1 Structural Design and Catalytic Applications of Homogenous And

Home , Diimide, Organogold chemistry

Structural Design and Catalytic Applications of Homogenous and Heterogeneous Organometallic Lewis Acids

Dissertation

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

in the Graduate School of The Ohio State University

Benjamin Russell Reiner

Graduate Program in Chemistry

The Ohio State University

2018

Dissertation Committee

Casey R. Wade, Advisor

Christo S. Sevov

Hannah Shafaat

Copyrighted by

Benjamin Russell Reiner

2018

Abstract

A broad goal of organometallic chemistry to use ligand and catalyst design to effect useful chemical transformations. Rational design of novel ligands has allowed us to study how binuclear, redox active, or supramolecular frameworks can engender new or enhanced chemical reactivity.

One of the major barriers to adopting gold into traditional two electron redox cycling is the high potential required to access gold(III). Use of a binuclear gold framework can eschew this issue by distributing the redox load over multiple metal sites. Phosphorus ylide gold dimers support a wide range of oxidation states which has allowed study into the reductive processes involved in the thermolysis of arylated or alkylated gold(III,III) complexes. Additionally, the binuclear framework allows access to a dicationic gold(II,II) complex that exhibits markedly superior Lewis acid mediated catalytic activity compared to gold complexes supported by other ligands or in other oxidation states.

While rylene imides have a variety of applications in electrical energy storage devices, their use as redox non-innocent ligands has been severely underexplored. Rigorous electrochemical and photochemical investigations have shown how reduced naphthalene diimide (NDI) species interact with charge dense redox inactive cations such as Mg2+ or

Li+. Moreover, chemical reduction revealed formation of a discrete dimeric complex that features strong coordination of the NDI oxygen atoms to the Mg centers. Electrochemical

i and structural studies into the interaction between rylene imides and transition metals was accomplished by using NDI or phthalimide ligands decorated with pyridyl-thiazole units.

Cyclic voltammetry of the ensuing Co2+ and Zn2+ complexes uncovered important structure function relationships between the redox state of the ligand and the accessibility of metal- borne reduction events.

Metal-organic frameworks (MOFs) are a hybrid class of material which allow translation of well-established principles in homogenous catalysis into a heterogenous context. Diphosphine pincer complexes facilitate a multitude of impressive catalytic processes but can suffer from deactivation processes that limit their long term activity.

Immobilization of diphosphine pincer complexes as linkers within MOFs can extend catalytic lifetime by suppressing deleterious side reactions. Lattice immobilized Pd PNNNP

N N − (P N P = 2,6-(HNPAr2)2C5H3N; Ar = p-C6H4CO2 ) pincer complexes demonstrate longer catalytic lifetimes in Lewis acid mediated catalysis compared to homogenous analogs.

Additionally, immobilized Pd PNNNP pincer complexes were found to exhibit size selective catalytic activity which was not for a homogenous analog.

Dedication

Tho' much is taken, much abides; and tho' We are not now that strength which in old days Moved earth and heaven, that which we are, we are; One equal temper of heroic hearts, Made weak by time and fate, but strong in will To strive, to seek, to find, and not to yield.

For Katie

iii

Acknowledgments

Science, like life, is never accomplished alone and to stand on the shoulders of giants and call yourself tall is morally unacceptable. Below are people who are smarter, wittier, stronger, more disciplined, more handsome, and more ethically sound than I could ever hope to be. Here be giants.

My research advisor, Prof. Casey Wade is a tremendously good scientist and his mentorship and friendship have driven me to be a markedly better chemist. His unwavering sense of scientific rigor and demands for perfection continue to resonate with me to this day. His patience in training and tolerating bad jokes have not gone unnoticed and never will. My undergraduate advisor, Prof. Jeffery Byers, never treated me any different than a graduate student which was both exciting and more often horrifying. I enjoyed this sort of trial-by-fire training and I owe an absurd amount of my knowledge and success to his early mentorship. Dr. Christine Goldman inspired me to stay in chemistry. Her enthusiasm and passion for teaching is underappreciated and she deserves a thousand thanks and a tripled salary. I’d also like to thank Dr. David Healey, my music instructor through college, for teaching me two things: 1) proper breath control determines quality of sound and 2) saying someone “has potential” should be wildly insulting because if you have potential you should use it. Dr. Sue Pochapsky trained me how to use and think about NMR spectroscopy and I always desperately appreciated any knowledge she would spare during cryogen fills.

I’d also like to acknowledge Prof. Christine Thomas for, purposely or not, being a wonderful mentor and friend.

During my graduate tenure, I have had the opportunity to work with some of the kindest and most intelligent people I’ve ever had the good fortune to meet. Sarah

Baranowski has a deliciously simple outlook on life and draws success from the air. Mark

Bezpalko was (and still is) easily bribed with banana bread and beer to solve crystal structures quickly and can play a mean game of soccer. Keith Fritzsching and I had more late night science talks than I care to admit but, for better and worse (usually worse), we always went back to lab. Cami Schneider, Jeff Slater, and Andrew Chen were unnecessarily helpful during the early lab set up. This included both helping us find chemicals and the good beer. Katie Gramigna is the smartest and sweetest person I know. Her friendship means the world to me and always will. I’m not sure I ever deserved her.

My roommate Jeff Mckee lived with me for seven years through college and most of graduate school. We ran the Boston Marathon twice together. For lack of the proper paperwork and kicking us into a higher tax bracket we could have been common law married. He was and is a great friend. Nate Walka and Connor Rooney are better people and teachers than I will ever fathom and I have always enjoyed our every conversation on teaching, writing, the role of the modern scientist, and why Muppet Treasure Island is the greatest movie ever made.

I’d like to thank our past Wade Lab postdocs and current lab members. I enjoyed working with Abebu Kassie and eating her Ethiopian food. I wish I got to do more of both.

Samantha Burgess was a brilliant labmate and Neil Mucha was always kind even when watching BC lose to Clemson live on national television.

To my parents: your support has never wavered or faltered. It’s your fault I’m a chemist. To Katie Bien: all these people had to tolerate me some of the time but you had to tolerate me all of the time. For this and all, thank you and I love you.

Vita

May 2013...... B.S. Chemistry, Boston College

2013 – 2015...... Graduation Teaching Assistant, Brandeis

University

2015 – 2017...... NMR Teaching Assistant, Brandeis

University

2018 – present...... Graduate Research Assistant, Department

of Chemistry, The Ohio State University

Publications

Reiner, B. R.; Mucha, N. T.;Rothstein, A; Temme S. J.; Duan, P.; Schmidt-Rohr, K.; Foxman, B. M.; Wade. C. R. Zirconium Metal-Organic Frameworks Assembled from Pd and Pt PNNNP Pincer Complexes: Synthesis, Postsynthetic Modification, and Lewis Acid Catalysis. Inorg. Chem. 2018. 57, 2663–2672

Reiner, B. R.; Foxman, B. M.; Wade, C.R. Electrochemical and Structural Investigation of the Interactions between Naphthalene Diimides and Metal Cations. Dalton Trans. 2017. 46, 9472 – 9480.

Reiner, B. R.; Bezpalko, M. W.;Foxman, B. M.; Wade, C.R. Lewis Acid Catalysis with Cationic Dinuclear Gold(II,II) and Gold(III,III) Phosphorus Ylide Complexes. Organometallics. 2016, 35, 2830 – 2935. (selected as Cover Art)

Fields of Study

Major Field: Chemistry

vii

Table of Contents

Abstract ...... i Dedication ...... iii Acknowledgments ...... iv Vita ...... vii Table of Contents ...... viii List of Tables ...... xi List of Figures ...... xii Chapter 1: Introduction ...... 1 1.1 Ligand Design and Catalysis ...... 1 1.2. The catalytic and stoichiometric activity of binuclear gold ylide complexes ...... 2 1.3 Rylene imides as redox non-innocent ligands ...... 5 1.4 Immobilization of diaryl pincer complexes in metal-organic frameworks ...... 10 Chapter 2: Reactivity of Alkylated and Arylated Dinuclear Gold Ylide Dimers ...... 15 2.1 Introduction ...... 15 2.2 Synthesis and physical characterization of alkylated dinuclear gold(III,III) phosphorus ylide complexes ...... 17 2.3 Thermolysis of 6a...... 19 2.4 Mechanistic investigation into the thermolysis of 8 and 9...... 25 2.5 Conclusions ...... 35 2.6 Experimental ...... 36 Chapter 3: Lewis Acid Catalysis with Cationic Dinuclear Gold(II,II) and Gold(III,III) Phosphorus Ylide Complexes ...... 40 3.1 Introduction ...... 40 3.2 Synthesis and structure of cationic dinuclear gold(II,II) and gold(III,III) phosphorus ylide complexes ...... 41 3.3 Catalytic Mukaiyama addition reactions...... 44

viii

3.4 Kinetic correlation to substrate exchange rate ...... 48 3.5 Catalytic hydroamination reactions ...... 54 3.6 Conclusions ...... 58 3.7 Experimental ...... 59 Chapter 4: Interactions of Naphthalene Diimides and Metal Cations ...... 68 4.1 Introduction ...... 68 4.2 Electrochemical investigation into the interactions of naphthalene diimides and metal cations...... 70 4.3 Photochemical study exploring the behavior of reduced naphthalene diimides ..... 75 4.4 Coordination chemistry of reduced naphthalene diimide–metal cation complexes 79 4.5 Conclusions ...... 85 4.6 Experimental ...... 86 Chapter 5: Structural and Electrochemical Investigation into Transition Metal–Rylene Imide Complexes...... 93 5.1 Introduction ...... 93 2+ 2+ 5.2 Structural and electrochemical study of Co and Zn –NDI complexes ...... 96 2+ 2+ 5.3 Structural and electrochemical extension to Co and Zn –phthalimide complexes ...... 105 5.5 Pd0–MNTP Adduct: A New Coordination Mode for NDIs ...... 109 5.4 Conclusions ...... 113 5.5 Experimental ...... 115 Chapter 6: Synthesis and Catalytic Activity of a Pd PNNNP Pincer Complex Immobilized in a Zr Metal-Organic Framework ...... 126 6.1 Introduction ...... 126 6.2 Synthesis and characterization of Zr MOF assembled from a Pd PNNNP pincer complex ...... 130 6.3 Postsynthetic halide exchange reactions ...... 137 6.4 Initial catalytic studies ...... 141

6.5 Activation with NOBF4 ...... 146

6.6 Catalytic activity and kinetics of 38–PdBF4...... 152 6.7 Catalytic pore selectivity ...... 157 6.8 Catalytic carbonyl-ene cyclization with citronellal ...... 159 6.9 Conclusions ...... 163 6.10 Experimental...... 165 ix

Bibliography ...... 174

List of Tables

Table 1. Gold catalyzed Mukaiyama additions...... 45 Table 2. Gold-catalyzed hydroamination...... 55 Table 3. Gold-catalyzed cascade reaction...... 58 Table 4. Crystallographic data and refinement parameters for 15-OTf2 and 5-OTf2...... 63 ·−/2− Table 5. Shifts in cathodic peak potential (DEpc) of the NDI redox couple upon addition of Li+ or Mg2+ in different solvents and formation constants log(b)...... 75 Table 6. Crystallographic data and refinement parameters for 27...... 91 Table 7. Crystallographic data and refinement parameters for 30, 31, 32, and 33...... 120 Table 8. Crystallographic data and refinement parameters for 33 and 34...... 121 Table 9. Hydroamination of o-alkynyl aniline 39...... 143 Table 10. Hydroamination of o-alkynyl aniline 39...... 154 Table 11. Cyclization of o-ethynyl aniline 42...... 159 Table 12. Cyclization of citronellal (46) to isopulegol (47)...... 162

List of Figures

Figure 1. Select examples of redox-non innocent ligands in catalysis...... 6 Figure 2. Rylene imides and their redox potentials...... 8 Figure 3. General structure of PZEZP pincer complexes ...... 13 Figure 4. Synthesis of 6a...... 16 Figure 5. Synthesis of binuclear gold A-frames...... 18 Figure 6. Reaction profile for the thermolysis of 6a in o-dichlorobenzene at 150 °C. .... 20 Figure 7. a) Reported isomerization pathways for binuclear gold PY complexes. b) Proposed isomerization – reductive elimination sequence for 6a...... 21 31 1 Figure 8. P{ H} spectra of the thermolysis of 6a in the presence of PPh3 or CuI...... 22 Figure 9. Synthesis of 8 and 9...... 24 Figure 10. Thermolysis reactions of 8 and 9...... 25 Figure 11. Reaction profile for the thermolysis of 8 monitored by 31P{1H} NMR spectroscopy...... 27 1 31 Figure 12. H{ P} NMR spectrum (CDCl3) of the thermolysis of 8 after 3 hours...... 28 1 Figure 13. H NMR spectrum (CDCl3) of 11 generated in situ via thermolysis of 8...... 30 1 Figure 14. H NMR spectrum (CDCl3) of 12 isolated from the thermolysis of 8 in DMSO solution...... 32 Figure 15. Reaction profile for the thermolysis of 9 monitored by 19F NMR spectroscopy...... 34 Figure 16. Synthesis of 15-OTf2 and 5-OTf2...... 42 Figure 17. Solid-state structures of 15-OTf2 (a) and 5-OTf2 (b)...... 44 Figure 18. Reaction profile for the addition of 16 to crotonaldehyde (17a) in the presence of 15-OTf2 (circles), 5-OTf2 (triangles), or Ph3PAuOTf (squares)...... 47 31 1 Figure 19. P{ H} NMR spectrum (25 °C) of IPrAuOTf, 5-OTf2, PPh3AuOTf, and 15- OTf2 in the presence of 2 equivalents of OPEt3 in CH2Cl2...... 49 31 Figure 20. P{1H} NMR spectrum of 15-OTf2 with 2 equivalents of PEt3 in CH2Cl2. . 50 31 Figure 21. P NMR spectra of 5-OTf2 in the presence of 2 -5 equivalents of Et3PO in CH2Cl2...... 51 31 1 Figure 22. Variable temperature P{ H} spectra of 5-OTf2 in the presence of 2 equivalents of Et3PO (left) and van’t Hoff plot (right)...... 52 Figure 23. Reaction profile for the addition of 16 to crotonaldehyde (17a) in the presence of 4 mol % 5-OTf2 (red circles) fit to an exponential function (black lines)...... 53 Figure 24. %Vbur steric maps for 15-OTf2 (left) and 5-OTf2 (right)...... 56 Figure 25. Cathodic sweeps of the cyclic voltammograms of the electrochemical titration of Dipp2NDI...... 71 xii

n+ + Figure 26. Plots of DEpc versus log[M ] for titration with Li in MeCN (a), DMF (b), or DMSO (d) and Mg2+ in DMF (c)...... 73 Figure 27. (a) UV-vis difference spectra measured for a THF solution of Dipp2NDI (8.9 µM), LiPF6 (180 µM), and Et3N (0.17 M) upon irradiation. Samples were irradiated at 365 nm for 0 – 480 s (green to blue to red colour transitions). The spectrum measured at t = 0 s served as a baseline. Also shown are images of the reaction vessel illustrating qualitative colour changes during the experiment. (b) UV-vis spectra measured on a THF solution of Dipp2NDI (8.9 µM), metal salt (180 µM), and Et3N (0.17 M). Samples were irradiated at 365 nm for 600 s...... 77 Figure 28. UV-vis spectra after UV irradiation (365 nm) of a solution of Dipp2NDI (8.9 µM), additive (180 µM), and Et3N (0.17 M) in MeCN (lef) or DMF (right) for 10 minutes...... 77 2– Figure 29. Proposed mechanism for photoreduction of Dipp2NDI to [Dipp2NDI] in the presence of Li+ or Mg2+ (Mn+)...... 78 Figure 30. (a) UV-vis spectrum of a solution of Dipp2NDI (20 µM) in THF; (b) UV-vis ·− spectrum of [Dipp2NDI] generated in situ by addition of NaHDMS (100 µM); UV-vis ·− spectra of in situ generated [Dipp2NDI] in the presence of NaOTf (200 µM) (c), LiPF6 (200 µM) (d), and MgNTf2 (200 µM) (e)...... 78 2– Figure 31. Synthesis of Mg complexes of [Dipp2NDI] ...... 79 Figure 32. UV-vis spectrum of 26 in THF solution...... 80 Figure 33. ATR-IR spectrum of Dipp2NDI (black) and 26 (red)...... 81 Figure 34. TGA data for 26 measured using a ramp rate of 2 °C/min...... 82 Figure 35. Solid-state structure of 27...... 84 2– Figure 36. Syn and anti resonance structures of [R2NDI] ...... 85 Figure 37. DFT optimized structures and single point energies of syn and anti H2[Dipp2NDI]...... 92 Figure 38. Cyclic voltammograms (CVs) of Dipp2NDI in the presence of 0 (red), 1 (purple), or 2 (blue) equivalents of Co(OTf)2THF2...... 94 Figure 39. Comparison of bond angles between a pyridyl-thiazole and bipyridine ligand backbone...... 95 Figure 40. Synthesis of M(MNTP)2OTf2...... 97 Figure 41. Cyclic voltammograms (CV) of a) MNTP (blue), Co(MNTP)2OTf2 (30, red), st and Zn(MNTP)2OTf2 (31, green) and b) 1 derivative CV plots with respect to current. 99 Figure 42. Cyclic voltammograms (CV) of Co(MNTP)2OTf2 (30) at various concentrations of MeCN solution (a) or scan rates (b)...... 100 Figure 43. Solid state structure of Co(MNTP)2OTf2 (30)...... 104 Figure 44. Synthesis of [M(PTP)2]OTf2...... 106 Figure 45. Cyclic voltammograms (CV) of PTP (blue), [Co(PTP)2]OTf2 (32, red), and [Zn(PTP)2]OTf2 (33, green)...... 106 Figure 46. Solid state structure of a) [Co(PTP)2]OTf2 (32), b) [Zn(PTP)2]OTf2 (33), and c) overlay of structures of 32 (red) and 33 (blue)...... 109 Figure 47. UV-vis spectrum of 34 before (solid line) and after exposure to air (dashed line)...... 110 Figure 48. Solid state structure of 34...... 112 xiii

Figure 49. Reactivity of 34 with organic oxidants...... 112 Figure 50. General structure of PZEZP pincer complexes ...... 127 Figure 51. Synthesis of pincer complex H3L-PdI...... 130 Figure 52. (a) PXRD patterns of 37-PdX and 38-PdX (Cu Kα radiation, λ = 1.54 Å). (b) Defect-free framework structure of 38-PdX (left) and view of a portion of the framework showing ovoidal pores (right)...... 132 Figure 53. DFT differential pore volume plots for 37-PdX (blue) and 38-PdX (black) obtained from the respective N2 adsorption isotherms (77 K)...... 133 Figure 54. Solid-state 31P NMR spectrum of 38-PdX with magic-angle spinning (MAS) and total suppression of spinning sidebands (TOSS)...... 134 Figure 55. Solid-state 13C NMR spectrum of 38-PdX...... 136 Figure 56. Postsynthetic halide ligand exchange reactions of 38-PdX...... 137 Figure 57. 31P{1H} NMR spectra of digested samples of 38–PdX, 38–PdI, 38–PdX + PhI(TFA)2, 38–PdTFA, and 38–PdOTf/AgOTf...... 139 Figure 58. PXRD patterns of 38–PdX, 38–PdI, 38–PdTFA, and 38–PdX/AgOTf (Cu Kα, λ = 1.54 Å)...... 141 31P 1 Figure 59. { H} NMR spectrum of an acid-digested (3/1 CF3CO2H/C6D6) sample of 38–PdPMe3...... 144 Figure 60. Cyclic voltammograms (CV) of tBuL–PdCl (red), tBuL–PdI (blue), and NOBF4 (green)...... 147 Figure 61. Synthesis of 38–PdBF4...... 148 Figure 62. XRF spectrum of 38–PdI after 0 (black), 1 (blue), or 2 (red) soaks with NOBF4 in MeCN solution...... 148 Figure 63. (a) PXRD patterns of 38-PdI and 38-PdBF4. Data was collected with Cu Kα radiation. (b) N2 adsorption isotherms (77 K) for 38-PdI (black) and 38-PdBF4 (red) after desolvation at 150 °C and ~10-4 torr for 12 h...... 149 31 1 Figure 64. P{ H} NMR spectra of digested samples of 38–PdBF4, 38–PdTFA, and 38–PdI...... 151 Figure 65. ATR-IR spectrum of 38–PdI (black) and 38–PdBF4 (red)...... 152 Figure 66. Reaction profile for the cyclization of 2 in the presence of a) 38-PdBF4 or b) t Bu4L-PdBF4...... 156 Figure 67. DFT differential pore volume plots for 38-PdBF4 obtained from the respective N2 adsorption isotherms (77 K)...... 158 Figure 68. Reaction profile for the cyclization of citronellal at 100 or 200 mM in the presence of 38–PdBF4...... 161

xiv

Chapter 1: Introduction1

1.1 Ligand Design and Catalysis

Organometallic chemistry strives to overcome the inherent bias of transition or main group metals by imposing unique steric or electronic constraints with bespoke ligand platforms. While nature has used 4.5 billion years to carefully tailor enzymes for complex chemical transformations, the organometallic chemist must make do with substantially less time. Motivated, inspired, or more aptly outright stealing common architectures from nature has guided organometallic chemistry toward impressive advances in catalysis, electron transport, and molecular recognition. Rational ligand design requires a holistic approach which encompasses consideration of donor atoms, denticity, lability, ligand field strength, steric parameters, redox activity, and geometrical constraints as well as a host of other key thermodynamic and kinetic parameters in order to predict and affect how a complex will engage with incoming substrates or products.

The work described in this thesis endeavors to demonstrate how coopting simple ligand design strategies from nature can be advantageous in both stoichiometric and catalytic applications. Use of multinuclear platforms, adjacent redox active sites, and immobilization in a supramolecular framework are all ligand design strategies that nature employs to orchestrate complex but critical multielectron chemical transformations.

1 This dissertation follows the style and format of the Journal of the American Chemical Society. 1

1.2. The catalytic and stoichiometric activity of binuclear gold ylide complexes

Perhaps owing to its cultural significance, the chemistry of gold has been developed much later than its neighbors on the periodic table. However, within the last four decades the chemistry of gold has become a popular topic of research and technology, particularly for applications in catalysis. Gold has impressive applications in medicine, quantum chemistry, photophysics, and heterogeneous and homogenous catalytic and stoichiometric transformations.1 Interest in gold has stemmed from its low toxicity, tolerance to oxygen and water, and low proclivity to beta-hydride elimination. Gold also exhibits some exotic properties as a consequence of pronounced relativistic effects. Although there have been notable achievements utilizing gold in a variety of disciplines the present discussion will focus on the chemistry of well-defined, single-site, homogenous gold complexes.

Gold has a number of properties that can be exploited to engender novel and powerful reactivity. The low oxophilicity of gold allows Au-based complexes to tolerate water, alcohols, and oxygen. Gold-alkyl complexes also exhibit a low proclivity toward beta-hydride elimination3 although recent work by the Bourissou and coworkers has shown this reluctance only extends to gold(I) complexes.2,3 Beta-hydride elimination can be deleterious in cross coupling reactions by mediating off-cycle processes and decreasing catalytic efficiency. One of the major barriers to using gold for traditional two electron redox cycling in cross coupling reactions is the high oxidative barrier required to access gold(III). The two electron oxidation potential of Pd(0) is +0.915 V versus SHE while the analogous oxidation of Au(I) is +1.410 V versus SHE, nearly 500 mV higher.4

Consequently, while the oxidative addition of aryl iodides to Pd(0) is a well-known

2 reaction, analogous reactivity with gold complexes has only recently been reported.5–7 The guiding design principle has been preorganization of the gold(I) center to accommodate the square-planar geometry of the ensuing oxidative addition product. Additionally, recent reports by Corma and coworkers have shown that a similar effect can be reached using gold nanoparticles where adsorption of aryl iodide to the gold surface serves to lower the barrier to oxidative addition by 20 kcal/mol.8

Pronounced relativistic effects provide a source for several interesting properties of gold. Relativistic contraction of the 6s orbitals and concomitant expansion of the 5d orbitals are partially responsible for the preference of Au(I) toward low coordination number, linear geometries, and high electronegativity as well as the presence aurophilic interactions. The coordination chemistry of Au(I) is significantly impacted by a preference toward linearity.

Consequently, reaction of Au(I) sources with bidentate ligands often afford bridging dinuclear structures9–11 rather than chelate complexes.5,12,13 Low coordinate complexes of

Au(I) can also associate to form multinuclear or polymeric structures supported by direct gold-gold contacts. Steric effects permitting, the metal-metal distances are shorter than the sum of the van der Waals radii (3.7 Å) and range from 2.7 - 3.3 Å. The strength of these aurophilic interactions are on the order of a hydrogen bond (7 – 12 kcal/mol).14 The basis for this interaction is electron correlation between the formally closed shell components akin to a very strong van der Waals interaction. Relativistic stabilization of the 6s orbitals and destabilization of the 5d orbitals permits partial occupation of the 6s orbitals affording a pseudo open shell d orbital manifold.15

Identifying the oxidation state responsible for catalysis in gold-mediated processes has become an emergent challenge as new reactions are uncovered and investigated in more detail. This problem is compounded by reports of divergent reactivity between different oxidation states of gold.16–21 The soft π–acidic gold(I) and hard oxophilic gold(III) furnish different isomers in various intramolecular cyclizations or skeletal rearrangements based on preferential binding to alkynes or pendant Lewis basic functionality.

The aurophilicity of gold frequently encourages the formation of multinuclear complexes. Binuclear gold(I,I) complexes supported by bisphosphines have been shown to exhibit enhanced catalytic activity compared to their mononuclear counterparts in part owing to the thermodynamic stabilization imparted by formation of a Au–Au bond. Toste and coworkers have demonstrated that access to gold(II,II) intermediates in a heteroarylation of alkenes lowers the barrier to reductive elimination by roughly 15 kcal/mol as compared to the analogous reaction with a mononuclear complex.22 A similar strategy was employed for the allylation of aryl boronic acids where a gold(II,II) complex was implicated as a key catalytic intermediate. Gold complexes bearing bisphosphines are also competent photocatalysts. [Au2(µ-dppm)2]Cl2 (dppm = 1,1- bis(diphenylphosphino)methane) can mediate amine-alkyne coupling,23 perfluoroalkylation of hydrazones,24 atom transfer radical polymerization,25 semipinacol rearrangement,26 and intramolecular cyclization to furnish functionalize indoles27 or pyrroles.28

Gold(I,I) complexes bearing dinucleating anionic ligands including ylides, amidates, and thiolates demonstrate impressive oxidative chemistry.29,30 In contrast to their

4 mononuclear equivalents, these binuclear gold complexes are nucleophilic at the metal

– center. In particular, phosphorus ylide ligands of the form [R2P(CH2)2] (R = aryl or alkyl) endow the resultant binuclear gold complexes with enhanced stability owing to the highly covalent Au-C bonds. Moreover, charge delocalization helps mitigate decomposition pathways via the Kochi-type electron transfer mechanism.31 Phosphorus ylides confer stability across a wide range of oxidation states. Binuclear gold complexes in the gold(I,I), gold(II,II), gold(I,III), and gold(III,III) oxidation states have been structurally characterized.29,30 These complexes activate a number of small molecules including alkyl halides, halogens, nitromethane and undergo an impressive 2-center 4-electron oxidation with dihalomethanes.9,31–36 The narrow bite angle of phosphorus ylide ligands holds the gold centers in close proximity, permitting facile activation of small molecules through formation of Au–Au bonds. Nucleophilic attack of electrophiles via an SN2-type pathway is the dominant mechanism for oxidation.33,37

While the oxidative chemistry of binuclear gold(I,I) complexes is well-established, reactions involving reductive coupling of alkyl or aryl components across the bimetallic core is almost entirely unknown. Schmidbaur and coworkers detail the only report which shows formal C–C bond formation and this reaction requires harsh thermolytic conditions.36 Chapters 2 and 3 are focused on exploring the reductive chemistry and Lewis acid catalytic activity of dinuclear gold complexes.

1.3 Rylene imides as redox non-innocent ligands

Over the last fifty years redox non-innocent ligands have become a staple of the organometallic toolbox. Jørgensen first coined the term “non-innocent” to describe species

5 that contain metal centers with ambiguous oxidation states.38 Any ligand that allows mixing of the electron density in the frontier molecular orbitals can be considered non-innocent.39

Synthetic and biological catalytic systems often employ redox non-innocent ligands to facilitate multielectron transfer and transport through ligand–metal cooperativity.

Figure 1. Select examples of redox-non innocent ligands in catalysis.

Two common redox non-innocent ligand scaffolds for catalytic applications are pyridyl diimine (PDI) and aminoquinone based platforms (Figure 1). Chirik and coworkers have reported formally Fe(0) complexes bearing PDI ligands that are competent for a wide range of transformations including hydrosilylation and [2 + 2] cyclizations.40–42 Spectroscopic data is consistent with an Fe(II) center supported by a two electron reduced ligand.

Mechanistic data suggests that preservation of the metal center in the ferrous state may extend catalyst lifetime by avoiding sensitive Fe(0) species that are prone to demetellation.

Additionally, Bart and coworkers detail a similar strategy with U(IV) – PDI complexes.43,44

Two electron oxidative addition chemistry is permitted at the metal center since the reducing equivalents are derived from the PDI ligand. The Heyduk lab has demonstrated that an aminoquinone based ligand (N2O2) supports oxidative chemistry at Zr. Oxidation of a formally Zr(IV)–N2O2 complex with PhICl2 results in oxidation of the ligand and addition of chloride ligands to the Zr center. Ligand based redox chemistry facilitates two electron chemistry at Zr including the catalytic disproportionation of hydrazine to azobenzene.45

Rylene diimides (RDIs) are a class of electron deficient aromatic compounds that exhibit rich photo and electrochemistry. They have been heavily explored for use in light harvesting, energy storage devices, and supramolecular assembilies.46–49 The proclivity for strong π–π stacking in solution often limits the solubility of these complexes however installation of bulky substitutents can suppress aggregation. Additionally, the redox potentials of the NDI and PMDI derivatives match well with the reduction potentials of common small molecules such as H2, CO2, and O2 as well as catalytically active base metals like Co, Fe, or Ni (Figure 2). The electrochemistry is typically well behaved, exhibiting reversible one electron redox events by cyclic voltammetry. RDIs should be well poised for use as redox-non innocent ligands on transition metals but to the best of our knowledge no such reports exist.

Figure 2. Rylene imides and their redox potentials.

A general problem when utilizing redox-non innocent ligands in combination with redox active metal centers is identification of where redox events occur. While a suite of physical characterization techniques have been developed to help answer these questions, redox ambiguity in these systems remains a persistent challenge. The oxidation states of

RDIs have distinct spectroscopic UV-vis signatures that can aid in identifying whether the redox state of the metal or ligand has changed. Moreover, construction of RDIs is extremely modular, allowing derivatization at the arene core or nitrogen substituents. The optical properties can be finely tuned over the visible and near-IR spectrum and the redox potentials can be adjusted over nearly a 0.5 V range.50 RDIs are also well known to undergo photoreduction to stable radical anion species (RDI·−) in the presence of sacrificial electron donors. This behaviour has been exploited for the study of photoinduced charge separation51,52 and photooxidation of DNA.53–58 Visible light-absorbing perylene diimides

59–62 (PDIs) have also been investigated as photosensitizers for catalytic water splitting, CO2 reduction,63 and reduction of aryl halides.64,65

Despite thorough investigation into the electrochemical, photochemical, and energy storage properties of RDIs, coordination chemistry with metal cations remains underexplored.66 RDI-transition metal conjugates are typically constructed for the purpose of exploring the excited state behavior of the complexes. RDIs can substantially extend the lifetime of charge separated excited states owing in part to the stability of the RDI·−.

However, in these complexes direct interaction between the RDI moiety and the metal center is often geometrically impossible. Studies providing evidence for the direct coordination of RDIs to metal cations are rare. Dichtel and coworkers report the redox potentials of NDIs immobilized in a porous thin film can be attenutated over a 700 mV range in the presence of charge dense cations such as Li+ and Mg2+. The same effects were not observed for larger cations such as K+ or tetrabutylammonium. Cation dependent stabilization of the electrogenerated NDI2– species indicates interaction between cations and the reduced NDI. However, no fine structural detail to describe these interactions has even been reported. Recently, Würthner and coworkers reported the structure of the sodium salt of a perylene diimide (PDI) dianion with cyano and chloro substituents in the perylene core.67 The structure features a two dimensional network assembled via coordination of the oxygen atoms and nitrile substituents of the PDI dianion to the Na+ ions.

Chapters 4 and 5 describe the direct interactions between reduced NDI species and both redox inactive and redox active metal cations. Applications toward energy storage and electrocatalysis are discussed.

1.4 Immobilization of diaryl pincer complexes in metal-organic frameworks

Metal–organic frameworks (MOFs) provide attractive scaffolds for heterogeneous catalysis owing to their well-defined structures, inherent porosity, and impressive modularity.68–71 The exquisite tunability of MOFs is born out of the near infinite combination of organic linkers and secondary building units (SBUs) that comprise these materials. Organic linkers typically consist of arenes decorated with carboxyl, pyrazole, or triazole subunits while SBUs can consist of multinuclear metal clusters or chains. The formation of a coordinative bond between the Lewis acidic metals at the SBUs and Lewis basic functional groups on the organic linker drive formation of a coordination polymer.

This chemoselectivity in this self-assembly process is a guiding design principle that has allowed incorporation of framework and non-framework metal ions for catalytic applications.

MOFs are typically assembled under solvothermal conditions using aprotic polar solvents such as dimethylformamide. Brønsted acid modulators are used to promote the formation of crystalline phases. Tuning the pKa of the reaction solution allows organic linkers to bind reversibly to SBUs which permits the formation of crystalline phases under pseudo-thermodynamic control. While the “black box” self-assembly reactions inherent to standard MOF synthesis can preclude any rational critique of existing synthetic methods, some advances have been made in understanding how assembly conditions affect (and effect) defect formation in MOFs.72–77 Defects are generally categorized according to the absent structural unit, namely missing linkers or missing nodes. These defects can be sites for catalysis and often do not degrade the bulk crystallinity of the material. Zr MOFs

10 frequently exhibit missing linker defects where the conjugate base of the acid modulator used during synthesis replaces an organic linker at the SBUs. Bulk compositional analysis is helpful in determining the quantity of defect sites within a material, however understanding the distribution or apportionment of defect sites remains a challenging prospect.

Postsynthetic modification (PSM), originally coined in reference to posttranslational modification of proteins, has become a powerful means of functionalized MOFs post- assembly.78 PSM encompasses metal or linker exchange reactions in the MOF as well as functional group transformations at organic linkers. Metal exchange at the SBU can be used to install exogeneous metal ions that are active for catalysis. Exchange of organic linkers can be used to concentrate or dilute catalytically active species in a material where a high density of sites can be detrimental toward the efficient mass transport of substrates and products.

Incorporation of catalytically active metal species into MOFs is often achieved using one of three major strategies: exchange or tethering at the SBU, binding at the organic linker, or by encapsulation within the pores of the material.79 Incorporating catalytically active species at the SBU can jeopardize the structural integrity of the MOF, however recent work has demonstrated that functionalization of SBUs under appropriate synthetic conditions can preserve bulk crystallinity.80–83 Encapsulation of the catalyst within MOF pores can reduce substrate access and lead to leaching of catalytically relevant species.

However, Byers and coworkers have recently demonstrated a unique approach organometallic catalysts that prevents leaching over multiple recycling experiments.84

Appending metal-based catalysts at the organic linkers is a favorable strategy because leaching and structural integrity issues are eschewed while maintaining efficient substrate ingress and egress from catalytically active sites.

Homogenous diphosphine pincer complexes have been shown to catalyze a myriad of reactions.85,86,95–104,87,105,88–94 Diphosphine pincer ligands (Figure 3) are most often categorized according to the identity of the central donor (E) or linker side arms (Z). E is commonly C or N and Z is commonly CH2, OH, or NH2. Moving from an X-type to L-type central donor as well as attenuating the electron density at the phosphorus arms via the identity of Z can have profound effects on the reactivity of the resulting complex.

The versatility of diphosphine pincer complexes in catalytic applications makes them ideal candidates for heterogenization. Nonetheless, heterogenization of pincer complexes is still relatively rare. Ir(III) POCOP complexes have been grafted onto g–alumina106 or mounted at the SBUs of NU-1000107 via a p-phenol group on the pincer central donor.

Additionally, Ir(III) POCOP dihydride complexes can be immobilized on silica via protonation of the metal hydride.108 Chiral Zn(II) pybox and Ru(II) terpy complexes can be grafted onto partially dehydroxylated silica via a Zn-OH linkage formed after hydrolysis of the corresponding metal alkyl.109,110 However, grafting techniques that involve hydrolysis or protonation of acidic metal fragments risk suppressing the catalytic activity of the complex as one coordination site must always be occupied by Lewis basic functionality from the heterogeneous support.

Immobilization of diphosphine pincer complexes within a MOF can allow the study of site isolation and pore shape/size effects on catalysis in a well-defined structure. However, 12 the synthesis of diphosphine pincer complexes that can be incorporated into a MOF is challenging because the ligands must be endowed with multiple donor functionalities.

Lewis basic moieties such as carboxylic acids are necessary for chemoselective binding to framework metal ions while phosphine donors must bind chemoselectively to the catalytically active metal center. Our group and the Humphrey group have reported the successful installation of PCP,111,112 POCOP,113 and PNNNP114 Pd(II) complexes into Zr or

Co MOFs via binding of pendant carboxylic acids to the SBU.

Figure 3. General structure of PZEZP pincer complexes

Characterization of MOFs requires a suite of physical characterization methods.

Common techniques include elemental and gas sorption analysis, single crystal and powder

X-ray diffraction, and solid state NMR, EPR, UV-Vis, FTIR, and X-ray fluorescence (XRF) spectroscopy. Additionally, under highly acidic or basic conditions MOFs may disassemble and the resulting solutions can be analyzed by the usual array of solution-state characterization techniques. The crystallinity of MOFs differentiates them from many other heterogeneous platforms and allows the use of X-ray diffraction for precise structural characterization. Langmuir-type adsorption curves extracted from gas sorption analysis are valuable in determining information about porosity and host-guest binding characteristics. 13

Information about overall composition can be garnered from elemental analysis and XRF and solid state NMR spectroscopy. Chapter 6 describes work in which postsynthetic modification activates a MOF immobilized Pd2+ PNNNP complex toward Lewis acid catalysis. Implications of pore selective transformations as well as complicated reaction kinetics are also discussed.

Chapter 2: Reactivity of Alkylated and Arylated Dinuclear Gold Ylide Dimers

2.1 Introduction

As discussed in Chapter 1, one of the major barriers to gold-mediated cross coupling chemistry is the high oxidation potential required to access gold(III) which is ~ 0.5 V positive of the Pd(0)/Pd(II) couple (1.41 V vs 0.91 V versus SHE).4 This reluctance to undergo traditional two electron chemistry is often overcome using strong oxidants such as hypervalent iodines or XeF2. While oxidative addition of aryl iodides to Pd(0) is a well- established reaction,115 the first example with gold was only reported four years ago.5 In their seminal report, Bourissou and coworkers employed a carborane diphosphine ligand with a small bite angle to pre-organize a gold(I) center toward oxidative addition. More recent reports have shown that fluorinated bipyridine7 as well as hemilabile P,N bidentate ligands6 can effect similar reactivity. Additionally, intramolecular oxidative addition reactions can be accomplished by phosphine directing groups.12

An alternative strategy to access high valent gold species is to utilize multinuclear complexes where the redox load can be distributed over multiple metal sites. A study by the Toste lab found two electron oxidation of a binuclear gold complex bearing a bridging diphosphine ligand was shifted 140 mV more positive than a mononuclear counterpart owing to the formation of a Au–Au bond.22 This behavior is consistent with work conducted by Fackler, Schmidbaur, and others showing that binuclear gold “A-frame” complexes supported by phosphorus ylide (PY) ligands (4) are competent for the oxidative addition of alkyl and allyl halides, reactivity that has not be observed for mononuclear equivalents.10,30,116–119 Formation of Au–Au bonds is often proposed as the thermodynamic 15 driving force for this oxidative chemistry and the strongly donating nature of the PY ligands stabilizes high valent complexes.9,120 Binuclear gold complexes bearing PY ligand have been characterized in a wide array of oxidation states including gold(I,I), gold(II,II), gold(I,III), and gold(III,III).29,30

Schmidbaur and coworkers discovered a particularly interesting reaction upon thermolysis of an alkylated gold(III,III) complex. Oxidation of 4a was accomplished using dibromomethane to cleanly afford gold(III,III) complex 5a via a 2–center, 4–electron oxidative addition (Figure 4). Subsequent alkylation with MeLi furnished 6a, which upon thermolysis in the solid state, evolved propane and regenerated 4a.36 The liberation of propane is the result of a formal methylene insertion reaction. General methods for methylene insertion that do not require operationally challenging reagents such as diazomethane remains an outstanding challenge in synthetic chemistry. Unfortunately, the thermolysis of 6a was never studied in any greater detail in order to explore whether the formation of propane proceeded via a radical process or an alternative reductive elimination mechanism.

Figure 4. Synthesis of 6a. Conditions: (i) CH2Br2, neat, 12 h; (ii) MeLi, C6H6, 2 h

Motivated by broad mechanistic questions and the prospect of developing a system for catalytic methylene insertion reactions, we set out to revisit the seminal chemistry reported by Schmidbaur and coworkers. Moreover, binuclear gold complexes bearing PY ligands have shown remarkable stability which should permit observation of complexes that mimic key organometallic intermediates. The behavior of alkylated or arylated binuclear gold complexes bear relevance to gold mediated cross coupling reactions but have never been rigorously explored. We considered that replacing the phenyl substituents at the phosphorus center of the PY ligands with electron withdrawing or sterically bulky groups might encourage productive reductive elimination via electronic or steric pressure. This chapter describes the synthesis of a small family of novel binuclear gold PY complexes, their thermolysis to induce reductive coupling, and the liberation of bis(perfluorophenyl)methane via sequential C–C bond forming events.

2.2 Synthesis and physical characterization of alkylated dinuclear gold(III,III) phosphorus ylide complexes

Synthesis of the PY ligands was carried out by sequential alkylation of the appropriate chlorophosphine (1a–c) with methyl lithium and iodomethane to afford phosphonium salts

(2a–c). Double deprotonation of the phenyl and tolyl derivatives, 2a and 2b respectively, furnished phosphorus ylides, 3a and 3b, as white microcrystalline solids (Figure 5). The lithium salts were quenched with one equivalent of AuCl(PMe3) to deliver binuclear gold complexes 4a and 4b in good yield (~ 90%). Rigorous exclusion of light was necessary to avoid adventitious photoreduction to gold(0).

