0-0 Bond Formation and the Transition Metal Chemistry of P-Diketiminate, Siloxide, and Triamide Ligands

0-0 Bond Formation and the Transition Metal Chemistry of p-Diketiminate, Siloxide, and Triamide Ligands

MASSACHUSETTS INSTITUTE Michael Pesek Marshak OF TECHNOLOG Y

B. A. Cornell University | JUN 1 5 2112 Ithaca, NY 2006 LIBRARIES

SUBMITTED TO THE DEPARTMENT OF CHEMISTRY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTORATE OF PHILOSOPHY

AT THE ARCHIVES MASSACHUSETTS INSTITUTE OF TECHNOLOGY

JUNE 2012

Signature of Author...... - . ... Michael P. Marshak Department of Chemistry Ar 120, 2012

C ertified by...... Ihaniel G. Nocera The Henry Dreyfus Professor of Energy and Professor of Chemistry 7) Thesis Supervisor

A ccepted by ...... V...... Robert W. Field Chairman, Departmental Committee on Graduate Students This doctoral thesis has been examined by a Committee of the Department of Chemistry as follows:

Richard R. Schrock Frederick G. Keyes Professor of Chemistry Committee Chairman

~A/ 7

Daniel G. Nocer The Henry Dreyfus Professor of Energy andProfessor of Chemistry Thesis Supervisor

Christopher C. Cummins Professor of Chemistry

2 0-0 Bond Formation and the Transition Metal Chemistry of p-Diketiminate, Siloxide, and Triamide Ligands

Michael Pesek Marshak

Submitted to the Department of Chemistry on May 11, 2012 in Partial Fulfillment of the Requirements for the Degree of Doctorate of Philosophy in Chemistry

Abstract

The electrochemical splitting of water into hydrogen and oxygen has been proposed as an alternative means to store electrical energy. The limiting aspect of this reaction is the oxygen forming reaction, which can be catalyzed by transition metal species with varying degrees of efficiency. This thesis examines the characteristics of oxygen bonding that complicate the 0-0 bond formation reaction, and examines ligand platforms that can stabilize high valent metal oxo intermediates. Siloxide ligands were used to generate a series of 4-coordinate Cr, Mn, Fe, and Co complexes, but these could not support the corresponding high valent oxo species. Instead, a remarkably stable 4-coordinate Crv tetrasiloxide complex was isolated. Chelating triamide ligands were explored, which might generate pseudotetrahedral metal oxo species, but complexes of various metal ions could not be reliably isolated. Planar bis-p- diketiminate (NacNac) complexes were synthesized by reaction of acetonitrile with Fe and Co mesityl species. The Co complex undergoes a ligand centered oxidation event to yield the first structurally characterized NacNac radical cation. In contrast to known redox noninnocent ligand platforms, no significant changes in C-C or C-N bond lengths are observed by X-ray crystallography. DFT calculations and the electronic and structural characterization of the oxidized and reduced Co complex confirm these conclusions.

Thesis Supervisor: Daniel G. Nocera

Title: The Henry Dreyfus Professor of Energy and Professor of Chemistry

3 This thesis is dedicated to my parents Daniel and Joan, and to my brother Andrew, for the endless love and support they have given me throughout my lfe.

4 Table of Contents

A bstract...... 3

Table of C ontents...... 5

List of Figures...... 10

List of Tables...... 13

List of Schem es...... 15

Chapter 1: The Complexity of 0-0 Bond Formation ...... 16

1.1 Oxygen ...... 17

1.2 Biological and Synthetic W ater Oxidation ...... 19

1.2.1 Photosystem II...... 19

1.2.2 Synthetic W ater Oxidation Catalysts...... 19

1.3 The M echanism of 0-0 Bond Form ation ...... 21

1.3.1 a V ersus 7 Bonding in 02...... 21

1.3.2 "Sphixis"...... 21

1.3.3 Electrochemical Aspects and Electron Transfer...... 25

1.3.4 Therm odyam ics ...... 26

1.3.5 Role of the Catalyst ...... 27

1.4 Ligand Design and Scaffolds for Oxidizing Species ...... 32

1.5 References ...... 34

Chapter 2: Synthesis and Redox Activity of bis-p-Diketiminate Cobalt...... 38

5 2.1 Introduction...... 39

2.2 Results ...... 41

2.2.1 Reaction with Acetonitrile...... 41

2.2.2 Characterization of 1 ...... 42

2.2.3 Electrochem istry and Synthesis of 2 ...... 50

2.2.4 Synthesis of Fe(1-mesitylbutane-1,3-diimine) 2 (3) ...... 55

2.3 Discussion ...... 57

2.4 Conclusions ...... 61

2.5 Experim ental Section ...... 64

2.5.1 General Considerations...... 64

2.5.2 Physical M ethods...... 64

2.5.3 X-ray Crystallographic Details...... 65

2.5.4 Computational M ethods ...... 65

2.5.5 Preparation of Co(1-mesitylbutane-1,3-diimine) 2 (1)...... 66

2.5.6 Preparation of [Co(1 -mesitylbutane- 1,3-diimine) 2(THF)2] [PF 6] (2)...... 67

2.5.7 Preparation of Fe(1 -mesitylbutane- 1,3-diim ine) 2 (3)...... 67

2.5.8 X-ray Crystallographic Tables...... 68

2.5.9 Cartesian Coordinates for Optim ized Structures ...... 70

2.6 References ...... 74

Chapter 3: First Row Transition Metal Siloxide Chemistry of the "DTBMS" Ligand...... 79

3.1 Introduction...... 80

3.1.1 Electronic Considerations...... 80

6 3.1.2 G eom etric Considerations ...... 82

3.1.3 Synthetic Considerations...... 84

3.2 Titanium /V anadium Siloxides ...... 86

3.2.1 Attem pted Syntheses ...... 86

3.3 Chrom ium Siloxides...... 87

3.3.1 Synthesis of Chrom ium Siloxide Species...... 87

3.3.2 Redox Chem istry ...... 92

3.3.3 Spectroscopy...... 92

3.3.4 D FT Calculations of 3 ...... 95

3.3.5 D FT calculations of hypothetical (M e3 SiO)3 CrO ...... 96

3.3.6 Comparison to Solid State Cr4...... 97

3.4 M anganese Siloxides...... 99

3.4.1 Synthesis...... 99

3.5 Iron Siloxides...... 101

3.5.1 Synthesis and Characterization of Fe(II) Species...... 101

3.5.2 Synthesis and Characterization of Fe(III) Species ...... 104

3.6 Cobalt Siloxides ...... 107

3.6.1 Background...... 107

3.6.2 Synthesis...... 107

3.7 Concluding Rem arks...... 109

3.8 Experim ental Section ...... 110

3.8.1 General Synthetic Considerations ...... 110

7 3.8.2 Physical Methods...... 110

3.8.3 Crystallographic Procedures...... 111

3.8.4 Computational M ethods ...... 111

t 3.8.5 Preparation of [Cr(OSi Bu2M e)3]n (1) ...... 112

t 3.8.6 Preparation of Na(THF)Cr(OSi Bu2Me) 4 (2) ...... 112

3.8.7 Preparation of Cr(OSitBu2 Me) 4 (3) ...... 112

3.8.15 X-ray Crystallographic Tables...... 115

3.8.16 Cartesian Coordinates for Optimized Structures ...... 118

3.9 References ...... 121

Chapter 4: Tripodal Amine Ligands Towards the Stabilization of High-Valent Metal Oxo

Com plexes...... 126

4.1 Introduction...... 127

4.1.1 Tri(am inom ethyl)ethane (TAM E)...... 128

4.1.2 Triaminocyclohexane (TACH) ...... 129

4.1.3 Synthetic Strategies For Extension to Late Transition Metals ...... 129

4.2 Results ...... 131

4.3 Other TACH Derivatives ...... 136

4.4 Conclusions ...... 137

4.5 Experim ental Section ...... 138

4.5.1 Preparation of L1...... 138

4.5.2 Preparation of LiVO ...... 138

4.5.3 Preparation of [K(12-crown-4) 2][LiL1FeOH] ...... 138

8 4.5.4 X-ray Crystallographic Table ...... 140

4.6 References ...... 141

Acknow ledgem ents ...... 144

9 List of Figures

Figure 1.1 Latimer diagram illustrating the one, two, and four electron reduction potentials of 02 and related species at pH 7 (adapted from ref. 58). Potentials exceeding 0.815 are highlighted in red...... 25

Figure 1.2 Frost diagram for the oxidation of H2 0 to 02 at pH = 7, normalized to the 4e- pathway potential. (inset) The same Frost diagram without normalization.4...... 274

Figure 2.1 Infrared spectrum of 1 and 1-d12 showing the isotopic shift in the N-H vs. N-D stretch ing mode...... 42

Figure 2.2 (a) X-ray crystal structure of the two independent molecules in 1 with 50% probability ellipsoids. Hydrogen atoms were omitted for clarity (except for N-H). (b) X-ray crystal structure of la with 50% probability ellipsoids. Hydrogen atoms were omitted for clarity (except for N -H ) ...... 43

Figure 2.3 Solid State temperature dependence of the magnetic moment of 1 recorded on a SQU ID magnetom eter at 1000 G ...... 45

Figure 2.4 Relaxed potential energy surface scan of the N5-N2-N3-N4 dihedral angle of 1 in both low spin (S = 1/2) and high spin (S = 3/2) configurations. The intersection of these plots indicates that crossover occurs at a dihedral angle of approximately 57 degrees w ith a 6 kcal/mol energy barrier...... 46

Figure 2.5 EPR of 1 in toluene glass at 77 K. Simulation yields g-values of gi = 2.808, g2 = 2 .002, and g3 = 1.956...... 47

Figure 2.6 Singly occupied molecular orbital (SOMO) of 1 showing mixing of NacNac nonbonding n orbital with metal dyz, isovalue 0.03...... 48

Figure 2.7 Spin density plot of 1 showing unpaired electron localized on the metal based dyz orbital (99% ), isovalue 0.008...... 48

Figure 2.8 UV-Vis absorption spectrum of 1 in pentane...... 49

Figure 2.9 The dominant donor and acceptor molecular orbitals for the four prominent absorptions in 1, obtained from spin-unrestricted TD-DFT calculations...... 50

10 Figure 2.10 Cyclic voltammogram of 1, 1 mM in THF with 0.1 M Bu 4NPF6 supporting electrolyte at a scan rate of 100 mV/s and referenced vs. ferrocene. Oxidation is centered at - 434 m V ...... 5 1

Figure 2.11 X-ray crystal structure of 2 with 50% probability ellipsoids. An additional uncoordinated THF molecule and hydrogen atoms were omitted for clarity (except for N - H )...... 52

Figure 2.12 UV-Vis-NIR Absorption Spectrum of 2 in THF...... 53

Figure 2.13 The dominant donor and acceptor molecular orbitals for the two prominent absorptions in 2, obtained from spin-unrestricted TD-DFT calculations...... 54

Figure 2.14 Spin density plot of 2 showing spin density both on the metal and ligand, isovalue = 0 .0 0 8 ...... 55

Figure 2.15 X-ray crystal structure of one of the two independent molecules of 3 with 50% probability ellipsoids. Hydrogen atoms were omitted for clarity (except for N-H)...... 56

Figure 3.1 X-ray crystal structure of 2 with thermal ellipsoids shown at the 50% probability level...... 89

Figure 3.2 X-ray crystal structure of 3 with thermal ellipsoids shown at the 50% probability level...... 91

Figure 3.3 UV-Vis-NIR Absorption spectrum of 3...... 93

Figure 3.4 Infrared Spectrum of Cr(DTBMS) 4 (3)...... 95

Figure 3.5 Qualitative ligand field diagram for 3...... 96

Figure 3.6 Calculated spin density plot of (Me3SiO)3CrO ...... 97

Figure 3.7 X-ray crystal structure of 5 with thermal ellipsoids shown at the 50% probability level...... 100

Figure 3.8 X-ray crystal structure of 6 with thermal ellipsoids shown at the 50% probability level...... 102

Figure 3.9 X-ray crystal structure of 7 with thermal ellipsoids shown at the 50% probability level. Hydrogen atoms and 1.5 molecules of toluene were omitted for clarity...... 103

Figure 3.10 X-ray crystal structure of 8 with thermal ellipsoids shown at the 50% probability level. The tert-butyl methyl groups and hydrogen atoms (except for O-H) were om itted for clarity ...... 105

11 Figure 3.11 X-ray crystal structure of 10 with thermal ellipsoids shown at the 50% probability lev el...... 10 8

Figure 4.1 X-ray crystal structure of Li with thermal ellipsoids at the 50% probability level. The tosyl groups (except for a-carbon) and hydrogen atoms (except for 1,3,5 C-H, and N- H ) w ere rem oved for clarity...... 131

Figure 4.2 'H NMR Spectrum (CH 3CN) of L1 ...... 132

Figure 4.3 1H NMR Spectrum (toluene) of L1VO...... 133

Figure 4.4 X-ray crystal structure of [K(12-crown-4) 2][LiL1FeOH] with 50% probability

ellipsoids. The K(12-crown-4) 2, 2 molecules of THF, and hydrogen atoms were rem oved for clarity...... 134

12 List of Tables

Table 1.1 Homolytic Bond Dissociation Enthalpies of Oxygen Species.3...... 29

Table 2.1 Selected bond lengths (A) and angles (0) of 1...... 44

Table 2.2 Selected bond lengths (A) and angles (*) of la...... 44

Table 2.3 Selected bond lengths (A) and angles (*) of 2...... 52

Table 2.4 Selected bond lengths (A) and angles (0) of 3...... 56

Table 2.5 Change in Co-N, C-N, and C-C Bond Lengths in 1, la, and 2...... 60

Table 2.6 X-ray Crystallographic Table for 1, la, and 2...... 68

Table 2.7 X-ray Crystallographic Table for 3...... 69

Table 2.8 Cartesian coordinates for the DFT optimized geometry for the S = 1/2 model of 1 (E = - 1993.63267829 hartrees)...... 70

Table 2.9 Cartesian coordinates for the DFT optimized geometry for the S = 3/2 model of 1 (E = - 1993.63221780 hartrees) ...... 71

Table 2.10 Cartesian coordinates for the DFT optimized geometry for the S = 1 model of 2 (E = -2225.80065097 hartrees)...... 72

Table 3.1 Selected bond lengths (A) and angles (0) of (DTBMS)4CrNa(THF) (2)...... 89

Table 3.2 Selected bond lengths (A) and angles (0) of Cr(DTBMS) 4 (3)...... 91

Table 3.3 Electronic absorption data of Cr(DTBMS)4 (3)...... 93

Table 3.4 Selected bond lengths (A) and angles (0) of Na(1 5-crown-5)Mn(DTBMS) 3 (5)...... 100

Table 3.5 Selected bond lengths (A) and angles (0) of [Fe(DTBMS) 2]2-(THF) (6)...... 102

Table 3.6 Selected bond lengths (A) and angles (0) of Na(1 5-crown-5)Fe(DTBMS) 3 (7)...... 103

Table 3.7 Selected bond lengths (A) and angles (0) of [Na(DTBMS) 3FeOH]2 (8)...... 105

13 Table 3.8 Selected bond lengths (A) and angles (0) of Na(15-crown-5)Co(DTBMS) 3 (10)...... 108

Table 3.9 X-ray Crystallographic Table for 2, 3, and 5...... 115

Table 3.10 X-ray Crystallographic Table for 6, 7, and 8...... 116

Table 3.11 X -ray Crystallographic Table for 10...... 117

Table 3.12 Cartesian coordinates for the DFT optimized geometry for the model of 3

Cr(OSiMe3)4 (E = -2982.62944260 hartrees)...... 118

Table 3.13 Cartesian coordinates and atomic spin densities for the DFT optimized geometry of

(Me3 SiO)3 CrO (E = -2573.34690776 hartrees). Cr-O = 1.57819 A...... 119

Table 4.1 Selected bond lengths (A) and angles (*) of [K(12-crown-4) 2][LiL1FeOH]...... 134

Table 4.2 X-ray Crystallographic Table for L, and [K(12-crown-4) 2][LiL1 FeOH]...... 140

14 List of Schemes

Scheme 1.1 Possible 0-0 bond formation mechanisms and their respective bond enthalpies...... 30

Scheme 1.2 Ligand Field Diagrams for Felv-Oxo Species in Different Geometries 19'44 ...... 33

Scheme 2.1 Proposed Mechanism for the Formation of 1 (adapted from ref. 26 and 29)...... 42

Scheme 2.2 Qualitative MO diagram illustrating n interactions of NacNac ligands with planar Co in 1 ...... 59

Scheme 2.3 Highest Occupied Molecular Orbitals of NacNac, Salen, and Porphyrin...... 61

Scheme 3.1 Synthetic methods for metalation of siloxide ligands...... 85

15 Chapter 1: The Complexity of 0-0 Bond Formation

16 1.1 Oxygen

By weight, oxygen is the third most abundant element in the universe, the second most abundant element on Earth, and the first most abundant element in the human body.1- 3 Molecular oxygen (02) makes up roughly 21% of the earth's atmosphere, and nearly all of it was produced biologically from water though photosynthesis. The production of 02 from water requires the energy input of at least 350 kJ/mol (pH = 7),4 and photosynthetic organisms are able to capture energy harvested from the sun to carry out this transformation. The primary byproducts of 02 synthesis are 4 protons (He) and 4 electrons (e), which are used to make NADPH and ATP, and ultimately to transform CO2 into glucose. The result of this photosynthetic water splitting reaction is the fundamental chemical imbalance that allows for all higher forms of life to exist. The broad diversity of plants and animals seen today is a direct result of the solar energy captured by photosynthesis.

Chemotrophic bacteria are considered to be among the first organisms on earth, and use hydrogen, sulfur, ammonia, hydrogen sulfide, or reduced minerals as a source of energy to reduce carbon dioxide to organic material.'- 9 As a result, early life was limited by the rate and availability of reduced matter that came from oceanic vents and other places deep within the earth's core. The arrival of phototrophic bacteria allowed the utilization of energy outside of the planet for the first time. These bacteria utilize sunlight to drive redox reactions, the most important of which is photosynthesis.

Just like pre-photosynthetic chemotrophic bacteria, modem civilization is now facing an energy challenge that pits our demand for reducing equivalents (fossil fuels) found deep within the earth's crust with the rate in which they can be extracted. It has been suggested current energy demand already has or is about to exceed the exploitable photosynthetic capacity of the planet.10 As a result, a more efficient system to capture and store light energy in chemical bonds is essential for civilization to continue to progress.11 12

17 Despite being one of the very first and most fundamental chemical reactions to be utilized by nature, the synthesis of 02 from water is still not well understood. Chemists have been successful in forming bonds between nearly all elements, but rational, well-behaved syntheses of an 0-0 bonds are comparatively rare. Understanding the mechanistic details of this bond formation process could allow refinement of current 02 production strategies or suggest new, more efficient ways to catalyze this essential reaction.

18 1.2 Biological and Synthetic Water Oxidation

1.2.1 Photosystem I

Photosystem II is an enzyme responsible for production of oxygen from water. X-ray crystallographic studies have elucidated the structure of the enzyme at increasing levels of detail, and have provided the arrangement of many components of the active site, oxygen evolving complex

13 8 (0EC). '- The most recent structure reveals a CaMn 30 4 cube with a fourth Mn ion attached by an oxo bridge to form a 'chair' like structure.' 8 While the exact mechanism of 0-0 bond formation is still heavily debated, two favored mechanisms to form the 0-0 bond include that hydroxide coordinated to a calcium ion attacks a terminal Mn(V) oxo, and a terminal Mn(V) oxo coupling to a bridging oxo. 19-22 An alternate mechanism involves the coupling of two water molecules oriented like a quasi-Zundel ion

23 (H20-H-OH-Mn). In this case, the 4H* and 4e- are lost through simultaneous proton coupled electron transfer (PCET) to Mn oxo units and basic amino acid residues. Due to the nature of the active site that cycles through five different electronic states and the ambiguity concerning different protonation states of the species, the mechanism for 02 generation in the OEC continues to attract and inspire much attention across the scientific community. 24

1.2.2 Synthetic Water Oxidation Catalysts

Most metal oxides will electrochemically generate oxygen from water with various overpotentials.' 2 -" Ruthenium and iridium oxide are among the most active, however their cost and natural abundance limits their widespread use. 28 ,29 The Nocera lab has investigated various amorphous

30-32 cobalt and nickel oxides which can function over a wide pH range. While these systems are among the most promising for large scale production of 02, the solid amorphous nature of these catalysts makes the mechanistic processes pertaining to 0-0 bond formation a challenge to elucidate at the atomic level.33

19 Homogeneous systems could offer this level of mechanistic detail, particularly if catalytically relevant intermediates could be identified and characterized. Meyer's work using molecular ruthenium catalysts suggests that Ru(V) oxo species are generated just prior to 0-0 bond formation.34 , 5

Unfortunately the reactivity of these intermediates makes their complete characterization particularly challenging.

Although chemical reactions that involve net 0-0 bond formation are rare, many high valent metal oxides are known to decompose thermally to give 02. For example, KMnO 4 and Cr0 3 will decompose at temperatures over 200 *C to give 02 and reduced metal oxide species. 02 can also be

(caution!). 3 6,37 generated from KMnO 4 by the addition of acids such as BF 3 or concentrated sulfuric acid

There are few examples of chemical synthesis of a peroxide species that form as the main reaction

product. Inorganic peroxides such as S052 , S2082-, and C2 0 62- can be synthesized electrochemically, but these require extremely high potentials (> 20 V) and involve the intermediate formation of hydroxyl radicals (-OH).38,39 Milstein has reported the formation of 02 by photolysis of a ruthenium bis-hydroxide complex that likely involves the formation of hydrogen peroxide as an intermediate, but this reaction only gives 49% conversion with the formation of unidentified byproducts. 40 Aside from the examples above, peroxide containing species most commonly synthesized from substrates that already have intact 0-0 bonds, either by reduction of 02 or by transfer of a peroxide moiety from one molecule to another.

The current interest in, yet scarcity of reports of 0-0 a bond formation might suggest an intrinsic difference in this process from other chemical bond formation. Section 1.3 will explore this enigma in the context of electrochemical potentials and thermodynamic bond enthalpies. Nonetheless, from a basic chemical perspective, 0-0 bonding is unique because unlike most other chemical bonds, the 7c bond in 02 is much stronger than the a bond.4 1 This means that there is little energetic stabilization in forming the peroxide a bond as an intermediate during oxidation of water, and complexes that do oxidize water to peroxide will have more than enough energy to further oxidize peroxide to 02.

20 1.3 The Mechanism of 0-0 Bond Formation

1.3.1 a Versus rr Bonding in 02

A discussion of 0-0 bond formation should begin stating the implicit assumption that there is nothing so unique about oxygen bonding that makes the general theory of chemical bonding invalid. In this light one can take inspiration for the theory of the two electron bond, which describes this in the context of ionic, covalent, and resonance components.4 4 In terms of a bonding, the second row elements show an increase in bond strength from Li-Li to C-C, which can be explained by the decreasing atomic radius of the atoms. The trend then reverses and the N-N and 0-0 a bond strengths decrease before recovering slightly with F-F. This phenomenon is usually attributed to the repulsion effect of lone pair electrons on the latter three atoms. Shaik has explored this lone pair weakening effect within the context of 7c bonds to show that they are not impacted since they are geometrically orthogonal to each other.4' As a result, the homoatomic 7c bond strengths increase as expected from bond overlap arguments C(72.0) < N

(90.7) < 0 (91.1 kcal/mol). This 0-0 7r bond strength suggests that the a bond contributes very little to the overall 0=0 double bond (118.9 kcal/mol).

1.3.2 "Sphixis"

The breaking of a chemical bond usually follows either a homolytic or heterolytic process, which simply describes whether the two electrons that make up the bond are evenly split between the atoms

(homolysis) or if both end up on one atom to give charged ions (heterolysis). Although this elegantly describes most bond cleavage processes, there is no antonym to describe the analogous bond formation reactions. Since the suffix -lysis derives from the Greek word "lysimo" (loosening), it seems reasonable to use the Greek word "sphiximo" (tightening) for bond formation. Thus the word homosphixis means that each atom uses one electron for bond formation, while heterosphixis means that one atom uses two electrons to bond to an atom that uses none. These two terms can be used to describe bond formation

21 reactions very broadly, but they are at times more specific than terms already in use. For example, the dihydride of Vaska's complex, [Ir](H) 2 can undergo reductive elimination by homosphixis to give H2, while the methyl iodide complex [Ir](CH3)(I) undergoes heterosphixis to give CH3I ([Ir] =

(Ph3 P)2(CO)IrCl). While both reactions are described as reductive eliminations, these new terms can provide a simple way to describe the different reaction mechanisms.

The formation of bonds in organic chemistry usually follows arrow-pushing conventions

(heterosphixis), where atoms take on nucleophilic or electrophilic character. In addition, bonds can be formed by free radical processes that involve homosphixic bond formation. In this context, one might imagine 0-0 bonds to form in a similar fashion, and as a result two leading mechanistic proposals involve an acid/base mechanism where a nucleophilic oxide species attacks an electrophilic oxo

19 2 3 34 44 (heterosphixis) and radical coupling of two oxo species (homosphixis). , , ,

2- 0 1- 1-

0: 0 : 0-0: Heterosphixis

@0 00 lw 1- 1- 1- 1- : .' '. i: 0-0: Homosphixis

While acid/base chemistry is ubiquitous in organic reaction mechanisms, the requirement of an oxygen atom to behave as an electrophile is complicated by the fact that it is the second most electronegative element, next to fluorine. This means that the only way to formally oxidize an oxygen atom is if it bonds either to fluorine, another oxygen atom, or an extremely oxidizing metal. Furthermore, the heterosphixis mechanism requires the oxygen atom to be oxidized not by one, but two electrons which would put it in the 0 oxidation state for peroxide formation. This is the case for molecules such as HOF, which does in fact readily generate 02 at or above room temperature.

22 A number of mononuclear metal oxo species have been suggested to undergo nucleophilic attack

to form an intermediate peroxo species.4 5 - 7 For example, Akermark showed that a manganese corrole

complex could be oxidized with tBuOOH to give a (cor)MnvO which generates 02 upon addition of

aqueous Bu 4NOH. The proposed mechanism involves nucleophilic attack of the hydroxide onto the

(cor)MnvO to give a hydroperoxide intermediate, (cor)Mn'(OOH). This hydroperoxide species is then

oxidized by a second (cor)MnvO to give 02. If this heterosphixic mechanism is operable, it would suggest

that other nucleophiles in the system should also be capable of forming a (cor)Mnm'(0-Nuc) bond.

Instead, the complex is stable to nucleophilic attack from the acetonitrile and tBuO- present in the

reaction mixture. The fact that tBuOOH is heterolyzed to form the initial oxo species suggests that if a

Mnm(OOH) intermediate were to exist, it too should cleave to form (cor)MnO. Furthermore, the

(cor)MnO species is 5-coordinate, so if hydroxide were to attack the most electrophilic atom in

(cor)MnO, it would likely be the extremely oxidized Mnv center rather than the oxo. An alternative to the

proposed mechanism might involve homosphixis, where two (cor)MnvO react with H 20 to form

(cor)MnI"(OH) and (cor)Mn"(OOH). This mechanism is consistent with kinetic data that indicate that the

reaction rate is accelerated by tethering two (cor)Mn species together.4 8

Mononuclear ruthenium polypyridine catalysts are another example where nucleophilic attack on

23,35,49 an oxo has been invoked as a mechanism for 0-0 bond formation. Most notably, Meyer has

2 proposed a mechanism for water oxidation at pH = 1 using [Ru(tpy)(bpm)(OH 2)] + (tpy = 2,2';6',2"-

terpyridine, bpm = 2,2'-bipyrimidine), measured the kinetics of most mechanistic steps, and characterized

some intermediate species by electrochemical and spectroscopic methods.3" The rate determining step

2 under catalytic conditions was found to be loss of 02 from [Ru(tpy)(bpm)0 2] +, while the 0-0 bond

formation rate was found to be roughly an order of magnitude faster. Calculations performed on the

system also suggest that nucleophilic attack of water on a Ru(V) oxo is the operative mechansim, though the potential that this is calculated to occur is significantly higher than the experimental value.49 On the other hand, Sakai et al. studied a series of similar ruthenium catalysts and proposed a radical coupling mechanism for 0-0 bond formation.5 0 The report showed a substantial decrease in rate when

23 4 [Ce(S0 4 )4 ] ~ was used as an oxidant rather than [Ce(N0 3)6]2-. Since cerium nitrate species form a hydroxide complex 51 in solution, broken-symmetry calculations were performed on [Ce(N0 3 )s(OH)]2-, which revealed that its electronic structure can be described as a singlet biradical with substantial spin density on the oxygen atom of the hydroxide. As a result, hydroxyl radical transfer from Ce to a Ru(V) oxo species was suggested as the 0-0 bond formation pathway. Although the different Ru catalysts used by Sakai and Meyer could operate by different mechanisms, the possibility that [Ce(N0 3)5(OH)]2- could act as a hydroxyl radical transfer reagent rather than an outer-sphere oxidant cannot be ruled out in

Meyer's system, since Ce(N0 3)6]2- was utilized as an oxidant for his mechanistic studies.

