Synthesis and reactivity of palladium complexes that contain redox-active verdazyl ligands

(1)

by

Corey A. Sanz

B.Sc., University of Victoria, 2011

A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of

DOCTOR OF PHILOSOPHY in the Department of Chemistry

(2)

Supervisory Committee

Synthesis and Reactivity of Palladium Complexes that Contain Redox-Active Verdazyl Ligands

by

Corey A. Sanz

B.Sc., University of Victoria, 2011

Supervisory Committee

Dr. Robin G. Hicks, (Department of Chemistry)

Supervisor

Dr. David J. Berg, (Department of Chemistry)

Departmental Member

Dr. Neil Burford, (Department of Chemistry)

Departmental Member

Dr. Jay T. Cullen, (School of Earth and Ocean Sciences)

(3)

Abstract

This thesis presents the synthesis, characterization and reactivity of a series of palladium complexes that contain redox-active verdazyl ligands. This work was motivated by the possibility of discovering new and interesting reactivity that may eventually lead to the development of new chemical reactions.

A bidentate verdazyl radical ligand that contains an aryl phosphine was synthesized. Reaction of this ligand with (PhCN)2PdCl2 yielded a square planar (verdazyl)PdCl2 complex.

Structural and spectroscopic data suggest that this compound consists of a ligand-centered radical coordinated to a Pd(II) center. The radical complex was chemically reduced by one-electron to generate a binuclear chloride-bridged [(verdazyl)PdCl]2 complex. In this reduced complex, both

metals were still Pd(II) and the verdazyl ligand was determined to be in its singly reduced, monoanionic charge state. The original radical PdCl2 complex could be regenerated via

one-electron oxidation of the reduced complex using PhICl2. The verdazyl ligands in the reduced

complex could also be reversibly protonated to generate “leuco” verdazyl complex (verdazyl-H)PdCl2. Reaction of the radical (verdazyl)PdCl2 complex with water triggers a ligand-centered

redox disproportionation reaction.

A series of bis(verdazyl) palladium complexes were synthesized using a bidentate pyridine-substituted verdazyl ligand. Reaction of two equivalents of radical ligand with (CH3CN)4Pd2+

yielded a (verdazyl)2Pd(solvent)2+ complex (solvent = CH3CN or DMSO). In this complex, one

verdazyl radical ligand chelates to palladium and the other binds as a monodentate ligand. Two-electron reduction of this complex generated a (verdazyl)2Pd complex in which two monoanionic

(4)

reaction of 0.5 equivalents of Pd(0)2(dba)3 with two equivalents of radical ligand. In this reaction,

the metal is oxidized by two electrons and each ligand is reduced by a single electron. Two-electron oxidation of the reduced complex in the presence of DMSO yielded the original bis(radical)complex, (verdazyl)2Pd(DMSO)2+. Chlorination of the reduced complex using one

equivalent of PhICl2 (two-electron oxidation) resulted in dissociation of one verdazyl ligand to

afford a 1:1 mixture of free verdazyl : (verdazyl)PdCl2, in which both of the verdazyls are neutral

radicals. Reaction of the reduced complex with 0.5 equivalents of PhICl2 (one-electron oxidation)

yielded a (verdazyl)2PdCl complex that contained a bidentate reduced verdazyl ligand and a

monodentate radical ligand. All three of the oxidation reactions described above adhere to ligand-centered redox chemistry. Reaction of the reduced (verdazyl)2Pd complex with excess HCl

resulted in protonation of both the anionic verdazyl ring and the pyridyl group to generate a leuco/pyridinium tetrachloropalladate salt, (verdazyl-H2)2(PdCl4). The protonated salt could be

converted back to the original (verdazyl)2Pd complex by deprotonation with water.

Palladium complexes of a tridentate NNN-chelating verdazyl ligand were prepared and their redox chemistry was explored. Reaction of the radical ligand with (CH3CN)4Pd2+ yielded

radical complex (verdazyl)Pd(NCCH3)2+. The tridentate ligand was also prepared in its reduced,

leuco form (verdazyl-H). Reaction of the leuco verdazyl with (CH3CN)2PdCl2 generated HCl and

a (verdazyl)PdCl complex in which the ligand is in its monoanionic charge state. The reduced (verdazyl)PdCl complex was reacted with AgBF4 to afford (verdazyl)Pd(NCCH3)+ via chloride

abstraction; the verdazyl remained in its reduced charge state following the reaction. Both reduced complexes (chloro and acetonitrile) were oxidized by a single electron to afford the corresponding radical complexes. These radical complexes could be reduced by a single electron to regenerate the original reduced complexes. Like the previous two projects, all of the redox chemistry was

(5)

ligand-centered. The reactivity of these complexes with primary amines was also explored. Reaction of radical complex (verdazyl)Pd(NCCH3)2+ with n-butylamine resulted in one-electron

reduction of the verdazyl ligand. We were unable to determine the mechanism of the reaction, but the reactivity that was observed demonstrates the potential for verdazyl-palladium complexes to be used in the design of new radical reactions.

(6)

2.3.4 Electrochemical Properties of 2.9 ... 50 2.4 Chemical Oxidation of 2.9 ... 52 2.5 Reversible Protonation of 2.9 ... 54 2.6 Characterization of Complex 2.14 ... 54 2.6.1 X-Ray Structure of 2.14 ... 54 2.6.2 Spectroscopic Characterization of 2.14 ... 58 2.7 Chemical Oxidation of 2.14 ... 59

2.8 Reactivity of Radical Complex 2.6 with Water ... 60

2.9 Conclusions ... 62

2.10 Experimental ... 63

2.10.1 General Methods ... 63

(8)

Chapter 3: Redox Reactions of Palladium Bis(2-Pyridyl)-Verdazyl Complexes ... 74

3.1.1 Reduction of Dichloropalladium Complex 1.44 ... 75

3.2 Synthesis and Characterization of Bis(2-Pyridyl) Verdazyl-Radical Palladium Complexes ... 76

3.2.1 Synthesis of Bis(2-Pyridyl) Verdazyl-Radical Palladium Complexes ... 76

3.2.2 X-Ray Structure of Bis(Radical) Complex 3.4 ... 77

3.2.3 Spectroscopic Characterization of 3.3 and 3.4 ... 79

3.3 Synthesis and Characterization of a Neutral Bis(Verdazyl) Palladium Complex ... 83

3.3.1 X-Ray Structure of Neutral Complex 3.2 ... 84

3.4 Redox Reactions of Bis(Verdazyl) Complexes ... 88

3.5 Electrochemical Properties of Bis(Verdazyl) Complexes 3.4 and 3.2 ... 90

3.6 Chlorination of Bis(Anion) Complex 3.2 ... 94

3.6.1 X-Ray Structure of Bis(Verdazyl) Chloro Complex 3.5 ... 95

3.6.2 Spectroscopy and Electrochemistry of 3.5 ... 97

3.7 Reversible Protonation of 3.2 ... 101

3.7.1 X-Ray Structure and Spectroscopic Properties of Pyridinium Tetrachloropalladate Salt 3.6... 102

3.8 Air Oxidation of 3.6 ... 105

(9)

3.10.2 Syntheses... 108

Chapter 4: Redox Chemistry of Palladium Complexes Containing a Tridentate Verdazyl Ligand ... 116