The 4-trifluoromethylphenyl derivative, 2c, was less well-behaved. Attempts to treat

2c with methyl lithium afforded an intractable mixture. Similar decomposition was 17 observed when nBuLi and tBuLi were employed as bases for deprotonation of the phosphonium salts. Anionic oligomerization of the PY ligand or nucleophilic displacement of the 4-trifluoromethyl moiety are believed to be responsible for the observed decomposition. The use of a large non-nucleophilic base, sodium bis(trimethylsilyl)amide, in combination with a chelating tertiary amine, tetramethylenediamine, and dilute reaction conditions were necessary to mitigate these effects. The intermediate PY ligand was generated in situ and treated directly with PMe3AuCl to afford complex 2c in 78% yield.

Figure 5. Synthesis of binuclear gold A-frames. Conditions: (i) MeLi, -20 °C, 2 h; (ii) MeI, Et2O, 12 h; (iii) MeLi, Et2O, 6 h; (iv) AuPMe3Cl, THF/toluene, 12 h, dark; (v) NaHMDS, TMEDA, -78 °C, THF, 2 h; (vi) PMe3AuCl, THF, 12 h, dark

Complexes 4a-c are stable to air, moisture, and light under ambient conditions. The

31P{1H} NMR spectra featured sharp singlets in the 34.1 – 35.3 ppm region, consistent with literature precedent. The reported solid-state crystal structure of 4a shows that the 8- membered metallacycle adopts a chair conformation.31 However, fast ring flipping of the metallacycle on the NMR time scale gives rise to a doublet in the 1H NMR spectra, 18 attributed to the diastereotopic methylene protons a to the phosphorus atoms. Indeed, the

1H NMR spectra of 4a – c contain diagnostic doublets resonating at 1.34 – 1.41 ppm as well as the expected signals from the aromatic substituents.

2.3 Thermolysis of 6a.

Treatment of complex 4a with neat dibromomethane afforded gold(III,III) complex 5a via formal 2–center, 4–electron oxidative addition. Subsequent alkylation with 2 equivalents of methyl lithium cleanly afforded 6a (Figure 4).

The 31P{1H} NMR spectrum of 6a features one sharp singlet at 38.9 ppm consistent with reports in the literature. The a methylene protons appear as two doublets of doublets at 1.95 and 1.75 ppm in the 1H NMR spectrum confirming that the metallacycle adopts a boat conformation in accordance with the solid state structure.36 Additionally, the methyl groups produce a diagnostic singlet resonance at -0.34 ppm. The rest of the spectrum exhibits signals with the expected chemical shifts and coupling constants.

The thermolysis of 6a in o-dichlorobenzene at 150 °C was monitored by 31P{1H} NMR

(Figure 6) over the course of 20 minutes. Disappearance of 6a was accompanied by growth of 4a as a consequence of formal loss of a propane equivalent. Trace amounts of an intermediate or side product were detected at 32.0 ppm. This resonance is attributed to 7,33 which can be formed by ejection of the bridging methylene and subsequent valence isomerization. Ethylene is the postulated fate of the bridging methylene. Propane was detected by GC-MS but mobile phase contaminants including CO and N2 precluded analysis of smaller hydrocarbons, such as ethane, evolved from complex 7. Quantification of propane (m/z of 44) was convoluted by overlap with CO2 (m/z of 44). 19

Figure 6. Reaction profile for the thermolysis of 6a in o-dichlorobenzene at 150 °C.

The high temperature required for productive thermolysis prompted us to study how 6a might undergo reductive elimination. The X-ray crystal structure of 6a shows square planar gold(III) centers in which the bridging methylene and methyl groups are positioned trans to one another.36 Considering that concerted reductive elimination requires two substituents in a cis arrangement, 6a must isomerize to proceed via this exit pathway.

Figure 7. a) Reported isomerization pathways for binuclear gold PY complexes. b) Proposed isomerization – reductive elimination sequence for 6a.

Fackler and coworkers have reported that gold(I,I) and gold(II,II) PY complexes can undergo ligand exchange and isomerization. While the exact mechanism is not well understood, k2 – h3 ligand isomerization or formation of monomer species via rupture of the binuclear metallacycle have been implicated (Figure 7a).121,122 Moreover, Brønsted and

Lewis acids are known to increase the rate of ligand exchange while Brønsted bases decrease this rate.121 Although the greater Lewis acidity of the Au(III) centers in 6a may reduce the propensity for ligand dissociation, we envisioned that a similar mechanism could be operative for isomerization prior to reductive elimination. The gold centers in 6a would likely adopt a T-shaped geometry upon formation of an open coordination site via k2

– h3 ligand isomerization (Figure 7b). T-shaped trialkyl gold(III) complexes are well- known to undergo T–T isomerizations and isomerization between T-shaped configurations occurs 100 times faster than reductive elimination.123 Sequential reductive elimination 21 from these T-shaped intermediates would allow 6a to produce propane via productive two electron chemistry without invoking radical intermediates. The persistence of these T- shaped gold species can be probed by formation of Lewis adducts. Lewis bases are expected to bind at the open coordination site while Lewis acids should prolong the lifetime of the open site via coordination to an available ylide donor group (Figure 7b).

31 1 Figure 8. P{ H} spectra of the thermolysis of 6a in the presence of PPh3 or CuI.

Thus, we studied the thermolysis of 6a in the presence of various Lewis base and Lewis acid additives (Figure 8). 6a undergoes quantitative conversion to 4a over 20 minutes at

150 °C in the absence of any additive. However, addition of 1 equivalent of PPh3 severely retards this reaction and no conversion of 6a was observed under otherwise identical

22 conditions. Additionally, the 31P{1H} NMR spectrum of the reaction shows the signal attributed to PPh3 exhibits a peak width at half max of 35 Hz and the resonance associated with 6a broadens from 4 Hz to 13 Hz, suggesting that bound and free PPh3 are in dynamic equilibrium on the NMR timescale. Thus, we hypothesize that PPh3 sequesters incipient open coordination sites at gold suppressing the necessary isomerization for reductive elimination. Attempts to increase the persistence of open coordination sites at gold by treatment with Lewis acids were met with limited success. AlCl3 and B(C6F5)3 led to no reaction while THTAuCl (THT = tetrahydrothiophene) and IPrAuCl (IPr = 1,3-bis(2,6- diisopropylphenyl)imidazol-2-ylidene) resulted in decomposition of 6a as indicated by severe loss of signal intensity in the 31P{1H} spectra as well as formation of gold(0).

Thermolysis of 6a in the presence of CuI proceeded to complete conversion at 80 °C, nearly half the reaction temperature required in the absence of any additive. However, CuI may play a more complex role than that of a Lewis acid, including acting as a radical mediator or alkyl/methylene transfer agent. In order to interrogate these possibilities, careful analysis of the fate of the organic components of the reaction is necessary. However, quantitative and qualitative analysis is hampered by the volatility of the organic products liberated from 6a. Consequently, we sought to substitute the methyl groups of 6a with larger alkyl or aryl groups in order to facilitate identification of the organic components of the reaction.

Installation of larger organic groups at gold was challenging owing to the propensity for reduction of the gold centers of 5a. The use of organolithium or Grignard reagents often led to the formation of gold(II,II) or gold(I,III) complexes. Attempts to transmetallate from

23 copper, boron, silicon, or tin reagents were met with similar complications. However, successful installation of phenyl and pentafluorophenyl groups was accomplished via addition of the corresponding organolithium to 5a at low temperature to afford 8 and 9, respectively (Figure 9).

Figure 9. Synthesis of 8 and 9.

2.4 Mechanistic investigation into the thermolysis of 8 and 9.

Figure 10. Thermolysis reactions of 8 and 9.

The thermolysis of 8 and 9 was conducted in the presence and absence of CuI. Both compounds were surprisingly resistant to thermolysis in the absence of CuI, showing little conversion over 12 hours at 150 – 200 °C. Trace amounts of 1-(2- chlorobenzyl)pentafluorobenzene, likely formed via capture of transient pentafluorobenzyl radicals by o-dichlorobenzene, were detected by GC-MS following thermolysis of 9

(Figure 10a). Attempts to intercept radical intermediates with TEMPO or 9,10- dihydroanthracene were unsuccessful and reactions conducted in the presence of radical initiators such as AIBN did not lead to higher conversion of 8 or 9.

Thermolysis of 8 at 60 °C in CDCl3 in the presence of CuI affords an intermediate species 10 en route to complex 11 (Figure 10b). While 10 could not be isolated, after 3 hours a sufficient concentration of the compound was generated, allowing for characterization by 1H NMR spectroscopy (Figure 11). Intermediate 10 may be a CuI adduct of 8 since it only forms in the presence of CuI. Another plausible structure of 10 consistent with the 1H NMR data is shown in Figure 11 and Figure 12. Advanced NMR techniques or characterization by single crystal X-ray diffraction will be necessary to fully elucidate the structure of intermediate 10, but the 1H NMR features described below indicate a complex with limited symmetry at the binuclear metallacycle.

Figure 11. Reaction profile for the thermolysis of 8 monitored by 31P{1H} NMR spectroscopy.

Analysis of the mixture of 8, 10, and 11 after 3 hours of thermolysis was deconvoluted by using broadband 31P decoupling on the 1H NMR spectrum (Figure 12). The four diastereotopic proton resonances at 2.55, 2.17, 1.88, and 1.70 ppm attributed to 10 indicate a complex with limited symmetry at the binuclear metallacycle. The observed P–H

2 coupling constants are consistent with other binuclear gold PY complexes ( JPH = 12.5 –

13.1 Hz). Additionally, 10 exhibits two resonances at 2.15 and ~1.65 ppm (the upfield resonance is occluded by nearby signals) that do not collapse upon decoupling and are ascribed to benzylic protons.

1 31 Figure 12. H{ P} NMR spectrum (CDCl3) of the thermolysis of 8 after 3 hours.

The observation of two benzylic signals suggests dynamic motion along that moiety is limited on the NMR timescale and is indicative of insertion of an aryl fragment into the

Au–CH2 bond. This is affirmed by the observation of a well-resolved set of aryl resonances 28 at 7.06 ppm and 6.85 ppm which are attributed to the newly formed benzylic phenyl ring.

An additional diagnostic signal at 1.26 ppm is attributed to a spectroscopically distinct PY–

CH2.

A relatively pure 1H NMR spectrum of 11 is produced after thermolyzing 8 for 12 hours

(Figure 13). The 1H NMR spectrum of 11 reveals only 5 signals (vs. 7 exhibited by 10) associated with the metallacycle and no diastereotopic protons are observed. This apparent increase in symmetry is likely a consequence of fast ring flipping of a chair conformation on the NMR timescale. Again the spectrum exhibits two benzylic protons (2.66 and 1.59 ppm) and a well resolved set of aryl resonances (6.69 and 6.61 ppm) consistent with preservation of the benzylic moiety. In analogy to 10, the spectrum of complex 11 also exhibits a spectroscopically distinct PY–CH2 signal at 1.25 ppm. Integration against residual pentane shows no additional protons have been captured by reaction with adventitious acid or water. Additionally, heating 11 for extended reaction times or higher temperatures only resulted in decomposition of the complex. Trace amounts of biphenyl were detected by GC-MS.

1 Figure 13. H NMR spectrum (CDCl3) of 11 generated in situ via thermolysis of 8.

The thermolysis of 8 in the presence of CuI was repeated in DMSO in order to explore how solvent might affect reductive elimination. The reaction was monitored by 31P{1H}

1 and H NMR spectroscopy. While thermolysis in CDCl3 furnished complex 11 via C–C coupling, thermolysis in DMSO furnished complex 12 as the major product via formal ejection of the bridging methylene unit (Figure 10c). Increasing the reaction temperature to 150 °C led to elimination of stoichiometric amounts of biphenyl along with regeneration

+ of complex 4a. Significant decomposition resulting in formation of [Ph2Me2P] salts was also observed.

Complex 12 was isolated by precipitation from the reaction mixture (Figure 14). The

1H NMR spectrum features two doublets at 1.79 and 1.35 ppm. Additionally, two distinct sets of phenyl resonances are observed at 7.16 and 7.14 ppm and 6.58 and 6.50 ppm. The symmetry about the binuclear core is consistent with a pseudo-chair conformation. The

31P{1H} NMR spectrum exhibits one sharp peak at 32.5 ppm.

The reactivity of 8 under thermolytic conditions appears solvent dependent. Reactions conducted in CDCl3 lead to insertion of a phenyl group into a Au–CH2 bond. Conversely, reactions performed in DMSO afford products consistent with formal methylene extrusion.

This disparate reactivity is likely a consequence of the ability of the solvent to stabilize incipient coordination sites at gold as well as affect the activity and speciation of putative copper aryl/methylene complexes. Concerted reductive elimination via sequential two electron chemistry is an attractive thought experiment, but the operative mechanism for these reactions is likely much more complex and the intermediacy of radical-borne species is certainly plausible. More rigorous mechanistic studies will be required to help elucidate and predict how binuclear gold PY complexes engage with a variety of possible exit pathways.

1 Figure 14. H NMR spectrum (CDCl3) of 12 isolated from the thermolysis of 8 in DMSO solution.

13 14

While thermolysis of 8 only forges one new C–C bond by stoichiometric reaction, applications toward catalysis or mechanistic investigations remain a promising possibility.

Gold mediated cross coupling reactions are still rare despite several recent advances in gold redox chemistry. The dearth of sp2–sp3 gold mediated cross coupling reactions is caused by the difficulty in activating sp2 C–X bonds in the absence of strong oxidants or harsh reaction conditions. The binuclear gold framework eschews this issue by coupling activated sp2 and sp3 bonds in a Kumada-Corriu type reaction under mild conditions. Few 32 scaffolds allow observation of distinct intermediates such as complex 11 which is critical in elucidating the mechanistic underpinnings of new transformations.

The thermolysis of 9 in the presence of CuI was also conducted in order to evaluate how the electron deficient pentafluorophenyl groups might affect reductive elimination

(Figure 10d). The thermolysis was monitored by 19F NMR spectroscopy and was carried out in DMSO at 60 °C with one equivalent of CuI (Figure 15). The 19F NMR spectrum of

9 features three signals, a pentet at 162.4 ppm, a triplet at 160.9 ppm, and a broad singlet at 118.8 ppm in a 2:1:2 ratio. The singlet, attributed to the ortho fluorines, is likely broad owing to restricted rotation around the Au–aryl bond vector. After four hours, the 19F NMR spectrum showed formation of a spectroscopically distinct species with three well-resolved signals at -164.2, -156.5, and -140.9 ppm in a 2:1:2 ratio. The multiplicity associated with each resonance is consistent with free rotation in solution which supports extrusion of an organic compound from 9. 31P{1H} NMR showed regeneration of complex 4a and bis(perfluorophenyl)methane was detected by GC-MS confirming the diaryl methane had been successfully liberated from the organometallic complex.

Figure 15. Reaction profile for the thermolysis of 9 monitored by 19F NMR spectroscopy.

No intermediates were detected during the reaction from 9 to 4a although a small amount of decomposition of 9 was observed by 31P{1H} NMR spectroscopy. Although no tractable intermediates were observed, we expect 8 and 9 to both follow the same reaction coordinate. Judicious choice of arene coupling partner may allow observation or isolation of relevant intermediates based on the accompanying donor/acceptor ability. Although the methylene insertion reaction to form bis(perfluorophenyl)methane has not been observed to be catalytic, the simultaneous construction of two C–C bonds is still a notable synthetic

34 feat. Despite the possibility of radical borne intermediates, diarylmethanes are still formed over other plausible byproducts including toluene or biaryl fragments.

2.5 Conclusions

A small family of binuclear gold(I,I) PY complexes was synthesized in order to revisit early work performed by Schmidbaur and coworkers involving reductive elimination to form two C-C bonds. Thermolysis of 4a revealed the reaction is accelerated by addition of CuI, however analysis of the ensuing organic products was hampered by their volatility.

Gold(III,III) complexes 8 and 9 were synthesized with larger phenyl and pentafluorophenyl coupling partners such that the liberated thermolysis products would be easier to identify.

Complex 8 exhibited interesting solvent dependent reactivity where sp2–sp3 coupling was observed in CDCl3 while formal methylene extrusion and subsequent gold(II,II) – gold(I,III) isomerization was observed in DMSO. Thermolysis of 9 afforded the desired diarylmethane, bis(perfluorophenyl)methane, which constitutes an example of formal methylene insertion chemistry.

Forging two C–C bonds in one reaction is a powerful strategy for elaboration of synthetically relevant intermediates. Preliminary mechanistic insight into the coupling chemistry of binuclear gold PY complexes was possible owing to the stability imposed by the highly donating PY ligands. Formation of diarylmethanes likely proceeds via two sequential C–C bond forming events through intermediates similar to 11. Future work will focus on the precise role CuI plays in the reductive chemistry at gold and how various exit pathways, e.g. methylene extrusion vs C–C coupling, can be leveraged for productive bond formation. CuI likely acts as an aryl/methylene transfer agent but more rigorous studies are

35 required to confirm its role in the reaction. Moreover, cross-over experiments will be valuable in identifying whether C–C bond formation occurs via intra or intermolecular reaction and thus whether extension of the chemistry to cross coupling reactions will be possible. Finally, evaluation of the activity of analogs 4b and 4c may reveal how subtle electronic changes at the phosphorus substituents can help engender more versatile coupling chemistry.

2.6 Experimental

General Considerations. All manipulations were carried out using a nitrogen-filled glovebox or standard Schlenk techniques unless otherwise noted. All glassware was oven dried in a 150 °C oven before use. Solvents were degassed by sparging with ultra-high purity argon and dried via passage through columns of drying agents using a solvent

36 purification system from Pure Process Technologies. Au2[PPh2(CH2)2] (4a),

36 124 124 125 Au2Br2[PPh2(CH2)2](µ-CH2) (5a), PMe3AuCl, PPh3AuCl, and 9 were prepared according to literature procedures. All other chemicals were purchased from commercial vendors and used without further purification. NMR spectra were recorded at ambient temperature on a Varian Inova 400 MHz instrument. 1H chemical shifts were referenced to the residual solvent chemical shifts. 31P NMR chemical shifts were referenced to 85%

19 H3PO4, and F NMR chemical shifts were referenced to 99% F3CCO2H. GC-MS analysis was performed using an Agilent 7890A GC equipped with a HP-5 capillary column (30 m,

0.25 mm i.d., 0.25 µm film thickness) and a mass spectrometer 5975C as detector. The carrier gas was helium, at a flow rate of 1 mL/min. For MS detection an electron ionization system was used with an ionization energy of 70 eV.

Synthesis of Ar2PMe2I (Ar = 4-methylbenzene, 2b; Ar = 4-trifluoromethylbenzene,

2c). A solution of 1b or 1c (1.0 mmol) in Et2O (6.0 mL) was treated with MeLi (0.66 mL,

1.6 M in Et2O, 1 equiv.) at -20 °C. The reaction was allowed to warm to room temperature and then stirred for 1 hour. The reaction was filtered over a short plug of Celite and then concentrated in vacuo. Without further purification the resulting residue was dissolved in

Et2O (7.0 mL) and treated with MeI (160 mg, 1.1 mmol) at 25 °C. The reaction stirred for

12 hours. The resulting precipitate was collected by filtration and washed with copious amounts of Et2O (3 × 25 mL). The isolated powder was dried in vacuo to afford 2 as a white crystalline powder (1.0 mmol, 99 %).

1 3 3 2b. H NMR (400 MHz, CDCl3): δ 7.72 (dd, 4H, JPH = 13.0 Hz, JHH = 8.0 Hz, Ar-H),

4 3 2 7.46 (dd, 4H, JPH = 5.4 Hz, JHH = 8.0 Hz, Ar-H), 2.80 (d, 6H, JPH = 13.8 Hz, P-CH3),

31 1 2.46 (s, 6H, Ar-CH3). P{ H} NMR (161.8 MHz, CDCl3): δ 20.1 (s).

1 4 3 2c. H NMR (400 MHz, CDCl3): δ 7.99 (dq, 4H, JFH = 4.5 Hz, JHH = 7.5 Hz, Ar-H),

3 3 3 7.36 (t, 4H, JPH = 7.5 Hz, JHH = 7.5 Hz, Ar-H), 2.93 (d, 6H, JPH = 14.0 Hz, P-CH3).

31 1 19 P{ H} NMR (161.8 MHz, CDCl3): δ 20.9 (s). F NMR (376.4 MHz, CDCl3): δ -100.0

(s).

Synthesis of [Au2(µ-{CH2}2PAr2)2] (Ar = 4-methylbenzene) (4b). A suspension of

2b (200 mg, 0.540 mmol) in Et2O (10 mL) was treated dropwise with MeLi (0.675 mL, 1.6

M in Et2O, 2 equiv.) at 25 °C. The reaction stirred for 6 hours. The precipitate was collected by filtration and washed with Et2O (3 × 5 mL). The isolated powder was dried in vacuo for

30 minutes to afford 3b as a white crystalline powder (100 mg, 75%). Without further purification, a solution of 3b (100 mg, 0.403 mmol) in THF (2 mL) was cooled to -20 °C

37 then added dropwise to a stirring solution of PMe3AuCl (123.8 mg, 0.400 mmol) in toluene

(2 mL) at -20 °C in a 20 mL scintillation vial wrapped in black electrical tape to rigorously exclude light. The reaction stirred at 20 °C for 1 hour, was allowed to warm to room temperature, and then stirred for 24 hours. The solution turned from clear yellow to clear and colorless over the course of the reaction. The reaction was concentrated in vacuo and the resulting residue was washed with MeOH (3 × 5 mL). The reaction was then extracted in CH2Cl2, filtered over a short plug of Celite, and concentrated in vacuo to afford 4b as a

1 white crystalline powder (120 mg, 85 %). H NMR (400 MHz, CDCl3): δ 7.53 (dd, 8H,

3 3 3 JPH = 10.1 Hz, JHH = 7.7 Hz, Ar-H), 7.07 (d, 8H, JHH = 7.7 Hz, Ar-H), 2.35 (s, 12H, Ar-

2 31 1 CH3), 1.33 (d, 8H, JPH = 12.7 Hz, PY-CH2). P{ H} NMR (161.8 MHz, CDCl3): δ 34.2

(s).

Synthesis of [Au2(µ-{CH2}2PAr2)2] (Ar = 4-trifluoromethylbenzene) (4c). A vigorously stirring suspension of 2c (250 mg, 0.523 mmol) in Et2O (50 mL) was cooled to

-78 °C and then treated dropwise with a solution of NaHDMS (201.7 mg, 1.1 mmol) and

TMEDA (6 mg, 0.052 mmol) in THF (6 mL). The reaction stirred for 1 hour at -78 °C and was then treated with a solution of PPh3AuCl in THF (10 mL) and toluene (10 mL). The reaction was allowed to warm to room temperature and stirred for 12 hours. . The reaction was concentrated in vacuo and the resulting residue was washed with MeOH (3 × 5 mL).

The reaction was then extracted in CH2Cl2, filtered over a short plug of Celite, and concentrated in vacuo to afford 4c as a white crystalline powder (223 mg, 78 %). 1H NMR

3 3 3 (400 MHz, CDCl3): δ 7.77 (t, 8H, JPH = 8.8 Hz, JHH = 8.5 Hz, Ar-H), 7.58 (d, 8H, JHH =

2 31 1 8.8 Hz, Ar-H), 1.40 (d, 8H, JPH = 12.2 Hz, PY-CH2). P{ H} NMR (161.8 MHz, CDCl3):

19 δ 35.3 (s). F NMR (376.4 MHz, CDCl3): δ -63.3 (s).

Synthesis of [Au2Ph2µ-CH2(µ-{CH2}2PPh2)2] (8). A stirring suspension of 5a (100 mg, 0.101 mmol) in benzene (3.0 mL) was treated dropwise with PhLi (0.100 mL, 1.9 M in dibutyl ether) at 25 °C. The reaction stirred for 30 minutes and was then filtered over a short plug of Celite. The reaction was concentrated in vacuo and the resulting residue was triturated with pentane to afford 8 as a pale yellow crystalline solid (85 mg, 85 %). 1H

2 2 NMR (400 MHz, CDCl3): δ 7.96 – 7.04 (m, 30H, Ph), 2.59 (t, 4H, JPH = 13.0 Hz, JHH =

2 2 11.4 Hz, Py-CH2), 2.43 (s, 2H, µ-CH2), 1.68 (t, 4H, JPH = 13.0 Hz, JHH = 11.4 Hz, Py-

31 1 CH2). P{ H} NMR (161.8 MHz, CDCl3): δ 37.0 (s).

General Procedure for Thermolysis Reactions. An NMR tube was charged with 6a

(8.6 mg, 0.1 mmol), a Lewis acid additive (1 equiv., 0.1 mmol), and diluted with o- dichlorobenzene, CHCl3, or DMSO (0.6 mL). The reaction was then submerged in a preheated sand bath at 150 °C and monitored by 31P NMR.

Chapter 3: Lewis Acid Catalysis with Cationic Dinuclear Gold(II,II) and Gold(III,III) Phosphorus Ylide Complexes2

3.1 Introduction

Over the past decade, the use of homogeneous gold catalysts for C–C and C–X bond- forming reactions has increased remarkably.2,126–129 Interest in gold-based catalysis stems from its tolerance to oxygen and water, low toxicity, and low proclivity for β-hydride elimination. Moreover, much of the catalytic efficacy of gold is grounded in its soft Lewis acidity that allows for the activation of unsaturated organic substrates. Complexes containing gold(I) and gold(III) centers have been shown to activate alkenes and alkynes for nucleophilic addition.130–132 However, divergent reactivity between gold(I) and gold(III) has been observed in some cases and attributed to differences in hard/soft Lewis acidity.133–135

For example, Gevorgyan et al. have reported that Et3PAuCl and AuCl3 both catalyze the cyclization of haloallenyl ketones, but generate different isomers of the halofuran products.133 More recently, Toste et al. demonstrated that the gold(III) complex

[IPrAu(biphenyl)]OTf (IPr = N,N'-(2,6-diisopropylphenyl)imidazol-2-ylidene) activates

α,β-unsaturated carbonyl compounds for conjugate addition.136 However, the gold(I) analogue, IPrAuOTf, exhibited no catalytic activity for the reaction, affirming the contrast in hard/soft Lewis acidity between gold(I) and gold(III) centers.

2 Reprinted in part with permission from, “Lewis Acid Catalysis with Cationic Dinuclear Gold(II,II) and Gold(III,III) Phosphorus Ylide Complexes.” Reiner, B. R.; Bezpalko, M. W.;Foxman, B. M.; Wade, C.R. Organometallics. 2016, 35, 2830 – 2935. Copyright 2016 American Chemical Society 40

While the catalytic properties of gold(I) and gold(III) have been well-established, those of gold(II) have been relatively unexplored.22 Owing to a d9 electron configuration, gold(II) complexes are typically isolated in multinuclear frameworks containing covalent Au–Au bonds. Several dinucleating ligand platforms have been used to support gold(II) complexes.24,28,120,137,138 Among these, phosphorus ylide (PY) ligands have been shown to support rich redox chemistry. Dinuclear gold PY complexes with the metals in the (I,I),

(II,II), (I,III) and (III,III) oxidation states have been structurally characterized.9,36,116,119,139

The strong σ-donating character of the bridging PY ligands confers excellent stability across the range of oxidation states. This is notable considering that organogold(III) complexes are often unstable toward reductive elimination.140–142 Consequently, we envisaged the dinuclear gold PY framework as a platform for exploring the catalytic activity of well-defined gold(II) and gold(III) centers. While several examples of cationic gold(II,II) and gold(III,III) PY complexes have been reported, few contain combinations of weakly coordinating anions and labile ancillary ligands expected to result in active

Lewis acid catalysts.125,143 This chapter describes the synthesis, characterization, and Lewis acid catalytic activity of dicationic gold(II,II) and gold(III,III) PY complexes containing weakly coordinating OTf– counteranions and MeCN ancillary ligands.

3.2 Synthesis and structure of cationic dinuclear gold(II,II) and gold(III,III) phosphorus ylide complexes

The dinuclear gold PY complexes 4a and 5a were synthesized according to literature

35,119 procedures as described in Chapter 2. Reaction of 4a with PhICl2 afforded the gold(II,II) complex 15–Cl2 (Figure 16). The dicationic complex 15-OTf2 was isolated as a bright yellow solid after treatment of 15-Cl2 with AgOTf in acetonitrile solution. Notably, 41 the use of weakly coordinating solvents such as CH2Cl2 or THF for this reaction resulted in a mixture of decomposition products, suggesting that the dication is stabilized by

31 1 coordination of MeCN. The P{ H} NMR spectrum of 15-OTf2 in CD2Cl2 displays a single resonance at 34.5 ppm, shifted slightly upfield from that of 15-Cl2 (36.0 ppm). The

1H NMR spectrum features aromatic resonances consistent with equivalent Ph groups and a doublet at 1.95 ppm assigned to the CH2 groups of the PY ligand. The presence of a singlet resonance at 2.2 ppm is consistent with two molecules of coordinated MeCN. 15-

OTf2 is stable to air and moisture but exhibits moderate light sensitivity.

Figure 16. Synthesis of 15-OTf2 and 5-OTf2.

Treatment of 5a with AgOTf in CH2Cl2/MeCN solution generated the dicationic complex 5-OTf2 as a colorless solid. The complex is stable to air, moisture, and light in

31 1 both solution and the solid state. The P{ H} NMR spectrum of 5-OTf2 in CD3CN displays a single resonance at 35.1 ppm, shifted slightly upfield from that of 5-Br2 (37.5 ppm). The 1H NMR spectrum shows several diagnostic features that indicate the complex 42 retains a rigid, boat-like conformation in solution.35 These features include two sets of aromatic resonances consistent with inequivalent Ph groups. Two virtual triplets, centered at 1.88 ppm and 2.46 ppm, suggest a diastereotopic relationship between the methylene protons of the PY ligand. Singlet resonances at 2.68 ppm and 1.97 ppm are assigned to the

µ-CH2 ligand and two molecules of coordinated MeCN, respectively.

The solid state structures of 15-OTf2 and 5-OTf2 were determined by single crystal X- ray diffraction. X-ray quality crystals of both complexes were obtained by vapor diffusion of Et2O into MeCN solutions. 15-OTf2 crystallizes in the space group !1 with one-half molecule of [15]2+, one OTf– ion, and one MeCN molecule in the asymmetric unit. The structure of 15-OTf2 shows that the 8-membered metallacycle adopts a chair conformation with both gold(II) centers in square planar geometries (Figure 17a). The observed equivalence of the Ph groups by 1H NMR spectroscopy at room temperature suggests rapid ring flipping of the metallacycle in solution. The Au–CPY distances (2.106(2) and 2.111(2)

Å) are comparable to those found in related Au(II-II) complexes.32 Notably, the Au–Au intermetallic distance of 2.5580(2) Å is among the shortest observed for dinuclear gold complexes bearing PY ligands.144

2+ – 5-OTf2 crystallizes in the space group !1 with two molecules of [5] and four OTf ions in the asymmetric unit (Figure 17b). The complex adopts a boat conformation with the gold(III) centers in square planar coordination environments. Each gold coordination sphere contains the PY ligands in a trans arrangement, the µ-CH2 ligand, and a coordinated

MeCN molecule. The Au–CPY (2.031(17) – 2.148(16) Å) and Au–C29 (1.988(14) –

2.060(16) Å) distances are similar to those observed in the previously reported structure of 43

5a.35 The Au–Au distances (3.0673(10) and 3.065(18) Å) are within the sum of the van der Waals radii (3.32 Å), but longer than the sum of the covalent radii (2.88 Å). This may suggest the presence of weak aurophilic interactions, but these are rare for gold(III) and most prevalent with face-to-face pairing of square planar complexes.145–147

Figure 17. Solid-state structures of 15-OTf2 (a) and 5-OTf2 (b). All hydrogen atoms, solvent molecules and OTf– counterions have been omitted for clarity. Pertinent metrical parameters can be found in the text.

3.3 Catalytic Mukaiyama addition reactions

With 15-OTf2 and 5-OTf2 in hand, we sought to investigate their activity as Lewis acid catalysts. Cationic gold(I) and gold(III) complexes have demonstrated contrasting reactivity for the Lewis acid catalyzed addition of silyl enol ethers to α,β-unsaturated

136 carbonyl compounds. Thus, we decided to screen 15-OTf2 and 5-OTf2 as catalysts for the addition of silyl enol ether 16 to crotonaldehyde (17a) and cyclohexenone (17b). We also sought to compare the activity of the gold PY complexes with gold(I) Lewis acids.

+ Although the cationic gold(I) PY complex [Au(CH2PPh3)] would seem suitable for this

44 purpose, it has previously demonstrated poor catalytic activity owing to decomposition.148

Therefore the commonly employed Lewis acid catalysts Ph3PAuOTf and IPrAuOTf were chosen for comparison.126

Entry Catalyst Cat. Loading Substrate % Yield 18a % Yield 19a

1 15-OTf2 4 % 17a 68 27

2 5-OTf2 4 % 17a 70 26 c 3 Ph3PAuOTf 4 % 17a 62 31 4 IPrAuOTfc 4 % 17a 0 0

5 15-Cl2 4 % 17a 0 0 6 5a 4 % 17a 0 0 7 AgOTf 4 % 17a 0 0

8 15-OTf2 4 % 17b 0 88

9 5-OTf2 4 % 17b 0 45 b 10 Ph3PAuOTf 4 % 17b 0 20 11 IPrAuOTfb 4 % 17b 0 0

12 15-OTf2 2 % 17a 67 27 b 13 Ph3PAuOTf 4 % 17a 60 30

14 15-OTf2 0.5 % 17a 65 26 b 15 Ph3PAuOTf 0.5 % 17a 0 0 Table 1. Gold catalyzed Mukaiyama additions. Reaction conditions: 17a/b (0.1 mmol), 16 a 1 (0.1 mmol), CH2Cl2 (0.5 mL), C6D6 (0.1 mL), 25 °C, 12 h. Determined by H NMR with respect to an internal standard (trimethoxybenzene) at 12 h. bGenerated in situ from LAuCl/AgOTf.

The addition reactions were carried out at room temperature in CH2Cl2 solution with 4

1 mol % of catalyst and were monitored by H NMR spectroscopy (Table 1). 15-OTf2 provided nearly complete conversion of 17a after 1 h, while the reaction with 5-OTf2 required over 10 h to reach complete conversion (Figure 18). Analysis of the first order

-3 -1 -5 -1 rate plots revealed initial observed rate constants (kobs) of 1.1 × 10 s and 8.4 × 10 s with 15-OTf2 and 5-OTf2, respectively. Furthermore, a mixture of aldol (18a) and conjugate addition (19a) products was observed in a ca. 2.5:1 ratio for both catalysts

(entries 1 and 2).

The reaction of 16 and 17a was carried out under the same conditions using in situ generated Ph3PAuOTf and IPrAuOTf as catalysts (entries 3 and 4). While Ph3PAuOTf provided nearly complete conversion of 17a within 1.5 h, no reaction was observed with

IPrAuOTf after 12 h. The product selectivity with Ph3PAuOTf (18a:19a = 2:1) was similar to that observed for 15-OTf2. Complexes 5a and 15-Cl2 showed no catalytic activity for the reaction, confirming that halide exchange for the more weakly coordinating OTf– anion is necessary to activate the gold centers for catalysis (entries 5 and 6). In addition, AgOTf exhibited no catalytic activity under the reaction conditions (entry 7).

With cyclohexenone (17b) as a substrate, only the conjugate addition product (19b) was observed. 15-OTf2 provided 90% conversion of cyclohexenone after 12 h, while the reaction with 5-OTf2 reached only 50% conversion in 12 h (entries 8 and 9). Analysis of

-4 -1 the first order rate plots revealed initial observed rate constants (kobs) of 1.5 × 10 s and

-5 -1 1.1 × 10 s with 15-OTf2 and 5-OTf2, respectively. While Ph3PAuOTf provided 25% conversion after 12 h, no reaction was observed with IPrAuOTf after 12 h (entries 10 and

11). These results follow the same trend observed with crotonaldehyde, but the overall reaction rates are slowed by roughly a factor of 10.

Figure 18. Reaction profile for the addition of 16 to crotonaldehyde (17a) in the presence of 15-OTf2 (circles), 5-OTf2 (triangles), or Ph3PAuOTf (squares).

Since the adjacent gold centers in 15-OTf2 may both participate in substrate activation and catalysis, we evaluated its catalytic efficiency at 2 mol % (4 mol % Au) loading (entry

12). Analysis of the reaction profile revealed that 2 mol % 15-OTf2 provides a similar rate as 4 mol % Ph3PAuOTf for the addition reaction with crotonaldehyde (Figure 18). This suggests that Ph3PAuOTf and 15-OTf2 have similar catalytic efficiency on a per gold basis.

However, when the catalyst loadings were lowered to 0.5 mol %, 15-OTf2 provided complete conversion of crotonaldehyde after 12 h, while no reaction was observed with

Ph3PAuOTf (entries 14 and 15). This indicates that 15-OTf2 is less prone to deactivation

149 by trace impurities than Ph3PAuOTf. 47

3.4 Kinetic correlation to substrate exchange rate

The Gutmann-Beckett method has been used to gauge the relative Lewis acidity of gold complexes.136,150,151 We decided to employ this method in order to gain insight into the origin of the different catalytic activities of the gold complexes for Mukaiyama additions.

Solutions of 15-OTf2 and 5-OTf2 in CH2Cl2 were treated with two equivalents of Et3PO,

31 and the difference in P NMR chemical shift (∆δ) between a Et3PO internal standard (δ =

50) and gold-coordinated Et3PO species was calculated as a measure of relative Lewis acid

31 strengths. The P NMR spectrum of 5-OTf2/Et3PO shows three distinct ylide resonances that correspond to an equilibrium mixture of the parent complex and the 1:1 and 2:1 Et3PO adducts (Figure 19). A sharp resonance at 73 ppm (∆δ = 23) was assigned to coordinated

31 Et3PO in the 2:1 adduct. The room temperature P NMR spectrum of 15-OTf2/Et3PO exhibits a sharp ylide resonance at 32 ppm and a very broad signal around 53 ppm that suggests the rate of Et3PO exchange is close to the time scale of the NMR experiment

(Figure 19). The broad signal decoalesced upon cooling the sample to -45 °C, and resonances corresponding to coordinated Et3PO appeared at 75 and 76 ppm (∆δ = 25-26).

Additional ylide resonances also began to emerge around 32 ppm, but were too poorly resolved to determine the speciation at this temperature. When a solution of IPrAuOTf in

31 CH2Cl2 was treated with one equivalent of Et3PO, a sharp P NMR signal corresponding to the adduct appeared at 80 ppm (∆δ = 30). At room temperature, an equimolar mixture of Ph3PAuOTf and Et3PO gives rise to a single PPh3 resonance at 26 ppm and a broad signal around 72 ppm (Figure 20). The broad signal did not fully decoalesce upon cooling to -45 °C. The ∆δ values determined from this study suggest that 15-OTf2 and 5-OTf2

48 exhibit comparable Lewis acidities (∆δ = 25-26) and IPrAuOTf is a slightly stronger Lewis acid (∆δ = 30). No meaningful comparison can be made with Ph3PAuOTf since peak decoalescence was not observed. Interestingly, the ligand exchange equilibria and dynamics of these complexes are very different. At room temperature, Et3PO ligand exchange occurs more rapidly for 15-OTf2 and Ph3PAuOTf than for 5-OTf2 and IPrAuOTf.

While catalytic efficiency for the enone additions seems to better correlate with ligand exchange rates than apparent Lewis acidity, further studies will be necessary to fully elucidate the steric and electronic factors that influence catalytic activity.

31 1 Figure 19. P{ H} NMR spectrum (25 °C) of IPrAuOTf, 5-OTf2, PPh3AuOTf, and 15- OTf2 in the presence of 2 equivalents of OPEt3 in CH2Cl2.

31 Figure 20. P{1H} NMR spectrum of 15-OTf2 with 2 equivalents of PEt3 in CH2Cl2.

Motivated by questions concerning the modes of substrate activation by the dinuclear

31 1 catalysts, variable temperature P{ H} spectra of 5-OTf2/Et3PO were collected at regular

31 1 intervals from 228 – 298 K. Each P{ H} spectrum of 5-OTf2/Et3PO is well-resolved, featuring peaks attributed to an equilibrium mixture of the parent complex and the 1:1 and

2:1 Et3PO adducts in the region from 32 – 36 ppm as confirmed by titration experiment

(Figure 21).

31 Figure 21. P NMR spectra of 5-OTf2 in the presence of 2 -5 equivalents of Et3PO in CH2Cl2.

31 1 Figure 22. Variable temperature P{ H} spectra of 5-OTf2 in the presence of 2 equivalents of Et3PO (left) and van’t Hoff plot (right).

Measurement of spin-lattice relaxation times revealed similar parameters for each resonance (T1 ~3.5 s) allowing semi-quantitative determination of equilibrium concentration by simple integration of the inverse-gated, 1H decoupled 31P NMR spectra.

Equilibrium constants as well as the reaction enthalpy and entropy were extracted by van’t

Hoff analysis. Figure 22 (right) shows good linear correlations between ln(K) and 1/T.

From the least squares fits of these data, K1 and K2 at 298 K were calculated to be 357 ±

30 M-1 and 149 ±11 M-1, respectively. This is consistent with values estimated by titration experiments (Figure 21). DH (~ -28 kJ/mol) and DS (~ -50 J/mol) for both equilibrium constants were within error of each other. If Et3PO is considered a reasonable substrate mimic, then the difference between first and second binding constants for 5-OTf2 suggests

52 that substrate binding at one gold may temper the Lewis acidity at the adjacent gold.

Additionally, similar reaction entropies for each binding constant demonstrate the catalyst structure is not altered significantly upon substrate binding.

Figure 23. Reaction profile for the addition of 16 to crotonaldehyde (17a) in the presence of 4 mol % 5-OTf2 (red circles) fit to an exponential function (black lines).

The van’t Hoff analysis is consistent with successive rather than simultaneous substrate activation at the gold centers of 5-OTf2. This is further corroborated by examination of the reaction of crotonaldehyde (17a) and silyl enol ether 16 in the presence of 5-OTf2. The experimental data was fit satisfactorily using a single term exponential function consistent with a well-behaved first order reaction (Figure 23). If 5-OTf2 operated via simultaneous substrate activation, then a shift in reaction order should be evident at low substrate concentrations when concurrent substrate activation is unlikely. However, the reaction

53 trace does not show any obvious kinetic aberration. Overall, the assembled data suggests the gold centers of 5-OTf2 do not function as cooperative catalytic sites. However, firm conclusions cannot be constructed without more rigorous kinetic analysis. Unfortunately,

31 the poor resolution of P spectra of 15-OTf2/Et3PO mixtures does not allow integration of the individual Et3PO adducts thus precluding extension of the van’t Hoff analysis described above.

