Figure 3.13b), suggesting that only minor structural changes take place in the excited state; however, the shift is greater for the complexes with the bim donor. This difference in the rigidochromic effects in polystyrene suggests that the bim-based cMa complexes undergo a greater structural distortion in their ICT state than do the carbazole-based materials. The emission energies from the cMa complexes parallel the trend observed for the absorption energies of the ICT state (Figure 3.13). Compounds ligated with the electrophilic PZI and PAC ligands emit orange-red whereas those with BZI and BZAC luminesce in the blue spectral region. For Au derivatives sharing a common carbene, the spectra with BCz are red-shifted relative to their Cz analogs, whereas the bim congeners emit at higher energies. Addition of methyl or methoxy groups to the bim ligand destabilizes the HOMO and leads to a concomitant red shift in emission relative to the parent 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 complex (Figure 3.15). 83 300 350 400 450 500 550 600 0 5,000 10,000 15,000 20,000 25,000 Extinction Coeff. (M -1 cm -1 ) Wavelength (nm) Au BZI bim Au BZI Mbim Au BZI Obim 0.0 0.5 1.0 1.5 2.0 normalized Intensity (arb. units) Figure 3.15 Absorption and emission spectra for (BZI)Au(amide), amide = bim, Mbim and Obim. Extinction spectra are recorded in toluene solution and emission spectra in 1% doped polystyrene thin films. The 𝑀 𝐵𝐶𝑧 𝑃𝐴𝐶 complexes display only minor variation in emission energy with respect to the identity of the metal center. The luminescence properties 𝐴𝑢 𝑏𝑖 𝑚 𝐼𝑝𝑟 are consistent with assignment to a combined ICT/ligand-triplet transition, as observed for 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 , 2 leading to a slow radiative rate relative to the ICT emitters. 84 Table 3.7 Photophysical parameters for cMa complexes in toluene (tol) solution and polystyrene (PS) thin film (1% by weight). max (PL) (nm) PL (%)  ( μs) kr (10 6 s -1 ) knr (10 6 s -1 ) tol PS tol PS tol PS tol PS tol PS 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 502 459 >95 93 0.71 1.24 a 1.4 -- 0.04 -- 𝐴𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 500 484 >95 >95 0.56 0.72 1.7 1.4 0.07 <0.01 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 595 594 47 75 0.56 0.95 0.84 0.79 0.95 0.26 𝐴𝑔 𝐵𝐶𝑧 𝑃𝐴𝐶 610 588 10 51 0.58 0.26 0.17 1.0 1.5 0.33 𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 588 586 42 76 0.44 0.74 0.95 1.0 1.3 0.33 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 600 570 62 92 0.41 0.45 1.5 2.0 0.94 0.22 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 546 546 73 >95 0.74 0.81 1.0 1.2 0.36 0.04 𝐴𝑢 𝐶𝑧 𝑀𝐴𝐶 ref 35 -- 508 -- 85 -- 0.8 -- 1.0 -- 0.18 𝐴𝑢 𝐶𝑧 𝐶𝐴𝐴𝐶 b , ref 35 -- 472 -- >95 -- 1.14 -- 0.88 -- < 0.01 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 480 479 >95 >95 0.69 1.98 a 1.5 -- <0.01 -- 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 ref 2 448 432 94 >95 1.11 2.28 a 0.85 -- 0.05 -- 𝐴𝑢 𝑏𝑖𝑚 𝑃𝑍𝐼 600 552 31 91 0.21 0.24 1.5 3.8 3.2 0.37 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 562 532 30 81 0.17 0.27 1.8 3.0 4.2 0.7 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 548 506 19 88 0.17 0.40 1.1 2.2 4.8 0.3 𝐴𝑢 𝑏𝑖𝑚 𝐶𝐴𝐴𝐶 514 476 87 >95 0.63 0.55 1.4 1.8 0.02 <0.01 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 484 452 >95 >95 0.43 0.28 2.3 3.7 <0.01 <0.01 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 454 429 >95 >95 0.42 0.25 2.3 4.0 <0.01 <0.01 𝐴𝑢 𝑏𝑖𝑚 𝐼𝑝𝑟 340 400 <1 <5 c 13 c 0.004 c c 𝐴𝑢 𝑀𝑏𝑖𝑚 𝐵𝑍𝐼 460 436 >95 >95 0.42 0.29 2.4 3.4 <0.01 <0.01 𝐴𝑢 𝑂𝑏𝑖𝑚 𝐵𝑍𝐼 480 450 92 >95 0.37 0.38 a 2.5 -- 0.2 -- a The lifetime given is the weighted average of a biexponential fit; 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 : 0.79 (73%), 2.47 (27%); 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 : 0.84 (57%), 3.40 (43%); 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 : 0.74 (46%); 3.6 (54%); 𝐴𝑢 𝑂𝑏𝑖𝑚 𝐵𝑍𝐼 : 0.25 (80%), 0.90 (20%). b The CAAC ligand on this complex has a menthyl instead of the adamantyl group. c Values cannot be accurately determined. 85 Cooling solutions of the 𝑀 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 complexes to 77 K leads to a marked change in the luminescence spectrum with a structured band appearing to the blue of the room temperature spectrum (Figure 3.13b). This rigidochromic transformation upon freezing the solvent has previously been shown to come about from destabilization of the ICT state to an energy higher than that of the triplet state localized on the carbazolide ligand ( 3 Cz). 2, 34, 35 The ICT state is destabilized to a lesser extent in a rigid polystyrene matrix such that it remains the lowest energy state upon cooling to 77 K. Cooling both toluene and polystyrene samples of 𝑀 𝐵𝐶𝑧 𝑃𝐴𝐶 leads to a blue shift of the ICT band, but this change in energy is not large enough to access the 3 Cz state. In contrast, the energy of the triplet state for bim (ET = 365 nm) is much higher than that of carbazole (ET = 415 nm) (Figure 3.16). 220 240 260 280 300 320 340 360 0.0 0.2 0.4 0.6 0.8 1.0 Normalized Absorbance (arb.U.) Wavelength (nm) BimH MeBimH MeOBimH (a) 300 350 400 450 500 0.0 0.2 0.4 0.6 0.8 1.0 Normalized Intensity (arb. Units) Wavelength (nm) HBim em HBim ex MeBim em MeBim ex MeOBim em MeOBim ex (b) Figure 3.16 The absorption spectrum of bim precursors in toluene (a). The excitation and emission spectra of bim in MeTHF at 77K(b). The ligand localized triplet states of the carbene ligands are also in the UV region so the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes emit from largely featureless ICT transitions with Gaussian line shapes at all temperatures (Figure 3.13d). These cMa complexes also observe solvent dependent emission shown below with PAC and BZAC compounds in Figure 3.17. 86 450 500 550 600 650 700 750 800 850 0.0 0.2 0.4 0.6 0.8 1.0 Normalized Emission Wavelength (nm) DCM MeTHF Tol 0.5 wt% PS MeCyc Cu PAC BCz 350 400 450 500 550 600 650 700 750 0.0 0.2 0.4 0.6 0.8 1.0 Emission Intensity (A.U.) Wavelength (nm) DCM MeTHF Tol 0.5 wt% PS MeCyc Cu BZAC BCz 450 500 550 600 650 700 750 800 850 900 0.0 0.2 0.4 0.6 0.8 1.0 Emission Intensity (Counts) Wavelength (nm) DCM MeTHF Tol 0.5 wt% PS MeCyc Ag PAC BCz 450 500 550 600 650 700 750 800 850 0.0 0.2 0.4 0.6 0.8 1.0 Emission Intensity (A.U.) Wavelength (nm) DCM Tol MeTHF MeCyc 0.5 wt% PS Au PAC BCz 350 400 450 500 550 600 650 700 0.0 0.2 0.4 0.6 0.8 1.0 Normalized Emission Intensity (A.U.) Wavelength (nm) DCM MeTHF Toluene 0.5wt%, PS MeCyclohex Au BZAC Cz 350 400 450 500 550 600 650 700 750 800 0.0 0.2 0.4 0.6 0.8 1.0 Emission Intensity (A.U.) Wavelength (nm) DCM MeTHF Tol 0.5 wt% PS MeCyc Au BZAC BCz 400 450 500 550 600 650 700 750 800 850 0.0 0.2 0.4 0.6 0.8 1.0 Emission (AU) Wavelength (nm) DCM MeTHF Tol 0.5wt%, PS MeCyclohex Au PAC Cz Figure 3.17 Solvent dependent emission of 𝑀 𝐴𝑚𝑖𝑑𝑒 𝐶𝑎𝑟𝑏𝑒𝑛𝑒 in DCM, MeTHF, Tol, PS film, and MeCyclohexane 87 The photoluminescence efficiency for several of the cMA complexes are high (PL > 0.95) in both fluid toluene and rigid polystyrene, and all have microsecond to sub-microsecond emission lifetimes. The Ag complex has a low PL in solution which may be related to photodecomposition of this derivative. Lower values for PL in toluene solution owe largely to increased nonradiative rates in fluid solutions, likely ligand rotation and/or excimer/exciplex formation. The radiative rates for the bim based cMa complexes are 1.8-4 x 10 6 s -1 , which are some of the highest values reported for TADF emitters and lead to radiative lifetimes as fast as 250 ns. In the cases where an analogous carbazole based complex is available for comparison, the radiative rate for the bim based cMa complexes are two-fold faster than the Cz analogs. Unfortunately, the near degeneracy in energy between the ICT and 3 Cz states in 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 and 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 leads to non-first order behavior in the luminescence decay traces for these derivatives precluding direct comparisons with the bim analogs. 3.12 Temperature Dependent Photophysics As discussed in the introduction, the radiative rate of TADF from luminophores with effective spin orbit coupling (SOC) is controlled by the radiative rate from the S 1 state and the equilibrium constant between the T1 and S1 states (Eq. 3.4). For 𝑀 𝐶𝑧 𝑀𝐴𝐶 and 𝑀 𝐶𝑧 𝐶𝐴𝐴𝐶 complexes, the principal factor that leads to faster TADF rates for the silver complexes over the copper and gold analogs is that the silver complex has larger equilibrium constant, owing to its smaller E ST (Eq. 3.5). The modeling presented above suggests that the EST values for bim based cMa complexes should be lower than their carbazolide counterparts, which may account for their faster TADF rates. To validate this conjecture, experimental temperature dependent photophysical measurements were performed between 4-300 K to determine the two parameters that control 𝑘 𝑟 TADF (i.e., 𝑘 𝑟 S 1 and EST). The temperature dependant photophysics was experimentally led by 88 Jonas Schaab (Section 3.7). These measurements were conducted using polystyrene thin films doped at ~1 wt% with the cMa complex. Fitting the temperature dependent lifetimes to a three- level model gives values for EST, the zero-field splitting (ZFS) and the radiative rate from the S1 and T1 states. 11, 12, 35, 36 The high PL in these cMa complexes allows us to neglect nonradiative processes in our model. In this analysis, the ZFS is the energy difference between the two closely spaced triplet sublevels and the highest energy triplet sublevel. The energy spacing between the two lowest sublevels could not be determined from data obtained at temperatures down to 4 K (which is the limit of our cryogenic system) using Eq. 3.12. The fits to the data are shown below in Table 3.8 for each complex, and the energy and rate data derived from the fits is given in Table 3.8. 𝜏 = 2 + 𝑒 −∆𝐸 (𝐼𝐼𝐼 −𝐼 ,𝐼𝐼 ) 𝑘 𝐵 𝑇 + 𝑒 −∆𝐸 (𝑆 1 −𝐼 ,𝐼𝐼 ) 𝑘 𝐵 𝑇 2( 1 𝑘 𝐼 ,𝐼𝐼 ) + ( 1 𝑘 𝐼𝐼𝐼 )𝑒 −∆𝐸 (𝐼𝐼𝐼 −𝐼 ,𝐼𝐼 ) 𝑘 𝐵 𝑇 + ( 1 𝑘 𝑓𝑙 )𝑒 −∆𝐸 (𝑆 1 −𝐼 ,𝐼𝐼 ) 𝑘 𝐵 𝑇 𝐸𝑞 .3.12 89 Table 3.8 Photophysical parameters for cMa complexes in toluene (tol) solution and polystyrene (PS) thin film (1% by weight) extracted from temperature dependent measurements. ZFS (meV / cm -1 )  10% EST (meV / cm -1 )  3% 𝜏 𝑆 1 (ns)  9% 𝜏 𝑇 1 (s)  5% 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 a 65 / 520 18 a 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 b 70 / 570 18 b 𝐴𝑔 𝐵𝐶𝑧 𝑃𝐴𝐶 b 26 / 210 35 b 𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 0.9 / 7 72 / 580 14 36 𝐴𝑢 𝐶𝑧 𝑀𝐴𝐶 1.2 / 10 87 / 700 13 28 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 0.9 / 7 54 / 440 21 82 𝐴𝑢 𝑏𝑖𝑚 𝑃𝑍𝐼 1.1 / 9 31 / 250 26 76 𝐴𝑢 𝑏𝑖𝑚 𝑃𝐴𝐶 1.0 / 8 45 / 360 17 38 𝐴𝑢 𝑏𝑖𝑚 𝑀𝐴𝐶 2.0 / 16 51 / 410 14 38 𝐴𝑢 𝑏𝑖𝑚 𝐶𝐴𝐴𝐶 1.1 / 9 53 / 430 18 16 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 1.2 / 10 41 / 330 19 19 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 1.5 / 12 41 / 330 15 19 𝐴𝑢 𝑀𝑏𝑖𝑚 𝐵𝑍𝐼 1.5 / 12 44 / 350 16 11 𝐴𝑢 𝑂𝑏𝑖𝑚 𝐵𝑍𝐼 1.2 / 10 41 / 330 19 17 a The lifetime given is the weighted average of a biexponential fit; 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 : 0.79 (73%), 2.47 (27%); 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 : 0.84 (57%), 3.40 (43%); 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 : 0.74 (46%); 3.6 (54%); 𝐴𝑢 𝑂𝑏𝑖𝑚 𝐵𝑍𝐼 : 0.25 (80%), 0.90 (20%). b Values could not be accurately determined. The 𝑀 𝐵𝐶𝑧 𝑃𝐴𝐶 complexes show the same trends in TADF parameters observed previously for the MAC and CAAC analogs. 35 The copper and gold complexes give similar values for EST and S1 radiative rates, but the silver complex gives a smaller EST and slower 𝑘 𝑟 S 1 , leading to comparable rates of TADF for complexes with the three different metals. The 𝑀 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes show similar values for 𝑘 𝑟 S 1 to their carbazole analogs, but the EST values for the bim based cMa complexes are uniformly lower than those of the analogous carbazole based materials by 30% or more. The fast 𝑘 𝑟 TADF for the bim based cMa complexes is thus due largely to a significant increase 90 in Keq caused by the small EST. The origin of this decrease in EST for the bim complexes will be discussed in the following discussion section, but there is a clear connection to the smaller NTO for the bim based complexes. The values of ZFS for the 𝑀 𝐶𝑧 𝑀𝐴𝐶 and 𝑀 𝐶𝑧 𝐶𝐴𝐴𝐶 complexes in Table 3.8 are much lower than values presented in previous reports. 35, 36 Our earlier measurements used a cryostat that had poor thermal regulation below 80 K, which led to incorrect data collected at the low temperatures needed to extract the ZFS parameters. The present results were collected using a cryogenic system that is more reliable, providing reproducible data on repeated heating and cooling cycles. Thus, the ZFS values reported here have high accuracy. 3.13 Discussion The ICT transition in the cMa complexes is essentially an electron transfer from the amide group to the carbene. While the transition utilizes the full spatial extent of the HOMO and LUMO (Figure 3.12), one can approximate this process as being a charge transfer from the nitrogen lone pair of the amide to the vacant p-orbital of the N-heterocyclic carbene. With this line of reasoning, it is apparent that if the amide is kept constant, the energy of the vacant p-orbital on the carbene will determine the energy of the ICT state. The energy of the LUMO in the carbene ligands chosen for the present study span a range of values (Table 3.6, Figure 3.11). The LUMO energy in N-heterocyclic carbene ligands is destabilized by electron donation from the nitrogen(s) adjacent 91 Figure 3.18 (a) The ICT transition and the nature of the interaction between the N lone pairs and carbene p-orbital of the N-heterocycle carbene (NHC) for the cMa complexes are illustrated. In this representation the plane of the NHC ligand is perpendicular to the page. (b) LUMO orbitals are shown for 𝑀 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛 𝑒 complexes. The dipp groups are only shown for the IPr and CAAC complexes. In the other complexes the dipp groups are not involved in the LUMO. The LUMO energies (in eV) from DFT calculations are given below the acronym of each ligand. to the p-orbital of the carbene [Figure 3.18a]. Orbital contours from DFT calculations illustrate the antibonding nature of the N-C carbene bond in the LUMO [Figure 3.18 (b)], consistent with the energy diagram presented in Figure 3.18a. The carbene p-orbital will be destabilized by greater the participation of the nitrogen(s) in this MO, thus raising the transition energy of the ICT state. The CAAC ligand has only a single N-atom, whereas the others have two N-atoms destabilizing the carbene p-orbital, which explains the significant stabilization of the CAAC LUMO relative to that of the closest analog (Ipr). The benzannulated phenyl ring of BZI accepts electron density from the imidazolium nitrogen atoms and leads to less mixing of the N lone pair with the carbene p-orbital, stabilizing the LUMO relative to Ipr. Nitrogen substitution into the arene ring of BZI (forming PZI) leads to even greater interaction with the imidazolium nitrogens, stabilizing the carbene LUMO further. The BZAC ligand would be expected to have a LUMO between that of Ipr and BZI, since only one nitrogen is attenuated by the benzannulation, however, the ring expansion from a five- to a six-membered ring stabilizes the carbene p-orbital, 69 leading to similar 92 LUMO energies for BZI and BZAC. The carbonyl groups of MAC and PAC compete effectively for the nitrogen lone pair, therefore the LUMO energies for these two carbenes are lower than for BZAC. It is evident upon comparison of BZAC and MAC that the carbonyl group leads to a much greater stabilization of the LUMO than does benzannulation. The lowest LUMO energy is for PAC since the -system is stabilized from both benzannulation and the carbonyl group. An important set of observations that deserves further discussion are the short lifetimes for the cMa complexes coordinated to the bim ligand. In all cases, the radiative rate for the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complex is a factor of two faster that the analogous 𝐴𝑢 𝐶𝑧 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complex. This difference occurs despite both derivatives having similar excited state energies and extinction coefficients. Based on modeling studies (Table 3.6), the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes have lower NTO and EST values than their 𝐴𝑢 𝐶𝑧 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 counterparts (excluding the Ipr complex, which emits form a 3 bim state). The latter trend is mirrored in the data from variable temperature photophysical measurements that show the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes do indeed have lower EST values than their 𝐴𝑢 𝐶𝑧 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 analogs (Table 3.8). Thus, since values of 𝑘 𝑟 S 1 are similar for complexes with both types of amide, the principal source of the fast 𝑘 𝑟 TADF values for the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes are the low values for EST. A key difference between Cz and bim in the cMa complexes is the location of the center of positive charge in the ICT excited states of the two amides (Section 3.6, Figure 3.8). The positions calculated for the electron and hole for the 𝑀 𝑎𝑚𝑖𝑑𝑒 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes discussed here are illustrated in Figure 3.19. The two charges lie within the plane of the carbene and amide ligand, respectively. For Cz the center of positive charge lies near the N atom, whereas in 𝑀 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes the three nitrogen atoms of the core guanidinium unit disperse the positive charge and shift the center away from the nitrogen atom bound to the metal. The position of the electron in each carbene for the 93 ICT excited state is nearly independent on the choice of Cz or bim, and similarly the position of the hole for each amide is unaffected by the choice of carbene. The center of electron charge resides within the C-M bond for Ipr and CAAC, whereas the charge shifts into the carbene ligand for BZI, BZAC, MAC, PAC and PZI, see Figure 3.19a. 1x10 6 2x10 6 3x10 6 4x10 6 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 RED amide = bim BLACK amide = Cz PAC MAC CAAC BZAC BZI PZI d(h + , e - ) (Å) k TADF r (s -1 ) Au carbene amide (b) Figure 3.19 (a) The centers of negative charge (blue spheres) and positive charge (red spheres) are shown on the molecular frame for the ICT excited states. Ar = 2,6-diisopropylphenyl and Ad = adamantyl. (b) The rate of TADF emission (𝑘 𝑟 𝑇𝐴𝐷𝐹 ) at room temperature for a doped polystyrene film is plotted as a function of the hole/electron separation distance. This is consistent with the enhanced delocalization in the LUMO by -extending the carbene backbone. The distance between the centers of positive and negative charge, d(h + , e - ), for each cMa is given in Table 3.9. Figure 3.19b shows the relationship of d(h + , e - ) for 𝑀 𝑎𝑚𝑖𝑑𝑒 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes to 𝑘 𝑟 𝑇𝐴𝐷𝐹 . Values for d(h + , e - ) are spaced further apart for the bim complexes than the Cz-based ones. 94 Table 3.9 Hole/electron separation distances for (carbene)Au(amide) complexes. 𝐴𝑢 𝑎𝑚𝑖𝑑𝑒 𝐶𝑎𝑟𝑏𝑒𝑛𝑒 d(h + ,e - ) (Å) Cz bim Ipr 3.94 4.31 BZI 5.15 5.51 BZAC 5.08 5.36 CAAC 4.43 4.74 MAC 5.16 5.47 PAC 5.55 5.86 PZI 6.54 6.96 Radiative rates for both bim and Cz complexes show a dependence on d(h + , e - ), with larger separations leading to faster 𝑘 𝑟 𝑇𝐴𝐷𝐹 (Figure 3.19b); however, the degree to which 𝑘 𝑟 𝑇𝐴𝐷𝐹 increases with d(h + , e - ) differ for the two derivatives. In all cases where data are available for a given carbene, the 𝑘 𝑟 𝑇𝐴𝐷𝐹 values are markedly faster for the bim-based cMa complexes than for the Cz- based analogs (see dashed lines in Figure 3.19b). The 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 and 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 complexes were not included in Figure 3.19b because the triplet excited states for these two compounds have a mixed ICT/ligand-localized character. 2 The energies of the ICT and 3 Cz states are nearly degenerate in both of these complexes, which promotes mixing of the states and a concomitant increase in the measured lifetime, so one cannot unequivocally determine 𝑘 𝑟 𝑇𝐴𝐷𝐹 . Although the energy separation between the 3 ICT and ligand localized triplet states for the 𝑀 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 and 𝑀 𝐶𝑧 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes with PAC, MAC, CAAC and PZI carbenes is large enough to prevent mixing between these excited states, the bim complexes still display a marked increase in the radiative rate relative to their carbazole analogs. The large d(h + , e - ) separations for the bim complexes are consistent with lower NTO values, and thus the lower EST values, for the bim complexes relative to their carbazole counterparts. At first glance an increase in d(h + , e - ) might also be expected to decrease 𝑘 𝑟 𝑆 1 , as 95 observed upon substituting Ag + (with a large ionic radius) in place of Cu + and Au + for the 𝑀 𝐵 𝐶𝑧 𝑃𝐴𝐶 complexes, as well as for the 𝑀 𝐶𝑧 𝑀𝐴𝐶 and 𝑀 𝐶𝑧 𝐶𝐴𝐴𝐶 derivatives. 35 However, this is not the case for 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes since 𝑘 𝑟 𝑆 1 remains nearly constant (in the 15-20 ns range) for complexes with either carbazole or bim ligands (Table 3.8). Thus, the decrease in EST brought about by the enhanced d(h + , e - ) is the principal factor leading to the fast 𝑘 𝑟 TADF values for the 𝐴𝑢 𝑏𝑖𝑚 𝑐𝑎𝑟𝑏𝑒𝑛𝑒 complexes. Interestingly, considering all of the complexes other than 𝐴𝑢 𝑏𝑖𝑚 𝑃𝑍𝐼 , values for 𝑘 𝑟 𝑆 1 are near constant and do not show an increase by the cube of the emission energy as predicted by the Einstein relationship, 70 even though the complexes span a range of emission energies from 2.2- 2.9 eV. It is also noteworthy that even though 𝐴𝑢 𝑏𝑖𝑚 𝑃𝑍𝐼 has a substantially larger d(h + , e - ) than 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐼 and 𝐴𝑢 𝑏𝑖𝑚 𝐵𝑍𝐴𝐶 , all three complexes have comparable rates for 𝑘 𝑟 𝑇𝐴𝐷𝐹 . For the PZI based cMa complexes, low values for EST are compensated by S1 lifetimes that are slower than the other Au(I) derivatives. This suggests that the d(h + , e - ) value observed for 𝐴𝑢 𝑏𝑖𝑚 𝑃𝑍𝐼 (6.8 Å) is approaching a limit for enhancing 𝑘 𝑟 𝑇𝐴𝐷𝐹 , and we therefore anticipate that any further increase in d(h + , e - ) will continue to increase the radiative lifetime for the S1 state. 3.14 Future Work In this chapter, we demonstrated an empirical relationship between kr(TADF) and the separation of the hole and electron wavefunctions for two-coordinate cMas. The trend is simple; the further the hole and electron are “pulled” from one another, the faster the TADF rate ends up being (Figure 3.19). One can think of this as a photon sling shot. The underlying mechanism is that ST decreases as d(h + ,e - ) increases, which optimizes Keq. However, larger d(h + ,e - ) also decreases 𝑘 𝑟 S 1 for a given donor as predicted by Eq. 3.3 and is experimentally evident in Table 3.8. Despite this tradeoff, an important observation of Figure 3.19 is that the TADF rate does not appear to 96 plateau with respect to the hole-electron separation for a given donor. Thus, the complexes studied in Chapter 3 fall into a d(h + ,e - ) regime where Keq is dominating kr(TADF). However, there must be a limit as d(h + ,e - ) → ∞ at which the drop in 𝑘 𝑟 S 1 dominates the increase in Keq resulting in a plateau of kr(TADF). The natural question is: if we continue to design molecules with larger hole- electron separation, will we observe the plataue in kr(TADF)? The PZI carbene has the furthest electron localization from the metal (Figure 3.19), which makes it a great candidate for testing this hypothesis. A key feature of PZI is aza-substitution of the carbene backbone which appears to efficiently pull electron density away from the carbene carbon p orbital. Following the observation of the effect of aza-substitution, several other carbenes have been computationally explored and the results are demonstrated below. Figure 3.20 Hole-electron separation calculations for a series of substituted BZAC-Au-Cz complexes. The electron is visualized on the chemdraw based on the calculated <re-> coordinates from QChem. The computations were performed at the same level of theory as the preceding hole-electron calculations. As one can see, aza-substitution on BZAC analogues has a direct pull effect on the electron distribution. Going from 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 to 𝐴𝑢 𝐶𝑧 𝑃 ℎ−𝐵𝑍𝐴𝐶 shows a modest influence of (d = 0.07 Å) on d(h + ,e - ). However, 𝐴𝑢 𝐶𝑧 𝑝𝑦 −𝐵𝑍𝐴𝐶 has a notable change of 0.63 Å with respect to 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 . This further confirms that -extension of the electron wavefunction is more efficiently achieved by 97 functionalizing with groups that have nitrogen heteroatoms rather than using purely carbon-based functional groups. This is similar to what was observed with PZI in comparison to BZI, BZAC, or MAC. While these proposed structures increase d(h + ,e - ) from the acceptor side, this problem can also be addressed by designing donors that pull the hole further from the metal. A key difference between carbazole and bim is the number of conjugated nitrogen atoms in the pi-system. The hole density of 𝐴𝑢 𝐵𝑖𝑚 𝐶𝑎𝑟𝑏𝑒𝑛𝑒 vs. 𝐴𝑢 𝐶𝑧 𝐶𝑎𝑟𝑏𝑒𝑛𝑒 shows predominant localization over the imine nitrogen p-orbitals (Figure 3.21). This suggests that aza-substitution may also be an effective strategy for “pulling” the hole density further away from the metal center. Figure 3.21 The HOMO of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐼 (left) and 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 (right) shown with the iso-value set to 0.35. Another interesting detail of Figure 3.19 is that there appear to be two outlier points in the kr(TADF) vs. d(h + ,e - ) plot, namely 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 and 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 . These complexes have the highest kr(TADF) despite them not having the largest hole-electron separation. The carbazole analogues of these complexes could not be plotted because cryo-measurements of Keq and 𝑘 𝑟 𝑆 1 resulted in bi-exponential decay due to the relatively lower laying carbazole 3 LE. Most of the complexes with large d(h + ,e - ) in Figure 3.19 are green, yellow, or orange emitters. 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 and 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 happen to 98 be the only complexes in Figure 3.19 that emit blue photons. S. J. Strickler and Robert A. Berg showed that kr has a cubic dependance on the transition energy as shown in Eq. 13 𝑘 𝑟 = 𝐶 ⟨𝑣̃ 𝑎𝑣𝑒 ⟩ 3 ∫𝘀 𝑑𝑙𝑛 𝑣̃ 𝐸𝑞 .13 Where C is a constant, and “𝑣̃ 𝑎𝑣 𝑒 ” is the frequency of the transition in wavenumbers, and the integral is over the molar absorptivity spectrum over the energy axis. Thus, it is reasonable to suspect that 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 and 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 are outliers due to their relatively large emission 𝑣̃ 𝑎𝑣𝑒 . Previous work has shown that the ICT absorption and emission 𝑀 𝐶𝑧 𝐶𝑎𝑟𝑏𝑒𝑛𝑒 can be blue shifted by cyano substitutions in the 3 and 6 position. 35, 36 Interestingly, this does not have a significant impact on the location of the hole because the CN groups do not efficiently mix with the  system of the HOMO (Figure 3.22). Figure 3.22 State analysis of 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 and 𝐴𝑢 𝐶𝑧𝐶𝑁 2 𝑃𝑍𝐼 . Thus, CN substitutions can be strategically incorporated into molecular design to significantly modify the emission energy with lesser impact on the hole-electron separation. In fact, CN substitution in the case of 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 results in a slight decrease in d(h + ,e - ) which should yield a slower kr(TADF) than the parent complex. Thus, any experimental improvement in kr(TADF) upon 3,6- 99 CN substitution would strongly suggest that emission energy does play a significant role in kr(TADF). The role of emission energy in kr(TADF) for cMas is still under investigation, but preliminary results were achieved by synthesizing 𝐶𝑢 𝐶𝑧𝐶𝑁 2 𝑃𝐴𝐶 . The results were compared to 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 because 𝐶𝑢 𝐶𝑧 𝑃𝐴𝐶 has not been made yet. The PL spectra are compared below in Figure 3.23. 350 400 450 500 550 600 650 700 Intensity Wavelength (nm) Cu PAC CzCN2 Abs Cu PAC CzCN2 em Au PAC Cz Abs Au PAC Cz em Figure 3.23 PL spectra of 𝐶𝑢 𝐶𝑧𝐶𝑁 2 𝑃𝐴𝐶 and 𝐴𝑢 𝐶 𝑧 𝑃𝐴𝐶 in toluene. The blue shift in absorption and emission energy is consistent with published reports of cyano- substituted carbazole donors for cMas, which is also computationally evident in Figure 3.22. 36 While the absorption and emission spectra do not observe significant shifting between Cu and Au analogues, Au complexes typically have faster kr than their copper analogues by a factor of ~ 1.3. The kr(TADF) of 𝐶𝑢 𝐶𝑧𝐶𝑁 2 𝑃𝐴𝐶 was measured to be 1.4 × 10 6 s -1 . Thus, 𝐴𝑢 𝐶𝑧𝐶𝑁 2 𝑃𝐴𝐶 is predicted to have kr ~ 1.8 × 10 6 s -1 . This TADF rate is much faster than the kr of 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 which was measured to be 1.0 × 10 6 s -1 . This suggests that emission energy must be incorporated into the kr(TADF) vs. 100 d(h + ,e - ) model as an additional independent variable. While this result is only a single data point, it merits future investigation with complexes such as 𝐴𝑢 𝐵𝑖𝑚𝐶𝑁 2 𝑃𝐴𝐶 , 𝐴𝑢 𝐶𝑧𝐶𝑁 2 𝑃𝑍𝐼 , and 𝐴𝑢 𝐵𝑖𝑚𝐶𝑁 2 𝑃𝑍𝐼 . 101 3.15 Synthesis Of All Complexes & Precursors Metal triflates were used to perform ring closures of PrePAC and PreBZAC where M = Cu(II), Ag(I), and Na + . (PAC)M’(X) was achieved for M’ = Cu(I), Ag(I), and Au(I) where X = Cl - or BF4 - . BZAC M’’(Cl) was achieved for M’’ = Cu(I) and Au(I). (BZI)Au(D) was prepared for D = bim, Mbim, and Obim. The bim donor ligands were prepared according to the manuscript. 