The Development of a Metal-Free Catalytic Method for the Selective Hydroxylation of Aliphatic C–H Bonds

THE DEVELOPMENT OF A METAL-FREE CATALYTIC METHOD FOR THE SELECTIVE HYDROXYLATION OF ALIPHATIC C–H BONDS

A DISSERTATION SUBMITTED TO THE DEPARTMENT OF CHEMISTRY AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Nichole Danielle Litvinas April 2010

This work is licensed under a Creative Commons Attribution- Noncommercial 3.0 United States License. http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/qr271bk1476

ii I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Justin Du Bois, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Robert Waymouth

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Paul Wender

Approved for the Stanford University Committee on Graduate Studies. Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file in University Archives.

iii

Dissertation Abstract

The diversity and structural intricacies of molecules needed for pharmaceutical, biological, and materials applications have challenged chemists to develop transformative chemical processes that greatly facilitate small molecule synthesis. C–H Bond functionalization represents one such class of reaction types, and is a general problem for reaction discovery that has witnessed an explosion of interest within the past 10 years. Inspired largely by Nature’s ability to conduct site- and stereoselective C–H bond oxidation reactions, we have been driven to design small molecule catalysts that can emulate such processes. Our focus has taken aim at the problem of C–H bond hydroxylation, efforts that have yielded a novel, non-metal-based catalytic system for the selective oxidation of 3° C–H bonds. These findings contrast the large body of literature detailing C–H hydroxylation reactions through transition-metal mediated catalysis. 1,2,3-Benzoxathiazine-2,2-dioxide-based heterocycles have been shown to function as catalysts for C–H hydroxylation with H2O2 operating as the terminal oxidant. The evolution of this catalytic process, which capitalizes on the unique reactivity of an oxaziridine intermediate, was made possible using density functional theory to help guide reagent design. In addition, kinetic analysis of the stoichiometric oxygen-atom transfer reaction has provided insight into the principal features that influence oxaziridine reactivity. This information coupled with the discovery that reactions could be conducted under aqueous reaction conditions with H2O2 has resulted in a markedly improved process for 3° C–H hydroxylation. The reaction occurs stereospecifically and with predictable chemoselectivity in substrates possessing more than one 3° C–H center. The enhanced performance of this catalytic process has been ascribed to the hydrophobic aggregation of the benzoxathiazinane catalyst and hydrocarbon substrate, which serves to accelerate the kinetically slow hydroxylation event.

iv Acknowledgements The production of this document is due in no small part to my family, friends, and coworkers at Stanford. I need to express my greatest gratitude to my advisor, Justin Du Bois, for patiently teaching me how to be a scientist. He taught me many things but most importantly to trust myself—a characteristic I never knew I was missing but that I had to develop to successfully complete this dissertation. My colleagues in the Du Bois group have been truly invaluable. Rose Conrad was a wonderful friend and mentor to me; she patiently guided through a rough first few years with lots of encouraging words. I am greatly indebted to Scott Wolckenhauer who was an excellent teacher and even better friend. I relied upon my classmates, Brian Andresen, John Mulcahy, and David Zalatan, for support and comfort as we tackled the many challenges of graduate school—I truly couldn’t have done it without them. The months spent writing this document would have been unbearable without Brian’s companionship. My most recent set of podmates, Dave Olson, Arun Thottumkara, and Mark Harvey, have been wonderful to work with. Research proposal writing was actually fun when they were around to brainstorm with me. I am lucky to have had many excellent mentors in the lab, particularly Alan Whitehead and Ben Brodsky. The Du Bois lab has always been a wonderful place to work. My friends at Stanford have been the best I have ever had and I am grateful to them for their support and companionship. Alicia Gutierrez, Ashley Jaworski, Brian Trantow, Mark Harvey, and Jay Fitzgerald can make me laugh like no one else. Finally, I am extremely grateful to my family, who has cautiously watched me go through tough times. My parents have done everything imaginable to support me and I can’t thank them enough. They have taught me what it means to be selfless and I hope I can pass that trait along someday.