3.5 Catalytic hydroamination reactions

Next, we decided to probe the ability of 15-OTf2 and 5-OTf2 to activate C–C π bonds by screening their activity for the hydroamination of phenylacetylene (20) with aryl amines

21a – c. These reactions were carried out in the presence of 4 mol % of each catalyst and

1 monitored by H NMR spectroscopy for the formation of imines (22a – c, Table 2). 5-OTf2 and IPrAuOTf proved to be the most efficient catalysts when toluidine (21a) was used as the nucleophile (entries 4 and 10). The catalytic activity of 5-OTf2 decreased with the more sterically hindered amines 21b and 21c (entries 5 and 6) while 15-OTf2 and Ph3PAuOTf exhibited improved activity with the larger aryl amine substrates. IPrAuOTf showed decreased conversion with substrate 21b, but good activity for the larger aniline 21c. This peculiar trend was reproducible over several runs and with different batches of catalyst and substrate.

b Entry Catalyst Ar % Yield %Vbur

1 15-OTf2 21a 54 53

2 15-OTf2 21b 77 53

3 15-OTf2 21c 81 53

4 5-OTf2 21a 63 58

5 5-OTf2 21b 34 58

6 5-OTf2 21c 32 58 c 7 Ph3PAuOTf 21a 25 30 c 8 Ph3PAuOTf 21b 54 30 c 9 Ph3PAuOTf 21c 64 30 10 IPrAuOTfc 21a 69 48 11 IPrAuOTfc 21b 39 48 12 IPrAuOTfc 21c 79 48 Table 2. Gold-catalyzed hydroamination. Reaction conditions: phenylacetylene (0.1 mmol), aryl amine (0.1 mmol), catalyst (0.004 mmol), CH2Cl2 (0.5 mL), C6D6 (0.1 mL), 40 °C, 12 h. adipp = 2,6-diisopropylphenyl. bDetermined by 1H NMR with respect to an internal standard (1,3,5-trimethoxybenzene) at 12 h. cGenerated in situ from LAuCl/AgOTf.

Amine coordination to Lewis acid catalyst sites is competitive with alkyne activation in hydroamination reactions.152–157 The rigid boat conformation of the metallacycle in 5-

OTf2 results in increased steric crowding around the gold coordination sites compared to

158 15-OTf2 which is supported by the steric parameter, percent buried volume (%Vbur).

%Vbur is defined as the percent of the total volume of a sphere occupied by a ligand and is

55 calculated using crystallographic data. %Vbur for 5-OTf2 and 15-OTf2 was calculated to be

58% and 53% respectively (Figure 24). This steric crowding may suppress strong binding of the relatively unhindered amine 21a to the gold centers in 5-OTf2, resulting in higher catalytic activity. In the reactions employing sterically bulky 21b and 21c, amine coordination should be less competitive with alkyne activation for all catalysts. Indeed, this would explain the increased catalytic activity of 15-OTf2 and Ph3PAuOTf with these substrates. However, 5-OTf2 exhibits the opposite trend, suggesting that increased steric hindrance may block nucleophile access to the activated alkyne.

Figure 24. %Vbur steric maps for 15-OTf2 (left) and 5-OTf2 (right).

From the results above, 15-OTf2 and 5-OTf2 exhibit greater versatility as catalysts for

Mukaiyama additions and hydroamination of alkynes than Ph3PAuOTf and IPrAuOTf. We considered that this versatility might be leveraged for use in cascade reactions. In order to

56 test this hypothesis, we screened the activity of the gold catalysts for the cascade reaction of ortho-substituted aniline 23 and cyclohexenone (17b). The reaction involves intramolecular hydroamination to afford 2-phenylindole (24) and subsequent intermolecular conjugate addition to cyclohexenone to generate 2,3-substituted indole 25.

(Table 3).159–161

The reaction was carried out at 40 °C in a 1:1 mixture of THF and CH2Cl2 solution with

4 mol % of catalyst, and the product distributions were determined by 1H NMR spectroscopy and GC-MS. While the reaction with 15-OTf2 delivered indole 25 in good yield, the reactions with 5-OTf2 and Ph3PAuOTf as catalysts afforded low yields of the cascade product (entries 3 – 5). Only a trace amount of 25 was observed in the reaction with IPrAuOTf after 12 h, but indole 24 was generated in near-quantitative yield (entry 6).

These results are in agreement with previous experiments. While 15-OTf2 and 5-OTf2 showed comparable activity as hydroamination catalysts, 15-OTf2 is a better catalyst for enone additions. This trend is reflected in the higher yield of 25 delivered by 15-OTf2.

Moreover, while IPrAuOTf proves to be a good hydroamination catalyst, it is not competent for the conjugate addition reaction necessary to furnish indole 25. Notably, conducting the reaction in the presence of HOTf did not afford either indole 24 or 25 in good yield, suggesting that adventitious acid formed during the reaction is not responsible for the observed catalysis (entry 2).

Entry Catalyst % Yield 24a % Yield 25a 1 none 0 0 2 HOTf 4 4

b 3 15-OTf2 25 75 (68)

4 5-OTf2 78 14

c 5 Ph3PAuOTf 37 8 6 IPrAuOTfc 95 (90)b 2

Table 3. Gold-catalyzed cascade reaction. Reaction conditions: cyclohexenone (0.1 mmol), aniline (0.1 mmol), catalyst (0.004 mmol), CH2Cl2 (1.0 mL), THF (1.0 mL), 40 °C, 12 h. aDetermined by GC-MS with respect to an internal standard (trimethoxybenzene) at 12 h. bIsolated yield in parentheses. cGenerated in situ from LAuCl/AgOTf.

3.6 Conclusions

In summary, this chapter describes the synthesis, characterization, and catalytic activity of cationic dinuclear gold(II,II) and gold(III,III) complexes supported by phosphorus ylide ligands. To the best of our knowledge, this is the first reported example of Lewis acid catalysis at gold(II) centers. 15-OTf2 and 5-OTf2 catalyze both Mukaiyama addition and alkyne hydroamination reactions, offering similar or improved catalytic activity compared to the gold(I) complexes Ph3PAuOTf and IPrAuOTf. The ability of 15-OTf2 and 5-OTf2 to activate alkynes and enones suggests both soft carbophilic and hard oxophilic Lewis

58 acid character. This hybrid Lewis acidity was further demonstrated in a cascade reaction sequence involving intramolecular hydroamination followed by intermolecular enone addition to generate a 2,3-substituted indole. 15-OTf2 proved to be an good catalyst for this cascade sequence while 5-OTf2 and the commonly-used gold(I) catalysts Ph3PAuOTf and IPrAuOTf were ineffective.

3.7 Experimental

Au2Br2[PPh2(CH2)2](µ-CH2) (5–Br2) PPh3AuCl, IPrAuCl, and 1-phenyl-1- trimethylsiloxyethylene (16) were prepared according to literature procedures.

35,119,124,162,163 Toluidine and mesitylamine were purified by sublimation and vacuum distillation, respectively. All other chemicals were purchased from commercial vendors and used without further purification. NMR spectra were recorded at ambient temperature on a Varian Inova 400 MHz instrument. 1H and 13C NMR chemical shifts were referenced to the residual solvent chemical shifts. 31P NMR chemical shifts were referenced to 85%

19 H3PO4, and F NMR chemical shifts were referenced to 99% F3CCO2H. Solvent suppressed 1H NMR spectra were collected using the WET1D sequence with default parameters.164 Briefly, spectra were collected using selective pulses of 86 ms with the

SEDUCE pulse shape. Gradient pulses were 2 ms in duration, had amplitude ratios of

8:4:2:1, and were each followed by an addition 2 ms delay prior to the next RF pulse. The recycle delay between scans was 30 s, 16K points were collected, and the acquisition time was 2.5 s. GC-MS analysis was performed using an Agilent 7890A GC equipped with a

HP-5 capillary column (30 m, 0.25 mm i.d., 0.25 µm film thickness) and a mass spectrometer 5975C as detector. The carrier gas was helium, at a flow rate of 1 mL/min.

For MS detection an electron ionization system was used with an ionization energy of 70 eV.

Crystallography. All operations were performed on a Bruker-Nonius Kappa Apex2 diffractometer, using graphite-monochromated MoKa radiation. All diffractometer manipulations, including data collection, integration, scaling, and absorption corrections were carried out using the Bruker Apex2 software. Preliminary cell constants were obtained from three sets of 12 frames. Data collection was carried out at 120K using a detector distance of 60 mm. The details of the data collection and structure refinement for

2-OTf2 and 3-OTf2 are included in Table 4.

Solution and Refinement for 5-OTf2. Data collection was carried out using a frame time of 10 sec. The optimized strategy used for data collection consisted of four phi and six omega scan sets, with 0.5° steps in phi or omega; completeness was 99.7%. A total of

3641 frames were collected. Final cell constants were obtained from the xyz centroids of

9922 reflections after integration.

From the systematic absences, the observed metric constants and intensity statistics, space group P1# was chosen initially; subsequent solution and refinement confirmed the correctness of this choice. The structure was solved using SuperFlip, and refined (full- 60 matrix-least squares) using the Oxford University Crystals for Windows program. The asymmetric unit contains one-half molecule of the complex and one acetonitrile solvate

(for the complex, Z = 1; Z’ = ½). All non-hydrogen atoms were refined using anisotropic displacement parameters. After location of H atoms on electron-density difference maps, the H atoms were initially refined with soft restraints on the bond lengths and angles to regularize their geometry (C---H in the range 0.93--0.98 Å and Uiso (H) in the range 1.2-

1.5 times Ueq of the parent atom), after which the positions were refined with riding constraints. The CF3 substituent on the triflate counter-ion was found to be disordered and was modeled as a two-component disorder. Occupancies of major and minor component atoms, F(10, 20, 30) and F(11, 21, 31), respectively, were constrained to sum to 1.0; the major component occupancy refined to a value of 0.565 (8). The modeled disordered components were refined using isotropic displacement parameters. Distance and angle restraints were applied to the disordered CF3 substituent. The final least-squares refinement converged to R1 = 0.0168 (I > 2s(I), 6217 data) and wR2 = 0.0414 (F2, 6346 data, 269 parameters).

Solution and Refinement for 5-OTf2. Data collection was carried out using a frame time of 50 sec. The optimized strategy used for data collection consisted of four phi and five omega scan sets, with 0.5° steps in phi or omega; completeness was 98.9%. A total of

2414 frames were collected. Final cell constants were obtained from the xyz centroids of

9886 reflections after integration. From the systematic absences, the observed metric constants and intensity statistics, space group P1# was chosen initially; subsequent solution and refinement confirmed the correctness of this choice. CELL_NOW, using 1053

61 reflections, revealed twinning, with 832 reflections fitting the major component, and the remaining 210 fitting the minor component. The twin law, a 180º rotation about b*, was (-

1 0 0 / 0.943 1 1 / 0 0 -1). The data were integrated as a TLQS twin. Refinements using both HKLF4 and HKLF5 files were examined; the HKLF5 file showed better results and was used in the final refinement. The structure was solved using SuperFlip and the remaining atoms were located on electron density difference maps. The structure was refined (full-matrix-least squares) using the Oxford University Crystals for Windows program. The asymmetric unit contains two molecules of the complex, and has significant void volume, examined using the SQUEEZE procedure (see below) (for the complex, Z =

4; Z’ = 2). All non-hydrogen atoms were refined using anisotropic displacement parameters. After location of H atoms on electron-density difference maps, the H atoms attached to ordered atoms were initially refined with soft restraints on the bond lengths and angles to regularize their geometry (C---H in the range 0.93--0.98 Å and Uiso (H) in the range 1.2-1.5 times Ueq of the parent atom), after which the positions were refined with riding constraints. It became apparent during the structure solution that the structure contained significant void volume (392.8 Å3 per unit cell (9.3%)). Electron density difference maps did not reveal significant electron density in the void areas, and a “test”

PLATON SQUEEZE run suggested only a small improvement in R1 (0.0025). It was decided not to process the data using SQUEEZE: the data quality and/or probable desolvation of the original acetonitrile solvate contributed to the inability to further model the contents of the void areas.

5–OTf2 15–OTf2 chemical formula C34H34Au2F6N2O6P2S2 C35H36Au2F6N2O6P2S2 fw 1200.65 1214.67 T (K) 120 120 λ (Å) 0.71073 0.71073 a (Å) 8.4651(5) 14.562(3) b (Å) 11.4271(6) 18.699(4) c (Å) 12.8696(7) 19.343(4) α (deg) 68.237(2) 115.350(7) β (deg) 71.193(2) 104.860(7) γ (deg) 87.701(2) 103.390(7) V (Å3) 1089.97(11) 4234.9(15) space group P-1 P-1 Z, Z’ 1, ½ 4, 2

3 Dcalcd (g/cm ) 1.83 1.91 µ (cm–1) 6.966 7.164 R1 (I > 2σ (I))a 0.0168 0.0825 wR2 (all data)a 0.0414 0.2362

Table 4. Crystallographic data and refinement parameters for 15-OTf2 and 5-OTf2.

Synthesis of [AuPY]2Cl2 (15-Cl2). This compound was prepared using a modified

165 literature procedure. A solution of PhICl2 (17 mg, 0.062 mmol) in CH2Cl2 (1.0 mL) was added dropwise to a stirring solution of Au2[PPh2(CH2)2] (1) (51 mg, 0.062 mmol) in

CH2Cl2 (1.0 mL). The reaction mixture initially turned bright green, but the color faded to

63 a pale yellow by the end of the addition. The reaction was stirred for an additional 15 minutes. The reaction mixture was concentrated to half of the original volume and diethyl ether (5 mL) was added, resulting in formation of a bright yellow precipitate. The precipitate was collected by centrifugation and washed with additional diethyl ether (3 × 5 mL). The precipitate was dried in vacuo to afford a crystalline, bright yellow powder. Yield:

50 mg, 90% yield. The 1H and 31P NMR spectral data matched that reported in the literature.119,122

Synthesis of [Au2(µ-PY)2(MeCN)2]OTf2 (15–OTf2). The following reaction was carried out in a 20 mL scintillation vial wrapped in black electrical tape to rigorously exclude light. A solution of 15-Cl2 (20 mg, 0.022 mmol) in CH2Cl2 (1.0 mL) was added dropwise to a stirring suspension of AgOTf (12 mg, 0.047 mmol) in MeCN (1.0 mL). The reaction was stirred for an additional 30 minutes, and then concentrated in vacuo. The resulting residue was extracted with CH2Cl2 (5 mL) and the mixture was filtered through a

0.45 micron PTFE syringe filter to give a clear, orange solution. Acetonitrile (0.1 ml) was added to the filtrate furnishing a yellow solution, and the reaction was concentrated in vacuo, resulting in precipitation of the product as a bright yellow crystalline solid. X-ray quality crystals were obtained by vapor diffusion of diethyl ether into a concentrated

1 solution of 15–OTf2 in acetonitrile. Yield: 21 mg, 84%. H NMR (400 MHz, CD2Cl2): δ

7.61 – 7.53 (m, 12H, Ar-H), 7.50 – 7.46 (m, 8H, Ar-H), 2.20 (s, 6H, MeCN), 1.96 (d, 8H,

2 31 1 19 JPH = 10.2 Hz, PY-CH2). P{ H} NMR (161.8 MHz, , CD2Cl2): δ 34.5 (s). F NMR

13 1 (376.4 MHz, , CD2Cl2): δ -77.3 (s). C{ H} NMR (100.5 MHz, CD2Cl2): δ 133.32 (s),

2 2 131.2 (d, JPC = 9.1 Hz), 131.1 (s), 129.9 (d, JPC = 11.4 Hz), 129.9 (s), 14.6 (d, JPC = 60.3

Hz), 2.8 (s). Anal. Calcd for C30H28Au2F6O6P2S2 + 1.5 C2H3N: C, 33.58; H, 2.78; N, 1.78.

Found: C, 33.33; H, 2.82; N, 1.81 (approx. 0.5 equiv. acetonitrile were lost in drying).

Synthesis of [Au2(µ-PY)2(µ-CH2)(MeCN)2]OTf2 (5–OTf2). A solution of 5-Br2 (30 mg, 0.03 mmol) in CH2Cl2 (1.0 mL) was added dropwise to a stirring suspension of AgOTf

(16.0 mg, 0.062 mmol) in CH2Cl2 (1.0 mL). Following completion of the addition, the reaction was stirred for an additional 30 minutes. The reaction was filtered through a 0.45 micron PTFE syringe filter and then concentrated in vacuo to afford the product as an off- white powder. X-ray quality crystals were obtained from vapor diffusion of diethyl ether

1 into a concentrated solution of 5–OTf2 in acetonitrile. Yield: 31 mg, 94%. H NMR (400

3 MHz, CD3CN): δ 7.64 – 7.48 (m, 16H, Ar-H), 7.46 (t, 4H, , JPH = 7.4 Hz, Ar-H), 2.68 (s,

2 3 2H, µ-CH2), 2.46 (dd, 4H, JPH = 12.9 Hz, JHH = 12.9 Hz, PY-CH2), 1.97 (s, 6H, MeCN),

2 2 31 1 1.88 (dd, 4H, JPH = 12.1 Hz, JHH = 12.1 Hz, PY-CH2),. P{ H} NMR (161.8 MHz,

19 13 1 CD3CN): δ 35.1 (s). F NMR (376.4 MHz, CD3CN): δ -77.8 (s). C{ H} NMR (100.5

2 MHz, CD3CN): δ 138.8 (s), 137.9 (s), 137.7 (d, JPC = 70.2 Hz), 135.8 (t, JPC = 4.5 Hz),

2 135.4 (t, JPC = 4.5 Hz), 134.7 (m), 129.4 (d, JPC = 90.8 Hz), 19.5 (d, JPC = 48.8 Hz), 18.4

(s). Anal. Calcd for C31H30Au2F6O6P2S2 + 1.33 C2H3N: C, 34.05; H, 2.89; N, 1.57. Found:

C, 33.75; H, 3.17; N, 1.58 (approx. 0.66 equiv. acetonitrile were lost in drying).

Generation of PPh3AuOTf and IPrAuOTf. Catalyst stock solutions (13.3 mM) of

PPh3AuOTf and IPrAuOTf were prepared by addition of AgOTf (0.040 mmol) to CH2Cl2 solutions (3.0 ml) of PPh3AuCl and IPrAuCl (0.040 mmol), respectively. The mixtures were allowed to stir for 30 minutes, and the precipitated AgCl was removed by filtration

65 using a 0.45 micron PTFE syringe filter.149,166–168 The resulting catalyst solutions were used immediately.

General Procedure for Mukaiyama Addition Reactions. In a N2-filled glovebox, a vial was charged with 16 (0.1 mmol), 17a/b (0.1 mmol), CH2Cl2 (0.2 mL), and C6D6 (0.1 mL). The appropriate amount of catalyst or catalyst stock solution was added (see Table 1) and CH2Cl2 was added to reach a total reaction volume of 0.6 mL. For runs that involved monitoring reaction progress, hexamethylbenzene (0.01 – 0.05 mmol) was added as an internal standard and the reaction mixtures were transferred to NMR tubes. For all other runs, the reactions were quenched after 12 h with 10:1 MeOH:1M HCl (0.1 mL), filtered over a plug of silica, and concentrated to dryness. The crude residues were then extracted with CDCl3 and a known amount of trimethoxybenzene (0.01 – 0.05 mmol) was added as an internal standard. The product yields were determined by integration of the 1H NMR spectra.

General Procedure for Hydroamination Reactions. In a N2-filled glovebox, a vial was charged with 20 (0.1 mmol), 21a/b/c (0.1 mmol), CH2Cl2 (0.2 mL), and C6D6 (0.1 mL). The appropriate amount of catalyst or catalyst stock solution was added (see Table 2) and CH2Cl2 was added to reach a total reaction volume of 0.6 mL. A known amount of trimethoxybenzene (0.01 – 0.05 mmol) was then added as an internal standard. The reaction mixtures were transferred to an NMR tube and heated at 40 °C for 12 h. The yields were determined by 1H NMR spectroscopy.

General Procedure for Cascade Reaction of 23 and 17b. In a N2-filled glovebox, a vial was charged with 23 (0.1 mmol), 17b (0.1 mmol), catalyst (0.004 mmol), CH2Cl2 (1.0

66 mL), and THF (1.0 mL). The reaction mixture was heated at 40 °C for 12 h in a sealed vial.

The reaction was filtered over a plug of silica and concentrated to dryness. The residue was extracted with CDCl3 and a known amount of trimethoxybenzene (0.01 – 0.05 mmol) was added as an internal standard. The product yield was determined by 1H NMR spectroscopy and GC-MS.

Chapter 4: Interactions of Naphthalene Diimides and Metal Cations3

4.1 Introduction

The ability of redox-inactive metal cations to promote electron transfer has recently attracted considerable attention.169,170,179,180,171–178 Moreover, mechanistic studies on metal ion-coupled electron transfer (MCET), defined in analogy to proton-coupled electron transfer (PCET), have been carried out with both organic and inorganic electron acceptors.181,182 In general, interactions between an electron acceptor and metal cation result in a positive shift in the reduction potential of the acceptor. This effect has been rigorously investigated for quinones and other carbonyl-containing molecules183–186 and was recently exploited to facilitate photoinduced charge accumulation187 and increase the energy density of electrochemical energy storage (EES) materials.188

Rylene diimides (RDIs) have been widely studied in EES189–198 and light harvesting applications46–48,50,66 owing to their modest reduction potentials, tunable light absorption, and chemical stability. RDIs typically exhibit two reversible one-electron reductions that are reminiscent of those observed for quinones. Despite this similarity, investigations into the influence of metal cations on the electro- and photochemical behaviour of RDIs remain scarce.199–203 Dichtel and coworkers recently reported that Li+ and Mg2+ induce anodic shifts of the second reduction potential of naphthalene diimides (NDI) incorporated in a polymer thin film.204 The effect was especially pronounced with Mg2+ for which the two

3 Reprinted in part with permission from, “Electrochemical and Structural Investigation of the Interactions between Naphthalene Diimides and Metal Cations.” Reiner, B. R.; Foxman, B. M.; Wade, C.R. Dalton Trans. 2017. 46, 9472 – 9480. Reproduced by permission of the Royal Society of Chemistry.

68 one-electron redox processes were observed to merge into a single two-electron process.

Although dramatic effects were observed for the NDI-based polymer, Li+ and Mg2+ had no effect on the redox behavior of the NDI monomer in DMSO solution.

RDIs are also well known to undergo photoreduction to radical anion species

(RDI·−) in the presence of electron donors. This behaviour has been exploited for the study of photoinduced charge separation51,52 and photooxidation of DNA.53–58 Visible light- absorbing perylene diimides (PDIs) have also been investigated as photosensitizers for

59–62 63 64,65 catalytic water splitting, CO2 reduction, and reduction of aryl halides. Despite the myriad of photochemical studies involving formation of RDI·− species, photogeneration of doubly reduced RDI2− species has rarely been reported.205,206 Consequently, new strategies for photochemical generation of RDI2− species may better position RDIs for use in multielectron photocatalysis.207,208

Considering the paucity of previous studies and implications toward energy storage and multi-electron transformations, we decided to more fully investigate the interaction of redox-inactive metal cations with RDIs. This chapter describes how Li+ and Mg2+ prompt the appearance of two-electron redox processes for N,N’-bis(2,6- diisopropylphenyl)naphthalene diimide (Dipp2NDI) in organic solvents of moderate dielectric constant and donor ability. Electrochemical titrations show that this effect is

2− driven by strong interactions between the doubly reduced diimide ([Dipp2NDI] ) and metal cations. Moreover, Li+ and Mg2+ have been found to facilitate efficient

2− photoreduction of Dipp2NDI to [Dipp2NDI] in the presence of a sacrificial electron

·− donor by inducing disproportionation of [Dipp2NDI] . Finally, chemical reduction of

2+ Dipp2NDI has led to X-ray structural characterization of a dimeric Mg complex of

2− 2+ [Dipp2NDI] . The complex features strong coordination of Mg by the oxygen atoms of the reduced NDI and is the first structurally characterized example of an NDI2− species.

4.2 Electrochemical investigation into the interactions of naphthalene diimides and metal cations

+ + Changes in the electrochemical behaviour of Dipp2NDI upon addition of Na , Li or

Mg2+ were monitored using cyclic voltammetry (CV). All experiments were carried out in anhydrous organic solvents, namely acetonitrile (MeCN), tetrahydrofuran (THF), N,N- dimethylformamide (DMF), or dimethylsulfoxide (DMSO) with 0.1 M TBAPF6 as the

− − supporting electrolyte. Anhydrous NaOTf (OTf = CF3SO3 ), Li(PF6), and Mg(NTf2)2

− − (NTf2 = (CF3SO2)2N ) were used as sources of the respective metal cations and were chosen for their solubility as well as the weakly coordinating nature of the counter anions.

The cathodic sweeps of the voltammograms measured upon titration of Dipp2NDI with

Li(PF6) and Mg(NTf2)2 are shown in Figure 25. In MeCN solution, Dipp2NDI (1 mM)

0/·− ·−/2− exhibits two reversible reductions at E1/2 = −0.92 (NDI ) and −1.38 V (NDI ) versus

0/+ ·−/2− Fc . Titration with LiPF6 (0-150 equiv.) produces a progressive anodic shift of the NDI

0/·− reduction wave (Epc = −1.45 V) to the point that it merges with the NDI wave to give a

0/·− single two-electron reduction at Epc = −0.83 V. (Figure 25A). The NDI reduction wave

+ (Epc = −0.94 V) remains unchanged in the presence of up to 10 equiv. of Li . In contrast, the addition of one equiv. of Mg(NTf2)2 to an MeCN solution of Dipp2NDI results in the

0/2− appearance of a single two-electron reduction wave (NDI ) at Epc = −0.97 V (Figure

25B).

Figure 25. Cathodic sweeps of the cyclic voltammograms of the electrochemical titration of Dipp2NDI. Dipp2NDI in MeCN, THF, DMF, or DMSO solution was titrated with up to 150 equiv. of LiPF6 or Mg(NTf2)2. Addition of 0 to 10 equiv. of the metal salts is shown in red to blue colour transitions; 10 to 150 equiv. is shown in blue to green colour transitions. All measurements were performed with 1 mM Dipp2NDI and 0.1 M TBAPF6 supporting electrolyte at a scan rate of 100 mV/s.

The presence of up to 100 equiv. of NaOTf resulted in only a slight shift of the NDI·−/2−

0/·− ·−/2− reduction wave (ΔEpc = 0.03 V) in MeCN. In THF solution, reversible NDI and NDI redox couples are observed at E1/2 = −1.04 and −1.69 V, respectively. Titration with up to

·−/2− 10 equiv. of LiPF6 or Mg(NTf2)2 results in large anodic shifts of the NDI couple, but the effects are quite different from those observed in MeCN solution. A single, two-electron reduction appears at Epc = −1.17 V (Figure 25C) after addition of only 6 equiv. of LiPF6.

0/·− ·−/2− Titration with up to 10 equiv. of Mg(NTf2)2 results in overlapping NDI and NDI reduction waves at Epc = −1.14 and −1.28 V, respectively (Figure 25D).

+ 2+ The effects of Li and Mg on the redox properties of Dipp2NDI are greatly diminished in DMSO and DMF solution. In DMF, the NDI·−/2− reduction wave experiences only small anodic shifts upon titration with up to 100 equiv. of Li+ or Mg2+ (Figure 25 E,F), and almost no shift is observed in DMSO solution (Figure 25 G,H). The NDI0/·− and

·−/2− NDI redox processes show reversibility in all solvents. However, changes in peak shape of some of the anodic waves in THF and MeCN suggest adsorption of reduced species on the electrode surface.3d

From the CV titration data, it is clear that small, charge dense cations can influence the redox behaviour of NDIs in solution. The magnitude of this influence is highly dependent on the solvent medium and can be correlated with solvent dielectric constant and donor properties. The most pronounced shifts in the NDI·−/2− reduction waves are observed in

THF, which has the lowest dielectric constant in the series (εr = 7.52) and is more weakly coordinating than DMF or DMSO. On the other hand, up to 100 equiv. of Li+ or 50 equiv.

2+ ·−/2− of Mg have almost no effect on the NDI reduction wave in DMSO solution (εr = 46.7).

Different effects are also observed for Li+ and Mg2+. Specifically, Li+ induces a larger shift of the NDI·−/2− reduction wave than Mg2+ in THF, DMF, and DMSO solution. This situation is reversed in MeCN solution where only one equiv. of Mg2+ is necessary to observe a shift leading to the appearance of a single two-electron reduction wave. These observations suggest that other factors such as cation solvation and solvent donor ability also likely play important roles.

0.05916 0.05916 ∆% = log(3) − 6 log ([9:;]) &' . .

Equation 1. Calculation for formation constant b. DEpc = Shift in peak cathodic peak potential; z = Number of electrons transferred, b = formation constant; q = Stoichiometry of complexation.

n+ + Figure 26. Plots of DEpc versus log[M ] for titration with Li in MeCN (a), DMF (b), or DMSO (d) and Mg2+ in DMF (c).

Previous electrochemical studies with p-quinones have indicated strong binding of Li+

2+ 186,209,210 or Mg to quinone radical anions. However, in the case of Dipp2NDI, the fact that the NDI0/·− redox couple remains largely unaffected by the presence of moderate amounts

+ 2+ ·− of Li or Mg suggests that [Dipp2NDI] does not interact strongly with these metal cations. Instead, the large anodic shifts of the NDI·−/2− reduction wave indicate that the

2− dominant interactions occur between [Dipp2NDI] and the metal cations.

2− + Formation constants (b) for complexation of [Dipp2NDI] with Li in MeCN, DMF, and DMSO and with Mg2+ in DMF were estimated using Equation 1.211 The rapid appearance of two-electron reduction processes for the titrations in THF and with Mg2+ in

MeCN preclude analysis of those data. Moreover, the analyses were carried out using the cathodic peak potentials (Epc) rather than half-wave potentials (E1/2) owing to changes in peak shape of the anodic waves in some of the titrations, possibly due to adsorption of the reduced species on the surface of the electrode.186 Figure 26 shows good linear correlations

·−/2− n+ between the shifts in potential of the NDI reduction waves (ΔEpc) and log[M ]. From the least squares fits of this data and Equation 1, formation constants (logβ) of 10.5, 4.7, and 1.8 were calculated for Li+ complexation in MeCN, DMF, and DMSO, respectively

(Table 5). The large differences in these formation constants, which span almost 10 orders of magnitude, reflect the profound influence of solvent on cation complexation. In addition, the formation constant for complexation of Li+ in DMF is two orders of magnitude greater than for Mg2+ in the same solvent. Although a larger formation constant might be expected for the more charge dense Mg2+, it should also experience stronger solvent-cation interactions. Thus, the lower formation constant can be rationalized by the need to

74 overcome stronger solvation effects. Although Equation 1 can also provide information about the stoichiometry of complexation (q), changes in diffusion properties of the ion- paired species have hampered the extraction of meaningful data from this analysis.

·−/2− Cation Equiv. Solvent DEpc(NDI )/V log(b) Li 10 MeCN 0.24 10.5 Li 150 MeCN 0.62 10.5 Mg 10 MeCN 0.49 N.D. Li 10 THF 0.68 N.D. Mg 10 THF 0.48 N.D. Li 100 DMF 0.14 4.7 Mg 50 DMF 0.07 2.7 Li 100 DMSO 0.06 1.8 Mg 50 DMSO N/A N.D.

·−/2− Table 5. Shifts in cathodic peak potential (DEpc) of the NDI redox couple upon addition of Li+ or Mg2+ in different solvents and formation constants log(b).

4.3 Photochemical study exploring the behavior of reduced naphthalene diimides

The electrochemical titration data show that in MeCN or THF solution, Li+ and Mg2+

2− facilitate two-electron reduction of Dipp2NDI to [Dipp2NDI] at similar or slightly less negative potentials than the unperturbed NDI0/·− redox couple. This prompted us to

2− + consider that [Dipp2NDI] might be accessed by photoreduction in the presence of Li or

Mg2+ and a suitable sacrificial electron donor. The photoreduction of NDIs to generate

[NDI]·− is quite common, but examples of photochemical generation of [NDI]2− species are rare.206

A THF solution of Dipp2NDI (8.9 µM), LiPF6 (180 µM), and Et3N (0.17 M) was irradiated with a UV lamp array (365 nm) and product generation was monitored by UV- vis spectroscopy (Figure 27a). Before irradiation, the UV-Vis spectrum exhibits π–π* transitions corresponding to Dipp2NDI at λmax = 359 and 379 nm. During the first 30 s of irradiation, these bands decrease in intensity with concomitant emergence of absorbances

·− 212–214 consistent with the formation of [Dipp2NDI] at λmax = 474, 690, and 772 nm. After

·− 30 s the [Dipp2NDI] absorption bands decrease steadily in intensity as new bands

2− diagnostic of [Dipp2NDI] appear at λmax = 413, 550, and 598 nm. The concentration of

2− [Dipp2NDI] increases up to 480 s at which point it is the dominant species in solution.

The solution undergoes noticeable color changes from nearly colorless (NDI) to brown

(NDI·−) to red (NDI2−) as the dominant species present change over the course of the experiment (Figure 27a). Photolysis experiments carried out with Mg(NTf2)2 (180 µM) in

2− place of LiPF6 also led to formation of [Dipp2NDI] via a similar evolution of major

·− species. However, [Dipp2NDI] was observed as the major photoproduct in the absence of any metal cation additive as well as in the presence of NaOTf (Figure 27b). In line with the electrochemical titration data, photoreductions conducted in MeCN showed similar

·− results to those in THF while irradiation in DMF resulted in formation of [Dipp2NDI] as

·− the major product (Figure 28). The rapid appearance of [Dipp2NDI] followed by gradual

2− + 2+ formation of [Dipp2NDI] supports a mechanism in which Li and Mg drive

·− disproportionation of photogenerated [Dipp2NDI] rather than direct two-electron reduction (Figure 29). This mechanism is supported by the ready disproportionation of

·− chemically-generated [Dipp2NDI] upon addition of LiPF6 or Mg(NTf2)2. Similar

76 experiments performed with Na+ did not show any evidence of disproportionation (Figure

30).

Figure 27. (a) UV-vis difference spectra measured for a THF solution of Dipp2NDI (8.9 µM), LiPF6 (180 µM), and Et3N (0.17 M) upon irradiation. Samples were irradiated at 365 nm for 0 – 480 s (green to blue to red colour transitions). The spectrum measured at t = 0 s served as a baseline. Also shown are images of the reaction vessel illustrating qualitative colour changes during the experiment. (b) UV-vis spectra measured on a THF solution of Dipp2NDI (8.9 µM), metal salt (180 µM), and Et3N (0.17 M). Samples were irradiated at 365 nm for 600 s.

Figure 28. UV-vis spectra after UV irradiation (365 nm) of a solution of Dipp2NDI (8.9 µM), additive (180 µM), and Et3N (0.17 M) in MeCN (lef) or DMF (right) for 10 minutes. 77

2– Figure 29. Proposed mechanism for photoreduction of Dipp2NDI to [Dipp2NDI] in the presence of Li+ or Mg2+ (Mn+).

Figure 30. (a) UV-vis spectrum of a solution of Dipp2NDI (20 µM) in THF; (b) UV-vis ·− spectrum of [Dipp2NDI] generated in situ by addition of NaHDMS (100 µM); UV-vis ·− spectra of in situ generated [Dipp2NDI] in the presence of NaOTf (200 µM) (c), LiPF6 (200 µM) (d), and MgNTf2 (200 µM) (e).

4.4 Coordination chemistry of reduced naphthalene diimide–metal cation complexes

Despite extensive study of the photo- and electrochemical properties of rylene diimides, structurally characterized examples of reduced radical anion and dianion species remain rare.215,216 Recently, Würthner and coworkers reported the structure of the sodium salt of a perylene diimide (PDI) dianion with cyano and chloro substituents in the perylene core.67,217 The structure features a two dimensional network assembled via coordination of the oxygen atoms and nitrile substituents of the PDI dianion to the Na+ ions. Given the

+ 2+ effects of Li and Mg on the electrochemical behavior of Dipp2NDI, we decided to explore the chemical reduction of Dipp2NDI in hopes of elucidating the structures of the

2− complexes formed between [Dipp2NDI] and these cations.

2– Figure 31. Synthesis of Mg complexes of [Dipp2NDI] .

As a result, we found that reaction of Dipp2NDI with isopropyl magnesium chloride

2− (iPrMgCl, 5 equiv.) in THF solution generates a Mg complex of [Dipp2NDI] as a red microcrystalline solid (26, Figure 31). The isolated solid is only sparingly soluble in polar

79 aprotic solvents such as THF, CH2Cl2, MeCN, DMSO, and DMF. This poor solubility has precluded characterization by NMR spectroscopy. However, the UV-Vis absorption spectrum of 26 in THF displays absorption bands in the visible region (λmax = 534, 590 nm) that are characteristic of NDI2− (Figure 32).212–214 Moreover, the ATR-IR spectrum features a carbonyl stretching band at 1608 cm−1, which is redshifted from the carbonyl band

2− observed for the neutral Dipp2NDI and consistent with data reported for other NDI species (Figure 33).218

Figure 32. UV-vis spectrum of 26 in THF solution.

Figure 33. ATR-IR spectrum of Dipp2NDI (black) and 26 (red).

The compound is highly air and moisture sensitive in solution but relatively robust in the solid state, maintaining its characteristic red colour for several hours under ambient conditions. Elemental and thermogravimetric analysis data (Figure 34) are consistent with an empirical formula of (Dipp2NDI)(MgCl)2(THF)4. This analytical data along with structural considerations (vide infra) have led us to formulate the tentative structure of 26 shown in Figure 31. Importantly, reaction of 26 with excess MeOH results in clean regeneration of Dipp2NDI, indicating that the chemical reduction does not lead to degradation or functionalization of the NDI molecule.

Figure 34. TGA data for 26 measured using a ramp rate of 2 °C/min.

A small number of single crystals were obtained when the reaction of Dipp2NDI with iPrMgCl was carried out under dilute conditions. X-ray diffraction analysis revealed formation of the centrosymmetric dimer 27 via coordination of adjacent oxygen atoms of

2− 2+ [Dipp2NDI] to two Mg ions (Figure 35). The structure also contains highly disordered solvent molecules, presumably THF, that could not be successfully modelled and their contribution to the structure factors was removed using the SQUEEZE routine in PLATON.

2− The coordinated THF molecules and portions of the [Dipp2NDI] groups are also severely disordered, precluding detailed analysis of some of the bond distances and angles. The coordination geometry of the Mg centers of 27 are best described as distorted square pyramidal (τ5 = 0.44) with THF molecules occupying three of the coordination sites. The

Mg1−O4 (1.961(3) Å) and Mg1−O7 (1.964(4) Å) bond distances are slightly longer than 82 the Mg−OAr distances (1.879−1.923 Å) observed in five-coordinate aryloxide complexes of the form Mg(OAr)2(THF)3 (OAr = OC6H3(CHMe2)2−2,6; OC6H2Cl3−2,4,6;

219 OC6H2(CF3)3−2,4,6), but shorter than those typically found for κ-O coordination of neutral amide ligands (2.045−2.067 Å).220,221 These comparisons suggest that

2− [Dipp2NDI] exhibits a donor ability intermediate between aryloxides and amides, although steric factors could also play a role in bond lengthening. The C−O and C−C bonds within the imide groups are consistent with the presence of the syn resonance form of

2− [Dipp2NDI] (Figure 36). Specifically, the C13−C35 (1.396(6) Å) and C32−C50

(1.393(7) Å) bond lengths suggest increased double bond character compared to the

C30−C37 (1.439(12) Å) and C26−C27 (1.440(14) Å) distances. As expected, the analogous C−C distances observed in solid state structures of Dipp2NDI are slightly longer and fall in the range 1.472−1.491 Å.222,223

We speculate that formation of 27 could proceed by the elimination of MgCl2 from 26 or a similar species under dilute conditions (Figure 31). Consequently, we considered that the bulk microcrystalline solid obtained under concentrated reaction conditions might contain a physical mixture of the single crystal phase of 27 and MgCl2. However, the peaks observed in powder X-ray diffraction patterns acquired for the microcrystalline solid do not match those in the pattern simulated from the single crystal structure of 27. Thus, while the exact structure and composition of 26 has not been elucidated, the product does not appear to contain the single crystal phase of 27.

Figure 35. Solid-state structure of 27. All hydrogen atoms and disordered atoms have been omitted for clarity. Coordinated THF molecules are shown in wireframe.

Given that NDI2− species are usually represented by the anti resonance form shown in

2− Scheme 3, the syn coordination mode of the [Dipp2NDI] groups in the solid state structure of 27 caught our attention. While kinetic favourability and/or crystal packing

2− preferences likely facilitate formation of dimer 27, the electronic structure of [Dipp2NDI] might also play a role. Bard and Kovak have previously proposed that the syn resonance structure of NDI2− should be favoured over the anti form owing to a greater preservation of aromaticity in the Kekulé structure of the former.224,225 This consideration prompted us

84 to use DFT calculations to determine the relative stability of syn and anti isomers of a protonated model complex, H2[Dipp2NDI] (Figure 37). The single point energy of the syn isomer after geometry optimization was found to be 14 kJ/mol lower than that of the anti

2− isomer. Thus, [Dipp2NDI] exhibits a slight thermodynamic preference for syn coordination that may help direct the formation of 27.

2– Figure 36. Syn and anti resonance structures of [R2NDI]

4.5 Conclusions

In summary, cyclic voltammetry studies show that Li+ and Mg2+ can induce large anodic shifts of NDI redox processes in solution, even to the extent that two-electron reduction is observed at similar or slightly less negative potentials than the unperturbed

NDI0/·− redox couple. This effect mainly originates from strong interactions between the metal cations and NDI2− while interactions with the neutral (NDI) and radical anion (NDI·−) species appear to play a less significant role. The strength of metal cation binding is highly solvent dependent, and the effects subside with increasing solvent dielectric constant and

85 donor ability. Although these results may not be completely unexpected, they disclose the importance of metal cation and solvation effects on the redox properties of RDIs and should be considered with the use of RDIs in EES applications. Moreover, the cation effects revealed in the electrochemical study have been leveraged toward the unusual photochemical generation of NDI2−. UV-Vis spectroscopy experiments indicate that the formation of NDI2− proceeds via metal cation-induced disproportionation of photogenerated NDI·−. This novel strategy for photogeneration of RDI2− species may be exploited to further study the role of this class of molecules for light-driven, multi-electron transformations. Lastly, chemical reduction of Dipp2NDI with iPrMgCl has led to the first structurally characterized example of a two-electron reduced NDI. Complex 27 adopts a

2+ 2− dimeric structure with strong coordination to Mg via adjacent O atoms of [Dipp2NDI] .

Elucidation of the structure of 27 along with DFT calculations indicate that NDI2− prefers a syn versus anti coordination arrangement, and this insight should serve as a guide for the future design of novel metal-organic materials based on reduced NDIs.

4.6 Experimental

General considerations. All manipulations were carried out using a nitrogen-filled glovebox or standard Schlenk techniques unless otherwise noted. All glassware was oven- dried in a 150 °C oven before use. THF and MeCN were degassed by sparging with ultrahigh-purity argon and dried via passage through columns of drying agents using a solvent purification system from Pure Process Technologies. DMF and DMSO were purified by stirring over CaH2 followed by vacuum distillation onto 4Å Linde-type molecular sieves.