1 (PAC)M’(D’) and (BZAC)M’’(D’) were isolated where D’ = Cz, BCz, and bim. Ar = 3,6 – diisopropylphenyl. All chemicals, if not otherwise stated were used as received from chemical supplier. All inert reactions were done in dry nitrogen atmosphere. Flasks, cannula, stir bars and stoppers were dried prior usage at 140 °C. The detailed syntheses of all precursors and complexes are detailed below. PAC OTf: 1,3-bis(2,6-diisopropylphenyl)-1l4-quinazolin-4(3H)-one, trifluoromethanesulfonate salt (PAC OTf) was synthesized based on a modification of a literature procedure. 71 A 250 mL Schlenk flask with a stir bar was charged with 8.5 g (16.9 mmol) (E)-2-bromo-N-(2,6- diisopropylphenyl)-N-(((2,6-diisopropylphenyl)imino)methyl)benzamide (PrePAC), 4.36 g (25.34 mmol) NaOTf, and 10.8 g (29.73 mmol) Cu(OTf)2 was connected to a reflux condenser and topped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solids followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced with a rubber septum against positive N2 pressure, and 100 mL of dry DMSO was canula transferred into the reaction vessel yielding a blue suspension. The reaction mixture was heated to 160 °C with an oil bath for 12 h (The suspension fully dissolved after the temperature passed 100 °C). The mixture became dark purple overnight. The mixture was raised out of the oil bath and 100 mL DI water was added after the mixture reached room temperature. Products were extracted from the crude 102 mixture with dichloromethane three times. The dichloromethane phases were combined and washed with 200 mL of brine a total of five times. The dichloromethane phase was further dried by mixing MgSO4 and performing vacuum filtration to re-collect the dichloromethane filtrate. The solution was concentrated on the rotovap and the product was precipitated by addition of excess hexane. The product was vacuum filtered and further washed with hexane to afford PAC OTf as a light blue powder (2.55 g, 30% yield); the light blue color likely comes from trace inorganic impurities the NMR matched the literature spectrum. 1 H NMR (400 MHz, acetone-d6): 10.98, (s, 1H); 8.63, (ddd, 1H); 8.24 (ddd, 1H); 8.10 (ddd, 1H); 7.81 (m, 1H); 7.65 (m, 4H); 7.52 (m, 3H); 7.38 (ddd, 1H); 3.03 (sept, 2H); 2.92 (sept, 2H); 1.29 (dd, 12H); 1.20 (d, 6H); 1.14 (d, 6H). BZAC OTf: 1,3-bis(2,6-diisopropylphenyl)-3,4-dihydroquinazolin-1-ium trifluoromethanesulfonate (BZAC OTf) was synthesized based on a modification of a literature procedure. 71 A 250 mL Schlenk flask with a stir bar was charged with 15g (28.11 mmol) (E)-N-(2-bromobenzyl)-N,N'- bis(2,6-diisopropylphenyl)formimidamide (PreBZAC), 8.46 g NaOTf (49.19 mmol), 20.33 g Cu(OTf)2 (56.22 mmol), and 7.22 g AgOTf (28.11 mmol) was connected to a reflux condenser and topped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solids followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced with a rubber septum against positive N2 pressure, and 100 mL of dry DMSO was canula transferred into the reaction vessel yielding a blue suspension. The reaction mixture was heated to 160 °C with an oil bath for 12 h (The suspension fully dissolved after the temperature passed 100 °C). The mixture was cooled to room temperature and 100 mL DI water was added. The product was extracted from the crude mixture with dichloromethane three times. The dichloromethane phases were combined and washed with 200 mL of brine a total of five times. The dichloromethane phase was further 103 dried by mixing MgSO4 and performing vacuum filtration to re-collect the dichloromethane filtrate. The solution was concentrated on the rotovap and the product was precipitated by addition of excess hexane. The product was vacuum filtered and further washed with hexane to afford a peach colored BZAC OTf powder (5.93 g, 35% yield). The NMR matched the literature spectrum. 1 H NMR (400 MHz, acetone-d6): 9.10 (s, 1H); 7.69 (m, 1H); 7.61 (dd, 1H); 7.54 (m, 2H); 7.50 (m, 3H); 7.45 (m, 2H); 6.60 (m, 1); 5.48 (s, 2H); 3.46 (sept, 2H); 3.17 (sept, 2H); 1.39 (d, 6H); 1.30 (dd, 12H); 1.15 (d, 6H). PZI-HCl: Figure 3.24 Synthesis of PZI-HCl. The first step was synthesized according to the publication of Chi-Ming Che. 72 The second step was done similar to Chi-Ming Che’s PZIPr-BF4, 72 but varied in the following way to yield the Cl-salt: Pyrazine-di(2,6-diisopropylaniline) (1.29 g, 3.0 mmol, 1.0 eq) dissolved in 300 ml Triethylorthoformate and acetic acid (0.17 ml, 3 mmol, 1.0 eq). Triethylorthoformate and ethanol was slowly, over the course of 4-5 h, nearly fully distilled off at 150 ˚C. Reaction was cooled to room temperature, trimethylsilylchloride (43 mL, 340 mmol, 113 eq) and 50 mL of fresh triethylorthoformate was added and the reaction mixture was heated to 70 ˚C overnight. All solvents were removed until a solid crude was left behind, which was first washed with diethyl ether followed with ispropyl alcohol, yielding the pure off-white product in 75% yield. 104 1 H NMR (400 MHz, CD3CN) δ 10.52 (s, 1H), 8.97 (s, 3H), 7.78 (d, J = 7.8 Hz, 2H), 7.59 (d, J = 7.9 Hz, 4H), 2.40 (sept, J = 6.8 Hz, 4H), 1.19 (dd, J = 20.5, 6.8 Hz, 24H). 𝑪𝒖 𝑪𝒍 𝑷𝑨𝑪 : A 250 mL Schlenk flask with a stir bar was charged with 1.1g (1.78 mmol) of PAC-OTf and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 100 mL of dry THF from our solvent purification system was directly added through the septum. The PAC-OTf dissolved within 5 minutes. Next, 4.6 mL (2.32 mmol) of 0.5M potassium hexamethylsilylamide in THF was added through the septum and the reaction was stirred for an additional 5 minutes. The rubber septum was removed against positive N 2 pressure and 240 mg of CuCl was added (2.43 mmol). The dark green reaction mixture was covered with aluminum foil and stirred for 12 hrs. Removing the aluminum foil revealed a brown suspension which was filtered through Celite. The filtrate was concentrated on the rotovap and the product was precipitated out by addition of excess hexane to yield 1,3-bis(2,6-diisopropylphenyl)-4-oxo- 3,4-dihydroquinazolin-1-ium-2-ide copper(I) chloride 𝑪𝒖 𝑪𝒍 𝑷𝑨𝑪 as an off-white solid (390 mg, 39% yield). 1 H NMR (400 MHz, acetone-d6): 8.43 (ddd, 1H); 7.98 (ddd, 1H); 7.79 (ddd, 1H); 7.67 (m, 1H); 7.52 (m, 3H); 7.39 (m, 2H); 6.95 (ddd, 1); 2.91 (sept, 2H); 2.80 (sept, 2H); 1.33 (dd, 12H); 1.18 (d, 6H); 1.09 (d, 6H). 105 𝑨𝒈 𝑩 𝑭 𝟒 𝑷𝑨𝑪 : The synthesis and workup was the same as (PAC)Cu(Cl) except 318 mg (0.52 mmol) PAC- OTf was used, 129 mg (0.67 mmol) AgBF4 and 1.3 mL (0.67 mmol) 0.5 M potassium hexamethylsilylamide in THF was used to yield 1,3-bis(2,6-diisopropylphenyl)-4-oxo-3,4- dihydroquinazolin-1-ium-2-ide silver(I) tetrafluoroborate 𝑨𝒈 𝑩 𝑭 𝟒 𝑷𝑨𝑪 (100 mg, 29% yield) (400 MHz, acetone-d6): 8.45(ddd, 1H); 8.02 (ddd, 1H); 7.83 (t, 1H); 7.68 (t, 0.95); 7.54 (m, 3H); 7.42 (m, 2H); 7.00 (m, 1H); 2.93 (sept, 2H); 2.78 (sept, 2H); 1.32 (d, 12H); 1.18 (d, 6H); 1.09 (d, 6H). 𝑨𝒖 𝑪𝒍 𝑷𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 500 mg (0.81 mmol) of PAC-OTf and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 50 mL of dry THF from our solvent purification system was directly added through the septum. The PAC-OTf dissolved within 5 minutes. Next, 2.1 mL (1.05 mmol) of 0.5 M potassium hexamethylsilylamide in THF was added through the septum and the reaction was stirred for an additional 5 minutes. The rubber septum was removed against positive N2 pressure and 307 mg of AuS(CH3)2Cl was added (1.05 mmol). The dark red reaction mixture was covered with aluminum foil and stirred for 12 hrs. Removing the aluminum foil revealed a gold solution which was filtered through celite and dried using the Schlenk line vacuum yielding a light brown solid. The solid was washed with ethanol and vacuum filtered to yield 1,3-bis(2,6- diisopropylphenyl)-4-oxo-3,4-dihydroquinazolin-1-ium-2-ide gold(I) chloride 𝑨𝒖 𝑪𝒍 𝑷𝑨𝑪 as a white powder (440 mg, 78 % yield). 106 (400 MHz, acetone-d6): 8.43 (ddt, 1H); 7.99 (dddd, 1H); 7.81 (m, 1H); 7.61 (m, 1H); 7.52 (m, 3H); 7.39 (m, 2H); 6.98 (ddd, 1H); 2.98 (sept, 2H); 2.78 (sept, 2H); 1.38 (dd, 12H); 1.17 (dd, 6H); 1.07 (d, 6H). 𝐂𝐮 𝐂𝐥 𝐁𝐙𝐀𝐂 : A 200mL Schlenk flask with a stir bar was charged with 1.0 g (1.66 mmol) of BZAC-OTf and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 100 mL of dry THF from our solvent purification system was directly added through the septum. The BZAC-OTf dissolved within 5 minutes yielding a transparent solution. Next, 3.07 mL (2.14 mmol) of 0.7 M potassium hexamethylsilylamide in THF was added through the septum and the reaction was stirred for an additional 5 minutes. The solution immediately became dark green as the potassium hexamethylsilylamide was added but quickly returned to transparent. The rubber septum was removed against positive N2 pressure and 223 mg of CuCl was added (2.24 mmol). The reaction mixture was covered with aluminum foil and stirred for 12 h. Removing the aluminum foil revealed an orange solution which was filtered through Celite. The filtrate was concentrated on the rotovap yielding a dark orange solid. The crude solid was washed with ethanol to afford 1,3-bis(2,6-diisopropylphenyl)-1,4-dihydroquinazolin-3-ium-2-ide copper(I) chloride 𝑪𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 as a white solid (694 mg, 76% yield). (400 MHz, acetone-d6: 7.53 (dd, 1H); 7.39 (m, 5H); 7.25 (m, 3H); 6.33 (m, 1H); 4.96 (s, 2H); 3.37 (sept, 2H); 3.17 (sept, 2H); 1.36 (m, 18H); 1.12 (dd, 6H) 107 𝑨𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 : The synthesis and workup was the same as (BZAC)Cu(Cl) except with 390 mg (0.65 mmol) of BZAC-OTf, 1.0 mL (0.70 mmol) of 0.7 M potassium hexamethylsilylamide in THF, and 189 mg (0.73 mmol) of AuS(CH3)2Cl. The reaction mixture was covered with aluminum foil and stirred for 12 h. Removing the aluminum foil revealed a brown suspension which was filtered through celite to yield a clear, green filtrate. The filtrate was concentrated on the rotovap yielding a dark green solid. The crude solid was washed with ethanol to afford 1,3-bis(2,6- diisopropylphenyl)-1,4-dihydroquinazolin-3-ium-2-ide gold(I) chloride 𝑨𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 as a white solid (285 mg, 60% yield). (400 MHz, acetone-d6: 7.54 (m, 1H); 7.46 (dd, 1H); 7.40 (m, 2H); 7.36 (dd, 2H); 7.27 (m, 3H); 6.35 (m, 1H); 5.00 (d, 2H); 3.35 (sept, 2H); 3.13 (sept, 2H); 1.42 (dd, 12H); 1.35 (d, 6H); 1.10 (dd, 6H). 𝑨𝒖 𝑪𝒍 𝑷𝒁𝑰 : PZI-HCl (400 mg, 0.838 mmol, 1.0 eq) were pump purged and dissolved in bubble degassed THF. 0.7 M potassium hexamethylsilylamide in THF (1.26 mL, 0.880 mmol, 1.05 eq) was added and solution steered for 1 hour. Au(Me2S)Cl (272 mg, 0.922 mmol, 1.10 eq) was added and solution was stirred overnight. The reaction mixture was filtered through Celite to yield a dark red filtrate, the solvents were removed, and the residue was dissolved in minimal CH2Cl2 and precipitated with hexanes to yield the light brown 𝑨𝒖 𝑪𝒍 𝑷𝒁𝑰 product in 34% yield (190 mg, 0.282 mmol) 1H NMR (400 MHz, acetone-d6: 7.54 (m, 1H); 7.46 (dd, 1H); 7.40 (m, 2H); 7.36 (dd, 2H); 7.27 (m, 3H); 6.35 (m, 1H); 5.00 (d, 2H); 3.35 (sept, 2H); 3.13 (sept, 2H); 1.42 (dd, 12H); 1.35 (d, 6H); 1.10 (dd, 6H). 𝑨𝒖 𝑪𝒛 𝑷𝑨𝑪 : 108 A 100 mL Schlenk flask with a stir bar was charged with 57 mg (0.34 mmol) 1H-carbazole and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 38 mL of dry THF from our solvent purification system was directly added through the septum. The carbazole dissolved after 5 minutes of stirring which gave a transparent solution, and 0.19 mL (0.37 mmol) of 2 M sodium tert-butoxide was added through the septum. The reaction was stirred for 30 minutes. The mixture went from transparent to slightly green over the duration of stirring. The rubber septum was removed against positive N2 pressure and 238 mg of 𝑨𝒖 𝑪𝒍 𝑷𝑨𝑪 was added (0.34 mmol). The solution was immediately yellow emissive under UV lamp upon addition of 𝑨𝒖 𝑪𝒍 𝑷𝑨𝑪 . The reaction stirred for an additional 12 h yielding a dark green suspension. The solution was filtered through celite and dried using the rotovap. The resultant gel- like solid was redissolved in 80mL of a 50/50 dichloromethane/hexane mixture and re-dried on the rotovap three times which changed the texture of the solid from gel-like to powdery. The dried solid was washed copiously with methanol and collected via vacuum filtration to yield 𝑨𝒖 𝑪𝒛 𝑷𝑨𝑪 as a yellow powder (129 mg, 46% yield). 1 H NMR (400 MHz, acetone-d6) δ 8.47 (ddd, J = 7.9, 1.6, 0.5 Hz, 1H), 8.02 (ddd, J = 8.6, 7.3, 1.6 Hz, 1H), 7.96 (t, J = 7.8 Hz, 1H), 7.87 – 7.77 (m, 4H), 7.72 (d, J = 7.9 Hz, 2H), 7.59 (d, J = 7.8 Hz, 2H), 7.05 (dt, J = 8.5, 0.8 Hz, 1H), 6.96 (ddd, J = 8.2, 7.0, 1.3 Hz, 2H), 6.78 (ddd, J = 7.8, 7.0, 1.0 Hz, 2H), 6.13 (dt, J = 8.1, 0.9 Hz, 2H), 3.04 (sept, J = 6.9 Hz, 2H), 2.94 (sept, J = 7.0 Hz, 2H), 1.33 (dd, J = 13.3, 6.8 Hz, 12H), 1.22 (d, J = 6.8 Hz, 6H), 1.12 (d, J = 6.8 Hz, 6H). 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 206.3, 206.3, 206.3, 201.4, 160.0, 150.6, 147.5, 147.2, 142.9, 137.9, 137.8, 132.5, 131.5, 129.9, 129.7, 127.0, 125.8, 125.0, 124.2, 120.1, 120.0, 119.8, 117.0, 115.3, 30.6, 30.4, 30.2, 30.2, 30.0, 30.0, 29.8, 29.6, 29.5, 24.9, 24.8, 24.7, 24.4. CHN: C: 62.76%; H: 5.75%; N: 4.87%; calculated: C: 63.69%; H: 5.59%; N: 5.06% 109 𝑨𝒖 𝑪𝒛 𝑩𝒁𝑨𝑪 : The same synthesis and workup as 𝑨𝒖 𝑪𝒛 𝑷𝑨𝑪 was used except with 61 mg (0.36 mmol) 1H- carbazole, 0.2 mL (0.4 mmol) of 2 M sodium tert-butoxide, and 250 mg (0.36 mmol) of (BZAC)Au(Cl). The dark grey suspension was immediately bright blue emissive under UV lamp upon addition of (BZAC)Au(Cl). The final 𝑨𝒖 𝑪𝒛 𝑩𝒁𝑨𝑪 was an off-white powder (157 mg, 53% yield). 1 H NMR (400 MHz, acetone-d6) δ 7.87 – 7.70 (m, 4H), 7.61 (d, J = 7.8 Hz, 2H), 7.56 (d, J = 7.7 Hz, 2H), 7.38 – 7.25 (m, 3H), 6.91 (ddd, J = 8.2, 7.0, 1.3 Hz, 2H), 6.76 (ddd, J = 7.9, 7.0, 1.0 Hz, 2H), 6.49 – 6.41 (m, 1H), 6.11 (dt, J = 8.2, 0.9 Hz, 2H), 5.12 (d, J = 0.9 Hz, 2H), 5.12 (s, 2H), 3.50 (sept, J = 6.9 Hz, 2H), 3.29 (sept, J = 6.9 Hz, 2H), 1.39 (dd, J = 12.3, 6.9 Hz, 18H), 1.17 (d, J = 6.8 Hz, 6H). 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 206.3, 206.3, 206.3, 206.2, 196.8, 150.7, 148.4, 147.0, 142.0, 138.4, 135.9, 131.4, 130.9, 130.1, 128.3, 127.5, 126.5, 124.8, 124.0, 119.7, 119.2, 117.5, 116.5, 115.2, 52.5, 30.6, 30.4, 30.2, 30.0, 29.8, 29.8, 29.7, 29.5, 25.3, 25.2, 25.1, 24.7. Crystal structure data is presented in the next section (therefore no CHN) 110 𝑪𝒖 𝑩𝑪𝒛 𝑷𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 182 mg (0.65 mmol) 3,6-di-tert- butyl-9H-carbazole and 62 mg (0.65mmol) sodium tert-butoxide and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 38 mL of dry THF from our solvent purification system was directly added through the septum. The carbazole dissolved after 5 minutes of stirring which gave a transparent solution. The reaction was stirred for 30 minutes. The rubber septum was removed against positive N2 pressure and 350 mg (0.62 mmol) of (𝑪𝒖 𝑪𝒍 𝑷𝑨𝑪 was added. The solution was immediately red in color, and red emissive under UV lamp upon addition of 𝑪𝒖 𝑪𝒍 𝑷𝑨𝑪 . The reaction stirred for an additional 12 h yielding a dark brown suspension. The solution was filtered through Celite yielding a clear red solution. The resultant gel-like solid was redissolved in 80 mL of a 50/50 dichloromethane/hexane mixture and re-dried on the rotovap three times which changed the texture of the solid from a gel to a yellow-emissive powder. The dried solid was dissolved in ethanol and the filtrate was collected via vacuum filtration. Water was added to the ethanol solution which caused the precipitation of a yellow powder. The yellow powder was vacuum filtered and washed with methanol which afforded 𝑪𝒖 𝑩𝑪𝒛 𝑷𝑨𝑪 as a yellow powder (40 mg, 8% yield). The relatively low yield is likely due to the unconventional workup; the cMa complexes are not typically isolated by precipitation from ethanol using water. The product is also quite soluble in methanol which was used as a rinse solvent. 1 H NMR (400 MHz, acetone-d6) δ 8.47 (dd, J = 8.0, 1.4 Hz, 1H), 8.01 (td, J = 7.3, 1.9 Hz, 2H), 7.89 – 7.78 (m, 4H), 7.76 (d, J = 7.9 Hz, 2H), 7.62 (d, J = 7.8 Hz, 2H), 6.95 (td, J = 8.8, 8.4, 1.9 Hz, 3H), 5.64 (dd, J = 8.5, 0.6 Hz, 2H), 3.08 (sept, J = 6.8 Hz, 2H), 2.99 (sept, J = 6.8 Hz, 2H), 1.34 (s, 18H), 1.30 – 1.23 (m, 18H), 1.16 (d, J = 6.8 Hz, 6H). 111 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 206.3, 206.3, 206.3, 206.3, 206.3, 206.2, 159.7, 149.7, 147.5, 147.2, 142.9, 138.3, 137.5, 137.5, 132.7, 131.6, 129.7, 129.6, 127.2, 126.1, 125.1, 121.4, 120.5, 119.7, 115.5, 115.3, 35.1, 32.9, 30.7, 30.6, 30.4, 30.2, 30.0, 29.8, 29.6, 29.5, 25.4, 24.9, 24.5, 24.3. Crystal structure data present in next section (therefore no CHN) 𝑨𝒈 𝑩𝑪𝒛 𝑷𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 145 mg (0.52 mmol) 3,6-di-tert- butyl-9H-carbazole and 50 mg (0.52 mmol) sodium tert-butoxide and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 40 mL of dry THF from our solvent purification system was directly added through the septum. The solids dissolved after 30 minutes of stirring which gave a transparent solution. The rubber septum was removed against positive N2 pressure and 326 mg (0.49 mmol) of 𝑨𝒈 𝑩 𝑭 𝟒 𝑷𝑨𝑪 was added. The solution was immediately red in color, upon addition of 𝑨𝒈 𝑩 𝑭 𝟒 𝑷𝑨𝑪 . The reaction stirred for an additional 12 h yielding a red suspension. The crude solution was immediately dried on the rotovap. The resultant gel-like solid was redissolved in 80 mL of a 50/50 dichloromethane/hexane mixture and re-dried on the rotovap three times which changed the texture of the solid from a gel to a yellow-emissive powder. The dried yellow solid was dissolved in ether and filtered through celite to yield a bright red solution. The ether solution was dried on the rotovap and the resulting yellow solid was washed copiously with methanol over a vacuum filtration setup to afford 𝑨𝒈 𝑩𝑪𝒛 𝑷𝑨𝑪 as a yellow powder (116 mg, 28% yield). 1 H NMR (400 MHz, acetone-d6) δ 8.51 (dd, J = 7.9, 1.3 Hz, 1H), 8.06 (ddd, J = 8.7, 7.3, 1.6 Hz, 1H), 7.96 (t, J = 7.8 Hz, 1H), 7.91 – 7.78 (m, 3H), 7.74 (d, J = 7.8 Hz, 2H), 7.61 (d, J = 7.8 Hz, 2H), 7.48 – 7.36 (m, 1H), 7.09 (d, J = 8.4 Hz, 1H), 7.05 (dd, J = 8.5, 2.0 Hz, 2H), 6.07 (dd, J = 8.5, 0.6 Hz, 2H), 3.07 (sept, J = 7.2 Hz, 2H), 2.95 (sept, J = 6.7 Hz, 2H), 1.38 – 1.07 (m, 42H). 112 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 206.3, 206.2, 150.3, 150.2, 147.2, 146.9, 137.8, 137.6, 132.7, 131.6, 130.1, 129.8, 127.1, 126.1, 124.9, 121.5, 120.6, 120.3, 115.6, 115.1, 35.1, 32.9, 30.8, 30.6, 30.4, 30.2, 30.0, 29.8, 29.6, 29.5, 25.3, 24.8, 24.8, 24.6.] CHN: C: 72.57%; H: 7.48%; N: 4.74%; calculated: C: 73.22%; H: 7.33%; N: 4.93% 𝑨𝒖 𝑩𝑪𝒛 𝑷𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 84 mg (0.30 mmol) 3,6-di-tert- butyl-9H-carbazole and 32 mg (0.33mmol) sodium tert-butoxide and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 38 mL of dry THF from our solvent purification system was directly added through the septum. The reaction was stirred for 30 minutes yielding a transparent solution. The rubber septum was removed against positive N2 pressure and 210 mg (0.30 mmol) of 𝑨𝒖 𝑪𝒍 𝑷𝑨𝑪 was added. The solution was immediately red in color, upon addition of 𝑨𝒖 𝑪𝒍 𝑷𝑨𝑪 and was yellow luminescent under the UV lamp. The reaction stirred for an additional 12 h yielding a red suspension. The crude solution was filtered through Celite and concentrated on the rotovap, then precipitated with addition of excess hexane. The resulting yellow solid was washed copiously with methanol over a vacuum filtration setup to afford 𝑨𝒖 𝑩𝑪𝒛 𝑷𝑨𝑪 as a yellow powder (136 mg, 48% yield). 1 H NMR (400 MHz, acetone-d6) δ 8.48 (ddd, J = 7.9, 1.6, 0.5 Hz, 1H), 8.07 – 7.94 (m, 2H), 7.86 (dt, J = 2.4, 1.2 Hz, 2H), 7.87 – 7.78 (m, 2H), 7.74 (d, J = 7.8 Hz, 2H), 7.60 (d, J = 7.8 Hz, 2H), 7.10 – 7.01 (m, 3H), 6.10 (dd, J = 8.5, 0.7 Hz, 2H), 3.04 (sept, J = 6.8 Hz, 2H), 2.94 (sept, J = 6.9 Hz, 2H), 1.37 – 1.32 (m, 30H), 1.23 (d, J = 6.8 Hz, 6H), 1.14 (d, J = 6.8 Hz, 6H). 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 206.3, 206.3, 206.3, 206.3, 206.2, 149.2, 147.4, 147.1, 139.2, 132.5, 131.4, 129.6, 126.9, 125.8, 124.9, 121.7, 120.0, 115.7, 114.7, 35.1, 32.8, 30.6, 30.4, 30.2, 30.0, 29.8, 29.6, 29.4, 24.9, 24.8, 24.6, 24.5; CHN: C: 67.23%; H: 6.67%; N: 4.54%; calculated: C: 66.30%; H: 6.63%; N: 4.46% 113 𝑪𝒖 𝑩𝑪𝒛 𝑩𝒁𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 165 mg (0.36 mmol) 3,6-di-tert- butyl-9H-carbazole and 65 mg (0.43 mmol) sodium tert-butoxide and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 50 mL of dry THF from our solvent purification system was directly added through the septum. The reaction was stirred for 30 minutes yielding a transparent solution. The rubber septum was removed against positive N2 pressure and 375 mg (0.41 mmol) of 𝑪𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 was added. The solution immediately became lime green upon addition of 𝑪𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 and was bright green luminescent under the UV lamp. The reaction stirred for an additional 12 h while covered with aluminum foil which yielded a green suspension. The crude solution was immediately dried on the rotovap, then dissolved in diethyl ether and filtered through Celite. The ether phase was dried on the rotovap, then redissolved in 50 mL acetonitrile. Hexane was used to extract the product out of the acetonitrile phase (15 extractions with 50 mL hexane). The hexane phases were combined and dried on the rotovap. The solid was redissolved in 40mL methanol and four additional 20mL hexane extractions were performed, this time keeping the methanol phase and discarding the hexane phases. The methanol phase was dried on the Schlenk line yielding 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 as an off-white powder (26 mg, 6% yield). 1 H NMR (400 MHz, acetone-d6) δ 7.86 (t, J = 7.8 Hz, 1H), 7.81 – 7.73 (m, 4H), 7.63 (d, J = 7.8 Hz, 2H), 7.58 (d, J = 7.8 Hz, 2H), 7.38 – 7.22 (m, 3H), 6.91 (dd, J = 8.6, 2.1 Hz, 2H), 5.58 (dd, J = 8.6, 0.7 Hz, 2H), 5.06 (s, 2H), 3.54 (sept, J = 6.9 Hz, 2H), 3.33 (sept, J = 6.8 Hz, 2H), 1.42 (d, J = 7.0 Hz, 6H), 1.33 (s, 18H), 1.29 (dd, J = 6.9, 2.0 Hz, 12H), 1.18 (d, J = 6.8 Hz, 6H). 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 206.3, 206.3, 206.3, 206.2, 206.1, 149.3, 148.4, 147.0, 142.0, 138.6, 138.3, 136.0, 131.4, 130.8, 130.1, 128.3, 127.4, 126.5, 124.7, 121.5, 119.2, 117.5, 115.5, 114.6, 52.5, 35.1, 32.9, 30.6, 30.4, 30.2, 30.0, 29.8, 29.7, 29.5, 25.3, 25.2, 25.1, 24.7. CHN: C: 72.32%; H: 7.76%; N: 4.63%; calculated: C: 72.37%; H: 7.56%; N: 4.78% (CHN-Analysis includes one cocrystalized CH2Cl2 molecule per 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 molecule.) 114 𝑨𝒖 𝑩𝑪𝒛 𝑩𝒁𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 102 mg (0.36 mmol) 3,6-di-tert- butyl-9H-carbazole and 42 mg (0.43 mmol) sodium tert-butoxide and capped with a glass stopper. A vacuum of ~300 mTorr was pulled on the solid followed by backfilling with nitrogen for a total of three cycles. The glass stopper was replaced by a rubber septum and 50 mL of dry THF from our solvent purification system was directly added through the septum. The reaction was stirred for 30 minutes yielding a transparent solution. The rubber septum was removed against positive N2 pressure and 285 mg (0.41mmol) of 𝑨𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 was added. The solution immediately became lime green upon addition of 𝑨𝒖 𝑪𝒍 𝑩𝒁𝑨𝑪 and was bright blue luminescent under the UV lamp. The reaction stirred for an additional 12 h while covered with aluminum foil which yielded a green suspension. The crude solution was immediately dried on the rotovap, then dissolved in diethyl ether and filtered through Celite. The ether phase was dried on the rotovap, then redissolved in 50 mL acetonitrile. Hexane was used to extract the product out of the acetonitrile phase (15 extractions with 50 mL hexane). The hexane phases were combined and dried on the rotovap. The solid was redissolved in 40 mL methanol and four additional 20 mL hexane extractions were performed, this time keeping the methanol phase and discarding the hexane phases. The methanol phase was dried on the Schlenk line yielding 𝑨𝒖 𝑩𝑪𝒛 𝑩𝒁𝑨𝑪 : as an off-white powder (82 mg, 24% yield). 1 H NMR (400 MHz, acetone-d6) δ 7.88 – 7.79 (m, 3H), 7.74 (t, J = 7.8 Hz, 1H), 7.61 (d, J = 7.8 Hz, 2H), 7.56 (d, J = 7.8 Hz, 2H), 7.37 – 7.30 (m, 1H), 7.34 – 7.24 (m, 2H), 7.01 (dd, J = 8.6, 2.1 Hz, 2H), 6.48 – 6.41 (m, 1H), 6.06 (dd, J = 8.5, 0.6 Hz, 2H), 5.13 – 5.08 (m, 2H), 3.50 (sept, J = 6.8 Hz, 2H), 3.28 (sept, J = 6.8 Hz, 2H), 1.45 – 1.36 (m, 18H), 1.33 (s, 18H), 1.17 (d, J = 6.8 Hz, 6H); 13 C{ 1 H} NMR (100 MHz, acetone-d6) δ 205.1, 205.1, 205.1, 205.1, 204.9, 148.2, 147.3, 145.8, 140.9, 137.5, 137.2, 134.8, 130.2, 129.7, 129.0, 127.2, 126.3, 125.3, 123.6, 120.3, 118.1, 116.4, 114.4, 113.5, 51.3, 34.0, 31.7, 29.5, 29.3, 29.1, 28.9, 28.7, 28.5, 28.3, 24.2, 24.0, 24.0, 23.6. 115 CHN: C: 66.80%; H: 6.94%; N: 4.49%; calculated: C: 67.3%; H: 6.95%; N: 4.53% 𝑨𝒖 𝑪𝒛 𝑷𝒁𝑰 : A 25 mL Schlenk flask with a stir bar with 52 mg (0.311 mmol, 1.1 eq) 1H-carbazole was pump purged and bubble degassed dry THF (10 mL) was added via cannula transfer. 0.155 mL (0.311 mmol, 1.1 eq) 2 M sodium tert-butoxide (NaOtBu) solution was added dropwise. After 1h stirring, 𝐴𝑢 𝐶𝑙 𝑃𝑍𝐼 (190 mg, 0.282 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washed with dichloromethane and solvent was removed. Product was precipitated from dichloromethane by adding hexanes. Solid was filtered and washed with diethyl ether, yielding to the colorless 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 in 90% yield (205 mg, 0.282 mmol). Under a UV light the solid is orange/green emissive. 1 H NMR (400 MHz, CDCl3) δ 8.59 (s, 2H), 7.95 (ddd, J = 7.7, 1.4, 0.7 Hz, 2H), 7.78 (t, J = 7.9 Hz, 2H), 7.53 (d, J = 7.8 Hz, 4H), 7.07 (ddd, J = 8.2, 7.0, 1.3 Hz, 2H), 6.92 (ddd, J = 7.9, 7.0, 1.1 Hz, 2H), 6.65 (dt, J = 8.1, 0.9 Hz, 2H), 2.48 (sept, J = 6.8 Hz, 4H), 1.33 (d, J = 6.8 Hz, 12H), 1.17 (d, J = 6.9 Hz, 12H). 13 C NMR (100 MHz, CDCl3) δ 149.2, 146.7, 141.1, 140.4, 131.5, 130.4, 124.7, 123.9, 123.5, 119.4, 116.1, 113.4, 65.9, 29.7, 24.3, 24.0, 23.8, 15.3. CHN: C: 56.63%; H: 5.15%; N: 7.88%; calculated: C: 56.76%; H: 5.22%; N: 7.88% (CHN-Analysis includes one cocrystalized CH2Cl2 molecule per 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 molecule.) 𝑨𝒖 𝑩𝒊𝒎 𝑩𝒁𝑨𝑪 : A 100 mL Schlenk flask with a stir bar was charged with 258 mg (1.25 mmol, 1.05 eq) 1H- bim ligand was pump purged and bubble degassed dry THF (50 mL) was added via cannula transfer. 120 mg (1.25 mmol, 0.624 mL, 1.05 eq) sodium tert-butoxide (NaOtBu) was added as 2 M solution. After 1 h stirring, 𝐴𝑢 𝐶𝑙 𝐵𝑍𝐴𝐶 (815 mg, 1.19 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and removing of solvent yielded to a solid, which was precipitated from dichloromethane /hexanes. If necessary, washing 116 again with pure dichloromethane, dissolves 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 but, does not dissolve remaining bim ligand. 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 could be isolated as an colorless powder in 88% yield (902 mg, 1.05 mmol). Under a UV light the solid is blue emissive. This compound was crystallized with vapor diffusion of hexanes into a dichloromethane solution of the compound. Crystallographic data can be obtained in the next Chapter. 1 H NMR (400 MHz, CDCl3): δ 7.67 (d, J = 7.8 Hz, 1H), 7.58 (d, J = 8.4 Hz, 2H), 7.53 – 7.45 (m, 4H), 7.41 (d, J = 7.8 Hz, 2H), 7.23 – 7.19 (m, 2H), 7.18 – 7.11 (m, 2H), 6.98 (dd, J = 7.5, 1.2 Hz, 1H), 6.90 (dd, J = 7.5, 1.5 Hz, 2H), 6.51 – 6.46 (m, 1H), 6.05 (dd, J = 7.4, 1.6 Hz, 1H), 4.94 (s, 2H), 3.32 (d, J = 6.8 Hz, 2H), 3.13 (d, J = 6.8 Hz, 2H), 1.44 (dd, J = 6.8, 3.4 Hz, 12H), 1.40 (d, J = 6.8 Hz, 6H), 1.16 (d, J = 6.8 Hz, 6H). 13 C{ 1 H} NMR (101 MHz, CDCl3) δ 194.5, 162.5, 149.3, 146.7, 145.4, 145.3, 140.4, 136.8, 135.0, 130.6, 130.1, 129.3, 128.4, 127.1, 126.7, 126.5, 125.6, 121.1, 120.6, 117.4, 117.2, 117.1, 117.1, 116.5, 113.5, 109.1, 108.9, 51.7, 29.0, 28.9, 25.1, 24.9, 24.6. Crystal structure data present in next section (therefore no CHN) 𝑨𝒖 𝑩𝒊𝒎 𝑩𝒁𝑰 : A 50 mL Schlenk flask with a stir bar with 56.5 mg (0.25 mmol, 1.0 eq) 1H-bim ligand was pump purged and bubble degassed dry THF (20 mL) was added via cannula transfer. 24 mg (0.25 mmol, 1.0 eq) sodium tert-butoxide (NaOtBu) was added as 2M solution. After 1 h stirring, 𝐴𝑢 𝐶𝑙 𝐵𝑍𝐼 (167 mg, 0.25 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and recrystallization from dichloromethane with layered pentane yielded to the colorless 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 in 84% yield (110 mg, 0.21 mmol). Under a UV light the solid is blue emissive. 1 H NMR (400 4MHz, CDCl3) δ 7.68 (t, J = 7.8 Hz, 2H), 7.62 (d, J = 7.7 Hz, 1H), 7.58 (d, J = 7.5 Hz, 1H), 7.53 – 7.40 (m, 7H), 7.15 (dq, J = 6.1, 3.5, 2.8 Hz, 3H), 7.06 – 6.87 (m, 3H), 6.38 (d, J = 7.9 Hz, 1H), 2.54 (sept, J = 6.7 Hz, 4H), 1.40 (d, J = 6.9 Hz, 13H), 1.14 (d, J = 6.8 Hz, 13H). 13 C{ 1 H} NMR (101 MHz, CDCl3) δ 183.0, 162.3, 148.7, 146.8, 145.4, 135.0, 131.5, 131.3, 128.3, 127.0, 125.5, 125.0, 121.5, 121.0, 117.7, 117.1, 117.0, 113.5, 112.2, 109.2, 29.3, 24.9, 24.1. CHN: C: 62.54%; H: 5.52%; N: 8.29%; calculated: C: 62.78%; H: 5.51%; N: 8.32% 117 𝑨𝒖 𝑩𝒊𝒎 𝑪𝑨𝑨𝑪 : A 50 mL Schlenk flask with a stir bar with 60 mg (0.29 mmol, 1.0 eq) 1H-bim ligand was pump purged and bubble degassed dry THF (20 mL) was added via cannula transfer. 28 mg (0.29 mmol, 1.0 eq) sodium tert-butoxide (NaOtBu) was added as 2 M solution. After 1 h stirring, 𝐴𝑢 𝐶𝑙 𝐶𝐴𝐴𝐶 (180 mg, 0.29 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and recrystallization in hot acetone yielded to the colorless 𝐴𝑢 𝐵𝑖𝑚 𝐶𝐴𝐴 𝐶 in 77% yield (192 mg, 0.31 mmol). Under a UV light the solid is blue emissive. This compound was crystallized with vapor diffusion of hexanes into a dichloromethane solution of the compound. Crystallographic data can be obtained in the next section. 