Open NLD Thesis FINAL.Pdf

The Pennsylvania State University

The Graduate School

Eberly College of Science

DEVELOPMENT OF PHOSPHORUS-CATALYZED REDUCTIONS VIA P(III)/P(V)

CYCLING WITH STRAINED PHOSPHORUS CATALYSTS

A Dissertation in

Chemistry

Nicole L. Dunn

 2016 Nicole L. Dunn

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

May 2016

The dissertation of Nicole L. Dunn was reviewed and approved* by the following:

Alexander T. Radosevich Assistant Professor of Chemistry Dissertation Advisor Chair of Committee

Kenneth S. Feldman Professor of Chemistry Chemistry Graduate Program Chair

Scott T. Phillips Associate Professor of Chemistry Martarano Career Development Professorship

Joshua D. Lambert Associate Professor of Food Science

*Signatures are on file in the Graduate School

iii

ABSTRACT

Phosphorus catalysis is a growing field that is currently based upon the utility of three major catalytic cycles: 1) three coordinate phosphine to four coordinate phosphonium, 2) three coordinate phosphine to four coordinate phosphine oxide, and 3) four coordinate phosphonium to five coordinate phosphorane. In this thesis, a new mode of phosphorus catalysis will be described, wherein the catalytic cycle relies upon the transition from three-coordinate phosphorus to five- coordinate phosphorane. Specifically, we have developed two systems where geometrically constrained phosphorus(III) compounds catalyze the reduction of unsaturated substrates. In the first system, a planar, T-shaped phosphorus(III) compound is transformed into a five-coordinate hydridophosphorane with no observable intermediates; catalytic reduction of azobenzene is observed with both species in the presence of a reductant. In the second example, a cyclic phosphine catalyzes the reductive transposition of allylic halides. We believe that the demonstrated ability of phosphorus compounds to cycle between three-coordinate and five-coordinate complexes marks the development of new class of phosphorus-catalyzed reactions that could have broad applications in synthetic chemistry.

TABLE OF CONTENTS

List of Figures ...... viii

List of Tables ...... xv

Acknowledgements ...... xvi

Chapter 1 Phosphorus Reactivity and Related Catalytic Applications ...... 1

Abstract ...... 1 1.1. P(III) Phosphines ...... 1 1.1.1 Nucleophilic Reactions ...... 2 1.1.2 Oxophilic Reactions ...... 4 1.1.3 Oxidative Addition ...... 6 1.2. Tetracoordinate P(V) Phosphonium Salts ...... 7 1.2.1 Electrophilic Phosphonium Catalysis ...... 8 1.3. Pentacoordinate P(V) Species ...... 9 1.3.1 Reductive Elimination from P(V) ...... 9 1.4. Reversible Oxidative Addition and Reductive Elimation ...... 10 1.4.1 Intramolecular Examples ...... 11 1.4.2. Intermolecular Examples ...... 11 1.5. Conclusions ...... 12 1.6 References ...... 13

Chapter 2 Oxidative Addition and Reductive Elimination at a T-shaped, Planar Phosphorus Compound Provides Phosphorus Transfer Hydrogenation Catalysts ...... 17

Abstract ...... 17 2.1 Introduction ...... 17 2.1.1 Structural and Electronic Requirements for Two Electron Redox at Phosphorus ...... 18 2.1.2 Characteristics of a T-shaped P(III) Compound ...... 19 2.2 Synthesis of Dihydridophosphorane 2-3 ...... 22 2.2.1 Synthesis of Dihydridophosphorane 2-3 from Dichlorophosphorane 2-10..... 22 2.2.2 Synthesis of Dihydridophosphorane 2-3 from Three-Coordinate Species 2- 1 ...... 23 2.3 Characterization of 2-3 ...... 26 2.3.1 NMR Spectroscopy ...... 26 2.3.2 X-ray Crystallography ...... 27 2.3.3 Computations of P-H Bonds ...... 28 2.3.4 Structural Equilibration ...... 29 2.3.5 Synthesis and Characterization of 2-3 Isotopologues ...... 31 2.4 Reactivity of 2-3 ...... 33

2.4.1 Hydrogen Release from Dihydridophosphorane 2-3 to Reform Three Coordinate Compound 2-1 ...... 33 2.4.2 Catalytic Reduction of Azobenzene with Phosphorus Catalysts ...... 34 2.5 Conclusions ...... 36 2.6 Experimental ...... 37 2.6.1 General Materials and Methods ...... 37 2.6.2 Synthetic Procedures and Data ...... 37 2.6.3 Kinetics Data for Arrhenius Analysis of Hydrogen Transfer from 2-3 to Azobenzene...... 45 2.6.4 Computational Details...... 47 2.6.5 Crystallographic Details...... 49 2.6.6 NMR Spectra ...... 50 2.7 References ...... 54

Chapter 3 Mechanistic Studies of Hydrogen Transfer to a P(III) Center from Ammonia Borane ...... 58

Abstract ...... 58 3.1 Introduction ...... 58 3.1.1 Importance of Hydrogenation ...... 58 3.1.2 Hydrogen Addition to a Metal Center ...... 59 3.1.3 Hydrogen Addition at a Non-metal Center ...... 60 3.1.4 Hydrogen activation at Phosphorus ...... 61 3.2 Stoichiometric conversion of 2-1 to 2-3: Hydrogen Transfer from Ammonia Borane ...... 62 3.2.1 Kinetics of Hydrogen Transfer from Ammonia-Borane to 2-1...... 62 3.2.2 Eyring Analysis ...... 64 3.2.3 Kinetic Isotope Effects...... 65 3.2.4 Isotopic Labeling at 2-3 ...... 67 3.2.5 11B NMR Monitoring...... 68 3.3 Reactions of 2-1 with other Hydrogen Sources ...... 70 3.3.1 Reaction of 2-1 with Substituted Amine Boranes ...... 70 3.3.2 Reaction of 2-1 with Nonpolar Hydrogen Sources ...... 71 3.4 Mechanistic Discussion ...... 71 3.5 Conclusions ...... 78 3.6 Experimental ...... 78 3.6.1 General Materials and Methods...... 78 3.6.2 Synthetic Procedures and Data ...... 79 3.6.3 NMR Monitoring...... 80 3.6.4 Kinetics ...... 84 3.7 References ...... 91

Chapter 4 Investigations of Dihydridophosphorane 2-3 as a Hydrogen Source ...... 94

Abstract ...... 94 4.1 Introduction ...... 94 4.1.1 Acidity of P-H Compounds ...... 94 4.1.2 Radical Abstraction from P-H bonds ...... 96 4.1.3 Hydricity of P-H bonds ...... 96

4.2 Hydrogen Transfer to Organic Substrates from Dihydridophosphorane 2-3 ...... 98 4.2.1 Imines ...... 98 4.2.2 Aldehydes and Ketones ...... 99 4.2.3 Alkenes ...... 100 4.2.4 Reduction of Triple Bonds ...... 102 4.2.5 Reduction of Single Bonds ...... 103 4.2.6 Hydrogen Transfer to Frustrated Lewis Pairs and Carbenes ...... 103 4.3 Monoatomic Hydrogen Transfer from Dihydridophosphorane 2-3 ...... 104 4.3.1 Hydride Transfer from Dihydridophosphorane 2-3 to Cations...... 104 4.3.2 Proton Transfer from Dihydridophosphorane 2-3...... 106 4.3.3 Hydrogen Atom Transfer from 2-3...... 107 4.3.4 Electron Transfer from 2-3...... 108 4.4 Mechanism of Hydrogen Transfer from Dihydridophosphorane 2-3 to Azobenzene ...... 109 4.4.1 Reaction Kinetics ...... 110 4.4.2 Isotopic effects on rate and labeling...... 112 4.4.3 Hammett correlation of substituent effects...... 113 4.4.4 Mechanistic Discussion ...... 114 4.5 Conclusions ...... 118 4.6 Experimental Section ...... 118 4.6.1 Synthesis of Substrates ...... 118 4.6.2 Reduction Products ...... 125 4.6.3 Electrochemistry ...... 129 4.6.4 EPR Spectroscopy ...... 131 4.6.5 NMR Monitoring of Isotopic Labelling ...... 132 4.6.6 Kinetics ...... 133 4.6.7 Spectral Data ...... 139 4.7 References ...... 148

Chapter 5 Regioselective Reductive Transposition of Allylic Bromides via P(III)/P(V) Redox Cycling...... 153

Abstract ...... 153 5.1 Introduction ...... 153 5.2 Optimization of Catalytic Reaction ...... 157 5.2.1 Substrate Optimization ...... 157 5.2.2 Reductant Optimization ...... 159 5.2.3 Catalyst Optimization ...... 160 5.3 Mechanistic Studies ...... 167 5.3.1 NMR Studies ...... 168 5.3.2 Synthesis and Characterization of Isolable Hydridophosphorane ...... 169 5.3.3 Deuterium Labeling Studies ...... 172 5.3.4 Pulse-Chase Studies ...... 173 5.3.5 Conversion of Hydridophosphorane to Products ...... 174 5.4 Conclusions ...... 175 5.5 Experimental ...... 176 5.5.1 General Methods and Materials ...... 176 5.5.2 Synthetic Procedures ...... 177 5.5.3 Optimization and Procedure for Catalytic Allylic Reduction ...... 188

vii

5.5.4 VT-NMR Experiments ...... 200 5.5.5 Computational Data ...... 204 5.5.6 Crystallographic Data ...... 207 5.5.7 Spectral Data ...... 209 5.6 References ...... 231 Appendix Permissions ...... 234

viii

LIST OF FIGURES

Figure 1-1. General reactivity of phosphines with electrophiles leads to the formation of phosphonium salts ...... 2

Figure 1-2. Rauhut-Currier reaction of 1-1 catalyzed by tributylphosphine...... 3

Figure 1-3: Phosphine catalyzed addition of nucleophiles to ɑ,β-unsaturated carbonyl compounds. A) The role of the phosphorus catalyst in activating the nucleophile, and B) reaction of the nucleophile with substrate...... 3

Figure 1-4. Generalized Wittig olefination reaction...... 4

Figure 1-5. Catalytic modification of the Wittig reaction reported by O’Brien and coworkers...... 5

Figure 1-6. General addition of an oxidant to a phosphine...... 6

Figure 1-7. A) The formation of a phosphorane from phosphite 1-16. B) The formation of a phosphonium salt from 1-19...... 7

Figure 1-8. Nucleophilic addition to a phosphonium salt to form a five-coordinate phosphorane...... 7

Figure 1-9. Catalytic cycle of hydrodefluorination of fluoroalkanes with fluorophosphonium catalyst 1-21...... 8

Figure 1-10. Ligand coupling at a pentacoordinate phosphorus complex...... 9

Figure 1-11. Ligand coupling at phosphorane 1-30 to form 2,2’-bipyridine...... 10

Figure 1-12. Ligand coupling at phosphorane 1-33 to form phosphine 1-34...... 10

Figure 1-13. Equilibrium between a σ5 hydridophosphorane 1-35 and a σ3 phosphite 1-36.... 11

Figure 1-14. A solution of cyclic disulfide 1-38 and phosphonite 1-37 is in equilibrium with phosphorane 1-39...... 12

Figure 1-15. The addition of fluorine to phosphine in a cryogenic matrix proceeds through hydridophosphorane 1-40 and produces two fluorophosphines 1-41 and 1-42...... 12

Figure 2-1. Structural distortion from pyramidal to T-shaped geometries may ease the transition to five coordinate trigonal bipyramidal structures...... 18

Figure 2-2. A) Compound 2-1. B) Depiction of perpendicular p orbital LUMO and lone pair HOMO at phosphorus. C) Computations of HOMO-2 at 2-1 and D) LUMO at 2- 1...... 19

Figure 2-3. Synthesis of 2-1...... 20

Figure 2-4. Oxidative addition of 1 with A) 1,2-dicarbonyl chloranil, B) methanol, and C) propylamine...... 21

Figure 2-5. Report of hydrogen release from a dihydridophosphorane...... 21

Figure 2-6. Synthesis of dihydridophosphorane 2-3 from dichlorophosphorane 2-10...... 22

Figure 2-7. Proposed mechanism of the formation of 2-3 and 2-1 from 2-10...... 23

Figure 2-8. Attempted synthesis of 2-3 from 2-1 and H2...... 24

Figure 2-9. Attempted synthesis of 2-3 from 2-1 and H2...... 25

Figure 2-12. Possible equilibrium between five coordinate 2-3 and three coordinate 2-12. ... 29

Figure 2-13. 2-3 reacts with borane-dimethylsulfide and maintains a five coordinate structure...... 30

Figure 2-14. Synthetic routes 2-3 and 2-3-d2 by dehydrogenation of ammonia borane and isotope exchange...... 32

31 Figure 2-15. P NMR of 2-3-d2...... 33

Figure 2-16. Attempted release of hydrogen gas from 2-3...... 33

Figure 2-17. Reduction of azobenzene (2-4) to diphenylhydrazine (2-14) with 2-3 as the reductant...... 34

Figure 2-18. Stoichiometric transitions between 2-1 and 2-3...... 35

Figure 2-20. Synthesis of 2-1 and 2-10. (a) Br2, Et2O, 93%; (b) BnNH2, PhMe, 80 °C, 61%; (c) H2, Pd/C, MeOH, 88%; (d) K3PO4, H2O, CH2Cl2, 88%; (e) PCl3, Et3N, C5H12, 71%; (f) PCl5, Et3N, C5H12, 77% ...... 38

Figure 2-20. Plots of ln(k/T) vs. (1/T) from the data in the table at right...... 47

Figure 2-21. Summary of computational results. Relative electronic energies for stationary points are noted in bold (zero-point corrected energies in parentheses)...... 48

Figure 3-1. Catalyzed hydrogenation of alkenes...... 59

Figure 3-2. A) Orbital overlap between a vacant d orbital at a transition metal and an H2 σ bond. B) Orbital overlap between a filled d orbital at a transition metal and an H2 σ* orbital. C) Hydrogen activation at Vaska’s complex...... 60

Figure 3-3. A) Orbital overlap between a singlet carbine and H2. B) Hydrogen activation at an alkylaminocarbene...... 61

Figure 3-4. Hydrogen transfer from AB to 2-1...... 62

Figure 3-5. Plot of ln[2-1] vs. time, showing that the consumption of 2-1 is first order in 2- 1. R2 = 0.994...... 63

Figure 3-6. Plot of kobs of the reaction with various concentrations of AB, which shows a linear increase in rate with increasing [AB]. R2 = 0.84. Error bars show the standard deviation for triplicate runs, other data points were single runs...... 64

Figure 3-7. Eyring plot of hydrogen transfer from AB to 2-1 between 40-60 °C...... 65

Figure 3-8. Kinetic isotope effects are observed for isotopic substitution for both N-H and B-H during hydrogen transfer from AB to 2-1...... 66

Figure 3-9. Isotopic scrambling between N-H and B-H at AB in the presence of 2-1 or 2-3 is not observed by 2H NMR...... 66

Figure 3-10. A) The route proposed by Sakaki for hydrogen transfer to 2-1. B) σ3-σ5 isomerization to form 2-7 is not observed by 2H NMR...... 67

Figure 3-11. A) 11B NMR of AB heated at 60 °C for 24 h. B) 11B NMR of AB and one equiv 2-1 heated at 60 °C for 24 h. C) Boron species observed...... 69

Figure 3-12. Polar hydrogen sources for transfer hydrogenation...... 71

Figure 3-13. Nonpolar hydrogen sources for transfer hydrogenation...... 71

Figure 3-14. Dissociation of AB...... 72

Figure 3-15. Isotopic scrambling between N-H and B-H at AB in the presence of 2-1 or 2- 3 is not observed by 2H NMR...... 73

Figure 3-16. Reaction of 2-1 with ammonia, followed by reaction of 3-24 with AB...... 74

Figure 3-17. N-H activation of AB with 2-1, followed by reaction of 3-25 with AB...... 75

Figure 3-18. A) Formation of 2-3 from 2-1 and AB is promoted by catalytic base. B) Proposed reaction sequence...... 76

Figure 3-19. Initial proton transfer to 2-1 forms 3-27, which is not transformed into 2-3...... 77

Figure 3-20. Proposed mechanism of hydrogen transfer from AB to 2-1. An equilibrium proton transfer forms complex 3-23, followed by hydride transfer to form 2-3...... 78

Figure 3-21: 11B NMR of 2-1, 20 equiv. AB, 200 equiv. cyclohexene, heated to 60 C for 4 h. Peak at -23 is AB, peak at 47 is Cy2BNH2, peak at -5 may be CyBHNH2...... 80

Figure 3-22: 31P NMR of 1, 20 equiv. AB, 200 equiv. cyclohexene, heated to 60 C for 4 h. Triplet at -45 is 2-3, no 2-1 remains...... 81

2 Figure 3-23: H NMR of 2-1 with ND3BH3 in THF at t=0...... 82

2 Figure 3-24: H NMR of 2-1 with ND3BH3 in THF after 4 h, 60 °C shows only P-D signals from 1•[D]2...... 82

2 Figure 3-25: H NMR of 2-3 with ND3BH3 in THF at t=0...... 83

2 Figure 3-26: H NMR of 2-3 with ND3BH3 in THF after 4 h, 60 °C, which shows only P- D peaks from 1•[D]2...... 83

Figure 3-27: A comparison of the various treatments of the concentration of 2-1 vs time shows that the best fit is ln[2-1], and the reaction is first order in 2-1...... 86

Figure 3-28: kobs vs [AB] shows that the reaction is first order in AB...... 87

Figure 4-1. Examples of the base-promoted nucleophilicity of P-H bonds. A) reaction of phosphine 4-1 with alkyl halide 4-3. B) Reaction of hydridophosphorane 4-5 with imine 4-6...... 95

Figure 4-2. Hydrophosphination of A) limonene 4-8 with phosphine, and B) dibutyldisulfide with hydridophosphorane 4-10...... 96

Figure 4-3. A) Reduction of benzophenone with P-hydrido-1,3,2-diazaphospholane. B) Release of hydrogen from hydridophosphorane 4-15. C) Reduction of aryl aldehydes by 4-17...... 97

Figure 4-4. Imine reduction with 2-3...... 99

Figure 4-5. Proposed reactivity of aldehydes with 2-3...... 100

Figure 4-6. Scope of alkene reduction with 2-3...... 101

Figure 4-7. The reduction of maleic acid with catalyst 4-29 under an atmosphere of deuterium gas...... 101

Figure 4-8. Reduction of dimethylmaleate shows that cis-addition predominates...... 102

Figure 4-9. Reduction of dimethylmaleate shows that cis-addition predominates...... 104

Figure 4-10. Hydride abstraction from 2-3 with trityl cation 4-36...... 105

Figure 4-11. Unsuccessful attempts to isolate putative phosphonium intermediate 3-23 through A) hydrogen abstraction with triphenylcarbenium tetraphenylborate, and B)

xii

hydrogen abstraction from stabilized derivative 4-39 with triphenylcarbenium tetrafluoroborate...... 106

Figure 4-12. Deprotonation of 2-3 by DBU is followed by hydrogen release, thus reforming 2-1...... 107

Figure 4-13. Hydrogen atom transfer from 2-3 to 2 equiv of galvinoxyl radical 4-43...... 107

Figure 4-14. CV of 2-3 shows two oxidation potentials, as well as cobaltcenium hexafluorophosphate couple at ca. –1.4 V...... 108

Figure 4-15. Synthesis of radical cation 4-45 from the reaction of 2-3 with silver triflate...... 109

Figure 4-16. EPR signal from 4-45, generated by the reaction of 2-3 with AgOTf in C6D6 at room temperature. The signal is overlayed with a simulated generated by Alexey Silakov...... 109

Figure 4-17. Hydrogen transfer from hydridophosphorane 2-3 to azobenzene 2-4...... 110

Figure 4-18. Consumption of 2-3 vs. time under pseudo-first order conditions shows the reaction is first order in 2-3, R2 = 0.999...... 111

Figure 4-19. Linear dependence of observed rate on concentration of azobenzene (2-4). R2 = 0.997...... 111

Figure 4-20. Synthesis of radical cation 4-45 from the reaction of 2-3 with silver triflate...... 113

Figure 4-21. Reduction of substituted azobenzenes allows for the construction of a Hammett plot...... 114

Figure 4-22. Hammett plot for hydrogen transfer from 2-3 to substituted azobenzenes. ρ = 1.2, R2 = 0.90. All replicates shown as separate points...... 114

Figure 4-23. Potential reaction mechanisms for azobenzene reduction, including A) initial proton transfer, B) hydrogen atom transfer, C) electron transfer, and D) hydride transfer...... 115

Figure 4-24. A) Conversion of azobenzene is unaffected by AIBN addition. B) Reduction of 2-allylazobenzene 4-52 proceeds cleanly without cyclization to 4-53...... 116

Figure 4-25. A) One electron reduction of cyclopropyl imine 4-54. B) Reduction of a cyclopropanated imine (4-54) with 2-3 proceeds cleanly to amine 4-58 without ring fission...... 117

Figure 4-26: Plot of 2-3 with colbatacenium...... 130

Figure 4-27: Plot of azobenzene with ferrocene...... 131

2 Figure 4-28: H NMR of azobenzene with 2-3-d2 in benzene, t=0...... 132

xiii

2 Figure 4-29: H NMR of azobenzene with 2-3-d2 in benzene, after 4 h, 60 °C...... 133

Figure 5-1. Allylic reduction of 5-1 to provide 1,4-diene 5-2...... 154

Figure 5-2. α-functionalization and allylic reduction to provide trans-olefins...... 154

Figure 5-3. Stereoselective allylic reduction...... 154

Figure 5-4. α-functionalization and allylic reduction to provide trans-olefins...... 156

Figure 5-5. Proposed catalytic cycle for phosphine-catalyzed allylic reductions...... 157

Figure 5-6. Conversion of cinnamyl substrates to triphenylphosphonium salts is successful when X = Cl, Br. X = I, OTS are unstable, while X = OAc, OCOCF3 are unreactive.... 158

Figure 5-7. Selectivity of aluminum hydride reductants. Yield and product composition were determined by GC analysis against dodecane...... 160

Figure 5-8. Use of an ortho-amide to increase rate of iminophosphorane hydrolysis...... 163

Figure 5-9. A) Reduction of 5-8 produces 5-12b. B) Reduction of 5-32 produces 5-11. C) Competition experiment between 5-8 and 5-32 proves that 5-32 is reduced preferentially in an environment with limiting reductant. Product mixture ratio was determined by GC analysis against dodecane as an internal standard. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society...... 166

Figure 5-11. Potential mechanisms for the reduction of allylic phosphonium salts, showing SN2’ reduction (Path A) or intramolecular rearrangement of hydridophosphorane intermediate 5-37 (Path B)...... 167

Figure 5-12. Reaction scheme and NMR data showing A) phosphonium salt 5-40 prior to addition of LAH, B) intermediate hydridophosphorane 5-37, observed at –70 °C in 31P NMR, C) phosphine 5-35, observed upon warming to 0 °C. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society...... 169

Figure 5-13. A) Synthesis and characterization of stable hydridophosphorane 5-42 (major diastereomers of salt and phosphorane shown). 31P NMR of 5-42 B) decoupled and C) coupled to hydrogen...... 171

Figure 5-14. Intermolecular hydride transfer observed from hydridophosphorane 5-42...... 172

Figure 5-15. Deuterium studies show A) delivery of hydride selectively to the γ-position, and B) no scrambling of α-protons under the reaction conditions. Adapted with

xiv

Figure 5-16. Pulse-chase studies show exchange between aluminum and phosphorous hydrides...... 174

Figure 5-19. Graphical representation and computed relative energies for the polytopal rearrangement of 5-43. All geometries optimized at M06-2X/6-311++G(2d,2p). Energies (enthalpies) reported in kcal/mol. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society...... 205

Figure 5-20. Computed relative energies for concerted group transfer rearrangements. All geometries optimized at M06-2X/6-311++G(2d,2p). Energies (enthalpies) reported in kcal/mol relative to 5-43. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society...... 206

Figure 5-21. Thermal ellipsoid plot (50%) of 1-allyl-2,2,3-trimethyl-1- phenylphosphetanium bromide. Hydrogen atoms and counter-ion are omitted for clarity. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society...... 207

LIST OF TABLES

Table 2-1. Selected bond lengths for 2-1 and 2-3 in Å ...... 28

Table 2-2. Selected bond angles for 2-1 and 2-3 reported in degrees...... 28

Table 2-3. Reduction of 2-4 in the presence of AB and phosphorus promoters. aDetermined using 1H NMR. Adapted with permission from J. Am. Chem. Soc., 2012, 134, 11330. Copyright 2012 American Chemical Society...... 35

Table 4-1. Unique IR signals from the cis and trans products. Italicized values show the overlap between the experimental results and the cis product...... 102

Table 4-2. Eyring analysis for the kinetic runs in benzene and acetonitrile...... 112

Table 5-1. Products for the reduction of cinnamyl compounds with LAH...... 155

Table 5-2. Substrate optimization table with LiAlH(OtBu)3 and catalyst 5-31. Yield and product composition were determined by GC analysis against dodecane as an internal standard...... 158

Table 5-3. Reduction of cinnamyl bromide with various phosphine catalysts. Yield and product compositions were determined by GC analysis against dodecane as an internal standard...... 161

Table 5-4. Reduction of cinnamyl bromide with various phosphine catalysts. Yield and product composition were determined by GC analysis against dodecane as an internal standard...... 164

Table 5-6. Effect of hydride source on regioselectivity...... 194

Table 5-7. Effect of leaving group on regioselectivity...... 196

Table 5-8. Effect of equivalent of LAH on the reaction of 5-31...... 202

Table 5-9. Data obtained from the Eyring analysis of the reduction of 5-31...... 203

Table 5-10. Sample and crystal data for 1-allyl-2,2,3-trimethyl-1-phenylphosphetanium bromide...... 208

Table 5-11. Atomic coordinates and equivalent isotropic atomic displacement parameters (Å2) for 1-allyl-2,2,3-trimethyl-1-phenylphosphetanium bromide...... 209

xvi

ACKNOWLEDGEMENTS

I would like to thank Dr. Alexander Radosevich for challenging me to push boundaries: in science and in my own life. I appreciate Dr. Kenneth Feldman, Dr. Scott Philips and Dr. Joshua

Lambert for taking the time to serve on my committee, and also send my regards to Dr. Mike

Hickner and Dr. Ryan Elias, who have attended my committee meetings in the past.

To all the Radosevich group members, past and present, thank you for sharing lab space, glassware, and dark humor. I can’t imagine getting through the work day without you. Everyone from the original cast deserves a special shout-out: we learned a lot about chemistry together in those first few years.

To my mom and dad and the rest of my family, thank you for supporting me through all my challenges. I especially appreciate Julie and Meredith, for reminding me of the life outside of school, and all the volunteers at the State College YMCA rock wall, for giving me a positive outlet while I was writing my thesis.

Finally, I would like to thank Anthony for sticking with me through all the ups and downs of graduate school. There is no one else I’d rather be lost in the woods with.

Chapter 1

Phosphorus Reactivity and Related Catalytic Applications

Abstract

My thesis will describe two novel examples of phosphorus catalysis, both of which involve phosphorus cycling through three and five coordinate states. In this chapter, the basic reactivity of phosphorus compounds will be described, as will the catalytic applications of this reactivity. A wide range of stable phosphorus compounds have been synthesized, most typically in either +3 or

+5 oxidation states, and with 2-6 bonds around the central phosphorus atom. This structural diversity provides for a range of reactivities, including the nucleophilic and oxophilic properties of three coordinate P(III) compounds, and the electrophilic behavior of four coordinate phosphonium cations, all of which have been applied in catalytic systems. Transitions from three-coordinate phosphorus compounds to five-coordinate species, as well as the reverse, have also been reported.

The application of this reactivity in catalytic cycles will be presented in Chapter 2 and Chapter 5.

1.1. P(III) Phosphines

Three coordinate phosphorus (III) compounds contain three σ bonds and a lone pair of electrons; they typically exhibit a pyramidal geometry. Three-coordinate phosphorus complexes can form additional bonds to access four coordinate states; phosphorus is also able to accommodate the formation of five and six coordinate hypervalent compounds. The lone pair on pyramidal phosphines provides these compounds with characteristic nucleophilic and oxophilic reactivities.

2 1.1.1 Nucleophilic Reactions

The most fundamental example of the nucleophilic behavior of phosphines is their propensity to participate in nucleophilic substitution reactions with alkyl halides.1 This reaction will expel a leaving group on the electrophile and form a four coordinate, positively charged phosphonium salt (Figure 1-1). Nucleophilic phosphorus compounds will also add to electron-poor double bonds in a similar fashion. Phosphorus nucleophilicity has been applied to a wide range of reactions.2

Figure 1-1. General reactivity of phosphines with electrophiles leads to the formation of phosphonium salts

The nucleophilic behavior of phosphines has been harnessed for catalysis3 in a classic example reported by Rauhut and Currier.4 They observed that electron deficient alkenes, such as

1-1, dimerized in the presence of tributylphosphine (Figure 1-2). The mechanism of this reaction has been elucidated, and is initiated by the addition of the phosphine to the alkene to form phosphonium enolate 1-2.5 Species 1-2 will then undergo C-C bond formation with a second equivalent of alkene 1-1 to form compound 1-3. Proton transfer at species 1-3 is followed by elimination of the phosphine catalyst and the formation of the coupling product, 1-5.

Figure 1-2. Rauhut-Currier reaction of 1-1 catalyzed by tributylphosphine.

The Rauhut-Currier reaction has also been applied to cross reactions with multiple alkene reagents.6 Similar phosphine reactivity is operative as well in the Morita-Baylis-Hillman reaction, which further expands the scope of C-C bond formation enabled by phosphorous(III) compounds to include alkene-aldehyde coupling.7 Phosphines can also promote the addition of nucleophiles to ɑ,β-unsaturated carbonyl compounds, such as 1-1.89 In such a reaction, the phosphonium enolate

1-2 will still be formed; however, in this case 1-2 will deprotonate a pro-nucelophile instead of undergoing a bond formation reaction (Figure 1-3A). The nucleophile will then add to another equivalent of 1-1, ultimately providing product 1-8 (Figure 1-3B).

4 1.1.2 Oxophilic Reactions

Phosphine oxophilicity is derived from the strength of the P=O bond, at 130-140 kcal/mol.10 Depending upon the stability of the phosphine, some species may oxidize slowly (or rapidly) in the presence of ambient atmospheric conditions. Several key reactions feature the formation of the P=O bond as a driving force, including the Staudinger synthesis of amines,11 the

Mitsuonbu reaction for functionalization of alcohols,12 the Appel reaction,13 and Wittig olefination reaction.14 In the case of the Wittig reaction, the transformation begins with the formation of a phosphonium ylide: first a phosphine is alkylated with a functional group of interest to form a phosphonium salt, and then treatment of the phosphonium salt with a base leads to ɑ-deprotonation to form an ylide.15 The phosphonium ylide then reacts with a carbonyl compound, yielding a four- membered oxaphosphetane. Cycloreversion of the oxaphosphetane results in the formation of a strong P=O bond in a phosphine oxide and a C=C alkene bond (Figure 1-4). Various phosphorus compounds can promote the reaction, and control of the reaction conditions can lead to E/Z selectivity in the alkene product, making the Wittig reaction and related modifications particularly useful and synthetically relevant.16

Figure 1-4. Generalized Wittig olefination reaction.

5 While the formation of phosphine oxide in the Wittig olefination and other similar reactions is a necessary step, the separation of a stoichiometric by-product is inconvenient and inefficient.

The drawbacks of stoichiometric phosphorus by-products have driven the development of a phosphine-catalyzed Wittig olefination reaction. Catalysis in this system requires that the phosphine oxide, once formed, is reduced in situ to the phosphine catalyst. A solution to the challenge of reducing the strong P=O bond under mild conditions has recently been reported by

O’Brien and coworkers.17 The use of a cyclic phospholane catalyst (1-9), accompanied by diphenylsilane as a reductant, allows for a convenient one-pot phosphorus-catalyzed Wittig reaction (Figure 1-5). The cyclic phosphine is key to catalytic reactivity, as cyclic phosphine oxides are reduced more readily than their acyclic counterparts, allowing for efficient catalyst turnover.18

Reduction rates of phosphine oxides increase as ring size decreases, due to the relief of ring-strain.19

Other examples of phosphine oxide catalysis20 have also been presented for Appel21 and aza-

Wittig22 reactions.

Figure 1-5. Catalytic modification of the Wittig reaction reported by O’Brien and coworkers.

6 1.1.3 Oxidative Addition

Three coordinate phosphorus species have the capability of expanding their coordination sphere to form five and six-coordinate23 hypervalent compounds. This reactivity is most often observed in the presence of strong oxidizing agents, such as halogens and peroxides (Figure 1-6).24

However it has also been observed in the presence of milder X-H bonds and 1,2-dicarbonyls.

Figure 1-6. General addition of an oxidant to a phosphine.

One example of a conversion between three-coordinate phosphorus and five coordinate species is the Kukhtin-Ramirez reaction, in which the reaction of a phosphite with a 1,2-dicarbonyl compound forms a pentaoxyphosphorane.25 This addition is believed to occur in a stepwise fashion; evidence for such a mechanism includes the fact that dipolar forms of the Kukhtin-Ramirez adducts can also be observed, depending upon the substituents of the five coordinate species (Figure

1-7).26 This reactivity can be demonstrated by the addition of phosphite 1-16 to 1,2-dicarbonyl 1-

15 to form phosphorane 1-17; the same reaction with amine-substituted 1-19 instead yields phosphonium salt 1-20. The addition of phosphines to 1,2-dicarbonyls has been applied in the synthesis of epoxides27 and cyclopropanes28 and the ɑ-functionalization of carbonyl compounds.29

Figure 1-7. A) The formation of a phosphorane from phosphite 1-16. B) The formation of a phosphonium salt from 1-19.

1.2. Tetracoordinate P(V) Phosphonium Salts

Tetracoordinate P(V) phosphonium salts typically have tetrahedral geometries. The phosphorus atom is positively charged, and therefore exhibits electrophilic behavior. The phosphonium LUMO is a σ* bond orbital,30 therefore the addition of a nucleophile to a phosphonium results in the formation of a hypervalent 3-centered-4-electron bond in a neutral five- coordinate phosphorane.

Figure 1-8. Nucleophilic addition to a phosphonium salt to form a five-coordinate phosphorane.

8 1.2.1 Electrophilic Phosphonium Catalysis

Stephan and coworkers recently reported the hydrodefluorination of fluoroalkanes catalyzed by fluorophosphonium salts in the presence of a silane reductant (Figure 1-9). 31 The fluorophosphonium catalysts were designed to increase the electrophilicity at phosphorus. This design was successful: fluoride substitution at phosphorus significantly lowers the energy of the

LUMO such that fluorophosphonium ion 1-21 is a better fluoride acceptor than tris(pentafluorophenyl)borane, a common Lewis acid catalyst. To initiate the hydrodefluorination of fluoralkanes such as 1-22, the phosphonium catalyst abstracts fluoride and carbenium ion 1-23 is formed, along with fluorophosphorane 1-24. Carbenium species 1-23 then is reduced to the neutral hydrocarbon species 1-26 via hydride addition from a silane reductant. Finally, the newly formed silicon cation abstracts fluoride from fluorophosphorane 1-24 to reform phosphonium catalyst 1-21. The electrophilic properties of phosphonium salts also have been utilized in the catalysis32 of Diels-Alder reactions,33 olefin hydrosilylation,34 and C-X bond formation.35

Figure 1-9. Catalytic cycle of hydrodefluorination of fluoroalkanes with fluorophosphonium catalyst 1-21.

9 1.3. Pentacoordinate P(V) Species

Pentacoordinate P(V) species (i.e., phosphoranes) typically exhibit trigonal bipyramidal geometries, with more electronegative substituents occupying the axial positions in a 3-center-2- electron bond. Phosphoranes undergo primarily two modes of reactivity: they may behave as Lewis acids due to their low-lying σ* bond orbital,36 or they may undergo bond forming ligand coupling reactions to form tricoordinate phosphorus species.

1.3.1 Reductive Elimination from P(V)

We have previously summarized reports wherein phosphonium salts and transient phosphoranes have the coordination number at phosphorus decreased by one, through the action of an exogenous reagent. However, phosphoranes also undergo ligand coupling such the coordination number at phosphorus decreases by two (Figure 1-10). These reactions can occur with and without the influence of additional promoters.

Figure 1-10. Ligand coupling at a pentacoordinate phosphorus complex.

In one report, Oae and coworkers have shown that carbon-carbon bond formation is observed when phosphine oxides such as 1-29 are treated with strong nucleophiles.37 For example, the reaction of 1-29 with methyllithium, followed by an aqueous workup, allows the isolation of bipyridine 1-31 in 54% yield; pyridine is also observed. This ligand coupling reaction is presumed to proceed from phosphorane intermediate 1-30. Similar results are obtained upon the hydrolysis of 2-pyridine substituted phosphonium salts.38

Figure 1-11. Ligand coupling at phosphorane 1-30 to form 2,2’-bipyridine.

Additionally, carbon-carbon bond forming reactions have been reported between ligands at phosphorus in the absence of addition reagents.39 These reactions typically require high temperatures and yield a variety of products; for example, the thermolysis of pentaphenylphosphorane yields a mixture which contains biphenyl, triphenylphosphine, and benzene, among other species. Concurrent styrene polymerization has implicated a radical mechanism in these reactions, and may explain for the complex mixtures they produce.40 However, in the case of cyclic phosphorane 1-33, brief heating yields phosphine 1-34 selectively in 86% yield.41

Figure 1-12. Ligand coupling at phosphorane 1-33 to form phosphine 1-34.