226,227 Dipp2NDI was prepared according to literature procedure. IPrMgCl was purchased

86 from Sigma Aldrich as a 2 M solution in 2-MeTHF and was used without further purification. All other chemicals were purchased from commercial vendors and used without further purification. NMR spectra were recorded at ambient temperature on a

Varian Inova 400 MHz instrument. 1H NMR chemical shifts were referenced to residual solvent chemical shifts. Solvent-suppressed 1H NMR spectra were collected using the

WET1D sequence with default parameters.164 AT R-IR spectra were measured using a

Nicolet IR 200 with a diamond ATR accessory. TGA data were recorded on a TGAQ500 instrument controlled by TA software. Powder X-ray diffraction patterns were collected on a Rigaku Miniflex 600 diffractometer using Nickel-filtered Cu-Kα radiation (λ = 1.5418

Å). Elemental analyses were performed at Atlantic Microlab, Inc., Norcross, GA.

Electrochemistry. Cyclic voltammetry measurements were carried out in a nitrogen- filled glovebox in a one compartment cell using a CH Instruments 600C electrochemical analyzer. A glassy carbon electrode and platinum wire were used as the working and auxiliary electrodes, respectively. A silver wire was used as a pseudoreference electrode, and potentials are reported relative to an internal ferrocene reference. Solutions of electrolyte (0.1M TBAPF6) and analyte (1 mM) were also prepared in the glovebox.

Photolysis. Photoreduction experiments were carried out in gas-tight cuvettes using a

Rayonet 1070 photochemical reactor with 365 nm lamps. Solutions of Dipp2NDI (8.9 µM), additive (180 µM), and Et3N (0.17 M) were prepared in the glovebox and transferred to a gas-tight cuvette. The solutions were irradiated at room temperature and UV−vis spectra were recorded at the indicated time points using a Cary 50 UV−vis spectrophotometer.

Synthesis of 26. A solution of Dipp2NDI (100 mg, 0.17 mmol) in THF (15 mL) was treated with iPrMgCl (0.50 mL, 2 M solution in 2-MeTHF) at room temperature. The reaction mixture was stirred briefly and then allowed to stand for 1 hour. The resulting bright red solid was collected by centrifugation, washed with THF (2 × 5 mL), and dried in vacuo to give 85 mg (est. 80% yield) of 26. Anal. Calcd. for Chemical Formula:

Dipp2NDI(MgClTHF2)2 + 1 THF (C58H78Cl2Mg2N2O9): C, 65.30; H, 7.37; N, 2.63. Found:

C, 65.15; H, 7.81; N, 2.62.

Synthesis of 27. iPrMgCl (0.50 mL, 2 M in 2-MeTHF) was added dropwise to a solution of Dipp2NDI (100 mg, 0.17 mmol) in THF (60 mL). The resulting red solution was left undisturbed in a sealed vial at room temperature for 24 h to afford a mixture of flocculent red powder and dense red crystals. The red crystals were manually separated from the mixture and analysed by single crystal X-ray diffraction.

X-ray Crystallography. All operations were performed on a Bruker-Nonius Kappa

Apex2 diffractometer, using graphite-monochromated Mo Kα radiation. All diffractometer manipulations, including data collection, integration, scaling, and absorption corrections, were carried out using the Bruker Apex2 software.228 Crystallographic parameters are summarized in Table 6. Preliminary cell constants were obtained from three sets of 12 frames. Data collection was carried out at 120K, using a frame time of 120 sec and a detector distance of 60 mm. The optimized strategy used for data collection consisted of one phi and one omega scan set, with 0.5° steps in phi or omega; completeness was 99.9%.

A total of 771 frames were collected. Final cell constants were obtained from the xyz centroids of 5037 reflections after integration.

From the systematic absences, the observed metric constants and intensity statistics, space group Pbca was chosen initially; subsequent solution and refinement confirmed the correctness of this choice. The structure was solved using SIR-92, and refined (full-matrix- least squares) using the Oxford University Crystals for Windows program. The asymmetric unit contains one-half molecule of the complex as well as (likely) two highly disordered

THF solvate molecules, modeled using the SQUEEZE procedure (see below) (for the complex, Z = 4; Z’ = ½). All ordered non-hydrogen atoms were refined using anisotropic displacement parameters. After location of H atoms on electron-density difference maps, the H atoms were initially refined with soft restraints on the bond lengths and angles to regularize their geometry (C---H in the range 0.93--0.98 Å and Uiso (H) in the range 1.2-

1.5 times Ueq of the parent atom), after which the positions were refined with riding constraints. During the structure solution, electron density difference maps revealed that there were highly disordered solvate molecules, which could not be successfully modeled.

Thus, from observations and reaction history, the remaining disordered solvate was assigned as THF in a volume of 2907.5 Å3 per unit cell (26.1%). It appeared that the cavity area contained about 16 solvate molecules, located near the centers of symmetry at (½,0,0);

(0,0,½); (½,½,½); (0,½,0). Modeling with or without restraints was unsuccessful, as was step by step acquisition of peaks using successive electron density difference maps. Thus, the structure factors were modified using the PLATON SQUEEZE technique, in order to produce a “solvate-free” structure factor set. PLATON reported a total electron density of 712 e- per unit cell, likely representing about two THF molecules in the asymmetric unit (16 per unit cell would be 640 e-). Use of the SQUEEZE technique resulted in a decrease of ca. 6.3 %

89 in R. Disorder in the THF molecules bound to Mg, as well as in the NDI dianion ligand bound to Mg, was partially resolved; occupancies of the major and minor components were constrained to sum to 1.0. For the bound THF ligands: [ligand 1: C(2,3) major / (C102,103) minor, occupancies: 0.587/0.413(18)]; [ligand 2: C106,107 major / C(6,7) minor, occupancies:

0.55/0.45(2)]; [ligand 3: C(11,12) major / (C111,112) minor, occupancies: 0.61/0.39(2)]. For the NDI dianion ligand bound to Mg: [O(5,6) + C(26-30, 37) major / O(105,106) + C(126-130,

137) minor, occupancies: 0.578/0.422(7)]. The final least-squares refinement converged to

2 R1 = 0.0892 (I > 2s(I), 4323 data) and wR2 = 0.2899 (F , 8241 data, 531 parameters).

90 chemical formula C100H124Mg2N4O14 fw 1654.72 T (K) 120 λ (Å) 0.71073 a (Å) 21.5029(15) b (Å) 19.1115(14) c (Å) 27.069(2) α (deg) 90 β (deg) 90 γ (deg) 90 V (Å3) 11124.1(14) space group Pbca Z, Z’ 4, ½ 3 b Dcalcd (g/cm ) 0.99 µ (cm–1) 0.075 R1 (I > 2σ (I))a 0.0892 wR2 (all data)a 0.2899

bDoes not include squeezed solvent

Table 6. Crystallographic data and refinement parameters for 27.

Computational Details. DFT structural optimizations for syn- and anti-H2[Dipp2NDI] were carried out using Gaussian09.229 All calculations were performed with the B3LYP hybrid functional230,231 and 6-31++G** basis set for all atoms.232–236 The crystallographically reported structure of Dipp2NDI with geometrically placed H atoms in the syn or anti positions were used as starting points.222,223

Figure 37. DFT optimized structures and single point energies of syn and anti H2[Dipp2NDI]. 92

Chapter 5: Structural and Electrochemical Investigation into Transition Metal– Rylene Imide Complexes

5.1 Introduction

Understanding the details of electronic communication among adjacent redox active centers is critical in the study of electron transfer and transport. Many synthetic and biological systems as well as energy storage devices are grounded in effective orchestration of multi–electron events. As introduced in the previous chapter, rylene diimides (RDIs) are a class of redox active aromatic compounds that have been widely studied for applications in energy storage and fundamental electron transfer processes.189–198 Encouraged by the substantial effect of simple alkali earth metals on the redox potentials of NDIs described in

Chapter 4, we were inspired to investigate the interactions between reduced NDI species and transition metals. Additionally, the electrochemical behavior of transition metal–NDI complexes is fundamentally interesting considering the availability of redox activity of both components in similar potential regions. The redox couples of catalytically active base metals including Fe, Ni, and Co are well matched with the reduction potential of NDIs suggesting electron transfer between these species may be facile. To this end, electrochemical titration of Dipp2NDI with CoOTf2THF2 was performed (Figure 38).

However, only negligible shifts in the NDI reduction waves were observed and no clear metal-based reduction event were assigned. The NDI0/•− redox couple showed no shift with up to 2 equivalents of Co2+ while the NDI•−/2− couple only experienced a small anodic shift of 85 mV. Moreover, the NDI redox couples became quasi-reversible upon addition of

Co2+, suggestive of species adsorbed to the electrode. The observed poor redox behavior

93 prompted us to consider alternative strategies for bringing together RDIs and transition metal species.

Figure 38. Cyclic voltammograms (CVs) of Dipp2NDI in the presence of 0 (red), 1 (purple), or 2 (blue) equivalents of Co(OTf)2THF2. Samples were collected in 0.1 M n [ Bu4N][PF6] in MeCN (scan rate = 100 mV/s). All potentials are referenced to the ferrocene (Fc)/ferrocenium (Fc+) redox couple.

In order to facilitate metal–NDI communication, we envisioned anchoring a binding unit proximal to the NDI which could capture transition metal ions. Transition metal–NDI complexes are often synthesized and studied in the context of excited state dynamics where the metal center can act as a photosensitizer. Metal centers are frequently placed distal to

NDI fragments, mitigating any direct interaction.66 We hypothesized that installing NDI moieties proximal to a metal center would facilitate metal–NDI interactions. Moreover,

94 modulation of redox potentials by non-covalent means is important to consider in the rational design of redox-active materials.

We were inspired by polypyridyl ligand systems which have been shown to support impressive electrocatalytic activity at base metals including cobalt.237–239 Thus, we considered decoration of rylene imide fragments with pyridyl-thiazole units might be a potent manifold for investigating communication between adjacent redox active sites.

Pyridyl-thiazole is synthetically easier to access than the analogous bipyridine backbone.

The wider N-C-RDI bond angle in the pyridyl-thiazole unit (Figure 39) should draw the

RDI moiety away from the metal fragment which reduces steric crowding. Moreover, the pyridyl-thiazole unit bears prolific coordination chemistry allowing facile access to a broad family of complexes.240–247

Figure 39. Comparison of bond angles between a pyridyl-thiazole and bipyridine ligand backbone.

Judicious choice of both the metal ion and rylene imide fragment should allow elucidation of valuable structure-function relationships. This chapter describes the synthesis, structure, and electrochemical study of cobalt and zinc metal complexes bearing

95 rylene imide derived ligands. Also described is the isolation of a Pd(0)–NDI adduct that is well poised for the synthesis of bimetallic complexes and exhibits an interesting oxidatively triggered isomerization.

2+ 2+ 5.2 Structural and electrochemical study of Co and Zn –NDI complexes

The ligand, MNTP, (Figure 40) was prepared in 85% yield by condensation of pyridyl–thiazole amine 28 with mesityl naphthalene imide anhydride 29 in DMF at 130

1 °C. The H NMR spectrum in CDCl3 solution features diagnostic singlets at 8.89 and 8.39 ppm ascribed to the core NDI peaks and thiazole proton, respectively. The rest of the spectrum exhibits signals with the expected chemical shifts and coupling constants.

MNTP was treated with a solution of Co(OTf)2THF2 in MeCN to afford Co(MNTP2)OTf2

(30) as an off white microcrystalline solid in 83% yield. The 1H NMR spectrum in MeCN solution exhibits eight broad paramagnetic peaks (FWHM ~ 1600 Hz) spanning a spectral window from -15 – 95 ppm. The number of resonances supports a symmetric species.

Solution state Evan’s method shows a magnetic moment of 4.77 µB, consistent with an S =

3/2 ground state. The UV-Vis spectrum exhibits π–π* transitions corresponding to the NDI moiety at λmax = 360 nm and 380 nm. The similarity of the UV-vis spectrum of 30 to those measured for Dipp2NDI (λmax = 359 nm and 379 nm) and free MNTP (λmax = 358 nm and

378 nm) indicate negligible electronic coupling between the metal and NDI in the neutral state.

Figure 40. Synthesis of M(MNTP)2OTf2. Conditions: (i) DMF, 130 °C, 16 h; (ii) M(OTf)2 (M = Co, Zn), MeCN, 24 hr.

The redox behavior of 30 was explored by cyclic voltammetry (CV) to evaluate the effect of the Co2+ cation on the redox potentials of the NDI. MNTP exhibits two reversible

0/•− •−/2− 0/+ reductions at E1/2= −0.878 V (NDI ) and −1.321 V (NDI ) versus Fc (Figure 41a) which is consistent with the reduction potentials of other aromatic NDIs in the literature.48

The CV of 30 contains four quasi-reversible features with Epc = –0.680 V, –0.816 V, –1.09

V, and –1.16 V as well as two ill-defined features at –2.10 V and –2.28 V. The sharp current deflection at –1.16 V is likely due to adsorption at the electrode surface. At higher scan rates the feature exhibits characteristic diffusive behavior – i.e. the peak current is proportional to the square root of scan rate (Figure 42b). A dilution study was performed to garner additional information about the adsorbed species. A 1.4 mM MeCN solution of

30 was diluted by consecutive additions of MeCN. As shown in Figure 42a, sequential dilution shows a loss in intensity of the cathodic peak observed at –1.16 V with a

97 concomitant progressive anodic shift up to 111 mV. This is consistent with strong adsorption occurring after the third reductive event.

Metal-based reduction of Zn polypyridyl complexes is expected to occur outside the

248 potential window of the NDI-based redox processes. Consequently, Zn(MNTP)2OTf2

(31), was synthesized In order to aid in the assignment of the reduction waves observed in the CV of 30.

31 was synthesized by reaction of MNTP and ZnOTf2 in MeCN solution (Figure 40).

The diamagnetic NMR spectrum in MeCN solution is consistent with a symmetric species.

Diagnostic resonances include the NDI core protons which resolve into two doublets resonating at 8.81 and 8.64 ppm and the a pyridyl proton which shifts 0.64 ppm upfield upon binding Zn2+. The remaining resonances display the expected chemical shifts and coupling constants. The UV-Vis spectrum is similar to that observed for 30.

Figure 41. Cyclic voltammograms (CV) of a) MNTP (blue), Co(MNTP)2OTf2 (30, red), st and Zn(MNTP)2OTf2 (31, green) and b) 1 derivative CV plots with respect to current. n Collected in 0.1 M [ Bu4N][PF6] in MeCN (scan rate = 100 mV/s). All potentials are referenced to the ferrocene (Fc)/ferrocenium (Fc+) redox couple.

Figure 42. Cyclic voltammograms (CV) of Co(MNTP)2OTf2 (30) at various n concentrations of MeCN solution (a) or scan rates (b). Collected in 0.1 M [ Bu4N][PF6] in MeCN (scan rate = 100 mV/s unless otherwise noted). All potentials are referenced to the ferrocene (Fc)/ferrocenium (Fc+) redox couple.

The CV of 31 contains three quasi-reversible features at Epc = –0.680 V, –0.767 V, and

–1.10 V as well as two ill-defined features at –2.24 V and –2.41 V. The third reductive peak at –1.10 V is likely a poorly resolved two electron feature. At low concentrations, the CV of 30 also only exhibits three reduction waves because the 3rd and 4th NDI–based reductions coalesce into one broad feature (Figure 42a). Differential pulse voltammetry will be required to confirm that the feature at –1.10 V in the CV of 31 is the result of a 2–electron event.

The set of cathodic peaks observed from –0.6 to –1.3 V for both 30 and 31 are assigned as sequential NDI–based 1–electron reductions while the ill–defined cathodic features are tentatively assigned as either metal or pyridine based reductions.249–251 This is line with the observed electrochemistry for NDI–based macrocycles where electronic coupling between 100 distinct NDI units is responsible for the splitting of reduction waves into resolved one- electron processes.192,252–259 Unlike work presented by Stoddart and coworkers where the

NDI•−/2− redox processes of oligomeric NDI macrocycles occurs at more negative potentials than an analogous NDI monomer, the formation of NDI dianions in 30 and 31 occurs at a potential 200 mV more positive than the NDI•−/2− redox couple of MNTP. This is likely due to the Lewis acidity and charge of the Co2+ or Zn2+ metal centers offsetting the Coulombic repulsion inherent to reduced species. The ability of M(MNTP2)OTf2 complexes (M = Co2+, Zn2+) to reversibly accept up to four electrons at mild potentials signals the importance of NDI–metal center proximity and heralds these complexes as potent charge carriers in energy storage or synthetic catalytic applications.

0 0 •− Notably, a 240 mV anodic shift is observed between the [NDI2] /[NDI NDI ] reduction

0/•− waves of 30 and 31 (Epc = –0.680 V) and the NDI reduction wave of MNTP (Epc = -

0.919 V). This is contrast to our previous work with Li+ and Mg2+ which demonstrated that

0/•− charge dense cations have a neglibible effect on the NDI redox process of Dipp2NDI.

260 Preorganization of the NDI moiety within the coordination sphere of the metal ion likely facilitates interactions between the metal ion and reduced NDI species. Additionally, CV shows the peak separation between the first and second NDI-based reduction potentials differ between 30 (D = 136 mV) and 31 (D = 87 mV). This is clearly demonstrated by plotting the first derivative of current with respect to time against applied potential (Figure

41b). It is important to note that the potentials shown in Figure 41b are not the Epc of each peak but rather the inflection point from the cathodic wave.

101

The Robin-Day classification is used to describe the amount of delocalization in mixed valent complexes. Class I describes a system that has completely localized redox states while Class III describes a system with completely delocalized redox states. Class II describes an intermediate case. Complexes can be categorized according to the equilibrium

261 constant (Kc) between the mixed valent and homovalent species (Equation 2). Kc can be extracted after rearranging the Nernst equation (Equation 3). Accordingly, the calculated Kc for 30 and 31 is 200 and 30, respectively, corresponding to a Robin-Day

Class I system where redox states are largely localized at each NDI.

Equation 2. Comproportionation reaction of 30 or 31 to form a mixed valent complex.

F∆EE> K = exp B H > RT Equation 3. Rearranged Nernst Equation. F = Faraday constant; R = gas constant; T = temperature (K);

The identity of the metal or differences in structure could both affect the electrochemical behavior. In order to ascertain the origin of this effect, the solid state structures of 30 and 31 were studied by single crystal X-ray diffraction. X-ray quality crystals of 30 were obtained by vapor diffusion of Et2O into MeCN solution. 30 crystallizes

+ in the space group P21/c with one molecule of [30] in the asymmetric unit accompanied 102 by one inner sphere and one outer sphere OTf– ion. The structure features the cobalt center in an octahedral coordination environment. The cobalt coordination sphere contains

MNTP pyridine units in a trans arrangement and MNTP thiazole units in a cis arrangement

(Figure 43). The remaining coordination sites are occupied by a coordinated MeCN molecule and a coordinated OTf– ion. Co–N bond distances (2.094(5) – 2.140(5) Å) are similar to other Co2+ bipyridine complexes (range of Co2+–N distances = 1.821 – 2.380

Å).4 The outer sphere OTf– ion resides 2.914 Å above the NDI core which may indicate a weak anion–π interaction. Saha and coworkers have previously characterized a NDI–

– PdOTf2 coordination polymer which exhibits a similar interaction between the OTf ions and NDI fragments (NDI–OTf– = 2.83 Å).262 Despite the increased interest in anion–π interactions in both synthetic and biological contexts, crystallographic characterization of such interactions remains rare.

Complex formation is likely driven in part by intramolecular π–π donor–acceptor interactions between the pyridine and NDI moieties. NDI donor–acceptor complexes are well reported in the literature48,194,263,264 and are usually expressed by a polar/pi model (i.e. maximization of complementary quadrupole–quadrupole interactions). The electron poor

NDI unit has a high quadrupole moment (Qzz ≥ 14B) and the relatively electron rich

265,266 pyridine unit has a lower quadrupole moment (Qzz ~ 2). Intramolecular distances of

3.356 and 3.442 Å between the two moieties support a donor acceptor interaction.

4 Based on a 2018 search of the Cambridge Structural Database. 103

Figure 43. Solid state structure of Co(MNTP)2OTf2 (30). All hydrogen atoms, solvent molecules, and disordered atoms have been omitted for clarity. The coordinated MeCN molecule and OTf– counterion are shown in the wireframe. (b) Idealized view to illustrate the anion–π interaction between the outer sphere OTf– counterion and the NDI surface.

Crystals of 31 produced a weak diffraction pattern that precludes a detailed analysis of bond distances and angles. However, data refinement unambiguously defines the connectivity and topology of the molecule. 31 crystallizes in the space group P21/c with one molecule of [31]+ in the asymmetric unit accompanied by one inner sphere and one outer sphere OTf– ion. The coordination environment around the octahedral zinc center is identical to the cobalt center in 30. The preliminary structure solution for 31 overlays well with 30. Thus, differences in the observed electrochemical behavior between 30 and 31 are

104 likely not due to structural differences but rather the identity of the metal ion. Investigation into the exact origin of this electrochemical variance is underway and is likely related to

2+ 2+ the difference in Lewis acidity between the Co and Zn metal centers. Additionally, exploration of how open and closed shell metal centers interact with reduced NDI species may also provide valuable insight.

2+ 2+ 5.3 Structural and electrochemical extension to Co and Zn –phthalimide complexes

We anticipated that the M(MNTP)2OTf2 complexes would exhibit metal– and ligand– based reductions that would facilitate their use for synthetic applications employing multielectron transformations. However, neither 30 nor 31 showed clear metal based reductions in their respective voltammograms. The NDI–based ligands can accept two electrons at mild potentials which may preclude access to a metal based reduction owing to Coulombic effects. Smaller rylenes such as phthalimides can only accept one electron at mild potentials which may allow organometallic complexes derived from these fragments to engage in both ligand and metal borne redox chemistry. We decided to pursue the synthesis and subsequent characterization of complexes bearing phthalimide derived ligands in order to compare the coordination and electrochemistry to the NDI–based congeners.

The phthalimide derived ligand, PTP, (Figure 40) was prepared in 71% yield by condensation of pyridyl–thiazole amine 28 with phthalic anhydride in acetic acid at reflux.

Treatment with Co(OTf)2THF2 or ZnOTf2 in MeCN solution afforded [Co(PTP)2]OTf2

1 (32) and [Zn(PTP)2]OTf2 (33) respectively as microcrystalline solids. The H NMR spectrum of 32 features seven paramagnetic peaks (FWHM ~ 30 Hz) spanning a spectral window from 0 – 95 ppm. The number of resonances supports a symmetric species.

105

Solution state Evan’s method shows a magnet moment of 4.29 µB, consistent with an S =

3/2 ground state. The 1H NMR spectrum of 33 is consistent with a symmetric diamagnetic species. Diagnostic resonances include the thiazole proton which shifts downfield upon binding Zn2+. The remaining resonances display the expected coupling constants and chemical shifts.

Figure 44. Synthesis of [M(PTP)2]OTf2. Conditions: (i) AcOH, 130 °C, 16 h; (ii) M(OTf)2 (M = Co, Zn), MeCN, 24 hr.

Figure 45. Cyclic voltammograms (CV) of PTP (blue), [Co(PTP)2]OTf2 (32, red), and n [Zn(PTP)2]OTf2 (33, green). Collected in 0.1 M [ Bu4N][PF6] in MeCN (scan rate = 100 mV/s). All potentials are referenced to the ferrocene (Fc)/ferrocenium (Fc+) redox couple. 106

The redox behavior of 32 and 33 were explored by CV in order to evaluate the effect of metal ions on the phthalimide redox chemistry. PTP exhibits two irreversible reductions

0/•− •−/2− 0/+ at Epc= −1.70 V (PTP ) and −2.44 V (PTP ) versus Fc (Figure 45) which is consistent with the reduction potentials of other phthalimides reported in the literature.267

The CV of 32 features four broad irreversible reductions with Epc = −1.02 V, −1.15 V,

−1.35, and −1.64 V. The CV of 33 is similar to 32, displaying broad irreversible reductions.

However, the reduction at –1.35 V is notably absent suggesting that current deflection at this potential in the CV of 32 may be due to a metal–based reduction. The reductions of 32 are thus assigned as two sequential one electron PTP–based reductions followed by a CoII/I reduction. The final reduction is likely a ligand borne PTP•−/2− couple considering the presence of this current deflection in both the voltammogram of 32 and 33. However, this would require a large anodic shift (~ 800 mV) from the free ligand reduction potential.

Additionally, the putative reduction of Co in the CV of 32 doesn’t result in the expected cathodic shift of the assigned PTP•−/2− redox couple. This could be due to changes in structure or speciation upon reduction. Infrared spectroelectrochemistry may be valuable in helping to elucidate the origin of these cathodic features. Attempts to scan to more negative potentials led to fouling of the working electrode and poor reproducibility between scans likely owing to adsorption of reduced species. The irreversibility of the redox waves of 32 and 33 can be attributed to the instability of the phthalimide core which is known to undergo self–dimerization upon reduction.267–269 While voltammetry did not reveal an obvious metal based reduction of metal complexes bearing NDI derived ligands, the CV of 32 does include a cathodic feature indicative of a metal–based reduction. The

107

NDI moieties in 30 accept four electrons at –1.16 V which likely forces the CoII/I redox couple to more negative potentials owing to Coulombic repulsion. In contrast, the phthalimide moieties in 32 only accept two electrons at –1.16 V which may allow access to the CoII/I redox couple at more mild potentials.

32 and 33 were studied by single crystal X-ray diffraction in order to elucidate structural differences between their NDI-based congeners. X-ray quality crystals of 32 and 33 were obtained by vapor diffusion of Et2O into MeCN solutions. Both molecules crystallize in

2+ 2+ 2+ the space group !1# with one molecule of [M(PTP)2] (M = Co , Zn ), two outer sphere

OTf– ions, and two interstitial MeCN molecules in the asymmetric unit (Figure 46). The octahedral metal coordination environment is occupied by meridonal k3–N,N’,O coordination of two PTP ligands and the thiazole units are in a trans arrangement. Co–N distances (2.045 – 2.159 Å) and Zn – N distances (2.055 – 2.247 Å) are consistent with other reported divalent metal bipyridine complexes (range for Co2+–N = 1.821 – 2.380 Å; range for Zn2+–N = 1.915 – 2.364 Å).5 Unlike the solid state structures of the

M(MNTP)2OTf2 complexes, the [M(PTP)2]OTf2 structures reveal coordination via the rylene imide oxygen. The decreased ring size of the phthalimide would be expected to draw the imide oxygens away from the metal center; however, the increased nucleophilicity of the heterocycle must offset this geometric constraint.

5 Based on a 2018 search of the Cambridge Structural Database. 108

Figure 46. Solid state structure of a) [Co(PTP)2]OTf2 (32), b) [Zn(PTP)2]OTf2 (33), and c) overlay of structures of 32 (red) and 33 (blue). All hydrogen atoms, solvent molecules, OTf– counterions, and disordered atoms have been omitted for clarity.

Additionally, the phthalimide moiety is considerably less π–acidic than the NDI as evidenced by the difference in the first ligand reduction potential (Epc = -0.95 V vs –1.70

V). The lower Lewis acidity and consequently lower quadrupole moment would not be expected to help template isomer formation via π– π stacking. The solid state structures of

32 and 33 overlay nearly identically (Figure 46c) indicating that the metal identity has little effect over the ensuing structural isomer.

5.5 Pd0–MNTP Adduct: A New Coordination Mode for NDIs

0 Pd –enone adducts such as Pd2dba3 (dba = dibenzylideneacetone) are well-established

Pd0 sources and serve a regular role in traditional cross coupling chemistry. However despite the analogy to Lewis acidic enones such as dba, NDI complexes of Pd0 have never been reported. We were motivated by how the redox activity of the NDI moiety might

109 affect the structural chemistry of the ensuing Pd0–adduct and whether partial charge transfer from Pd to the NDI fragment might be observable or triggered in the presence of

Lewis acids or light.

Pd–NDI complex 34 was synthesized by addition of Pd(PPh3)4 to a THF solution of

MNTP (Figure 49). The reaction immediately afforded a magenta solution which was analyzed by UV-vis spectroscopy. The spectrum exhibits π–π* transitions at λmax = 356 and

376 nm. Additionally, two broad features can be observed at λmax = 446 and 547 nm.

Notably absent from the spectrum are diagnostic absorbances at 474 or 590 nm indicative of reduction of the NDI fragment to NDI•− or NDI2– respectively. Upon brief exposure of the UV-vis cell to air, the solution color immediately bleaches. The resulting UV-vis spectrum shows complete loss of the low energy absorbances, prompting their assignment as MLCT bands which is consistent with the high π–acidity of the NDI.

Figure 47. UV-vis spectrum of 34 before (solid line) and after exposure to air (dashed line). 110

Crystals suitable for single crystal X-ray diffraction were grown under dilute reaction conditions in THF. 34 crystallizes in the space group !1# with one molecule of 34 and one molecule of interstitial THF in the asymmetric unit. The structure reveals a (PPh3)2Pd–NDI adduct with the Pd bound h2 at the periphery of the NDI core (Figure 48). To the best of our knowledge this is the first report of such a binding mode for any rylene imide. Pd adopts a slightly distorted square planar geometry (t = 0.99) occupied by two cis PPh3 ligands and one C–C π bond of the NDI aromatic core. The N,N’ ligand chelating functional group of the MNTP ligand remains available for metal binding which poises this molecule for the controlled synthesis of homo or heterobimetallic complexes. The C10–C11 vs C18–C19 bond lengths should be the most diagnostic structural elements for determining the relative charge transfer from Pd to the NDI core. The C10–C11 bond length (1.43 Å) is slightly longer than the C18–C19 bond length (1.39 Å) which is indicative of weak activation of

0 0 the π bond. Thus, 34 is best described as (PPh3)2Pd MNTP . Moreover, the C10–C11 bond

0 270,271 length is consistent with other reported (PPh3)2Pd –enone adducts in the literature.

The UV-vis data supports this assignment as brief exposure of 34 to air or moisture easily

0 releases the weakly bound (PPh3)2Pd fragment.

111

Figure 48. Solid state structure of 34. All hydrogen atoms, solvent molecules, and disordered atoms have been omitted for clarity.

The poor solubility of 34 precluded characterization by solution state NMR. However, addition of organic oxidants produced soluble byproducts that could be analyzed. Exposure of a DMSO solution of 34 to iodobenzene liberated free ligand and furnished the expected

II 0 (PPh3)2Pd (Ph)(I), affirming 34 behaves as (PPh3)2Pd precursor.

Figure 49. Reactivity of 34 with organic oxidants. 112

Treatment of a DMSO suspension of 34 with 3 equivalents of PhICl2 cleanly generated

35 with stoichiometric amounts of PPh3PCl2 (Figure 49). Assignment of 35 was confirmed by independent synthesis. Redox triggered migration of the Pd center from the soft Lewis acidic NDI core to the hard Lewis basic pyridyl-thiazole unit was intriguing. The development of switchable especially photoswitchable, bistable molecules have potential applications in optical molecular information storage.272,273 Considering NDIs can be photoreduced by mild sacrificial reductants such as Et3N (Chapter 4), we were inspired to attempt a phototriggered isomerization of complex 34. The electron rich Pd0 center was expected to serve as a suitable reductant for the NDI to generate a zwitterionic Pd2+ – NDI2– species such as 36. Unfortunately, prolonged irradiation of a THF suspension of 34 at 365

•− nm in the absence or presence of NDI disproportionation agents such as LiPF6 or MgNTf2 did not show evidence of electron transfer from Pd to the NDI fragment. The ancillary PPh3 ligands may hinder isomerization from 34 to 36 by competing with the N,N’ ligand chelate for capture of incipient Pd2+ species. Phosphine scavengers may be necessary to help drive this reaction. For example, a strong Lewis acid such as pentafluorophenylborane could serve dual purposes by acting as both an NDI•− disproportionation agent and phosphine ligand scavenger. Nucleophilic imide oxygens from NDI2– may also be competitive nucleophiles for photogenerated Pd2+ intermediates.

5.4 Conclusions

Rylene imide derived ligands, MNTP and PTP, were used to synthesize Co2+ and Zn2+ complexes and the electrochemistry and coordination chemistry of the resulting complexes were investigated. M(MNTP)2OTf2 complexes display complex electrochemical behavior

113 featuring four sequential one-electron ligand based reductions. The ability of these complexes to accept four electrons at mild potentials (-1.16 V) signals their potential as potent charge carriers in energy storage or synthetic catalytic applications. Notably, while the CV of Co(MNTP)2OTf2 complex 30 did not show a clear metal–based reduction event,

II/I the CV of [Co(PTP)2]OTf2 complex 32 did exhibit current deflection attributable to a Co redox event. The difference in electrochemical behavior between 30 and 32 was ascribed to the number of electrons that can be stored in the reduced ligands at mild potentials.

While MNTP can be easily reduced by two electrons, PTP can only be reduced by one electron at mild potentials leading to more accessible metal-borne redox events owing to decreased Coloumbic repulsion between reduced metal and rylene imide fragments.

Additionally, the coordination chemistry of the M(MNTP)2OTf2 and [M(PTP)2]OTf2 complexes differed substantially. The more nucleophilic and less π–acidic phthalimide ligands afforded a tridentate chelate around the divalent metal centers. Conversely, the high

π–acidity of the NDI ligands template the formation of structures poised for strong π–π interactions and furnish bidentate chelates around the metal centers. The structural differences in these isomers could be relevant in synthetic applications where reduced rylene imides could help deliver substrates to the catalytically active metal center.

Treatment of MNTP with Pd(PPh3)4 afforded complex 34 which features a novel binding mode between the NDI and the Pd0 fragment. Pd migration from the NDI core to the N,N’ ligand chelate to furnish 35 was triggered by oxidation with PhICl2. Attempts to phototrigger the same isomerization were unsuccessful. However, an exhaustive study has

114 not been completed and use of phosphine scavengers may be necessary to effect this interesting transformation.

Study of the interaction between adjacent redox-active centers is instrumental in the continued rational development of redox-active materials. Rylene imides present a modular scaffold for tuning the interaction between redox active metal centers and reduced ligand species. Future work will focus on leveraging the observed structure-function relationships toward the electrocatalytic reduction of small molecules.

5.5 Experimental

274 275 purification system from Pure Process Technologies. Co(OTf)2THF2 and 28 were synthesized according to literature procedure. All other chemicals were purchased from commercial vendors and used without further purification. NMR spectra were recorded at ambient temperature on a Varian Inova 400 MHz instrument. 1H and 13C NMR chemical shifts were referenced to the residual solvent chemical shifts.

115 and potentials are reported relative to an internal ferrocene reference. Solutions of electrolyte (0.1M TBAPF6) and analyte (1 mM) were also prepared in the glovebox.

Crystallography. All operations were performed on a Bruker-Nonius Kappa Apex2 diffractometer, using graphite-monochromated MoKa radiation. All diffractometer manipulations, including data collection, integration, scaling, and absorption corrections were carried out using the Bruker Apex2 software.S1 Preliminary cell constants were obtained from three sets of 12 frames. Data collection was carried out at 120K using a detector distance of 60 mm. The details of the data collection and structure refinement for

30, 31, 32, and 33 are included in Table 7.

Solution and Refinement for 30. A total of 903 frames were collected. The total exposure time was 22.57 hours. The frames were integrated with the Bruker SAINT software package using a narrow-frame algorithm. The integration of the data using a monoclinic unit cell yielded a total of 40684 reflections to a maximum θ angle of 23.03°

(0.91 Å resolution), of which 9351 were independent (average redundancy 4.351, completeness = 99.9%, Rint = 14.96%, Rsig = 15.75%) and 4812 (51.46%) were greater than 2σ(F2). Due to the very weak diffraction of the crystal, it was not possible to obtain data to a higher resolution. The final cell constants of a = 20.1925(9) Å, b = 21.1333(9) Å, c = 17.2230(8) Å, β = 114.339(3)°, volume = 6696.4(5) Å3, are based upon the refinement of the XYZ-centroids of 2776 reflections above 20 σ(I) with 4.482° < 2θ < 41.16°. Data were corrected for absorption effects using the Multi-Scan method (SADABS). The ratio of minimum to maximum apparent transmission was 0.847. The calculated minimum and maximum transmission coefficients (based on crystal size) are 0.9300 and 0.9680.

116

The structure was solved and refined using the Bruker SHELXTL Software Package within APEX3 1 and OLEX2, using the space group P21/c, with Z = 4 for the formula unit, C70H49CoF6N11O14S4. Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in geometrically calculated positions with Uiso = 1.2Uequiv of the parent atom (Uiso = 1.5Uequiv for methyl). Two fluorine atoms of the outer-sphere triflate anion were found to be disordered over two positions. The occupancies were allowed to refine freely with constraints (EADP) on the thermal displacement parameters of the disordered atoms, converging at a 54/46 ratio for the major and minor positions, respectively. The final anisotropic full-matrix least-squares refinement on F2 with 971 variables converged at R1 = 6.31%, for the observed data and wR2 = 14.88% for all data.

The goodness-of-fit was 1.001. The largest peak in the final difference electron density synthesis was 0.393 e-/Å3 and the largest hole was -0.484 e-/Å3 with an RMS deviation of

0.082 e-/Å3. On the basis of the final model, the calculated density was 1.557 g/cm3 and

F(000), 3212 e-.

Solution and Refinement for 32. Cell parameters were retrieved using the SAINT

(Bruker, V8.38A, 2016) software and refined using SAINT (Bruker, V8.38A, 2016) on

9880 reflections, 18 % of the observed reflections. Data reduction was performed using the

SAINT (Bruker, V8.38A, 2016) software which corrects for Lorentz polarisation. The final completeness is 100.00 % out to 30.553° in Q. A multi-scan absorption correction was performed using SADABS-2014/5 (Bruker, 2016) was used for absorption correction. wR2(int) was 0.0399 before and 0.0304 after correction. The ratio of minimum to maximum transmission is 0.8630. The x/2 correction factor is 0.00150. The absorption coefficient µ of

117 this material is 0.698 mm-1 at this wavelength (l = 0.71073Å) and the minimum and maximum transmissions are 0.6892 and 0.7461. The structure was solved in the space group P-1 (# 2) by Intrinsic Phasing using the XT (Sheldrick, 2015) structure solution program and refined by Least Squares using version 2016/6 of XL (Sheldrick, 2008). All non-hydrogen atoms were refined anisotropically. Hydrogen atom positions were calculated geometrically and refined using the riding model. The CF3 group of one triflate anion was found to be disordered over two positions, as was one of the two acetonitrile solvent molecules. The relative occupancies of the two positions was freely refined, converging at 58/42 for the CF3 and 62/38 for the MeCN. Constraints (EADP) were used on the thermal displacement parameters of the MeCN atoms.

Solution and Refinement for 33. The total exposure time was 6.29 hours. A total of

2264 frames were collected. The frames were integrated with the Bruker SAINT software package using a narrow-frame algorithm. The integration of the data using a triclinic unit cell yielded a total of 37092 reflections to a maximum θ angle of 27.50° (0.77 Å resolution), of which 9716 were independent (average redundancy 3.818, completeness = 99.8%, Rint

= 3.24%, Rsig = 3.32%) and 7926 (81.58%) were greater than 2σ(F2). The final cell constants of a = 11.8116(4) Å, b = 12.8887(4) Å, c = 14.7358(5) Å, α = 108.043(2)°, β =

93.446(2)°, γ = 95.248(2)°, volume = 2114.78(12) Å3, are based upon the refinement of the

XYZ-centroids of 9906 reflections above 20 σ(I) with 4.538° < 2θ < 54.86°. Data were corrected for absorption effects using the Multi-Scan method (SADABS). The ratio of minimum to maximum apparent transmission was 0.907. The calculated minimum and maximum transmission coefficients (based on crystal size) are 0.6580 and 0.8260. The

118 structure was solved and refined using the Bruker SHELXTL Software Package within

APEX3 1 and OLEX2, using the space group P-1, with Z = 2 for the formula unit,

C38H24F6N8O10S4Zn. Non-hydrogen atoms were refined anisotropically. Hydrogen atoms were placed in geometrically calculated positions with Uiso = 1.2Uequiv of the parent atom

(Uiso = 1.5Uequiv for methyl). Both molecules of co-crystallized acetonitrile solvent were disordered over two positions, as was the CF3 portion of one triflate anion. The relative occupancies of each group of disordered atoms was freely refined, with EADP constraints used on C34/C34A, C36/C36A and C37/C37A. The final anisotropic full-matrix least- squares refinement on F2 with 683 variables converged at R1 = 3.58%, for the observed data and wR2 = 9.21% for all data. The goodness-of-fit was 1.040. The largest peak in the final difference electron density synthesis was 0.656 e-/Å3 and the largest hole was -0.433 e-/Å3 with an RMS deviation of 0.062 e-/Å3. On the basis of the final model, the calculated density was 1.665 g/cm3 and F(000), 1072 e-.

119

30 31 32 chemical formula CoC66H43F6N9O14S4 ZnC66H43F6N9O14S4 CoC36H23F6N5O10S4 fw 1569.37 1575.84 1053.82 T (K) 120 120 120 λ (Å) 0.71073 0.71073 0.71073 a (Å) 20.1925(9) 20.094(7) 11.9335(3) b (Å) 21.1333(9) 21.233(7) 12.8132(3) c (Å) 21.1333(9) 17.231(6) 14.6547(4) α (deg) 90 90 107.9400(10) β (deg) 114.339(3) 114.201(7) 92.2530(10) γ (deg) 90 90 96.0700(10) V (Å3) 6696.4(5) 6706.(4) 2113.77(9) space group P21/c P21/c P-1 Z, Z’ 4, 1 4, 1 2, 1

3 Dcalcd (g/cm ) 1.557 1.566 1.656 µ (cm–1) 0.475 0.582 0.698 R1 (I > 2σ (I))a 0.0631 0.0632 0.0435 wR2 (all data)a 0.1488 0.1783 0.1125

Table 7. Crystallographic data and refinement parameters for 30, 31, 32, and 33.

120

33 34 chemical formula ZnC36H23F6N5O10S4 PdC67H49N4O4P2S fw 1060.26 117.58 T (K) 120 120 λ (Å) 0.71073 0.71073 a (Å) 11.8116(4) 12.7522(4) b (Å) 12.8887(4) 13.4540(4) c (Å) 14.7358(5) 19.8902(6) α (deg) 108.043(2) 77.715(2) β (deg) 93.446(2) 73.152(2) γ (deg) 95.248(2) 67.151(2) V (Å3) 2114.78(12) 2990.27(17) space group P-1 P-1 Z, Z’ 2, 1 2,1

3 Dcalcd (g/cm ) 1.665 1.386 µ (cm–1) 0.873 0.455 R1 (I > 2σ (I))a 0.0358 0.0556 wR2 (all data)a 0.0921 0.1280

Table 8. Crystallographic data and refinement parameters for 33 and 34.

Synthesis of mesityl-naphthaleneimide anhydride (28). A 500 mL round bottom fitted with a vented additional funnel was charged naphthalene dianhydride (5.4 g, 20.1 mmol) and diluted with DMF (200 mL). The reaction was placed in a preheated oil bath and stirred for 30 minutes at 130 °C under N2. A solution of mesitylamine (1.5 g, 11.1

121 mmol) and DMF (100 mL) was added dropwise at a rate of ~3 sec/drop at 130 °C.