4.2 Synthesis and Characterization of a Palladium Complex Containing a Tridentate Radical Verdazyl Ligand... 117

4.2.1 X-Ray Structure of Radical Complex 4.3 ... 119

4.3 Synthesis and Characterization of a Monochloro Verdazyl-Palladium Complex ... 124

4.3.1 X-Ray Structure of Verdazyl-Anion/Chloro Complex 4.7 ... 126

4.4 Synthesis of an Acetonitrile Complex with a Reduced Tridentate Verdazyl Ligand ... 129

4.4.1 X-Ray Structure of Acetonitrile Complex 4.8 ... 130

4.5 Electrochemical Properties of 4.7 and 4.8 ... 133

4.6 Chemical Redox Reactions of Palladium Complexes Containing a Tridentate Verdazyl Ligand ... 137

4.7 Characterization of Radical/Chloro Complex 4.11 ... 138

4.7.1 X-Ray Structure of 4.11 ... 138

4.8 Reactivity of Verdazyl-Palladium Complexes with Primary Amines ... 143

4.8.1 Synthesis and Characterization of Butylamine Complex 4.12 ... 146

(10)

4.8.3 Attempted Oxidation of Butylamine Complex 4.12 ... 148

4.10.1 General Methods ... 151

4.10.2 Syntheses... 153

Chapter 5: Summary and Future Work ... 162

References ... 171

Appendix A: Crystallographic Parameters ... 178

Appendix B: Complete Listing of Bond Lengths and Angles ... 184

Appendix C: NMR Spectra ... 214

(11)

List of Figures

Figure 1.1: The active site of galactose oxidase ... 9 Figure 2.1: X-ray structures of radicals 2.1 (left) and 2.5 (right). Hydrogen atoms have been

omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 31

Figure 2.2: X-ray structure of palladium complex 2.6. Hydrogen atoms have been omitted for

clarity and the terminal phenyl groups are drawn as wireframes. Thermal ellipsoids are shown at the 50% probability level. ... 33

Figure 2.3: Electronic spectra of free ligand 2.1 (black line) and dichloropalladium complex 2.6

(red line) in dichloromethane. ... 34

Figure 2.4: EPR spectrum of phosphine-verdazyl 2.1 in dichloromethane (black trace) and the

corresponding simulated spectrum (red trace). The spectrum was recorded at room temperature (295 K) ... 35

Figure 2.5: EPR spectrum of palladium complex 2.6 in CH2Cl2 (black trace) and the

corresponding simulated spectrum (red trace). The spectrum was recorded at room temperature (295 K) ... 37

Figure 2.6: Cyclic voltammogram of 2.1 (black trace), 2.5 (blue trace), and 2.6 (red trace). The

vertical axis is current and the arrows show both the direction and starting point for each scan. All CVs were recorded as 1 mM solutions in dichloromethane at a scan rate of 100 mV/s (all solutions contain 0.1 M Bu4NBF4 as electrolyte)... 39

Figure 2.7: X-ray structure of binuclear complex 2.9. The three different orientations highlight

the square planar geometry around the metal (top left), atom labelling around the verdazyl ring (top right), and puckering of the verdazyl core (bottom). Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 42

Figure 2.8: Singly occupied molecular orbital of a verdazyl radical ... 43 Figure 2.9: Schematic diagram showing the electrons that would have π symmetry in the verdazyl

core if the ring were planar. Left: radical complex 2.6 (seven π electrons); right: reduced complex

2.9 (eight π electrons). ... 44 Figure 2.10: (a) Electronic spectrum of 2.9 in dichloromethane. (b) Electronic spectra of 2.9 in

different solvents to illustrate the solvatochromism associated with the major visible absorption band. Legend: dichloromethane (black trace), toluene (red), isopropanol (blue), acetone (green), and acetonitrile (purple). ... 47

(12)

Figure 2.11: Cyclic voltammogram of 2.9 in (a) dichloromethane and (b) acetonitrile. The vertical

axis is current and the arrows show both the direction and starting point for each scan. Conditions: 1 mM analyte, 100 mV/s scan rate, 0.1 M Bu4NBF4. ... 51

Figure 2.12: X-ray structure of leuco-verdazyl complex 2.14. The perspective on the right is meant

to highlight the distortion of the tetrazine core from planarity. All hydrogen atoms except H2 have been omitted for clarity and the terminal phenyl groups are drawn as wireframes. Thermal ellipsoids are shown at the 50% probability level. ... 55

Figure 2.13: Schematic diagram showing the electrons that would have π symmetry in the

verdazyl core of 2.14 if the ring were planar (eight π electrons) ... 57

Figure 3.1: X-ray structure of bis(radical) complex 3.4. Hydrogen atoms and BF4 ions have been

Figure 3.2: EPR spectra of 3.3 (red line) and 3.4 (black line) in dichloromethane. The bold lines

represent the experimental spectra and the thin lines represent simulated spectra. Both spectra were recorded at room temperature (295 K)... 81

Figure 3.3: Electronic spectra of acetonitrile complex 3.3 (red line) and DMSO complex 3.4 (black

line) in dichloromethane. ... 82

Figure 3.4: X-ray structure of 3.2 showing atom labelling (left) and the orientation of the isopropyl

groups with respect to the palladium ion’s square plane (right). Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 84

Figure 3.5: Electronic spectrum of 3.2 in dichloromethane ... 87 Figure 3.6: Cyclic voltammogram of (a) 3.4 and (b) 3.2 in dichloromethane. The vertical axis is

current and the arrows show both the direction and starting point for each scan. Conditions: 1 mM analyte, 100 mV/s scan rate, 0.1 M Bu4NBF4. ... 91

Figure 3.7: X-ray structure of 3.5 showing atom labelling (left) and the shape of the verdazyl ring

in each ligand (right). Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 96

Figure 3.8: EPR spectrum of 3.5 in dichloromethane (black trace) and the corresponding simulated

spectrum (red trace). The spectrum was recorded at room temperature (295 K) ... 98

Figure 3.9: Electronic spectrum of 3.5 in dichloromethane ... 99 Figure 3.10: Cyclic voltammogram of 3.5 in dichloromethane. The vertical axis is current and the

arrow shows both the direction and starting point of the scan. Conditions: 1 mM analyte, 100 mV/s scan rate, 0.1 M Bu4NBF4. ... 100

(13)

Figure 3.11: X-ray structure of 3.6. Only one of the two crystallographically equivalent

leuco-pyridinium cations has been shown for clarity. All hydrogen atoms except H2 and H5 have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 103

Figure 4.1: X-ray structure of radical/acetonitrile complex 4.3. Hydrogen atoms and BF4 ions have

been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 120

Figure 4.2: EPR spectrum of radical/acetonitrile complex 4.3 in acetonitrile (black trace) and the

corresponding simulated spectrum (red trace) ... 122

Figure 4.3: Electronic spectrum of radical/acetonitrile complex 4.3 in acetonitrile. ... 123 Figure 4.4: Electronic spectra of free radical 4.1 (dashed line) and nickel complex 4.5 (solid line)

in acetonitrile. Reproduced from Brook et al.64 with permission from The Royal Society of Chemistry. ... 124

Figure 4.5: X-ray structure of verdazyl-anion/chloro complex 4.7 showing atom labelling (left)

and the shape of the verdazyl ring (right). Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 127

Figure 4.6: Electronic spectrum of verdazyl-anion/chloro complex 4.7 in acetonitrile. ... 129 Figure 4.7: X-ray structure of verdazyl-anion/acetonitrile complex 4.8 showing atom labelling