1.4. Reversible Oxidative Addition and Reductive Elimation

We have observed that there are a variety of oxidative addition transformations with phosphorus, as well as several examples of ligand coupling. There is also a class of phosphorus

(III) compounds that are in equilibrium with P(V) species. Most of these examples are

11 intramolecular additions where pendant X-H bonds interact with the phosphorus center to form spirophosphoranes, however intermolecular examples are also observed.

1.4.1 Intramolecular Examples

Spirophosphoranes are five-coordinate phosphorus species which have two bidentate ligands at phosphorus. If the spirophosphorane contains both P-H bonds and P-X bonds, then the potential for an equilibrium exists between the five-coordinate phosphorane (1-35) and a three- coordinate phosphine (1-36) with a pendant X-H group (Figure 1-13). The ability of spirophophoranes to tautomerize from σ5 phosphoranes to σ3 phosphorus species has been widely studied.42 While specific species have different equilibriums, both σ3 and σ5 species have been observed by IR,43 NMR,44 and x-ray crystallography.45

Figure 1-13. Equilibrium between a σ5 hydridophosphorane 1-35 and a σ3 phosphite 1-36.

1.4.2. Intermolecular Examples

While intermolecular reversible bond activation at phosphorus is not as commonly observed as the intramolecular version, an example has been published wherein a cyclic disulfide adds to a phosphite, as seen in Figure 1-14.46 The equilibrium of this system depends upon the

12 nature of the solvent: in acetonitrile phosphorane 1-39 is favored; however, in most other solvents phosphonite 1-37 is observed preferentially.

Figure 1-14. A solution of cyclic disulfide 1-38 and phosphonite 1-37 is in equilibrium with phosphorane 1-39.

The reaction of a three-coordinate complex to a pentacoordinate state, followed by the reversion to novel three-coordinate products has also been reported. Specifically, phosphine reacts with fluorine in a cryogenic matrix, and difluorophosphorane 1-40 is formed.47 Also observed are the three-coordinate products of fluorophosphine 1-41 and difluorophosphine 1-42, which are formed through the evolution of hydrogen fluoride and hydrogen from 1-40, respectively.

Figure 1-15. The addition of fluorine to phosphine in a cryogenic matrix proceeds through hydridophosphorane 1-40 and produces two fluorophosphines 1-41 and 1-42.

1.5. Conclusions

We have observed that nucleophilic, electrophilic, and phosphine oxide catalysis are well- developed fields of organocatalysis. However, while examples of the conversion between three and five coordinate phosphorus species are known, this reactivity has not been modified for catalytic applications. We believe that this field has the potential for growth; and is an important

13 area for development due to its analogy to transition metal catalyzed reactions. Transition metal catalysis often depends on the ability of the metal to undergo facile changes in oxidation state; in particular two-electron redox cycling is the key to fundamental processes such as oxidative addition and reductive elimination. In these processes both the electron count and coordination number of the central atom are altered by two; oxidative addition and reductive elimination are vital processes for catalysic hydrogenations, cross-couplings and C-H bond activations. The ability of phosphorus to undergo similar fundamental two-electron, two-ligand reactions raises the possibility that phosphorus could replace transition metals in some catalytic applications. This would be beneficial for the reduction of the economic and environmental costs of transition metal catalysis.

In fact in the following chapters we will report on two systems of phosphorus catalysis which cycle between three and five coordinate complexes for substrate reduction. One system cycles directly between a three-coordinate phosphorus compound and a P(V) dihydridophosphorane reductant that transfers an equivalent of hydrogen to various substrates

(Chapter 2-4). The second system operates by transitioning through three coordinate, four coordinate and five coordinate phosphorus complexes (Chapter 5). We believe these developments will lead to new applications and strategies for phosphorus-catalyzed reactions.

1.6 References

1 White, D. A.; Baizer, M. M. Tetrahedron Lett. 1973, 37, 3597

2 Valentine, D. H.; Hillhouse, J. H. Synthesis 2003, 317.

3 Methot, J.L.; Roush, W.R. Adv. Synth. Catal. 2004, 346, 1035.

4 Aroyan, C.E.; Dermenci, A.; Miller, S.J. Tetrahedron 2009 65, 4069.

5 Hill, J.S.; Isaacs, N.S. J. Phys. Org. Chem. 1990, 3, 285.

6 McClure, J.D. J. Org. Chem. 1970, 35, 3045.

7 (a) Morita, K.; Suzuki, Z.; Hirose, H. Bull. Chem. Soc. Jpn.,1968, 41, 2815. (b) Basavaiah, D.;

Rao, A.J.; Satyanarayana, T. Chem. Rev., 2003, 103, 811.

8 (a) White, D.A.; Baizer, M. M. Tet. Lett. 1973, 14, 3597. (b) Gomez-Bengoa, E.; Cuerva, J.M.;

Mateo, C.; Echavarren, A. J. Am. Chem. Soc. 1996, 118, 8553

9 Trost, B.M.; Li, C.-J. J. Am. Chem. Soc. 1994, 116, 3167.

10 Quin, L.D. A Guide to Organophosphorus Chemistry, Wiley: New York, 2000, p 95.

11 (a) Stuchwisch, C.G. Sythesis 1973, 469. (b) Gololobov, Y.G.; Zhmurova, I.N.; Kasukhin, L.F.

Tetrahedron 1981, 37, 437.

12 (a) Mitsunobu, O.; Yamada, M. Bull. Chem. Soc. Jpn. 1967, 40, 2380. (b) Hughes, D.L. Org.

React. 1992, 42, 335.

13 (a) Downie, I. M.; Holmes, J.B.; Lee, J.B. Chem. Ind. 1966, 900. (b) Appel, R. Angew. Chem.

1975, 87, 863.

14 (a) Wittig, G.; Schollkopf, U. Chem. Ber. 1954, 97, 1318. (b) Edmonds, M.; Abell, A. Modern

Carbonyl Olefination 2004, 1.

15 Vedejs, E.; Marth, C.F. J. Am. Chem. Soc. 1990, 112, 3905.

16 (a) Schlosser, M. Top. Stereochem. 1970, 5, 1. (b) Boutagy, J.; Tohmas, R. Chem. Rev. 1974, 74,

87. (c) Still, W.C.; Gennari, C. Tet. Lett. 1983, 24, 4405.

17 O’Brien, C. J.; Tellez, J. L.; Nixon, Z. S.; Kang, L. J.; Carter, A. L.; Kunkel, S. R.; Przeworski,

K. C.; Chass, G. C. Angew. Chem., Int. Ed. 2009, 48, 6836.

18 van Kalkeren, H. A.: Leenders, S. H. A. M.: Hommersom, C. R. A.: Rutjes, F. P. J. T.; van Delft,

F. L. Chem. Eur. J. 2011, 17, 11290.

19 Keglevich, G.; Fekete, M.; Chuluunbaatar, T.; Dobo, A.; Harmat, V.; Toke, L. J. Chem. Soc.,

Perkins Trans. 1 2000, 4451.

20 (a) Marsden, S. P. In Sustainable Catalysis; Dunn, P. J., Hii, K. K., Krische, M. J., Williams, M.

T., Eds.; John Wiley & Sons, Inc.: New York, 2013; p 339. (b) Van Kalkeren, H. A.; van Delft, F.

L.; Rutjes, F. P. J. T. ChemSusChem 2013, 6, 1615. (c) An, J.; Denton, R. M.; Lambert, T. H.;

Nacsa, E. D. Org. Biomol. Chem. 2014, 12, 2993.

21 Denton, R. M.: An, J.: Adeniran, B. Chem. Commun. 2010, 3025

22 Marsden, S. P.; A. E. McGonagle, A.E.; McKeever-Abbas, B. Org. Lett. 2008, 10, 2589.

23 Timosheva, N.V.; Chandrasekara, A.; Day, R.O.; Holmes, R.R. J. Am. Chem. Soc. 2002, 124,

7035.

24 Denney, D.B.; Jones, D.H. J. Am. Chem. Soc. 1969, 91, 5821.

25 (a) Ramirez, F.; Desai, N.B. J. Am. Chem. Soc. 1960, 82, 2652. (b) Ramirez, F. Acc. Chem. Res.

1968, 1, 168.

26 (a) Ramirez, F.; Patwardhan, A.V.; Kugler, H.J.; Smith, C.P. Tetrahedron 1968, 24, 2275. (b)

Ramirez, F.; Gulati, A.S.; Smith, C.P. J. Org. Chem. 1968, 33, 13.

27 (a) Mark, V. J. Am. Chem. Soc. 1963, 85, 1884; (b) Newman, M. S.; Blum, S. J. Am. Chem. Soc.

1964, 86, 5598; (c) Ramirez, F.; Gulati, A.S.; Smith, C.P. J. Org. Chem. 1968, 33, 13; d) Liu, X.;

Verkade, J.G. J. Org. Chem. 2000, 65, 4560.

28 (a) Corre, E.; Foucaud, A. J. Chem. Soc. D 1971, 570; (b) Fauduet, H.; Burgada, R. Synthesis

1980, 642.

29 (a) Miller, E. J.; Zhao, W.; Herr, J. D.; Radosevich, A. T. Angew. Chem., Int. Ed. 2012, 51, 10605.

(b) Zhao, W.; Fink, D. M.; Labutta, C. A.; Radosevich, A. T. Org. Lett. 2013, 15, 3090. (c) Zhao,

W.; Yan, P.K.; Radosevich, A.T. J. Am. Chem. Soc. 2015, 137, 616.

30 Swamy, K. C. K.; Kumar, N.S. Acc. Chem. Res., 2006, 39, 324.

31 Welch, G. C.; Juan, R. R. S.; Masuda, J. D.; Stephan, D. W. Science, 2006, 314, 1124.

32 Bayne, J.M.; Stephan, D.W. Chem. Soc. Rev. 2016, DOI: 10.1039/C5CS00516G

33 Terada, M.; Kouchi, M. Tetrahedron, 2006, 62, 401.

34 Perez, M.; Hounjet, L.J.; Caputo, C.P.; Dobrovetsky, R.; Stephan, D.W. J. Am. Chem. Soc. 2013,

135, 18308.

35 Werner, T. Adv. Synth. Catal. 2009, 351, 1469.

36 (a) Holmes, R. R.; Gallagher, W. P. Inorg. Chem. 1963, 2, 432. (b) Yong, C.Y.; Kennepohl,

D.K.; Cavell, R.G. Chem. Rev. 1996, 96, 1917.

37 (a) Uchida, Y.; Onoue, K.; Tada, N.; Nagao, F.; Oae, S. Tet. Lett. 1989, 30, 567. (b) Newkome,

G.R.; Hager, D.C. J. Am. Chem. Soc. 1978, 100, 5567.

38 Berlin, K.D.; Butler, G.B. Chem. Rev. 1960, 60, 243.

39 (a) Wittig, G.; Rieber, M. Liebigs Ann. Chem. 1949, 562, 187. (b) Hellwinkel, D. Chem. Ber.

1965, 98, 576.

40 Wittig, G.; Geissler, G. Liebigs Ann. Chem. 1953, 580, 44.

41 Wittig, G.; Maercker, A. Chem. Ber. 1964, 97, 747.

42 Burgada, R. Phosphorus and Sulfur 1976, 2, 237.

43 Grechkin, N.P.; Shagidullin, R.R.; Gubanova, G.S. Bulletin of the Academy of Sciences of the

USSR 1968, 17, 1700

44 Germa, H.; Willson, M.; Burgada, R. C.R. Acad. Sci. Paris C 1970, 270, 1426.

45 Nakafuji, S.-Y.; Kobayashi, J.; Kawashima, T.; Schmidt, M.W. Inorg. Chem. 2005, 44, 6500.

46 Kimura, Y.; Kokura, T.; Saegusa, T. J. Org. Chem. 1983, 48, 3815.

47 Andrews, L.; Withnall, R. Inorg. Chem. 1989, 28, 494.

Chapter 2

Oxidative Addition and Reductive Elimination at a T-shaped, Planar Phosphorus Compound Provides Phosphorus Transfer Hydrogenation Catalysts

Abstract

In this chapter, we describe a planar, T-shaped phosphorus (III) compound (2-1) that adds two hydrogen atoms at a single phosphorus center by dehydrogenation of ammonia borane (2-2,

AB). The resulting dihydridophosphorane (2-3) is found to transfer hydrogen to unsaturated substrates. This reactivity has been applied to the reduction of azobenzene (2-4), which can be completed using either P(III) (2-1) or P(V) (2-3) compounds as catalysts.1 This phosphorus- catalyzed reduction adds to a limited but growing repertoire of hydrogen transfer mechanisms catalyzed by non-metals, and uniquely involves hydrogen transfer to and from a single main group atom.

2.1 Introduction

The elementary reactions of oxidative addition and reductive elimination that occur at metal complexes have been harnessed by synthetic chemists to effect a variety of synthetically useful transformations.2 Chapter 1 has detailed how certain phosphorus compounds also undergo oxidative addition3 and reductive elimination reactions.4 While some of these reactions have been observed to be reversible,5 no catalytic applications had yet been presented. My goal was to enable reversible oxidative addition and reductive elimination at a phosphorus center, and apply this reactivity in a catalytic fashion.

2.1.1 Structural and Electronic Requirements for Two Electron Redox at Phosphorus

In order to achieve the goal of two-electron redox catalysis at phosphorus, we sought to lower the activation barrier to both oxidative addition and reductive elimination. We believed that the key to oxidative addition at phosphorus would be 1) the orientation of vacant coordination sites, and 2) the orientation of the HOMO and LUMO.

It is known that three-coordinate pyramidal phosphines can be converted to five-coordinate hypervalent phosphoranes by the addition of a suitable oxidizing agent. Such a transformation requires structural reorganization to accommodate the transition from pyramidal to trigonal bipyramidal geometries about the phosphorus center. Concerted mechanisms for this transformation are governed by the Woodward-Hoffman selection rules. Specifically, reductive elimination (and by extension, oxidative addition) from a trigonal bipyramidal phosphorane must involve axial-axial or equatorial-equatorial coupling, as determined by computational studies.6

Geometric distortion of a pyramidal phosphine to a T-shaped phosphorus(III) compound would overlay onto a trigonal bipyramid with two vacant equatorial sites (Figure 2-1). Oxidative addition to this T-shaped phosphorus compound could in principle occur with minimal structural rearrangement according to the Woodward-Hoffman orbital overlap rules. This hypothesis led us to investigate 2-1, a unique planar, T-shaped phosphorus (III) compound reported by Arduengo and coworkers.7

Figure 2-1. Structural distortion from pyramidal to T-shaped geometries may ease the transition to five coordinate trigonal bipyramidal structures.

Concerted oxidative addition at metal centers requires orthogonal filled and unfilled d orbitals.8 Herein we will examine the orbital configuration for planar, T-shaped compound 2-1,

19 shown in Figure 2-2. T-shaped geometries arise from hypervalent bonding, and, in fact, the hypervalent nature of the P-O bonds in 2-1 is supported by crystallographic data.7 The oxygens on phosphorus occupy a 3-center-4-electron bond: phosphorus contributes a p orbital and 2-electrons to this bond.9 Three phosphorus orbitals remain: a sp orbital which bonds with nitrogen, a vacant p orbital, and a lone pair in the sp orbital.10 Calculations by Alexander Radosevich have confirmed that 2-1 contains a LUMO with p orbital character, and a HOMO-2 with lone pair character. The orthogonal orientation of HOMO and LUMO should facilitate concerted oxidative addition at 2-1.

For comparison, an idealized trigonal pyramidal geometry for a phosphine would have the HOMO

(lone pair) and LUMO (σ* P-C orbital) at an angle of ~70°.11

Figure 2-2. A) Compound 2-1. B) Depiction of perpendicular p orbital LUMO and lone pair HOMO at phosphorus. C) Computations of HOMO-2 at 2-1 and D) LUMO at 2-1.

2.1.2 Characteristics of a T-shaped P(III) Compound

T-shaped compound 2-1 is formed from the reaction of diketoamine 2-5 with phosphorus trichloride in the presence of triethylamine (Figure 2-3).7 X-ray crystallography of this compound

20 has confirmed that 2-1 has a planar T-shaped structure, which is unusual for a trivalent phosphorus species. The angle ∠OPO angle is ca. 169°, just shy of linear; for comparison, an idealized trigonal pyramid would have angles of 109.5°.11 Notably, the P-O bond lengths are 1.84 and 1.81 Å, which are comparable to apical P-O bonds in pentacoordinate phosphoranes.12 The elongation indicates that 2-1 demonstrates hypervalent characteristics. Overall, this crystallographic analysis indicates that 2-1 is a T-shaped, planar, and hypervalent σ3 phosphorus species. These characteristics fit our requirements to allow for concerted oxidative addition with minimal structural reorganization of the compound.

Figure 2-3. Synthesis of 2-1.

Previously, compound 2-1 has been reported to undergo oxidative additions with several classes of compounds, including 1,2-dicarbonyl compounds and alcohols (Figure 2-4). The reaction of 2-1 with 1,2-dicarbonyl o-chloranil yields phosphorane 2-6, where the O,N,O ligand scaffold bends to accommodate a new apical-equatorial spanning ring system.7b In contrast, 1 reacts rapidly with methanol to form σ5 phosphorane 2-7, but this species is not stable and undergoes further rearrangement to 2-8.7b Similar reactivity is observed with other alcohols, including isopropanol and p-cresol. More recently, our group has reported that 2-1 undergoes oxidative addition with ammonia and other primary amines to form stable σ5 phosphoranes such as

2-9.13 In this example of oxidative addition, the newly formed P-N and P-H bonds occupy equatorial positions, and the O,N,O-backbone is not disturbed from its planar orientation.

Figure 2-4. Oxidative addition of 1 with A) 1,2-dicarbonyl chloranil, B) methanol, and C) propylamine.

Arduengo also reported that the dihydridophosphorane 2-3, which is derived from 2-1, evolved hydrogen and reformed 2-1 (Figure 2-5).7c We seized upon this report to investigate the possibility of reversible hydrogen activation at a phosphorus center through oxidative addition and reductive elimination.

Figure 2-5. Report of hydrogen release from a dihydridophosphorane.

2.2 Synthesis of Dihydridophosphorane 2-3

2.2.1 Synthesis of Dihydridophosphorane 2-3 from Dichlorophosphorane 2-10

Arduengo had previously reported that the treatment of dichlorophosphorane 2-10 with lithium aluminum hydride provided dihydridophosphorane 2-3.7c Exact conditions were not provided, so we first reacted 2-10 with 2 equivalents of LiAlH4 in THF at room temperature. This procedure led only to decomposition; no phosphorus species were observed by 31P NMR. At this point we began to screen the reaction of 2-10 with various hydride donors (Figure 2-6). Treatment of 2-10 with diisobutylaluminumhydride (DIBAL) also lead to decomposition. The failure of both

DIBAL and LiAlH4 to produce any tractable results suggested that aluminum hydrides are too reactive to be compatible with 2-10.

Figure 2-6. Synthesis of dihydridophosphorane 2-3 from dichlorophosphorane 2-10.

Next, we tested borohydride reductants using the same reaction conditions. Sodium borohydride proved to be unreactive with the substrate in THF, so we turned instead to more soluble borohydrides. The reaction of 2-10 with tetrabutylammonium borohydride produced 2-3, as indicated by a triplet at -45 ppm by 31P NMR, which is consistent with the reported NMR shift of the dihydridophosphorane.7c However, attempts to scale up this reaction and isolate the product led to low yields (<20%) and significant impurities.

The reaction of 2-10 with sodium cyanoborohydride also successfully produces 2-3; however, three-coordinate compound 2-1 is also observed. This reaction is quite exothermic, so conditions were altered such that sodium cyanoborohydride was added to a solution of 2-10 cooled to –35 °C. Reducing the temperature did not eliminate side product 2-1, but provided 2-1 and 2-3 in a ratio of 1:2. We suspected that the formation of 2-1 is caused by a side reaction with 2-11, formed via a single hydride replacement at 2-10 (Figure 2-7). If this hypothesis was correct, then increasing the equivalents of hydride would ensure that a second hydride displacement occurs prior to decomposition of intermediate 2-11. We were gratified to find that a significant excess of reductant (15-20 equivalents) leads to a more favorable product ratio, with as little as 5% of 2-1 observed. This reaction can be performed on gram scale, and the desired product 2-3 is isolated after concentrating the reaction to a solid and extracting with pentane.

Figure 2-7. Proposed mechanism of the formation of 2-3 and 2-1 from 2-10.

2.2.2 Synthesis of Dihydridophosphorane 2-3 from Three-Coordinate Species 2-1

While we were successful at preparing dihydridophosphorane 2-3 from the corresponding dichlorophosphorane, we were also particularly interested in exploring the transformation of three-

24 coordinate compound 2-1 to 2-3 via hydrogen addition at phosphorus. We attempted to react 2-1 with both hydrogen gas and other hydrogen sources to form 2-3.

We first examined the ability of 2-1 to activate diatomic hydrogen. Reacting 2-1 under a hydrogen atmosphere, even at elevated temperature (50 °C and 110 °C) and pressures (1-40 atm), did not produce 2-3 (Figure 2-8). The addition of weak bases (proton sponge, DBU, DABCO or

KHMDS) under 1 atm hydrogen to effect heterolytic bond cleavage, as demonstrated by DuBois, also failed to effect hydrogen addition at phosphorus.14

Figure 2-8. Attempted synthesis of 2-3 from 2-1 and H2.

Since hydrogen gas proved an unsuitable reagent for the desired transformation, we turned to other commonly used hydrogen sources. For example, ammonium formate is used as a source of hydrogen in transfer hydrogenations.15 These reactions are typically catalyzed by metal catalysts such as palladium on carbon, and cause ammonium formate to decompose to hydrogen gas, ammonia, and carbon dioxide. However, the reaction of ammonium formate with 2-1 in the absence of a catalyst was not productive, as no hydrogen transfer was observed.

Next, we turned to ammonia borane (2-2, AB) as a hydrogen source. AB is a coordination compound of ammonia and borane. It is of interest for hydrogen storage and release purposes due to its high hydrogen content by weight (~20%).16 Hydrogen loss from AB occurs in both the solid17 and solution18 states upon heating; however, the compound is a stable solid at room temperature.

We found that the reaction of 2-1 with AB furnishes 2-3 (Figure 2-9). The reaction requires heating, but proceeds at temperatures from 40-80 °C. Excess ammonia borane (15 equivalents) is required

to drive the reaction to completion in a reasonable time frame. Isolation of 2-3 using these conditions can be achieved with 75% yield.

Figure 2-9. Attempted synthesis of 2-3 from 2-1 and H2.

In considering the transformation from 2-1 to 2-3, we were struck by the unique reactivity of 2-1 with AB compared to other hydrogen sources. This result, combined with the rare ability of the system to transfer an equivalent of hydrogen to a single non-metal atom, led us to stringently examine our reaction conditions. Previous reports have found that trace transition metals can contaminate common reagents, and may be responsible for the reactivity of “non-metal catalyzed” reactions.19 To determine whether trace metal contamination was a factor in our case, we compared the reactivity of 2-1 with various AB sources, including 1) commercially available 90% technical grade and 97% grade AB from Sigma-Aldrich used without purification, as well as 2) commercial sources of AB that had been purified by sublimation or recrystallization (from either diethyl ether or methanol), or 3) AB samples which had been synthesized from sodium borohydride and ammonium sulfate. All samples of AB transferred hydrogen to 1, and the rate of conversion in each case was nominally similar. If a trace impurity is responsible for the observed reactivity, it is soluble in organic solvents, easily sublimed, and present in various grades of AB, as well as in either ammonium chloride or sodium borohydride.

2.3 Characterization of 2-3

2.3.1 NMR Spectroscopy

The proton NMR spectrum of dihydridophosphorane 2-3 presents three unique signals in d3-acetonitrile (Figure 2-10A). The t-butyl groups have an 18 proton singlet at 1.06 ppm. The vinylic protons appear at 5.82 ppm as a doublet with a coupling constant of 34 Hz. This splitting

3 pattern is due to JP-H coupling. The hydrogens on phosphorus are the most diagnostic, and are present as a doublet at 7.96 ppm, with a coupling constant of 670 Hz. The value of the coupling

1 constant is indicative of JP-H coupling, and confirms that these hydrogen are bound directly to phosphorus. The number of unique signals indicate that the molecule is symmetrical, with both hydrogens on phosphorus equivalent in solution at room temperature.

Phosphorus NMR of 2-3 displays a triplet of triplets centered at –43.7 ppm, with coupling constants of 34 and 670 Hz (Figure 2-10B). The shift is indicative a five-coordinate phosphorane.20

The coupling constants mirror those observed in the proton spectrum, and therefore most likely

1 3 arise from JP-H and JP-H coupling. The provenance of the coupling is further confirmed as the proton-decoupled phosphorus NMR collapses into a singlet.

2.3.2 X-ray Crystallography

A crystal of dihydridophosphorane 2-3 (Figure 2-11, Tables 2-1 and 2-2) was prepared by sublimation of the solid compound. X-ray diffraction performed by Hemant Yennawar provides structural detail; overall, the structure deviates slightly from an idealized trigonal bipyramid. The bond angle ∠OPO is 170.5°, which shows very little change from the same bond angle observed in compound three-coordinate compound 2-1 (169°). The P-N bond is 1.676 Å, while the P-O bonds are 1.721 and 1.699 Å. The P-N bond length and OPO bond angle values are nearly identical to those found in 2-1, substantiating our theory that the unique planar, T-shaped phosphorus compound could undergo oxidative addition with very few structural changes. Interestingly, the

P-O bond distances have shortened by over 0.1 Å. Typically, apical bonds in 5-coordinate phosphoranes would lengthen compared to their 3-coordinate analogues. Observing the opposite effect supports the theory that the P-O bonds in 2-1 are already hypervalent.

P-H bonds were located on the difference Fourier map to provide additional information about the inner sphere bonds of 2-3. The hydrides are inequivalent in the solid state, with bond lengths of 1.31 and 1.29 Å, and ∠PNH of 116.1 and 134.8°.

Table 2-1. Selected bond lengths for 2-1 and 2-3 in Å Bond 2-1 2-3 P-N 1.68 1.676 P-O1 1.84 1.721 P-O2 1.81 1.600 P-H1 -- 1.29 P-H2 -- 131

Table 2-2. Selected bond angles for 2-1 and 2-3 reported in degrees. Angle 2-1 2-3 ONO 169 170.5 NPO1 83.5 87.6 NPO2 84.2 88 NPH1 -- 134.8 NPH2 -- 116.1

2.3.3 Computations of P-H Bonds

X-ray crystallography indicates that the two hydrogens bonded to phosphorus are inequivalent; however, no distinction between the two is observed in 1H NMR. Calculations

29 performed by Alexander Radosevich provide geometry optimization at the B3LYP/6-311+G* level of theory. These computations show that a distorted square pyramid with inequivalent P-H bonds is a local minima, in agreement with X-ray data. However, a trigonal bipyramidal structure with equivalent P-H bonds is less than 1 kcal/mol higher in energy. P-H rocking leads to conversion between these two structures, and is accessible at room temperature, as represented in the NMR data.

2.3.4 Structural Equilibration

It has been previously reported that σ5 hydridophosphoranes are often in equilibrium with

σ3 phosphorus species with pendant X-H groups. Such an equilibrium could be envisioned between five-coordinate 2-3 and three-coordinate compound 2-12 (Figure 2-12). In some cases, this equilibrium has been observed via NMR spectroscopy,21 IR spectroscopy,22 or solid state structures.23 In other compounds, the σ3 phosphorus species is not observed directly; however, its existence via equilibrium is implied by reactivity that is indicative of a phosphorus atom lone pair.24

Figure 2-12. Possible equilibrium between five coordinate 2-3 and three coordinate 2-12.

5 NMR spectroscopy and x-ray crystallography of compound 2-3 show only the σ species at room temperature and up to 80 °C. We have therefore attempted several reactions to determine whether 2-3 is in equilibrium with 2-12, which would have a lone pair at phosphorus to participate in various nucleophilic reactions.

To determine whether 2-3 is capable of acting as a nucleophile, it was exposed to methyl iodide, borane-dimethylsulfide, and selenium. No reaction was observed between 2-3 and methyl iodide, even at elevated temperatures; a three coordinate phosphorus species would be alkylated to form a methylphosphonium salt.25 Likewise, 2-3 does not form a phosphorus selenide in the presence of selenium; a σ3 species would form the heavy analogue of a phosphine oxide.26 While

2-3 does react with borane-dimethylsulfide, the resulting compound 2-13 is still a five coordinate phosphorus complex: the product has a 31P NMR signal of a triplet at –54.3 ppm, with a coupling constant of 581 Hz. This small upfield deviation in chemical shift from the parent complex (at ca.

–44 ppm) to the product is consistent with borane coordination at nitrogen, not at phosphorus

(Figure 2-13).27 It should be noted that no further reaction is seen, even in the presence of 10 equivalents of borane complex; in contrast, a σ3 species would form a phosphine-borane Lewis complex.

Figure 2-13. 2-3 reacts with borane-dimethylsulfide and maintains a five coordinate structure.

Collectively, all spectral, structural, and reactivity data indicates that 2-3 exists solely as a phosphorane, and is not in equilibrium with σ3 phosphorus species 2-12. This conclusion is supported by previous findings of Burgada, which indicated that planar, aromatic phosphacycles favor the σ5 configuration.28 There are also reports of other σ5 hydridophosphoranes which do not equilibrate with σ3 complexes.29

2.3.5 Synthesis and Characterization of 2-3 Isotopologues

During our investigations of 2-3, we were interested in synthesizing and characterizing both the monodeuterated and dideuterated isotopologues. While we successfully obtained 2-3-d2, attempts to prepare the mixed isotopologue 2-3-d from 2-1 have failed. Instead, 2-3-d2 is the sole phosphorus containing product obtained from the reaction of 2-1 with D3NBH3, isolated in good chemical yield (Figure 2-14). Complementarily, when H3NBD3 is employed, only 2-3 is obtained.

In short, the nitrogen substituent governs the outcome of the mixed isotopic experiments. We believe this result to be a function of facile P–H exchange between the pentacoordinate phosphoranes and the labile N–H(D) of the ammonia-boranes. Specifically, 2-3 can be converted to 2-3-d2 by treatment with excess D3NBH3 but not H3NBD3. Indeed, a similar isotopic exchange can be observed upon treatment of 2-3 with excess BnND2. Similar reactivity is reported with other hydridophosphoranes.30

Figure 2-14. Synthetic routes 2-3 and 2-3-d2 by dehydrogenation of ammonia borane and isotope exchange.

1 When 2-1 is treated with ND3BH3 the dideuterophosphorane 2-3-d2 is formed. The H

NMR spectra of 2-3-d2 is nearly identical to the spectra for 2-3. The absence of signals for the

31 hydrogens bonded to phosphorus distinguish the complexes. The P NMR signal for 2-3-d2 is shifted slightly upfield to –46.3 ppm (Figure 2-15), as compared to the signal for 2-3. The 31P{1H}

NMR features a quintet formed from the coupling of phosphorus with two equivalent 2H nuclei,

1 1 31 3 such that JP-D = 103 Hz. In the H coupled P NMR, the signal is a quintet of triplets, with a JP-H

= 34 Hz, which is the same magnitude observed for the coupling of phosphorus with the vinylic

2 2 protons in 2-3. When the reaction of 2-1 and ND3BD3 is monitored by H NMR, the H nuclei on

1 phosphorus are observed as a doublet at 8.03 ppm with a JP-D = 103 Hz, which is nearly identical to that in the hydrogen isotopologue. It is notable that no other 2H signals were observed, indicating that no proton exchange occurs under the reaction conditions.

31 Figure 2-15. P NMR of 2-3-d2.

2.4 Reactivity of 2-3

2.4.1 Hydrogen Release from Dihydridophosphorane 2-3 to Reform Three Coordinate Compound 2-1

Arduengo had previously stated that hydrogen gas could be released from 2-3; however,

7c no experimental detail was provided. In our hands, heating samples of 2-3 in solvent did occasionally produce signals for 2-1 as identified by 31P NMR. However, these results were not reproducible. Upon scale-up, no evolution of hydrogen could be observed by GC analysis or volumetric measurements (Figure 2-16).

Figure 2-16. Attempted release of hydrogen gas from 2-3.

Since we were unable to produce hydrogen directly from 2-3, we sought instead to evaluate the ability of the compound to reduce compounds which could act as hydrogen acceptors.

Azobenzene 2-4 proved to be an appropriate substrate: in the presence of 2-3, 63% of azobenzene is converted to diphenylhydrazine 2-14 after 19 h at 40 °C, as determined by NMR integration

(Figure 2-17). In the presence of excess azobenzene, 2-3 is converted cleanly to 2-1.

Figure 2-17. Reduction of azobenzene (2-4) to diphenylhydrazine (2-14) with 2-3 as the reductant.

2.4.2 Catalytic Reduction of Azobenzene with Phosphorus Catalysts

Section 2.2.2 detailed that 2-1 is converted into 2-3 via hydrogen transfer from AB.

Additionally, Section 2.4.1 detailed that hydrogen transfer from 2-3 to azobenzene 2-4 cleanly provides 2-1 with concomitant reduction of the unsaturated bond in 2-4 (Figure 2-18). The application of these reactions sequentially should therefore allow for either 2-1 or 2-3 to be used as a transfer hydrogenation catalyst.

Figure 2-18. Stoichiometric transitions between 2-1 and 2-3.

Treating azobenzene with 10 mol% 2-1 and 4 equivalents of AB in acetonitrile at 40 °C

1 for 24 h provides 2-14 in 80% yield, as measured by H NMR (Table 2-3). Using 10 mol% of 2-3 as a catalyst provides 81% yield after 48 h. Comparatively, treating azobenzene with AB alone does not promote any reaction. Additionally, the addition of stoichiometric amounts of other phosphorus complexes, including PPh3 and P(OMe)3, does not promote the reduction of azobenzene. Only the addition of stoichiometric P(NMe2)3 provides a discernable amount of 2-14, albeit with a significantly decreased yield of 24%.31 These results indicate that compounds 2-1 and

2-3 act uniquely as phosphorus-based hydrogen transfer catalysts.

Catalyst Catalyst Temp (°C) Time (h) Yielda Loading -- -- 80 24 0% 2-1 10% 40 24 80%

2-3 10% 40 48 81%

PPh3 100% 80 24 0%

P(OMe)3 100% 80 24 0%

P(NMe2)3 100% 80 24 24%

Monitoring the catalytic reduction of azobenzene using 10 mol% 2-1 by 31P NMR shows that at midpoints of the reaction (~3 h) both 2-1 and 2-3 are observed, indicating that both species are involved in the catalytic cycle. This result is expected based on the observed stoichiometric reactivity of hydrogen transfer with both compounds.

2-1 2-3

2.5 Conclusions

We have shown that a planar, T-shaped phosphorus complex 2-1 accepts hydrogen from

AB. Two hydrogen atoms are transferred to phosphorus in a rare example of H2 addition to a single nonmetal atom. The dihydride formed from this reaction can reduce azobenzene 2-4 and reform compound 2-1. This reactivity allows either species to be used as phosphorus catalysts for the transfer hydrogenation of azobenzene. These catalytic systems are added to a small group of non- metal hydrogenation catalysts. The mechanisms of hydrogen transfer to phosphorus will be

37 explored in Chapter 3. Chapter 4 will provide mechanistic detail about hydrogen transfer from phosphorus to substrates and improvements to the scope of substrates amenable to this reaction.

2.6 Experimental

2.6.1 General Materials and Methods

Ammonia-borane (90%, technical grade) was purchased from Aldrich and purified by recrystallization from ethanol prior to use. Phosphorous (V) chloride was sublimed prior to use.

Diethyl ether (Et2O), methylene chloride (CH2Cl2), tetrahydrofuran (THF), and toluene were dried according to the method of Grubbs32 as modified by Bergman.33 All other commercially available reagents were purchased from suppliers and used without further purification. All reactions were carried out under nitrogen either in a double port glovebox (Innovative Technology) or with a

Schlenk manifold vented through an oil bubbler unless otherwise noted. All glassware was oven- dried at 120 °C prior to use. 1H, 13C, and 31P NMR spectra were recorded with Bruker DPX-300,

AV-360, and DRX-400 spectrometers. 1H NMR spectra were referenced to residual solvent peaks

31 (CDCl3: 7.26 ppm, CD3CN: 1.94 ppm, DMSO-d6 2.50 ppm). P NMR spectra were referenced to an external standard (H3PO4: 0.0 ppm). Mass spectrometric data were obtained from the Proteomics and Mass Spectrometry Core Facility operated by the Huck Institute of Life Sciences at the

Pennsylvania State University.

2.6.2 Synthetic Procedures and Data

Phosphorus compounds 2-1 and 2-10 were prepared by minor modifications to the literature method7b according to the following sequence (Figure 2-18):

Figure 2-20. Synthesis of 2-1 and 2-10. (a) Br2, Et2O, 93%; (b) BnNH2, PhMe, 80 °C, 61%; (c) H2, Pd/C, MeOH, 88%; (d) K3PO4, H2O, CH2Cl2, 88%; (e) PCl3, Et3N, C5H12, 71%; (f) PCl5, Et3N, C5H12, 77%

2-15: The following procedure is adapted from the literature.34 To a solution of

pinacolone (80 mL, 640 mmol) in ethyl ether (80 mL) was added several drops of

bromine at ambient temperature. The dark solution became colorless and was subsequently cooled in an ice bath. Bromine (36 mL, 702 mmol) was added dropwise over an hour so as to maintain the internal temperature below 10 °C. Following the addition, the reaction mixture was stirred an additional 30 min, then 200 mL of water was added and stirred for 10 min. The reaction mixture was warmed to room temperature and treated with solid sodium bicarbonate (59 g, 702 mmol); the resulting phasic layers partitioned. The ether layer was washed with brine (1x50 mL), aqueous sodium bicarbonate (3x50 mL), dried (Na2SO4) and concentrated in vacuo.

Purification by vacuum distillation yielded 2-15 as a light yellow oil (107 g, 598 mmol, 93%). 1H

NMR (300 MHz, CDCl3): δ 4.18 (2H, s), 1.23 (9H, s) ppm.