Table of Contents

Chapter 1. Catalytic Hydroxylation of C–H Bonds...... 1

1.1 Introduction ...... 1 1.2 Challenges and Goals in Developing C–H Hydroxylation Catalysts...... 3 1.3 Hydroxylation Reactions in Nature ...... 4 1.4 Cytochrome P450 and Metalloenzyme Mimics ...... 5 1.4.1 Cytochrome P450s ...... 5 1.4.2 A Survey of “Outer-Sphere” Transition-Metal Catalyzed C–H Hydroxylation Methods ...... 9 1.4.3 Summary of outer-sphere transition-metal catalyzed C–H hydroxylation methods ...... 15 1.5 Flavin-Containing Monooxygnease and Organocatalytic C–H Hydroxylation16 1.5.1 Flavin-Containing Monooxygenase ...... 16 1.5.2 A Survey of Potentially-Catalytic Organic Oxidation Methods ...... 18 1.6 Conclusion ...... 33

Chapter 2. Benzoxathiazines for Selective O-Atom Transfer ...... 34

2.1 Introduction ...... 34 2.1a Background...... 34 2.1b Project Goals and Methods ...... 37 2.2 Oxaziridine Synthesis...... 39 2.2a Heterocycle synthesis...... 40 2.2b Synthesis of expanded-ring benzoxathiazine-derived oxaziridines ...... 42 2.2c Incorporation of aromatic substituents ...... 43 2.3 Overview of Computational Studies and Kinetics Analysis of Oxaziridines for Stoichiometric Epoxidation ...... 46 2.3a DFT Calculations ...... 46 2.3b Kinetics method ...... 48 2.4 Evaluation of benzoxathiazine-derived oxaziridines as stoichiometric oxidants ...... 50 2.4a Heteroatom and the Heterocycle ...... 50 2.4b Heterocycle ring size ...... 52 2.4c Halogenation of aromatic ring ...... 54

vi 2.4d Hammett analysis ...... 56 2.4e C5 modification ...... 57 2.4 Conclusions ...... 58

Chapter 3. A Catalytic Method for the Hydroxylation of C–H Bonds ...... 100

3.1 Introduction ...... 100 3.2 Protocol A: Diaryldiselenide and Urea Hydrogen Peroxide...... 101 3.2a Diaryldiselenides: background and synthesis...... 101 3.2b Improved catalyst design ...... 104 3.2c Results with diaryldiselenide cocatalyst...... 107 3.3 Protocol B: Acetic Acid and Aqueous Hydrogen Peroxide ...... 109 3.3a Hydrophobic Effects in Catalysis ...... 110 3.3b Results...... 112 3.3c Mechanistic Insights...... 118 3.4 Conclusions ...... 119

Appendix A. X-ray Crystallographic Data for Chapter 2...... 152

Appendix B. X-ray Crystallographic Data for Chapter 3...... 164

vii

List of Figures

Figure 1.1. C–H hydroxylation of (+)-artemisinin...... 2 Figure 1.2. C–H hydroxylation of a late-stage bryostatin precursor...... 3 Figure 1.3. A simplified view of the active site of cytochrome P450...... 6 Figure 1.4. The proposed catalytic cycle for cytochrome P450-mediated hydroxylation of an aliphatic substrate...... 7 Figure 1.5. The first oxygenation step in the biosynthesis of taxol...... 8 Figure 1.6. An example of the impressive site-selectivity exhibited by a CYP450 in the biosynthesis of taxol...... 8 Figure 1.7. A synthetic CYP450 mimic developed by Groves and coworkers...... 9 Figure 1.8. A synthetic CYP450 mimic developed by Mansuy and coworkers...... 10