A homosphixic mechanism for 0-0 bond formation offers a much more thermodynamically favorable energetic scheme, since each oxygen atom is oxidized by only one electron. Although hydroxyl

1 radicals are extremely reactive and rapidly dimerize to form H2 0 2 at a rate of approximately 1010 M- s", these species can be stabilized by coordination to a metal center, as in the case of Cev.38 Oxyl radicals can also be stabilized by oxidizing metals, and are the catalytically active species in many oxidation catalysts,52 57 or they can be stabilized by nitrogen to form nitroxyl species such as TEMPO. There is an increasing consensus in the literature that 0-0 bond formation proceeds through a radical coupling

(homosphixic) mechanism with metal oxyl species as intermediates.2 3 This seems to stem from the simple fact that an 0-0 bond is not ionic, as there is not difference in electronegativity between the atoms. While

C-C bonds can be made though heterosphixis, the synthesis of a symmetric bond as the one found in ethane would require equivalents of -CH 3 and *CH3. This reaction is very exothermic compared with the homosphixis of two - CH3 radicals to give ethane. In contrast, the ionic bond of LiF can be formed by

0 heterosphixis of Li* and F or by homosphixis of Li and F2. These simple examples suggest that heterosphixis is the favored mechanism for forming ionic bonds while homosphixis is required to synthesize symmetric covalent bonds such as 02 with low overpotential.

24 1.3.3 Electrochemical Aspects and Electron Transfer

Electrochemical oxidation of water to molecular oxygen is complicated by the fact that it requires the simultaneous removal of four electrons in a concerted pathway to achieve the thermodynamic minimum of energy required. Oxidation of water by one electron to produce a hydroxyl radical is extremely unfavorable, and even oxidation by two electrons to form hydrogen peroxide requires an overpotential of 0.53 V. The thermodynamic 02 reduction potentials for the 1, 2, and 4 electron transfer pathways at pH = 7 are given in figure 1.1.58

02 02- H20 2 H20 + H0 2 H20 -0.33 V +0.89 V +0.38 V +2.32 V i e~/ 1 H*

+0.281 V +1.349 V 2 e- / 2 H+

+0.815 V 4 e- / 4 H+

While the overpotentials for the individual le~ steps could be overcome by coordination to a catalyst, Krishtalik has suggested that in order to attain the thermodynamic potential of 0.815 V for each individual step, a base stronger than water is required to accept the Hi lost. In addition, the overpotential for a stepwise mechanism must include corrections for inner and outer sphere reorganization energies, local concentration effects, diffusion entropy, coulombic repulsion, and the binding energies of intermediate species to the catalyst. In essence, this means that even if a catalyst could perfectly neutralize the activation energy for each e- step through the formation of strong bonds, there still are a number of other ways the system would not operate at the thermodynamic potential of the simultaneous 4e- pathway.

25 1.3.4 Thermodyamics

Since most discussion of the electrochemical water oxidation reaction focuses on electrochemical potentials and overpotentials, it is easy to lose sight of the fact that the standard reduction potentials are actually determined by thermodynamic analysis of reaction free energies, AG*. Using the data from figure

1.1, it is possible to create an oxidation state free energy diagram (Frost diagram), which shows the energies of the different reaction pathways (figure 1.2, inset).4 By subtracting the value of 0.815 V per e~ transferred, this diagram can be normalized to illustrate the thermodynamic activation energies for the le- and 2e- steps (figure 1.2).

This normalized Frost diagram emphasizes the substantial activation energy required for the reaction to proceed in le- or 2e- steps, and illustrates some important points regarding water oxidation and oxygen reduction. First, the le- oxidation of water to hydroxyl radical should proceed further to

generate 02 since it has sufficient energy to oxidize H 2 0 2 and 02-. In addition, it shows a very small minimum at H202, suggesting that the only way to generate H202 from water without subsequent 02 formation is by a 2e- oxidation process between 1.35 and 1.42 V (75 mV overpotential). This shallow potential energy well explains why hydrogen peroxide that generated during water le~ oxidation pathways only exists in low concentrations. On the other hand, the reduction of 02 by two le- steps

allows the facile synthesis of H 2 0 2, since the third le- reduction requires a much higher potential.

Superoxide (02-) can be formed by le- reduction of 02, but it is susceptible to disproportionation in solution to give H20 2 and 02. Finally, it suggests that reduction of 02 by 3e- or H20 2 by le- cleaves the

0-0 bond, creating a very reactive hydroxyl radical that is the active species in Fenton oxidation reactions.5 9

26 2.4 -O . 1 Electron Pathway HO2

2.2 - 2 Electron Pathway H 2O2 50

2.0 4 Electron Pathway > 1.8 - .40 7o > 1.6 - 1.4 - -(- -30 1.2 - H2O 02 -%

( 1 .0 - 20 * 0.8-

0.6 - * 10

0.2 - .2 2 H20 0.0 0 0 1 2 3 4 Oxidation Number

Figure 1.2 Frost diagram for the oxidation of H2 0 to 02 at pH = 7, normalized to the 4e- pathway potential. (inset) The same Frost diagram without normalization.4

1.3.5 Role of the Catalyst

When considering a catalyst that would carry out the oxidation of water to 02, it is important to emphasize that while a catalyst might stabilize a le- oxidized species by forming a bond whose energy

exactly matches the activation energy required (e.g. AG = 0 for M + H20 +-+ M-OH + H* + e-), the energy of that bond must be overcome in the activation energy of the following step. Krishtalik warns that if considering a mechanism for formation of H202 involving the dimerization of two metal bound hydroxyl radicals, the energy required for the reaction to be thermoneutral does not come anywhere close to being matched with a reasonable absorption energy for H202. In other words, the function of an ideal water oxidation catalyst should not be to bind reactive oxygen intermediates, but to store oxidizing

27 equivalents and base until four electrons and four protons can be simultaneously removed from two water

molecules. This analysis has influenced Dau to propose a radically different mechanism for the biological

23 60 OEC. , ,6' Dau suggests that when the manganese ions in the OEC are oxidized, g-OH species become

deprotonated to form tt-0 to accommodate the increasing charge on the complex. These g-0, rather than

being the substrate oxygen atoms that form 02, serve a base and allow protons in water molecules to be

removed as electrons are transferred to the manganese and/or tyrosyl moieties during the 0-0 bond

formation step. This proton coupled electron transfer (PCET) process takes advantage of the fact that the

oxidation of a metal complex in aqueous solution is accompanied by a decrease in pKa of the associated

ligands. This increased acidity would be detrimental to the reaction if water had already begun to be

oxidized, since the reaction becomes less favorable by 59 mV/pH. Instead, the excess H diffuses out of the reaction center, and the resulting Mn-pt-0 unit has both oxidizing power and, upon electron transfer, becomes a strong base with which to effect water oxidation efficiently.

Aside from the OEC, this mechanism of PCET is also potentially applicable to cobalt oxide thin

6 2 films. Just like Mn1 in the OEC of photosystem II, Co"I forms g-OH moieties that undergo a large pKa change upon oxidation to CoIv. The increased presence of these moieties in the more active amorphous

films of CoOx, rather than in less active crystalline Co 20 3 support this theory. While an 0-0 bond formation mechanism by which a terminal CoJv-oxyl radical is attacked by a Colv-OH may also be active in this system, rapid hydrogen atom transfer from the resulting Co"I-OOH species to a neighboring Co-p-

0 unit would allow the system to avoid discrete formation of this high energy hydroperoxide intermediate, and allow the formation of a stronger 0=0 multiple bond in the resulting Co01--02' or Co"-

02 species.

28 Table 1.1 Homolytic Bond Dissociation Enthalpies of Oxygen Species.6 3

. Ox.state BDE Species of "0" (kcal/mol)

02 0 118.9 02 -0.5 94.6 HO-O' -0.5 65.6 HO-0- -1 50.4 (gas) 46.9 (aq.) HO-OH -1 46.8 HO-H -2 118.2

The data in table 1.1 show that the 0-0 bond in H 20 2 is roughly 72 kcal/mol weaker than both the 0=0 double bond in 02 and the O-H bond in water. This weak bond strength suggests that in water oxidation mechanisms, peroxide intermediates will be high in energy unless coordination to a metal can stabilize the peroxide bond. While these values could change upon coordination to a metal, they serve as a rough framework to qualitatively discuss possible 0-0 bond formation mechanisms. Scheme 1.1 shows four possible mechanisms and their respective bond enthalpies represented by the bonds formed minus the bonds broken. The coupling of two metal oxos is shown in A,34 6'4 and the enthalpy change is represented by the enthalpy of the 0-0 bond formed minus the multiple bonding from each oxo. Since 0-0 single bonds are typically weak, this mechanism is expected to be favorable only if there is very weak metal-oxo multiple bonding. The stability of species such as MnO4 and FeO42- is likely due to the extremely strong multiple bonding that occurs in these complexes, despite their favorable redox potentials for water oxidation. In B, a metal bound hydroxide couples with an oxo species.33 This results in both a M-O a as well as 7c bond to be broken to form an 0-0 bond. Since transition metals typically form strong M-OH a bonds, dissociative mechanisms that result in open metal coordination sites are predicted to be extremely unfavorable. Mechanisms C and D describe the hydrogen atom transfer from water by surface bound oxide and hydroxide species. Both form 0-0 and MO-H bonds by cleaving M-O ir and HO-H bonds.

Since late transition metals often form weak M-O 7r bonds and strong MO-H bonds, these could be very thermodyamically favorable pathways. By adding two water molecules as substrates, mechanism E can eliminate the thermodynamic cost of breaking strong bonds, except for the required cleaving of the O-H

29 bonds in water. While this thermodynamic treatment is crude, and does not take into account oxidation state changes in the species, it suggests that the formation of strong metal-oxo bonds should inhibit the formation of a relatively weak 0-0 bond. In addition, water is likely to be a more thermodynamically favorable substrate for 0-0 bond formation than a species bound to a metal because it can bypass the activation energy required to dissociate the product. The 10' difference in rate of water oxidation by RuOx

2 and [Ru(tpy)(bpz)(OH 2)] + might imply that the former does not bind or dissociate intermediate species

(mechanism E), because the latter is shown to be limited by the rate of 02 dissociation from Ru.

Scheme 1.1 Possible 0-0 bond formation mechanisms and their respective bond enthalpies.

0 0 0-0 II I I AH =(0-0) - 2(M-O)rr A M|| M M M

OOH 0 OH B || AH = (0-0) - (M=O) M M M M

HO' H OH

0 OH 0 0 AH = [(0-0) + (MO-H)] I I - [2(M-O)r + (HO-H)] M M M M

HO' H OH

0 OH O OH2 AH = [(0-0) + (MO-H)] D O I I - [(M-O)Tr + (HO-H)] M M M M

0 H O OH O E OH 2 AH = [4(MO-H) + (0=0)] M M -[(M-O)T + 4(HO-H)] 0 H

30 In conclusion, this section describes several features of 0-0 bond formation and attempts to integrate various literature theories into a better understanding of different mechanistic pathways. First, it points to Shaik's work to explain why 0-0 a bonds are weak in comparison with the 7r bond. Next, two new words are created to describe possible bond formation events. Heterosphixis is argued to be particularly disfavored in the formation of a nonpolar 0-0 bond, and therefore homosphixis is likely to occur in all but the most extreme systems. Next, electrochemical potentials of oxygen intermediates are used to create a new 'normalized' Frost diagram, which aids in visualization of the relative activation energies of intermediate species. Finally, the role of the catalyst is examined using bond dissociation energies to describe the energy required for 0-0 bond formation mechanisms. Dau's suggestion that terminal and p-oxo moieties might act as basic sites for protons in water is used to construct a new water oxidation mechanism (E) where two water substrates are associated with metal catalyst through only hydrogen bonds. If the catalyst does not bind to the water substrates, rapid 0-0 bond formation and release of 02 could occur. While different systems likely proceed by other mechanisms, mechanism E might be operable in heterogeneous and biological systems that facilitate rapid oxidation of water to 02.

31 1.4 Ligand Design and Scaffolds for Oxidizing Species

In order to carefully study possible mechanisms of 0-0 bond formation, the synthesis of molecular metal species that could stabilize oxo intermediates along reaction pathways was targeted.

Although early transition metals often form very stable oxo species, 65 later metals such as MnV, FeiV, or

Co'v were pursued, as they are expected to have the oxidation potential to carry out 0-0 bond formation processes. In contrast with earlier transition metals, these metals will have d-electrons in the ligand field, and stabilization of these electrons is important in the formation of a strong metal-oxo bond.

Metals can form up to three bonds with an oxygen atom: one a and two 7. If considering the bond along the z axis, oxo bonding destabilizes the dZ , dxz, and dyz orbitals. Scheme 1.2 (adapted from refs. 19 and 44) shows qualitative ligand field d-orbital diagrams for a representative FeIv species in various ligand field geometries. 19' 44 For a low spin FeIv-oxo complex, four valence electrons of iron will lie in the xy plane. In order to not perturb these orbitals and retain the strong oxo triple bond, ligands that coordinate below this plane could provide linear, planar, or trigonal oxo complexes. FeIv(N) complexes in trigonal ligand fields show that this strong bonding can be accommodated by chelating ligands that range in LL-Fe-N from 1000 to 120*,6667 which suggests that a strong ligand field could compensate for departures from pseudotetrahedral bond angles of 109.5*. On the other hand, tetragonal or trigonal bipyramidal geometries will destabilize these orbitals causing d, and dyz, which are 7r-antibonding with respect to the oxo, to become populated. These geometries will lower the metal-oxo bond order to two and potentially transfer radical character on the oxo.

32 Scheme 1.2 Ligand Field Diagrams for Fe'v Oxo Species in Different Geometries1 9' 44

0 0 0 0 0 0 FIVIII L., I L LI~L FeIv FelV FelV FeV FelV L-Fi FeIV L L L L L L0 L L LV" I L L L 1 L

dz 2 ...... -4--- -

dxz, dyz ----- :4 dx 2-y 2, dxy ±

Many transition metals are able to attain very high oxidation states by coordinating ligands in

tetrahedral ligand fields. The stable species Crv042-, Mn"04-, and FeV042- are the highest known

oxidation states for each of these respective metal species.36 These species, however, do not contain

formal metal-oxo triple bonds since the metal cannot accommodate the 12 bonds needed to bond with all

of the oxos present. The 9 valence orbitals of a transition metal limits the bond order in these complexes to 9/4 (2.25) in Cr0 2 and 2 2 4 MnO4 , and a bond order of 2.0 in d Fe0 4 -

Betley and Peters reported a pseudotetrahedal FeIv(N) complex that dimerizes to form an N 2 * 68 species. While this reaction has also been observed with pseudooctahedral RuvI and OsV1 nitride

species, 69-71 the presence of four d-electrons in the FeIv(N) suggests that an analogous FeIv(O) species might perform 0-0 bond formation. 19' 44 Although the bond strength arguments in the previous section suggest that this is very unlikely to occur, the properties of an oxo complex in this geometry would be interesting to investigate for more favorable oxidation reactions. Chapters 3 and 4 of this thesis describe the use of siloxide ligands and tripodal amide ligands to generate pseudotetrahedral species, while chapter

2 explores the unusual synthesis and redox activity of bis-p-diketiminate complexes.

33 1.5 References

(1) Croswell, K. The Alchemy of the Heavens: Searchingfor Meaning in the Milky Way; 1ST ed.; Anchor, 1996.

(2) Morgan, J. W.; Anders, E. Proc.Nat. Acad. Sci. 1980, 77, 6973-6977.

(3) Zumdahl, S. S.; Zumdahl, S. A. Chemistry; 6th ed.; Brooks Cole, 2002.

(4) Armstrong, F. A. Phil. Trans. R. Soc. B 2008, 363, 1263-1270.

(5) Stearns, S.; Hoekstra, R. Evolution; 2nd ed.; Oxford University Press, USA, 2005.

(6) Buick, R. Phil. Trans. R. Soc. Lond. B. 2008, 363, 2731-2743.

(7) Olson, J. M. PhotosynthesisResearch 2006, 88, 109-117.

(8) Kashefi, K.; Lovley, D. R. Science 2003, 301, 934-934.

(9) Catling, D. C.; Zahnle, K. J.; McKay, C. P. Science 2001, 293, 839-843.

(10) Diamond, J. Collapse: How Societies Choose to Fail or Succeed: Revised Edition;Revised.; Penguin (Non-Classics), 2011.

(11) Lewis, N. S.; Nocera, D. G. Proc. Nat.Acad. Sci. 2006, 103, 15729-15735.

(12) Blankenship, R. E.; Tiede, D. M.; Barber, J.; Brudvig, G. W.; Fleming, G.; Ghirardi, M.; Gunner, M. R.; Junge, W.; Kramer, D. M.; Melis, A.; Moore, T. A.; Moser, C. C.; Nocera, D. G.; Nozik, A. J.; Ort, D. R.; Parson, W. W.; Prince, R. C.; Sayre, R. T. Science 2011, 332, 805-809.

(13) Zouni, A.; Witt, H.-T.; Kern, J.; Fromme, P.; Krauss, N.; Saenger, W.; Orth, P. Nature 2001, 409, 739-743.

(14) Kamiya, N.; Shen, J.-R. Proc. Nat. Acad. Sci. 2003, 100, 98-103.

(15) Ferreira, K. N.; Iverson, T. M.; Maghlaoui, K.; Barber, J.; Iwata, S. Science 2004, 303, 1831-1838.

(16) Guskov, A.; Kern, J.; Gabdulkhakov, A.; Broser, M.; Zouni, A.; Saenger, W. Nature Structural & Molecular Biology 2009, 16, 334-342.

(17) Broser, M.; Gabdulkhakov, A.; Kern, J.; Guskov, A.; Mfnh, F.; Saenger, W.; Zouni, A. J Biol. Chem. 2010, 285, 26255-26262.

34 (18) Umena, Y.; Kawakami, K.; Shen, J.-R.; Kamiya, N. Nature 2011, 473, 55-60.

(19) Betley, T. A.; Surendranath, Y.; Childress, M. V.; Alliger, G. E.; Fu, R.; Cummins, C. C.; Nocera, D. G. Phil. Trans. R Soc. B 2008, 363, 1293-1303.

(20) Siegbahn, P. E. M. Chemistry - A EuropeanJournal 2008, 14, 8290-8302.

(21) Siegbahn, P. E. M. Phil. Trans. R. Soc. B 2008, 363, 1221-1228.

(22) Barber, J.; Siegbahn, P. E. M.; Nocera, D.; Dau, H.; Armstrong, F.; Rich, P. Phil. Trans. R. Soc. B. 2008, 363, 1228-1228.

(23) Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser, P. ChemCatChem 2010, 2, 724- 761.

(24) Gatt, P.; Stranger, R.; Pace, R. J. J. Photochem. Photobio.B. 2011, 104, 80-93.

(25) Grimes, C. A.; Varghese, 0. K.; Ranjan, S. Light, water, hydrogen: the solargeneration of hydrogen by water photoelectrolysis; Springer, 2008.

(26) Merrill, M. D.; Dougherty, R. C. J Phys. Chem. C 2008, 112, 3655-3666.

(27) Rossmeisl, J.; Qu, Z.-W.; Zhu, H.; Kroes, G.-J.; Norskov, J. K. J. ElectroanalyticalChem. 2007, 607,83-89.

(28) Trasatti, S. ElectrochimicaActa1984,29, 1503-1512.

(29) Cook, T. R.; Dogutan, D. K.; Reece, S. Y.; Surendranath, Y.; Teets, T. S.; Nocera, D. G. Chem. Rev. 2010, 110, 6474-6502.

(30) Kanan, M. W.; Nocera, D. G. Science 2008, 321, 1072-1075.

(31) Kanan, M. W.; Surendranath, Y.; Nocera, D. G. Chem. Soc. Rev. 2009, 38, 109-114.

(32) Dinca, M.; Surendranath, Y.; Nocera, D. G. Proc. Nat. Acad. Sci. 2010, 107, 10337-10341.

(33) Surendranath, Y.; Kanan, M. W.; Nocera, D. G. J. Am. Chem. Soc. 2010, 132, 16501-16509.

(34) Liu, F.; Concepcion, J. J.; Jurss, J. W.; Cardolaccia, T.; Templeton, J. L.; Meyer, T. J. Inorg. Chem. 2008, 47, 1727-1752.

(35) Concepcion, J. J.; Tsai, M.-K.; Muckerman, J. T.; Meyer, T. J. J. Am. Chem. Soc. 2010, 132, 1545-1557.

35 (36) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. Advanced Inorganic Chemistry, 6th Ed.; Wiley-Interscience: New York, 1999.

(37) Yiu, S. M.; Man, W. L.; Wang, X.; Lam, W. W. Y.; Ng, S. M.; Kwong, H. K.; Lau, K. C.; Lau, T. C. Chem. Comm. 2011, 47, 4159-4161.

(38) Serrano, K.; Michaud, P. A.; Comninellis, C.; Savall, A. ElectrochimicaActa 2002, 48, 431-436.

(39) Dinnebier, R. E.; Vensky, S.; Stephens, P. W.; Jansen, M. Angew. Chem. Int. Ed. 2002, 41, 1922- 1924.

(40) Kohl, S. W.; Weiner, L.; Schwartsburd, L.; Konstantinovski, L.; Shimon, L. J. W.; Ben-David, Y.; Iron, M. A.; Milstein, D. Science 2009, 324, 74-77.

(41) Galbraith, J. M.; Blank, E.; Shaik, S.; Hiberty, P. C. Chemistry - A European Journal2000, 6, 2425-2434.

(42) Pauling, L. The Nature of the Chemical Bond and the Structure of Molecules and Crystals: An Introduction to Modern Structural Chemistry; 3rd ed.; Cornell University Press, 1960.

(43) Cotton, F. A.; Nocera, D. G. Acc. Chem. Res. 2000, 33, 483-490.

(44) Betley, T. A.; Wu, Q.; Van Voorhis, T.; Nocera, D. G. Inorg. Chem. 2008, 47, 1849-1861.

(45) Zong, R.; Thummel, R. P. J. Am. Chem. Soc. 2005, 127, 12802-12803.

(46) Tseng, H.-W.; Zong, R.; Muckerman, J. T.; Thummel, R. Inorg. Chem. 2008, 47, 11763-11773.

(47) Gao, Y.; Akermark, T.; Liu, J.; Sun, L.; Akermark, B. J. Am. Chem. Soc. 2009, 131, 8726-8727.

(48) Gao, Y.; Liu, J.; Wang, M.; Na, Y.; Akermark, B.; Sun, L. Tetrahedron 2007, 63, 1987-1994.

(49) Wang, L.-P.; Wu, Q.; Van Voorhis, T. Inorg. Chem. 2010, 49, 4543-4553.

(50) Yoshida, M.; Masaoka, S.; Abe, J.; Sakai, K. Chemistry - An Asian Journal2010, 5, 2369-23 78.

(51) Yu, P.; Hayes, S. A.; O'Keefe, T. J.; O'Keefe, M. J.; Stoffer, J. 0. J.Electrochem. Soc. 2006, 153, C74-C79.

(52) Bach, R. D.; Dmitrenko, 0. J. Am. Chem. Soc. 2006, 128, 1474-1488.

(53) Kobayashi, K.; Ohtsu, H.; Wada, T.; Kato, T.; Tanaka, K. J Am. Chem. Soc. 2003, 125, 6729- 6739.

(54) Lundberg, M.; Blomberg, M. R. A.; Siegbahn, P. E. M. Inorg. Chem. 2003, 43, 264-274.

36 (55) Siegbahn, P. E. M.; Crabtree, R. H. J. Am. Chem. Soc. 1998, 121, 117-127.

(56) Vennat, M.; Herson, P.; Brdgeault, J.; Shul'pin, G. B. Eur. J. Inorg. Chem. 2003, 2003, 908-917.

(57) Groysman, S.; Villagrin, D.; Nocera, D. G. Inorg. Chem. 2010, 49, 10759-10761.

(58) Wood, P. M. Biochem J 1988, 253, 287-289.

(59) Fenton, H. J. H. J. Chem. Soc. Trans. 1894, 65, 899.

(60) Dau, H.; Haumann, M. PhotosynthesisResearch 2005, 84, 325-331.

(61) Dau, H.; Grundmeier, A.; Loja, P.; Haumann, M. Phil. Trans. R. Soc. B 2008, 363, 123 7-1244.

(62) Risch, M.; Klingan, K.; Ringleb, F.; Chernev, P.; Zaharieva, I.; Fischer, A.; Dau, H. ChemSusChem 2012, 5, 542-549.

(63) Luo, Y.-R. Comprehensive Handbook of Chemical Bond Energies; 1st ed.; CRC Press, 2007.

(64) Wang, L.-P.; Van Voorhis, T. J. Phys. Chem. Lett. 2011, 2, 2200-2204.

(65) Holm, R. H. Chem. Rev. 1987, 87, 1401-1449.

(66) Vogel, C.; Heinemann, F. W.; Sutter, J.; Anthon, C.; Meyer, K. Angew.Chem. Int. Ed. 2008, 47, 2681-2684.

(67) Scepaniak, J. J.; Fulton, M. D.; Bontchev, R. P.; Duesler, E. N.; Kirk, M. L.; Smith, J. M. J. Am. Chem. Soc. 2008, 130, 10515-10517.

(68) Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2004, 126, 6252-6254.

(69) Man, W.-L.; Tang, T.-M.; Wong, T.-W.; Lau, T.-C.; Peng, S.-M.; Wong, W.-T. J. Am. Chem. Soc. 2003, 126,478-479.

(70) Buhr, J. D.; Taube, H. Inorg. Chem. 1979, 18, 2208-2212.

(71) Seymore, S. B.; Brown, S. N. Inorg. Chem. 2002, 41, 462-469.

37 Chapter 2: Synthesis and Redox Activity of bis-p- Diketiminate Cobalt

Submitted to Journalof the American Chemical Society

38 2.1 Introduction

The oxidation or reduction metal often causes significant structural changes to reflect the

differing electronic structure of the system. Catalytic processes that involve a change in oxidation state thus confront reorganization energy barriers.1 - One method to facilitate multielectron redox processes

and minimize the inherent geometric and electronic energy barriers that transition metals must incur is through the use of ligands that can participate in the redox chemistry. Although organic ligands are not

immune to structural changes upon oxidation or reduction, the resulting organic radicals are usually delocalized over multiple atoms, minimizing reorganizational energy.

Redox active ligands can be broadly divided into two classes: one in which the redox potential of the ligand is only weakly coupled to the electronics of the metal, and one in which the ligand strongly affects the redox properties of the metal. In early transition metal complexes, ionic bonding prevents strong coupling of the redox events of the ligand to the metal center. This permits redox inactive metals such as zirconium or aluminum to participate in transformations traditionally carried out only by redox active metal species.'' The increased electronegativity of late transition metals often produces complexes with electrons in d-orbitals that are closer in energy to redox-active orbitals in the ligand. In this case, the metal and ligand orbitals can mix to give molecular orbitals that delocalize the corresponding electrons.