1 H NMR (400 MHz, CDCl3) δ 7.66 – 7.60 (m, 2H), 7.58 (dd, J = 7.9, 1.1 Hz, 2H), 7.40 (d, J = 7.8 Hz, 2H), 7.18 (dd, J = 7.7, 1.2 Hz, 1H), 7.04 (dd, J = 7.5, 1.1 Hz, 1H), 6.95 (dd, J = 7.5, 1.1 Hz, 1H), 6.86 (dd, J = 7.6, 1.3 Hz, 1H), 5.89 (d, J = 7.8 Hz, 1H), 4.28 (d, J = 12.9 Hz, 2H), 3.48 (s, 1H), 2.87 (sept, J = 6.7 Hz, 2H), 2.54 (s, 1H), 2.43 (s, 2H), 2.04 (d, J = 24.3 Hz, 7H), 1.91 (s, 3H), 1.43 (s, 6H), 1.35 (dd, J = 13.8, 6.7 Hz, 12H). 13 C{ 1 H} NMR (126 MHz, CDCl3) δ 149.2, 145.9, 145.2, 136.6, 129.9, 128.5, 127.2, 125.4, 121.4, 121.2, 117.8, 116.8, 113.6, 110.0, 109.3, 109.1, 77.7, 64.2, 59.8, 48.2, 39.2, 37.1, 35.7, 34.5, 29.5, 29.4, 29.2, 29.1, 28.9, 28.8, 28.6, 28.0, 27.5, 26.0, 22.7, 20.0, 13.7; Crystal structure data present in next section (therefore no CHN) 𝑨𝒖 𝑩𝒊𝒎 𝑴𝑨𝑪 : A 25 mL Schlenk flask with a stir bar with 43 mg (0.206 mmol, 1.0 eq) 1H-bim ligand was pump purged and bubble degassed dry THF (10 mL) was added via cannula transfer. 0.10 mL (0.206 mmol, 1.0 eq) 2M sodium tert-butoxide (NaOtBu) solution was added dropwise. After 1h stirring, 𝐴𝑢 𝐶𝑙 𝑀𝐴𝐶 (140 mg, 0.206 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and removing solvent. Product was purified by layered recrystallization of dichloromethane and pentane. Removing of solvent yielded 118 to the lightly yellow 𝐴 𝑢 𝐵𝑖𝑚 𝑀𝐴𝐶 in 85% yield (148 mg, 0.174 mmol). Under a UV light the solid is sky-blue emissive. 1 H NMR (400 MHz, CDCl3) δ 7.64 – 7.54 (m, 3H), 7.48 (dd, J = 7.5, 1.4 Hz, 1H), 7.43 (dd, J = 8.0, 1.0 Hz, 1H), 7.39 (dd, J = 7.8, 1.1 Hz, 4H), 7.14 (dd, J = 7.8, 1.2 Hz, 1H), 6.98 (dd, J = 7.6, 1.1 Hz, 1H), 6.90 (td, J = 7.6, 1.4 Hz, 1H), 6.85 (td, J = 7.6, 1.4 Hz, 1H), 5.95 (dd, J = 7.7, 1.3 Hz, 1H), 3.86 (s, 2H), 3.25 (sept, J = 6.9 Hz, 2H), 3.01 (sept, J = 6.8 Hz, 2H), 1.62 (s, 6H), 1.44 (d, J = 6.8 Hz, 6H), 1.39 (dd, J = 6.9, 3.7 Hz, 12H), 1.24 (d, J = 6.8 Hz, 6H); 13 C{ 1 H} NMR (101 MHz, CDCl3) δ 203.9, 171.4, 162.2, 149.0, 145.7, 145.1, 144.5, 140.1, 135.8, 130.7, 130.5, 128.4, 127.1, 125.8, 124.9, 121.3, 120.8, 117.5, 117.2, 116.8, 113.4, 109.1, 109.0, 62.1, 38.1, 29.5, 29.1, 25.0, 24.7, 24.5, 23.9; Crystal structure data present in next section (therefore no CHN) 𝑨𝒖 𝑩𝒊𝒎 𝑷𝑨𝑪 : A 25 mL Schlenk flask with a stir bar with 62 mg (0.300 mmol, 1.05 eq) 1H-bim ligand was pump purged and bubble degassed dry THF (12 mL) was added via cannula transfer. 0.150 mL (0.300 mmol, 1.05 eq) 2M sodium tert-butoxide (NaOtBu) solution was added dropwise. After 1h stirring, 𝐴𝑢 𝐶𝑙 𝑃𝐴𝐶 (200 mg, 0.285 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and removing solvent. Product was purified by layered recrystallization of dichloromethane and pentane. Removing of solvent yielded to the yellow 𝐴𝑢 𝐵𝑖 𝑚 𝑃𝐴𝐶 in 75% yield (186 mg, 0.214 mmol). Under a UV light the solid is green emissive. 1 H NMR (400 MHz, acetone-d6) δ 8.49 (ddd, J = 7.9, 1.6, 0.5 Hz, 1H), 8.05 (ddd, J = 8.6, 7.3, 1.6 Hz, 1H), 7.86 (td, J = 7.6, 1.1 Hz, 1H), 7.79 (ddt, J = 8.3, 7.4, 0.5 Hz, 1H), 7.69 (ddd, J = 7.8, 1.3, 0.7 Hz, 1H), 7.67 – 7.60 (m, 4H), 7.51 (d, J = 7.8 Hz, 2H), 7.36 (ddd, J = 8.0, 1.1, 0.6 Hz, 1H), 7.13 – 7.05 (m, 2H), 6.99 – 6.87 (m, 3H), 6.31 – 6.25 (m, 1H), 3.03 (sept, J = 6.8 Hz, 2H), 2.92 (sept, J = 6.8 Hz, 2H), 1.45 (dd, J = 15.1, 6.8 Hz, 12H), 1.22 (d, J = 6.8 Hz, 6H), 1.12 (d, J = 6.8 Hz, 6H), 0.14 (d, J = 6.6 Hz, 4H). 13 C{ 1 H} NMR (100 MHz, CDCl3) δ 158.7, 145.3, 145.1, 141.7, 135.9, 131.6, 130.6, 129.4, 128.3, 125.9, 124.9, 121.1, 120.6, 118.8, 118.5, 117.4, 117.0, 116.6, 113.3, 109.0, 108.9, 77.2, 29.4, 29.1, 24.5, 24.4, 24.3, 23.9, 1.0; Crystal structure data present in next section (therefore no CHN) 119 𝑨𝒖 𝑩𝒊𝒎 𝒊𝑷𝒓 : A 25 mL Schlenk flask with a stir bar with 70 mg (0.337 mmol, 1.05 eq) 1H-bim ligand was pump purged and bubble degassed dry THF (10 mL) was added via cannula transfer. 0.17 mL (0.338 mmol, 1.05 eq) 2M sodium tert-butoxide (NaOtBu) solution was added dropwise. After 1h stirring, 𝐴𝑢 𝐶𝑙 𝑖𝑃𝑟 (200 mg, 0.322 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and removing solvent. Product was first washed with Diethyl ether and then dissolved in as little dichloromethane as possible. Remaining bim ligand, does not dissolve. Removing of solvent yielded to the colorless 𝐴𝑢 𝐵𝑖𝑚 𝑖𝑃𝑟 in 80% yield (205 mg, 0.259 mmol). Under a UV light the solid is blue emissive. 1 H NMR (400 MHz, acetone-d6) δ 7.92 (s, 2H), 7.73 – 7.65 (m, 2H), 7.60 (ddd, J = 8.3, 7.2, 0.5 Hz, 2H), 7.50 – 7.44 (m, 4H), 7.29 (ddd, J = 8.0, 1.2, 0.7 Hz, 1H), 7.06 (ddd, J = 8.0, 7.3, 1.3 Hz, 1H), 6.97 – 6.87 (m, 3H), 6.68 – 6.59 (m, 1H), 2.83 (sept, J = 6.8 Hz, 4H), 1.46 (d, J = 6.9 Hz, 12H), 1.30 (d, J = 6.9 Hz, 12H). 13 C NMR (100 MHz, CDCl3) δ 207.0, 145.8, 145.6, 134.1, 130.8, 124.5, 124.3, 123.5, 121.4, 120.8, 117.5, 116.9, 116.8, 113.3, 109.0, 77.2, 30.9, 29.0, 28.8, 24.5, 24.1, 24.0. CHN: C: 60.75%; H: 5.59%; N: 8.73%; calculated: C: 60.68%; H: 5.60%; N: 8.85% 𝑨𝒖 𝑩𝒊𝒎 𝑷𝒁𝑰 : A 25 mL Schlenk flask with a stir bar with 65 mg (0.311 mmol, 1.1 eq) 1H-bim ligand was pump purged and bubble degassed dry THF (10 mL) was added via cannula transfer. 0.155 mL (0.311 mmol, 1.1 eq) 2M sodium tert-butoxide (NaOtBu) solution was added dropwise. After 1 h stirring, 𝐴𝑢 𝐶𝑙 𝑃𝑍𝐼 (190 mg, 0.282 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washed with dichloromethane and solvent was removed. Product was precipitated from dichloromethane by adding hexanes. Solid was filtered and washed with diethyl ether, yielding to the colorless 𝐴𝑢 𝐵𝑖𝑚 𝑃𝑍𝐼 in 76% yield (182 mg, 0.216 mmol). Under a UV light the solid is green emissive. 120 1 H NMR (400 MHz, acetone-d6) δ 8.76 (s, 2H), 7.81 – 7.66 (m, 4H), 7.61 (d, J = 7.8 Hz, 4H), 7.33 (d, J = 7.9 Hz, 0H), 7.09 (td, J = 7.9, 1.2 Hz, 1H), 7.01 – 6.92 (m, 3H), 6.58 – 6.49 (m, 1H), 2.67 (sept, J = 6.9 Hz, 4H), 1.45 (d, J = 6.8 Hz, 12H), 1.15 (d, J = 6.8 Hz, 12H). 13 C NMR (100 MHz, acetone-d6) δ 161.7, 148.7, 146.9, 144.7, 142.2, 140.3, 131.5, 130.7, 128.1, 126.7, 124.7, 121.4, 121.3, 118.1, 116.9, 116.7, 112.9, 109.2, 109.2, 29.2, 23.9, 23.7, 23.3, 23.2. CHN: C: 59.83%; H: 5.24%; N: 11.56%; calculated: C: 59.78%; H: 5.26%; N: 11.62% 𝑨𝒖 𝑴𝒆𝑩𝒊𝒎 𝑩𝒁𝑰 : A 25 mL Schlenk flask with a stir bar with 33 mg (0.149 mmol, 1.0 eq) 1H-Mbim ligand was pump purged and bubble degassed dry THF (10 mL) was added via cannula transfer. 0.075 mL (0.149 mmol, 1.0 eq) 2 M sodium tert-butoxide (NaOtBu) solution was added dropwise. After stirring for 1 h, 𝐴𝑢 𝐶𝑙 𝐵𝑍𝐼 (100 mg, 0.149 mmol, 1.0 eq) was added and reaction was stirred overnight. Filtration through Celite, washing with dichloromethane and recrystallization in hot acetone yielded to the colorless 𝐴𝑢 𝑀𝑒𝐵𝑖𝑚 𝐵𝑍𝐼 in 76% yield (116 mg, 0.217 mmol). Under a UV light the solid is blue emissive. This compound was crystallized with vapor diffusion of hexanes into a dichloromethane solution of the compound. Crystallographic data can be obtained in the next section. The NMR spectra showed a 50/50 mixture of both possible tautomers. The proton integrations are for both tautomers together. 1 H NMR (400 MHz, CDCl3) δ 7.68 (t, J = 7.8 Hz, 2H), 7.62 (d, J = 7.7 Hz, 1H), 7.57 (d, J = 7.0 Hz, 1H), 7.53 – 7.40 (m, 7H), 7.15 (dq, J = 6.1, 3.5, 2.8 Hz, 3H), 7.06 – 6.87 (m, 3H), 6.38 (d, J = 7.2 Hz, 1H), 5.30 (s, 1H), 2.54 (sept, J = 6.7 Hz, 4H), 1.40 (d, J = 6.9 Hz, 13H), 1.14 (d, J = 6.8 Hz, 13H). 13 C{ 1 H} NMR (101 MHz, CDCl3) δ 183.0, 162.3, 148.7, 146.8, 145.4, 135.0, 131.5, 131.3, 128.3, 127.0, 125.5, 125.0, 121.5, 121.0, 117.7, 117.1, 117.0, 113.5, 112.2, 109.2, 29.3, 24.9, 24.1 Crystal structure data present in next section (therefore no CHN) 121 𝑨𝒖 𝑴𝒆𝑶𝑩𝒊𝒎 𝑩𝒁𝑰 : A 25 mL round-bottom flask with stir bar was charged with 𝐴𝑢 𝐶𝑙 𝐵𝑍𝐼 (200 mg, 0.30 mmol, 1.0 eq), 71 mg (0.30 mmol, 1.0 eq) 1H-Obim ligand and fine ground K2CO3 (125 mg, 0.9 mmol, 3.0 eq) was dissolved in minimal acetone and stirred for 24 h at room temperature. Solution was filtered through Celite, washed with dichloromethane and dried under vacuum. Sonicating in diethylether and collecting the precipitate yielded to the colorless 𝐴𝑢 𝑀𝑒𝑂𝐵𝑖𝑚 𝐵𝑍𝐼 in 85% yield (140 mg, 0.255 mmol). Under a UV light the solid is blue emissive. The NMR spectra showed a 50/50 mixture of both possible tautomers. The proton integrations are for both tautomers together. [ 1 H NMR (400 MHz, CDCl3) δ 7.67 (t, J = 7.8 Hz, 1H), 7.62 – 7.56 (m, 2H), 7.53 – 7.45 (m, 4H), 7.41 (d, J = 7.8 Hz, 2H), 7.23 – 7.19 (m, 2H), 7.18 – 7.11 (m, 2H), 6.98 (td, J = 7.5, 1.2 Hz, 1H), 6.90 (pd, J = 7.4, 1.5 Hz, 2H), 6.51 – 6.46 (m, 1H), 6.05 (dd, J = 7.4, 1.6 Hz, 1H), 4.94 (s, 2H), 3.32 (sept, J = 6.8 Hz, 2H), 3.13 (sept, J = 6.8 Hz, 2H), 1.44 (dd, J = 6.8, 3.4 Hz, 12H), 1.40 (d, J = 6.8 Hz, 6H), 1.16 (d, J = 6.8 Hz, 6H). 13 C{ 1 H} NMR (101 MHz, CDCl3) δ 194.5, 162.5, 149.3, 146.7, 145.4, 145.3, 140.4, 136.8, 135.0, 130.6, 130.1, 129.3, 128.4, 127.1, 126.7, 126.5, 125.6, 121.1, 120.6, 117.4, 117.2, 117.1, 117.1, 116.5, 113.5, 109.1, 108.9, 51.7, 29.0, 28.9, 25.1, 24.9, 24.6.; CHN: C: 61.79%; H: 5.56%; N: 7.96%; calculated: C: 61.99%; H: 5.55%; N: 8.03% 3.16 NMR & Mass Spec of All Reported Complexes in Ch 3 All NMR measurements were performed with a Varian 400-MR 2-Channel NMR Spectrometer. MALDI TOF experiments were performed using a Bruker Auto Flex Speed MALDI. The Bruker peptide calibration standard II with the HCCA matrix was used as an internal callibrant. Measurements were performed in a reverse positive method. The instrument was calibrated for each compound using the peptide standard and was double-checked after each analyte was measured to confirm that no drift occurred during the measurement. All compounds 122 displayed a clear match between calculated and experimental isotope patterns. The weighted average of all isotope signals was compared to the calculated molecular weight and reported below. 123 Figure 3.25 1 H NMR of PAC OTf in acetone-d6 Figure 3.26 1 H NMR of BZAC OTf (400 MHz in acetone-d6). 124 Figure 3.27 1 H NMR of 𝐶𝑢 𝐶𝑙 𝑃𝐴𝐶 (400 MHz in acetone-d6). Figure 3.28 1 H NMR of 𝐴𝑔 𝐵 𝐹 4 𝑃𝐴𝐶 (400 MHz in acetone-d6). 125 Figure 3.29 1 H NMR of 𝐴𝑢 𝐶𝑙 𝑃𝐴𝐶 (400 MHz in acetone-d6). Figure 3.30 1 H NMR of 𝐶𝑢 𝐶𝑙 𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). 126 Figure 3.31 1 H NMR of 𝐴𝑢 𝐶𝑙 𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). 127 Figure 3.32 1 H and 13 C{ 1 H} NMR of 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 (400 MHz in acetone-d6). 128 Figure 3.33 Mass spectrometry (MALDI) of 𝐴𝑢 𝐶𝑧 𝑃𝐴𝐶 vs Bruker Peptide Standard 0.0 0.5 1.0 1.5 2.0 2.5 4 x10 Intens. [a.u.] 829 830 831 832 833 834 m/z 129 Figure 3.34 1 H and 13 C{ 1 H} NMR of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). 130 Figure 3.35 Mass spectrometry (MALDI) of 𝐴𝑢 𝐶𝑧 𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard 0.0 0.5 1.0 1.5 4 x10 Intens. [a.u.] 815.0 815.5 816.0 816.5 817.0 817.5 818.0 m/z 131 Figure 3.36 1 H and 13 C{ 1 H} NMR of 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 (400 MHz in acetone-d6). 132 Figure 3.37 Mass spectrometry (MALDI) of 𝐶𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 vs Bruker Peptide Standard 0.00 0.25 0.50 0.75 1.00 1.25 4 x10 Intens. [a.u.] 806 807 808 809 810 811 812 813 m/z 133 Figure 3.38 1 H and 13 C{ 1 H} NMR of 𝐴𝑔 𝐵𝐶𝑧 𝑃𝐴𝐶 (400 MHz in acetone-d6). 134 Figure 3.39 Mass spectrometry (MALDI) of 𝐴𝑔 𝐵𝐶𝑧 𝑃𝐴𝐶 vs Bruker Peptide Standard 0 500 1000 1500 2000 2500 Intens. [a.u.] 848 850 852 854 856 858 m/z 135 Figure 3.40 1 H and 13 C{ 1 H} NMR of 𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 (400 MHz in acetone-d6). 136 Figure 3.41 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝐶𝑧 𝑃𝐴𝐶 vs Bruker Peptide Standard 0 1000 2000 3000 4000 5000 Intens. [a.u.] 940 942 944 946 948 m/z 137 Figure 3.42 1 H and 13 C{ 1 H} NMR of 𝐴𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). 138 Figure 3.43 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard 0 500 1000 1500 2000 2500 3000 Intens. [a.u.] 926.0 926.5 927.0 927.5 928.0 928.5 929.0 929.5 930.0 930.5 m/z 139 Figure 3.44 1 H and 13 C{ 1 H} NMR of 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 (400 MHz in acetone-d6). 140 Figure 3.45 Mass spectrometry (MALDI) of 𝐶𝑢 𝐵𝐶𝑧 𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard 0 500 1000 1500 2000 Intens. [a.u.] 790 792 794 796 798 800 m/z 141 Figure 3.46 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝑃𝑍𝐼 in acetone-d6 142 Figure 3.47 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝑃𝐴𝐶 in CDCl3 143 Figure 3.48 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝑖𝑚 𝑃𝐴𝐶 vs Bruker Peptide Standard 0.0 0.5 1.0 1.5 2.0 2.5 4 x10 Intens. [a.u.] 867 868 869 870 871 872 873 874 875 876 m/z 144 Figure 3.49 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝑀𝐴𝐶 in CDCl3 145 Figure 3.50 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝑖𝑚 𝑀𝐴𝐶 vs Bruker Peptide Standard 0 1 2 3 4 4 x10 Intens. [a.u.] 846 848 850 852 854 856 m/z 146 Figure 3.51 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝐶𝐴𝐴𝐶 in CDCl3 147 Figure 3.52 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝑖𝑚 𝐶𝐴𝐴𝐶 vs Bruker Peptide Standard 0 1000 2000 3000 4000 Intens. [a.u.] 779 780 781 782 783 784 785 786 787 m/z 148 Figure 3.53 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 in CDCl3 Figure 3.54 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐴𝐶 vs Bruker Peptide Standard 856.290 855.183 0.0 0.5 1.0 1.5 2.0 4 x10 Intens. [a.u.] 854 855 856 857 858 859 860 861 m/z 149 Figure 3.55 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 in CDCl3 150 Figure 3.56 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝑖𝑚 𝐵𝑍𝐼 vs Bruker Peptide Standard 0 2000 4000 6000 8000 Intens. [a.u.] 840 841 842 843 844 845 846 847 m/z 151 Figure 3.57 1 H and 13 C{ 1 H} spectra for 𝐴𝑢 𝑀𝑒𝐵𝑖𝑚 𝐵𝑍𝐼 in CDCl3 152 Figure 3.58 Mass spectrometry (MALDI) of 𝐴𝑢 𝑀𝑒𝐵𝑖𝑚 𝐵𝑍𝐼 vs Bruker Peptide Standard 0 2000 4000 6000 8000 Intens. [a.u.] 853 854 855 856 857 858 859 860 861 m/z 153 Figure 3.59 1 H and 13 C{1H} spectra for 𝐴𝑢 𝑀𝑒𝑂𝐵𝑖𝑚 𝐵𝑍𝐼 in CDCl3 Figure 3.60 Mass spectrometry (MALDI) of 𝐴𝑢 𝑀𝑒𝑂𝐵𝑖𝑚 𝐵𝑍𝐼 vs Bruker Peptide Standard 0 2000 4000 6000 8000 Intens. [a.u.] 870.0 870.5 871.0 871.5 872.0 872.5 873.0 873.5 874.0 874.5 875.0 m/z 154 Figure 3.61 1 H and 13 C{1H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝑖𝑃𝑟 in CDCl3 155 Figure 3.62 Mass spectrometry (MALDI) of 𝐴𝑢 𝐵𝑖𝑚 𝑖𝑃𝑟 vs Bruker Peptide Standard 0 2 4 6 4 x10 Intens. [a.u.] 790 792 794 796 798 800 m/z 156 Figure 3.63 1 H and 13 C{1H} spectra for 𝐴𝑢 𝐵𝑖𝑚 𝑃𝑍𝐼 in acetone-d6 157 Figure 3.64 1 H and 13 C{1H} spectra for 𝐴𝑢 𝐶𝑧 𝑃𝑍𝐼 in acetone-d6 158 3.17 Chapter 3 References (1) Muniz, C. N.; Schaab, J.; Razgoniaev, A.; Djurovich, P. I.; Thompson, M. E. π-Extended Ligands in Two-Coordinate Coinage Metal Complexes. Journal of the American Chemical Society 2022, 144 (39), 17916-17928. DOI: 10.1021/jacs.2c06948. (2) Hamze, R.; Idris, M.; Muthiah Ravinson, D. S.; Jung, M. C.; Haiges, R.; Djurovich, P. I.; Thompson, M. E. Highly Efficient Deep Blue Luminescence of 2-Coordinate Coinage Metal Complexes Bearing Bulky NHC Benzimidazolyl Carbene. Frontiers in Chemistry 2020, 8 (401), Original Research. DOI: 10.3389/fchem.2020.00401. (3) Idris, M.; Kapper, S. C.; Tadle, A. C.; Batagoda, T.; Muthiah Ravinson, D. S.; Abimbola, O.; Djurovich, P. I.; Kim, J.; Coburn, C.; Forrest, S. R.; et al. Blue Emissive fac/mer-Iridium (III) NHC Carbene Complexes and their Application in OLEDs. Advanced Optical Materials 2021, 9 (8), 2001994, https://doi.org/10.1002/adom.202001994. DOI: https://doi.org/10.1002/adom.202001994 (acccessed 2023/03/29). (4) Tadle, A. C.; El Roz, K. A.; Soh, C. H.; Sylvinson Muthiah Ravinson, D.; Djurovich, P. I.; Forrest, S. R.; Thompson, M. E. Tuning the Photophysical and Electrochemical Properties of Aza- Boron-Dipyridylmethenes for Fluorescent Blue OLEDs. Advanced Functional Materials 2021, 31 (27), 2101175, https://doi.org/10.1002/adfm.202101175. DOI: https://doi.org/10.1002/adfm.202101175 (acccessed 2023/03/29). (5) Ha, D.-G.; Tiepelt, J.; Fusella, M. A.; Weaver, M. S.; Brown, J. J.; Einzinger, M.; Sherrott, M. C.; Van Voorhis, T.; Thompson, N. J.; Baldo, M. A. Dominance of Exciton Lifetime in the Stability of Phosphorescent Dyes. Advanced Optical Materials 2019, 7 (21), 1901048, https://doi.org/10.1002/adom.201901048. DOI: https://doi.org/10.1002/adom.201901048 (acccessed 2023/03/29). (6) Ma, J.; Idris, M.; Li, T. Y.; Ravinson, D. S. M.; Fleetham, T.; Kim, J.; Djurovich, P. I.; Forrest, S. R.; Thompson, M. E. Symmetric “Double Spiro” Wide Energy Gap Hosts for Blue Phosphorescent OLED Devices. Advanced Optical Materials 2022, 10 (2), 2101530, https://doi.org/10.1002/adom.202101530. DOI: https://doi.org/10.1002/adom.202101530 (acccessed 2023/03/29). (7) Zhu, Z.-L.; Gnanasekaran, P.; Yan, J.; Zheng, Z.; Lee, C.-S.; Chi, Y.; Zhou, X. Efficient Blue Electrophosphorescence and Hyperphosphorescence Generated by Bis-tridentate Iridium(III) Complexes. Inorganic Chemistry 2022, 61 (23), 8898-8908. DOI: 10.1021/acs.inorgchem.2c01026. (8) Sajoto, T.; Djurovich, P. I.; Tamayo, A. B.; Oxgaard, J.; Goddard, W. A., III; Thompson, M. E. Temperature Dependence of Blue Phosphorescent Cyclometalated Ir(III) Complexes. Journal of the American Chemical Society 2009, 131 (28), 9813-9822. DOI: 10.1021/ja903317w. (9) Kuo, H.-H.; Zhu, Z.-l.; Lee, C.-S.; Chen, Y.-K.; Liu, S.-H.; Chou, P.-T.; Jen, A. K. Y.; Chi, Y. Bis-Tridentate Iridium(III) Phosphors with Very High Photostability and Fabrication of Blue- Emitting OLEDs. Advanced Science 2018, 5 (9), 1800846, 159 https://doi.org/10.1002/advs.201800846. DOI: https://doi.org/10.1002/advs.201800846 (acccessed 2023/03/29). (10) Escudero, D. Quantitative prediction of photoluminescence quantum yields of phosphors from first principles. Chemical Science 2016, 7 (2), 1262-1267, 10.1039/C5SC03153B. DOI: 10.1039/C5SC03153B. (11) Ravinson, D. S. M.; Thompson, M. E. Thermally assisted delayed fluorescence (TADF): fluorescence delayed is fluorescence denied. Materials Horizons 2020, 7 (5), 1210-1217, 10.1039/D0MH00276C. DOI: 10.1039/D0MH00276C. (12) Li, T.-y.; Schaab, J.; Djurovich, P. I.; Thompson, M. E. Toward rational design of TADF two- coordinate coinage metal complexes: understanding the relationship between natural transition orbital overlap and photophysical properties. Journal of Materials Chemistry C 2022, 10 (12), 4674-4683. DOI: 10.1039/d2tc00163b. (13) Ziolo, R. F.; Lipton, S.; Dori, Z. The photoluminescence of phosphine complexes of d 10 metals. Journal of the Chemical Society D: Chemical Communications 1970, (17). DOI: 10.1039/c29700001124. (14) Antipas, A.; Dolphin, D.; Gouterman, M.; Johnson, E. C. Porphyrins. 38. Redox potentials, charge transfer transitions, and emission of copper, silver, and gold complexes. Journal of the American Chemical Society 1978, 100 (24), 7705-7709. DOI: 10.1021/ja00492a044. (15) Sorrell, T. N.; Borovik, A. S. Absorption, emission, and photophysical properties of copper(I) pyrazole complexes and their carbonyl adducts. Inorganic Chemistry 1987, 26 (12), 1957-1964. DOI: 10.1021/ic00259a029. (16) Mercs, L.; Albrecht, M. Beyond catalysis: N-heterocyclic carbene complexes as components for medicinal, luminescent, and functional materials applications. Chem. Soc. Rev. 2010, 39 (6), 1903-1912, 10.1039/B902238B. DOI: 10.1039/b902238b. (17) Visbal, R.; Gimeno, M. C. N-heterocyclic carbene metal complexes: photoluminescence and applications. Chem. Soc. Rev. 2014, 43 (10), 3551-3574, 10.1039/C3CS60466G. DOI: 10.1039/c3cs60466g. (18) Wang, H. M. J.; Chen, C. Y. L.; Lin, I. J. B. Synthesis, Structure, and Spectroscopic Properties of Gold(I)−Carbene Complexes. Organometallics 1999, 18 (7), 1216-1223. DOI: 10.1021/om980718b. (19) King, C.; Khan, M. N. I.; Staples, R. J.; Fackler, J. P. Luminescent mononuclear gold(I) phosphines. Inorganic Chemistry 1992, 31 (15), 3236-3238. DOI: 10.1021/ic00041a013. (20) Hong, X.; Cheung, K.-K.; Guo, C.-X.; Che, C.-M. Luminescent organometallic gold(I) complexes. Structure and photophysical properties of alkyl-, aryl- and µ-ethynylene gold(I) complexes. J. Chem. Soc., Dalton Trans. 1994, (13), 1867-1871, 10.1039/DT9940001867. DOI: 10.1039/dt9940001867. 160 (21) Wang, H. M. J.; Vasam, C. S.; Tsai, T. Y. R.; Chen, S.-H.; Chang, A. H. H.; Lin, I. J. B. Gold(I) N-Heterocyclic Carbene and Carbazolate Complexes. Organometallics 2005, 24 (4), 486- 493. DOI: 10.1021/om049221u. (22) Gimeno, M. C.; Laguna, A.; Visbal, R. N-Heterocyclic Carbene Coinage Metal Complexes as Intense Blue-Green Emitters. Organometallics 2012, 31 (20), 7146-7157. DOI: 10.1021/om300571m. (23) Gomez-Suarez, A.; Nelson, D. J.; Thompson, D. G.; Cordes, D. B.; Graham, D.; Slawin, A. M.; Nolan, S. P. Synthesis, characterization and luminescence studies of gold(I)-NHC amide complexes. Beilstein J. Org. Chem. 2013, 9, 2216-2223. DOI: 10.3762/bjoc.9.260. (24) Romanov, A. S.; Di, D.; Yang, L.; Fernandez-Cestau, J.; Becker, C. R.; James, C. E.; Zhu, B.; Linnolahti, M.; Credgington, D.; Bochmann, M. Highly photoluminescent copper carbene complexes based on prompt rather than delayed fluorescence. Chem. Commun. 2016, 52 (38), 6379-6382. DOI: 10.1039/C6CC02349E and 10.1039/C8CC90135J (25) Hamze, R.; Jazzar, R.; Soleilhavoup, M.; Djurovich, P. I.; Bertrand, G.; Thompson , M. E. Phosphorescent 2-, 3- and 4-coordinate cyclic (alkyl)(amino)carbene (CAAC) Cu(I) complexes. Chem. Commun. 2017, 53, 9008-9011. DOI: 10.1039/C7CC02638B. (26) Shi, S.; Collins, L. R.; Mahon, M. F.; Djurovich, P. I.; Thompson, M. E.; Whittlesey, M. K. Synthesis and characterization of phosphorescent two-coordinate copper(i) complexes bearing diamidocarbene ligands. Dalton Transactions 2017, 46 (3), 745-752, 10.1039/C6DT04016K. DOI: 10.1039/C6DT04016K. (27) Romanov, A. S.; Becker, C. R.; James, C. E.; Di, D.; Credgington, D.; Linnolahti, M.; Bochmann, M. Copper and Gold Cyclic (Alkyl)(amino)carbene Complexes with Sub- Microsecond Photoemissions: Structure and Substituent Effects on Redox and Luminescent Properties. Chemistry – A European Journal 2017, 23 (19), 4625-4637. DOI: 10.1002/chem.201605891. (28) Yersin, H.; Czerwieniec, R.; Shafikov, M. Z.; Suleymanova, A. F. TADF Material Design: Photophysical Background and Case Studies Focusing on CuI and AgI Complexes. Chemphyschem 2017, 18 (24), 3508-3535. DOI: 10.1002/cphc.201700872 (acccessed 2018/06/05). (29) Gernert, M.; Balles-Wolf, L.; Kerner, F.; Müller, U.; Schmiedel, A.; Holzapfel, M.; Marian, C. M.; Pflaum, J.; Lambert, C.; Steffen, A. Cyclic (Amino)(aryl)carbenes Enter the Field of Chromophore Ligands: Expanded π System Leads to Unusually Deep Red Emitting CuI Compounds. Journal of the American Chemical Society 2020, 142 (19), 8897-8909. DOI: 10.1021/jacs.0c02234. (30) Li, T.-y.; Muthiah Ravinson, D. S.; Haiges, R.; Djurovich, P. I.; Thompson, M. E. Enhancement of the Luminescent Efficiency in Carbene-Au(I)-Aryl Complexes by the Restriction of Renner–Teller Distortion and Bond Rotation. Journal of the American Chemical Society 2020, 142 (13), 6158-6172. DOI: 10.1021/jacs.9b13755. 161 (31) Di, D.; Romanov, A. S.; Yang, L.; Richter, J. M.; Rivett, J. P. H.; Jones, S.; Thomas, T. H.; Abdi Jalebi, M.; Friend, R. H.; Linnolahti, M.; et al. High-performance light-emitting diodes based on carbene-metal-amides. Science 2017, 356 (6334), 159. DOI: 10.1126/science.aah4345. (32) Conaghan, P. J.; Menke, S. M.; Romanov, A. S.; Jones, S. T. E.; Pearson, A. J.; Evans, E. W.; Bochmann, M.; Greenham, N. C.; Credgington, D. Efficient Vacuum-Processed Light-Emitting Diodes Based on Carbene–Metal–Amides. Adv. Mater. 2018, 30 (35), 1802285. DOI: 10.1002/adma.201802285 (acccessed 2019/09/07). (33) Romanov, A. S.; Jones, S. T. E.; Yang, L.; Conaghan, P. J.; Di, D.; Linnolahti, M.; Credgington, D.; Bochmann, M. Mononuclear Silver Complexes for Efficient Solution and Vacuum-Processed OLEDs. Advanced Optical Materials 2018, 6 (24), 1801347. DOI: doi:10.1002/adom.201801347. (34) Hamze, R.; Peltier, J. L.; Sylvinson, D.; Jung, M.; Cardenas, J.; Haiges, R.; Soleilhavoup, M.; Jazzar, R.; Djurovich, P. I.; Bertrand, G.; et al. Eliminating nonradiative decay in Cu(I) emitters: > 99% quantum efficiency and microsecond lifetime. Science 2019, 363 (6427), 601-606. DOI: 10.1126/science.aav2865. (35) Hamze, R.; Shi, S.; Kapper, S. C.; Muthiah Ravinson, D. S.; Estergreen, L.; Jung, M.-C.; Tadle, A. C.; Haiges, R.; Djurovich, P. I.; Peltier, J. L.; et al. “Quick-Silver” from a Systematic Study of Highly Luminescent, Two-Coordinate, d10 Coinage Metal Complexes. Journal of the American Chemical Society 2019, 141 (21), 8616-8626. DOI: 10.1021/jacs.9b03657. (36) Shi, S.; Jung, M. C.; Coburn, C.; Tadle, A.; Sylvinson M. R, D.; Djurovich, P. I.; Forrest, S. R.; Thompson, M. E. Highly Efficient Photo- and Electroluminescence from Two-Coordinate Cu(I) Complexes Featuring Nonconventional N-Heterocyclic Carbenes. Journal of the American Chemical Society 2019, 141 (8), 3576-3588. DOI: 10.1021/jacs.8b12397. (37) Romanov, A. S.; Yang, L.; Jones, S. T. E.; Di, D.; Morley, O. J.; Drummond, B. H.; Reponen, A. P. M.; Linnolahti, M.; Credgington, D.; Bochmann, M. Dendritic Carbene Metal Carbazole Complexes as Photoemitters for Fully Solution-Processed OLEDs. Chem. Mater. 2019, 31 (10), 3613-3623. DOI: 10.1021/acs.chemmater.8b05112. (38) Chotard, F.; Sivchik, V.; Linnolahti, M.; Bochmann, M.; Romanov, A. S. Mono- versus Bicyclic Carbene Metal Amide Photoemitters: Which Design Leads to the Best Performance? Chem. Mater. 2020, 32 (14), 6114-6122. DOI: 10.1021/acs.chemmater.0c01769. (39) Romanov, A. S.; Jones, S. T. E.; Gu, Q.; Conaghan, P. J.; Drummond, B. H.; Feng, J.; Chotard, F.; Buizza, L.; Foley, M.; Linnolahti, M.; et al. Carbene metal amide photoemitters: tailoring conformationally flexible amides for full color range emissions including white-emitting OLED. Chemical Science 2020, 11 (2), 435-446, 10.1039/C9SC04589A. DOI: 10.1039/c9sc04589a. (40) Feng, J.; Yang, L.; Romanov, A. S.; Ratanapreechachai, J.; Reponen, A.-P. M.; Jones, S. T. E.; Linnolahti, M.; Hele, T. J. H.; Köhler, A.; Bässler, H.; et al. Environmental Control of Triplet Emission in Donor–Bridge–Acceptor Organometallics. Adv. Funct. Mater. 2020, 30 (9), 1908715. DOI: 10.1002/adfm.201908715. 162 (41) Conaghan, P. J.; Matthews, C. S. B.; Chotard, F.; Jones, S. T. E.; Greenham, N. C.; Bochmann, M.; Credgington, D.; Romanov, A. S. Highly efficient blue organic light-emitting diodes based on carbene-metal-amides. Nature Communications 2020, 11 (1), 1758. DOI: 10.1038/s41467-020- 15369-8. (42) Yersin, H. Triplet emitters for OLED applications. Mechanisms of exciton trapping and control of emission properties. In Transition Metal And Rare Earth Compounds III, Topics In Current Chemistry, Vol. 241; 2004; pp 1-26. (43) Thompson, M. E.; Djurovich, P. E.; Barlow, S.; Marder, S. Organometallic Complexes for Optoelectronic Applications. In Comprehensive Organometallic Chemistry III, 1 ed.; O'Hare, D. Ed.; Applications III: Functional Materials, Environmental and Biological Applications, Vol. 12; Elsevier Ltd., 2007; pp 102-194. (44) Bolink, H. J.; Coronado, E.; Costa, R. D.; Lardiés, N.; Ortí, E. Near-Quantitative Internal Quantum Efficiency in a Light-Emitting Electrochemical Cell. Inorganic Chemistry 2008, 47 (20), 9149-9151. DOI: 10.1021/ic801587n. (45) Mahoro, G. U.; Fernandez-Cestau, J.; Renaud, J.-L.; Coto, P. B.; Costa, R. D.; Gaillard, S. Recent Advances in Solid-State Lighting Devices Using Transition Metal Complexes Exhibiting Thermally Activated Delayed Fluorescent Emission Mechanism. Advanced Optical Materials 2020, 8 (16), 2000260. DOI: 10.1002/adom.202000260. (46) Kalyanasundaram, K. Photophysics, photochemistry and solar energy conversion with tris(bipyridyl)ruthenium(II) and its analogues. Coord. Chem. Rev. 1982, 46, 159-244. DOI: 10.1016/0010-8545(82)85003-0. (47) Kim, S.; Toste, F. D. Mechanism of Photoredox-Initiated C–C and C–N Bond Formation by Arylation of IPrAu(I)–CF3 and IPrAu(I)–Succinimide. Journal of the American Chemical Society 2019, 141 (10), 4308-4315. DOI: 10.1021/jacs.8b11273. (48) Jazzar, R.; Soleilhavoup, M.; Bertrand, G. Cyclic (Alkyl)- and (Aryl)-(amino)carbene Coinage Metal Complexes and Their Applications. Chem. Rev. 2020, 120 (9), 4141-4168. DOI: 10.1021/acs.chemrev.0c00043. (49) Keefe, M. H.; Benkstein, K. D.; Hupp, J. T. Luminescent sensor molecules based on coordinated metals: a review of recent developments. Coord. Chem. Rev. 2000, 205 (1), 201-228. DOI: 10.1016/S0010-8545(00)00240-X. (50) Lo, K. K.-W.; Louie, M.-W.; Zhang, K. Y. Design of luminescent iridium(III) and rhenium(I) polypyridine complexes as in vitro and in vivo ion, molecular and biological probes. Coord. Chem. Rev. 2010, 254 (21), 2603-2622. DOI: 10.1016/j.ccr.2010.01.014. (51) Liu, Q.; Wu, H.; Zhao, Z.; Wei, D. Macrometallocycle binuclear NHC silver(I) complex with bridging azobenzene: Synthesis, structure and recognition for hydrogen sulfate. Tetrahedron 2019, 75 (23), 3128-3134. DOI: 10.1016/j.tet.2019.04.066. 163 (52) Yang, Z.; Mao, Z.; Xie, Z.; Zhang, Y.; Liu, S.; Zhao, J.; Xu, J.; Chi, Z.; Aldred, M. P. Recent advances in organic thermally activated delayed fluorescence materials. Chem. Soc. Rev. 2017, 46 (3), 915-1016, 10.1039/C6CS00368K. DOI: 10.1039/c6cs00368k. (53) Li, W.; Li, B.; Cai, X.; Gan, L.; Xu, Z.; Li, W.; Liu, K.; Chen, D.; Su, S.-J. Tri-Spiral Donor for High Efficiency and Versatile Blue Thermally Activated Delayed Fluorescence Materials. Angew. Chem. Int. Ed. 2019, 58 (33), 11301-11305. DOI: 10.1002/anie.201904272 (acccessed 2020/10/11). (54) Li, P.; Chan, H.; Lai, S.-L.; Ng, M.; Chan, M.-Y.; Yam, V. W.-W. Four-Coordinate Boron Emitters with Tridentate Chelating Ligand for Efficient and Stable Thermally Activated Delayed Fluorescence Organic Light-Emitting Devices. Angew. Chem. Int. Ed. 2019, 58 (27), 9088-9094. DOI: 10.1002/anie.201903332 (acccessed 2019/09/08). (55) Hu, J.; Li, Q.; Wang, X.; Shao, S.; Wang, L.; Jing, X.; Wang, F. Developing Through-Space Charge Transfer Polymers as a General Approach to Realize Full-Color and White Emission with Thermally Activated Delayed Fluorescence. Angew. Chem. Int. Ed. 2019, 58 (25), 8405-8409. DOI: 10.1002/anie.201902264 (acccessed 2020/10/11). (56) Luo, Y.; Li, S.; Zhao, Y.; Li, C.; Pang, Z.; Huang, Y.; Yang, M.; Zhou, L.; Zheng, X.; Pu, X.; et al. An Ultraviolet Thermally Activated Delayed Fluorescence OLED with Total External Quantum Efficiency over 9%. Adv. Mater. 