2-16:To a solution of bromopinacolone 2-15 (107 g, 598 mmol) in

nitrogen-sparged toluene (650 mL) was added benzylamine (65 mL, 595

mmol) in one portion. The reaction mixture was heated at 80 °C with stirring for 62 h, during which time a voluminous white solid precipitated. The solid was collected on a Buchner funnel, washed with water (2x100 mL), and recrystallized in ethanol to give S2-2 as

1 a white crystalline solid (69.4 g, 181 mmol, 61%). H NMR (360 MHz, CDCl3): δ 7.66 (2H, m),

7.46 (3H, m), 4.94 (2H, m), 4.72 (4H, m), 1.14 (18H, s).

2-17: To a suspension of 2-16 (20.0 g, 52.1 mmol) in nitrogen-sparged

methanol (180 mL) was added 10% palladium on carbon (200 mg, 0.19

mmol). The atomsphere was exchanged for hydrogen via three evacuation/backfill cycles, and the heterogeneous mixture was stirred under H2 (1 atm) for 18 hours. The reaction mixture was then filtered over celite, and the filtrate was concentrated.

Recrystallization of the residue in ethanol gave 2-17 as a white solid (13.5 g, 45.9 mmol, 88%). 1H

NMR (360 MHz, CDCl3): δ 4.31 (4H, s), 1.23 (18H, s).

2-5: In a 1 L round bottom flask with magnetic stirbar, 2-17 (19.8 g, 67.2

mmol) in water (300 mL) and dichloromethane (150 mL) was cooled in

an ice bath. An aqueous solution of tribasic potassium phosphate (17 g,

80.2 mmol, 150 mL, 0.5 M) was added dropwise over 30 minutes, then the solution was stirred for an additional 2 hours in an ice bath. The aqueous layer was extracted with dichloromethane (3x40 mL), then the dichloromethane extracts were washed with water (3x40 mL). The organic layer was dried over sodium sulfate and concentrated in vacuo to give 2-5 as a yellow solid (12.6 g, 259.2

1 mmol 88%), which was stored at –35 °C before use. H NMR (400 MHz, CDCl3): δ3.62 (4H, s),

2.83 (1H, br s), 1.15 (18H, s).

2-1: To a solution of phosphorus(III) chloride (1.1 mL, 12.6 mmol) in pentane (60 mL)

under nitrogen at –78 °C was added 2-5 (2.67 g, 12.5 mmol) in pentane (60 mL)

dropwise. Triethylamine (5.2 mL, 37.6 mmol) in pentane (20 mL) was then added

dropwise, and the resulting mixture was warmed to room temperature and stirred for

an additional 2.5 hours. The reaction mixture was concentrated to a residue in vacuo, triturated with pentane and filtered over celite. The filtrate was concentrated to give 2-1 as an off- white solid (2.14 g, 8.9 mmol, 71%), which can be further recrystallized from a concentrated

1 pentane solution at –35 °C. H NMR (400 MHz, CDCl3): δ 7.40 (2H, d, J=9.6 Hz), 1.27 (18H, s).

31 1 13 P{ H} NMR (145 MHz, CDCl3): 187.0 ppm. C NMR (75 MHz, CDCl3): δ 169.61, 110.63 (d,

J=5.6 Hz), 33.96 (d, J=6.6 Hz), 27.94 ppm.

2-10: To a solution of phosphorus (V) chloride (3.13 g, 15.1 mmol) in

dichloromethane (60 mL) under nitrogen at –78 °C was added 2-5 (3.20 g, 15.1

mmol) in dichloromethane (30 mL) dropwise. Triethylamine (6.3 mL, 45.5 mmol)

in dichloromethane (15 mL) was then added dropwise and the solution was warmed to room temperature and stirred for 12 hours. The solvent was removed in vacuo, and the residue was dissolved in pentane and filtered. The filtrate was concentrated to give 2-10 as a brown

1 solid (3.61 g, 11.6 mmol, 77%). H NMR (300 MHz, CDCl3) δ 6.06 (2H, d, J=35 Hz), 1.14 (18H,

31 s). PNMR (145 MHz, CDCl3): –23.1 (t, J=35 Hz) ppm.

2-3: Dihydridophosphorane 2-3 was prepared by two distinct methods. Method A:

To a solution of 2-10 (1.52 g, 4.87 mmol) in tetrahydrofuran (50 mL) under nitrogen

at –35 °C was added sodium cyanoborohydride (7.25 g, 115.0 mmol) in one portion.

The reaction mixture was stirred at –35 °C for 30 min. The solvent was then removed in vacuo, and the resulting residue was triturated with pentane. The suspension was filtered and the filtrate evaporated to give 2-3 as a pale yellow solid (702 mg, 2.89 mmol, 59%) in

41 ca. 95% purity (contains ~3% 1 by 1H NMR integration). Method B: To a solution of 2-1 (350 mg,

1.45 mmol) in acetonitrile (17.5 mL) under nitrogen was added borane-ammonia complex (1.06 g,

34.2 mmol). The reaction mixture was heated to 40 °C for 21 h. Upon cooling to room temperature, the solvent was removed in vacuo. The residue was triturated with pentane and the resulting suspension was filtered over celite. The filtrate was evaporated to give 2-3 as an off-white solid

(266 mg, 1.09 mmol, 75%). X-ray quality crystals were produced by vacuum sublimation at 35 °C

1 and ca. 1 mmHg onto a water-chilled cold finger. H NMR (400 MHz, CD3CN) δ7.96 (2H, d, J =

31 670 Hz) (9.07, 6.84), 5.82 (2H, d, J = 34 Hz), 1.06 (18H, s) ppm. P NMR (145 MHz, CDCl3): –

1 3 13 43.7 (tt, JPH = 670 Hz, JPH = 34 Hz) ppm. C NMR (75 MHz, CDCl3): δ 151.32 (d, J=5.1 Hz),

+ 101.10 (d, J=17.9 Hz), 31.98 (d, J=2.6 Hz), 27.73 HRMS (EI) calc’d for [C12H22NO2P] : 243.1388; found: 243.1387.

2-3-d2: Available from two different methods: Method A: A solution of 2-3 (10 mg,

0.04 mmol) and AB-d6 (30 mg, 0.8 mmol) in d8-THF (0.6 mL) is heated for 8 hr at

60 °C. Method B: A solution of 2-3 (10 mg, 0.04 mmol) and benzylamine-N-d2 (9

1 mg, 0.08 mmol) in C6D6 (0.5 mL) is heated for 1 hr at 80 °C. H NMR (400 MHz,

31 1 CD3CN) δ5.82 (2H, d, J = 34 Hz), 1.06 (18H, s) ppm. P NMR (145 MHz, CDCl3): –46.3 (tt, JPH

3 13 = 670 Hz, JPH = 34 Hz) ppm C NMR (75 MHz, CDCl3): δ 151.29 (dd, J=5.1, 2.1 Hz), 101.24 (d,

J=18.0 Hz), 31.97 (d, J=2.4 Hz), 27.72 ppm.

2-3-d: A solution of 2-3-d2 (40 mg, 0.16 mmol) and 2-3 (40 mg, 0.16 mmol) in C6D6

31 (1 mL) was heated to 60 °C for 1 h. P NMR shows a statistical mixture of 2-3, 2-

3-d, and 2-3-d2

2-13: To 2-3 (20 mg, 0.08 mmol) in C6D6 (1 mL) was added borane-dimethylsulfide (7.6 mg, 0.1

31 mmol) and left at room temp 1 h. P NMR (145 MHz, C6D6): 54.3 ppm (t, J=574 Hz).

To determine that 2-3 does not react with methyl iodide or selenium, 2-3 (0.1 mmol) was mixed

31 with 10 equivalents of reagent in CDCl3 (1 mL), and the reaction was monitored by P NMR.

After 1 h at room temperature, the reagents were heated to 60 °C for 8 h. No reaction was observed in either case.

35 2-2: NH3BH3 was synthesized from a literature procedure. To powdered ammonium sulfate (5.28 g, 40 mmol) and sodium borohydride (1.5 g, 40 mmol) was added dry THF, with an inlet of nitrogen and an outlet via a bubbler which was vented directly to the hood exhaust to prevent the buildup of hydrogen. The flask was heated to 40 °C, then stirred for four hours and monitored by crude boron

NMR (NH3BH3 = –24 ppm, NaBH4 = –40). Upon completion, the reaction was then filtered and

1 filtrate concentrated en vacuo, to provide 2-2 as a white solid (1.03 g, 81%). H NMR (CD3CN):

11 3.55 (3H, t, br), 1.30 (3H, quar, br). B NMR (CD3CN): –23 (quar).

36 ND3BH3: NH3BH3 (2.24 g, 72 mmol) was dissolved in D2O (20 mL). The mixture was stirred five minutes, then the solvent was removed in vacuo. The procedure was repeated twice to yield a white solid, which was dissolved in THF and dried over sodium sulfate, then concentrated to yield

1.63 g (66%). 1H NMR shows >99% deuterium incorporation. 1H NMR (CD3CN): 1.44 (broad quartet)

NH3BD3: Performed as 2-2, with ammonium sulfate (3.17 g, 24 mmol) and sodium borodeuteride

(1 g, 24 mmol). Following work-up, sample was dissolved in H2O (10 mL) and evaporated at reduced pressure, repeated 2x, dissolved in THF and dried over sodium sulfate to give a white solid

(694 mg, 85%). 1H NMR shows 90% BD3 incorporation. 1H NMR (CD3CN):

ND3BD3: NH3BD3 (500 mg, 15 mmol) was dissolved in D2O (5 mL). The mixture was stirred five minutes, then the solvent was removed in vacuo. The procedure was repeated twice to yield a white solid, which was dissolved in THF and dried over sodium sulfate, then concentrated to yield 423 mg (84%).

2.6.2.1 Hydrogen Transfer Experiments.

Activation of Ammonia-Borane. In a sealable NMR tube, compound 2-1 (10 mg, 0.04 mmol, 1.0 equiv) and ammonia-borane (32 mg, 1.04 mmol, 26 equiv.) were dissolved in 0.5 mL of CD3CN containing trimethyl phosphate (5 µL, 0.04 mmol) as an internal standard. The reaction mixture was heated to 40 °C for 16 h, at which time 1H and 31P NMR shows complete consumption of 2-1 and conversion to 2-3. Representative 31P NMR spectra at various time points are shown in Figure

S2-8.

Hydrogen Transfer to Stoichiometric Azobenzene. In a sealable NMR tube, dihydridophosphorane

2-3 (10 mg, 0.04 mmol, 1.0 equiv.) and azobenzene 2-4 (7.5 mg, 0.04 mmol, 1.0 equiv.) were dissolved in 0.5 mL of CD3CN containing trimethyl phosphate (5 µL, 0.04 mmol) as an internal standard. A 1H and 31P NMR spectrum of this mixture was recorded as t = 0 h. The reaction mixture was then heated to 40 °C for 19 h, at which time 1H and 31P NMR integration shows 72 ± 5% conversion of 2-3 to 2-1, and 63 ± 5% conversion of 2-4 to 2-14.

Hydrogen Transfer to Excess Azobenzene. In a sealable NMR tube, dihydridophosphorane 2-3 (10 mg, 0.04 mmol, 1.0 equiv.) and azobenzene 2-4 (75 mg, 0.40 mmol, 10.0 equiv.) were dissolved in

1 0.5 mL of CD3CN containing trimethyl phosphate (5 µL, 0.04 mmol) as an internal standard. A H and 31P NMR spectrum of this mixture was recorded as t = 0 h. The reaction mixture was then heated to 40 °C for 3.5 h, at which time 1H and 31P NMR showed complete conversion of 2-3 to 2-

2.6.2.2 Catalytic Transfer Hydrogenation Experiments.

Control experiment without phosphorus additives. Ammonia borane 2-2 (51 mg, 1.65 mmol, 4.0 equiv) and azobenzene 2-4 (75 mg, 0.41 mmol) in d3-acetonitrile (1 mL) under nitrogen were heated in a sealed J. Young NMR tube to 80 °C for 24 h, at which time 1H NMR integration shows ca. 5% conversion of azobenzene to diphenylhydrazine.

General procedure for experiments with 2-1 and 2-3. In a sealable J. Young NMR tube, the phosphorus compound (0.04 mmol, 0.1 equiv), ammonia borane 2-2 (51 mg, 1.64 mmol, 3.9 equiv) and azobenzene 2-4 (77 mg, 0.42 mmol, 1.0 equiv.) were heated to 40 °C for 48 h in d3-acetonitrile

(1 mL). Reaction progress was assessed by 1H NMR integration of the crude reaction mixture.

Representative 31P NMR spectra at various time points are shown in Figure S2-11.

General procedure for experiments with PPh3, P(OMe)3 and P(NMe2)3. In a sealable J. Young

NMR tube, the phosphorus compound (0.22 mmol, 1.0 equiv), ammonia borane 2-2 (25 mg, 0.88 mmol, 3.9 equiv) and azobenzene 2-4 (38 mg, 0.22 mmol) were heated to 80 °C for 24 h in d3- acetonitrile (0.5 mL). Reaction progress was assessed by 1H NMR integration of the crude reaction mixture.

2.6.3 Kinetics Data for Arrhenius Analysis of Hydrogen Transfer from 2-3 to Azobenzene.

Kinetic runs were conducted on 0.025 M solutions of 2-3 in CDCl3 with 20-fold excess of azobenzene in sealed J. Young NMR tubes over a temperature range from 30 °C to 60 °C. Reaction progress was monitored by integration of the tert-butyl 1H NMR resonance of 2-3.

T = 30 °C

Time (s) [2-3] (M) ln [2-3]

0 0.014646 -4.22357

660 0.013784 -4.28426

1320 0.013096 -4.33548

1980 0.012381 -4.39157

2640 0.011903 -4.43095

3300 0.011372 -4.47658

3960 0.010728 -4.53494

4620 0.010155 -4.58979

5280 0.009753 -4.63017

5940 0.009215 -4.68697

6600 0.008882 -4.72368

7260 0.008348 -4.78577

T= 50 °C

Time (s) [2-3] (M) ln [2-3]

0 0.01388 -4.27728

410 0.012165 -4.40916

820 0.010373 -4.5685

1230 0.008979 -4.71291

1640 0.007822 -4.85085

2050 0.006936 -4.97109

2460 0.006109 -5.09804

2870 0.005332 -5.23397

3280 0.004821 -5.33469

3690 0.004249 -5.46103

4100 0.003798 -5.57328

4510 0.003339 -5.70221

T= 40 °C

Time (s) [2-3] (M) ln [2-3]

0 0.014139 -4.25885

360 0.013361 -4.31541

720 0.012457 -4.38547

1080 0.011799 -4.43975

1440 0.010973 -4.5123

1800 0.010251 -4.58041

2160 0.009599 -4.64607

2520 0.009173 -4.69153

2880 0.008653 -4.74981

3240 0.008031 -4.82442

3600 0.007538 -4.88786

3960 0.007274 -4.92341

T= 60 °C

Time (s) [2-3] (M) ln [2-3]

0 0.013776 -4.28486

240 0.012011 -4.42194

480 0.010356 -4.57017

720 0.008981 -4.71265

960 0.007934 -4.83657

1200 0.007054 -4.95416

1440 0.006217 -5.08042

1680 0.00564 -5.17784

1920 0.005025 -5.29331

2160 0.003911 -5.54385

2400 0.003709 -5.59696

2640 0.0034 -5.68391

Figure 2-20. Plots of ln(k/T) vs. (1/T) from the data in the table at right.

2.6.4 Computational Details.

Calculations were performed with the Gaussian09 suite37 running on Intel X5670 2.93 GHz processors located at the Research Computing and Cyberinfrastructure Unit of Information

Technology Services at the Pennsylvania State University. We employed a series of truncated models of compound 2-3 in which the –C(CH3)3 groups have been omitted (Figure S3). All

48 structures were fully optimized at the B3LYP/6-311+G* level of theory. Stationary points were characterized by frequency analysis at the same level to ensure local minima for intermediates and first-order saddle points for transition structures. The imaginary frequency of the located transition structure was then animated using Gaussview to ensure that it corresponded with the desired reaction coordinate. An IRC calculation further established that the transition structure 2-3b connects to minima for 2-3a.

Figure 2-21. Summary of computational results. Relative electronic energies for stationary points are noted in bold (zero-point corrected energies in parentheses).

A geometry optimization identifies a Cs-symmetric distorted square pyramidal structure as the local minimum for model dihydridophosphorane 2-3a, in accord with the crystallographic observation. An ascent in symmetry locates a trigonal bipyramidal structure (2-3b) as a first-order saddle point on the potential energy surface. However, this C2v-symmetric stationary point, corresponding to a transition structure along the –PH2 rocking coordinate for 2-3a, resides at just

+0.7 kcal/mol (+0.2 kcal/mol). The exceptionally low barrier to interconversion of the two hydrogen positions via the pathway 2-3a2-3b2-3a is therefore consistent with the observed equivalence at ambient temperature of the two P–H hydrides on the NMR time scale.

2.6.5 Crystallographic Details.

A colorless block shaped crystal with approximate dimensions 0.15 x 0.18 x 0.22 mm was used for the X-ray crystallographic analysis. The X-ray intensity data were measured at 173(2) K, cooled by Rigaku-MSC X-Stream 2000, on a Bruker SMART APEX CCD area detector system equipped with a graphite monochromator and a MoKα fine-focus sealed tube (l = 0.71073Å) operated at 1600 watts power (50 kV, 32 mA). The detector was placed at a distance of 5.8 cm from the crystal. A total of 1850 frames were collected with a scan width of 0.3º in w and an exposure time of 10 seconds/frame. The total data collection time was about 8 hours. The frames were integrated with the Bruker SAINT software package using a narrow-frame integration algorithm. The integration of the data using a Orthorhombic unit cell yielded a total of 21920 reflections to a maximum q angle of 28.22° (0.90 Å resolution), of which 3609 were independent, completeness = 99.8%, Rint = 0.0448, Rsig = 0.0331 and 2954 were greater than 2 (I). The final cell constants: a = 11.5987(18)Å, b = 10.7814(16)Å, c = 23.414(4)Å, α = 90°, β = 90°, γ = 90°, volume = 2927.9(8)Å3, are based upon the refinement of the XYZ-centroids of 3944 reflections above 20 (I) with 2.472° < θ <25.243°. Analysis of the data showed negligible decay during data collection. Data were corrected for absorption effects using the multiscan technique (SADABS).

The ratio of minimum to maximum apparent transmission was 0.8464. The structure was solved and refined using the Bruker SHELXTL (Version 6.1) Software Package, using the space group

Pbca, with Z = 8 for the formula unit, C12H22NO2P. All non-hydrogen atoms were refined using anisotropic displacement parameters. Hydrogen atoms attached to the phosphorus atom were located on ΔF maps, and refined using isotropic displacement parameters. The remaining hydrogen atoms were fixed at calculated geometric positions and allowed to ride on the corresponding carbon atoms. The final anisotropic full-matrix least-squares refinement on F2 with 157 variables converged at R1 = 5.88%, for the observed data and wR2 = 13.82% for all data. The goodness-of- fit was 1.118. The largest peak on the final difference map was 0.419 e-/Å3 and the largest hole

50 was -0.240 e-/Å3. Based on the final model, the calculated density of the crystal is 1.104 g/cm3 and F(000) amounts to 1056 electrons.

2.6.6 NMR Spectra

2.7 References

1 Dunn, N. L., Ha, M., and Radosevich, A. T. J. Am. Chem. Soc., 2012, 134, 11330.

2 Tsuji, J. Transition Metal Reagents and Catalysts: Innovations in Organic Synthesis. Wiley:

Chichester, 2000.

3 Ramirez, F. Acc. Chem. Res. 1968, 1, 168.

4 Finet, J.-P. Ligand Coupling Reactions with Heteroatomic Compounds. Elsevier: Oxford, 1998.

5 Burgada, R. Phosphorus and Sulfur 1976, 2, 237.

6 Hoffman, R.; Howell, J. M.; Muetterties, E. L. J. Am. Chem. Soc. 1972, 94, 3047.

7 (a) Culley, S. A.; Arduengo, A. J. J. Am. Chem. Soc. 1984, 106, 1164. (b) Arduengo, A. J.; Stewart,

C. A.; Davidson, F.; Dixon, D. A.; Becker, J. Y.; Culley, S. A.; Mizen, M. B. J. Am. Chem. Soc.

1987, 109, 627. (c) Arduengo, A. J.; Stewart, C. A. Chem. Rev. 1994, 94, 1215.

8 Hartwig, J. Organotransition Metal Chemistry: From Bonding to Catalysis; University Science:

Sausalito, 2010, p 266.

9 Akiba, K.-Y. Hypervalent Compounds. In Chemistry of Hypervalent Compounds; Akiba, K.-Y.,

Ed.; Wiley: New York; 1999; p 4.

10 It should be noted that Arduengo and co-workers have evidence to suggest that 2-1 is a 10-P-3 compound in the ground state, with electron donation from the ligand to phosphorus. However, we believe the observed reactivity is better predicted by the 8-P-3 electronic arrangement; additionally the 2-elelctron reduction of phosphorus is not observed in five-coordinate analogues.

See: Arduengo, A. J., III; Dixon, D. A. Electron Rich Bonding at Low

Coordination Main Group Element Centers. In Heteroatom

Chemistry: ICHAC-2; Block, E., Ed.; VCH New York, 1990, p47.

11 Anslyn, E.V.; Dougherty, D.A. Modern Physical Organic Chemistry. University Science:

Sausalito, 2006.

12 Holmes, R. R. Pentacoordinated Phosphoranes Vol. 1: Structure and Spectroscopy. ACS

Monograph 175; American Chemical Society: Washington, DC, 1980.

13 McCarthy, S. M.; Lin, Y. C.; Davarajan, D.; Change, J. W.; Yennawar, H. P.; Rioux, R. M.; Ess,

D. H.; Radosevich, A. T. J. Am. Chem. Soc., 2014, 136, 4640.

14 Curtis, C. J.; Miedaner, A.; Ellis, W. W.; DuBois, D. L. J. Am. Chem. Soc. 2002, 124, 1918.

15 Ram, S.; Ehrenkaufer, R. E., Synthesis 1988, 91.

16 Staubitz, A.; Robertson, A. P. M.; Manners, I. Chem. Rev. 2010, 110, 4079.

17 Hu, M. G.; Geanangel, R. A.; Wendlandt, W. W. Thermochim. Acta. 1978, 23, 249.

18 Wang, J. S.; Geanangel, R. A. Inorg. Chim. Acta 1988, 148, 185.

19 Thome, I.; Nijs, A.; Bolm, C. Chem. Soc. Rev. 2012, 41, 979.

20 Gorenstein, D.G. Phosphorus-31 NMR: Principles and Applications. Academic Press: Orlando,

1984.

21 Germa, H.; Willson, M.; Burgada, R. C.R. Acad. Sci. Paris C 1970, 270, 1426.

22 Grechkin, N. P.; Shagidullin, R .R.; Gubanova, G. S. Bull. Acad. Sci. USSR 1968, 17, 1700

23 Nakafuji, S.-Y.; Kobayashi, J.; Kawashima, T.; Schmidt, M. W. Inorg. Chem. 2005, 44, 6500.

24 (a) Kobayashi, J.; Goto, K.; Kawashima, T.; Schmidt, M. W.; Nagase, S. Chem. Eur. J. 2006 12,

3811. (b) Contreras, R.; Houlla, D; Klabe, A.; Wolf, R. Tetrahedron Lett. 1981 22, 3953 (c)

Granoth, I.; Martin, J.C. J. Am. Chem. Soc. 1979, 101, 4623.

25 Pearson, R.G.; Sobel, H.; Songstad, J. J. Am. Chem. Soc. 1968, 90, 319.

26 Fourmy, K; Mallet-Ladeira, S.; Dechy-Caberet, O.; Gouygou, M. Organometallics 2013, 32,

1571.

27 (a) Bouvier, F.; Dupart, J. M.; Riess, J. G. Inorg. Chem. 1988, 27, 427. (b) Dupart, J. M.;

LeBorgne, G.; Pace, S.; Riess, J. G. J. Am. Chem. Soc. 1985, 107, 1202. (c) Vannoorenberge, Y.;

Buono, G. J. Am. Chem. Soc. 1990, 112, 6142.

28 Burgada, R.; Laurence, C. J. Organomet. Chem. 1974, 66, 255.

29 (a) Contreras, R.; Murillo, A.; Uribe, G.; Klaebe, A. Heterocycles 1985, 23, 2187 (b) Granoth,

I.; Martin, J. C. J. Am. Chem. Soc. 1979, 101, 4623.

30 Granoth, I.; Martin, J. C. J. Am. Chem. Soc. 1979, 101, 4623.

31 No hydridophosphorane species is observed upon the reaction of P(NMe2)3 with 2-2. Instead, we believe that P(NMe2)3 acts as a nucleophilic promoter, forming a zwitterion with 2-4 to initiate hydrogen transfer.

32 Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers, F. J. Organometallics

1996, 15, 1518.

33 Alaimo, P. J.; Peters, D. W.; Arnold, J.; Bergman, R. G. J. Chem. Ed. 2001, 78, 64.

34 Zhdanko, A. G.; Gulevich, A. V.; Nenajdenko, V. G. Tetrahedron 2009, 65, 4692.

35 Ramachandra, P. V.; Gagare, P. D. Inorg. Chem. 2007, 46, 7810.

36 Puhakainen, K.; Stoyanov, E.; Evans, M.J.; Leinenweber, K.; Haussermann, U. Journal of Solid

State Chemistry 2010, 183, 1785.

37 Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;

Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.;

Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.;

Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;

Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers,

E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.;

Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.;

Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;

Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski,

V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.;

Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision A.1. Gaussian, Inc.,

Wallingford CT, 2009.

Chapter 3

Mechanistic Studies of Hydrogen Transfer to a P(III) Center from Ammonia Borane

Abstract

In Chapter 2, the transfer of hydrogen from ammonia borane (2-2, AB) to three-coordinate phosphorus species 2-1 was described. This system is unique, as an equivalent of hydrogen is transferred to a single non-metal center, increasing both the coordination number and oxidation state of the central phosphorus atom by two. While this general reaction type is well-represented at transition metal centers, there are few examples at non-metals. Next, we wished to interrogate the mechanism of hydrogen transfer from AB to 2-1, especially as a means of extending this unique reactivity with other hydrogen sources. Herein we describe that the overall transfer of a dihydrogen equivalent to phosphorus in 2-1 evolves via a rapid proton and hydride addition.

3.1 Introduction

3.1.1 Importance of Hydrogenation

Hydrogenation is an essential reaction in the toolbox of synthetic organic chemists.

1 2 Molecular hydrogen (H2) has a pKa of 35, and a bond dissociation enthalpy of 104 kcal/mol; these values make both homolytic and heterolytic activation of dihydrogen particularly challenging.

While hydrogenation of alkenes is an exothermic reaction, the activation barrier is prohibitively high without a catalyst.3 However, an array of transition metal catalysts allow for hydrogenation reactions to proceed easily, often at room temperatures and under an atmosphere of hydrogen gas

(Figure 3-1).4 Hydrogenation reactions can also be performed with other hydrogen sources, such as Hantzsch esters, isopropanol, or formic acid.5 These synthetic equivalents of hydrogen provide

59 access to H2 without requiring the handling of gaseous reagents. While transition metal catalyzed hydrogenations are widely used in organic synthesis, alternative methods of hydrogenation are sought due to the scarcity, toxicity, and cost of transition metal catalysts.

Figure 3-1. Catalyzed hydrogenation of alkenes.

3.1.2 Hydrogen Addition to a Metal Center

Hydrogen activation is a field that has long been dominated by transition metal catalysis.

While there are several mechanisms by which metals activate hydrogen, here we will focus on hydrogen activation at a central atom via oxidative addition. In these cases, hydrogen initially interacts with metals primarily via donation from the H-H σ bond to an unfilled d orbital on the metal center (Figure 3-2A).6 Kubas has notably isolated and characterized stable metal-dihydrogen complexes arising from this interaction.7 To affect bond cleavage and form a metal dihydride, donation from a filled metal d-orbital to the H-H σ* bond is required (Figure 3-2B). Vaska’s complex 3-1 is an example where a dihydrogen complex is not observed; instead, dihydride 3-2 is formed (Figure 3-2C).8 In order to undergo an oxidative addition reaction with hydrogen, transition metals require also two vacant coordination sites, as well as a vacant and filled orbital.

3.1.3 Hydrogen Addition at a Non-metal Center

While there are several prominent non-metal systems that have also been shown to activate hydrogen, such as germynes9 and frustrated Lewis pairs,10 only a handful of examples exhibit the addition of hydrogen to a single central atom. In initial reports, reactivity was observed only with extremely unstable species, including gallium,11 aluminum,12 and triplet carbene compounds,13 which could only be isolated and reacted within cryogenic matrices. Recently and notably, singlet alkylaminocarbenes14 and heavier germylene analogues15 have been found to react with hydrogen under more accessible reaction conditions.

The orbital configuration of singlet germylenes and carbenes bears a resemblance to transition metal orbitals, with an electron pair perpendicular to an empty p-orbital (Figure 3-3A).

Due to this orbital similarity, activation of hydrogen by germylenes is computed to first require hydrogen donation into the p orbital, similar to transition metal activations. Conversely, computational studies of hydrogen activation by carbenes suggest a distinct mechanism: the carbene acts primarily as a nucleophile, donating electron density from the lone pair to H-H σ*.

This donation creates a transient hydride species, which attacks the empty p orbital, although the activation process is largely concerted. Calculations also suggest that alkylaminocarbenes such as

61 3-3 are particularly suitable to this form of activation due to their high energy lone pair, and low energy p orbital (Figure 3-3B).

Figure 3-3. A) Orbital overlap between a singlet carbine and H2. B) Hydrogen activation at an alkylaminocarbene.

3.1.4 Hydrogen activation at Phosphorus

The phosphorus atom in 2-1 exhibits some similarities to both carbenes and transition metal centers. As detailed in Chapter 2, Alexander Radosevich has calculated that phosphorus in this compound has both a high energy lone pair and an orthogonal nonbonding p orbital, the qualitative arrangement of which resembles both transition metal and main group compounds. As further shown in Chapter 2, phosphorus compound 2-1 is capable of accepting hydrogen from AB (2-2),16 a unique hydrogen source that contains protic and hydridic hydrogens, but which is a stable solid under atmospheric conditions.17 Herein we wish to elucidate the mechanism of hydrogen transfer to phosphorus compound 2-1 from AB (Figure 3-4).

62 Figure 3-4. Hydrogen transfer from AB to 2-1.

3.2 Stoichiometric conversion of 2-1 to 2-3: Hydrogen Transfer from Ammonia Borane

In order to better understand the transformation shown in Figure 3-4, we probed the reaction with a variety of kinetic experiments to determine the rate equation of the reaction, the activation parameters, and the kinetic isotope effects. In addition, we investigated the reaction through labeling studies.

3.2.1 Kinetics of Hydrogen Transfer from Ammonia-Borane to 2-1.

To begin the investigation into the formation of 2-3, kinetic studies of the reaction of 2-1 and excess ammonia-borane (20 equiv relative to 2-1) were conducted in d8-THF at 60 °C. The

1 concentrations of 2-1 and 2-3 was determined by H NMR integration of the vinylic protons of each compound. Due to the large excess of ammonia borane, [AB] was considered to be constant throughout the course of the reaction, therefore rate will depend only upon the concentration of 2-

1.18 A plot of ln[2-1] vs. time is linear over three half-lives, indicating the reaction is first order in

2-1 (Figure 3-5).

63 -2 0 2000 4000 6000 8000 10000 12000 14000 -2.5

-3

-3.5

]

- 2

-4 ln[

-4.5

-5

-5.5

-6 time (s)

Figure 3-5. Plot of ln[2-1] vs. time, showing that the consumption of 2-1 is first order in 2-1. R2 = 0.994.

To determine the order in AB, reaction rates were measured as the concentration of AB was varied from 0.07-1.55 M in AB.19 Above this concentration limit, an uncatalyzed bimolecular decomposition of AB via hydrogen loss complicates the kinetic analysis of the reaction. The plot of kobs as a function of the concentration of ammonia-borane (Figure 3-6) is linear, indicating that the reaction is first order in ammonia-borane. The rate constant k is obtained from the slope of this plot, such that k=1.49E-04 L*mol-1*s-1 and a rate law for the formation of 2-3 from 2-1 and ammonia-borane that is consistent with the data in Figures 5 and 6 is given in Eq. 3-1.

Eq. 3-1

rate = 1.49E-04 [2-1] [AB]

64 0.00025

0.0002

0.00015

)

1 -

(s 0.0001

obs obs k

0.00005

0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

-0.00005 [Ammonia Borane](M)

A kinetic analysis was also attempted using superstoichiometric equivalents of 2-1 relative to AB; the rate was measured by monitoring [AB] vs time with 11B NMR spectroscopy.

Unfortunately, the reaction kinetics could only be monitored between 1-5 equivalents of 2-1. The narrow range of relative concentrations makes it difficult to accurately determine the order of either reactant in this system; however, qualitatively the data suggests that the reaction rate depends on the concentration of both AB and 2-1.

3.2.2 Eyring Analysis

An Eyring analysis20 of hydrogen transfer from AB to 2-1 was performed by analyzing reaction rates between 40-60 °C in d8-THF (Figure 3-7). A linear regression of the resulting Eyring plot provides activation parameters of the reaction such that ΔH‡ = 27.5 (±2.6) kcal/mol and ΔS‡ =

6.1 (±8.2) cal/mol*K. These values indicate a moderate activation enthalpy, and a small increase in entropy at the transition state.

-12 0.00295 0.003 0.00305 0.0031 0.00315 0.0032 0.00325 -13

-14

-15 ln(k/T) -16

-17

-18 1/T (1/K)

Figure 3-7. Eyring plot of hydrogen transfer from AB to 2-1 between 40-60 °C.

3.2.3 Kinetic Isotope Effects.

Isotope effects were determined for the reaction of 2-1 and the isotopologues of ammonia borane (Section 2.3.5). The results show a strong primary kinetic isotope effect for the N-H bonds of ammonia borane at 4.3 ± 1.1 (Figure 3-8), indicating that the N-H bond is breaking during the transition state. A smaller effect for substitution of the B-H bonds at 2.2 ± 0.7.21 The effect for substituting both nitrogen and boron positions with deuterium closely matches the effects for substitution on nitrogen alone, at 4.2 ± 1.0. The combined effects likely indicate a stepwise hydrogen transfer, with N-H bonds reacting first.

Figure 3-8. Kinetic isotope effects are observed for isotopic substitution for both N-H and B-H during hydrogen transfer from AB to 2-1.

It is notable that no scrambling of deuterium at boron and nitrogen was observed under

2 reaction conditions via H NMR. This result was determined by monitoring ND3BH3 after heating the compound for 4 hours at 60 °C either alone or in the presence of 5 mol% 2-1 or 2-3. While we know that there is typically an exchange between N-H and P-H atoms that should facilitate isotopic exchange at nitrogen (see section 2.3.5), these results indicate that: 1) due to the large excess of d3-

AB, the exchange is negligible and below detectable limits for 2H NMR, and 2) deuterium scrambling is not responsible for the observed KIEs.

Figure 3-9. Isotopic scrambling between N-H and B-H at AB in the presence of 2-1 or 2-3 is not observed by 2H NMR.

67 3.2.4 Isotopic Labeling at 2-3

3 In Section 2.3.4, the potential isomerization of 2-3 to a σ species like 3-5 was investigated,

5 and it was determined that 2-3 remains in a σ state under all conditions examined. However, the results that we obtained do not address the possibility that the reaction of 2-1 with AB could proceed through a σ3 species. A mechanism wherein hydrogen would be first transferred across a P-O bond to form 3-5 has been suggested computationally (Figure 3-10A).22,22 To address this mechanistic

2 possibility, we have treated 2-1 with ND3BH3 to provide 2-3-d2 and monitored the reaction via H

NMR (Figure 3-10B). The formation of 3-5 should lead to rapid scrambling of hydrogen between oxygen and carbon positions due to enol-keto tautomerization, which has been computed in a similar system to have a lower barrier than the final step of Sakaki’s mechanism.23 2H NMR spectra show that deuterium is only observed at phosphorus, not at carbon, which indicates that either 1) hydrogen is delivered directly to phosphorus, or 2) tautomerization of 3-5 to 3-7 does not occur.

Figure 3-10. A) The route proposed by Sakaki for hydrogen transfer to 2-1. B) σ3-σ5 isomerization to form 2-7 is not observed by 2H NMR.

68 3.2.5 11B NMR Monitoring.

As reported in the literature, heating a solution of AB in the absence of additives produces

18 H2 and various polymeric N-B derivatives. To determine if these polymerization reactions are required for hydrogen transfer to 2-1, the reaction was monitored by 11B NMR. After heating AB alone (0.8 M) for 24 h at 60 °C in THF, primarily AB remains (δ –21 ppm); however, B- cyclodiborazanylaminoborohydride (3-8, δ -5, -12, -24 ppm) and cyclotriborazane (3-9, δ –12 ppm) are observed, along with traces of polyborazylene (3-10, δ 28 ppm) and borazine (3-11, δ 32 ppm)

(Figure 3-11A).

The addition of 1 equivalent of 2-1 in a parallel experiment leads to the formation of significant precipitate, while heating AB in THF alone causes sparse white precipitate. The reaction with 2-1 also alters the speciation of AB polymers. Similar to the solution of AB alone, heating AB in the presence of 2-1 leaves some unreacted AB; 3-9 and 3-10 are also observed

(Figure 3-11B). However, new signals from diammoniate of diborane (3-12, –38 ppm) as well as an unidentified species at 10 ppm in the 11B NMR spectra. This observation indicates that phosphorus species 2-1 affects the nature of dehydrogenation products from AB, likely by directly mediating dehydrogenation.