Figure 1.9. C–H hydroxylation with 1 mol% [Fe(TPA)(CH3CN)2](ClO4)2...... 11 Figure 1.10. Selective tertiary C–H bond hydroxylation with Fe(S,S-PDP)...... 12 Figure 1.11. Improved efficiency for selective tertiary C–H bond hydroxylation with Fe((S,S,R)-mcpp)...... 13 Figure 1.12. Catalytic C–H hydroxylation with MTO...... 14 Figure 1.13. C–H hydroxylation of a steroid substrate with catalytic chromium...... 14 Figure 1.14. Site-selective ruthenium-catalyzed C–H hydroxylation...... 15 Figure 1.15. Proposed mechanism for C–H hydroxylation catalyzed by FMO...... 17 Figure 1.16. An example of a C–H hydroxylation reaction catalyzed by FMO...... 17 Figure 1.17. Hydroxylation of RemO with an FMO...... 18 Figure 1.18. An example of a synthetic FMO...... 19 Figure 1.19. Catalytic substrate oxidation with a synthetic FMO using molecular oxygen as the terminal oxidant...... 19 Figure 1.20. Catalytic olefin epoxidation with hexafluoroacetone and hydrogen peroxide...... 20 Figure 1.21. Dioxiranes can mediate a range of oxygenation reactions...... 21 Figure 1.22. Catalytic cycle for dioxirane-mediated olefin epoxidation...... 22 Figure 1.23. The formation of a stable hydrated ketone inhibits catalytic turnover...... 23 Figure 1.24. Dioxiranes react with C–H bond in a stereospecific manner...... 23 Figure 1.25. Radical clock experiments suggest that dioxirane-mediated C–H hydroxylation occurs in a concerted fashion...... 24 Figure 1.26 Four different transition structures for the epoxidation of ethylene with dioxdirane...... 25

viii Figure 1.27. The transition structures for olefin epoxidation and C–H hydroxylation with dioxiranes are similar...... 26 Figure 1.28. Hydroxylation of a single C–H bond with TFDO...... 26 Figure 1.29. Selective C–H hydroxylation at C25 is attributed to optimal stereoalignment with the dioxirane. Figure taken from reference (39)...... 27 Figure 1.30. The short half-life of dioxiranes in the presence of Oxone precludes the development of a catalytic C–H hydroxylation reaction mediated by dioxiranes...... 28 Figure 1.31. Stoichiometric intramolecular oxygen-atom transfer with an in situ generated dioxirane...... 28 Figure 1.32. Comparison of physical characteristics of dioxiranes and oxaziridines...... 29 Figure 1.33. Olefin epoxidation with an oxaziridine is characterized by a spiro, asynchronous transition structure...... 30 Figure 1.34. Oxaziridines can effect oxygenation reactions with nucleophilic substrates. 31 Figure 1.35. Rare examples of oxaziridine-mediated catalytic sulfide oxidation...... 32 Figure 1.36. A polyfluorinated oxaziridine effects the hydroxylation of a steroidal C–H bond...... 33 Figure 1.37. The only known organocatalytic system for C–H hydroxylation...... 33 Figure 2.1. Resnati and Davis established precedent for developing a catalytic oxaziridine-mediated C–H hydroxylation reaction...... 35 Figure 2.2. A proposed catalytic cycle for oxaziridine-mediated C–H hydroxylation...... 37 Figure 2.3. Comparison of the Davis oxaziridine with the Du Bois oxaziridine...... 38 Figure 2.4. Structural elements to investigate...... 38 Figure 2.5. A strategy for evaluating potential catalysts: synthesis, kinetic analysis, and DFT calculations...... 39 Figure 2.6. A two-step process for the conversion of salicylaldehydes to benzoxathiazine- derived oxaziridines...... 40 Figure 2.7. An N-phosphinoyl oxaziridine is capable of effecting olefin epoxidation...... 41 Figure 2.8. Synthesis of phosphorous oxaziridines 2.6 and 2.7...... 42 Figure 2.9. Crystal structure of N-phosphinoyl oxaziridine 2.7...... 42 Figure 2.10. Synthesis of seven-membered ring N-sulfonyl oxaziridine...... 43 Figure 2.11. Attempted synthesis of a benzoxathiazecine heterocycle...... 43 Figure 2.12. The Riemer-Tiemann formylation reaction was employed to access salicylaldehydes from substituted phenols...... 44 Figure 2.13. Crystal structure of 2.30...... 45 Figure 2.14. Cross-coupling route to substituted benzoxathiazine-derived oxaziridines. 46