(left) and the shape of the verdazyl ring (right). Hydrogen atoms and the BF4 anion have been

Figure 4.8: Electronic spectra of verdazyl-anion/acetonitrile complex 4.8 (black line) and

verdazyl-anion/chloro complex 4.7 (red line) in acetonitrile. ... 133

Figure 4.9: Cyclic voltammogram of 4.8 (red trace), 4.7 (black trace), and 4.1 (blue trace). The

vertical axis is current and the arrows show both the direction and starting point for each scan. All CVs were recorded as 1 mM solutions in acetonitrile at a scan rate of 100 mV/s (all solutions contain 0.1 M Bu4NBF4 as electrolyte)... 134

Figure 4.10: X-ray structure of radical/chloro complex 4.11. Hydrogen atoms and the BF4 ion

have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 140

Figure 4.11: EPR spectrum of radical/chloro complex 4.11 in acetonitrile (black trace) and the

corresponding simulated spectrum (red trace) ... 142

Figure 4.12: Electronic spectrum of radical/chloro complex 4.11 (black line) and

radical/acetonitrile complex 4.3 (red line) in acetonitrile. ... 143

Figure 4.13: Cyclic voltammogram of n-butylamine complex 4.12 in acetonitrile. The vertical

axis is current and the arrow shows both the direction and starting point for the scan. The CV was recorded as a 1 mM solution at a scan rate of 100 mV/s (the solution also contains 0.1 M Bu4NBF4

(14)

Figure B-1: X-ray structure of 2.1. Hydrogen atoms have been omitted for clarity. Thermal

ellipsoids are shown at the 50% probability level. ... 184

Figure B-4: X-ray structure of 2.9. The top image shows atom labeling of the asymmetric unit and

the bottom image shows the full structure. Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 190

Figure B-5: X-ray structure of 2.14. All hydrogen atoms except H2, H2A, and H2B have been

Figure B-6: X-ray structure of 3.4. Hydrogen atoms have been omitted for clarity. BF4 ions have

also been removed for clarity (2:1 BF4:Pd in the structure). Thermal ellipsoids are shown at the

50% probability level. ... 194

Figure B-7: X-ray structure of 3.2. The material crystallized as two half-molecules residing on

different inversion centers; the two images represent the two molecules that crystallized. Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 197

Figure B-9: X-ray structure of 3.11. All hydrogen atoms except H2 and H5 have been omitted for

clarity. Thermal ellipsoids are shown at the 50% probability level. ... 202

Figure B-12: X-ray structure of 4.8. The material crystallized as two separate molecules; the two

images represent the two molecules that crystallized. Hydrogen atoms have been omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. ... 208

(15)

Figure C-1: 1H NMR spectrum of 2.4 in CDCl3. The peak at 1.5 ppm is from residual water. 214

Figure C-2: 13_C{1_{H} NMR spectrum of 2.4 in CDCl}

3. The triplet at 77 ppm is from the NMR

solvent. ... 215

Figure C-3: 1H NMR spectrum of 2.9 in CD3CN (temperature = 240 K). The peak at 1.9 ppm is

from the NMR solvent and the peak at 3.4 is from residual water. ... 216

Figure C-4: 13C{1H} NMR spectrum of 2.9 in CD2Cl2 (temperature = 240 K). The peak at 54 ppm

is from the NMR solvent... 217

Figure C-5: 1H NMR spectrum of 2.14 in CD2Cl2. The peak at 5.3 ppm is from the NMR solvent

and the peaks at 1.5 and 2.0 ppm are from residual acetonitrile and water, respectively. ... 218

Figure C-6: 1H NMR spectrum of 3.2 in CDCl3. The peak at 7.3 ppm is from the NMR solvent

and the peak at 1.6 is from residual water. ... 219

Figure C-7: Expanded regions of the 1_{H NMR spectrum of 3.2 (7.3-9.3 ppm, 3.4-4.9 ppm) ... 220}

Figure C-8: Expanded region of the 1H NMR spectrum of 3.2 (methyl protons, 1.0-1.5 ppm) 221

Figure C-9: COSY spectrum of 3.2 in CDCl3 (360 MHz). ... 222

Figure C-10: 13C{1H} NMR spectrum of 3.2 in CDCl3. The triplet at 77 ppm is from the NMR

solvent. ... 223

Figure C-11: 1H NMR spectrum of 3.6 in DMSO-d6. The peak at 2.5 ppm is from the NMR

solvent. ... 224

Figure C-12: Expanded regions of the 1_{H NMR spectrum of 3.6 (7.4-8.8 ppm, 4.4-7.6 ppm) . 225}

Figure C-13: 1H NMR spectrum of 3.6 in CD3CN (temperature = 233 K). The peak at 2.0 ppm is

from the NMR solvent and the peaks at 1.8, 2.5, and 3.7 are from residual solvent. ... 226

Figure C-14: 13C{1H} NMR spectrum of 3.6 in DMSO-d6. The peak at 40 ppm is from the NMR

solvent. ... 227

Figure C-15: 1H NMR spectrum of 4.6 in CDCl3... 228

Figure C-16: Expanded aromatic region of the 1H NMR spectrum of 4.6 (6.7-8.9 ppm). The peak at 7.26 is from the NMR solvent. ... 229

solvent. ... 230

Figure C-18: 1_{H NMR spectrum of 4.7 in CDCl}

3. The peak at 7.3 ppm is from the NMR solvent

(16)

Figure C-19: Expanded aromatic region of the 1H NMR spectrum of 4.7 (6.7-9.1 ppm). The peak at 7.26 is from the NMR solvent. ... 232

solvent. ... 233

Figure C-21: 1H NMR spectrum of 4.8 in CD3CN. The peak at 1.9 ppm is from the NMR solvent

and the peak at 2.3 ppm is from residual water. ... 234

Figure C-22: Expanded aromatic region of the 1H NMR spectrum of 4.8 (6.9-8.3 ppm). ... 235

Figure C-23: 13C{1H} NMR spectrum of 4.8 in CD3CN. The peaks at 1 and 118 ppm are from the

NMR solvent. ... 236

Figure C-24: 1H NMR spectrum of 4.12 in CDCl3. The peak at 7.3 ppm is from the NMR solvent.

... 237

Figure C-25: Expanded regions of the 1_{H NMR spectrum of 4.12 (6.8-8.7 ppm, 0.5-5.3 ppm).}

... 238

solvent. ... 239

Figure D-1: EPR spectrum of phosphine-oxide verdazyl 2.5 in dichloromethane (black trace) and

the corresponding simulated spectrum (red trace). The spectrum was recorded at room temperature (295 K). ... 240

Figure D-2: Electronic spectrum of phosphine-oxide verdazyl 2.5 in dichloromethane. ... 241 Figure D-3: Cyclic voltammogram of 2.14 in dichloromethane. The vertical axis is current and

the arrow shows both the direction and starting point for the scan. Conditions: 1 mM analyte, 100 mV/s scan rate, 0.1 M Bu4NBF4. ... 242

Figure D-4: Electronic spectrum of a crude sample of [1.44][FeCp*2] in dichloromethane. .... 242

Figure D-5: Electronic spectra of leuco compounds 2.14 (black line) and 3.6 (red line) in

acetonitrile... 243

Figure D-6: Cyclic voltammogram of 4.3 (red trace) and 4.11 (black trace). The vertical axis is

current and the arrows show both the direction and starting point for each scan. Both CVs were recorded as 1 mM solutions in acetonitrile at a scan rate of 100 mV/s (both solutions contain 0.1 M Bu4NBF4 as electrolyte). ... 244