The addition of cyclohexene to a solution of AB prevents the formation of AB polymers

24 by trapping aminoborane (NH2BH2) as dicyclohexylborane amine (NH2BCy2). In our experiment, 1 equivalent of 2-1 was reacted with 20 equivalents of AB in the presence of 200

11 equivalents of cyclohexene. B NMR confirms that only AB, NH2BCy2 and a small amount of

NH2BHCy are observed; however, 2-1 is still converted efficiently to 2-3. This observation indicates that hydrogen transfer comes from AB itself, and is not dependent upon its polymeric derivatives, a conclusion that is supported by the reaction’s first order dependence upon AB.

Figure 3-11. A) 11B NMR of AB heated at 60 °C for 24 h. B) 11B NMR of AB and one equiv 2-1 heated at 60 °C for 24 h. C) Boron species observed.

70 3.3 Reactions of 2-1 with other Hydrogen Sources

We investigated the ability of other hydrogen sources to transfer hydrogen to 2-1 and subsequently form 2-3. We have examined both polar and nonpolar hydrogen sources, with a focus on substituted amine-boranes. AB was found to have unique properties which allow for hydrogen transfer to 2-1.

3.3.1 Reaction of 2-1 with Substituted Amine Boranes

To determine whether the reaction of 2-1 with AB is general to amine boranes, we studied the reactivity of 2-1 with a suite of amine boranes substituted at both boron and nitrogen (Figure 3-

12). We first synthesized NH3BPh3, which was designed to contain only protic hydrogens.

Unsurprisingly, treating 2-1 with 20 equivalents of NH3BPh3 does not result in hydrogen transfer, and in fact no reaction is observed.

We next examined substitution at nitrogen, and reacted 2-1 with Me2NHBH3 and t-

BuNH2BH3. In both cases, no reaction was observed. Concerned that the steric bulk at nitrogen in these species was impeding reactivity, we synthesized n-BuNH2BH3 to minimize sterics and better mimic AB; however, again no reaction with 2-1 was observed. The N-alkylated analogues of AB are often described as having similar hydrogen release reactivity as AB and better solubility profiles. However, there are several differences which may arise due to alkylation. The addition of alkyl groups to amines increases their basicity, which increases the strength of the B-N dative bond;25 however, additional steric bulk also has been found to decrease B-N bond strength.26

Additionally, while the pKa values of most amine boranes are not reported, comparison to simple

27 amines indicates that increasing alkyl substitution on nitrogen leads to an increase in pKa. These characteristics could impede hydrogen transfer to 2-1.

71 Figure 3-12. Polar hydrogen sources for transfer hydrogenation.

3.3.2 Reaction of 2-1 with Nonpolar Hydrogen Sources

Reacting 2-1 with common nonpolar organic hydrogen transfer reagents was similarly unsuccessful. 1,2,3,4-tetrahydroquinoline (3-13), 1,2-dihydroquinoline (3-14), 2,3-dihydrofuran

(3-15), 2,3-dihydroindole (3-16), and Hantzch ester 3-17 (Figure 3-13) were used as hydrogen sources; in each case the loss of hydrogen from the cyclic systems would lead to aromatization and provide a driving force for hydrogen transfer. It is notable that dihydroquinoline 3-14 was so unstable that it disproportionated to quinoline and tetrahydroquinoline 3-13 under reaction conditions; despite this result, hydrogen was not transferred to 2-1.

Figure 3-13. Nonpolar hydrogen sources for transfer hydrogenation.

3.4 Mechanistic Discussion

There are several mechanisms that can be envisioned for hydrogen transfer from AB to 2-

1 (Figure 3-4). It is well documented that AB releases hydrogen gas upon heating; however, we have previous shown that 2-1 does not react with H2. We will therefore focus here on other possible reactivity of AB. The polarity of AB means that it contains both protic and hydridic hydrogen atoms. In addition, the nature of the coordinative bond between N and B means that it is also feasible that AB could dissociate in solution, providing a source of ammonia or borane (Figure 3-

14).28 The theoretical activation energy of B-N dissociation is 25.9 kcal/mol, similar to the

‡ 29 calculated barrier of the 2-1 to 2-3 transformation (ΔG = 30.6 kcal/mol at 60 °C).

Figure 3-14. Dissociation of AB.

With the reactivity of AB in mind, we have proposed six initial mechanistic possibilities for hydrogen transfer to 2-1 (Figure 3-15). Compound 2-1 could activate a B-H bond on borane or

AB in an inner sphere mechanism (Path A). Similarly, 2-1 could activate an N-H bond from ammonia or AB (Path B). Two concerted mechanisms can be envisioned, with hydrogen transfer from AB occurring either across a P-O bond (Path C) or directly to the phosphorus atom (Path D).

Finally, we consider that initial addition of hydride (Path E) or proton (Path F) at phosphorus could occur. Each mechanism will be examined in detail.

Kinetic studies indicate that the rate determining step is first order in both AB and 2-1. The reaction has a moderate enthalpic barrier and a positive activation entropy, which indicates increased disorder in a rate determining step. The positive activation entropy could be caused by bond lengthening, breaking, or solvent reorganization predominating the entropic factor.30 Isotopic substitution at AB shows a large KIE at nitrogen, a smaller one at boron, and that substitution at both atoms gives an KIE equal to nitrogen substitution alone. These results indicate a stepwise mechanism, which includes a change in RDS dependent upon the isotopologue, as well as an initial

N-H activation.

Figure 3-15. Isotopic scrambling between N-H and B-H at AB in the presence of 2-1 or 2-3 is not observed by 2H NMR.

Path A: B-H Activation: Of the collected experimental data for the RDS of hydrogen transfer, the positive activation entropy may indicate a dissociation of ammonia and borane during or before the RDS. However, attempts to react 2-1 with B-H bonds have proven largely unsuccessful, and oxidative addition of B-H at P is computed to have a high barrier.21 The reaction

74 of 2-1 with BH3•THF leads only to decomposition, so we do not believe B-H activation to be operative in the hydrogen transfer reaction.

Path B: N-H Activation: Our group’s previous report has shown that complex 2-1 is capable of activating the N-H bond of amines.31 In fact, 2-1 reacts rapidly with ammonia at low temperatures. With this in mind, initial B-N dissociation followed by N-H activation of ammonia could initiate hydrogen transfer (Figure 3-14). As stated above, dissociation of ammonia from AB could be indicated by the positive activation entropy, additionally KIE indicate that N-H activation is involved in the RDS. To further probe this possibility, the ability of 3-24 to transform into 2-3 was examined.

Treatment of 3-24 with a variety of hydride donors and boranes (LiBH4, LAH, DIBAL, catechol borane, dicyclohexyl borane) did not affect the desired transformation to 2-3. On the other hand, the reaction of 3-24 with an excess of ammonia borane does form 2-3 (Figure 3-16).

However, this transformation proceeds to only 20% conversion in 24 h, where 2-1 exposed to identical conditions is converted quantitatively in 6 hours. In addition, 2-19 is observed if a reaction of 2-1 and AB is heated for an additional 12 hours after 2-1 is completely converted to 2-3, along with other unidentified side products. These results indicate an off-path equilibrium between the two species, and do not support 3-24 being an intermediate in the direct conversion between 2-1 and 2-3.

Figure 3-16. Reaction of 2-1 with ammonia, followed by reaction of 3-24 with AB.

While a similar N-H bond activation at AB to form 3-25 can be envisioned (Figure 3-17), activation of ammonia with 2-1 exhibits a third order dependency on NH3, and the basicity of NH3 is an important factor. The kinetics do not support a similar mechanism in this system, and AB is

75 a poor Lewis base in comparison. Additionally, 2-1 does not react with AB analogues with substitution at boron, such as H3NBPh3. Initial N-H activation of AB is therefore not considered a plausible mechanism.

Figure 3-17. N-H activation of AB with 2-1, followed by reaction of 3-25 with AB.

Path C: 1,2-Addition: Computations have suggested that hydrogen could be delivered across the P-O bond, forming 3-20, a three-coordinate phosphorus hydride with a pendant enol, which would then rearrange to yield the observed 2-3.21,22 We have detailed in Section 2.3.4 that

32 no evidence of the 3-coordinate tautomer of 2-3 has been observed. Our isotopic labeling studies

(Section 3.2.4) further indicate that 3 to 5 coordinate isomerism is not operative during or after hydrogen transfer from AB to 2-1 (Figure 3-4). These results lead us to disfavor an initial hydrogen transfer across the P-O bond.

Path D: Concerted Mechanism: A double hydrogen transfer from AB to 2-1 can be envisioned. Such a transfer would proceed through a concerted transition state, which would display a multiplicative KIE. Our kinetic isotope effects indicate a stepwise reaction instead, thus eliminating this concerted pathway.

Path E: Hydride Delivery: Since AB can also be a hydride source, we tested this proposed pathway by treating 2-1 with various hydride sources. Strong donors (e.g., LAH) cause rapid decomposition, while weaker hydride sources (e.g., LiAlH(OtBu)3, NaH) do not react with 2-1 at room temperature. A stable intermediate was not observed in any of these experiments.

However, treating AB with catalytic lithium hexamethyldisilazide (LiHMDS), then adding

2-1 leads to clean conversion to 2-3 at room temperature (Figure 3-18A) (in comparison, treatment of 2-1 with AB in the absence of a strong base requires heating to 40 °C to induce hydrogen

76 transfer). The use of other bases, such as methylmagnesium bromide and sodium hydride, in the presence of AB also provides room temperature conversion to 2-3, although side products are also observed with these bases. Using a full equivalent of LiHMDS leads only to decomposition of all phosphorus species. This decomposition suggests that treatment with hydride in the presence of a proton source is key to optimal reactivity. LiHMDS presumably deprotonates AB and forms

- anionic NH2BH3 3-26 in situ; this anion then delivers hydride to 2-1, which in turn abstracts a

- proton from an AB complex to reform NH2BH3 (Figure 3-18B). Mimicking such conditions with alternative sources of both proton and hydride is challenging, as the rapid release of hydrogen precludes any reactivity with 2-1.

Figure 3-18. A) Formation of 2-3 from 2-1 and AB is promoted by catalytic base. B) Proposed reaction sequence.

Path F: Proton Delivery: 2-1 has demonstrated a basicity similar to amines in THF, and is partially protonated in the presence of 2,6-dimethylpyridine hydrochloride, which has a pKa value of 7.2 in THF. Treatment of 2-1 with HCl sources provides only the C-protonated species 3-27.

This result is in agreement with the previously reported reaction of 2-1 with triflic acid.33

Investigations of protonated species 3-27 show that no transformation to 2-3 occurs with hydride or B-H sources such as pinacolborane, 9-borabicyclononane, catecholborane, sodium borohydride,

77 sodium cyanoborohydride, sodium hydride, lithium aluminum hydride, lithium tris(t- butoxy)aluminum hydride, or triethylsilane (Figure 3-19).

Figure 3-19. Initial proton transfer to 2-1 forms 3-27, which is not transformed into 2-3.

34 Ammonia borane is deprotonated in the presence of proton sponge, which has a pKa value of 11.1 in THF. We have evidence of 2-1 abstracting a proton from ammonium species with a pKa as high as 7.3 in THF. In light of these values, equilibrium proton transfer may occur between 2-1

- and AB at elevated temperatures, forming anionic 3-26 ( H2NBH3), which has been shown to be a significantly stronger hydride donor than AB.35 This theory is supported by the rapid room temperature conversion of 2-1 to 2-3 in the presence of AB and strong bases. While we have shown that C-protonated 3-27 cannot be converted to 2-3, we cannot say that a transient reactive species is not formed under the reaction conditions (any conversion to 3-27 could be reverted back to 2-1 by the action of a weak base, including 3-26). In addition, proton transfer from N followed by hydride transfer from B is consistent with the observed KIE’s at AB. In comparison, this mode of activation draws upon common reactivity at phosphorus: a phosphorus(III) compound forms a bond with an electrophile (in this case, a proton), and then a nucleophile (hydride) attacks the cationic intermediate. The experimental evidence best supports a mechanism of proton then hydride transfer from AB to 2-1.

Figure 3-20. Proposed mechanism of hydrogen transfer from AB to 2-1. An equilibrium proton transfer forms complex 3-23, followed by hydride transfer to form 2-3.

3.5 Conclusions

We have previously reported that T-shaped, planar phosphorus (III) compound 2-1 is transformed to dihydridophosphorane 2-3 in the presence of the hydrogen source AB. This transformation adds to a small body of organic compounds which accept hydrogen at a single nonmetal center, similar to oxidative addition of hydrogen at transition metals. A thorough study of the transition from 2-1 to 2-3 has been completed, and has shown that hydrogen transfer to 2-1 is highly dependent upon the hydrogen source. AB appears to be ideal due to its ability to act as both proton and hydride donor upon heating while remaining stable under ambient conditions. We believe that the ability of a phosphorus compound to react with discrete electrophilic and nucleophilic reagents in sequence provides a broad opportunity to expand catalysis using P(III) and

P(V) species. In fact, a method of intramolecular reductions at phosphorus using this stepwise activation method will be fully described in Chapter 5.

3.6 Experimental

3.6.1 General Materials and Methods.

Phosphorous (V) chloride was sublimed prior to use. Diethyl ether (Et2O), methylene chloride

36 (CH2Cl2), tetrahydrofuran (THF), and toluene were dried according to the method of Grubbs as

79 modified by Bergman.37 All other commercially available reagents were purchased from suppliers and used without further purification. All reactions were carried out under nitrogen either in a double port glovebox (Innovative Technology) or with a Schlenk manifold vented through an oil bubbler unless otherwise noted. All glassware was oven-dried at 120 °C prior to use. 1H, 13C, and

31P NMR spectra were recorded with Bruker DPX-300, AV-360, and DRX-400 spectrometers. 1H

NMR spectra were referenced to residual solvent peaks (CDCl3: 7.26 ppm, CD3CN: 1.94 ppm,

31 DMSO-d6 2.50 ppm). P NMR spectra were referenced to an external standard (H3PO4: 0.0 ppm).

Mass spectrometric data were obtained from the Mass Spectromotry Laboratory at the University of Illinois at Urbana-Champaign.

3.6.2 Synthetic Procedures and Data

Reaction of 2-1, AB and LiHMDS: To ammonia borane (8 mg, 0.25 mmol) in THF (1 mL) was added LiHMDS (0.1 M in THF, 0.1 mL, 0.01 mmol). After 5 minutes, 2-1 (12 mg, 0.05 mmol) was added. NMR taken immediately shows complete conversion to 2-3.

nBuNH2BH3:38 To a solution of n-butylamine (2 mL, 20 mmol) in THF at -78 C was added dropwise BH3THF (1M in THF, 20 mL, 20 mmol) over 30 minutes, the solution was allowed to warm to room temp and then the solvent was removed en vacuo to reveal a white solid which melted just above room temp. 1H NMR (CDCl3): 3.75 (2H, br), 2.8 (2H, m), 2-1 (br, BH) 1.58 (2H, m),

1.36 (2H, m), 0.92 (3H, t). 11B NMR (CDCl3): -19.7

39 NH3BPh3: To the aqueous solution of triphenylborane sodium hydroxide (7%, 10 mL,

2.65 mmol) under nitrogen was added aqueous ammonium hydroxide (30%, 20 mL, 150 mmol) at room temperature and stirred 1 h; the precipitate was collected and washed with copious water, then dissolved in ethyl acetate. Organics were mixed with sodium sulfate and celite, then filtered again and concentrated, then diluted slightly with toluene and the

80 precipitate collected and dried as a white solid (467 mg, 68%). 1H NMR (CDCl3): 3.75 (2H, br), 2.8 (2H, m), 2-1 (br, BH) 1.58 (2H, m), 1.36 (2H, m), 0.92 (3H, t). 11B NMR (CDCl3): -19.7

3.6.3 NMR Monitoring

11B NMR Monitoring

The experiments were prepared as described below under a nitrogen atmosphere and sealed in a quartz NMR tube, then monitored by 11B NMR. Control: Ammonia borane (25 mg, 0.80 mmol) was dissolved in proteo-THF (1 mL) and heated to 60 °C for 24 h. With 2-1: Ammonia borane (25 mg, 0.80 mmol) and 2-1 (10 mg, 0.04 mmol) were dissolved in proteo-THF (1 mL) and heated to

60 °C for 24 h.

With cyclohexene: Ammonia borane (25 mg, 0.80 mmol), 2-1 (10 mg, 0.04 mmol) and cyclohexene

(656 mg, 8 mmol) were dissolved in proteo-THF (0.5 mL) and heated to 60 °C for 4 h, at which time 31P NMR confirmed complete conversion of 2-1 to 2-3.

Figure 3-21: 11B NMR of 2-1, 20 equiv. AB, 200 equiv. cyclohexene, heated to 60 C for 4 h. Peak at -23 is AB, peak at 47 is Cy2BNH2, peak at -5 may be CyBHNH2.

Figure 3-22: 31P NMR of 1, 20 equiv. AB, 200 equiv. cyclohexene, heated to 60 C for 4 h. Triplet at -45 is 2-3, no 2-1 remains. 2H NMR Monitoring

Conversion of 2-1 to 2-3: A solution of 2-1 (10 mg, 0.04 mmol) and ND3BH3 (26 mg, 0.80 mmol) in 1 mL THF was sealed under a nitrogen atmosphere and heated to 60 °C for 4 h. The reaction was monitored by 2H NMR and reference externally to the 3.58 ppm peak of THF.

H-D exchange at phosphorus: A solution of 2-3 (10 mg, 0.04 mmol) and ND3BH3 (26 mg,

0.80 mmol) in 1 mL THF was sealed under a nitrogen atmosphere and heated to 60 °C for 4 h. The reaction was monitored by 2H NMR and reference externally to the 3.58 ppm peak of THF.

2 Figure 3-23: H NMR of 2-1 with ND3BH3 in THF at t=0.

2 Figure 3-24: H NMR of 2-1 with ND3BH3 in THF after 4 h, 60 °C shows only P-D signals from 1•[D]2.

2 Figure 3-25: H NMR of 2-3 with ND3BH3 in THF at t=0.

2 Figure 3-26: H NMR of 2-3 with ND3BH3 in THF after 4 h, 60 °C, which shows only P-D peaks from 1•[D]2.

84 3.6.4 Kinetics

Order in 2-1

A solution of 2-1 (10 mg, 0.04 mmol) and ammonia borane (25 mg, 0.8 mmol) in d8-THF

(0.6 mL) was heated for 3.5 h at 60 °C in a temperature controlled NMR. The spectra were referenced to the residual THF peak at 3.58 ppm and the relative concentrations of 2-1 and 2-3 were determined by integration of the vinylic peaks at between 7.71-7.42 and 5.93-5.60 ppm respectively.

85 [2-1] vs time 0.07

0.06

0.05

0.04

0.03 1] 1] (M)

- y = -5E-06x + 0.0555 [2 0.02 R² = 0.927

0.01

0 0 2000 4000 6000 8000 10000 12000 14000 -0.01 time (s)

ln[2-1] vs time 0 0 2000 4000 6000 8000 10000 12000 14000 -1

-2

-3

1] -

ln[2 -4

-5 y = -0.0003x - 2.3943 R² = 0.9938 -6

-7 time (s)

86 1/[2-1]2 vs time 120000

100000

80000

2 60000 y = 4.4474x - 17920 1] - R² = 0.581

1/[2 40000

20000

0 0 2000 4000 6000 8000 10000 12000 14000

-20000 Time (s)

Figure 3-27: A comparison of the various treatments of the concentration of 2-1 vs time shows that the best fit is ln[2-1], and the reaction is first order in 2-1.

Order in Ammonia Borane

A solution of 2-1 (10 mg, 0.04 mmol) and ammonia borane (x mmol) in d8-THF (0.6 mL) was heated for 3.5 h at 60 °C in a temperature controlled NMR. The spectra were referenced to the residual THF peak at 3.58 ppm and the relative concentrations of 2-1 and 2-3 were determined by integration of the vinylic peaks at between 7.71-7.42 and 5.93-5.60 ppm respectively.

kobs vs [AB]

5.00E-04

4.00E-04 y = 0.0002x - 0.0001 R² = 0.899 3.00E-04

obs 2.00E-04 k

1.00E-04

0.00E+00 0 0.5 1 1.5 2 2.5 -1.00E-04 [AB] (M)

Figure 3-28: kobs vs [AB] shows that the reaction is first order in AB. Equivalents k obs [AB] (M) [AB]^2 [AB]^3 of AB

5 3.02E-05 0.341667 0.116736 0.039885

7.5 5.18E-05 0.5125 0.262656 0.134611

10 7.01E-05 0.683333 0.466944 0.319079

12.5 7.64E-05 0.854167 0.729601 0.623201

15 8.68E-05 1.025 1.050625 1.076891

17.5 1.90E-04 1.195833 1.430017 1.710062

20 2.07E-04 1.366667 1.867778 2.55263

22.5 2.04E-04 1.5375 2.363906 3.634506

25 3.35E-04 1.708333 2.918403 4.985605

30 4.61E-04 2.05 4.2025 8.615125

Kinetic Isotope Effects

A solution of 2-1 (10 mg, 0.04 mmol) and ammonia borane (0.8 mmol) in d8-THF (0.6 mL) was heated for at 60 °C in a temperature controlled NMR. The spectra were referenced to the residual

THF peak at 3.58 ppm and the relative concentrations of 2-1 and 2-3 were determined by integration of the vinylic peaks at between 7.71-7.42 and 5.93-5.60 ppm respectively.

a b Species k trial 1 k trial 2 k trial 3 k avg σk KIE σKIE

NH3BH3 1.16E-04 1.81E-04 1.50E-04 1.49E-04 3.25E-05 1.00 0.31

ND3BH3 3.10E-05 3.45E-05 3.91E-05 3.49E-05 4.06E-06 4.27 1.06

c NH3BD3 6.89E-05 8.96E-05 6.61E-05 7.49E-05 1.28E-05 2.24 0.69

ND3BD3 3.67E-05 3.68E-05 3.25E-05 3.53E-05 2.45E-06 4.22 0.97 a b Where σk is the standard deviation for the three trials. σKIE is determined by error propagation

휎퐻 휎퐷 such that 휎 = 퐾퐼퐸√( ) ( ), where σH/D is the standard deviation for the proteo or deutero 퐾퐼퐸 퐻 퐷

c species and H or D is the average k for those species. The KIE for NH3BD3 is adjusted to

푘 −푘 (푥−1) account for only 90% incorporation of D at boron such that 푘 ′ = 퐷 퐻 , where x is the 퐷 푥 mole fraction of deuterium incorporation. *Note: if N and B dueteration causes an additive effect, you would expect the ND3BD3 species to have a KIE value of 5.25-7.77; multiplicative effects would give a KIE of 5.76-13.3.

Eyring Analysis of 1 and Ammonia Borane

A solution of 2-1 (10 mg, 0.04 mmol) and ammonia borane (25 mg, 0.80 mmol) in d8-THF

(0.5 mL) was heated in a temperature controlled NMR between 40-60 °C. The spectra were referenced to the residual THF peak at 3.58 ppm and the relative concentrations of 2-1 and 2-3 were determined by integration of the vinylic peaks between 7.71-7.42 and 5.93-5.60 ppm respectively.

89 Temp (K) k (s-1M-1) 1/T (1/K) ln(k/T)

313 7.78E-06 0.003195 -17.5103

318 3.27E-05 0.003145 -16.0914

323 2.26E-05 0.003096 -16.4754

328 6.01E-05 0.003049 -15.5119

333 1.71E-04 0.003003 -14.4819

-12 0.00295 0.003 0.00305 0.0031 0.00315 0.0032 0.00325 -13

-14

-15

ln(k/T) y = -13814x + 26.774 -16 R² = 0.8684

-17

-18 1/T (1/K)

3.7 References

1 Karty, J. Organic Chemistry: Principles and Mechanisms. W. W. Norton: New York, 2014, p

301.

2 Blanksby, S. J.; Ellison, G. B. Acc. Chem. Res., 2003, 36, 255.

3 McMurry, J. Organic Chemistry, 7th ed. Brooks/Cole: Belmont, 2008, p 187.

4 Harmon, R. E.; Gupta, S. K.; Brown, D. J. Chem. Rev. 1973, 73, 21.

5 Brieger, G.; Nestrick, T .J. Chem. Rev. 1974, 74, 567.

6 Hartwig, J. Organotransition Metal Chemistry: From Bonding to Catalysis. University Science:

Sausalito, CA, 2010, p 266.

7 Kubas, G. J.; Ryan, R. R.; Swanson, B. I.; Vergamini, P. J.; Wasserman, H. J. J. Am. Chem. Soc.

1984, 106, 451.

8 Vaska, L.; DiLuzio, J. W. J. Am. Chem. Soc. 1962, 84, 679.

9 Spikes, G. H.; Gettinger, J. C.; Power, P. P. J. Am. Chem. Soc. 2005, 127, 12232.

10 (a) Welch, G.C.; Stephan, D. W. J. Am. Chem. Soc. 2007, 129, 1880. (b) Welch, G. C.; San Juan,

R. R.; Masuda, J. D.; Stephan, D. W. Science 2006, 314, 1124.

11 Xiao, Z. L.; Hauge, R. H.; Margrave, J. L. Inorg. Chem. 1993, 32, 64.

12 Himmel, H. J.; Vollet, J. Organometallics 2002, 21, 5972.

13 Zuev, P. S.; Sheridan, R. S. J. Am. Chem. Soc. 2001, 123, 12434.

14 Frey, G.D.; Lavallo, V.; Donnadieu, B.; Schoeller, W.W.; Bertrand, G. Science 2007, 316, 439.

15 Peng, Y.; Guo, J. D.; Ellis, B. D.; Zhu, Z.; Fettinger, J. C.; Nagase, S.; Power, P. P. J. Am. Chem.

Soc. 2009, 131, 16272.

16 Dunn, N. L.; Ha, M., Radosevich, A. T. J. Am. Chem. Soc., 2012, 134, 11330.

17 Staubitz, A.; Robertson, A. P. M.; Manners, I. Chem. Rev. 2010, 110, 4079.

18 Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic Chemistry. University Science:

Sausalito, CA, 2005.

19 Shaw, W. J.; Linehan, J. C.; Szymczak, N. K.; Heldebrant, D. J.; Yonker, C.; Camaioni, D. M.;

Baker, R. T; Autrey, T. Angew. Chem. Int. Ed. 2008, 47, 7493.

20 Eyring, H. J. Chem. Phys. 1935, 3, 107.

21 Melander, L. C. S.; Saunders: W. Reaction Rates of Isotopic Molecules. Krieger: Malabar, FL,

1987.

22 Zeng, G.; Maeda, S.; Taketsugu, T.; Sakai, S. Angew. Chem. Int. Ed. 2014, 53, 4633.

23 Pal, A.; Vanka, K. Inorg. Chem. Article ASAP, DOI: 10.1021/acs.inorgchem.5b01074

24 Pons, V.; Baker, R. T.; Szymczak, N. K.; Heldebrant, D. J.; Linehan, J. C.; Matus, M. H.; Grant,

D. J.; Dixon, D. A. Chem. Commun. 2008, 6597.

25 Flores-Segura, H.; Torres, L. A. Structural Chem. 1996, 7, 363.

26 Flores-Segura, H.; Torres, L. A. Structural Chem. 1997, 8, 227.

27 Smith, D.A. Metabolism, Pharmacokinetics and Toxicity of Functional Groups. Royal Society of Chemistry: London, 2010, p 175.

28 Nguyen, M. T.; Nguyen, V. S.;Matus, M. H.; Gopakumar, G. Dixon, D. A. J. Phys. Chem. A

2007, 111, 679.

29 (a) Dixon, D. A.; Gutowski, M. J. Phys. Chem. A 2005, 109, 5129. (b) Plumley, J. A.; Evanseck,

J. D. J. Phys. Chem. A 2007, 111, 13472.

30 Potter, R. G.; Camaioni, D. M.; Vasiliu, M.; Dixon, D. A. Inorg. Chem. 2010, 49, 10512.

31 McCarthy, S. M.; Lin, Y. C.; Davarajan, D.; Change, J. W.; Yennawar, H. P.; Rioux, R. M.; Ess,

D. H.; Radosevich, A. T. J. Am. Chem. Soc., 2014, 136, 4640.

32 Wolf, R. Pure & Applied Chem. 1980, 52, 1141.

33 Arduengo, A. J.; Stewart, C. A.; Davidson, F.; Dixon, D. A.; Becker, J. Y.; Culley, S. A.; Mizen,

M. B. J. Am. Chem. Soc. 1987, 109, 627.

34 Himmelberger, D. W.; Yoon, C. W.; Bluhm, M. E.; Carroll, P. J.; Sneddon, L. J. J. Am. Chem.

Soc. 2009, 131, 14101.

35 Heiden, Z.M.; Lathem, A.P. Organometallics 2015, 34, 1818.

36 Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers, F. J. Organometallics

1996, 15, 1518.

37 Alaimo, P. J.; Peters, D. W.; Arnold, J.; Bergman, R. G. J. Chem. Ed. 2001, 78, 64.

38 Yamamoto, Y.; Miyamoto, K.; Umeda, J.; Nakatani, Y.; Yamamoto, T.; Miyaura, N. J.

Organomet. Chem. 2006, 691, 4909.

39 Layton, W.J.; Niedenzu, K.; Niedenzu, P.M.; Trofimenko, S. Inorg. Chem. 1985, 24, 1454.

Chapter 4

Investigations of Dihydridophosphorane 2-3 as a Hydrogen Source

Abstract

In Chapter 2, the ability of dihydridophosphorane 2-3 to transfer hydrogen to azobenzene was detailed. Here we present a further investigation of the reactivity of the P-H bonds on 2-3.

The ability of 2-3 to reduce additional unsaturated substrates has been explored. Furthermore, we present the potential for phosphorus-hydrogen bonds to act as proton, hydride, hydrogen atom and electron donors. Finally, the mechanism of hydrogen transfer from 2-3 to azobenzene was studied, and experimental results indicate that initial hydride transfer is the rate determining step, followed by proton transfer.

4.1 Introduction

To better understand the reactivity of dihydridophosphorane 2-3, we have surveyed the reactivity of other compounds which contain P-H bonds, including 3-coordinate phosphines, 5- coordinate phosphoranes and 6-coordinate species as relevant. Phosphorus and hydrogen have very similar electronegativities, at 2.06 and 2.22, respectively; therefore the P-H bond is relatively non- polarized.1 This characteristic leads to the range of reactivity of P-H compounds, including heterolytic bond cleavage to produce either proton or hydride, and homolytic cleavage to provide phosphorus radicals. 2

4.1.1 Acidity of P-H Compounds

Compounds with P-H bonds are traditionally viewed as convenient precursors to a wide variety of useful phosphorus species. In particular, various phosphines have become extremely

95 valuable as ligands in transition metal catalysis for their ability to tune the reactivity and selectivity of transformations.3 Functionalization of P-H compounds often depends on their acidity.

Deprotonation and metalation with a strong base such as butyllithium provides a phosphorus anion, which can then react with a variety of electrophiles (Figure 4-1). This mode of reactivity is a common method of providing P(III) and P(V) compounds. It is especially useful when installing multiple substituents at phosphorus, and acts as a complement to bond formation with nucleophiles at P-X bonds.4 This strategy is used in the synthesis of chiral chelating ligands like DIOP; diphenylphosphine 4-1 is first treated with butyllithium, then added to a solution of halogenated substrate 4-3 to form DIOP 4-4 (Figure 1A).5 Similarly, hydridophosphoranes also react with electrophiles; for example, compound 4-7 is formed by the reaction of hydridophosphorane 4-5 with imine 4-6, in this case no external base is required for reactivity (Figure 1B).6

Figure 4-1. Examples of the base-promoted nucleophilicity of P-H bonds. A) reaction of phosphine 4-1 with alkyl halide 4-3. B) Reaction of hydridophosphorane 4-5 with imine 4-6.

96 4.1.2 Radical Abstraction from P-H bonds

Compounds with P-H bonds can also react with multiple bonds via hydrophosphination

(Figure 4-2).7 This class of reaction is often used in the synthesis of functionalized phosphorus compounds. Hydrophosphination can be promoted by radical initiators, light, or heat.8 In addition, a wide range of multiple bonds can react with P-H bonds via hydrophosphination, including alkenes, alkynes, aldehydes and imines; stereoselective reactions are also accessible in the presence of chiral catalysts.9 An example of alkene hydrophosphination is the addition of phosphine to limonene 4-8, which is catalyzed by AIBN.10 Similarly, the alkylthiolation of hydridophosphorane

4-10 can be effected by exposure to UV light to form phosphorane 4-11.11

Figure 4-2. Hydrophosphination of A) limonene 4-8 with phosphine, and B) dibutyldisulfide with hydridophosphorane 4-10.

4.1.3 Hydricity of P-H bonds

In contrast to functionalizations of P-H bonds where the hydrogen atom acts primarily as a proton, several examples of hydridic behavior have been reported. While this reactivity has not been widely utilized, examples are known for three-coordinate, five coordinate and six-coordinate

97 phosphorus species. For example, tri-coordinate diazaphospholane 4-12 (Figure 4-3A) has been shown to 1) transfer hydride to triphenylcarbenium cations, 2) undergo halogen-hydrogen exchange with a variety of E-X bonds, and 3) reduce ketones such as 4-13 to form diazaphospholane 4-14.12

The unique reactivity of this species may involve electron donation from the nitrogen ligands, leading to increased hydricity. Hydrogen elimination has been observed from hydridophosphoranes with pendant alcohols, such as 4-15, as shown in Figure 4-3B.13 This reactivity confirms that the P-H bond acts as a hydride in the presence of a proton source (in this case, an intramolecular one).14 Hexacoordinate dihydrophosphate 4-17 also has been found to reduce aldehydes and ketones in good yields (Figure 4-3C).15

Figure 4-3. A) Reduction of benzophenone with P-hydrido-1,3,2-diazaphospholane. B) Release of hydrogen from hydridophosphorane 4-15. C) Reduction of aryl aldehydes by 4-17.

Keeping the diverse reactivity of P-H bonds in mind, we set out to probe the behavior of dihydridophosphorane 2-3, especially in regard to its ability to transfer hydrogen to substrates.

4.2 Hydrogen Transfer to Organic Substrates from Dihydridophosphorane 2-3

We had initially reported that dihydridophosphorane 2-3 transfers hydrogen to azobenzene and concomitantly reforms planar three-coordinate compound 2-1. We sought to broaden the substrate scope for reductions using 2-3, and herein we present successful hydrogenations of imines and alkenes. The reduction of aldehydes resulted in undesired side-reactions, while attempted reductions of ketones and triple bonds were unsuccessful.

4.2.1 Imines

The reaction of imines with 2-3 proceeded to provide the corresponding amines in moderate yields (Figure 4-4). The reduction of C-N π bonds with 2-3 is exemplified by the reaction of the compound with tosylated benzaldimine 4-20a. One equivalent of 2-3 was added to 4-20a, and the reaction mixture was refluxed for 60 h in chloroform. The corresponding amine 5-20a was isolated in 71% yield; in most cases the unreacted imine starting compounds were observed and could be separated by column chromatography. We have observed that yield is independent of imine substituents: 4-bromo and 4-methoxy benzaldimines (4-20b and 4-20c) are reduced and isolated in 63% and 62% yield, respectively. In addition, both N-tosyl and N-para-methoxyphenol imines (4-20a and 4-20d) are reduced successfully. It is notable that although the yield for cinnamaldehyde derivated 4-20e is relatively poor, the imine is selectively reduced over the alkene.

Finally, ketimines such as 4-20f are reduced as effectively as aldimines.16

Figure 4-4. Imine reduction with 2-3.

4.2.2 Aldehydes and Ketones

Since imine reduction with 2-3 was successful, we also attempted the reduction of carbonyl compounds. We found that the reaction of benzaldehyde with 2-3 leads to full consumption of the dihydridophosphorane, as expected. However, 2-1 is not formed, and in situ 1H NMR spectroscopy did not show the expected peaks for benzylic alcohol. Similarly anomalous results were also obtained with p-fluorobenzaldehyde, and p-trifluoromethylbenzaldehyde. To further probe this reactivity, 2-3 and p-trifluoromethylbenzaldehyde 4-22 was heated and monitored by 31P NMR at intervals. The triplet of 2-3 is initially consumed, and a doublet at -21.7 ppm (J = 802 Hz) appears transiently; eventually only a singlet at 117 ppm is observed.

We suspected that upon the reduction of benzaldehyde to the desired alcohol (4-23), the product reacted further with 2-1 via O-H addition. The O-H adduct, 4-24 would then undergo a tautomerization to a three coordinate complex, 4-25, as has been reported to occur with MeOH

(Figure 4-5). We attempted to confirm this theory by reacting p-trifluoromethylbenzyl alcohol 4-

23 with 2-1 to observe both the transient five coordinate species and the resulting three coordinate compound. However, the reaction of 2-1 with 4-23 at room temperature is more complex than

100 expected, with doublets appearing at -31 ppm (d, J = 859 Hz) and 8 ppm (d, J = 706 Hz), and singlets at 123 and 140 ppm. These shifts are similar to those observed from the reduction of aldehyde with 2-3, however the exact values obtained indicate that the compounds are not identical.

Attempts to fully isolate and purify these compounds have been unsuccessful.

Figure 4-5. Proposed reactivity of aldehydes with 2-3.

In contrast to the reactivity observed with aryl aldehydes, the ketone acetophenone is not reduced in the presence of 2-3, even upon heating to 80 °C for 12 hours.