2020, 32 (32), 2001248. DOI: 10.1002/adma.202001248 (acccessed 2020/10/11). (57) Izumi, S.; Higginbotham, H. F.; Nyga, A.; Stachelek, P.; Tohnai, N.; Silva, P. d.; Data, P.; Takeda, Y.; Minakata, S. Thermally Activated Delayed Fluorescent Donor–Acceptor–Donor– Acceptor π-Conjugated Macrocycle for Organic Light-Emitting Diodes. Journal of the American Chemical Society 2020, 142 (3), 1482-1491. DOI: 10.1021/jacs.9b11578. (58) Gernert, M.; Müller, U.; Haehnel, M.; Pflaum, J.; Steffen, A. A Cyclic Alkyl(amino)carbene as Two-Atom π-Chromophore Leading to the First Phosphorescent Linear CuI Complexes. Chem. Eur. J. 2017, 23 (9), 2206-2216. DOI: https://doi.org/10.1002/chem.201605412. (59) Jones, P. F.; Calloway, A. R. Temperature effects on the intramolecular decay of the lowest triplet state of benzophenone. Chem. Phys. Lett. 1971, 10 (4), 438-443. DOI: 10.1016/0009- 2614(71)80329-9. (60) Strickler, S. J.; Berg, R. A. Relationship between Absorption Intensity and Fluorescence Lifetime of Molecules. The Journal of Chemical Physics 1962, 37 (4), 814-822. DOI: 10.1063/1.1733166. (61) Chen, T.; Zheng, L.; Yuan, J.; An, Z.; Chen, R.; Tao, Y.; Li, H.; Xie, X.; Huang, W. Understanding the Control of Singlet-Triplet Splitting for Organic Exciton Manipulating: A Combined Theoretical and Experimental Approach. Sci. Rep. 2015, 5, 10923. DOI: 10.1038/srep10923. (62) Sheldrick, G. Crystal structure refinement with SHELXL. Acta Crystallographica Section C 2015, 71 (1), 3-8. DOI: doi:10.1107/S2053229614024218. 164 (63) Shao, Y.; Gan, Z.; Epifanovsky, E.; Gilbert, A. T. B.; Wormit, M.; Kussmann, J.; Lange, A. W.; Behn, A.; Deng, J.; Feng, X.; et al. Advances in molecular quantum chemistry contained in the Q-Chem 4 program package. Molecular Physics 2015, 113 (2), 184-215. DOI: 10.1080/00268976.2014.952696. (64) Hermann, G.; Pohl, V.; Tremblay, J. C.; Paulus, B.; Hege, H.-C.; Schild, A. ORBKIT: A modular python toolbox for cross-platform postprocessing of quantum chemical wavefunction data. J. Comput. Chem. 2016, 37 (16), 1511-1520. DOI: https://doi.org/10.1002/jcc.24358. (65) Python wrapper for Cubature: adaptive multidimensional integration; 2021. (accessed. (66) Yang, J.-G.; Song, X.-F.; Wang, J.; Li, K.; Chang, X.; Tan, L.-Y.; Liu, C.-X.; Yu, F.-H.; Cui, G.; Cheng, G.; et al. Highly Efficient Thermally Activated Delayed Fluorescence from Pyrazine- Fused Carbene Au(I) Emitters. Chemistry – A European Journal 2021, 27 (71), 17834-17842. DOI: 10.1002/chem.202102969 (acccessed 2022/03/11). (67) Karon, K.; Lapkowski, M. Carbazole electrochemistry: a short review. J. Solid State Electrochem. 2015, 19 (9), 2601-2610. DOI: 10.1007/s10008-015-2973-x. (68) Tzouras, N. V.; Martynova, E. A.; Ma, X.; Scattolin, T.; Hupp, B.; Busen, H.; Saab, M.; Zhang, Z.; Falivene, L.; Pisanò, G.; et al. Simple Synthetic Routes to Carbene-M-Amido (M=Cu, Ag, Au) Complexes for Luminescence and Photocatalysis Applications. Chemistry – A European Journal 2021, 27 (46), 11904-11911. DOI: 10.1002/chem.202101476. (69) Weinstein, C. M.; Junor, G. P.; Tolentino, D. R.; Jazzar, R.; Melaimi, M.; Bertrand, G. Highly Ambiphilic Room Temperature Stable Six-Membered Cyclic (Alkyl)(amino)carbenes. J. Am. Chem. Soc. 2018, 140 (29), 9255-9260. DOI: 10.1021/jacs.8b05518 From NLM PubMed-not- MEDLINE. (70) Hilborn, R. C. Einstein coefficients, cross sections, f values, dipole moments, and all that. Am. J. Phys. 1982, 50 (11), 982-986. DOI: 10.1119/1.12937. (71) Shen, W.; Li, J.; Zhang, C.; Shi, M.; Zhang, J. Copper, Silver and Sodium Salt-Mediated Quaternization by Arylation: Syntheses of N-Heterocyclic Carbene Precursors and 6-H- Phenanthridine Derivatives. Chem. Asian J. 2016, 11 (13), 1883-1886. DOI: 10.1002/asia.201600563 From NLM PubMed-not-MEDLINE. (72) Tang, R.; Xu, S.; Lam, T.-L.; Cheng, G.; Du, L.; Wan, Q.; Yang, J.; Hung, F.-F.; Low, K.-H.; Phillips, D. L.; et al. Highly Robust CuI-TADF Emitters for Vacuum-Deposited OLEDs with Luminance up to 222 200 cd m−2 and Device Lifetimes (LT90) up to 1300 hours at an Initial Luminance of 1000 cd m−2. Angewandte Chemie International Edition 2022, 61 (33), e202203982. DOI: https://doi.org/10.1002/anie.202203982. 165 4 Chapter 4: Two-Coordinate Coinage Metal Complexes as Photosensitizers for Solar Fuels Solar fuels is a renewable energy technology that uses energy emitted by the sun to directly convert abundant molecules into fuel (i.e. converting 2H2O to 2H2 + O2). The task of photo- converting a molecule into fuel is typically accomplished by pairing two types of molecules that work in concert, namely a “photosensitizer” and an “electrocatalyst”. Electrocatalysts are able to convert a target molecule into a valuable product (Carbon dioxide to carbon monoxide or formic acid, methane to methanol, water to hydrogen, etc.). 1-5 However, an electrocatalyst is only active once it is charged. A photosensitizer has the ability to absorb light and transfer charge onto a target molecule. Thus, a photosensitizer can be used to collect light from the sun and transfer charge onto an electrocatalyst, which then becomes active for converting a target molecule into fuel. The most common photosensitizers are organometallic complexes that are based on the rarest metals on Earth such as platinum, ruthenium, and iridium. These scarce metal complexes have excellent figures of merit for photosensitzer chemistry (see more detail in section 4.1), but cannot be scaled up for implementation as a renewable energy feedstock. However, this technology could be scalable if photosensitizers based on abundant metals were designed to have similar performance to the scarce metal complexes. Successful efforts have been demonstrated with nickel, copper, and zinc. 6 Among these abundant metals, copper has been identified as an outstanding candidate for molecular photosensitizers. The few reported copper sensitizers in the literature are all four- coordinate complexes. Four-coordinate complexes have concomitant restrictions in their use as photosensitizers including susceptibility to Jahn-Teller distortion and a restrictive ligand design (see section 4.4 for more details). 166 In this work, we made the first evaluation of two-coordinate, carbene metal amide complexes based on copper, silver, and gold as solar photosensitizers to convert water into hydrogen. Three new molecules were designed and their photophysical properties were evaluated in a variety of solvents to determine the structure-function properties of this class of molecules. We established the design principles for the carbene ligand, the amide ligand, and we determined the role of the metal. We find that these complexes offer competitive figures of merit in comparison to ruthenium and iridium complexes and offer a higher degree of photophysical tunability than the four-coordinate copper sensitizers. 4.1 Solar Photosensitizer Design Principles The first important design principle for a solar photosensitizer is the ability to efficiently absorb photons emitted by the sun. The emission spectrum of the sun shows that the highest intensity occurs near the blue end of the visible spectrum (Figure 4.1), and relatively low intensity in the ultra-violet and infrared regimes. One might initially think that the absorption spectrum of the ideal solar photosensitizer should be catered to the emission spectrum shown in Figure 4.1 for minimal waste of light. There are several issues with this logic. First, designing a molecule with an absorption spectrum that perfectly matches the emission spectrum of the sun is incredibly challenging. Second, and more importantly, if the lowest energy absorption is in the infrared, then the excited state will only have ≤ 0.9 V to drive photocatalysis upon relaxation to the S1/T1 manifolds. However, designing molecules that efficiently absorb blue light (λmax,em), corresponding to the S0→S1 transition, ensures that significant light is captured from the sun and the excited state will have ~2.8 V to drive photocatalysis. In other words, the goal is to optimize the extinction coefficient near the blue edge of the visible spectrum. 167 Figure 4.1 The solar emission spectrum. 7 Photosensitizers absorb photons and use their excited state to perform intermolecular charge or energy transfer to a target molecule. The target molecule is general and is often referred to as a “quencher” because it quenches the excited state by accepting charge or energy. The two mechanisms are shown below. 𝑃𝑆 ∗ + 𝑄 → 𝑃𝑆 +/− + 𝑄 −/+ 𝐸𝑞 .4.1 Eq. 4.1 describes photoredox from the excited photosensitizer (PS*) to the target molecule (Q). 𝑃𝑆 ∗ + 𝑄 → 𝑃𝑆 + 𝑄 ∗ 𝐸𝑞 .4.2 Eq. 4.2 describes energy transfer from PS* to Q. There are no selection rules for photoredox, but energy transfer must conserve spin. In other words, 1 PS* can only generate 1 Q*, and 3 PS* can only generate 3 Q*. In solar fuels, it is important that the photosensitizer performs photoredox to the quencher instead of energy transfer, because an electrocatalyst is only activated by charge, not by energy transfer. 168 There are several ways to understand the driving force of photoredox in Eq. 4.1 with varying degrees of accuracy. The simplest way to think about the driving force is with a 0 th order model that only considers the HOMO and LUMO of the PS and the Q molecules shown in Figure 4.2. It assumes no change in the MO energies between PS and PS*, and only accounts for a change in the occupancy. Figure 4.2 A 0 th order approximation of the energetic driving force of photo-reduction (a) and photo-oxidation (b). HO represents HOMO and LU represents LUMO. Note that the HO and LU of PS* are no longer accurate labels in the excited state, but they are used here for simplicity. The PS* has a vacancy in the HOMO which can accept electrons, and an electron in the LUMO (really the SOMO) which means it is both easier to reduce and oxidize than the ground state PS. This is because the first electron that will be ionized is no longer in the HOMO, rather in the higher lying LUMO, and it is much easier to put an electron in the lower laying HOMO than it is the higher laying LUMO. If the energy of the quencher LUMO is lower than that of the PS*, then it is energetically downhill for the electron to move from the PS* LUMO to the quencher LUMO with an energetic driving force of Ered (Figure 4.2a). Note that the HOMO of the quencher in Figure 4.2a is fully occupied and lower in energy than the HOMO of PS*, which prevents reduction of PS*. Similarly, photo-oxidation of the quencher is shown in Figure 4.2b where it is energetically 169 favorable for an electron to move from the quencher HOMO to the PS* HOMO. Note that in both cases, the HOMO-LUMO energy gap is larger for Q than it is for PS*, which ensures that energy transfer from PS* to Q is unfavorable as it violates the conservation of energy. This model in Figure 4.2 is useful for developing an initial picture for photoredox, but this model suffers from many oversimplifications. For example, this model predicts that the excited state oxidation potential (E +/ *) is equal to the ground state reduction potential (E 0/- ), and that the excited state reduction potential (E* /- ) is equal to the ground state oxidation potential (E +/0 ). This of course cannot be true because the ground and excited states have different molecular geometries and charge distributions, and different energetic perturbation from solvent/spin. A more accurate treatment of photoredox can be achieved by incorporating the experimentally determined molecular state transitions from the absorption and emission spectra to replace the gas phase HOMO-LUMO energy gap. Furthermore, the absolute energies of the HOMO and LUMO should be replaced with the ground state redox potentials because they incorporate solvent effects and provide a reference potential. Such a model is known as a “Latimer Diagram” and is shown below in Figure 4.3. 170 Figure 4.3 A Latimer diagram showing the relationship between the ground state and excited state redox potentials. E00 represents the S0→S1 transition energy. The Latimer diagram predicts that the excited state oxidation and reduction potentials are given by Eq. 4.3 and 4.4. 𝐸 +/∗ = 𝐸 +/0 − 𝐸 00 𝐸𝑞 .4.3 𝐸 ∗/− = 𝐸 0/− + 𝐸 00 𝐸𝑞 .4.4 Note that E +/0 and E 0/- are typically measured from electrochemical experiments such as cyclic voltammetry, and E00 is the energy at the point of overlap between the lowest energy absorption highest energy emission band (to best estimate the energy of the S0 →S1 transition). This treatment also changes the reference from vacuum to the reference electrode used to determine the ground state redox potentials. The free energy of photoredox depends on the excited state redox potentials of the photosensitizer and the ground state redox potentials of the quencher as shown in Eq. 4.5 and 4.6. ∆𝐺 = −𝑛𝐹 (𝐸 𝑄 0/− − 𝐸 𝑃𝑆 +/∗ ) 𝐸𝑞 .4.5 171 ∆𝐺 = −𝑛𝐹 (𝐸 𝑄 +/0 − 𝐸 𝑃𝑆 ∗/− ) 𝐸𝑞 .4.6 A visual diagram is shown below vs. the Fc +/0 couple in Figure 4.4. Figure 4.4 An accurate diagram of energetically favorable (a) and unfavorable (b) photoredox between the excited state photosensitizer (PS*) and the quenching molecule (Q). The axis here represent potential vs. Fc +/0 . Figure 4.4 is an alternative way to understand excited state photoredox between PS* and Q. Plotting this on the potential axis is a convenient way to visualize that electron transfer proceeds from the higher laying state to the lower laying state, similar to the picture in Figure 4.2. This figure demonstrates that a potent photoreducing agent has a very negative E +/ *, whereas a potent photo-oxidant has a very positive E* /- . Figure 4.4b shows that photoredox does not proceed when the reduction potential of Q is more negative than the excited state oxidation potential of the photosensitizer, or when the oxidation potential is more positive than the excited state of the excited state reduction potential of the photosensitizer. Another important design principle for photosensitizers is stability of PS + and PS - . Consider the photo-oxidation of PS* to generate PS + and Q - . If PS + is unstable, then it will decompose before 172 it can return to the ground state PS. Decomposition of PS + would halt the photocatalytic activity because the PS would be consumed as equivalents of fuel are generated. The same is true for stability of PS - in photo-reduction of PS* to generate Q + and PS - . Stability of the cation and anion forms of the PS can be evaluated with cyclic voltammetry (CV). In CV, an analyte molecule (A) is dissolved in an electrolyte solution connected to electrodes. In an oxidation experiment, the potential at the working electrode is increased until electrons are pulled off of the analyte resulting in oxidation (formation of A + ). The current generated by oxidizing “A” is measured. Next, the potential is decreased until the electrons on the electrode have an energetic driving force to return to A + and reform the neutral analyte “A”. If A + is stable, then one expects to measure the same amount of current going back into A + as was initially removed from A. This is known as a reversible oxidation. However, if A + is not stable and reacts or decomposes to a biproduct (B), then the current that was initially removed from A cannot be returned. Thus, no current will be measured as the potential is decreased. This is known as irreversible oxidation. Reversible and irreversible oxidation in a CV experiment are illustrated below in Figure 4.5. 173 Figure 4.5 A schematic for reversible (a) and irreversible (b) oxidation in cyclic voltammetry. The excited state lifetime () of the PS is an important design parameter because PS* must persist long enough to diffuse over to the electrocatalyst and transfer charge. If the radiative rate (kr) and/or non-radiative rate (knr) of the molecule is fast, then PS* will be too short-lived to drive photocatalysis. Another way of thinking about this is that the rate of electron transfer (keT) must be much larger than kr or knr for efficient photoredox. A simple kinetic scheme for photo-oxidation of PS* to the electrocatalyst quenching molecule is shown below in Figure 4.6. 174 Figure 4.6 Kinetic scheme of the deactivation of the excited state photosensitizer (PS*) via normal radiative/non-radiative decay (kr + knr) or photoreducing the electrocatalyst (Q) with rate constant keT. The kinetic scheme outlined in Figure 4.6 incorporates all of the rate constants mentioned above. This is represented mathematically in Eq. 4.7 & 4.8 𝑑 [𝑃𝑆 ∗ ] 𝑑𝑡 = −(𝑘 𝑟 + 𝑘 𝑛𝑟 )[𝑃𝑆 ∗ ] − 𝑘 𝑒𝑇 [𝑃𝑆 ∗ ][𝑄 ] 𝐸𝑞 .4.7 which is simplified to: 𝑑 [𝑃𝑆 ∗ ] 𝑑𝑡 = −((𝑘 𝑟 + 𝑘 𝑛𝑟 )+ 𝑘 𝑒𝑇 [𝑄 ])[𝑃𝑆 ∗ ] 𝐸𝑞 .4.8 The expression in Eq. 4.8 gives an integrated rate law with a rate constant k r + knr + keT[Q] and shows that the rate of electron transfer depends on the quencher concentration. This can be conceptually reasoned by considering a solution of PS and Q. If Q is at high concentration, the average distance between Q and PS will be lower in a homogenously distributed solution, which means that it will take PS* less time to diffuse to Q to perform photoredox. Thus, the rate of electron transfer increases with [Q]. The rate constant keT depends on the thermodynamics of electron transfer that was discussed in the previous section. If PS* is a very potent photoreducing agent, there will be a large driving force for electron transfer which will manifest as a high keT. The work described later in the chapter will show that keT has a maximum value of ~1 × 10 10 M -1 s -1 which corresponds to diffusion-limited electron transfer (when the slowest rate of photo-redox is 175 intermolecular diffusion). Table 4.1 includes different electrocatalyst concentrations and the corresponding rate of electron transfer based on a highly potent photo-reducing agent with keT = 1 × 10 10 M -1 s -1 . Table 4.1 Table of electron transfer rates and corresponding lifetimes at various concentrations. [Q] (mM) keT[Q] (10 9 s -1 ) teT (ns) 0.01 0.0001 10,000 0.1 0.001 1000 1 0.01 100 10 0.1 10 100 1 1 1000 10 0.1 Reasonable electrocatalyst concentrations are in the 10 M - 10 mM regime, which corresponds to lifetime constant of 10 s – 10 ns respectively. Thus, PS* should have an excited state lifetime of at least 10 ns for photoredox to be competitive with concomitant radiative and non-radiative relaxation. A quenching efficiency can also be calculated by comparing the electron transfer rate to all deactivating processes as shown in Eq. 4.9 𝛷 𝑒𝑇 = 𝑘 𝑒𝑇 [𝑄 ] 𝑘 𝑟 + 𝑘 𝑛𝑟 + 𝑘 𝑒𝑇 [𝑄 ] 𝐸𝑞 .4.9 The kr and knr for cMa complexes that perform TADF in solution are typically on the order of 10 6 s -1 -10 7 s -1 which corresponds to  = 1-0.1 s. Thus, if the concentration of [Q] is at 1 mM, eT will be ~50% efficient. Efficiencies > 90% are achieved at [Q] ≥ 10 mM. Photoredox efficiencies 176 ≥ 90% can be achieved at even lower concentrations of electrocatalyst with increased  of the photosensitizer. For example, if  = 100 s and keT = 10 10 M -1 s -1 , then [Q] can be as low as 20 M and still have eT = 91%. This further confirms that a longer excited state lifetime is a desirable figure of merit for photosensitizers. 4.2 Existing Photosensitizers: Case Studies Section 4.1 identified the important design principles for photosensitizers including high molar absorptivity in the visible spectrum, potent photoredox potentials (E +/ *, E* /- ), stabile ground state cationic (PS + ) and/or anionic (PS - ) species, and long excited state lifetimes ≥ 10ns. Case studies of complexes that satisfy most of these requirements are organized in Table 4.2. Scarce metal complexes such as Ru(bpy)3 2+ [bpy = 2,2’-bipyridine] and Ir(ppy)3 [ppy = 2-phenylpyridine] were identified in the late 1970s as promising photosensitizers due to their exceptional stability and long excited state lifetimes. 8 Platinum photosensitizers such as Pt( t Bu3tpy)([CtCC6H4]H [tpy = 2,2';6',2"-terpyridine] were also developed by the Castellano group which were successfully applied to photo-reduction of water to make H2. 9 Tri-coordinate gold systems such as Au III (H^Cnp^N^Cnp)(NHC)](OTf) [HCnp^N^CnpH = 4-R-2,6-di-naphthalen-2-yl-pyridine; R = C6H4-4-OCH3: NHC = 1,3-dimethylimidazol-2-ylidene) have been employed in oxidation of secondary amines, but suffer from absorption that is restricted to UV photons. 10 177 Table 4.2 Photosensitizer figures of merit for reported organometallic complexes. Entries with (--) were not reported in the literature source. Photosensitizer (10 3 M -1 cm- 1 )/ max(nm) E +/ * (V vs. SCE) E* /- (V vs. SCE)  (s) Ref Ru(bpy)3 2+ 20/450 -0.81 0.77 1.1 11 Ir(ppy)3 20/373 -1.61 0.59 2.0 12 Ir(ppy)2(bpy) 10/400 -0.85 0.68 0.38 13 Pt( t Bu3tpy) ([CtCC6H4]H 4/400 -- 1.50 0.51 9 Au III (H^Cnp^N^Cnp) (NHC)](OTf) 16/410 -- 1.20 0.6 10 ZnTMPyP 15/890 -0.8 0.5 1300 14, 15 [Cu(dsbtmp)2](PF6) 8/440 -1.0 -0.5 2.0 16 Noble metal free photosensitizers have also been developed such as ZnTMPyP [TMPyP = 5,10,15,20-tetra-p-N-methylpyridinio-chloride] which makes use of the macrocyclic porphyrin ligand to stabilize a ligand to metal charge transfer excited state with  = 1.3 ms. The excited state lifetime of Zn porphyrin complexes are long and the absorption energy extends through the visible into the NIR with notable ~10 4 M -1 cm -1 . The low energy absorption features directly lower the energy of the excited state which yields lower excited state redox potentials, but they have still been incorporated in solar fuels production. 17 The McMillin and Castellano groups have invested significant efforts in optimizing four-coordinate Cu(I) photosensitizers such as copper porphyrin and phosphine complexes. 16, 18-21 One highlighted complex is [Cu(dsbtmp)2](PF6) [dsbtmp = 2,9- di(sec-butyl)-3,4,7,8-tetra-methyl-1,10-phenanthroline], which has an excited state lifetime in the microsecond regime and absorbs in the blue region of the visible spectrum. 16 While examples of four-coordinate Cu(I) complexes have been demonstrated in the literature, two-coordinate Cu(I) complexes have yet to be considered. The following chapters will highlight the work that I did 178 with Claire Archer, Jack Applebaum, Anushan Alagaratnam, and Jonas Schaab to evaluate cMas as photosensitizers and to compare them to four-coordinate Cu(I) systems. 4.3 Exploring Two-Coordinate Coinage Metal Complexes as Photosensitizers Herein we report two-coordinate carbene-metal-amide (cMa, M = Cu(I) and Au(I)) complexes can be used as sensitizers to promote the light driven reduction of water to hydrogen. The cMa complexes studied here absorb visible photons (vis > 10 3 M -1 cm -1 ), maintain long excited state lifetimes ( ~ 0.2-1 s) and perform stable photo-induced charge transfer to a target substrate with high photoreducing potential (E +/ * up to -2.33 V vs. Fc +/0 based on a Rehm-Weller analysis). We pair these coinage metal complexes with a cobalt-glyoxime electrocatalyst to photocatalytically generate hydrogen and compare the performance of the copper- and gold-based cMa complexes. We also find that these two-coordinate complexes presented can perform photo- driven hydrogen production from water without the addition of the cobalt-glyoxime electrocatalyst. In this “catalyst free” system the cMa sensitizer partially decomposes to give metal nanoparticles that catalyze water reduction. This work identifies two-coordinate coinage metal complexes as promising abundant metal, solar fuels photosensitizers that offer exceptional tunability and photoredox properties. 4.4 Introducing Two-Coordinates Generating fuels via solar photocatalysis is a highly active research field in modern chemistry. 22, 23 As mentioned in section 4.1, molecular light-driven chemistry requires two primary processes: light absorption and chemical interaction with a substrate to store the light energy in the form of a fuel. These tasks are often divided between two different molecules. Light absorption occurs at the photosensitizer (PS), which transfers charge onto an electrocatalyst that performs the 179 chemical transformation. Complexes such as Ru(bpy)3 2+ and Ir(ppy)3 are commonly used photosensitizers owing to their high molar absorptivity ( ~10 4 M -1 cm -1 ), long-lived excited states, and their high stability in oxidized and reduced forms. 24, 25 These properties enable the complexes to efficiently deliver charge to a variety of electrocatalysts from their excited state. The main drawback of Ru/Ir compounds is that they are among the rarest metals on earth, which precludes their use in scalable and sustainable energy technology. 26 Chemists have responded to the challenge of developing sustainable solutions by exploring metal complexes of earth abundant metals as viable alternatives. Significant efforts have been put into first row transition metals, among which copper has been a rising candidate to match the performance of scarce metals. 6, 27, 28 The few reported classes of copper(I) photosensitizers use complexes with a four-coordinate structure. Prominent classes pioneered by the McMillin group are cationic Cu(I)(N^N)2 or Cu(I)(N ^ N)(P^P) complexes (where N ^ N is a bis-imine ligand, such as a phenanthroline or bipyridine derivative, and P^P = is a bis-phosphine derivative). 19, 29, 30 These complexes display visible absorption assigned to metal-to-ligand charge-transfer (MLCT) transitions that have molar absorptivities greater than 10 3 M -1 cm -1 and lifetimes for emission that vary from 10 2 to 10 5 ns. Compounds such as Cu(dmp)2 + (dmp = 2,9-dimethyl-1,10-phenanthroline) and Cu(dap)(Xantphos) (dap = 2,9-bis(4-methoxyphenyl)-1,10-phenanthroline, Xantphos = (9,9- dimethyl-9H-xanthene-4,5-diyl)bis(diphenylphosphane)) have been successfully employed in photoredox catalysis. 19, 20, 28, 31-34 However, an inherent restriction in the photophysical properties in these Cu(I) derivatives is Jahn-Teller distortion of the MLCT state in fluid solution. 35 The accompanying rapid non-radiative decay will often lead to short lifetimes of the excited state, thereby diminishing the effectiveness of the compounds as photosensitizers. Significant efforts from the McMillin and Castellano groups have shown that installing bulky functional groups at 180 the 2,9-position of phenanthroline and methylating in the 3,4,7,8-positions significantly inhibits Jahn-Teller distortion, extending the excited state lifetime to the microsecond regime. 18, 21, 36-38 Two-coordinate coinage metal complexes with carbene-metal-amide (cMa) structure have recently been studied for applications in organic LEDs but have not been investigated for photoredox catalysis. 11, 39-48 These two-coordinated compounds have several attributes that are advantageous to their application as photosensitizers. The lowest energy absorption band of cMa complexes is typically an interligand charge transfer (ICT) transition with a hole localized on the carbazole and an electron on the carbene. Thus, the absorption energy is easily tuned throughout the visible spectrum by the choice of either carbene or amide, which is of considerable benefit in comparison to the four-coordinate complexes where the energy of the MLCT state is primarily adjusted by varying the energy of the ligand-localized LUMO. 39, 42, 43, 49 The redox potentials of the cMa complexes can also be varied over a wide range by judicious selection of the carbene and substituents on the amide. Finally, the ICT state is not susceptible to Jahn-Teller distortion, thus leading to low rates for non-radiative decay and high photoluminescence efficiency. However, it is valid to question whether these two-coordinate complexes have sufficient stability to serve as effective photosensitizers. The ligands are monodentate making them kinetically labile. The molecules are also coordinatively unsaturated, which could allow for deleterious non-radiative decay of the excited state via solvent coordination. Solvent coordination and exciplex formation is a well-documented issue in Cu(I) complexes that results in a decrease of excited state lifetime even for the sterically protected four-coordinate complexes. 16 181 Figure 4.7 Structure of the three photosensitizers investigated in this work: M RCz MAC [M = Cu or Au, R = tert- butyl (B) or phenyl (Ph)] (iPr = isopropyl). M Cz MAC were reported previously and were used for comparison in this study. 43 In this work, we probe the excited state photoredox capabilities of several two-coordinate cMa complexes in a variety of solvents to evaluate the suitability of these compounds as a new class of Cu(I) and Au(I) photosensitizers. The carbene and carbazole ligands for the complexes were chosen here to ensure visible light absorption with long excited state lifetimes (Figure 4.7) While all of the complexes are stable to electrochemical reduction, it is well known that carbazole is unstable to oxidation as it oligomerizes at the 3,6-positions. 50 Thus, alkyl or aryl substituents at the 3,6-sites were installed to increase the stability of the oxidized cMa complex. Oxidation potentials for the excited states (E +/ *) were estimated from electrochemical and steady state photophysical measurements and confirmed using a Rehm-Weller analysis for each photosensitizer. The role of the metal was investigated by comparing properties of Cu(I) and Au(I) derivatives with a common amide (3,6-di-tert-butylcarbazole), and that of the amide was probed by varying the substituent groups in Cu RCz MAC (R = tert-butyl or phenyl). We find that two-coordinate cMa complexes are a viable new class of photosensitizers for solar photochemistry, demonstrating the photo-driven production of hydrogen from water, both with and without an added cobalt- glyoxime electrocatalyst. 182 4.5 Experimental: Synthesis All reactions were performed under a N2 atmosphere using a Schlenk line. Tetrahydrofuran (THF), dichloromethane, and toluene were dried over alumina using a solvent purification system by SG Water USA, LLC. Dimethylformamide (DMF) and acetonitrile (MeCN) were purchased as dry solvents from VWR. Precursor complexes 1,3-bis(2,6-diisopropylphenyl)-5,5-dimethyl-4- oxo-3,4,5,6-tetrahydropyrimidin-1-ium-2-ide M(I) chloride (M Cl MAC [M = Cu(I) or Au(I)]) were prepared according to literature precedent. 