(17)

List of Schemes

Scheme 1.1: (a) Dioxolenes and related RALs. (b) The triad of oxidation states that are available

to 1.2-1.4 ... 3

Scheme 1.2: Resonance structures of a copper-iminosemiquinone complex ... 5

Scheme 1.3: Average bond lengths (Å) of dioxolene ligands19 ... 7

Scheme 1.4: (a) Charge states available to a bis(imino)pyridine ligand. (b) Proposed mechanism for the cycloaddition reaction catalyzed by 1.15 ... 11

Scheme 1.5: Ligand-centered reductive elimination from a Zr(IV) center ... 12

Scheme 1.6: Ligand-centered oxidative addition ... 13

Scheme 1.7: Negishi-type cross coupling using RALs as electron reservoirs ... 14

Scheme 1.8: A selected part of the catalytic cycle reported by Grützmacher37_{for alcohol oxidation.} The stoichiometric oxidant is benzoquinone. ... 16

Scheme 1.9: Dihydrogen oxidation catalyzed by an iridium complex ... 17

Scheme 1.10: Radical mechanism for CH amination using palladium complex 1.34 ... 21

Scheme 1.11: The redox chemistry of verdazyl radicals ... 22

Scheme 2.1: Synthesis of verdazyl radical 2.1 and phosphine-oxide 2.5 ... 29

Scheme 2.2: Synthesis of palladium dichloro complex 2.6 ... 30

Scheme 2.3: Proposed mechanism to account for the irreversibility of electrochemical oxidation of radical 2.1 ... 39

Scheme 2.4: Chemical reduction of verdazyl-radical complex 2.6... 41

Scheme 2.5: Two resonance structures of a reduced verdazyl ligand bound to palladium(II) .... 45

Scheme 2.6: Chlorination of binuclear complex 2.9 using PhICl2... 53

Scheme 2.7: Reversible protonation of the anionic verdazyl ligands in complex 2.9 ... 54

Scheme 2.8: Oxidation of a leuco-verdazyl ligand using PhICl2. ... 60

(18)

Scheme 2.10: Reaction of 2.1 with (CH3CN)2PdCl2 in wet acetonitrile ... 61

Scheme 2.11: Reduction of palladium(II) to palladium(0) using a tertiary phosphine ... 62

Scheme 3.1: Reduction of 1.44 using decamethylferrocene ... 76

Scheme 3.2: Synthesis of bis(verdazyl) palladium complexes 3.3 and 3.4 ... 77

Scheme 3.3: Reaction of verdazyl radical 3.1 with Pd2(dba)3 (2:1 ratio of verdazyl:palladium) 83 Scheme 3.4: Reduction and oxidation of bis(radical) complex 3.4 and bis(anion) complex 3.2, respectively ... 89

Scheme 3.5: Reactions of 3.2 with iodobenzene dichloride ... 95

Scheme 3.6: Reversible protonation of 3.2 using anhydrous HCl ... 102

Scheme 3.7: Oxidation of 3.6 in solutions exposed to air ... 105

Scheme 4.1: Synthesis of verdazyl radical 4.1 according to Brook et. al.64 ... 117

Scheme 4.2: Synthesis of palladium complex 4.3 using a crude sample of verdazyl radical 4.1 ... 118

Scheme 4.3: Synthesis of leuco verdazyl 4.6 using a procedure adapted from Brook and coworkers64 ... 125

Scheme 4.4: Synthesis of verdazyl-anion/chloro complex 4.7 from leuco verdazyl 4.6 ... 126

Scheme 4.5: Synthesis of acetonitrile complex 4.8 from 4.7 and AgBF4 ... 130

Scheme 4.6: Chemical redox reactions of palladium complexes containing a tridentate verdazyl ligand and either an acetonitrile ligand (top) or a chloro ligand (bottom). ... 138

Scheme 4.7: Reactivity of radical/acetonitrile complex 4.3 with n-butylamine ... 145

Scheme 4.8: Two possible resonance forms of amido/verdazyl radical complex 4.13 ... 146

Scheme 4.9: Synthesis of n-butylamine complex 4.12 from verdazyl-anion/acetonitrile complex 4.8... 147

Scheme 4.10: Attempts to deprotonate n-butylamine complex 4.12 ... 147

Scheme 4.11: Oxidation of n-butylamine complex 4.12 using DDQ/HBF4 to generate 4.3 ... 149

Scheme 4.12: Reaction of n-butylamine complex 4.12 with [Magic Blue][SbCl6] to generate radical/chloro complex 4.11... 150

(19)

Scheme 5.1: Proposed synthesis of pyrrolidine complex 5.3 via an intramolecular radical

cyclization reaction ... 166

Scheme 5.2: Proposed example of CH amination using a radical verdazyl/amido palladium

complex ... 168

Scheme 5.3: Proposed ligand-centered reductive elimination of biphenyl from a

(20)

List of Tables

Table 2.1: Selected bond lengths (Å) for 2.1, 2.5 and 2.6 ... 31

Table 2.2: EPR parameters for phosphine-verdazyl 2.1 and palladium complex 2.6. Hyperfine coupling constants are given in G. ... 36

Table 2.3: Selected bond lengths (Å) for 2.9 and 2.6 ... 42

Table 2.4: Wavelength of maximum absorption for the major visible band in the electronic spectra of 2.9 in various solvents ... 48

Table 2.6: CO stretching frequencies (cm-1) of relevant verdazyl ligands ... 59

Table 3.2: CO stretching frequencies (cm-1) of 3.1, 1.44, 3.3, and 3.4 ... 80

Table 3.3: EPR Parameters for 3.3 and 3.4. Hyperfine coupling constants are given in G. ... 81

Table 3.4: Selected bond lengths (Å) for 3.2, 3.4 and [1.41]- ... 85

Table 3.6: EPR Parameters for 3.5. Hyperfine coupling constants are given in G. ... 98

Table 3.7: Electrochemical data for 3.4, 3.2 and 3.5. Potentials are in V vs Fc/Fc+ in CH2Cl2 solution ... 100

Table 3.8: Selected bond lengths (Å) for 3.6, 2.14, 3.4 and 3.2 ... 104

Table 4.2: EPR Parameters for 4.3. Hyperfine coupling constants are given in G. ... 122

Table 4.5: Electrochemical parameters for radicals 4.1, 4.9 and 4.10 and palladium complexes 4.7, 4.8, 1.44, and 1.46. Potentials are in V vs Fc/Fc+. The CVs are in either acetonitrile (4.1, 4.7, 4.8, and 4.9) or dichloromethane (4.10, 1.44, and 1.46). ... 135

(21)

Table 4.6: Selected bond lengths (Å) for chloro complexes 4.11 and 4.7 and acetonitrile complexes

4.3 and 4.8 ... 140

Table 4.7: EPR Parameters for 4.11 and 4.3. Hyperfine coupling constants are given in G. .... 142

Table 4.8: Electrochemical data for 4.12, 4.7 and 4.8. Potentials are in V vs Fc/Fc+ in CH3CN solution. ... 149

Table A-1: Crystallographic parameters for 2.1, 2.5, and 2.6 ... 178

Table A-2: Crystallographic parameters for 2.9 and 2.14... 179

Table B-1: Bond lengths (Å) for 2.1 ... 184

Table B-2: Bond Angles (º) for 2.1 ... 185

(22)

Table B-15: Bond lengths (Å) for 3.5 ... 200 Table B-16: Bond Angles (º) for 3.5 ... 201 Table B-17: Bond lengths (Å) for 3.11 ... 202 Table B-18: Bond Angles (º) for 3.11 ... 203 Table B-19: Bond lengths (Å) for 4.3 ... 204 Table B-20: Bond Angles (º) for 4.3 ... 205 Table B-21: Bond lengths (Å) for 4.7 ... 206 Table B-22: Bond Angles (º) for 4.7 ... 207 Table B-23: Bond lengths (Å) for 4.8 ... 209 Table B-24: Bond Angles (º) for 4.8 ... 210 Table B-25: Bond lengths (Å) for 4.11 ... 212 Table B-26: Bond Angles (º) for 4.11 ... 213 Table D-1: EPR parameters for 2.1, 2.5, and 2.6. Hyperfine coupling constants are given in G.