Following the addition the reaction stirred for 12 h at 130 °C and an additional 12 h at room temperature. The reaction solvent was removed in vacuo and the residue was extracted with EtOAc (100 mL) and passed over a plug of silica using EtOAc (~ 100 mL) as an eluent and concentrated to dryness. The crude product was subjected to flash column chromatography using 1:1 CH2Cl2 as the eluent system. 28 was isolated as a yellow

1 microcrystalline powder (1.6 g, 37 % yield). H NMR (400 MHz, CDCl3): δ 8.89 (d, 2H,

3 3 JHH = 7.6 Hz, naph-H), δ 8.86 (d, 2H, JHH = 7.6 Hz, naph-H), δ 7.26 (s, 2H, mes-H), δ

2.37 (s, 3H, mes-H), δ 2.09 (s, 6H, mes-H).

Synthesis of MNTP. A 100 mL round bottom flask fitted with a condenser was charged with 28 (500 mg, 2.82 mmol), 29 (980 mg, 2.54 mmol), and diluted with DMF (40 mL).

The reaction was placed in a preheated oil bath and stirred for 12 hours at 120 °C. The reaction was allowed to cool to room temperature and the reaction solvent was removed in vacuo. The residue was dissolved in a minimal amount of CH2Cl2 (~2 mL) and added dropwise to Et2O (~20 mL). The resulting powder was collected by filtration and washed with Et2O. The material was subjected to flash column chromatography and eluted with

8:2 CH2Cl2:EtOAc. MNTP was isolated as a yellow microcrystalline powder (Yield: 1.180

1 3 g, 85%). H NMR (400 MHz, CDCl3): δ 8.89 (s, 4H, naph-H), δ 8.66 (d, 1H, JHH = 4.9

3 3 Hz, py), δ 8.39 (s, 1H, thiazole), δ 8.09 (d, 1H, JHH = 8.0 Hz, py), δ 7.7 (dt, 1H, JHH = 8.8

4 3 Hz, JHH = 1.7 Hz, py), δ 7.27 (t, 1H, JHH = 4.8 Hz, py), δ 7.08 (s, 2H, mesityl), δ 2.38 (s,

3H, mesityl), δ 2.18 (s, 6H, mesityl).

122

Synthesis of Co(MNTP)2OTf2 (30). A solution of Co(OTf)2THF2 (90 mg, 0.179 mmol) in MeCN (3 mL) was added to a stirring solution of MNTP (200 mg, 0.367 mmol) in MeCN (3 mL). The reaction was stirred at room temperature for 12 hours under inert atmosphere. The reaction was concentrated to ¼ volume and Et2O (5 mL) was added resulting in precipitation of a yellow-orange solid. The solid was collected by filtration, washed with Et2O, and dried under reduced pressure to afford 30 (0.190 g, 83 % yield).

Crystals of 30 suitable for X-ray diffraction were grown via slow diffusion of Et2O into a

1 concentrated MeCN solution at room temperature. H NMR (400 MHz, CD3CN): δ 89.9

(br), 70.9 (br), 51.0 (br), 42.9 (br), 22.4 (br), 2.60 (br), 1.97 (br), -13.6 (br).

Synthesis of Zn(MNTP)2OTf2 (31). Zn(OTf)2 (33.5 mg, 0.092 mmol) and MNTP

(100 mg, 0.184 mmol) were stirred in MeCN (3 mL) at 80 °C until a homogenous solution was achieved. This solution was allowed to cool to room temperature and then filtered via

0.45 µm syringe filter into a clean 4 dram vial. The solution was layered with Et2O (3 mL) and left to stand for 12 hours. The resulting precipitate was washed with Et2O and dried under reduced pressure to afford 31 (0.140 mg, 71 %). Crystals of 31suitable for X-ray diffraction were grown via slow diffusion of Et2O into a concentrated MeCN solution at

1 3 room temperature. H NMR (400 MHz, CD3CN): δ 8.81 (d, 2H, JHH = 7.6 Hz, naph-H), δ

3 3 8.64 (d, 2H, JHH = 7.6 Hz, naph-H), δ 8.59 (s, 1H, thiazole), δ 8.50 (d, 1H, JHH = 5.3 Hz,

3 3 py), δ 7.53 (d, 1H, JHH = 8.1 Hz, py), δ 7.15 (s, 2H, mesityl), δ 7.11 (t, 1H, JHH = 7.9 Hz,

3 py), δ 6.81 (d, 1H, JHH = 6.4 Hz, py), δ 2.41 (s, 3H, mesityl), δ 2.32 (s, 6H, mesityl).

Synthesis of PTP. A 25 mL round bottom flask fitted with a condenser was charged with 28 (200 mg, 1.13 mmol), phthalic anhydride (184 mg, 1.24 mmol), and diluted with

123 acetic acid (10 mL). The reaction was place in a preheated oil path and stirred for 16 hours at 130 °C. The reaction was allowed to cool to room temperature and the reaction solvent was removed in vacuo. The residue was dissolved in water and extracted into CH2Cl2. The combined organic layers were washed with NaHCO3 until effervescence ceased. The organic layer was then washed with 1.0 M NaOH, H2O, and brine then dried over MgSO4 concentrated in vacuo to afford PTP as an off-white microcrystalline powder (Yield: 0.250

1 3 3 g, 71%). H NMR (400 MHz, CDCl3): δ 8.61 (d, 1H, JHH = 4.8 Hz, py), δ 8.19 (d, 1H, JHH

3 4 = 7.9 Hz, py), δ 8.10 (s, 1H, thiazole), δ 8.01 (dd, 2H, JHH = 5.5 Hz, JHH = 3.0 Hz,

3 4 phthalimide), δ 7.84 (dd, 2H, JHH = 5.5 Hz, JHH = 3.0 Hz, phthalimide), δ 7.77 (dt, 1H,

3 4 3 4 JHH = 7.7 Hz, JHH = 1.7 Hz, py), δ 7.23 (dd, 1H, JHH = 7.5 Hz, JHH = 1.6 Hz, py).

Synthesis of [Co(PTP)2]OTf2 (32). A solution of Co(OTf)2THF2 (82 mg, 0.165 mmol) in MeCN (3 mL) was added to a stirring suspension of PTP (100 mg, 0.325 mmol) in

MeCN (3 mL) affording a pale orange homogenous solution. The reaction was stirred at room temperature for 12 hours under inert atmosphere. The reaction was concentrated to

¼ volume and Et2O (5 mL) was added resulting in precipitation of a yellow-orange solid.

The solid was collected by filtration, washed with Et2O, and dried under reduced pressure to afford 32 (0.090 g, 72 % yield). Crystals of 32 suitable for X-ray diffraction were grown

1 via slow diffusion of Et2O into a concentrated MeCN solution at room temperature. H

NMR (400 MHz, CD3CN): δ 91.6 (br), 80.6 (br), 74.8 (br), 42.3 (br), 15.0 (br), 7.1 (br),

5.1 (br). µeff (CD3CN): 4.29 µB.

Synthesis of [Zn(PTP)2]OTf2 (33). Zn(OTf)2 (60 mg, 0.165 mmol) and PTP (100 mg,

0.325 mmol) were stirred in MeCN (3 mL) at 80 °C until a homogenous solution was

124 achieved. This solution was allowed to cool to room temperature and then filtered via 0.45

µm syringe filter into a clean 20 mL scintillation vial. The solution was layered with Et2O

(3 mL) and left to stand for 12 hours. The resulting precipitate was washed with Et2O and dried under reduced pressure to afford 33 (0.185 mg, 84 %). Crystals of 33 suitable for X- ray diffraction were grown via slow diffusion of Et2O into a concentrated MeCN solution

1 at room temperature. H NMR (400 MHz, CD3CN): δ 8.53 (s, 1H, thiazole), δ 8.34 (d, 1H,

3 3 4 JHH = 8.0 Hz, py), δ 8.17 (m, 3H, py), δ 7.94 (dd, 2H, , JHH = 5.5 Hz, JHH = 3.0 Hz,

3 4 3 phthalimide), δ 7.89 (dd, 2H, JHH = 5.5 Hz, JHH = 3.0 Hz, phthalimide), δ 7.51 (t, 2H, JHH

= 6.1 Hz, py).

2 Synthesis of h -MNTPPd(PPh3)2 OTf2 (34). Pd(PPh3)4 (106 mg, 0.092 mmol) dissolved in THF (~ 2 mL) was added dropwise to a stirring solution of MNTP (50 mg,

0.092 mmol) in THF (~2 mL). The reaction stirred for five minutes at room temperature quickly affording a magneta colored precipitate. The solid was washed with THF (3 × 5 mL) and dried in vacuo to furnish 34 as a magenta microcrystalline powder (90 mg, 83% yield). Crystals of 34 suitable for X-ray diffraction were grown via slow diffusion of pentane into a dilute reaction mixture (1:1 THF/MeCN, ~ 2mg/mL) at room temperature.

λmax = 356, 376, 446, and 547 nm.

125

Chapter 6: Synthesis and Catalytic Activity of a Pd PNNNP Pincer Complex Immobilized in a Zr Metal-Organic Framework6

6.1 Introduction

Transition metal complexes supported by arene-based diphosphine pincer ligands

(PZEZP, Figure 1) have been found to catalyze a wide range of organic transformations and small molecule activation reactions.85,86,95–104,87,105,88–94 Much of the appeal and success of

PZEZP pincer ligand platforms arises from their tunability and the stability conferred by chelation. We and others have been interested in incorporating transition metal PZEZP complexes into porous metal-organic frameworks (MOFs).111–113,276 In addition to offering ease of product separation and recyclability, heterogenization of homogeneous catalysts can potentially improve catalyst lifetime and activity via site isolation or other immobilization effects. MOFs have attracted considerable interest as heterogeneous catalyst platforms owing to their well-defined structures, inherent porosity, and amenability to functionalization. 79,277–282

We recently reported the synthesis and characterization of a Zr MOF, 37–PdX, assembled from linkers based on Pd POCOP pincer complexes.113 The MOF exhibits markedly better catalytic activity for transfer hydrogenation of aldehydes than an analogous homogeneous Pd POCOP complex, and the difference in activity has been attributed to inhibition of catalyst decomposition pathways in the MOF. Humphrey and coworkers have employed a similar design strategy to prepare a Co MOF containing Pd

6Reprinted in part from, “Zirconium Metal-Organic Frameworks Assembled from Pd and Pt PNNNP Pincer Complexes: Synthesis, Postsynthetic Modification, and Lewis Acid Catalysis.” Reiner, B. R.; Mucha, N. T.;Rothstein, A; Temme S. J.; Duan, P.; Schmidt-Rohr, K.; Foxman, B. M.; Wade. C. R. Inorg. Chem. 2018. 57, 2663–2672. Copyright 2018 American Chemical Society 126

C C – P C P linkers and shown that the material activates CO2 under mild conditions after Cl

– 111 /CH3 ligand exchange at the Pd pincer sites. Stoddart, Farha, and co-workers have used solvent-assisted ligand incorporation to immobilize Ir POCOP complexes within the mesoporous Zr MOF NU-1000, and the resulting material was found to catalyze the gas phase hydrogenation of ethylene.276

Figure 50. General structure of PZEZP pincer complexes

Although the identities of the central arene donor (E) and phosphine linker groups (Z) can have a profound impact on reactivity, PZEZP pincer complexes often exhibit similar molecular structures.92,283–285 This has led us to consider that an isostructural series of

MOFs might be obtained using PZEZP pincer metallolinkers with different arene donors, linker groups, or even chelated metal species. In order to explore this possibility and the scope of immobilization and site isolation effects, we have been investigating the assembly of Zr MOFs from linkers based on PNNNP pincer complexes. This chapter reports the synthesis of a porous Zr MOF, 38-PdX, constructed from a Pd PNNNP pincer complex. This material adopts the same cubic framework structure as 37–PdX, providing initial evidence that the isoreticular principle can be applied to pincer MOFs.

127

Historically, pore space manipulation in MOFs has aimed to tune host-guest

286 287 interactions between the MOF and a gas (e.g. CO2, hydrogen, ethylene) , water , or small unsaturated substrates for sieving or chromatography applications.68,288,289 However, there are very few reports that demonstrate pore selective catalysis where product distribution is governed by pore shape or size. Wang and coworkers have demonstrated pore selective Aldol oligomerizations using MIL-101 encapsulated phosphotungstic acid

(PTA) to catalyze the self-condensation of cyclic ketones.290 PTA molecules effectively restrict the accessible space within the MOF cages, leading to a higher selectivity for monocondensation products with increased PTA loadings in the MOF. Additionally, Smit and coworkers have developed a free energy relationship between the selectivity of propene dimerization and the pore size of nanoporous MOF catalysts.291 Linear selectivity increases in materials where pore size is commensurate with the size of the linear isomers. The formation of branched products is suppressed owing to steric hindrance.

Pore selective catalysis is distinct from immobilization or site isolation effects where suppression of bimolecular pathways leads to an alternate selectivity than that observed in solution. Moreover, divorcing pore size or shape effects from the inherent heterogeneous catalyst selectivity can be challenging without an accurate homogenous analog.

Prompted by the paucity of reports detailing pore selective catalytic reactions with

MOFs, we sought to explore whether an adequately activated pincer complex immobilized in a MOF lattice could be utilized for pore selective Lewis acid catalysis. Postsynthetic halide ligand exchange reactions at the Pd centers of 38-PdX were investigated as a means of activating the MOF for Lewis acid-catalyzed transformations. Silver salts are ubiquitous

128 and operationally simple halide abstraction agents but precipitation of insoluble AgX salts can potentially pollute a heterogeneous catalyst. As a result, halide abstraction becomes uniquely challenging in MOFs owing to the need for soluble byproducts that can be easily separated from the heterogeneous catalyst. We previously reported the use of PhI(TFA)2

– – (TFA = O2CCF3) for I /TFA ligand exchange in 37–PdX and considered this reagent to be

113 advantageous because it produces PhI and I2 as soluble byproducts. Treatment of 38–

PdX with NaI followed by PhI(O2CCF3)2 furnishes 38–PdTFA which shows superior catalytic activity to as synthesized material. However, 38–PdTFA suffers from poor recyclability and compositional analysis revealed I–/TFA– exchange was incomplete leading to a low density of catalytic sites. This issue has been eschewed by use of a stronger

– oxidant namely, NOBF4. Treatment of 38–PdX with NaI followed by NOBF4 completes I

– – – /BF4 exchange and affords 38–PdBF4. BF4 is a more non-coordinating anion than TFA leading to a markedly improved catalyst compared to 38–PdTFA. The catalytic activity of

38–PdBF4 was chronicled by the carbonyl-ene cyclization of citronellal and the intramolecular hydroamination of 2-(butyn-1-yl)aniline. The immobilized Lewis acidic

2+ Pd sites within 38–PdBF4 are well poised for pore selective catalysis. Indeed, 38–PdBF4 exhibits a high degree of selectivity for the intramolecular hydroamination of 2- ethynylaniline over a self-dimerization process. Comparison to a homogenous analog, t Bu4L-PdBF4, demonstrates this selectivity is not retained in solution. This result underscores the inherent value of leveraging the microenvironment of the MOF toward selectivity that cannot be enforced in solution.

129

6.2 Synthesis and characterization of Zr MOF assembled from a Pd PNNNP pincer complex

t t The ligand Bu4L was prepared by the P–N coupling of ClP(C6H4CO2 Bu)2 with 2,6- diaminopyridine in the presence of triethylamine (Figure 51).292 The Pd–Cl pincer

t t complex, Bu4L-PdCl, was obtained by reaction of Bu4L with an equimolar amount of

– – PdCl2(cod) s(cod = 1,5-cyclooctadiene) in CH2Cl2 solution. Subsequent Cl /I exchange

t 31 1 t with NaI generated the pincer complex Bu4L-PdI. The P{ H} NMR spectrum of Bu4L-

t PdCl exhibits a single resonance at 67.8 ppm while the signal for the Bu4L-PdI complex is shifted slightly downfield to 75.6 ppm. The 1H NMR spectra of these complexes display all expected resonances.

Figure 51. Synthesis of pincer complex H3L-PdI. Conditions: (i) NEt3 toluene, 80 °C, 16 h; (ii) PdCl2(cod), CH2Cl2, 16 h; (iii) NaI, acetone, 1.5 h; (iv) CF3CO2H, CH2Cl2, 16 h; (v) pyridine, acetone, 0.5 h

t Bu4L-PdI was deprotected by reaction with trifluoroacetic acid (HTFA) in CH2Cl2 solution. The crude product was then treated with pyridine (1.5 equiv.) in acetone solution resulting in elimination of one equiv. of HI and isolation of the zwitterionic complex H3L-

130

31 1 PdI. The P{ H} NMR spectrum of H3L-PdI exhibit one broadened resonance centered at 69.9 ppm. Single crystals of H3L-PdI were obtained from a saturated methanol solution, and the X-ray crystal structure supports deprotonation of one of the carboxylate groups concomitant with loss of the outer sphere I– to form the zwitterionic complex.

Solvothermal reaction of H3L-PdI with ZrCl4 in a 4/1 v/v mixture of N,N- dimethylformamide (DMF) and acetic acid (AcOH) generates 38–PdX as a bright yellow microcrystalline powder. Notably, attempts to use the deprotected chloride derivatives (i.e.

H3L-PdCl) for MOF synthesis resulted only in the formation of dark, amorphous solids. A similar trend was previously observed for the synthesis of Zr MOFs from H4[POCOP-PdX]

(X = Cl, I) complexes.113 The contrasting behavior of the Pd–Cl and Pd–I complexes likely reflects the difference in leaving group ability between Cl– and I– ligands.293 The more strongly bound I– ligands ostensibly act as protecting groups, suppressing ligand exchange and subsequent decomposition processes under the solvothermal reaction conditions.

The powder X-ray diffraction (PXRD) data show that 38–PdX is isostructural with 37-

PdX (Figure 52). Full-pattern decomposition of the data was performed using Pawley refinement and provided cubic unit cell parameters of a = 16.76 Å. 38–PdX is stable under ambient conditions, but experiences loss of crystallinity upon methanol solvent exchange followed by drying in vacuo. Optimized workup conditions include washing with DMF and acetone followed by soaking (ca. 16 h) in acetone prior to drying under reduced pressure. Following this procedure, a sample of 38–PdX was desolvated by heating under

−4 reduced pressure (10 Torr) at 150 °C for 16 h. The N2 adsorption measurement (77 K) gave a calculated Brunauer–Emmett–Teller (BET) surface area of 922 m2 g-1 which is

131 comparable to that previously observed for 37–PdX (1164 m2 g-1). Pore size distribution analyses for 37-PdX and 38–PdX were calculated using the non-local density functional theory (NLDFT) method.294–296 NLDFT models adsorption isotherms by taking into account both adsorbate-adsorbent interactions as well as adsorbate-adsorbate interactions that can become thermodynamically relevant in microporous materials (pore diameter < 2 nm). This model shows similar major pore distributions for 37-PdX and 38–PdX around

10-12 Å that are consistent with the proposed structure. However, 38–PdX exhibits a broad distribution of larger pores in the 14-20 Å range that could be attributable to the presence of missing linker defects (Figure 53).

Figure 52. (a) PXRD patterns of 37-PdX and 38-PdX (Cu Kα radiation, λ = 1.54 Å). (b) Defect-free framework structure of 38-PdX (left) and view of a portion of the framework 12+ showing ovoidal pores (right). Blue octahedra represent [Zr6O4(OH)4] building units.

132

Combustion analysis (C, H, N, Cl, I) and ICP-OES (Zr, Pd, Pt) were used to determine the composition of a desolvated sample of 38–PdX. ICP-OES provided Zr:Pd ratios of

2.9:1. This value reflects a Pd deficiency based on the formula expected from the

N N framework structure (Zr6O4(OH)4[MX(P N P)I]3) and is likely the result of missing linker-type defects (vide infra).73–76,297–301 The elemental analysis data also provided

Pd:I:Cl molar ratios of 1:0.82:0.88 indicating that a significant amount of Cl– is introduced from the use of ZrCl4 for the MOF synthesis. An overall 1:2 M:halide ratio is expected if the halides provide charge balance of the cationic pincer complexes. Consequently, the experimentally determined ratios reflect the partial absence of charge balancing outer sphere halides. We have not clearly identified the remaining charge-balancing species, but

302 it is possible that acetate or deprotonated Zr6 SBUs may fulfill this role.

Figure 53. DFT differential pore volume plots for 37-PdX (blue) and 38-PdX (black) obtained from the respective N2 adsorption isotherms (77 K). 133

Figure 54. Solid-state 31P NMR spectrum of 38-PdX with magic-angle spinning (MAS) and total suppression of spinning sidebands (TOSS).303

38–PdX has been characterized by solid and solution state NMR spectroscopy as well as electrospray mass spectrometry (ESI-MS) in order to gain further insight into its structure and composition. The MAS 31P NMR spectrum of 38–PdX shows a broad, asymmetric signal centered at 72 ppm (Figure 54). The spectrum also show the presence of a minor species giving rise to a signal around 25 ppm. The chemical shift of the minor species is in line with that observed for the PNNNP pincer ligand in solution, but could also arise from a small amount of ligand decomposition.113 Although ligand decomposition is likely to result in the presence of oxidized phosphine species, no P=O stretching bands could be clearly identified in the ATR–IR spectrum.

A sample of 38–PdX was digested with a 3:1 v:v mixture of trifluoroacetic acid (HTFA)

31 1 and C6D6, and the resulting solution was analyzed by P and H NMR spectroscopy and

ESI-MS. The 31P{1H} NMR spectrum exhibit two major signals appearing at 74.0 and 69.9

134

(Figure 57). The signals appear in ~3:1 ratio, which was reproducible over several samples.

The 1H NMR spectra of the digested sample also clearly show two sets of resonances consistent with the presence of two distinct pincer complexes. The solid and solution state

NMR data for the MOF closely resembles that obtained for the H3L-PdCl and H3L-PdI complexes in solution, indicating that the pincer complex linker remains intact both within the MOF and in solution after acid digestion. ESI-MS analysis confirmed the identity of the species present in the digested solution as the Pd–I and Pd–Cl pincer complexes. The mass spectrum of 38–PdX shows two major parent ions with m/z = 795 and 886 amu that

+ + are assigned to H4[L-PdCl] and H4[L-PdI] , respectively. Together, this data supports mixed occupancy of the palladium coordinated halide.

In addition to the resonances assigned to the pincer complexes, the 1H NMR spectrum of the digested sample of 38–PdX displays signals attributable to acetic acid (HOAc) at

1.86 ppm. Since desolvated samples were used for digestion, HOAc is not likely to be present as a pore-occluding guest molecule, but rather incorporated into the framework structure as OAc–. Assignment of framework bound acetate is further supported by a quantitative solid-state 13C NMR spectrum that shows signals consistent with acetate groups at 22 and 179 ppm (Figure 55). The CH3 resonance at 22 ppm exhibits significant line broadening and partial dipolar dephasing that is in contrast to the sharp signal of a highly mobile species, assigned to residual acetone, at 30 ppm. These characteristics confirm the solid-like behavior of the OAc– species in the MOF. Integration of the solution-

1 13 + state H and solid-state C NMR spectra provides H4[L-PdX] : OAc ratios of ~1:1. This data combined with the Zr:Pd ratios determined by ICP-OES analysis suggest that the

135 empirical formulas of 38-PdX is best given as {Zr6O4(OH)4(OAc)2.4[L-PdX]2.4}Y2.4. The presence of OAc– or other modulator-derived anions often signals the presence of missing linker defects in Zr MOFs.76,297–300 It seems likely that 38-PdX retains the same framework structure as 37-PdX, but contain a larger number of disordered missing linker defects.

However, given the level of uncertainty associated with structure determination from powder X-ray diffraction data, we cannot rule out the possibility of alternate framework structures. Efforts to obtain samples suitable for single crystal X-ray diffraction have not yet been successful.

Figure 55. Solid-state 13C NMR spectrum of 38-PdX. The spectrum was recorded with ComPmultiCP at 14 kHz magic-angle spinning (MAS). Thick black line: Spectrum of all carbon; thin red line: spectrum of carbon not bonded to hydrogen, or with large-amplitude segmental motions. “ssb” spinning sideband.

136

6.3 Postsynthetic halide exchange reactions

Exchange of halide ligands for more weakly coordinating anions is often necessary to activate homogeneous organometallic complexes for catalysis. We previously observed

– – that treatment of 37–PdX with PhI(TFA)2 facilitates I /TFA ligand exchange, activating the MOF for transfer hydrogenation catalysis.113 This halide exchange strategy is advantageous because it produces PhI and I2 as soluble byproducts that can be easily separated from the MOF. Consequently, we sought to determine if a similar approach could be used for halide ligand exchange in 38–PdX.

Figure 56. Postsynthetic halide ligand exchange reactions of 38-PdX.

137

In our initial screening, 38–PdX was treated with a solution of PhI(TFA)2 in MeCN and washed copiously with MeCN (Figure 56). 31P{1H} NMR analysis of the acid-digested

+ product indicated that the H4[L-PdCl] pincer complex was the major species in solution

(Figure 57). This result is consistent with oxidative exchange of I– for TFA– followed by coordination of the outer sphere Cl– ions that remain present in 38–PdX. In order to circumvent the formation of Pd–Cl species, we subjected 38–PdX to a Cl–/I– ligand exchange reaction prior to treatment with PhI(TFA)2. Accordingly, 38–PdI was generated by treating 38–PdX with an aqueous solution of NaI (Figure 56). The 31P{1H} NMR spectrum of an acid-digested sample of 38–PdI exhibits a single major resonance at 74.0 ppm, and the 1H NMR spectrum displays a single set of well-resolved signals (Figure 57).

The spectra match those expected for the Pd–I pincer complex and indicate quantitative

Cl–/I– exchange. Moreover, the ESI-MS spectrum of the digested product contains a major

+ signal corresponding to the H4[L-PdI] parent ion at m/z = 886. Next, 38–PdI was treated with a CH2Cl2 solution of PhI(TFA)2 (4 equiv. per Pd), and after 24 h the supernatant

31 1 solution had turned a pink color, signaling the formation of I2 (Figure 56). The P{ H}

NMR spectrum of an acid-digested sample of the resulting material, 38-PdTFA, features two singlets centered at 74.0 and 71.1 ppm that appear in a ~1:1 ratio (Figure 57). The

+ downfield signal is consistent with the presence of unreacted H4[L-PdI] species, while the chemical shift of the new resonance at 71.1 ppm closely matches that observed for the

t analogous homogeneous complex, Bu4L-PdTFA, in the same solvent mixture.

Consequently, the data indicates ~50 % conversion of the Pd–I sites to Pd–TFA. Efforts to increase conversion by resubjecting the product to PhI(TFA)2/CH2Cl2 solutions were

138 largely unsuccessful. After the second treatment, only a faint color change was observed for the reaction supernatant and 31P NMR analysis showed no significant changes in the

Pd–I:Pd–TFA product ratio.

Figure 57. 31P{1H} NMR spectra of digested samples of 38–PdX, 38–PdI, 38–PdX + PhI(TFA)2, 38–PdTFA, and 38–PdOTf/AgOTf.

The precipitation of highly insoluble AgX byproducts makes silver-based reagents convenient for halide abstraction reactions involving soluble organometallic complexes.

However, these reagents are potentially problematic for use with MOFs since the solid

AgX cannot be easily separated. Considering the ubiquity of silver-based precatalyst activation, we wanted to investigate the effectiveness of silver salts for halide exchange in

139

– – 38–PdX. A suspension of 38–PdX in MeCN was treated with AgOTf (OTf = O3SCF3 ) and heated at 60 °C for 12 h to generate 38–PdX/AgOTf. PXRD analysis confirms that

38–PdX/AgOTf retains crystallinity after the reaction. However, the pattern also contains peaks corresponding to crystalline AgCl and AgI (Figure 58), and these byproducts could not be removed by washing with common organic solvents. The 31P{1H} NMR spectrum of an acid-digested sample of 38–PdX/AgOTf (CF3CO2H:C6D6, 3:1 v:v) exhibits two major resonances at 69.9 and 74.0 ppm (Figure 57). These resonances closely resemble those expected for the Pd–Cl and Pd–I species, albeit appearing in a different ratio (Pd–

t Cl:Pd–I ≈ 5.7:1) than was observed for 38–PdX (~1:3). The homogenous complex Bu4L-

t PdOTf was prepared by reaction of Bu4L-PdCl with AgOTf in MeCN solution and

31 1 observed to give rise to a P{ H} NMR signal at 75.5 ppm in the CF3CO2H:C6D6 solvent mixture. Thus, the 31P{1H} NMR spectrum of the digested sample of 38–PdX/AgOTf does not reflect the presence of Pd–OTf species in solution. Nevertheless, the appearance of crystalline AgI and AgCl in the PXRD patterns of 38–PdX/AgOTf indicates that AgOTf facilitates some degree of X–/OTf– halide exchange, but likely only with the outer sphere halide ions.

140

Figure 58. PXRD patterns of 38–PdX, 38–PdI, 38–PdTFA, and 38–PdX/AgOTf (Cu Kα, λ = 1.54 Å).

6.4 Initial catalytic studies

Indoles are common motifs in natural products and pharmaceuticals.304,305 They can be efficiently assembled using a variety of strategies including Lewis acid-catalyzed intramolecular cyclization of o-alkynyl substituted anilines.306 Moreover, Pd and Pt PNP pincer complexes have found success as Lewis acid catalysts for intramolecular hydrofunctionalization reactions.307–314 In most cases, these catalysts are activated by

141 exchanging halide ligands for more weakly coordinating anions. These considerations prompted us to compare the catalytic activity of the Pd PNNNP MOFs activated with

PhI(TFA)2 and AgOTf for intramolecular hydroamination reactions. 2-(butyn-1-yl)aniline

(39) was chosen as a benchmark substrate, and catalytic reactions were carried out in 1,4- dioxane at 95 °C with 5 mol % catalyst based on Pd (Table 9). Product yields were determined by integration of the 1H NMR spectra with respect to an internal standard.

Under the catalytic conditions, 38–PdX and 38–PdI afforded 2-ethylindole 40 in 77 % and 40 % yield respectively (entries 1 and 2) after 12 h. 38–PdX exhibits remarkably good catalytic activity despite the presence of I– and Cl– ligands. The lower catalytic activity observed for 38–PdI is consistent with the presence of only the less labile I– ligands. To our surprise, 38–PdX/AgOTf furnished 40 in slightly lower yield (69 %) than 38–PdX.

For comparison, a sample of 38–PdI/AgOTf was prepared by treating 38–PdI with AgOTf and found to deliver 40 in 35 % yield under the same catalytic conditions. The similar catalytic activities of 38–PdX/AgOTf and 38–PdI/AgOTf to the parent MOFs 38–PdX and 38–PdI provide further support that AgOTf is not effective for X–/OTf– exchange at the Pd–X sites within these materials. Thus, formation of the AgX species observed by

PXRD analysis should be the result of exchange of only the outer sphere halide counter ions. Moreover, the slight decrease in catalytic activity of the AgOTf-treated MOFs compared to 38–PdX and 38–PdI may be attributed to pore occlusion by the insoluble

AgX species. 38–PdTFA proved to be the most active of the MOF catalysts, generating indole 40 in 93 % yield (entry 5). The large difference in catalytic activity between 38–

PdI/AgOTf and 38–PdTFA clearly illustrates the superiority of PhITFA2 as an activating

142 reagent. However, 38–PdTFA exhibits only modestly better activity than 38–PdX and significantly diminished activity upon attempted recycling (entry 6). Trace amounts of trifluoroacetamide 41 were also observed in the catalytic reactions employing 38–PdTFA.

The consumption of TFA anions and a proton of unknown origin in this off-cycle reaction pathway may lead to catalyst deactivation and account for the modest activity and poor recyclability of 38–PdTFA. This notion is further supported by a marked difference in

t t catalytic activity between the homogeneous complexes Bu4L-PdOTf and Bu4L-PdTFA

(vide infra).

Entry Catalyst % Yield 40a % Yield 41a TONb 1 38–PdX 77 – 19 2 38–PdI 40 – 10 3 38–PdX/AgOTf 69 – 17 4 38–PdI/AgOTf 35 – 9 5 38–PdTFA 93 trace 23 6 38–PdTFA (run 2) 48 – 12

7 38–PdPMe3 <5 – <1 t c 8 Bu4L–PdOTf 94 – 19 t 9 Bu4L–PdTFA 52 5 10

Table 9. Hydroamination of o-alkynyl aniline 39. Reaction conditions: substrate (0.1 mmol), catalyst (0.005 mmol Pd), 1,4-dioxane, 95 °C, 12 h. aDetermined by 1H NMR with respect to an internal standard (hexamethylbenzene). bTurnover numbers (TON) were calculated per Pd using the empirical formula for 38–PdX that accounts for missing linker defects. cReaction time was 1 h.

143

31P 1 Figure 59. { H} NMR spectrum of an acid-digested (3/1 CF3CO2H/C6D6) sample of 38– + PdPMe3. The resonance at δ -2.94 ppm is assigned to [HPMe3] .

Given the evidence for a large number of defect sites in 38–PdX, we sought to confirm that the Pd pincer sites are responsible for the observed catalysis. We hypothesized that the strongly donating L-type ligand PMe3 should selectively coordinate and block access to the Pd sites, shutting down the contribution of these sites to the observed catalysis.

Consequently, 38–PdPMe3 was synthesized by soaking 38–PdX in a MeCN solution of

31 1 PMe3 (Figure 56). After washing, the P{ H} NMR spectrum of an acid-digested sample

1 features a doublet and triplet resonance in a 2:1 ratio centered at 78.1 and d –9.7 ppm ( J31P–

2+ P = 25.7 Hz), respectively, that are assigned to H4[L-Pd(PMe3)] species (Figure 59).

+ + Signals attributed to H4[L-PdI] (74.0 ppm) and [HPMe3] (–2.90 ppm) also appear in a

~1:1 ratio. The presence of the latter species indicates either incomplete conversion (~ 75 144

%) of the Pd–I sites to Pd–PMe3 or PMe3 ligand dissociation induced by the strongly acidic digestion procedure. Regardless, 38–PdPMe3 delivered only trace amounts of 40 under the catalytic conditions (entry 7), substantiating that the Pd pincer sites are primarily responsible for catalysis. Similarly, catalytic reactions carried out in the presence of UiO–

67 or absence of any catalyst showed no appreciable formation of the desired indole product.

t t The catalytic activities of the homogeneous complexes Bu4L-PdOTf and Bu4L-PdTFA

t were also investigated (Table 9, entries 8 and 9). Bu4L-PdOTf was markedly superior to

t 38–PdTFA, generating indole 40 in 94 % yield after 1 h while Bu4L-PdTFA furnished

40 in only 52 % yield after 12 h. Similar to 38–PdTFA, a stoichiometric amount (~5 %)

t of trifluoroacetamide 41 was identified as a side product in the reaction employing Bu4L-

PdTFA. Thus, we surmise that consumption of TFA– via the formation of 41 leads to

t catalyst deactivation and is responsible for the poor activity of Bu4L-PdTFA and lack of recyclability of 38–PdTFA. Further studies will be necessary to elucidate the mechanism of catalyst deactivation, but we hypothesize that the process stems from the greater basicity of TFA– versus OTf– and may involve deprotonation of the NH linker groups of the PNNNP ligand. This hypothesis would suggest that a more active and recyclable pincer MOF catalyst could be obtained by oxidative halide ligand exchange with a less basic anion such as OTf–. However, attempts to carry out oxidative I–/OTf– exchange using in situ-generated

PhI(OTf)2 have been unsuccessful thus far, resulting in materials that suffer significant losses in crystallinity and exhibit poor catalytic activity. We believe this to be a limitation

145 of the ill-defined nature of PhI(OTf)2, which is usually generated in situ from

315 PhI(O2CCH3)2 and Me3SiOTf.

6.5 Activation with NOBF4

The scarcity of iodinanes supported by weakly coordinating anions prompted us to reconsider the identity of the oxidant used for oxidative ligand exchange in 38–PdI.

Careful examination of the literature revealed an abundance of oxidants that have sufficient thermodynamic driving force for iodide oxidation.316 However, attempts to use redox agents with seemingly sufficient oxidative potential such as acetyl ferrocenium salts (~200

– – mV positive of I /I2 couple) were met with limited success, showing poor I exchange at

Pd. Thus, we considered oxidation of Pd–I species might require more forcing conditions than oxidation of free iodide.

The redox behavior of homogenous analogs, tBuL–PdCl and tBuL–PdI, was evaluated by cyclic voltammetry (CV) in MeCN solution to identify the requisite potential for iodide oxidation (Figure 60). tBuL–PdCl exhibits one irreversible reduction at –1.36 V and one irreversible oxidation at 0.83 V while the CV of tBuL–PdI features two ill-defined irreversible reductive features at –0.94 V and –1.24 V and two quasireversible oxidations at 0.04 V and 0.34 V.

The ill-defined character of the reductive features is likely due to halide labilization upon reduction of the complex. The oxidative features at 0.04 and 0.34 V are assigned to outer sphere and inner sphere iodide oxidation respectively. These assignments are supported by the absence of similar redox features in the voltammogram of tBuL–PdCl.

– Notably, the inner sphere I /I2 redox couple is 480 mV positive of the reported potential for

146

– 316 the free I /I2 couple in the same solvent which helps explain the apparent recalcitrance of the anion toward oxidation.

Figure 60. Cyclic voltammograms (CV) of tBuL–PdCl (red), tBuL–PdI (blue), and n NOBF4 (green). Collected in 0.1 M [ Bu4N][PF6] in MeCN (scan rate = 100 mV/s). All potentials are referenced to the ferrocene (Fc)/ferrocenium (Fc+) redox couple.

Based on this data, NOBF4 (E1/2 = 0.870 V vs ferrocene) was identified as a potentially

– – – suitable oxidant for I / BF4 oxidative ligand exchange of the Pd–I species. Moreover, BF4 is one of the smallest weakly coordinating anions with a crystallographic volume of 53

3 317 – – 3 3 Å . For comparison, OTf and PF6 occupy 85 Å and 75 Å respectively. Smaller counter anions are advantageous in MOF based materials where congested pores can preclude efficient ingress and egress of substrates to and from catalytic centers.

147

Figure 61. Synthesis of 38–PdBF4.

Figure 62. XRF spectrum of 38–PdI after 0 (black), 1 (blue), or 2 (red) soaks with NOBF4 in MeCN solution.

38–PdBF4 was synthesized by successive treatment of 38–PdI with MeCN solutions of NOBF4 (2 equiv. per Pd, Figure 61). After the first soak, the supernatant turned a bright orange color, signaling the formation of INOx species. Analysis of the isolated solid by X- 148 ray fluorescence (XRF) showed a substantial decrease in the signal arising from the I Ka1 emission line. A pale orange supernatant resulted following the second soak and XRF analysis of the isolated solid showed nearly complete disappearance of the signal associated with I (Figure 62). PXRD analysis confirms 38–PdBF4 retains bulk crystallinity after the reaction (Figure 63a). Moreover, N2 adsorption measurements (77 K) revealed a

Type I Langmuir isotherm and gave a calculated Brunauer−Emmett−Teller (BET) surface

2 -1 area of 733 m g demonstrating preservation of microporosity (Figure 63b). 38–PdBF4

2 -1 exhibits a smaller N2 accessible surface area than 38–PdI (922 m g ) which is consistent

– 3 – 3 with substitution of I (35 Å ) for the larger BF4 (53 Å ).

Figure 63. (a) PXRD patterns of 38-PdI and 38-PdBF4. Data was collected with Cu Kα radiation. (b) N2 adsorption isotherms (77 K) for 38-PdI (black) and 38-PdBF4 (red) after desolvation at 150 °C and ~10-4 torr for 12 h.

A sample of 38–PdBF4 was digested with a 3:1 v:v mixture of trifluoroacetic acid

1 31 (HTFA) and C6D6, and the resulting solutions were analyzed by H and P NMR 149 spectroscopy. The 31P{1H} NMR spectrum features two singlets centered at 76.2 and 70.2 ppm that appear in a ~3.5:1 ratio as well as two minor resonances (<5 %) at 75.2 and 74.5 ppm that appear in a 1:1 ratio (Figure 64). Addition of water results in disappearance of the downfield signal at 76.2 ppm with a concomitant increase of the intensity of the signal at 70.2 ppm. The minor resonances are unaffected. Accordingly, the signal at 76.2 ppm is

+ assigned to H4[L-PdBF4] species, while the signal at 70.2 ppm is attributed to H4[L-

+ + PdOH] resulting from hydrolysis of H4[L-PdBF4] under digestion conditions. The minor resonances are attributed to pincer species resulting from partial iodination of the pyridyl backbone. The corresponding 1H NMR spectra are consistent with these assignments. The

31 1 t P{ H} NMR spectrum of the analogous homogeneous complex Bu4L-PdBF4, synthesized via NOBF4, also displays two peaks in the HTFA:C6D6 solvent mixture resonating at 76.0 and 71.0 ppm. However, the same complex features only one 31P resonance at 72.3 ppm when the spectrum is collected in CH2Cl2. This further supports the proclivity of Pd-BF4 species to hydrolysis under the highly acidic digestion conditions.

150

31 1 Figure 64. P{ H} NMR spectra of digested samples of 38–PdBF4, 38–PdTFA, and 38– PdI.

– Oxidation of I with NOBF4 produces NO which can sequester metal coordination sites and hinder catalysis. However, the IR spectrum of 38–PdBF4 does not feature any stretches attributable to metal nitrosyls. A stretch at 1222 cm-1 that is not present in the parent MOF can be assigned to a B–F stretching band (Figure 65). Overall, the collected data indicates full conversion of the Pd–I sites to Pd–BF4 without any substantial loss in structural integrity.

151

Figure 65. ATR-IR spectrum of 38–PdI (black) and 38–PdBF4 (red).

6.6 Catalytic activity and kinetics of 38–PdBF4.

– – We hypothesized substitution of TFA for a less basic anion such as BF4 (pKaHBF4 = -

318 7.7, pKaHTFA = -0.25) would afford a more catalytically robust material. With 38–PdBF4 in hand we sought to evaluate its catalytic activity. The intramolecular cyclization of 2-

(butyn-1-yl)aniline (39) was chosen as a benchmark reaction based on previous studies.114

Catalytic reactions were carried out in 1,4-dioxane at 95 °C with 5 mol % catalyst based on Pd (Table 10). Product yields were determined by integration of the 1H NMR spectra with respect to an internal standard.

Under the catalytic conditions, 38–PdI and 38–PdX delivered 2-ethylindole 40 in 27% and 41% yield respectively after 4 hours (entries 5 and 6), which is consistent with 152 previously observed trends.114 38–PdTFA afforded 40 in only 54% yield (entry 4). 38–

PdBF4 proved to be the best catalyst in the series, furnishing 40 in 99% yield (entry 1).