43 The amide precursors 3,6-di-tert-butyl-9H-carbazole and 3,6-di-phenyl-9H-carbazole were purchased from AK Scientific and Tokyo Chemical Industry, respectively. Sodium tert-butoxide was purchased from Sigma Aldrich. Reference compounds Cu Cz MAC and Au Cz MAC were prepared according to literature procedures. 43 All purchased chemicals were used without further purification. The general synthesis of all reported compounds involved adding solid carbazole (or substituted carbazoles) and NaO t Bu to a freshly dried Schlenk-flask connected to a Schlenk line and a nitrogen atmosphere was established. THF was injected against positive N2 pressure, and the mixture stirred at room temperature for one hour to generate free carbazolide. The M Cl MAC (M = Cu or Au) complex was next added against positive N2 pressure. The flask was then covered with aluminum foil and allowed to stir at room temperature overnight. All subsequent workup was performed in air. The workup involved filtering the reaction solution through Celite and removing the solvent with a rotary evaporator (rotovap). The crude solid was dissolved in 50/50 MeCN/hexane by volume, and product was extracted into the MeCN phase. The MeCN solution was evaporated on the rotovap to yield a sticky yellow gel. The gel was redissolved in 20/80 CH2Cl2/hexane by volume and the solvent removed on the rotovap, which converted the gel-like solid to a powder. The process of dissolving in 20/80 CH2Cl2/hexane and solvent removal was 183 repeated three times to yield a fine yellow powder. Lastly, the solid was washed copiously with either methanol or diethyl ether to yield analytically pure samples. The detailed synthesis and analytical data for each cMa sensitizer is given in section 4.18 4.6 Computational Parameters Density functional theory (DFT) and time dependent DFT (TDDFT) calculations were performed in QChem v5.0 to predict the structural and electronic properties of the cMa photosensitizers in the ground and excited states. Geometry optimizations were performed using the B3LYP method, LACVP basis, and Fit-LACVP effective core potential (ECP). 51 TDDFT calculations were performed using the LACVP basis, fit-LACVP ECP, and the CAM-B3LYP exchange (attenuation parameter = 0.2) with RPA enabled. Orbital composition analysis and wavefunction overlap of the HOMO and LUMO were evaluated using Multiwfn. 52 4.7 Electrochemical Methods Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements were performed with a VeraSTAT potentiostat using a glassy carbon working electrode, platinum counter electrode, and silver wire as a pseudo-reference electrode. Solutions were prepared in dry, N2 sparged solvents with 0.1M tetrabutylammonium hexafluorophosphate electrolyte concentration. CV and DPV measurements were performed with scan rates of 0.1 Vs -1 , and 10 mVs -1 , respectively. Measurements were repeated using ferrocene or decamethylferrocene as an internal standard; all potentials were referenced to the ferrocenium/ferrocene (Fc +/0 ) couple. 4.8 Photophysical Measurements Absorption spectra were recorded using a Hewlett-Packard 4853 diode array spectrometer. Photoluminescence (PL) spectra were recorded using a Photon Technology International 184 QuantaMaster spectrofluorimeter. Excited state lifetime measurements were performed using time correlated single photon counting (TCSPC) with an IBH Fluorocube apparatus interfaced to a Horiba FluoroHub+ controller. Quantum yield (QY) measurements were performed using a Hamamatsu C9920 integrating sphere with a xenon lamp excitation source. Samples for PL, TCSPC, and QY measurements were prepared in a custom made long, glass cuvette that allowed the solutions to be sparged with N2 and sealed with a Teflon valve. 4.9 Stern-Volmer Quenching Experiments Quenching of the excited state via photo-induced electron transfer from PS * to Q was measured by TCSPC. Deactivation of the excited state via energy transfer was avoided by choosing quenchers that had energies for their T1 state higher than that of the PS. Five solutions were prepared with a constant photosensitizer concentration (~10 M) and increasing quencher concentration. The optical spectra were measured to ensure that the absorption features of the PS did not change, showing little or no interaction of PS and Q is present in the ground state (no static quenching). Oxygen-free conditions were established by sparging solutions with N2 to prevent quenching of PS * by oxygen. Excited state lifetimes () were measured as a function of increasing quenching concentration ([Q]) and compared to the lifetime without quencher (0). The Stern- Volmer equation was used to extract the rate constant kq corresponding to quenching of the excited state. τ 0 τ = 1+ 𝑘 q τ 0 [Q] 𝐸𝑞 .4.10 Plots of  vs. [Q] yield a linear line with R 2 values near unity. For a given Stern-Volmer plot, the slope was divided by 0 to yield values for kq. 185 4.10 Photocatalysis Methods The cobalt-glyoxime electrocatalyst [Co(dmgH)2pyCl] was synthesized by Claire Archer according to literature procedure. 53 The sacrificial reductant 1,3-dimethyl-2-phenyl-2,3-dihydro- 1H-benzo[d]imidazole (BIH) was purchased from AmBeed . 54 Solutions were prepared in a N2- filled glovebox. A 50 mL Schlenk flask was charged with either Au BCz MAC or Cu PhCz MAC , Co(dmgH)2pyCl, and 224 mg BIH. All solids were dissolved in 5 mL of THF/water (5% water by volume). The flask was capped with a rubber septum. The flask was placed in front of a 470 nm LED (model number M470L5) with focusing optics (model number SM1U25-A & SM1RC) purchased from Thorlabs as shown below in Figure 4.8. Figure 4.8 Experimental Setup for photocatalysis The flask was wrapped in foil apart from a small area for the focused blue light to enter the reaction flask. The flask was placed ~6 inches from the LED with an incident power of 486 mW and the reaction was stirred with a magnetic stir bar. 186 Photocatalytic hydrogen production was quantified in Professor Surya Prakash’s lab with help from Anushan Alagaratnam using a Thermo Finnigan Trace GC 2000 with a thermal conductivity detector and nitrogen as a carrier gas. The GC internal temperature was 70 °C at the time of sample injection. The temperature was maintained for 7 minutes followed by an increase to 100 °C at a rate of 15 °C/min. The final temperature of 100 °C was maintained for two minutes. A calibration curve was created by injecting different volumes of 5% H2 in N2 gas using a Hamilton gastight 500 L syringe and integrating the area of the hydrogen peak in the chromatogram. A linear relationship with R 2 > 0.99 was established in the integrated area vs. volume of hydrogen injected as seen in Figure 4.9. 5 10 15 20 25 5 E6 10 E6 15 E6 20 E6 25 E6 Integrated Area (Signal) Volume H 2 injected (L) Equation y = a + b*x Intercept 0 ± -- Slope 1053930.4218 ± 13795.6870 4 R-Square (COD) 0.99932 Figure 4.9 GC H2 calibration curve. Samples were illuminated for a set amount of time to generate hydrogen. The LED was then turned off and gas was drawn from the photoreaction headspace through the rubber septum and injected into the GC. The integral of the hydrogen peak in the chromatogram was used to determine the 187 volume of hydrogen in the syringe based on the calibration curve. This quantity was related to the volume of hydrogen produced in the reaction, which allowed us to calculate turn over numbers (TONs) with respect to the electrocatalyst (moles H2/moles Co(dmgH)2pyCl) and/or sensitizer (moles H2/moles M amide MAC ). 4.11 Computational Results The electronic structure of the complexes in the ground state was examined using DFT calculations for the cMa complexes. All materials have their highest occupied molecular orbital (HOMO) localized on the carbazole, and lowest unoccupied molecular orbital (LUMO) localized on the carbene (Figure 4.10). Both orbitals show only a small contribution from the metal d orbital involved with -bonding to the coordinating atom of each ligand. The energies for the HOMO in the cMa complexes are altered by substituents on the carbazole ligand, whereas the energies for the LUMO remain effectively constant because the complexes have the same carbene ligand (Table 4.3). The energy of the HOMO for the Cu analogues is destabilized in the order Cu BCz MAC > Cu Cz MAC > Cu PhCz MAC , which is consistent with the electron donating nature of the t Bu group and electron withdrawing nature of the phenyl group. Changing the metal from copper to gold also stabilizes the HOMO, which is ascribed to the higher ionization potential of Au(I) relative to Cu(I). 55 The choice of metal ion does not affect the LUMO energies due to the nature of the dative carbene-metal bond compared to the relatively ionic M + - Cz - interaction. 188 Figure 4.10 The HOMO [red/blue] and LUMO [cyan/beige] wavefunctions (left), and hole [red] and electron [blue] NTO wavefunctions (right) of Cu BCz MAC displayed with isovalue = 0.1. Hydrogen atoms and the 3,6-diisopropylphenyl groups were omitted for clarity. The HOMO/LUMO densities of Cu PhCz MAC and Au BCz MAC are qualitatively the same. Table 4.3 Computational results of each photosensitizer. Here, f corresponds to the oscillator strength of the transition, and h+,e- is the overlap between the hole and electron wavefunctions. HOMO (eV) LUMO (eV) S0→S1 (eV/f) S0→T1 (eV) EST a (eV) % metal (HO/LU) h+,e- (S1) (%) Cu BCz MAC -4.08 -1.99 2.41/0.12 2.19 0.22 4.4/7.8 27 Cu Cz MAC -4.17 -1.99 2.48/0.11 2.26 0.22 4.6/7.8 26 Cu PhCz MAC -4.23 -2.08 2.51/0.14 2.32 0.19 3.3/7.8 25 Au BCz MAC -4.22 -1.98 2.59/0.19 2.33 0.26 4.1/9.0 31 Au Cz MAC -4.32 -1.99 2.66/0.16 2.41 0.25 4.3/9.1 36 Analysis of the excited states using TDDFT shows that the hole and electron wavefunctions in the S0 → S1 transition are well described by the HOMO and LUMO, respectively (Figure 4.10) . Localization of the HOMO and LUMO on the donor and acceptor ligands characterizes the interligand charge transfer excited state. The energy of the ICT state mirrors the trend in HOMO energy since the energy of the LUMO is largely invariant in these complexes. The energy of the 189 T1 state lies close to the S1 state because of the large spatial separation between the hole and electron centers of charge for two-coordinate cMa complexes. The small exchange energies, along with strong spin-orbit coupling imparted by the metal, are known to promote rapid T1 → S1 intersystem crossing and thermally assisted delayed fluorescence (TADF) in two-coordinate complexes with coplanar ligands. 40, 56 Comparing Cu BCz MAC and Cu Cz MAC to their gold analogues shows an increase in oscillator strength for the ICT transition for the gold analogs. This property is consistent with the trend in HOMO/LUMO overlap (h+,e-), which suggests the Au(I) metal center provides greater orbital overlap than Cu(I). This difference in overlap is interesting because the metal d orbital contributes a small, albeit important, point of overlap between the HOMO and LUMO. To further investigate this feature, the contribution of the metal atom in each orbital was computed using the Hirshfeld method. 52 The product of the metal orbital contribution to the HOMO and LUMO (%MHOMO × %MLUMO) is greater for the gold complexes (36-39%) with respect to their copper analogs (26-36%). This difference leads to an increase in oscillator strength for the ICT transition in the Au derivative. 4.12 Photophysical Properties The photophysical properties of the cMa complexes were examined in polar and non-polar solvents. All compounds display broad, featureless absorption bands in the visible spectrum which are assigned to an ICT transition (Figure 3). These bands have values for molar absorptivity ( = 4000-8000 M -1 cm -1 ) consistent with charge transfer transitions. 57-61 The energy of the ICT band blue-shifts in the order Cu BCz MAC < Cu PhCz MAC < Au BCz MAC . This hypsochromic shift is due to the stabilization of the HOMO in Cu PhCz MAC and Au BCz MAC , as discussed above. The Cu PhCz MAC complex has a higher molar absorptivity than Cu BCz MAC stemming from -extension by the aryl substituents of 3,6- 190 diphenylcarbazole. The molar absorptivity for Au BCz MAC is greater than that of Cu BCz MAC owing to the increase in NTO overlap induced by the more diffuse 5d orbitals of gold relative to 3d orbitals on copper (Table 4.3). Luminescence spectra for the cMa complexes are broad and featureless, with a large Stokes shift, and are assigned to TADF emission from an ICT state (Figure 4.11b). The quantum yields () are 0.17 and 0.15 for Cu BCz MAC and Au BCz MAC with excited state lifetimes () of 350 ns and 250 ns in THF, respectively. The decrease in lifetime for the gold complex is due to the simultaneous increase in the radiative and non-radiative rates (kr and knr, see Table 4.4). Replacing the tert-butyl substituents with phenyl groups in Cu PhCz MAC results in a decrease of knr which increases both the photoluminescence efficiency ( = 0.43) and lifetime ( = 460 ns). 350 375 400 425 450 475 500 0 2000 4000 6000 8000 Cu MAC BCz Cu MAC PhCz Au MAC BCz  (M -1 cm -1 ) Wavelength (nm) (a) 450 500 550 600 650 700 750 800 0.0 0.2 0.4 0.6 0.8 1.0 Intensity (Arb. U) Wavelength (nm) Cu MAC BCz Cu MAC PhCz Au MAC BCz (b) Figure 4.11 a) Molar absorptivity and (b) the emission of all compounds in THF. The complexes display intense UV absorption bands which are assigned to high energy -* transitions localized on the carbene and amide ligands. The absorption feature at 350nm is identical for the two di-tert-butylcarbazole complexes, which indicates that it is a feature of the carbazole ligand. The di-phenyl-carbazole analogue has different absorption features with higher , which is consistent with the enhanced conjugation from the phenyl rings. 191 250 300 350 400 450 500 550 0 10000 20000 30000 40000 Cu MAC PhCz Cu MAC BCz Au MAC BCz  (M -1 cm -1 ) Wavelength (nm) Figure 4.12 Molar absorptivity spectra in THF for all reported photosensitizers. Photophysical properties vary in solvents of varying polarity (toluene, THF, CH2Cl2, MeCN, and DMF). The solvent dependent absorption and emission spectra are shown below (Figure 4.13 - Figure 4.15). 192 350 400 450 500 550 0.0 0.2 0.4 0.6 Intensity (Arb. U.) Wavelength (nm) Tol THF DCM MeCN DMF (a) 500 550 600 650 700 750 0.0 0.2 0.4 0.6 0.8 1.0 Tol THF DCM MeCN DMF Emission (Arb. U.) Wavelength (nm) (b) Figure 4.13 Solvent dependent PL of Cu BCz MAC : (a) absorption, (b) emission. 350 400 450 500 550 0.0 0.2 0.4 0.6 0.8 1.0 Tol THF DCM MeCN DMF Absorbance (Arb. U.) Wavelength (nm) (a) 500 550 600 650 700 750 0.0 0.2 0.4 0.6 0.8 1.0 Tol THF DCM MeCN DMF Intensity (Arb. U.) Wavelength (nm) Au MAC BCz (b) Figure 4.14 Solvent dependent PL of Au BCz MAC : (a) absorption, (b) emission. 375 400 425 450 475 500 525 0.0 0.1 0.2 0.3 0.4 0.5 Tol THF DCM MeCN DMF Absorbance (Arb. U.) Wavelength (nm) (a) 500 550 600 650 700 750 0.0 0.2 0.4 0.6 0.8 1.0 Tol THF DCM MeCN DMF Intensity (Arb. U.) Wavelength (nm) Cu MAC PhCz (b) Figure 4.15 Solvent dependent PL of Cu PhCz MAC : (a) absorption, (b) emission. The absorption band for the ICT transition blue shift with increasing solvent polarity, whereas the emission spectra red shift. This solvatochromism has been previously explained by the large 193 change in dipole between the ground and excited state of these molecules. 41 Here we observe that solvatochromism is markedly more pronounced in the absorption spectrum than the emission spectrum. An important consideration for a photosensitizer is the lifetime of the excited state, which must persist long enough for the PS * to diffuse to an electrocatalyst and transfer charge. The diffusion limited rate constant of electron transfer (kq) plateaus at ~1x10 10 M -1 s -1 for highly endergonic electron transfer processes. 62-64 The total intermolecular electron transfer rate depends directly on the concentration of the quenching agent ([Q]), i.e. the electrocatalyst. Quencher concentrations of 1-10 mM will lead to quenching rates (kq = 1x10 7 s -1 to 1x10 8 s -1 ) that correspond to lifetimes for the quenched PS ( = 100-10 ns) that define lower limits for the useful effective lifetime of the photosensitizer (Section 4.1). 194 Table 4.4 Photophysical parameters of all reported complexes in toluene, THF, CH2Cl2, DMF, and MeCN. solvent max abs (nm) max em (nm)  (ns)  kr (10 5 s -1 ) knr (10 5 s -1 ) Cu BCz MAC Toluene     4.7 9.2 THF 435 603 350 0.17 4.7 24 CH2Cl2 425 608 190 0.10 5.2 47 DMF 402 607 130 0.04 3.1 74 MeCN 403 607 120 0.04 3.4 81 Cu PhCz MAC Toluene 445 556 710 0.70 9.9 4.2 THF 424 578 460 0.43 9.4 12 CH2Cl2 413 584 330 0.29 8.7 21 DMF 393 * 584 190 0.14 7.6 47 MeCN 393 * 584 190 0.16 8.4 44 Au BCz MAC Toluene 449 568 450 036 8.0 14 THF 427 584 250 0.15 6.0 34 CH2Cl2 421 600 130 0.08 6.0 70 DMF 398 598 75 0.05 6.7 130 MeCN 400 600 69 0.04 5.8 140 * Values cannot be accurately determined because the ICT absorption band overlaps with the ligand localized transitions at higher energy. The excited state lifetimes and quantum yields also vary as a function of solvent polarity as shown in Table 4.4. The excited state lifetimes for all photosensitizers decrease with increasing solvent polarity owing to increases in non-radiative decay. Complexes in toluene solution have the longest excited state lifetimes and the most intense absorption of visible light. The enhanced absorption of visible light due to the red shift of the ICT transition with decreasing solvent polarity. However, non-polar solvents can lead to additional kinetic barriers for ion separation after intermolecular charge transfer. Of note is the fact that the copper complexes have the longest emission lifetimes in any given solvent. This property can be advantageous in improving the photocatalytic efficiency 195 by making electron transfer quenching of the excited state (kq) a more competitive deactivation pathway than the non-productive processes (kr and knr). 4.13 Electrochemical properties Electrochemical measurements were performed to evaluate the ground state redox potentials and reversibility of the cationic and anionic species of the cMa complexes in a variety of solvents that are suitable for solar fuels production, i.e. THF, DMF and MeCN. 65-67 The redox potentials measured in CH2Cl2, THF, DMF, and MeCN are given in Table 4.5. The electrochemical reversibility is strongly influenced by the identity of the solvent, whereas the redox potentials are near solvent independent. Reduction is reversible for all cMa complexes across all solvents and the potential remains effectively constant. This behavior is consistent with the LUMO being localized on the MAC carbene ligand common to all the complexes. The trend in oxidation potentials also follows the trends computed for the energy of the HOMO. The value of E +/0 for Cu Cz MAC is greater than that of Cu BCz MAC and Cu PhCz MAC but less than that of the Au analog. 196 Table 4.5 Electrochemical potentials of cMa complexes in various solvents. Potentials are in Volts vs. Fc +/0 couple. See the manuscript for the full CV and DPV for all complexes. 68 E +/0 a E 0/- a E +/0 a E 0/- a CH 2 Cl 2 Cu BCz MAC 0.06 rev b THF Cu BCz MAC 0.06 rev -2.54 Cu PhCz MAC 0.17 rev b Cu PhCz MAC 0.12 rev -2.62 Au BCz MAC 0.13 rev b Au BCz MAC 0.18 rev -2.62 Cu Cz MAC 0.19 irr b Cu Cz MAC 0.18 rev -2.61 Au Cz MAC 0.22 irr b Au Cz MAC 0.28 rev -2.55 DMF Cu BCz MAC 0.06 irr -2.54 MeCN Cu BCz MAC 0.04 irr -2.51 Cu PhCz MAC 0.12 irr -2.54 Cu PhCz MAC 0.13 irr -2.52 Au BCz MAC 0.18 rev -2.52 Au BCz MAC 0.21 rev -2.47 * Reductions are outside of the solvent window. rev = reversible, irr = irreversible The cMa complexes with unsubstituted carbazolyl ligands (Cu Cz MAC and Au Cz MAC ) undergo irreversible oxidation in all solvents. This response is to be expected since the HOMO localized on carbazolyl ligand irreversibly oxidizes due to reaction after proton loss at the 3,6-positions. 50 Weak cathodic waves at ~ -0.25 V to 0.2 V for these cMa complexes that appear after the second cycle are thus assigned to oxidation of the oligomerization products as the features are absent in derivatives substituted on the 3,6-position. Hence, tert-butyl or phenyl groups at these sites leads to enhanced oxidative stability in CH2Cl2 and THF on the timescale of seconds. However, all the copper cMa complexes are irreversibly oxidized in MeCN and DMF. In contrast, Au BCz MAC maintains a notable return wave in all solvents which indicates greater stability of gold complexes as monocations compared to copper cMa complexes. These results may suggest that decomposition occurs by nucleophilic attack at the metal-amide linkage in MeCN, and that this bond is stronger for Au-amides than for the Cu analogs. Based on these experiments, we chose to perform quenching studies in THF because it sustains reversible oxidation and reduction for all 197 photosensitizers. Furthermore, THF gives the strongest absorptivity for cMa complexes among polar solvents, while also being sufficiently polar to facilitate efficient charge separation after electron transfer reactions. -0.50 -0.25 0.00 0.25 0.50 0.75 Current Potential vs. (V vs. Fc +/0 ) Cu MAC BCz Cu MAC PhCz Au MAC BCz Cu MAC Cz Au MAC Cz (a) -3.0 -2.5 -0.5 0.0 0.5 Cu MAC BCz Cu MAC PhCz Au MAC BCz Cu MAC Cz Au MAC Cz Current Potential vs. Fc + /Fc (V) (b) Figure 4.16 Cyclic voltammetry of all complexes in CH2Cl2 (a) and THF (b). The reduction event was outside of the solvent window for CH2Cl2 therefore, only the oxidative sweeps are shown here. 4.14 Redox Properties of the Excited State Molecules in the excited state are more potent oxidizing and reducing agents than in their ground state (Section 4.1). Redox potentials in the excited state are related to the redox potentials in the ground state and the energy gained by absorbing a photon. These potentials are approximated using Equations 4.3 and 4.4 (Figure 4.3). 69 Table 4.6 Ground and excited state redox potentials for the cMa complexes in THF. a E +/0 (V) E 0/- (eV) E00 (eV) E +/ * (V) E* /- (V) Cu BCz MAC 0.01 -2.65 2.44 -2.43 -0.21 Cu PhCz MAC 0.12 -2.65 2.57 -2.45 -0.08 Au BCz MAC 0.13 -2.60 2.51 -2.38 -0.09 a Potentials are referenced to internal Fc +/0 . 198 Using Equations 2 and 3, the cMa complexes are predicted to be potent photo-reducing agents, with near equal values of E +/ * (-2.4 V vs. Fc +/0 ), whereas being only mild photo-oxidants. The near degeneracy of E +/ * reflects the trend in increasing energy of E00 (Cu BCz MAC > Cu PhCz MAC > Au BCz MAC ) being offset by an increase in oxidation potential for each complex. This behavior is to be expected in a system where the energy of the S1 state is dictated by a transition involving the energy of a spatially separated HOMO and LUMO. Substituents that stabilize/destabilize the oxidation potential will correspondingly raise/lower the E00 energy. Thus, the value for E +/ * is largely governed by the choice of carbene. Analogously, the value for E* /- is approximately governed by the choice of metal and amide donor since these moieties determine the energy of the HOMO. Photo-oxidative quenching of the cMa complexes was therefore investigated due to their remarkably high E +/ * values. The thermodynamic driving force of quenching an excited state via electron transfer (G) depends on the values of E +/ * of the photosensitizer (E PS +/∗ ), E 0/- of the quencher (E Q 0/− ), and work terms (Ws) associated with coulombic interactions of separating charged species in the solvent medium (Eq. 4.11).  ∆𝐺 = (E PS +/∗ − E Q 0/− )+ W 𝑠 𝐸𝑞 .4.11 As the quenching molecule becomes harder to reduce, the rate of electron transfer is expected to decrease sharply when ∆G > 0. The initial step for electron transfer from the excited photosensitizer (PS * ) to a quenching molecule (Q) requires the two species to diffuse together to form an encounter complex [PS*,Q] with rate constant kd (Figure 4.17). This encounter complex can relax through concomitant radiative and non-radiative processes (kr and knr), or it can be quenched by transferring an electron or energy to the quencher. Here, we focus only on quenching 199 via charge transfer by choosing quenchers that have energies for their S1 and T1 states well above those of the cMa complexes, precluding quenching by energy transfer. Electron transfer occurs with a rate constant keT, and the resultant charge separated pair [PS + ,Q - ] can either recombine by back electron transfer with rate constant kbeT, or return to the ground state with rate constant kre. The charge separated pair can also diffuse away from each other as individual species (PS + + Q - ) with rate constant ks. Figure 4.17 Kinetic scheme for the oxidative quenching of a photosensitizer. Rehm and Weller showed that the net quenching rate constant kq given by Eq. 4.12 must consider the equilibrium which forms the encounter complex (kd/k-d = V), and charge recombination (kre): 𝑘 𝑞 = 𝑘 𝑑 1 + ( 𝑘 𝑑 ∆𝑉 12 × 𝑘 𝑟𝑒 )(exp( ∆𝐺 ‡ 𝑅𝑇 ) + exp( ∆𝐺 𝑅𝑇 )) 𝐸𝑞 .4.12 where R is the ideal gas constant, T is temperature, and G is given by Eq. 4.11. 71 The free energy term (G ‡ ) accounts for the reorganization energy of (PS * ,Q) to (PS + ,Q - ) and the solvent reorganization required for electron transfer (Gr). 200 ∆𝐺 ‡ = [( ∆𝐺 2 ) 2 + (∆𝐺 𝑟 ) 2 ] 1 2 + ∆𝐺 2 𝐸𝑞 .4.13 Rate constants of luminescence quenching for each complex were derived from Stern-Volmer analysis of experiments performed in THF using quenchers with varying reduction potentials; quinoxaline (QNX), N-methylpthalimide (NMeP), N,N-dimethyl-4-nitroaniline (DMNA), nitrobenzene (NB), and perinapthenone (PN). The reduction potentials for each quencher is given in Table 4.7 and plots of Stern-Volmer quenching data are given in the Section 4.17. An example plot is shown here for quenching the excited state of Au BCz MAC with NMeP in Figure 4.18. −1 0 1 2 3 4 5 6 7 8 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4   [Q] (mM)    =  + x 8 M -1 s -1  0 [Q] Au MAC BCz NMeP Figure 4.18 Stern-Volmer quenching if of Au BCz MAC with NMeP in THF. The Stern-Volmer data rigorously demonstrates that the excited state PS is quenched by NMeP, but it does not uniquely identify whether it is electron transfer or energy transfer. However, the lowest triplet energy of NMeP is ~3.0 eV and the 3 ICT of Au BCz MAC is ≤ 2.4 eV, thus energy transfer would be at least 500 meV uphill in energy and would violate energy conservation without additional perturbation to the system. 72 To further verify this, I collaborated with Steven 201 Bradforth’s group and performed transient absorption measurements of Au BCz MAC and NMeP in THF with Michael Kellogg and Austin Mencke. 73 This allowed us to spectroscopically track the process of photoredox with time resolution. The excited PS* quickly intersystem crosses to the 3 ICT followed by electron transfer from PS* to NMeP. This generates PS(+) and NMeP(-); only PS(+) has absorption features in the 400-900nm window. Thus, transient absorption measurements confirm that the quenching is from electron transfer because initial spectral signatures correspond to the triplet of Au BCz MAC , and the Au BCz MAC cation spectrum grows in as the triplet spectrum decays (Figure 4.19a). If the excited state PS was quenched by energy trasnfer, then Au BCz MAC (+) would not be observed. The absorption spectrum of Au BCz MAC (+) was obtained by Austin Mencke using pulse radiolysis at Brookhaven National Lab under advisorship from Mathew Bird. The cationic spectrum was further corroborated by Nina B. Reyes at the National Renewable Energy Lab (Figure 4.19b). 400 500 600 700 800 900 0.0 0.2 0.4 0.6 0.8 1.0 Absorption (OD) Wavelength (nm) V = 0 V = + V = - 605 nm 660 nm 730 nm Au MAC BCz (b) Figure 4.19 Transient absorption of Au BCz MAC and NMeP in THF excited at 355nm. The blue dashed line is the absorption spectrum of Au BCz MAC and the dot-dashed line is the Au BCz MAC cation spectra obtained from bulk elecrolysis (a). The cationic and anion absorption spectra of Au BCz MAC (b). 202 Table 4.7 Stern-Volmer quenching rate constants (kq) for the cMa complexes measured in THF. Reduction potentials of the quenching molecules were measured using differential pulse voltammetry and are reported vs. Fc +/0 . 68 E 0/- (V) kq(Au BCz MAC ) (10 9 M -1 s -1 ) kq(Cu BCz MAC ) (10 9 M -1 s -1 ) kq(Cu PhCz MAC ) (10 9 M -1 s -1 ) QNX -2.18 -- 0.064 0.14 NMeP -1.92 0.8 2.3 3.1 DMNA -1.90 2.1 2.8 3.5 NB -1.62 8.0 9.9 11 PN -1.50 9.5 11 11 Analysis of the Stern-Volmer data for a given quenching molecule shows that rate of kq is fastest for Cu PhCz MAC and slowest for Au BCz MAC (Table 4.7). This behavior mirrors the energy of E +/ * for the complexes estimated using Eq. 4.3. An increase in kq is also apparent as E 0/- for the quencher becomes less negative. This change is due to an increase in the driving force of electron transfer from PS * to Q (G ‡ ) with increasing electron affinity of the quencher. When the driving force becomes significantly large, kq plateaus at 1.1x10 10 s -1 , which corresponds to diffusion-limited quenching. In this regime, the slowest step in Scheme 1 is diffusion of PS* to Q to form the encounter complex (PS * ,Q). Thus, the value for kd in the Rehm-Weller fit was set to 1.1x10 10 s -1 . The reverse process k-d is typically near equivalent to kd, so V12 was set to unity. This yielded the following parameter values with less than 1% error (Table 4.8). Table 4.8 Parameters from Rehm-Weller fit of kq vs. E 0/- of the quenching molecule. E +/ * (V vs. Fc +/0 ) Gr (eV) WS (eV) kre (s -1 ) 𝐂𝐮 𝐁𝐂𝐳 𝐌𝐀𝐂 -2.28 0.25 0.1 4.0 x 10 3 𝐂𝐮 𝐏𝐡𝐂𝐳 𝐌𝐀𝐂 -2.33 0.25 0.1 4.0 x 10 3 𝐀𝐮 𝐁𝐂𝐳 𝐌𝐀𝐂 -2.23 0.25 0.1 4.0 x 10 3 The trend in excited state oxidation potential is consistent with the Stern-Volmer quenching experiments. For a given quenching molecule in a given solvent, the measured kq always increases 203 in the order of Au BCz MAC < Cu BCz MAC < Cu PhCz MAC . The Rehm-Weller analysis suggests that this is due to the difference in reductive potency of the excited state E +/ * which follows the same trend: Au BCz MAC < Cu BCz MAC < Cu PhCz MAC . The solvent reorganization term Gr is within the typical range compared to other reported systems. 