... 240

Table D-2: Electrochemical parameters for free radical 4.1, acetonitrile complexes 4.3 and 4.8,

and chloro complexes 4.11 and 4.7. Potentials are in V vs Fc/Fc+ and the CVs are in acetonitrile. ... 244

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

List of Abbreviations

(CH3CN)2PdCl2 bis(acetonitrile)dichloropalladium(II) (PhCN)2PdCl2 bis(benzonitrile)dichloropalladium(II) [(CH3CN)4Pd][BF4]2 tetrakis(acetonitrile)palladium(II) tetrafluoroborate °C degrees Celsius Å angstroms

a hyperfine coupling constant AgBF4 silver tetrafluoroborate

Anal. Calc. analytical calculated BF4 tetrafluoroborate

Boc tert-butoxycarbonyl

Boc2O di-tert-butyl dicarbonate

Bu butyl CD2Cl2 dichloromethane-d2 CD3CN acetonitrile-d3 CDCl3 chloroform-d cm centimeters cm-1 wavenumber Cp cyclopentadienyl

CV cyclic voltammetry or cyclic voltammogram d doublet (NMR peak descriptor)

(32)

DDQ 2,3-dichloro-5,6-dicyano-1,4-benzoquinone DDQH2 2,3-dichloro-5,6-dicyano-1,4-hydroquinone

DFT density functional theory DMSO dimethyl sulfoxide DMSO-d6 dimethyl sulfoxide-d6

dmtacn 1,4-dimethyl-1,4,7-triazacyclononane DPPH 2,2-diphenyl-1-picrylhydrazyl

e- electron

Eº redox potential

Eºox / Eox oxidation potential Eºred / Ered reduction potential

EPR electron paramagnetic resonance ESI electrospray ionization

Et ethyl

Fc/Fc+ _{ferrocene/ferrocenium}

FeCp*2 decamethylferrocene

FT-IR Fourier transform infrared

g g-value or gram

G Gauss

GHz gigahertz

HBF4 fluoroboric acid

(33)

HRMS high-resolution mass spectrometry Hz hertz I nuclear spin in vacuo in a vacuum iPr isopropyl J coupling constant K Kelvin KBr potassium bromide L neutral ligand or litre

LIFDI liquid injection field desorption ionization LiHMDS lithium bis(trimethylsilyl)amide

m medium (IR descriptor) or multiplet (NMR)

M molar

m/z mass-to-charge ratio

Magic Blue tris(4-bromophenyl)aminium

MeCN acetonitrile Mes mesityl MHz megahertz mL millilitre mM millimolar mmol millimole Mp melting point

(34)

MS mass spectrometry mV/s millivolts per second

NEt3 triethylamine

nm nanometre

NMR nuclear magnetic resonance NO nitrosyl or nitric oxide

NO+ nitrosonium

O2 molecular oxygen

Pd2(dba)3 tris(dibenzylideneacetone)dipalladium(0)

Ph phenyl

PhCN benzonitrile

PhICl2 iodobenzene dichloride

ppm parts per million

Q-TOF quadrupole time of flight R organic functional group RAL redox-active ligand

s singlet (NMR), strong (IR) SbCl6 hexachloroantimonate

SCE saturated calomel electrode SOMO singly occupied molecular orbital

T temperature

(35)

TEMPO 2,2,6,6-tetramethylpiperidine 1-oxyl thf tetrahydrofuran UV ultraviolet V volt Vd verdazyl Vis visible δ chemical shift

ε molecular extinction coefficient

λ wavelength

λmax wavelength of maximum electronic absorption

(36)

Acknowledgements

I would like to first acknowledge my supervisor Dr. Robin Hicks for his continued support throughout my degree. Robin has served as an amazing mentor and has helped shape me into the scientist that I am today. He gave me the freedom to pursue my own ideas, which has allowed me to develop valuable skills as an independent researcher. Many thanks Robin, I am very grateful to have had you as my supervisor.

I also want to thank all of the members of the Hicks group (past and present): Genevieve Boice, Dr. Emma Davy, Dillon Hofsommer, Erica Hong, Cooper Johnston, Shaun MacLean, Dr. Graeme Nawn, and Nick Richard. We have had some great conversations in the office; from science to sports, our discussions are always entertaining and often informative. In addition, I would like to acknowledge all of the undergraduate researchers who have helped contribute to the work in this thesis: Trevor Bolduc, Rehan Higgins, Elena Liles, Zach Mckay, Alex Wang, and Cam Zheng. I would also like to thank my UVic friends who have supported me throughout my degree: Dr. Aman Bains, Fraser Burns, Graham Garnett, Sundiata Kly, Aiko Kurimoto, and many more.

I would like to acknowledge members of the staff in the department of chemistry at UVic. I have to thank Andrew MacDonald and Shubha Hosalli for keeping our instruments and computers running as well as Chris Secord and Jeff Trafton for their excellent work repairing our vacuum pumps. I also have to thank Ori Granot for his help with mass spectrometry and Chris Barr for his fantastic NMR services. Thanks to the senior lab instructors who have supported me as both an undergraduate and a graduate student: Dr. Peter Marrs, Kelli Fawkes and Dr. Dave Berry. I would also like to thank all of the office staff for their help throughout my degree.

(37)

I also have to thank all of the faculty members in the chemistry department. Many of you have played an important role in my education as both teachers and researchers. In particular, I would like to thank Dr. David Berg, Dr. Neil Burford, Dr. Natia Frank, Dr. Tom Fyles, Dr. Scott McIndoe, Dr. Lisa Rosenberg, and Dr. Frank van Veggel. Also, I have to thank Dr. Fraser Hof and his research group for being so accepting of me at their weekly “Beer Friday” events.

I would also like to acknowledge those who have contributed to this thesis from other institutions. Dr. Pierre Kennepohl and his graduate student Weiying He (University of British Columbia) recorded the EPR spectra in Chapter 4. I would also like to thank Dr. Robert McDonald and Dr. Michael J. Ferguson at the University of Alberta for collecting X-Ray crystallography data for me near the beginning of my degree. I also have to thank Dr. Brian Patrick at the University of British Columbia for collecting X-Ray crystallography data for me. Brian has been extremely helpful in teaching me how to solve my own crystal structures and providing assistance whenever I came across a tough problem.

Lastly, I have to thank my family and friends who have always been there for me. My mom, dad and brother have been some of the most influential people in my life and I cannot thank you enough for your love and support.

(38)

Chapter 1: Introduction and Background

1.1 Introduction to Redox-Active Ligands

The redox chemistry of transition metals has been a focus of fundamental and applied research for over a century. A fair amount of effort has been put forth to understand electron transfer processes to and from metal complexes and this has resulted in the discovery of many synthetically useful reactions. The ligands that make up the surrounding framework of a coordination complex traditionally play a supporting role and the metal atom carries out the redox chemistry. But in some cases the ligands also play an active role in the redox chemistry of a metal complex.

In 1966, Jørgenson made a distinction between redox-active and redox-inert ligands, stating that “ligands are innocent when they allow oxidation states of the central metal atom to be defined”.1_{Nitric oxide (NO) was given as an example of a “suspect ligand” due to the ambiguity}

in assignment of oxidation states for metal nitrosyl complexes. Over the next decade, several research groups would help validate his theory through a series of publications investigating the electronic structure of dithiolene ligands (1.1).2 For the most part, the ligands were found to exist in their radical, monoanionic charge state and as a result they were labelled as “non-innocent ligands”.3_{More specifically, a non-innocent ligand is a ligand in a metal complex where the}

(39)

The field of non-innocent ligands began to grow as more chemists became intrigued by the challenges associated with characterizing their electronic structure. Research in this area eventually developed into studies on the redox chemistry of these ligands, giving rise to a new term: redox-active ligands (RALs). In this thesis, the term redox-active ligand will be used to refer to any ligand that can exist in two or more different charge states. More recently, focus has shifted towards applications of this new subset of ligands in stoichiometric and catalytic reactions.4 RALs can also potentially find use in materials-based applications such as energy storage,5_magnetism,6

and molecular conductors7 or switches8.

1.2 Classes of Redox-Active Ligands

Many RALs are adapted from the structure of ortho catecholate ligand 1.2 (Scheme 1.1a). Amidophenolato (1.3) and phenylenediamido (1.4) ligands benefit from an increased degree of tunability through substitution at the nitrogen atom(s). These types of ligands can access three different oxidation states which are separated by single-electron redox steps (Scheme 1.1b).

(40)

Scheme 1.1: (a) Dioxolenes and related RALs. (b) The triad of oxidation states that are available

to 1.2-1.4

Other types of RALs include pyridine diimines (1.5)9, porphyrins (1.6)10, carbenes11, and some amines12, as well as small molecules such as dioxygen (O2) and NO. Although there is no

accepted definition, the term RAL is typically used to refer to a ligand that can be chemically or electrochemically oxidized or reduced within a “moderate” potential window (“moderate” typically refers to ±2 V relative to the Fc/Fc+ redox couple). Under this definition, there are many ligands that can be regarded as redox-active. A recent essay by Kaim13_{highlights some}

(41)

1.3 Physical Oxidation States

It is commonly accepted practice to assign a formal oxidation state to a transition metal by looking at the charge that would remain on the metal after all of the ligands have been removed in their regular closed-shell configuration (i.e. with all of their electron pairs).15 When the ligand is redox-active, however, it cannot be assumed that it exists in this “normal” configuration and thus the formal oxidation state may not provide an accurate description of the electronic structure of the complex. To account for this, Wieghardt articulated the concept of a physical oxidation state in which the oxidation state of a metal is assigned by using physical methods to count the number of d-electrons.16 Formal and physical oxidation states are the same in most cases, but there can often be a discrepancy when one or more ligands are redox-active. Consider resonance structures

1.8, 1.9 and 1.10 for the copper complex shown in Scheme 1.2.16 The formal oxidation state of copper would be Cu(III) because the amidophenolate ligand would be treated as a dianion (1.8). However, physical methods such as X-ray crystallography established a physical oxidation state of Cu(II) and a monoanionic radical charge state for the ligand (a combination of resonance structures 1.9 and 1.10).16 In this thesis, the term “charge state” refers to the overall charge that

(42)

would be left on a ligand after the metal has been removed in its physical oxidation state (analogous to the term “oxidation state” but referring to the state of a ligand).

Scheme 1.2: Resonance structures of a copper-iminosemiquinone complex

Vanadium complex 1.11 also raises an interesting point in terms of oxidation state assignment.3 The formal oxidation state assignment of V(VI) would suggest an improbable electron configuration of 3p5. A more reasonable explanation is to invoke the non-innocence of dithiolene ligands. In fact, a physical oxidation state of V(IV) has been experimentally determined for 1.11 using various different physical methods, which indicates that the ligands have been oxidized to some extent.17_{The presence of a RAL does necessarily imply that the formal and}

physical oxidation states of a metal complex will differ, but the possibility of a discrepancy necessitates a physical investigation into the electronic structure of a “suspect” metal complex.

(43)

There are a number of different methods that can provide insight into the physical oxidation state of a metal. One of the primary tools for determining a physical oxidation state is X-ray crystallography. This method is more frequently used to probe the charge state of the non-innocent ligand, which in turn provides the oxidation state of the metal. When a ligand is oxidized or reduced, its bond metrics typically change by measureable amounts. In this respect, the length of certain bonds can be used to determine the charge state of a particular ligand. In fact, standard bond lengths for each charge state have been reported for some of the more popular RALs in order to facilitate the process of assigning physical oxidation states.16, 18 For example, the average bond lengths in dioxolene ligands (in different charge states) have been determined from X-ray structures with various different metals (Scheme 1.3).19 Progressing from left to right, the CO bond order decreases and as a result these bonds lengthen. In addition, the aromatic ring in the fully reduced state (far right) renders all of the CC bond lengths nearly equivalent, which is not the case in the other two charge states.

(44)

Scheme 1.3: Average bond lengths (Å) of dioxolene ligands19

The physical oxidation state of a metal is sometimes also referred to as a spectroscopic oxidation state. In many cases, the oxidation state of a metal is a measurable quantity than can be determined using spectroscopic methods.16 In this regard, a spectroscopic oxidation state is the oxidation state of a metal that has been experimentally determined using spectroscopic techniques. For example, electronic spectroscopy can be used to determine the number of d-electrons in a metal which also provides the physical oxidation state.20 A more specific case is the assignment of oxidation state in iron complexes using 57Fe Mössbauer spectroscopy.21 Electronic spectroscopy can also provide information about the charge state of a ligand. In the case of complex 1.12, the monoanionic iminosemiquinone ligands were assigned as such partly due to the presence of a ligand-to-ligand charge transfer band in the electronic spectrum.16 Magnetic measurements also confirmed the presence of two radical ligands and a paramagnetic Cu(II) ion (d9).16

More recently, computational chemistry has emerged as a powerful tool in the elucidation of physical oxidation states. By comparing experimental data with calculated structures or spectra,

(45)

researchers are able to confirm their computational model and gain valuable insight into the electronic structure of a metal complex and its surrounding ligands. This type of approach was applied by Neese and coworkers in a study involving a series of metal complexes containing redox-active dithiolene and phenylenediamine ligands.19 Density functional theory (DFT) or ab-initio methods were used to generate a computational model that accurately described the observed trends in the experimental spectra.

1.4 The Role of Redox-Active Ligands in Stoichiometric and Catalytic

Reactions

One of the most attractive features of RALs is their ability to store and transfer electrons during chemical reactions. Many catalysts rely on “classical” transition metal redox chemistry, where the metal is oxidized and reduced. However, the past decade has seen the development of many systems in which the ligands play an active role in the redox chemistry of the reaction. This new approach allows the ligand to act as the electron reservoir, resulting in reactivity pathways that differ from traditional catalysts which rely on metal-centered redox chemistry. The use of RALs in various different types of chemical reactions has been the subject of several recent reviews.4, 22 While these reactions may be new and innovative, they have not replaced traditional catalysts or reagents. The field of RAL reactivity is still in its infancy and it has yet to make an impact on modern organic synthesis.

There is a growing interest in studying RALs bound to first-row metals with the ultimate goal of creating cheaper catalysts.23 Noble metals are expensive, but the two-electron redox chemistry that they tend to favour make them a good choice in catalysis. First-row metals are much

(46)

cheaper, but they are known for single-electron redox steps which can result in reactivity that is difficult to control.23 One solution to this problem is to pair first-row metals with RALs to generate a complex that is capable of performing multi-electron chemistry. The fungal enzyme galactose oxidase uses this type of approach in the aerobic oxidation of D-galactose to its corresponding aldehyde.24_{A modified tyrosine residue in the active site of this enzyme exists as a neutral,}

open-shell phenoxyl radical (Figure 1.1). Substrate oxidation is accomplished by delivering a single electron to both the radical ligand and the central copper ion.25

Figure 1.1: The active site of galactose oxidase

Several structural mimics of galactose oxidase have been reported in the literature.26 Complex 1.13 catalyzes the aerobic oxidation of primary alcohols which includes difficult substrates such as ethanol and methanol.27 The importance of the RAL in this process is demonstrated by the reactivity of zinc analogue 1.14 which contains a Zn(II)ion. This d10 metal is redox-inert, making the salen ligand the only component that can accept electrons from the incoming alcohol.

(47)

The bis(imino)pyridine framework depicted in Scheme 1.4a has become a popular ligand in the design of new RAL-iron catalysts, particularly those involving C-C bond formation. The cycloaddition reaction shown in Scheme 1.4b is thermally forbidden without the aid of a catalyst or photochemical excitation.28 In the presence of 1.15, however, the reaction proceeds through a mechanism where the ligand acts as an electron reservoir. The catalytic cycle involves the temporary storage of two electrons on the ligand which allows for the new C-C bonds to form.29

As an added bonus, the redox activity of the ligand allows the iron atom to remain in the 2+ oxidation state throughout the cycle, avoiding Fe(0) intermediates which are known sources of catalyst decomposition. 1.15 has also been used as a catalyst in several other reactions which follow similar ligand-centred redox mechanisms.22a

(48)

Scheme 1.4: (a) Charge states available to a bis(imino)pyridine ligand. (b) Proposed mechanism

for the cycloaddition reaction catalyzed by 1.15

Reductive elimination is an important process in many catalytic cycles, but metal-centered reduction can also be a source of catalyst decomposition. The final step of the proposed cycloaddition mechanism in Scheme 1.4b is formally a reductive elimination. Unlike traditional

(49)

reductive elimination reactions, however, the electrons from the alkyl ligands were transferred to the redox-active pyridine diimine ligand while the iron atom remained in the same oxidation state, Fe(II). Heyduk and coworkers have also observed a ligand-centered reductive elimination reaction using zirconium complex 1.16 (Scheme 1.5).30 Two-electron oxidation generates intermediate 1.17 which eliminates biphenyl upon warming. The electron-transfer processes are localized on the ligands while the Zr(IV) center is redox inert.

Scheme 1.5: Ligand-centered reductive elimination from a Zr(IV) center

Heyduk has also experimented with ligand-centered oxidative addition using the zirconium-iminosemiquinone system. Exposing 1.18 to chlorine gas resulted in a two-electron oxidation of the metal complex to afford dichloro complex 1.19 (Scheme 1.6).31 The d0 Zr(IV)

(50)

atom cannot be oxidized any further so clearly the amidophenolate ligands must have played a role in the redox reaction. The monoanionic charge state of the ligands in 1.19 were confirmed using X-ray crystallography and EPR spectroscopy. This type of “oxidative addition” has also been reported for a uranium complex with the same ligand system where PhICl2 was used as the

oxidant.32

Scheme 1.6: Ligand-centered oxidative addition

Carbon-carbon cross coupling reactions have become paramount in synthetic organic chemistry and for the most part these reactions require some sort of precious metal catalyst. Soper and coworkers have designed a system that is able to perform Negishi-type cross coupling using amidophenolate-cobalt complex 1.20 (Scheme 1.7).33 The electrons required for this reaction come from the ligands while the metal remains in the Co(III) oxidation state. The metal-carbon bond forming step to generate 1.21 is best described as a nucleophilic attack by the copper complex on the electrophilic alkyl halide. Each ligand is oxidized by a single electron to supply the electrons for the newly formed Co-C bond. Reaction with an organozinc compound afforded the C-C coupled product and the ligands were reduced back to their original dianionic charge state.

(51)

Scheme 1.7: Negishi-type cross coupling using RALs as electron reservoirs

The involvement of RALs in chemical reactions is not limited to electron storage in redox reactions. For example, oxidation or reduction of a ligand can alter the electronic structure of the metal it is bound to and potentially change the reactivity at this metal centre. This concept has been used in the design of “redox switches” where a RAL, often tagged with a metallocene, can be oxidized or reduced to modulate the reaction rate. One of the first examples of this was demonstrated using the cobaltocene-tagged rhodium complex 1.22 in olefin hydrogenation.34 Single electron reduction of the doubly charged 1.22 complex led to a 16-fold increase in reaction rate for the hydrogenation of cyclohexene. In a similar example, the polymerization of lactide using titanium complex 1.23 can be controlled by changing the charge state of the two ferrocene units.35_{Two-electron oxidation of neutral 1.23 to the bis-ferrocenium (2+) analog using Ag(I)}

(52)

slowed down the rate of polymerization substantially. The original rate could be restored by treating the solution with decamethylferrocene (FeCp*2), which is assumed to regenerate the

original neutral complex.

1.5 Chemical Reactions Involving Redox-Active Ligands Bound to

Noble Metals

The pursuit of base metal catalysts that can carry out multi-electron chemistry has led to the development of many RAL complexes of the first-row transition metals.23 This search for cheaper catalysts has left the area of noble metal RAL complexes relatively unexplored. However, there are still several examples of these types of complexes and some have been shown to be active catalysts.

Traditional iridium complexes that catalyze the oxidation of alcohols typically cycle between Ir(I) and Ir(III) oxidation states and rarely proceed through a radical mechanism.36 In contrast, Grützmacher and coworkers have reported an iridium complex that catalyzes the oxidation of alcohols without invoking the redox chemistry of the metal (Scheme 1.8).37 The diamide ligand was found to play a very active role, participating in both the redox chemistry and the hydrogen atom transfer process. The beta hydrogen atom from the alkoxide ligand in 1.24 is

(53)

transferred to the radical amine ligand, representing a key step in the reaction mechanism. This highlights an interesting area of catalysis known as ligand cooperativity, where a ligand directly participates in a reversible chemical transformation. Galactose oxidase also makes use of this principle when a phenoxyl radical ligand abstracts a hydrogen atom from the alkoxide substrate. In this regard, ligand cooperativity is somewhat of a bioinspired approach to catalysis and it is receiving an increased amount of attention38 as chemists continue their search for new reaction pathways.

Scheme 1.8: A selected part of the catalytic cycle reported by Grützmacher37_{for alcohol}

oxidation. The stoichiometric oxidant is benzoquinone.

Catalytic oxidation of H2 has also been reported using an iridium-RAL complex.

Compound 1.25 does not bind H2 and shows no catalytic activity on its own; however, oxidation

of the ligand to its radical iminosemiquinone form increases the Lewis acidity of the metal centre (Scheme 1.9).39 The oxidized complex (1.26) is now able to bind H2 which is subsequently

deprotonated by a non-coordinating base (2,6-di-tert-butylpyridine). The electrons that were formally derived from H2 are used to reduce the radical ligand and regenerate 1.25.

(54)

Scheme 1.9: Dihydrogen oxidation catalyzed by an iridium complex

Tanaka and coworkers have reported some ruthenium complexes that utilize RALs for oxidation chemistry. Binuclear catalyst 1.27 contains redox-active dioxolene ligands that facilitate the electrochemical oxidation of water to O2 with over 33000 turnovers.40 Replacing the dioxolene

ligands with bipyridine renders the complex inactive towards water oxidation, which demonstrates the importance of the dioxolene ligands in this reaction. Unlike traditional ruthenium-based water oxidation catalysts that access high energy Ru(IV)/(V) oxidation states,41_{1.27 shuttles between}

Ru(II)/(III) states with the support of ligand-centered oxidations.42 Tanaka has also studied alcohol oxidation using ruthenium complex 1.28, which again relies on the redox chemistry of a dioxolene ligand to carry out the reaction.43 1.28 was found to oxidize benzyl alcohol to benzaldehyde under electrolysis at a potential of +0.4 V (vs SCE).

(55)

Palladium has received very little attention in the context of RALs despite its widespread use in catalysis and coordination chemistry. This metal accounts for a number of different catalysts that can be utilized in a variety of different organic transformations. However, there are very few instances where the reactivity of a palladium-RAL complex has been explored. A few key examples will be given below.

External oxidants have been utilized in reactions catalyzed by palladium for a long time, a notable example being copper(I) chloride in the Wacker process.44 Possibly some of the earliest examples of palladium interacting with an organic redox-active molecule were reported by Bäckvall and coworkers when benzoquinone was incorporated into various oxidation reactions.45 These studies (and some more recent examples46_{) likely involve intermediates with a quinone}

(56)

weakly bound to palladium (similar reports47 have been published using TEMPO ((2,2,6,6-tetramethylpiperidin-1-yl)oxyl, 1.29). There are also examples where a dioxolene has been covalently linked to a phosphine (1.30)48 or N-heterocyclic carbene ligand (1.31)49.

Palladium complexes of the more conventional RALs exist as well, although most of the reports only study the redox activity of these ligand electrochemically. Early work by Wieghardt focused on the synthesis and electrochemistry of a series of bis-RAL complexes (1.32).16, 50 This work was eventually expanded to include chemical oxidation and reduction experiments to generate an electron-transfer series for each set of ligands.51_{The redox events were found to be}

ligand-centered and the palladium atom remained in the Pd(II) oxidation state. Since these initial reports by Wieghardt there have been several examples of RALs bound to palladium52 including some from our group53. However, these studies only explore the fundamental redox chemistry of the ligands. There are a few instances where the reactivity of a palladium-RAL complex was also

(57)

considered.11, 54 For example, the carbene ligand in 1.33 has been shown to act as an electron reservoir during electrocatalytic CO2 reduction, which reportedly circumvents a major pathway

for catalyst deactivation.55

A recent pair of reports by van der Vlugt and coworkers demonstrate how radical reactivity can be achieved using a palladium complex (1.34).56 The amidophenolate ligand plays a key role in catalyzing the C-H amination of organic azide 1.35 (Scheme 1.10). Single-electron transfer from this ligand to the coordinated azide generates an imidyl radical after the loss of N2. Intramolecular

hydrogen atom abstraction followed by a radical rebound step results in the formation of pyrrolidine 1.36, which is cleaved off the metal using di-tert-butyl dicarbonate (Boc2O). The

radical reaction pathway observed in this catalytic cycle contrasts the traditional two-electron chemistry of palladium and represents the first example of a radical-type transformation catalyzed by a RAL-Pd complex.

(58)

Scheme 1.10: Radical mechanism for CH amination using palladium complex 1.34

1.6 Verdazyl Radicals as Redox-Active Ligands

Verdazyl radicals are a class of paramagnetic organic compounds in which the unpaired electron is delocalized across four nitrogen atoms (1.37). This delocalization accounts for the high level of stability in these compounds which makes them easy to isolate and handle despite their radical nature. In 2007 our group began investigating the redox chemistry of these radicals.57 They can be reversibly reduced by a single electron to arrive at a monoanionic charge state or reversibly oxidized by one electron to generate a closed-shell cation (Scheme 1.11). The general structure of

(59)

a verdazyl allows for various different substitutions around the ring. Electron-donating and electron-withdrawing substituents affect the oxidation and reduction potentials, which allows for a high degree of electronic tunability.57

Scheme 1.11: The redox chemistry of verdazyl radicals

The coordination chemistry of verdazyls has been studied extensively and the primary focus of past research has been on the development of new molecular magnets.58 More recently, our group has been studying the redox chemistry of verdazyl-metal complexes. The redox-inert zinc(II) ion in compounds 1.38 and 1.39 creates an opportunity to examine the redox chemistry of the ligands directly. Cyclic voltammetry experiments revealed that the redox properties of the free verdazyls were retained upon coordination (i.e. each radical could be reversibly oxidized or reduced).59 The oxidation and reduction potentials in both complexes shift to more positive potentials with respect to the free ligand.

(60)

When a verdazyl is coordinated to a redox-active metal the possibility arises for ligand non-innocence. This is the case in ruthenium complex 1.40, which was the first example of a verdazyl acting as a non-innocent ligand.60 Charge transfer from the ruthenium centre to the verdazyl π* orbital was found to be strongly dependent on the nature of the ancillary diketonate ligands. In the case of 1.41 (R = CF3), there was very little non-innocence observed and the formal

representation of Vd(0)-Ru(II) was an accurate description of the electronic structure (where x represents the charge state of the verdazyl ligand and y represents the physical oxidation state of the metal in Vd(x)-Ru(y)). In 1.40 (R = CH3), the more electron-rich acetylacetonate ligands alter

the electronic structure significantly. The resonance structure Vd(–)-Ru(III) becomes significant in this case due to the effective transfer of electron density from the electron-rich metal to the verdazyl ligand. Similar results were observed for binuclear compounds 1.42 where the verdazyl is a bridging ligand between two ruthenium atoms.61 Again, the extent of electron-withdrawing character on the diketonate ligands dictated the amount of non-innocence that was observed.

The coordination chemistry of verdazyl radicals with palladium has also been explored. Compounds 1.43, 1.44, and 1.45 were originally synthesized with the intent of measuring

(61)

intermolecular magnetic interactions.62 In a subsequent publication palladium complexes 1.44,

1.46, and 1.47 were characterized electrochemically.63 In all three cases, the two observed redox events were found to be ligand-centered (one oxidation and one reduction) while the 2+ oxidation state of palladium remained unchanged. The reversibility of the reduction waves in these compounds lends support to the possibility of chemically reducing the radical ligand. In theory, one-electron reduction of these complexes would generate a verdazyl ligand that exists in its monoanionic charge state.

1.7 Thesis Objectives

Over the past decade the Hicks group has made a fair amount of progress in establishing verdazyl radicals as a new class of RAL. Other group members have explored the redox activity of these ligands electrochemically57, 59, 63, but an in-depth chemical investigation is still lacking.