38–PdBF4 could also be recycled up to five times without any substantial decrease in activity (entry 2). Notably, the catalyst remained competent with loadings as low as 0.5 mol % delivering 40 in 84% yield (entry 3). We attribute the considerable increase in catalytic activity of 38–PdBF4 over 38–PdTFA to two factors: an increase in the density

– – of catalytic sites owing to complete I /BF4 exchange and suppression of a basic anion mediated catalyst deactivation pathway.

t The catalytic activity of the homogeneous complex Bu4L-PdBF4 was also investigated. The complex was synthesized via oxidative halide exchange with NOBF4 to

+ t eschew any complications associated with adventitious Ag based co-catalysis. Bu4L-

PdBF4 was markedly superior to 38–PdBF4, generating indole 40 in 99% yield after 4 h at room temperature (entry 7). 38–PdBF4 however, exhibited no catalytic activity at this temperature (entry 8).

153

Entry Catalyst % Yield 40a TONb Temp. (°C)

1 38–PdBF4 99 20 95

2 38–PdBF4 (run 5) 92 18 95 c 3 38–PdBF4 84 168 95 4 38–PdTFA 54 11 95 5 38–PdX 41 8 95 6 38–PdI 27 5 95 t 7 Bu4L–PdBF4 99 20 25

8 38–PdBF4 0 0 25

Table 10. Hydroamination of o-alkynyl aniline 39. Reaction conditions: substrate (0.1 mmol), catalyst (0.005 mmol Pd), 1,4-dioxane, 95 °C, 4 h. aDetermined by 1H NMR with respect to an internal standard (hexamethylbenzene). bTurnover numbers (TON) were calculated per Pd using the empirical formula for 38–PdX that accounts for missing linker defects. cReaction conducted with 0.5 mol % Pd for 12 hours.

In order to better understand the large discrepancy in activity between 38–PdBF4 and t Bu4L-PdBF4, the cyclization of 39 was monitored by GC-FID in the presence of each catalyst.

The reaction catalyzed by 38–PdBF4 features two distinct kinetic regimes (Figure

66a). Initially, the reaction proceeds at a relative fast reaction rate followed by a regime that appears zero order in substrate. This is corroborated by a high linear correlation between the substrate concentration ([39]) and time in that region. This unusual behavior is reproducible over multiple runs. Morris and coworkers report a similar kinetic 154 phenomena when monitoring the hydrolysis of a chemical warfare simulant by UiO-66.319

Reaction profiles required two term exponential models to reach adequate fits. The extracted rate constants differed by two orders of magnitude which was ascribed to the varying accessibility of catalytic Zr sites as the reaction progressed. The observed kinetic profile collected for the consumption of 39 was treated similarly and a two term exponential model led to satisfactory fit statistics. The extracted observed reaction rate

-1 -1 constants (kobs), 0.017 min and 0.25 min , differ by an order of magnitude. We attribute the large difference to the reactivity of catalyst sites near the surface of the MOF crystallites versus internal sites. MOF surface followed by reaction at internal sites. Internal sites are subject to mass transport limitations as a function of both substrate diffusion and adsorption/desorption processes. However, substrate flux at the material surface is likely not kinetically relevant. Even a two term exponential fit does not accurately model the phenomenological reaction behavior. Analysis is likely complicated by the varying accessibility of the Pd pincer sites moving from the surface of the crystallites into the bulk.

Unfortunately, measuring reaction rate as a function of crystallite size is difficult since 38–

PdX has only been obtained as a fine microcrystalline powder that is not amenable to grinding. Nonetheless, identification of mass transport limitations, is consistent with reactions occurring, at least in part, within the pores of 38–PdBF4.

t The reaction catalyzed by Bu4L-PdBF4 displays the expected exponential decay from pseudo-first order kinetics (Figure 66b). However, a plot of ln[39] vs time shows clear deviation from linearity. We attribute this behavior to decomposition of the catalyst over time, which is supported by the evolution of multiple species by 31P NMR spectroscopy as

155 well as the formation of metal mirrors on reaction vessels following catalytic reactions.

t Decomposition of Bu4L-PdBF4 likely leads to the formation of additional kinetically relevant catalytic species, which convolute a simplistic analysis. Nonetheless, diffusion- limited behavior is not evident in the observed kinetics profile for reactions catalyzed by t Bu4L-PdBF4.

Figure 66. Reaction profile for the cyclization of 2 in the presence of a) 38-PdBF4 or b) t Bu4L-PdBF4

156

t Overall, the cyclization of 39 catalyzed by 38–PdBF4 or Bu4L-PdBF4 display gross

differences in the observed kinetic behavior. The reaction catalyzed by 38–PdBF4 displays a region that appears zero order in substrate indicative of mass transport limitations.

t However, reactions catalyzed by Bu4L-PdBF4 do not display this behavior, which suggests that Pd sites within the pores of 38–PdBF4 are kinetically relevant for catalysis. A more thorough investigation will be required to interrogate the finer mechanistic details of reactions catalyzed by 38–PdBF4.

6.7 Catalytic pore selectivity

We considered the mass transport limitations inherent to 38–PdBF4 could be leveraged toward pore selective transformations. We hypothesized intramolecular reactions could be favored over intermolecular reactions owing to slow diffusion kinetics as well as the limited pore size available for reactions occurring within the cages of 38–PdBF4.

t Moreover, the homogenous analog, BuL–PdBF4, is not constrained by these limitations and would likely exhibit little control over the reaction. 2-ethynyl aniline, 42, was chosen as a model substrate. 42 can undergo an intramolecular hydroamination to form indole 43 or a tandem intermolecular hydroamination–annulation sequence to afford quinoline

320–322 44. Pore size distribution analyses for 38-PdBF4 using the NLDFT method show major pore distributions around 10-12 Å (Figure 67). While 42 should be easily accommodated within 38–PdBF4 owing to a relatively small kinetic diameter (~7 Å), arrangement of two molecules of 42 within the pores of 38–PdBF4 to afford quinoline 44

t would seem unlikely. Initial attempts to catalyze the cyclization of 42 with BuL–PdTFA or 38-PdTFA were unsuccessful. Only stoichiometric amounts of trifluoroacetamide 45

157

(~10 %) were observed which is derived from capture of adventitious trifluoroacetic acid

– (Table 11, entries 2 and 4). We expected pincer complexes supported by the less basic BF4 anion would not undergo this deleterious side process. Catalytic reactions were carried out in 1,4-dioxane at 95 °C with 5 mol % catalyst based on Pd (Table 11). Product yields were determined by GC-FID.

Figure 67. DFT differential pore volume plots for 38-PdBF4 obtained from the respective N2 adsorption isotherms (77 K).

t After 24 hours under the catalytic conditions, BuL–PdBF4 promoted quantitative conversion of 42 delivering a ~1:1 mixture of indole 43 and quinoline 44. Monitoring the reaction over time showed the selectivity did not change substantially as a function of conversion. The conversion of ethynyl aniline 42 was substantially slower than butynyl aniline 39 likely owing to substrate inhibition associated with the less sterically congested 158 aniline. Nonetheless, after 60 h, 38–PdBF4 furnished 43 in 65% yield (entry 1) with no detectable amounts of 44. A small amount of aminoacetophenone (<5 %) was also observed as a consequence of adventitious hydration of the starting material. Observation of 43 as the sole major product demonstrates a rare example of pore selective reactivity wherein immobilization of a catalytic species within a MOF lattice substantially enhances reaction selectivity. Moreover, this result illustrates that the microenvironment of the MOF can be used to engender unique selectivity that is not easily enforced in solution.

% % Selectivity entry catalyst conversion yield 43a (43:44:45)

1 1–PdBF4 70 65 65:0:0 2 1–PdTFA 10 0 0:0:10 t b 3 Bu4L–PdBF4 100 45 45:55:0 t 4 Bu4L–PdTFA 10 0 0:0:10 Table 11. Cyclization of o-ethynyl aniline 42. Reaction conditions: substrate (0.1 mmol), catalyst (0.005 mmol Pd), 1,4-dioxane, 95 °C, 60 h. aDetermined by GC–FID with respect to an internal standard (hexamethylbenzene). bReaction time was 24 h.

6.8 Catalytic carbonyl-ene cyclization with citronellal

The successful synthesis of 38–PdBF4 which features substrate accessible non- framework Lewis acidic sites prompted us to compare the catalytic activity to other known

159 heterogeneous Lewis acids. The carbonyl-ene cyclization of citronellal, 46, is frequently used as a benchmark for measuring the Lewis acidity in heterogeneous platforms.323–327

The diastereoselective cyclization of citronellal to isopulegol, 47, is a key reaction in the synthesis of menthol, a naturally occurring extract of mint leaves with valuable fragrance and pharmaceutical properties. Moreover, recyclable heterogeneous catalysts are attractive substitutes for ZnBr2 which acts as a stoichiometric promoter in the industrial process but is considered a severe marine toxin. Catalytic reactions were carried out in toluene at 100

°C with 10 mol % catalyst based on Pd (Table 12). Product yields were determined by GC-

FID.

After 30 minutes under the catalytic conditions, 38–PdBF4 promoted quantitative conversion of citronellal and delivered isopulegol in 64% yield (entry 1). Only pulegol isomers were observed and selectivity toward the isopulegol diastereomer, 47, was 67%.

Selectivity did not improve when conducting the reaction at lower temperatures. Other

323,326 324 MOF catalysts including Cu3BTC2 and UiO-66 have shown comparable selectivity. 38–PdBF4 was recycled three times without any loss in activity (entry 2), supporting immobilized Pd2+ sites as the actuating catalyst. Adventitious catalytic species such as homogenous Pd or proton sources would be expected to be washed away as a consequence of the recycling process. Notably, the parent MOF, 38–PdX, shows limited activity for this reaction affording 9 in only 16% yield (entry 4).

The catalytic activity of 38–PdBF4 for the carbonyl-ene cyclization was further evaluated by kinetic analysis. The reaction proceeded too quickly at 10 mol % Pd, so kinetic measurements were performed at 0.5 mol % Pd. At 100 mM substrate concentration

160 the full reaction profile was fit satisfactorily with a single exponential component.

However, at 200 mM substrate concentration, the kinetic profile could only be adequately fit using a two term exponential model (Figure 68). At high substrate to catalyst loadings, internal sites are likely engaged earlier in the reaction, leading to complex reaction kinetics.

The change in reaction behavior upon doubling the substrate concentration underlines the inherent complexity in modeling the kinetics of heterogeneous porous materials. Substrate concentration, mass transport processes, and the intrinsic catalytic activity of each active site must be considered for an accurate kinetic model.

Figure 68. Reaction profile for the cyclization of citronellal at 100 or 200 mM in the presence of 38–PdBF4.

161

% % entry catalyst TONc conversion 46 yield 47a

1 38–PdBF4 100 64 10

2 38–PdBF4 (run 3) 100 63 10 3 38–PdX 17 16 2 b 4 38–PdBF4 96 60 192 t 5 Bu4L–PdBF4 100 73 10 t b 6 Bu4L–PdBF4 71 51 140 7 None 0 0 0

Table 12. Cyclization of citronellal (46) to isopulegol (47). Reaction conditions: substrate (0.05 mmol), catalyst (0.005 mmol Pd), toluene, 100 °C, 3 h. bDetermined by GC–FID with respect to an internal standard (hexamethylbenzene). cReaction conducted with 0.5 mol % for 3 hours. dTurnover numbers (TON) were calculated per Pd using the empirical formula for 38–PdX that accounts for missing linker defects. TON refer to total amount of pulegol isomers formed.

Reactions run at 200 mM have a qualitatively higher initial rate than reactions run at

100 mM, however sampling granularity led to large errors in extracting an initial TOF.

Nonetheless, a conservative estimate can be obtained by determining conversion at the first time point. After 1 minute, the reaction conducted at 200 mM progressed 10.2%, corresponding to an initial TOF of ~1200 hr-1. This is the highest reaction rate for citronellal cyclization reported for any MOF based catalyst by three orders of magnitude.325 While operative catalytic sites in both Cu2BTC3 and UiO-66 are derived from open coordination 162 sites at the secondary building unit, 38–PdBF4 contains catalytic sites at the organic linker.

The high activity of 38–PdBF4 in comparison highlights the value of unveiling non- framework Lewis acidic sites in MOF based catalysts.

Immobilization within the MOF lattice may endow 38–PdBF4 with a higher catalytic

t lifetime compared to its homogenous counterpart, Bu4L–PdBF4. To this end, the

t cyclization of citronellal was conducted in the presence of 0.5 mol % Bu4L–PdBF4. After

3 hours, only 71% of the starting material had been converted, delivering 51% of the desired isopulegol diastereomer, 47 (entry 5). In comparison, 38–PdBF4 furnishes 47 in

60% yield concomitant with nearly quantitative citronellal conversion in the same time

t period (entry 4). During the course of the reaction, Bu4L–PdBF4 may decompose via adventitious reduction to Pd(0), a common pathway that plagues Pd pincer catalyzed cross coupling reactions.85 Moreover, the resultant zero valent Pd species are expected to exhibit inferior Lewis acidity than the parent pincer complex. Immobilization of the Pd PNPNP complex within the MOF lattice may suppress these deleterious reductive processes which decrease catalyst lifetime. The exact origin of these immobilization effects are under current study and will be reported in due time.

6.9 Conclusions

A carboxylate-functionalized Pd PNNNP pincer complex has been used for the assembly of a porous Zr MOF, 38–PdX. PXRD analysis shows that the MOF adopts a cubic framework structure and is isostructural to the previously reported 37–PdX. This result is perhaps not surprising given the structural similarities between the PNNNP and POCOP pincer ligand frameworks, but heralds the ability to synthesize isoreticular or multivariate

163

MOFs using PZEZP pincer complexes. Despite similar solvothermal synthesis conditions, spectroscopic characterization and elemental analysis indicate that 38–PdX contains a larger number of missing linker defects than 37–PdX and is best formulated as

{Zr6O4(OH)4(OAc)2.4[PNNNP-PdX]2.4}Y2.4 (X/Y = Cl, I).

Sequential postsynthetic halide exchange reactions employing NaI followed by either

PhI(TFA)2 or NOBF4 were used to activate 38–PdX toward Lewis acid mediated catalysis.

– – NOBF4 proved to be the superior iodide exchange agent and achieved complete I /BF4 substitution. Moreover, compositional analysis and catalytic studies showed oxidative halide exchange is a more effective activator of the MOF toward catalysis than treatment with AgOTf.

38-PdBF4 enjoys high recyclability and catalytic activity giving rise to turnover numbers of nearly 200 for the intramolecular hydroamination of 2-(butyn-1-yl)aniline and the carbonyl-ene cyclization of citronellal. 38-PdBF4 also demonstrated an initial TOF ~

1200 h-1 for the cyclization of citronellal which surpasses any reported MOF based catalyst by three orders of magnitude. Notably, 38-PdBF4 demonstrated markedly superior

– catalytic activity to 38-PdTFA. Substitution of O2CCF3 for the less basic and less

– – coordinating BF4 successfully suppressed a previously observed O2CCF3 mediated decomposition process. While 38-PdTFA exhibited a large drop in activity following one catalytic reaction with 2-(butyn-1-yl)aniline, 38-PdBF4 was recycled up to five times without any substantial loss in yield.

2+ Finally, the Pd sites immobilized within the MOF lattice of 38-PdBF4 were used to promote the pore selective cyclization of 2-ethynylaniline. This substrate can undergo

164 either intramolecular hydroamination to afford indole 43 or a tandem intermolecular hydroamination–annulation sequence to afford quinoline 44. While a direct homogenous analog did not enforce selectivity over the reaction, 38-PdBF4 delivered the intramolecular product 43 exclusively. The steric constraints of the MOF pores likely hinder formation of larger intermolecular reaction products. 38-PdBF4 demonstrates a rare example of pore selective catalysis where the microenvironment of the MOF governs product distribution.

The tunable pore space endemic to MOF based materials has not been heavily explored as a selectivity inducing agent for organic transformations. Ongoing work is focused on understanding how catalytic pore selectivity mediated by 38-PdBF4 and similar

PincerMOFs can be extended to other more challenging chemical reactions.

6.10 Experimental

General Considerations. ZrCl4 (Sigma Aldrich), N,N-dimethylformamide (DMF,

99.9%, EMD), and glacial acetic acid (Macron) used for synthetic preparations were used

t 19 56 as received unless otherwise noted. ClP(C6H5-COO Bu)2, PdCl2(cod), PtCl2(cod),

328 329 330 butynyl aniline 39, ethynyl aniline 42, and PhITFA2 were prepared as described in the literature. All other solvents and reagents were purchased from commercial suppliers and used as received. Routine powder X-ray diffraction patterns for phase identification were collected on a Rigaku Miniflex 600 diffractometer using Nickel-filtered Cu-Kα radiation (λ = 1.5418 Å). High-resolution synchrotron powder diffraction data (PXRD) were collected at 295 K using beamline 11-BM at the Advanced Photon Source (APS,

Argonne National Laboratory, Argonne, IL) using an average wavelength of 0.414536 Å.

Nitrogen adsorption isotherms were measured at 77K (liquid nitrogen bath) using a

165

Micromeritics 3Flex Surface Characterization Analyzer. Prior to analysis, samples (100-

200 mg) were heated under reduced pressure until the outgas rate was less than 2 mTorr/minute. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) measurements were performed by Robertson Microlit Laboratories (Ledgewood, NJ).

Elemental microanalyses were performed by Atlantic Microlab (Norcross, GA) or

Robertson Microlit Laboratories (Ledgewood, NJ). ESI-MS experiments were performed using a Micromass ZQ4000 single quadrupole mass detector. XRF spectra were collected on an Innov-X Systems X-500 spectrometer using unfiltered Co-Kα radiation (λ = 1.7892

Å). GC-MS analysis was performed using an Agilent 7890B GC system equipped with the

HP-5 Ultra Inert column (30 m, 0.25 mm, 0.25 µm), and a FID detector. For MS detec-tion an electron ionization system was used with an ioniza-tion energy of 70 eV.

Solution-state NMR spectra were measured using either a Varian Inova or MR 400

MHz spectrometer or Bruker DPX 400 MHz spectrometer (101 MHz operating frequency for 13C, 162 MHz operating frequency for 31P, and 376 MHz operating frequency for 19F).

For 1H and 13C{1H} NMR spectra, the solvent resonance was referenced as an internal

31 1 standard. For P{ H} NMR spectra, 85% H3PO4 was used as an external standard (0 ppm).

19 For F NMR spectra, 1% CF3COOH was used as an external standard (-76.55 ppm).

Solvent-suppressed 1H NMR spectra were collected using the WET1D sequence with default parameters or 180° water selective excitation sculpting with default parameters and pulse shapes.164,331 Solid-State NMR experiments were performed on a Bruker (Billerica,

MA) DSX-400 spectrometer at a resonance frequency of 400 MHz for 1H and 162 MHz

166 for 31P and 100 MHz for 13C, using magic-angle spinning (MAS) probe in double- resonance mode.

t Synthesis of Bu4L. Under an inert atmosphere, a 100 mL Schlenk flask was charged with 2,6-diaminopyridine (0.156 g, 1.43 mmol), triethylamine (0.360 g, 3.56 mmol), and toluene (25 mL). The reaction mixture was cooled to 0 °C and a solution of

tBu ClP(C6H4COO )2 (1.344 g, 3.19 mmol) in toluene (20 mL) was added dropwise. The flask was then sealed and heated at 80 °C for 16 h. After cooling, the pale yellow solution was filtered, and the solvent was removed under reduced pressure. The resulting sticky yellow powder was recrystallized with toluene/pentane (1:4) to give the desired product as a white

31 1 microcrystalline powder (1.04 g, 1.18 mmol, 83%). P{ H} NMR (162.0 MHz, CDCl3): δ

1 25.4 (s). H NMR (400 MHz, CDCl3): δ 7.95 (d, J = 7.7 Hz, 8H), 7.46 (t, J = 7.6 Hz, 8H),

7.34 (t, J = 8.0 Hz, 1H), 6.44 (d, J = 7.8 Hz, 2H), 4.99 (d, J = 7.6 Hz, 2H), 1.58 (s, 36H).

13 1 2,6 C{ H} NMR (101 MHz, CDCl3): δ 165.4 (s, 4C, CO2), 157.2 (d, J = 20.5 Hz, 2C, Py ),

144.4 (d, J = 14.4 Hz, 4C, Ph), 140.0 (s, C, Py4), 133.0 (s, 4C, Ph4), 131.1 (d, 8C, J = 21.0

Hz, Ph2,6), 129.5 (d, 8C, J = 6.6 Hz, Ph3,5), 99.9 (d, 2C, J = 14.4 Hz, 2C, Py3,5), 81.4 (s, 4C,

167

31 1 C(CH3)3), 28.3 (s, 12C, C(CH3)3). P{ H} NMR (162.0 MHz, CDCl3): δ 25.4 (s). Anal.

Calcd. for C49H57N3O8P2: C, 67.04; H, 6.54; N, 4.79; Found: C, 67.30; H, 6.55; N, 4.83.

t Synthesis of Bu4L-PdCl. A solution of PdCl2(cod) (0.164 g, 0.575 mmol) in CH2Cl2

(3 mL) was added to a stirring solution of tBu4L (0.505 g, 0.575 mmol) in CH2Cl2 (8 mL).

The reaction was stirred at room temperature for 4 h under an inert atmosphere before removing the volatiles under reduced pressure. The resulting orange residue was triturated with a small amount of pentane (3 mL) resulting in formation of a bright yellow solid. The solid was collected by filtration, washed with pentane (3 × 3 mL), and dried under reduced

1 pressure to yield tBu4L-PdCl (0.574 g, 95% yield). H NMR (400 MHz, CDCl3): δ 11.17

3 3 3 (s, 2H, NH), δ 8.10 (dd, JH-P = 14.2 Hz, JH-H = 6.4 Hz, 8H, benzoate Ar-H), 7.92 (d, JH-

3 3 H = 8.1 Hz, 8H, benzoate Ar-H), 7.18 (t, JH-H = 6.4 Hz, 1H, pyridine Ar-H), 7.11 (d, JH-H

t 13 1 = 7.8 Hz, 2H, pyridine Ar-H), 1.50 (s, 36H, Bu). C{ H} NMR (101 MHz, CDCl3): δ

164.3 (s, 4C, CO2), 160.9 (t, JC-P = 7.7 Hz, 2C, Ar), 142.6 (s, C, Ar), 135.5 (s, 4C, Ar),

134.0 (t, JC-P = 28.6 Hz, 4C, Ar), 132.2 (t, JC-P = 8.1 Hz, 8C, Ar), 130.0 (t, JC-P = 6.1 Hz,

31 1 8C, Ar), 102.2 (br, 2C, Ar), 81.9 (s, 4C, C(CH3)3), 28.2 (s, 12C, C(CH3)3). P{ H} NMR

(162.0 MHz, CDCl3): δ 67.8 (s). Anal. Calcd. for C49H57Cl2N3O8P2Pd: C, 55.77; H, 5.44;

N, 3.98; Found: C, 55.96; H, 5.61; N, 3.79.

t Synthesis of Bu4L-PdI. A solution of tBu4L-PdCl (1.105 g, 1.05 mmol) in acetone (8 mL) was treated with a solution of NaI (0.317 g, 2.12 mmol) in acetone (1 mL). Immediate formation of NaCl was observed, and the reaction was allowed to stir at room temperature for 1 h. The solvent was removed under reduced pressure, and the dark red residue was extracted with CH2Cl2 (5 mL) and filtered through a 0.45 µm PTFE syringe filter. The

168 filtrate was concentrated in vacuo to afford the desired product as an orange solid (1.248

1 g, 96%). H NMR (400 MHz, CDCl3): δ 9.60 (s, 2H, NH), 8.09–7.99 (m, 16H, benzoate

3 3 Ar-H), 7.40 (t, JH-H = 7.5 Hz, 1H, pyridine Ar-H), 7.30 (d, JH-H = 8.2 Hz, 2H, pyridine

t 13 1 Ar-H), 1.53 (s, 36H, Bu). C{ H} NMR (101 MHz, CDCl3): δ 164.4 (s, 4C, CO2), 159.7

(t, JC-P = 7.2 Hz, 2C, Ar), 142.6 (s, C, Ar), 136.0 (s, 4C, Ar), 133.44 (t, JC-P = 8.0 Hz, 8C,

Ar), 133.40 (t, JC-P = 29.4 Hz, 4C, Ar), 130.0 (t, JC-P = 6.1 Hz, 8C, Ar), 102.3 (br, 2C, Ar),

31 1 82.1 (s, 4C, C(CH3)3), 28.2 (s,12C, C(CH3)3). P{ H} NMR (162.0 MHz, CDCl3): δ 75.6

(s). Anal. Calcd. for C49H57I2N3O8P2Pd: C, 47.53; H, 4.64; N, 3.39; Found: C, 47.57; H,

4.79; N, 3.45;

t Synthesis of Bu4L-PtI. The compound was prepared from tBu4L-PtCl (0.863 g, 0.754 mmol) following the same procedure used for tBu4L-PdI. The reaction yielded 0.951 g

1 (95%) of tBu4L-PtI as a reddish-orange solid. H NMR (400 MHz, CDCl3): δ 10.09 (s, 2H),

3 8.09–7.99 (m, 16H, benzoate Ar-H), 7.42 (t, JH-H = 7.7 Hz, 1H, pyridine Ar-H), 7.31 (d,

3 t 13 1 JH-H = 8.0 Hz, 2H, pyridine Ar-H), 1.53 (s, 36H, Bu). C{ H} NMR (101 MHz, CDCl3):

δ 164.4 (s, 4C, CO2), 159.6 (t, JC-P = 6.6 Hz, 2C, Ar), 141.7 (s, C, Ar), 136.0 (s, 4C, Ar),

133.6 (t, JC-P = 7.8 Hz, 8C, Ar), 133.1 (t, JC-P = 33.9 Hz, 4C, Ar), 129.9 (t, JC-P = 6.3 Hz,

31 1 8C, Ar), 101.8 (br, 2C, Ar), 82.1 (s, 4C, C(CH3)3), 28.2 (s, 12C, C(CH3)3). P{ H} NMR

1 (162.0 MHz, CDCl3): δ 68.1 (d, JPt–P = 2682.8 Hz). Anal. Calcd. for C49H57I2N3O8P2Pt:

C, 44.36; H, 4.33; N, 3.17; Found: C, 44.31; H, 4.52; N, 3.14.

Synthesis of H3L-PdI. Trifluoroacetic acid (1 mL) was added to a solution of tBu4L-

PdI (0.542 g, 0.438 mmol) in CH2Cl2 (5 mL) resulting in a color change of the solution from ruby to dark purple. The solution was stirred for 16 h at room temperature before

169 removing the solvent using a rotary evaporator. Deionized water (12 mL) was added resulting in precipitation of a bright orange solid. The solid was collected by vacuum filtration and washed with deionized water (3 × 5 mL) and CHCl3 (~20 mL). The solid was dried under reduced pressure to afford 0.393 g of the crude product (H4[L-PdI]I). The solid was suspended in acetone (5 mL), and a solution of pyridine (0.032 g, 0.405 mmol) in acetone (2 mL) was added, resulting in a color change of the supernatant from ruby to bright orange. The solution was stirred for 30 mins at room temperature. The mixture was centrifuged and the supernatant was decanted. The solid was washed successively with acetone (3 × 5 mL) and then water (~ 20 mL) until no color persisted in the filtrate. The resulting bright orange solid was dried under vacuum to afford H3[L-PdI] (0.284 g, 0.321

1 3 mmol, 86% yield). H NMR (400 MHz, DMSO): δ 12.71 (br, 2H, NH), 8.07 (d, JH-H = 7.8

3 3 Hz, 8H, benzoate Ar-H), δ 7.94 (dd, JH-P = 12.9 Hz, JH-H = 6.1 Hz, 8H, benzoate Ar-H),

3 3 7.41 (t, JH-H = 7.7 Hz, 1H, pyridine Ar-H), 6.28 (d, JH-H = 7.2 Hz, 2H, pyridine Ar-H).

13 1 C{ H} NMR (101 MHz, DMSO): δ 166.7 (s, 4C, CO2), 162.9 (br, 2C, Ar), 141.8 (br, C,

Ar), 136.4 (br, 4C, Ar), 134.3 (br, 8C, Ar), 132.5 (dd, JC-P = 13.0 Hz, JC-P = 6.4 Hz, 4C,

31 1 Ar), 129.5 (t, JC-P = 5.4 Hz, 8C, Ar), 98.8 (br, 2C, Ar). P{ H} NMR (162.0 MHz, DMSO):

δ 69.9 (s). Anal. Calcd. for H3L-PdI·(H2O); C33H26IN3O9P2Pd: C, 43.85; H, 2.90; N, 4.65;

Found: C, 44.18; H, 3.07; N, 4.55.

Synthesis of 38-PdX. Anhydrous ZrCl4 (0.030 g, 0.129 mmol) was suspended in acetic acid (1.6 mL) and DMF (4.4 mL) in a 20 mL screw-top scintillation vial. A solution of

H3L-PdI (0.043 mmol) in DMF (2 mL) was added and the vial was sealed with Teflon- lined screw-top cap (Qorpak® CAP-00554). The reaction mixture was sonicated for 5 min

170 to ensure complete dissolution of the solids. The vial was then placed in a programmable oven at room temperature oven and heated to 120 °C for 16 h. After reaching room temperature, the solvent was decanted from the precipitated solid. The solid was washed with DMF (3 × 15 mL) and acetone (3 × 10 mL) and dried in vacuo (0.01 torr) at room temperature for 2 h to afford 0.035 mg of product. 31P{1H} NMR (162.0 MHz, 3/1 v/v

CF3COOH/C6D6): δ 74.0 (s); 69.9 (s). Anal. Calcd. for

{Zr6O4(OH)4[PdClC33H21N3O8P2]2.4(CH3COO)2.4}I2.4; C, 34.68; H, 2.16; N, 3.37; I, 10.18;

Cl, 2.84, Zr, 17.68; Pd, 8.54; Found: C, 32.70; H, 2.47; N, 3.11; I, 6.54; Cl, 1.95; Zr, 16.60;

Pd, 6.68.

Synthesis of 38-PdI. A solution of NaI (0.098 mg, 1.31 mmol) in deionized water (5 mL) was added to a suspension of 38–PdX (0.160 g) in deionized water (5 mL) in a 20 mL scintillation vial. The vial was sealed and placed in an oven at 60 °C for 2 h. After cooling, the mixture was centrifuged, and the supernatant was decanted. A fresh solution of NaI

(0.098 mg, 1.31 mmol) in deionized water (5 mL) was added and the reaction was again heated at 60 °C for 2 h. The solid was collected by centrifugation and washed successively with water (3 × 10 mL) and acetone (3 × 10 mL) and dried briefly in vacuo to afford 38–

PdI as a yellow microcrystalline powder (0.150 g). 31P{1H} NMR (162.0 MHz, 3/1 v/v

CF3COOH/C6D6): δ 74.0 (s).

Synthesis of 38–PdTFA. In a N2-filled glovebox, solution of PhITFA2 (0.026 g, 0.06 mmol) in CH2Cl2 (1 mL) was added to a suspension of 38–PdI (0.050 g) in CH2Cl2 (5 mL) in a 20 mL scintillation vial. The vial was sealed and left gently stirring at room temperature. After 12 h, the reaction mixture was centrifuged and the pink purple

171 supernatant was decanted. The solid was washed with CH2Cl2 (3 × 5 mL) and dried briefly in vacuo. 38–PdTFA was isolated as a pale yellow, microcrystalline powder (0.050 mg).

31 1 P{ H} NMR (162.0 MHz, 3/1 v/v CF3COOH/C6D6): δ 74.0 (s); 71.1 (s).

Reaction of 38–PdX and 38–PdI with AgOTf. In a N2-filled glovebox, a solution of

AgOTf (0.023 g, 0.090 mmol) in MeCN (1 mL) was added to a suspension of 38–PdX or

38–PdI (0.050 g) in MeCN (5 mL) in a 20 mL scintillation vial. The vial was sealed and left gently stirring at 60 °C. After 12 h, the reaction mixture was centrifuged, and the supernatant was decanted. The solid was washed with MeCN (3 × 5 mL) and dried briefly in vacuo. 38–PdX/AgOTf and 38–PdI/AgOTf were isolated as yellow microcrystalline powders (0.050 mg).

Synthesis of 38–Pd(PMe3). In a N2-filled glovebox, a solution of PMe3 (0.097 g, 0.127 mmol) in MeCN (1 mL) was added to a suspension of 38–PdX (0.020 g) in MeCN (2 mL) in a 20 mL scintillation vial. The vial was sealed and left gently stirring at room temperature. After 12 h, the reaction mixture was centrifuged and the supernatant was decanted. The solid was washed with MeCN (3 × 5 mL) and dried briefly in vacuo. 38–

31 1 Pd(PMe3) was isolated as a yellow microcrystalline powder (0.020 mg). P{ H} NMR

1 2+ (162.0 MHz, 3/1 v/v CF3COOH/C6D6): δ 78.1 (d, JP-P = 25.7 Hz, H4[L-Pd(PMe3)] ) 74.0

+ + 1 2+ (s, H4[L-PdI] ); 2.9 (s, HPMe3 ); -9.7 (t, JP-P = 25.7 Hz, H4[L-Pd(PMe3)] ).

Synthesis of 38–PdBF4. In a N2-filled glovebox, solution of NOBF4 (0.009 g, 0.072 mmol) in MeCN (1 mL) was added to a suspension of 1–PdI (0.100 g) in MeCN (5 mL) in a 20 mL scintillation vial. The vial was sealed and left gently stirring at room temperature.

After 12 h, the reaction mixture was centrifuged and the orange supernatant was decanted.

172

A fresh solution of NOBF4 (0.009 g, 0.072 mmol) in MeCN (6 mL) was added and the reaction again stirred at room temperature for 12 hours. The solid was collected by centrifugation, washed with MeCN (3 × 5 mL), and dried briefly in vacuo. 38–PdBF4 was isolated as an off-white microcrystalline powder (0.100 mg). 31P{1H} NMR (162.0 MHz,

3/1 v/v CF3COOH/C6D6): δ 76.2 (s); 75.2 (s); 74.5 (s); 70.2 (s).

General Procedure for Intramolecular Hydroamination Reactions: In a N2-filled glovebox, a vial was charged with 39 (0.1 mmol), 5 mol % catalyst, 1,4-dioxane (0.4 mL),

C6D6 (0.1 mL), and a known amount of hexamethyl benzene (0.02 – 0.04 mmol) as an internal standard. The reaction mixture was transferred to an NMR tube and heated at 95

°C for 12 hr. The product yields were determined by 1H NMR spectroscopy.

General Procedure for Carbonyl-Ene Reactions: In a N2-filled glovebox, a 1 dram screw-top vial fitted with a Teflon lined cap was charged with 46 (0.05 mmol), 10 mol % catalyst, toluene (2 mL), and a known amount of hexamethyl benzene (0.01 – 0.04 mmol) as an internal standard. The reaction mixture was heated at 100 °C for 30 minutes. The product yields were determined by quantitative GC-FID.

173

Bibliography

(1) Raubenheimer, H. G.; Schmidbaur, H. The Late Start and Amazing Upswing in Gold Chemistry. J. Chem. Educ. 2014, 91, 2024–2036.

(2) Joost, M.; Amgoune, A.; Bourissou, D. Reactivity of Gold Complexes towards Elementary Organometallic Reactions. Angew. Chem. Int. Ed. 2015, 54, 15022– 15045.

(3) Rekhroukh, F.; Estevez, L.; Mallet-Ladeira, S.; Miqueu, K.; Amgoune, A.; Bourissou, D. β-Hydride Elimination at Low-Coordinate Gold(III) Centers. J. Am. Chem. Soc. 2016, 138, 11920–11929.

(4) Bratsch, S. G. Standard Electrode Potentials and Temperature Coefficients in Water at 298.15 K. J. Phys. Chem. Ref. Data 1989, 18, 1–21.

(5) Joost, M.; Zeineddine, A.; Estévez, L.; Mallet-Ladeira, S.; Miqueu, K.; Amgoune, A.; Bourissou, D. Facile Oxidative Addition of Aryl Iodides to Gold(I) by Ligand Design: Bending Turns on Reactivity. J. Am. Chem. Soc. 2014, 136, 14654–14657.

(6) Zeineddine, A.; Estévez, L.; Mallet-Ladeira, S.; Miqueu, K.; Amgoune, A.; Bourissou, D. Rational Development of Catalytic Au(I)/Au(III) Arylation Involving Mild Oxidative Addition of Aryl Halides. Nat. Commun. 2017, 8, 565.

(7) Harper, M. J.; Arthur, C. J.; Crosby, J.; Emmett, E. J.; Falconer, R. L.; Fensham- Smith, A. J.; Gates, P. J.; Leman, T.; McGrady, J. E.; Bower, J. F.; et al. Oxidative Addition, Transmetalation, and Reductive Elimination at a 2,2′-Bipyridyl-Ligated Gold Center. J. Am. Chem. Soc. 2018, 140, 4440–4445.

(8) Corma, A.; Juárez, R.; Boronat, M.; Sánchez, F.; Iglesias, M.; García, H. Gold Catalyzes the Sonogashira Coupling Reaction without the Requirement of Palladium Impurities. Chem. Commun. 2011, 47, 1446–1448.

(9) Fackler, J. P. Metal-Metal Bond Formation in the Oxidative Addition to Dinuclear Gold(I) Species. Implications from Dinuclear and Trinuclear Gold Chemistry for the Oxidative Addition Process Generally. Polyhedron 1997, 16, 1–17.

(10) Schmidbaur, H.; Franke, R. Organogold Chemistry. XVII. Synthesis and Reactions of the Gold(I)-Dimethylphosphonium-Bis-Methylid Dimer. Inorganica Chim. Acta 174

1975, 13, 85–89.

(11) Usón, R.; Laguna, A.; Laguna, M.; Fraile, M. N.; Jones, P. G.; Sheldrick, G. M. Mono- and Bi-Nuclear Gold(I), Gold(II ), and Gold(III) Perhalogenoaryl Complexes with the Ligand Bis(Diphenylphosphino)Amine. Crystal and Molecular Structure of µ-[Bis(Diphenylphosphino)Amine]- Dichlorobis(Pentafluorop. J. Chem. Soc., Dalt. Trans. 1986, No. 2, 291–296.

(12) Guenther, J.; Mallet-Ladeira, S.; Estevez, L.; Miqueu, K.; Amgoune, A.; Bourissou, D. Activation of Aryl Halides at Gold(I): Practical Synthesis of (P,C) Cyclometalated Gold(III) Complexes. J. Am. Chem. Soc. 2014, 136, 1778–1781.

(13) Gimeno, M. C.; Laguna, A. Three- and Four-Coordinate Gold(I) Complexes. Chem. Rev. 1997, 97, 511–522.

(14) Schmidbaur, H. The Aurophilicity Phenomenon: A Decade of Experimental Findings, Theoretical Concepts and Emerging Applications. Gold Bull. 2000, 33, 3–10.

(15) Pyykkö, P. Theoretical Chemistry of Gold. Angew. Chem. Int. Ed. 2004, 43, 4412– 4456.

(16) Ferrer, C.; Escribano-Cuesta, A.; Echavarren, A. M. Synthesis of the Tetracyclic Core Skeleton of the Lundurines by a Gold-Catalyzed Cyclization. Tetrahedron 2009, 65, 9015–9020.

(17) Ferrer, C.; Amijs, C. H. M.; Echavarren, A. M. Intra- and Intermolecular Reactions of Indoles with Alkynes Catalyzed by Gold. Chem. - A Eur. J. 2007, 13, 1358– 1373.

(18) Ranieri, B.; Escofet, I.; Echavarren, A. M. Anatomy of Gold Catalysts: Facts and Myths. Org. Biomol. Chem. 2015, 13, 7103–7118.

(19) McGonigal, P. R.; de León, C.; Wang, Y.; Homs, A.; Solorio-Alvarado, C. R.; Echavarren, A. M. Gold for the Generation and Control of Fluxional Barbaralyl Cations. Angew. Chem. Int. Ed. 2012, 51, 13093–13096.

(20) Ung, G.; Soleilhavoup, M.; Bertrand, G. Gold(III)- versus Gold(I)-Induced Cyclization: Synthesis of Six-Membered Mesoionic Carbene and Acyclic (Aryl)(Heteroaryl) Carbene Complexes. Angew. Chem. Int. Ed. 2013, 52, 758–761.

(21) Zhang, L. Tandem Au-Catalyzed 3,3-Rearrangement−[2 + 2] Cycloadditions of Propargylic Esters: Expeditious Access to Highly Functionalized 2,3-Indoline- Fused Cyclobutanes. J. Am. Chem. Soc. 2005, 127, 16804–16805.

(22) Tkatchouk, E.; Mankad, N. P.; Benitez, D.; Goddard, W. A.; Toste, F. D. Two 175

Metals Are Better Than One in the Gold Catalyzed Oxidative Heteroarylation of Alkenes. J. Am. Chem. Soc. 2011, 133, 14293–14300.

(23) Xie, J.; Shi, S.; Zhang, T.; Mehrkens, N.; Rudolph, M.; Hashmi, A. S. K. A Highly Efficient Gold-Catalyzed Photoredox α-C(Sp 3 )-H Alkynylation of Tertiary Aliphatic Amines with Sunlight. Angew. Chem. Int. Ed. 2015, 54, 6046–6050.

(24) Xie, J.; Zhang, T.; Chen, F.; Mehrkens, N.; Rominger, F.; Rudolph, M.; Hashmi, A. S. K. Gold-Catalyzed Highly Selective Photoredox C(Sp 2 )−H Difluoroalkylation and Perfluoroalkylation of Hydrazones. Angew. Chem. Int. Ed. 2016, 55, 2934–2938.

(25) Nzulu, F.; Telitel, S.; Stoffelbach, F.; Graff, B.; Morlet-Savary, F.; Lalevée, J.; Fensterbank, L.; Goddard, J.-P.; Ollivier, C. A Dinuclear Gold(I) Complex as a Novel Photoredox Catalyst for Light-Induced Atom Transfer Radical Polymerization. Polym. Chem. 2015, 6, 4605–4611.

(26) Zidan, M.; McCallum, T.; Thai-Savard, L.; Barriault, L. Photoredox Meets Gold Lewis Acid Catalysis in the Alkylative Semipinacol Rearrangement: A Photocatalyst with a Dark Side. Org. Chem. Front. 2017, 4, 2092–2096.

(27) Kaldas, S. J.; Cannillo, A.; McCallum, T.; Barriault, L. Indole Functionalization via Photoredox Gold Catalysis. Org. Lett. 2015, 17, 2864–2866.

(28) Revol, G.; McCallum, T.; Morin, M.; Gagosz, F.; Barriault, L. Photoredox Transformations with Dimeric Gold Complexes. Angew. Chem. Int. Ed. 2013, 52, 13342–13345.

(29) Mohamed, A. A.; Abdou, H. E.; Fackler Jr., J. P. Coordination Chemistry of Gold(II) with Amidinate, Thiolate and Ylide Ligands. Coord. Chem. Rev. 2010, 254, 1253–1259.

(30) Fackler, J. P. Forty-Five Years of Chemical Discovery Including a Golden Quarter-Century. Inorg. Chem. 2002, 41, 6959–6972.

(31) Stein, J.; Fackler, J. P.; Paparizos, C.; Chen, H. W. Structure and Properties of Transition-Metal Ylide Complexes. 2. Organometallic Complexes of Gold(III). J. Am. Chem. Soc. 1981, 103, 2192–2198.