24, 71 The value is consistent for all photosensitizers which can be explained by the similar structural motif of the photosensitizers; they are all two-coordinate carbene-metal amides. The solvent work term of 0.1 eV is reasonable and consistent between all photosensitizers which is to be expected, because WS is an intrinsic property of the solvent. 62 Lastly, kre is consistent between all photosensitizers and suggests the charge separated pair (PS + ,Q - ) recombines to the ground state (PS,Q) with a time constant of 250µs. The charge recombination of (PS + ,Q - ) was also investigated by looking at kinetic traces of the photoredox of Au BCz MAC and NMeP monitored by transient absorption spectroscopy (Figure 4.20). 400 500 600 700 800 900 -4 -2 0 2 4 6 8 6 ns 100 ns 12 ns 200 ns 50 ns 2000 ns 1.5 ns, psTA Absorption PR Cation 0 200 400 600 800 1000 0.0 0.2 0.4 0.6 Abs/PL Intensity (arb. u.) Time (s) 650 nm TA 650 nm PL 750 nm TA 750 nm PL Figure 4.20 Transient absorption of Au BCz MAC in THF with ~0.1 mM NMeP. 73 The initial time-trace at 6ns is assigned to the T1 of Au BCz MAC . This absorption decays as a new feature grows in with a peak max of 730nm. The Au BCz MAC (+) spectrum is long lasting; the kinetic trace at 750nm lasts decays with a time constant of ~ 300µs, which may be assigned to conversion of 204 (PS + ,Q - ) to (PS,Q). This is fairly consistent with the kre achieved from the Rehm-Weller fit, which yields a time constant of 250µs. Values of E +/ * for each complex were determined from fits of the quenching rates to Eq. 4.12 (Figure 4.21). Values for E +/ * can be obtained from fits to the data in Figure 4.21 (-2.28 V for Cu BCz MAC , -2.33 V for Cu PhCz MAC , and -2.23 V for Cu BCz MAC ). Thus, the trend in kq parallels the same trend in oxidation potential of the complexes (Cu PhCz MAC > Cu BCz MAC > Au BCz MAC ). This analysis suggests that the small difference in E +/ * values found for the three complexes is consistent with the predictions of Eq. 4.3, yet 100-170 meV lower than the estimates of -2.4 V. The disparity can be accounted for by considering that E00 used in Eq. 4.3 is based on the S0 → S1 transition. Intersystem crossing is extremely rapid in the cMa complexes (kISC > 10 9 s -1 ) and they are likely quenched out of the long-lived triplet excited state (T1). 43 The energy of the T1 state for Au Cz MAC was reported to be 90 meV lower than the S1 state, which accounts for most of the difference in E +/ *. 39 Similar EST values are expected for the other derivatives because experimental values for EST in cMa complexes do not vary significantly between Cu and Au congeners, and substituents to the amide ligand do not contribute significantly to the hole NTO. 205 −1.5 −1.6 −1.7 −1.8 −1.9 −2.0 −2.1 −2.2 1E7 1E8 1E9 1E10 Cu MAC BCz E +/ * = -2.28 Cu MAC PhCz E +/ * = -2.33 Au MAC BCz E +/ * = -2.23 k q (M -1 s -1 ) Quencher Reduction Potential (V vs. Fc +/0 ) Figure 4.21 Rehm-Weller analysis of the cMa complexes in THF. Several quenching studies were also performed in toluene, but values of kq for a given photosensitizer-quencher pair were consistently lower when compared to values found in THF (See Figure 4.31 and Table 4.9). This effect is likely due to the non-polar nature of toluene which increases the work term (Ws in Eq. 4.13) and consequently the energy barrier (∆Gr in Eq. 4.11) for charge separation. Non-polar media also inhibits physical separation of the charged species in the cage complex (PS + ,Q - ), which can allow the rate of kbet to become competitive with ket. This effect is expected to lower the overall efficiency of charge transfer, and thus kq. To confirm this hypothesis, I performed transient absorption measurements with Michael Kellogg in toluene. The transient absorption of Au BCz MAC with NMeP in toluene (Figure 4.22) shows notably different features than the spectra in THF (Figure 4.19). Instead of the 3 PS decaying and the PS(+) cation growing in, a single spectra is observed which appears to be a combination of the 3 PS and PS(+) which decays uniformly over hundreds of nanoseconds. This result supports that photoredox in toluene results in a (PS + ,Q - ) ion pair that does not separate, rather remains in close proximity where back 206 electron transfer to form ( 3 PS,Q) is possible. The equilibrium between (PS + ,Q - ) and ( 3 PS,Q) is quickly established which accounts for the uniformly decaying spectra that has signatures of the triplet and cation (Figure 4.22), and consequently lower kq values (Table 4.9). Figure 4.22 Transient absorption of Au BCz MAC with NMeP in toluene excited at 355nm. The absorption spectrum of the ground state complex is corresponds to the dashed blue line, and the navy blue dot-dashed line is the absorption of the cation determined by bulk electrolysis. 73 Table 4.9 Stern-Volmer results in toluene. Vred (V vs. Fc +/0 ) kq(Cu BCz MAC ) (M -1 s -1 10 9 ) kq(Cu PhCz MAC ) (M -1 s -1 10 9 ) Quin -2.18 0.04 0.08 NMeP -1.92 0.71 1.38 Alternatively, several quenching studies were performed in MeCN (Table 4.10). Quenching rates for a given photosensitizer-quencher pair are faster in highly polar MeCN than in THF since the work term becomes negligible in MeCN, which yields a larger driving force in G and a concomitant increase in kq. However, the benefit from the faster quenching rate is offset by the 207 hypsochromic shift for the ICT transition in MeCN, which decreases the amount light absorbed within the visible portion of the solar spectrum, and the significantly shorter excited state lifetime. Table 4.10 Stern-Volmer results in MeCN. Vred (V vs. Fc +/0 ) kq(Cu BCz MAC ) (M -1 s -1 10 9 ) Quin -2.18 1.4 NMeP -1.92 4.7 4.15 Generating Solar Fuels: Photoreduction of Water We sought to demonstrate the viability of cMa complexes as solar photosensitizers by carrying out the hydrogen evolution reaction (HER) using a well-established electrocatalyst for water reduction, Co(dmgH)2pyCl. We chose this compound because it uses an earth abundant metal, is readily synthesized and has been extensively studied for HER. 67, 74-77 Two reduction events at -1.0 V and -1.7 V vs. Fc +/0 in Co(dmgH)2pyCl have been ascribed to Co III/II and Co II/I couples which generate the active species for water reduction. 78 Therefore, our Rehm-Weller analysis predicts that all three cMa complexes can photoreduce Co(dmgH)2pyCl with a driving force ≥ 500 meV. The Au BCz MAC and Cu PhCz MAC complexes were chosen as they have near equal values for E* /- (Table 4.6), which makes for direct comparisons of gold to copper. A Stern-Volmer study using Au BCz MAC confirmed that the excited cMa complex was effectively quenched by Co(dmgH)2pyCl (kq = 2.5x10 9 M -1 s -1 , Figure 4.29). The measured kq is near the diffusion limit which is consistent with data in Figure 4.21. BIH was chosen as the sacrificial reductant as it has an oxidation potential (E +/0 = 0.33 V) sufficient to reduce Cu PhCz MAC and Au BCz MAC in the excited state (E* /- ~ -0.08 V and -0.09 V respectively). 79 For example, BIH was recently found to efficiently reduce photo-excited Au BCz MAC (kq = 2.5x10 8 M -1 s -1 ), 73 albeit at a rate ten-fold slower than 208 Co(dmgH)2pyCl. Thus, two pathways for photodriven HER are possible. The excited cMa* complex could first reduce Co(dmgH)2pyCl followed by subsequent reduction of cMa + by BIH (route I) or the excited cMa* could be reduced first by BIH and the newly formed cMa - subsequently reduces Co(dmgH)2pyCl (route II). Initial HER photolysis was performed using Au BCz MAC , Co(dmgH)2pyCl and BIH in THF/water solution (5% v/v). The concentrations of Co(dmgH)2pyCl (0.3 mM) and BIH (200 mM) used here are expected to quench Au BCz MAC * at rates of 7.5x10 4 s -1 and 5x10 7 s -1 , respectively. The measured lifetime of Au BCz MAC * in the THF/water solution is 100 ns (kr + knr = 1.0x10 7 s -1 ) so we expect HER to proceed exclusively by route II since quenching by the catalyst will be uncompetitive with unimolecular decay of the excited sensitizer. Slow formation of hydrogen was observed upon irradiating these solutions with 470 nm light for 90 minutes, giving only 1.5 turnovers based on the sensitizer, but hydrogen production increased significantly after the first two hours, giving 30 turnovers at seven hours (Figure 4.23). Figure 4.23 (a) TON versus irradiation time for Au BCz MAC (160 M), Co(dmgH)2pyCl (290 M), and BIH (200 mM) in 5 mL of the THF/water mixture (5% v/v). (b) Photo-HER in the presence of a mercury droplet for Au BCz MAC and Cu PhCz MAC . 209 Control experiments show that no hydrogen was generated if the photosensitizer was omitted (Figure 4.24). Further, BIH generates protons upon oxidation which could potentially contribute to HER. 80 However, irradiation of Au BCz MAC , Co(dmgH)2pyCl and BIH for one hour in anhydrous THF produced less hydrogen (by a factor of five) than in the water/THF mixture under identical conditions (Figure 4.25), thus confirming water as the primary source of H2. 210 0.0 0.5 1.0 1.5 2.0 2.5 3.0 200000 400000 600000 800000 1000000 Intensity Time 150 M Au MAC BCz 300 M Co(dmgH 2 )pyCl 0.2M BIH 5% H 2 O 95% THF (vol) 470nm LED, 1hr H 2 N 2 N 2 0.0 0.5 1.0 1.5 2.0 2.5 3.0 200000 400000 600000 800000 1000000 Intensity Time 300 M Co(dmgH 2 )pyCl 0.2M BIH 5% H 2 O 95% THF (vol) 470nm LED, 1hr N 2 Figure 4.24 Irradiation of the sample with (left) and without (right) the photosensitizer as a control experiment. This proves that the photosensitizer is necessary to generate H2. 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 0 100000 200000 300000 400000 500000 Intensity Time No H 2 O With H 2 O H 2 N 2 Figure 4.25 Control experiment of photo driven HER with and without water after one hour of irradiation in a 355nm photoreactor chamber. Interestingly, a control reaction without Co(dmgH) 2pyCl, i.e., Au BCz MAC and BIH in THF/water, showed behavior analogous that shown in Figure 4.23a for the system catalyzed with cobalt- glyoxime. Hydrogen was produced at roughly the same rate as seen in Figure 4.23a for the first two hours, followed by a faster rate, giving a TON of ~10 after 20 hours (Figure 4.26). A similar experiment irradiating Cu PhCz MAC in the presence of only BIH showed the same behavior (~10 TON of H2 after 20 hours, Figure 4.26). The increased rate of hydrogen production after an initial 211 induction period suggests that the excited sensitizer or the reduced sensitizer (formed by reduction of M RCz MAC * by BIH) decomposed to form a catalytically active material. 0 200 400 600 800 1000 1200 1400 0 2 4 6 8 10 12 TON (PS) Time (min) 0 200 400 600 800 1000 1200 1400 2 4 6 8 10 12 TON (PS) Time (min) Figure 4.26 Photocatalysis without Co(dmgH)2pyCl. Samples were prepared with 160 mM Au BCz MAC (left), and 63 mM Cu PhCz MAC (right) in wet THF (5% water by vol), 200mM BIH, and irradiated with a 470nm LED. To further investigate the photostability of two-coordinate cMa complexes, the absorption spectrum was monitored at different time intervals of exposure to the 470nm LED shown in Figure 4.8. These experiments show that the sensitizers slowly bleach during photolysis. For example, the intensity of the ICT absorption band of Au BCz MAC decreases 5% upon irradiation for 9 hours with 470 nm light in THF/water (Figure 4.27). Alternatively, decomposition of M BCz MAC after photoreduction by BIH could be the source of the electrocatalyst. 212 300 350 400 450 500 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Absorbance Wavelength (nm) 0 hr 1 hr 2 hr 3 hr 15 hr 24 hr (a) 0 5 10 15 20 25 66 68 70 72 74 [Ps] (M) Time (hrs) [PS] = -0.38t + 74.13 M R 2 = 0.999 (b) Figure 4.27 Photostability of Au BCz MAC in air free water/THF (5% v/v) irradiated with a 470nm LED at ~486 mW intensity. The absorption spectrum (a) and the concentration of Au BCz MAC calculated using Beer’s law (b). We suspect that a likely HER catalyst produced by decomposition of the sensitizer could be metal nanoparticles, which would be consistent with the observed incubation period. Nanoparticle metal complexes containing 13 Au atoms have been prepared upon reduction of Au Cl carbene complexes with NaBH4. 81 To test the hypothesis that metal nanoparticles are the catalyst generated by irradiation of the sensitizer and BIH, photolyses of M amide MAC and BIH in THF/water were performed with elemental mercury since the presence of mercury has been shown to suppress catalysis from metal nanoparticles. 9 Photolysis performed with the addition of a mercury drop led to a marked decrease in H2 production. Only 0.8 and 0.2 TONs, corresponding to >90% reduction in hydrogen production, was observed after ~23 hours of irradiation for Cu PhCz MAC and Au BCz MAC , respectively (Figure 4.28). 213 1.6 1.8 2.0 2.2 2.4 0.0 E06 0.5 E06 1.0 E06 1.5 E06 2.0 E06 2.5 E06 3.0 E06 3.5 E06 Intensity Time Cu no Hg Cu with Hg H 2 N 2 1.6 1.8 2.0 2.2 2.4 0.2 E06 0.4 E06 0.6 E06 0.8 E06 Intensity Time Au no Hg Au without Hg H 2 N 2 Figure 4.28 Mercury poisoning tests after ~23 hrs of irradiation without Co(dmgH)2pyCl for 𝐶𝑢 𝑃 ℎ𝐶𝑧 𝑀𝐴𝐶 (left), and 𝐴𝑢 𝐵𝐶𝑧 𝑀𝐴𝐶 (right). The concentrations of the photosensitizer these experiments were ~ 300 mM for the copper experiments, and ~150 mM for the gold experiments. In contrast, photolysis with added Co(dmgH)2pyCl in the presence of mercury led to a linear rise in H2 production from the start of the reaction (Figure 4.23b). Both cMa complexes drive catalytic HER under these conditions with similar performance after 22 hours of irradiation. However, Au BCz MAC exhibits higher stability compared to Cu PhCz MAC over several days of photocatalysis as the Au derivative can drive HER at a consistent rate over the 69 hours (~0.5 TON/hr). These results suggest the intriguing possibility that the cMa complexes can also serve as a robust single component for photocatalytic HER. 4.16 Conclusion Photophysical investigations into the excited state redox properties of two-coordinate coinage cMa complexes demonstrate that these compounds are viable as a new class of photosensitizers for production of solar fuels. Figures of merit were investigated in a variety of solvents for three two-coordinated Cu(I) and Au(I) cMa complexes. All reported complexes efficiently absorb visible light, are potent photoreducing agents and have long-lived excited states. Electrochemical studies show that substituents at the 3,6-position of carbazole stabilizes the 214 cationic species generated upon oxidation. The Cu(I) complexes have lower molar absorptivities and electrochemical stability than the Au(I) analog, but both have comparable excited state reducing potential ( -2.3 V vs. Fc +/0 ), which are higher than values found in commonly used photosensitizers such as Ru(bpy)3 2+ (E +/ * = -1.12 V vs. Fc +/0 ) and Ir(ppy)3 (E +/* = - 2.14 V vs. Fc +/0 ). Two-coordinate cMa complexes also have excited state reduction potentials greater than Cu(I) bis phenanthroline systems, (E +/* = -1.8 V vs. Fc +/0 ). 25 We find that the choice of carbene largely determines the value for E + /* in cMa complexes, while the amide determines the value for E* /- which is a result of the ICT nature of the excited state. The facile ability to tailor absorption and redox properties of the ICT state in two-coordinate complexes offers enhanced tunability compared to Cu(I) complexes relying on an MLCT excited state. Two-coordinate cMa complexes can be modified to make even more potent photo-reducing agents using carbenes with more negative reduction potentials than the carbene used here; however, such a change will shift the lowest energy absorption bands into the near-UV spectrum. The ICT transitions can be shifted back into the visible spectrum by pairing these carbenes with amides that have lower oxidation potentials, thus lowering the energy of the ICT state. Alternatively, amides such as cyano-substituted carbazole can be used to increase the E* /- for photo-oxidative applications. Therefore, these cMa complexes can be designed to activate any given electrocatalysts with significant control over the energy of the visible absorption transition, which makes them a promising new class of photosensitizers to produce solar fuels. This work is the first demonstration of photocatalytic HER driven by copper and gold cMa complexes, using a well-studied cobalt-glyoxime HER electrocatalyst. Future studies are anticipated to reach much higher turnover numbers as the conditions are optimized. The TONs can likely be significantly increased by optimizing the choice of electrocatalyst, light source, pH, and 215 sacrificial reductant. 5, 82, 83 We find that Au BCz MAC is more stable over prolonged irradiation compared to Cu PhCz MAC , which mirrors the added electrochemical stability of gold cMa complexes compared to copper analogues. It is important to note that the photocatalytic HER described above was carried out in the presence of elemental mercury, to eliminate the involvement of metal particles in the electrocatalysis. When the mercury and cobalt-glyoxime catalyst are left out of the system, both Au BCz MAC and Cu PhCz MAC produce hydrogen at a level of ca. 20 turnovers after 20 hours. An induction period is followed by a steep increase in HER activity. The fact that added mercury suppresses this catalysis is strong evidence that the electrocatalysis is due to gold and copper nanoparticles, respectively. Based on the work of Crudden, et al., 81 we suspect that the metal particles are carbene capped metal particles, formed in the presence of the sacrificial reductant. The work reported here involves only a single carbene ligand. We are in the process of exploring this “self-catalyzed” cMa based HER with other carbenes to explore the activity and stability of the metal particle electrocatalysts with other carbene capping agents. 216 4.17 Stern-Volmer Quenching Data 0.0 0.2 0.4 0.6 0.8 1.0 1.00 1.02 1.04 1.06 1.08 1.10 1.12  0 / [Q] (mM) Au MAC BCz Co(dmgH) 2 (py)Cl    =  + 9x10 9 M -1 s -1  0 [Q] Figure 4.29 Stern-Volmer study of Au BCz MAC quenched by Co(dmgH)2pyCl in THF. The kq is near diffusion limited as predicted by the Rehm-Weller analysis. 217 𝐀𝐮 𝐁𝐂𝐳 𝐌𝐀𝐂 : 0.0 0.5 1.0 1.5 2.0 2.5 1 2 3 4 5 6 7    [Q] (mM) Au MAC BCz PN    =  + 9x 9 M -1 s -1  0 [Q] 0 1 2 3 4 5 0 2 4 6 8 10 12  0 / [Q] (mM) Au MAC BCz    =  + 9x10 9 M − s -1  0 [Q] NapA -2 0 2 4 6 8 10 12 14 0 5 10 15 20 25 30  0 / [Q] (mM)    =  + 8x10 9 M -1 s -1  0 [Q] Au MAC BCz NB 0 1 2 3 4 5 1.0 1.5 2.0 2.5 3.0 3.5  0 / [Q] (mM) Au MAC BCz DMNA    =  + x10 9 M -1 s -1  0 [Q] −1 0 1 2 3 4 5 6 7 8 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4   [Q] (mM)    =  + x 8 M -1 s -1  0 [Q] Au MAC BCz NMeP 𝑪𝒖 𝑩𝑪𝒛 𝑴𝑨𝑪 : 218 0 1 2 3 4 5 0 5 10 15 20    Q [M] Cu MAC BCz PN    =  + x10 10 M -1 s -1  0 [Q] -2 0 2 4 6 8 10 12 14 0 10 20 30 40 50    [Q] (M) Cu MAC BCz NB    =  + 99x10 9 M -1 s -1  0 [Q] 0 1 2 3 4 5 1 2 3 4 5 6    [Q] (M) Cu MAC BCz DMNA    =  + 6x10 9 M -1 s -1  0 [Q] 0 1 2 3 4 5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0    [Q] (mM) Cu MAC BCz NMeP    =  + 9x10 7 M -1 s -1  0 [Q] 0 1 2 3 4 1.00 1.02 1.04 1.06 1.08 1.10    [Q] (mM) Cu MAC BCz QI    =  + 68x10 7 M -1 s -1  0 [Q] 𝑪𝒖 𝑷𝒉𝑪𝒛 𝑴𝑨𝑪 : 219 0 1 2 3 4 5 0 5 10 15 20 25 30  0 / [Q] (mM) Cu MAC PhCz PN    =  + x10 10 M -1 s -1  0 [Q] 0 4 8 12 0 10 20 30 40 50 60 70  0 / [Q] (mM) Cu MAC PhCz NB    =  + x10 10 M -1 s -1  0 [Q] 0 1 2 3 4 5 0 1 2 3 4 5 6 7 8 9  0 / [Q] (mM) Cu MAC PhCz DMNA    =  + x10 9 M -1 s -1  0 [Q] 0 1 2 3 4 5 0 1 2 3 4 5 6 7 8    [Q] (mM) Cu MAC PhCz NMeP    =  + x10 9 M -1 s -1  0 [Q] 0 1 2 3 4 1.00 1.05 1.10 1.15 1.20 1.25 1.30    [Q] (mM) Cu MAC PhCz QI    =  + x10 8 M -1 s -1  0 [Q] Figure 4.30 Stern-Volmer plots measured in THF. The corresponding photosensitizer and electron accepting quencher are displayed on the plots. 220 𝑪𝒖 𝑩𝑪𝒛 𝑴𝑨𝑪 : 0 1 2 3 4 1.0 1.5 2.0 2.5 3.0  0 / [Q] (mM) Cu MAC BCz NMeP    =  + x10 8 M -1 s -1  0 [Q] 0 1 2 3 4 0.98 1.00 1.02 1.04 1.06 1.08 1.10 1.12 1.14   / [Q] (mM) Cu MAC BCz QI    =  + x10 7 M -1 s -1  0 [Q] 𝑪𝒖 𝑷𝒉𝑪𝒛 𝑴𝑨𝑪 : 0 1 2 3 4 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0   / [Q] (mM) Cu MAC PhCz NMeP    =  + 8x10 9 M -1 s -1  0 [Q] 0 1 2 3 4 1.00 1.05 1.10 1.15 1.20 1.25    [Q] (mM) Cu MAC PhCz QI    =  + 89x10 7 M -1 s -1  0 [Q] Figure 4.31 Stern-Volmer plots measured in Toluene. The corresponding photosensitizer and electron accepting quencher are displayed on the plots 221 0 1 2 3 4 5 6 7 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5  0 / [Q] (mM) Cu MAC BCz NMeP    =  + x10 9 M -1 s -1  0 [Q] 0 1 2 3 4 5 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8    [Q] (mM) Cu MAC BCz Quin    =  + x10 9 M -1 s -1  0 [Q] Figure 4.32 Stern-Volmer plots measured in MeCN. The corresponding photosensitizer and electron accepting quencher are displayed on the plots. 4.18 Synthesis & Characterization The cMa compounds were synthesized using methods that follow literature precedent. 43 The copper complexes were isolated with >63% yield, and Au BCz MAC was isolated with 31% yield. The final products are all indefinitely air stable as solids. The composition was characterized by 1 H and 13 C NMR, CHNS elemental analysis Figure 4.33 Synthetic scheme of the new cMa complexes Cu BCz MAC , Au BCz MAC and Cu PhCz MAC . 222 Synthesis of 1,3-bis(2,6-diisopropylphenyl)-5,5-dimethyl-4-oxo-3,4,5,6- tetrahydropyrimidin-1-ium-2-ide-Cu(I)-3,6-di-tert-butylcarbazol-9-ide (𝐂𝐮 𝐁𝐂𝐳 𝐌𝐀𝐂 ): Following the general procedure, a 150 mL Schlenkflask was charged with 512 mg 3,6-di-tert-butyl-9-H- carbazole (1.83 mols) and 185 mg sodium tert-butoxide (1.92 mols). The solids were dissolved in 80 mL THF, and 1.0 g (1.83 mols) of MAC CuCl was added. The workup yielded a yellow powder. The powder was washed with methanol which dissolved some product but removed all remaining impurities, yielding 890 mg (64% yield) of Cu BCz MAC as a yellow powder which is bright yellow emissive under 365 nm UV radiation. 1 H NMR (400 MHz, acetone-d6) δ: 7.81-7.72 (m,4H), 7.58 (d, J = 7.6 Hz, 2H), 7.52 (d, J = 7.9 Hz, 2H), 6.90 (dd, J = 8.6 Hz, 2.1 Hz, 2H), 5.52 (d, J = 8.8 Hz, 2H), 4.25 (s, 2H), 3.55 (hept, J = 7.0 Hz, 2H), 3.29 (hept, J = 7.2 Hz, 2H), 1.66 (s, 6H), 1.42 (d, J = 6.8 Hz, 6H), 1.32 (s, 18H), 1.28-1.22 (m, 18H). 13 C NMR (100 MHz, acetone-d6) δ: 211.0, 206.3, 206.3, 172.3, 149.5, 147.4, 146.4, 141.5, 138.0, 137.5, 131.3, 130.9, 126.7, 125.8, 124.9, 121.2, 115.3, 115.0, 62.2, 39.0, 35.0, 32.8, 32.5, 30.6, 30.4, 30.2, 30.0, 29.8, 29.8, 29.6, 29.5, 29.4, 25.5, 24.9, 24.7, 24.6, 24.5. CHN (C: 76.61%; H: 8.40%; N: 5.47%; calculated: C: 76.15%; H: 8.44%; N: 5.33%). Synthesis of 1,3-bis(2,6-diisopropylphenyl)-5,5-dimethyl-4-oxo-3,4,5,6- tetrahydropyrimidin-1-ium-2-ide-Cu(I)-3,6-di-phenylcarbazol-9-ide (𝐂𝐮 𝐏𝐡𝐂𝐳 𝐌𝐀𝐂 ): Following the general procedure, a 250 mL Schlenkflask was charged with 585 mg 3,6-diphenyl-9-H-carbazole (1.83 mols) and 185 mg (1.92 mols) NaO t Bu. The solids were dissolved in 100 mL THF, and 1.0 g (1.83 mols) of MAC CuCl was added to yield a light yellow powder after the general workup. The powder was washed copiously with with diethyl ether to afford 1020 mg (67% yield) of Cu PhCz MAC as a shiny yellow powder which emits intense turquoise under 365 nm UV radiation. 1 H NMR (400 MHz, acetone-d6) δ: 8.17 (dd, J = 0.7Hz, 2.0 Hz, 2H), 7.87-7.80 (m, 2H), 7.67-7.62 (m, 6H), 7.58 (d, J = 7.8 Hz, 2H), 7.41-7.36 (m, 4H), 7.23-7.18 (m, 4H), 5.66 (dd, J = 0.6 Hz, J = 8.3Hz), 4.30 (s, 2H), 3.59 (hept, J = 6.9Hz, 2H), 3.33 (hept, J = 7.0 Hz, 2H), 1.68 (s, 6H), 1.4 (d, J=6.6 Hz, 6H), 1.33-1.24 (m, 18H). 13 C NMR (100 MHz, acetone-d6) δ: 210.7, 206.3, 206.3, 206.3, 172.3, 151.2, 147.5, 146.4, 144.2, 141.5, 137.5, 131.5, 131.1, 129.8, 129.5, 129.4, 127.9, 127.5, 127.3, 126.8, 126.2, 126.1, 125.9, 125.9, 123.5, 119.7, 118.2, 116.1, 62.1, 39.0, 32.4, 30.5, 30.4, 30.3, 30.2, 30.0, 29.8, 29.6, 29.5, 29.4, 29.3, 25.5, 24.9, 24.7, 24.5. CHN (C: 77.70%; H: 6.94%; N: 5.14%; calculated: C: 78.27%; H: 7.06%; N: 5.07%) 223 Synthesis of 1,3-bis(2,6-diisopropylphenyl)-5,5-dimethyl-4-oxo- 3,4,5,6-tetrahydropyrimidin-1-ium-2-ide-Au(I)-3,6-di-tert-butylcarbazol-9-ide (𝐀𝐮 𝐁𝐂𝐳 𝐌𝐀𝐂 ): Following the general procedure, a 250 mL Schlenkflask was charged with 411 mg (1.47 mols) 3,6-di-tertbutyl-9-H-carbazole and 142 mg NaO t Bu (1.62 mols). The solids were dissolved in 120 mL THF and 1.0g (1.47 mols) MAC AuCl was added. The general workup afforded 425 mg (31% yield) of Au BCz MAC as a mustard yellow powder that emits bright yellow under 365 nm radiation. 1 H NMR (400 MHz, acetone-d6) δ: 7.83 (d, J = 2 Hz, 2H), 7.78-7.69 (m, 2H), 7.56 (d, J = 7.7 Hz, 2H), 7.50 (d, J = 7.8 Hz, 2H), 7.0 (dd, J = 1.7 Hz, J = 9.0 Hz, 2H), 6.00 (d, J = 8.6 Hz, 2H), 4.25 (s, 2H), 3.50 (hept, J = 6.8 Hz, 2H), 3.25 (hept, J = 6.8 Hz, 2H), 1.68 (s, 6H), 1.41 (d, J = 6.7 Hz, 6H), 1.37 (d, J = 6.8 Hz, 6H), 1.34 (d, J = 7.4 Hz, 6H), 1.33 (s, 18H), 1.22 (d, J = 6.8 Hz, 6H). 13 C NMR (100 MHz, acetone-d6) δ 206.3, 206.3, 206.2, 205.7, 173.0, 149.1, 147.4, 146.2, 141.7, 138.8, 137.7, 131.1, 130.8, 126.4, 125.4, 124.7, 121.5, 115.5, 114.4, 62.1, 39.1, 35.0, 35.0, 32.8, 32.6, 30.6, 30.6, 30.4, 30.4, 30.2, 30.0, 29.8, 29.8, 29.6, 29.6, 29.4, 25.0, 24.7, 24.7, 24.6, 24.3, 14.8. CHN (C: 64.59%; H: 7.10%; N: 4.63%; calculated: C: 65.13%; H: 7.22%; N: 4.56%). Crystal Structures. Molecular structures of Cu BCz MAC , Au BCz MAC and Cu PhCz MAC were solved by Jonas Schaab using single crystal X-ray diffraction. Structural drawings of the three complexes are given in Figure 4.34. The Au BCz MAC complex crystallized with a disorder of the C=O and the two methyl groups whereas Cu BCz MAC and Cu PhCz MAC do not show any disorder in the MAC ligands in their structures. Values for the bond lengths and angles were similar to parameters found in related cMa derivatives. 39-43 The three cMa complexes have linear coordination geometries as the C-M-N bond angles are all near ~180°. The carbene and carbazole ligands have near coplanar orientations, with the torsion angles between that vary between 6.2° to 19.6° (Table 4.11). It has been shown that coplanar ligand orientation is important for achieving appreciable oscillator strength of the ICT transition. 11, 44, 84 Derivatives of cMa complexes with ligands having large torsion angles display 224 ICT bands with notably low oscillator strengths because the  systems of the ligands have poor spatial overlap. 11, 85, 86 The full crystallographic data set is available in the manuscript. 68 Figure 4.34 Single crystal structures of all new cMa complexes. Table 4.11 Crystallographic Parameters for Cu BCz MAC , Au BCz MAC and Cu PhCz MAC compound C-M (A) M-N (A) C-M-N (˚) Torsion (˚) NC-M-NC C-N-C(O) (˚) C-N-C(H2) (˚) Disorder ratio Cu BCz MAC 1.875(2) 1.849(2) 178.71 19.62 126.41 124.03 -/- Au BCz MAC 1.977(4) 1.997(3) 176.23 14.66 124.76 124.03 60/40 (CH3) 50/50 (C=O) Cu PhCz MAC 1.882(2) 1.854(2) 175.11 2.22 6.52 125.89 123.46 -/- Accession Codes. CCDC 2238026 2237840, and 2237833 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033. 225 1 H and 13 C NMR Spectra Figure 4.35 1 H and 13 C NMR of 𝐶𝑢 𝐵𝐶𝑧 𝑀𝐴𝐶 in acetone-d6. The peaks at 2.05ppm and 2.8ppm in the 1 H NMR correspond to acetone and water respectively. 226 Figure 4.36 1 H and 13 C NMR of 𝐶𝑢 𝑃 ℎ𝐶𝑧 𝑀𝐴𝐶 in acetone-d6. The peaks at 5.2ppm, 2.8ppm, and 2.05ppm in the 1 H NMR correspond to DCM, water, and acetone-d6 respectively. 227 Figure 4.37 1 H and 13 C NMR of Au BCz MAC in acetone-d6. The peaks at 2.05ppm and 2.8ppm in the 1 H NMR correspond to acetone and water respectively. 228 4.19 Future Work for Solar Fuels with cMa Complexes The ability of Au BCz MAC and Cu BCz MAC to photocatalytically convert water to hydrogen without an added electrocatalyst was an unexpected and exciting result. We showed that Au BCz MAC is not photostable under the 500 mW exposure of the 470 nm LED in a water/THF (5% v/v) solution that was sparged with nitrogen (Figure 4.27). We assigned the decomposition product to metal nanoparticles based on the significant suppression of H2 in a solution with 0.3M BIH and a drop of mercury. However, we did not do any real characterization of the photodecomposition product(s) of Au BCz MAC . The assumption that decomposition leads to formation of metallic nanoparticles was suggested because mercury efficiently amalgamates with other metals. However, there is no conclusive evidence that small metal-rich organometallic clusters would not also be absorbed into elemental mercury. Whatever the decomposition product may be, it displays significant catalytic activity towards water reduction. For example, HER is estimated to activate at ~3-5 hours of irradiation for Au BCz MAC without an added electrocatalyst. At 5 hours, the complex has decomposed by ~3% (using the equation in Figure 4.27b). This corresponds to the loss of 0.95 nmol of material which can generate up to 36 mols of H2 after 3 days of irradiation. Assuming that the decomposition product is stoichiometric with the Au BCz MAC complex (i.e. only one metal is present in the decomposition product), this corresponds to TON of ~3,800 over 3 days of radiation with a TOF ≥ 52 equivalents of H2 per hour. These numbers are already exciting, but if the decomposition product is revealed to be nanoclusters, then the TON must be multiplied by the number of metal atoms in the cluster. If we are forming nanoclusters such as the ones seen in Crudden’s lab (12 metal atoms), then the TON would be 45,600 after 3 days. 81 If the catalytically active product is larger copper nanoparticles (10 to ≥ 100,000), then the TON would be in the (456,000 to ≥ 4,560,000,000) range. Thus, it is worth performing further characterization and 229 isolation of the decomposition product, and testing the HER capabilities in a more controlled manner. A primary advantage of Au cMa complexes over Cu analogues as photosensitizers is the added electrocatalytic and photocatalytic stability. We suspect that the M + Cz - bond is labile upon ground state oxidation and photoexcitation. A potential solution to this issue is to covalently link the carbene and amide ligands to form a macrocyclic complex. This will prevent dissociation of the metal-carbazolide bond. One approach that I started with Jack Applebaum was to perform an ester-linkage as shown below. Figure 4.38 Two macrocycle targets for cMa complexes with enhanced stability. M Cz BZI }2 (left) and M BCz BZI }1 (right). The structures of these complexes were largely designed around the availability of 1,8 and 2,7 – dibromocarbazole precursors. One parameter to consider is the length of the ester chain which is governed by “n” in Figure 4.38. Geometry optimizations were performed in QChem using the B3LYP method and LACVP basis/ECP with varying ester chain lengths. Letting n = 1 for Cu BCz BZI }1 resulted in significant bending of the C:-M-N angle (~30°). However, letting n = 2 resulted in a coplanar ligand system and the S1 and T1 states were ICT in nature. Extending the ester chain 230 length to n = 3 resulted in twisting of the carbene ligand with respect to the carbazole ligand, which negatively impacts the bond strength and oscillator strength for ICT transitions. Finally, letting n = 4 resulted in complete dissociation of the carbene-metal bond. Interestingly, the gold analogue (Cu BCz BZI }1) observed coplanarity for n = 4, which is a direct result of the increased ionic radius between Cu(I) and Au(I). The structures for M Cz BZI }2 were also evaluated computationally to predict the length for the ester chain. Cu Cz BZI }2 showed coplanar ligands for n = 2, and twisted ligands for n = 3. The geometry optimization of Au Cz BZI }2 yielded coplanar ligands for n = 3. The retrosynthesis of M BCz BZI }1 starts from functionalizing BZI by adding a protected alkyl bromo ester to a solution of deprotonated BZI (Figure 4.39). The second ester is added to BZI by a simple SN2 reaction between the alkyl bromide group and the imine nitrogen of BZI. The carbene is then metalated by mixing with Ag2O; a standard route of metalating a carbene HCl. The silver can then be replaced by Cu(I) or Au(I) by transmetallation. 