When the metal and ligand electron distribution approaches equivalence, the formal oxidation state of the complex can become ambiguous, since both metal and ligand orbitals are only partially populated.'

In pursuit of attaining facile multielectron redox chemistry, our group has investigated

3 17 porphyrin, 8-10 corrole, 1-1 porphyrinogen,14- and PNP pincer18 systems containing redox noninnocent ligand systems. These properties have been exploited to facilitate the two- and four-electron processes of hydrogen and oxygen generation, respectively. One common motif in many redox active ligands is the presence of 7c bonding in the ligand.19 The delocalization that occurs in 7c bonding yields a small HOMO-

39 LUMO gap, placing one of these within the d-orbital manifold of the metal. Upon oxidation or reduction, electrons can be gained by or removed from ligand based orbitals rather than metal d-orbitals.

Most 7r-based redox-active ligands studied to date contain an even number of atoms in the framework to generate an equal number of 7c bonding and antibonding orbitals. 2 0 By changing the electronic occupancy of these orbitals, the ligand atomic bond distances are expected to change significantly. For example, bipyridine can coordinate to transition metals in its neutral, monoanionic, or

2 dianionic form. ' Detailed X-ray structural analysis has shown that the reduction of bipyridine corresponds to a successive increase in the Cl-Cl' bond length by roughly 0.05 A for each electron added. Thus, one of the most common methods to determine the oxidation state of such ligands has been through X-ray crystallography.

p-diketiminate (NacNac) ligands have an odd number of atoms in the n system. Despite their ubiquitous presence in the literature,22 only recently have Khusniyarov, Wieghardt, et al. discovered

NacNac ligands to be redox-active. 2 3 Pseudo-tetrahedral nickel bis-NacNac forms an intensely colored complex upon its oxidation at -20 'C. Based on the strong visible absorptions and corresponding DFT calculations, this oxidized complex has been described as a Ni(II) NacNac radical cation rather than a

Ni(III) complex. Due the odd number of orbitals in the n system, the HOMO of a NacNac ligand is a i nonbonding orbital. Corresponding oxidation of this orbital implies that no ligand bond length changes should be detectable by X-ray crystallography, leading to the moniker of "hidden noninnocence" to describe a class of ligands that undergo redox events without accompanying observable structural changes.

This chapter describes a planar cobalt bis-NacNac system that is synthesized through reaction with acetonitrile. The convenient synthesis of NacNac from organometallic precursors provides a new route to generate late transition metal NH functionalized NacNac complexes. Oxidation of the cobalt complex yields the first crystallographically characterized NacNac based radical in a metal complex. The nature of this cation radical is compared to acac, salen, and porphyrin ligand systems.

40 2.2 Results

2.2.1 Reaction with Acetonitrile

2 4 2 5 Treatment of either Li(THF) 4Co(mesityl) 3 or [Co(mesityl) 2] 2 with excess acetonitrile yields a brown crystalline solid of (1-mesitylbutane-1,3-diimine) 2Co (1). While this reaction proceeds slowly at or below room temperature, the reaction rate can be accelerated by mild heating. The 1H NMR spectrum shows four broad features indicating paramagnetism, which have been assigned as mesityl and methyl resonances. There is no evidence for the a-C-H or N-H resonances of the ligand. This can be rationalized by the delocalization of the SOMO on these atoms (vide infra).

1/2 Mes

Os IoHc (1) NH HN Mes

Reaction with CD 3CN shows deuterium incorporation at the in-plane methyl group by 2H NMR, and the IR reveals the presence of N-D stretching absorptions (figure 2.1). This pattern of deuterium incorporation suggests that 1 is formed by insertion of acetonitrile into the cobalt-mesityl bond, followed by attack on a second coordinated acetonitrile (scheme 2.1). This mechanism of acetonitrile insertion has been proposed in a scandium-methyl, 26 yttrium-ary1, 27 and a chromium-methyl species,28,29 each of which form corresponding NacNac derivatives. Reaction of [Co(mesityl) 2] 2 with propionitrile or phenylacetonitrile did not yield isolable NacNac complexes.

41 Scheme 2.1 Proposed Mechanism for the Formation of 1 (adapted from ref. 26 and 29).

11 N Mes Mes CH 3CN o LnCo-Mes 30- LnCo-Mes ow LnCo-N & LnCo-NH

CH3CN

HN N ,NMes N Mes LnCo-NH LnCo-NH LnCo-NH

2451 cm

1-d 1'1i 1-....

3297 cm

3274 cm . I . . I . I . i . 3500 3450 3400 3350 3300 3250 3200 3150 2650 2600 2550 2500 2450 2400 2350 2300 1 v / cm- v / cm"

Figure 2.1 Infrared spectrum of 1 and 1-d12 showing the isotopic shift in the N-H vs. N-D stretching mode.

2.2.2 Characterization of I

X-ray crystal structures of 1 have been obtained from crystals grown from solutions of both acetonitrile (la) and ether (1) (figure 2.2). Both structures show a square planar cobalt center ligated by two p-diketiminate ligands, with average Co-N bond distances of 1.87 A and 1.86 A respectively (tables

2.1 and 2.2). 1 crystallizes in a solvent-free lattice with 2 independent molecules in the asymmetric unit.

30 Each molecule lies on an inversion center in a fashion reminiscent of the structure of (Me6acac) 2Ni.

42 Each molecule also displays a planar geometry of the core unit (non-mesityl), with nearly identical

LN-Co-N bond angles of 90.00.

In the case of la, two molecules of acetonitrile are present and show hydrogen bonding to the

N-H functionalities, rather than coordination to the cobalt center. This hydrogen bonding induces a slight interligand ZN-Co-N bond angle contraction to 89.40 which is compensated by an intraligand bond angle expansion to 90.6*, retaining a planar CoN 4 molecular core.

(a) (b)

N(3)

43 Table 2.1 Selected bond lengths (A) and angles (0) of 1.

Bond Lengths

Co(1)-N(1) 1.8642(11) Co(1)-N(2) 1.8539(12) Co(2)-N(3) 1.8667(12) Co(2)-N(4) 1.8565(11) N(1)-C(2) 1.3290(18) N(2)-C(4) 1.3333(18) N(3)-C(15) 1.3263(17) N(4)-C(17) 1.3332(17) C(2)-C(3) 1.4007(19) C(3)-C(4) 1.3963(18) C(15)-C(16) 1.4047(19) C(16)-C(17) 1.3899(19)

Bond Angles

N(1)-Co(1)-N(2) 90.00(5) N(1)-Co(1)-N(2A) 90.00(5) N(3)-Co(1)-N(4) 90.15(5) N(3)-Co(1)-N(4A) 89.85(5) N(1)-C(2)-C(3) 122.30(12) N(2)-C(4)-C(3) 122.59(13) N(3)-C(15)-C(16) 121.83(13) N(4)-C(17)-Cr(16) 122.90(12) C(2)-C(3)-C(4) 122.59(13) C(15)-C(16)-C(17) 122.81(12)

Table 2.2 Selected bond lengths (A) and angles (*) of la.

Bond Lengths

Co(1)-N(1) 1.8679(11) Co(1)-N(2) 1.8585(10) N(1)-C(2) 1.3296(16) N(2)-C(4) 1.3317(16) C(2)-C(3) 1.3988(17) C(3)-C(4) 1.3959(17)

Bond Angles

N(1)-Co(1)-N(2) 90.60(5) N(1)-Co(1)-N(2A) 89.40(5) N(1)-C(2)-C(3) 122.25(11) N(2)-C(4)-C(3) 122.86(11) C(2)-C(3)-C(4) 123.09(12)

44 2.50

2.25

2.00 *.,. * * * * * * ** * * 4

1.75

1.50

: 1.25

=- 1.00

0.75

0.50

0.25

0.00 I 0 50 100 150 200 250 300 T/K

Figure 2.3 Solid State temperature dependence of the magnetic moment of 1 recorded on a SQUID magnetometer at 1000 G.

SQUID magnetometry was carried out on a solid crystalline sample of 1 and shows simple paramagnetic behavior down to 2 K. The magnetic moment of the sample, pf, was found to be 1.98 BM and not variant with temperature (figure 2.3). On the other hand, in benzene solution, complex 1 shows a room temperature magnetic moment of 2.4(1) BM measured by Evans' method.31 These differing values in solution and in the solid state could suggest a planar-tetrahedral isomerization process, which is accompanied by spin crossover. Since molecules in a crystalline sample of 1 are not free to rotate, they remain in a square planar geometry with a low-spin (S = 1/2) electronic configuration. In solution, however, the NacNac ligands may freely rotate with respect to each other, allowing the high spin isomer to contribute to the magnetic moment. DFT calculations suggest that the pseudotetrahedral high spin (S =

3/2) species is less than 1 kcal/mol higher in energy, than the low spin complex. Relaxed potential energy

45 surface scans along the N-N-N-N dihedral angle of both high spin and low spin configurations gives a rotation barrier of about 6 kcal/mol for the two spin-energy surfaces (figure 2.4). The average Co-N bond length is 0.033 A longer in the calculated (1.892 A) vs. experimental (1.859 A) structure of 1, so the low spin isomer might be more stable than the calculations suggest. The solution based magnet moment compares closely to that of cobalt porphryin (2.4 BM32), but the disagreement in magnetic data may reflect some other phenomenon. No spectroscopic evidence of the high spin isomer has been identified, however similar cobalt species have been shown to exhibit planar-tetrahedral spin crossover

3 3 3 6 isomerization. -

12-

10-

6- 2' 0 4- Spin

I I I I 0 20 40 60 80 100 Dihedral Angle / deg

46 g-factor 3.2 3 2.8 2.6 2.4 2.2 2 1.8

200 250 300 350 400 Field / mT

Figure 2.5 EPR of 1 in toluene glass at 77 K. Simulation yields g-values of gi = 2.808, g2 = 2.002, and g3 = 1.956.

The EPR spectrum of 1 shows a broad signal at g = 2.0 at room temperature which splits to a rhombohedral signal at 77 K (toluene glass, figure 2.5). The observed g-values of gi = 2.808, g2 = 2.002,

37 38 3 9 4 0 The gi g3 = 1.956 are typical of related low spin planar Col acacen, salen, and amben , complexes. value indicates a lack of apical coordination and significant electron delocalization along one axis. The g3

2 the value that is less than 2 corresponds to complexes characterized as having a B 2 (dy,)' ground state in

with the electronic structure of Co 1 complexes give C2 , point group. Equations correlating g-values

2 1 roughly a 2300 cm-1 energy gap between the ground state (dyz)' and the first excited state (dZ) .41 The

2 EPR data thus suggest that the unpaired electron in 1 occupies a dyz orbital, rather than a dz orbital, which

42 is commonly found in cobalt (II) oxime and porphyrin complexes. DFT calculations confirm that the

47 SOMO of 1 is in fact of b2 symmetry (figure 2.6), and the spin density is exclusively metal based (>99%, figure 2.7).

Figure 2.6 Singly occupied molecular orbital (SOMO) of 1 showing mixing of NacNac non-bonding x orbital with metal dyz, isovalue 0.03.

0.99 I

Figure 2.7 Spin density plot of 1 showing unpaired electron localized on the metal based dyz orbital (99%), isovalue 0.008.

1 shows four intense transitions in the UV (figure 2.8), which TD-DFT calculations suggest correspond to n --+ a* and charge transfer transitions in the bis-NacNac ligand system (figure 2.9). Since the n-system of 1 can be considered a subset of a porphyrin, these transitions arise from similar transitions

that give rise to the Soret and Q-bands. The descent in symmetry from D4h to C 2v, however, splits the

48 additional degenerate eg LUMO of a porphyrin into b1 and b2 type orbitals in complex 1, giving rise to the transitions.

25000 264 nm

298 nm 20000

339 nm

15000 377 nm E 10000

5000 417 nm

0- 300 400 500 600 700 800 X/ nm

Figure 2.8 UV-Vis absorption spectrum of 1 in pentane.

49 68"o 67P 69a 68a

379 nm 322 nm 297 nm 263 nm MLCT LMCT MLCT ITr *

660 63j3 66a 64a

Figure 2.9 The dominant donor and acceptor molecular orbitals for the four prominent absorptions in 1, obtained from spin-unrestricted TD-DFT calculations.

2.2.3 Electrochemistry and Synthesis of 2

Figure 2.10 displays the cyclic voltammogram of 1. One reversible oxidation event is observed within the solvent windows of dichloromethane, acetonitrile, and THF. Curiously, reduction to a Co(I) species was not observed, unlike most other planar Co(II) complexes. Similar redox behavior of 1 is

observed in CH 2C12, but the reversible oxidation is shifted 248 mV more positive than in THF. While differences in solvent dielectric and electrolyte ion pairing could shift the observed potential, this phenomenon is likely due to coordination of THF to Co in the oxidized complex. Alternately, THF could hydrogen bond to the N-H protons of the ligand, stabilizing the planar isomer relative to the pseudotetrahedral high spin isomer.

50 0.08

0.06

0.04

0.02

E 0.00 -

a, -0.02

S-0.04

-0.06

-0.08

-0.10

0 -500 -1000 -1500 -2000 -2500 Potential (mV)

Figure 2.10 Cyclic voltammogram of 1, 1 mM in THF with 0.1 M Bu 4NPF6 supporting electrolyte at a scan rate of 100 mV/s and referenced vs. ferrocene. Oxidation is centered at -434 mV.

To characterize the oxidized species, 1 was treated with ferrocenium hexafluorophosphate in THF to yield complex 2, which was isolated in 90% yield as a dark blue crystalline solid. Oxidation was also attempted in the absence of coordinating solvents, and although reaction in dichloromethane or benzene solutions proceeds similarly, a dark brown solid precipitated from the resulting dark blue solutions over the course of an hour. Formation of the brown precipitate was accelerated during attempts to crystallize this species as well as upon cooling. Since the PF6 counterion might react with the oxidized complex, reaction of 1 with FcBArF in non-coordinating solvents was attempted. This reaction also resulted in formation of brown precipitates.

The X-ray crystal structure of 2 reveals a THF molecule bound apically to the cobalt center to give an approximate square-pyramidal geometry (figure 2.11). Another THF molecule is hydrogen bound

51 to two of the N-H moieties in a fashion reminiscent of la. Surprisingly, oxidation of 1 increased the

Co-N bond distances by 0.01 A to 1.87 A in 2 (table 2.3). This effect is perhaps due to weaker metal- ligand 7r interactions since the planes defining the NacNac ligands are each bent 50 away from the apical

THF molecule. The structure also reveals that one of the NacNac ligands has rotated 1800 with respect to the other to give a cis arrangement of the mesityl group, perhaps indicative of the aforementioned spin crossover isomerization of 1.

0(30)

Figure 2.11 X-ray crystal structure of 2 with 50% probability ellipsoids. An additional uncoordinated THF molecule and hydrogen atoms were omitted for clarity (except for N-H).

Table 2.3 Selected bond lengths (A) and angles (0) of 2.

Bond Lengths

Co(1)-N(1) 1.8695(16) Co(1)-N(2) 1.8677(16) Co(1)-N(3) 1.8708(17) Co(1)-N(4) 1.8716(16) Co(1)-O(30) 2.1246(14) N(1)-C(2) 1.324(2) N(2)-C(4) 1.328(2) N(3)-C(15) 1.325(2) N(4)-C(17) 1.328(2) C(2)-C(3) 1.399(3) C(3)-C(4) 1.397(3) C(15)-C(16) 1.397(3) C(16)-C(17) 1.393(3)

52 Bond Angles

N(1)-Co(1)-N(2) 90.29(7) N(1)-Co(1)-N(3) 88.79(7) N(2)-Co(1)-N(4) 88.94(7) N(3)-Co(1)-N(4) 89.83(7) N(1)-Co(1)-O(30) 94.88(6) N(2)-Co(l')-(30) 95.03(7) N(3)-Co(1)-O(30) 96.03(7) N(4)-Co(1)-O(30) 96.28(6) N(1)-C(2)-C(3) 122.33(17) N(2)-C(4)-C(3) 122.65(17) N(3)-C(15)-C(16) 122.07(18) N(4)-C(17)-C(16) 122.88(18) C(2)-C(3)-C(4) 123.24(17) C(15)-C(16)-C(17) 123.14(18)

15000

10000

C~3 5000

00 200 400 600 800 1000 1200 1400 /nm

Figure 2.12 UV-Vis-NIR Absorption Spectrum of 2 in THIF.

A broad and intense band at 742 nm gives rise to the observed blue color while a broader and weaker absorption is observed at 1273 nm (figure 2.12). Whereas these bands correspond to MLCT and

IVCT transitions, respectively, in the tetrahedral Ni complex, the molecular orbitals that give rise to these absorptions in 2 are highly covalent in character, populating both the metal and ligand (figure 2.13).

53 Calculations and magnetic data (pf = 3.2(1), by Evans' method) suggest that this paramagnetic species is a ground state triplet, which corresponds to one unpaired electron in d 2 and one in an orbital that is shared between a metal dyz orbital and a NacNac ligand i orbital.

8613 863

1213 nm 695 nm

Figure 2.13 The dominant donor and acceptor molecular orbitals for the two prominent absorptions in 2, obtained from spin-unrestricted TD-DFT calculations.

The calculated spin density plot of 2 (figure 2.14) corresponds to one unpaired spin on dz and

roughly 60% unpaired spin on dyz The remaining 40% is delocalized on both NacNac ligands with a

small contribution on the oxygen atom of the coordinated THF. In comparison with 1, the spin of the

metal dyz orbital in 2 is delocalized on the ligand due to increased covalency of the metal-ligand

interaction in the cationic complex.

54 0.04

0.08 0.06

0.16 0.16

0.08 1.61 0.08

Figure 2.14 Spin density plot of 2 showing spin density both on the metal and ligand, isovalue = 0.008.

2.2.4 Synthesis of Fe(1-mesitylbutane-1,3-diimine) 2 (3)

In order to investigate the generality of acetonitrile insertion and condensation to form p- diketiminates, an ether solution of [Fe(mesityl) 2]2 was treated with CH3CN at -40 'C. If the solution was kept in the dark and cold, a small number of dark brown crystals separated from the solution. The crystals were analyzed by X-ray crystallography and found to be Fe(1-mesitylbutane-1,3-diimine) 2 (3) (figure

2.15). The Fe analog of 1 crystallizes in the same space group, with approximately the same unit cell parameters. The only significant difference between 3 and 1 are the Fe-N bonds (table 2.4), which have

expanded to an average of 1.91 A, compared to the Co-N average of 1.86 A.

The square planar geometry of 3 is intriguing since the lack of axial coordination might suggest

that the dZ orbital is occupied as it is in 1. While an Fe(II) species would not be expected to be oxidizing

enough for the NacNac ligand to be formally oxidized, the dark color might be indicative of low energy a

-+ 7E* transition in the ligand framework. Unfortunately, 3 proved to be extremely sensitive to light and

55 heat, in addition to water and oxygen, and decomposition prevented the isolation of 3 in pure form to permit further analysis.

Figure 2.15 X-ray crystal structure of one of the two independent molecules of 3 with 50% probability ellipsoids. Hydrogen atoms were omitted for clarity (except for N-H).

Table 2.4 Selected bond lengths (A) and angles (*) of 3.

Bond Lengths Fe(1)-N(1) 1.9069(11) Fe(1)-N(2) 1 8966(10) Fe(2)-N(3) 1.9052(10) Fe(2)-N(4) . 8923(1 1) N(1)-C(2) 1.3273(15) N(2)-C(4) 1.3344(15) N(3)-C(15 ) 1.3269(16) N(4)-C(17) 3333(16) C(2)-C(3) 1.4035(17) C(3)-C(4) 3929(17) C(15)-C(1 6) 1.4013(17) C(16)-C(17) .1022(17) Bond Angles

N(1)-Fe(1 )-N(2) 89.59(4) N(1)-Fe(1)-N(2A) 90.41(4) N(3)-Fe(1)-N(4) 90.38(5) N(3)-Fe(1)-N(4A) 89 .62(5) N(1)-C(2) -C(3) 122.33(11) N(2)-C(4)-C(3) 123.22(11) N(3)-C(15 )-C(16) 122.69(11) N(4)-C(17)-Cr(16) 123.02(12) C(2)-C(3) -C(4) 123.67(11) C(15)-C(16)-C(17) 123.24(12)

56 2.3 Discussion

The insertion of acetonitrile might imply that the Cor-Mesityl bond is relatively polar to allow

for insertion of the nitrile functionality. This suggestion, however, contrasts with the covalent bonding often observed in late transition metal-aryl complexes.4 3 In addition, a polar M-Ar bond would be

strongly basic, and potentially favor a deprotonation pathway rather than insertion. Deprotonation of acetonitrile (pKa (DMSO) = 31.344) induces oligomerization reactions, and has been shown to produce

isolable NacNac metal species derived from trimerization (eq 2).45-" While the fate of the extra equivalent of LiMes has not been determined when synthesizing 1 from Li(THF) 4CoMes 3, it seems reasonable that this may indeed deprotonate and oligomerize acetonitrile. The retention of two mesityl groups in 1 and the high isolated yields from both mesityl precursors strongly suggests that nitrile

insertion is the dominant pathway (eq 3).

M-R 3 CH 3CN M/ CN (2) Deprotonation HN-

M-R 2 CH3CN M (3) Insertion HN- R

Acetonitrile insertion into a late metal-aryl bond has only been observed in a few circumstances: insertion into a palladium ortho-phenol species,51 and insertion into a nickel-mesityl to form a nickel-

52 imine species. In both cases, thallium salts of weakly coordinating anions induce abstraction of a coordinated halide and promote the binding of acetonitrile to the cationic metal complex. In contrast, the

inherent coordinative unsaturation of [Co(mesityl) 2] 2 likely induces binding of acetonitrile to form

(MeCN) 2CoMes 2, which can then undergo metal-mediated insertion and condensation processes.

57 Nitrile insertion into late transition metal-carbon bonds is rarely observed, especially given studies featuring the functional tolerance of nitriles in cross-coupling and acrylonitrile polymerization reactions. In fact, based on the few known examples, it appears that nitrile insertion into late metal-carbon bonds has two critical requirements: A cationic or electron deficient metal center, and an open site for nitrile coordination in a position cis to the alkyl or aryl moiety.

Although a planar structure has been suggested in cobalt bis-NacNac complexes, it is intriguing that no solvent molecules appear to bind in apical positions. Indeed, there is no change in the absorption spectrum in pentane, THF, and acetonitrile solvents. The absence of axial ligands can be rationalized by the electronic structure of 1. Scheme 2.2 outlines a qualitative molecular orbital diagram of 1 by mixing a standard a-only d-orbital splitting diagram for a square planar d7 metal complex (left) with the frontier 7r orbitals of the NacNac framework (right). The a nonbonding orbital (with respect to

2 2 1 the ligand) interacts with a dyz orbital of the metal to change the SOMO from A1 (dz ) symmetry (left) to

2 1 2 an orbital of B2 (dyz) symmetry (center). Thus, the dz orbital is filled in 1, and the cobalt center takes on a planar geometry. That the SOMO is higher in energy than dz2 is consistent with the metal complex being resistant to electrochemical reduction to a Co, species. Orbital population analysis from DFT calculations indicates that the 3dyz orbital of Co is in fact only half occupied, as is indicated pictorially by the spin density plot of 1 (figure 2.7).

58 Scheme 2.2 Qualitative MO diagram illustrating 7c interactions of NacNac ligands with planar Co in 1.

MN4 a-only N4 2-only

dxy ------

4dyz - ,nb

dxz J -- nb dyz , 4 dx2_2 ----- x- nb

IF dyz + ,nb

The changes in electronic structure upon oxidation of 1 are governed by the binding of THF in 2.

Apical coordination of THF in 2 causes a corresponding rise in energy of the dz2 orbital. Thus 2 can be described as possessing a triplet ground state with SOMOs in both dz2 and dyz type orbitals. The 3d orbitals of Co(III) are expected to become increasingly electronegative upon oxidation, resulting in a descent in energy. This causes an increase in ligand character of the dyz SOMO. As a result, the spin density of 2 no longer resides solely on the metal, but also on the ligand (figure 2.14). The spin density value of 1.61 on Co indicates the while the dz2 type SOMO is mostly metal-centered, the dyz SOMO is roughly 60% metal based, 40% ligand based. Although oxidation state formalisms break down when considering redox-active ligands, the Co center in 2 could be best described as having an oxidation state of roughly 2.5.

59 Table 2.5 Change in Co-N, C-N, and C-C Bond Lengths in 1, la, and 2.

Bond 1 and la (avg.) 2 (avg.) Bond Length Change

Co-N 1.862(1) 1.870(2) 0.008(2) C-N 1.330(2) 1.326(2) -0.004(2) C-C 1.398(2) 1.397(2) -0.001(2)

The bond length changes of each molecule in 1, la, and 2 are presented in table 2.4. The Co-N bond lengths have increased by an average of 0.008(2) A, which is too little if the oxidation were to occur at the Co center. The C-N bond lengths of the ligand decrease by an average of 0.004(2) A, an observation predicted by previous DFT calculations.2 3 The C-C bond lengths do not change within error upon oxidation. The lack of significant Co-N bond length change along with the paramagnetic nature of 2 suggests that the cobalt center did not experience a formal change in oxidation state upon oxidation of 1 to 2. Further support for the oxidation of the NacNac ligand framework is provided by the intense visible absorptions and calculated spin density. As predicted by simple molecular orbital theory, the oxidation of a NacNac ligand displays no significant structural change.

Undoubtedly, one feature of 2 that allows isolation of an oxidized NacNac species is the presence of not one, but two coplanar NacNac ligands. If the ligand based radical were localized on only one at a time, structural changes could be unobservable due to disorder in the crystal structure. The presence of a low energy absorption in the NIR suggests that this is not the case. Electronic communication between the ligands is conveyed through the dyz orbital of the metal center, so that the molecule can be described as a class III delocalized radical cation.56 Such behavior has been observed in an electronically related

(salen)Ni* species."

60 2.4 Conclusions

Since NacNac does not change structure upon oxidation, the presence of oxidized NacNac moieties can only be discerned through spectroscopic and calculation methods. Due to their ubiquitous presence in the literature, there is a possibility that the nature of some metal NacNac species were previously misidentified. In these cases one would expect a metal in a formal oxidation state one level higher than the structure and reactivity warranted. While it is difficult to suggest redox behavior in these

species without supporting calculations, there are some recently reported NacNac species that contain unusual electronic and spectroscopic properties.'' 9

Scheme 2.3 Highest Occupied Molecular Orbitals of NacNac, Salen, and Porphyrin.

bis-NacNac b1 bis-NacNac b2

Salen b1 Salen b2

Porphyrin a2u Porphyrin alu

61 Acac complexes are expected to behave similarly to NacNac congeners, but the 7n non-bonding orbital is much less covalent, so it should be more difficult to oxidize. Cotton's unimolecular

(tBu 2acac)2Ni species is perhaps a prime example where oxidation might produce an acac radical species. Although this may prove too reactive to characterize structurally, acac complexes with significant spin density on the ligand do exist."

Salen complexes have a similar 7r system (scheme 2.3), and indeed redox non-innocence has been

observed in some nickel, palladium, and platinum species.57,61-63 In these cases, the N 20 2 coordination environment of the salen lowers the symmetry of noninnocent 7r orbital so that small but observable bond length changes do occur upon oxidation. On the other hand, cobalt salen complexes tend to coordinate additional ligands axially, raising the dZ orbital in energy above dy. This axial coordination and the more electropositive nature of earlier transition metals likely stifles significant salen based redox-activity in other species.

Depending on the presence of certain meso substituents, the porphyrin HOMO could possess either ai, or a2u symmetry (scheme 2.3). While the ai, orbital would generate observable bond length

changes if oxidized, the a2 u orbital, which is related to the HOMO of the bis-NacNac ligand framework of

1 and 2, would not. In fact, complexes featuring porphyrin radical cations often show few structural changes relative to the neutral systems, though there often exists a strong tendency for intermolecular interactions in the solid state.64 Nevertheless, porphyrins are likely another class of molecules that possess

"hidden noninnocence."