(32) Trzcinska-Bancroft, B.; Khan, M. N. I.; Fackler, J. P. Oxidative Addition of Nitroalkanes to Dinuclear Gold(I) Ylide Complexes. The Characterization by x- Ray Crystallography of Au2(Ylide)2(NO2)2, a Nitro Complex of Gold(II). Organometallics 1988, 7, 993–996.

(33) Fackler, J. P.; Basil, J. D. Oxidative Addition of Methyl Iodide to a Dinuclear

176

Gold(I) Complex. The x-Ray Crystal Structure of Bis[.Mu.- (Dimethyldimethylenephosphoranyl-C,C)]-Iodomethyldigold(II)(Au-Au), Au2[(CH2)2P(CH3)2]2(CH3)I. Organometallics 1982, 1, 871–873.

(34) Raptis, R. G.; Murray, H. H.; Staples, R. J.; Porter, L. C.; Fackler, J. P. Structural Isomers of [Au(CH2)2PPh2]2Br4. 2. Crystal Structures of Cis/Cis- [Au(CH2)2PPh2]2Br4 and the Cationic A-Frame [(.Mu.- Br)(Au(CH2)2PPh2)2Br2][IBr2]. Inorg. Chem. 1993, 32, 5576–5581.

(35) Jandik, P.; Schubert, U.; Schmidbaur, H. Methylene Bridging of Two Gold Atoms through Double Oxidative Addition of Methylene Dihalides to a Cyclic Ylide Complex. Angew. Chemie Int. Ed. English 1982, 21, 73–73.

(36) Schmidbaur, H.; Hartmann, C.; Riede, J.; Huber, B.; Mueller, G. Alkylation of Methylene- and Ylide-Bridged Binuclear Gold(III) Complexes. Organometallics 1986, 5, 1652–1656.

(37) Murray, H. H.; Fackler, J. P.; Mazany, A. M. Bridging (Methylene)Gold(III) Ylide Dimers: The Reactivity of and a Possible Intermediate Involved in Its Generation. Organometallics 1984, 3, 1310–1311.

(38) Jørgensen, C. K. Differences between the Four Halide Ligands, and Discussion Remarks on Trigonal-Bipyramidal Complexes, on Oxidation States, and on Diagonal Elements of One-Electron Energy. Coord. Chem. Rev. 1966, 1, 164–178.

(39) Wolczanski, P. T. Flipping the Oxidation State Formalism: Charge Distribution in Organometallic Complexes As Reported by Carbon Monoxide. Organometallics 2017, 36, 622–631.

(40) Tondreau, A. M.; Atienza, C. C. H.; Weller, K. J.; Nye, S. A.; Lewis, K. M.; Delis, J. G. P.; Chirik, P. J. Iron Catalysts for Selective Anti-Markovnikov Alkene Hydrosilylation Using Tertiary Silanes. Science 2012, 335, 567–570.

(41) Sylvester, K. T.; Chirik, P. J. Iron-Catalyzed, Hydrogen-Mediated Reductive Cyclization of 1,6-Enynes and Diynes: Evidence for Bis(Imino)Pyridine Ligand Participation. J. Am. Chem. Soc. 2009, 131, 8772–8774.

(42) Bouwkamp, M. W.; Bowman, A. C.; Lobkovsky, E.; Chirik, P. J. Iron-Catalyzed [2π + 2π] Cycloaddition of α,ω-Dienes: The Importance of Redox-Active Supporting Ligands. J. Am. Chem. Soc. 2006, 128, 13340–13341.

(43) Kiernicki, J. J.; Fanwick, P. E.; Bart, S. C. Utility of a Redox-Active Pyridine(Diimine) Chelate in Facilitating Two Electron Oxidative Addition Chemistry at Uranium. Chem. Commun. 2014, 50, 8189.

177

(44) Anderson, N. H.; Odoh, S. O.; Yao, Y.; Williams, U. J.; Schaefer, B. A.; Kiernicki, J. J.; Lewis, A. J.; Goshert, M. D.; Fanwick, P. E.; Schelter, E. J.; et al. Harnessing Redox Activity for the Formation of Uranium Tris(Imido) Compounds. Nat. Chem. 2014, 6, 919–926.

(45) Blackmore, K. J.; Lal, N.; Ziller, J. W.; Heyduk, A. F. Catalytic Reactivity of a Zirconium(IV) Redox-Active Ligand Complex with 1,2-Diphenylhydrazine. J. Am. Chem. Soc. 2008, 130, 2728–2729.

(46) Würthner, F.; Stolte, M. Naphthalene and Perylene Diimides for Organic Transistors. Chem. Commun. 2011, 47, 5109.

(47) Suraru, S.-L.; Würthner, F. Strategies for the Synthesis of Functional Naphthalene Diimides. Angew. Chem. Int. Ed. 2014, 53, 7428–7448.

(48) Al Kobaisi, M.; Bhosale, S. V.; Latham, K.; Raynor, A. M.; Bhosale, S. V. Functional Naphthalene Diimides: Synthesis, Properties, and Applications. Chem. Rev. 2016, 116, 11685–11796.

(49) Zhao, X.; Xiong, Y.; Ma, J.; Yuan, Z. Rylene and Rylene Diimides: Comparison of Theoretical and Experimental Results and Prediction for High-Rylene Derivatives. J. Phys. Chem. A 2016, 120, 7554–7560.

(50) Bhosale, S. V.; Jani, C. H.; Langford, S. J. Chemistry of Naphthalene Diimides. Chem. Soc. Rev. 2008, 37, 331–342.

(51) Šolomek, T.; Powers-Riggs, N. E.; Wu, Y.-L.; Young, R. M.; Krzyaniak, M. D.; Horwitz, N. E.; Wasielewski, M. R. Electron Hopping and Charge Separation within a Naphthalene-1,4:5,8-Bis(Dicarboximide) Chiral Covalent Organic Cage. J. Am. Chem. Soc. 2017, 139, 3348–3351.

(52) Langford, S. J.; Latter, M. J.; Woodward, C. P. Progress in Charge Transfer Systems Utilizing Porphyrin Donors and Simple Aromatic Diimide Acceptor Units. Photochem. Photobiol. 2006, 82, 1530–1540.

(53) Matsugo, S.; Kawanishi, S.; Yamamoto, K.; Sugiyama, H.; Matsuura, T.; Saito, I. Bis(Hydroperoxy)Naphthaldiimide as a“Photo-Fenton Reagent”: Sequence- Specific Photocleavage of DNA. Angew. Chemie Int. Ed. English 1991, 30, 1351– 1353.

(54) Núñez, M. E.; Noyes, K. T.; Gianolio, D. A.; McLaughlin, L. W.; Barton, J. K. Long-Range Guanine Oxidation in DNA Restriction Fragments by a Triplex- Directed Naphthalene Diimide Intercalator †. Biochemistry 2000, 39, 6190–6199.

(55) Rogers, J. E.; Weiss, S. J.; Kelly, L. A. Photoprocesses of Naphthalene Imide and

178

Diimide Derivatives in Aqueous Solutions of DNA. J. Am. Chem. Soc. 2000, 122, 427–436.

(56) Vicic, D. A.; Odom, D. T.; Núñez, M. E.; Gianolio, D. A.; McLaughlin, L. W.; Barton, J. K. Oxidative Repair of a Thymine Dimer in DNA from a Distance by a Covalently Linked Organic Intercalator. J. Am. Chem. Soc. 2000, 122, 8603–8611.

(57) Kawai, K.; Takada, T.; Nagai, T.; Cai, X.; Sugimoto, A.; Fujitsuka, M.; Majima, T. Long-Lived Charge-Separated State Leading to DNA Damage through Hole Transfer. J. Am. Chem. Soc. 2003, 125, 16198–16199.

(58) Kawai, K.; Osakada, Y.; Fujitsuka, M.; Majima, T. Effects of Reaction Rate of Radical Anion of a Photosensitizer with Molecular Oxygen on the Photosensitized DNA Damage. Chem. Commun. 2006, No. 37, 3918.

(59) Chen, S.; Li, Y.; Wang, C. Visible-Light-Driven Photocatalytic H 2 Evolution from Aqueous Suspensions of Perylene Diimide Dye-Sensitized Pt/TiO 2 Catalysts. RSC Adv. 2015, 5, 15880–15885.

(60) Chen, S.; Jacobs, D. L.; Xu, J.; Li, Y.; Wang, C.; Zang, L. 1D Nanofiber Composites of Perylene Diimides for Visible-Light-Driven Hydrogen Evolution from Water. RSC Adv. 2014, 4, 48486–48491.

(61) Monti, A.; de Ruiter, J. M.; de Groot, H. J. M.; Buda, F. A Dynamic View of Proton-Coupled Electron Transfer in Photocatalytic Water Splitting. J. Phys. Chem. C 2016, 120, 23074–23082.

(62) Vagnini, M. T.; Smeigh, A. L.; Blakemore, J. D.; Eaton, S. W.; Schley, N. D.; D’Souza, F.; Crabtree, R. H.; Brudvig, G. W.; Co, D. T.; Wasielewski, M. R. Ultrafast Photodriven Intramolecular Electron Transfer from an Iridium-Based Water-Oxidation Catalyst to Perylene Diimide Derivatives. Proc. Natl. Acad. Sci. 2012, 109, 15651–15656.

(63) Martinez, J. F.; La Porte, N. T.; Mauck, C. M.; Wasielewski, M. R. Photo-Driven Electron Transfer from the Highly Reducing Excited State of Naphthalene Diimide Radical Anion to a CO 2 Reduction Catalyst within a Molecular Triad. Faraday Discuss. 2017, 198, 235–249.

(64) Ghosh, I.; Ghosh, T.; Bardagi, J. I.; Konig, B. Reduction of Aryl Halides by Consecutive Visible Light-Induced Electron Transfer Processes. Science 2014, 346, 725–728.

(65) Zeng, L.; Liu, T.; He, C.; Shi, D.; Zhang, F.; Duan, C. Organized Aggregation Makes Insoluble Perylene Diimide Efficient for the Reduction of Aryl Halides via Consecutive Visible Light-Induced Electron-Transfer Processes. J. Am. Chem. 179

Soc. 2016, 138, 3958–3961.

(66) Castellano, F. N. Transition Metal Complexes Meet the Rylenes. Dalt. Trans. 2012, 41, 8493.

(67) Schmidt, D.; Bialas, D.; Würthner, F. Ambient Stable Zwitterionic Perylene Bisimide-Centered Radical. Angew. Chem. Int. Ed. 2015, 54, 3611–3614.

(68) Li, J.-R.; Kuppler, R. J.; Zhou, H.-C. Selective Gas Adsorption and Separation in Metal–organic Frameworks. Chem. Soc. Rev. 2009, 38, 1477.

(69) Zhou, H. C.; Long, J. R.; Yaghi, O. M. Introduction to Metal-Organic Frameworks. Chem. Rev. 2012, 112, 673–674.

(70) Cohen, S. M.; Zhang, Z.; Boissonnault, J. A. Toward “MetalloMOFzymes”: Metal–Organic Frameworks with Single-Site Metal Catalysts for Small-Molecule Transformations. Inorg. Chem. 2016, 55, 7281–7290.

(71) Kong, X.; Deng, H.; Yan, F.; Kim, J.; Swisher, J. A.; Smit, B.; Yaghi, O. M.; Reimer, J. A. Mapping of Functional Groups in Metal-Organic Frameworks. Science 2013, 341, 882–885.

(72) Yuan, S.; Qin, J.-S.; Zou, L.; Chen, Y.-P.; Wang, X.; Zhang, Q.; Zhou, H.-C. Thermodynamically Guided Synthesis of Mixed-Linker Zr-MOFs with Enhanced Tunability. J. Am. Chem. Soc. 2016, 138, 6636–6642.

(73) Gutov, O. V.; Hevia, M. G.; Escudero-Adán, E. C.; Shafir, A. Metal-Organic Framework (MOF) Defects under Control: Insights into the Missing Linker Sites and Their Implication in the Reactivity of Zirconium-Based Frameworks. Inorg. Chem. 2015, 54, 8396–8400.

(74) Shearer, G. C.; Chavan, S.; Bordiga, S.; Svelle, S.; Olsbye, U.; Lillerud, K. P. Defect Engineering: Tuning the Porosity and Composition of the Metal-Organic Framework UiO-66 via Modulated Synthesis. Chem. Mater. 2016, 28, 3749–3761.

(75) Atzori, C.; Shearer, G. C.; Maschio, L.; Civalleri, B.; Bonino, F.; Lamberti, C.; Svelle, S.; Lillerud, K. P.; Bordiga, S. Effect of Benzoic Acid as a Modulator in the Structure of UiO-66: An Experimental and Computational Study. J. Phys. Chem. C 2017, 121, 9312–9324.

(76) Shearer, G. C.; Chavan, S.; Ethiraj, J.; Vitillo, J. G.; Svelle, S.; Olsbye, U.; Lamberti, C.; Bordiga, S.; Lillerud, K. P. Tuned to Perfection: Ironing out the Defects in Metal-Organic Framework UiO-66. Chem. Mater. 2014, 26, 4068– 4071.

(77) Bai, Y.; Dou, Y.; Xie, L.-H.; Rutledge, W.; Li, J.-R.; Zhou, H.-C. Zr-Based Metal– 180

organic Frameworks: Design, Synthesis, Structure, and Applications. Chem. Soc. Rev. 2016, 45, 2327–2367.

(78) Cohen, S. M. The Postsynthetic Renaissance in Porous Solids. J. Am. Chem. Soc. 2017, 139, 2855–2863.

(79) Wang, C.; Liu, D.; Lin, W. Metal–Organic Frameworks as A Tunable Platform for Designing Functional Molecular Materials. J. Am. Chem. Soc. 2013, 135, 13222– 13234.

(80) Ji, P.; Song, Y.; Drake, T.; Veroneau, S. S.; Lin, Z.; Pan, X.; Lin, W. Titanium(III)-Oxo Clusters in a Metal–Organic Framework Support Single-Site Co(II)-Hydride Catalysts for Arene Hydrogenation. J. Am. Chem. Soc. 2018, 140, 433–440.

(81) Ji, P.; Manna, K.; Lin, Z.; Urban, A.; Greene, F. X.; Lan, G.; Lin, W. Single-Site Cobalt Catalysts at New Zr 8 (µ 2 -O) 8 (µ 2 -OH) 4 Metal-Organic Framework Nodes for Highly Active Hydrogenation of Alkenes, Imines, Carbonyls, and Heterocycles. J. Am. Chem. Soc. 2016, 138, 12234–12242.

(82) Ji, P.; Solomon, J. B.; Lin, Z.; Johnson, A.; Jordan, R. F.; Lin, W. Transformation of Metal–Organic Framework Secondary Building Units into Hexanuclear Zr- Alkyl Catalysts for Ethylene Polymerization. J. Am. Chem. Soc. 2017, 139, 11325– 11328.

(83) Ji, P.; Manna, K.; Lin, Z.; Feng, X.; Urban, A.; Song, Y.; Lin, W. Single-Site Cobalt Catalysts at New Zr 12 (µ 3 -O) 8 (µ 3 -OH) 8 (µ 2 -OH) 6 Metal–Organic Framework Nodes for Highly Active Hydrogenation of Nitroarenes, Nitriles, and Isocyanides. J. Am. Chem. Soc. 2017, 139, 7004–7011.

(84) Li, Z.; Rayder, T. M.; Luo, L.; Byers, J. A.; Tsung, C.-K. Aperture-Opening Encapsulation of a Transition Metal Catalyst in a Metal–Organic Framework for CO 2 Hydrogenation. J. Am. Chem. Soc. 2018, 140, 8082–8085.

(85) Selander, N.; Szabó, K. J. Catalysis by Palladium Pincer Complexes. Chem. Rev. 2011, 111, 2048–2076.

(86) Zhang, H.; Lei, A. Palladium(Iv) Chemistry Supported by Pincer Type Ligands. Dalt. Trans. 2011, 40, 8745.

(87) Gunanathan, C.; Milstein, D. Bond Activation and Catalysis by Ruthenium Pincer Complexes. Chem. Rev. 2014, 114, 12024–12087.

(88) Younus, H. A.; Ahmad, N.; Su, W.; Verpoort, F. Ruthenium Pincer Complexes: Ligand Design and Complex Synthesis. Coord. Chem. Rev. 2014, 276, 112–152.

181

(89) Zell, T.; Milstein, D. Hydrogenation and Dehydrogenation Iron Pincer Catalysts Capable of Metal–Ligand Cooperation by Aromatization/Dearomatization. Acc. Chem. Res. 2015, 48, 1979–1994.

(90) Younus, H. A.; Su, W.; Ahmad, N.; Chen, S.; Verpoort, F. Ruthenium Pincer Complexes: Synthesis and Catalytic Applications. Adv. Synth. Catal. 2015, 357, 283–330.

(91) Bauer, G.; Hu, X. Recent Developments of Iron Pincer Complexes for Catalytic Applications. Inorg. Chem. Front. 2016, 3, 741–765.

(92) Kumar, A.; Bhatti, T. M.; Goldman, A. S. Dehydrogenation of Alkanes and Aliphatic Groups by Pincer-Ligated Metal Complexes. Chem. Rev. 2017, 117, 12357–12384.

(93) Garbe, M.; Junge, K.; Beller, M. Homogeneous Catalysis by Manganese-Based Pincer Complexes. European J. Org. Chem. 2017, 2017, 4344–4362.

(94) Tanaka, R.; Yamashita, M.; Nozaki, K. Catalytic Hydrogenation of Carbon Dioxide Using Ir(III)-Pincer Complexes. J. Am. Chem. Soc. 2009, 131, 14168– 14169.

(95) Arashiba, K.; Miyake, Y.; Nishibayashi, Y. A Molybdenum Complex Bearing PNP-Type Pincer Ligands Leads to the Catalytic Reduction of Dinitrogen into Ammonia. Nat. Chem. 2011, 3, 120–125.

(96) Kang, P.; Meyer, T. J.; Brookhart, M. Selective Electrocatalytic Reduction of Carbon Dioxide to Formate by a Water-Soluble Iridium Pincer Catalyst. Chem. Sci. 2013, 4, 3497.

(97) Filonenko, G. A.; Van Putten, R.; Schulpen, E. N.; Hensen, E. J. M.; Pidko, E. A. Highly Efficient Reversible Hydrogenation of Carbon Dioxide to Formates Using a Ruthenium PNP-Pincer Catalyst. ChemCatChem 2014, 6, 1526–1530.

(98) Scheuermann, M. L.; Semproni, S. P.; Pappas, I.; Chirik, P. J. Carbon Dioxide Hydrosilylation Promoted by Cobalt Pincer Complexes. Inorg. Chem. 2014, 53, 9463–9465.

(99) Rösler, S.; Obenauf, J.; Kempe, R. A Highly Active and Easily Accessible Cobalt Catalyst for Selective Hydrogenation of C=O Bonds. J. Am. Chem. Soc. 2015, 137, 7998–8001.

(100) Mastalir, M.; Glatz, M.; Pittenauer, E.; Allmaier, G.; Kirchner, K. Sustainable Synthesis of Quinolines and Pyrimidines Catalyzed by Manganese PNP Pincer Complexes. J. Am. Chem. Soc. 2016, 138, 15543–15546.

182

(101) Ma, Q.-Q.; Liu, T.; Li, S.; Zhang, J.; Chen, X.; Guan, H.; Yang, Q.; Eberhardt, N. A.; Leahy, K. M.; Krause, J. A.; et al. Highly Efficient Reduction of Carbon Dioxide with a Borane Catalyzed by Bis(Phosphinite) Pincer Ligated Palladium Thiolate Complexes. Chem. Commun. 2016, 52, 14262–14265.

(102) Adams, G. M.; Weller, A. S. POP-Type Ligands: Variable Coordination and Hemilabile Behaviour. Coord. Chem. Rev. 2018, 355, 150–172.

(103) Anaby, A.; Feller, M.; Ben-David, Y.; Leitus, G.; Diskin-Posner, Y.; Shimon, L. J. W.; Milstein, D. Bottom-Up Construction of a CO2-Based Cycle for the Photocarbonylation of Benzene, Promoted by a Rhodium(I) Pincer Complex. J. Am. Chem. Soc. 2016, 138, 9941–9950.

(104) Bertini, F.; Glatz, M.; Gorgas, N.; Stöger, B.; Peruzzini, M.; Veiros, L. F.; Kirchner, K.; Gonsalvi, L. Carbon Dioxide Hydrogenation Catalysed by Well- Defined Mn(I) PNP Pincer Hydride Complexes. Chem. Sci. 2017, 8, 5024–5029.

(105) Benito-Garagorri, D.; Kirchner, K. Modularly Designed Transition Metal PNP and PCP Pincer Complexes Based on Aminophosphines: Synthesis and Catalytic Applications. Acc. Chem. Res. 2008, 41, 201–213.

(106) Huang, Z.; Brookhart, M.; Goldman, A. S.; Kundu, S.; Ray, A.; Scott, S. L.; Vicente, B. C. Highly Active and Recyclable Heterogeneous Iridium Pincer Catalysts for Transfer Dehydrogenation of Alkanes. Adv. Synth. Catal. 2009, 351, 188–206.

(107) Rimoldi, M.; Nakamura, A.; Vermeulen, N. A.; Henkelis, J. J.; Blackburn, A. K.; Hupp, J. T.; Stoddart, J. F.; Farha, O. K. A Metal–organic Framework Immobilised Iridium Pincer Complex. Chem. Sci. 2016, 7, 4980–4984.

(108) Rimoldi, M.; Fodor, D.; van Bokhoven, J. A.; Mezzetti, A. A Stable 16-Electron Iridium(Iii) Hydride Complex Grafted on SBA-15: A Single-Site Catalyst for Alkene Hydrogenation. Chem. Commun. 2013, 49, 11314.

(109) Huang, E. K.; Cheung, W.-M.; Chan, K.-W.; Lam, F. L.-Y.; Hu, X.; Zhang, Q.-F.; Williams, I. D.; Leung, W.-H. Alkylruthenium Complexes Containing Polypyridyl Ligands: Synthesis, Characterization, and Immobilization on Silica. Eur. J. Inorg. Chem. 2013, 2013, 2893–2899.

(110) Ternel, J.; Delevoye, L.; Agbossou-Niedercorn, F.; Roisnel, T.; Gauvin, R. M.; Thomas, C. M. Design of a Well-Defined, Silica-Supported Chiral Zn Scaffold for Enantioselective Catalysis. Dalt. Trans. 2010, 39, 3802.

(111) He, J.; Waggoner, N. W.; Dunning, S. G.; Steiner, A.; Lynch, V. M.; Humphrey, S. M. A PCP Pincer Ligand for Coordination Polymers with Versatile Chemical 183

Reactivity: Selective Activation of CO 2 Gas over CO Gas in the Solid State. Angew. Chem. Int. Ed. 2016, 55, 12351–12355.

(112) He, J.; Bohnsack, A. M.; Waggoner, N. W.; Dunning, S. G.; Lynch, V. M.; Kaska, W. C.; Humphrey, S. M. 1-D and 2-D Phosphine Coordination Materials Based on a Palladium(II) PCP Pincer Metalloligand. Polyhedron 2018, 143, 149–156.

(113) Burgess, S. A.; Kassie, A.; Baranowski, S. A.; Fritzsching, K. J.; Schmidt-Rohr, K.; Brown, C. M.; Wade, C. R. Improved Catalytic Activity and Stability of a Palladium Pincer Complex by Incorporation into a Metal-Organic Framework. J. Am. Chem. Soc. 2016, 138, 1780–1783.

(114) Reiner, B. R.; Mucha, N. T.; Rothstein, A.; Temme, J. S.; Duan, P.; Schmidt-Rohr, K.; Foxman, B. M.; Wade, C. R. Zirconium Metal–Organic Frameworks Assembled from Pd and Pt P N N N P Pincer Complexes: Synthesis, Postsynthetic Modification, and Lewis Acid Catalysis. Inorg. Chem. 2018, 57, 2663–2672.

(115) Hartwig, J. F. Organotransition Metal Chemistry: From Bonding to Catalysis; University Science Books: Mill Valley, 2010.

(116) Schmidbaur, H.; Jandik, P. Au(I)-Au(II) and Au(I)-Au(III) Oxidation in Ylide Complexes with Chlorinating Agents. Inorganica Chim. Acta 1983, 74, 97–99.

(117) Schmidbaur, H.; Hartmann, C. Methylen-Verbrückte Zweikernige Gold(III)- Komplexe Mit Terminalen Und Verbrückenden Ylid-Liganden. Angew. Chemie 1986, 98, 573–574.

(118) Schmidbaur, H.; Hartmann, C.; Reber, G.; Müller, G. Isovalente Und Gemischtvalente Ylid-Komplexe Des Golds: Aufbau Dreikerniger Verbindungen Mit Doppelschaufelradstruktur. Angew. Chemie 1987, 99, 1189–1191.

(119) Basil, J. D.; Murray, H. H.; Fackler, J. P.; Tocher, J.; Mazany, A. M.; Trzcinska- Bancroft, B.; Knachel, H.; Dudis, D.; Delord, T. J.; Marler, D. Experimental and Theoretical Studies of Dinuclear Gold(I) and Gold(II) Phosphorus Ylide Complexes. Oxidative Addition, Halide Exchange, and Structural Properties Including the Crystal and Molecular Structures of [Au(CH2)2PPh2]2 and [Au(CH2)2PPh2]2(CH3)Br. J. Am. Chem. Soc. 1985, 107, 6908–6915.

(120) Laguna, A.; Laguna, M. Coordination Chemistry of Gold(II) Complexes. Coord. Chem. Rev. 1999, 193–195, 837–856.

(121) Heinrich, D. D.; Staples, R. J.; Fackler, J. P. Activation of Dinuclear Gold(I) Ylide Complexes by Lewis Acids. Isomerization of Trans-[Au(CH2)2PPhMe]2, and the Crystal Structures of [Au(CH2)2PPhMe]2, [Au(CH2)2PPhMe]2[SO2]2, [Au(CH2)2PPh2]2[SO2]2 and [Au(CH2)2PPhMe]2[S2CNEt2]2. Inorganica 184

Chim. Acta 1995, 229, 61–75.

(122) Fackler, J. P.; Trzcinska-Bancroft, B. Isomerization of a Symmetrical Metal-Metal Bonded Gold(II) Ylide Dimer to a Mixed-Valence Gold(III)/Gold(I) Species. Organometallics 1985, 4, 1891–1893.

(123) Komiya, S.; Albright, T. A.; Hoffmann, R.; Kochi, J. K. Reductive Elimination and Isomerization of Organogold Complexes. Theoretical Studies of Trialkylgold Species as Reactive Intermediates. J. Am. Chem. Soc. 1976, 98, 7255–7265.

(124) Müller, T. E.; Green, J. C.; Mingos, D. M. P.; McPartlin, C. M.; Whittingham, C.; Williams, D. J.; Woodroffe, T. M. Complexes of Gold(I) and Platinum(II) with Polyaromatic Phosphine Ligands. J. Organomet. Chem. 1998, 551, 313–330.

(125) Bardají, M.; Gimeno, M. C.; Jiménez, J.; Laguna, A.; Laguna, M.; Jones, P. G. Neutral or Cationic (µ-Methylene)Bisylide DigoldIII Complexes. J. Organomet. Chem. 1992, 441, 339–348.

(126) Gorin, D. J.; Sherry, B. D.; Toste, F. D. Ligand Effects in Homogeneous Au Catalysis. Chem. Rev. 2008, 108, 3351–3378.

(127) Roşca, D.-A.; Wright, J. A.; Bochmann, M. An Element through the Looking Glass: Exploring the Au–C, Au–H and Au–O Energy Landscape. Dalt. Trans. 2015, 44, 20785–20807.

(128) Hashmi, A. S. K.; Dr, G. J. H. Gold Catalysis. Angew. Chem. Int. Ed. 2006, 45.

(129) Pflästerer, D.; Hashmi, A. S. K. Gold Catalysis in Total Synthesis – Recent Achievements. Chem. Soc. Rev. 2016, 45, 1331–1367.

(130) Dorel, R.; Echavarren, A. M. Gold(I)-Catalyzed Activation of Alkynes for the Construction of Molecular Complexity. Chem. Rev. 115, 9028–9072.

(131) Shen, H. C. Recent Advances in Syntheses of Heterocycles and Carbocycles via Homogeneous Gold Catalysis. Part 1: Heteroatom Addition and Hydroarylation Reactions of Alkynes, Allenes, and Alkenes. Tetrahedron 2008, 64, 3885–3903.

(132) de Frémont, P.; Singh, R.; Stevens, E. D.; Petersen, J. L.; Nolan, S. P. Synthesis, Characterization and Reactivity of N -Heterocyclic Carbene Gold(III) Complexes. Organometallics 2007, 26, 1376–1385.

(133) Sromek, A. W.; Rubina, M.; Gevorgyan, V. 1,2-Halogen Migration in Haloallenyl Ketones: Regiodivergent Synthesis of Halofurans. J. Am. Chem. Soc. 127, 10500– 10501.

(134) Morita, N.; Yasuda, A.; Shibata, M.; Ban, S.; Hashimoto, Y.; Okamoto, I.; 185

Tamura, O. Gold(I)/(III)-Catalyzed Synthesis of Cyclic Ethers; Valency- Controlled Cyclization Modes. Org. Lett. 2015, 17, 2668–2671.

(135) Reetz, M. T.; Sommer, K. Gold-Catalyzed Hydroarylation of Alkynes. European J. Org. Chem. 2003, 2003, 3485–3496.

(136) Wu, C.-Y.; Horibe, T.; Jacobsen, C. B.; Toste, F. D. Stable Gold(III) Catalysts by Oxidative Addition of a Carbon–carbon Bond. Nature 2015, 517, 449–454.

(137) Mirzadeh, N.; Bennett, M. A.; Bhargava, S. K. Cycloaurated Complexes of Aryl Carbanions: Digold(I), Digold(II) and Beyond. Coord. Chem. Rev. 2013, 257, 2250–2273.

(138) Levin, M. D.; Toste, F. D. Gold-Catalyzed Allylation of Aryl Boronic Acids: Accessing Cross-Coupling Reactivity with Gold. Angew. Chem. Int. Ed. 2014, 53, 6211–6215.

(139) Schmidbaur, H.; Franke, R. Phosphorus Ylides as Structural Units in Organogold Compounds. Angew. Chemie Int. Ed. English 1973, 12, 416–417.

(140) Wolf, W. J.; Winston, M. S.; Toste, F. D. Exceptionally Fast Carbon–carbon Bond Reductive Elimination from Gold(III). Nat. Chem. 2014, 6, 159–164.

(141) Kawai, H.; Wolf, W. J.; Dipasquale, A. G.; Winston, M. S.; Toste, F. D. Phosphonium Formation by Facile Carbon-Phosphorus Reductive Elimination from Gold(III). J. Am. Chem. Soc. 2016, 138, 587–593.

(142) Hashmi, A. S. K.; Blanco, M. C.; Fischer, D.; Bats, J. W. Gold Catalysis: Evidence for the In-Situ Reduction of Gold(III) During the Cyclization of Allenyl Carbinols. European J. Org. Chem. 2006, 2006, 1387–1389.

(143) Laguna, A.; Laguna, M.; Jimenez, J.; Lahoz, F. J.; Olmos, E. Mixed-Valent Linear Chains of Gold Atoms. X-Ray Structure of [{(2,4,6- C6F3H2)Au(CH2PPh2CH2)2Au}2Au(CH2PPh2CH2)2Au](ClO4)2. Organometallics 1994, 13, 253–257.

(144) Porter, L. C.; Knachel, H.; Fackler Jnr, J. P. Structure of a Mononuclear Gold(I) Complex Containing a Covalently Bound Ylide Ligand. Acta Crystallogr. Sect. C Cryst. Struct. Commun. 1986, 42, 1125–1128.

(145) Hayoun, R.; Zhong, D. K.; Rheingold, A. L.; Doerrer, L. H. Gold(III) and Platinum(II) Polypyridyl Double Salts and a General Metathesis Route to Metallophilic Interactions. Inorg. Chem. 45, 6120–6122.

(146) Doerrer, L. H. Steric and Electronic Effects in Metallophilic Double Salts. Dalton Trans. 39, 3543–3553. 186

(147) Klapötke, T. M.; Krumm, B.; Galvez-Ruiz, J.-C.; Nöth, H. Highly Sensitive Ammonium Tetraazidoaurates(III). Inorg. Chem. 2005, 44, 9625–9627.

(148) Cordón, J.; López-de-Luzuriaga, J. M.; Monge, M. Experimental and Theoretical Study of the Effectiveness and Stability of Gold(I) Catalysts Used in the Synthesis of Cyclic Acetals. Organometallics 2016, 35, 732–740.

(149) Kumar, M.; Hammond, G. B.; Xu, B. Cationic Gold Catalyst Poisoning and Reactivation. Org. Lett. 2014, 16, 3452–3455.

(150) Tlahuext-Aca, A.; Hopkinson, M. N.; Daniliuc, C. G.; Glorius, F. Oxidative Addition to Gold(I) by Photoredox Catalysis: Straightforward Access to Diverse ( C , N )-Cyclometalated Gold(III) Complexes. Chem. - A Eur. J. 2016, 22, 11587– 11592.

(151) Yang, H.; Gabbaı̈ , F. P. Activation of a Hydroamination Gold Catalyst by Oxidation of a Redox-Noninnocent Chlorostibine Z-Ligand. J. Am. Chem. Soc. 2015, 137, 13425–13432.

(152) Lavallo, V.; Nd, W. J.; Tham, F. S.; Quinlivan, S. Perhalogenated Carba-Closo- Dodecaborate Anions as Ligand Substituents: Applications in Gold Catalysis. Angew. Chemie 2013, 125.

(153) Katari, M.; Rao, M. N.; Rajaraman, G.; Ghosh, P. Computational Insight into a Gold(I) N-Heterocyclic Carbene Mediated Alkyne Hydroamination Reaction. Inorg. Chem. 51, 5593–5604.

(154) Zeng, X.; Soleilhavoup, M.; Bertrand, G. Gold-Catalyzed Intermolecular Markovnikov Hydroamination of Allenes with Secondary Amines. Org. Lett. 2009, 11, 3166–3169.

(155) Lavallo, V.; Frey, G. D.; Donnadieu, B.; Soleilhavoup, M.; Bertrand, G. Homogeneous Catalytic Hydroamination of Alkynes and Allenes with Ammonia. Angew. Chem. Int. Ed. 2008, 47, 5224–5228.

(156) Khin, C.; Hashmi, A. S. K.; Rominger, F. Gold(I) Complexes of P,N Ligands and Their Catalytic Activity. Eur. J. Inorg. Chem. 2010, 2010, 1063–1069.

(157) Zhdanko, A.; Maier, M. E. Mechanistic Study of Gold(I)-Catalyzed Hydroamination of Alkynes: Outer or Inner Sphere Mechanism? Angew. Chem. Int. Ed. 2014, 53, 7760–7764.

(158) Poater, A.; Cosenza, B.; Correa, A.; Giudice, S.; Ragone, F.; Scarano, V.; Cavallo, L. SambVca: A Web Application for the Calculation of the Buried Volume of N- Heterocyclic Carbene Ligands. Eur. J. Inorg. Chem. 2009, No. 13 SPEC. ISS.,

187

1759–1766.

(159) Dyker, G.; Muth, E.; Hashmi, A. S. K.; Ding, L. Gold(III) Chloride-Catalyzed Addition Reactions of Electron-Rich Arenes to Methyl Vinyl Ketone. Adv. Synth. Catal. 2003, 345, 1247–1252.

(160) Hashmi, A. S. K.; Ramamurthi, T. D.; Rominger, F. On the Trapping of Vinylgold Intermediates. Adv. Synth. Catal. 2010, 352, 971–975.

(161) Hashmi, A. S. K.; Schwarz, L.; Choi, J.-H.; Frost, T. M. A New Gold-Catalyzed C−C Bond Formation. Angew. Chem. Int. Ed. 2000, 39, 2285–2288.

(162) Lin, J.-M.; Liu, B.-S. A Practical Method for the Preparation of Trimethylsilyl Enol Ethers. Synth. Commun. 1997, 27, 739–749.

(163) Collado, A.; Gómez-Suárez, A.; Martin, A. R.; Slawin, A. M.; Nolan, S. P. Straightforward Synthesis of [Au(NHC)X] (NHC = N-Heterocyclic Carbene, X = Cl, Br, I) Complexes. Chem. Commun. 2013, 49, 5541–5543.

(164) Smallcombe, S. H.; Patt, S. L.; Keifer, P. A. WET Solvent Suppression and Its Applications to LC NMR and High-Resolution NMR Spectroscopy. J. Magn. Reson. Ser. A 1995, 117, 295–303.

(165) Méndez, L. A.; Jiménez, J.; Cerrada, E.; Mohr, F.; Laguna, M. A Family of Alkynylgold(III) Complexes [Au I (µ-{CH 2 } 2 PPh 2 ) 2 Au III (C⋮CR) 2 ] (R = Ph, t Bu, Me 3 Si): Facile and Reversible Comproportionation of Gold(I)/Gold(III) to Digold(II). J. Am. Chem. Soc. 2005, 127, 852–853.

(166) Wang, D.; Cai, R.; Sharma, S.; Jirak, J.; Thummanapelli, S. K. “Silver Effect” in Gold(I) Catalysis: An Overlooked Important Factor. J. Am. Chem. Soc. 2012, 134.

(167) Homs, A.; Escofet, I.; Echavarren, A. M. On the Silver Effect and the Formation of Chloride-Bridged Digold Complexes. Org. Lett. 2013, 15, 5782–5785.

(168) Lu, Z.; Han, J.; Hammond, G. B.; Xu, B. Revisiting the Influence of Silver in Cationic Gold Catalysis: A Practical Guide. Org. Lett. 2015, 17, 4534–4537.

(169) Grajeda, J.; Kita, M. R.; Gregor, L. C.; White, P. S.; Miller, A. J. M. Diverse Cation-Promoted Reactivity of Iridium Carbonyl Pincer-Crown Ether Complexes. Organometallics 2016, 35, 306–316.

(170) Reath, A. H.; Ziller, J. W.; Tsay, C.; Ryan, A. J.; Yang, J. Y. Redox Potential and Electronic Structure Effects of Proximal Nonredox Active Cations in Cobalt Schiff Base Complexes. Inorg. Chem. 2017, 56, 3713–3718.

(171) Bae, S. H.; Lee, Y.-M.; Fukuzumi, S.; Nam, W. Fine Control of the Redox 188

Reactivity of a Nonheme Iron(III)-Peroxo Complex by Binding Redox-Inactive Metal Ions. Angew. Chem. Int. Ed. 2017, 56, 801–805.

(172) Delgado, M.; Ziegler, J. M.; Seda, T.; Zakharov, L. N.; Gilbertson, J. D. Pyridinediimine Iron Complexes with Pendant Redox-Inactive Metals Located in the Secondary Coordination Sphere. Inorg. Chem. 2016, 55, 555–557.

(173) Kakuda, S.; Rolle, C. J.; Ohkubo, K.; Siegler, M. A.; Karlin, K. D.; Fukuzumi, S. Lewis Acid-Induced Change from Four- to Two-Electron Reduction of Dioxygen Catalyzed by Copper Complexes Using Scandium Triflate. J. Am. Chem. Soc. 2015, 137, 3330–3337.

(174) Park, J.; Morimoto, Y.; Lee, Y.-M.; Nam, W.; Fukuzumi, S. Unified View of Oxidative C–H Bond Cleavage and Sulfoxidation by a Nonheme Iron(IV)–Oxo Complex via Lewis Acid-Promoted Electron Transfer. Inorg. Chem. 2014, 53, 3618–3628.

(175) Dolinar, B. S.; Berry, J. F. Lewis Acid Enhanced Axial Ligation of [Mo 2 ] 4+ Complexes. Inorg. Chem. 2013, 52, 4658–4667.

(176) Herbert, D. E.; Lionetti, D.; Rittle, J.; Agapie, T. Heterometallic Triiron- Oxo/Hydroxo Clusters: Effect of Redox-Inactive Metals. J. Am. Chem. Soc. 2013, 135, 19075–19078.

(177) Yoon, H.; Lee, Y.-M.; Wu, X.; Cho, K.-B.; Sarangi, R.; Nam, W.; Fukuzumi, S. Enhanced Electron-Transfer Reactivity of Nonheme Manganese(IV)–Oxo Complexes by Binding Scandium Ions. J. Am. Chem. Soc. 2013, 135, 9186–9194.

(178) Tsui, E. Y.; Tran, R.; Yano, J.; Agapie, T. Redox-Inactive Metals Modulate the Reduction Potential in Heterometallic Manganese–oxido Clusters. Nat. Chem. 2013, 5, 293–299.

(179) Morimoto, Y.; Park, J.; Suenobu, T.; Lee, Y.-M.; Nam, W.; Fukuzumi, S. Mechanistic Borderline of One-Step Hydrogen Atom Transfer versus Stepwise Sc 3+ -Coupled Electron Transfer from Benzyl Alcohol Derivatives to a Non-Heme Iron(IV)-Oxo Complex. Inorg. Chem. 2012, 51, 10025–10036.

(180) Fukuzumi, S.; Morimoto, Y.; Kotani, H.; Naumov, P.; Lee, Y.-M.; Nam, W. Crystal Structure of a Metal Ion-Bound Oxoiron(IV) Complex and Implications for Biological Electron Transfer. Nat. Chem. 2010, 2, 756–759.

(181) Fukuzumi, S.; Ohkubo, K. Metal Ion-Coupled and Decoupled Electron Transfer. Coord. Chem. Rev. 2010, 254, 372–385.

(182) Fukuzumi, S.; Ohkubo, K.; Morimoto, Y. Mechanisms of Metal Ion-Coupled

189

Electron Transfer. Phys. Chem. Chem. Phys. 2012, 14, 8472.

(183) Pletcher, D.; Thompson, H. A Microelectrode Study of the Influence of Electrolyte on the Reduction of Quinones in Aprotic Solvents. J. Chem. Soc. Faraday Trans. 1998, 94, 3445–3450.

(184) Oyama, M.; Marken, F.; Webster, R. D.; Cooper, J. A.; Compton, R. G.; Okazaki, S. Ion Pair Formation between the Electrogenerated 2,3-Dichloro-5,6-Dicyano- 1,4-Benzoquinone Dianion and the Sodium Ion at Platinum Surfaces. J. Electroanal. Chem. 1998, 451, 193–201.

(185) Nagaoka, T.; Okazaki, S.; Fujinaga, T. Ion-Pair Effects on the Electroreduction of Carbonyl Compounds in N,-Dimethylformamide. J. Electroanal. Chem. Interfacial Electrochem. 1982, 133, 89–99.

(186) Wain, A. J.; Wildgoose, G. G.; Heald, C. G. R.; Jiang, L.; Jones, T. G. J.; Compton, R. G. Electrochemical ESR and Voltammetric Studies of Lithium Ion Pairing with Electrogenerated 9,10-Anthraquinone Radical Anions Either Free in Acetonitrile Solution or Covalently Bound to Multiwalled Carbon Nanotubes. J. Phys. Chem. B 2005, 109, 3971–3978.