87 The carbene-M-halide for M = Cu(I) or Au(I) is expected to be stable to base, thus a simple base deprotection with K2CO3 can be used in step 4 to achieve the metalated carbene appended with alcohol substituents. On the donor side, 1,8-dibromo-3,6-di-tert-butyl-9H-carbazole is readily available from chemical vendors. This can be converted to the dicyano analogue (i) by refluxing in NMP with excess CuCN. 88 The cyano substituents can be converted to carboxylic acid groups by refluxing the dicyanocarbazole in aqueous solution with excess NaOH. 89 Finally, this could be converted to a di-acid chloride by refluxing the carboxylic acid in DCM and thionyl chloride. 231 Figure 4.39 Retrosynthesis of M BCz BZI }1. An identical approach can be used to make M Cz BZI }2. Di-terbutyl substituents in the 3,6 positions would be ideal for this complex, but the carbazole precursor would have to be made instead of purchased. 232 Figure 4.40 Retrosynthesis of M Cz BZI }2. These complexes would likely display enhanced photo- and electrochemical stability. Another added benefit of covalently the linkers would be an increased oscillator strength of the ICT transition since the covalent linkage between the carbene and amide  systems would prevent the ligand  systems from twisting out of a mutual plane. The increase in oscillator strength would lead to greater extinction coefficient which would make the complexes more efficient at absorbing photons: a desirable material property for solar fuels. Another future direction is to try make the solar fuels chemistry more green by doing the process in aqueous solution instead of THF spiked with water (5% v/v). An interesting surfactant entitled “Savie” has been discovered by Joseph Kincaid in the Lipshutz group at UCSB that has 233 an organic tail that is decorated with ether and amide groups. 90 The organic groups that make up the non-polar component of Savie allow for small organic reaction capsules upon formation of micelles. Kincaid et al. showed that an impressive variety of organic reactions could be performed in aqueous solution with Savie micelles. It would be interesting to incorporate this technology into the photo-reduction of water using cMa complexes (Figure 4.41) Figure 4.41Schematic of photocatalysis in aqueous solution with cMa complexes hosted in Savie micelles. cMa* represents the excited cMa complex, ECat represents an electrocatalyst molecule, and BIH is the sacrificial reductant. The first experiment would be to probe the solubility of cMa complexes in an aqueous solution of Savie micelles. Since cMa complexes are highly soluble in most organic solvents, they are expected to be soluble in Savie micelles. The next important experiment is to measure the absorption and emission spectrum, as well as the excited state lifetime, and quantum yield. The absorption of cMa complexes is blue shifted in polar media (Figure 4.13-Figure 4.15), thus it would be interesting to probe the ICT absorption band in the aqueous micelle. The excited state lifetime 234 of cMa complexes has also been shown to drop as much as 90% between polar and non-polar solvent environments (Table 4.4). It is imperative that the lifetime remain longer than 10ns at minimum, to accomplish photoredox. The kinetics of photoredox may be obscured from a traditional models because the concentration of the photosenstizer, electrocatalyst, and sacrificial reductant are now with respect to the micelle, rather than the volume of the reaction flask. If the excited state lifetimes are long enough to do photoredox, then an electrocatalyst and sacrificial reductant could be co-solublized in the micelle for photo-HER experiments. Successful results would demonstrate solar fuels generation in aqueous media, which would improve the sustainability and reduce the hazards associated with this technology. 4.20 Chapter 4 References (1) Albo, J.; García, G. Enhanced visible-light photoreduction of CO2 to methanol over Mo2C/TiO2 surfaces in an optofluidic microreactor. Reaction Chemistry & Engineering 2021, 6 (2), 304-312, 10.1039/D0RE00376J. DOI: 10.1039/D0RE00376J. (2) Ma, W.; Xie, S.; Zhang, X.-G.; Sun, F.; Kang, J.; Jiang, Z.; Zhang, Q.; Wu, D.-Y.; Wang, Y. Promoting electrocatalytic CO2 reduction to formate via sulfur-boosting water activation on indium surfaces. Nature Communications 2019, 10 (1), 892. DOI: 10.1038/s41467-019-08805-x. (3) Nguyen-Phan, T.-D.; Hu, L.; Howard, B. H.; Xu, W.; Stavitski, E.; Leshchev, D.; Rothenberger, A.; Neyerlin, K. C.; Kauffman, D. R. High current density electroreduction of CO2 into formate with tin oxide nanospheres. Scientific Reports 2022, 12 (1), 8420. DOI: 10.1038/s41598-022-11890-6. (4) Pan, H.; Barile, C. J. Electrochemical CO2 reduction to methane with remarkably high Faradaic efficiency in the presence of a proton permeable membrane. Energy & Environmental Science 2020, 13 (10), 3567-3578, 10.1039/D0EE02189J. DOI: 10.1039/D0EE02189J. (5) Xie, L.; Tian, J.; Ouyang, Y.; Guo, X.; Zhang, W.; Apfel, U.-P.; Zhang, W.; Cao, R. Water- Soluble Polymers with Appending Porphyrins as Bioinspired Catalysts for the Hydrogen Evolution Reaction. Angewandte Chemie International Edition 2020, 59 (37), 15844-15848, https://doi.org/10.1002/anie.202003836. DOI: https://doi.org/10.1002/anie.202003836 (acccessed 2023/03/02). (6) Hockin, B. M.; Li, C.; Robertson, N.; Zysman-Colman, E. Photoredox catalysts based on earth- abundant metal complexes. Catalysis Science & Technology 2019, 9 (4), 889-915, 10.1039/C8CY02336K. DOI: 10.1039/C8CY02336K. 235 (7) Thuillier, G.; Hersé, M.; Labs, D.; Foujols, T.; Peetermans, W.; Gillotay, D.; Simon, P. C.; Mandel, H. The Solar Spectral Irradiance from 200 to 2400 nm as Measured by the SOLSPEC Spectrometer from the Atlas and Eureca Missions. Solar Physics 2003, 214 (1), 1-22. DOI: 10.1023/A:1024048429145. (8) Chandrasekaran, K.; Natarajan, P. Photosensitization of superoxo-bridged dicobalt(III) cations by tris-(2,2′-bipyridyl)ruthenium(II) in the excited state. Journal of the Chemical Society, Chemical Communications 1977, (21), 774b-775, 10.1039/C3977000774B. DOI: 10.1039/C3977000774B. (9) Wang, X.; Goeb, S.; Ji, Z.; Pogulaichenko, N. A.; Castellano, F. N. Homogeneous photocatalytic hydrogen production using π-conjugated platinum (II) arylacetylide sensitizers. Inorganic chemistry 2011, 50 (3), 705-707. (10) To, W.-P.; Tong, G. S.-M.; Lu, W.; Ma, C.; Liu, J.; Chow, A. L.-F.; Che, C.-M. Luminescent Organogold(III) Complexes with Long-Lived Triplet Excited States for Light-Induced Oxidative C H Bond Functionalization and Hydrogen Production. Angewandte Chemie International Edition 2012, 51 (11), 2654-2657, https://doi.org/10.1002/anie.201108080. DOI: https://doi.org/10.1002/anie.201108080 (acccessed 2023/04/01). (11) Yang, J.-G.; Song, X.-F.; Cheng, G.; Wu, S.; Feng, X.; Cui, G.; To, W.-P.; Chang, X.; Chen, Y.; Che, C.-M.; et al. Conformational Engineering of Two-Coordinate Gold(I) Complexes: Regulation of Excited-State Dynamics for Efficient Delayed Fluorescence. ACS Applied Materials & Interfaces 2022, 14 (11), 13539-13549. DOI: 10.1021/acsami.2c01776. (12) Yuan, Y.-J.; Yu, Z.-T.; Cai, J.-G.; Zheng, C.; Huang, W.; Zou, Z.-G. Water Reduction Systems Associated with Homoleptic Cyclometalated Iridium Complexes of Various 2- Phenylpyridines. ChemSusChem 2013, 6 (8), 1357-1365, https://doi.org/10.1002/cssc.201300451. DOI: https://doi.org/10.1002/cssc.201300451 (acccessed 2023/04/01). (13) Goldsmith, J. I.; Hudson, W. R.; Lowry, M. S.; Anderson, T. H.; Bernhard, S. Discovery and High-Throughput Screening of Heteroleptic Iridium Complexes for Photoinduced Hydrogen Production. Journal of the American Chemical Society 2005, 127 (20), 7502-7510. DOI: 10.1021/ja0427101. (14) Hogan, C.; Harris, A.; Bond, A.; Sly, J.; Crossley, M. Electrochemical studies of porphyrin- appended dendrimers. Physical chemistry chemical physics : PCCP 2006, 8, 2058-2065. DOI: 10.1039/b516281e. (15) Kalyanasundaram, K.; Grätzel, M. Light Induced Redox Reactions of Water Soluble Porphyrins, sensitization of hydrogen generation from water by zincporphyrin derivatives. Helvetica Chimica Acta 1980, 63 (2), 478-485, https://doi.org/10.1002/hlca.19800630219. DOI: https://doi.org/10.1002/hlca.19800630219 (acccessed 2023/04/01). (16) McCusker, C. E.; Castellano, F. N. Design of a Long-Lifetime, Earth-Abundant, Aqueous Compatible Cu(I) Photosensitizer Using Cooperative Steric Effects. Inorganic Chemistry 2013, 52 (14), 8114-8120. DOI: 10.1021/ic401213p. 236 (17) Harriman, A.; Porter, G.; Richoux, M.-C. Colloidal platinum catalysts for the reduction of water to hydrogen, photosensitised by reductive quenching of zinc porphyrins. Journal of the Chemical Society, Faraday Transactions 2: Molecular and Chemical Physics 1981, 77 (10), 1939- 1948, 10.1039/F29817701939. DOI: 10.1039/F29817701939. (18) Rosko, M. C.; Wells, K. A.; Hauke, C. E.; Castellano, F. N. Next Generation Cuprous Phenanthroline MLCT Photosensitizer Featuring Cyclohexyl Substituents. Inorganic Chemistry 2021, 60 (12), 8394-8403. DOI: 10.1021/acs.inorgchem.1c01242. (19) McMillin, D. R.; Buckner, M. T.; Ahn, B. T. A light-induced redox reaction of bis(2,9- dimethyl-1,10-phenanthroline)copper(I). Inorganic Chemistry 1977, 16 (4), 943-945. DOI: 10.1021/ic50170a046. (20) Cuttell, D. G.; Kuang, S.-M.; Fanwick, P. E.; McMillin, D. R.; Walton, R. A. Simple Cu(I) Complexes with Unprecedented Excited-State Lifetimes. Journal of the American Chemical Society 2002, 124 (1), 6-7. DOI: 10.1021/ja012247h. (21) Eggleston, M. K.; McMillin, D. R.; Koenig, K. S.; Pallenberg, A. J. Steric Effects in the Ground and Excited States of Cu(NN)2+ Systems. Inorganic Chemistry 1997, 36 (2), 172-176. DOI: 10.1021/ic960698a. (22) Kärkäs, M. D.; Verho, O.; Johnston, E. V.; Åkermark, B. Artificial Photosynthesis: Molecular Systems for Catalytic Water Oxidation. Chem. Rev. 2014, 114 (24), 11863-12001. DOI: 10.1021/cr400572f. (23) Cheung, K. P. S.; Sarkar, S.; Gevorgyan, V. Visible Light-Induced Transition Metal Catalysis. Chem. Rev. 2022, 122 (2), 1543-1625. DOI: 10.1021/acs.chemrev.1c00403. (24) Bevernaegie, R.; Wehlin, S. A. M.; Piechota, E. J.; Abraham, M.; Philouze, C.; Meyer, G. J.; Elias, B.; Troian-Gautier, L. Improved Visible Light Absorption of Potent Iridium(III) Photo- oxidants for Excited-State Electron Transfer Chemistry. Journal of the American Chemical Society 2020, 142 (6), 2732-2737. DOI: 10.1021/jacs.9b12108. (25) Prier, C. K.; Rankic, D. A.; MacMillan, D. W. C. Visible Light Photoredox Catalysis with Transition Metal Complexes: Applications in Organic Synthesis. Chem. Rev. 2013, 113 (7), 5322- 5363. DOI: 10.1021/cr300503r. (26) Larsen, C. B.; Wenger, O. S. Photoredox Catalysis with Metal Complexes Made from Earth- Abundant Elements. Chemistry – A European Journal 2018, 24 (9), 2039-2058. DOI: 10.1002/chem.201703602 (acccessed 2023/01/08). (27) Chan, A. Y.; Perry, I. B.; Bissonnette, N. B.; Buksh, B. F.; Edwards, G. A.; Frye, L. I.; Garry, O. L.; Lavagnino, M. N.; Li, B. X.; Liang, Y.; et al. Metallaphotoredox: The Merger of Photoredox and Transition Metal Catalysis. Chem. Rev. 2022, 122 (2), 1485-1542. DOI: 10.1021/acs.chemrev.1c00383. 237 (28) Hossain, A.; Bhattacharyya, A.; Reiser, O. Copper's rapid ascent in visible-light photoredox catalysis. Science 2019, 364 (6439). DOI: 10.1126/science.aav9713 From NLM PubMed-not- MEDLINE. (29) Lazorski, M. S.; Castellano, F. N. Advances in the light conversion properties of Cu(I)-based photosensitizers. Polyhedron 2014, 82, 57-70. DOI: 10.1016/j.poly.2014.04.060. (30) Zhang, Y.; Schulz, M.; Wächtler, M.; Karnahl, M.; Dietzek, B. Heteroleptic diimine– diphosphine Cu(I) complexes as an alternative towards noble-metal based photosensitizers: Design strategies, photophysical properties and perspective applications. Coord. Chem. Rev. 2018, 356, 127-146. DOI: 10.1016/j.ccr.2017.10.016. (31) Alkan-Zambada, M.; Hu, X. Cu Photoredox Catalysts Supported by a 4,6-Disubstituted 2,2′- Bipyridine Ligand: Application in Chlorotrifluoromethylation of Alkenes. Organometallics 2018, 37 (21), 3928-3935. DOI: 10.1021/acs.organomet.8b00585. (32) Yu, Z.-J.; Chen, H.; Lennox, A. J. J.; Yan, L.-J.; Liu, X.-F.; Xu, D.-D.; Chen, F.; Xu, L.-X.; Li, Y.; Wu, Q.-A.; et al. Heteroleptic copper(I) photosensitizers with carbazole-substituted phenanthroline ligands: Synthesis, photophysical properties and application to photocatalytic H2 generation. Dyes and Pigments 2019, 162, 771-775. DOI: 10.1016/j.dyepig.2018.10.067. (33) Hernandez-Perez, A. C.; Collins, S. K. Heteroleptic Cu-Based Sensitizers in Photoredox Catalysis. Acc. Chem. Res. 2016, 49 (8), 1557-1565. DOI: 10.1021/acs.accounts.6b00250. (34) Sakaki, S.; Kuroki, T.; Hamada, T. Synthesis of a new copper(i) complex, [Cu(tmdcbpy)2]+ (tmdcbpy = 4,4′,6,6′-tetramethyl-2,2′-bipyridine-5,5′-dicarboxylic acid), and its application to solar cells. J. Chem. Soc., Dalton Trans. 2002, (6), 840-842, 10.1039/B109518H. DOI: 10.1039/B109518H. (35) Bersuker, I. B. Pseudo-Jahn–Teller Effect—A Two-State Paradigm in Formation, Deformation, and Transformation of Molecular Systems and Solids. Chem. Rev. 2013, 113 (3), 1351-1390. DOI: 10.1021/cr300279n. (36) Cunningham, C. T.; Cunningham, K. L. H.; Michalec, J. F.; McMillin, D. R. Cooperative Substituent Effects on the Excited States of Copper Phenanthrolines. Inorganic Chemistry 1999, 38 (20), 4388-4392. DOI: 10.1021/ic9906611. (37) Lavie-Cambot, A.; Cantuel, M.; Leydet, Y.; Jonusauskas, G.; Bassani, D. M.; McClenaghan, N. D. Improving the photophysical properties of copper(I) bis(phenanthroline) complexes. Coord. Chem. Rev. 2008, 252 (23), 2572-2584. DOI: 10.1016/j.ccr.2008.03.013. (38) Dietrich-Buchecker, C. O.; Marnot, P. A.; Sauvage, J.-P.; Kirchhoff, J. R.; McMillin, D. R. Bis(2,9-diphenyl-1,10-phenanthroline)copper(I): a copper complex with a long-lived charge- transfer excited state. J. Chem. Soc., Chem. Commun. 1983, (9), 513-515, 10.1039/C39830000513. DOI: 10.1039/C39830000513. 238 (39) Muniz, C. N.; Schaab, J.; Razgoniaev, A.; Djurovich, P. I.; Thompson, M. E. π-Extended Ligands in Two-Coordinate Coinage Metal Complexes. Journal of the American Chemical Society 2022, 144 (39), 17916-17928. DOI: 10.1021/jacs.2c06948. (40) Hamze, R.; Peltier, J. L.; Sylvinson, D.; Jung, M.; Cardenas, J.; Haiges, R.; Soleilhavoup, M.; Jazzar, R.; Djurovich, P. I.; Bertrand, G.; et al. Eliminating nonradiative decay in Cu(I) emitters: >99% quantum efficiency and microsecond lifetime. Science 2019, 363 (6427), 601. DOI: 10.1126/science.aav2865. (41) Hamze, R.; Idris, M.; Muthiah Ravinson, D. S.; Jung, M. C.; Haiges, R.; Djurovich, P. I.; Thompson, M. E. Highly Efficient Deep Blue Luminescence of 2-Coordinate Coinage Metal Complexes Bearing Bulky NHC Benzimidazolyl Carbene. Frontiers in Chemistry 2020, 8, 401, 10.3389/fchem.2020.00401. (42) Shi, S.; Jung, M. C.; Coburn, C.; Tadle, A.; Sylvinson M. R, D.; Djurovich, P. I.; Forrest, S. R.; Thompson, M. E. Highly Efficient Photo- and Electroluminescence from Two-Coordinate Cu(I) Complexes Featuring Nonconventional N-Heterocyclic Carbenes. Journal of the American Chemical Society 2019, 141 (8), 3576-3588. DOI: 10.1021/jacs.8b12397. (43) Hamze, R.; Shi, S.; Kapper, S. C.; Muthiah Ravinson, D. S.; Estergreen, L.; Jung, M.-C.; Tadle, A. C.; Haiges, R.; Djurovich, P. I.; Peltier, J. L.; et al. “Quick-Silver” from a Systematic Study of Highly Luminescent, Two-Coordinate, d10 Coinage Metal Complexes. Journal of the American Chemical Society 2019, 141 (21), 8616-8626. DOI: 10.1021/jacs.9b03657. (44) Tang, R.; Xu, S.; Lam, T.-L.; Cheng, G.; Du, L.; Wan, Q.; Yang, J.; Hung, F.-F.; Low, K.-H.; Phillips, D. L.; et al. Highly Robust CuI-TADF Emitters for Vacuum-Deposited OLEDs with Luminance up to 222 200 cd m−2 and Device Lifetimes (LT90) up to 1300 hours at an Initial Luminance of 1000 cd m−2. Angew. Chem. Int. Ed. 2022, 61 (33), e202203982. DOI: 10.1002/anie.202203982. (45) Bertrand, B.; Romanov, A. S.; Brooks, M.; Davis, J.; Schmidt, C.; Ott, I.; O'Connell, M.; Bochmann, M. Synthesis, structure and cytotoxicity of cyclic (alkyl)(amino) carbene and acyclic carbene complexes of group 11 metals. Dalton Transactions 2017, 46 (45), 15875-15887, 10.1039/C7DT03189K. DOI: 10.1039/C7DT03189K. (46) Chotard, F.; Sivchik, V.; Linnolahti, M.; Bochmann, M.; Romanov, A. S. Mono- versus Bicyclic Carbene Metal Amide Photoemitters: Which Design Leads to the Best Performance? Chem. Mater. 2020, 32 (14), 6114-6122. DOI: 10.1021/acs.chemmater.0c01769. (47) Conaghan, P. J.; Matthews, C. S. B.; Chotard, F.; Jones, S. T. E.; Greenham, N. C.; Bochmann, M.; Credgington, D.; Romanov, A. S. Highly efficient blue organic light-emitting diodes based on carbene-metal-amides. Nature Communications 2020, 11 (1), 1758. DOI: 10.1038/s41467-020- 15369-8. (48) Gernert, M.; Balles-Wolf, L.; Kerner, F.; Müller, U.; Schmiedel, A.; Holzapfel, M.; Marian, C. M.; Pflaum, J.; Lambert, C.; Steffen, A. Cyclic (Amino)(aryl)carbenes Enter the Field of Chromophore Ligands: Expanded π System Leads to Unusually Deep Red Emitting CuI 239 Compounds. Journal of the American Chemical Society 2020, 142 (19), 8897-8909. DOI: 10.1021/jacs.0c02234. (49) Romanov, A. S.; Jones, S. T. E.; Gu, Q.; Conaghan, P. J.; Drummond, B. H.; Feng, J.; Chotard, F.; Buizza, L.; Foley, M.; Linnolahti, M.; et al. Carbene metal amide photoemitters: tailoring conformationally flexible amides for full color range emissions including white-emitting OLED. Chemical Science 2020, 11 (2), 435-446, 10.1039/C9SC04589A. DOI: 10.1039/C9SC04589A. (50) Karon, K.; Lapkowski, M. Carbazole electrochemistry: a short review. J. Solid State Electrochem. 2015, 19 (9), 2601-2610. DOI: 10.1007/s10008-015-2973-x. (51) Epifanovsky, E.; Gilbert, A. T. B.; Feng, X.; Lee, J.; Mao, Y.; Mardirossian, N.; Pokhilko, P.; White, A. F.; Coons, M. P.; Dempwolff, A. L.; et al. Software for the frontiers of quantum chemistry: An overview of developments in the Q-Chem 5 package. The Journal of Chemical Physics 2021, 155 (8), 084801. DOI: 10.1063/5.0055522 (acccessed 2023/01/08). (52) Lu, T.; Chen, F. Multiwfn: A multifunctional wavefunction analyzer. J. Comput. Chem. 2012, 33 (5), 580-592. DOI: 10.1002/jcc.22885 (acccessed 2023/01/08). (53) Liebeskind, L. S.; Leeds, J. P.; Baysdon, S. L.; Iyer, S. Convergent, high-yield, regioselective synthesis of substituted benzoquinones from maleoylcobalt complexes and alkynes. Journal of the American Chemical Society 1984, 106 (21), 6451-6453. DOI: 10.1021/ja00333a075. (54) Hou, S.; Yang, H.; Cheng, B.; Zhai, H.; Li, Y. Cobaloxime-catalyzed hydration of terminal alkynes without acidic promoters. Chem. Commun. 2017, 53 (51), 6926-6929, 10.1039/C7CC03919K. DOI: 10.1039/C7CC03919K. (55) Neogrdy, P.; Kell, V.; Urban, M.; Sadlej, A. J. Ionization potentials and electron affinities of Cu, Ag, and Au: Electron correlation and relativistic effects. International Journal of Quantum Chemistry 1997, 63 (2), 557-565. DOI: 10.1002/(sici)1097-461x(1997)63:2<557::Aid- qua25>3.0.Co;2-3. (56) Li, T.-y.; Schaab, J.; Djurovich, P. I.; Thompson, M. E. Toward rational design of TADF two- coordinate coinage metal complexes: understanding the relationship between natural transition orbital overlap and photophysical properties. Journal of Materials Chemistry C 2022, 10 (12), 4674-4683. DOI: 10.1039/d2tc00163b. (57) Pellegrin, Y.; Sandroni, M.; Blart, E.; Planchat, A.; Evain, M.; Bera, N. C.; Kayanuma, M.; Sliwa, M.; Rebarz, M.; Poizat, O.; et al. New Heteroleptic Bis-Phenanthroline Copper(I) Complexes with Dipyridophenazine or Imidazole Fused Phenanthroline Ligands: Spectral, Electrochemical, and Quantum Chemical Studies. Inorganic Chemistry 2011, 50 (22), 11309- 11322. DOI: 10.1021/ic2006343. (58) Ruthkosky, M.; Castellano, F. N.; Meyer, G. J. Photodriven Electron and Energy Transfer from Copper Phenanthroline Excited States. Inorganic Chemistry 1996, 35 (22), 6406-6412. DOI: 10.1021/ic960503z. 240 (59) Das, S.; Rodrigues, R. R.; Lamb, R. W.; Qu, F.; Reinheimer, E.; Boudreaux, C. M.; Webster, C. E.; Delcamp, J. H.; Papish, E. T. Highly Active Ruthenium CNC Pincer Photocatalysts for Visible-Light-Driven Carbon Dioxide Reduction. Inorganic Chemistry 2019, 58 (12), 8012-8020. DOI: 10.1021/acs.inorgchem.9b00791. (60) Huckaba, A. J.; Shirley, H.; Lamb, R. W.; Guertin, S.; Autry, S.; Cheema, H.; Talukdar, K.; Jones, T.; Jurss, J. W.; Dass, A.; et al. A Mononuclear Tungsten Photocatalyst for H2 Production. ACS Catalysis 2018, 8 (6), 4838-4847. DOI: 10.1021/acscatal.7b04242. (61) Tinker, L. L.; Bernhard, S. Photon-Driven Catalytic Proton Reduction with a Robust Homoleptic Iridium(III) 6-Phenyl-2,2′-bipyridine Complex ([Ir(C/\N/\N)2]+). Inorganic Chemistry 2009, 48 (22), 10507-10511. DOI: 10.1021/ic900777g. (62) Ma, B.; Djurovich, P.; Thompson, M. Excimer and electron transfer quenching studies of a cyclometalated platinum complex. Coord. Chem. Rev. 2005, 249, 1501-1510. DOI: 10.1016/j.ccr.2005.02.004. (63) Rosspeintner, A.; Kattnig, D. R.; Angulo, G.; Landgraf, S.; Grampp, G. The Rehm–Weller Experiment in View of Distant Electron Transfer. Chemistry – A European Journal 2008, 14 (20), 6213-6221. DOI: 10.1002/chem.200701841 (acccessed 2023/01/26). (64) Fukuzumi, S.; Fujita, M.; Otera, J.; Fujita, Y. Electron-transfer oxidation of ketene silyl acetals and other organosilanes. Mechanistic insight into Lewis acid mediated electron transfer. Journal of the American Chemical Society 1992, 114 (26), 10271-10278. (65) Fischer, S.; Rösel, A.; Kammer, A.; Barsch, E.; Schoch, R.; Junge, H.; Bauer, M.; Beller, M.; Ludwig, R. Diferrate [Fe2(CO)6(μ-CO){μ-P(aryl)2}]− as Self-Assembling Iron/Phosphor-Based Catalyst for the Hydrogen Evolution Reaction in Photocatalytic Proton Reduction—Spectroscopic Insights. Chemistry – A European Journal 2018, 24 (60), 16052-16065. DOI: 10.1002/chem.201802694 (acccessed 2023/01/08). (66) Natali, M. Elucidating the Key Role of pH on Light-Driven Hydrogen Evolution by a Molecular Cobalt Catalyst. ACS Catalysis 2017, 7 (2), 1330-1339. DOI: 10.1021/acscatal.6b03087. (67) Khnayzer, R. S.; McCusker, C. E.; Olaiya, B. S.; Castellano, F. N. Robust Cuprous Phenanthroline Sensitizer for Solar Hydrogen Photocatalysis. Journal of the American Chemical Society 2013, 135 (38), 14068-14070. DOI: 10.1021/ja407816f. (68) Collin, M.; Claire, A.; Jack, A.; Anushan, A.; Jonas, S.; Peter, D.; Mark, T. Two-Coordinate Coinage Metal Complexes as Solar Photosensitizers. ChemRxiv. Cambridge: Cambridge Open Engage;, 2023. (69) Jones, W. E., Jr.; Fox, M. A. Determination of Excited-State Redox Potentials by Phase- Modulated Voltammetry. The Journal of Physical Chemistry 1994, 98 (19), 5095-5099. DOI: 10.1021/j100070a025. 241 (70) Scandola, F.; Balzani, V. Free-energy relationships for electron-transfer processes. Journal of the American Chemical Society 1979, 101 (20), 6140-6142. DOI: 10.1021/ja00514a057. (71) Rehm, D.; Weller, A. Kinetics of Fluorescence Quenching by Electron and H-Atom Transfer. Isr. J. Chem. 1970, 8 (2), 259-271. DOI: 10.1002/ijch.197000029. (72) Sakamoto, M.; Kim, S. S.; Fujitsuka, M.; Majima, T. Reversible Intramolecular Triplet−Triplet Energy Transfer in Benzophenone-N-methylphthalimide Dyad. The Journal of Physical Chemistry A 2008, 112 (7), 1403-1407. DOI: 10.1021/jp075885g. (73) Kellogg, M.; Mencke, A.; Muniz, C.; Ahammad, T.; Cardoso-Delgado, F.; Baluyot-Reyes, N.; Sewell, M.; Bird, M.; Bradforth, S.; Thompson, M. Intra- and inter-molecular charge transfer dynamics of carbene-metal-amide photosensitizers. 2023. DOI: 10.26434/chemrxiv-2023-6tqv9. (74) Khnayzer, R. S.; Olaiya, B. S.; El Roz, K. A.; Castellano, F. N. Homogeneous Photocatalytic H2 Production Using a RuII Bathophenanthroline Metal-to-Ligand Charge-Transfer Photosensitizer. ChemPlusChem 2016, 81 (10), 1016-1016. DOI: 10.1002/cplu.201600424 (acccessed 2023/03/02). (75) O'Neill, J. S.; Kearney, L.; Brandon, M. P.; Pryce, M. T. Design components of porphyrin- based photocatalytic hydrogen evolution systems: A review. Coord. Chem. Rev. 2022, 467, 214599. DOI: 10.1016/j.ccr.2022.214599. (76) Robertson, W. D.; Bovell, A. M.; Warncke, K. Cobinamide production of hydrogen in a homogeneous aqueous photochemical system, and assembly and photoreduction in a (βα)8 protein. JBIC Journal of Biological Inorganic Chemistry 2013, 18 (6), 701-713. DOI: 10.1007/s00775- 013-1015-3. (77) Lazarides, T.; McCormick, T.; Du, P.; Luo, G.; Lindley, B.; Eisenberg, R. Making Hydrogen from Water Using a Homogeneous System Without Noble Metals. Journal of the American Chemical Society 2009, 131 (26), 9192-9194. DOI: 10.1021/ja903044n. (78) Dempsey, J. L.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Hydrogen Evolution Catalyzed by Cobaloximes. Acc. Chem. Res. 2009, 42 (12), 1995-2004. DOI: 10.1021/ar900253e. (79) Hasegawa, E.; Takizawa, S.; Seida, T.; Yamaguchi, A.; Yamaguchi, N.; Chiba, N.; Takahashi, T.; Ikeda, H.; Akiyama, K. Photoinduced electron-transfer systems consisting of electron-donating pyrenes or anthracenes and benzimidazolines for reductive transformation of carbonyl compounds. Tetrahedron 2006, 62 (27), 6581-6588. DOI: 10.1016/j.tet.2006.03.061. (80) Pellegrin, Y.; Odobel, F. Sacrificial electron donor reagents for solar fuel production. Comptes Rendus Chimie 2017, 20 (3), 283-295. DOI: 10.1016/j.crci.2015.11.026. (81) Narouz, M. R.; Takano, S.; Lummis, P. A.; Levchenko, T. I.; Nazemi, A.; Kaappa, S.; Malola, S.; Yousefalizadeh, G.; Calhoun, L. A.; Stamplecoskie, K. G.; et al. Robust, Highly Luminescent Au(13) Superatoms Protected by N-Heterocyclic Carbenes. J. Am. Chem. Soc. 2019, 141 (38), 14997-15002. DOI: 10.1021/jacs.9b07854 From NLM PubMed-not-MEDLINE. 242 (82) Kohler, L.; Niklas, J.; Johnson, R. C.; Zeller, M.; Poluektov, O. G.; Mulfort, K. L. Molecular cobalt catalysts for H2 generation with redox activity and proton relays in the second coordination sphere. Inorganic Chemistry 2018, 58 (2), 1697-1709. (83) Barrozo, A.; Orio, M. Molecular electrocatalysts for the hydrogen evolution reaction: input from quantum chemistry. ChemSusChem 2019, 12 (22), 4905-4915. (84) Föller, J.; Marian, C. M. Rotationally Assisted Spin-State Inversion in Carbene–Metal– Amides Is an Artifact. The Journal of Physical Chemistry Letters 2017, 8 (22), 5643-5647. DOI: 10.1021/acs.jpclett.7b02701. (85) Wang, H.-J.; Liu, Y.; Yu, B.; Song, S.-Q.; Zheng, Y.-X.; Liu, K.; Chen, P.; Wang, H.; Jiang, J.; Li, T.-Y. A Configurationally Confined Thermally Activated Delayed Fluorescent Two- Coordinate CuI Complex for Efficient Blue Electroluminescence. Angew. Chem. Int. Ed. 2023, 62 (7), e202217195. DOI: 10.1002/anie.202217195 (acccessed 2023/03/02). (86) Gu, Q.; Chotard, F.; Eng, J.; Reponen, A.-P. M.; Vitorica-Yrezabal, I. J.; Woodward, A. W.; Penfold, T. J.; Credgington, D.; Bochmann, M.; Romanov, A. S. Excited-State Lifetime Modulation by Twisted and Tilted Molecular Design in Carbene-Metal-Amide Photoemitters. Chem. Mater. 2022, 34 (16), 7526-7542. DOI: 10.1021/acs.chemmater.2c01938. (87) Garner, M. E.; Niu, W.; Chen, X.; Ghiviriga, I.; Abboud, K. A.; Tan, W.; Veige, A. S. N- heterocyclic carbene gold(i) and silver(i) complexes bearing functional groups for bio-conjugation. Dalton Transactions 2015, 44 (4), 1914-1923, 10.1039/C4DT02850C. DOI: 10.1039/C4DT02850C. (88) Zou, J.; Berg, D. J.; Stuart, D.; McDonald, R.; Twamley, B. Carbazole-bis(oxazolines) as Monoanionic, Tridentate Chelates in Lanthanide Chemistry: Synthesis and Structural Studies of Thermally Robust and Kinetically Stable Dialkyl and Dichloride Complexes. Organometallics 2011, 30 (18), 4958-4967. DOI: 10.1021/om200588y. (89) Lu, Z.; Gao, Y.; Chen, H.; Liu, Z.; Sun, L. Water oxidation catalyzed by a charge-neutral mononuclear ruthenium(iii) complex. Dalton Transactions 2017, 46 (4), 1304-1310, 10.1039/C6DT04160D. DOI: 10.1039/C6DT04160D. (90) Kincaid, J. R. A.; Wong, M. J.; Akporji, N.; Gallou, F.; Fialho, D. M.; Lipshutz, B. H. Introducing Savie: A Biodegradable Surfactant Enabling Chemo- and Biocatalysis and Related Reactions in Recyclable Water. Journal of the American Chemical Society 2023, 145 (7), 4266- 4278. DOI: 10.1021/jacs.2c13444.