4.2.3 Alkenes

Next, we explored the reduction of alkenes, and found that 2-3 is also capable of reducing

C-C π bonds to alkanes, provided they are sufficiently electron deficient (Figure 4-6). We first reacted styrene 4-26a with 2-3; however, no hydrogen transfer occurred. Adding a single methyl ester to the alkene (4-26b) did not alter this result. On the other hand, 1,1-diester 4-26c was successfully reduced in 20% yield. 1,2-diester 4-27d is also successfully reduced in 36% yield. A comparison of 4-26c and 4-26d suggests that disubstitution with both 1,1- and 1,2- electron withdrawing groups will allow reductions. However, the addition of alkyl groups to a similarly activated α,β-diester, 4-26e, again prohibits reactivity. These results indicate that 2-3 is more reactive with electron deficient alkenes.16

101

Figure 4-6. Scope of alkene reduction with 2-3.

In order to probe the stereochemistry of hydrogen addition to the alkenes, we explored the reduction of substrates that would provide diastereomeric products that can be identified spectroscopically. We modified an experiment that Wilkinson used previously to determine the selectivity of rhodium-catalyzed hydrogen transfer. Wilkinson reduced maleic acid 4-28 with a rhodium catalyst 4-29 under an atmosphere of deuterium gas (Figure 4-7). He reported that the diasteromers of the product could be distinguished by IR spectroscopy, and further than cis-product

4-30 was formed.

Figure 4-7. The reduction of maleic acid with catalyst 4-29 under an atmosphere of deuterium gas.

In our case, we selected deuterium labeled dimethylmaleate 4-31 as the substrate. This choice allows us to use 2-3 as the reductant; the reaction of 2-3-d2 with dimethylmaleate was

17 observed to be prohibitively slow. Additionally, a 2.5 fold excess of 2-3 was used, as unreacted

4-31 complicates the IR analysis. The reaction solution was refluxed in chloroform for 24 h, then purified with column chromatography to provide product 4-32 as a yellow oil in 44% yield (Figure

102 4-8). The IR of the product was compared to analytically pure samples of cis and trans isomers, which were synthesized from the reduction of dimethylmaleate and dimethylfumarate with deuterium in the presence of palladium on carbon. IR spectra of the reaction product shows five

IR peaks which are unique to the cis-isomer 4-23, while there are no peaks matching the trans- isomer 4-24 (Table 4-1). This result indicates that hydrogen transfer from 2-3 to alkenes occurs in a concerted fashion, or a stepwise transfer occurs that is more rapid than bond rotation.

Figure 4-8. Reduction of dimethylmaleate shows that cis-addition predominates.

Table 4-1. Unique IR signals from the cis and trans products. Italicized values show the overlap between the experimental results and the cis product. Compound Wavenumbers (cm-1) 4-32 2911.16, 2848.05, 1294.63, 1260.78, 1169.63, 1107.48, 1031.69, 975.49, 918.07, 858.51, 832.07, 784.17, 733.1 4-33 2907.69, 2846.43, 1302.62, 1256.98, 1167.39, 1120.81, 1023.55, 980.31, 916.03, 847.09, 813.04, 776.69 Experiment 2875.5, 1258.56, 1171.40, 1030.28, 860.95, 829.54, 785.52.

4.2.4 Reduction of Triple Bonds

We subsequently tested the reduction of electron-poor alkynes, such as dimethylacetylene dicarboxylate with 2-3; however, the expected alkenes and alkanes were not observed. Further investigation determined that alkyne substrates reacted unpredictably with three coordinate species

2-1, leading to complex reaction mixtures. The observed reactivity indicates that if some alkynes are indeed reduced in the presence of 2-3, the formed 2-1 would then also react with the remaining substrate.

103 The reduction of nitriles could in principle yield primary amines. This reactivity is observed in the case of metal catalyzed hydrogenations or upon treatment of nitriles with strong hydride donors followed by protic workup. However, no reaction was observed between 2-3 and benzonitrile or acetonitrile.

4.2.5 Reduction of Single Bonds

We also sought to reduce weak single bonds, such as disulfides, which have a BDE of ~60 kcal/mol.18 Disulfides have previously been reduced to their substituent thiols via hydridic reagents in protic solvents,19 as well as hydrogen transfer from hydrazine.20 However, these substrates proved wholly unreactive in the presence of 2-3.

The ring opening of epoxides was also attempted, as they are typically observed to be reactive with nucleophiles, including hydrides, due to the strain inherent in the small ring size.

Additionally, aziridines have previously been reported to react with hydridophosphoranes.21 We selected epoxynorbornene to emphasize the ring strain in a bicyclic system and further destablilize the substrate; however, no reactivity was observed with 2-3.

4.2.6 Hydrogen Transfer to Frustrated Lewis Pairs and Carbenes

Attempts to transfer hydrogen from 2-3 to frustrated Lewis pairs was unsuccessful: no reaction was observed with tris(t-butyl)phosphine tris(pentafluorophenyl)borane, or 2,2,6,6- tetramethylpiperidine tris(pentafluorophenyl)borane. A similar reaction with 1,3-di(t- butyl)imidizolidin-2-ylidene tris(pentafluorophenyl)borane 4-34 did not furnish the expected hydrogenated Lewis pair, but instead transferred hydrogen to the lone pair of the carbene to provide imidazolidine 4-35 (Figure 4-9).

104

Figure 4-9. Reduction of dimethylmaleate shows that cis-addition predominates.

4.3 Monoatomic Hydrogen Transfer from Dihydridophosphorane 2-3

While 2-3 has been shown to act as a source of H2 with unsaturated compounds, there is also the possibility of 2-3 transferring hydrides, protons, hydrogen atoms, or electrons. These reactions were subsequently studied with appropriate substrates.

4.3.1 Hydride Transfer from Dihydridophosphorane 2-3 to Cations.

The transfer of hydride from phosphorus to carbocations has been proposed previously.12,22

The hydricity of 2-3 was demonstrated by its reaction with triphenylcarbenium tetrafluoroborate 4-

36 (Figure 4-10). Hydride transfer occurs within seconds at room temperature to form triphenylmethane, as evidenced by the color change of the reagents. A novel phosphorus species is also observed, which has been identified as fluorohydridophosphorane 4-37. The phosphorus signal appears at -30 ppm, denoting a five-coordinate phosphorane; it appears as a doublet of doublet of triplets, where J = 35, 940, 1060 Hz. The 35 and 940 Hz constants arise from proton coupling, and are eliminated in 31P{1H} NMR. To confirm the existence of the phosphorus- fluorine bond, 19F NMR spectra of the compound was also obtained. The signal appears at 60 ppm

2 1 23 as a doublet of doublets, with JH-F = 104 Hz and JP- F = 1049 Hz. The abstraction of fluoride from the tetrafluoroborate counterion indicates that the putative phosphonium intermediate 3-23 has a high Lewis acidity; fluoride abstraction by phosphonium salts is precedented.24

105

Figure 4-10. Hydride abstraction from 2-3 with trityl cation 4-36.

Several attempts to observe phosphonium intermediate 3-23, which we have proposed as a key intermediate in Section 3.4, were made. Triphenylcarbenium tetraphenylborate 4-38 was selected as a substrate to eliminate the source of fluoride. In this case, hydride transfer occurs just as rapidly to produce triphenylmethane; however, only phosphorus decomposition was observed

(Figure 4-11A). This result indicates that free phosphonium species 3-23 is quite unstable. To stabilize the phosphonium, we introduced proximal bulky aryl groups with ortho-fluoro substituents, which have previously been shown to stabilize positively charged silicon atoms.25 In order to introduce these stabilizing groups, the O,N,O ligand scaffold was exchanged for a N,N,N ligand, which allows for incorporation of N-aryl substituents in 4-39 (Figure 4-11B).

Unfortunately, this strategy was not successful in our system, and a hydridofluorophosphorane 4-

40 species was still formed when the compound was exposed to triphenylcarbenium tetrafluoroborate.

106

Figure 4-11. Unsuccessful attempts to isolate putative phosphonium intermediate 3-23 through A) hydrogen abstraction with triphenylcarbenium tetraphenylborate, and B) hydrogen abstraction from stabilized derivative 4-39 with triphenylcarbenium tetrafluoroborate.

4.3.2 Proton Transfer from Dihydridophosphorane 2-3.

Yi-Chun Lin in our group has determined that 2-3 in acetonitrile is deprotonated in the presence of DBU (4-41) to form 2-1 and hydrogen gas (Figure 12). This reactivity presumably proceeds by initial deprotonation of 2-3 to the conjugate base 3-22. The conjugate base is unstable, such that 3-22 and the conjugate acid of DBU (4-42) then react to release hydrogen gas, which has been detected by GC.

The conjugate acid of DBU has a pKa of 24.3 in acetonitrile, and weaker bases do not react with 2-3. These results allow us to bracket the pKa of 2-3 as approximately 24 in acetonitrile. In

26 the literature, a hydridophosphorane with a pKa of 16 in acetonitrile has been reported, while

27 others appear to have higher pKa values.

107

Figure 4-12. Deprotonation of 2-3 by DBU is followed by hydrogen release, thus reforming 2-1.

4.3.3 Hydrogen Atom Transfer from 2-3.

Compound 2-3 reacts with 2 equivalents of the free radical galvinoxyl (4-43) cleanly to produce 2-1 and 2 equivalents of the resulting alcohol 4-44, indicating that 2-3 is capable of two subsequent hydrogen atom transfers (Figure 13). This reactivity indicates that the products 4-44 and 2-1 are more stable than the reactants. The stability change may be partially attributed to the formation of the O-H bond in 4-44, which has a bond strength of 78.6 kcal/mol.28 This value is larger than the P-H bond strength computed for a similar hydridophosphorane (3-24), at 70.6 kcal/mol.29

Figure 4-13. Hydrogen atom transfer from 2-3 to 2 equiv of galvinoxyl radical 4-43.

108 4.3.4 Electron Transfer from 2-3.

A cyclic voltammagram of 2-3 (0.2 uM in MeCN with 0.1M NBu4PF6, 100 mV/sec scan rate, cobaltacenium hexafluorophosphate used as the internal reference) shows two irreversible one electron oxidation potentials at –126 mV and 711 mV adjusted to Fc/Fc+. The second oxidation potential is close to the value reported for 2-1, which is 705 mV vs. Fc/Fc+.30 A similar trend, where pentacoordinate phosphorus complexes have lower oxidation potentials than their tricoordinate analogues, has recently been documented by our group.31

10 Current (mA) Current

0 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5

-10 Potential vs Fc/Fc+ (V)

Figure 4-14. CV of 2-3 shows two oxidation potentials, as well as cobaltcenium hexafluorophosphate couple at ca. –1.4 V.

The first oxidation potential of 2-3 allows 2-3 to be oxidized by Ag+, which has a reduction potential of ~1.0 V vs SHE32 (compared to an oxidation potential of 0.51 V for 2-3). Treatment of

2-3 with 1 equivalent of silver triflate in benzene at room temperature leads to rapid precipitation of silver (0) and produces a yellow solution (Figure 4-15). Filtration of the sample allows for collection of an EPR spectrum (Figure 4-16), which was obtained by Alexey Silakov. The spectrum

109 presents as a complex quartet at g = 2.00656. The hyperfine splitting can be attributed to the contributions of one 14N (23.37 MHz), one 31P (26.98 MHz) and two 1H (3.8737 MHz). The hyperfine splitting indicates that the signal likely arises from radical cation 4-45, which has not undergone any bond scission.

Figure 4-15. Synthesis of radical cation 4-45 from the reaction of 2-3 with silver triflate.

Simulated Data Experimental Data

3.35E+03 3.36E+03 3.37E+03 3.38E+03 3.39E+03 3.40E+03 Time (s)

Figure 4-16. EPR signal from 4-45, generated by the reaction of 2-3 with AgOTf in C6D6 at room temperature. The signal is overlayed with a simulated generated by Alexey Silakov.

4.4 Mechanism of Hydrogen Transfer from Dihydridophosphorane 2-3 to Azobenzene

Since dihydridophosphorane 2-3 has been shown to react with a diverse range of substrates and elementary reactions, we focused on determining the mechanism of hydrogen transfer to azobenzene, as it relates to catalytic hydrogen transfer with this

110 species. Specifically, we have determined the rate equation, activation parameters, kinetic isotope effects and Hammett effects of the general reaction.

Figure 4-17. Hydrogen transfer from hydridophosphorane 2-3 to azobenzene 2-4.

4.4.1 Reaction Kinetics

The reaction of 2-3 with 20 equivalents of azobenzene 2-4 in d6-benzene was monitored

1 by H NMR to determine the concentrations of 2-3 and 2-1 at 70 °C. The large excess of azobenzene allows us to approximate that the concentration of substrate is constant throughout the course of the reaction. Therefore the rate will depend only on the concentration of 2-3. A standard

2 comparison of the natural log of the concentration of 2-3 over time shows a linear correlation (R

= 0.999) (Figure 4-18), indicating that the reduction of azobenzene is first order in 2-3.

In order to determine the order in azobenzene, we measured kobs of the hydrogen transfer from 2-3 while varying the concentration of azobenzene (Figure 4-19). We used 1, 5, 10, 15, 20 and 100 equivalents of 2-4, with the absolute concentration of azobenzene in the experiments between 0.016 and 1.6 M. The reactions were again performed in d6-benzene at 60 °C. A plot of kobs versus the concentration shows a linear correlation, which indicates that the reaction is first order in azobenzene as well. The rate constant k is obtained from the slope of this plot, such that k=5.17E-04 L*mol-1*s-1. The rate of reaction is given by Eq. 4-1, and is second order overall.

Eq. 4-1 rate = 5.17E-04[2-3][2-4]

111

-3 0 500 1000 1500 2000 2500 -3.5

] -4

- 2

ln[ -4.5

-5

-5.5 Time (s)

Figure 4-18. Consumption of 2-3 vs. time under pseudo-first order conditions shows the reaction is first order in 2-3, R2 = 0.999.

9.00E-04 8.00E-04 7.00E-04

6.00E-04

) 1

- 5.00E-04 (s

obs 4.00E-04 k 3.00E-04 2.00E-04 1.00E-04 0.00E+00 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Concentration of 2-4 (M)

Figure 4-19. Linear dependence of observed rate on concentration of azobenzene (2-4). R2 = 0.997.

Activation parameters were determined through an Eyring analysis, in which the rate of the reaction was measured at varying temperatures in both benzene and acetonitrile. A solution of

2-3 and 20 equivalents of azobenzene 2-4 was monitored by 1H NMR in a thermostatted NMR between 30-60 °C for d3-acetonitrile and 30-70 °C for d6-benzene. Linear regression analysis of

112 an Eyring plot allows for the determination of the activation entropy and activation enthalpy of both solvents, from which the Gibbs energy of activation at 293 K was extrapolated (Table 4-2).

In both solvents the activation entropy is between -22 and -24 cal/mol*K, indicative of a bimolecular transition state, and in agreement with the previously determined rate equation (Eq. 4-

1) for this reaction. The activation enthalpy shows a modest solvent effect, at 16.9 kcal/mol for benzene and 12.4 kcal/mol for acetonitrile. These values translate to a lower Gibbs energy of activation for acetonitrile at 19.5 kcal/mol vs. 23.5 kcal/mol in benzene. This 4 kcal/mol difference may be indicative of a polar transition state, which is stabilized by the more polar acetonitrile solvent.

Table 4-2. Eyring analysis for the kinetic runs in benzene and acetonitrile. Solvent ΔH‡ (kcal/mol) ΔS‡ (cal/molK) ΔG‡, 293K (kcal/mol)

d6-benzene 16.9 (±0.9) -22.4 (±2.9) 23.5

d3-acetonitrile 12.4 -24.1 19.5

4.4.2 Isotopic effects on rate and labeling.

A study of the comparative rates of reduction of azobenzene with 2-3 and 2-3-d2 gives a kinetic isotope effect of 3.8 ± 0.5. This large positive value is consistent with a primary isotope effect, indicating that a P-H bond is cleaved during the rate determining step of the reaction. Our inability to cleanly synthesize 2-3-d precludes further analysis as to whether one or both P-H bonds are cleaved during the rate determining step.

2 It is notable that when azobenzene is reduced by 2-3-d2 and monitored via H NMR, the deuterium signal shifts cleanly from the phosphorus atom to the nitrogen of the diphenylhydrazine product. No deuterium incorporation is observed at the vinylic position of the backbone. This result indicates that a three to five coordinate isomerization of species 2-1•[D]2 is not relevant

113 during azobenzene reduction (Figure 4-20), although this mechanism has been suggested by a computational study.33

Figure 4-20. Synthesis of radical cation 4-45 from the reaction of 2-3 with silver triflate.

4.4.3 Hammett correlation of substituent effects.

A Hammett study compared the rate of hydrogen transfer from 2-3 to symmetrically substituted azobenzenes (Figure 4-22 and 4-23). A plot of the rate of reaction for a particular substrate is plotted against the σ value of its substituent. The results of our experiment show that the rate of reduction is increased with increased electron withdrawing character of the substituent, which is consistent with negative charge formation on the substrate during the rate limiting step.

114 Figure 4-21. Reduction of substituted azobenzenes allows for the construction of a Hammett plot.

1.5 4,4'-CF3

1 3,3'-Br )

0 0.5 3,3'-OMe 4,4'-Me 3,3'-Me

log(k/k 0 4,4'-Cl 4,4'-H 4,4'-F -0.5 4,4'-OMe -1 -0.5 0 0.5 1 1.5 σ

Figure 4-22. Hammett plot for hydrogen transfer from 2-3 to substituted azobenzenes. ρ = 1.2, R2 = 0.90. All replicates shown as separate points.

4.4.4 Mechanistic Discussion

We have shown that 2-3 reduces azobenzene 2-4 to diphenylhydrazine 2-14, and in fact can do so catalytically in the presence of ammonia borane. In order to determine how this reduction proceeds, we have studied the reactivity of 2-3 with azobenzene as well as other substrates.

Experimental results indicate that the rate determine step between 2-3 and azobenzene is first order in both reagents, and accelerated by electron-withdrawing groups. The experimental activation entropy indicates a bimolecular transition state, and the activation enthalpy is reduced in polar solvents, indicating a charged transition state. The reaction proceeds without any observable intermediates. Reactions with a variety of substrates have shown that 2-3 is capable of the elementary transfer of proton, hydride, hydrogen atom and electron in suitable systems. Each of these elementary reactions could be envisioned as a first step for the overall reduction of azobenzene by 2-3, as shown in Figure 4-23.

115

Figure 4-23. Potential reaction mechanisms for azobenzene reduction, including A) initial proton transfer, B) hydrogen atom transfer, C) electron transfer, and D) hydride transfer.

Path A: Proton Transfer

Initial proton transfer can be eliminated as a potential mechanism, as azobenzene is not a strong base. In fact, the conjugate acid of azobenzene, 4-47, is a stronger acid than HCl in THF.34

Comparatively, 2-3 is not deprotonated by moderate amine bases in acetonitrile. The difference in acidities means that proton transfer from 2-3 to azobenzene would not be spontaneous. Moreover, if proton-transfer was a rate limiting step, the Hammett correlation would be expected to show a positive charge buildup in the transition state, which was not observed.

Path B: Hydrogen Atom Transfer

Hydrogen atom transfer is a potential mechanism for azobenzene reduction. Historically,

P-H bonds are known to participate in radical deoxygenation and dehalogenation reactions, although these reactions typically require radical initiators.35 Experimentally, the reaction of dihydridophosphorane 2-3 with azobenzene 2-4 has been monitored by EPR, and no signal was observed. Furthermore, no dehalogenation was observed when halogenated azobenzenes or imines

116 were used as substrates. The addition of radical-initiator AIBN to a typical reaction also does not increase the reduction rate of azobenzene (Figure 4-24A). To further probe for the presence of radical intermediate 4-49, allyl-substituted azobenzene 4-52 was synthesized (Figure 4-24B). The reduction with 2-3 produces the related hydrazobenzene 4-53 cleanly, with no cyclization of the pendant allyl group. While this observation does not indicate with certainty that no radical reaction occurs, it does indicate that a second intermolecular hydrogen atom transfer occurs more quickly than cyclization of a nitrogen radical with the allyl group. Collectively, our observations disfavor the reaction proceeding through a radical pathway.

Figure 4-24. A) Conversion of azobenzene is unaffected by AIBN addition. B) Reduction of 2- allylazobenzene 4-52 proceeds cleanly without cyclization to 4-53.

Pathway C: Electron Transfer

In order to evaluate electron transfer from dihydridophosphorane 2-3 to azobenzene 2-4, we calculated the cell potential of the electron transfer. First, the half-reactions are referenced to the standard hydrogen electrode (SHE); the one-electron reduction of azobenzene is -1.15 V vs

SHE, and the one-electron oxidation of 2-3 is 0.51 V vs SHE. The cell potential is merely the sum

117 of the half-potentials, such that Ecell of -0.64 V. A Gibbs energy of 14.7 kcal/mol can be derived from ∆퐺° = −푛퐹퐸푐푒푙푙. This value indicates electron transfer would not be sponatenous alone, but may be a rate limiting step.

To further study the possibility of initial electron transfer, cyclopropy imine 4-54 was synthesized as a radical clock. An initial electron transfer to 4-54 would form a carbon radical adjacent to the cyclopropy group, which is known to rapidly rearrange to give an n-propyl group, observed in 4-57 (Figure 4-25A).36 The reaction of 4-54 with 2-3 left the cyclopropyl group intact, which indicates an initial electron transfer is unlikely (Figure 4-25B). Additionally, while 4-45 may be synthesized and observed by EPR, the reduction of azobenzene is EPR silent. These results disfavor initial electron transfer from 2-3 to azobenzene.

Figure 4-25. A) One electron reduction of cyclopropyl imine 4-54. B) Reduction of a cyclopropanated imine (4-54) with 2-3 proceeds cleanly to amine 4-58 without ring fission.

Pathway D: Hydride Transfer

In a fourth mechanistic pathway, 2-3 could act primarily as a hydride donor. The capacity of 2-3 to undergo hydride transfer has been confirmed by its reactivity with trityl cations (Section

4.3.1). This mechanism is also supported by the Hammett correlation, which shows that reduction rates are faster with electron-poor compounds, and indicates negative charge on the substrate during the transition state.

118 4.5 Conclusions

We have presented a full study of the reactivity of dihydridophosphorane 2-3. Compound

2-3 transfers hydrogen to a variety of substrates, including azobenzenes, imines, ketones, carbenes and alkenes, favoring electron poor π-bonds. Depending upon the substrate, 2-3 is also capable of transferring a proton, hydrogen atom, electron or hydride. Mechanistic studies have shown that hydride transfer is the most likely route for hydrogen transfer to azobenzene, and is therefore active in the catalytic reduction of azobenzene with 2-3 described in Chapter 2. We believe that the elucidation of the reduction mechanism of 2-3 will allow for the development of other phosphorus reductants.

4.6 Experimental Section

4.6.1 Synthesis of Substrates

4.6.1.1 Imines

4-20a37: To sodium p-toluenesulfinate (3.6 g, 20 mmol) in water (30 mL) was

added p-toluenesulfonamide (3.4 g, 20 mmol), then formic acid (30 mL) and

benzaldehyde (2 mL, 20 mmol). Reaction stirred at room temp 12 h. Resulting suspension was filtered, and the white precipitate was rinsed with water (30 mL) and hexanes (30 mL). Precipitate was then dissolved in DCM (100 mL) and stirred with water (30 mL) and concentrated aqueous sodium bicarbonate (30 mL) for 30 minutes. DCM was separated, then aqueous layers washed with DCM and combined organics dried over sodium sulfate and concentrated to give imine as a white solid requiring no further purification (1.77 g, 6.8 mmol,

1 34%). H NMR (300 MHz, CDCl3): δ 9.06 (1H, s), 7.98-7.91 (4H, m), 7.65 (1H, t, J=7. 3 Hz), 7.51

13 (2H, t, J=7.7 Hz), 7.38 (2H, d, J=8.0 Hz), 2.47 (3H, s). C NMR (75 MHz, CDCl3): δ170.15,

144.62, 134.99, 134.94, 132.23, 131.22, 129.79, 129.10, 128.01, 21.59.

119 4-20b38: To zinc (1.14 g, 17.5 mmol) under nitrogen was added a solution of

benzyl bromide (1.39 mL, 11.7 mmol) in THF (15 mL), stirred for 15 min.

P-toluenesulfonamide (1 g, 5.8 mmol), 4-bromobenzaldehyde (1.06 mL, 8.8

mmol) were added and stirred rt overnight. The reaction was diluted with brine (15 mL) and extracted with ethyl acetate (3x5 mL), dried over sodium sulfate and concentrated to a white solid which was recystallized from hexanes/ethyl acetate to yield 5d (900

1 mg, 46%). H NMR (300 MHz, CDCl3): δ 8.98 (1H, s), 7.88 (2H, d, J=8.3 Hz), 7.78 (2H, d,

J=8.5 Hz), 7.63 (2H, d, J=8.5 Hz), 7.35 (2H, d, J=8.0 Hz), 2.44 (3H, s). 13C NMR (75 MHz,

CDCl3): δ 168.93, 144.95, 134.93, 132.72, 132,53, 131.32, 130.39, 130.00, 128.28, 21.81 .

4-20c8: To zinc (1.14 g, 17.5 mmol) under nitrogen was added a solution

of benzyl bromide (1.39 mL, 11.7 mmol) in THF (15 mL), stirred for 15

min. P-toluenesulfonamide (1 g, 5.8 mmol), p-anisaldehyde (1.06 mL, 8.8

1 mg, 44%). H NMR (300 MHz, CDCl3): δ 8.94 (1H, s), 7.87 (4H, d, J=7.0 Hz), 7.32 (2H, d,

13 J=8.0 Hz), 6.96 (2H, d, J=8.8 Hz), 3.88 (3H, s), 2.43 (3H, s). C NMR (75 MHz, CDCl3): δ

169.32, 165.37, 144.37, 135.78, 133.84, 129.83, 127.99, 125.27, 114.77, 55.79, 21.74.

4-20d: To benzaldehyde (0.4 mL, 3.9 mmol) in toluene (3 mL) was

added p-anisidine (492 mg, 4 mmol) and stirred at rt overnight, then

sodium sulfate was added and stirred another 30 min, diluted with toluene

and filtered. Filtrate was concentrated and recrystallized from ethanol to

1 39 yield a white solid (497 mg, 2.36 mmol, 60%). H NMR (360 MHz, CDCl3) : δ 8.53 (1H, s),

7.93 (2H, m), 7.50 (3H, m), 7.27 (2H, d, J=9.0 Hz), 6.98 (2H, d, J=6.9 Hz), 3.88 (3H, s). 13C

NMR (75 MHz, CDCl3): δ 158.30, 144.85, 136.47, 131.03, 128.73, 128.60, 122.25, 114.38, 55.43

120

4-20e: To cinnamaldehyde (1.25 mL, 10 mmol) in ethanol (10 mL)

was added p-anisidine (1.23 g, 10 mmol) and stirred 12 h at room

temperature. Suspension was filtered and solid was recrystallized from

ethanol to form bright yellow plates (1.64 g, 6.9 mmol, 69%). 1H

40 NMR (300 MHz, CDCl3) : δ 8.32 (1H, t, J=1.3 Hz), 7.57 (2H, d, J=8.2 Hz), 7.39 (3H, quart,

J=7.8 Hz), 7.23 (2H, d, J=6.8 Hz), 7.13 (2H, d, J=4.9 Hz), 6.93 (2H, d, J=8.9 Hz), 3.85 (3H, s).

13 C NMR (75 MHz, CDCl3): δ159.31, 158.33, 144.42, 142.90, 135.66, 129.29, 128.82, 128.66,

127.31, 122.21, 114.34, 55.32.

4-20f:41 In a dry flask palladium diacetate (260 mg, 0.35 mmol) and p-

toluenesulfonamide (1.2 g, 7 mmol) were suspended in benzonitrile (9 mL) and

placed under an oxygen atmosphere. The reaction was heated 60 ⁰C and then styrene (4.8 mL, 42 mmol) was added and the reaction stirred for 24 hours. The reaction was cooled and solvent was removed via vacuum distillation to yield a brown semisolid purified by column chromatography 5% EtOAc/hexanes on silica gel to yield the product as a light yellow

1 42 solid (609 mg, 2.23 mmol, 32%). H NMR (300 MHz, CDCl3) : δ 7.91 (4H, t, J=8.4 Hz), 7.51

(1H, t, J=6.1 Hz), 7.44-7.33 (4H, m), 2.99 (3H, s), 2.45 (3H, s).

4-54:43 To cyclopropylphenylketone (2.0 g, 13.7 mmol) in toluene (10

mL) was added p-anisidine (1.68 g, 13.7 mmol) and molecular sieves and

refluxed overnight. Reaction was filtered and filtrate was concentrated.

Recyrstallization from hexanes provides yellow crystals as a 3:5 mixture

1 44 of stereoisomers (381 mg, 11%). H NMR (300 MHz, CDCl3): δ: major: 7.39 (2H, m), 7.21 (3H, m), 6.63 (2H, d, J=8.7 Hz), 6.52 (2H, d, J=8.7 Hz), 3.70 (3H, s),

1.99 (1H, m), 1.22 (2H, m), 0.99 (2H, m). minor: 7.72 (2H, m), 7.16 (3H, m), 6.92 (4H, m), 3.82

(3H, s), 1.89 (1H, m), 0.83 (2H, m), 0.58 (2H, m). 13C: 173.03, 171.18, 156.36, 155.82, 144.57,

121 144.42, 139.14, 138.52, 129.63, 128.62, 128.49, 128.41, 128.35, 128.32, 122.60, 114.37, 114.13,

55.85, 55.68, 20.57, 14.29, 9.72, 8.34.

4-26c:45 To benzaldehyde (0.71 mL, 7 mmol) in toluene (50 mL) was added

dimethyl malonate (0.87 mL, 7.7 mmol), piperidine (0.07 mL, 0.7 mmol), and

acetic acid (0.04 mL, 0.7 mmol), then refluxed 48 h with a Dean Stark trap.

Cooled, concentrated and purified by column chromatography (5% EtOAc/Hex on silica gel), followed by Kruegelrohr distillation to remove impurities gave the product as a tan oil (628 mg,

41%). 1H NMR (300 MHz, CDCl3)46: δ 7.78 (1H, s), 7.40 (5H, m), 3.85 (6H, s). 13C NMR (75

MHz, CDCl3): δ167.21, 164.55, 143.00, 132.81, 130.79, 129.46, 128.97, 125.53, 52.78, 52.75.

4-26d:47 To potassium carbonate (760 mg, 5.5 mmol) in DMA (6.5

mL) was added maleic acid (500 mg, 4.3 mmol), stirred 40 min rt, cooled to

0 C then benzyl bromide (1 mL, 8.4 mmol) added dropwise, warmed to rt and stirred 48 h, diluted with 10 mL water, extracted with ether (3x10 mL), combined extracts washed with brine (3x20 mL), dried over sodium sulfate and concentrated, purified by CC (hexanes to elute impurities, 10% E/H to elute product) as a colorless oil (960 mg, 75%). 1H NMR (300

48 MHz, CDCl3) : δ 7.34 (10H, s), 6.29 (2H, s), 5.14 (4H, s)

4-26e:49,50 To Cis-1,2,3,6-tetrahydro naphthalic anhydride (S6, 10 g, 66 mmol) was added phosphorus pentoxide (200 mg, 1.4 mmol), the mixed solids were placed under a nitrogen atmosphere and heated to ~220 ⁰C overnight. Upon cooling the solid was recrystallized from dry toluene (20 mL, 5 mL for second batch) to give S7 as a tan solid (5.96 g, 60%). 1H NMR

122 13 (CDCl3): δ 2.47-2.42 (4H, m), 1.85-1.81 (4H, m). C NMR (CDCl3): δ 165.02, 145.28, 20.87,

20.73.

To S7 (760 mg, 5 mmol) in toluene (50 mL) under air was added copper bromide (143 mg,

1 mmol) and silver fluoride (1.9 g, 15 mmol), in one portion. Phenyltrimethoxysilane (1.41 mL,

7.5 mmol) was then added dropwise, and the reaction was heated to reflux 36 h. The reaction was cooled, filtered over celite and concentrated, the oil was purified by column chromatography silica,

10% EtOAc/Hex to give 4-26e as a yellow oil (367 mg, 36%). 1H NMR (CDCl3): δ 7.39 (2H, t,

J=7.9 Hz), 7.23 (1H, t, J=7.9 Hz), 7.16 (2H, d, J=7.9 Hz), 3.77 (3H, s), 2.50 (2H, m), 2.40 (2H, m),

1.74 (4H, m). 13C NMR (CDCl3): δ 169.08, 167.45, 150.99, 136.57, 135.65, 129.87, 126.34,

121.84, 52.69, 26.84, 26.69, 21.60. HRMS (ESI): Calc’d: 261.1127; Obs’d: 261.1128 (M+1)

4.6.1.2 Azobenzenes

General Procedures for Azobenzene Synthesis

Method A:51 To aniline (10-30 mmol) in toluene (4 mL per mmol aniline), copper bromide (0.06 equiv) and pyridine (0.18 equiv) were added; the reaction was stirred vigorously in air and heated to 60 °C for 20 hours. The solution was cooled, solvent removed en vacuo and residue purified by column chromatography on silica gel with 2% ethyl acetate/hexanes as the eluent, where the desired product is the first spot to elute as a bright orange solution, and is concentrated to a bright orange powder.

Method B:52 To 4-fluoroaniline (0.85 mL, 9 mmol) in 50% ethanol/water (25 mL) was added

KOH (1.03 g, 18.4 mmol) in 50% ethanol/water (25 mL) and then solid potassium hexacyanoferrate (12.35 g, 37.5 mmol), portionwise. The solution was refluxed 12 h, then cooled, filtered over celite and washed with DCM. The filtrate was extracted with DCM, and the organic layer was dried over sodium sulfate, concentrated and purified by column chromatography with hexanes as eluent on silica gel to yield an orange solid.

123

S10: Prepared by Method A from m-toluidine. Yield 1.00 g, 45%.

1 53 H NMR (300 MHz, CDCl3): δ 7.75 (4H, s), 7.44 (2H, t, J=8 Hz), 7.32

(2H, d, J=8 Hz), 2.49 (6H, s).

S11: Prepared by Method A from m-anisidine. Yield 1.03 g, 47%. 1H

54 NMR (CDCl3): δ 7.60 (2H, d, J=8 Hz), 7.46 (4H, m), 7.07 (2H, d, J=8

Hz), 3.92 (6H, s).

S12: Prepared by Method A from 3-bromoaniline, reaction time 60 h.

1 55 Yield 1.30 g, 77%. H NMR (CDCl3): δ 8.05 (2H, t, J=1.8 Hz), 7.88

(2H, d, J=8 Hz), 7.62 (2H, dd, J=1 Hz, 6 Hz), 7.42 (2H, t, J=8 Hz).

S13: Prepared by Method A from p-toluidine. Yield 1.77 g, 79%.

1 56 H NMR (CDCl3): δ 7.81 (4H, d, J=8 Hz), 7.31 (4H, d, J=8 Hz), 2.43

(6H, s).

S14: Prepared by Method A from p-anisidine. Yield 1.03 g,

1 57 28%. H NMR (CDCl3): δ 7.88 (4H, d, J=9 Hz), 7.00 (4H, d, J=9 Hz),

3.89 (6H, s)

124

S15: Prepared by Method A from p-trifluoromethylaniline.

1 58 Yield 1.90 g, 85%. H NMR (CDCl3): δ 8.04 (4H, d, J=8 Hz), 7.80

(4H, d, J=8 Hz)

S16: Prepared by Method B from 4-fluoroaniline. Yield 245 mg,

1 59 25%. H NMR (CDCl3): δ 7.96 (4H, t), 7.22 (4H, t)

S17: Prepared by Method B from 4-chloroaniline. Yield 151 mg,

1 60 15%. H NMR (CDCl3): δ 7.87 (4H, d, J=9 Hz), 7.49 (4H, d, J=9 Hz)

4-52: To 2-allylaniline (250 mg, 1.88 mmol) in acetic acid (2 mL) was

added nitrosobenzene (201 mg, 1.88 mmol) and stirred 24 h at rt. Solution was

diluted with water (10 mL) and extracted with DCM (3x10 mL). Combined organics were dried and concentrated, purified by column chromatography (hexanes) to provides

1 an orange oil (147 mg, 35%). H NMR (CDCl3): δ 7.39 (4H, t, J=7.5 Hz), 7.14 (2H, t, J=7.4 Hz),

13 6.81 (4H, d, J=8.3 Hz), 2.17 (6H, s). C NMR (CDCl3): δ 153.07, 150.21, 140.08, 137.92, 131.33,

130.99, 130.54, 129.19, 127.10, 123.12, 122.97, 115.82, 115.54, 35.69 HRMS (ESI) calc’d for

C15H15N2 (M+1): 223.1235; found: 223.1236.

125

4.6.2 Reduction Products

Under a nitrogen atmosphere substrate (0.25 mmol) and 2-3 (61 mg, 0.25 mmol) were dissolved in chloroform (7.5 mL) and heated in a sealed tube for 60 h at 80 ⁰C, then cooled, concentrated and products isolated by column chromatography.

1 61 4-21a: 41 mg, 71%. H NMR (300 MHz, CDCl3): δ 7.78 (2H, d, J=8.2 Hz), 7.55-

7.05 (7H, m), 4.76 (1H, br s), 4.14 (2H, d, J=6.2 Hz), 2.46 (3H, s). 13C NMR (75

MHz, CDCl3): δ 143.41, 136.81, 136.44, 129.69, 128.56, 127.86, 127.70, 127.13,

47.11, 21.52.

1 25 4-21b: 53 mg, 63%. H NMR (300 MHz, CDCl3): δ 7.71 (2H, d,

J=8.3 Hz), 7.36 (2H, d, J=8.4 Hz), 7.28 (2H, d, J=8.4 Hz), 7.06 (2H, d, J=8.4

Hz), 5.12 (1H, t, J=6.1 Hz), 4.05 (2H, d, J=6.3 Hz), 2.43 (3H, s). 13C NMR (75

MHz, CDCl3): δ 143.79, 136.81, 135.53, 131.80, 129.87, 129.65, 127.22,

121.86, 46.66, 21.67.