(187) Bonn, A. G.; Wenger, O. S. Photoinduced Charge Accumulation by Metal Ion- Coupled Electron Transfer. Phys. Chem. Chem. Phys. 2015, 17, 24001–24010.

(188) Hernández-Burgos, K.; Rodríguez-Calero, G. G.; Zhou, W.; Burkhardt, S. E.; Abruña, H. D. Increasing the Gravimetric Energy Density of Organic Based Secondary Battery Cathodes Using Small Radius Cations (Li + and Mg 2+ ). J. Am. Chem. Soc. 2013, 135, 14532–14535.

(189) Song, Z.; Zhan, H.; Zhou, Y. Polyimides: Promising Energy-Storage Materials. Angew. Chem. Int. Ed. 2010, 49, 8444–8448.

(190) Kim, D. J.; Je, S. H.; Sampath, S.; Choi, J. W.; Coskun, A. Effect of N- Substitution in Naphthalenediimides on the Electrochemical Performance of Organic Rechargeable Batteries. RSC Adv. 2012, 2, 7968.

(191) Vadehra, G. S.; Maloney, R. P.; Garcia-Garibay, M. A.; Dunn, B. Naphthalene Diimide Based Materials with Adjustable Redox Potentials: Evaluation for Organic Lithium-Ion Batteries. Chem. Mater. 2014, 26, 7151–7157.

(192) Chen, D.; Avestro, A.-J.; Chen, Z.; Sun, J.; Wang, S.; Xiao, M.; Erno, Z.; Algaradah, M. M.; Nassar, M. S.; Amine, K.; et al. A Rigid Naphthalenediimide Triangle for Organic Rechargeable Lithium-Ion Batteries. Adv. Mater. 2015, 27, 2907–2912.

190

(193) Hernández, G.; Casado, N.; Coste, R.; Shanmukaraj, D.; Rubatat, L.; Armand, M.; Mecerreyes, D. Redox-Active Polyimide–polyether Block Copolymers as Electrode Materials for Lithium Batteries. RSC Adv. 2015, 5, 17096–17103.

(194) Xu, F.; Jin, S.; Zhong, H.; Wu, D.; Yang, X.; Chen, X.; Wei, H.; Fu, R.; Jiang, D. Electrochemically Active, Crystalline, Mesoporous Covalent Organic Frameworks on Carbon Nanotubes for Synergistic Lithium-Ion Battery Energy Storage. Sci. Rep. 2015, 5, 8225.

(195) Zeigler, D. F.; Candelaria, S. L.; Mazzio, K. A.; Martin, T. R.; Uchaker, E.; Suraru, S.-L.; Kang, L. J.; Cao, G.; Luscombe, C. K. N-Type Hyperbranched Polymers for Supercapacitor Cathodes with Variable Porosity and Excellent Electrochemical Stability. Macromolecules 2015, 48, 5196–5203.

(196) Veerababu, M.; Varadaraju, U. .; Kothandaraman, R. Reversible Lithium Storage Behaviour of Aromatic Diimide Dilithium Carboxylates. Electrochim. Acta 2016, 193, 80–87.

(197) Chen, M.; Yang, C.; Xu, Z.; Tang, Y.; Jiang, J.; Liu, P.; Su, Y.; Wu, D. A Facile Self-Assembly Strategy towards Naphthalene Diimide/Graphene Hybrids as High Performance Organic Cathodes for Lithium-Ion Batteries. RSC Adv. 2016, 6, 13666–13669.

(198) Schon, T. B.; McAllister, B. T.; Li, P.-F.; Seferos, D. S. The Rise of Organic Electrode Materials for Energy Storage. Chem. Soc. Rev. 2016, 45, 6345–6404.

(199) Okamoto, K.; Mori, Y.; Yamada, H.; Imahori, H.; Fukuzumi, S. Effects of Metal Ions on Photoinduced Electron Transfer in Zinc Porphyrin?Naphthalenediimide Linked Systems. Chem. - A Eur. J. 2004, 10, 474–483.

(200) Fukuzumi, S.; Ohkubo, K.; Ortiz, J.; Gutiérrez, A. M.; Fernández-Lázaro, F.; Sastre-Santos, Á. Formation of a Long-Lived Charge-Separated State of a Zinc Phthalocyanine-Perylenediimide Dyad by Complexation with Magnesium Ion. Chem. Commun. 2005, No. 30, 3814.

(201) Tan, L.; Zhang, G.; Zhang, D.; Zhu, D. Linear and Cyclic Tetrathiafulvalene– Naphthalenediimide Donor–Acceptor Molecules: Metal Ions-Promoted Electron Transfer. J. Org. Chem. 2011, 76, 9046–9052.

(202) Pan, M.; Lin, X.-M.; Li, G.-B.; Su, C.-Y. Progress in the Study of Metal–organic Materials Applying Naphthalene Diimide (NDI) Ligands. Coord. Chem. Rev. 2011, 255, 1921–1936.

(203) Jiang, J.-J.; Yan, C.; Pan, M.; Wang, Z.; Deng, H.-Y.; He, J.-R.; Yang, Q.-Y.; Fu, L.; Xu, X.-F.; Su, C.-Y. Structural Conformation and Optical and Electrochemical 191

Properties of Imidazolyl-Substituted Naphthalenediimide and Its HgII, CdII, and CuII Halide Complexes. Eur. J. Inorg. Chem. 2012, 2012, 1171–1179.

(204) DeBlase, C. R.; Hernández-Burgos, K.; Rotter, J. M.; Fortman, D. J.; dos S. Abreu, D.; Timm, R. A.; Diógenes, I. C. N.; Kubota, L. T.; Abruña, H. D.; Dichtel, W. R. Cation-Dependent Stabilization of Electrogenerated Naphthalene Diimide Dianions in Porous Polymer Thin Films and Their Application to Electrical Energy Storage. Angew. Chem. Int. Ed. 2015, 54, 13225–13229.

(205) de Miguel, M.; Álvaro, M.; García, H.; Céspedes-Guirao, F. J.; Fernández-Lázaro, F.; Sastre-Santos, Á. Selective Photoinduced Single or Double Electron Reduction of Perylenebisimides. J. Photochem. Photobiol. A Chem. 2012, 231, 28–32.

(206) Skaisgirski, M.; Guo, X.; Wenger, O. S. Electron Accumulation on Naphthalene Diimide Photosensitized by [Ru(2,2′-Bipyridine) 3 ] 2+. Inorg. Chem. 2017, 56, 2432–2439.

(207) Pellegrin, Y.; Odobel, F. Molecular Devices Featuring Sequential Photoinduced Charge Separations for the Storage of Multiple Redox Equivalents. Coord. Chem. Rev. 2011, 255, 2578–2593.

(208) Hammarström, L. Accumulative Charge Separation for Solar Fuels Production: Coupling Light-Induced Single Electron Transfer to Multielectron Catalysis. Acc. Chem. Res. 2015, 48, 840–850.

(209) Okoshi, M.; Yamada, Y.; Yamada, A.; Nakai, H. Theoretical Analysis on De- Solvation of Lithium, Sodium, and Magnesium Cations to Organic Electrolyte Solvents. J. Electrochem. Soc. 2013, 160, A2160–A2165.

(210) Okoshi, M.; Yamada, Y.; Komaba, S.; Yamada, A.; Nakai, H. Theoretical Analysis of Interactions between Potassium Ions and Organic Electrolyte Solvents: A Comparison with Lithium, Sodium, and Magnesium Ions. J. Electrochem. Soc. 2017, 164, A54–A60.

(211) Fukuzumi, S.; Nishizawa, N.; Tanaka, T. Effects of Magnesium(II) Ion on Hydride-Transfer Reactions from an NADH Model Compound to p-Benzoquinone Derivatives. The Quantitative Evaluation Based on the Reaction Mechanism. J. Chem. Soc. Perkin Trans. 2 1985, No. 3, 371.

(212) Lingane, J. J. Interpretation of the Polarographic Waves of Complex Metal Ions. Chem. Rev. 1941, 29, 1–35.

(213) Nesset, M. J. M.; Shokhirev, N. V.; Enemark, P. D.; Jacobson, S. E.; Walker, F. A. Models of the Cytochromes. Redox Properties and Thermodynamic Stabilities of Complexes of “Hindered” Iron(III) and Iron(II) Tetraphenylporphyrinates with 192

Substituted Pyridines and Imidazoles. Inorg. Chem. 1996, 35, 5188–5200.

(214) Nguyen, M.; Phan, T.; Caemelbecke, E. Van; Wei, X.; Bear, J. L.; Kadish, K. M. Synthesis and Characterization of (3,1) Ru 2 (F 3 Ap) 4 (NCS) and (3,1) Ru 2 (F 3 Ap) 3 (F 2 Oap)(NCS) Where F 3 Ap Is the 2-(2,4,6-Trifluoroanilino)Pyridinate Anion. Inorg. Chem. 2008, 47, 4392–4400.

(215) Gosztola, D.; Niemczyk, M. P.; Svec, W.; Lukas, A. S.; Wasielewski, M. R. Excited Doublet States of Electrochemically Generated Aromatic Imide and Diimide Radical Anions. J. Phys. Chem. A 2000, 104, 6545–6551.

(216) Wade, C. R.; Li, M.; Dincă, M. Facile Deposition of Multicolored Electrochromic Metal-Organic Framework Thin Films. Angew. Chem. Int. Ed. 2013, 52, 13377– 13381.

(217) Kumar, S.; Ajayakumar, M. R.; Hundal, G.; Mukhopadhyay, P. Extraordinary Stability of Naphthalenediimide Radical Ion and Its Ultra-Electron-Deficient Precursor: Strategic Role of the Phosphonium Group. J. Am. Chem. Soc. 2014, 136, 12004–12010.

(218) Seifert, S.; Schmidt, D.; Würthner, F. An Ambient Stable Core-Substituted Perylene Bisimide Dianion: Isolation and Single Crystal Structure Analysis. Chem. Sci. 2015, 6, 1663–1667.

(219) Sazanovich, I. V.; Alamiry, M. A. H.; Best, J.; Bennett, R. D.; Bouganov, O. V.; Davies, E. S.; Grivin, V. P.; Meijer, A. J. H. M.; Plyusnin, V. F.; Ronayne, K. L.; et al. Excited State Dynamics of a Pt II Diimine Complex Bearing a Naphthalene- Diimide Electron Acceptor. Inorg. Chem. 2008, 47, 10432–10445.

(220) Zechmann, C. A.; Boyle, T. J.; Rodriguez, M. A.; Kemp, R. A. Solvent Influences on the Molecular Aggregation of Magnesium Aryloxides: Synthesis, Crystal Structure, and Solution Characterization of Mg(OAr)2(L)3 (OAr=DMP, DIP, TCP) and [Mg(DIP)2]3. Polyhedron 2000, 19, 2557–2564.

(221) Roesky, H. W.; Scholz, M.; Noltemeyer, M. Über Reaktionen Des 2,4,6- Tris(Trifluormethyl)Phenols Mit Verbindungen von Hauptgruppen- Und Nebengruppen-Elementen (Li, Na, Mg, Ca, Ba, Ge, Sn Und Ti, W, Mn, Cd). Chem. Ber. 1990, 123, 2303–2309.

(222) Grepioni, F.; Wouters, J.; Braga, D.; Nanna, S.; Fours, B.; Coquerel, G.; Longfils, G.; Rome, S.; Aerts, L.; Quéré, L. Ionic Co-Crystals of Racetams: Solid-State Properties Enhancement of Neutral Active Pharmaceutical Ingredients via Addition of Mg2+ and Ca2+ Chlorides. CrystEngComm 2014, 16, 5887.

(223) Zhang, J.; Liu, Y.; Gong, L.; Xu, G.; Zhang, C. Four Distinct Anionic 193

Heterothiometallic W/S/Ag Polymeric Clusters Directed by Solvent-Coordinated Metal Cations. Polyhedron 2016, 109, 67–74.

(224) Wu, X.-X.; Zhang, S.-L.; Ng, S. W. 2,7-Bis(2,6-Diisopropylphenyl)Benzo[ Lmn ][3,8]Phenanthroline-1,3,6,8(2 H ,7 H )-Tetraone. Acta Crystallogr. Sect. E Struct. Reports Online 2011, 67, o3173–o3173.

(225) Tong, L. H.; Pengo, P.; Clegg, W.; Lowe, J. P.; Raithby, P. R.; Sanders, J. K. M.; Pascu, S. I. Complexes of Aryl-Substituted Porphyrins and Naphthalenediimide (NDI): Investigations by Synchrotron X-Ray Diffraction and NMR Spectroscopy. Dalt. Trans. 2011, 40, 10833.

(226) Lee, S. K.; Zu, Y.; Herrmann, A.; Geerts, Y.; Müllen, K.; Bard, A. J. Electrochemistry, Spectroscopy and Electrogenerated Chemiluminescence of Perylene, Terrylene, and Quaterrylene Diimides in Aprotic Solution. J. Am. Chem. Soc. 1999, 121, 3513–3520.

(227) Viehbeck, A. Electrochemical Properties of Polyimides and Related Imide Compounds. J. Electrochem. Soc. 1990, 137, 1460.

(228) Apex2. Bruker Analytical X-ray Systems: Madison, WI 2006.

(229) M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Had, J. C. and D. J. F. Gaussian 09. Gaussian, Inc.: Wallingford, CT 2009.

(230) Becke, A. D. Density-functional Thermochemistry. III. The Role of Exact Exchange. J. Chem. Phys. 1993, 98, 5648–5652.

(231) Lee, C.; Yang, W.; Parr, R. G. Development of the Colle-Salvetti Correlation- Energy Formula into a Functional of the Electron Density. Phys. Rev. B 1988, 37, 785–789.

(232) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer, P. V. R. Efficient Diffuse Function-Augmented Basis Sets for Anion Calculations. III. The 3-21+G Basis Set for First-Row Elements, Li-F. J. Comput. Chem. 1983, 4, 294–301.

(233) Frisch, M. J.; Pople, J. A.; Binkley, J. S. Self-consistent Molecular Orbital Methods 25. Supplementary Functions for Gaussian Basis Sets. J. Chem. Phys. 1984, 80, 3265–3269.

(234) Hariharan, P. C.; Pople, J. A. The Influence of Polarization Functions on Molecular Orbital Hydrogenation Energies. Theor. Chim. Acta 1973, 28, 213–222.

194

(235) Hariharan, P. C.; Pople, J. A. Accuracy of AH n Equilibrium Geometries by Single Determinant Molecular Orbital Theory. Mol. Phys. 1974, 27, 209–214.

(236) Hehre, W. J.; Ditchfield, R.; Pople, J. A. Self—Consistent Molecular Orbital Methods. XII. Further Extensions of Gaussian—Type Basis Sets for Use in Molecular Orbital Studies of Organic Molecules. J. Chem. Phys. 1972, 56, 2257– 2261.

(237) Zee, D. Z.; Chantarojsiri, T.; Long, J. R.; Chang, C. J. Metal–Polypyridyl Catalysts for Electro- and Photochemical Reduction of Water to Hydrogen. Acc. Chem. Res. 2015, 48, 2027–2036.

(238) Heinemann, F.; Karges, J.; Gasser, G. Critical Overview of the Use of Ru(II) Polypyridyl Complexes as Photosensitizers in One-Photon and Two-Photon Photodynamic Therapy. Acc. Chem. Res. 2017, 50, 2727–2736.

(239) Queyriaux, N.; Jane, R. T.; Massin, J.; Artero, V.; Chavarot-Kerlidou, M. Recent Developments in Hydrogen Evolving Molecular Cobalt(II)–polypyridyl Catalysts. Coord. Chem. Rev. 2015, 304–305, 3–19.

(240) Frija, L. M. T.; Pombeiro, A. J. L.; Kopylovich, M. N. Coordination Chemistry of Thiazoles, Isothiazoles and Thiadiazoles. Coord. Chem. Rev. 2016, 308, 32–55.

(241) Rice, C. R.; Wörl, S.; Jeffery, J. C.; Paul, R. L.; Ward, M. D. New Multidentate Ligands for Supramolecular Coordination Chemistry: Double and Triple Helical Complexes of Ligands Containing Pyridyl and Thiazolyl Donor Units. J. Chem. Soc. Dalt. Trans. 2001, No. 5, 550–559.

(242) Huxel, T.; Leone, S.; Lan, Y.; Demeshko, S.; Klingele, J. 2-Amino-4-(2- Pyridyl)Thiazole as Chelating Ligand: A Dinuclear Oxido-Bridged Ferric Complex and Mononuclear 3d Metal Complexes. Eur. J. Inorg. Chem. 2014, 2014, 3114–3124.

(243) Tsang, C.-S.; Yee, C.-C.; Yiu, S.-M.; Wong, W.-T.; Kwong, H.-L. Single-, Double- and Triple-Stranded Helical Metal Complexes of Chiral Tetradentate Pyridyl-Thiazole Ligands: Stereoselective Synthesis, Spectroscopic Properties and Structural Characterization. Polyhedron 2014, 83, 167–177.

(244) Luconi, L.; Lyubov, D. M.; Rossin, A.; Glukhova, T. A.; Cherkasov, A. V.; Tuci, G.; Fukin, G. K.; Trifonov, A. A.; Giambastiani, G. Organolanthanide Complexes Supported by Thiazole-Containing Amidopyridinate Ligands: Synthesis, Characterization, and Catalytic Activity in Isoprene Polymerization. Organometallics 2014, 33, 7125–7134.

(245) Menzel, R.; Kupfer, S.; Mede, R.; Weiß, D.; Görls, H.; González, L.; Beckert, R. 195

Arylamine-Modified Thiazoles as Donor-Acceptor Dyes: Quantum Chemical Evaluation of the Charge-Transfer Process and Testing as Ligands in Ruthenium(II) Complexes. European J. Org. Chem. 2012, 2012, 5231–5247.

(246) Menzel, R.; Weiß, D.; Täuscher, E.; Beckert, R.; Görls, H. Formation, X-Ray Structure and Magnetic Properties of Two Carboxy-Bridged Copper(II) Complexes with Tridentate Bipyridine- and Phenanthroline-Based 1,3-Thiazoles Ligands. Inorg. Chem. Commun. 2012, 18, 65–68.

(247) Carlson, L. J.; Welby, J.; Zebrowski, K. A.; Wilk, M. M.; Giroux, R.; Ciancio, N.; Tanski, J. M.; Bradley, A.; Tyler, L. A. Spectroscopic Differences between Heterocyclic Benzothiazoline, -Thiazole and Imine Containing Ligands and Comparison of the Co and Cu Pyridine Benzothiazole and Imine Complexes. Inorganica Chim. Acta 2011, 365, 159–166.

(248) Arora, H.; Philouze, C.; Jarjayes, O.; Thomas, F. CoII, NiII, CuII and ZnII Complexes of a Bipyridine Bis-Phenol Conjugate: Generation and Properties of Coordinated Radical Species. Dalt. Trans. 2010, 39, 10088.

(249) Grice, K. A.; Kubiak, C. P. Recent Studies of Rhenium and Manganese Bipyridine Carbonyl Catalysts for the Electrochemical Reduction of CO2; 2014; pp 163–188.

(250) Riplinger, C.; Sampson, M. D.; Ritzmann, A. M.; Kubiak, C. P.; Carter, E. A. Mechanistic Contrasts between Manganese and Rhenium Bipyridine Electrocatalysts for the Reduction of Carbon Dioxide. J. Am. Chem. Soc. 2014, 136, 16285–16298.

(251) Machan, C. W.; Stanton, C. J.; Vandezande, J. E.; Majetich, G. F.; Schaefer, H. F.; Kubiak, C. P.; Agarwal, J. Electrocatalytic Reduction of Carbon Dioxide by Mn(CN)(2,2′-Bipyridine)(CO) 3 : CN Coordination Alters Mechanism. Inorg. Chem. 2015, 54, 8849–8856.

(252) Wu, Y.; Frasconi, M.; Gardner, D. M.; McGonigal, P. R.; Schneebeli, S. T.; Wasielewski, M. R.; Stoddart, J. F. Electron Delocalization in a Rigid Cofacial Naphthalene-1,8:4,5-Bis(Dicarboximide) Dimer. Angew. Chem. Int. Ed. 2014, 53, 9476–9481.

(253) Nalluri, S. K. M.; Liu, Z.; Wu, Y.; Hermann, K. R.; Samanta, A.; Kim, D. J.; Krzyaniak, M. D.; Wasielewski, M. R.; Stoddart, J. F. Chiral Redox-Active Isosceles Triangles. J. Am. Chem. Soc. 2016, 138, 5968–5977.

(254) Avestro, A.-J.; Gardner, D. M.; Vermeulen, N. A.; Wilson, E. A.; Schneebeli, S. T.; Whalley, A. C.; Belowich, M. E.; Carmieli, R.; Wasielewski, M. R.; Stoddart, J. F. Gated Electron Sharing Within Dynamic Naphthalene Diimide-Based Oligorotaxanes. Angew. Chem. Int. Ed. 2014, 53, 4442–4449. 196

(255) Wu, Y.; Krzyaniak, M. D.; Stoddart, J. F.; Wasielewski, M. R. Spin Frustration in the Triradical Trianion of a Naphthalenediimide Molecular Triangle. J. Am. Chem. Soc. 2017, 139, 2948–2951.

(256) Liu, Z.; Nalluri, S. K. M.; Stoddart, J. F. Surveying Macrocyclic Chemistry: From Flexible Crown Ethers to Rigid Cyclophanes. Chem. Soc. Rev. 2017, 46, 2459– 2478.

(257) Kim, D. J.; Hermann, K. R.; Prokofjevs, A.; Otley, M. T.; Pezzato, C.; Owczarek, M.; Stoddart, J. F. Redox-Active Macrocycles for Organic Rechargeable Batteries. J. Am. Chem. Soc. 2017, 139, 6635–6643.

(258) Lipke, M. C.; Cheng, T.; Wu, Y.; Arslan, H.; Xiao, H.; Wasielewski, M. R.; Goddard, W. A.; Stoddart, J. F. Size-Matched Radical Multivalency. J. Am. Chem. Soc. 2017, 139, 3986–3998.

(259) Wu, Y.; Nalluri, S. K. M.; Young, R. M.; Krzyaniak, M. D.; Margulies, E. A.; Stoddart, J. F.; Wasielewski, M. R. Charge and Spin Transport in an Organic Molecular Square. Angew. Chem. Int. Ed. 2015, 54, 11971–11977.

(260) Reiner, B. R.; Foxman, B. M.; Wade, C. R. Electrochemical and Structural Investigation of the Interactions between Naphthalene Diimides and Metal Cations. Dalt. Trans. 2017, 46, 9472–9480.

(261) Ventura, K.; Smith, M. B.; Prat, J. R.; Echegoyen, L. E.; Villagrán, D. Introducing Students to Inner Sphere Electron Transfer Concepts through Electrochemistry Studies in Diferrocene Mixed-Valence Systems. J. Chem. Educ. 2017, 94, 526– 529.

(262) Guha, S.; Goodson, F. S.; Clark, R. J.; Saha, S. Deciphering Anion–π-Acceptor Interactions and Detecting Fluoride Using a Naphthalenediimide-Based Pd(II) Coordination Polymer. CrystEngComm 2012, 14, 1213–1215.

(263) Jin, S.; Ding, X.; Feng, X.; Supur, M.; Furukawa, K.; Takahashi, S.; Addicoat, M.; El-Khouly, M. E.; Nakamura, T.; Irle, S.; et al. Charge Dynamics in A Donor- Acceptor Covalent Organic Framework with Periodically Ordered Bicontinuous Heterojunctions. Angew. Chem. Int. Ed. 2013, 52, 2017–2021.

(264) Emmett, L.; Prentice, G. M.; Pantoş, G. D. Donor–acceptor Interactions in Chemistry. Annu. Reports Sect. “B” (Organic Chem. 2013, 109, 217.

(265) Gorteau, V.; Bollot, G.; Mareda, J.; Matile, S. Rigid-Rod Anion–π Slides for Multiion Hopping across Lipid Bilayers. Org. Biomol. Chem. 2007, 5, 3000.

(266) Doerksen, R. J.; Thakkar, A. J. Quadrupole and Octopole Moments of

197

Heteroaromatic Rings. J. Phys. Chem. A 1999, 103, 10009–10014.

(267) Farnia, G.; Romanin, A.; Capobianco, G.; Torzo, F. Electrochemical Reduction Mechanism of Phthalimide and Some of Its N-Derivatives in DMF. J. Electroanal. Chem. Interfacial Electrochem. 1971, 33, 31–44.

(268) Cai, Y.; Koshino, N.; Saha, B.; Espenson, J. H. Kinetics of Self-Decomposition and Hydrogen Atom Transfer Reactions of Substituted Phthalimide N -Oxyl Radicals in Acetic Acid. J. Org. Chem. 2005, 70, 238–243.

(269) UEDA, C.; NOYAMA, M.; OHMORI, H.; MASUI, M. Reactivity of Phthalimide- N-Oxyl : A Kinetic Study. Chem. Pharm. Bull. (Tokyo). 1987, 35, 1372–1377.

(270) Ogoshi, S.; Yoshida, T.; Nishida, T.; Morita, M.; Kurosawa, H. Coordination of Lewis Acid to η 2 -Enonepalladium(0) Leading to Continuous Structure Variation from η 2 -Olefin Type to η 3 -Allyl Type. J. Am. Chem. Soc. 2001, 123, 1944– 1950.

(271) Buchner, M. R.; Bechlars, B.; Wahl, B.; Ruhland, K. Influence of Electronically and Sterically Tunable Cinnamate Ligands on the Spectroscopic, Kinetic, and Thermodynamic Properties of Bis(Triphenylphosphine)Palladium(0) Olefin Complexes. Organometallics 2012, 31, 588–601.

(272) Castet, F.; Rodriguez, V.; Pozzo, J.-L.; Ducasse, L.; Plaquet, A.; Champagne, B. Design and Characterization of Molecular Nonlinear Optical Switches. Acc. Chem. Res. 2013, 46, 2656–2665.

(273) Kawata, S.; Kawata, Y. Three-Dimensional Optical Data Storage Using Photochromic Materials. Chem. Rev. 2000, 100, 1777–1788.

(274) Kim R. Dunbar. Useful Reagents and Ligands. In Inorganic Syntheses; John Wiley & Sons, Inc., 2002; pp 75–121.

(275) Mjambili, F.; Njoroge, M.; Naran, K.; De Kock, C.; Smith, P. J.; Mizrahi, V.; Warner, D.; Chibale, K. Synthesis and Biological Evaluation of 2-Aminothiazole Derivatives as Antimycobacterial and Antiplasmodial Agents. Bioorg. Med. Chem. Lett. 2014, 24, 560–564.

(276) Rimoldi, M.; Nakamura, A.; Vermeulen, N. A.; Henkelis, J. J.; Blackburn, A. K.; Hupp, J. T.; Stoddart, J. F.; Farha, O. K. A Metal–organic Framework Immobilised Iridium Pincer Complex. Chem. Sci. 2016, 7, 4980–4984.

(277) Lee, J.; Farha, O. K.; Roberts, J.; Scheidt, K. A.; Nguyen, S. T.; Hupp, J. T. Metal– organic Framework Materials as Catalysts. Chem. Soc. Rev. 2009, 38, 1450.

(278) Chughtai, A. H.; Ahmad, N.; Younus, H. A.; Laypkov, A.; Verpoort, F. Metal– 198

organic Frameworks: Versatile Heterogeneous Catalysts for Efficient Catalytic Organic Transformations. Chem. Soc. Rev. 2015, 44, 6804–6849.

(279) Cohen, S. M.; Zhang, Z.; Boissonnault, J. A. Toward “MetalloMOFzymes”: Metal-Organic Frameworks with Single-Site Metal Catalysts for Small-Molecule Transformations. Inorg. Chem. 2016, 55, 7281–7290.

(280) Rogge, S. M. J.; Bavykina, A.; Hajek, J.; Garcia, H.; Olivos-Suarez, A. I.; Sepúlveda-Escribano, A.; Vimont, A.; Clet, G.; Bazin, P.; Kapteijn, F.; et al. Metal–organic and Covalent Organic Frameworks as Single-Site Catalysts. Chem. Soc. Rev. 2017, 46, 3134–3184.

(281) Gui, B.; Meng, X.; Chen, Y.; Tian, J.; Liu, G.; Shen, C.; Zeller, M.; Yuan, D.; Wang, C. Reversible Tuning Hydroquinone/Quinone Reaction in Metal–Organic Framework: Immobilized Molecular Switches in Solid State. Chem. Mater. 2015, 27, 6426–6431.

(282) Meng, X.; Gui, B.; Yuan, D.; Zeller, M.; Wang, C. Mechanized Azobenzene- Functionalized Zirconium Metal-Organic Framework for on-Command Cargo Release. Sci. Adv. 2016, 2, e1600480–e1600480.

(283) Choi, J.; MacArthur, A. H. R.; Brookhart, M.; Goldman, A. S. Dehydrogenation and Related Reactions Catalyzed by Iridium Pincer Complexes. Chem. Rev. 2011, 111, 1761–1779.

(284) Nawara-Hultzsch, A. J.; Hackenberg, J. D.; Punji, B.; Supplee, C.; Emge, T. J.; Bailey, B. C.; Schrock, R. R.; Brookhart, M.; Goldman, A. S. Rational Design of Highly Active “Hybrid” Phosphine-Phosphinite Pincer Iridium Catalysts for Alkane Metathesis. ACS Catal. 2013, 3, 2505–2514.

(285) Shaffer, D. W.; Johnson, S. I.; Rheingold, A. L.; Ziller, J. W.; Goddard, W. A.; Nielsen, R. J.; Yang, J. Y. Reactivity of a Series of Isostructural Cobalt Pincer Complexes with CO2, CO, and H+. Inorg. Chem. 2014, 53, 13031–13041.

(286) Li, H.; Wang, K.; Sun, Y.; Lollar, C. T.; Li, J.; Zhou, H.-C. Recent Advances in Gas Storage and Separation Using Metal–organic Frameworks. Mater. Today 2018, 21, 108–121.

(287) Furukawa, H.; Gándara, F.; Zhang, Y.-B.; Jiang, J.; Queen, W. L.; Hudson, M. R.; Yaghi, O. M. Water Adsorption in Porous Metal–Organic Frameworks and Related Materials. J. Am. Chem. Soc. 2014, 136, 4369–4381.

(288) Adil, K.; Belmabkhout, Y.; Pillai, R. S.; Cadiau, A.; Bhatt, P. M.; Assen, A. H.; Maurin, G.; Eddaoudi, M. Gas/Vapour Separation Using Ultra-Microporous Metal–organic Frameworks: Insights into the Structure/Separation Relationship. 199

Chem. Soc. Rev. 2017, 46, 3402–3430.

(289) Qiu, S.; Xue, M.; Zhu, G. Metal–organic Framework Membranes: From Synthesis to Separation Application. Chem. Soc. Rev. 2014, 43, 6116–6140.

(290) Deng, Q.; Nie, G.; Pan, L.; Zou, J.-J. J.; Zhang, X.; Wang, L. Highly Selective Self-Condensation of Cyclic Ketones Using MOF-Encapsulating Phosphotungstic Acid for Renewable High-Density Fuel. Green Chem. 2015, 17, 4473–4481.

(291) Liu, Y. M.; Smit, B. Predicting Product Distribution of Propene Dimerization in Nanoporous Materials. ACS Catal. 2017, 7, 3940–3948.

(292) Benito-Garagorri, D.; Becker, E.; Wiedermann, J.; Lackner, W.; Pollak, M.; Mereiter, K.; Kisala, J.; Kirchner, K. Achiral and Chiral Transition Metal Complexes with Modularly Designed Tridentate PNP Pincer-Type Ligands Based on N-Heterocyclic Diamines. Organometallics 2006, 25, 1900–1913.

(293) Cattalini, L.; Cusumano, M.; Ricevuto, V.; Trozzi, M. Fast Reactions at Planar Four-Co-Ordinate Complexes. Part II. The Leaving-Group Effect in Palladium(II) Complexes of 3-Azapentane-1,5-Diamine. J. Chem. Soc. Dalt. Trans. 1975, 1 SRC-Go, 771.

(294) Lastoskie, C.; Gubbins, K. E.; Quirke, N. Pore Size Heterogeneity and the Carbon Slit Pore: A Density Functional Theory Model. Langmuir 1993, 9, 2693–2702.

(295) Lastoskie, C.; Gubbins, K. E.; Quirke, N. Pore Size Distribution Analysis of Microporous Carbons: A Density Functional Theory Approach. J. Phys. Chem. 1993, 97, 4786–4796.

(296) Ravikovitch, P. I.; Haller, G. L.; Neimark, A. V. Density Functional Theory Model for Calculating Pore Size Distributions: Pore Structure of Nanoporous Catalysts. Adv. Colloid Interface Sci. 1998, 76–77, 203–226.

(297) Katz, M. J.; Brown, Z. J.; Colón, Y. J.; Siu, P. W.; Scheidt, K. A.; Snurr, R. Q.; Hupp, J. T.; Farha, O. K. A Facile Synthesis of UiO-66, UiO-67 and Their Derivatives. Chem. Commun. 2013, 49, 9449.

(298) Øien, S.; Wragg, D.; Reinsch, H.; Svelle, S.; Bordiga, S.; Lamberti, C.; Lillerud, K. P. Detailed Structure Analysis of Atomic Positions and Defects in Zirconium Metal-Organic Frameworks. Cryst. Growth Des. 2014, 14, 5370–5372.

(299) Trickett, C. A.; Gagnon, K. J.; Lee, S.; Gándara, F.; Bürgi, H.-B.; Yaghi, O. M. Definitive Molecular Level Characterization of Defects in UiO-66 Crystals. Angew. Chem. Int. Ed. 2015, 54, 11162–11167.

(300) Wu, H.; Chua, Y. S.; Krungleviciute, V.; Tyagi, M.; Chen, P.; Yildirim, T.; Zhou, 200

W. Unusual and Highly Tunable Missing-Linker Defects in Zirconium Metal − Organic Framework UiO-66 and Their Important Effects on Gas Adsorption. J. Am. Chem. Soc. 2013, 135, 10525–10532.

(301) Taddei, M. When Defects Turn into Virtues: The Curious Case of Zirconium- Based Metal-Organic Frameworks. Coord. Chem. Rev. 2017, 343, 1–24.

(302) Klet, R. C.; Liu, Y.; Wang, T. C.; Hupp, J. T.; Farha, O. K. Evaluation of Brønsted Acidity and Proton Topology in Zr- and Hf-Based Metal–organic Frameworks Using Potentiometric Acid–base Titration. J. Mater. Chem. A 2016, 4, 1479–1485.

(303) Dixon, W. T.; Schaefer, J.; Sefcik, M. D.; Stejskal, E. O.; McKay, R. A. Total Suppression of Sidebands in CPMAS C-13 NMR. J. Magn. Reson. 1982, 49, 341– 345.

(304) Vitaku, E.; Smith, D. T.; Njardarson, J. T. Analysis of the Structural Diversity, Substitution Patterns, and Frequency of Nitrogen Heterocycles among U.S. FDA Approved Pharmaceuticals. J. Med. Chem. 2014, 57, 10257–10274.

(305) Horton, D. A.; Bourne, G. T.; Smythe, M. L. The Combinatorial Synthesis of Bicyclic Privileged Structures or Privileged Substructures. Chem. Rev. 2003, 103, 893–930.

(306) Humphrey, G. R.; Kuethe, J. T. Practical Methodologies for the Synthesis of Indoles. Chem. Rev. 2006, 106, 2875–2911.

(307) Yang, Z.; Xia, C.; Liu, D.; Liu, Y.; Sugiya, M.; Imamoto, T.; Zhang, W. P- Stereogenic PNP Pincer-Pd Catalyzed Intramolecular Hydroamination of Amino- 1,3-Dienes. Org. Biomol. Chem. 2015, 13, 2694–2702.

(308) Pierson, J. M.; Ingalls, E. L.; Vo, R. D.; Michael, F. E. Palladium(II)-Catalyzed Intramolecular Hydroamination of 1,3-Dienes to Give Homoallylic Amines. Angew. Chem. Int. Ed. 2013, 52, 13311–13313.

(309) McGhee, A.; Cochran, B. M.; Stenmark, T. A.; Michael, F. E. Stereoselective Synthesis of 2,5-Disubstituted Morpholines Using a Palladium-Catalyzed Hydroamination Reaction. Chem. Commun. 2013, 49, 6800–6802.

(310) Cucciolito, M. E.; Damora, A.; Vitagliano, A. Catalytic Hydroalkylation of Olefins by Stabilized Carbon Nucleophiles Promoted by Dicationic Platinum(II) and Palladium(II) Complexes. Organometallics 2010, 29, 5878–5884.

(311) Cochran, B. M.; Michael, F. E. Synthesis of 2,6-Disubstituted Piperazines by a Diastereoselective Palladium-Catalyzed Hydroamination Reaction. Org. Lett. 2008, 10, 329–332.

201

(312) Cochran, B. M.; Michael, F. E. Mechanistic Studies of a Palladium-Catalyzed Intramolecular Hydroamination of Unactivated Alkenes: Protonolysis of a Stable Palladium Alkyl Complex Is the Turnover-Limiting Step. J. Am. Chem. Soc. 2008, 130, 2786–2792.

(313) Michael, F. E.; Cochran, B. M. Room Temperature Palladium-Catalyzed Intramolecular Hydroamination of Unactivated Alkenes. J. Am. Chem. Soc. 2006, 128, 4246–4247.

(314) Kawatsura, M.; Hartwig, J. F. Transition Metal-Catalyzed Addition of Amines to Acrylic Acid Derivatives. A High-Throughput Method for Evaluating Hydroamination of Primary and Secondary Alkylamines. Organometallics 2001, 20, 1960–1964.

(315) Aprile, A.; Iversen, K. J.; Wilson, D. J. D.; Dutton, J. L. Te(II)/Te(IV) Mediated C-N Bond Formation on 2,5-Diphenyltellurophene and a Reassignment of the Product from the Reaction of PhI(OAc)2 with 2 TMS-OTf. Inorg. Chem. 2015, 54, 4934–4939.

(316) Connelly, N. G.; Geiger, W. E. Chemical Redox Agents for Organometallic Chemistry. Chem. Rev. 1996, 96, 877–910.

(317) Clegg, J. K.; Cremers, J.; Hogben, A. J.; Breiner, B.; Smulders, M. M. J. J.; Thoburn, J. D.; Nitschke, J. R. A Stimuli Responsive System of Self-Assembled Anion-Binding Fe4L68+ Cages. Chem. Sci. 2013, 4, 68–76.

(318) Paenurk, E.; Kaupmees, K.; Himmel, D.; Kütt, A.; Kaljurand, I.; Koppel, I. A.; Krossing, I.; Leito, I. A Unified View to Brønsted Acidity Scales: Do We Need Solvated Protons? Chem. Sci. 2017, 8, 6964–6973.

(319) Wang, G.; Sharp, C.; Plonka, A. M.; Wang, Q.; Frenkel, A. I.; Guo, W.; Hill, C.; Smith, C.; Kollar, J.; Troya, D.; et al. Mechanism and Kinetics for Reaction of the Chemical Warfare Agent Simulant, DMMP( g ), with Zirconium(IV) MOFs: An Ultrahigh-Vacuum and DFT Study. J. Phys. Chem. C 2017, 121, 11261–11272.

(320) Praveen, C.; Perumal, P. Revisiting the Gold-Catalyzed Dimerization of 2- Ethynylanilines: A Room-Temperature and Silver-Free Protocol for the Synthesis of Multifunctional Quinolines. Synthesis (Stuttg). 2016, 48, 855–864.

(321) Sakai, N.; Annaka, K.; Konakahara, T. Direct Synthesis of Polysubstituted Quinoline Derivatives by InBr 3 -Promoted Dimerization of 2-Ethynylaniline Derivatives. J. Org. Chem. 2006, 71, 3653–3655.

(322) Sakai, N.; Annaka, K.; Fujita, A.; Sato, A.; Konakahara, T. InBr 3 -Promoted Divergent Approach to Polysubstituted Indoles and Quinolines from 2- 202

Ethynylanilines: Switch from an Intramolecular Cyclization to an Intermolecular Dimerization by a Type of Terminal Substituent Group. J. Org. Chem. 2008, 73, 4160–4165.

(323) Alaerts, L.; Séguin, E.; Poelman, H.; Thibault-Starzyk, F.; Jacobs, P. A.; De Vos, D. E. Probing the Lewis Acidity and Catalytic Activity of the Metal-Organic Framework [Cu3(Btc)2] (BTC = Benzene-1,3,5-Tricarboxylate). Chem. - A Eur. J. 2006, 12, 7353–7363.

(324) Vermoortele, F.; Vandichel, M.; Van De Voorde, B.; Ameloot, R.; Waroquier, M.; Van Speybroeck, V.; De Vos, D. E. Electronic Effects of Linker Substitution on Lewis Acid Catalysis with Metal-Organic Frameworks. Angew. Chem. Int. Ed. 2012, 51, 4887–4890.

(325) Cirujano, F. G. MOFs: Vs. Zeolites: Carbonyl ActIVation with M(IV) Catalytic Sites. Catal. Sci. Technol. 2017, 7, 5482–5494.

(326) Vandichel, M.; Vermoortele, F.; Cottenie, S.; De Vos, D. E.; Waroquier, M.; Van Speybroeck, V. Insight in the Activity and Diastereoselectivity of Various Lewis Acid Catalysts for the Citronellal Cyclization. J. Catal. 2013, 305, 118–129.

(327) Cirujano, F. G.; Llabrés I Xamena, F. X.; Corma, A. MOFs as Multifunctional Catalysts: One-Pot Synthesis of Menthol from Citronellal over a Bifunctional MIL-101 Catalyst. Dalt. Trans. 2012, 41, 4249–4254.

(328) Maaliki, C.; Chevalier, Y.; Thiery, E.; Thibonnet, J. Palladium and Copper Catalyzed Sonogashira Decarboxylative Coupling of Aryl Iodides and Alkynyl Carboxylic Acids. Tetrahedron Lett. 2016, 57, 3358–3362.

(329) Sakai, N.; Tamura, K.; Shimamura, K.; Ikeda, R.; Konakahara, T. Copper- Catalyzed [5 + 1] Annulation of 2-Ethynylanilines with an N , O -Acetal Leading to Construction of Quinoline Derivatives. Org. Lett. 2012, 14, 836–839.

(330) Zagulyaeva, A. A.; Yusubov, M. S.; Zhdankin, V. V. A General and Convenient Preparation of [Bis(Trifluoroacetoxy)Iodo] Perfluoroalkanes and [Bis(Trifluoroacetoxy)Iodo] Arenes by Oxidation of Organic Iodides Using Oxone and Trifluoroacetic Acid. J. Org. Chem. 2010, 75, 2119–2122.

(331) Hwang, T. L.; Shaka, A. J. Water Suppression That Works. Excitation Sculpting Using Arbitrary Wave-Forms and Pulsed-Field Gradients. J. Magn. Reson. - Ser. A 1995, 112, 275–279.

203