Figure 2.15. Predicted (left) and experimentally obtained (right) IR spectra of 2.2...... 48 Figure 2.16. Raw kinetics data obtained for oxaziridine 2.18 following a pseudo-first order kinetics scheme...... 49 Figure 2.17. Eyring analysis...... 50 Figure 2.18. Oxaziridine reagents related to 2.2...... 50 Figure 2.19. DFT calculations suggest that the phenolic oxygen plays a large role delocalizing negative charge in the transition structure...... 51 Figure 2.20. Hammett plot...... 57 Figure 3.1. Benzoxathiazines react with a terminal oxidant to form a N-sulfonyl oxaziridine capable of transferring oxygen to a C–H bond...... 100 Figure 3.2. A catalytic cycle for C–H hydroxylation mediated by a benoxathiazine-derived oxaziridine...... 100 Figure 3.3. Diaryldiselenides react with hydrogen peroxide to form a perseleninic acid...... 101 Figure 3.4. Electron-deficient diaryldiselenides are excellent precatalysts for simple olefin epoxidation reactions...... 102 Figure 3.5. C–H hydroxylation was briefly examined with a range of diaryldiselenides. 103 Figure 3.6. Diaryldiselenides are useful catalytic reagents for oxidizing a benzoxathiazine to a oxaziridine...... 103 Figure 3.7. A synthetic route to trifluoromethylated benzoxathiazines and related oxaziridines...... 106 Figure 3.8. A novel cross-coupling method developed by Fagnou and coworkers was used to install a pentafluorophenyl group at C6...... 107 Figure 3.9. Electron-deficient catalysts give diminished conversions in the catalytic reaction...... 107 Figure 3.10. A stable hydrate formed at C4 could inhibit oxaziridine formation under catalytic conditions...... 108 Figure 3.11. 1H NMR spectra of benzoxathiazine 3.16 (C6=pentafluorophenyl). Asterisks are used to demarcate the signals corresponding to the hydrated benzoxathiazine. After >24 h at <1 torr, the catalyst is entirely dehydrated (spectrum on the right)...... 108 Figure 3.12. Electron-deficient ketones form stable hydrates and the necessary exchange with monoperoxysulfate is subsequently slow...... 109 Figure 3.13. New catalytic conditions feature peracetic acid that is formed from hydrogen peroxide and acetic acid in situ...... 110

x Figure 3.14. A marked improvement in product conversion is observed with aqueous reaction conditions...... 110 Figure 3.15. A large increase in the rate of a Diels-Alder reaction is attributed to the hydrophobic effect...... 111 Figure 3.16. Competition studies show that oxidants that can π-stack with a substrate give high selectivity for that substrate...... 112 Figure 3.17. Hammett analysis...... 114 Figure 3.18. Reaction of aromatic substrates with an oxaziridine...... 117 Figure 3.19. The enantiopurity of the product diminishes when resubjected to the reaction conditions...... 118

List of Tables Table 1.1. Effect of ketone structure on the efficiency of catalytic olefin epoxidation...... 22 Table 2.1. Substrate scope for catalytic oxaziridine-mediated C–H hydroxylation...... 36 Table 2.2. Benzoxathiazines and benzoxathiazine-derived oxaziridines synthesized from substituted salicylaldehydes...... 45 Table 2.3. Comparison of theoretical and experimental structural parameters for oxaziridine 2.30. (DFT calculations with B3LYP/6-31G*)...... 47 Table 2.4. Comparison of activation barriers for ethylene epoxidation...... 52 Table 2.5. Comparison of performance of Davis’s cyclic and acyclic oxaziridines...... 53 Table 2.6. Comparison of observed rate constants and calculated activation barriers for a range of heterocycle ring sizes...... 54 Table 2.7. Comparison of rate of oxygen transfer for a variety of chlorinated oxaziridines...... 55 Table 2.8. Comparison of halogenated heterocycles of different ring sizes...... 56 Table 2.9. Hammett analysis...... 56 Table 2.10. Effect of C5 modification of the rate of expoxidation...... 57 Table 3.1. Substrate scope for C–H hydroxylation with a first generation benzoxathiazine catalyst...... 104 Table 3.2. The effect of C4 substitution of the rate of olefin epoxdiation...... 105

Table 3.3 Catalytic C–H hydroxylation with aqueous H2O2...... 113 Table 3.4. Substrate profile for reactions catalyzed by 3.16...... 116

xii Towards the Selective Catalytic Hydroxylation of C–H Bonds

Chapter 1. Catalytic Hydroxylation of C–H Bonds

1.1 Introduction The diversity and structural intricacies of molecules needed for pharmaceutical, biological, and materials applications have challenged chemists to develop transformative chemical processes that greatly facilitate small molecule synthesis. As such, C–H functionalization has emerged as an efficient method for assembling complex molecules. By definition, C–H functionalization methods do not require the presence of any unnecessary functional groups to increase molecular complexity; thus molecular scaffolds are rapidly accessed and subsequently modified at a late stage in a synthesis. Furthermore, in principle this approach could eliminate the historical requirement of cumbersome protecting group strategies and is often one of the most atom- and step- economical approaches to chemical synthesis. As an arsenal of C–H functionalization methods emerges, the logic of chemical synthesis is being revolutionized such that bond disconnections never before considered are now in the realm of the possible. These new synthetic strategies are reminiscent of the approach Nature takes to assembling complex metabolites: instead of a series of functional group interconversions, enzymes will often assemble the aliphatic skeleton of a molecule and then embellish the structure through a series of C–H hydroxylation and olefin epoxidation reactions.1 Although there are many advantages to the pursuit of a biologically-inspired synthetic strategy such as this, lack of access to selective oxidative reactions have severely limited its usage. Many examples of selective, late-stage C–H hydroxylation are found in metabolic pathways and as a result, many researchers have turned to Nature for inspiration when developing selective, catalytic C–H hydroxylation methods. For instance, the active site of the cytochrome P450s has inspired the development of a large number of non- organometallic metalloenzyme-mimics. Although considerable progress has been made with these bio-inspired methods, a general and selective catalytic method for C–H hydroxylation remains elusive. Fortunately, clues to developing such a method can be found in biological processes and Nature continues to serve as the chemist’s muse. The oxygen insertion reactions carried out by oxygenase enzymes like cytochrome P450s (CYP450s) and flavin-containing monoxygenases (FMOs) are difficult

1 Chen, K.; Baran, P. S. “Total synthesis of eudesmane terpenes by site-selective C–H oxidation.” Nature 2009, 459, 824-828.

1 Chapter 1

to mimic in practical synthetic reactions. The high levels of selectivity exhibited by these enzymes surpass conventional rules of reactivity; indeed, enzyme binding pockets and complex hydrogen bonding networks are necessary to orient a particular substrate to achieve the desired site- and stereoselectivity. In the absence of this type of substrate- catalyst molecular recognition, however, the selectivity displayed by synthetic catalysts is based solely on the inherent reactivity of different C–H bonds within a given molecule. Recently, considerable progress has been made towards general methods for intermolecular C–H hydroxylation reactions. White et al have demonstrated that an iron catalyst can selectively hydroxylate a single C–H bond in a natural product as complex as artemisinin.2 In this example, (+)-artemisinin, an anti-malarial agent, is hydroxylated to give (+)-10β-hydroxyartemisinin (Figure 1.1). Remarkably, the delicate endoperoxide functionality is maintained under the reaction conditions. This transformation has also been achieved enzymatically; when (+)-artemisinin is submitted to microbial cultures of Cunninghamella echinulata, (+)-10β-hydroxyartemisinin is produced in 47% yield. While the synthetic system slightly underperforms the enzymatic system, this result illustrates one potential benefit of achieving synthetic cytochrome P450-esque reactivity in the flask: complex substrates can be manipulated at a late stage to produce value-added structures, which have new physical and biological properties.

Me Me H H OH 5 mol% Fe(S,S-PDP) O Me 50 mol% AcOH x3 Me O O O O 1.2 equiv H2O2 O H H O O Me CH3CN, 30 min, rt Me O O 34%

Figure 1.1. C–H hydroxylation of (+)-artemisinin.

Wender and coworkers have showcased the benefits of a highly selective hydroxylation reaction using dimethyldioxirane (DMDO) for the late-stage modification of a bryostatin analogue (Figure 1.2).3 DMDO is a cyclic organic peroxide reagent capable of impressive selectivity that is similar to the enzyme class FMO, a non-metal-based oxygenase. The example illustrated in Figure 1.2 demonstrates the great utility of a late

2 Chen, M. S.; White, M. C. “A Predictably Selective Aliphatic C–H Oxidation Reaction for Complex Molecule Synthesis.” Science 2007, 158, 783-787. 3 Wender, P. A., Hilinski, M. K. & Mayweg, A. V. W. “Late-Stage Intermolecular C–H Activation for Lead Diversification: A Highly Chemoselective Oxyfunctionalization of the C-9 Position of Potent Bryostatin Analogues.” Org. Lett. 2005, 7, 79-82.