4-21c: 62% by NMR (77% of 59 mg isolated as mixture with imine). 1H

62 NMR (300 MHz, CDCl3): δ 7.74 (2H, d, J=8.2 Hz), 7.28 (2H, d, J=8.0

Hz), 7.09 (2H, d, J=8.6 Hz), 6.79 (2H, d, J=8.6 Hz), 4.85 (1H, t, J=5.9

Hz), 4.05 (2H, d, J=6.1 Hz), 3.77 (3H, s), 2.44 (3H, s). 13C NMR (75

MHz, CDCl3): δ 159.48, 143.66, 137.04, 129.91, 129.45, 128.42, 128.08, 127.37, 114.23, 55.46,

46.97, 21.73.

126

1 63 4-21d: 29 mg, 55%. H NMR (300 MHz, CDCl3): δ 7.42-7.28 (5H, m),

6.80 (2H, d, J=8.8 Hz), 6.63 (2H, d, J=8.9 Hz), 4.31 (2H, s), 3.77 (3H, s).

13C NMR (75 MHz, CDCl3): δ 162.43, 152.29, 142.55, 139.78, 128.71,

127.67, 127.29, 122.32, 115.00, 114.22, 55.91, 49.35, 29.83, 1.16.

1 64 4-21e: 14 mg, 23%. H NMR (300 MHz, CDCl3): δ 7.38-7.21 (5H, m),

6.79 (2H, d, J=8.9 Hz), 6.63 (3H, m), 6.34 (1H, m), 3.89 (2H, d, J=4.9

Hz), 3.75 (3H, s). 13C NMR (75 MHz, CDCl3): δ 152.42, 142.39,

137.03, 131.57, 128.70, 127.63, 126.44, 115.02, 114.55, 55.93, 47.36.

1 25 4-21f: 54 mg, 79%. H NMR (300 MHz, CDCl3): δ 7.61 (2H, d, J=8.1 Hz), 7.17-

7.07 (7H, m), 5.19 (1H, d, J=6.8 Hz), 4.44 (1H, t, J= 6.8 Hz), 2.37 (3H, s), 1.40 (3H, d, J=6.8 Hz). 13C NMR (75 MHz, CDCl3): δ 143.11, 142.21, 137.69, 129.71, 129.47, 128.53,

127.39, 127.14, 126.20, 53.72, 23.64, 21.54.

1 65 4-58: 52 mg, 83%. H NMR (300 MHz, CDCl3): δ 7.41-7.22

(5H, m), 6.66 (2H, d, J=9 Hz), 6.43 (2H, d, J=9 Hz), 4.05 (1H, br s), 3.66

(3H, s), 3.54 (1H, d, J=9Hz), 1.14 (1H, m), 0.60-0.30 (4H, m). 13C NMR

(75 MHz, CDCl3): δ 152.03, 143.70, 142.04, 128.61, 127.09, 126.61,

114.80, 114.73, 63.96, 55.84, 19.94, 4.35, 3.58.

127 4-27c: 11 mg, 20% 1H NMR (300 MHz, CDCl3):66 7.27-7.17(5H, m),

3.69 (6H, s), 3.21 (2H, d, J=7.6 Hz)

4-27d: 28 mg, 36% (accounting for ~10% integration of alkene by

NMR.) 1H NMR (300 MHz, CDCl3): 67 7.41 (10H, s), 5.16 (4H, s), 2.74

(4H, s).

4-53: 4-52 (50 mg, 0.225 mmol) and 2-3 (109 mg, 0.449) in

chloroform (5 mL) under nitrogen in a schlenk flask which was heated to 80 C

for 16 hours open to a nitrogen atmosphere, at which time the solution was concentrated and purified by CC hexanes to 5% Et2O/Hex, followed by acetonitrile/hexanes extractions (5x 5 mL/5mL)- the hydrazine was isolated upon concentrating the acetonitrile layers

1 as a clear oil (10 mg, 20%). H NMR (CDCl3): 7.24 (2H, t, J=8.1 Hz), 7.14 (2H, m), 7.01 (1H, d,

J=7.8 Hz), 6.84 (4H, d, J=7.9 Hz), 6.03 (1H, m), 5.79 (1H, s), 5.59 (1H, s), 5.18 (2H, m), 3.41 (2H, d, J=5.8 Hz). 13C NMR (75 MHz, CDCl3): δ 148.88, 146.58, 135.99, 130.31, 129.48, 127.97,

122.96, 119.96, 119.72, 116.82, 112.46, 111.79, 36.41 HRMS (ESI) calc’d for C15H17N2 (M+1):

225.1392; found: 225.1393.

4-32: From 4-31: Under a nitrogen atmosphere 4-31 (146 mg, 1 mmol) and 2-3

(515 mg, 2.1 mmol) were dissolved in chloroform (7.5 mL) and heated in a sealed tube for 60 h at 80 ⁰C, then cooled, concentrated and products isolated by column chromatography (10% EtOAc/Hex) to afford the product as a yellow oil. The spectral characteristics were compared to the two diastereomers, see below.

From Dimethyl maleate68: A mixture of deuterium oxide (30 mL) and Pd/C (70 mg) was placed under a hydrogen atmosphere, then sealed. The mixture was stirred vigorously overnight, then dimethyl maleate (0.63 mL, 5 mmol) in methanol (2 mL) was added. The reaction was stirred

128 8 h, diluted with diethyl ether and filtered over celite. The organic layer was dried and concentrated to give 4-32 as a clear oil which solidifies slowly upon standing (411 mg, 56% yield). 1H NMR

13 (CDCl3): 2.62 (2H, s), 3.70 (6H, s). C NMR (CDCl3): 172.64, 51.69, 28.39 (t, J=19.9 Hz). HRMS:

(M+Na) Calc’d: 171.0571; Obs’d: 171.0576. FT-IR: 3001.4, 2955.8, 2911.2, 2848.1, 1732.3,

1437.5, 1294.6, 1260.8, 1201.0, 1169.6, 1107.5, 1031.7, 975.5, 918.1, 858.5, 832.1, 784.2, 733.1 cm-1.

4-33: As above for 4-32, from dimethyl fumarate. 1H NMR (CDCl3): 2.62 (2H,

13 s), 3.70 (6H, s). C NMR (CDCl3): 172.70, 51.75, 28.45 (t, J=19.9 Hz). HRMS:

M+Na Calc’d: 171.0602 ; Obs’d: 171.0608. FT-IR: 3001.3, 2955.9, 2907.7, 2846.4, 1732.7,

1437.8, 1302.6, 1257.0, 1199.9, 1167.4, 1120.8, 1023.6, 980.3, 916.0, 847.1, 813.0, 776.7, 732.7 cm-1

4-44: To 2-3 (61 mg, 0.25 mmol) in benzene (7.5 mL) under a nitrogen

atmosphere was added galvinoxyl (211 mg, 0.5 mmol). The solution changed

upon mixing from black to orange. After 1 hr concentrated and purified by

column chromatography on silica gel with 2% Ethyl acetate/hexanes as eluent to

give 4-44 as an orange powder (198 mg, 94%). 1H NMR (CDCl3): 7.60 (1H, s),

7.35 (2H, s) 7.16 (1H, s), 7.00 (1H, s), 5.56 (1H, s), 1.47 (18H, s), 1.32 (18H, s). 13C NMR (CDCl3):

186.57, 155.63, 148.88, 146.97, 144.64, 136.56, 135.80, 130.05, 128.31, 127.74, 35.60, 35.09,

34.58, 30.40, 29.82, 29.68. HRMS (EI): (M+1): calc: 423.3263, obsd: 423.3263.

Triphenylmethane: Prepped in glovebox. 2-3 (62 mg, 0.25 mmol) and

triphenylcarbenium tetrafluoroborate (83 mg, 0.25 mmol) were stirred in 2 mL of acetonitrile at room temp under a nitrogen atmosphere for one hour, concentrated onto silica gel and purified by column chromatography (hexanes on silica gel) to give triphenylmethane as a white

129 solid (40 mg, 66%). 1H NMR (CDCl3): 7.30-7.16 (9H, m), 7.10 (6H, d, J=7.2 Hz), 5.55 (1H, s).

13C NMR (75 MHz, CDCl3): δ 144.02, 129.59, 128.44, 126.43, 56.95.

Other Procedures

4-40: From the reaction of 4-39 (20 mg, 0.036 mmol) with trityl tetrafluoroborate in d3-acetonitrile

(0.5 mL). 4-40 is indicated by the multiplet at -49 ppm in the 31P and the dd at -46 ppm in the F19

4.6.3 Electrochemistry

All cyclic voltammetry experiments were conducted at ambient temperature using a Pine

WaveNow XV potentiostat under a nitrogen atmosphere. A solution of analyte (1 mM) and tetrabutylammonium hexafluorophosphate (0.1 M) was prepared in 15 mL of acetonitrile. A 3.0 mm glassy carbon disc, polished with 0.05 micron alumina and rinsed with acetonitrile prior to use, served as the working electrode. A platinum wire served as the counter electrode, and non- aqueous Ag/Ag+ (10 mM in AgNO3 in acetonitrile) served as the reference electrode. At the end of each experiment, sublimed ferrocene (1.0 mM) was added as an internal reference.

130 30

10 Current (mA)

0 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5

-10 Potential vs Fc/Fc+ (V)

Figure 4-26: Plot of 2-3 with colbatacenium.

131

Figure 4-27: Plot of azobenzene with ferrocene

4.6.4 EPR Spectroscopy

EPR spectra were recorded using a Bruker ESP300 equiped with ER041MR MW bridge and a

ER 4102ST dual mode resonator. The EPR spectrum was recorded using the following settings:

MW power = 2 mW, MW frequency = 9.478 GHz, sweep width = 10.0 G, modulation amplitude

= 0.5 G, conversion time = 29.30 ms, and time constant = 40.96 ms. EPR simulation was performed utilizing "garlic" routine from EasySpin package69 for MatLab and aided by home- written automatic spectral fitting routine.

132 4.6.5 NMR Monitoring of Isotopic Labelling

Reduction of Azobenzene with 2-3-d2: A solution of 2-3-d2 (10 mg, 0.04 mmol) and azobenzene

(73 mg, 0.40 mmol) in C6H6 was sealed under a nitrogen atmosphere and heated to 60 °C for 4 h.

2 The reaction was monitored by H NMR and reference to C6D6

2 Figure 4-28: H NMR of azobenzene with 2-3-d2 in benzene, t=0.

133

2 Figure 4-29: H NMR of azobenzene with 2-3-d2 in benzene, after 4 h, 60 °C.

4.6.6 Kinetics

Kinetic Isotope Effects from 2-3-d2

From stock solutions which were stored over sieves at -35 °C under a nitrogen atmosphere for 12-

48 h prior to use, 2-3 or 2-3-d2 (3 mg, 0.012 mmol), azobenzene (23 mg, 0.126 mmol), and d- benzene (0.8 mL) were mixed under a nitrogen atmosphere and placed in a J. Young NMR into an NMR thermostatted at 60 °C. The rate was determined by comparing the relative concentrations of 2-1 and 2-3 by integration of the vinylic peaks.

-1 -1 -1 Kobs (s ) average (s ) std dev (s )

1 2 3

D 2.64E-05 2.32E-05 2.35E-05 2.44E-05 1.77E-06

H 8.09E-05 9.80E-05 9.40E-05 9.10E-05 8.94E-06

KIE 3.73E+00

134 error 4.56E-01

Azobenzene Reduction with 2-3 -4.2 0 5000 10000 15000 20000 25000 -4.4 -4.6 -4.8 y = -9.40E-05x - 4.18E+00 R² = 9.89E-01

-5

]

3 -

2 -5.2 y = -9.80E-05x - 4.19E+00 ln[ -5.4 R² = 9.92E-01 -5.6 -5.8 y = -8.09E-05x - 4.18E+00 -6 R² = 9.91E-01 -6.2 Time (sec)

Azobenzene Reduction with 2-3-d2 -4.2 0 5000 10000 15000 20000 25000 30000 35000 -4.3

-4.4 y = -2.34E-05x - 4.25E+00

-4.5 R² = 9.94E-01

] 2

d -4.6

3 - 2 -4.7 y = -2.64E-05x - 4.24E+00 ln[ R² = 9.95E-01 -4.8

-4.9

-5 y = -2.35E-05x - 4.26E+00 R² = 9.96E-01 -5.1 Time (sec)

Hammett Plot

From stock solutions which were stored over sieves at -35 °C under a nitrogen atmosphere for 12-

48 h prior to use, 2-3 (3 mg, 0.012 mmol), azobenzene (0.126 mmol), and d-benzene (0.8 mL) were mixed under a nitrogen atmosphere and placed in a J. Young NMR into an NMR

135 thermostatted at 60 °C. The rate was determined by comparing the relative concentrations of 2-3 and 2-1 by integration of the vinylic peaks.

1.5 4,4'-CF3

1 3,3'-Br

0.5 4,4'-Cl 3,3'-OMe

log(k/k0) 0 3,3'-Me 4,4'-Me 4,4'-H 4,4'-F -0.5

4,4'-OMe

-1 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 σ

sigma k (s-1) log(k/k0)

0 8.09E-05 -0.05093

0 9.80E-05 0.032344

0 9.40E-05 0.014246

-0.28 5.23E-05 -0.24038

-0.28 5.79E-05 -0.1962

-0.28 4.74E-05 -0.2831

-0.24 2.38E-05 -0.58231

-0.24 2.44E-05 -0.57149

0.3 9.06E-05 -0.00175

0.3 9.49E-05 0.018384

0.3 1.38E-04 0.180997

136 0.48 1.45E-04 0.202486

0.48 1.43E-04 0.196454

0.48 1.51E-04 0.220095

1.06 1.98E-03 1.337783

1.06 2.17E-03 1.377577

1.06 2.26E-03 1.395226

-0.12 7.19E-05 -0.10215

-0.12 7.93E-05 -0.05961

-0.12 5.63E-05 -0.20837

0.2 1.00E-04 0.041118

0.2 1.13E-04 0.094196

0.74 5.33E-04 0.767845

0.74 5.19E-04 0.756285

0.74 6.24E-04 0.836302

Order in 2-3

Under a nitrogen atmosphere, a solution of 2-3 (3 mg, 0.012 mmol) and azobenzene (45 mg, 0.245 mmol) in 0.5 mL of C6D6 was prepared from stock solutions of each reagent which were stored over sieves at -35 °C under a nitrogen atmosphere for 12-48 h prior to use. The solution was transferred into a J. Young NMR tube and scans were taken using an NMR thermostatted at 70 °C.

The rate was determined by comparing the relative concentrations of 2-1 and 2-3 by integration of the vinylic peaks.

Order in Azobenzene

Under a nitrogen atmosphere, a solution of 2-3 (3 mg, 0.012 mmol) and azobenzene (either

1, 5, 10, 15, 20 or 100 equivalents) in 0.8 mL of C6D6 was prepared from stock solutions of each reagent which were stored over sieves at -35 °C under a nitrogen atmosphere for 12-48 h prior to use. The solution was transferred into a J. Young NMR tube and scans were taken using an NMR

137 thermostatted at 60 °C. The rate was determined by comparing the relative concentrations of 2-1 and 2-3 by integration of the vinylic peaks.

concentration equiv k (s-1) [azo]^2 [azo]^3 azobenzene (M) 0.315934 20 1.46E-04 0.0998143 0.0315347 0.236951 15 1.13E-04 0.0561456 0.0133037 0.157967 10 7.20E-05 0.0249536 0.0039418 0.078984 5 4.82E-05 0.0062384 0.0004927 0.015797 1 3.67E-05 0.0002495 3.942E-06 1.58 100 8.24E-04 2.4964 3.944312

k vs. [Azobenz]2 9.00E-04 y = 3.02E-04x + 7.14E-05 8.00E-04 R² = 9.90E-01 7.00E-04

6.00E-04

) 1

- 5.00E-04 (s

obs 4.00E-04 k 3.00E-04 2.00E-04 1.00E-04 0.00E+00 0 0.5 1 1.5 2 2.5 3 [Azobenz]2 (M2)

Eyring Plot of 2-3 with Azobenzene

°C. The rate was determined by comparing the relative concentrations of 2-1 and 2-3 by integration of the vinylic peaks.

138 -10 0.00285 0.0029 0.00295 0.003 0.00305 0.0031 0.00315 0.0032 0.00325 0.0033 0.00335

-11

-12

-13 ln(k/T) y = -8490.8x + 12.486 -14 R² = 0.9877

-15

-16 1/T (1/K)

Temp (K) 1/T (1/K) k (s-1M-1) ln(k/T)

303 0.0033 5.24E-05 -15.5708

313 0.003195 1.24E-04 -14.7401

323 0.003096 4.23E-04 -13.5466

333 0.003003 7.02E-04 -13.0701

343 0.002915 1.51E-03 -12.3316

139 4.6.7 Spectral Data

140

141

142

143

144

145

146

147

148

4.7 References

1 McMurry, J. Organic Chemistry, 7th ed.; Brooks/Cole: Belmont, CA, 2008, p 36.

2 Studer, A.; Amrein, S. Synthesis 2002, 7, 835.

3 Hartwig, J. Organotransition Metal Chemistry: From Bonding to Catalysis, University Science:

Sausalito, CA, 2010, p 36.

4 Ford-Moore, A. H.; Perry, B. J. Org. Synth. 1951, 31, 111.

5 Werz, U.; Brune, H.-A. J. Organomet. Chem. 1989, 363, 377.

6 Laurenco, C.; Bernard, D.; Burgada, R. C. R. Acad. Sci. Paris, C 1974, 278, 1301.

7 Marque, S.; Tordo, P. Top. Curr. Chem. 2005, 250, 43.

8 Semenzin, D.; Etemad-Moghadam, G.; Albouy, D.; Diallo, O.; Koenig, M. J. Org. Chem. 1997,

62, 2414.

149

9 (a) Han, L.-B.; Zhao, C.-Q.; Onozawa, S.; Goto, M.; Tanaka, M. J. Am. Chem. Soc. 2002, 124,

3842. (b) Han, L.-B.; Zhao, C.-Q. J. Org. Chem. 2005, 70, 10121.

10 Robertson, A.; Bradaric, C.; Frampton, C. S., McNulty, J.; Capretta, A. Tetrahedron Lett. 2001,

42, 2609.

11 Bentrude, W. G.; Kawashima, T.; Keys, B. A.; Garroussian, M.; Heide, W.; Wedegnertner, D.A.

J. Am. Chem. Soc. 1987, 109, 1227.

12 (a) Gudat, D.; Haghverdi, A.; Nieger, M. Angew. Chem. Int. Ed. 2000 39, 3084. (b) Burck, S.;

Gudat, D.; Nieger, M.; Du Mont, W.-W. J. Am. Chem. Soc. 2006, 128, 3946.

13 (a) Kajiyama, K.; Kojima, S.; Akiba, K. Y. Tetrahedron Lett. 1996, 37, 8409. (b) Kajiyama, K.;

Kojima, S.; Nakamoto, M.; Akiba, K. Y. J. Am. Chem. Soc. 1996, 188, 12866.

14 Alder, R. W.; Read, D. Angew. Chem. Int. Ed. 2000, 39, 2879.

15 Miyake, H.; Kano, N.; Kawashima, T. J. Am. Chem. Soc. 2009, 131, 16622.

16 Neither imine nor electron poor alkenes are appropriate substrates for phosphorus-catalyzed reductions; 2-2 can reduce these substrates directly. (a) Yang, X.; Fox, T.; Berke, H. Chem.

Commun. 2011, 47, 2053. (b) Yang, X.; Zhao, L.; Fox, T.; Wang, Z.-X.; Berke, H. Angew. Chem.

Int. Ed. 2010, 49, 2058.

17 (a) Osborn, J.A.; Jardine, F.H.; Young, J.F.; Wilkinson, G. J. Chem. Soc. A 1966, 1711. (b)

Childs, C. R. Jr.; Block, K. J. Org. Chem. 1961, 26, 1630.

18 Jursic, B. S. Int. J. Quant. Chem. 1997, 62, 291.

19 Ookawa, A.; Yokoyama, S.; Soai, K. Synth. Commun. 1986, 16, 819.

20 Maiti, S. N.; Spevak, P.; Singh, M. P.; Micetich, R. G.; Reddy, A. V. N. Synth. Commun. 1988,

18, 575.

21 Grechkin, N. P.; Grishina, L. N. Bull. Acad. Sci. USSR 1970, 19, 1549.

22 (a) Bercaw, J. E.; Grubbs, R. H.; Hazari, N.; Labinger, J. A.; Li, X. Chem. Commun. 2007, 2974.

(b) Shevchenko, V.; Engel, R. Heteroat. Chem. 1997, 8, 501.

23 Arduengo, A. J.; Breker, J.; Davidson, F.; Kline, M. Heteroat. Chem. 1993, 4, 213

24 Caputo, C. B.; Hounjet, L. J.; Dobrovetsky, R.; Stephan, D. W. Science 2013, 341, 1374

150

25 (a) Duttwyler, S.; Do, Q.-Q.; Linden, A.; Baldridge, K. K.; Siegel, J. S. Angew. Chem., Int. Ed., 2008, 47,

1719. (b) Romanato, P.; Duttwyler, S.; Linden, A.; Baldridge, K. K.; Siegel, J. S. J. Am. Chem. Soc. 2010,

132, 7828.

26 (a) Lensink, C.; Xi, S. K.; Daniels, M.; Verkade, J. G. J. Am. Chem. Soc. 1989, 111, 3478. (b)

Ross, M. R.; Martin, J. C. J. Am. Chem. Soc. 1981, 103, 1234. c) Ovchinnikov, V. V.; Pudovik, M.

A.; Galkin, V. I.; Cherkasov, R. A.; Pudovik, A. N. Bull. Acad. Sci. USSR 1977, 26, 39.

27 Milbrath, D. S.; Verkade, J. G. J. Am. Chem. Soc. 1977, 99, 6607

28 J. Am. Chem. Soc. 1969, 91, 3883.

29 McCarthy, S. M.; Lin, Y. C.; Davarajan, D.; Change, J. W.; Yennawar, H. P.; Rioux, R. M.; Ess,

D. H.; Radosevich, A. T. J. Am. Chem. Soc., 2014, 136, 4640.

30 Arduengo, A. J.; Stewart, C. A.; Davidson, F.; Dixon, D. A.; Becker, J. Y.; Culley, S. A.; Mizen,

M. B. J. Am. Chem. Soc. 1987, 109, 627

31 Pistner, A. Unpublished results.

32 Song, L.; Trogler, W. C. Angew. Chem. Int. Ed. 1992, 31, 770.

33 Zeng, G.; Maeda, S.; Taketsugu, T.; Sakai, S. Angew. Chem. Int. Ed. 2014, 53, 4633.

34 Klotz, I. M.; Fiess, H. A.; Ho, J. Y. C.; Mellody, M. J. Am. Chem. Soc. 1954, 76, 5136.

35 (a) Leca, D.; Fensterbank, L.; Lacote, E.; Malacria, M. Chem. Soc. Rev. 2005, 34, 858. (b) Barton,

D. H. R.; Jang, D. O.; Jaszbereny, J. C. Tetrahedron Lett. 1992, 33, 5704.

36 (a) Bellamy, A. J.; Campbell, E. A.; Hall, I. R. J. Chem. Soc., Perkin Trans. 2 1974, 1347. (b)

Yang, D.; Tanner, D.D. J. Org. Chem. 1986, 51, 2267.

37 Rodriguez-Solla, H.; Concellón, C.; Alvaredo, N.; Soengas, R.G. Tetrahedron 2012, 68, 1736.

38 Fan, R.; Pu, D.; Wen, F.; Ye, Y.; Wang, X. J. Org. Chem. 2008, 73, 3623.

39 Cano, R.; Ramón, D.J.; Yus, M. J. Org. Chem., 2011, 76, 5547.

40 Bennett, J.S.; Charles, K.L.; Miner, M.R.; Heuberger, C.F.; Spina, E.J.; Bartels, M.F.; Foreman,

T. Green Chem. 2009, 11, 166.

41 Brice, J.L.; Harang, J.E.; Timokhin, V.I.; Anastasi, N.R.; Stahl, S.S. J. Am. Chem. Soc. 2005,

127, 2868.

151

42 Matsubara, R.; Berthiol, F.; Kobayashi, S. J. Am. Chem. Soc. 2008, 130, 1804.

43 Malkov, A. V.; Vrankova, K.; Stoncius, S.; Kocovsky, P. J. Org. Chem. 2009, 74, 5839.

44 Matsubara, R.; Berthiol, F.; Kobayashi, S. J. Am. Chem. Soc. 2008, 130, 1804.

45 Goldberg, A.F.G.; O’Connor, N.R.; Craig, R.A.; Stoltz, B.M. Org. Lett. 2012, 14, 5314.

46 Mase, N.; Horibe, T. Org. Lett. 2013, 15, 1854.

47 Thaqi, A.; McCluskey, A.; Scott., J.L. Tet. Lett. 2008, 49, 6962.

48 Matsubara, R.; Berthiol, F.; Kobayashi, S. J. Am. Chem. Soc. 2008, 130, 1804.

49 Fredj, A.B.; Bagieu-Beucher, M.; Rejeb, S.B.; Lakhdar, Z.B. Helv. Chim. Acta 2004, 87, 1527.

50 Luo, F.; Pan, C.; Qian, P.; Cheng, J. J. Org. Chem. 2010, 75, 5379 .

51 Zhang, C.; Jiao, N. Angew. Chem. Int. Ed. 2010, 49, 6174.

52 Leyva, A.; Medina, C.; Moctezuma, E.; Leyva, E. Can. J. Chem. 2004, 82, 1712

53 Hu, L.; Cao, X.; Shi, L.; Qi, F.; Guo, Z.; Lu, J.; Gu, H. Organic Letters, 2011, 13, 5640.

54 Thorwirth, R.; Bernhardt, F.; Stolle, A.; Ondruschka, B.; Asghari, J. Chem. Euro. J. 2010, 16,

13236.

55 Ma, H.;Li, W.; Wang, J.; Xiao, G.; Gong, Y.; Qi, C.; Feng,Y.; Li, X.; Bao, Z.; Cao, W.; Sun, Q.;

Veaceslav, C.;Wang, F.; Lei, Z. Tetrahedron 2012, 68, 8358.

56 Sakai, N.; Asama, S.; Anai, S.; Konakahara, T. Tetrahedron 2014, 70, 2027.

57 Zhang, C.; Jiao, N. Angew. Chem. Int. Ed. 2010, 49, 6174.

58 Sakai, N.; Asama, S.; Anai, S.; Konakahara, T. Tetrahedron 2014, 70, 2027.

59 Leyva, A.; Medina, C.; Moctezuma, E.; Leyva, E. Can. J. Chem. 2004, 82, 1712.

60 Sakai, N.; Asama, S.; Anai, S.; Konakahara, T. Tetrahedron 2014, 70, 2027.

61 Muether, K.; Mohr, J.; Oestreich, M. Organometallics 2013 , 32, 6643.

62 Kiruthika, S. E.; Perumal, P.T. Org. Lett., 2014 , 16, 484.

63 Qin, C.; Shen, T.; Tang, C.; Jiao, N. Angew. Chem. Int. Ed. 2012 , 51, 6971

64 Wang, C.; Pettman, A.; Basca, J.; Xiao, J. Angew. Chem. Int. Ed. 2010 , 49, 7548.

65 Li, C.; Wang, C.; Villa-Marcos, B.; Xiao, J. J. Am. Chem. Soc. 2008 , 130, 14450.

66 Goldys, A. M.; Nez, M. G.; Dixon, D. J. Org. Lett. 2014 , 16, 6294.

152

67 Kan, R.O. J. Am. Chem. Soc. 1964, 86, 5180 .

68 Kurita, T.; Aoki, F.; Mizumoto, T.; Maejima, T.; Esaki, H.; Maegawa, T.; Monguchi, Y.; Sajiki, H. Chem. Eur. J. 2008, 14, 3371. 69 Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42.

Chapter 5

Regioselective Reductive Transposition of Allylic Bromides via P(III)/P(V) Redox Cycling

The work described in this chapter was performed by the author of this dissertation in collaboration with Kyle Reichl. Kyle studied phosphacycles as catalysts and optimized the catalytic reaction conditions. The author of this dissertation studied the electron deficient catalysts and performed all of the mechanistic studies.

Abstract

Chapter 3 describes efforts towards the development of hydridophosphoranes as reducing agents for unsaturated substrates. Based on these results, we sought to expand this reactivity to other systems, and were particularly interested in previous reports of a unique phosphorus-mediated allylic-reduction. Herein we report a phosphorus-catalyzed reductive transposition of allylic halides. We have further shown that this reaction proceeds through a hydridophosphorane intermediate. The catalytic reaction mechanism uniquely draws upon sequential transitions between σ3, σ4, and σ5 phosphorus species. This work expands the synthetic utility of hydridophosphoranes in phosphorus-catalyzed reductions.

5.1 Introduction

Phosphorus promoted allylic reductions have been reported in the literature several times over the past fifty years. Van Tamelen first published that the treatment of allylic phosphonium salts, such as 5-1, with lithium aluminum hydride (LAH) unexpectedly delivers the hydride γ to the phosphorus substituent. The result is a reductive allylic transposition, in this case providing 1,4-

154 dienes such as 5-2 in a convenient divergent synthesis from 1,5-dienes (Figure 5-1).1 In a related study by Tunemoto, the reduction of phosphonates was preceded by deprotonation and functionalization, to provide trans-olefins upon reduction (Figure 5-2).2 This allylic reduction methodology has since been used synthetically for allylic transposition in the formation of 5-7 where the reaction provides a stereoselective reduction for a study of the C ring of taxol (Figure 5-

3). 3

Figure 5-1. Allylic reduction of 5-1 to provide 1,4-diene 5-2.

Figure 5-2. α-functionalization and allylic reduction to provide trans-olefins.

Figure 5-3. Stereoselective allylic reduction.

155 Nojima and co-workers have published a comparative study of the reduction of allylic species, and have determined the effect of the leaving group on product distribution (Table 5-1).

They have found that cinnamyl bromide (5-8), chloride, and tosylate are reduced exclusively ɑ to the leaving group, to form β-methylstyrene (5-11).4 Cinnamyl ammonium salts (5-9) and sulfonium salts also are reduced predominantly at the ɑ position; however, cinnamyl phosphonates and phosphonium salts (5-10) are reduced selectively at the γ position to form allylbenzene (5-12).

Nojima also confirmed that the selectivity of the phosphonium reductions is maintained for a wide range of substrates.

Table 5-1. Products for the reduction of cinnamyl compounds with LAH.

The unique reactivity of allylic phosphonium salts has several potential explanations. One possibility is that electronic perturbations of the substrate alter the selectivity of hydride delivery: while most substituents produce SN2 reactivity, phosphorus derivatives provide SN2’ reactivity.

Alternatively, Gallagher has suggested that a similar reaction occurs via the formation of a hydridophosphorane. Specifically, the production of toluene (5-16) from benzylic phosphonium salt 5-13 is hypothesized to go through hydridophosphorane intermediate 5-14. Compound 5-14 then undergoes intramolecular hydrogen transfer from phosphorus to the proximal alkene (Figure

5-4).5 Although formation of hydridophosphoranes from phosphonium salts and metal hydrides

156 has precedent,6 no mechanistic evidence has been reported in allylic reductions of phosphonium salts.

Figure 5-4. α-functionalization and allylic reduction to provide trans-olefins.

From the reports of stoichiometric phosphorus-promoted allylic reductions, we envisioned a catalytic variant. The proposed catalytic cycle would commence with the reaction of phosphine

5-17 with allylic electrophile 5-18 to form allylic phosphonium salt 5-19. Compound 5-19 would react with a hydride donor to form hydridophosphorane 5-20, which would rearrange to form 5-21, the product of allylic reduction, and regenerate the phosphine catalyst. To design the desired catalytic system, a phosphine catalyst must be identified such that a phosphonium salt is formed rapidly via reaction of the phosphine with the substrate. Therefore, the phosphine must be nucleophilic. Additionally, the phosphonium salt must be reduced selectively in the presence of the allylic halide, which requires an electrophilic phosphonium species. Notably, this catalytic cycle would rely on sequential reactivities of three phosphorus species: 1) a nucleophilic phosphine,

2) an electrophilic phosphonium and 3) rearrangement of a phosphorane. Successful development of such a catalytic cycle would represent a new motif for phosphorus catalysis.

157

Figure 5-5. Proposed catalytic cycle for phosphine-catalyzed allylic reductions.

5.2 Optimization of Catalytic Reaction

To determine the ideal conditions for reductive transposition of allylic halides catalyzed by phosphines, we surveyed a variety of substrates, reductants, and phosphine catalysts.

5.2.1 Substrate Optimization

Allylic bromides were found to be ideal substrates for phosphorus-catalyzed allylic reduction, as has been previously reported.4 Cinnamyl derivatives were selected for screening in the presence of a stoichiometric amount of triphenylphosphine to determine successful conversion to triphenylphosphonium salt 5-22, the first step required for a catalytic allylic reduction.

Triphenylphosphonium salt 5-22 was obtained from cinnamyl chloride and cinnamyl bromide as

158 expected; however, cinnamyl iodide and cinnamyl tosylate substrates were unstable, and acetate and trifluoromethylacetate were unreactive as leaving groups (Figure 5-6).

Figure 5-6. Conversion of cinnamyl substrates to triphenylphosphonium salts is successful when X = Cl, Br. X = I, OTS are unstable, while X = OAc, OCOCF3 are unreactive.

To select between chloride and bromide substrates, the cinnamyl derivatives were screened under optimized catalytic reaction conditions (substrate and 10% catalyst 5-31 were heated in toluene at 90 °C for 16 h while a solution of LiAlH(OtBu)3 in tetrahydrofuran was added dropwise over 14 h). The undesired ɑ-reduction to 5-12 predominates for cinnamyl chloride (5-23, 85%), while preferred γ-reduction to 5-11 is observed in 94% for cinnamyl bromide (5-8, Table 5-2).

These results indicate that rapid conversion of allylic halides to phosphonium salts in the presence of phosphine is required for the desired transformation in this system, and therefore we focused our attention on optimizing the reaction for allylic bromides.

Table 5-2. Substrate optimization table with LiAlH(OtBu)3 and catalyst 5-31. Yield and product composition were determined by GC analysis against dodecane as an internal standard.

159 5.2.2 Reductant Optimization

Our initial aim was to find a reductant that reduced the phosphonium salt exclusively in the presence of cinnamyl bromide, which would allow for a simple one-pot reaction. When we screened the reaction of cinnamyl bromide (5-8) with various reductants in the absence of phosphine (0.1 mmol 5-8, 2.5 equivalents of reductant, 1 h, 60 °C, 1 mL 1:1 THF–PhMe), we found that aluminum hydrides (e.g., LAH, LiAlH(OtBu)3) react rapidly to provide 5-12, while a variety of borohydrides (sodium borohydride, sodium acetoxyborohydride, sodium triacetoxyborohydride, sodium trimethoxyborohydride) do not reduce 5-8 at all under these conditions. To determine if borohydride reductants were suitable for the reduction of allylic phosphonium salts, they were screened against electron poor phosphonium salt 5-31. No reduction was observed with any of the borohydrides.

Since reductants which are compatible with the cinnamyl bromide substrate are similarly unreactive with allylic phosphonium salts, focus shifted to slow addition of strong reductants.

Gratifyingly, Kyle Reichl found that slow addition of strong reductants provided optimal yield and selectivity for γ-reduction of allylic substrates. Slow addition maintains a low concentration of reductant compared to both substrate and phosphonium salts, and allows selective reduction of the most reactive electrophile in a complex mixture. In fact, screening of aluminum hydride reductants under standard conditions (cinnamyl bromide and 10% catalyst 5-31 were heated in toluene at 90

°C for 16 h while a solution of reductant in tetrahydrofuran was added dropwise over 14 h) shows that lithium aluminum hydride and LiAlH(OtBu)3 showed selectivity 94:6 for γ-reduction of cinnamyl bromide, while DIBAL-H was less selective. LiAlH(OtBu)3 was selected as the ideal reductant due to its selectivity for γ-reduction and improved solubility compared to LAH.

160

Figure 5-7. Selectivity of aluminum hydride reductants. Yield and product composition were determined by GC analysis against dodecane.

5.2.3 Catalyst Optimization

A variety of phosphine catalysts were screened for the reduction of cinnamyl bromide in toluene at 90 °C for 16 h while a solution of LiAlH(OtBu)3 in tetrahydrofuran was added dropwise over 14 h. Catalysts were typically screened as the corresponding allylic salts, which are air stable solids and therefore easier to handle than the often air-sensitive free phosphines. These allylic salts are reduced in the presence of the reductant to produce the free phosphine and propene.

To screen a range of phosphine catalysts, a panel of phosphines with both electronic and structural perturbations were examined (Table 5-3). We hypothesized that the selectivity of hydride transfer would be dictated primarily by the electrophilicity of the phosphonium intermediate of the reaction. To favor γ-selectivity, the phosphonium salt should be more electrophilic than the allylic bromide substrate, causing more rapid reduction of phosphonium salt compared to substrate. The electrophilicity of phosphonium salts should track well with the electronic properties of the parent

1 7 phosphine, which can be measured by the analysis of phosphine selenide JP-Se coupling. The

161

1 coupling constant has been used to judge the electronic character of phosphines, where a larger JP-

Se value indicates greater s-character of the lone-pair of electrons on phosphorus, and therefore a less nucleophilic phosphine.8 Using this information, we selected phosphines which covered a

1 1 wide range of JP-Se values. The most nucleophilic phosphine (i.e., with the smallest JP-Se values), trimethylphosphine 5-24, gave 5-11 to 5-12 selectivity of 11:89, which was nearly identical to the selectivity observed in an uncatalyzed reaction. We theorized that the α-reduction is favored because the derived phosphonium salts are not reduced selectively in the presence of substrate.

Increasing phosphine electrophilicity does increase the selectivity for the desired product to a point,

1 as demonstrated by the increasing ratio of 5-11 to 5-12 in concert with JP-Se values, up to 36:64 using catalyst 5-27. However, 5-11 never becomes the major product; in fact, the most electronegative phosphine tested, 5-28 is less selective. Compound 5-12 is observed as the major product likely because the rate of phosphine alkylation with the substrate is decreased, leaving free substrate to be reduced to 2.

Table 5-3. Reduction of cinnamyl bromide with various phosphine catalysts. Yield and product compositions were determined by GC analysis against dodecane as an internal standard.

162

Since electron-poor phosphines did not prove optimal catalysts, we sought other ways to improve the reactivity. Our proposed catalytic cycle involves the transition from phosphonium salts to hydridophosphoranes. Therefore, we focused on structural effects to promote this transition. For example, Abe and coworkers reported that the ortho-functionalization of aryl substituents at phosphorus was crucial for rapid nucleophilic addition to a phosphorane center via neighboring group participation (Figure 5-8).9 We had hypothesized that the interaction between

163 a proximal donating group and the positive charge at phosphorus would reorganize a tetrahedral phosphonium center into a five coordinate trigonal bipyramidal species, and therefore allow facile displacement of the donor group by a nucleophile. To test this principle in our system, we synthesized a triphenylphosphonium catalyst with a single o-methoxymethyl ether substituent (5-

25). The ether provides a donor oxygen to facilitate the formation of a five-membered phosphacycle, but does not contain amide or ester groups that would be reduced under the reaction conditions. Unfortunately, 5-25 was less effective than triphenylphosphine (5-24). The ortho- donating group does not improve selectivity, and the catalyst instead follows the general trend for nucleophilicity, wherein the substituent provides a more electron-rich phosphorus center, as

1 observed by a smaller JP-Se value and corresponding lower yield of 5-11.

Figure 5-8. Use of an ortho-amide to increase rate of iminophosphorane hydrolysis.

When electronic perturbations and neighboring group participation proved to be ineffective strategies for catalytic allylic reductions, we instead turned to the unique reactivity of cyclic phosphines, or phosphacycles. Since four-coordinate cyclic phosphorus compounds are distorted closer to the required trigonal bipyramidal structure of a hydridophosphorane intermediate, the

164 reaction barrier to forming this intermediate should be decreased. We theorized that this strategy could also be used for intermolecular hydride transfer to phosphonium salts. In fact, previous research has shown that phosphacyclic P(V) compounds have enhanced reactivity with nucleophiles.10

More recently, phosphacycles have been featured in phosphine oxide catalysis, which relies upon in situ reduction of phosphine oxides to regenerate phosphines.11 The phosphine oxides of small rings are reduced more readily than acyclic species.12 Phosphine oxide reduction by silanes proceeds first by association of silicon and oxygen, which leads to the formation of silanol and free phosphine.13 Computational modeling suggests that a hydridophosphorane intermediate is formed, and that intramolecular hydride transfer from silicon to phosphorus is the rate limiting step of the reaction.14 The addition of a hydride to a tetracoordinate phosphorus complex is also proposed for the catalytic transpositive reduction of allylic halides, therefore we screened phosphacycles as catalysts.

Compared to acyclic phosphine catalysts, which provide 5-12 preferentially, several phosphacycles reduce cinnamyl bromide selectively to 5-11 (Table 5-4). Aryl phospholane 5-29 provides a ratio of 5-11:5-12 of 19:81; however, phospholane 5-30, which replaces two aryl groups with alkyl groups provides the opposite selectivity: 54:46. Increasing ring strain with phosphetane catalyst 5-31 gives excellent selectivity for the γ-reduction: 94:6 in favor of 5-11. The inversion of selectivity between cyclic and acyclic phosphine catalysts suggests that increasing ring strain lowers the barrier of hydride addition to phosphonium salts. This geometric reorganization allows the preferential reduction of phosphonium salts compared to the substrate.

Table 5-4. Reduction of cinnamyl bromide with various phosphine catalysts. Yield and product composition were determined by GC analysis against dodecane as an internal standard.

165

Competition experiments confirm that cinnamyl phosphetanium salt 5-32 is an effective catalyst because it is selectively reduced by LiAlH(OtBu)3 in the presence of cinnamyl bromide

(Figure 5-9). Control experiments have shown that 5-11 results predominantly from the reaction with the phosphetanium salt 5-32, while the reduction of 5-8 yields 5-12 exclusively. When 5-32 and 5-8 were mixed in equimolar quantities with limiting reductant, the γ-reduction was the predominant product, composing 94%, of the product mixture, while only 6% ɑ-reduction was observed by GC analysis.

166

Similarly, allyl phosphetanium salt 5-32 proved to be the preferred hydride acceptor compared to an acyclic phosphine with similar substituents, allyldibutylphenylphosphonium salt 5-

33 (Figure 5-10). An equimolar mixture of the two salts with a limiting amount of reductant was analyzed by 31P NMR; only the reduced phosphetane 5-35 was observed in 92% yield. No dibutylphenylphosphine 5-34 was observed, indicating that 5-33 was not reduced at all in the presence of 16. This result indicates that the small ring size of the phosphacycle has a considerable impact on the electrophilicity of the phosphetanium salt.

5.3 Mechanistic Studies

The catalytic reduction of allylic substrates with a phosphine catalyst and reductant has several mechanistic possibilities. Initially, the phosphine catalyst will react with cinnamyl bromide to form a phosphonium salt, 5-32. This phosphonium salt then could be reduced via an SN2’ reduction with the reductant (Figure 5-11, Path A), or the hydride could be transferred directly to the phosphorus to produce a hydridophosphorane intermediate, 5-37, which then undergoes intramolecular hydride transfer to produce 5-11 and 5-35 (Figure 5-11, Path B). To differentiate between these possibilities, isotopic labeling, spectroscopic studies, kinetics and computational experiments were performed.

Figure 5-11. Potential mechanisms for the reduction of allylic phosphonium salts, showing SN2’ reduction (Path A) or intramolecular rearrangement of hydridophosphorane intermediate 5-37 (Path B).

168 5.3.1 NMR Studies

To investigate the possible mechanisms further, we focused on studying the reaction of allylphosphetanium salt 5-40 with hydride by 31P NMR. Tetraphenylborate was selected as the counterion for 5-40 to ensure full solubility of the salt in THF. At room temperature, the reduction is extremely rapid and no intermediates are observed. However, cooling the reaction down to –70

°C allowed us to observe a doublet at –78 ppm, J = 243 Hz (Figure 5-12). The chemical shift of the new signal is in agreement with a pentacoordinate phosphorus atom, and the doublet likely

1 arises from JP-H coupling. While this coupling value can range from 100-1000 Hz, alkyl substituted phosphorus compounds fall on the lower end of that spectrum.15 This NMR signal is attributed to the hydridophosphorane intermediate 5-37. As the sample was brought up to 0 °C, only free phosphetane 5-35 was observed, indicating the reaction was completed upon warming.

169

5.3.2 Synthesis and Characterization of Isolable Hydridophosphorane

To confidently identify the intermediate observed in the low temperature trials, a fully saturated phosphetanium salt, 5-41, was synthesized. In this system, hydride transfer to the positively charged phosphorus center can occur to form a hydridophosphorane, but the hydridophosphorane cannot undergo further isomerization as observed in unsaturated species.

Phosphetanium salt 5-41 was treated with LiAlH4 in pentane with 0.5% THF at room temperature; after 30 minutes, the reaction mixture was filtered. The filtrate was concentrated to give a colorless oil in 35% yield which we have identified as hydridophosphorane 5-42.

170 The 31P NMR spectra of 5-42 shows signals for two diastereomers at –81 and –76.5 ppm,

1 which are proton-coupled doublets with JP-H coupling constants of 248 and 239 Hz respectively

(Figure 5-13). Each signal resolves to a singlet in 31P{1H} NMR, confirming that the profile is due to phosphorus-hydrogen coupling. Additionally, the hydride signal is seen in 1H NMR at 5.78 and

5.87 ppm, again appearing as doublets with coupling constants of 248 and 240 Hz. The coupling constants observed typically indicate an axial P-H bond, whereas equatorial P-H bonds have larger coupling constants.16 However, phosphoranes which contain phosphetanes are distorted towards

1 square pyramidal geometries, making it difficult to determine hydrogen position by JP-H coupling constants alone.17 High resolution mass spectrometry via electron impact ionization confirms the expected mass of 312.2007. Compound 5-42 is not crystalline, precluding further structural study via X-ray diffraction. The phosphorus signals of hydridophosphorane 5-42 coincide well with the signals we observed in low temperature NMR, supporting our initial hypothesis that we could observe the hydridophosphorane intermediate 5-37 in situ.

171

B) C)

Figure 5-13. A) Synthesis and characterization of stable hydridophosphorane 5-42 (major diastereomers of salt and phosphorane shown). 31P NMR of 5-42 B) decoupled and C) coupled to hydrogen.

To confirm the reactivity of the stable hydridophosphorane, we reacted 5-42 with

phosphonium salt 5-32 in d6-benzene (Figure 5-14). After 1 h, the starting materials had been

largely consumed and free phosphine 5-35 was observed, along with saturated phosphonium salt

5-41. Allylbenzene 5-11 was observed in the 1H NMR. These results suggest that the

hydridophosphorane is a competent intermolecular reductant for the phosphonium salts.

172 Figure 5-14. Intermolecular hydride transfer observed from hydridophosphorane 5-42.

5.3.3 Deuterium Labeling Studies

We performed deuterium labeling studies for the conversion of phosphetanium salt 5-38 to

2-allylnaphthalene 5-39 in order to track the incorporation of the hydride. The reduction of 5-38

γ with lithium aluminum deuteride (LiAlD4) provided species 5-39-d with deuterium incorporation exclusively at the γ position (Figure 5-15A). This result is in accord with Nojima’s previous study.4

ɑ Studies with phosphonium salt 5-38-d2 confirm that the phosphonium salt is not deprotonated

ɑ under the reaction conditions, as observed by the isolation of product 5-39-d2 , which has not undergone any H/D scrambling at the terminal position of the alkene (Figure 5-15B). This observation eliminates the possibility of ylide formation under the reaction conditions, which is observed with other phosphonium salts and hydrides,18 but would not lead to the observed products.

Both results are in accord with the formation and rearrangement of a hydridophosphorane species.

Figure 5-15. Deuterium studies show A) delivery of hydride selectively to the γ-position, and B) no scrambling of α-protons under the reaction conditions. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society.

173 5.3.4 Pulse-Chase Studies

To determine if the γ-hydrogen in the product comes from the hydridophosphorane intermediate or the reductant, phosphetanium salt 5-38 was treated first with 5 equivalents of

LiAlD4 at –78 °C (excess reductant is required to fully convert the phosphonium to the phosphorane) (Figure 5-16A). Low temperature NMR confirmed that all salt had been converted to phosphorane. Next, another 5 equivalents of LiAlH4 was added at –78 °C. The reaction was warmed to room temperature, concentrated, and purified via column chromatography to provide the allyl product 24 with 66% D content. The incorporation of 1H into the γ-position indicates that the reductant may be involved in the conversion of hydridophosphorane to products.

To probe this pulse-chase reaction further, fully saturated phosphonium salt 5-41

P underwent the same treatment as reported above: first, excess LiAlD4 was added to form 5-42-d , then LiAlH4 was added (Figure 5-16B). The reaction was warmed in a thermostatted NMR at 20

°C intervals and monitored by 1H NMR for the appearance of the P-H bond. A fully saturated salt was chosen to avoid overlap with vinylic protons, as well to allow for easier observation at higher temperatures without conversion to the intermediate of interest. At –60 °C, the P-H bond was indeed observed, and, at 0 °C, was integrated to 15% of phosphorane 5-42. This experiment shows that hydride exchange can occur between phosphorus and aluminum, and that the pulse-chase experiment is therefore not valid for determining the source of hydrogen in this system.

174 Figure 5-16. Pulse-chase studies show exchange between aluminum and phosphorous hydrides.

5.3.5 Conversion of Hydridophosphorane to Products

To further study the conversion of hydridophosphorane intermediate to phosphine and allylbenzene, kinetic experiments were performed. Rates for the transformation of 5-43 to 5-35 and propene were collected at temperatures between –45 and –70 °C (Figure 5-17). Linear regression analysis of the Erying plot provides activation parameters for the transformation of ΔH‡

= +15.2(0.8) kcal/mol, and ΔS‡ = –3(4) cal/mol*K. The negative activation entropy indicates an ordered transition structure, such as would be expected for an intramolecular reaction.

Intramolecular pericyclic reactions typically have less negative activation entropies than their intermolecular counterparts due to pre-association of reactant moieties.19 The activation enthalpy compares well to value determined for this system computationally: 11.9 kcal/mol. DFT calculations (M06-2X/6-311++G(2d,2p)), which were constrained by maintaining the phosphetane ring in an axial-equatorial configuration, show that the lowest-energy transition state involves both

P-C bond breaking and C-H bond formation in the same step (Figure 5-18). The transformation takes place via a square pyramidal transition state, with the phenyl group in the apical position.

175

-10 4.30E-034.40E-034.50E-034.60E-034.70E-034.80E-034.90E-035.00E-03 -11

-12

-13 ln(k/T)

-14

-15 y = -7633.9x + 22.254 R² = 0.9662 -16 1/T (1/K)

5.4 Conclusions

In summary, a catalytic transpositive reduction of allylic bromides has been developed.

This work has revealed a novel type of phosphorus catalysis, where three-coordinate, four- coordinate and five-coordinate phosphorus states are accessed in sequence, and adds to a small

176 repertoire of reagents and catalysts that can effect such a transformation.20,21,22 The ideal catalyst for this transformation is a phosphetane, and the structural and electronic perturbations caused by the ring size are crucial to optimal reactivity. The reaction proceeds first by the reaction of the phosphetane and substrate to form a phosphetanium salt. Through mechanistic studies, we have determined that a hydride source then reacts with the phosphetanium salt to form a five coordinate hydridophosphorane, and computational studies suggest that the transfer of hydrogen from phosphorus to carbon occurs via a concerted transition state. We anticipate that this mode of reactivity, which proceeds through phosphine, phosphonium, and phosphorane species, will prove useful to expanding the role of phosphorus-catalyzed bond formation.

5.5 Experimental

5.5.1 General Methods and Materials

All reactions were carried out under an N2 atmosphere using dry glassware and standard

Schlenk techniques. Column chromatography was performed using 230-400 mesh silica gel purchased from Silicycle as the stationary phase. All NMR spectra were obtained in CDCl3, CD2Cl2

1 or C6D6 using a Bruker DPX-300, AMX-360, DRX-400, or AVIII-HD500 spectrometer. H spectra were referenced to an internal TMS standard, 13C spectra to the internal solvent peak (δ 77.16 ppm

31 for CDCl3, δ 49.00 ppm for CD3OD, δ 53.84 ppm for CD2Cl2) and P spectra to an external sample of 85% H3PO4 (δ 0.0 ppm). High resolution EI and ESI mass spectra were obtained from the Mass

Spectrometry Laboratory at the School of Chemical Sciences, University of Illinois at Urbana-

Champaign.

All reagents were purchased through Sigma-Aldrich, Alfa Aesar, or Oakwood Chemical, and used as received unless otherwise noted. Dimethylphenylphosphine,23 1-phenylphospholane,24

177 P-phenyl-5H-dibenzophosphole,25 1-ethyl-2,2,3-trimethylphosphetane26, and their quaternization to allylic salts were prepared based on previous literature procedures.25

Gas chromatography was performed on a Shimadzu GC-2010 equipped with an FID detector, AOC-20i autoinjector, and Shimadzu SHRXI-5MS capillary column (30 m x 0.25 mm x

0.25 μm). Quantitative measurements were calibrated with samples of authentic analytes.

5.5.2 Synthetic Procedures

5.5.2.1 Representative Procedure for the Synthesis of Phosphetanium Bromide Salts (RP1)

1-allyl-2,2,3-trimethyl-1-phenylphosphetanium bromide (5-31): To a

solution of 1-phenyl-2,2,3-trimethylphosphetane-1-oxide27 (~5:1

trans:cis) (1.41 g, 6.8 mmol) in toluene (25 mL) was added triethylamine

(1.0 mL, 7.0 mmol, 1.0 eq.) and trichlorosilane (0.7 mL, 6.9 mmol, 1.0 eq.) at ambient temperature.

The resulting mixture was then stirred at 80 °C overnight. After cooling in an ice bath, the reaction was quenched with 10% w/w aqueous NaOH solution (40 mL), and the aqueous layer extracted with additional toluene. The combined organic layers were washed with saturated aqueous

NaHCO3 and NaCl solutions and dried over anhydrous MgSO4. Filtration and removal of solvent yielded crude phosphetane as a mixture of diastereomers, which was immediately dissolved in allyl bromide (5 mL). Stirring overnight at ambient temperature yielded a thick precipitate, which was collected by filtration and washed with diethyl ether to afford the desired product as a white solid

(1.52 g, 4.8 mmol, 72%). An approximately 2.6:1 trans:cis diastereomeric mixture was used in all catalytic experiments. Recrystallization from methanol/diethyl ether can be used to provide a single diastereomer, which was unambiguously assigned as trans via x-ray crystallography. 1H NMR

(CD3OD, 400 MHz): δ = 1.14 (d, JP-H = 21 Hz, 3H, CH3), 1.24 (d, J = 6.7 Hz, 3H, CH3), 1.66 (d, J

178

= 21 Hz, 3H, CH3), 2.74-2.88 (m, 1H), 2.95-3.09 (m, 1H), 3.26-3.38 (m, 1H), 3.82 (dd, J = 15, 7.4

Hz, 2H, PCH2), 5.27-5.43 (m, 2H, C=CH2), 5.53-5.67 (m, 1H, C=CH), 7.75-7.83 (m, 2H, Ar), 7.84-

7.91 (m, 1H, Ar), 8.04-8.14 (m, 1H, Ar). Minor (cis) diastereomer: 1.19 (d, J = 21 Hz, 3H, CH3),

1.57 (d, J = 21 Hz, 3H, CH3), 3.90 (dd, J = 15, 7.3 Hz, 2H, PCH2), 5.68-5.78 (m, 1H, C=CH), 7.92-

13 8.00 (m, 2H, Ar); C NMR (CD3OD, 101 MHz): 15.5 (d, JP-C = 24 Hz), 17.5 (d, JP-C = 1.2 Hz),

23.3 (d, JP-C = 46 Hz), 24.1 (d, JP-C = 4.0 Hz), 28.4 (d, JP-C = 25 Hz), 40.9 (d, JP-C = 16 Hz), 47.8 (d,

JP-C = 44 Hz), 117.1 (d, JP-C = 63 Hz), 124.5 (d, JP-C = 12 Hz), 125.1 (d, JP-C = 12 Hz), 131.3 (d, JP-

C = 12 Hz), 134.2 (d, JP-C = 9.0 Hz), 136.0 (d, JP-C = 3.0 Hz). Minor (cis) isomer: 15.4 (d, JP-C = 24

Hz), 18.5 (d, JP-C = 2.5 Hz), 22.6 (2 unresolved resonances), 25.4 (d, JP-C = 32 Hz), 39.9 (d, JP-C =

16 Hz), 47.5 (d, JP-C = 45 Hz), 118.7 (d, JP-C = 56 Hz), 124.5 (d, JP-C = 11 Hz), 125.1 (d, JP-C = 12

31 Hz), 131.2 (d, JP-C = 12 Hz), 133.8 (d, JP-C = 9.7 Hz), 136.2 (d, JP-C = 3.2 Hz); P NMR (CD3OD,

+ 146 MHz): 46.8 (cis, minor), 48.9 (trans, major); HRMS (ESI) Calcd. For [C15H22P ]: 233.1459,

Found: 233.1454.

2,2,3-trimethyl-1-phenyl-1-(3-phenylpropyl)phosphetanium

bromide (5-41): Prepared using RP1 on a 36 mmol scale. The crude

phosphetane was diluted with toluene (20 mL), then 1-bromo-3- phenylpropane (22 mL, 144 mmol) was added and heated to 90 °C for 8 h under a nitrogen atmosphere. Filtration afforded the desired product as a white solid (7.69 g, 55% yield over 2 steps).

Recrystallization from methanol/ether at room temperature yielded a single diastereomer for

1 characterization. H NMR (CDCl3, 360 MHz): δ = 1.11-1.30 (m, 6H), 1.53-1.69 (m, 1H), 1.69 (d,

J = 21 Hz, 3H, CH3), 1.77-1.93 (m, 1H), 2.65-2.87 (m, 3H), 3.25 (t, J = 11 Hz, 3H), 3.45-3.62 (m,

13 1H), 7.07-7.25 (m, 5H, Ar), 7.66-7.82 (m, 3H, Ar), 8.13-8.26 (m, 2H, Ar); C NMR (CDCl3, 101

MHz): 14.9 (d, JP-C = 8.7 Hz), 15.1 (d, JP-C = 9.3 Hz), 17.3, 22.2 (d, JP-C = 26 Hz), 23.1 (d, JP-C = 44

Hz), 23.7 (d, JP-C = 6.4 Hz), 35.5 (d, JP-C = 14 Hz), 39.8 (d, JP-C = 16 Hz), 46.1 (d, JP-C = 47 Hz),

115.2 (d, JP-C = 62 Hz), 126.1, 128.2, 130.0 (d, JP-C = 12 Hz), 132.8, 132.8 (d, JP-C = 20 Hz), 134.3,

179

31 + 139.7; P NMR (CDCl3, 146 MHz): 52.5; HRMS (ESI) Calcd. For [C21H28P ]: 311.1929, Found:

311.1923.

2,2,3-trimethyl-1-phenyl-1-(3-phenylprop-2-enyl)phosphetanium

bromide (5-40): Prepared using RP1 on a 10 mmol scale. The crude

phosphetane was diluted with toluene (5 mL), then 1-bromo-3- phenylprop-2-ene (2.0 g, 10 mmol) was added and heated to 90 °C for 8 h under a nitrogen atmosphere. Filtration afforded the desired product as a white solid (1.20 g, 31% yield over 2 steps).

Recrystallization from methanol/ether at room temperature yielded a single diastereomer for

1 characterization. H NMR (CDCl3, 360 MHz): δ = 1.23 (d, J = 6.7 Hz, 3H, CH3), 1.30 (d, J = 21

Hz, 3H, CH3), 1.90 (d, J = 21 Hz, 3H, CH3), 2.58-2.75 (m, 1H), 3.08-3.24 (m, 1H), 3.44 (q, J = 13

Hz, 1H), 4.36-4.50 (m, 1H), 4.55-4.71 (m, 1H), 5.77-5.92 (m, 1H, C=CH), 6.82 (dd, J = 16, 5.8 Hz,

1H, C=CH), 7.18 (s, 5H, Ar), 7.64-7.78 (m, 3H, Ar), 8.26 (dd, J = 12, 7.1 Hz, 2H, Ar); 13C NMR

(CDCl3, 101 MHz): 15.2 (d, JP-C =24 Hz), 17.6, 23.0 (d, JP-C = 46 Hz), 24.1 (d, JP-C = 3.8 Hz), 28.9

(d, JP-C = 23 Hz), 39.6 (d, JP-C = 15 Hz), 47.5 (d, JP-C = 43 Hz), 113.7 (d, JP-C = 13 Hz), 115.8 (d, JP-

C = 62 Hz), 126.3 (d, JP-C = 1.9 Hz), 128.1, 128.4, 130.0 (d, JP-C = 12 Hz), 133.3 (d, JP-C = 8.9 Hz),

31 134.4 (d, JP-C = 2.9 Hz), 135.6 (d, JP-C = 4.2 Hz), 138.8 (d, JP-C = 13 Hz); P NMR (CDCl3, 146

+ MHz): 48.5; HRMS (ESI) Calcd. For [C21H26P ]: 309.1772, Found: 309.1777.

2,2,3-trimethyl-1-phenyl-1-(3-naphthylprop-2-enyl-1-)

phosphetanium bromide (5-38): Prepared using RP1 on a

10 mmol scale. The crude phosphetane was diluted with 5 mL toluene, then (E)-2-(3-bromoprop-1-en-1-yl)naphthalene (1.12 g, 4.5 mmol) was added and heated to 80 °C for 8 h under a nitrogen atmosphere. Filtration afforded the desired product as a white

1 solid, (1.35 g, 68% yield). H NMR (CDCl3, 360 MHz): δ = 1.23 (d, J = 6.7 Hz, 3H, CH3), 1.30 (d,

J = 21 Hz, 3H, CH3), 1.80 (d, J = 21 Hz, 3H, CH3, minor isomer), 1.95 (d, J = 21 Hz, 3H, CH3),

2.56-2.70 (m, 1H), 3.00-3.14 (m, 1H), 3.39-3.56 (m, 1H), 4.50-4.64 (m, 1H), 4.66-4.80 (m, 1H),

180 5.84-5.99 (m, 1H, C=CH), 6.99 (dd, J = 16, 5.0 Hz, 1H, C=CH), 7.25-7.31 (m, 1H, Ar), 7.35-7.42

13 (m, 2H, Ar), 7.52 (s, 1H, Ar), 7.54-7.77 (m, 6H, Ar), 8.17-8.27 (m, 2H, Ar); C NMR (CDCl3, 75

MHz): 15.5 (d, JP-C = 24 Hz), 17.8, 23.1 (d, JP-C = 46 Hz), 24.3 (d, JP-C = 4.2 Hz), 29.2 (d, JP-C = 23

Hz), 39.8 (d, JP-C = 15 Hz), 47.8 (d, JP-C = 43 Hz), 114.0 (d, JP-C = 13 Hz), 116.1 (d, JP-C = 62 Hz),

123.0, 126.3, 126.4, 127.1 (d, JP-C = 3.3 Hz), 127.6, 128.1, 128.2, 130.2 (d, JP-C = 12 Hz), 133.1,

133.2, 133.3 (d, JP-C = 8.9 Hz), 134.5 (d, JP-C = 3.2 Hz), 139.1 (d, JP-C = 13 Hz) (1 resonance

31 unresolved); P NMR (CDCl3, 146 MHz): 48.3 (major), 47.2 (minor); HRMS (ESI) Calcd. For

+ [C25H28P ]: 359.1929, Found: 359.1932.

2,2,3-trimethyl-1-phenyl-1- (3-naphthylprop-2-enyl-1-d2)

α phosphetanium bromide (5-38-d2 ): Prepared using RP1 on

a 10 mmol scale. The crude phosphetane was diluted with toluene (5 mL), then S26 (0.74 g, 3 mmol) was added and heated to 80 °C for 8 h under a nitrogen atmosphere. Filtration afforded the desired product as a white solid, which is a mixture of

1 diastereomers (0.965 g, 73% yield). H NMR (CDCl3, 360 MHz): δ = 1.23 (d, J = 6.7 Hz, 3H, CH3),

1.30 (d, J = 21 Hz, 3H, CH3), 1.80 (d, J = 21 Hz, 3H, CH3, minor isomer), 1.95 (d, J = 21 Hz, 3H,

CH3), 2.56-2.70 (m, 1H), 3.00-3.14 (m, 1H), 3.39-3.56 (m, 1H), 5.92 (dd, J = 16, 5.6 Hz, 1H,

C=CH), 7.00 (dd, J = 16, 5.9 Hz, 1H, C=CH), 7.25-7.31 (m, 1H, Ar), 7.33-7.42 (m, 2H, Ar), 7.52

13 (s, 1H, Ar), 7.54-7.77 (m, 6H, Ar), 8.17-8.27 (m, 2H, Ar); C NMR (CDCl3, 101 MHz): 15.4 (d,

JP-C = 24 Hz), 17.8, 23.1 (d, JP-C = 46 Hz), 24.3, 39.9 (d, JP-C = 15 Hz), 47.7 (d, JP-C = 43 Hz), 114.0

(d, JP-C = 13 Hz), 116.1 (d, JP-C = 62 Hz), 123.0, 126.2, 126.4, 127.0, 127.6, 128.1, 128.2, 130.2 (d,

JP-C = 12 Hz), 133.1, 133.2, 133.4 (d, JP-C = 8.9 Hz), 134.6, 139.1 (d, JP-C = 13 Hz) (2 resonances

31 unresolved); P NMR (CDCl3, 146 MHz): 48.3 (major), 47.0 (minor); HRMS (ESI) Calcd. For

+ [C25H26D2P ]: 361.2054, Found: 361.2049.

181 5.5.2.2 Representative Procedure for the Synthesis of Phosphetanium Tetraphenylborate Salts (RP2)

1-allyl-2,2,3-trimethyl-1-phenylphosphetanium tetraphenylborate

(5-41-BPh4): Phosphetanium salt 5-31 (0.936 g, 3.0 mmol) was

dissolved in water (30 mL), sodium tetraphenylborate (1.13 g, 3.3 mmol, 1.1 equiv) was added and the solution was stirred rapidly at ambient temperature for 30 minutes. White precipitate formed immediately and was filtered, washed with water (50 mL) and

1 dried to give the product as a white powder (1.32 g, 80%). H NMR (CD2Cl2, 360 MHz): δ = 1.08

(d, J = 21 Hz, 3H, CH3), 1.13 (d, J = 6.8 Hz, 3H, CH3), 1.36 (d, J = 21 Hz, 3H, CH3), 2.00-2.85 (m,

5H), 5.15-5.45 (m, 3H), 7.03 (t, J = 7.1 Hz, 4H, Ar), 7.17 (t, J = 7.3 Hz, 8H, Ar), 7.28-7.37 (m, 2H,

13 Ar), 7.46-7.57 (m, 8H, Ar), 7.65-7.85 (m, 3H, Ar); C NMR (CD2Cl2, 125 MHz): major peaks:

15.3 (d, JP-C = 24 Hz), 18.5 (d, JP-C = 2.6 Hz), 22.2 (d, JP-C = 45 Hz), 22.6 (d, JP-C = 3.9 Hz), 25.0 (d,

JP-C = 31 Hz), 39.2 (d, JP-C = 16 Hz), 46.4 (d, JP-C = 44 Hz), 116.4 (d, JP-C = 56 Hz), 122.2, 122.6 (d,

JP-C = 11 Hz), 125.7 (d, JP-C = 7.9 Hz), 126.1, 130.7 (d, JP-C = 12 Hz), 131.9 (d, JP-C = 9.2 Hz), 135.7

31 (d, JP-C = 3.2 Hz), 136.3, 164.3 (q, JB-C = 49 Hz); P NMR (CD2Cl2, 146 MHz): 46.2 (minor), 44.8

(major).

2,2,3-trimethyl-1-phenyl-1-(3-phenylprop-2-enyl)phosphetanium

tetraphenylborate (5-40): Prepared using RP2 from a single

diastereomer of 5-32 (0.33 g, 0.86 mmol). Product was isolated as a

1 white solid (0.291 g, 54% yield). H NMR (CD2Cl2, 400 MHz): δ = 0.98-1.15 (m, 6H), 1.40 (d, J =

21 Hz, 3H, CH3), 2.12 (q, J = 14 Hz, 1H), 2.43-2.64 (m, 2H), 2.80-3.00 (m, 2H), 5.47-5.62 (m, 1H,

C=CH), 6.37 (dd, J = 16, 4.6 Hz, 1H, C=CH), 6.87-6.99 (m, 4H, Ar), 7.07 (t, J =7.0 Hz, 8H, Ar),

7.21-7.29 (m, 2H, Ar), 7.31-7.38 (m, 3H, Ar), 7.39-7.50 (m, 10 H, Ar), 7.62-7.71 (m, 2H, Ar), 7.75-

13 7.88 (m, 1H, Ar); C NMR (CD2Cl2, 125 MHz): 15.5 (d, JP-C = 23 Hz), 17.9 (d, JP-C = 1.7 Hz), 23.0

(d, JP-C = 45 Hz), 24.1 (d, JP-C = 4.0 Hz), 27.7 (d, JP-C = 24 Hz), 40.0 (d, JP-C = 16 Hz), 47.0 (d, JP-C

182

= 43 Hz), 112.2 (d, JP-C = 12 Hz), 114.8 (d, JP-C = 62 Hz), 122.3, 126.1 (q, JB-C = 2.5 Hz), 126.9 (d,

JP-C = 1.9 Hz), 129.1, 129.3, 130.9 (d, JP-C = 12 Hz), 132.5 (d, JP-C = 8.5 Hz), 135.5 (d, JP-C = 4.2

31 Hz), 135.7 (d, JP-C = 2.9 Hz), 136.4, 140.0 (d, JP-C = 12 Hz), 164.4 (q, JB-C = 49 Hz); P NMR

(CD2Cl2, 146 MHz): 45.8.

allyldibutylphenylphosphonium tetraphenylborate (5-33): Following

literature precedent,28 to dichlorophenylphosphine (2.0 mL, 15 mmol) in

ether at -78 °C under a nitrogen atmosphere was added n-BuLi (12 mL,

2.5 M in hexanes, 30 mmol) dropwise. The reaction was stirred at room temperature overnight, then filtered under a nitrogen atmosphere and the filtrate concentrated to yield crude dibutylphenylphosphine as a colorless oil. To this oil was added neat allyl chloride (4 mL, 45 mmol) under a nitrogen atmosphere and stirred at room temperature overnight, then triturated with pentane to give the product 5-33 as a white solid. 5-33 (0.500 g, 1.7 mmol) was converted via RP2 in 2% acetone/water (50 mL) to give desired product as a white solid (0.84 g,

1 85%). H NMR (CD2Cl2, 360 MHz): δ = 1.00 (t, J = 7.2 Hz, 6H), 1.23-1.40 (m, 4H), 1.40-1.54 (m,

4H), 1.81-1.98 (m, 4H), 2.52 (dd, J = 15, 6.4 Hz, 2H), 5.25-5.48 (m, 3H), 7.00 (t, J = 7.1 Hz, 4H,

Ar), 7.14 (t, J = 7.2 Hz, 8H, Ar), 7.30-7.40 (m, 2H, Ar), 7.44-7.55 (m, 8H, Ar), 7.62-7.72 (m, 2H,

13 Ar), 7.82 (t, J = 7.5 Hz, 1H, Ar); C NMR (CD2Cl2, 101 MHz): 13.5, 19.1 (d, JP-C = 49 Hz), 23.6,

24.0 (d, JP-C = 15 Hz), 25.1 (d, JP-C = 49 Hz), 116.1 (d, JP-C = 80 Hz), 122.3, 122.7 (d, JP-C = 12 Hz),

125.6 (d, JP-C = 12 Hz), 126.1, 130.9 (d, JP-C = 12 Hz), 131.7 (d, JP-C = 8.5 Hz), 135.5, 136.4, 164.5

31 + (q, JB-C = 49 Hz); P NMR (CD2Cl2, 146 MHz): 26.6; HRMS (ESI) Calcd. For [C17H28P ]:

263.1929, Found: 263.1934.

2,2,3-trimethyl-1-hydrido-1-phenyl-1-(3-phenylpropyl)-λ5-

phosphetane (5-42): To a suspension of phosphonium salt 5-41

(2.09 g, 5.34 mmol) in pentane (150 mL) and THF (1 mL) under a nitrogen atmosphere was added LiAlH4 (0.244 g, 6.4 mmol) and the resulting mixture stirred for

183 0.5 h at ambient temperature. The salt dissolved, and the solution was filtered over a pad of celite under a nitrogen atmosphere, then concentrated. The concentrate was redissolved in pentane and again filtered over celite under a nitrogen atmosphere and concentrated to give a clear, tacky liquid

1 which is a 1:4 mixture of diastereomers (0.588 g, 35% yield). H NMR (CD3CN, 360 MHz): δ =

0.86-1.19 (m, 9H), 1.50-1.70 (m, 3H), 1.78-1.90 (m, 1H), 2.00-1.25 (m, 2H), 2.45-2.95 (m, 3H),

5.42 (1H, d, J = 248 Hz. 1H, major isomer), 5.48 (dd, J = 240, 3.1 Hz, 1H, minor isomer), 7.03 (d,

J = 6.8 Hz, 2H, Ar), 7.13 (t, J = 7.3 Hz, 1H, Ar), 7.18-7.25 (m, 2H, Ar), 7.30-7.55 (m, 3H, Ar),

13 7.64-7.74 (m, 2H, major isomer, Ar), 7.83-7.92 (m, 2H, minor isomer, Ar); C NMR (CD3CN, 101

MHz, predominantly major diastereomer): 15.4 (d, JP-C = 31 Hz), 17.6, 22.0, 27.9, 28.5, 32.4, 37.1

(d, JP-C = 13 Hz), 59.7 (d, JP-C = 6.6 Hz), 66.1 (d, JP-C = 62 Hz), 126.6, 128.9 (d, JP-C = 13 Hz),

31 129.1, 129.2, 131.1 (d, JP-C = 3.4 Hz), 133.5 (d, JP-C = 14 Hz), 134.5 (d, JP-C = 11 Hz), 143.1; P

NMR (C6D6, 146 MHz): 76.3 (d, JP-H = 239 Hz, minor isomer), 80.4 (d, JP-H = 248 Hz, major isomer). HRMS (EI) Calcd. For C21H29P: 312.20069, Found: 312.20066.

5-47: To thiophene (2.9 mL, 36 mmol) in THF (25 mL) at -78 C was added nBuLi

(2.5 M in hexanes, 12 mL, 30 mmol) dropwise over 1 hr. Solution was warmed

to room temp for 30 min, then cooled to -78 again for dropwise addition of phosphorus trichloride (0.87 mL, 10 mmol in 5 mL THF). Solution was warmed to room temp and stirred 4 hours then quenched with aqueous ammonium chloride solution, then mixture was extracted with ether 3x, then washed with brine and dried over sodium sulfate, concentrated to a brown oil which was purified by column chromatography hexanes on silica gel to a yellow oil

(1.245 g, 45%). 1H NMR (CDCl3): 7.58 (3H, m), 7.35 (3H, m), 7.09 (3H, m). 31P NMR (CDCl3):

-46.4 ppm.

184

1 5-27: 952 mg, 79%. 1.15 eq allyl bromide. Yellow solid. H NMR (CDCl3):

8.59 (3H, t), 8.20 (3H, m), 7.62 (3H, m), 5.82 (1H, m), 5.49 (2H, m), 4.68 (2H,

31 dd). P NMR (CDCl3): 4.1 ppm. HRMS: M+ C15H14S3P Calculated: 320.9995, found: 320.9995.

5-48: To furan (2.2 mL, 30 mmol) in THF (25 mL) at -78 C was added nBuLi (2.5

M in hexanes, 10 mL, 25 mmol) dropwise over 1 hr. Solution was warmed to

room temp for 30 min, then cooled to -78 again for dropwise addition of phosphorus trichloride (0.72 mL, 8.3 mmol). Solution was warmed to room temp and stirred 4 hours then quenched with aqueous ammonium chloride solution, then mixture was extracted with ether 3x, then washed with brine and dried over sodium sulfate, concentrated to a brown oil which was purified by column chromatography hexanes on silica gel to a white solid (1.16 g, 50%) 1H

NMR (CDCl3): 7.65 (3H, m), 6.80 (3H, m), 6.40 (3H, m) pp. 31P NMR (CDCl3): -78.0 ppm.

HRMS (M+1, C12H10O2P) Found: 233.0361 Calculated:233.0368.

5-28: 1H NMR (CDCl3): 8.02 (3H, m), 7.94 (3H, m), 6.76 (3H, m) pp, 5.71 (1H,

13 m), 5.52 (1H, m), 5.30 (1H, m), 4.60 (2H, m). C NMR (CDCl3): 153.4, 131.6,

130.1, 126.5, 121.6, 113.3, 28.4. 31P NMR (CDCl3): -15.0 ppm. HRMS (EI,

M+, C15H14O3P) calculated: 273.0681, found: 273.0681.

5-49:29 To a solution of 2-bromobenzylbromide (5.22 g, 20.9 mmol) in dry

methanol (10 mL) under nitrogen atmosphere at room temp was added a sodium methoxide solution (25% in methanol, 6 mL, 26 mmol) over five minutes. Solution was refluxed

4 h, concentrated en vacuo and dissolved in ethyl acetate, washed water (3x) and brine, dried

185 sodium sulfate and concentrated to a clear oil (3.57 g, 85%). 1H NMR (CDCl3): 7.45 (1H, d), 7.32

(1H, d), 7.28 (1H, t), 7.13 (1H, t), 4.52 (2H, s), 3.46 (3H, s).

5-50: To a solution of 1-bromo-2-methoxymethylbenzene (3.57 g, 17. 8 mmol) in

THF (40 mL) at -78 was added BuLi (2.5 M in hexanes, 7.1 mL, 17.8 mmol) dropwise over 30 minutes, stirred cold an additional 30 minutes. Chlorodiphenylphosphine (3.3 mL, 17.8 mmol) was added at -78 over 30 minutes, stirred 30 minutes cold and warmed to room temp over an hour. Concentrated then partitioned in water and ethyl acetate, organics washed with water and brine and dried over sodium sulfate, concentrated to a clear oil, purified by CC DCM, to give a white solid (3.46 g, 64%). 1H NMR (CDCl3): 7.50-6.85 (14 H, m), 4.63 (2H, s), 3.26 (3H, s). 31P NMR (CDCl3): -15.8.

5-25: 6 equiv allyl bromide. White oil triturated with pentane to give white solid.

2.45 g, 57%. 1H NMR (CDCl3): 7.88-7.58 (14 H, m), 5.60 (2H, m), 5.35 (1H,

13 m), 4.75 (2H, quart), 4.34 (2H, s), 2.80 (3H, s). C NMR (CDCl3): 143.5, 136.9,

135.2, 134.4, 132.9, 131.7, 130.2, 129.4, 125.5, 124.3, 120.2, 116.2, 73.2, 57.6, 30.2 ppm. 31P

NMR (CDCl3): 22.9 ppm. HRMS: M+ C23H24OP Caclulated: 342.1565, Found: 347.1562.

5.5.2.3 Synthesis of Substrates

Representative Procedure for the Synthesis of Allylic Esters (RP3)

ethyl (E)-3-(naphthalene-2-yl)acrylate (5-44): To an ice-cold

suspension of NaH (60% dispersion in mineral oil, 1.23 g, 31 mmol, 1.2

eq.) in THF (30 mL) was added neat triethyl phosphonoacetate (6.1 mL,

186 31 mmol, 1.2 eq.) dropwise, and the mixture stirred for an additional 0.5 h at ambient temperature after all visible solid dissolved. A solution of 2-naphthaldehyde (3.91 g, 25 mmol) in THF (15 mL) was then added and the solution stirred at ambient temperature. Upon complete consumption of the starting material, the reaction was quenched with saturated aqueous NH4Cl solution (20 mL), and the aqueous layer extracted with EtOAc. The combined organic layers were washed with saturated aqueous NaCl solution, dried over anhydrous Na2SO4, and concentrated. Column chromatography

(silica gel, hexanes to 5:1 hexanes:EtOAc) afforded the desired product as a white solid (4.26 g, 19

30 1 mmol, 75%). Spectral data agrees with previous literature characterization. H NMR (CDCl3, 400

MHz): δ = 1.36 (t, J = 7.1 Hz, 3H, CH3), 4.29 (d, J = 7.1 Hz, 2H, OCH2), 6.55 (d, J = 16 Hz, 1H,

C=CH), 7.48-7.54 (m, 2H, Ar), 7.64-7.69 (m, 1H, Ar), 7.80-7.88 (m, 4H, Ar), 7.93 (s, 1H, Ar).

Representative Procedure for the Reduction of Allylic Alcohols (RP4)

(E)-3-(naphthalen-2-yl)prop-2-en-1-ol (5-45): To a solution of 5-44

(4.07 g, 20 mmol) in THF (80 mL) at -78 °C was added DIBAL-H (1.0

M in hexanes, 42 mL, 2.3 eq.) dropwise, and the solution stirred at this temperature for 2 h. The reaction was then stirred for an additional 0.5 h at 0 °C before quenching with saturated aqueous

NH4Cl solution (30 mL) to yield a thick emulsion. Saturated aqueous Rochelle’s salt (100 mL) and

EtOAc (200 mL) were added and the mixture vigorously stirred until two layers were observed and separated. The aqueous layer was extracted with additional EtOAc, and the combined organic layers washed with saturated aqueous NaCl solution and dried over anhydrous MgSO4. Filtration over a plug of silica gel followed by removal of solvent afforded the desired product as a white solid (3.12 g, 17 mmol, 94%). No further purification attempted. Spectral data agrees with previous

31 1 literature characterization. H NMR (CDCl3, 400 MHz): δ = 2.07 (br, 1H, OH), 4.32 (d, J= 5.3

Hz, 2H, OCH2), 6.43 (dt, J = 16, 5.4 Hz, 1H, C=CH), 6.71 (d, J = 16 Hz, 1H, C=CH), 7.38-7.46

(m, 2H, Ar), 7.54 (d, J = 8.6 Hz, 1H, Ar), 7.66 (s, 1H, Ar), 7.70-7.80 (m, 3H, Ar).

187

(E)-3-(naphthalen-2-yl)prop-2-en-1-d2-1-ol (5-45-d2): Following

literature precedent,32 to a suspension of LiD (0.200 g, 22.2 mmol) in

THF (5 mL) under a nitrogen atmosphere at ambient temperature is added dropwise dimethylaluminum chloride (0.9 M in heptane, 25 mL, 22.2 mmol). The solution was heated to 60 °C with reflux apparatus for 4 h, then cooled to room temperature. The resultant soluble solution was transferred via cannula slowly to a solution of S3 (2.28 g, 10 mmol) in toluene

(50 mL) at -78 °C. The solution was allowed to warm to ambient temperature and stirred for 8 h, then cooled to 0 °C and quenched slowly with aqueous saturated ammonium chloride. The resultant slurry is filtered over celite and rinsed with ethyl acetate (50 mL). The aqueous layer is extracted with ethyl acetate (20 mL), then the combined organic layers were dried over sodium sulfate, concentrated, and purified by column chromatography (silica gel, 10% to 25% EtOAc:hexanes) to

1 give the desired product as a white solid (0.970 g, 52%). H NMR (CDCl3, 400 MHz): 1.77 (s, 1H,

OH), 6.45 (d, J = 16 Hz, 1H, C=CH), 6.74 (d, J = 16 Hz, 1H, C=CH), 7.40-7.48 (m, 2H, Ar), 7.56

13 (d, J = 8.6 Hz, 1H, Ar), 7.69 (s, 1H, Ar), 7.73-7.81 (m, 3H, Ar); C NMR (CDCl3, 101 MHz): 63.1

(pentet, JC-D = 22 Hz), 123.7, 126.0, 126.4, 126.6, 127.8, 128.1, 128.3, 128.9, 131.4, 133.1, 133.6,

134.2; HRMS (EI): Calcd. for C13H10D2O: 186.10137, Found: 186.10333.

Representative Procedure for the Synthesis of Allylic Bromides (RP5)

(E)-2-(3-bromoprop-1-en-1-yl)naphthalene (5-46): To an ice-cold

solution of 5-45 (1.71 g, 9 mmol) in diethyl ether (10 mL) was added neat PBr3 (0.96 mL, 1.1 eq.) dropwise, and the mixture stirred at this temperature for 1 h. The reaction was then poured into ice cold saturated aqueous NaHCO3 (20 mL), and the aqueous layer extracted with diethyl ether. The combined organic layers were washed with saturated aqueous

NaHCO3 and water, dried over anhydrous MgSO4, and filtered. Removal of solvent afforded the desired product as a white solid (1.73 g, 7.0 mmol, 75%). No further purification attempted. 1H

188

NMR (CDCl3, 400 MHz): δ = 4.21 (d, J = 7.8 Hz, 2H, BrCH2, 6.51 (dt, J = 16, 7.8 Hz, 1H, C=CH),

6.79 (d, J = 16 Hz, 1H, C=CH), 7.42-7.50 (m, 2H, Ar), 7.58 (dd, J = 8.6, 1.2 Hz, 1H, Ar), 7.73 (s,

13 1H, Ar), 7.76-7.83 (m, 3H, Ar); C NMR (CDCl3, 101 MHz): 33.7, 123.6, 125.7, 126.4, 126.6,

127.3, 127.8, 128.2, 128.5, 133.4, 133.5, 133.6, 134.8; HRMS (EI) Calcd. For C13H11Br: 246.0044,

Found: 246.0046.

(E)-2-(3-bromo-3-d2-prop-1-en-1-yl)naphthalene (5-46-d2):

Prepared analogously from 5-45-d2 (0.90 g, 4.8 mmol) using RP5.

1 Isolated as a white solid (0.87 g, 3.5 mmol, 72%). H NMR (CDCl3,

360 MHz): δ = 6.50 (d, J = 16 Hz, 1H, C=CH), 6.79 (d, J = 16 Hz, 1H, C=CH), 7.42-7.51 (m, 2H,

13 Ar), 7.57 (dd, J = 8.6, 1.8 Hz, 1H, Ar), 7.71-7.84 (m, 4H, Ar); C NMR (CDCl3, 125 MHz): 33.3

(pentet, JC-D = 23 Hz), 123.6, 125.5, 126.4, 126.6, 127.3, 127.8, 128.2, 128.5, 133.4 (2), 133.6,

134.9; HRMS (EI) Calcd. For C13H9D2Br: 248.01696, Found: 248.01595.

5.5.3 Optimization and Procedure for Catalytic Allylic Reduction

5.5.3.1 General Procedure for the Screening of Organophosphorus Catalyst

t A solution of LiHAl(O Bu)3 (2.5 mmol, 2.5 eq.) in THF (6 mL) was added dropwise over

15 h via syringe pump to a 90 °C mixture of distilled cinnamyl bromide (1.0 mmol), organophosphonium catalyst (0.1 mmol, 0.1 eq.), and dodecane (1.0 mmol, 1.0 eq.) in PhMe (4 mL). After complete addition, the reaction mixture was stirred for an additional 1 h before cooling to ambient temperature. A 1.0 mL aliquot was removed and filtered over a small pad of silica gel, which was rinsed with dichloromethane (6 mL). The combined filtrate was then diluted to 12 mL with additional dichloromethane and analyzed directly by GC (Instrument settings: 0.1 μL injection volume, 280 °C injection temperature, 3.0 mL/min column flow, 8.0 mL/min total flow. Oven

189 temperature settings: Hold at initial 50 °C for 3 minutes, then increase at 5 °C/min to 130 °C followed by 10 °C/min to 250 °C). Retention times: 8.6 min (allylbenzene (5-11)), 11.3 min (trans-

β-methylstyrene (5-12)), 16.6 min (dodecane).

Ratio Entry R P Yield 3 (5-11:5-12)

1 none 80 9 : 91 2 5-26 78 18 : 82 3 5-24 85 11 : 89 4 5-29 95 19 : 81 5 5-30 95 54 : 46 6 5-31 96 94 : 6

Chromatograms for data tabulated in Table 5-5. Table 5-5, Entry 1

190

Table 5-5, Entry 2

191 Table 5-5, Entry 3

Table 5-5, Entry 4

192

Table 5-5, Entry 5

193 Table 5-5, Entry 6

194 5.5.3.2 Reductant Screening and Data

Table 5-6. Effect of hydride source on regioselectivity.

Yield Entry Reductant 5-11:5-12a (%)a,b

1 LiAlH4 95 94:6

2 LiAlH(OtBu)3 96 94:6 3 DIBAL-H 42c 30:70 d 4 [NBu4][BH4] trace - e 5 PhSiH3 - - e 6 PhSi2H2 - - e 7 Et3SiH - - a Determined by GC analysis. b Combined Yield. c propylbenzene observed in 4% yield. d Complex mixture e Reduction of allylphosphetanium not observed.

Table 5-6 summarizes the results of reductant optimization (using the procedure described in Section 5.5.3.1) described in Section 5.2.2. LiAlH4 (entry 1) was found to be equally effective as LiAlH(OtBu)3 (entry 2) in promoting the transposition; however, the neutral aluminium hydride

DIBAL-H resulted in reduced 5-11:5-12 ratios and propylbenzene byproduct in 4% yield (entry 3).

- Although [BH4] was capable of reducing allylphosphetanium salts in stoichiometric experiments in MeOH, applying this reductant to our catalytic conditions resulted in a complex mixture with trace 5-11 and 5-12 formation (entry 4). We attribute this to potential catalyst inhibition from BH3 produced in situ. Several silanes were incapable of reducing allylphosphetanium salts in

195 stoichiometric experiments (entries 5-7), prompting us to choose LiAlH(OtBu)3 as our model reductant over LiAlH4 due to increased safety and milder hydricity.

Chromatograms for data tabulated in Table 5-6. Table 5-6, Entry 1

196 Table 5-6, Entry 3

5.5.3.3 Effect of Leaving Group and Data

Table 5-7. Effect of leaving group on regioselectivity.

Yield Entry Leaving Group 5-11:5-12a (%)a,b

1 Cl 98 15:85 2 Br 96 94:6 c 3 CO2CF3 trace -

197

a Determined by GC analysis. b Combined Yield. c Cinnamyl alcohol major product as determined by 1H NMR.

Table 5-7 summarizes the results of leaving group optimization (using the procedure described in Section 5.5.3.1) described in Section 5.2.1. Allylic chlorides were found to be poor substrates for our conditions, providing 5-12 as the major product (entry 1). We attribute this reversal in selectivity to sluggish displacement of chloride by 5-31, allowing direct reduction to outcompete catalytic transposition by 5-31. Alkylation of 5-13 in neat allyl chloride was significantly slower in comparison to allyl bromide, furthering this notion. Reduction to the corresponding allylic alcohol was observed when a trifluoroacetate leaving group was employed

(entry 3), indicating preferential reduction of the electron-deficient ester moiety. From these results, allylic bromides were chosen as the optimal substrate for our allylic transposition conditions.

Chromatograms for data tabulated in Table 5-7. Table S2, Entry 1

198

To probe the effect of various hydride sources on the allylic transposition, the procedure outlined in Section III. A was followed, and the results are summarized below (Table S1).

5.5.3.4 General Procedure for the Phosphetane-Catalyzed Reduction of Allylic Bromides

t A solution of LiHAl(O Bu)3 (2.5 mmol, 2.5 eq.) in THF (6 mL) was added dropwise over

15 h via syringe pump to a 90 °C mixture of allylic bromide (1.0 mmol) and 5-31 (0.1 mmol, 0.1 eq.) in PhMe (4 mL). After complete addition, the reaction mixture was stirred for an additional 1 h before cooling to ambient temperature and quenching with saturated aqueous NH4Cl solution (2 mL). The aqueous layer was extracted with diethyl ether, and the combined organic layers washed with saturated aqueous NaCl solution and dried over Na2SO4. Removal of solvent followed by column chromatography (silica gel) afforded the desired product as a mixture of isomers. Ratios of

γ to α reduction products determined by 1H NMR analysis of clearly distinguishable peaks

(generally allylic methylene of γ product, allylic methyl of α product) unless otherwise stated.

33 1 2-allylnaphthalene (5-39): Colorless oil. H NMR (CDCl3, 400 MHz): δ

= 3.53 (d, J = 6.6 Hz, 2H, CH2), 5.07-5.17 (m, 2H, C=CH2), 5.97-6.10 (m,

1H, C=CH), 7.31 (d, J = 8.4 Hz, 1H, Ar), 7.37-7.46 (m, 2H, Ar), 7.60 (s, 1H, Ar), 7.72-7.81 (m,

3H, Ar). Minor Isomer: 1.91 (d, J = 6.5 Hz, 3H) 6.34 (dq, J = 16, 6.5 Hz), 6.54 (d, J = 16 Hz);

HRMS (EI) Calcd. For C13H12: 168.0939, Found: 168.0932.

2-(1-d-allyl)naphthalene (5-39-dγ): To a suspension of napthyl

phosphetanium salt 5-38 (0.438 g, 1 mmol) in THF (10 mL) was added

LiAlD4 (0.046 g, 1.1 mmol) and the reaction was stirred at ambient temperature for 30 min, then quenched with concentrated aqueous ammonium chloride solution (5 mL). The reaction was diluted with ether (20 mL), and combined organics washed with brine (2 x

199 10 mL), dried over sodium sulfate and concentrated to give a clear liquid, which was purified by column chromatography (silica gel, hexanes) to give the desired product as a colorless oil (0.093

1 mg, 55%). No other isomers or products were observed in this reaction. H NMR (CDCl3, 400

MHz): δ = 3.51 (br s, 1H, ArCHD), 5.07-5.16 (m, 2H C=CH2), 6.03 (ddd, J = 17, 10, 7.0 Hz, 1H,

C=CH), 7.31 (d, J = 7.9 Hz, 1H, Ar), 7.37-7.46 (m, 2H Ar), 7.61 (s, 1H, Ar), 7.72-7.81 (m, 3H,

13 Ar); C NMR (CDCl3, 101 MHz): 40.5 (t, JC-D = 19 Hz), 116.5, 125.7, 126.4, 127.1, 127.8, 127.9,

128.1, 128.4, 132.6, 134.1, 137.7, 138.0; HRMS (EI) Calcd. For C13H11D: 169.1002, Found:

169.1001.

α 2-(3-d2-allyl)naphthalene (5-39-d2 ): To a suspension of napthyl

α phosphetanium salt 5-38-d2 (0.441 g, 1 mmol) in THF (10 mL) was

t added LiAlH(O Bu)3 (0.305 g, 1.2 mmol) and the reaction was stirred at ambient temperature for 3 h, then quenched with concentrated aqueous ammonium chloride solution (5 mL). The reaction was diluted with ether (20 mL), and combined organics washed with brine (2 x 10 mL), dried over sodium sulfate and concentrated to give a clear liquid, which was purified by column chromatography (silica gel, hexanes) to give the desired product as a colorless

1 oil (0.128 g, 75%). No other isomers or products were observed in this reaction. H NMR (CDCl3,

360 MHz): δ = 3.54 (d, J = 6.6 Hz, 2H), 6.03 (br s, 1H, C=CH), 7.32 (dd, J = 8.4, 1.8 Hz, 1H, Ar),

13 7.39-7.48 (m, 2H, Ar), 7.63 (s, 1H, Ar), 7.75-7.84 (m, 3H, Ar); C NMR (CDCl3, 101 MHz): 40.4,

115.7 (pentet, J = 24 Hz), 125.4, 126.1, 126.8, 127.5, 127.6, 127.8, 128.1, 132.3, 133.8, 137.2,

137.7; HRMS (EI) Calcd. For C13H10D2: 170.10645, Found: 170.10652.

200 5.5.4 VT-NMR Experiments

5.5.4.1 Observation of Hydridophosphorane Intermediate

To cinammyl phosphetanium 5-40 (single diastereomer, 0.047 g, 0.075 mmol) dissolved in THF (0.5 mL) in a J. Young NMR tube sealed under a nitrogen atmosphere and cooled in a dry ice/acetone bath was added a solution of LiAlH4 (0.014 g, 0.375 mmol) in THF (0.5 mL) via a septum. The NMR tube was resealed and removed from the cooling bath, shaken twice to mix the reagents and placed in an NMR probe thermostatted at -80 °C. The reaction course was monitored via 31P NMR, with spectra obtained sequentially at increasing 20 °C intervals.

5.5.4.2 Order in Phosphorus

To 5-31-BPh4 (6:1 trans:cis mix of diastereomers, 0.0414 g, 0.075 mmol) and triphenylphosphine (0.005 g, 0.02 mmol) dissolved in THF (0.5 mL) in a J. Young NMR tube sealed under a nitrogen atmosphere and cooled in a dry ice/acetone bath was added a solution of

LiAlH4 (0.014 g, 0.375 mmol) in THF (0.5 mL) via a septum. The NMR tube was resealed and removed from the cooling bath, shaken twice to mix the reagents and placed in an NMR probe

31 thermostatted at -55 °C. The reaction course was monitored by P NMR (D1 = 5 sec); only the major diastereomer was integrated. The decay of 5-31-BPh4 has a linear fit for ln[5-31-BPh4] over time, indicating the reaction is first order in 5-31-BPh4.

201

Zero Order Fit 0.06

0.05

0.04

0.03

31] (M) 31] - [5 0.02

0.01 y = -1.60E-05x + 4.54E-02 R² = 9.49E-01 0 0 500 1000 1500 2000 2500 3000 Time (s)

First Order Fit 0 0 500 1000 1500 2000 2500 3000

-1

-2

] 31

- -3

5 ln[

-4

-5 y = -7.33E-04x - 2.89E+00 R² = 9.91E-01

-6 Time (s)

202

Second Order Fit 160

140

120 y = 4.06E-02x + 1.83E+00 100 R² = 9.08E-01

] (1/M) ] 80

31 -

5 60 1/[ 40

0 0 500 1000 1500 2000 2500 3000 Time (s)

5.5.4.3 Order in Reductant

To 5-31-BPh4 (6:1 trans:cis mix of diastereomers, 0.0207 g, 0.038 mmol) and triphenylphosphine (0.005 g, 0.02 mmol) dissolved in THF (0.5 mL) in a J. Young NMR tube sealed under a nitrogen atmosphere and cooled in a dry ice/acetone bath was added a solution of

LiAlH4 (2.5-20 equiv) in THF (0.5 mL) via a septum. The NMR tube was resealed and removed from the cooling bath, shaken twice to mix the reagents and placed in an NMR probe thermostatted

31 at -55 °C. The reaction course was monitored by P NMR (D1=5 sec), only the major diastereomer was integrated. The rate does not depend on the concentration of the reductant.

Table 5-8. Effect of equivalent of LAH on the reaction of 5-31. Equivalents of k (s-1) LAH obs 2.5 8.85E-04 5 3.61E-04 10 9.34E-04 10 9.53E-04

203

10 1.40E-03 15 2.62E-04 20 1.22E-03

5.5.4.4 Eyring Kinetic Experiments

LiAlH4 (0.014 g, 0.375 mmol) in THF (0.5 mL) via a septum. The NMR tube was resealed and removed from the cooling bath, shaken twice to mix the reagents and placed in an NMR probe

31 thermostatted at the desired temperature. The reaction course was monitored by P NMR (D1 = 5 sec), only the major diastereomer was integrated.

Table 5-9. Data obtained from the Eyring analysis of the reduction of 5-31. Temp (K) k (s-1) 1/T (1/ K) ln(k/T) 228 4.09E-03 4.39E-03 -1.09E+01 223 1.05E-03 4.48E-03 -1.23E+01 218 7.33E-04 4.59E-03 -1.26E+01 213 1.75E-04 4.69E-03 -1.40E+01 208 1.19E-04 4.81E-03 -1.44E+01 203 5.22E-05 4.93E-03 -1.52E+01

204 5.5.5 Computational Data

5.5.5.1 General Computational Information

Geometries were optimized in Gaussian 0934 using the M06-2X35 density functional with

6-311++G(2d,2p) basis set. Geometry optimizations were performed in the gas phase without symmetry constraint. Stationary points were characterized by frequency calculations to confirm their identity as either local minima (zero imaginary frequencies) or first-order saddle points (one imaginary frequency). For transition structure TS, an intrinsic reaction coordinate (IRC) calculation was conducted to ensure connection along the potential energy surface to 21 and products.

205 5.5.5.2 Relative Energies of Polytopal Isomers of 5-43

206 5.5.5.3 Transition structures for concerted group transfer

207 5.5.6 Crystallographic Data

Figure 5-21. Thermal ellipsoid plot (50%) of 1-allyl-2,2,3-trimethyl-1-phenylphosphetanium bromide. Hydrogen atoms and counter-ion are omitted for clarity. Adapted with permission from J. Am. Chem. Soc., 2015, 137, 5292. Copyright 2015 American Chemical Society.

A colorless block shaped crystal with approximate dimensions 0.19 x 0.25 x 0.29 mm, was used for the X-ray crystallographic analysis. The X-ray intensity data were measured at 298(2) K, on a Bruker SMART APEX CCD area detector system equipped with a graphite monochromator and a MoK fine-focus sealed tube ( = 0.71073Å) operated at 1600 watts power (50 kV, 32 mA).

The detector was placed at a distance of 5.8 cm from the crystal. A total of 1850 frames were collected with a scan width of 0.3º in  and an exposure time of 10 seconds/frame. The total data collection time was about 8 hours. The frames were integrated with the Bruker SAINT software package using a narrow-frame integration algorithm. The integration of the data using a

208

Orthorhombic unit cell yielded a total of 14351 reflections to a maximum  angle of 28.30  (0.90

Å resolution), of which 3679 were independent, completeness = 99.8%, Rint = 0.0264, Rsig =

0.0292 and 3070 were greater than 2(I). The final cell constants: a = 13.3991(15)Å, b =

13.9186(15)Å, c = 8.4200(9)Å,  = 90°,  = 90°,  = 90°, volume = 1570.3(3)Å3, are based upon the refinement of the XYZ-centroids of 5656 reflections above 20(I) with 2.826° < <26.884°.

Analysis of the data showed negligible decay during data collection. Data were corrected for absorption effects using the multiscan technique (SADABS). The ratio of minimum to maximum apparent transmission was 0.6060. The structure was solved and refined using the Bruker

SHELXTL (Version 6.1) Software Package, using the space group Pna2(1), with Z = 4 for the formula unit, C15 H22 Br P. Hydrogen atoms were placed geometrically, except for one on C8, and rode the parent atom. The hydrogen atom on C8 was located in difference Fourier map and refined isotropically. The final anisotropic full-matrix least-squares refinement on F2 with 161 variables converged at R1 = 3.35%, for the observed data and wR2 = 8.72% for all data. The goodness-of-fit was 0.998. The largest peak on the final difference map was 0.521 e-/Å3 and the largest hole was -0.217 e-/Å3. Based on the final model, the calculated density of the crystal is 1.325 g/cm3 and F(000) amounts to 648 electrons.

Table 5-10. Sample and crystal data for 1-allyl-2,2,3-trimethyl-1-phenylphosphetanium bromide. ______Crystallization solvents Methanol/diethyl ether Crystallization method slow diffusion

Empirical formula C15 H22 Br P Formula weight 313.21 Temperature 298(2) K Wavelength 0.71073 Å Crystal size 0.29 x 0.25 x 0.19 mm Crystal habit colorless block Crystal system Orthorhombic Space group Pna2(1)

209

Unit cell dimensions a = 13.3991(15) Å = 90° b = 13.9186(15) Å = 90° c = 8.4200(9) Å  = 90° Volume 1570.3(3) Å3 Z 4 Density (calculated) 1.325 g/cm3 Absorption coefficient 2.699 mm-1 F(000) 6 48 ______

Table 5-11. Atomic coordinates and equivalent isotropic atomic displacement parameters (Å2) for 1-allyl-2,2,3-trimethyl-1-phenylphosphetanium bromide.

U(eq) is defined as one third of the trace of the orthogonalized Uij tensor. ______x y z U(eq) ______Br1 0.079487(19) 0.315045(17) 0.51175(7) 0.06052(11) C1 0.1721(2) 0.0231(2) 0.7307(4) 0.0500(6) C2 0.0808(2) 0.0525(2) 0.6715(5) 0.0639(9) C3 -0.0014(2) -0.0069(3) 0.6852(5) 0.0710(9) C4 0.0077(3) -0.0950(2) 0.7587(4) 0.0672(9) C5 0.0968(2) -0.1223(2) 0.8198(5) 0.0670(9) C6 0.1808(2) -0.0648(2) 0.8055(4) 0.0585(7) C7 0.3904(2) 0.0540(2) 0.7919(4) 0.0601(8) C8 0.3673(3) 0.1157(3) 0.9393(4) 0.0692(9) C9 0.2952(3) 0.1906(2) 0.8598(4) 0.0596(8) C10 0.3467(3) 0.2788(3) 0.7916(5) 0.0838(12) C11 0.2035(3) 0.2194(3) 0.9578(6) 0.0932(14) C12 0.4577(3) 0.1551(4) 1.0295(8) 0.1008(13) C13 0.2855(2) 0.1361(2) 0.4998(4) 0.0614(7) C14 0.2798(4) 0.0468(4) 0.3921(5) 0.0980(14) C15 0.3479(5) 0.0086(4) 0.3344(8) 0.131(2) P1 0.27813(5) 0.09993(5) 0.70392(9) 0.04843(17) ______

5.5.7 Spectral Data

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231 5.6 References

1 Axelrod, E.; Milne, G. M.; Van Tamelen, E. E. J. Am. Chem. Soc. 1970, 92, 2139.

2 Kondo, K.; Negishi, A.; Tunemoto, D. Angew. Chem. Int. Ed. Engl. 1974, 13, 407.

3 Utsugi, M.; Miyano, M.; Nakada, M. Org. Lett. 2006, 8, 2973;

4 Hirabe, T.; Nojima, M.; Kusabayashi, S. J. Org. Chem. 1984, 49, 4084.

5 Donoghue, N.; Gallagher, M. J. Chem. Commun. 1998, 1973.

6 Brophy, J. J.; Gallagher, M. J. Aust. J. Chem. 1969, 22, 1399. B) Hellwinkel, D. Angew. Chem.,

Int. Ed. 1966, 5, 968

7 Pinnell, R.P.; Megerle, C.A.; Manatt, S.L.; Kroon, P.A. J. Am. Chem. Soc. 1973, 95, 977.

8 Allen, D.W.; Taylor, B.F. J. Chem. Soc., Dalton Trans. 1982, 51.

9 Saneyoshi, H.; Ochikubo,T., Mashimo,T., Hatano, K., Ito, Y., Abe, H. Org. Lett. 2014, 16, 30.

10 (a) Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70. (b) Hudson, R. F.; Brown, C. Acc. Chem.

Res. 1972, 5, 204.

11 (a) O’Brien, C. J.; Tellez, J. L.; Nixon, Z. S.; Kang, L. J.; Carter, A. L.; Kunkel, S. R.; Przeworski,

K. C.; Chass, G. A. Angew. Chem., Int. Ed. 2009, 48, 6836. (b) van Kalkeren, H. A.; Leenders, S.

H. A. M.; Hommersom, C. R. A.; Rutjes, F. P. J. T.; van Delft, F. L. Chem.-Eur. J. 2011, 17, 11290.

19, 5854.

12 Van Kalkeren, H.A.; van Delft, F.L.; Rutjes, F. P. J. T. Pure Appl. Chem. 2013, 85, 817.

13 (a) Marsi, K. L. J. Am. Chem. Soc. 1969, 91, 4724. (b) Marsi, K. L. J. Org. Chem. 1974, 39, 265.

14 Krenske, E. H. J. Org. Chem. 2012, 77, 3969.

15 Kühl, O. Phosphorus-31 NMR Spectroscopy: A Concise Introduction for the Synthetic Organic and Organometallic Chemist. Springer: Berlin, 2008.

232

16 (a) Ross, M.R.; Martin, J.C. J. Am. Chem. Soc. 1981, 103, 1234. (b) Kajiyama, K.; Kojima, S.;

Akiba, K.Y. Tetrahedron Lett. 1996, 37, 8409.

17 (a) Althoff, W.; Day, R. O.; Brown, R. K.; Holmes, R. R. Inorg. Chem. 1978, 17, 3265. (b)

Howard, J. A.; Russell, D. R.; Trippett, S. J. Chem. Soc. Chem. Commun. 1973, 856. (c) Holmes,

R. R. J. Am. Chem. Soc. 1978, 100, 433.

18 Maryanoff, B.E.; Reitz, A.B. Chem. Rev. 1989, 89, 863

19 Oppolzer, W.; Sneickus, V. Angew. Chem. Int. Ed. 1978, 17, 476

20 (a) Castro, C. E.; Kray, W. C. J. Am. Chem. Soc. 1963, 85, 2768. (b) Omoto, M.; Kato, N.; Sogon,

T.; Mori, A. Tetrahedron Lett. 2001, 42, 939. (c) Abad, A.; Agulló, C.; Cuñat, A. C.; Navarro, I.

Tetrahedron Lett. 2001, 42, 8965. (d) Abad, A.; Agulló, C.; Cuñat, A. C.; de Alfonso Marzal, I.;

Gris, A.; Navarro, I.; de Arellano, C. R. Tetrahedron 2007, 63, 1664.

21 (a) Myers, A. G.; Zheng, B. Tetrahedron Lett. 1996, 37, 4841. (b) Movassaghi, M.; Piizzi, G.;

Siegel, D. S.; Piersanti, G. Angew. Chem. Int. Ed. 2006, 45, 5859. (c) Movassaghi, M.; Ahmad, O.

K. J. Org. Chem. 2007, 72, 1838. (d) Siegel, D. S.; Piizzi, G.; Piersanti, G.; Movassaghi, M. J. Org.

Chem. 2009, 74, 9292.

22 (a) Tsuji, J.; Yamakawa, T. Tetrahedron Lett. 1979, 20, 613. (b) Tsuji, J.; Shimizu, I.; Minami,

I. Chem. Lett. 1984, 6, 1017. (c) Oshima, M.; Shimizu, I.; Yamamoto, A.; Ozawa, F.

Organometallics 1991, 10, 1221. (d) Lautens, M.; Paquin, J.-F. Org. Lett. 2003, 5, 3391. (e) Chau,

A.; Paquin, J.-F.; Lautens, M. J. Org. Chem. 2006, 71, 1924. (f) Movassaghi, M.; Ahmad, O. K.

Angew. Chem., Int. Ed. 2008, 47, 8909.

23 Mika, L. T.; Orha, L.; Farkas, N.; Horvath, I. T. Organometallics 2009, 28, 1593.

24 Sun, X.-M.; Manabe, K.; Lam, W. W.-L.; Shiraishi, N.; Kobayashi, J.; Shiro, M.; Utsumi, H.;

Kobayashi, S. Chem. Eur. J. 2005, 11, 361.

25 Byrne, P. A.; Gilheany, D. G. J. Am. Chem. Soc. 2012, 134, 9225.

26 Wetzel, R. B.; Kenyon, G. L. J. Am. Chem. Soc. 1974, 96, 5189.

233

27 Cremer, S. E.; Chorvat, R. J. J. Org. Chem. 1967, 32, 4066.

28 Owen, J.S.; Chan, E. M; Liu, H.; Alivisatos, A. P. J. Am. Chem. Soc. 2010, 132, 18206.

29 Dutta, D.K.; Deb, B.; Hua, G.; Woollins, J.D. J. Mol. Cat. A: Chemical 2012, 353, 7. 30 Zeitler, K. Org. Lett. 2006, 8, 637.

31 Charette, A. B.; Molinaro, C.; Brochu, C. J. Am. Chem. Soc. 2001, 123, 12168.

32 Whitesides, G.M.; San Filippo, J. J. Am. Chem. Soc. 1970, 92, 6611.

33 Seomoon, D.; Lee, P. H. J. Org. Chem. 2008, 73, 1165.

34 Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;

Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.;

Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.;

Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;

Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers,

E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.;

Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, M. J.; Klene, M.; Knox, J.

E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;

Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski,

V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.;

Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision B.01; Gaussian, Inc.:

Wallingford, CT, 2009.

35 (a) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215. (b) Zhao, Y.; Truhlar, D. G. Acc.

Chem. Res. 2008, 41, 157.

Appendix

Permissions

235

VITA Nicole Dunn

Nicole Dunn was born in 1988 and grew up in Brewster, NY. Upon graduating from

Brewster High School in 2006, Nicole matriculated at Colgate University. During her undergraduate education, Nicole took advantage of several research opportunities which inspired her to pursue further education. Nicole received a B.A. in Biochemistry from Colgate, and attended

Pennsylvania State University for graduate school in Chemistry. At Penn State, Nicole joined the fledgling lab of Dr. Alexander Radosevich and studied phosphorus-catalyzed reductions.