2 1 represents a classic case where the ground state of the low-spin cobalt (II) molecule is B2 rather

2 than A1 due to 7c donation from the ligands. The interaction between these two states strongly dictates the coordination and reactivity of the molecule. 1 does not bind ligands axially, and this feature likely shuts down its reactivity with oxygen as well as electrochemical reduction to Co(I). Given the importance of planar cobalt species for mediating radical polymerization,65 electrocatalytic hydrogen production, 66' 67 and reversible oxygen transport,68-70 understanding and modulating the interaction of these two low lying

62 electronic states in each catalyst is paramount for future ligand design and improvement of these catalytic reactions.

63 2.5 Experimental Section

2.5.1 General Considerations

All reactions were performed in a nitrogen filled glovebox or by using standard air-free schlenk

tedhniques. Solvents were dried by passage through a column of activated alumina and stored over 4 A

5 7 molecular sieves. Li(THF) 4 CoMes3 , [CoMes2] 2, and CP2FeBAr ' were prepared according to literature

procedures. Cp 2FePF6 was purchased from Sigma-Aldrich. Elemental analyses were performed by

Midwest Microlab LLC.

2.5.2 Physical Methods

IR spectra of powdered samples were recorded on a PerkinElmer Spectrum 400 FT-IR/FT-NIR

Spectrometer outfitted with a Pike Technologies GladiATR attenuated total reflectance accessory with a monolithic diamond crystal stage and pressure clamp. NMR and EPR spectra were recorded at the MIT

Department of Chemistry Instrumentation Facility on a Varian Mercury 300 NMR Spectrometer and a

Bruker EMX EPR spectrometer (X band, G=9.860 GHz). Magnetic data were collected using a Quantum

Design MPMS-XL SQUID magnetometer. Measurements for 1 were obtained for microcrystalline sample restrained in a frozen eicosane matrix within a polycarbonate capsule. Dc susceptibility measurements were collected in the temperature range 2-300 K under a dc field of 1000 Oe. Dc magnetic susceptibility data were corrected for diamagnetic contributions from the sample holder and eicosane, as well as for the core diamagnetism of each sample (estimated using Pascal's constants). Cyclic voltammetry experiments were performed in a nitrogen filled glovebox using dry solvents containing 0.1 M Bu 4NPF6 as a supporting electrolyte. A three compartment cell was employed possessing a glassy carbon electrode as the working electrode, Pt wire as the auxiliary electrode, and Ag/AgCl as a reference electrode. CVs were monitored with scan rates of 10-100 mV/s employing iR compensation, and referenced to Cp 2Fe/Cp2Fe*.

64 2.5.3 X-ray Crystallographic Details

Single crystals of 1 and la were obtained by cooling a concentrated solution of ether and acetonitrile respectively to -40 'C. Single crystals of 2 were obtained by vapor diffusion of hexane into a concentrated THF solution at room temperature. Crystals were removed from the supernatant liquid and transferred onto a microsocope slide covered in Paratone N oil. Selected crystals were affixed to a

MiTeGen MicroMounts polyimide tip and mounted on a Bruker three circle goniometer platform equipped with an APEX detector. A graphite monochromator was employed for wavelength selection of the Mo Ka radiation () = 0.71073 A). The data were processed and refined using the program SAINT supplied by Siemens Industrial Automation. Structures were solved by direct methods in SHELXS and refined by standard difference Fourier techniques in the SHELXTL program suite (6.10 v., Sheldrick G.

M., and Siemens Industrial Automation, 2000). Hydrogen atoms were placed in calculated positions using the standard riding model and refined isotropically. Hydrogen atoms bound to nitrogen were clearly visible in the difference map and refined freely. Refinement of 1 yielded two crystallographically independent molecules, although they were nearly identical to each other. In 2, the four carbon atoms of the THF molecule that is involved in hydrogen bonding are disordered and were modeled using the part command. The minor partition (26%) was restrained with simu and delu commands. The PF6 ion was also disordered and the minor partition (7%) was modeled isotropically. Unit cell parameters, morphology, and solution statistics for the structures of 1, la, and 2 are summarized in Table 4.5.

2.5.4 Computational Methods

Density functional theory (DFT) calculations were performed with the hybrid functional Becke-3 parameter exchange functional2-14 and the Lee-Yang-Parr nonlocal correlation functional (B3LYP)75 as implemented in the Gaussian 03, Revision B.05 software package. 76 The triple-( basis sets with one-set of polarization functions (TZVP) 77 were used for cobalt, nitrogen, and oxygen atoms, and the double-( basis sets with one-set of polarization functions (SVP) 7 8 were used on the carbon and hydrogen atoms. The

65 calculations were performed on truncated models of 1 and 2 where the mesityl groups were replaced with methyl groups. All geometries were confirmed as local minima structures by calculating the Hessians and checking that no negative eigenvalues were present. TD-DFT calculations were performed to calculate the first 50 excited states of each model. Cartesian coordinates of the optimized geometries and absolute energies are provided in section 2.5.8.

2.5.5 Preparation of Co(1-mesitylbutane-1,3-diimine) 2 (1).

(a) Li(THF) 4 CoMes 3 (100 mg, 0.14 mmol) was dissolved in 5 mL of THF and cooled to -40 'C.

5 mL of acetonitrile was added dropwise. After standing for 12 h, the solution was warmed to room temperature and then heated to 70 *C for 12 h. The reaction was cooled and the solvent was removed by vacuum. A dark brown semicrystalline solid was extracted with 3 times with 5 mL of diethyl ether, filtered, reduced in volume to 2 mL, and placed in a freezer at -40 'C. Two crops of crystals were collected to give 57 mg of 1 as an orange-yellow crystalline solid (88%). 1H NMR (300 MHz, CD 6): 6 -

4.64 (s, 2H), 1.83 (s, 3H), 0.09 (br. s., 6H), -37 (br. s. 3H). per = 2.4(1) BM (Evans' method,31'79 CD,

18 'C). UV-vis (pentane): max, nm (s, M'cm ') 264 (17000), 298 (20600), 377 (12700), 417 (3600). IR

(paratone-n, N-H, cmf): 3297, 3273. Anal. Calcd (Found) for C26H34 N4 Co: C, 67.7 (67.7); H, 7.4 (7.5);

N 12.1 (12.1).

(b) The procedure is identical to (a) except that [CoMes 2] 2 (50 mg, 0.084 mmol) was used to yield

27 mg of 1 (70%).

2 CH 3CN. 'H NMR (300 MHz, C6D6): 6 = 4.64 (s, 2H), 1.83 (s, 3H), 0.09 (br. s., 6H). H NMR (300 MHz,

C6D6): 6 = -36.8 (s. 3H). IR (paratone-n, N-H): 2451 cmf.

66 2.5.6 Preparation of [Co(1-mesitylbutane-1,3-diimine) 2(THF)2][PF6] (2)

2 25 mg (0.054 mmol) of 1 was dissolved in 3 mL diethyl ether. Solid Cp2FePF6 (18 mg, 0.054 mmol) was added and the resulting green solution was stirred for 2 h. The solution was the filtered, and hexane was added to precipitate dark blue crystals of 2. The crystals were washed with ether and dried under vacuum to yield 37 mg of 2 (91%). 1H NMR (300 MHz, C6D6 ): 8 = 6.42 (s. 3H), 5.79 (s. 3H), 5.19

(s. 3H), 1.32 (s, 1H), 0.42 (s, 1H), -77.44 (br.s. 3H). peff = 3.2 BM (Evans' method, C6D6, 18 *C). UV- vis-NIR (THF): ma, nm (s, M-Tcm-1) 740 nm, 1264 nm. IR (paratone-n, N-H, cm'): 3330, 3266. Anal.

Calcd (Found) for C34H5oN40 2F6PCo: C, 54.4 (54.6); H, 6.7 (6.6); N, 7.5 (7.5).

2.5.7 Preparation of Fe(1-mesitylbutane-1,3-diimine) 2 (3)

Precooled acetonitrile (0.5 mL) was added to a solution of [Fe(mesityl) 2] 2 in diethyl ether at -40

'C. The solution was kept in the dark at -40 'C for 10 days at which point dark brown crystals had separated from solution along with an insoluble dark black powder. Upon warming or brief exposure to light, the crystals decomposed to a viscous oil and a black powder. 'H NMR (C6 D6 ) showed multiple diamagnetic products, as well as several broad peaks indicating one or more paramagnetic species.

67 2.5.8 X-ray Crystallographic Tables

Table 2.6 X-ray Crystallographic Table for 1, la, and 2.

1 la 2

Formula C H CoN 26 34 4 C30H40CoN 6 C 38H 58CoF6N 4O 3P fw, g/mol 461.50 543.61 822.78 Temperature 100 K 100 K 100 K

cryst. syst. Triclinic Triclinic Monoclinic

space group P-1 P-1 P2 1/n

color Orange Brown Blue

a (A) 9.0455(8) 8.5629(5) 14.4724(13) b (A) 11.5983(10) 8.6093(5) 15.0318(13) c (A) 11.9082(10) 10.7283(6) 19.4774(17) a (deg) 87.059(1) 110.247(1) 90 p (deg) 73.004(1) 97.545(1) 106.142(2) y (deg) 89.521(1) 96.697(1) 90 V (k) 1193.15(18) 724.35(7) 4070.2(6) Z 2 1 4 no. refl. 26851 16251 81202 unique refl. 6675 4026 11469

Rint 0.0400 0.0348 0.0672 Ria (all data) 0.0423 0.0362 0.0839 wR2b(all data) 0.0939 0.0863 0.1111 RI [(I> 2a)] 0.0335 0.0325 0.0425 wR2 [I > 2a)] 0.0882 0.0863 0.0940 GOFC 1.037 1.049 1.005 2 2 2 2 2 1 2 2 aRI = ElIFo - |Fe||/E|F. bwR2 = (E(w(F - F ) )/E(w(F0 ) )) . cGOF = (I w(F _ 1 2 Fe22/1(n -p)) / where n is the number of data and p is the number of parameters refined.

68 Table 2.7 X-ray Crystallographic Table for 3.

Formula C 26H 34 FeN 4 fw, g/mol 458.42

Temperature 100(2) K cryst. syst. Triclinic space group P-1 color Dark Brown a (A) 9.0592(6) b (A) 11.6280(7) c (A) 11.9161(7) a (deg) 86.5470(10) p (deg) 73.4000(10) y (deg) 88.8570(10) V(A 3 ) 1200.74(13) Z 2 no. refl. 32199 unique refl. 6739

Rint 0.0330 R1' (all data) 0.0387 wR2b(all data) 0.1003 RI [(I> 2y)] 0.0321 wR2 [I> 2a)] 0.0963

GOFc 1.048 2 2 2 2 2 1 2 2 aRI = Y|IF 0| - |Fc||/YIF. bwR 2 = (I(w(F0 - F ) )/E(w(F ) )) / . cGOF = (Z w(F _ Fe22/1(n -p))"2 where n is the number of data and p is the number of parameters refined.

69 2.5.9 Cartesian Coordinates for Optimized Structures

Table 2.8 Cartesian coordinates for the DFT optimized geometry for the S = 1/2 model of 1 (E = -1993.63267829 hartrees).

Atom x y z

Co 0.000000 -0.000007 -0.000138 N 1.340451 1.334989 0.000260 N -1.340460 -1.335012 0.000048 N -1.340459 1.334994 -0.000427 N 1.340481 -1.334996 -0.000401 C -2.664733 -1.237190 0.000142 C -3.477649 2.514542 0.000140 C -2.664718 1.237200 -0.000190 C -3.328933 -0.000002 0.000065 C -3.477655 -2.514537 0.000254 C 2.664740 1.237186 0.000219 C 2.664732 -1.237195 -0.000261 C 3.477659 -2.514537 0.000190 C 3.477619 2.514567 0.000197 C 3.328941 0.000010 -0.000017 H -4.557425 2.313208 -0.002700 H -3.245617 3.123112 0.890595 H -3.241458 3.126382 -0.886957 H 4.557434 -2.313211 -0.002909 H 3.245802 -3.122901 0.890839 H 3.241288 -3.126597 -0.886702 H 1.032005 2.306484 0.000512 H -1.032023 -2.306509 0.000151 H -1.031988 2.306479 -0.000750 H 1.032016 -2.306484 -0.000590 H -4.418370 0.000002 0.000191 H -3.243897 -3.124571 -0.888738 H -4.557436 -2.313210 0.000730 H -3.243165 -3.124918 0.888817 H 3.244926 3.123775 -0.889651 H 4.557407 2.313290 0.002153 H 3.242021 3.125763 0.887897 H 4.418378 0.000007 0.000041

70 Table 2.9 Cartesian coordinates for the DFT optimized geometry for the S = 3/2 model of 1 (E = -1993.63221780 hartrees).

Atom x y z

0.000000 -0.000006 -0.000013 -1.375994 1.019210 -1.014164 1.376009 -1.014649 -1.018725 1.375977 1.014615 1.018759 -1.375994 -1.019283 1.014084 2.691496 -0.891887 -0.895254 3.574384 1.762591 1.769583 2.691469 0.891922 0.895259 3.319376 0.000043 -0.000013 3.574440 -1.762568 -1.769538 -2.691483 0.895608 -0.891552 -2.691483 -0.895614 0.891545 -3.574410 -1.769382 1.762759 -3.574412 1.769405 -1.762738 -3.319376 0.000021 0.000019 4.643533 1.589338 1.586961 3.363125 2.829894 1.587537 3.371846 1.566785 2.836172 -4.643572 -1.592331 1.583888 -3.366289 -2.836086 1.573698 -3.368729 -1.580581 2.830005 -1.120979 1.718422 -1.710266 1.121002 -1.711063 -1.717627 1.120951 1.710993 1.717691 -1.120974 -1.718536 1.710142 4.409228 0.000076 -0.000029 3.363481 -2.829873 -1.587152 4.643585 -1.589008 -1.587186 3.371627 -1.567107 -2.836138 -3.366278 2.836102 -1.573652 -4.643573 1.592360 -1.583860 -3.368744 1.580627 -2.829990 -4.409228 0.000051 0.000050

71 Table 2.10 Cartesian coordinates for the DFT optimized geometry for the S = 1 model of 2 (E = -2225.80065097 hartrees).

Atom x y z

0.000008 -0.62069 0.000007 1.336965 -0.81546 -1.33031 -1.33695 -0.81545 1.330331 -1.3322 -0.76152 -1.34602 1.332222 -0.76145 1.34604 -2.66008 -0.90474 1.225719 -3.46879 -0.96014 -2.51749 -2.65341 -0.84258 -1.25188 -3.31622 -0.87159 -0.01296 -3.47053 -1.10093 2.48478 2.660102 -0.90475 -1.2257 2.653425 -0.84251 1.251901 3.46881 -0.96003 2.517519 3.470553 -1.10096 -2.48475 3.316242 -0.87157 0.012984 0.001423 2.409299 -1.19359 -0.00148 2.409325 1.193526 -0.31095 3.820513 0.701548 0.310871 3.820502 -0.70164 -2.6E-05 1.581815 -2. 1E-05 1.026233 -0.90676 -2.29647 -1.02621 -0.90673 2.29649 -1.01784 -0.78493 -2.31484 1.017862 -0.78482 2.314856 -4.42476 -0.42581 -2.42347 -2.93105 -0.56273 -3.39116 -3.70289 -2.01985 -2.71461 -4.40386 -0.93364 -0.01796 -3.53986 -2.17649 2.720687 -4.49456 -0.722 2.365489 -3.00984 -0.59588 3.347466 4.424766 -0.42568 2.423477 2.931059 -0.5626 3.391175 3.702928 -2.01972 2.71467 3.539845 -2.17652 -2.72066 4.494592 -0.72207 -2.36546 3.009876 -0.5959 -3.34744 4.403884 -0.93361 0.017987 -0.99482 2.342659 -1.66367

72 H 0.752308 2.005595 -1.88858 H 0.994763 2.342708 1.663608 H -0.75235 2.005626 1.888535 H 0.105334 4.594258 1.362255 H -1.40006 3.976376 0.641302 H 1.399985 3.976381 -0.6414 H -0.10542 4.594227 -1.36236

73 2.6 References

(1) Marcus, R. A. J Chem. Phys. 1956, 24, 966-978.

(2) Marcus, R. A. Disc. FaradaySoc. 1960, 29, 21.

(3) Marcus, R. A. J Chem. Phys. 1965, 43, 679-701.

(4) Marcus, R. A.; Sutin, N. Biochimica et Biophysica Acta. 1985, 811, 265-322.

(5) Blackmore, K. J.; Ziller, J. W.; Heyduk, A. F. Inorg. Chem. 2005, 44, 5559-5561.

(6) Myers, T. W.; Berben, L. A. Inorg. Chem. 2012, 51, 1480-1488.

(7) Green, M. L. H. J. Organomet. Chem. 1995, 500, 127-148.

(8) Chang, C. J.; Chng, L. L.; Nocera, D. G. J. Am. Chem. Soc. 2003, 125, 1866-1876.

(9) McGuire, R.; Dogutan, D. K.; Teets, T. S.; Suntivich, J.; Shao-Horn, Y.; Nocera, D. G. Chem. Sci. 2010, 1, 411-414.

(10) Soper, J. D.; Kryatov, S. V.; Rybak-Akimova, E. V.; Nocera, D. G. J. Am. Chem. Soc. 2007, 129, 5069-5075.

(11) Dogutan, D. K.; McGuire, R.; Nocera, D. G. J. Am. Chem. Soc. 2011, 133, 9178-9180.

(12) Dogutan, D. K.; Stoian, S. A.; McGuire, R.; Schwalbe, M.; Teets, T. S.; Nocera, D. G. J. Am. Chem. Soc. 2011, 133, 131-140.

(13) Schwalbe, M.; Dogutan, D. K.; Stoian, S. A.; Teets, T. S.; Nocera, D. G. Inorg. Chem. 2011, 50, 1368-1377.

(14) Bachmann, J.; Nocera, D. G. J. Am. Chem. Soc. 2004, 126, 2829-2837.

(15) Bachmann, J.; Nocera, D. G. J. Am. Chem. Soc. 2005, 127, 4730-4743.

(16) Bachmann, J.; Nocera, D. G. Inorg. Chem. 2005, 44, 6930-6932.

(17) Bachmann, J.; Hodgkiss, J. M.; Young, E. R.; Nocera, D. G. Inorg. Chem. 2007, 46, 607-609.

(18) Radosevich, A. T.; Melnick, J. G.; Stoian, S. A.; Bacciu, D.; Chen, C.-H.; Foxman, B. M.; Ozerov, 0. V.; Nocera, D. G. Inorg. Chem. 2009, 48, 9214-922 1.

74 (19) Kaim, W. Inorg. Chem. 2011, 50, 9752-9765.

(20) Chirik, P. Inorg. Chem. 2011, 50, 9737-9740.

(21) Scarborough, C. C.; Sproules, S.; Weyhermueller, T.; DeBeer, S.; Wieghardt, K. Inorg. Chem. 2011, 50, 12446-12462.

(22) Bourget-Merle, L.; Lappert, M. F.; Severn, J. R. Chem. Rev. 2002, 102, 3031-3065.

(23) Khusniyarov, M. M.; Bill, E.; Weyhermuller, T.; Bothe, E.; Wieghardt, K. Angew. Chem. Int. Ed. 2011, 50, 1652-1655.

(24) Seidel, W.; Burger, I. Z Chem. 1977, 17, 31.

(25) Theopold, K. H.; Silvestre, J.; Byrne, E. K.; Richeson, D. S. Organometallics 1989, 8, 2001-2009.

(26) Bercaw, J. E.; Davies, D. L.; Wolczanski, P. T. Organometallics 1986, 5, 443-450.

(27) Barroso, S.; Cui, J.; Carretas, J. M.; Cruz, A.; Santos, I. C.; Duarte, M. T.; Telo, J. P.; Marques, N.; Martins, A. M. Organometallics2009,28, 3449-3458.

(28) Richeson, D. S.; Mitchell, J. F.; Theopold, K. H. J. Am. Chem. Soc. 1987, 109, 5868-5870.

(29) Richeson, D. S.; Mitchell, J. F.; Theopold, K. H. Organometallics 1989, 8, 2570-2577.

(30) Cotton, F. A.; Wise, J. J. Inorg. Chem. 1966, 5, 1200-&.

(31) Evans, D. F. J. Chem. Soc. 1959, 2003-2005.

(32) Dugad, L. B.; Marathe, V. R.; Mitra, S. Chem. Phys. Lett. 1985, 120, 239-242.

(33) Everett, G. W.; Holm, R. H. J. Am. Chem. Soc. 1965, 87, 5266-&.

(34) Everett, G. W.; Holm, R. H. J. Am. Chem. Soc. 1966, 88, 2442-&.

(35) Everett, G. W.; Holm, R. H. Inorg. Chem. 1968, 7, 776-&.

(36) Wolny, J. A.; Rudolf, M. F.; Ciunik, Z.; Gatner, K.; Wolowiec, S. J Chem. Soc. Dalton Trans. 1993, 1611-1622.

(37) Hoffman, B. M.; Basolo, F.; Diemente, D. L. J. Am. Chem. Soc. 1973, 95, 6497-6498.

(38) Busetto, C.; Cariati, F.; Fusi, A.; Gullotti, M.; Morazzon.F; Pasini, A.; Ugo, R.; Valenti, V. J Chem. Soc. Dalton Trans. 1973, 754-765.

75 (39) Malatesta, V.; McGarvey, B. R. Can. J. Chem. 1975, 53, 3791-3800.

(40) Urbach, F. L.; Bereman, R. D.; Topich, J. A.; Harihara, M; Kalbache, B; J. Am. Chem. Soc. 1974, 96, 5063-5069.

(41) Hitchman, M. A. Inorg. Chem. 1977, 16, 1985-1993.

(42) La Mar, G. N.; Walker, F. A. J Am. Chem. Soc. 1973, 95, 1790-1796.

(43) Crabtree, R. H. The OrganometallicChemistry of the TransitionMetals, 3rd Edition; 3rd ed.; Wiley-Interscience, 2000.

(44) Matthews, W. S.; Bares, J. E.; Bartmess, J. E.; Bordwell, F. G.; Cornforth, F. J.; Drucker, G. E.; Margolin, Z.; McCallum, R. J.; McCollum, G. J.; Vanier, N. R. J. Am. Chem. Soc. 1975, 97, 7006- 7014.

(45) Michelin, R. A.; Mozzon, M.; Bertani, R. Coord. Chem. Rev. 1996, 147, 299-33 8.

(46) Antunes, M. A.; Dias, M.; Monteiro, B.; Domingos, A.; Santos, I. C.; Marques, N. Dalton Trans. 2006.

(47) Kopp, M. R.; Neumller, B. Z. anorg.allgem. Chem. 1999, 625, 1246-1248.

(48) Nishijo, J.; Nishi, N. Eur. J. Inorg. Chem. 2006, 2006, 3022-3027.

(49) Redshaw, C.; Jana, S.; Shang, C.; Elsegood, M. R. J.; Lu, X.; Guo, Z. X. Chem. Comm. 2010, 46.

(50) Yuan, J.; Hughes, R. P.; Rheingold, A. L. Organometallics2009, 28, 4646-4648.

(51) Vicente, J.; Bad, J. A.; Lopez-Saez, M. J.; Jones, P. G. Angew. Chem. Int. Ed. 2005, 44, 6001- 6004.

(52) Ceder, R. M.; Muller, G.; Ordinas, M.; Ordinas, J. I. Dalton Trans. 2007, 83-90.

(53) McGeachin, S. Can. J Chem. 1968, 46, 1903-&.

(54) Hitchcock, P. B.; Lappert, M. F.; Liu, D. S. J. Chem. Soc. Chem. Comm. 1994, 1699-1700.

(55) Hitchcock, P. B.; Lappert, M. F.; Linnolahti, M.; Sablong, R.; Severn, J. R. J. Organomet. Chem. 2009, 694, 667-676.

(56) Robin, M. B.; Day, P. In Adv. Inorg. Chem.; Academic Press, 1968; Vol. Volume 10, pp. 247-422.

(57) Storr, T.; Wasinger, E. C.; Pratt, R. C.; Stack, T. D. P. Angew. Chem. Int. Ed. 2007, 46, 5198- 5201.

76 (58) Conradie, J.; Ghosh, A. J Chem. Theo.Comp. 2007, 3, 689-702.

(59) Shaffer, D. W.; Ryken, S. A.; Zarkesh, R. A.; Heyduk, A. F. Inorg. Chem. 2011, 50, 13-2 1.

(60) Das, A.; Scherer, T. M.; Chowdhury, A. D.; Mobin, S. M.; Kaim, W.; Lahiri, G. K. Inorg. Chem. 2012, 51, 1675-1684.

(61) Shimazaki, Y.; Yajima, T.; Tani, F.; Karasawa, S.; Fukui, K.; Naruta, Y.; Yamauchi, 0. J Am. Chem. Soc. 2007, 129, 2559-2568.

(62) Shimazaki, Y.; Stack, T. D. P.; Storr, T. Inorg. Chem. 2009, 48, 8383-8392.

(63) Kurahashi, T.; Fujii, H. J Am. Chem. Soc. 2011, 133, 8307-83 16.

(64) Kadish, K. M.; Smith, K. M.; Guilard, R. The Porphyrin Handbook: Inorganic, organometallic and coordinationchemistry; Elsevier, 1999.

(65) Debuigne, A.; Poli, R.; Jerome, C.; Jerome, R.; Detrembleur, C. Prog.Poly. Sci. 2009, 34, 211- 239.

(66) Dempsey, J. L.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Acc. Chem. Res. 2009, 42, 1995- 2004.

(67) Artero, V.; Chavarot-Kerlidou, M.; Fontecave, M. Angew. Chem. Int. Ed. 2011, 50, 723 8-7266.

(68) Crumblis.Al; Basolo, F. Science 1969, 164, 1168-&.

(69) Crumblis.Al; Basolo, F. J. Am. Chem. Soc. 1970, 92, 55-&.

(70) Tine, M. R. Coord. Chem. Rev. 2012, 256, 316-327.

(71) Le, B. J.; Jiao, H.; Meyer, W. E.; Hampel, F.; Gladysz, J. A. J Organomet. Chem. 2000, 616, 54- 66.

(72) Becke, A. D. Phys. Rev. A. 1988, 38, 3098-3 100.

(73) Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377.

(74) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.

(75) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B. 1988, 37, 785-789.

(76) Gaussian 03, Revision B.05; Frisch, M.; Trucks, G.; Schlegel, H.; Scuseria, G.; Robb, M.; Cheeseman, J.; Montgomery, J.; Vreven, T.; Kudin, K.; Burant, J.; Millam, J.; Iyengar, S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G.; Nakatsuji, H.;

77 Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, 0.; Nakai, H.; Klene, M.; Li, X.; Knox, J.; Hratchian, H.; Cross, J.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R.; Yazyev, 0.; Austin, A.; Cammi, R.; Pomelli, C.; Ochterski, J.; Ayala, P.; Morokuma, K.; Voth, G.; Salvador, P.; Dannenberg, J.; Zakrzewski, V.; Dapprich, S.; Daniels, A.; Strain, M.; Farkas, 0.; Malick, D.; Rabuck, A.; Raghavachari, K.; Foresman, J.; Ortiz, J.; Cui, Q.; Baboul, A.; Clifford, S.; Cioslowski, J.; Stefanov, B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, .; Martin, R.; Fox, D.; Keith, T.; Laham, A.; Peng, C.; Nanayakkara, A.; Challacombe, M.; Gill, P.; Johnson, B.; Chen, W.; Wong, M.; Gonzalez, C.; Pople, J. Gaussian, Inc., Pittsburgh PA, 2003.

(77) Schafer, A.; Huber, C.; Ahlrichs, R. J Chem. Phys. 1994, 100, 5829-5835.

(78) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-2577.

(79) Schubert, E. M. J Chem. Educ. 1992, 69, 62.