Abstract (if available)

Abstract Generating renewable fuel from sunlight and designing molecules for organic-LED (OLED) displays are two prominent fields of modern research. While these technologies are very different in what they aim to accomplish, they are united by their fundamental starting point: namely, the excited state of a molecule. This dissertation highlights accomplishments in both fields. Chapter 1 provides a top-level overview of solar fuels and OLED technology. The technical details of the molecular photophysics that govern these technologies are presented in Chapter 2. Chapter 3 highlights work that was published in J. Am. Chem. Soc on the discovery of blue emissive molecules that have record fast radiative rates involving triplet states. An explanation of the structural design that enables fast photon emission is given, and future work to further improve the radiative rate is proposed. Chapter 4 highlights a new class of molecular photosensitizers for solar fuels that feature abundant metal complexes that was published in J. Am. Chem. Soc. We demonstrate figures of merit that are comparable to scarce metal complexes such as ruthenium and iridium. The photosensitzers are paired with a cobalt electrocatalyst and sacrificial reductant to convert water to hydrogen upon irradiation with a blue LED.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Simple complexes: synthesis and photophysical studies of luminescent, monovalent, 2-coordinate carbene-coinage metal complexes and higher coordination geometries

PDF

Catalysis for solar fuels production: molecular and materials approaches

PDF

Design of catalysts for the transformation of abundant small molecules into solar fuels

PDF

Synthesis, photo- and electroluminescence of three- and two-coordinate coinage metal complexes featuring non-N-heterocyclic carbene and non-conventional N-heterocyclic carbene ligands

PDF

Photophysical properties of luminescent iridium and coinage metal complexes

PDF

Utilizing n-heterocyclic chromophores for solar energy harvesting.

PDF

Synthesis, structural and photophysical characterization of phosphorescent three-coordinate Cu(I)-N-heterocyclic carbene complexes

PDF

Molecular design strategies for blue organic light emitting diodes

PDF

Cobalt phosphino-thiolate complexes: applications towards electrocatalytic small molecule conversion to synthetic fuels

PDF

Molecular design for organic photovoltaics: tuning state energies and charge transfer processes in heteroaromatic chromophores

PDF

Potential-induced formation and dissociation of adducts at interfaces and adsorption of phenols onto metal electrodes

PDF

Organic solar cells: molecular electronic processes and device development

PDF

Studies on direct oxidation formic acid fuel cells: advantages, limitations and potential

PDF

Efficient ternary blend bulk heterojunction solar cells with tunable open-circuit voltage

PDF

Molecular and morphological effects on the operational parameters of organic solar cells

PDF

Elucidating local electric fields and ion adsorption using benzonitrile: vibrational spectroscopic studies at interfaces and in bulk

PDF

Synthesis and characterization of novel donor-acceptor systems for solar water splitting

PDF

Organic materials for linear and nonlinear optical microdevices

PDF

Two-dimensional metal dithiolene metal-organic frameworks as conductive materials for solar energy conversion

PDF

Porphyrin based near infrared‐absorbing materials for organic photovoltaics

Asset Metadata

Creator Muniz, Collin Nicholas (author)

Core Title Two-coordinate coinage metal complexes for solar fuels and organic LED chemistry

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Degree Conferral Date 2023-08

Publication Date 07/10/2023

Defense Date 04/19/2023

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag carbene,hydrogen,inorganic chemistry,OAI-PMH Harvest,OLED,photophysics,solar fuels,water reduction

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Thompson, Mark (committee chair), Cronin, Stephen (committee member), Dawlaty, Jahan (committee member), Inkpen, Michael (committee member)

Creator Email collin.n.muniz@gmail.com,collinmu@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113263167

Unique identifier UC113263167

Identifier etd-MunizColli-12056.pdf (filename)

Legacy Identifier etd-MunizColli-12056

Document Type Thesis

Format theses (aat)

Rights Muniz, Collin Nicholas

Internet Media Type application/pdf

Type texts

Source 20230710-usctheses-batch-1065 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu