New Developments in Transition Metal Alkyl Chemistry

David Armstrong

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Chemistry University of Toronto

David Armstrong

Doctor of Philosophy

Department of Chemistry University of Toronto

2018 Abstract In the first section of this thesis, the synthesis and reactivity of the first reported palladium(IV) aryldiazenido complex is explored, as well as its relevance toward the potential for catalytic coupling cycles based on the Pd(II)/Pd(IV) redox pair. The synthesis was achieved by two-electron

- oxidation of the anionic palladium(II) complex KPd(Tp*)Me2 (Tp* = hydridotris(3,5- dimethylpyrazolyl)borate) by an aryldiazonium cation. The resulting Pd(IV) complex could be cleanly isolated in good yields, is stable for weeks as a solid when kept cold, and decays in solution over a period of days at room temperature. The decomposition was studied in great depth by both

NMR and EPR spectroscopy and it was determined that one-electron pathways were involved.

From the synthesis of the Pd(IV) aryldiazenido complex, side reactions were observed whereby free Tp*- ligand could be modified via hydride abstraction by an aryldiazonium cation to form the elusive tris(pyrazolyl)borane which could be trapped and isolated as its acetone adduct in a

Frustrated Lewis-pair type reaction. The resulting compound can act as a bidentate N-donor ligand with unusual axial steric bulk. Addition of a second equivalent of aryldiazonium cation further abstracts a 3,5-dimethylpyrazolide group to form a cationic intermediate which can again be trapped by acetone to form a C2-chiral cation.

In the later chapters of this thesis, extensive work towards the synthesis of 1- and 2- adamantyl anion equivalents was explored, as well as their utility in the synthesis of transition metal adamantyl complexes. The development of new reliable synthetic methods to produce adamantyl magnesium and zinc compounds (including the unknown diadamantylzincs) was achieved, producing clean and shelf stable compounds. The applications of these adamantyl anions toward organometallic chemistry was investigated. It was observed that the diadamantylzinc compounds were particularly potent in transmetallation reactions, cleanly producing several new transition metal and main group adamantyl complexes. While compounds containing mercury, gold, bismuth, tungsten, and platinum have now been synthesized, those of particular interest have the potential for C-H bond activation chemistry and could open access to new compound classes of facially trifunctionalized adamantanes.

iii

Acknowledgments

I would like to extend my gratitude toward Professor Ulrich Fekl for providing me the opportunity to not only complete this degree, but to take charge and make this research my own.

Thanks to my parents Kim and Keith for instilling me with a desire to never stop learning, and to my sister Heather for always pushing me to be better.

Thanks to the members of the Fekl lab past and present. My experience would not have been the same were it not for the atmosphere of friendship and support that has been ever present here.

Lastly, I would like to thank Fioralba Taullaj for not only her constant support personally, but also for bringing an attitude of hard work and dedication to the lab which really inspired me to push beyond my bounds.

Table of Contents

Acknowledgments...... iv

Table of Contents ...... v

List of Tables ...... ix

List of Figures ...... x

List of Schemes ...... xiii

List of Abbreviations ...... xvi

Contributions...... xviii

Chapter 1 - Introduction ...... 1

1.1 History of Metal Alkyl Chemistry ...... 1

1.2 Why Are We Interested in Metal Alkyls? ...... 2

1.3 Decomposition of Metal Alkyls ...... 5

1.4 Synthesis of Metal Alkyls ...... 9

1.5 C-H Bond Activation and Functionalization ...... 13

1.6 Computational Chemistry ...... 16

1.7 Organization of this Thesis ...... 18

1.8 References for Chapter 1 ...... 20

Chapter 2 - The First Palladium(IV) Aryldiazenido Complex – Relevance for C-C Coupling ....23

2.1 Abstract ...... 23

2.2 Introduction ...... 23

2.3 Results and Discussion ...... 28

2.4 Conclusion ...... 40

2.5 Experimental Section ...... 40

2.5.1 General Specifications ...... 40

2.5.2 Synthesis of 1,5-Cyclooctadiene Dimethyl Palladium(II) {Pd(COD)Me2)} ...... 41

2.5.3 Synthesis of Pd(Tp*)Me2(pmbd) ...... 41 v

2.5.4 X-ray structure determination of Pd(Tp*)Me2(pmbd) ...... 42

2.5.5 Thermolysis of Pd(Tp*)Me2(pmbd) in C6D6 ...... 44

2.5.6 Thermolysis of Pd(Tp*)Me2(pmbd) in Acetone-d6 ...... 46

2.5.7 Isolation and characterization of Pd(Tp*)Me3 (2.2) ...... 48

2.5.8 X-ray structure determination of Pd(Tp*)Me3 ...... 50

2.6 References for Chapter 2 ...... 52

Chapter 3 - Facile Transformations of hydridotris(3,5-dimethylpyrazolyl)borate – An

Intramolecular Frustrated Lewis Pair...... 55

3.1 Abstract ...... 55

3.2 Introduction ...... 55

3.3 Results and Discussion ...... 59

3.4 Conclusions ...... 64

3.5 Experimental Section ...... 65

3.5.1 General Specifications ...... 65

3.5.2 Synthesis of the acetone adduct of tris(3,5-dimethylpyrazol-1-yl)borane (Compound 3.1) ...... 65

3.5.3 Synthesis of the bis-acetone adduct of bis(3,5-dimethylpyrazol-1-yl)borinium (Compound 3.2) ...... 67

3.5.4 Complexation of ZnBr2 with 3.1 ...... 69

3.6 References for Chapter 3 ...... 69

Chapter 4 - Synthesis and Characterization of 1- and 2-Adamantyl Anions ...... 70

4.1 Abstract ...... 70

4.2 Introduction ...... 70

4.3 Results and Discussion ...... 74

4.4 Conclusions ...... 88

4.5 Experimental Section ...... 90

4.5.1 General Specifications: ...... 90 vi

4.5.2 Isolation of 1-adamantylzinc bromide as its THF adduct [1-AdZnBr(THF)x] (4.1a, THF-solvate) ...... 91

4.5.3 Isolation of 2-adamantylzinc bromide as its THF adduct [2-AdZnBr(THF)x] (4.1b, THF-solvate) ...... 91

4.5.4 Synthesis of 1-adamantylmagnesium bromide as its THF adduct [1- AdMgBr(THF)x] (4.2a, THF-solvate)...... 91

4.5.5 Synthesis of 2-adamantylmagnesium bromide as its THF adduct [2- AdMgBr(THF)x] (4.2b, THF-solvate) ...... 92

4.5.6 Synthesis of bis(1-adamantyl)zinc [1-Ad2Zn] (4.3a) ...... 92

4.5.7 Synthesis of bis(2-adamantyl)zinc [2-Ad2Zn] (4.3b) ...... 93

4.5.8 Synthesis of 2,2’bipyridyl-bis(1-adamantyl)zinc [1-Ad2Zn(bipy)] (4.4a) ...... 94

4.5.9 Synthesis of 2,2’bipyridyl-bis(2-adamantyl)zinc [2-Ad2Zn(bipy)] (4.4b) ...... 96

4.5.10 Synthesis of bis(1-adamantyl)mercury [1-Ad2Hg] (4.5a) ...... 98

4.5.11 Synthesis of bis(2-adamantyl)mercury [2-Ad2Hg] (4.5b): ...... 98

4.5.12 Synthesis of bis(2-adamantyl)bismuth(III) bromide [2-Ad2BiBr] (4.6b) ...... 99

4.5.13 Attempted Synthesis of bis(1-adamantyl)bismuth(III) bromide [1-Ad2BiBr] (4.6a) ...... 101

4.5.14 Generation of 1-adamantylgold(triphenylphosphine) [1-AdAu(PPh3)] (4.7a) ....101

4.5.15 Synthesis of 2-adamantylgold(triphenylphosphine) [2-AdAu(PPh3)] (4.7b) ...... 102

4.5.16 Generation of 1-adamantylgold(tricyclohexylphosphine) [1-AdAu(PCy3)] (4.8a) ...... 103

4.5.17 Synthesis of 2-adamantylgold(tricyclohexylphosphine) [2-AdAu(PCy3)] (4.8b) ...... 103

4.6 References for Chapter 4 ...... 104

Chapter 5 - Synthesis and Reactivity of Transition Metal Adamantyl Complexes ...... 108

5.1 Abstract ...... 108

5.2 Introduction ...... 108

5.3 Results and Discussion ...... 114

5.4 Conclusion ...... 128

vii

5.5 Experimental ...... 129

5.5.1 General Specifications: ...... 129

5.5.2 Transmetallation of (1-Ad)2Zn onto WCl6...... 129

5.5.3 Synthesis of 1,5-cyclooctadieneplatinum(II)(2-adamantyl)chloride [(COD)Pt(2-Ad)Cl] ...... 130

5.6 References for Chapter 5 ...... 133

Chapter 6 – General Conclusions and Future Directions...... 135

6.1 General Conclusions ...... 135

6.2 Palladium(IV) aryldiazenido complexes ...... 136

6.3 Tris(pyrazolyl)borane-based FLP ...... 138

6.4 Adamantyl anion equivalents ...... 140

6.5 Transition metal adamantyl complexes ...... 141

6.6 References for Chapter 6 ...... 142

viii

List of Tables

Table 2.1 Comparison of the redox behavior of palladium and platinum…………….………… 28

Table 2.2 Crystal data and structure refinement for Pd(Tp*)Me2(pmbd)………….……………. 43

Table 2.3 Crystal data and structure refinement for Pd(Tp*)Me3………………….………...….. 51

Table 3.1 Crystal data and structure refinement for 3.1……………………….………………... 64

Table 3.2 Crystal data and structure refinement for 3.2………………………….……………... 68

Table 4.1 Crystal data and structure refinement for 4.4a……………………….………………. 95

Table 4.2 Crystal data and structure refinement for 4.4b……………………….………………. 97

Table 4.3 Crystal data and structure refinement for 4.6b……………………….……………… 100

Table 5.1 DFT (B3LYP-D3/SDD) calculated thermodynamics for γ-CH activation of group III

tris(2-adamantyl) complexes………………………………………………………………. 116

Table 5.2 DFT (B3LYP-D3/SDD) calculated thermodynamics for γ-CH activation of group III

tris(2-adamantyl) complexes containing explicit THF solvation in order to improve the

favorability…………………………………………………….………………………...... 117

Table 5.3 Comparison of experimental EA to some expected stoichiometries for a tungsten

adamantyl…………………………………………………..….………………………...... 123

Table 5.4 Crystal data and structure refinement for (COD)Pt(2-Ad)Cl………….……………. 132

List of Figures

Figure 1.1 A simplified representation of olefin hydrogenation at a metal surface…….………... 3

Figure 2.1 Common Aryl cation equivalents (halides, triflates, tosylates, acetates)………..…… 24

Figure 2.2 Multiple potential binding modes of aryldiazonium cations to a metal center…….… 26

Figure 2.3 Anisotropic displacement plot (20% probability ellipsoids) of 2.1………………….. 30

15 1 Figure 2.4 N– H gHMBC NMR spectrum of Pd(Tp*)Me2(pmbd) (2.1)…………………….… 32

Figure 2.5 Anisotropic displacement plot (20% probability ellipsoids) for 2.2……………….... 34

Figure 2.6 X-band EPR spectrum of the thermolysis mixture generated from 2.1 in the presence

of a PBN radical trap…………………………………………….……………………… 39

1 Figure 2.7 H NMR spectrum of Pd(Tp*)Me2(pmbd) in C6D6………....………………………. 45

1 Figure 2.8 H NMR spectrum of Pd(Tp*)Me2(pmbd) in C6D6 after heating for 1 h at 70 °C…... 46

1 Figure 2.9 H NMR spectrum of Pd(Tp*)Me2(pmbd) in acetone-d6………..…………….……... 47

1 Figure 2.10 H NMR spectrum of Pd(Tp*)Me2(pmbd) in acetone-d6 after heating for

1 h at 70 °C………………………………………………………………………….…... 48

1 Figure 2.11 H NMR spectrum of Pd(Tp*)Me3 (2.2) in C6D6 at 25 °C…………...……….……. 49

13 Figure 2.12 C{H} NMR spectrum of Pd(Tp*)Me3 (2.2) in C6D6 at 25 °C………...…….……. 50

Figure 3.1 Structures of the two isolated side products from the synthesis of Pd(IV)

aryldiazenido complexes………………………………………………………………... 56

Figure 3.2 Common Lewis acids and bases and Lewis Adduct………………………….……... 57

Figure 3.3 A demonstration of steric “frustration” in Lewis pairs……………………….……... 57

- Figure 3.4 Hydridotris(pyrazolyl)borate ligand class (R1 = R2 = H for Tp , R1 = R2 = CH3 for

Tp*-) and binding modes of Tp ligands which favour square planar and octahedral geometries, respectively………………………………………………………………… 58

+ Figure 3.5 Anisotropic displacement plots of 3.1 and 3.2 (cation in 2-BF4) showing 30%

probability ellipsoids……………………………………………………………………. 62

Figure 3.6 DFT Computed structure of the ZnCl2 adduct of compound 3.1, showing the potential

for single-face steric protection……………………………………….………………… 63

Figure 4.1 An atom of choice (blue) bound to a 1-adamantyl group (left) and to a 2-adamantyl

group (right), highlighting the structural similarity to a tert-butyl (left, 1-adamantyl) and

to an axially locked cyclohexyl (right, 2-adamantyl)…………………….……………... 74

Figure 4.2 UV-Vis spectra of 4.4a (red) and 4.4b (blue) in toluene…………….………………. 77

Figure 4.3 Anisotropic displacement plots (50% probability ellipsoids) for 4.4a (left) and 4.4b

(right)…………………………………………………………………….……………… 78

1 Figure 4.4 H NMR spectrum of 5a in C6D6 at 25 °C…………………………...………………. 80

13 1 Figure 4.5 C{ H} NMR spectrum of 5a in C6D6 at 25 °C…………………….………………. 81

1 Figure 4.6 H NMR spectrum of 5b in C6D6 at 25 °C………………………….……………….. 82

13 1 Figure 4.7 C{ H} NMR spectrum of 5b in C6D6 at 25 °C……………………….……………. 83

Figure 4.8 Anisotropic displacement plot (50% probability ellipsoids) for 6b….……………… 85

1 Figure 4.9 H NMR spectrum of 4.6b in C6D6 at 25 °C…………………………….………….. 86

13 1 Figure 4.10 C{ H} NMR spectrum of 4.6b in C6D6 at 25 °C……………………….………… 87

Figure 5.1 The famous “Schleyer synthesis” is now common route for the commercial synthesis

of adamantane………………………………………………………………….………. 109

Figure 5.2 Examples of 1-, 2-, and 3-dimensional “diamondoid” structures………….………. 109

Figure 5.3 Examples of FDA approved adamantane based pharmaceuticals…………….……. 111

Figure 5.4 Hypothetical classes of adamantane based compounds…………………….……… 112

Figure 5.5 DFT (B3LYP-D3/SDD) optimized structure of La(2-Ad)3 showing agostic

interactions…………………………………………………………………………….. 115

Figure 5.6 Ruthenium alkyl complexes which have undergone γ-CH activation…….………... 121

Figure 5.7 Examples of isolable tungsten(VI) metalacyclobutane complexes……….……….. 123

Figure 5.8 Attempted syntheses of Cl4W(2-Ad)2, from which no metal adamantyl could be

isolated………………………………………………………………………………… 123

1 Figure 5.9 H NMR (C6D6, 400MHz) spectrum of the crude reaction mixture of (1-Ad)2Zn and xi

WCl6 in toluene with proposed assignment…………………………………….……… 125

Figure 5.10 Anisotropic displacement plot (50% probability ellipsoids) for

(COD)Pt(2-Ad)Cl……………………………………………………………….……... 127

1 Figure 5.11 H NMR Spectrum of (COD)Pt(2-Ad)Cl in C6D6 at 25 °C………………….…… 131

xii

List of Schemes

Scheme 1.1 Overall reaction scheme for metal mediated coupling reactions……………….…… 2

Scheme 1.2 Representative example of Zr(IV) catalyzed olefin polymerization……………..….. 3

Scheme 1.3 Homogeneous olefin hydrogenation by Wilkinson's catalyst…………………….…. 4

Scheme 1.4 Common decomposition products of metal alkyl complexes on exposure to air or

moisture…………………………………………………………………………………... 5

Scheme 1.5 Commonly observed mechanism of β-hydride elimination……………………….… 5

Scheme 1.6 β-hydride elimination in alkylplatinum(II) complexes…………………………….... 6

Scheme 1.7 An example of β-methyl elimination from a Zirconium neopentyl complex……….. 7

Scheme 1.8 Metal 1-norbornyl and adamantyl complexes resist β-hydride elimination……….… 8

Scheme 1.9 A generic scheme for the reductive elimination of metal alkyls………………….…. 8

Scheme 1.10 General scheme for metal-alkyl homolytic bond cleavage……………………….... 9

Scheme 1.11 Common condition for the synthesis of metal alkyls via transmetallation……….... 9

Scheme 1.12 General scheme for an oxidative addition reaction……………………………….. 10

Scheme 1.13 Oxidative addition of alkyl bromides to a Pd(0) phosphine complex…….………. 11

Scheme 1.14 Example of the concerted oxidative addition mechanism………………….……... 11

Scheme 1.15 A Common example of the “nucleophilic” oxidative addition mechanism….…… 11

Scheme 1.16 Examples of cyclometallation of metal neopentyl complexes……………….…… 12

Scheme 1.17 Ethylene insertion into a molybdenum hydride, forming a metal alkyl……….….. 12

Scheme 1.18 The first report of organometallic C-H bond activation……………………….….. 14

Scheme 1.19 The famous Shilov cycle for catalytic C-H bond functionalization of methane….. 14

Scheme 1.20 The Periana cycle for conversion of methane to methyl sulfate………………….. 15

Scheme 2.1 General palladium catalyzed cross coupling cycle……………………………….... 22

Scheme 2.2 The two most common mechanisms of oxidative addition……………………….... 25

Scheme 2.3 Suggested catalytic coupling cycle using aryldiazonium cations………………….. 27

xiii

+ Scheme 2.4 Oxidative addition of ArN2 to Pt(II)…………………………..………………...... 28

Scheme 2.5 Synthesis of PdTp*Me2(pMBD) (2.1)…………………….……….……………….. 29

Scheme 2.6 A hypothetical 2-electron mechanism for the decomposition of 2.1 to form 2.2 and

organic products………………………………………………………………………… 36

Scheme 2.7 The preferred radical mechanism for the decomposition of 2.1 to form 2.2 and

organic products………………………………………………………………………… 38

Scheme 3.1 A potential pathway for the generation and trapping of a tris(pyrazolyl)borane…... 58

Scheme 3.2 DFT-computed (B3LYP/6-31G**) thermodynamics for the polymerization of

tris(pyrazolyl)borane (top) vs. FLP-like reactivity with acetone (bottom)…….………... 59

Scheme 3.3 Formation of 3.1 (A) and 3.2+ (B)………………………………….………………. 61

Scheme 4.1 Selected literature preparations of adamantyl anions…….………………………… 71

Scheme 4.2 The unusual reactivity of bromoadamantane with active magnesium…….... 72

Scheme 4.3 Reductive insertion of zinc metal in bromoadamantane, a reliable method... 73

Scheme 4.4 New routes to 1-adamantyl (labelled a) and 2-adamantyl (labelled b) anion

equivalents…………………………………………………….………………………… 75

Scheme 4.5 Synthesis of Ad2Zn from reduction of AdZnBr with the Schlenk equilibrium

shown………………………………………………………….…………………. 76

Scheme 4.6 Synthesis of adamantyl of Adamantyl Grignard reagent from Ad2Zn

with the Schlenk equilibrium shown…………………………….……………….. 76

Scheme 4.7 Demonstrated new routes to 1- and 2-adamantyl metal complexes…….….. 84

Scheme 5.1 The standard procedure for oxidation of adamantane to adamantanone.…. 110

Scheme 5.2 Conditions for 1-brominations of adamantane…………………….……… 110

Scheme 5.3 Potential applications of a single-face substituted 2,2’,2’’-adamant-triyl….…….. 112

Scheme 5.4 γ-CH activation reactions in dialkylplatinum(II) complexes…………….……….. 113

Scheme 5.5 Suggested decomposition mechanisms of homoleptic group (IV) neopentyls….... 113

Scheme 5.6 Proposed reaction scheme for synthesizing di- and trianionic adamantyls…….…. 114 xiv

Scheme 5.7 Proposed method for γ-CH activation of group III tris(2-adamantyl) complexes... 115

Scheme 5.8 Potential use of solvation energy to improve the thermodynamics of activation.... 116

Scheme 5.9 Generic synthetic scheme for group III metal adamantyl complexes……….……. 117

Scheme 5.10 Potential pathways to γ-CH activated adamantyls using metal complexes of group

(V) and later………………………………………………………………….………… 119

Scheme 5.11 Attempted syntheses of Cl2Ta(2-Ad)3, from which no metal adamantyl could be

isolated………………………………………………………………………………… 120

6 Scheme 5.12 Attempted syntheses of (η -C6Me6)(PMe3)Ru(2-Ad)2………………….……….. 121

Scheme 5.13 Classical synthetic methods for hexamethyltungsten(VI)…………….…………. 121

Scheme 5.14 Proposed reaction scheme based on 1H NMR of the observed products…...…….. 124

Scheme 5.15 First successful synthesis of an isolable adamantyl platinum complex….……… 126

Scheme 6.1 Proposed catalytic cycle for C-C homocoupling of aryldiazonium cations to

substituted biphenyls…………………………………...……………………………… 137

Scheme 6.2 Proposed formation of a C-N-N scorpionate ligand from compound 3.1 via C-H bond

activation by platinum….……………………………………………………………… 138

List of Abbreviations

1Ad 1-adamantyl substituent

2Ad 2-adamantyl substituent

Ad Both 1- and 2-adamantyl

+ ArN2 aryldiazonium cation

B3LYP Becke’s 3-parameter exchange with Lee-Yang-Parr correlation bipy 2,2”-bipyridine bs broad singlet

COD 1,5-cyclooctadiene

Cp- cyclopentadienyl

Cp*- pentamethylcyclopentadienyl

CVD Chemical vapour deposition

DFT Density functional theory diox 1,4-dioxane

DME 1,2-dimethoxyethane

DMPE 1,2-bis(dimethylphosphino)ethane

E Total energy

EA Elemental analysis

EPR Electron paramagnetic resonance fac Facial

FDA Food and Drug Administration

FLP Frustrated Lewis pair

GAMESS General Atomic and Molecular Electronic Structure System

Ĥ Hamiltonian operator

HMBC Heteronuclear multiple bond correlation

xvi

IR Infra-red

LLCT Ligand-to-ligand charge transfer

MHz Megahertz

MS Mass spectrometry nm nanometer

NMR Nuclear magnetic resonance

Np neopentyl

OTf- trifluoromethanesulfonate (triflate)

PPh3 triphenylphosphine ppm parts per million pz pyrazole

Ψ Psi (wave function) red reductant tBu tertiary-butyl

THF tetrahydrofuran

TMS trimethylsilyl

Tp- tris(pyrazolyl)borate

Tp*- tris(3,5-dimethylpyrazolyl)borate

UV-Vis Ultraviolet-visible

xvii

Contributions

The content of this thesis has depended not only on my own work, but the hard work and dedications of other members of the Fekl lab past and present. The work discussed in Chapter 2 was initiated by Marzieh Daryanavard, a transfer student from Isfahan University of Technology, Iran. She synthesized and performed the initial characterization of the Pd(IV) aryldiazenido complex mentioned therein, as well as isolating out the two compounds discussed in Chapter 3 from the crude reaction mixtures. From there, I optimized and performed the preparatory scale synthesis of the Pd(IV) complexes and performed further thermolysis reactions to identify the decomposition products, as well as elucidate the mechanism via NMR and EPR experiments. I was also responsible for determining the route of formation of the FLP compounds discussed in Chapter 3, as well as computational investigation, preparatory synthesis, characterization, and observation of the use of compound 3.1 as a ligand for zinc.

The synthesis and crystallization of adamantyl bismuth compounds mentioned in Chapter 4 was performed by an undergraduate student Kamalpreet Singh, and the preparatory synthesis, crystallization, characterization, and stability testing of adamantyl gold complexes was carried out by Graduate student Fioralba Taullaj. The first synthesis and crystallization of (COD)Pt(2- Ad)Cl which is discussed in Chapter 4 was simultaneously performed by both Fioralba Taullaj and myself.

All single crystal X-Ray diffraction and structure solution was performed by Dr. Alan J. Lough.

xviii 1

Chapter 1 Introduction

1.1 History of Metal Alkyl Chemistry

The study of metal alkyl chemistry dates back to the mid 19th century with the initial discovery of diethylzinc by Edward Frankland who was, at the time, attempting to isolate free ethyl radicals by reaction of ethyl iodide with zinc metal. He later made note that:1

“On pouring a few drops of water upon the residue, a green-blue flame, several feet long, shot out of the tube, causing great excitement among those present.”

This great excitement has indeed become the attitude toward many of the discoveries made in the field of metal alkyl chemistry. Following the first discovery of diethylzinc it was over 50 years until the discovery of alkylmagnesium compounds by Victor Grignard,2 who was later awarded the 1912 Nobel Prize in chemistry for this work (which happened to be his doctoral thesis). The first report of an organolithium compound came in 1917 when Wilhelm Schlenk performed exchange reactions from dialkylmercury compounds.3 Later work by many other researchers produced alkyl compounds of most of the p-block elements (aluminum, silicon, and tin, to name a few).4 Surprisingly, the first report of a transition metal (the d10 metal Zn2+ being excluded) alkyl compound was in 19095 when William Pope and Stanley Peachey synthesized trimethylplatinum iodide by reaction of “methylmagnesium iodide on platinic chloride.” Following this development, it took decades for the field of transition metal alkyl chemistry to develop any further.6 With the widespread difficulties encountered by those attempting to synthesize transition metal alkyls, a view was developed that metal-carbon bond strengths must be too weak to support the existence of such compounds. We now know that this is untrue and that kinetics, rather than thermodynamics are responsible for the apparent instability of transition metal-carbon bonds. The variety of readily accessible decomposition pathways is responsible for both the general instability of transition metal alkyls, and their endlessly variable reactivity.7

1.2 Why Are We Interested in Metal Alkyls?

Applications of metal alkyl complexes (beyond general interest in bonding and stability) are generally related to the functionalization of organic molecules. This includes polymerization and metathesis catalysis by early transition metal alkyl complexes, as well as coupling, hydrogenation, and other types of functionalizations by late transition metal alkyls. Homoleptic metal alkyls are also often used as precursors in chemical vapour deposition (CVD) systems.

Probably the most ubiquitous use of metal alkyl complexes is in coupling catalysis, through which C-C, C-N, and C-O bonds can be made in a much more facile manner than would otherwise be possible. Advancements in metal mediated coupling chemistry have had staggering impacts on organic synthesis and its applications.8 An important thing to note is that the success of metal alkyl based catalysis is heavily weighted on such metal alkyl intermediates being only metastable. High reaction rates require that no individual step in a catalytic cycle be excessively low in energy (a thermodynamic sink), as such sinks introduce unnecessarily large energetic barriers. For this reason, coupling catalysts are most often based on metals whose alkyl complexes are of moderate stability. Such metals include palladium, nickel, copper, and iron, with palladium being by far the most common. A general scheme of a catalytic coupling cycle is shown in Scheme 1.1.

Scheme 1.1. Overall reaction scheme for metal mediated coupling reactions.

This type of metal mediated coupling catalysis has been so important to the development of organic synthesis that the Nobel prize in chemistry was awarded to some of the founders of the field in 2010.9 Such reactions are often very difficult in the absence of a coupling catalyst, but are also required for fine chemical synthesis. Applications of this reactivity are further discussed in chapter 2.

A highly industrially relevant use of metal alkyl complexes is in olefin polymerization catalysis. An archetypical polymerization catalyst would be a methyl zirconocene complex cation, typically generated in situ by alkyl abstraction with methylaluminoxane (MAO), as shown in Scheme 1.2.

Scheme 1.2. Representative example of Zr(IV) catalyzed olefin polymerization.

The process of olefin polymerization involves subsequent insertion reactions, each of which results in an alkyl complex at one end of the growing polymer chain. Olefin polymerization is responsible for the production of many materials including fibers, plastics, and synthetic rubbers.

Hydrogenation catalysis is another example of the utility of metastable metal alkyls. The hydrogenation of olefins is done on the largest scale by solid state catalysts, but can also be achieved by molecular complexes. Common metals used in solid state hydrogenation catalysts are nickel, palladium, and platinum. An example of a mechanism proposed for this hydrogenation process is given in Figure 1.1.10

Figure 1.1. A simplified representation of olefin hydrogenation at a metal surface.

The process proceeds through several steps including: binding of an olefin to the metal surface, oxidative addition of H2, addition of hydrogen into the olefin (notably forming a surface metal alkyl), and final elimination of the product alkane.

The analogous reaction using a molecular catalyst is mechanistically very similar to this, the most notable catalyst for which being Wilkinson’s catalyst, or Rh(Ph3P)3Cl (Scheme 1.3). The active catalyst is formed by dissociation of a PPh3 ligand, which is then followed by oxidative addition 11 of H2 and binding of the substrate olefin, analogous to the solid state reaction.

Scheme 1.3. Homogeneous olefin hydrogenation by Wilkinson's catalyst.

A less related, but still prevalent application of metal alkyl complexes is in their utility for chemical vapour deposition (CVD). This process can be used to deposit films (typically metal, metal oxide or nitride) on surfaces by decomposition of a volatile metal precursor (thermally or via a co- reactant). This requirement can often be achieved through the use of homoleptic metal alkyl complexes (most often with small alkyl ligands, to improve volatility). CVD applications require that these metal complexes have modest thermal stability, such that decomposition can be achieved in a controlled manner. Common CVD precursor compounds include compounds of the form 12 MMe3 (M = Al, Ga, In), and of the form MCp2Me2 (M = Ti, Zr, Hf) or MCp3 (M = Y, La).

1.3 Decomposition of Metal Alkyls

By nature, many metal alkyl compounds contain a highly ionic metal-carbon bond and as such, are highly sensitive to reaction with even mild acids (the most common one being water). Reaction of metal alkyls with water produces (often violently) metal hydroxides and oxides, and the parent alkanes. Similarly, reaction of many metal alkyls with oxygen produces metal alkoxides or peroxides, which may be quenched with water to form alcohols (shown in Scheme 1.4).

Scheme 1.4. Common decomposition products of metal alkyl complexes on exposure to air or moisture.

The most common decomposition pathway for transition metal alkyl complexes under inert conditions (and conveniently the reason why the greatest number of stable alkyl complexes are those with a methyl ligand) is β-elimination. The mechanism for the most frequently observed β- hydride elimination is shown in Scheme 1.5 below.

Scheme 1.5. Commonly observed mechanism of β-hydride elimination.

Observing this mechanism, it can be seen that several conditions must be met for a metal alkyl to decompose in this manner: 1. The beta position of the alkyl ligand must have a hydrogen substituent (a property that is distinctly absent in metal-methyl complexes). 2. The metal must have a vacant site in proximity to the β-hydride (or at least have a hemilabile group which can create a vacant site).

3. The M-C-C-H unit must be able to come in close proximity and form a planar 4- membered transition state.

A common example of β-hydride elimination can be seen in alkylplatinum(II) complexes. Dissociation of the ligand cis to the alkyl group allows for β-hydride elimination by providing an open site, and generating olefin and a platinum(II) hydride (Scheme 1.6).13

Scheme 1.6. β-hydride elimination in alkylplatinum(II) complexes.

While β-hydride elimination is most commonly observed in organometallic chemistry, other types of elimination, such as β-halide and even β-methyl elimination have been observed.14 In the example below, an analogous system to the one shown above for β-hydride elimination contains a neopentyl ligand rather than an ethyl. The effect of this is that the alkyl group on the coordinatively unsaturated intermediate contains no β-hydrogens, and therefore undergoes β-methyl elimination to form a zirconium methyl complex and releases isobutylene (Scheme 1.7). It should be noted that such reactions are substantially less common than β-hydride elimination.

Scheme 1.7. An example of β-methyl elimination from a Zirconium neopentyl complex.

Common methods of avoiding the decomposition of metal alkyl complexes via β-hydride elimination are to either sterically protect the metal such that there are no available vacant coordination sites to undergo β-hydride elimination, or to “protect” the alkyl ligand by not having any available β-hydrogens. Examples of alkyl ligands which are resistant to β-hydride elimination include neopentyl and trimethylsilylmethyl groups, which contain no β-hydrogen atoms at all. Alternatively, cage type alkyl ligands such as 1-norbornyl or adamantyl contain β-hydrogens but remain resistant to elimination due to the inability of the M-C-C-H unit to co-planarize. In these cases, the resistance to β-hydride elimination is assisted by the thermodynamic unfavourability of forming a highly strained bridge-head alkene, which is described by Bredt’s rule (shown in Scheme 1.8).

Scheme 1.8. Metal 1-norbornyl and adamantyl complexes resist β-hydride elimination.

While these methods of avoiding β-hydride elimination do typically work, they do not always result in stable metal alkyl complexes, as a number of other possible decomposition pathways exist.

The next most common pathway of decomposition is through reductive elimination (shown in Scheme 1.9).

Scheme 1.9. A generic scheme for the reductive elimination from metal alkyls.

Reductive elimination involves the coupling of two ligands and a corresponding two-electron reduction of the metal. This is particularly common because reductive elimination can form a great number of different C-X bonds, and many metal alkyl complexes will contain other ligands which can undergo reductive elimination with the alkyl group.15

Contrary to the early assumptions that metal-carbon bonds were particularly weak, the homolytic cleavage of metal-carbon bonds is relatively rare and typically induced by heat or UV irradiation (Scheme 1.10). This decomposition pathway is, however still present in metal alkyl complexes, particularly those with a readily accessible one-electron redox pair.16 Perhaps the most relevant example of this is in 5’-deoxyadenosylcobalamin (an active form of vitamin B12), which undergoes homolytic cleavage of a Co-C bond.17

Scheme 1.10. General scheme for metal-alkyl homolytic bond cleavage.

1.4 Synthesis of Metal Alkyls

Metal alkyl complexes can be formed in a number of different ways, many of which can be considered as the reverse reaction of some of the common decomposition pathways. The simplest route to metal alkyl complexes is by transmetallation (demonstrated in Scheme 1.11).

Scheme 1.11. Common condition for the synthesis of metal alkyls via transmetallation.

The ease of preparation of both anionic alkyl precursors (alkylzinc, Grignard, alkyllithium reagents) and metal halides makes transmetallation by far the most straightforward method through which to produce transition metal alkyl complexes. The success in these transmetallation reactions

10 is generally dictated by relative metal-carbon and metal-halide (provided a halide is the leaving group) bond strengths between the transition metal of interest and the metal contained in the “anion” precursor, with bond strengths following the general trend of M-Zn > M-Mg > M-Li. Other “anion” precursors based on boron and aluminum are also known but much less commonly used to make stable transition metal alkyls. It is worth noting that while the most thermodynamically favorable transmetallation reactions occur through the use of alkyl lithium reagents, this is not always desirable. Excessive energy release (resulting from a very large ΔG of transmetallation) can promote unwanted side reactions or result in complete decomposition of the product. In these instances, it can often be desirable to use milder transmetallating agents (i.e. those with a stronger M-C bond) such that the ΔG of transmetallation is smaller in magnitude. Good examples of these mild transmetallating agents include alkylzinc compounds (of the general form

R-ZnX) and boronic acids (of the general form R-B(OR)2). The implications of this will be further discussed in chapters 3 and 4.

The second most common route to transition metal alkyls (oxidative addition, Scheme 1.12) happens to be the reverse reaction of one of the common routes to decomposition (reductive elimination).

Scheme 1.12. General scheme for an oxidative addition reaction.

Oxidative addition involves the cleavage of a (usually polar) sigma bond, coupled with 2-electron oxidation of a metal. The process generally increases the d-electron count of the metal by 2, as well as increasing the coordination number of the metal by 2. A typical example is the oxidative addition of alkyl bromides to palladium(0), shown in Scheme 1.13.18

Scheme 1.13. Oxidative addition of alkyl bromides to a Pd(0) phosphine complex.

This reaction is a reliable method for the synthesis of metal alkyl complexes in the absence of an “anionic” alkyl precursor but rather an alkyl cation equivalent. Oxidative addition can occur via multiple mechanisms, depending on the nature of the substrate. Radical mechanisms can be observed from combinations of metal and substrate for which one-electron pathways are readily available. The concerted mechanism (shown below in Scheme 1.14) is commonly observed in the oxidative addition of aryl halides and C-H bonds and relies on the initial formation of a side-on sigma coordination complex, followed by concerted oxidation and bond cleavage.19

Scheme 1.14. Example of the concerted oxidative addition mechanism.

The second common mechanism of oxidative addition is the nucleophilic or “SN2-like” mechanism (Scheme 1.15), which occurs most frequently for highly polar substrates.20

Scheme 1.15. A common example of the “nucleophilic” oxidative addition mechanism.

It is important to note that the concerted mechanism produces a complex in which the components of the “R-X” fragment are coordinated cis to each other whereas the nucleophilic mechanism does not necessarily do so. A specialized method for producing metal alkyls by oxidative addition is cyclometallation, in which the oxidative addition substrate is coordinated to the metal by a pendant group which results in the formation of a metalacycle following oxidative addition. Cyclometallation can also occur via a sigma bond metathesis mechanism (both shown in Scheme 1.16 below).21

Scheme 1.16. Examples of cyclometallation of metal neopentyl complexes.

The final method through which metal alkyl complexes are produced is through insertion, the most common type of which being olefin insertion. Similarly to oxidative addition, olefin insertion is the reverse reaction of β-elimination (see Scheme 1.17). Insertion requires initial coordination of an olefin to a metal bearing a nucleophilic ligand (such as a hydride).

Scheme 1.17. Ethylene insertion into a molybdenum hydride, forming a metal alkyl.

The use of olefin insertion to form stable metal alkyl complexes is fairly uncommon, as the β- eliminated species is often lower in energy.

1.5 C-H Bond Activation and Functionalization

Among the most unique and interesting reactions that can be performed by organometallic complexes is carbon-hydrogen (C-H) bond activation. C-H bonds are extremely abundant in organic molecules, making them a desirable target for functionalization. However, due to their low bond polarity and high dissociation energies, the activation of C-H bonds is particularly challenging. The most common mechanisms of organometallic C-H bond activation are the same as those discussed in the formation of metal alkyl complexes, namely, oxidative addition and sigma bond metathesis.

There are a huge number of current and potential applications for C-H bond activation chemistry, most of which are related to the functionalization of inert molecules. This is important for fine chemical synthesis, but the single greatest potential application for this type of chemistry is in the efficient utilization of volatile hydrocarbons, particularly methane. Methane is a principal component of natural gas, but it can be very difficult to handle and transport due to its high vapour pressure and low flash point. Liquefying natural gas is not a trivial process and requires specialized facilities. Because of this, massive amounts of methane contained in oil and gas wells are simply disposed of by “flaring” (burning), as methane is an even more potent greenhouse gas than the 22 CO2 which is produced by burning it. Small scale chemical processing through which methane could be converted into liquid products such as methanol are highly desirable as they could be performed on-site. The least energy intensive options for this process would involve direct C-H bond activation of methane and this has become one of the primary goals of the field.

The earliest report of an organometallic C-H bond functionalization was in 1955 when Shunsuke Murahashi achieved the cobalt-catalyzed synthesis of N-phenyl phthalimide from benzaldehyde anil via oxidative insertion of CO (Scheme 1.18).23

Scheme 1.18. The first report of organometallic C-H bond activation.

The first example of a C-H bond activation to form an isolable metal alkyl complex was demonstrated in 1965 by Joseph Chatt with the oxidative addition of aromatic C-H bonds by 24 reduced Ru(DMPE)Cl2 to form Ru(DMPE)H(2-naphthyl). This work was closely followed in

1969 by the seminal work of Alexander Shilov who demonstrated that K2PtCl4 could induce isotope scrambling of methane in D2O. Shilov later demonstrated that K2PtCl4 could catalytically convert methane into methanol and methyl chloride in water by using sacrificial K2PtCl6 as the oxidant (Scheme 1.19).25

Scheme 1.19. The famous Shilov cycle for catalytic C-H bond functionalization of methane.

A similar system was later developed by Roy Periana, through which methane could be converted to methyl sulfate via oxidation by SO3, catalyzed by either mercury or platinum (see Scheme 1.20).26

Scheme 1.20. The Periana cycle for conversion of methane to methyl sulfate.

Many modifications to these core systems have been made, providing small iterative improvements, but all of them suffer from poor turnover numbers.27 One of the biggest issues with the catalytic conversion of methane (or other alkanes) to alcohols by direct C-H bond activation is that the C-H bonds of the product alcohol are weaker and typically easier to activated than the parent alkane, making over-oxidation problematic. While alcohols would be considered a value- added product from small alkanes, the over-oxidized products such as CO, CO2, and formic acid are not desirable.

Many organometallic systems for C-H bond functionalization have since been developed, but doing so selectively remains the greatest challenge.28 An important part in the development of these systems is gaining an understanding of their mechanisms. Unfortunately, kinetic and mechanistic investigations of these system can be difficult due to the conditions employed (high temperature, strong acids, strong oxidants, etc.) which complicate the use of common analytical techniques (Mass spec, NMR, etc.). In these situations, computational chemistry can become highly advantageous.

1.6 Computational Chemistry

The field of computational chemistry sits at the frontiers of chemistry, quantum mechanics, and computer science. It can essentially be broken down into the branches of quantum mechanics and molecular mechanics, depending on the desired purpose of the calculations. Quantum mechanical calculations are based fundamentally on Schrödinger’s equation (ĤΨ=EΨ) and can predict the electronic and chemical properties of molecules to a high degree of accuracy, but can be very time consuming. Ab initio calculations in particular, which don’t contain any element-specific empirical parameters, are exceptionally slow. Molecular mechanics is based on empirically derived information and can provide some information about very large molecules, even proteins (albeit to a lower degree of accuracy) in relatively short periods of time.29 The utility of these methods is highly appreciated in research environments, with the 2013 Nobel Prize in Chemistry being awarded to some of the founders of computational chemistry.30

Many experimental parameters can be predicted by computational chemistry methods, some of which include equilibrium geometries, bond energies, dipole moments, thermodynamic and kinetic reaction parameters, and optical and electronic spectra. In principle, all experimental parameters are accessible by quantum mechanics calculations. The degree of accuracy with which the predictions can be made varies greatly depending on the methods used. The fundamental aspects that makes computational methods useful are “approximations.” Approximations are an inherent necessity for a few different reasons. The first (and most fundamental) of these reasons is that Schrödinger’s equation can only be analytically solved for a system containing up to one nucleus and one electron, where the nucleus is treated as a potential field, rather than a particle (this is commonly known as the many-body problem). Systems which contain only a single electron are obviously not relevant for solving real-world chemical problems.

In order to simplify calculations of molecules, the most commonly employed approximation is the Born-Oppenheimer approximation. This proposes that a wave function Ψ can be broken down into two components: Ψnuclear and Ψelectronic, allowing the deconvolution of the spatial variables of the nuclei and electrons in a system, effectively reducing one large calculation into two smaller ones. An example would be the calculation of a water molecule, which contains 3 nuclei and 10 electrons. Since each of these particles possesses three spatial coordinates, a full calculation of this

17 system would require 39 variables. By implementing the Born-Oppenheimer approximation, the computation can be performed in two different steps. The first step is solving the electronic wave function (now requiring only 30 variables) relative to the nuclei in fixed positions, a reasonable assumption given that nuclear motion is orders of magnitude slower than the motion of electrons. The second part of the calculation would solve a Schrödinger equation for the nuclear terms, now requiring a mere 9 variables, and using the previous electronic solution as a potential in this calculation. Further, the nuclear wave function is often not even needed, and sufficient chemical insight can often be obtained by using a classical treatment of the nuclei. A second related simplification can be made by separating the total energy of a system into its principal components, namely: kinetic energies of electrons and nuclei, electron-electron repulsion, nuclear-nuclear repulsion, and electron-nuclear attraction.

Because the electronic Hamiltonian operator Ĥ contained in the electronic Schrödinger equation

Ĥelecφelec=Eφelec depends only on spatial coordinates of the electrons, an additional coordinate needs to be included to account for electron spin in order to accurately describe the electrons in a system. The way that this is most often done is by introducing Slater determinants, which use spin orbitals (orbitals with both spatial and spin coordinates) to describe each of the single-electron wave functions for a chemical system. The Slater determinant is essentially a product but takes into account the Pauli exclusion principle.

The most commonly used ab initio method is known as the Hartree-Fock method. This type of calculation uses an approximate Hamiltonian operator such that individual electrons are considered to be non-interacting. The effect of this is that the energy of the electronic system is approximated by the kinetic and nuclear attraction energy of the electrons, and the averaged interelectron repulsions, where each electron is repelled by the average coulombic field produced by the remaining electrons in the system. The Hartree-Fock approximation assumes that the wave function for a system can be fully approximated by using a single Slater determinant. This approximation permits calculations to be performed relatively quickly, but it is not by any means an exact solution to the Schrödinger equation. In order to improve upon the accuracy of these results, so-called post-Hartree Fock methods have been developed. These include, but are not limited to, configuration-interaction, coupled cluster, and Møller-Plesset perturbation theory. Each

18 of these methods performs iterative corrections to the Hartree-Fock solution which improves the accuracy of the results, but correspondingly increases computational expense.

A computational method which was later developed and is fundamentally different from Hartree- Fock and post-Hartree-Fock methods is known as Density Functional Theory (DFT).31 This method uses the electron density to predict the physical properties of chemical systems. This is achieved through the use of functionals – a function of another function – where the function in this case is the spatially dependent electron density. The benefit of this method is that electron density, unlike the position or momentum of individual electrons, is a quantum mechanical observable. Electron density depends only on the three spatial coordinates, whereas the wave function of an N-electron system is dependent on 3N variables (plus an additional binary spin variable for each electron). The complication of the DFT method is that there is no well-defined method for constructing density functionals, so a number of approximations need to be made. The development of this method required a great degree of development and refinement, but has now become one of the most reliable, accurate, and computationally inexpensive methods for predicting the physical properties of chemical systems.32 In this work, density functional theory was exclusively used.

1.7 Organization of this Thesis

This thesis focuses broadly on new developments to the field of transition metal alkyl chemistry. A number of different avenues were explored, including palladium-based coupling reactions, development of methods for the synthesis of anion equivalents of the adamantyl fragment, and their applications toward the synthesis of new transition metal adamantyl complexes. The reactivity and further applications of these complexes is also discussed.

In Chapter 2, The First Palladium(IV) Aryldiazenido Complex – Relevance for C-C Coupling, the synthesis and reactivity of the first reported palladium(IV) aryldiazenido complex is explored. The synthesis was achieved by two-electron oxidation of the anionic palladium(II) complex

KPdTp*Me2 by an aryldiazonium cation. The resulting Pd(IV) complex could be cleanly isolated in good yields and characterized crystallographically, as well as by NMR. The complex is stable

19 for weeks as a solid when kept cold and decays in solution over a period of days at room temperature. The decomposition was studied in great depth by both NMR and EPR and it was determined that radical pathways were involved.33 The future relevance of this chemistry toward a Pd(II)/Pd(IV) based catalytic coupling cycle is discussed.

In Chapter 3, Facile Transformations of hydridotris(3,5-dimethylpyrazolyl)borate – An Intramolecular Frustrated Lewis Pair, the reactivity of a new FLP compound generated through a hydride abstraction from the hydridotris(3,5-dimethylpyrazolyl)borate anion to generate the latent hydridotris(3,5-dimethylpyrazolyl)borane, is explored. The hydride abstraction is achieved using 1 equivalent of aryldiazonium cation, and the intermediate borane species is quickly trapped by small molecules such as acetone. A second equivalent of aryldiazonium cation further abstracts a 3,5-dimethylpyrazolide group to form a cationic intermediate which again is trapped by small molecules.34 Initial reactions were focused on the trapping of acetone, with later work focusing on the activation of other small molecules such as CO2, CS2, MeCN, and benzophenone.

In Chapter 4, Synthesis and Characterization of 1- and 2-Adamantyl Anions, extensive synthetic work towards the synthesis of adamantyl anion equivalents was explored. It has been established that common synthetic methods for generating adamantyl anions such as Grignard or lithium- halogen exchange reactions are highly unreliable. The development of new reliable synthetic methods to produce adamantyl magnesium and zinc compounds (including the unknown diadamantylzincs) was achieved, producing clean and shelf stable compounds. These compounds were extensively characterized by NMR, EA, MS, and IR.35

In Chapter 5, Synthesis and Reactivity of Transition Metal Adamantyl Complexes, the applications of the adamantyl anions described in chapter 4 toward organometallic chemistry was investigated. It was observed that the diadamantylzinc compounds were particularly potent in transmetallation reactions, cleanly producing several new transition metal and main group adamantyl complexes. While compounds containing mercury, gold, bismuth, tungsten, and platinum have now been synthesized, those of particular interest are capable of C-H bond activation chemistry and hold the potential to open access to new compound classes of facially trifunctionalized adamantanes.

1.8 References for Chapter 1

1 (a) Seyferth, D. Organometallics 2001, 20, 2940. (b) Frankland, E. Experimental Researches in Pure, Applied and Physical Chemistry; John van Voorst: London, 1877.

2 Grignard, V. Compt. Rend. 1900, 130, 1322.

3 Schlenk, W.; Holtz, J. Ber. Dtsch. Chem. Ges. 1917, 50, 262.

4 (a) Hallwachs, W.; Schafarik, A. Liebigs Ann. Chem. 1859, 109, 206. (b) Ziegler, K. Angew. Chem. 1956, 68, 721. (c) Friedel, C.; Crafts, J. M. Ann. 1863, 127, 28. (d) Frankland, E. Ann. 1849, 71, 213. (e) Frankland, E. J. Chem. Soc. 1850, 2, 297.

5 Pope, W. J.; Peachey, S. J. J. Chem. Soc., Transactions. 1909, 95, 571.

6 Crabtree, R. H. The Organometallic Chemistry of the Transition Elements, 6th ed. Wiley: Hoboken, New Jersey, 2014.

7 Cotton, F. A. Chem. Rev. 1955, 55, 551.

8 Goodson, F. E.; Cichowicz, M. B. Organic Synthesis Using Metal-mediated Coupling Reactions: Encyclopedia of Inorganic and Bioinorganic Chemistry, Wiley, 2011.

9 https://www.nobelprize.org/nobel_prizes/chemistry/laureates/2010/

10 Zaera, F. ACS Catal. 2017, 7, 4947.

11 (a) Osborn, J. A.; Wilkinson, G. Inorg. Synth. 1967, 10, 67. (b) Meakin, P.; Jesson, J. P.; Tolman, C. A. J. Am. Chem. Soc. 1972, 94, 3240. (c) Duckett, S. B.; Newell, C. L.; Eisenberg, R. J. Am. Chem. Soc. 1994, 116, 10548. (d) Halpern, J. Inorg. Chim. Acta. 1981, 50, 11.

12 Malandrino, G.; ed. Jones, A. C.; Hitchman, M. L. Chemical Vapour Deposition: Precursors, Processes and Applications. RSC Publishing, Cambridge, 2008.

13 Alibrandi, G.; Scolaro, L. M.; Minniti, D.; Romeo, R. Inorg. Chem. 1990, 29, 3467.

14 Yang, P.; Baird, M. C. Organometallics 2005, 24, 6005.

15 (a) Goldberg, K. I.; Yan, J. Y.; Winter, E. L. J. Am. Chem. Soc. 1994, 116, 1573. (b) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of Organotransition Metal Chemistry; University Science Books: Mill Valley, CA, 1987.

16 Halpern, J. Bonding Energetics in Organometallic Compounds. ACS Symposium Series, 1990, Chapter 7, 100.

17 Kräutler, B. Met. Ions Life Sci. 2009, 6, 1.

18 Stille, J. K.; Lau, K. S. Y. J. Am. Chem. Soc. 1976, 98, 5841.

19 Crabtree, R. H. The Organometallic Chemistry of the Transition Elements, 6th ed. Wiley: Hoboken, New Jersey, 2014.

20 Monaghan, P. K.; Puddephatt, R. J. J. Chem. Soc. Dalton. Trans. 1988, 595.

21 (a) DiCosimo, R.; Moore, S. S.; Sowinski, A. F.; Whitesides, G. M. J. Am. Chem. Soc. 1982, 104, 124. (b) Bruno, J. W.; Smith, G. M.; Marks, T. J.; Fair, C. K.; Schultz, A. J.; Williams, J. M. J. Am. Chem. Soc. 1986, 108, 40.

22 https://www.epa.gov/ghgemissions/overview-greenhouse-gases

23 Murahashi, S. J. Am. Chem. Soc. 1955, 77, 6403.

24 Chatt, J.; Davidson, J. M. J. Chem. Soc. 1965, 154, 843.

25 Shilov, A. E.; Shul’pin, G. B. Chem. Rev. 1997, 97, 2879.

26 (a) Periana, R.A.; Taube, D.J.; Evitt, E.R.; Loffler, D.G.; Wentrcek, P.R.; Voss, G.; Masuda, T. Science. 1993, 259, 340. (b) Periana, R. A.; Taube, D. J.; Gamble, S.; Taube, H.; Satoh, T.; Fujii, H. Science. 1998, 280, 560.

27 Labinger, J. A. Chem. Rev. 2017, 117, 8483.

28 Goldman, A. S.; Goldberg, K. I. ACS Symposium Series 885, Activation and Functionalization of C-H Bonds, 2004, 1.

29 Ramachandran, K. I.; Deepa, G.; Namboori, K. Computational Chemistry and Molecular Modeling: Principles and Applications. Springer: Berlin, 2008.

30 https://www.nobelprize.org/nobel_prizes/chemistry/laureates/1998/

31 Kohn, W. Electronic Structure of Matter - Wave Functions and Density Functionals. Nobel Lecture, 1999.

32 Szabo, A.; Ostlund, N. S. Modern Quantum Chemistry: Introduction to Advanced Electronic Structure Theory. McGraw-Hill: New York, 1989.

33 Armstrong, D.; Daryanavard, M.; Lough, A. J.; Fekl, U. Dalton Trans. 2017, 46, 4004.

34 Daryanavard, M.; Armstrong, D.; Lough, A. J.; Fekl, U. Eur. J. Inorg. Chem. 2015, 20, 3232.

35 Armstrong, D.; Taullaj, F.; Singh, K.; Mirabi, B.; Lough, A. J.; Fekl, U. Dalton Trans. 2017, 46, 6212.

Chapter 2 The First Palladium(IV) Aryldiazenido Complex – Relevance for C-C Coupling

2.1 Abstract

Overall two-electron oxidative addition of the aryldiazonium cation para- + II − − methoxybenzenediazonium (pmbd ) to [Pd (Tp*)Me2] (Tp* = hydridotris-(3,5- dimethylpyrazolyl)borate) produces the first characterized palladium(IV) aryldiazenido complex, IV − [(Tp*)Pd Me2(pmbd)] (pmbd = para-methoxybenzenediazenido). Thermolysis in benzene forms IV a mixture of products that contains Pd (Tp*)Me3 and the C–C coupled product, 4,4’- dimethoxybiphenyl. The use of acetone as the thermolysis solvent changes the product distribution towards the formation of anisole. While some involvement of two-electron pathways in the decomposition cannot be ruled out, radicals, through one-electron steps, are involved in the major pathways for decomposition. The involvement of radicals is confirmed by the solvent dependent nature of the product distribution and by observing that the added N-tert-butyl-α-phenylnitrone efficiently traps aryl radicals.

2.2 Introduction

Palladium catalysts have become a staple of the modern synthetic laboratory. The variety of syntheses which have been made possible by palladium mediated reactions is staggering, and it is no surprise that a Nobel prize in chemistry was awarded to some of the originators of the field of palladium catalyzed cross coupling reactions.36 While the variety in palladium coupling chemistry is huge, the common ground among these reactions is the need for an equally massive variety of anion and cation equivalents of alkyl- and aryl- groups. A stereotypical palladium catalyzed cross- coupling cycle is shown in Scheme 2.1 below.

Scheme 2.1. General palladium catalyzed cross coupling cycle.

While there is a great degree of variability in the above depiction of cross-coupling catalysis, it is generally true that a formal carbocation equivalent is coupled to a formal carbanion (or heteroanion) equivalent. By far the most common carbocation equivalents are those with an electronegative leaving group (halides, triflates, etc., see Figure 2.1).

Figure 2.1. Common aryl cation equivalents (halides, triflates, tosylates, acetates).

The oxidative addition of such compounds typically proceeds through one of two pathways (Scheme 2.2), and this type of reactivity is quite well understood. Radical pathways have also been suggested for some substrates but are less common.37

Scheme 2.2. The two most common mechanisms of oxidative addition.

In contrast to the more common aryl halides, aryldiazonium ions present an intriguing alternative source of aryl cation equivalents. While aryldiazonium ions have been used as aryl cation equivalents since 1890,38 it has only been since the 1980's that a mechanistic understanding of their organic reactivity has been appreciated, and their reactivity with metal complexes is still not completely understood. The binding of aryldiazonium ligands to metals is highly diverse, as shown in Figure 2.2. In the absence of metals, aryldiazonium ions decay very slowly to form aryl cation fragments and as such, this type of reactivity has limited utility.39 Reactions involving aryl radical intermediates are often observed in the presence of one-electron reductants such as Cu(I).40 From a mechanistic standpoint, relatively little is understood about the reactions of aryldiazonium ions with transition metals that undergo two-electron processes.

Figure 2.2. Multiple potential binding modes of aryldiazonium cations to a metal center.

Pd-catalyzed reactions using aryldiazonium cations are known and include Suzuki and Heck-type coupling reactions as well as carbonylations.41 It is however, important to note that the vast majority of cross-coupling is catalyzed by a Pd(0)/Pd(II) redox pair, and the alternative Pd(II)/Pd(IV) pair is relatively unexplored.42 For coupling reactions involving other aryl cation equivalents, the involvement of Pd(II)/Pd(IV) has shown some potential, but due to the strongly oxidizing nature of aryldiazonium cations, we hypothesized that they could prove advantageous for use in Pd(II)/Pd(IV) coupling catalysis (Scheme 2.3).

Scheme 2.3. Suggested catalytic coupling cycle using aryldiazonium cations.

Oxidative addition of an aryldiazonium cation to a Pd(II) complex would simultaneously result in the oxidation of Pd(II) to Pd(IV) and the reduction of aryldiazonium to a metal-bound aryldiazenido ligand. This step could prove to be important for future Pd(II)/Pd(IV) coupling catalysis involving aryldiazonium ions as coupling reagents. It is worth noting that the oxidative addition of an aryldiazonium cation to Pt(II) was first demonstrated by our group (Scheme 2.4).43 It was yet unknown whether aryldiazonium ions were sufficiently oxidizing to undergo oxidative addition at Pd(II).

+ Scheme 2.4. Oxidative addition of ArN2 to Pt(II).

While the existing platinum aryldiazenido complex suggested the feasibility of a similar reaction at palladium, it is important to note that oxidative addition of an aryldiazonium equivalent to Pt(II) should inherently be more favourable than the analogous Pd(II) system, as Pt(IV) is substantially less oxidizing than Pd(IV), as shown in Table 2.1.44

Table 2.1. Comparison of the redox behaviour of palladium and platinum. Redox Pair Reduction Potential 2- - - [PtCl ] + 2 e ⇌ Pt + 4 Cl + 0.755 V 4 2- - 2- - [PtCl ] + 2 e ⇌ [PtCl ] + 2 Cl + 0.680 V 6 4 2- - - [PdCl ] + 2 e ⇌ Pd + 4 Cl + 0.591 V 4 2- - 2- - [PdCl ] + 2 e ⇌ [PdCl ] + 2 Cl + 1.288 V 6 4

2.3 Results and Discussion

Since it is significantly more difficult to oxidize Pd(II) to Pd(IV) than Pt(II) to Pt(IV), it was necessary to create a model system which would strongly favor the stabilization of a Pd(IV) complex. For the purposes of investigating coupling reactivity, a Pd(II) alkyl precursor is ideal, and the use of a tripodal nitrogen donor ligand would best stabilize the octahedral geometry of a Pd(IV) complex.45 We selected hydrotris(3,5-dimethylpyrazolyl)borate (Tp*-) as the ligand for our

29 model system for its strong preference for octahedral geometries as well as its negative charge to - charge-balance a high valent Pd(IV) complex. The known precursor [Pd(Tp*)Me2] was 46 synthesized using a heavily modified procedure starting from Pd(COD)Cl2. The full synthesis is shown in Scheme 2.5.

Scheme 2.5. Synthesis of Pd(Tp*)Me2(pMBD) (2.1).

A new method for the synthesis of (COD)PdMe2 was developed for the purposes of avoiding the traditional work-up which required the use of sodium cyanide to quench the dimethylcuprate reagent. It was found that the addition of 2.5 equivalents of MeMgBr to Pd(COD)Cl2 in dimethoxyethane solvent at -50 °C, followed by filtration through celite resulted in good yields of - Pd(COD)Me2 without the need for aqueous work-up or cyanide quenching. [(Tp*)PdMe2] could then be generated in situ by mixing Pd(COD)Me2 in a 1:1 ratio with KTp* in acetone, followed by the addition of para-methoxybenzenediazonium tetrafluoroborate (pmbd-BF4) which resulted in a rapid colour change from colourless to dark orange. The Pd(IV) aryldiazenido complex

Pd(Tp*)Me2(pmbd) (compound 2.1) could be isolated using this method in up to 94% yield.

Crystals suitable for single crystal X-ray diffraction were grown from a saturated Et2O solution at -35 °C. The structure is given in Figure 2.3 below.

Figure 2.3. Anisotropic displacement plot (20% probability ellipsoids) of 2.1. Selected bond distances (Å) and angles (deg): Pd1−C23, 2.036(2); Pd1−C24, 2.042(3); Pd1−N1, 2.212(2); Pd1−N3, 2.212(2); Pd1−N5, 2.174(2); Pd1−N7, 2.009(2); N7−N8, 1.212(3); N8−C16, 1.456(4); C23−Pd1−C24, 87.98(14); N7−Pd1−C23, 91.26(11); N7−Pd1−C24, 88.87(12); N5−Pd1−N7, 175.20(10); N7−N8−C16, 115.5(3); Pd1−N7−N8, 119.8(2).

The structure of 2.1 (anisotropic displacement plot in Fig. 1) shows that the Pd(IV) ion is surrounded by two Me ligands in a cis geometry, a κ3-Tp*− facially coordinating the nitrogen ligand, and a para-methoxybenzenediazenido (pmbd−) ligand bound in doubly-bent mode. A

31 previously reported platinum(IV) aryldiazenido complex also exhibited a doubly-bent mode for the aryldiazenido ligand, which was shown by 15N-NMR spectroscopy, in the absence of a crystal structure.47 The crystal structure of 2.1, displaying an only slightly distorted octahedral geometry for Pd and a clearly doubly-bent geometry for pmbd, impressively demonstrates that Pd(II) has been oxidized to Pd(IV) and the previously reported linear pmbd+ has been reduced to pmbd−. The Pd–Me bond lengths (2.036(3) and 2.042(3) Å) fall in the range of Pd–Me bond lengths usually observed for similar Pd(IV) complexes supported by hydridotris(pyrazolyl)borate ligands.48 The Pd–N bond lengths related to the tripodal ligand are within the range of common Pd–N bond lengths49 of 2.10–2.25 Å for the bonds related to pyrazolyl nitrogen atoms but the Pd–N(7) bond, for the nitrogen belonging to the pmbd− ligand, is very short, at 2.009(2) Å. The N7–N8–C16 and Pd1–N7–N8 angles are 115.5(3) and 119.8(2)°, respectively, for the aryldiazenido ligand in a doubly-bent coordination mode. Complex 2.1 was also characterized by elemental analysis and 1H and 13C{1H} NMR spectroscopy (see experimental section). The 1H and 13C{1H} NMR spectra are in accord with the occurrence of two pyrazole ring environments in a 2:1 ratio for 2.1. The two methyl ligands of 2.1 give rise to a singlet at 2.37 ppm in the 1H NMR spectrum, and at 26.74 ppm in the 13C{H} NMR spectrum.

The aromatic hydrogens in the p-MeO–C6H4N=N fragment are straightforwardly assigned to two signals (apparent doublets but actually AA′BB′ spin systems), at 6.63 and 7.56 ppm for 2.1.

15 1 Figure 2.4. N– H gHMBC NMR spectrum of Pd(Tp*)Me2(pmbd) (2.1) in C6D6 at 25 °C.

The spectroscopic data for the Tp*− ligand are consistent with those for related complexes with − − 50 15 Tp and Tp* ligands. The doubly bent nature of the Ar–N2 unit is confirmed using N NMR spectroscopy in solution. Due to the low natural abundance and low sensitivity of 15N, 15N–1H gHMBC (see Figure 2.4) was used for 2.1 to indirectly detect 15N signals for those nitrogen atoms that show significant coupling to 1H. The C-bound nitrogen of the aryldiazenido unit (N8 in Figure. 2.4) is observed. The two aryl hydrogens ortho to N (in the aryldiazenido fragment) couple sufficiently strongly with 15N to allow for the detection of a nitrogen NMR signal for this C-bound nitrogen (Nα). Since 15N is detected through coupling with 1H, the Pd-bound nitrogen (Nβ) is

33 invisible with this technique. The resonance of Nα is observed at 117.5 ppm, which is consistent with the monoanionic (reduced) aryldiazenido in doubly-bent mode. It is known that Nα of doubly- bent aryldiazenido units occurs within the 40 to 165 ppm range, with most examples between +130 and +165 ppm. In contrast, Nα of almost linear or singly-bent aryldiazenido units occurs in a separate region, in the −250 to 5 ppm range, where most values range from −250 to −190 ppm.51 All NMR data indicate that the solution structure of complex 2.1, including the bent nature of the aryldiazenido, is identical to the structure determined by X-ray diffraction. The thermal reactivity of 2.1 was the next aspect to be investigated. The thermolysis of 2.1 (for example at 70 °C in C6D6 for 1 h, or at room temperature for 48 h) leads to the formation of a very complex reaction mixture in solution (concomitant with the formation of a dark insoluble precipitate) that contains one organometallic product and one major organic product, as determined by 1H NMR spectroscopy. The organic product was identified as 4,4′-dimethoxybiphenyl, formed in 36% yield (0.18 mol of the biaryl product per mol of pmbd in the Pd complex 2.1). The identity was confirmed by 1H NMR against a genuine reference sample. The 4,4′-dimethoxybiphenyl was also isolated from the reaction mixture and characterized crystallographically, matching the unit cell parameters reported 52 in the literature. The organometallic product was identified as Pd(Tp*)Me3, compound 2.2. At the end of the decomposition reaction, 34% of 2.1 had been converted into 2.2, with the balance of the palladium complex “missing” almost certainly residing in the insoluble precipitate. Compound 2.2 is easily identifiable by NMR spectroscopy due to its high symmetry (NMR in experimental section), and the chemical shifts are very similar to those observed for the 53 isostructural platinum complex Pt(Tp*)Me3. Compound 2.2 was also isolated by recrystallization from Et2O solution at −35 °C, and characterized by elemental analysis and X-ray crystallography.

Figure 2.5. Anisotropic displacement plot (20% probability ellipsoids) for 2.2. Selected bond distances (Å) and angles (deg): Pd1−C16, 2.0439(16); Pd1−C17, 2.0397(15); Pd1−C18, 2.0399(15); Pd1−N1, 2.2304(13); Pd1−N3, 2.2122(12); Pd1−N5, 2.1925(12); C16−Pd1−C17, 87.65(7); C16−Pd1−C18, 87.96(7); C17−Pd1−C18, 87.49(7).

The structure of 2.2 (Figure 2.5) is octahedral around the Pd(IV) center, with three Me groups and 3 a κ -Tp*−N-bound ligand. The Pd–Me bond distances are ca. 2.04 Å, as in Pd(Tp)Me3 and other IV 54 Pd Me3 complexes. The three Me–Pd–Me angles are all close to 90° (all within the 87.49(7)–

87.96(7)° range) confirming a symmetric pseudo-C3v geometry similar to Pd(Tp)Me3, and +,55 [Pd(Me3tacn)Me3] (Me3tacn = N,N′,N″-trimethyl trimethyl- 1,4,7-triazacyclononane).

Complex 2.2 is stable for weeks in C6D6 and acetone-d6 solutions at room temperature. The ligand redistribution to form 2.2 from 2.1 is very intriguing. It is well known that methyl transfer can easily occur from d6 metal complexes to d8 metal complexes, an overall 2-electron process (transfer

35 of a methyl cation equivalent from the more oxidized metal to the more reduced metal).56 It is thus possible that decomposition occurs through a mechanism involving two-electron steps, including the eventual formation of 4,4′-dimethoxybiphenyl by reductive elimination from a Pd(IV) diaryl or Pd(II) diaryl, triggered by the initial formation of a d8 Pd(II) complex. Since a trace of ethane is only observed at later stages of the decomposition, it can be ruled out that the d8 complex is generated by initial reductive elimination of ethane from 2.1. If the decomposition of 2.1 should occur through 2-electron steps, it would likely involve reversibility in the formation of the oxidative addition product 2.1, i.e., in equilibrium, some reductive elimination of aryldiazonium 8 − should occur from 2.1 to generate the d complex [Pd(Tp*)Me2] . Such a mechanistic scenario is summarized in Scheme 2.6.

Scheme 2.6. A hypothetical 2-electron mechanism for the decomposition of 2.1 to form 2.2 and organic products.

Since the decomposition would be triggered by the formation of ionic intermediates from neutral ones, such a mechanism, if operating, would imply that the reaction should occur much faster in more polar solvents. We decided to switch from C6D6 to acetone-d6. Compound 2.1 decomposed at about the same rate (reaction being complete within 1 h at 70 °C, or 48 h at room temperature), which argues against the mechanism shown in Scheme 2.5. Furthermore, the product distribution changes: while 2.2 is still being formed, the yield of 4,4′-dimethoxybiphenyl decreases from 36% to 5%, and anisole-d1 forms as a new major product (49%), along with 46% unidentified aromatics. Since acetone has relatively weak C–H bonds, it can act as a radical trap, and the formation of anisole-d1 upon thermolysis in acetone-d6 may point to a radical mechanism. It is thus suggested that the decomposition of 2.1 leads to the formation of the 4-methoxyphenyl radical, which undergoes further decomposition in a solvent-dependent manner (Scheme 2.7).

Scheme 2.7. The preferred radical mechanism for the decomposition of 2.1 to form 2.2 and organic products.

Palladium(III) species have been previously identified as intermediates in oxidations of some Pd(II) compounds.57 Examples of reductive C–C bond formation, as well as disproportionation by methyl ligand transfer have been previously reported for Pd(III) methyl complexes.58 The III formation of a [Pd (Tp*)Me2] intermediate via the elimination of an aryldiazene radical from 2.1 is a plausible pathway leading to the formation of both 2.2 and 4,4′-dimethoxybiphenyl. The overall decomposition behavior of 2.1 did not change or speed up when the thermolysis in acetone − was performed in the presence of one equivalent of added [Pd(Tp*)Me2] (potassium salt), which further argues against an ionic mechanism and indirectly supports the radical mechanism. Finally, IV the decomposition of a pure sample of Pd (Tp*)Me2(pmbd) in the presence of a N-tert-butyl-α- phenylnitrone (PBN) spin trap was performed. An EPR spectrum was obtained that matched the literature EPR data59 of the PBN trapped 4-methoxyphenyl radical (Figure 2.6).

Figure 2.6. X-band EPR spectrum of the thermolysis mixture generated from 2.1 in the presence of a PBN radical trap. The signal obtained yields g = 2.0063, A = 14.4 G, 2.3 G, indicating the trapped 4-methoxyphenyl radical.

It is known that aryldiazene radicals lose N2 very rapidly to form aryl radicals, and biaryl formation is common if aryldiazene radicals are inefficiently trapped.60 Thus, the combined evidence, solvent dependence and EPR of a trapped radical, strongly suggests that a mechanism involving the formation of aryldiazene radicals operates in the thermal decomposition of 2.1, as shown in Scheme 2.6.

2.4 Conclusion

In summary, the first aryldiazenido complex of palladium(IV) was synthesized and its thermal reactivity was investigated. The finding demonstrates that it is possible for an organopalladium(II) complex to be oxidized by an aryldiazonium ion to the Pd(IV) oxidation state. Applying this knowledge to catalytic systems that involve palladium and hard ligands that stabilize Pd(IV), it seems now quite possible that aryldiazenido complexes of Pd(IV) could be involved in some catalytic systems. What types of reactivity are available to aryldiazenido complexes of Pd(IV)? Broadly speaking, this will most certainly depend on the exact ligand sphere of palladium, and we plan to study the ligand effect in future work. For the complex that has been synthesized so far,

[(Tp*)PdMe2(pmbd)], it can be concluded that it easily forms aryldiazene radicals.

2.5 Experimental Section

2.5.1 General Specifications

All manipulations were carried out under an inert (N2 or Ar) atmosphere using standard glove box (M Braun UniLab) and Schlenk-type techniques except where noted, and using oven-dried glassware. All reagents for which synthesis is not given were commercially available from Sigma Aldrich, Alfa Aesar, or STREM and were used as received without further purification. Solvents were purified prior to use by passing through a column of activated alumina using an MBRAUN

SPS. NMR solvents were obtained from Cambridge Isotope Laboratories. Diethyl ether (Et2O), dimethoxyethane (DME), tetrahydrofuran (THF), and benzene-d6 (C6D6) were further purified using by vacuum distillation from purple sodium benzophenone ketyl. Acetone and acetone-d6 were dried/degassed over molecular sieves (3Å) before use. Pd(COD)Me2 was prepared by a modified literature procedure by addition of two equivalents of commercial methylmagnesium bromide solution (Aldrich, 3 M in Et2O) to a slurry of Pd(COD)Cl2 in dimethoxyethane at – 30 °C. All NMR spectra were recorded on a Bruker Avance III 400 MHz spectrometer. Standard VNMR pulse sequences for heteronuclear multiple bond correlation (gHMBC) were used. The 2-

15 1 3 D experiment ( N- H gHMBC) was also optimized for an average JNH coupling of 3.5 Hz. Chemical shifts are reported in ppm and referenced to residual solvent resonance peaks. 15N reference standard was CH3NO2. Abbreviations for the multiplicity of NMR signals are s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), bs (broad singlet). Elemental analyses were conducted at Guelph (Chemisar) Laboratories, Guelph, ON, Canada.

2.5.2 Synthesis of 1,5-Cyclooctadiene Dimethyl Palladium(II) {Pd(COD)Me2)}

285 mg (1 mmol) of (COD)PdCl2 was slurried in 30 mL of DME and cooled to -40 °C. To this slurry was added 700 μL (2.1 mmol) of a 3M Et2O solution of CH3MgBr. This was allowed to stir for 30 min. while warming to 0 °C and then a few drops of 1,4-dioxane were added to precipitate MgBrCl as the dioxane coordination polymer. The slurry was cooled again to -40 °C and filtered through a cooled plug of silica gel. The solvent was removed from the resulting clear solution in vacuo at -30 °C, leaving a white powder which was kept under an argon atmosphere in a freezer at -36 °C. (Yield: 198 mg, 81%)

1 H NMR (C6D6, 400MHz, δ (ppm)): 5.18 (s, 4H, COD-CH), 1.84 (bs, 8H, COD-CH2), 0.92 (s, 6H, Pd–Me)

2.5.3 Synthesis of Pd(Tp*)Me2(pmbd)

A solid sample of Pd(COD)Me2 (61.1 mg, 0.25 mmol) was placed into a cooled 25 mL air-free round-bottom flask equipped with a magnetic stirring bar. The flask was cooled to -40 °C, after which 5 mL of acetone was added. A solution of KTp* (84.1 mg, 0.25 mmol) in 2 mL acetone was then added at -40 °C and the reaction mixture was allowed to stir while coming to room temperature under Ar over a period of 1 h. A solution of pmbd-BF4 (55.5 mg, 0.25 mmol) in 3 mL acetone was added at room temperature, which immediately resulted in a rapid color change within minutes from colorless to dark orange without any intermediate colour observed. The solvent was removed by evaporation under high vacuum at room temperature and the resulting solid residue was transferred to the glove box, extracted with 20 mL of diethyl ether, filtered using a 0.2 μm

Whatman disk filter and 30 mL glass syringe, and evaporated to dryness to give an orange solid of Pd(Tp*)Me2(pmbd) (1) in 94% (133.6 mg, 0.23 mmol).

1 H NMR (C6D6, 400MHz, δ (ppm)): 2.073 (s, 6H, pz–Me), 2.233 (s, 3H, pz–Me), 2.267 (s, 6H, pz–Me), 2.279 (s, 3H, pz–Me), 2.373 (s, 6H, Pd–Me), 3.163 (s, 3H, Ph–OMe), 5.619 (s, 1H, pz– CH), 5.636 (s, 2H, pz–CH), 6.630 (d, 2H, o-Ph), 7.561 (d, 2H, m-Ph).

13 C{H} NMR (C6D6, 101.6 MHz, δ (ppm)): 12.771 (2C, pz–Me), 12.982 (2C, pz–Me), 13.171 (1C, pz–Me), 13.864 (1C, pz–Me), 26.738 (2C, Pd–Me), 54.955 (1C, Ph–OMe), 106.950 (2C, pz– CH), 107.915 (1C, pz–CH), 114.583 (2C, o-Ph), 124.114 (2C, m-Ph), 140.608 (1C, p-Ph), 142.998 (2C, pz–C(Me)), 144.192 (1C, pz–C(Me)), 148.905 (2C, pz–C(Me)), 150.553 (1C, pz–C(Me)), 161.192 (1C, ipso-Ph).

Anal. Calcd for C24H35BN8OPd (568.81 g mol-1): C, 50.67; H, 6.20; N, 19.69. Found: C, 50.52; H, 6.01; N, 19.47.

2.5.4 X-ray structure determination of Pd(Tp*)Me2(pmbd)

X-ray quality crystals for 1 were obtained from a concentrated diethyl ether solution after three days at -35 °C. An orange single crystal with dimensions of 0.10 x 0.10 x 0.03 mm was chosen for the single-crystal X-ray diffraction study. Data were collected61 on a Nonius-Kappa CCD diffractometer using graphite monochromatized Mo Kα radiation (λ = 0.71073 Å) at 150(2) K using the ω-scan technique. Unit cell parameters were refined and data were reduced with Denzo- SMN.62 The absorption correction performed was semi-empirical from equivalents. The structure was solved using direct methods63 and refined, full-matrix least-squares on F2, using SHELXTL V6.1.64 Crystal data is given in Table 2.2.

Table 2.2. Crystal data and structure refinement for Pd(Tp*)Me2(pmbd). ______

Chemical formula C24H35BN8OPd Formula weight 568.81 Temperature (K) 150(2) Crystal system Triclinic Space group P-1 a (Å) 7.9636(7) b (Å) 12.4983(12) c (Å) 13.5112(12) β (°) 87.268(2) V (Å3) 1329.8(2) Z 2 Calculated density (Mg/m3) 1.421 Crystal size (mm) 0.10 x 0.10 x 0.03 F(000) 588 θ range (°) 1.51 to 27.49 Reflections collected 22570 Independent reflections (Rint) 6073 [R(int) = 0.0540] Data / restraints / parameters 6073 / 0 / 329 Goodness-of-fit on F2 1.020

Final R indices R1 = 0.0372, wR2 = 0.0790

R indices (all data) R1 = 0.0548, wR2 = 0.0862 Range of h, k, l −10/8, −16/16, −17/17 ______

2.5.5 Thermolysis of Pd(Tp*)Me2(pmbd) in C6D6

A solution of 3 mg of Pd(Tp*)Me2(pmbd) (2.1) in C6D6 was prepared under N2 atmosphere, placed into a J. Young tube. The reaction mixture was heated at 70 °C and periodically analyzed by 1H NMR Spectroscopy (see Figures 2.8 and 2.9). 2.1 disappeared to give one major organometallic product, Pd(Tp*)Me3 (2), and one major organic product, 4,4'-dimethoxybiphenyl, in ~34% and ~36% yield, respectively, after heating for 1 h at 70 °C. The yields of products were determined by 1H NMR integration relative to an internal standard (The silicone ‘grease’ signal was used as the internal standard, which was set to an integration of 1.0 in each spectrum and integration intensities of the least obscured peaks were used to determine NMR yields, namely aromatic C-H peaks for 4,4'-dimethoxybiphenyl and pyrazole C-H peaks for 2.2).

1 H NMR (C6D6, δ (ppm)): 4,4'-Dimethoxybiphenyl: 3.329 (s, 6H, Ph–OMe), 6.841 (d, 4H, o-Ph), 7.417 (d, 4H, m-Ph); 2.2: 1.859 (s, 9H, Pd–Me), 2.181 (s, 9H, pz–Me), 2.249 (s, 9H, pz–Me), 5.587 (s, 3H, pz–CH).

1 Figure 2.7. H NMR spectrum of Pd(Tp*)Me2(pmbd) in C6D6 before heating at 70 °C with integration toward the silicone grease signal as the internal standard.

1 Figure 2.8. H NMR spectrum of Pd(Tp*)Me2(pmbd) in C6D6 after heating for 1 h at 70 °C with integration toward the silicone grease signal as the internal standard.

2.5.6 Thermolysis of Pd(Tp*)Me2(pmbd) in Acetone-d6

A solution of 2.9 mg (0.01 mmol) of Pd(COD)Me2 in acetone-d6 was prepared under N2 atmosphere, placed into a J. Young tube. The solution was cooled to -50 °C and one equivalent

(3.4 mg, 0.01 mmol) of KTp* in acetone-d6 was added and allowed to react for 30 minutes to afford K[Pd(Tp*)Me2] in situ. One equivalent (2.2 mg, 0.01 mg) of [pmbd][BF4] in acetone-d6 1 was then added to give Pd(Tp*)Me2(pmbd) (1), and a H NMR spectrum was immediately recorded. The reaction mixture was heated at 70 °C and periodically analyzed by 1H NMR spectroscopy (see Figures 2.10 and 2.11). 2.1 disappeared to give one major organometallic product, Pd(Tp*)Me3 (2.2), and one major organic product, anisole-d1 in ~50% and ~49% yield, respectively, after heating for 1 h at 70 °C. 4,4'-dimethoxybiphenyl was formed in ~5% yield. The yields of products were determined by 1H NMR integration relative to an internal standard (The acetone-d5 residual solvent signal was used as the internal standard, which was set to an integration

47 of 1.0 in each spectrum and integration intensities of the least obscured peaks were used to determine NMR yields, namely methoxy C-H peaks for anisole-d1 and 4,4'-dimethoxybiphenyl, and pyrazole C-H peaks for 2.2).

1 H NMR (Acetone-d6, δ (ppm)):

1,5-COD: 5.51 (bs, 4H, CH). 2.33 (bs, 8H, CH2); Anisole-d1: 3.78 (s, 3H, Ph–OMe), 7.27 (m, 2H, Ph), 6.92 (d, 2H, Ph); 4,4'-Dimethoxybiphenyl: 3.82 (s, 6H, Ph–OMe), 7.53 (d, 4H, o-Ph), 6.99 (d, 4H, m-Ph); 2.2: 1.64 (s, 9H, Pd–Me), 2.28 (s, 9H, pz–Me), 2.37 (s, 9H, pz–Me), 5.77 (s, 3H, pz– CH).

1 Figure 2.9. H NMR spectrum of Pd(Tp*)Me2(pmbd) in acetone-d6 before heating at 70 °C with integration toward the solvent residual signal as the internal standard.

1 Figure 2.10. H NMR spectrum of Pd(Tp*)Me2(pmbd) in acetone-d6 after heating for 1 h at 70 °C with integration toward the solvent residual signal as the internal standard.

2.5.7 Isolation and characterization of Pd(Tp*)Me3 (2.2)

2.2 was isolated from the crude reaction mixture of thermolysis of 50 mg of Pd(Tp*)Me2(pmbd) in C6D6 at 70 °C after 1 h. The solvent of the reaction mixture was removed by evaporation under high vacuum at room temperature and the resulting solid residue was transferred to the glove box, extracted with 2 mL of diethyl ether, filtered using a 0.2 μm Whatman disk filter, and recrystallized at -35 °C to give the colorless crystals of Pd(Tp*)Me3 in ~43% yield (16.5 mg, 0.036 mmol). It was also characterized by X-ray crystallography, elemental analysis, and 1H and 13C{H} NMR spectroscopies (see Figures 2.11 and 2.12). It is worth noting that the isolated yield of 2.2 is greater than the reported NMR yield. This is likely due to the use of Et2O in the isolation of the product, since the product appears to have better solubility in Et2O than in C6D6.

1 H NMR (C6D6, δ (ppm)): 1.861 (s, 9H, Pd–Me), 2.183 (s, 9H, pz–Me), 2.250 (s, 9H, pz–Me), 5.588 (s, 3H, pz–CH). 13 C{H} NMR (C6D6, δ (ppm)): 12.863 (3C, Pd–Me), 13.365 (3C, pz–Me), 13.450 (3C, pz–Me), 107.322 (3C, pz–CH), 142.906 (3C, pz–C(Me)), 148.962 (3C, pz–C(Me)). Elemental Analysis:

Calculated for C18H31BN6Pd (448.70 g/mol): C, 48.14; H, 6.90; N, 18.72. Found: C, 47.95; H, 6.72; N, 18.56.

1 Figure 2.11. H NMR spectrum of Pd(Tp*)Me3 (2.2) in C6D6 at 25 °C.

13 Figure 2.12. C{H} NMR spectrum of Pd(Tp*)Me3 (2.2) in C6D6 at 25 °C.

2.5.8 X-ray structure determination of Pd(Tp*)Me3

X-ray quality crystals for 2.2 were obtained from a concentrated diethyl ether solution after a week at -35 °C. A colorless single crystal with dimensions of 0.38 x 0.23 x 0.15 mm was chosen for the single-crystal X-ray diffraction study. Data were collected on a Nonius-Kappa CCD diffractometer using graphite monochromatized Mo Kα radiation (λ = 0.71073 Å) at 110(2) K using the ω-scan technique. The structure refinement was performed using similar methods and techniques as used for 2.1. Crystal data are given in Table 2.3.

Table 2.3 Crystal data and structure refinement for Pd(Tp*)Me3. ______

Chemical formula C18H31BN6Pd Formula weight 448.70 Temperature (K) 110(2) Crystal system Orthorhombic Space group Pbca a (Å) 16.3066(9) b (Å) 9.8567(6) c (Å) 25.2093(14) β (°) 90 V (Å3) 1329.8(2) Z 8 Calculated density (Mg/m3) 1.471 S 16 F(000) 1856 θ range (°) 1.62 to 33.17 Reflections collected 32099 Independent reflections (Rint) 7727 [R(int) = 0.0282] Data / restraints / parameters 7727 / 0 / 248 Goodness-of-fit on F2 1.067

Final R indices R1 = 0.0271, wR2 = 0.0585

R indices (all data) R1 = 0.0354, wR2 = 0.0621 Range of h, k, l −25/20, −11/15, −38/38 ______

2.6 References for Chapter 2

36 (a) Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457. (b) Lyons, T. W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147. (c) Stille, J. K. Angew. Chem. Int. Ed. 1986, 25, 508. (d) Hassan, J.; Sevignon, M.; Gozzi, C.; et al. Chem. Rev. 2002, 102, 1359. (e) Chen, X.; Engle, K. M.; Wang, D.-H.; et al. Angew. Chem. Int. Ed. 2009, 48, 5094. (f) Littke, A. F.; Fu, G. C. Angew. Chem. Int. Ed. 2002, 41, 4176. (g) Nicolaou, K. C.; Bulger, P. G.; Sarlah, D. Angew. Chem. Int. Ed. 2005, 44, 4442. (h) Hartwig, J. F. Angew. Chem. Int. Ed. 1998, 37, 2046. (i) Kotha, S.; Lahiri, K.; Kashinath, D. Tetrahedron 2002, 58, 9633. (j) Martin, R.; Buchwald, S. L. Acc. Chem. Res. 2008, 41, 1461. (k) Corbet, J.-P.; Mignani, G. Chem. Rev. 2006, 106, 2651. (l) Surry, D. S.; Buchwald, S. L. Angew. Chem. Int. Ed. 2008, 47, 6338. (m) Negishi, E. I. Acc. Chem. Res. 1982, 15, 340. (n) Hartwig, J. F. Acc. Chem. Res. 1998, 31, 852.

37 Hall, T. L.; Lappert, M. F.; Lednor, P. W. J. Chem. Soc., Dalton Trans. 1980, 8, 1448.

38 Gattermann, L. Ber. Dtsch. Chem. Ges. 1890, 23, 1218.

39 Swain, C. G.; Sheats, J. E.; Harbison, K. G. J. Am. Chem. Soc. 1975, 97, 783.

40 Galli, C. Chem. Rev. 1988, 88, 765.

41 (a) Roglans, A.; Pla-Quintana, A.; Moreno-Mañas, M. Chem. Rev. 2006, 106, 4522. (b) Schmidt, B.; Elizarov, N.; Riemer, N.; Hölter, F. Eur. J. Org. Chem. 2015, 5826. (c) Colleville, A. P.; Horan, R. A. J.; Tomkinson, N. C. O. Org. Process Res. Dev. 2014, 18, 1128.

42 (a) Lyons, T. W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147; (b) Xu, L.-M.; Li, B.-J.; Yang, Z.; Shi, Z.-J. Chem. Soc. Rev. 2010, 39, 712; (c) Canty, A. J. Dalton Trans. 2009, 10409; (d) Chen, X.; Engle, K. M.; Wang, D. H.; Yu, J. Q. Angew. Chem., Int. Ed. 2009, 48, 5094; (e) Muniz, K. Angew. Chem., Int. Ed. 2009, 48, 9412; (f ) Luo, J.; Rath, N. P.; Mirica, L. M. Organometallics 2013, 32, 3343; (g) Bonney, K. J.; Schoenebeck, F. Chem. Soc. Rev. 2014, 43, 6609; (h) Sehnal, P.; Taylor, R. J. K.; Fairlamb, I. J. S. Chem. Rev. 2010, 110, 824.

43 De Crisci, A. G.; Hamer, G. K.; Fekl, U. J. Coord. Chem. 2010, 63, 2928

44 Harris, D. C. Quantitative Chemical Analysis. W. H. Freeman: New York, 2010. Appendix H – Standard Reduction Potentials.

45 Canty, A. J. Dalton Trans. 2009, 10409

46 Canty, A. J.; Jin, H.; Roberts, A. S.; Skelton, B. W.; White, A. H. Organometallics 1996, 15, 5713.

47 De Crisci, A. G.; Hamer, G. K.; Fekl, U. J. Coord. Chem. 2010, 63, 2928

48 (a) Canty, A. J.; Honeyman, R. T.; Roberts, A. S.; Traill, P. R.; Colton, R.; Skelton, B. W.; White, A. H. J. Organomet. Chem. 1994, 471, C8; (b) Canty, A. J.; Jin, H.; Roberts, A. S.; Skelton, B. W.; Traill, P. R.; White, A. H. Organometallics 1995, 14, 199.

49 (a) Canty, A. J. Dalton Trans. 2009, 10409. (b) Luo, J.; Rath, N. P.; Mirica, L. M. Organometallics 2013, 32, 3343.

50 (a) Wick, D. D.; Goldberg, K. I. J. Am. Chem. Soc. 1997, 119, 10235; (b) Wick, D. D.; Goldberg, K. I. J. Am. Chem. Soc. 1999, 121, 11900; (c) Fekl, U.; van Eldik, R.; Lovell, S.; Goldberg, K. I. Organometallics 2000, 19, 3535.

51 (a) Haymore, B. L.; Hughes, M.; Mason, J.; Richards, R. J. Chem. Soc. Dalton Trans. 1988, 2935; (b) Mason, J.; Larkworthy, L. F.; Moore, E. A. Chem. Rev. 2002, 102, 913.

52 Su, H., communication to the CCDC 2006: 4,4'-dimethoxybiphenyl: Refcode JEHDEG.

53 Fekl, U.; van Eldik, R.; Lovell, S.; Goldberg, K. I. Organometallics 2000, 19, 3535.

54 (a) Byers, P. K.; Canty, A. J.; Skelton, B. W.; White, A. H. J. Chem. Soc., Chem. Comm. 1987, 1093; (b) Byers, P. K.; Canty, A. J.; Skelton, B. W.; White, A. H. Organometallics 1990, 9, 826; (c) Canty, A. J.; Dedieu, A.; Jin, H.; Milet, A.; Skelton, B. W.; Trofimenko, S.; White, A. H. Inorg. Chim. Acta 1999, 287, 27; (d) Bayler, A.; Canty, A. J.; Edwards, P. G.; Skelton, B. W.; White, A. H. Dalton Trans. 2000, 3325. 54

55 Khusnutdinova, J. R.; Qu, F.; Zhang, Y.; Rath, N. P.; Mirica, L. M. Organometallics 2012, 31, 4627.

56 Aye, K.-T.; Canty, A. J.; Crespo, M.; Puddephatt, R. J.; Scott, J. D.; Watson, A. A. Organometallics 1989, 8, 1518.

57 (a) Derat, E.; Maestri, G. Comput. Mol. Sci. 2013, 3, 529; (b) Khusnutdinova, J. R.; Rath, N. P.; Mirica, L. M. J. Am. Chem. Soc. 2010, 132, 7303; (c) Tang, F.; Qu, F.; Khusnutdinova, J. R.; Rath, N. P.; Mirica, L. M. Dalton Trans. 2012, 41, 14046; (d) Qu, F.; Khusnutdinova, J. R.; Rath, N. P.; Mirica, L. M. Chem. Comm. 2014, 50, 3036; (e) Khusnutdinova, J. R.; Rath, N. P.; Mirica, L. M. Inorg. Chem. 2014, 53, 13112.

58 (a) Lanci, M. P.; Remy, M. S.; Kaminsky, W.; Mayer, J. M.; Sanford, M. S. J. Am. Chem. Soc. 2009, 131, 15618; (b) Liu, Q.; Dong, X.; Li, J.; Xiao, J.; Dong, Y.; Liu, H. ACS Catal. 2015, 5, 6111; (c) Khusnutdinova, J. R.; Rath, N. P.; Mirica, L. M. J. Am. Chem. Soc. 2012, 134, 2414; (d) Khusnutdinova, J. R.; Qu, F.; Zhang, Y.; Rath, N. P.; Mirica, L. M. Organometallics 2012, 31, 4627; (e) Tang, F.; Zhang, Y.; Rath, N. P.; Mirica, L. M. Organometallics 2012, 31, 6690.

59 Pan, K.; Lin, C.-R.; Ho, T.-I. Magn. Reson. Chem. 1993, 31, 632.

60 Cohen, T.; Lewarchik, R. J.; Tarino, J. Z. J. Am. Chem. Soc. 1974, 96, 7753.

61 Nonius COLLECT; Delft, The Netherlands, 1997-2003.

62 Otwinowski, Z.; Minor, W. In Methods in Enzymology, C. W. Carter and R. M. Sweet, Eds., Academic Press, New York, 1997, 276, 307.

63 Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. J. Appl. Crystallogr. 1994, 27,

1045.

64 Sheldrick, G. M. SHELXTL, Structure determination software suite, 6.1, Madison, WI, 2001.

Chapter 3 Facile Transformations of hydridotris(3,5-dimethylpyrazolyl)borate – An Intramolecular Frustrated Lewis Pair

3.1 Abstract

Two novel transformations were observed for the well-known hydridotris(3,5- dimethylpyrazolyl)borate (Tp*–). The uncommon Lewis base abstractor aryldiazonium cation + (ArN2 ; Ar = p-Br-C6H4 or p-MeO-C6H4) if used in equimolar amounts, removes a hydride from Tp*– in acetone to generate the unstable tris(3,5-dimethylpyrzolyl)borane in situ, which traps acetone in a Frustrated Lewis Pair (FLP) like reaction, producing (pz*)3B·(C3H6O) (compound 3.1), which contains a B-N-N-C-O heterocycle and can be characterized as a resonance hybrid + between a charge neutral borate and a donor-stabilized borane. If 2 equivalents of ArN2 are used, 3,5-dimethylpyrazolide (pz*-) can subsequently be abstracted from 3.1, and the resulting intermediate bifunctional borane adds another equivalent of acetone to form cation 3.2+ + {[(pz*)2B·(C3H6O)2] , obtained as 3.2-BF4}, a cationic compound that is the formal addition product of 2 equivalents of acetone to the bis(3,5-dimethylpyrazolyl)borinium cation. 3.2+ exists as a resonance hybrid between a cationic borate and a boronium cation. Compounds 3.1 and 3.2+ are characterized by 1H, 13C, and 11B NMR spectroscopy, mass spectrometry, and single-crystal X-ray diffraction. Compound 3.1 is potentially useful as a charge-neutral bidentate chelate ligand + with controlled sterics, and 3.2 is a C2-chiral cation.

3.2 Introduction

During the early investigations of the palladium(IV) aryldiazenido complexes discussed in chapter 2, reaction by-products could be isolated and characterized as decomposition products of excess free hydrotris(3,5-dimethylpyrazol-1-yl)borate ligand with acetone. These products were not obtained in the later experiments, which avoided excess Tp*-. In fact, these products turned out to

56 be entirely unrelated to any Pd chemistry and can, as I was able to show, be obtained simply from - + Tp* and ArN2 in acetone. From this reaction, two different products could be isolated. Compound 3.1 was the result of a hydride being abstracted from Tp*- and a molecule of acetone had been trapped in a frustrated Lewis pair (FLP)-type fashion. In compound 3.2+, an equivalent on 3,5- dimethylpyrazolide had been abstracted from 1, and a second molecule of acetone was trapped (see Figure 3.1). The origins of these compounds and their reactivity will be discussed in this chapter.

Figure 3.1. Structures of the two isolated side products from the synthesis of Pd(IV) aryldiazenido complexes. Crystallographic C-C bond lengths in the backbone of pyrazole rings bound to acetone are nearly identical in both cases, suggesting that both resonance structures have near equal contribution.

In the Lewis Theory of acids and bases, rather than being defined by their ability to donate or accept protons (as is commonly done by using the Brønsted definition), an acid is defined by its ability to accept a pair of electrons and a base by its ability to donate a pair of electrons. Lewis acids and bases react to form adducts in which a bonding interaction between the two centers is formed. The strength of this adduct depends on the Lewis acidity and basicity of the two components. Strictly speaking, any filled non-bonding orbital can be considered Lewis basic and any empty orbital Lewis acidic (see Figure 3.2). There are, however some archetypical compound classes which commonly demonstrate these properties. Amines and phosphines are among the most common Lewis bases, as they possess very basic lone pairs (while not carrying a formal charge). Lewis acids are most often represented by boranes and allanes, as they are often stable with an entirely unoccupied p-orbital.

Figure 3.2. Common Lewis acids and bases and Lewis Adduct.

Lewis acidic compounds can be useful for a wide range of reactivity, while Lewis bases are very often used as ligands for transition metals. The reactivity of Lewis pairs becomes substantially more interesting if the acid and base are sterically “frustrated” (Figure 3.3). By disallowing the formation of a formal Lewis adduct, frustrated Lewis pairs have demonstrated the ability to activate a wide range of small molecules.65 These compounds have found many uses in catalysis 66 and activation of small molecules such as H2 and CO2.

Figure 3.3. A demonstration of steric “frustration” in Lewis pairs.

Hydridotris(pyrazolyl)borate (Tp-) and its substituted analogues are common and highly useful ligands for transition metal complexes (see Figure 3.4). Their ability to act as either cis- coordinating bidentate or fac-coordinating tridentate ligands offers significant advantages, particularly in systems for which changes in coordination number are relevant.67 These ligands,

58 especially those containing bulkier substituents are robust against both hydrolysis and oxidation, and this stability is further improved by coordination to a metal atom.68

- Figure 3.4. Hydridotris(pyrazolyl)borate ligand class (R1 = R2 = H for Tp , R1 = R2 = CH3 for Tp*-) and binding modes of Tp ligands which favour square planar and octahedral geometries, respectively.

Reaction of these ligands with hydride abstracting reagents has the potential for the formation of a tris(pyrazolyl)borane (B(pz)3) species via abstraction of a hydride, but little is known about this type of reactivity. A tris(pyrazolyl)borane would have the potential to act as an intramolecular frustrated Lewis pair via a Lewis basic nitrogen and Lewis acidic boron, leading to interesting reactivity that has not yet been thoroughly explored. Two possible reaction pathways exist from an intermediately formed tris(pyrazolyl)borane. It can either be trapped by reaction with a substrate (Scheme 3.1), or by reaction with another equivalent of itself forming a dimer (or oligomer) (Scheme 3.2).

Scheme 3.1. A potential pathway for the generation and trapping of a tris(pyrazolyl)borane.

3.3 Results and Discussion

In order to determine the relative thermodynamics of dimerization/oligomerization vs substrate trapping, the energetics were calculated using density functional theory (see Scheme 3.2). Calculations were performed using GAMESS(US) at the B3LYP/6-31++G* level of theory.

Energy of oligomerization of tris(3,5-dimethylpyrazolyl)borane (B(pz*)3) was approximated by calculating the energy of dimerization, and the substrate molecule chosen for trapping was acetone. Geometry optimizations were performed in the gas phase and frequency corrections were performed in order to verify equilibrium geometries and adjust for free energy values.

Scheme 3.2. DFT-computed (B3LYP/6-31G**) thermodynamics for the polymerization of tris(pyrazolyl)borane (top) vs. FLP-like reactivity with acetone (bottom). The polymerization energy is approximated by the energy of dimerization. In the computed structure of the dimer, the central BNNBNN ring exhibits a slight chair conformation, with two of the four exterior pyrazole rings aligned with the central B-B axis and the other two twisted perpendicular to the axis due to steric constraints, such that no C2 or Cs symmetry exists. Thus, although the structure roughly approximates point group C2, the actual symmetry is only C1.

Calculations suggested that it would be more thermodynamically favorable for B(pz*)3 to form an FLP-like adduct with acetone than for the molecule to dimerize (or, by extension polymerize) by approximately 13 kcal/mol. This was congruent with observations made while working with palladium complexes of hydridotris(3,5-dimethylpyrazolyl)borate (Tp*) in the presence of aryldiazonium salts. Palladium(II) complexes of Tp* can be oxidized by aryldiazonium salts to form Pd(IV) aryldiazenido complexes without any reaction occurring at the ligand. In the presence of an excess of Tp*, however, it was curiously observed that free Tp* could also be oxidized by diazonium salts via a hydride abstraction (presumably to form intermediate B(pz*)3). These reactions were initially performed in acetone and as such, the isolated product was the FLP-type acetone adduct of B(pz*)3. Performing the reaction between Tp* and the tetrafluoroborate salts of

4-methoxybenzenediazonium and 4-bromobenzenediazonium ([pMBD][BF4] and [pBBD][BF4]) in a non-coordinating solvent such as benzene resulted in complex mixtures of largely insoluble products, suggested that in the absence of a polar substrate, the intermediately formed B(pz*)3 formed oligomeric or polymeric products.

NMR-scale reactions were performed in order to monitor the reactivity of aryldiazonium salts + (ArN2 , where Ar = p-Br-C6H4 or p-MeO-C6H4) with KTp* in acetone at room temperature. The + reaction of ArN2 with KTp* in a 1:1 molar ratio yields one main product (compound 3.1) within less than 1 minute. A colour change (from colourless to orange) appears complete within seconds of mixing, and immediate acquisition of an NMR spectrum confirms the reaction to be complete. + + – Similar reactions, where 2 equiv. of ArN2 are used, yield a different product (3.2 BF4 ) as the main product, again within less than 1 min, as ascertained by NMR spectroscopy.

Both species bear clear 1H NMR signatures based on the number and intensity of methyl and pyrazole CH signals (five signals with intensity 6:6:6:3:3 for methyl groups in 3.1, two signals with intensity 2:1 for pyrazole CH signals in 3.1; four signals with intensity 6:6:6:6 for methyl groups in 3.2-BF4, one signal with intensity 2 for pyrazole CH in 3.2-BF4). If the reaction is 1 performed in acetone-d6 instead of protio-acetone, the H NMR signals for the acetone-derived methyl groups are missing. 13C and 11B NMR spectra confirm the structural assignment from 1H NMR spectroscopy, and X-ray crystal structure analyses of crystalline samples provide further confirmation. The suggested mechanism is also shown in Scheme 3.3.

Scheme 3.3. Formation of 3.1 (A) and 3.2+ (B). In each case here, only a single resonance form has been drawn, but it can be assumed that both forms (as shown in Figure 3.1) are relevant throughout.

+ Addition of the first equivalent of ArN2 to KTp* results in initial hydride abstraction followed by trapping of acetone to form compound 3.1. Abstraction of 3,5-dimethylpyrazolate (pz*–) from 3.1 + – – by the second equivalent of ArN2 yields the cationic 3.2-BF4. Both steps involve ligand (H , pz* ) removal from the boron atom, and while in none of the steps the boron center is being oxidized, the first step may also be classified as an oxidation, since oxidation of the hydridic hydrogen atom occurs in this step. Abstraction of this hydride leads to the generation (3,5-dimethylpyrazolyl)- borane, a historically elusive compound which inherently possess the ability to act as a frustrated Lewis pair (FLP). While this type of bifunctional reactivity with a substrate has been reported for a mono(pyrazolyl) borane,69 the utility of a tris(pyrazolyl) borane had not been observed until this work. We were able to obtain both acetone-quenched bifunctional boranes, 3.1 and 3.2-BF4, on a preparative scale (56% for 3.1, 80% for 3.2-BF4), where selectivity is controlled by the number of equivalents of oxidant used. The yield in the synthesis of 3.1 was inherently lower due to the poor

+ solubility of KTp* and ArN2-BF4 in acetone, resulting in the undesired formation of 3.2 , even when the reaction was performed in a 1:1 ratio. The structures of 3.1 and 3.2+, from single-crystal X-ray crystallography, are shown in Figure 3.5, with selected parameters in the caption.

+ Figure 3.5. Anisotropic displacement plots of 3.1 and 3.2 (cation in 2-BF4) showing 30% probability ellipsoids. Selected distances [Å] and angles [°] for 3.1: B1–O1 1.462(2), B1–N1 1.532(2), B1–N4 1.570(2), B1–N5 1.530(2), N3–C16 1.502(2), C16–O1 1.411(2); O1–B1–N1 111.7(1), O1–B1–N4 99.0(1), O1–B1–N5 113.3(1), N1–B1–N5 112.0(1), N4–B1–N5 110.0(1). Selected distances [Å] and angles [°] for 3.2+: B1–O1 1.441(2), B1–N2 1.560(2), N1–C6 1.503(2), C6–O1 1.4104(17); O1–B1–O1’ 118.0(2), N2–B1– N2’ 113.3(2), O1–B1–N2 100.1(1).

+ Of particular interest is that 3.2 belongs to the chiral point group C2, both in solution (NMR) and in the crystal. In the crystal, the two boron-containing rings within one spirocyclic molecule are related by a crystallographic twofold axis. The chirality of 3.2+ is due to the dissymmetric spirocyclic configuration. Of course, due to the preparation employed (from achiral reagents only),

3.2+ is obtained as a racemate. Compound 3.1 is expected to act as a charge-neutral bidentate “N- N” chelate ligand towards transition metals, very similar to bis(pyrazolyl)methane it shares most geometric features with. The steric bulk arising from the bound acetone, however, would shield one side of a square-planar transition metal complex in a way that is hard to achieve with bis(pyrazolyl)methanes. Reaction of 3.1 with 1 equivalent of anhydrous ZnBr2 in acetone-d6 resulted in quantitative shifting of all 1H NMR peaks, where the CH signal for the two pyrazole rings that are expected to act as donors shifted most (Δδ = 0.46 ppm). The calculated structure of the ZnCl2 adduct of 3.1 is shown in Figure 3.6.

Figure 3.6. DFT Computed structure of the ZnCl2 adduct of compound 3.1, showing the potential for single-face steric protection.

A number of analogous reactions involving trapping agents other than acetone were generally + unsuccessful. Reactions of ArN2 with KTp* in the presence of H2, CO2, CS2, MeCN, and benzophenone all produced inseparable mixtures of products, from which analogous compounds to 3.1 and 3.2 could not be isolated. It is possible that the ability of the latent B(pz*)3 to act as a frustrated Lewis pair requires a relatively reactive substrate (such as acetone) in very large excess.

3.4 Conclusions

Two very rapid transformations were found when hydridotris(3,5-dimethylpyrazolyl)borate was treated with aryldiazonium cation. The tris(pyrazolyl)borane initially generated in situ was trapped by acetone in a manner that represents FLP reactivity. The product (3.1) was isolated. It is potentially useful as a charge-neutral bidentate chelate ligand. Compound 3.1 reacts further if + another equivalent of aryldiazonium is present to form a C2-chiral cation, 3.2 . Future work will more deeply explore the reaction of boron hydrides with aryldiazonium cations, as well as the activation of unsaturated molecules with the bifunctional boranes thus generated. The coordination chemistry of 3.1 and its analogues will also be investigated.

3.5 Experimental Section 3.5.1 General Specifications

All manipulations were carried out under an inert (Ar) atmosphere using standard glove box (M Braun UniLab) and Schlenk techniques except where noted, and using oven-dried glassware. All reagents for which synthesis is not given were commercially available from Sigma Aldrich, Alfa Aesar, or STREM, and were used as received without further purification. Solvents were purified using by vacuum distillation from purple sodium benzophenone ketyl or activated molecular sieves. NMR solvents were obtained from Cambridge Isotope Laboratories. NMR data was obtained on a Bruker Avance III 400 MHz spectrometer. Chemical shifts are reported in ppm and referenced to residual solvent resonance peaks.

3.5.2 Synthesis of the acetone adduct of tris(3,5-dimethylpyrazol-1- yl)borane (Compound 3.1)

336 mg (1 mmol) of potassium tri(3,5-dimethyl-1-pyrazolyl)borohydride was dissolved in 15 mL of acetone and cooled to -40 °C. A -40 °C solution of 222 mg (1 mmol) of 4- methoxybenzenediazonium tetrafluoroborate in 15 mL of acetone was added slowly, with stirring. A rapid colour change from colourless to orange was observed and the reaction was allowed to stir for 2 hours while coming to room temperature. The solvent was removed in vacuo leaving an orange residue which was extracted with 20 mL of toluene and filtered through silica gel. The solution was concentrated to ~4 mL and cooled to -28 °C in a freezer yielding orange crystals (crystal data in Table 3.1). (198 mg, 56% yield).

1 H NMR (Acetone-d6, 400 MHz, δ (ppm)): 6.24 (s, 1H, pz–CH), 5.67 (s, 2H, pz–CH), 2.48 (s, 3H, pz–Me), 2.18 (s, 3H, pz–Me), 2.04 (s, 6H, pz–Me) , 1.95 (s, 6H, acetone–Me) , 1.69 (s, 6H, pz– Me). 13 C{H} NMR (Acetone-d6, 101.6 MHz, δ (ppm)): 148.12, 144.54, 137.70, 112.18, 106.58, 104.22, 96.40, 29.99, 28.33, 14.11, 12.48, 12.23, 10.99. 11 B{H} NMR (Acetone-d6, 128.4 MHz, δ (ppm)): 3.43 (bs, 1B).

ESI-MS: Calcd for C18H27BN6O (354.23 g/mol): M/Z Calc’d: 355.23731, Found: 355.24

Table 3.1. Crystal data and structure refinement for 3.1. ______

Empirical formula C18H27BN6O Formula weight 354.27 Temperature 147(2) K Wavelength 1.54178 Å Crystal system Monoclinic

Space group P21/n Unit cell dimensions a = 8.6444(4) Å α= 90°.

b = 19.8405(10) Å β= 94.065(2)°.

c = 11.0084(5) Å γ = 90°.

Volume 1883.29(15) Å3 Z 4 Density (calculated) 1.249 Mg/m3 Absorption coefficient 0.641 mm-1 F(000) 760 Crystal size 0.13 x 0.12 x 0.10 mm3 Theta range for data collection 4.46 to 66.25°. Index ranges -10<=h<=7, -21<=k<=23, -12<=l<=13 Reflections collected 12217 Independent reflections 3186 [R(int) = 0.0258] Completeness to theta = 66.25° 96.4 % Absorption correction Semi-empirical from equivalents Max. and min. transmission 0.7538 and 0.6432 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 3186 / 0 / 243 Goodness-of-fit on F2 1.043

Final R indices [I>2sigma(I)] R1 = 0.0364, wR2 = 0.0911

R indices (all data) R1 = 0.0389, wR2 = 0.0932 ______

3.5.3 Synthesis of the bis-acetone adduct of bis(3,5-dimethylpyrazol-1- yl)borinium (Compound 3.2)

168 mg (0.5 mmol) of potassium tri(3,5-dimethyl-1-pyrazolyl)borohydride was dissolved in 10 mL of acetone and cooled to -40 °C. A -40 °C solution of 222 mg (1 mmol) of 4- methoxybenzenediazonium tetrafluoroborate in 15 mL of acetone was added slowly, with stirring. A rapid colour change from colourless to red was observed and the reaction was allowed to stir for 2 hours while coming to room temperature. The solvent was removed in vacuo leaving an orange residue which was extracted with 30 mL of toluene and filtered through a coarse filter frit. The solution was concentrated to ~8 mL and cooled to -28 °C in a freezer yielding red crystals (data in Table 3.2). (323 mg, 80% yield).

1 H NMR (Acetone-d6, 400 MHz, δ (ppm)): 6.56 (s, 2H, pz–CH), 2.61 (s, 6H, pz–Me), 2.14 (s, 6H, acetone–Me), 1.98 (s, 6H, acetone–Me) , 1.82 (s, 6H, pz–Me).

13 C{H} NMR (Acetone-d6, 101.6 MHz, δ (ppm)): 143.18, 141.68, 114.05, 96.78, 29.48, 27.71, 11.07, 10.80.

11 B{H} NMR (Acetone-d6, 128.4 MHz, δ (ppm)): 3.73 (bs, 1B, 3.2), -1.12 (s, 1B, BF4).

ESI-MS: Calcd for C16H26B2F4N4O2 (404.02 g/mol): M/Z Calc’d: 317.21, Found: 317.21

Table 3.2. Crystal data and structure refinement for 3.2. ______

Empirical formula C16H26B2F4N4O2 Formula weight 404.03 Temperature 147(2) K Wavelength 0.71073 Å Crystal system Orthorhombic

Space group P 21 21 2 Unit cell dimensions a = 10.808(2) Å α = 90°.

b = 13.350(3) Å β = 90°.

c = 6.9578(14) Å γ = 90°.

Volume 1003.9(4) Å3 Z 2 Density (calculated) 1.337 Mg/m3 Absorption coefficient 0.112 mm-1 F(000) 424 Crystal size 0.28 x 0.10 x 0.05 mm3 Theta range for data collection 2.42 to 27.46°. Index ranges -13<=h<=14, -6<=k<=17, -9<=l<=7 Reflections collected 5280 Independent reflections 2274 [R(int) = 0.0275] Completeness to theta = 27.46° 99.0 % Absorption correction Semi-empirical from equivalents Max. and min. transmission 0.7456 and 0.6580 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 2274 / 0 / 132 Goodness-of-fit on F2 1.055

Final R indices [I>2sigma(I)] R1 = 0.0334, wR2 = 0.0791

R indices (all data) R1 = 0.0392, wR2 = 0.0819 ______

3.5.4 Complexation of ZnBr2 with 3.1

In a J. Young NMR tube, 3.5 mg of 3.1 and 2.3 mg of anhydrous ZnBr2 were dissolved in acetone- 1 d6 and shaken for ~5 min. A H NMR was then collected.

1 H NMR (Acetone-d6, 400 MHz, δ (ppm)): 6.63 (s, 1H, pz–CH), 6.13 (s, 2H, pz–CH), 2.65 (s, 3H, pz–Me), 2.50 (s, 6H, pz–Me), 2.12 (s, 3H, pz–Me) , 1.78 (s, 6H, acetone–Me) , 1.73 (s, 6H, pz– Me).

3.6 References for Chapter 3

65 Stephan, D. W. Acc. Chem. Res. 2015, 48, 306.

66 Stephan, D. W. Dalton Trans. 2009, 17, 3129.

67 Trofimenko, S. Scorpionates: The Chemistry of Polypyrazolylborate Ligands. Imperial College Press, London, 1999.

68 Morawitz, T.; Zhang, F.; Bolte, M.; Bats, J. W.; Lerner, H.-W.; Wagner, M. Organometallics 2008, 27, 5067.

69 a) Theuergarten, E.; Schlüns, D.; Grunenberg, J.; Daniliuc, C. G.; Jones, P. G.; Tamm, M. Chem. Commun. 2010, 46, 8561. b) Theuergarten, E.; Schlösser, J.; Schlüns, D.; Freytag, M.; Daniliuc, C. G.; Jones, P. G.; Tamm, M. Dalton Trans. 2012, 41, 9101.

Chapter 4 Synthesis and Characterization of 1- and 2-Adamantyl Anions

4.1 Abstract

New routes to 1- and 2-adamantyl anion equivalents are described, starting from commercially available 1- and 2-adamantylzinc bromides and employing reducing metals (Mg; Li). Adamantylmagnesium bromides (both 1-AdMgBr and 2-AdMgBr) can reliably be produced via reaction of the corresponding adamantylzinc bromides with excess magnesium metal. Reactions of adamantylzinc bromides with stoichiometric lithium biphenylide or lithium 2,2’-bipyridylide afford the new diadamantylzinc species, 1-Ad2Zn and 2-Ad2Zn, isolable free of solvent and salt impurities. Addition of 2,2’-bipyridine (bipy) leads to the crystalline adducts 1-Ad2Zn(bipy) and

2-Ad2Zn(bipy), which were structurally characterized. The resulting adamantyl anions were used in order to generate the first adamantyl complexes of mercury (1- and 2-Ad2Hg), gold (1- and 2-

AdAu(PPh3), 1- and 2-AdAu(PCy3)) and bismuth (2-Ad2BiBr), of which 1- and 2-Ad2Hg, 2-

AdAu(PPh3), 2-AdAu(PCy3), and 2-Ad2BiBr were isolated. These include the first structurally characterized unsupported 2-adamantyl metal complexes.

4.2 Introduction

The adamantyl group, with its unique combination of lipophilicity, general inertness, biocompatibility, and structural rigidity, has become a key building block in the development of drug molecules,70 advanced materials,71 and ligands for transition metals.72 While many adamantane derivatives have been reported, the bulk of the synthetic methods used involve adamantyl electrophiles.73 The use of adamantyl anions in synthesis is a potentially powerful alternative approach but is rather underutilized.

The attempted syntheses of anionic adamantyl equivalents dates back to the 1970’s, with a great deal of work being performed by Gerard Molle in order to generate adamantyllithium and adamantyl Grignard reagents.74 It was discovered very early on that the behaviour of the adamantyl group in anion synthesis did not follow that of ordinary alkyl groups. Many

71 of the traditional reactions such as lithium-halogen exchange or the classic Grignard reaction were either entirely unsuccessful or complicated by large degrees of unwanted side reactions (such as C-C homocoupling or H atom abstraction from solvent).

Scheme 4.1. Selected literature preparations of adamantyl anions.

A number of reportedly successful syntheses of adamantyl lithium and adamantyl Grignard compounds have been published (see Scheme 4.1), but the yields are almost exclusively based on immediately quenching the product (deuterolysis, for example) and many of these reactions in my experience, have proven to be extremely irreproducible. A number of complications have been noted in various publications. In reactions involving lithium halogen exchange reactions of adamantyl halides with lithium metal, the sodium content of the metal was very important. It was reported by Molle that any deviation from 2% sodium content drastically reduced yields of the anionic product. Similarly, these reactions required

72 the inclusion of large amounts of crushed glass in order to remove lithium halide salts which adhered to the surface of the lithium metal and halted the progress of the reaction. The generally more reliable Grignard reaction also proved to be unreliable with adamantyl halide compounds. Reaction of bromoadamantane with magnesium turnings or highly active Rieke magnesium75 produces substantial quantities of homocoupled biadamantyls.76 Molle noted that the stirring speed had a large effect on the yields of these reactions, with the highest yields (ca. 60%) obtained in the absence of any mechanical stirring. The primary products of both of these reactions were adamantane and homocoupled biadamantyls (Scheme 4.2).

Scheme 4.2. The unusual reactivity of bromoadamantane with active magnesium.

A report was made in 1995 of the use of in-situ generated Rieke calcium to generate 1- adamantylcalcium bromide which was isolated by quenching with cyclohexanone at low temperature.77 This reaction was again found to have very poor reproducibility in our hands. Based on relative strengths of Ca-C and Li-C bonds, the general inaccessibility of alkyl calcium compounds, and the presence of an abundance of LiBr in the reaction mixture from generation of the Rieke calcium, it seems more likely that the active species is in fact 1- adamantyllithium.

A recent report by Knochel et al. (who also commented on the low reproducibility of the direct Grignard routes) detailing the synthesis of 1-adamantylzinc halides was a major step toward reproducible and high yielding methods for the preparation of adamantyl anion equivalents (Scheme 4.3).78

Scheme 4.3. Reductive insertion of zinc metal in bromoadamantane, a reliable method.

While important discoveries in the synthesis of organometallic adamantyl anion equivalents have been made, most reactions are either fraught with low yields or use reaction conditions that are hard to reproduce to the extent that synthetic use has been limited.79 High levels of contamination are also a problem, with large quantities of adamantane and biadamantyls being generated, which can be difficult to separate out, especially from organometallic products. This can even be observed in commercially available solutions of adamantylzinc bromides being contaminated with both zinc metal and adamantane. Beyond the synthetic difficulties which are common to adamantly anion chemistry, it is also important to note that most of the work on organometallic adamantyls focuses on the 1-adamantyl anion, and much less on the 2-adamantyl. While the 1- adamantyl fragment is sterically similar to a tert-butyl group, with additional bulk directly away from the binding site, the 2-adamantyl fragment provides a rather unique form of steric bulk. The conformation of the 2-adamantyl group is perhaps best described as a tied- back and conformationally locked cyclohexyl, such that each substituent always lies in an axial position, as depicted in Figure 4.1.

Figure 4.1. An atom of choice (blue) bound to a 1-adamantyl group (left) and to a 2-adamantyl group (right), highlighting the structural similarity to a tert-butyl (left, 1-adamantyl) and to an axially locked cyclohexyl (right, 2-adamantyl).

Starting from the commercially available 1- and 2-adamantylzinc bromides, new convenient methods for the synthesis and isolation of a number of carbanionic adamantyl compounds have been developed here. Reaction of these adamantyl anion equivalents with bismuth, gold, and mercury precursors leads to several new adamantyl metal complexes (1- and 2-Ad2Hg, 1- and 2-AdAu(PR3), and 2-Ad2BiBr), which include the first structurally characterized unsupported 2-adamantyl metal complexes.80,81 In this context, we refer to an adamantyl metal complex as “unsupported” if the metal is directly bonded to the adamantane cage only and to no secondary donor group. While unsupported 1-adamantyl metal complexes are more common, almost all unsupported adamantyl metal complexes (1- and 2-adamantyl), have been made using adamantyl electrophiles, for example through oxidative addition.82 All of the adamantyl metal complexes which are discussed in this chapter are among the rare examples of those formed through transmetallation.

4.3 Results and Discussion

Reductive transmetallation of adamantylzinc bromide (1-AdZnBr (4.1a) or 2-AdZnBr (4.1b)) provides a reliable route to alternative adamantyl anions. Reaction of adamantylzinc bromide with excess magnesium turnings furnishes the desired Grignard products (1- AdMgBr (4.2a) and 2-AdMgBr (4.2b)), with easily removable zinc metal precipitating out of solution (Scheme 4.4, I).

Scheme 4.4. New routes to 1-adamantyl (labelled a) and 2-adamantyl (labelled b) anion equivalents.

Using somewhat similar reductive techniques, it was discovered that reactions of adamantylzinc bromides with THF-solvated lithium biphenylide afford the new diadamantylzincs, namely 1-Ad2Zn (4.3a) and 2-Ad2Zn (4.3b), by removing ZnBr2 from the equilibrium via reduction to zinc metal (see Scheme 4.4, II). In this case, exactly one equivalent of Li biphenylide is required. The addition of excess Li biphenylide results in formation of adamantane, most likely through the corresponding adamantyl lithium compounds, which rapidly deprotonate THF to form adamantane.

We speculate that these transmetallations work by affecting the Schlenk equilibrium which is present in alkylzinc halide species (see Scheme 4.5). While the alkylzinc species itself is likely not susceptible to reduction, but rather the ZnBr2 which is constantly present in solution due to the Schlenk equilibrium. Gradual reduction of Zn(II) from solution as zinc metal can be achieved by using either Mg or Li metals. The stoichiometric reactions are shown below:

Scheme 4.5. Synthesis of Ad2Zn from reduction of AdZnBr with the Schlenk equilibrium shown. For Mg, 1,4-dioxane would be required to remove MgBr2 and form Ad2Zn.

In the case where an excess of the reducing metal is used, further reduction can be achieved (shown in Scheme 4.6 for excess Mg).

Scheme 4.6. Synthesis of adamantyl of Adamantyl Grignard reagent from Ad2Zn with the Schlenk equilibrium shown. The Schlenk equilibrium is suggested here as no direct reaction between Ad2Zn and Mg was ever observed.

The analogous reaction with an excess of lithium metal produces only adamantane, presumably via initial formation of the corresponding adamantyllithium, which is unstable in THF solution and is quenched.

Regarding further improvements of practical approaches to the synthesis of 4.3a and 4.3b, we found that the use of lithium 2,2’- bipyridylide83 (Scheme 4.4, III) instead of lithium biphenylide (Scheme 4.4, II) offers a significant advantage: no separation of the byproduct stemming from the lithium reductant is needed. If lithium biphenylide is used as the reductant, the biphenyl formed needs to be removed from the diadamantyl zinc product via sublimation. If lithium 2,2’-bipyridylide is used as the reductant, the 2,2’-bipyridine (bipy) that is formed will quantitatively precipitate from the reaction mixture as (bipy)LiBr, affording homoleptic and nitrogen ligand-free 4.3a or 4.3b.

The new routes are high yielding (ca. 80 %) and offer significant advantages over existing routes due to their high degree of reproducibility even when performed using non-activated

77 magnesium turnings (for the Grignards) or lithium chunks without the addition of sodium as an activator (for Ad2Zn synthesis). The new diadamantylzinc compounds can easily be isolated in pure solvent-free and salt-free form and are conveniently soluble in hydrocarbon solvents.

Finally, reaction of adamantylmagnesium bromide (synthesized as in Scheme 4.4, I) with adamantylzinc bromide (1:1) in the presence of dioxane (diox) also results in synthesis of

4.3a and 4.3b with [MgBr2(diox)]n precipitating from the reaction mixture, affording the cleanest and highest yielding (ca. 95%) route to 4.3a and 4.3b (Scheme 4.4, IV).

Addition of 2,2’-bipyridine to 4.3a or 4.3b cleanly affords the crystalline bipy adducts 4.4a and 4.4b. Bipy adducts of other dialkylzincs have been previously reported and characterized.84 Both 4.4a and 4.4b are strongly coloured (UV-Vis shown in Figure 4.2), which is expected of dialkylzinc bipyridyl complexes due to the ligand-to-ligand charge transfer85 between the alkyl group and bipyridine.86

3.5

2.5

1.5

Relative Relative Absorbance 1

0.5

0 250 300 350 400 450 500 550 600 650 700 Absorption Wavelength (nm)

Figure 4.2. UV-Vis spectra of 4.4a (red) and 4.4b (blue) in toluene.

Figure 4.3 shows the molecular structures of 4.4a and 4.4b from single crystal X-ray diffraction.

Figure 4.3. Anisotropic displacement plots (50% probability ellipsoids) for 4.4a (left) and 4.4b (right). Selected bond distances (Å) and angles (°) for 4.4a: Zn1–C11, 2.0223(19); Zn1–C21, 2.0209(19); Zn1–N1, 2.1724(16); Zn1–N2, 2.1754(16); C11–Zn1–C21, 132.13(8); N1–Zn1–N2, 74.83(6). The adamantyl containing C11 shows rotational disorder over two positions, and only one of the positions is shown. Selected bond distances (Å) and angles (°) for 4.4b: Zn1–C11, 2.0169(16); Zn1–C21, 2.0157(15); Zn1–N1, 2.2293(12); Zn1–N2, 2.2183(12); C11–Zn1–C21, 137.88(6); N1–Zn1–N2, 73.48(4).

4.4b contains a larger C-Zn-C angle of 137.9° compared to the 4.4a angle of 132.1°. This would usually indicate that the 2-adamantyl group can provide as much or more steric crowding at a metal centre than the 1-adamantyl group; however crystal packing effects are extremely important in dialkylzinc bipyridyl complexes, and the observed angle is not only dependent on the size of the alkyl group. For example, Me2Zn(bipy) has a C-Zn-C angle of o 87 o 88 127.96 , which is small as expected, whereas Et2Zn(bipy) has a C-Zn-C angle of 139.77 , which is larger than the C-Zn-C angles in both zinc diadamantyls reported here. This seems almost paradoxical, given that an ethyl, while larger than a methyl, is still much smaller

88 than an adamantyl. Inspection of the crystal structure of Et2Zn(bipy) shows that the effect is due to crystal packing: a neighbouring bipy packs into the C-Zn-C wedge of the complex. C-Zn-C bond deformations appear to be very soft and easily impacted by crystal packing.

Transmetallation of adamantyls onto more electronegative metals such as mercury, bismuth, and gold was accomplished, highlighting the utility of the new adamantyl anion equivalents. Both isomers of Ad2Hg can be synthesized by reaction of HgCl2 with any of the adamantyl anions (Scheme 2; 2:1 ratio transmetallating agent vs. Hg for AdZnBr and 1 AdMgBr, 1:1 for Ad2Zn). 1-Ad2Hg (4.5a) and 2-Ad2Hg (4.5b) were characterized via H and 13C NMR spectroscopy as well as elemental analysis (see experimental section). Yields of 78% and 80% were observed for 4.5a and 4.5b respectively. 1H and 13C NMR spectra of 4.5a and 4.5b are given in Figures 4.4-4.7, and are representative examples of the general appearance of 1- and 2-adamantyl NMR spectra. 1H spectra tend to have very broad signals and poor fine structure due to the higher order coupling and fixed dihedral angles of the adamantane cage.

1 Figure 4.4. H NMR spectrum of 4.5a in C6D6 at 25 °C. The signal at 1.74 ppm is identified as adamantane, and the signal at 0.29 ppm as silicone grease.

13 1 Figure 4.5. C{ H} NMR spectrum of 5a in C6D6 at 25 °C. Signals at 38.01 and 28.74 are from adamantane contaminant. Signal from 13C directly bonded to Hg cannot be seen.

1 Figure 4.6. H NMR spectrum of 5b in C6D6 at 25 °C. Total integration of 30H refers to two adamantyl ligands, confirmed by elemental analysis.

13 1 Figure 4.7. C{ H} NMR spectrum of 5b in C6D6 at 25 °C.

For bismuth, transmetallation was accomplished by reaction of BiBr3 with any of the three adamantyl anion precursors (Scheme 4.7), and isolable product was obtained for 2-adamantyl, in the form of 2-Ad2BiBr (4.6b) (Fig. 3). Attempts of isolating the analogous 1-Ad2BiBr (4.6a) have been unsuccessful due to rapid decomposition in solution. Complete decomposition of 6a occurs within minutes at low temperature (-40 o C) producing adamantane. In contrast, the 2-adamantyl analogue (4.6b) was found to be stable for up to 7 days at -35oC in solution; slow decomposition to form adamantane and bismuth metal was observed. However, a solid, crystalline sample was found to be stable for several weeks at room temperature.

Scheme 4.7. Demonstrated new routes to 1- and 2-adamantyl metal complexes.

While 4.6b can be made from BiBr3 by using either 4.1b, 4.2b, or 4.3b, not all precursors are equally advantageous. An isolated yield of 4.6b could not be obtained for the reactions using 4.1b or 4.2b due to the presence of a large amounts of impurities. This is likely due to THF coordination of the resulting zinc or magnesium salts which become extremely difficult to separate from the desired product yielding a wax or oily product. This problem can be resolved through the use of 4.3b in non-ether solvent, which produces 4.6b in 72% yield. 4.6b was analyzed via 1H and 13C NMR, as well as elemental analysis.

The molecular structure of 4.6b was obtained using single crystal X-ray diffraction (crystals for X-ray analysis were obtained by undergraduate student Kamalpreet Singh, Figure 4.8). Bond angles at bismuth (see legend to Figure 4.8) are between 90 o and 100 o, which is expected. Bismuth strongly favors angles close to 90o, due to involvement of almost pure p-orbitals in bonding.89 Likely due to this bond angle requirement and the size of the adamantyl group, formation of the tri-alkylated product 2-Ad3Bi was not observed, even in the presence of excess adamantyl anion equivalents.

Figure 4.8. Anisotropic displacement plot (50% probability ellipsoids) for 6b. Selected bond distances (Å) and angles (°): Bi1–C1, 2.310(5); Bi1–C11, 2.315(4); Bi1–Br1, 2.6879(6); C1–Bi1– C11, 95.28(16); C1–Bi1–Br1, 96.90(12); C11–Bi1–Br1, 92.57(12).

The apparent symmetry of 4.6b in solution is Cs. Bismuth is pyramidal and not prone to rapid inversion, such that there will be no apparent mirror plane through an individual adamantyl group. This is observed in both the 1H and 13C NMR spectra (Figures 4.9 and 4.10), in which ten inequivalent carbon atoms can be identified.

1 Figure 4.9. H NMR spectrum of 4.6b in C6D6 at 25 °C. (Contaminated with pentane, THF, and silicone grease).

13 1 Figure 4.10. C{ H} NMR spectrum of 4.6b in C6D6 at 25 °C. (Contaminated with adamantane)

Adamantyl complexes of gold could also be synthesized using these anionic precursors. Synthesis and characterization of the gold adamantyl complexes mentioned below was performed by

Fioralba Taullaj. 1-AdAu(PPh3) (7a) and 2-AdAu(PPh3) (4.7b) were generated through the reaction of either AdZnBr, AdMgBr or Ad2Zn with (triphenylphosphine)gold(I) chloride. The resulting compounds were analyzed by 1H, 13C, and 31P NMR spectroscopy. Similar to what was observed for the 1-adamantyl bismuth complex, 4.7a was much less stable than 4.7b and could not be isolated due to reduction (forming adamantane and gold metal), however 1H and 13C NMR spectroscopy confirmed that 4.7a can be generated in solution. 4.7b could be isolated in moderate yield (ca. 60%) when 4.3b was used in non-ether solvent. The use of 4.1b or 4.2b led to formation of an oil which showed large amounts of THF, likely coordinated to metal salts which could not be separated from the product. This observation confirms the usefulness of the new diadamantylzinc precursors over both Grignard and adamantylzinc bromide species.

Using a different phosphine, 2-AdAu(PCy3) (4.8b) was also synthesized by reaction of any of the adamantyl anions with (tricyclohexylphosphine)gold(I) chloride in moderate yield (ca. 65%). The

88 resulting complex was characterized by 1H, 13C, and 31P NMR spectroscopy. 4.8b showed greater stability when compared to 4.7b. In contrast, isolation of 1-AdAu(PCy3) (4.8a) was not possible due to reduction of gold by the more electron rich 1-adamantyl anion, as was observed for both 4.6a and 4.7a. However, 1H and 31P NMR data indicated that 4.8a was present in solution.

Synthesis of the 1-adamantyl complexes for bismuth and gold proved to be much more difficult than the synthesis of the 2-adamantyl analogues. The 1-adamantyl anion appears to act as a reducing agent, affording either gold or bismuth metal. Mercury, in contrast, was not reduced by the 1-adamantyl, likely due to lack of a low energy one-electron reduction pathway. Reduction of the 2-adamantyl complexes of bismuth and gold appeared to occur within 2-3 weeks. Transmetallation of adamantyls onto mercury, bismuth, and gold was possible using adamantylzinc bromide, adamantylmagnesium bromide, or diadamantylzinc as a precursor. Of the three possible transmetallating agents, diadamantylzinc gave the highest purities and yields. We associate the greater utility of diadamantyl zinc in transmetallation with the ability to isolate diadamantyl zinc species as solids, free of THF and salt impurities. This provides increased stoichiometric precision and added driving force due to precipitation of ZnX2 in the absence of coordinating solvents. Transmetallations using diadamantylzincs as precursors are therefore extremely favourable and can easily be performed in pure hydrocarbon solvents, including arenes and n-alkanes. The observed transmetallations are, to our knowledge, the first transmetallations of adamantyls to form stable, crystallographically characterized adamantyl metal complexes.

4.4 Conclusions

In conclusion, new and reliable routes to several carbanion equivalents for the 1- and 2- adamantyl fragments have been developed via reaction of commercially available 1- and 2-adamantylzinc bromides with reducing metals (Mg or Li). A new method for producing adamantyl Grignards is provided that generates good yields and completely avoids homocoupled side products, improving their synthetic utility. The first diadamantyl zinc compounds have been synthesized, both as homoleptic species and as the corresponding

2,2'-bipyridine adducts. These adamantyl anion equivalents can be made in good yield and isolated as pure compounds. The resulting adamantyl carbanion equivalents proved to be useful in transmetallations of adamantyl anions onto gold, bismuth, and mercury. 2-

Ad2BiBr, 2-AdAu(PPh3), 2-AdAu(PCy3) and 2-Ad2Zn(bipy) are the first unsupported 2- adamantyl metal complexes to be structurally characterized. The reactions of these anionic adamantyl precursors with other metals will be discussed in chapter 5.

4.5 Experimental Section 4.5.1 General Specifications:

All manipulations were carried out under an inert (N2 or Ar) atmosphere using standard glove box (M Braun UniLab) and Schlenk techniques except where noted, and using oven-dried glassware. All reagents for which synthesis is not given were commercially available from Sigma Aldrich, Alfa Aesar, or STREM, and were used as received without further purification. 1- and 2- Adamantylzinc bromide solutions (0.5 M in THF) were commercially available from Rieke metals via Sigma Aldrich. These compounds can alternatively be synthesized by reaction of 1- and 2- bromoadamantane with Rieke zinc in THF.90 Solvents were purified prior to use by passing through a column of activated alumina using an MBRAUN SPS. NMR solvents were obtained from Cambridge Isotope Laboratories. Diethyl ether (Et2O), tetrahydrofuran (THF), and benzene- d6 (C6D6) were further purified using by vacuum distillation from purple sodium benzophenone ketyl. Toluene and pentane were freshly distilled from activated molecular sieves. NMR data was obtained on a Bruker Avance III 400 MHz spectrometer. Chemical shifts are reported in ppm and referenced to residual solvent resonance peaks. Adamantane was present as a common contaminant in NMR spectra and was difficult to remove completely. Chemical shifts for adamantane are:

1 H NMR (C6D6, 400 MHz): 1.86 (bs, 4H, CH); 1.74 (bs, 12H, CH2) 13 C NMR (C6D6, 100.6 MHz): 38.02 (CH2); 28.76 (CH) It should be noted that α-carbon peaks for the 1-adamantyl species could not be located by 13C NMR. This is likely due to slow relaxation of the quaternary-like carbon bound to a metal center with no nuclear spin.

Some elemental analysis results for organometallic species were outside the range traditionally associated with analytical purity. Those data (deviation > 0.5 % points) are indicated with a “*” and they are provided here to illustrate the best values obtained.

4.5.2 Isolation of 1-adamantylzinc bromide as its THF adduct [1- AdZnBr(THF)x] (4.1a, THF-solvate)

2 mL (1 mmol) of 1-AdZnBr solution (0.5M in THF) were diluted to 5 mL with toluene. The solvent was removed in vacuo. The residue was extracted with 3 mL of toluene and placed in a freezer at -40 °C for 7 days, after which 1-AdZnBr(THF) had precipitated out as a soft, amorphous, white solid.

1 2 H NMR (C6D6, 400 MHz): 3.64 (bs, THF); 2.54 (s, 6H, CH2); 2.18 (d, 3H, CH2, JHH =10.8 Hz);

2.05 (d, 3H, CH2, coupling constant obscured by overlapping peaks); 2.02 (s, 3H, CH); 1.32 (bs, THF) 13 C NMR (C6D6, 100.6 MHz): 69.22 (THF); 45.37 (CH); 39.40 (CH2); 30.25 (CH2); 25.42 (THF)

4.5.3 Isolation of 2-adamantylzinc bromide as its THF adduct [2- AdZnBr(THF)x] (4.1b, THF-solvate)

The procedure performed was analogous to that for 4.1a, THF-solvate, above.

1 H NMR (C6D6, 400 MHz): 3.63 (bs, THF); 1.85-2.45 (m, 15H, 2-AdZnBr); 1.28 (bs, THF) 13 C NMR (C6D6, 100.6 MHz): 69.14 (THF); 43.76 (CH); 41.80 (CH2); 40.00 (CH2); 39.19 (CH);

34.16 (CH2); 30.05 (CH); 29.74 (CH); 25.34 (THF) .

4.5.4 Synthesis of 1-adamantylmagnesium bromide as its THF adduct [1- AdMgBr(THF)x] (4.2a, THF-solvate)

6 mL (3 mmol) of 1-AdZnBr solution (0.5M in THF) were stirred over excess magnesium turnings for 2 days. The mixture was allowed to settle over a period of 8 hours, and the supernatant was decanted off. The concentration of the solution was determined by titration against diphenyl ditelluride.91 (Yield 81%)

1 2 H NMR (C6D6, 400 MHz): 3.68 (bs, 12H, THF); 2.65 (s, 6H, CH2); 2.39 (d, 3H, CH2, JHH =11.4 2 Hz); 2.31 (d, 3H, CH2, JHH =11.7 Hz); 2.09 (s, 3H, CH); 1.35 (bs, THF)

13 C NMR (C6D6, 100.6 MHz): 69.14 (THF); 46.89 (CH); 40.67 (CH2); 30.40 (CH2); 26.51 (C- Mg); 25.39 (THF)

4.5.5 Synthesis of 2-adamantylmagnesium bromide as its THF adduct [2- AdMgBr(THF)x] (4.2b, THF-solvate)

The procedure performed was analogous to that for 2a, THF-solvate, above. (Yield 82%)

1 H NMR (C6D6, 400 MHz): 3.70 (bs, THF); 2.14-2.51 (m, 14H, 2-AdMgBr); 1.27 (bs, 1H, 2- AdMgBr, α-CH); 1.17 (bs, 6H, THF) 13 C NMR (C6D6, 100.6 MHz): 69.79 (THF); 43.51 (CH2); 41.48 (CH2); 39.92 (CH); 39.71 (CH);

35.12 (CH2); 30.97 (CH); 30.61 (CH); 25.06 (THF)

4.5.6 Synthesis of bis(1-adamantyl)zinc [1-Ad2Zn] (4.3a)

A) 154 mg (1 mmol) of biphenyl were dissolved in 15 mL of THF. 30 mg (4.3 mmol) of lithium pieces were added, and the solution was stirred for one hour, until a deep blue-green solution of lithium biphenylide was formed. The solution was transferred off of the excess lithium into a solution of 2 mL (1 mmol) of 1-adamantylzinc bromide solution (0.5 M in THF) which had been diluted to 10 mL with THF and cooled to -40°C. The mixture was allowed to come to room temperature and stirred for 24 hours, then the solvent was removed in vacuo, and the residue was heated to 40 °C under vacuum for 16 hours to remove the biphenyl by sublimation. The residue was extracted with 15 mL of toluene. The volume of the toluene solution was reduced until the solution became cloudy and was placed in a freezer at -40 °C for 2 days. The product was collected as an off-white microcrystalline powder. (Yield 130 mg, 77%)

B) 69 mg (0.4 mmol) of 2,2,-bipyridine were dissolved in 10mL of THF. 10mg (1.43 mmol) of lithium pieces were added, the solution was stirred for one hour, until a deep green solution of lithium 2,2’-bipyridylide was formed. The solution was transferred off the excess lithium into a solution of 0.8 mL (0.4 mmol) of 1-adamantylzinc bromide solution (0.5 M in THF) which had been diluted to 5mL with THF. The mixture was allowed to react with stirring for 24 hours, then

93 filtered through a coarse frit, followed by removal of the solvent in vacuo. The residue was extracted with 15 mL of toluene. The volume of the toluene solution was reduced until the solution became cloudy and was placed in the freezer at -40°C for 2 days. The product was collected as an off-white microcrystalline powder. (Yield 55 mg, 78%)

C) 10 mL of 1-AdZnBr solution (0.5 M in THF) was stirred over excess magnesium turnings at room temperature for 3 days. To this mixture was added 10 mL of 1-AdZnBr solution (0.5 M in THF) and 1 mL of dioxane with vigorous stirring. The mixture was stirred for 10 minutes and then filtered through a course frit. The solvent was removed in vacuo with gentle heating, then the residue was extracted with 80 mL of toluene. The toluene slurry was again filtered through a course frit and the solvent was removed in vacuo to give an off-white powder. This crude product is contaminated with ~5% adamantane by NMR integration, which can be removed by recrystallization with significant loss of yield. (Crude yield 1.61 g, 96%)

1 2 H NMR (C6D6, 400 MHz): 2.10 (m, 12H, CH2); 1.97 (d, 6H, CH2, JHH =12.2 Hz); 1.92 (d, 6H, 2 CH2, JHH =11.9 Hz); 1.80 (s, 6H, CH) 13 C NMR (C6D6, 100.6 MHz): 43.66 (CH); 39.10 (CH2); 29.55 (CH2)

HRMS (EI-TOF+): mass [M] calc’d for C20H30Zn 334.1639 Da, measured 334.1641 Da. -1 IR Absorptions: νmax/cm 2885 (CH2), 2841 and 2824 (CH), 1443 (CH2) MP 220 °C (decomp)

4.5.7 Synthesis of bis(2-adamantyl)zinc [2-Ad2Zn] (4.3b)

The procedure performed was analogous to that for 3a, above. Yield 86% for the lithium biphenylide method, 76% for the lithium 2,2’-bipyridylide method, and 95% for the Grignard method.

1 H NMR (C6D6, 400 MHz): 1.79-2.07 (m, 28H, 2-Ad2Zn); 1.38 (bs, 2H, 2-Ad2Zn, α-CH) 13 C NMR (C6D6, 100.6 MHz): 46.50 (CH); 42.13 (CH2); 40.91 (CH2); 38.56 (CH); 33.04 (CH2); 29.76 (CH); 29.41 (CH)

HRMS (EI-TOF+): mass [M] calc’d for C20H30Zn 334.1639 Da, measured 334.1644 Da.

-1 IR Absorptions: νmax/cm 2889 (CH2), 2833 (CH), 1443 (CH2) MP 230 °C (decomp)

4.5.8 Synthesis of 2,2’bipyridyl-bis(1-adamantyl)zinc [1-Ad2Zn(bipy)] (4.4a)

84 mg (0.25 mmol) of 1-Ad2Zn were dissolved in 5 mL of pentane and cooled to –40 °C. A solution of 40 mg (0.25 mmol) of 2,2’-bipyridyl in 5 mL of pentane at –40 °C was added dropwise. A dark blue solution formed from which olive crystals precipitated out of solution immediately. The solution was allowed to stand in a freezer at –40 °C overnight, after which crystals suitable for X- ray analysis were collected (crystal data in Table 4.1). (Yield 111 mg, 89%).

1 3 H NMR (C6D6, 400 MHz): 8.33 (d, 2H, bipy, JHH =5.3 Hz); 7.20 (m, 2H, bipy); 6.91 (t, 2H, bipy, 3 3 2 JHH =7.9 Hz); 6.57 (t, 2H, bipy, JHH =5.8 Hz); 2.39 (bs, 12H, CH2); 2.22 (d, 6H, CH2, JHH =11.0 2 Hz); 2.17 (d, 6H, CH2, JHH =12.1 Hz); 2.01 (bs, 6H, CH) 13 C NMR (C6D6, 100.6 MHz): 47.28 (CH); 40.22 (CH2); 30.68 (CH2)

** We were unable to acquire the peaks corresponding to the bipyridine ligand by 13C NMR although they were fully visible by 1H NMR. We attribute this to the presence of radical character localized on the bipy ligand.

*EA C30H38N2Zn (492.03): calc’d C 73.23, H 7.78, N 5.69; found C 69.90, H 7.39, N 5.32 (extremely air sensitive compound).

EI-TOF+ MS resulted in complete fragmentation to (1-Ad)2Zn and 2,2’-bipyridyl. -1 IR Absorptions: νmax/cm 2866 (CH2), 2848, 2829, and 2804 (CH), 1594 (Aromatic C=C), 1443

(CH2) MP 190 °C (decomp)

Table 4.1. Crystal data and structure refinement for 4.4a. ______

Empirical formula C30H38N2Zn Formula weight 491.99 Temperature 147(2) K Wavelength 1.54178 Å Crystal system Triclinic Space group P-1 Unit cell dimensions a = 10.7437(4) Å α= 86.964(2)°. b = 10.9461(5) Å β= 66.110(2)°. c = 11.8493(5) Å γ = 78.578(2)°. Volume 1248.30(9) Å3 Z 2 Density (calculated) 1.309 Mg/m3 Absorption coefficient 1.499 mm-1 F(000) 524 Crystal size 0.120 x 0.020 x 0.010 mm3 Theta range for data collection 4.082 to 67.389°. Index ranges -12<=h<=12, -13<=k<=13, -13<=l<=13 Reflections collected 20808 Independent reflections 4403 [R(int) = 0.0438] Completeness to theta = 67.389° 98.2 % Absorption correction Semi-empirical from equivalents Max. and min. transmission 0.7529 and 0.6541 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 4403 / 0 / 335 Goodness-of-fit on F2 1.041

Final R indices [I>2sigma(I)] R1 = 0.0318, wR2 = 0.0768

R indices (all data) R1 = 0.0398, wR2 = 0.0810 ______

4.5.9 Synthesis of 2,2’bipyridyl-bis(2-adamantyl)zinc [2-Ad2Zn(bipy)] (4.4b)

The procedure performed was analogous to that for 4a, above. The product was collected from a red solution as red crystals, which were suitable for X-ray analysis (Crystal data in Table 4.2). (Yield 93%)

1 3 3 H NMR (C6D6, 400 MHz): 8.47 (d, 2H, bipy, JHH =4.9 Hz); 7.62 (d, 2H, bipy, JHH =7.9 Hz); 3 6.97 (t, 2H, bipy, JHH =7.8 Hz); 6.57 (m, 2H, bipy) 2.00-2.30 (m, 28H, 2-Ad); 1.44 (bs, 2H, 2-Ad α-CH) 13 C NMR (C6D6, 100.6 MHz): 149.06 (bipy); 136.94 (bipy); 124.11 (bipy); 120.89 (bipy); 44.93 (2-Ad); 42.76 (2-Ad); 41.03 (2-Ad); 39.63 (2-Ad); 35.04 (2-Ad); 30.57 (2-Ad); 30.30 (2-Ad).

* EA C30H38N2Zn (492.03): calc’d C 73.23, H 7.78, N 5.69; found C 73.94, H 7.41, N 6.73

EI-TOF+ MS resulted in complete fragmentation to (2-Ad)2Zn and 2,2’-bipyridyl. -1 IR Absorptions: νmax/cm 2885 (CH2), 2829 (b, CH), 1594 (Aromatic C=C), 1437 (CH2) MP 180 °C (decomp)

Table 4.2. Crystal data and structure refinement for 4.4b. ______

Empirical formula C30H38N2Zn Formula weight 491.99 Temperature 147(2) K Wavelength 0.71073 Å Crystal system Triclinic Space group P-1 Unit cell dimensions a = 10.5819(10) Å α= 87.754(3)°. b = 11.0816(10) Å β= 66.311(3)°. c = 11.7087(11) Å γ= 78.315(3)°. Volume 1229.9(2) Å3 Z 2 Density (calculated) 1.329 Mg/m3 Absorption coefficient 1.019 mm-1 F(000) 524 Crystal size 0.150 x 0.150 x 0.050 mm3 Theta range for data collection 1.879 to 27.648°. Index ranges -13<=h<=13, -14<=k<=14, -15<=l<=15 Reflections collected 39635 Independent reflections 5701 [R(int) = 0.0387] Completeness to theta = 25.242° 100.0 % Absorption correction Semi-empirical from equivalents Max. and min. transmission 0.7456 and 0.7052 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 5701 / 0 / 298 Goodness-of-fit on F2 1.039

Final R indices [I>2sigma(I)] R1 = 0.0274, wR2 = 0.0587

R indices (all data) R1 = 0.0382, wR2 = 0.0625 ______

4.5.10 Synthesis of bis(1-adamantyl)mercury [1-Ad2Hg] (4.5a)

135 mg (0.5 mmol) of HgCl2 was slurried in 5 mL of toluene. 2 mL (1 mmol) of 1-AdZnBr (0.5 M in THF) was added slowly, with stirring. The mixture was stirred at room temperature for 24 h, and then the solvent was removed in vacuo. The residue was extracted with 10 mL of pentane and the volume was reduced in vacuo until the solution became cloudy. The solution was placed in a freezer at -40 °C and the product was isolated as a white powder. (Yield 184 mg, 78%)

1 2 H NMR (C6D6, 400 MHz): 2.09 (s, 6H, CH2); 1.89 (d, 3H, CH2, JHH =12.0 Hz); 1.87 (d, 3H, CH2, 2 JHH = 12.0 Hz); 1.68 (s, 3H, CH) 13 C NMR (C6D6, 100.6 MHz): 46.44 (CH); 38.95 (CH2); 31.60 (CH2)

* EA C20H30Hg (471.05): calc’d C 51.00, H 6.42; found C 56.18, H 7.21 (contaminated with adamantane)

HRMS (EI-TOF+): mass [M] calc’d for C20H30Hg 472.2054 Da, measured 472.2054 Da. -1 IR Absorptions: νmax/cm 2897 (CH2), 2845 (CH), 1447 (CH2) MP 250 °C (decomp)

4.5.11 Synthesis of bis(2-adamantyl)mercury [2-Ad2Hg] (4.5b):

The procedure performed was analogous to that for 5a, above. (Yield 80%)

1 H NMR (C6D6, 400 MHz): 1.80-2.10 (m, 30H, 2-Ad2Hg) 13 C NMR (C6D6, 100.6 MHz): 74.06 (CH); 42.77 (CH2); 40.89 (CH2); 38.66 (CH); 35.29 (CH2); 29.56 (CH); 29.25 (CH)

EA C20H30Hg (471.05): calc’d C 51.00, H 6.42; found C 51.25, H 6.38

HRMS (EI-TOF+): mass [M] calc’d for C20H30Hg 472.2054 Da, measured 472.2063 Da. -1 IR Absorptions: νmax/cm 2885 (CH2), 2841 (CH) MP 210 °C (decomp)

4.5.12 Synthesis of bis(2-adamantyl)bismuth(III) bromide [2-Ad2BiBr] (4.6b)

45 mg (0.1 mmol) of BiBr3 was slurried in 5 mL of diethyl ether and cooled to –40 °C. A solution of 34 mg (0.1 mmol) of 2-Ad2Zn in 5 mL of diethyl ether at –40 °C was added dropwise. The yellow solution gradually turned black and was allowed to come to room temperature with stirring and was stirred for 16h. The solution was filtered through a medium frit giving a clear yellow solution. The solvent was removed in vacuo to give an orange residue which was extracted with 10 mL of pentane. The pentane solution was filtered through a coarse frit and the solvent was removed in vacuo to give an orange powder. (Yield 40 mg, 72%) Crystals suitable for single crystal

XRD were obtained by slow evaporation of a pentane solution of 2-Ad2BiBr (data contained in Table 4.3).

1 H NMR (C6D6, 400 MHz): 3 2 4.22 (t, 1H, Bi-CH, JHH = 13.3 Hz); 3.10 (s, 1H, CH); 2.36 (d, 1H, CH2, JHH = 12.3 Hz); 2.22 (d, 2 2 1H, CH2, JHH = 12.2 Hz); 2.02 (d, 1H, CH2, JHH = 13.0 Hz); 1.98 (s, 2H, CH); 1.91 (s, 1H, CH); 2 1.77-1.56 (m, 6H, CH2); 1.44 (d, 1H, CH2, JHH = 12.2 Hz); 13 C NMR (C6D6, 100.6 MHz):

81.28 (1C, Bi-C); 45.54 (1C, CH2); 42.64 (1C, CH); 38.59 (1C, CH); 37.95 (1C, CH2); 37.44 (1C,

CH); 37.19 (1C, CH2); 33.62 (1C, CH); 28.51 (1C, CH2); 27.80 (1C, CH2)

EA C20H30BiBr (559.34): calc’d C 42.95, H 5.41; found C 43.32, H 5.49 + HRMS (EI-TOF ): mass [M] calc’d for C20H30BiBr 558.1335 Da, measured 558.1328 Da. -1 IR Absorptions: νmax/cm 2897 (CH2), 2842 (CH), 1446 (CH2) MP 120 °C (decomp)

100

Table 4.3. Crystal data and structure refinement for 4.6b. ______

Empirical formula C20H30BiBr Formula weight 559.33 Temperature 147(2) K Wavelength 0.71073 Å Crystal system Monoclinic Space group C2/c Unit cell dimensions a = 25.362(2) Å α = 90°. b = 6.4647(5) Å β = 93.377(3)°. c = 21.7366(16) Å γ = 90°. Volume 3557.7(5) Å3 Z 8 Density (calculated) 2.089 Mg/m3 Absorption coefficient 12.154 mm-1 F(000) 2144 Crystal size 0.100 x 0.015 x 0.015 mm3 Theta range for data collection 1.609 to 27.608°. Index ranges -32<=h<=32, -8<=k<=8, -23<=l<=28 Reflections collected 22873 Independent reflections 4112 [R(int) = 0.0617] Completeness to theta = 25.242° 100.0 % Absorption correction Semi-empirical from equivalents Max. and min. transmission 0.7456 and 0.4350 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 4112 / 0 / 199 Goodness-of-fit on F2 1.044

Final R indices [I>2sigma(I)] R1 = 0.0284, wR2 = 0.0543

R indices (all data) R1 = 0.0456, wR2 = 0.0588 ______

101

4.5.13 Attempted Synthesis of bis(1-adamantyl)bismuth(III) bromide [1- Ad2BiBr] (4.6a)

Under analogous reaction conditions as used for the synthesis of 4.6b but using 1-Ad2Zn, no 4.6a could be isolated. 1H and 13C NMR of the reaction contained a complex mixture of decomposition products, including large quantities of adamantane.

4.5.14 Generation of 1-adamantylgold(triphenylphosphine) [1- AdAu(PPh3)] (4.7a)

100 mg (0.20 mmol) of ClAu(PPh3) was slurried in 5 mL of THF. 0.500 mL (0.2mmol) of 1- AdMgBr (0.4 M in THF) was added slowly, with stirring. The mixture was stirred at room temperature for 24 h, and then the solvent was removed in vacuo. The residue was extracted with 10 mL of benzene and filtered through a coarse frit. The solvent was removed in vacuo to yield the product as an off white oil. Product decomposed rapidly and could not be isolated.

1 H NMR (C6D6, 400 MHz): 7.46 (broad s, 6H, PPh3); 7.32 (broad s, 8H, PPh3); 6.94 (broad s, 21H,

PPh3); 3.00 (s, 6H, CH2); 2.31 (d, 3H, CH2); 2.11 (s, 3H, CH); 2.01 (d, 3H, CH2). 31 * P NMR (C6D6, 161.98 MHz): 44.00 (PPh3); 40.00 (PPh3)

* The major phosphorus peak at 40 ppm corresponds to the major 1-AdAu(PPh3) species. A secondary peak from the decomposition appears at 44 ppm. This corresponds with the presence of 1 various triphenylphosphine species in the H NMR where broad singlets of the PPh3 peaks are observed with higher than expected integrations.

* EA C28H30AuP (594.49): calc’d C 56.57, H 5.09; found C 52.88, H 5.37

** EA Sample contaminated with MgBr2(THF)n which could not be sufficiently removed.

102

4.5.15 Synthesis of 2-adamantylgold(triphenylphosphine) [2-AdAu(PPh3)] (4.7b)

100 mg (0.20 mmol) of ClAu(PPh3) was slurried in 5 mL of THF. 0.670 mL (0.2mmol) of 2- AdMgBr (0.3 M in THF) was added slowly, with stirring. The mixture was stirred at room temperature for 24 h, and then the solvent was removed in vacuo. The residue was extracted with 10 mL of benzene and filtered through a coarse frit. The solvent was removed in vacuo to yield the product as an off-white oil.

Alternatively; 50 mg (0.10 mmol) of ClAu(PPh3) was slurried in 5 mL of Toluene. 17 mg (0.05 mmol) of 2-Ad2Zn was added slowly, with stirring. The mixture was stirred at room temperature for 2 h, and then the product was filtered through a medium frit. Solvent was removed in vacuo to yield the product as a light purple solid (minor amounts of gold metal). (Yield 62%). This complex could similarly be synthesized using 2-Ad2Zn(bipy).

1 3 3 H NMR (C6D6, 400 MHz): 7.48 (m, 6H, PPh3); 6.95 (m, 9H, PPh3); 3.17 (dt, JHP = 8.55 Hz, JHH 2 = 4.85 Hz, 1H, Au-CH); 2.99 (d, JHH = 11.44, 2H, CH2); 2.88 (s, 2H, CH); 2.07 (m, 10H, adamantyl) 31 P NMR (C6D6, 161.98 MHz): 43.44 (PPh3) 13 2 1 C NMR (C6D6, 100.6 MHz): 134.04 (d, JCP = 13.74 Hz , PPh3); 132.06, (d, JCP = 42.96 Hz , 4 3 2 PPh3); 130.17 (d, JCP = 1.72 Hz , PPh3); 128.57 (d, JCP = 10.19 Hz , PPh3); 63.19 (d, JCP = 97.24 Hz , Au-C); 42.72; 42.64; 39.11; 37.86; 37.82; 29.90; 29.66.

* EA C28H30AuP (594.49): calc’d C 56.57, H 5.09; found C 55.52, H 4.84, N 0.46 ** EA sample contaminated with a small amount (~2.5%) of 2,2’-bipyridyl

HRMS (EI-TOF+): mass [M] calc’d for C28H30AuP 594.1751 Da, measured 594.1734 Da. -1 IR Absorptions: νmax/cm 2885 (CH2), 2831 (b, CH), 1464 (Aromatic C=C), 1434 (CH2) MP 145 °C (decomp)

103

4.5.16 Generation of 1-adamantylgold(tricyclohexylphosphine) [1- AdAu(PCy3)] (4.8a)

50 mg (0.10 mmol) of ClAu(PCy3) was slurried in 5 mL of benzene. 17 mg (0.05 mmol) of 1-

Ad2Zn was added slowly, with stirring. The mixture was stirred at room temperature for 2 h, and then the product was filtered through a medium frit. Solvent was removed in vacuo to yield the product as a dark purple solid. The resulting compound was not isolable. NMR data indicates the presence of 1-AdAu(PCy3), but the product cannot be fully characterized. 1 H NMR (C6D6, 400 MHz): 2.92 (broad s, 6H, CH2); 2.31-0.86 (m, cyclohexyl and adamantyl);

1.87 (d, 3H, CH2); 1.68 (s, 3H, CH) 31 P NMR (C6D6, 161.98 MHz): 63.08; 55.02. **** The 31P NMR spectrum indicates that two products are present, not one.

4.5.17 Synthesis of 2-adamantylgold(tricyclohexylphosphine) [2- AdAu(PCy3)] (4.8b)

50 mg (0.10 mmol) of ClAu(PCy3) was slurried in 5 mL of benzene. 17 mg (0.05 mmol) of 2-

Ad2Zn was added slowly, with stirring. The mixture was stirred at room temperature for 2 h, and then the product was filtered through a medium frit. Solvent was removed in vacuo to yield the product as a dark purple solid likely due to presence of gold metal. (Yield 67%) Crystals suitable for single crystal XRD were obtained by slow evaporation of a benzene solution of 2-AdAu(PCy3). 1 H NMR (C6D6, 400 MHz): 2.97-1.04 (multiple signals that cannot be assigned, adamantyl and cyclohexyl); 1.89 (d, 3H, CH2); 1.87 (d, 3H, CH2); 1.68 (s, 3H, CH). 13 C NMR (C6D6, 100.6 MHz): 70.05; 64.40 (d, JCP = 94.69 Hz); 43.62 (d, JCP = 7.61 Hz); 42.11; 2 39.78; 38.40; 33.58 (d, JCP = 20.12 Hz); 30.81; 30.46 (d, JCP = 15.04 Hz); 27.45 (d, JCP = 10.80 Hz); 26.42; 25.07. 31 P NMR (C6D6, 161.98 MHz): 57.61 (PCy3)

HRMS (EI-TOF+): mass [M] calc’d for C28H48AuP 612.3159 Da, measured 612.3146 Da. -1 IR Absorptions: νmax/cm 2891 (CH2), 2831 (b, CH), 1444 (CH2) MP 165 °C (decomp)

104

4.6 References for Chapter 4

70 (a) Spasov, A. A.; Khamidova, T. V.; Bugaeva, L. I.; Morozov. I. S. Pharma. Chem. J. 2000, 34, 1. (b) Van der Schyf, C. J.; Geldenhuys, W. J. Neurotherapeutics 2009, 6, 175. (c) Wanka, L.; Iqbal, K.; Schreiner, P. R. Chem. Rev. 2013, 113, 3516. (d) Liu, J.; Obando, D.; Liao, V.; Lifa, T.; Codd. R. Eur. J. Med. Chem. 2011, 46, 1949.

71 (a) Clay, W. A.; Dahl, D. E. P.; Carlson, R. M. K.; Melosh, N. A.; Shen. Z.-X. Rep. Prog. Phys. 2015, 78, 016501. (b) Schwertfeger, H.; Fokin, A. A.; Schreiner, P. R. Angew. Chem. Int. Ed. 2008, 47, 1022. (c) Ranjbar, A.; Khazaei, M.; Venkataramanan, N. S.; Lee, H.; Kawazoe, Y. Phys. Rev. B. 2011, 83, 115401. (d) Muller, T.; Bräse, S. RSC Adv. 2014, 4, 6886. (e) Inomata, S.; Matsuoka, S.; Sakai, S.; Tajima, H.; Ishizone, T. Macromolecules 2012, 45, 4184.

72 (a) Hackett, M.; Whitesides, G. M. Organometallics 1987, 6, 403. (b) Aranyos, A.; Old, D. W.; Kiyomori, A.; Wolfe, J. P.; Sadighi, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1999, 121, 4369. (c) Chen, L.; Ren, P.; Carrow. B. P. J. Am. Chem. Soc. 2016, 138, 6392.

73 Zhou, Y.; Brittain, A. D.; Kong, D.; Xiao, M.; Meng, Y.; Sun. L. J. Mater. Chem. C. 2015, 3, 6947.

74 (a) Molle, G.; DuBois, J.-E.; Bauer, P. Tet. Lett. 1978, 34, 3177. (b) Wieringa, J. H.; Wynberg, H.; Strating, J. Synth. Comm. 1971, 1, 7. (c) Molle, G.; Bauer, P.; DuBois, J.-E. J. Org. Chem. 1983, 48, 2975. (d) Molle, G.; Bauer, P.; DuBois, J.-E. J. Org. Chem. 1982, 47, 4120. (a) Molle, G.; Bauer, P.; DuBois, J.-E. J. Org. Chem. 1982, 47, 4120. (b) Zhu, L.; Wehmeyer, R. M.; Rieke, R. D. J. Org. Chem. 1991, 56, 1445. (d) Rieke, R. D.; Hanson, M. V.; Brown, J. D. J. Org. Chem. 1996, 61, 2726. (e) Samann, C.; Dhayalan, V.; Schreiner, P. R.; Knochel, P. Org. Lett. 2014, 16, 2418.

75 Rieke, R. D.; Bales, S. E. J. Am. Chem. Soc. 1974, 96, 1775.

76 Molle, G.; Bauer, P.; Dubois, J. E. J. Org. Chem. 1982, 47, 4120.

105

77 Rieke, R. D.; Wu, T.-C.; Rieke, L. I. Org. Synth. 1995, 72, 147.

78 Sämann, C.; Dhayalan, V.; Schreiner, P. R.; Knochel, P. Org. Lett. 2014, 16, 2418.

79 (a) Wieringa, J. H.; Wynberg, H.; Strating, J. Syn. Comm. 1971, 1, 7. (b) Wieringa, J. H.; Strating, J.; Wynberg, H. Syn. Comm. 1972, 2, 191. (c) Wieringa, J. H.; Wynberg, H.; Strating, J. Tet. Lett. 1972, 20, 2071. (d) Molle, G.; Dubois, J. E.; Bauer, P. Tet. Lett. 1978, 34, 3177. (e) Molle, G.; Dubois, J. E.; Bauer, P. Syn. Comm. 1978, 8, 39. (f) Molle, G.; Bauer, P.; Dubois, J. E. J. Org. Chem. 1983, 48, 2975. (g) Molle, G.; Briand, S.; Bauer, P.; Dubois, J. E. Tetrahedron 1984, 40, 5113.

80 For supported metal 2-adamantyls, see: (a) Millard, M. D.; Moore, C. E.; Rheingold, A. L.; Figueroa, J. S. J. Am. Chem. Soc. 2010, 132, 8921. (b) Endo, K.; Grubbs, R. H. J. Am. Chem. Soc. 2011, 133, 8525. (c) Gerber, R.; Blacque, O.; French, C. M. ChemCatChem. 2009, 1, 393- 400. (d) Hartung, J.; Grubbs, R. H. J. Am. Chem. Soc. 2013, 135, 10183. (e) Keitz, B. K.; Endo, K.; Patel, P. R.; Herbert, M. B.; Grubbs, R. H. J. Am. Chem. Soc. 2012, 134, 693. (f) Gorman, T. W.; Gaunt, M. J.; Bond, A. D. CSD Communication 2016, refcode OJOKIJ (g) Herbert, M. B.; Suslick, B. A.; Liu, P.; Zou, L.; Dornan, P. K.; Houk, K. N.; Grubbs, R. H. Organometallics 2015, 34, 2858. (h) Diekmann, M.; Bockstiegel, G.; Lutzen, A.; Friedemann, M.; Saak, W.; Haase, D.; Beckhaus. R. Organometallics 2006, 25, 339. (i) Adler, C.; Bekurdts, A.; Haase, D.; Saak, W.; Schmidtmann, M.; Beckhaus, R. Eur. J. Inorg. Chem. 2014, 8, 1289. (j) Manen, M.; Lauterbach, N.; Dorfler, J.; Schmidtmann, M.; Saak, W.; Doye, S.; Beckhaus, R. Angew. Chem., Int. Ed. 2015, 54, 4383. (k) Scherer, A.; Haase, D.; Saak, W.; Beckhaus, R.; Meetsma, A.; Bouwkamp, M. W. Organometallics 2009, 28, 6969.

81 (a) For a recent report of an unsupported 2-adamantyl metal complex for which, however, no NMR or structural information could be obtained, see: Casitas, A.; Rees, J. A.; Goddard, R.; Bill, E.; DeBeer, S.; Fürstner, A. Angew. Chem. Int. Ed. 2017, 34, 10108. The 2-adamantyl iron complex reported was formed via transmetallation. (b) An unsupported 2-adamantyl chromium

106

complex has also been reported, however, no NMR or structural information could be obtained, see: Bochmann, M.; Wilkinson, G.; Young, G. B. J. Chem Soc., Dalton Trans. 1980, 10, 1879. The unsupported 2-adamantyl chromium complex was formed through transmetallation.

82 For unsupported metal 1-adamantyls, see: Bresciani-Pahor, N.; Marzilli, L. G.; Randaccio, L.; Toscano, P. J.; Zangrando, E. J. Chem. Soc., Chem. Comm. 1984, 1508. (b) Eckert, H.; Lenoir, D.; Ugi, I. J. Organometal. Chem. 1977, 141, C23. (c) Alaimo, P. J.; Arndsten, B. A.; Bergman, R. G. J. Am. Chem. Soc. 1997, 119, 5269. (d) Dreos-Garlatti, R.; Geremia, S.; Randaccio, L.; Ruffini, S.; Tauzher, G. J. Organometal. Chem. 1995, 487, C24. (e) Zuo, H.; Liu, Z.; Yang, W.; Zhou, Z.; Chan, K. S. Dalton Trans. 2015, 44, 20618.

83 (a) Kaim, W. Chem. Ber. 1981, 114, 3789. (b) Budzelaar, P. H. M. Eur. J. Inorg. Chem. 2012, 530. (c) Fedushkin, I. L.; Petrovskaya, T. V.; Girgsdies, F.; Nevodchikov, V. I.; Weimann, R.; Schumann, H.; Bochkarev, M. N. Russ. Chem. Bull. Int. Ed. 2000, 49, 1869.

84 (a) Thiele, K. H.; Rau, H.; Ehrhardt, U.; Wilcke, S. Z. Z. Anorg. Allg. Chem. 1969, 364, 270. (b) Thiele, K. H.; Dimitrov, V.; Thielemann, J. Z. anorg. Allg. Chem. 1981, 483, 145. (c) Dimitrov, V.; Thiele, K. H. Z. anorg. Allg. Chem. 1982, 494, 139. (d) Schenke, D.; Rehbaum, F.; Thiele, K. H. Z. anorg. Allg. Chem. 1988, 561, 73.

85 (a) Hasenzahl, S.; Kaim, W.; Stahl, T. Inorg. Chim. Acta 1994, 225, 23. (b) Noltes, J. G.; Boersma, J. J. Organomet. Chem. 1967, 9, 1.

86 (a) Noltes, J. G.; Van Den Hurk, J. W. G. J. Organomet. Chem. 1965, 3, 222. (b) Thiele, K. H.; Kӧhler, J. Z. anorg. Allg. Chem. 1965, 337, 260.

87 Wissing, E.; Kaupp, M.; Boersma, J.; Spek, A. L.; Van Koten, G. Organometallics 1994, 13, 2349.

88 Krahmer, J.; Beckhause, R.; Saak, W.; Haase, D. Z. anorg. Allg. Chem. 2008, 634, 1393.

107

89 (a) Jones, P. G.; Blaschette, A.; Henschel, D.; Weitze, A. Z. Kristallogr. Cryst. Mater. 1995, 210, 377. (b) Yamaguchi, S.; Shirasaka, T.; Tamao, K. Organometallics 2002, 21, 2555. (c) Schuster, O.; Schier, A.; Schmidbaur, H. Organometallics 2003, 22, 4079. (d) Whitmire, K. H.; Labahn, D.; Roesky, K. H.; Noltemeyer, M.; Sheldrick, G. M. J. Organomet. Chem. 1991, 402, 55. (e) Breunig, H. J.; Lork, E.; Rat, C. I.; Wagner, R. P. J. Organomet. Chem. 2007, 692, 3430.

90 Sämann, C.; Dhayalan, V.; Schreiner, P. R.; Knochel, P. Org. Lett. 2014, 16, 2418.

91 Aso, Y.; Yamashita, H.; Otsubo, T.; Ogura, F. J. Org. Chem., 1989, 54, 5627.

Chapter 5 Synthesis and Reactivity of Transition Metal Adamantyl Complexes

5.1 Abstract Investigations of the synthesis of transition metal adamantyl complexes were performed with the goal of achieving a γ-CH activation of a metallated 2-adamantyl. Preliminary computational work involving DFT calculations suggest that such reactivity could be thermodynamically feasible, but synthesis of metal adamantyls with the potential for C-H bond activation proved to be prohibitively difficult. Transmetallation of adamantyl anions onto group (III) metals using the available anion equivalents was entirely unsuccessful, and will likely require more potent transmetallating agents, such as the practically elusive adamantyl lithium. Reactions of diadamantylzinc precursors with a number of transition metals including Ta(V), Ru(II), W(VI), and Pt(II) resulted in successful transmetallations, but with rapid decomposition of the resulting complexes. The only isolable transition metal adamantyl complex with the potential for C-H bond activation is the newly synthesized (COD)Pt(2-Ad)Cl.

5.2 Introduction

The class of hydrocarbons referred to as diamondoids represents a unique class of compounds whose structures are closely related to that of diamond. The common element among these compounds is a cage-like polycyclic structure composed of exclusively sp3 carbon, terminated by hydrogen atoms. Because of the orientation of carbon atoms within these cages, idealized sp3 bond angles can be obtained without any distortion or ring strain, making diamondoid hydrocarbons anomalously stable. The simplest diamondoid hydrocarbon is adamantane (C10H16), the classical synthesis of which is shown in Figure 5.1 below.92

108 109

Figure 5.1. The famous “Schleyer synthesis” is now common route for the commercial synthesis of adamantane.

This reaction represents an unusual hydrocarbon rearrangement, which essentially proceeds with a Lewis acid catalyst and sufficient heat for a C10H16 fragment to rearrange to the thermodynamic minimum of adamantane. Higher polymantanes (diamantane, triamantane, etc., Figure 5.2) can be extracted from crude petroleum, and synthesized in modest to low yields.

Figure 5.2. Examples of 1-, 2-, and 3-dimensional “diamondoid” structures.

Because adamantane is so inert, reactions capable of functionalizing it are relatively limited. The two most common reactions from which further reactivity can be achieved are based are oxidation and bromination (see Schemes 5.1 and 5.2). Oxidation of adamantane by hot sulfuric acid results in the formation of adamantanone, which is the starting point for almost all 2-substituted 93 adamantyls (i.e. functionalized at the CH2 position).

110

Scheme 5.1. The standard procedure for oxidation of adamantane to adamantanone.

Bromination is selective for the 1- (i.e. CH) position. This reaction is carried out by heating adamantane in neat bromine, with the degree of bromination being controlled by which catalyst is used. Similar to adamantanone, 1-bromoadamantanes are the basis for a great deal of the further substitution of adamantane at the 1- position.94

Scheme 5.2. Conditions for 1-brominations of adamantane.

Applications of adamantane are wide-spread, ranging from polymers, to molecular electronics, to medicinal chemistry. The incorporation of adamantane appendages into polymers have been shown to improve many of their physical properties, including stiffness, thermal stability, and glass transition temperatures.95 Adamantane has been used to some extent as a tetrahedral building block in supramolecular structures such as metal-organic-frameworks, since a 1,3,5,7-tetrasubstituted adamantane behaves in many ways like a larger analogue to a carbon atom.96 The electronic effect of adamantane has been explored to some extent through their use in ligands for transition metals. The recently synthesized tri(1-adamantyl)phosphine has been determined to have σ-donation

111 strength on the order of NHC ligands, much higher than other phosphines thanks to the high degree of inductive donation from adamantane.97

A number of FDA approved drugs containing adamantane functional groups have been developed reaching as far back as the 1980’s, a few of which are shown in Figure 5.3. The adamantyl functional group has been referred to as the “lipophilic bullet,” with the two primary purposes of improving the lipophilicity and stability of existing pharmacophores.98

Figure 5.3. Examples of FDA approved adamantane based pharmaceuticals.

Inclusion of adamantane as an “add-on” to existing pharmaceuticals has proven to be a highly successful endeavor but is only a fraction of what is possible using the adamantane scaffold. While traditional organic chemistry can achieve a respectable number of adamantane derivatives, there remain many adamantane based molecules which simply do not exist. A class of compounds of particular interest to us are those that possess multiple bridge substituents on the same face of the cage in an all-axial arrangement, examples of which are shown in Figure 5.4.

112

Figure 5.4. Hypothetical classes of adamantane based compounds.

This may seem like a trivial substitution to perform and yet no examples of these compounds exist. To this end, the use of metal mediated C-H bond activation chemistry from 2-adamantyl metal complexes as a means to produce facially polysubstituted adamantanes was proposed (see Scheme 5.3).99

Scheme 5.3. Potential applications of a single-face substituted 2,2’,2’’-adamant-triyl.

The foundation of this proposal is the ability of transition metals to form metalacyclic complexes via C-H bond activation. In the case of adamantane, the activation event of interest would be a γ- CH activation, which has been observed in multiple different systems. The thermolysis of platinum bis(neopentyl) complexes is a classic example of γ-CH activation forming stable metalacycles, shown in Scheme 5.4.100

113

Scheme 5.4. γ-CH activation reactions in dialkylplatinum(II) complexes.

Similar reactivity has also been suggested in the decomposition of tetrakis(neopentyl) complexes of zirconium and hafnium (Scheme 5.5).101

Scheme 5.5. Suggested decomposition mechanisms of homoleptic group (IV) neopentyls.

The tendency of metals to form 4- and 5-membered metalacycles allows γ-CH activation to occur even in the presence of other accessible C-H bonds. The common requirement, however, for

114 successful γ-activation is the absence of β-hydrogen atoms, as β-hydride elimination is typically facile. While adamantane does contain β-hydrogens, there is a unique resilience to β-hydride elimination. Because of the rotational restriction imposed by the cage structure, the M-C-C-H unit in adamantane cannot become coplanar, presenting a kinetic barrier, while the product of β- elimination at adamantane (namely adamantene) contains a highly strained bridgehead olefin, which additionally provides a thermodynamic barrier. For these reasons, it is possible that activation of adamantane could occur preferentially at the γ- position.

5.3 Results and Discussion

Since γ-activation has been observed in both early and late transition metal systems, I initially investigated the thermodynamics of the activation of adamantane by different metals computationally. The general scheme which was calculated is shown in Scheme 5.6 below:

Scheme 5.6. Proposed reaction scheme for synthesizing di- and trianionic adamantyls. Metals with the potential for C-H bond activation via a sigma bond metathesis mechanism include group III and IV metals.

115

Figure 5.5. DFT (B3LYP-D3/SDD) optimized structure of La(2-Ad)3 showing agostic interactions.

The optimized geometry of tris(2-adamantyl)lanthanum(III) (La(2-Ad)3, Figure 5.5) exhibits an agostic interaction with the γ-CH of one of the bound adamantyls. This was encouraging, as this type of coordination could potentially mediate a γ-CH activation event. Since the C-H bonds of adamantane are so strong, pre-activation by an agostic interaction with a highly Lewis acidic metal could prove to be very useful in achieving a C-H bond activation reaction. Following this, the C- H bond activation reactions of the group III metals were computationally investigated. The general scheme for these reactions is shown in Scheme 5.7.

Scheme 5.7. Proposed method for γ-CH activation of group III tris(2-adamantyl) complexes.

The calculated thermodynamics for these reactions are given in Table 5.1.

116

Table 5.1. DFT (B3LYP-D3/SDD) calculated thermodynamics for γ-CH activation of group III tris(2-adamantyl) complexes.

Complex ΔEdiyl (kcal/mol) ΔEtriyl ΔGdiyl ΔGtriyl (kcal/mol) (kcal/mol) (kcal/mol)

Sc(2-Ad)3 +6.71 +8.58 -1.21 +3.13

Y(2-Ad)3 +9.33 +7.60 +0.66 +0.80

La(2-Ad)3 +9.14 +8.87 +0.33 -0.02

It was initially observed that C-H activation of group III metal 2-adamantyls from doubly and triply bound adamantyls were calculated to be thermodynamically unfavourable. On closer inspection, however, it can be seen that solvation energy can further improve the energetics of these reactions, as shown in Scheme 5.8 below, with calculated thermodynamic parameters included in Table 5.2.

Scheme 5.8. Potential use of solvation energy to improve the thermodynamics of activation.

117

Table 5.2. DFT (B3LYP-D3/SDD) calculated thermodynamics for γ-CH activation of group III tris(2-adamantyl) complexes containing explicit THF solvation in order to improve the favorability.

Complex ΔE’diyl ΔE’triyl ΔG’diyl ΔG’triyl (kcal/mol) (kcal/mol) (kcal/mol) (kcal/mol)

Sc(2-Ad)3(THF) -7.22 +0.25 -3.49 +0.79

Y(2-Ad)3(THF) -7.00 +5.71 -4.23 +5.88

La(2-Ad)3(THF) -5.40 -1.06 -3.85 +10.17

These figures suggest that group III metal 2-adamantyl complexes could, given the right conditions, be thermodynamically favourable to form adamant-di- and tri-yl complexes.

Due to their highly electropositive nature, group III alkyl complexes tend to be highly reactive and unstable. The synthesis of such complexes additionally tends to require very strong transmetallating agents, such as alkyl lithium or alkyl potassium reagents, as the kinetically labile and highly ionic M-C bonds being formed provide very poor transmetallation thermodynamics.102 In the absence of a reliable method for the synthesis of adamantyl lithium, attempts were made to synthesize group III adamantyl complexes from adamantyl magnesium bromide and diadamantyl zinc compounds. The general synthetic scheme for these attempted transmetallations is shown in Scheme 5.9.

Scheme 5.9. Generic synthetic scheme for group III metal adamantyl complexes.

Unfortunately, no conditions could be found with our existing adamantyl anion precursors that gave any transmetallation activity. Synthesis of these group III adamantyl complexes presumably requires a very hard anion source such as adamantyl lithium in order to be successful. Due to the

118 unreliable nature of synthesizing adamantyllithium as well as its very short lifetime in solution, this would be highly impractical.

Following this unsuccessful venture, we turned to later transition metals with stronger, more covalent metal-carbon bonds, but which still have the potential for C-H bond activation chemistry. This list includes many different metals, for which proposed reaction pathways are given in Scheme 5.10.

119

Scheme 5.10. Potential pathways to γ-CH activated adamantyls using metal complexes of group (V) and later.

Each of these proposed routes requires the synthesis of a metal complex containing two or more 2-Adamantyl ligands. There were two reasons for this design: 1) an anionic adamantyl, due to its massive amount of inductive donation and hyperconjugation, is more basic than practically any

120 other alkyl ligand, and 2) if the only alkyl substituents at the metal are 2-adamantyls, then the only possibility for a C-H activation event is at an adamantyl. Attempting to use practically any group other than an adamantyl as the base in a C-H bond activation reaction would likely be thermodynamically unfavourable, simply due to relative basicity. For these reasons, investigations were directed toward metals which are known to support multiple alkyl ligands. Among these metals, Tantalum(V) is one of the most well-studied.

103 104 Trimethyltantalum(V) dichloride (Ta(CH3)3Cl2) and pentamethyltantalum(V) (Ta(CH3)5) are both well-known complexes. Ta(CH3)3Cl2 was first synthesized in 1964 by addition of Me2Zn to

TaCl5 in pentane at -78 °C. This was a promising prospect, as (2-Ad)2Zn is now readily available to us. An analogous synthesis was attempted, as described in Scheme 5.11.

Scheme 5.11. Attempted syntheses of Ta(2-Ad)3Cl2, from which no metal adamantyl could be isolated.

1H NMR analysis of these reaction mixtures yielded exclusively adamantane, with no unreacted

(2-Ad)2Zn remaining. This suggests that the desired transmetallation reaction was successful, but the resulting complex was not sufficiently stable to persist long enough for analysis, even at low temperature.

γ-CH activation has been observed previously in ruthenium neopentyl complexes such as the one shown in Figure 5.6.105

121

Figure 5.6. Ruthenium alkyl complexes which have undergone γ-CH activation.

These cyclometallated complexes were synthesized by reaction of the (phosphine)(arene)ruthenium(II) dichloride complexes with two equivalents of neopentyl Grignard in pentane at room temperature. The intermediate bis(neopentyl) complexes could not be isolated as the cyclometallation reaction was facile. Based on these observations, an attempt was made to synthesize analogous 2-adamantyl Ru(II) complexes in hopes of observing a γ-CH activation event. The attempted synthesis is outlined in Figure 5.7.

6 Scheme 5.12. Attempted syntheses of Ru(η -C6Me6)(PMe3)(2-Ad)2.

6 The transmetallation precursor Ru(η -C6Me6)(PMe3)Cl2 was formed by addition of PMe3 to

6 [Ru(η -C6Me6)Cl2]2 in THF. This complex was treated with either (2-Ad)MgBr or (2-Ad)2Zn at low temperature. Addition of the adamantyl anion equivalent was typically followed by a colour change and solubilization of the poorly soluble dichloride complex. 1H NMR analysis revealed that adamantane had been formed as the major product, again suggesting that transmetallation had been achieved, but the resulting adamantyl complex was not sufficiently stable for characterization.

122

The results observed from the attempted syntheses of early transition metal, as well as Ta(V) and Ru(II) adamantyl complexes indicated that in order to isolate and characterize metal 2-adamantyls, it may be necessary to use metals which are particularly resilient to solvolysis or proton transfer pathways. To this end, tungsten(VI) was identified as a promising candidate for forming strong M-C bonds, as well as tolerating unusual bond angles since it is a d0 metal. The classical synthesis of hexamethyltungsten can be achieved using soft transmetallating agents 106 such as Al(CH3)3 or Zn(CH3)2, as shown in Scheme 5.13.

Scheme 5.13. Classical synthetic methods for hexamethyltungsten(VI).

Of particular interest to our work are the many examples of stable tungsten(VI) metalacyclobutane complexes, analogous to what could be observed in a 2,4-adamant-diyl, some examples of which are shown in Figure 5.7.107

123

Figure 5.7. Examples of isolable tungsten(VI) metalacyclobutane complexes.

Disappointingly, reaction of (2-Ad)2Zn with WCl6 (Figure 5.8) in hydrocarbon solvents did not produce an isolable tungsten adamantyl complex.

Figure 5.8. Attempted syntheses of WCl4(2-Ad)2, from which no metal adamantyl could be isolated.

Unlike previous attempts with other metals, attempted transmetallation of 2-adamantyls onto WCl6 did not produce stoichiometric adamantane. While adamantane was produced, the bulk of the crude products were an insoluble material which could not be characterized by NMR. Elemental analysis

124 also did not indicate any obvious stoichiometry. EA results for the product were: 24.59 %C, 1.82 %H and 5.44 %N.

Table 5.3. Comparison of experimental EA to some expected stoichiometries for a tungsten adamantyl. Compound %C %H %N Experimental 24.59 1.82 5.44 W(2-Ad)Cl5 24.20 3.05 0 W(2-Ad)Cl4 26.06 3.28 0 W(2-Ad)2Cl4 40.30 5.07 0 W(2-Ad)(o-tolyl)Cl4 36.99 4.02 0 W(2-Ad)(MeCN)Cl5 26.82 3.38 2.61 W(2-Ad)(MeCN)2Cl5 29.07 3.66 4.84 W(2-Ad)2(MeCN)Cl4 41.47 5.22 2.20 W(2-Ad)2(MeCN)2Cl4 42.50 5.35 4.13

Curiously, addition of (1-Ad)2Zn to WCl6 in toluene produced a product that was stable enough to be characterized by 1H NMR. From the spectrum, shown in Figure 5.9, it appears that the reaction solvent (toluene) may have been incorporated into the complex via a C-H bond activation, with the proposed overall reaction scheme given in Scheme 5.14.

Scheme 5.14. Proposed reaction scheme based on 1H NMR of the observed products.

125

1 Figure 5.9. H NMR (C6D6, 400MHz) spectrum of the crude reaction mixture of (1-Ad)2Zn and

WCl6 in toluene with proposed assignment.

Unfortunately, the structural assignment of this product could not be confirmed by elemental analysis or single crystal X-ray diffraction as it was isolated as a green oil, which decayed at ambient temperature to free adamantane.

The synthesis of metal adamantyl complexes via direct transmetallation was a difficult endeavor, but when working towards isolating stable metal alkyl complexes, one simply cannot ignore platinum. An attempt was made by the Wilkinson group in 1980 to synthesize a 2-adamantyl complex of platinum, but their efforts were unsuccessful, stating that:

“Reaction of this Grignard reagent [Mg(2-ad)Br] with metal halides has led to 108 disappointing results. With PtCl2(cod) only 2,2'-biadamantyl was recovered.”

126

We replicated these conditions by reacting freshly synthesized (2-Ad)MgBr with Pt(COD)Cl2, but rather than observing the formation of 2,2’-biadamantyl, we observed only adamantane. The same results were observed regardless of solvent or temperature conditions. Using (2-Ad)2Zn instead of the Grignard reagent produced the same results, forming only adamantane. Since no 2,2’- biadamantyl was observed, the latent platinum adamantyl species did not appear to be decomposing via reductive elimination. Each of these transmetallation attempts also resulted in black solutions, suggesting that the route of decomposition was more likely a homolytic Pt-C bond cleavage resulting in reduction of Pt(II) to Pt(0) and adamantane via H atom abstraction from solvent. The first major breakthrough in this synthesis came with the attempt to transmetallate from (2-Ad)2Zn onto Pt(COD)Cl2 in MeCN. The reaction, shown in Scheme 5.15 was attempted at NMR scale in MeCN-d3 in an attempt to observe any intermediate species.

Scheme 5.15. First successful synthesis of an isolable adamantyl platinum complex.

Rather than observing a latent reaction intermediate, it was instead discovered that both (2-Ad)2Zn 1 and Pt(COD)Cl2 have poor solubility in MeCN (far too low to observe any H NMR signal). Upon sitting for several days, however, large brown crystals were observed forming in the NMR tube. Single crystal X-ray diffraction of the crystals confirmed that the transmetallation was indeed successful and the reaction had produced the first ever stable adamantyl platinum complex, shown in Figure 5.10.

127

Figure 5.10. Anisotropic displacement plot (50% probability ellipsoids) for Pt(COD)(2-Ad)Cl. Selected bond distances (Å) and angles (°): Pt1–C9, 2.115(3); Pt1–Cl1, 2.3313(8); Pt1–C1, 2.145(3); Pt1–C2, 2.144(3); Pt1–C5, 2.379(3); Pt1–C6, 2.305(3); Cl1–Pt1–C9, 93.44(9).

In this instance, it is possible that due to the poor solubility of both reactants (as well as the product), the transmetallation reaction is sufficiently slow to allow the product to remain stable. Curiously, once this complex is isolated it is highly thermally stable under inert atmosphere.

Heating an NMR sample to 110 °C in toluene-d8 for several days did not result in any appreciable degree of decomposition. This observation relates to the difficulties of synthesizing adamantyl anion equivalents such as Grignard reagents, which are highly stable once synthesized, but tend to produce large quantities of reduced products. It would seem that many of these reaction intermediates are significantly less stable than their products and may relate to the stability or lifetime of adamantyl radical species.

Having observed the remarkable stability of Pt(COD)(2-Ad)Cl in benzene and toluene, when transmetallation reactions of adamantyl ligands onto Pt(COD)Cl2 in the same solvents resulted in complete decomposition, this suggested that the general lack of success in synthesizing metal adamantyl complexes is strongly related to the transmetallation event itself. It is suggested that the observed conditions in which Pt(COD)(2-Ad)Cl can be successfully isolated, namely using minimal MeCN solvent are relatively unique. Both reactants in this synthesis as well as the product have very poor solubility in MeCN, which could result in a slow enough rate of reaction, as well

128 as a minimal build-up of any unstable reaction intermediates from which decomposition may occur.

5.4 Conclusion

Methods for the production of currently unknown functionalizations of adamantane via transition metal mediated C-H bond activation appear to be computationally promising, but synthetically difficult. Thermodynamic calculations suggest that formation of 2,4-adamant-diyl and 2,4,6- adamant-triyl complexes of Group 3 metals could be possible, but the synthesis of such starting complexes via transmetallation has yet to be achieved. Stronger transmetallating agents such as adamantyllithium may be necessary for synthesis of early metal adamantyls, but reliable routes to these precursors remain elusive. A great deal of success has been found in the use of our newly synthesized diadamantylzinc compounds onto later transition metals, however the decomposition of reaction intermediates remains problematic, even at low temperatures. So far, the only successfully synthesized transition metal adamantyl complex with the potential for C-H bond activation activity is (COD)Pt(2-Ad)Cl.

129

5.5 Experimental

5.5.1 General Specifications: All manipulations were carried out under an inert (Ar) atmosphere using standard glove box (M Braun UniLab) and Schlenk techniques except where noted, and using oven-dried glassware. All reagents for which synthesis is not given were commercially available from Sigma Aldrich, Alfa Aesar, or STREM, and were used as received without further purification. Solvents were purified using by vacuum distillation from purple sodium benzophenone ketyl or activated molecular sieves. NMR solvents were obtained from Cambridge Isotope Laboratories. NMR data was obtained on a Bruker Avance III 400 MHz spectrometer. Chemical shifts are reported in ppm and referenced to residual solvent resonance peaks. Adamantane was present as a common contaminant in NMR spectra and was difficult to remove completely. Chemical shifts for adamantane are:

1 H NMR (C6D6, 400 MHz): 1.86 (bs, 4H, CH); 1.74 (bs, 12H, CH2) 13 C NMR (C6D6, 100.6 MHz): 38.02 (CH2); 28.76 (CH)

5.5.2 Transmetallation of (1-Ad)2Zn onto WCl6.

In a 20 mL scintillation vial, 40 mg (0.1 mmol) of WCl6 was dissolved in 7 mL of toluene, giving a deep blue solution, which was cooled to -60 °C. A cold solution 34 mg (0.1 mmol) of (1-Ad)2Zn in 7 mL of toluene was added with vigorous stirring. The colour changed rapidly from dark blue to deep red. Removal of the solvent in vacuo produced an oily green residue which was analyzed by NMR.

1 H NMR (C6D6, 400 MHz): 7.25 (d, 2H); 7.10 (d, 2H); 2.20 (s, 3H); 1.98 (bs, 3H); 1.88 (bs, 6H); 3 1.69 (bs, 6H, satellite peaks with proposed JW-H = 30.3 Hz)

130

5.5.3 Synthesis of 1,5-cyclooctadieneplatinum(II)(2-adamantyl)chloride [(COD)Pt(2-Ad)Cl]

In a J. Young NMR tube, 3.7 mg (0.01 mmol) of Pt(COD)Cl2 and 1.7 mg (0.005 mmol) of (2-

Ad)2Zn were suspended in 500 μL of CD3CN. The tube was shaken vigorously for 5 minutes and then allowed to stand at room temperature for 7 days. Large brown crystals suitable for X-ray analysis could be observed growing from the insoluble white powder at the base of the tube (Crystal data included in Table 5.3). The crystals could also be isolated and analyzed by NMR (Figure 5.11).

This synthesis could successfully be scaled up to several hundred mg without complication, as was discovered by Fioralba Taullaj. A slurry of 374 mg (1 mmol) of Pt(COD)Cl2 and 167 mg (0.5 mmol) of (2-Ad)2Zn in 5 mL of MeCN yielded the same product as a white powder after stirring for 3 days, in an isolated yield of 380 mg (80%), with impurities and side-products remaining dissolved in the supernatant, which could be removed by decanting. It was discovered that the best yields and purities were obtained when using minimal MeCN solvent, as some reduction to Pt0 could be observed when using larger solvent volumes.

1 2 2 H NMR (C6D6, 400 MHz): 5.59 (m, 2H, COD-CH, JPt-H = 25.5 Hz); 3.85 (m, 2H, COD C-H, JPt-

2 H = 74.2 Hz); 3.03 (bs, 1H, 2-Ad α-CH, JPt-H = 99.5 Hz); 2.55 (bd, 2H, Ad); 2.54 (bs, 2H, Ad);

2.11 (bs, 1H, Ad); 1.98-1.85 (m, 9H, Ad); 1.77 (m, 2H, COD-CH2); 1.61 (m, 2H, COD-CH2);

1.49-1.17 (m, 4H, COD-CH2).

131

1 Figure 5.11. H NMR Spectrum of (COD)Pt(2-Ad)Cl in C6D6 at 25 °C.

132

Table 5.4. Crystal data and structure refinement for (COD)Pt(2-Ad)Cl. ______

Empirical formula C18H27ClPt Formula weight 473.93 Temperature 150(2) K Wavelength 0.71073 Å Crystal system Monoclinic

Space group P21/c Unit cell dimensions a = 6.4868(3) Å α = 90°

b = 13.0312(6) Å β = 90.875(2)°

c = 18.3177(9) Å γ = 90°

Volume 1548.23(13) Å3 Z 4 Density (calculated) 2.033 Mg/m3 Absorption coefficient 9.224 mm-1 F(000) 920 Crystal size 0.150 x 0.100 x 0.050 mm3 Theta range for data collection 1.918 to 27.588°. Index ranges -8<=h<=8, -16<=k<=16, -22<=l<=23 Reflections collected 20940 Independent reflections 3574 [R(int) = 0.0388] Completeness to theta = 25.242° 100.0 % Absorption correction Semi-empirical from equivalents Max. and min. transmission 0.6456 and 0.4660 Refinement method Full-matrix least-squares on F2 Data / restraints / parameters 3574 / 0 / 181 Goodness-of-fit on F2 1.061

Final R indices [I>2sigma(I)] R1 = 0.0189, wR2 = 0.0385

R indices (all data) R1 = 0.0245, wR2 = 0.0402

133

5.6 References for Chapter 5

92 (a) Schleyer, P. von R. J. Am. Chem. Soc. 1957, 79, 3292. (b) Schleyer, P. von R.; Donaldson, M. M.; Nicholas, R. D.; Cupas, C. Org. Synth. 1973, 5, 16. (c) Mansoori, G. A. Molecular building blocks for nanotechnology: from diamondoids to nanoscale materials and applications. Springer, 2007, 48.

93 Geluk, H. W. and Keizer, V. G. Org. Synth. 1988, 6, 48.

94 Fort, Raymond C. Jr.; Schleyer, Paul Von R. Chem. Rev. 1964, 64, 277.

95 (a) Archibald, T. G.; Malik, A. A.; Baum, K.; Unroe, M. R. Macromolecules 1991, 24, 5261. (b) Hsiao, S. H.; Li, C. T. Macromolecules 1998, 21, 7213. (c) Pixton, M. R.; Paul, D. R. Polymer 1995, 16, 3165.

96 (a) Schwertfeger, H.; Fokin, A. A.; Schreiner, P. R. Angew. Chem. Int. Ed. 2008, 47, 1022. (b) Gunawan, M. A.; Hierso, J-C.; Poinsot, D.; Fokin, A. A.; Fokina, N. A.; Tkachenko, B. A.; Schreiner, P. R. New J. Chem. 2014, 38, 28. (c) Muller, T.; Brase, S. RSC Advances 2014, 4, 6886. (d) Clay, W. A.; Dahl, J. E. P.; Carlson, R. M. K.; Melosh, N. A.; Shen, Z-X. Rep. Prog. Phys. 2015, 78, 016501.

97 Chen, L.; Ren, P.; Carrow, B. P. J. Am. Chem. Soc. 2016, 138, 6392.

98 Wanka, L.; Iqbal, K.; Schreiner, P. R. Chem. Rev. 2013, 113, 3516.

99 Fekl, U. ACS PRF Grant Proposal, 2014.

100 DiCosimo, R.; Moore, S. S.; Sowinski, A. F.; Whitesides, G. M. J. Am. Chem. Soc. 1982, 104, 124.

101 Wu, Y-D.; Peng, Z-H.; Chan, K. W. K.; Liu, X.; Luinman, A. A.; Xue, Z. Organometallics 1999, 18, 2081.

134

102 (a) Zimmerman, M.; Anwander, R. Chem. Rev. 2010, 110, 6194. (b) Edelmann, F. T. Coord. Chem. Rev. 2017, 338, 27. (c) Kenward, A. L.; Piers, W. E.; Parvez, M. Organometallics 2009, 28, 3012.

103 Juvinall, G. L. J. Am. Chem. Soc. 1964, 86, 4202.

104 Schrock, R. R.; Meakin, P. J. Am. Chem. Soc. 1974, 96, 5288.

105 Diversi, P.; Ingrosso, G.; Lucherini, A.; Marchetti, F.; Adovasio, V.; Nardelli, M. J. Chem. Soc. Dalton Trans. 1991, 2, 203.

106 (a) Shortland, A. J.; Wilkinson, G. J. Chem. Soc. Dalton Trans. 1973, 8, 872. (b) Galyer, A. L.; Wilkinson, G. J. Chem. Soc., Dalton Trans. 1976, 21, 2235. (c) Kleinhenz, S.; Pfenning, V.; Seppelt, K. Chem. Eur. J. 1998, 9, 1687.

107 (a) Flook, M. M.; Jiang, A. J.; Schrock, R. R.; Muller, P.; Hoveyda, A. H. J. Am. Chem. Soc. 2009, 131, 7962. (b) Peryshkov, D. V.; Schrock, R. R. Organometallics 2012, 31, 7278.

108 Bochmann, M.; Wilkinson, G.; Young, G. B. J. Chem. Soc. Dalton Trans. 1980, 10, 1879.

135

Chapter 6 General Conclusions and Future Directions

6.1 General Conclusions

It can be broadly concluded that the chemistry of transition metal alkyls still contains many challenges and, despite the impression one may take home from textbooks, is not a mature field. Other groups have demonstrated the challenges that exist and the progress that can be made with very electron-poor alkyls (such as perfluoroalkyls).109 I, on the other hand, have demonstrated how difficult synthetic procedures still are for very electron-rich alkyls, in particular the adamantyl group. Both for synthetically useful adamantyl anion equivalents and for transmetallation onto other transition metals, this thesis contains ample evidence for how difficult this field still is synthetically and also significant examples where progress was achieved and where previously inaccessible compounds became synthetically accessible. If the field of transition metal alkyls is defined to include transition-metal-mediated cross-couplings involving alkyls, it is very evident that the field still has many challenges, linked to choice of precursor, selectivity, and so on. As a noteworthy contribution of this thesis, the decay pathways for a dialkyl-aryldiazenido-palladium(IV) complex (the first one ever reported), were investigated, culminating in the observation that radical pathways, through aryldiazene radicals, can be energetically accessible and may dominate the reactivity. In the following, more detailed conclusions are given and future directions are outlined, in the order of the individual chapters of the thesis.

136

6.2 Palladium(IV) aryldiazenido complexes

In an effort to explore the potential for catalytic C-C bond coupling reactivity in a Pd(II)/Pd(IV) system, the first aryldiazenido complex of palladium(IV) was synthesized via oxidative addition of an aryldiazonium cation to K[Pd(Tp*)Me2]. It was demonstrated that the thermal decomposition of the Pd(IV) complex resulted in the production of aryl radicals which could undergo either homocoupling or H-atom abstraction, depending on which solvent was used. While this reactivity is interesting, it did not exhibit the desired catalytic coupling ability. The question that remains to be answered is whether this system could be modified to become catalytic. The two major issues that currently prohibit catalysis are: the absence of a reductant, and the thermodynamic stability of the Pd(Tp*)Me3 by-product. In its current state, the addition of a reductant to this system could theoretically enable the homocoupling of aryldiazonium cations to form substituted biphenyls in a catalytic fashion, were it not for the unavoidable formation of Pd(Tp*)Me3, which acts as a thermodynamic sink. Future directions for this work should focus on modifications to the starting

Pd(II) complex. The solution may be as simple as replacing methyl ligands on K[Pd(Tp*)Me2] with aryl ligands, and introducing a sacrificial reductant to the system to recycle the Pd(IV) formed by oxidative addition of aryldiazonium back to Pd(II). A proposed catalytic cycle is shown in Scheme 6.1. Issues with compatibility between aryldiazonium cations and a sacrificial reductant may be difficult to avoid, and could limit the utility of these complexes to synthetic, rather than catalytic cycles.

137

Scheme 6.1. Proposed catalytic cycle for C-C homocoupling of aryldiazonium cations to substituted biphenyls.

138

6.3 Tris(pyrazolyl)borane-based FLP

Two very rapid transformations were observed when hydridotris(3,5-dimethylpyrazolyl)borate + was treated with aryldiazonium cation. Abstraction of a hydride by one equivalent of ArN2 generates an intermediate tris(pyrazolyl)borane which was trapped by acetone in an FLP type + reaction to form compound 3.1. Addition of a second equivalent of ArN2 then abstracts a + dimethylpyrazolide from 3.1 to give the C2-chiral cation 3.2 as a racemic mixture. Compound 3.2+ will have little utility as there is no clear way beyond chiral chromatography to separate the enantiomers. Compound 3.1, on the other hand, represents an interesting class of ligands, which act as N-N bidentate donors, but also forcing a methyl group close to the axis of a bound metal. The computed structure in Figure 3.6 demonstrates this unusual interaction which could potentially be used to further modify the ligand. Binding of 3.1 to a transition metal such as platinum could potentially induce a C-H bond activation of the axial methyl group to produce a new C-N-N scorpionate ligand, as shown in Scheme 6.2.

Scheme 6.2. Proposed formation of a C-N-N scorpionate ligand from compound 3.1 via C-H bond activation by platinum.

139

Additional future work should focus on determining whether the in situ generated tris(pyrazolyl)borane is capable of effectively trapping substrates other than acetone. Attempts which have been made so far have been unsuccessful, but it remains possible that with a sufficient excess of substrate trapping could be achieved.

140

6.4 Adamantyl anion equivalents

Anionic equivalents of adamantane have a long history of being difficult and unreliable to synthesize. Complications include the formation of C-C homocoupled products and large amounts of adamantane, as well as salt impurities (Li, Mg, or Zn salts) which can be difficult to separate. 1- and 2-adamantylzinc bromide solutions have become commercially available but we have determined by titration that even the labelled concentrations of these solutions are unreliable. I have developed new reliable methods for the synthesis of 1- and 2-adamantyl Girgnard and diadamantylzinc compounds and shown that they have great utility as reagents. In particular, diadamantylzinc is a convenient, shelf stable precursor which can be isolated as a powder which is free of salt and solvent impurities and is soluble in hydrocarbon solvents. This introduces the possibility of performing transmetallation reactions in the absence of ethereal solvents, which in most cases will produce insoluble zinc halide salts as the by product which can be easily removed. This was demonstrated in the synthesis of the first adamantyl complexes of gold and bismuth.

Future work in the synthesis of adamantyl anion equivalents could be directed towards the synthesis of highly ionic compounds such as adamantyllithium, since the existing adamantylzinc and magnesium compounds appear to have little to no ability to transmetalate onto Group 3 or 4 metals. This could potentially be achieved by an analogous reductive transmetallation approach that was discussed in Chapter 4. What is important to note is that neither 1- or 2- adamantyllithium are thermally stable in solvents such as diethyl ether or THF, so all reactions must be done in hydrocarbon solvents. Two starting materials of interest here would be (2-

Ad)2BiBr and potentially 1- or 2-adamantylmercury halides. Both of these classes of compounds have good solubility in hydrocarbon and aromatic solvents, and have a halide bound to the metal through which reduction by lithium metal may be kinetically initiated.

141

6.5 Transition metal adamantyl complexes

Metal mediated C-H bond activation of adamantane has the potential to provide routes to the selective functionalization of adamantane in ways that are simply not possible using organic chemistry. Based on DFT calculations, the formation of metal 2,4-adamant-diyl and 2,4,6- adamant-triyl complexes should be thermodynamically feasible but will not be synthetically trivial. The most promising candidates for C-H bond activation are Group 3 metals, as well as platinum and iridium. A 2-adamantylplatinum complex was synthesized and could eventually lead to a platinum 2,4-adamant-diyl.

There are many future directions for this work, but they can be generally grouped into: development of methods for transmetallation of adamantyl ligands onto transition metals without decomposition, synthesis of adamantyl metal complexes with the potential for C-H bond activation, and functionalization of activated adamantyls. As was observed from the synthesis of Pt(COD)(2-Ad)Cl, transmetallation reactions of adamantyl ligands are non-trivial. Only under very specific conditions (MeCN solvent in the case of Pt), does the transmetallation reaction proceed without decomposition of the product. This suggests that ligand and solvents systems should exist for other metals under which transmetallation can be successful, but there is no obvious way to identify such conditions. The end goals which were proposed in Scheme 5.3 have yet to be achieved and remain the focus of future work on this project.

142

6.6 References for Chapter 6

109 (a) Manuel, T. A.; Stafford, S. L.; Stone, F. G. A. J. Am. Chem. Soc. 1961, 83, 249. (b) Taw, F. L.; Scott, B. L.; Kiplinger, J. L. J. Am. Chem. Soc. 2003, 125, 14712. (c) Lee, G. M.; Leung, A. S. C.; Harrison, D. J.; Korobkov, I.; Hughes, R. P.; Baker, R. T. Organometallics 2017, 36, 2853. (d) Giffin, K.A.; Pua, L.A.; Piotrkowski, S.; Gabidullin, B.M.; Korobkov, I.; Hughes, R.P.; Baker, R.T. J. Am. Chem. Soc. 2017, 139, 4075. (e) Leclerc, M. C.; Bayne, J. M.; Lee, G. M.; Gorelsky, S. I.; Vasiliu, M.; Korobkov, I.; Harrison, D. J.; Dixon, D. A.; Baker, R. T. J. Am. Chem. Soc. 2015, 137, 16064. (f) Giffin, K. A.; Korobkov, I.; Baker, R. T. Dalton Trans. 2015, 44, 19587.