78 Chapter 3: First Row Transition Metal Siloxide Chemistry of the "DTBMS" Ligand

79 3.1 Introduction

In pursuit of an alternate platform to stabilize pseudotetrahedral metal oxo complexes, work in

our lab has explored the use of alkoxide ligands that can provide both trigonal and pseudotetrahedral

coordination environments.1 - To this end, a number of molecules have been prepared including both

V(ditox) 3 (ditox = 'Bu 2MeCO-), a reduced species that has been shown to bind dinitrogen, as well as an

oxidized complex, Cr(O)(ditox) 3. As opposed to what was anticipated for the Cr(V) pseudotetrahedral d' metal species, the reactivity and electronic structure of Cr(O)(ditox) 3 suggest that significant spin density resides on the oxo unit. This is believed to be a result of n donation from the oxygen atom of the alkoxide, which causes mixing between the ligand field orbitals of e symmetry in the C 3v coordination environment.

In pursuit of an alternate ligand platform that could stabilize pseudotetrahedral metal oxos complexes, the chemistry of the analogous ditertbutylmethylsiloxide (DTBMS) ligand was explored.

3.1.1 Electronic Considerations

As an anionic monodentate ligand, the bonding between siloxides and transition metals can be compared to that of alkoxide, amide, and alkyl species.'' Ligand field theory provides a basis for understanding the differences in these metal-ligand interactions by describing the stabilization energy both in ionic terms from crystal field theory and covalent terms from molecular orbital theory.' The outcome is a qualitative method that predicts the magnitude of metal-ligand interactions based on the spatial overlap of metal and ligand orbitals (the angular overlap model) and their difference in energy.

Comparing the interaction in M-C, M-N, and M-O bonds, one finds that the spatial overlap of orbitals increases across this series due to the decreasing covalent radius, although this effect is more than compensated for by the increasing difference in orbital energies. As a result, the metal-ligand interaction energy follows the trend M-C > M-N > M-0, as the bonding becomes increasingly ionic. Since M-C bonds are predicted to provide the strongest ligand fields, it follows that they will also be the most

80 reducing and could form stable complexes in unusually high oxidation states. This is precisely what is

observed in complexes of the type M(1-norboMyl) 4 which generates extremely high formal oxidation

states (M = Coy, CoIv, FeIV, MnIV, Crv)."'9 In contrast, metal alkoxides form complexes with

comparatively weak ligand fields and stabilize unusually reduced species. This strategy has been

exploited by Rothwell and Wolczanski to stabilize strongly reducing metal complexes. 0- 2 In between are

nitrogen based ligands, which have been shown to stabilize metal complexes in both high and low

oxidation states. This unique electronic property of M-N bonding, when carefully balanced, has permitted

the isolation and characterization of catalytically active Mo species in four different oxidation states with the same ligand coordination environment.13

To compare the electronic aspects of alkoxide and siloxide ligands, Wolczanski has suggested that an electropositive silicon nucleus makes the M-O bond more ionic than found in alkoxides." In

support of this argument, Krempner has drawn on crystallographic data to show that between similar

complexes, the M-O and Si-O bond lengths are inversely related, while the M-Si distance remains

remarkably constant.14 This is undoubtedly due to the electronic repulsion between the metal and silicon,

both of which contain partial positive charges.

For the same electrostatic reasoning outlined above, it has been argued that the M-O-Si bond

angle is not a reliable indicator of strong M-O bonding.14 The bond angle can be greatly influenced by

steric crowding about the metal and is frequently independent of M-O bond length. The M-O-Si angle is

related, however, to the number of 7r bonds the metal is capable of accepting. An angle greater than 1700 typically indicates that both lone pairs on oxygen can form a bonding interaction with the metal center.

For M-O-Si angles less than 1700, steric effects become more pronounced, complicating attempts to

distinguish intermediate levels of i donation.

81 M* -OSiR3 M=O- *SiR3 "pseudohalide" "masked oxo"

The M-O bonding in siloxides can be described by the two resonance extremes shown above, such that the siloxide behaves like a pseudo-halide, or that the silyl group masks a pseudo-oxo. The first description typically describes bonding in reduced early transition metal siloxides, while the oxo description can describe siloxides bound to more oxidizing metals. Transition metals with siloxide ligands are more electrophilic than the corresponding alkoxide complexes due to this electrostatic "oxide tug-of- war" between metal and silicon. Therefore, comparison of the M-O and Si-O bond distances can give insight into the electronic behavior of the metal complex.

3.1.2 Geometric Considerations

Siloxides have the potential to coordinate transition metals in a number of different geometries, depending mostly on the steric size of the ligand. As monodentate ligands, the steric bulk of siloxides can be described using cone angle approximations similar to those performed by Tolman on phosphine

5 6 ligands.1 ', ,' One complicating factor in determining the steric bulk of a ligand is the variation in the M-

O bond distance. Siloxides have a further complication in that variation can also occur in the O-Si bond

length and the M-O-Si bond angle. Use of the M-Si distance may permit a more accurate cone angle

measurement, however variation in this measurement between metal complexes complicates the

quantification of steric parameters.' 4

Although the steric properties of siloxide ligands are far from universal, their coordination

properties can often be predicted by analysis of the maximum number of siloxides that are sterically

permitted to coordinate to a metal. This maximum steric coordination number leads to the quantization of

the properties of siloxide ligands into several steric groups. Typically, when the siloxide to metal

stoichiometry is maximized, mononuclear species are formed. When this coordination number is less than

82 the maximum allowed, complexes tend to aggregate and the siloxides will often bridge multiple metal centers.

The bulkiest steric group includes the "bowl shaped" (tris(2,2",6,6"-tetramethyl-m-terphenyl-5'- yl)siloxide (TRMS), which forms the mononuclear bis-siloxide, Zr(TRMS) 2(CH 2Ph) 2L2 even in the presence of excess TRMS-H."'" The somewhat smaller tBu 3SiO- (silox) ligand is known to coordinate to transition metals with a maximum of three ligands around the metal. This unique property has led to the isolation of many stable species with trigonal planar geometry. Coordination of only two silox ligands to chromium leads to a dimeric species that still features trigonal chromium atoms, but the siloxides

coordinate in both terminal and bridging modes.1 9

tBu tBu Ar Ar tBu tBu tBu II I 0 0 0

TRMS Silox DTBMS

In contrast to the lower coordination geometries, tetrasiloxide complexes are fairly common. This

is undoubtedly related to the fact that many complexes of the type M(OSiR 3)4 (M = Ti, V, Cr) are

coordinatively saturated, and therefore mononuclear neutral complexes can often be isolated without

evidence of aggregation. The popular Ph3 SiO- ligand typically forms mononuclear tetrahedral species of

TiLV, VI, and VIv 20 ,2 1 but in at least one case, it was found in a pentacoordinate trigonal bipyramidal

2 2 complex, Ta(OSiPh 3). Siloxide ligands with very low steric requirements such as trimethylsiloxide

often lead to aggregation, and coordination numbers as high as 6 have been observed in rare earth

aggregates. 23,24

83 The isolation of a mononuclear pseudotetrahedral metal oxo usually requires a ligand system that can stabilize both three and four coordinate metal centers without the formation of aggregates. While there is much precedent to perform oxygen atom transfer reactions to and from M(silox)3 species (M = V,

25 Nb, Ta, Mo, W), the closely related 'Bu 2MeSiO~ (DTBMS) ligand was chosen for comparison to the

work in our group with the ditox ligand. Na(DTBMS) was prepared according to the literature procedure

t by treating 'Bu 2SiHCl with MeLi to form Bu 2MeSiH, which was then oxidized with KOH to form

t 16 Bu 2MeSiOH (DTBMS-H). The sodium salt is prepared by refluxing a toluene solution of DTBMS-H

with sodium, and it is conveniently crystallized from pentane.

3.1.3 Synthetic Considerations

Whereas siloxide ligands are generally robust ligands, even in a number of extreme

circumstances, 26,27 they are susceptible to specific degradation pathways. Nucleophilic attack at the silicon

center can result in transfer of the oxygen atom to the metal with loss of R3Si(Nuc). For example, reaction

28 of R3SiONa with MX, can yield MOX- 2, NaX, R3SiX (X = halide). In addition, bis-siloxide fragments,

29 M(OSiR 3)2 can lose (R 3 Si)2 0 to give a metal oxo species. In each of these cases, siloxides deliver an

oxygen atom to a metal without a change in oxidation state. In fact, the use of hexamethyldisiloxane is a

30 31 well-established method for converting metal halides to metal oxo species with the loss of Me3SiCl.

Although this may be a potentially useful synthetic strategy, it is also a potential degradation pathway that

can yield insoluble metal oxide species. In some anionic cases, siloxides form particularly weak bonds to

metals, and ligand lability can provide a kinetically competent degradation pathway for reactive

species. The degradation of siloxide complexes is usually observed in oxophilic metal halide

species, where the formation a strong M=O bond makes breaking the Si-O bond favorable. The tBu 3 SiO

ligand is perhaps unique in its resistance to degradation because it provides steric protection to the silicon

as well as the metal center.

84 Scheme 3.1 Synthetic methods for metalation of siloxide ligands.

n Na(OSiR 3) + MX, i M(OSiR 3)n + n NaX Salt Metathesis X = F, CI, Br, I

HOSiR 3 + MXn - . M(OSiR 3)n + n HX Silanolysis

X = CR 3, Ar, NR2 , OR, 0

n AcOSiR 3 + M(OR')n - w M(OSiR 3)n + n ROAc Transesterification

Transition metal siloxide species are synthesized by two main procedures: salt metathesis of metal halides and silanolysis of a more basic ligand (oxide, alkoxide, amide, alkyl, or aryl). Another reaction that has been uniquely employed is the trans-esterification of a silyl acetate with a metal alkoxide. Because siloxides lack the ability to reduce transition metals, it is reasonable to expect that both salt metathesis and silanolysis methods should lead to the corresponding siloxide species in high yield. In fact, the degradation pathways outlined above make the synthesis quite specific to the choice of metal starting material and siloxide. In most cases, the presence of halides and coordinating solvents will either permit or hinder the isolation of a molecular species. As such, the following sections are divided by transition metal, and each section gives a brief overview of syntheses of related siloxide species in the literature.

85 3.2 Titanium/Vanadium Siloxides

3.2.1 Attempted Syntheses

Among first-row transition metals, siloxide ligands applied to titanium and vanadium have claimed much synthetic attention for their potential use in catalytic oxidation reactions and as materials precursors. 20,29,34-42 These siloxide species are most often synthesized by silanolysis, since the use of metal halides often yields oxo products. The exception to this is the silox ligand, where salt metathesis methods have been exploited successfully for most cases.

The synthesis of 3 and 4 coordinate titanium and vanadium DTBMS complexes was attempted by both procedures. Salt metathesis reactions included the addition of Na(DTBMS) to TiCl 3(THF) 3, TiCl4,

VCl 3 (THF)3 , VCl4 (THF) 2 , VOF 3, and VOCl 3. Silanolysis reaction were also attempted by treating

Ti(NMe2)4, Ti(OiPr)4, V{N(SiMe3)2}3, and OV(OiPr) 3 with various equivalents of DTBMS-H and at

different temperatures. In all cases no crystalline products were isolated. The silanolysis of Ti(OPr) 4

yields a colorless oil which was characterized as mixtures of (DTBMS)nTi(OiPr)4- species by 1H NMR.

Treatment of VCl3(THF)3 with 4 eq. of Na(DTBMS) results in a royal blue species that is soluble in THF,

but the addition of hydrocarbon solvents results in the formation of Na(DTBMS) and a blue solid,

possibly [V(DTBMS) 3]n. Treatment of the THF solution with AgOTf, results in the immediate color

change to pale green, but this putative V(DTBMS) 4 species decomposes within minutes. Since siloxide

complexes of Ti"', Tilv, Vm,1. VIV, and Vv(oxo) have been well studied, it was surprising that the DTBMS

ligand did not yield analogous results. Nevertheless, these metal complexes are not expected to be strong

oxidants with the potential for 0-0 bond formation reactions.

86 3.3 Chromium Siloxides

3.3.1 Synthesis of Chromium Siloxide Species

Chromium siloxide species have been explored for their general coordination chemistry,

magnetism, and as catalyst precursors for heterogeneous ethylene polymerization. 7 ,43- Species have

been found in oxidation states from II to VI, but curiously no homoleptic Cr(V) siloxide/oxo species have

been prepared. Cr(II) and Cr(III) siloxides are generally synthesized by salt metathesis, while Cr(IV) and

3 Cr(VI) are prepared by silanolysis. Given the stability of Cr(O)(ditox) 3, it was expected a similar species

might be isolated with DTBMS by isolation of a Cr(III) species and performing an oxygen atom transfer

reaction.

The addition of Na(DTBMS) to a solution of CrC12 in THF gave sky blue solutions whose

formation did not appear to depend on the number of equivalents of ligand used. Over the course of days,

a fine black powder precipitated from solution that proved insoluble in common organic solvents. Heating

the reactions accelerated the formation of this precipitate, as did removal of the THY under vacuum or

dilution with toluene. This species was initially presumed to be an aggregate of formula [(DTBMS) 2Cr]n

or possibly [(DTBMS)CrCl]n based on previously reported Cr(II) siloxide chemistry.43 48 The latter

complex was ruled out based on recovery of 1.8 eq. of NaCl when 2 or more eq. of Na(DTBMS) were

added. The former was ruled out by elemental analysis, which revealed only trace amounts of carbon and

hydrogen. Assignment of the organic products by 1H NMR proved impossible since the reaction could not

be driven to completion, and the multitude of peaks observed was broadened by the underlying

paramagnetism.

Since all of the presumed NaCl could be accounted for, the lack of carbon or hydrogen implies a

CrOx species. The oxide could be formed via CrCl(DTBMS)(THF) 2, which is analogous to a similar

48 turquoise species reported. A second equivalent of Na(DTBMS) might form (DTBMS) 2Cr(THF) 2 which

t t could eliminate Bu 2MeSiOSi Bu 2Me to give CrOx. Alternately, Na(DTBMS) might attack the

87 coordinated siloxide in an SN2 type reaction to give the same products. A catalytic mechanism of chloride attack to yield tBu2MeSiCl as an intermediate to forming the silyl ether is also feasible. Oxygen-atom transfer from DTBMS to Cr(II) suggests that the formation of a stable oxide species may drive the reaction thermodynamically, however this has not prevented isolation of other Cr(II) siloxide species. 19,43,48

The addition of 3 eq. of Na(DTBMS) to CrCl 3(THF) 3 yielded a blue solution in THF, and a blue solid (1) that was insoluble in hydrocarbon solvents. Just like the V(III) analogue, this royal blue solid is thought to be a polymeric species [(DTBMS)3Cr] which is made soluble by coordination of THF. The blue color is consistent with the pseudotetrahedral complexes Cr(OCHIBu 2)3(THF) and

37 Li(THF)Cr(OCHBu 2)4, but contrasts with trigonal species like Cr(silox) 3 (green),

2 Cr{(OSi(O'Bu) 3} 3(TlF)2 (green), Cr{(OSi(O'Bu) 3} 3(HNEt 2)2 (purple)," Cr(OSi{C(SiMe 3)}Me2)

3 4 8 (turquoise),1 and Cr(ditox) 3 (turquois). The formulation of a tris-DTBMS chromium species is also suggested by elemental analysis.

While structural characterization of 1 has remained elusive, addition of 10 eq of Na(DTBMS) to this blue solid in 10:1 pentane-THF mixtures produces blue crystals of (DTBMS) 4CrNa(THF) (2) in low yield. This Cr(III) species is particularly labile, since dissolution of this species in benzene yields 1 equivalent of THF, 1 eq. of Na(DTBMS), and 1/n eq. of 1. This observation suggests that the crystalline solid (2) is in equilibrium with 1 as shown in equation 3.1.

[Na(THF)][Cr(DTBMS)- - [Cr(DTBMS) 3]n + Na(DTBMS) 3.1 2 1

88 Figure 3.1 X-ray crystal structure of 2 with thermal ellipsoids shown at the 50% probability level.

Table 3.1 Selected bond lengths (A) and angles (0) of (DTBMS) 4CrNa(THF) (2).

Bond Lengths

Cr(1 )-O(1) 1.8805(12) Cr(1)-O(2) 1.8885(13) Cr(1)-O(3) 1.8691(12) Cr(1)-O(4) 1.8432(12) Si(1)-O(1) 1.6206(14) Si(2)-O(2) 1.6133(13) Si(3)-O(3) 1.6299(11) Si(4)-O(4) 1.6130(14) O(1)-Na(1) 2.4245(14) O(2)-Na(1) 2.3530(15) O(3)-Na(1) 2.3604(13) O(5)-Na(1) 2.2794(14)

Bond Angles

O(1)-Cr(1)-O(2) 92.85(6) O(1)-Cr(1)-O(3) 95.46(5) O(1)-Cr(1)-O(4) 129.31(6) O(2)-Cr(1)-O(3) 94.68(6) O(2)-Cr(1)-O(4) 125.27(6) O(3)-Cr(1)-O(4) 111.07(5)

89 Si(1)-O(1)-Cr(1) 157.89(9) Si(2)-O(2)-Cr(1) 147.68(9) Si(3)-O(3)-Cr(1) 137.68(7) Si(4)-O(4)-Cr(1) 152.80(9)

The structure of 2 shown in figure 3.1 is closely related to ('Bu 2HCO)4CrLi(THF) except in this case the sodium ion coordinates to three siloxides instead of the lithium to two alkoxides.54 The geometry of 2 is a distorted tetrahedron where the three siloxide ligands that coordinate to the sodium are nearly 900 as is found in octahedral geometry (table 3.1). Placing a z-axis along the Cr-04 vector, this distortion

2 2 makes dx, and dyz high energy unfilled orbitals and dX 2 dz , and dxy as the singly occupied d-orbitals.

Therefore this distortion makes the ligand field of 2 appear like an octahedral field, which is most often observed for Cr(III) species.

Oxidation of 1 in THF solution with 0-atom transfer reagents such as Me3NO and PhIO yields a red-brown solution, which gives a solid that is soluble in pentane and other weakly polar solvents.

t Filtration and crystallization from pentane at -40*C gives Cr(OSi Bu 2Me) 4 (3) in 40% yield. The formation of 3 from O-atom transfer reagents suggests that the expected product Cr(O)(DTBMS) 3 is not formed or that it disproportionates to 3 and Cr(O)2(DTBMS)2 . Rational synthesis of 3 using 4 eq of

Na(DTBMS), 1 eq of CrCl3(THF) 3 and 1 eq of AgOTf yields 3 in 86% isolated yield as a brown crystalline solid.

The X-ray crystal structure of 3 shows a pseudo-tetrahedral Cr (IV) tetrasiloxide complex (figure

3.2). As is seen for other Cr(IV) complexes, there is a slight Jahn-Teller distortion which compresses two

O-Cr-O bond angles to 1080 and expands two to 1120 (table 3.2). The Cr-O bond lengths and angles are comparable to other Cr(IV) alkoxide and siloxide complexes, 475 4 as well as the discrete tetrahedra of

4 CrO4 - observed in the structure of Ba3 CrO5 (avg. Cr-O = 1.769(3) A)." The Cr-O bond lengths of 3

(avg. Cr-O = 1.764(1) A) as well as the Si-O-Cr angles (avg. LSi-O-Cr = 148.11(6)0) indicates significant 7c bonding in this complex and a more covalent Cr-OSi bond than found in most other metal-

90 siloxide complexes. The average Si-O bond length of 1.666 A is the longest observed for the DTBMS ligand.

Figure 3.2 X-ray crystal structure of 3 with thermal ellipsoids shown at the 50% probability level.

Table 3.2 Selected bond lengths (A) and angles (0) of Cr(DTBMS)4 (3).

Bond Lengths

Cr(1)-O(1) 1.7640(8) Cr(1)-0(2) 1.7647(8) Si(1)-O(1) 1.6659(8) Si(2)-0(2) 1.6658(8)

Bond Angles 0(1)-Cr(1)-O(1A) 112.61(5) 0(1)-Cr(1)-0(2) 108.02(4) 0(1)-Cr(1)-0(2A) 108.20(4) 0(2)-Cr(1)-0(2A) 111.83(6)

Si(1)-O(1)-Cr( 1) 148.35(5) Si(2)-0(2)-Cr(1) 147.88(6)

91 In contrast to 2, 3 has proven to be quite stable, even to water and air, and it can be conveniently recrystallized from hexanes on a bench top. In addition, it can be distilled at 80 *C under vacuum (100 mTorr). While Cr(IV) alkoxides are known to be stable to air and undergo slow hydrolysis, this surprising inertness is likely due to the crowded steric environment around the pseudotetrahedral Cr(IV) center.

3.3.2 Redox Chemistry

The reduction of 3 is irreversible as is the oxidation of THF solutions of 2. This suggests that the coordination of Na* or another alkali metal ion is essential for stabilization of the Cr(III) species. The

Cr(IV) species could not be oxidized within the solvent window, and 3 proved unreactive to chemical oxidation with Cp2FePF6 and (p-Br-C6H4)3N-SbCl6 . Reaction of 3 with XeF 2 gave a dark red solution, possibly (DTBMS) 2CrO2, but the product was not clean by 'H NMR and decomposed within minutes to give insoluble precipitates.

3.3.3 Spectroscopy

Qualitatively, the brown color of 3 is unlike other molecular Cr(IV) complexes, which can range

56 from royal blue in Cr(OtBu) 4, to green in Cr(NEt 2)4, or maroon in Cr(CH 2'Bu) 4 . -58 The only other report of a Cr(IV) siloxide, Cr(OSiEt3)4 was described as a viscous blue oil." Since the structure of 3 establishes a nearly tetrahedral geometry about chromium, Tanabe-Sugano analysis of the electronic absorption spectrum allows quantification of the ligand field parameters. The absorption spectrum of 3 is given in figure 3.3, and shows four relatively weak d-d bands that were assigned using the Tanabe-Sugano diagram for a d' Oh molecule (table 3.3). This diagram correlates to a d2 Td species by the hole formalism.5 9 The highest energy band is assigned as the 3T1(P) and is relatively weak because it

2 corresponds to a strong field state where both electrons in a (t 2 ) configuration. As is seen for other Cr(IV) species, 3 T1(F) is slightly split as a result of the Jahn-Teller distortion." The spin forbidden 1E transition

92 3 is weak but sharp, while the T2 (F) is broad and weak since it is forbidden by the electric dipole selection rule. Analysis of the transitions gives AT = 7940 cm 1 and B = 530 cm- 1.

1000

900

800

700

600 E 0O 500

400

300

200

100

0 0 200 400 600 800 1000 1200 1400 1600 1800 Wavelength / nm

Figure 3.3 UV-Vis-NIR Absorption spectrum of 3.

Table 3.3 Electronic absorption data of Cr(DTBMS) 4 (3).

1 in S/ n Energy / cm- s / Mcm i Assignment (approx.) Td symmetry

278 36000 3731 Charge transfer

540 18520 140 3T1(P)+_ 3A2(F)

800 12500 900 3T1(F)<3A2F)

3 1042 9600 (sh) 'E(D)+_ A2(F)

3 3 1240 8070 60 T2(F) - A2(F)

93 The ligand field parameters of 3 can be compared with other tetrahedral CrIv species such as

t Cr(O'Bu) 4, Cr(NEt2)4, and Cr(CH 2 Bu) 4. Bradley reports AT = 9430 cm' and B = 795 cm-1 for

5 1 Cr(OtBu) 4, while the presence of only one observable transition in the spectrum of Cr(NEt 2)4 could not be assigned by spectroscopy." Analysis of Cr(CH 2'Bu) 4 gives a stronger field of AT = 11500 cm-1 and B =

58 4 550 cm-1. These parameters place Cr(DTBMS) 4 as having the weakest known ligand field for a Cr * species, which of course is expected due to the weak donor ability of siloxide ligands.

The infrared spectrum of 3 (figure 3.4) shows no evidence of decomposition even after recrystallization from hexane/acetone solution, as it lacks a characteristic SiO-H absorption beyond 3000

1 cm- . Tetrahedral molecules are expected to exhibit strong symmetric and asymmetric bond stretching absorptions of T 2 symmetry." Frequency analysis of corresponding DFT calculations (section 3.3.4) suggest that the asymmetric Cr-O stretch should be observed near 890 cm-1, whereas typical Cr-O stretches for the related alkoxide complexes occur between 600 and 720 cm-1. Si-O bond stretches also show up within this region. If the four intense absorptions at 714 cm- 1, 764 cm-1, 817 cm-1, and 849 cm-1 are assigned as these stretching modes, this would suggest particularly strong Cr-O and Si-O bonds, consistent with a strong covalent interaction.

94 Wavenumber (cm') 4000 3500 3000 2500 2000 1500 1000 500 110 105 100 95 90 85 80 75 70 65 o- 60 55 50 45 40 35 30 25 20 15--

Figure 3.4 Infrared Spectrum of Cr(DTBMS) 4 (3).

3.3.4 DFT Calculations of 3

To better understand the electronic structure of 3, DFT calculations were performed on a truncated model where the tBu groups were converted to methyl. They reveal a standard Td ligand field

(figure 3.5) with a slight splitting of the half-filled SOMO of e symmetry by a Jahn-Teller distortion. The

T 2 set is also split to give e and b2 type orbitals. The average Co-O and Si-O bonds were 1.68 A and 1.78

A respectively, in good agreement with the crystal structure.

95 t -b2

-----. a e +bi

Figure 3.5 Qualitative ligand field diagram for 3.

3.3.5 DFT calculations of hypothetical (Me3SiO)3CrO

Since (DTBMS) 3CrO could not be isolated, calculations were performed on the truncated complex, (Me3SiO) 3CrO, to determine the spin density on the oxo relative to the (ditox) 3CrO species. The

calculated (Me 3SiO)3Cr-O bond distance is 1.58 A, which is shorter than that found in the structure of

(ditox)3CrO (1.65 A). Figure 3.5 shows spin density plot of (Me3 SiO) 3CrO the unpaired electron is

located 95% on the Cr atom with minimal contribution from the siloxide ligands or the oxo. These

calculations suggest that siloxides do, in fact, deliver more ionic character to the metal due to the

extremely weak ligand field bestowed upon the metal center. The long Cr-O bond length and calculated

radical character on (ditox) 3CrO suggests that it may be described as an antiferromagnetically coupled

Cr(IV) oxyl radical, as the one-electron oxidation reactivity of this complex indicates. This

characterization contrasts with other pseudotetrahedral Cr(V) complexes of the type (ArRN) 2(X)CrO (Ar

t = 2,5-C 6H3 FMe, R = Bu, X = 0, OC(O)Ph, OMe, and OSiPh3), which have been shown to be strongly

96 reducing, with short Cr-O bond distances suggesting formal triple bond character.60 In this system, covalent M-N bonding may transfer spin density to the ligands rather than the oxo. Since DFT calculations support that siloxide ligands do not elicit this radical behavior, the failure to isolate

(DTBMS) 3CrO may be ascribed the instability of tetrahedral Cr(V) species in ionic ligand environments.61

Figure 3.6 Calculated spin density plot of (Me3 SiO) 3 CrO.

3.3.6 Comparison to Solid State Cr**

Solid state materials doped with Cr** have received much attention as near infrared laser diode

62 63 4 3 materials. The lasing action of tetrahedral Cr * is comparable to that of octahedral Cr + because each

2 3 3 has a half-filled degenerate ground state, (e) and (t2g) respectively. The ruby laser, Cr +:Al2O 3, is a 3-

4 4 4 state laser system in which the octahedral chromium A 2 ground state is excited to T1 and T 2 states

2 which undergo spin crossover to a 2E state. Build up of this E state creates a population inversion, which

will undergo stimulated emission of coherent monochromatic light at 694 un. A Cr** ion in a tetrahedral

3 lattice has a similar excited state topology except that in this case it has a ground state of A2. Pumping of

3 3 1 the T1 and T2 bands can lead to spin crossover to a E state, which undergoes stimulated emission in a

97 similar fashion. While the ruby laser emits visible (red) light, in the case of Cr4 the weaker tetrahedral

ligand field makes this an interesting near infrared laser gain medium.

One of the most difficult challenges in the synthesis of a Cr** based laser diode is the doping of

3 materials with Cr species that resist reduction to Cr + upon introduction to solid materials. Nevertheless,

several different Cr4 doped materials have been prepared and isolated. Their ligand field parameters have

1 4 64 1 been characterized as AT = 9620 cm- and B = 480 cm 1 in Cr *:Ca 2GeO 4, AT= 9150 cm- and B = 515

4 65 1 4 66 cm-1 in Cr *:Y3Al5O1 2 (YAG), and AT = 10100 cm-1 and B = 860 cm for Cr *:Mg 2 SiO 4 (forsterite).

Spectroscopic analysis of these species is complicated by the fact that there are often significant

impurities of Cr3+ within lattice sites, and that the Cr4 sites are often severely distorted to lower

symmetries (D2d or C 2v). The fluorescence spectrum of these materials shows an infrared emission band

between 1000 and 1400 nm, which is characteristic of the Cr4* emission from the 2 E state.

With these parameters in mind, the solution fluorescence spectrum (pentane) of 3 was monitored

upon excitation at 400 nm and 800 nm, though no emission was observed at temperatures of 20*C,

-78*C, and -198 0 C (frozen toluene glass). This could be due to the ability of 3 to decay non-radiatively

through vibrational relaxation. In addition, since the 1E state is not the lowest excited state, higher energy

3 excited states can decay to the low energy T 1(F) state without spin crossover. Although 3 does not

fluoresce under these conditions, it is the first structural and spectroscopic characterization of a Cr4 *

siloxide species, and should aid in analyzing the spectra of the solid state species, perhaps becoming a

new source of Cr** for doping these materials.

98 3.4 Manganese Siloxides

3.4.1 Synthesis

Manganese siloxide complexes are rare, and only two homoleptic species have been reported in

t 54 5 3 the literature: [{( BuO) 3 SiO}2 Mn]2 and (trisilox)2Mn-LiCl (trisilox = Me 2 {(Me3 Si) 3 C} SiO). Addition of 2 eq. of Na(DTBMS) to MnCl 2(THF) 2 yields 2 eq. of NaCl and a colorless oil analyzing for

(DTBMS) 2Mn(THF) 2 (4). Treatment of MnCl2(THF) 2 with 4 eq. Na(DTBMS) in THF gives a colorless solution with a fine white precipitate (NaCl). In contrast to the case for chromium, no 4-coordinate manganese complex could be isolated. Instead, filtration and addition of 15-crown-5 gives colorless crystals of Na(15-crown-5)Mn(DTBMS) 3 (5) in 95% yield. Addition of other crown ethers such as 12-

crown-4, 18-crown-6, or dibenzo-18-crown-6 as well as salt metathesis with PPNC1, Et4NCl, or Ph4 PCl did not give similarly isolable species. Complexes 4 and 5 were subjected to oxidation with the following reagents/solvents/results: 02, THF, no reaction; PhIO, THF, no reaction; Me3NO, benzene, no reaction;

Cp 2FePF6, THF, no reaction; AgOTf, THF, no reaction; (4-bromophenyl) 3N-SbCl6, CH 2Cl 2 , insoluble precipitate (MnO2); and XeF 2, CH3CN, insoluble precipitate (MnO 2). These results are consistent with the observation that known manganese siloxide complexes are all found in the +2 oxidation state.

The X-ray crystal structure of 5 (figure 3.7) shows manganese with trigonal monopyramidal

geometry with three siloxide ligands in a nearly planar arrangement (Lo-MN--o = 357.10). Table 3.4 lists the relevant bond lengths and angles. Of note is the coordination of the sodium atom that elongates both the Mn-O(2) and Si(2)-O(2) bond relative to the respective lengths for O(1) and 0(3). 5 is the first

known tris-siloxide complex of manganese.

99 Figure 3.7 X-ray crystal structure of 5 with thermal ellipsoids shown at the 50% probability level.

Table 3.4 Selected bond lengths (A) and angles (*) of Na(15-crown-5)Mn(DTBMS) 3 (5).

Bond Lengths Mn(1)-O(1) 1.969(4) Mn(1)-0(2) 2.0435(12) Mn(1)-0(3) 1.960(3) Mn(1)-0(3 1) 2.3699(11) Si(l)-O(l) 1.606(5) Si(2)-0(2) 1.6170(14) Si(3)-0(3) 1.595(3) Na(2)-0(2) 2.2472(13) Bond Angles 0(1)-Mn(1)-0(2) 117.64(9) 0(1)-Mn(1)-0(3) 119.92(10) 0(1)-Mn(1)-O(31) 98.54(12) 0(2)-Mn(1)-0(3) 119.62(8) 0(2)-Mn(l)-0(3 1) 85.24(4) 0(3)-Mn(1)-0(3 1) 102.65(10) Si(1)-O(1)-Mn(1) 158.0(2) Si(2)-0(2)-Mn(1) 121.46(7) Si(3)-0(3)-Mn(1) 157.7(2) Mn(1)-O(2)-Na(2) 107.68(8)

100 3.5 Iron Siloxides

3.5.1 Synthesis and Characterization of Fe(II) Species

Iron siloxides have garnered significant attention as soluble analogues to heterogeneous silica supported iron catalysts. In this regard, iron silsesquioxanes have proven important for determining the coordination chemistry and reactivity of such species. 67- 9 Iron siloxide species are most commonly found

in the Fe(III) oxidation state, but some Fe(II) siloxides are also known. Syntheses of Fe(II) siloxides usually involve silanolysis of iron amide or mesityl starting materials to form trigonal species. Examples

3 2 of Fe(II) siloxides include [Fe{OSi(OBu) 3} 2]2 70 and Na(dme)Fe{OSi(SiMe 3)3} 3. On the other hand,

Fe(III) siloxides are typically synthesized by salt metathesis and typically take on a pseudotetrahedral

t geometry, as is the case of {(BuO)3SiO}3FeL (L= THF, Et2O) and the p-oxo complex

7 [Et 4N] 2[{('BuO)3SiO} 3Fe]O. 4'' Reaction of FeCl3 or [Et 4N][FeCl 4] with 3 eq. of Na(OSiMe 3) yields a

tris-siloxide dimer [Fe(OSiMe 3)3] 2 , while addition of 4 eq. of Na(OSiMe3) yields the tetrasiloxide

68 72 complexes NaFe(OSiMe 3)4 and [Et 4N][Fe(OSiMe 3)4].

Treatment of [Fe(mesityl) 2] 2 with 4 eq. of DTBMS-H gives [Fe(DTBMS) 2] 2, which can be

crystallized in the presence of trace amounts of THF to give [Fe(DTBMS) 2] 2-(THF) (6). Alternately, 6

can be prepared from FeCl2 and 2 eq. Na(DTBMS) in THF. The structure of 6 is shown in Figure 3.8 and

reveals one trigonal planar Fe(II) center and one trigonal monopyramidal Fe(II) center due to coordination

of THF, perhaps as a result of crystallization. As a result, the Fe-O bond lengths are longer for pyramidal

Fe(l) than for trigonal Fe(2) (table 3.5).

101 Figure 3.8 X-ray crystal structure of 6 with thermal ellipsoids shown at the 50% probability level.

Table 3.5 Selected bond lengths (A) and angles (*) of [Fe(DTBMS) 2]2-(THF) (6).

Bond Lengths

Fe(l)-O(1) 1.8353(7) Fe(1)-O(2) 2.0927(7) Fe(1)-O(3) 2.0292(7) Fe(1)-O(5) 2.1411(7) Fe(2)-O(2) 1.9293(6) Fe(2)-O(3) 1.9577(7) Fe(2)-O(4) 1.8116(7) Si(1)-O(1) 1.6166(7) Si(2)-O(2) 1.6495(7) Si(3)-O(3) 1.6538(7) Si(4)-O(4) 1.6223(8)

Bond Angles

O(1)-Fe(1)-O(2) 130.53(3) O(1)-Fe(1)-O(3) 138.89(3) O(2)-Fe(1)-O(3) 82.41 O(1)-Fe(1)-O(5) 105.56(3) O(2)-Fe(1)-O(5) 88.63(3) O(3)-Fe(1)-O(5) 97.92(3) O(2)-Fe(2)-O(3) 88.64(3) O(2)-Fe(2)-O(4) 135.51(3) O(3)-Fe(2)-O(4) 134.25(3)) Si(1)-O(1)-Fe(1) 162.93(5) Si(4)-O(4)-Fe(2) 158.89(5)

102 Addition of 3 eq. of Na(DTBMS) to FeCl 2 in THF gives a pale green solution of NaFe(DTBMS) 3.

This species can be crystallized upon addition of 15-crown-5 to give 7 in 89% yield as a pale green crystalline solid. The crystal structure (figure 3.9, table 3.6) shows that 7 has a similar structure to 5, but

in this case the 15-crown-5 does not coordinate to the Fe(II) center. This may be due to the fact that iron

is more electron-rich than manganese, and that one a interaction from each of the siloxide ligands would yield an 18-electron complex without the need for additional coordination of the crown ether.

Figure 3.9 X-ray crystal structure of 7 with thermal ellipsoids shown at the 50% probability level. Hydrogen atoms and 1.5 molecules of toluene were omitted for clarity.

Table 3.6 Selected bond lengths (A) and angles (*) of Na(15-crown-5)Fe(DTBMS) 3 (7).

Bond Lengths

Fe(1)-O(1) 1.8610(14) Fe(1)-O(2) 1.9154(12) Fe(1)-O(3) 1.8553(13) Na(1)-O(2) 2.3044(13)

103 Si(1)-O(1) 1.6003(14) Si(2)-O(2) 1.6202(12) Si(3)-O(3) 1.6082(13)

Bond Angles O(1)-Fe(1)-O(2) 114.30(6) O(1)-Fe(1)-O(3) 125.77(6) O(2)-Fe(1)-O(3) 119.93(6) Fe(1)-O(2)-Na(1) 111.58(6) Si(1)-O(1)-Fe(1) 163.93(10) Si(2)-O(2)-Fe(1) 135.63(7) Si(3)-O(3)-Fe(1) 163.71(10)

3.5.2 Synthesis and Characterization of Fe(III) Species

Synthesis of Fe(III) species was carried out by treatment of FeC13, FeBr3, or (Et 4N)(FeCl4) with

Na(DTBMS) in THF, ether, and dichloromethane. Regardless of the starting material or the number of equivalents of siloxide added, the only isolable complex from this reaction is Na2[(DTBMS) 3FeOH] 2 (8).

The yield of this species is maximized when 5 eq. of Na(DTBMS) per Fe were used, which might imply that the source of the oxygen atom is from the siloxide. On the other hand, the source of the hydrogen is more of a mystery. Synthesis of 8 in the presence of THF-d8 did not lead to a shift in the O-H stretching frequency, implying that the source of H is either from water or DTBMS-H.

The crystal structure of 8 (figure 3.10) reveals a dimeric species in which the hydroxide and two siloxide ligands are coordinated to the sodium ions to give a partial dicubane structure. The hydroxide displays the longest Fe-O bond length (table 3.7) because it is prevented from 7E bonding to the electropositive Fe(III) center by interaction with both sodium ions. Similarly, the two siloxides bound to sodium ions have correspondingly longer bond lengths than the one bound only to the Fe.

104 Figure 3.10 X-ray crystal structure of 8 with thermal ellipsoids shown at the 50% probability level. The tert-butyl methyl groups and hydrogen atoms (except for O-H) were omitted for clarity.

Table 3.7 Selected bond lengths (A) and angles (*) of [Na(DTBMS)3FeOH]2 (8).

Bond Lengths

Fe(1)-O( 1) 1.8225(8) Fe(1)-0(2) 1.8688(9) Fe(1)-O(3) 1.8648(9) Fe(1)-0(4) 1.9139(9) Na(1)-0(2) 2.4236(10) Na(1)-0(3A) 2.4882(10) Na(1)-0(4) 2.2831(11) Na(1)-0(4A) 2.3264(11) Si(1)-O(1) 1.6339(9) Si(2)-0(2) 1.6354(9) Si(3)-0(3) 1.6337(9)

Bond Angles

0(1)-Fe(1)-0(2) 117.85(4) 0(1)-Fe(1)-0(3) 116.73(4)

105 O(1)-Fe(1)-O(4) 112.35(4) O(2)-Fe(1)-O(3) 111.61(4) O(2)-Fe(1)-O(4) 96.21(4) O(3)-Fe(1)-O(4) 98.31(4) O(4)-Na(1)-O(4A) 88.44(4) Na(1)-O(4)-Na(1A) 91.56(4) O(3)-Fe(2)-O(4) 134.25(3)) Si(1)-O(1)-Fe(1) 162.93(5) Si(4)-O(4)-Fe(2) 158.89(5)

106 3.6 Cobalt Siloxides

3.6.1 Background

Siloxides of cobalt have been isolated in the +1, +2, and +3 oxidation states, however homoleptic siloxide complexes are only observed for Co(II). Species such as [Co(OSiPh3)2(THF)] 2 and

33 73 74 [Co{OSi(SiMe 3)3}2]2 are synthesized by silanolysis of Co{N(SiMe 3)2}2 in THF. , The Co(II) phosphine species, PhBP 3Co(OSiPh 3) and the cationic Co(III) species have been structurally characterized. When compared with the structurally similar Co(I) species, (Ph 3P)3 Co(OSiMe3 ),7 one may observe interesting bonding trends among the series. While the Co-O-Si bond angle is nearly 1800 in each case, the Co-O bond decreases from 1.86 A to 1.80 A and 1.77 A in the Co(I), Co(II), and Co(III) species respectively. Meanwhile the Si-O bond increases from 1.60 to 1.61 and 1.65 A. These changes seem to suggest that as the metal center becomes capable of adopting M-O multiple bond character, the

Si-O bond length increases rather dramatically. The diamagnetic Co(III) complex has been characterized as having two strong Co- 0t bonds, despite not having a formal a bond.

3.6.2 Synthesis

The addition of 2 eq. of Na(DTBMS) to CoCl2 in THF leads to the formation of salt precipitate and a blue solution, which upon filtration and removal of solvent yields a royal blue oil that analyzes for

Co(DTBMS) 2(THF) 2 (9). Addition of three or more equivalents of Na(DTBMS) yields a light blue solution, to which 15-crown-5 can be added to precipitate Na(15-crown-5)Co(DTBMS) 3 (10) as pale blue crystalline solid. Figure 3.11 displays the crystal structure of 10, which is found to be is isostructural to the manganese derivative 5, but with Co-O bond lengths that are shorter by an average of 0.07 A. The bonding in 10 is best described as strongly ionic, which is perhaps why the crown ether coordinates in 5 and 10 but not in 7. The relevant bond lengths and angles of 10 are given in table 3.8. 10 did not show

107 any electrochemical processes within the solvent window of TIHF or acetonitrile, and proved inert to chemical oxidants such as PhIO, Cp2FePF6, and (p-Br-C6H4) 3N.SbCl6.

Figure 3.11 X-ray crystal structure of 10 with thermal ellipsoids shown at the 50% probability level.

Table 3.8 Selected bond lengths (A) and angles (*) of Na(15-crown-5)Co(DTBMS) 3 (10).

Bond Lengths Co(1)-O(1) 1.8951(12) Co(1)-0(2) 1.9572(12) Co(1)-O(3) 1.8896(12) Co(1)-0(40) 2.2905(12) Si(1)-O(1) 1.6003(13) Si(2)-0(2) 1.6202(12) Si(3)-0(3) 1.6013(13) Na(1)-0(2) 2.2554(14) Bond Angles 0(1)-Co(1)-0(2) 118.42(5) O(1)-Co(l)-0(3) 117.11(6) 0(1)-Co(1)-0(40) 100.73(5) 0(2)-Co(1)-0(3) 121.22(5) O(2)-Co(1)-0(40) 85.95(5) 0(3)-Co(1)-0(40) 101.44(5)

Si(1)-0(1)-Co( 1) 159.90(9) Si(2)-0(2)-Co(1) 123.32(7) Si(3)-0(3)-Co(1) 156.25(9) Co(1)-0(2)-Na(1) 108.00(5)

108 3.7 Concluding Remarks

The DTBMS ligand has permitted the synthesis and characterization of new siloxide coordination complexes in the first row transition metal series. A ratio of two DTBMS ligands per metal yields oils with the formula (DTBMS) 2M(THF) 2 (M = Mn (5), Co (10)) or a dimeric solid in 6. Coordination of three

DTBMS ligands to M(III) yields a polymeric solid, 1 (M = Cr), or a hydroxide dimer, 8 (M = Fe). The coordinating properties of 15-crown-5 permitted the isolation of monomeric species of the type Na(15- crown-5)M(DTBMS) 3 (M = Mn, Fe, Co) in 5, 7, and 10. In the case of Cr, four DTBMS ligands were capable of coordination to yield anionic Cr(III) (2) or neutral Cr(IV) (3) complexes.

With the exceptions of 6 and 7, all of the metal complexes in this section are 4-coordinate. This suggests that the DTBMS ligand does not have the steric bulk to yield monomeric 3-coordinate complexes, which are observed with the analogous ditox or silox ligands.3, 4 8 Using the terminology introduced in section 3.1.2, the coordination properties of DTBMS can be described by a maximum steric coordination number of 4, since homoleptic complexes with fewer than 4 DTBMS ligands per metal form dimeric or polymeric species. When the maximum steric coordination number is present, an exceptionally stable Cr(IV) complex (3) can be isolated. The lack of redox processes in most of the DTBMS complexes is disappointing but not too surprising, since siloxide ligands do not typically stabilize oxidizing metal species.14 Although robust towards oxidation, DFT calculations support the notion that siloxide ligands ascribe considerable ionic character to metal complexes, which hinders their ability to support high valent metal oxo complexes.

109 3.8 Experimental Section

3.8.1 General Synthetic Considerations

All manipulations were carried out using modified Schlenk techniques under an atmosphere of N2 or in a MBraun glove box. Solvents for synthesis were of reagent grade or better and were dried according to standard methods or using a solvent purification system and stored over 4 A molecular sieves. Elemental analyses were performed by Midwest Microlabs LLC. Chemicals were purchased from

6 Aldrich, Strem, or Alfa Aesar, and were used as received unless otherwise noted. tBu 2MeSiONa, [Fe(1-

7 6 mesityl) 2] 2 were prepared by literature methods.

3.8.2 Physical Methods

NMR were recorded on solutions at 25 *C within the magnetic fields of Varian Mercury 300 and

Inova 501 spectrometers, which were located in the Department of Chemistry Instrumentation Facility

(DCIF) at MIT. Chemical shifts are reported using the standard 8 notation in ppm, and spectra were referenced to residual solvent peak. UV-Vis-NIR absorption spectra were collected on a Cary 5000 by

Varian Inc. IR spectra of powdered samples were collected on a PerkinElmer Spectrum 400 FT-IR/FT-

NIR Spectrometer outfitted with a Pike Technologies GladiATR attenuated total reflectance accessory with a monolithic diamond crystal stage and a pressure clamp. Cyclic voltammetry experiments were performed in a nitrogen filled glovebox using dry solvents containing 0.1 M Bu4NPF6 as a supporting electrolyte. A three compartment cell was employed possessing a glassy carbon electrode as the working electrode, Pt wire as the auxiliary electrode, and Ag/AgCl as a reference electrode. CVs were monitored with scan rates of 10-100 mV/s employing iR compensation, and referenced to Cp 2Fe/Cp2Fe*.

110 3.8.3 Crystallographic Procedures

X-ray diffraction experiments were performed on single crystals grown from pentane/THF 10:1

(2), pentane (3), THF (5), pentane/THF 10:1 (6), toluene (7 and 8), and THF (10). Crystals were removed from the supernatant liquid and transferred onto a microsocope slide covered in Paratone N oil. Selected

crystals were affixed to a MiTeGen MicroMounts polyimide tip and mounted on a Bruker three circle

goniometer platform equipped with an APEX detector. A graphite monochromator was employed for

wavelength selection of the Mo Ka radiation (k = 0.71073 A). The data were processed and refined by

using the program SAINT supplied by Siemens industrial Automation, Inc. The structures were solved by

direct methods (SHELXTL v6.10, Sheldrick, G. M., and Siemens Industrial Automation, Inc., 2000) in

conjunction with standard difference Fourier techniques. All non-hydrogen atoms were refined

anisotropically. Hydrogen atoms were placed in calculated positions unless otherwise noted. 2 was

modeled for disorder in the Si(1) fragment (27%), rotational disorder of the tBu groups of the Si(2)

fragment (37%), the Si(4) fragment (36%), and the THF (50%). 3 was modeled for a disordered Si(2)

fragment (35%). 5 was modeled for disordered Si(1) and Si(3) fragments (12%) and for Si(2) 15%. 7 had

two molecules of disordered toluene which were 48% disordered and constrained with the flat command.

8 was modeled for a 6% disorder of the Si(2) fragment. The hydroxide hydrogen (H4) was located during

the refinement and refined independently.

3.8.4 Computational Methods

Density functional theory (DFT) calculations were performed with the hybrid functional Becke-3

parameter exchange functional77 ~79 and the Lee-Yang-Parr nonlocal correlation functional80 (B3LYP) as

implemented in the Gaussian 03, Revision B.05 software package.81 The triple-( basis sets with one set of

polarization functions82 (TZVP) were used for chromium, silicon, and oxygen atoms, and the double-(

basis sets with one set of polarization functions83 (SVP) were used on the carbon and hydrogen atoms.

The calculations were performed on truncated models of 3 and (Me3SiO) 3CrO, in which the tert-butyl

111 groups of the DTBMS ligand were replaced with methyl groups. All geometries were confirmed as local minima structures by calculating the Hessians and checking that no negative eigenvalues were present.

TD-DFT calculations were performed to calculate the first 50 excited states of each model. Cartesian coordinates of the optimized geometries and absolute energies of 3 and (Me 3SiO) 3CrO are provided in section 3.8.16.

3.8.5 Preparation of [Cr(OSitBu2Me) 3]n (1)

A solution of Na(DTBMS) (200 mg, 1.02 mmol) in 2 mL THF was added to a purple suspension of CrCl 3(THF) 3 (125 mg, 0.333 mmol) in 2 mL THF to give a bright blue solution with a fine white precipitate (NaCl). The solution was stirred for 4 hours, filtered through celite, and reduced in volume under vacuum. Hexane was added to precipitate a blue solid which was washed with 3 x 5 mL hexane to give 180 mg of 1 (94%). Anal. Calcd. For C27H63CrSi 3O3: C, 56.69, H, 11.01. Found: C, 56.17; H, 10.78.

t 3.8.6 Preparation of Na(THF)Cr(OSi Bu 2Me) 4 (2)

A solution of Na(DTBMS) (100 mg, 0.51 mmol) in 2 mL hexane was added to 1 (30 mg, 0.052 mmol). 0.3 mL of THF was added to give a pale blue solution, which deposited blue crystals of 2 (9 mg,

20%) mixed with 1 over the course of 2 weeks. Anal. Caled. for C4oH 92 CrNaO5 Si 4 : C, 57.16; H, 11.03.

Found: C, 55.37; H, 10.54.

3.8.7 Preparation of Cr(OSitBu 2Me) 4 (3)

A THF solution of Na(DTBMS) (200 mg, 1.02 mmol) was added to a suspension of CrCl 3(THF)3

(95 mg, 0.253 mmol) in THF and stirred for 4 hours. The resulting blue solution was then treated with a

THF solution of AgOTf (67 mg, 0.026 mmol) to yield a dark brown solution. After 1 hour, the solvent was removed by vacuum, and the solid extracted with pentane (3 x 2 mL) and filtered through Celite. The pentane solution was condensed to 2 mL and cooled to -40 'C to deposit large brown crystals of 3 (160

112 1 mg, 86%). H NMR (300 MHz, CD 6) 6 = 2.8 (br. s. 18H, 'Bu), 5.0 (br. s. 3H, CH 3). pff = 2.85(5) BM

(Evans method, C6 D6 , 21 *C). Anal Caled. for C36H840 4 Si 4Cr: C, 58.01; H, 11.36. Found: C, 58.28; H

11.06.

3.8.8 Preparation of Mn(OSitBu 2Me) 2(THF) 2 (4)

A THF solution of Na(DTBMS) (100 mg, 0.51 mmol) was added to a suspension of

MnCl 2(THF) 2 (68 mg, 0.25 mmol) in THF and stirred for 12 hours. The resulting colorless solution was filtered, and the solvent removed under vacuum to yield a colorless oil (121 mg, 88%). Anal Calcd. For

C26H58MnO 2 Si 2 : C, 57.21; H, 10.71. Found: C, 57.53; H, 11.02.

t 3.8.9 Preparation of [Na(15-crown-5)][Mn(OSi Bu2Me)3 ] (5)

A THF solution of Na(DTBMS) (300 mg, 1.53 mmol) was added to a suspension of

MnCl 2(THF) 2 (130 mg, 0.48 mmol) in THF and stirred for 12 hours. The solvent was removed under vacuum, and the colorless solid was extracted with hexanes (3 x 2 mL) and filtered. 15-crown-5 (100 gL,

0.50 mmol) was added dropwise to precipitate a colorless microcrystalline solid. The crystals were

washed with pentane to yield 372 mg 5 (95%). Anal Calcd. for C3 7Hs3MnNaO8 Si 3 : C, 54.31; H, 10.22.

Found: C, 53.14; H, 9.88.

3.8.10 Preparation of [(tBu 2MeSiO) 2Fe] 2(THF) (6)

Na(DTBMS) (100 mg, 0.51 mmol) was added to a solution of FeCl2 (31 mg, 0.24 mmol) in THF and stirred for 4 hours. The solvent was removed under vacuum to give a pale yellow oil and colorless solid (NaCl), which was extracted with hexane and filtered. Evaporation of the hexane gave crude 7 which could be recrystallized at -40 'C from a mixture of pentane and THF (10:1) to give 170 mg 7

(79%). Anal Calcd. for C4 0H92 Fe 2O5Si4 : C, 54.77; H, 10.57. Found: C, 54.35; H, 10.25.

t 3.8.11 Preparation of [Na(15-crown-5)][Fe(OSi Bu2Me)31 (7)

A THF solution of Na(DTBMS) (675 mg, 3.44 mmol) was added to a solution of FeC12 (140 mg,

1.10 mmol) in THF and stirred for 12 hours. The solvent was removed under vacuum, and the resulting

113 pale blue solid was extracted with hexane (3 x 2 mL) and filtered. 15-crown-5 was added dropwise to separate a pale green oil. The hexane was decanted, washed with 2 x 1 mL hexane, and cooled to -40 'C to give 798 mg 8 a pale green solid (89%). The solid was tried under vacuum, and can be recrystallized

from toluene at -40 *C. Anal Calcd. for C3 7H83 FeNaOsSi 3 : C, 54.25; H, 10.21. Found: C, 52.33; H, 9.60.

3.8.12 Preparation of [Na] 2[(tBu2MeSiO) 3FeOH] 2 (8)

A THF solution of Na(DTBMS) (200 mg, 1.02 mmol) was added to a solution of FeBr 3 (74 mg,

0.25 mmol) in THF and stirred for 12 hours. The solvent was removed by vacuum, and the colorless solid was extracted with toluene and filtered. The toluene was condensed to 2 mL and cooled to deposit

colorless crystals of 9 (80 mg, 52%). Anal Calcd. for C54 H12 8Fe 2Na2 O8 Si6 : C, 52.65; H, 10.47. Found: C,

52.81; H 10.18.

3.8.13 Preparation of Co(OSitBu 2Me) 2(THF) 2 (9)

A THF solution of Na(DTBMS) (100 mg, 0.51 mmol) was added to a solution of CoCl 2(THF) 1.5

(121 mg, 0.25 mmol) in THF and stirred for 12 hours. The resulting blue solution was filtered and the

solvent was removed under vacuum to yield a blue oil. The oil was triturated with pentane 3 x 5 mL. Anal

Calcd. for C2 6H 58CoO4 Si 2 : C, 56.79; H, 10.63. Found: C, 56.31; H, 10.98.

t 3.8.14 Preparation of [Na(15-crown-5)][Co(OSi Bu2Me)3] (10)

A THF solution of Na(DTBMS) (300 mg, 1.53 mmol) was added to a suspension of

was removed by CoCl 2 (THF) 1 5 (114 mg, 0.48 mmol) in THF and stirred for 12 hours. The solvent

vacuum, and the resulting aqua blue solid was extracted with hexane and filtered to remove NaCl. 15-

crown-5 (100 uL, 0.50 mmol) was added dropwise to form a light green precipitate. The solid was filtered

and washed with pentane to yield 361 mg of 11 (91%). Anal. Caled. for C37 H83 CoNaO8 Si 3 : C, 54.05; H,

10.17. Found: C, 53.79; H, 9.95.

114 3.8.15 X-ray Crystallographic Tables

Table 3.9 X-ray Crystallographic Table for 2, 3, and 5.

2 3 5

Formula C 40H92CrNaO5 Si 4 C 36H 84CrO4Si4 C 37H 83MnNaO8 Si 3 fw, g/mol 840.49 745.39 818.23 Temperature 100(2) K 100(2) K 100(2) K

cryst. syst. Orthorhombic Monoclinic Monoclinic

space group Pbca C2/c P2(1)/n

color Blue Brown Colorless a (A) 21.8239(12) 22.888(2) 12.0120(9) b (A) 18.8675(10) 10.9647(9) 18.6736(14) c (A) 25.0854(14) 20.834(3) 21.2076(16) a(deg) 90 90 90 p (deg) 90 118.9420(10) 90.7080(10) y(deg) 90 90 90 V(A3) 10329.2(10) 4575.6(8) 4756.7(6) Z 8 4 4 no. refl. 225134 52150 97252 unique refl. 14525 7249 12882

Rint 0.0738 0.0376 0.0487 Ria (all data) 0.0694 0.0432 0.0537 wR2b(all data) 0.1217 0.0877 0.1093 RI [(I> 2a)] 0.0433 0.0321 0.0398 wR2 [I> 2a)] 0.1042 0.0799 0.0991 GOFc 1.052 1.074 1.038 2 2 2 2 2 1 2 cGOF = (E w(F 2 _ aR1 = 11Fo| - IFcI|/IIFol. bwR 2 = (I(w(F0 - F ) )/E(w(F0 ) )) . Fe221(n - p))112 where n is the number of data and p is the number of parameters refined.

115 Table 3.10 X-ray Crystallographic Table for 6, 7, and 8.

6 7 8

Formula C 40H 92Fe 2O 5Si4 C 47 5H 95FeNaO8 Si 3 C 27H6 4FeNaO 4Si 3 fw, g/mol 877.20 957.35 615.89 Temperature 100(2) K 100(2) K 100(2) K cryst. syst. Triclinic Triclinic Triclinic space group P-1 P-1 P-1 color Pale Yellow Pale Green Colorless a (A) 11.9118(6) 11.5787(10) 12.1314(12) b (A) 13.5204(7) 12.6964(11) 12.5601(12) c (A) 17.3906(8) 19.8210(17) 14.3997(14) a (deg) 68.0370(10) 82.502(2) 76.978(2)

0 (deg) 79.3680(10) 81.0680(10) 69.499(2) y (deg) 81.9930(10) 83.599(2) 62.242(2) V(A 3) 2545.3 2841.5(4) 1813.5(3) Z 2 2 2 no. refl. 72271 59396 44299 unique refl. 16134 14039 10819

Ri 0.0317 0.0432 0.0358

R1a (all data) 0.0313 0.0605 0.0459 wR2b(all data) 0.0688 0.1232 0.0878 RI [(I> 2a)] 0.0255 0.0438 0.0334 wR2 [I> 2a)] 0.0656 0.1116 0.0818 GOFC 1.035 1.036 1.054 2 2 2 2 2 2 2 aRI = I|Fo| - IFell/YIFoI. bwR2 = (I(w(F0 - F ) )/I(w(F ) ))1 . cGOF = (E w(F - Fe221(n - p))1 /2 where n is the number of data and p is the number of parameters refined.

116 Table 3.11 X-ray Crystallographic Table for 10.

Formula C 37H 83CoNaO8 Si 3 fw, g/mol 822.22

Temperature 100(2) K

cryst. syst. Monoclinic

space group P2(1)/n

color Pale Blue

a (A) 11.9662(8) b (A) 18.6544(13)

c (A) 21.0940(15)

a (deg) 90 p (deg) 90.7820(10) y (deg) 90

V(A 3) 4708.2(6) z 4

no. refl. 91432

unique refl. 13983 0.0430

Ria (all data) 0.0577 wR2b(all data) 0.1249 RI [(I> 2a)] 0.0438 wR2 [I > 2y)] 0.1131

GOFC 1.049 aR1 2 2 2 = ||F 0o| - |Fe||/EIFol. wR2 = (I(w(Fo - F ))/E(w(Foz ))" cGOF = (Y w(F0 - 1 2 IF22/1(n -1p)) / where n is the number of data and p is the number of parameters refined.

117 3.8.16 Cartesian Coordinates for Optimized Structures

Table 3.12 Cartesian coordinates for the DFT optimized geometry for the model of 3 Cr(OSiMe 3)4 (E = -2982.62944260 hartrees).

Atom x y z

0.003213 0.005729 -0.00242 -0.73109 -1.2248 -1.02851 0.844128 1.231154 -0.9489 1.176104 -0.78185 1.050623 -1.27918 0.79195 0.91575 1.542423 -2.23522 1.924843 -2.13746 -1.63293 -1.95927 2.33149 1.639303 -1.74306 -1.7366 2.226426 1.777268 -0.04048 -2.84739 2.778116 2.188423 -3.51448 0.678655 2.87247 -1.75417 3.190356 -1.60458 -3.05305 -3.10021 -3.49278 -2.18164 -0.74811 -2.68037 -0.10635 -2.95016 3.574327 2.17218 -0.40934 2.953745 0.116005 -2.69257 1.912268 3.06801 -2.92077 -3.28463 1.759155 2.772151 -0.30551 2.736828 2.918197 -2.10824 3.580384 0.49917 0.144579 -3.76701 3.349767 -0.43095 -2.08988 3.469538 -0.8221 -3.05837 2.036952 2.381875 -4.48294 1.160533 3.124039 -3.17506 0.215595 1.454293 -3.67048 -0.12178 2.493468 -0.98736 3.877616 3.187621 -2.62112 3.787078 3.759138 -1.34633 2.68824 -2.43875 -3.39615 -3.72757 -0.78891 -2.73691 -3.76242 -1.24765 -3.90867 -2.5123 -4.4376 -2.38776 -1.27011 -3.67658 -1.39853 -0.00175 -3.19581 -3.09318 -0.21348 -1.87596 0.236663 -3.61351

118 H -3.56176 -0.3258 -3.56867 H -2.9351 0.722423 -2.27725 H 3.225535 3.071127 0.115265 H 3.698928 1.375717 0.33484 H 4.558742 2.393862 -0.84397 H 3.895062 0.329431 -3.21674 H 2.21607 -0.2087 -3.43794 H 3.128638 -0.7225 -2.00652 H 1.491541 3.917074 -2.36698 H 1.170923 2.751501 -3.66507 H 2.805032 3.419426 -3.45611 H -3.66552 2.617768 3.34211 H -4.08367 1.41185 2.105039 H -3.06446 0.95044 3.480382 H -0.56008 3.633836 3.499708 H 0.598935 2.955047 2.335448 H -0.06335 1.930853 3.622715 H -2.96921 3.306341 -0.12447 H -2.33296 4.538837 0.98666 H -1.24659 3.726377 -0.16422

Table 3.13 Cartesian coordinates and atomic spin densities for the DFT optimized geometry of

(Me 3SiO) 3 CrO (E = -2573.34690776 hartrees). Cr-O = 1.57819 A.

Mulliken Atomic Atom x Y Z Spin Density

Cr -0.03216 -0.10512 -0.05939 0.958046 0 1.222236 -1.14212 0.620609 0.046962 0 0.223041 1.527503 0.49681 0.002918 0 -1.53935 -0.71982 0.610337 0.043601 0 -0.02674 -0.32434 -1.62227 -0.05777 Si 0.467799 3.154832 0.080375 0.001354 C -0.28616 3.478462 -1.61074 0.001012 C -0.37894 4.168875 1.415996 -0.00105 C 2.325323 3.44493 0.070255 -0.00182 Si 2.647524 -1.90244 0.118268 0.002779 C 2.182429 -3.3319 -1.0121 -0.0004 C 3.738845 -0.66443 -0.79019 0.003561 C 3.478276 -2.50901 1.692028 -0.00184 Si -3.0998 -1.15087 0.116738 0.00248

119 -3.82367 0.222104 -0.95038 0.003225 -2.98416 -2.75927 -0.85059 -0.00064 -4.08978 -1.36869 1.700065 -0.00155 -0.1277 4.526194 -1.91672 0.000301 0.163711 2.827824 -2.37762 0.000042 -1.37187 3.289778 -1.6039 -0.000071 -0.25386 5.248552 1.22858 -0.00015 0.043751 3.94238 2.407681 0.000048 -1.45869 3.95277 1.450532 0.000046 2.821309 2.809539 -0.68082 0.000057 2.556369 4.496542 -0.16853 -0.00028 2.763355 3.215042 1.054602 0.000074 3.081647 -3.86515 -1.36359 -0.00029 1.540345 -4.0555 -0.48488 0.000013 1.633377 -2.97068 -1.89619 0.000025 4.681538 -1.13533 -1.11599 0.000247 3.994302 0.19008 -0.14279 -0.00014 3.230122 -0.27596 -1.68742 -0.00018 2.824042 -3.21055 2.233592 0.000072 4.423936 -3.02877 1.463781 -0.00039 3.705017 -1.66826 2.366887 0.000068 -4.84534 -0.03235 -1.2791 0.000305 -3.87697 1.171782 -0.39339 -0.00015 -3.2118 0.386675 -1.8521 -0.00015 -3.98171 -3.08819 -1.18697 -0.00025 -2.55242 -3.55995 -0.22908 0.000037 -2.34727 -2.63793 -1.74126 0.000058 -4.10836 -0.43487 2.284366 0.000052 -5.13195 -1.65365 1.47812 -0.00031 -3.64706 -2.15462 2.332321 0.000063

120 3.9 References

(1) Betley, T. A.; Surendranath, Y.; Childress, M. V.; Alliger, G. E.; Fu, R.; Cummins, C. C.; Nocera, D. G. Phil. Trans. R Soc. B 2008, 363, 1293-1303.

(2) Betley, T. A.; Wu, Q.; Van Voorhis, T.; Nocera, D. G. Inorg. Chem. 2008, 47, 1849-186 1.

(3) Groysman, S.; Villagrin, D.; Nocera, D. G. Inorg. Chem. 2010, 49, 10759-10761.

(4) Groysman, S.; Villagran, D.; Freedman, D. E.; Nocera, D. G. Chem. Comm. 2011, 47, 10242- 10244.

(5) Frazier, B. A.; Wolczanski, P. T.; Lobkovsky, E. B. Inorg. Chem. 2009, 48, 11576-11585.

(6) Volpe, E. C.; Manke, D. R.; Bartholomew, E. R.; Wolczanski, P. T.; Lobkovsky, E. B. Organometallics2010, 29, 6642-6652.

(7) Cotton, F. A. ChemicalApplications of Group Theory; John Wiley & Sons: New York, 1990.

(8) Bower, B. K.; Tennent, H. G. J.Am. Chem. Soc. 1972, 94, 2512-2514.

(9) Byrne, E. K.; Theopold, K. H. J.Am. Chem. Soc. 1987, 109, 1282-1283.

(10) Rothwell, I. P. Acc. Chem. Res. 1988, 21, 153-159.

(11) Wolczanski, P. T. Polyhedron 1995, 14, 3335-3362.

(12) Wolczanski, P. T. Chem. Commun. 2009, 740-757.

(13) Schrock, R. R. Acc. Chem. Res. 2005, 38, 955-962.

(14) Krempner, C. Eur. J.Inorg. Chem. 2011, 2011, 1689-1698.

(15) Tolman, C. A. Chem. Rev. 1977, 77, 313-348.

(16) Chadeayne, A. R.; Wolczanski, P. T.; Lobkovsky, E. B. Inorg. Chem. 2004, 43, 3421-3432.

(17) Goto, K.; Okumura, T.; Kawashima, T. Chemistry Letters 2001, 30, 1258-1259.

(18) Shimo, I.; Okumura, T.; Goto, K.; Kawashima, T. J.Organomet. Chem. 2007, 692, 2115-2119.

(19) Sydora, 0. L.; Kuiper, D. S.; Wolczanski, P. T.; Lobkovsky, E. B.; Dinescu, A.; Cundari, T. R. Inorg. Chem. 2006, 45, 2008-2021.

121 (20) Johnson, B. F. G.; Klunduk, M. C.; Martin, C. M.; Sankar, G.; Teate, S. J.; Meurig Thomas, J. J. Organomet. Chem. 2000, 596, 221-225.

(21) Rost, M.; Grls, H.; Imhof, W.; Seidel, W.; Thiele, K. Z anorg.allgem. Chem. 1998, 624, 1994- 2000.

(22) Krinsky, J. L.; Anderson, L. L.; Arnold, J.; Bergman, R. G. Inorg. Chem. 2007, 47, 1053-1066.

(23) Mansfeld, D.; Mehring, M. Z anorg.allgem. Chem. 2005, 631, 2429-2432.

(24) McGeary, M. J.; Coan, P. S.; Folting, K.; Streib, W. E.; Caulton, K. G. Inorg. Chem. 1991, 30, 1723-1735.

(25) Veige, A. S.; Slaughter, L. M.; Lobkovsky, E. B.; Wolczanski, P. T.; Matsunaga, N.; Decker, S. A.; Cundari, T. R. Inorg. Chem. 2003, 42, 6204-6224.

(26) Neithamer, D. R.; LaPointe, R. E.; Wheeler, R. A.; Richeson, D. S.; Van Duyne, G. D.; Wolczanski, P. T. J. Am. Chem. Soc. 1989, 111, 9056-9072.

(27) Stensland, B.; Kierkegaard, P. Acta Chemica Scandinavica 1970, 24, 211-220.

(28) Marciniec; Maciejewski, H. Coord Chem Rev 2001, 223, 301-335.

(29) Preuss, F.; Werle, P. Z. Naturforsch., B: Chem. Sci. 2002, 57, 726-730.

(30) Jenkins, L. S.; Willey, G. R. J. Chem. Soc., Dalton Trans. 1979, 1697-1700.

(31) Gibson, V. C.; Kee, T. P.; Shaw, A. Polyhedron 1988, 7, 579-580.

(32) Kornev, A. .; Chesnokova, T. .; Semenov, V. .; Zhezlova, E. .; Zakharov, L. .; Klapshina, L..; Domrachev, G. .; Rusakov, V. .J Organomet. Chem. 1997, 547, 113-119.

(33) Chesnokova, T. A.; Zhezlova, E. V.; Kornev, A. N.; Fedotova, Y. V.; Zakharov, L. N.; Fukin, G. K.; Kursky, Y. A.; Mushtina, T. G.; Domrachev, G. A. J. Organomet. Chem. 2002, 642, 20-31.

(34) Sekiguchi, S.; Kurihara, A. Bull. Chem. Soc. Japan. 1969, 42, 1453-1454.

(35) Lugmair, C. G.; Tilley, T. D. Z Naturforsch. 2004, 59b, 1540-1547.

(36) Covert, K. J.; Wolczanski, P. T.; Hill, S. A.; Krusic, P. J. Inorg. Chem. 1992, 31, 66-78.

(37) Veith, M.; Voigt, T.; Huch, V. Z anorg allgem. Chem. 2006, 632, 243-247.

(38) Mansfeld, D.; Schiirmann, M.; Mehring, M. Appl. Organomet. Chem. 2005, 19, 1185-1188.

122 (39) Feher, F. J.; Blanski, R. L. Organometallics 1993, 12, 958-963.

(40) Vennat, M.; Bregeault, J.-M.; Herson, P. Dalton Trans. 2004, 908-913.

(41) Covert, K. J.; Neithamer, D. R.; Zonnevylle, M. C.; LaPointe, R. E.; Schaller, C. P.; Wolczanski, P. T. Inorg. Chem. 1991, 30, 2494-2508.

(42) Vennat, M.; Herson, P.; Br6geault, J.; Shul'pin, G. B. Eur. J. Inorg. Chem. 2003, 2003, 908-917.

(43) Qiu, P.; Cheng, R.; Liu, B.; Tumanskii, B.; Batrice, R. J.; Botoshansky, M.; Eisen, M. S. Organometallics2011, 30, 2144-2148.

(44) Muzart, J.; Ajou, A. N. J.Mol. Cat. 1993, 84, L15-L19.

(45) Stavropoulos, P.; Bryson, N.; Youinou, M. T.; Osborn, J. A. Inorg. Chem. 1990, 29, 1807-1811.

(46) Qiu, P.; Cheng, R.; Liu, Z.; Liu, B.; Tumanskii, B.; Eisen, M. S. J.Organomet. Chem. 2012, 699, 48-55.

(47) Fujdala, K. L.; Tilley, T. D. Chem. Mater. 2001, 13, 1817-1827.

(48) Sydora, 0. L.; Wolczanski, P. T.; Lobkovsky, E. B.; Buda, C.; Cundari, T. R. Inorg. Chem. 2005, 44, 2606-2618.

(49) Otto, M.; Wagener, R.; Hennig, H. InorganicaChimica Acta 1982, 64, L 1-L14.

(50) Terry, K. W.; Gantzel, P. K.; Tilley, T. D. Inorg. Chem. 1993, 32, 5402-5404.

(51) Alyea, E. C.; Basi, J. S.; Bradley, D. C.; Chisholm, M. H. J.Chem. Soc. A 1971, 772-776.

(52) Ciborska, A.; Chojnacki, J.; Wojnowski, W. Acta Cryst. E. 2007, 63, ml 103-mi 104.

(53) Blanchard, H.; Hursthouse, M. B.; Sullivan, A. C. J.Organomet. Chem. 1988, 341, 367-7 1.

(54) Bochmann, M.; Wilkinson, G.; Young, G. B.; Hursthouse, M. B.; Malik, K. M. A. J.Chem. Soc., Dalton Trans. 1980, 1863-1871.

(55) Liu, G.; Greedan, J. E.; Gong, W. J.Solid State Chem. 1993, 105, 78-9 1.

(56) Hagihara, N.; Yamazaki, H. J. Am. Chem. Soc. 1959, 81, 3160-3161.

(57) Basi, J. S.; Bradley, D. C.; Chisholm, M. H. J. Chem. Soc. A 1971, 1433-1436.

(58) Mowat, W.; Shortland, A. J.; Hill, N. J.; Wilkinson, G. J.Chem. Soc., Dalton Trans. 1973, 770- 778.

123 (59) Kiremire, E. M. R. . Chem. Educ. 1989, 66, 479.

(60) Odom, A. L.; Mindiola, D. J.; Cummins, C. C. Inorg. Chem. 1999, 38, 3290-3295.

(61) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. Advanced InorganicChemistry, 6th Ed.; Wiley-Interscience: New York, 1999.

(62) Petri'evid, V.; Gayen, S. K.; Alfano, R. R. Appl. Phys. Lett. 1988, 53, 2590-2592.

(63) Verdun, H. R.; Thomas, L. M.; Andrauskas, D. M.; McCollum, T.; Pinto, A. Appl. Phys. Lett. 1988, 53, 2593-2595.

(64) Reinen, D.; Kesper, U.; Atanasov, M.; Roos, J. Inorg. Chem. 1995, 34, 184-192.

(65) Eilers, H.; Hmmerich, U.; Jacobsen, S. M.; Yen, W. M.; Hoffman, K. R.; Jia, W. Phys. Rev. B 1994, 49, 15505-15513.

(66) Jia, W.; Liu, H.; Jaffe, S.; Yen, W. M.; Denker, B. Phys. Rev. B 1991, 43, 5234-5242.

(67) Liu, F.; John, K. D.; Scott, B. L.; Baker, R. T.; Ott, K. C.; Tumas, W. Angew. Chem. Int. Ed. 2000, 39, 3127-3130.

(68) Shapley, P. A.; Bigham, W. S.; Hay, M. T. InorganicaChimica Acta 2003, 345, 255-260.

(69) Hay, M. T.; Hainaut, B. J.; Geib, S. J. Inorg. Chem. Comm. 2003, 6, 431-434.

(70) Holland, A. W.; Li, G.; Shahin, A. M.; Long, G. J.; Bell, A. T.; Tilley, T. D. . Catal. 2005, 235, 150-163.

(71) Nozaki, C.; Lugmair, C. G.; Bell, A. T.; Tilley, T. D. J Am. Chem. Soc. 2002, 124, 13194-13203.

(72) Schmidbaur, H.; Richter, W. Chem. Ber. 1974, 107, 2427-2433.

(73) Sigel, G. A.; Bartlett, R. A.; Decker, D.; Olmstead, M. M.; Power, P. P. Inorg. Chem. 1987, 26, 1773-1780.

(74) Kornev, A. N.; Chesnokova, T. A.; Zhezlova, E. V.; Fukin, G. K.; Zakharov, L. N.; Domrachev, G. A. Dokl. Akad. Nauk 1999, 369, 203-206.

(75) Kownacki, I.; Kubicki, M.; Marciniec, B. Polyhedron 2001, 20, 3015-3018.

(76) Klose, A.; Solari, E.; Ferguson, R.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C. Organometallics 1993, 12, 2414-2416.

(77) Becke, A. D. PhysicalReview A 1988, 38, 3098-3100.

124 (78) Becke, A. D. J. Chem. Phys. 1993, 98, 1372-1377.

(79) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.

(80) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785-789.

(81) Gaussian 03, Revision B.05; Frisch, M.; Trucks, G.; Schlegel, H.; Scuseria, G.; Robb, M.; Cheeseman, J.; Montgomery, J.; Vreven, T.; Kudin, K.; Burant, J.; Millam, J.; Iyengar, S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, 0.; Nakai, H.; Klene, M.; Li, X.; Knox, J.; Hratchian, H.; Cross, J.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R.; Yazyev, 0.; Austin, A.; Cammi, R.; Pomelli, C.; Ochterski, J.; Ayala, P.; Morokuma, K.; Voth, G.; Salvador, P.; Dannenberg, J.; Zakrzewski, V.; Dapprich, S.; Daniels, A.; Strain, M.; Farkas, 0.; Malick, D.; Rabuck, A.; Raghavachari, K.; Foresman, J.; Ortiz, J.; Cui, Q.; Baboul, A.; Clifford, S.; Cioslowski, J.; Stefanov, B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R.; Fox, D.; Keith, T.; Laham, A.; Peng, C.; Nanayakkara, A.; Challacombe, M.; Gill, P.; Johnson, B.; Chen, W.; Wong, M.; Gonzalez, C.; Pople, J. Gaussian, Inc., Pittsburgh PA, 2003.

(82) Schafer, A.; Huber, C.; Ahlrichs, R. J Chem. Phys. 1994, 100, 5829-5835.

(83) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571-2577.

125 Chapter 4: Tripodal Amine Ligands Towards the Stabilization of High-Valent Metal Oxo Complexes

126 4.1 Introduction

Amide ligands have extremely rich transition metal coordination chemistry due to their steric and electronic diversity.' With the general formula -NR 2, they have two functionalities that can be manipulated to impose optimal steric and electronic properties on the coordinated transition metal. Amide ligands can participate in both ionic and covalent bonding, and as a result, they have the rare ability to stabilize transition metals in both high and low oxidation states.

The ability of amide ligands to stabilize high oxidation state transition metal oxo compounds is best demonstrated by Collins, who has prepared a series of tetra amide macrocyclic ligands (TAML).2 By carefully blocking common modes of decomposition, the ligands are robust in strongly oxidizing aqueous

environments.3 In this fashion, Collins and coworkers have isolated oxo complexes of Crv, Mnv, FeIv, and

Fev .4-9 Recently, Collins showed that (TAML)Fe complexes can oxidize water to 02 through chemical

oxidation with Cerv salts or NaIO 4.' The ability of these amide ligands to stabilize reactive tetragonal oxo

species has inspired our work to synthesize trigonal oxo complexes using tris amide ligands.

NH HN NH 2 NH 2 NH 2 NH T.- INH 2 NH 2

NH HN - NH2

TAML TAME TACH

Mono-amide ligands have been shown to support trigonal oxo complexes, but in very oxidizing

cases ligand dissociation can cause unwanted degradation. For example, treatment of Cr"' trisamides with

2 dioxygen gives (amide) 2CrV1(0) 2 complexes with loss of an amide radical."" While Cummins et al. have

gone on to show that these species can be functionalized to support trigonal CrV oxo complexes, one way

to possibly prevent this mode of decomposition is to utilize chelating amide ligands. These also have the

added benefit of enforcing a pseudo-tetrahedral geometry on the metal complex in a similar fashion to the

127 3-tert-butyl substituted tris-pyrazolyl borate system, which has been coined the "tetrahedral enforcer.""

In light of these criteria, chelating tris amide ligands were targeted using the tri(aminomethyl)ethane

(TAME) and triaminocyclohexane (TACH) backbones.

4.1.1 Tri(aminomethyl)ethane (TAME)

The tripodal TAME ligand was first synthesized by Stetter and B6ckmann in 1951 as a synthetic analog to urotropine (hexamethylenetetramine),1 4 however, it was rarely used as a ligand5 ' 16 until a higher yielding synthesis by Fleischer made large scale quantities readily available." In neutral form, TAME tends to coordinate pseudo-octohedral transition metals in afac- arrangement, with

Mn,19 and Co.20 The addition of bulky alkyl, silyl, and aryl groups on the amine functionality can block polymetallic chemistry, but the added sterics weaken the interaction of the metal to a neutral amine.

Consequently, deprotonation of the amine groups to form a tris-amide ligand has typically been used to synthesize molecular transition metal species.

Gade has synthesized a number of derivatives featuring a trianionic TAME backbone,21 22 which has allowed synthesis of the corresponding titanium complexes.'23,24 Titanium complexes of TAME-'Pr 3 displayed much decomposition and could not be isolated cleanly, while use of TAME-(SiMe 2R)3 (R =

Me, tBu) resulted in oils that could be purified by distillation, but decomposed upon standing at 5*C.

Nevertheless, crude solutions of these titanium complexes could be used in situ to form interesting heterobimetallic systems featuring unsupported Ti-M bonds (M = Fe, Co, Ru).2 5 Aside from titanium, the only other transition metal coordinated to an unsupported trianionic TAME ligand is a vanadium-N 2

26 2 dimer; however the synthesis of the Mo dimer [(TAME-'Pr 3)MoH] 2 has been insinuated."

128 4.1.2 Triaminocyclohexane (TACH)

TACH features a similar tripodal framework as TAME, but with a more rigid cyclohexane backbone. This structural motif allows the amino groups to adopt axial or equatorial positions, as the

cyclohexane backbone undergoes ring flips. While the amino groups typically stay equatorial in solution,

chelation to a metal ion can cause the amine groups to adopt an all-axial arrangement. Since there is not a

low energy pathway for a single amine to dissociate from the metal complex, TACH complexes are

considered more structurally rigid. The metalloadamantane structure is also thought to convey additional

rigidity compared to the TAME system. This has allowed Turculet and Tilley to isolate and structurally

characterize mononuclear TACH complexes of zirconium and titanium. 28 ,29 While amide complexes of

TACH have not been reported outside of group 4, neutral imine derivatives of TACH coordinated to

Fe(II) and Co(II) have been synthesized and structurally characterized.30''

4.1.3 Synthetic Strategies For Extension to Late Transition Metals

Given the scope of coordination chemistry and catalysis that other triamide ligands such as TREN

have afforded,32 the scarcity of TAME or TACH transition metal amide complexes beyond group 4

suggests that there may be a fundamental flaw involving the geometric or electronic properties of these

ligand frameworks. When one of these ligands coordinates to a metal center, three 6-membered

metallacyclohexane are formed in boat and chair conformations respectively. Although in both cases the

rings have rigid conformations, the magnitude of the chelate effect is partially offset by the fact that 6-

member chelate rings are considerably less stable than 5-membered rings.3-" The nature of this

instability has been described both by geometrical analysis as a function of M-N bond length and

molecular mechanics calculations. 36 This difference in 5 versus 6-membered chelate stability may explain

why TREN is a versatile transition metal chelate, while corresponding triaminopropylamine (TRPN)

complexes have not been reported beyond group 4.

129 It has been suggested that the bite angle of TAME and TACH ligands limits the chelate stability of these species.27 Although N-M-N angles are at or below 900 in most complexes, the geometric constraints alone cannot explain the formation of stable complexes featuring relatively large metal ions such as zirconium or thallium. 28,29,38 In addition, neutral forms of these ligands coordinate all of the first row transition metals.

Amide complexes of TAME and TACH feature metals that have particularly high reduction potentials and form strong ionic M-N bonds. Still, metalation of group 4 metal species generally gives yields well below 50%, despite the fact that corresponding monodentate amide complexes are typically synthesized in high yield. Gade has shown that titanium complexes of TAME-'Pr 3 are less stable than

23 TAME-(SiMe 3)3, which suggests that more strongly donating amide substituents decrease the stability of the resulting metal complexes. Furthermore, the presence of p-hydrogens in both ligand backbones potentially affords a facile means of degradation. In light of these electronic considerations and mindful of the peculiar void of mid and late transition metal complexes of TAME and TACH, several derivatives of TAME featuring electron withdrawing amide substituents became an initial target for research in our

9 4 0 lab.3 ,

130 4.2 Results

Preliminary metalation of electron withdrawing TAME derivatives with FeBr2 was reported," however the iron species could only be isolated as an impure viscous oil or powder. Addition of PhIO to this species in THF gave the free ligand and iron oxide. Since even the titanium derivatives of TAME could only be isolated as oils, the more rigid TACH backbone was targeted as an alternative. Synthesis

4 1 4 2 proceeds smoothly according to literature procedures to give the TACH free base in two steps. ,

43 The tri-tosyl derivative (L1) was prepared using a modified literature procedure by treating the

triamine free base with 4 eq. of Tos-Cl and 4 eq. Et3N in ether. The product was conveniently crystallized from ethanol to yield L1 in 94% yield as a colorless crystalline solid. Analysis of L1 matched previous reports, and the structure of Li was further characterized by X-ray crystallography (Figure 4.1).

The structure of L1 shows the central cyclohexane ring in a chair conformation with the sulfonamide groups in equatorial positions. The N-H bonds are each pointing approximately towards the same direction, though they do not appear to hydrogen bond with a neighboring ligand. In solution, the 1H

131 NMR shows that L1 retains this rigid conformation, each of the three unique protons of the cyclohexane backbone are resolved (Figure 4.2)

0Q~ F Tos GF S N D E TsN C

B A

E D

G C CHCNB A

8 7 6 5 4 3 2 1 ppm

1 Figure 4.2 H NMR Spectrum (CH 3CN) of L1 .

L1 is soluble in polar solvents such as THF, DMSO, acetonitrile, and dichloromethane, while insoluble in hydrocarbon or aromatic solvents. Addition of BuLi to L1 in THF gives a colorless solution

until 3 eq are added, upon which a milky suspension of Li 3L1 is produced. This species proved insoluble in all common solvents including THF, DMSO, and DMF. Nevertheless, this solid powder could be isolated or used in situ for metalation of metal-halide species.

Addition of solid VOF3 to a suspension of Li3L1 in THF yielded a pale yellow solution with formation of 3 eq. of LiF precipitate. The yellow-brown product L 1VO proved insoluble in non-polar organic solvents, but the addition of a small amount of THF greatly improved the solubility of this

1 species. The H NMR spectrum (figure 4.3) shows a C3 symmetric species that looks similar to that of L1

132 except for the absence of N-H resonances as expected. In addition, the hydrogen resonance (C) is shifted downfield by about 1 ppm, which is consistent with amide coordination to an electropositive metal

3 species. The 1 C NMR showed a similar downfield shift of the carbon atom bound to nitrogen to 67.0 ppm from 48.3 ppm.

0 III 11/ S Tos V

Tos N E

C Toluene F Toluene A

THF THF

E D Ether C ther B A

9 B 7 6 5 4 3 2 1 0 ppm

Figure 4.3 'H NMR Spectrum (toluene) of L1VO.

Numerous attempts were made to crystallize this species, but it appeared that loss of coordinating solvent induced the precipitation of a yellow powder. Crystallization was also attempted in the presence of LiCl, LiPF6 , and MgBr 2 to see if the presence of coordinating salts could stabilize the oxo species. The

51 V NMR spectrum showed a small peak at -1666 ppm, which may be indicative of a reduced vanadium species. Due to the numerous S=O stretches in the infrared spectrum, the V=O stretch could not be identified.

Metalation of L, with other transition metal salts proceeded smoothly but gave only solid powders. Since metalation with FeBr 2 was previously reported with a similar TAME tri-sulfonamide

133 system, the analogous reaction was attempted repeatedly. Addition of Li3Li to a THF solution of FeBr2 gave a pale yellow solution after filtration of 2 eq. of LiBr. Addition of 12-crown-4 gave an insoluble species, however in one case, a single pale green crystal was recovered from a concentrated THF solution, the structure of which is shown in figure 4.4.

Figure 4.4 X-ray crystal structure of [K(12-crown-4) 2][LiL1FeOH] with 50% probability ellipsoids. The K(12-crown-4) 2, 2 molecules of THE, and hydrogen atoms were removed for clarity.

Table 4.1 Selected bond lengths (A) and angles (0) of [K(12-crown-4) 2][LiL1 FeOH].

Bond Lengths Fe(1)-O( 1) 1.940(4) Fe(1)-N(1) 2.069(5) Fe(1)-N(2) 2.082(6) Fe(1)-N(3) 2.013(6) Li(1)-O(1) 1.987(13) Li(1A)-O(1) 1.972(13) N(1)-S(1) 1.555(5) N(2)-S(2) 1.561(6) N(3)-S(3) 1.566(6)

134 Bond Angles O(1)-Fe(1)-N(1) 118.7(2) O(1)-Fe(1)-N(2) 114.4(2) O(1)-Fe(1)-N(3) 135.0(2) N(1)-Fe(1)-N(2) 88.2(2) N(1)-Fe(1)-N(3) 92.9(2) N(2)-Fe(1)-N(3) 96.6(2)

Although the quality of the structure is poor (R = 0.12), the connectivity shows L1 coordinating to a dimeric Fe-O(H)Li fragment. The TACH amides coordinate as if capping the face of an octahedron, with N-Fe-N bond angles averaging 92.60 (table 4.1), and the nitrogen atoms appear to be sp 2 hybridized due to the electron withdrawing nature of the sulfonamide group. The lithium ions are coordinated in a tetrahedral fashion and bind sulfonate and hydroxide groups from each metal unit. These lithium ions appear to be critical to stabilize this unusual structure. Assignment as a hydroxide was assumed due to the long Fe-O bond length and the Fe-N bond lengths which are consistent with an Fe(II) oxidation state.

Although the source of the KOH is unknown, attempts to reproduce the synthesis of this species failed repeatedly. A rational synthesis of this complex was attempted by treatment of Li with KN(SiMe 3)2 followed by treatment with 2 equivalents of 12-crown-4 to give [K(12-crown-4) 2][H 2L 1]. This species

was then treated with V2 eq. of [Fe(mesityl) 2] 2 . Finally, 1 equivalent of anhydrous LiOH was added. Using this methodology under various reaction conditions did not yield the above species. The treatment of various [M][L 1Fe"] (M = Li, K, Mg, Fe, Ph4P, (Ph 3P)2N, Me4N) species with a THF solution of 0.1 M

t H20 and base (BuLi, KN(SiMe 3)2, LiO Bu) was also unsuccessful.

135 4.3 Other TACH Derivatives

Since salt metathesis with Li3L1 with metal halide salts as well as treatment of Li with

[Li(TIF) 4][Co(mesityl) 3, [Fe(mesityl) 2] 2, and metal amides M{N(SiMe 3)2}2(THF) 2 (M = Co, Fe, Mn)

gave suspected metallopolymers rather than monometalated species, new amide functionalities were

investigated. Use of mesityl- or 1,3,5-triisopropylphenyl-sulfonyl chloride did not give the corresponding

tris sulfonamide TACH systems possibly due to the extreme steric bulk of these derivatives. TACH could

28 be treated with C6F6 to make the perfluorophenyl derivative reported by Tilley. Metalation of this

derivative on transition metals outside of group 4 resulted in the formation of insoluble black precipitates,

which are likely reduced metal species arising from oxidation of the ligand. Treatment of TACH with

isobutyryl chloride gave the tris-amide molecule shown below. Isolation of metal species with this

triamide was also unsuccessful.

NH, C HN

Et3N, -Et3NHCI

HN NH, 0 N NH

136 4.4 Conclusions

Despite the rich coordination chemistry of amide ligands, the coordination chemistry of TACH and TAME amide derivatives is severely limited.' Use of alkyl or aryl functionalities appears to give an metal environment that is too reducing for transition metals beyond group 4, while the addition of carbamide or sulfonamide functionalities yields species that tend to oligomerize rather than chelate in the desired fashion. Furthermore, the presence of p-hydrogens inherent in the amine backbone provides a facile means of decomposition even if oxidizing species were to by formed. Betley et al. have synthesized derivatives of TACH and TAME featuring further amine functionality have been shown to support metal clusters, which feature unusual bonding, magnetism, and redox chemistry.44 -8 They have also synthesized tris-pyrrole ligands and isolated pseudotetrahedral complexes of Mn, Fe, Co, Ni, and Zn.49 While this ligand platform is promising for a variety of small molecule activation studies, it seems to be unstable towards oxidizing or reducing environments."0 While there now exists interesting multimetallic coordination chemistry featuring TAME and TACH ligand scaffolds, the derivatives described here appear unable to support mononuclear pseudo-tetrahedral metal oxo species in high oxidation states.

137 4.5 Experimental Section

4.5.1 Preparation of L1

TACH (0.525g, 4.06 mmol) was dissolved in 100 mL of diethylether and triethylamine (1.23g,

12.2 mmol). A solution of 4-tolunesulfonyl chloride (2.33g, 12.2 mmol) in 100 mL was added dropwise to give a cloudy solution. The reaction was stirred for 24 h, after which the ether was removed under vacuum. To the solid was added water (50 mL) and dichloromethane (20 mL), and the mixture was further extracted with dichloromethane (2 x 20 mL). The solvent was removed under vacuum to give a pale yellow oil which crystallized upon adding ethanol (20 mL). The solid was washed with ethanol (10 mL) and ether (3 x 10 mL), and dried under vacuum to give 2.35 g L, in 98% yield. Anal Calcd. for

C27 H33N 3 0 6 S3 : C, 54.78; H, 5.63; N, 7.10. Found: C, 54.66; H, 5.61; N, 7.47.

4.5.2 Preparation of L1VO

L, (100 mg, 0.168 mmol) was treated with "BuLi (0.51 mmol) in THF (4 mL) to give an opaque white colored solution. A THF solution of VOF3 (21 mg, 0.17 mmol) was added dropwise to give a pale yellow solution with a fine white precipitate of LiF. The solution was condensed to 0.5 mL and 4 mL of toluene was added. The solution was filtered and the solvent was removed to yield L1VO (98 mg, 89%) as

a yellow solid. 'H NMR(C 7 D8 ): 6 = 7.81 (d, 6H, J = 8.2 Hz), 6.76 (d, 6H, J = 8.2 Hz), 4.09 (br. s., 3H),

3 1.84 (s, 9H), 0.95 (d, 3H, J = 15 Hz), 0.61 (d, 3H, J = 15 Hz). 1 C NMR (DMSO-d6): 6 = 142.3, 139.1,

129.6, 126.0, 67.0, 25.1, 21.0. "V NMR (DMSO-d 6): 6 = -1666jka. Anal Calcd. for C2 7H 30N 30 7 S3V: C,

49.46; H, 4.61; N, 6.41. Found: C, 49.77; H, 4.25; N, 6.18.

4.5.3 Preparation of [K(12-crown-4) 2][LiL 1 FeOH]

L, (83 mg, 0.14 mmol) was treated with "BuLi (0.42 mmol) in THF (3 mL) to give an opaque

white colored solution. A THF solution of FeBr2 (30 mg, 0.14 mmol) was added dropwise to give a pale

138 yellow solution. A THF solution of 12-crown-4 (25mg, possibly contaminated with KOH) was added dropwise, and the reaction was left to stand for 12 h. A single pale green crystal was recovered from a mixture of LiBr (colorless) and a brown colored powder.

139 4.5.4 X-ray Crystallographic Table

Table 4.2 X-ray Crystallographic Table for L1 and [K(12-crown-4) 2][LiL1 FeOH].

Li [K(12-c-4)2][LiL1FeOH]

Formula C2 7H 33N 3 0 6 S 3 C 5 1H 79FeKLiN 30 17 S 3 fw, g/mol 591.74 1204.24 Temperature 100(2) K 100(2)K

cryst. syst. Monoclinic Monoclinic

space group P2(1) P2(1)/n

color Colorless Pale Green

a (A) 13.658(3) 19.210(3) b (A) 6.5508(12) 12.3369(19) c (A) 16.261(3) 24.958(4) a (deg) 90 90 p (deg) 106.462(3) 105.925(3) y(deg) 90 90 V(A 3 ) 1395.3(4) 5687.9(15) Z 2 4 no. refi. 31976 128225 unique refl. 8399 17318 Rmt 0.0835 0.2552 Ria (all data) 0.1383 0.2421 wR2b(all data) 0.3404 0.3400 RI [(I > 2;)] 0.1142 0.1161 wR2 [I> 2a)] 0.3246 0.2634 GOFC 1.090 1.012 2 2 2 2 2 2 aRI = ElIFo| - |Fe|/IIFol. bwR 2 = (E(w(Fz0 Fe))/E(w(F0 ) ))1. cGOF =( 2 2 2 w(Fo2 - Fe ) /(n - p))" where n is the number of data and p is the number of parameters refined.

140 4.6 References

(1) Lappert, M.; Protchenko, A.; Power, P.; Seeber, A. Metal Amide Chemistry; 2008.

(2) Bartos, M. J.; Gordon-Wylie, S. W.; Fox, B. G.; James Wright, L.; Weintraub, S. T.; Kauffmann, K. E.; MUnck, E.; Kostka, K. L.; Uffelman, E. S.; Rickard, C. E. F.; Noon, K. R.; Collins, T. J. Coord. Chem. Rev. 1998, 174, 361-390.

(3) Collins, T. J. Acc. Chem. Res. 1994, 27, 279-285.

(4) Collins, T. J.; Slebodnick, C.; Uffelman, E. S. Inorg. Chem. 1990, 29, 3433-3436.

(5) Collins, T. J.; Gordon-Wylie, S. W. J. Am. Chem. Soc. 1989, 111, 4511-4513.

(6) Collins, T. J.; Powell, R. D.; Slebodnick, C.; Uffelman, E. S. J. Am. Chem. Soc. 1990, 112, 899- 901.

(7) Collins, T. J.; Kostka, K. L.; Munck, E.; Uffelman, E. S. J Am. Chem. Soc. 1990, 112, 5637-5639.

(8) Chanda, A.; Shan, X.; Chakrabarti, M.; Ellis, W. C.; Popescu, D. L.; Tiago de Oliveira, F.; Wang, D.; Que, L.; Collins, T. J.; Mtinck, E.; Bominaar, E. L. Inorg. Chem. 2008, 47, 3669-3678.

(9) De Oliveira, F. T.; Chanda, A.; Banerjee, D.; Shan, X.; Mondal, S.; Que, L.; Bominaar, E. L.; Mfinck, E.; Collins, T. J. Science 2007, 315, 835-838.

(10) Ellis, W. C.; McDaniel, N. D.; Bernhard, S.; Collins, T. J. J. Am. Chem. Soc. 2010, 132, 10990- 10991.

(11) Ruppa, K. B. P.; Feghali, K.; Kovacs, I.; Aparna, K.; Gambarotta, S.; Yap, G. P. A.; Bensimon, C. 1 Chem. Soc., Dalton Trans. 1998, 1595-1606.

(12) Odom, A. L.; Mindiola, D. J.; Cummins, C. C. Inorg. Chem. 1999, 38, 3290-3295.

(13) Trofimenko, S. Scorpionates:Polypyrazolylborate Ligands and Their CoordinationChemistry; 1st ed.; World Scientific Publishing Company, 1999.

(14) Stetter, H.; Bockmann, W. Chem. Ber. 1951, 84, 834-839.

(15) Durham, D. J Inorg. Nuc. Chem. 1967, 29, 509.

(16) Kitajiri, N.; Arishima, T.; Takamoto, S. Nippon Kagaku Zasshi 1970, 91, 240-3.

(17) Fleischer, E. B.; Gebala, A. E.; Levey, A.; Tasker, P. A. J. Org. Chem. 1971, 36, 3042-3 044.

141 (18) Wieghardt, K.; Schmidt, W.; Endres, H.; Robert Wolfe, C. Chem. Ber. 1979, 112, 2837-2846.

(19) Hagen, K. S.; Westmoreland, T. D.; Scott, M. J.; Armstrong, W. H. J. Am. Chem. Soc. 1989, 111, 1907-1909.

(20) Williams, N. H.; Wyman, P. J. Chem. Soc., Perkin Trans. 2 2001, 2068-2073.

(21) Gade, L. H.; Becker, C.; Lauher, J. W. Inorg. Chem. 1993, 32, 2308-2314.

(22) Gade, L. H.; Mahr, N. J. Chem. Soc., Dalton Trans. 489-494.

(23) Friedrich, S.; Gade, L. H.; Edwards, A. J.; McPartlin, M. Chem. Ber. 1993, 126, 1797-1805.

(24) Friedrich, S.; Gade, L. H.; Li, W.; McPartlin, M. Chem. Ber. 1996, 129, 1287-1291.

(25) Friedrich, S.; Memmler, H.; Gade, L. H.; Li, W.-S.; Scowen, I. J.; McPartlin, M.; Housecroft, C. E. Inorg. Chem. 1996, 35, 2433-2441.

(26) Desmangles, N.; Jenkins, H.; Ruppa, K. B.; Gambarotta, S. InorganicaChimica Acta 1996, 250, 1-4.

(27) Gade, L. H. Acc. Chem. Res. 2002, 35, 575-582.

(28) Turculet, L.; Tilley, T. D. Organometallics2002, 21, 3961-3972.

(29) Turculet, L.; Tilley, T. D. Organometallics2004, 23, 1542-1553.

(30) Archibald, S. J.; Foxon, S. P.; Freeman, J. D.; Hobson, J. E.; Perutz, R. N.; Walton, P. H. J Chem. Soc., Dalton Trans. 2797-2799.

(31) Rocks, S. S.; Brennessel, W. W.; Machonkin, T. E.; Holland, P. L. Inorg. Chem. 2010, 49, 10914- 10929.

(32) Blackman, A. G. Polyhedron 2005, 24, 1-39.

(33) Zipp, S. G.; Zipp, A. P.; Madan, S. K. Coord. Chem. Rev. 1974, 14, 29-45.

(34) Paoletti, P.; Fabbrizzi, L. Coord. Chem. Rev. 1975, 15, 435-440.

(35) Comba, P. Coord. Chem. Rev. 1993, 123, 1-48.

(36) Hubin, T. J. Coord.Chem. Rev. 2003, 241, 27-46.

(37) Schrock, R. R.; Cummins, C. C.; Wilhelm, T.; Lin, S.; Reid, S. M.; Kol, M.; Davis, W. M. Organometallics1996, 15, 1470-1476.

142 (38) Hellmann, K. W.; Gade, L. H.; Steiner, A.; Stalke, D.; M61ler, F. Angew. Chem. Int. Ed. 1997, 36, 160-163.

(39) Betley, T. A.; Surendranath, Y.; Childress, M. V.; Alliger, G. E.; Fu, R.; Cummins, C. C.; Nocera, D. G. Phil. Trans. R. Soc. B 2008, 363, 1293-1303.

(40) Betley, T. A.; Wu, Q.; Van Voorhis, T.; Nocera, D. G. Inorg. Chem. 2008, 47, 1849-1861.

(41) Bowen, T.; Planalp, R. P.; Brechbiel, M. W. Bioorg. Med Chem. Lett. 1996, 6, 807-810.

(42) Yang, J. Y.; Shores, M. P.; Sokol, J. J.; Long, J. R. Inorg. Chem. 2003, 42, 1403-1419.

(43) Park, G.; Lu, F. H.; Ye, N.; Brechbiel, M. W.; Torti, S. V.; Torti, F. M.; Planalp, R. P. J. BiO. Inorg. Chem. 1998, 3, 449-457.

(44) Scharf, A. B.; Betley, T. A. Inorg. Chem. 2011, 50, 6837-6845.

(45) Powers, T. M.; Fout, A. R.; Zheng, S.-L.; Betley, T. A. J. Am. Chem. Soc. 2011, 133, 3336-3338.

(46) Harris, T. D.; Zhao, Q.; Sinchez, R. H.; Betley, T. A. Chem. Commun. 2011, 47, 6344-6346.

(47) Fout, A. R.; Zhao, Q.; Xiao, D. J.; Betley, T. A. J Am. Chem. Soc. 2011, 133, 16750-16753.

(48) Eames, E. V.; Harris, T. D.; Betley, T. A. Chem. Sci. 2012, 3, 407-415.

(49) Sazama, G. T.; Betley, T. A. Inorg. Chem. 2010, 49, 2512-2524.

(50) Sazama, G. T.; Betley, T. A. Organometallics2011, 30, 4315-4319.

143 Acknowledgements

I would like to thank my family for whom I am sincerely blessed. My parents brought me into this world, and have given me more love, guidance, and support than I can possibly imagine. Having now been through graduate school myself, I can't imagine the difficulty of starting a family under such circumstances. I want to thank my mom for exhibiting such patience with me, even when I was a lot to handle. Their achievements and sacrifices constantly inspire and motivate me to be a better person, and I hope that I can someday do the same for my children. I must also thank my brother Andrew for his wisdom and support throughout my life. I'm proud to call him my best friend. I want to thank my grandparents Daniel and Mary Pesek, whose love and support I appreciate and miss dearly. Although I never had the opportunity to know my grandfather Alfred Marshak, it is fascinating to read his publications, which I greatly admire. I thank my grandmother Celia Marshak, who I think is the only person on the planet that has life just about figured out. Finally, I thank Casandra for staying with me through some of the most difficult times of my life. Every moment I spend with her is amazing. The Nocera lab has been a wonderful and scientifically stimulating environment in which to have worked. I must first thank my friend Dino Villagrin, who I sat next to for almost 5 years. His patient advice got me through many difficult times. I also thank Ted Betley, whose research project inspired me to join the Nocera lab. Though I regret that I couldn't do more with it, I'm glad that it could be adapted to generate interesting polymetallic chemistry. I thank Stas Groysman for encouragement with the siloxide project. I also thank to Matt Chambers, who I envy for his ability to always reach a diplomatic solution to all problems always. I wish him the best next year in France with Nodmie, etj'espere qu'il put apprendre un peu defrangais la-bas. Thanks to Sebastian Stoian for his encouraging conversations about DFT and for fitting my EPR spectrum. Thanks to Danna Freedman for her help with getting magnetic data, and I wish her and Dave the best at their new jobs. Thanks to Yogi Surendranath and Changhoon Lee for help with electrochemistry, and for giving me a whole new way to think about chemistry. Thanks to Arthur, Joel, and Streece for getting me settled in the lab, and thanks to Emily Nytko-Lutz for not letting me get too comfortable. Thanks to Matt Kanan for teaching me how to talk to women, and to Alex Radosevich for teaching me how to dance with them. Thanks to Bob McGuire and Jon Melnick for convincing me that I should follow baseball again. Thanks to Arturo for going with me to a bunch of great concerts. Thanks to Tim Cook for all the 'code 4' and BUDK shenanigans, and to Tom Teets for all the rousing softball pep talks. Thanks to Jenny Yang and Liz Young for their continued friendship, in spite of the garden gnome. Thanks to Emily McLaurin always keeping me focused on the big picture. Thanks to Dilek Kiper for managing so much of the group functions, and for her constantly amusing pronunciation

144 of 3rd. Finally, thanks to Allison Kelsey for organizing almost everything and for her awesome attitude all the time. To the current crop of students, I hope the move to Harvard goes smoothly and you all continue to shine on. Outside of the Nocera group, I have probably spent the most time with members of the Cummins group. I would like to thank Glen Alliger for teaching me the meaning of perseverance, Alex Fox for his bright and sunny disposition, Chris Clough for making the best smoked ribs, Nazario Lopez for his epic karaoke ballads, and Matt Rankin for following me on a fishing trip to northern Minnesota in January. Thanks to Cesar Manna for help with one last 51V NMR. I also want to thank Daniel Tofan, Nazario, and Christian Hering for a successful cod fishing trip, and Alexandra Velian for always being enthusiastic about science. Finally, I would like to thank Jared Silvia for his pragmatism and Brandi Cossairt for her enthusiasm. You two are a perfect team. I would not be in the position that I am without all of the extraordinary teachers that I have known throughout my life. I am fortunate to have been given so much attention, encouragement, and motivation from such a talented and accomplished group of people. I thank Barbara McClintock for having the care and patience to teach basic plant biology to a 7-year-old, Martin Malloy for his encouragement and participation in my early interest in astrophotography, and Ruth Williams for imparting me with enthusiasm for mathematics, as well as for having been such a wonderful high school advisor and friend. Unconventional teaching methods can sometimes be frustrating to any student, but in light of much of the content of this thesis, I am grateful to Tom Shields for making me repeat, "a volt is a joule per coulomb," so many times that it became permanently implanted in my brain. Although I have never taken a formal engineering course, nothing has better prepared me for laboratory chemistry than the experience I had with a. j. howard navigating the complexities in the design and construction of a left- handed 'Marshak-o-caster.' I also thank Peter Julius for leading a powerful history class on the afternoon of September 11, which had 100% attendance, despite the school being closed. In college, I must thank Kurt Hirsekorn for steering my interest in chemistry to the Wolczanski lab, and teaching me the intricacies of inorganic and organometallic synthesis. Thanks to Andrew Chadeayne for imparting some of his enthusiasm for life, as well as his synthesis of the DTBMS ligand featured in Chapter 3. Thanks to Devon Rosenfeld and Orson Sydora for using Slayer to make late nights in the lab more interesting. Thanks to Elliott Hulley for always brining his phone to group meeting, and I wish him best of luck with Abi and little Calvin Michael. Congratulations to David and Brenda Kuiper for their synthesis of a "future world champion weightlifter, judo master, and Nobel Prize winning Chemist." Thanks to Emily Volpe for always taking the moral high ground. Thanks to Dave Manke for his constant support and praise, and for always giving me the right advice. Thanks to Tom Ruttledge for his

145 enthusiastic attitude, and for having my back through all of those tough moments. The only way I think I could repay him would be with a ticket to outer space. I thank Peter Wolczanski for doing everything in his power to make sure I will not have to say, "Do you want fries with that?" In all seriousness though, I am very fortunate to have stumbled into his lab, since it takes real character to be an accomplished researcher, yet still prioritize teaching and mentoring students. The topics and papers discussed in his course, Chem 605, much of which was carried out by the faculty at MIT, certainly influenced my decision to attend graduate school there. In that regard, I would like to thank my committee members, Christopher Cummins and Richard Schrock. One of the most rewarding experiences in graduate school was to read through their publications chronologically and see the process by which their respective research has evolved. Finally, I thank my research advisor Daniel Nocera. Working in his lab has exposed me to countless facets of chemistry, physics and biology, beyond those that I am directly involved. I am fortunate to have such an accomplished role model on which to base how I think and respond to scientific problems. It been an incredible experience that I feel has prepared me for the future better than I can yet realize. I am continuously inspired by his ability to impart his own enthusiasm for science on the many varied audiences who hear him speak. As my advisor, Dan has stood by me though many difficult circumstances, and in spite of his busy work schedule, he has always given me the time and attention I've needed. As a friend, he is among my closest confidants, and he is one of the most genuinely selfless people I know. I wish Dan continued success at Harvard, and wherever his life and travels take him.

146