2 Towards the Selective Catalytic Hydroxylation of C–H Bonds

stage C–H hydroxylation reaction. DMDO was employed to introduce a single hydroxyl group to a late-stage bryostatin intermediate. Subsequent manipulations of this alcohol provided access to a large number of bryostatin analogues. With a reagent like DMDO at their disposal, only one route to the complex intermediate was needed to access a large number of bryostatin analogues; this strategy stands in stark contrast to traditional synthetic routes that require functional groups to be engineered early in the synthesis.

O O O O O O O DMDO O Me OH HO O O acetone Me OH HO Me rt, 48h O O Me O OH O OH

C7H15 O CO2Me C7H15 O CO2Me 70%

Figure 1.2. C–H hydroxylation of a late-stage bryostatin precursor.

These examples by White and Wender, which demonstrate how the introduction of hydroxyl groups to complex molecules allows easy access to novel products, highlight the importance of developing biologically-inspired hydroxylation transformations. Studies of this sort, however, are still in their infancy and considerable work must be done to develop and refine the concepts governing selective C–H hydroxylations until such methods are generally applicable. This chapter is a survey of the various strategies employed by synthetic organic chemists to achieve catalytic C–H hydroxylation reactions. While this review is by no means comprehensive, it intends to document the recent evolution of both metallic- and organic-based hydroxylation catalysts. Some of the general challenges posed by the development of hydroxylation catalysts are described below while obstacles unique to certain systems are presented later. Overcoming these challenges will ultimately make catalytic C–H hydroxylation a permanent addition to the synthetic chemist’s toolbox.

1.2 Challenges and Goals in Developing C–H Hydroxylation Catalysts Any saturated hydrocarbon presents an extraordinary challenge in site-selectivity for C–H functionalization because no other functional group is as ubiquitous as the C–H bond. Although strategies involving directing groups are conceptually interesting and have found use in certain contexts, such processes are inherently limited in scope.

3 Chapter 1

One reason that a biologically-inspired approach to catalytic C–H hydroxylation is so appealing is the desire to mimic the stereospecificity exhibited by enzymatic oxygenases. For a synthetic transformation to find broad application, the resulting stereochemistry must be predictable. Stereospecific hydroxylation of a stereogenic tertiary C–H bond offers absolute stereocontrol and is an intrinsic feature of some catalytic hydroxylation processes. A significant problem for oxidation reactions occurs when catalysts are unstable to oxidative conditions. In the case of transition metal catalysts, excess oxidant can catalyze the oxidative dimerization (or multimerization) of the catalyst or degradation of the ligands. The fidelity of the ligand is particularly important because, depending on the terminal oxidant employed, metal centers that are not protected by a ligand framework will catalyze oxidant decomposition and the production of oxygen-centered radicals. Oxygen- centered radicals can wreak havoc on a selective oxygenation reaction because they will indiscriminately react with C–H bonds and other functional groups. Another problem for developing a C–H hydroxylation method is over-oxidation. Typically, once a C–H bond is hydroxylated; the product becomes more susceptible to oxidation than the starting material. A solution to this is to target tertiary C–H bonds where there is no possibility for over-oxidation. Employing a large excess of substrate to oxidant can help mitigate over-oxidation; however, this is not a useful solution when the substrate is synthetically valuable. Finally, a problem that plagues all of the methods discussed in the following chapter is a lack of substrate generality. While there are some remarkable examples of high selectivity with complex substrates (Figures 1.1 and 1.2), these cases are the exception, not the rule. The factors that govern the performance of these reactions, particularly the fate of the catalyst and oxidant when the reaction goes awry, are not well understood. A small sample of substrates are used when exploring virtually every C–H hydroxylation method and, given the ubiquity of the C–H bond, this is surprising, although perhaps telling of the limitations in the current state-of-the-art. Of note is the fact that most catalytic systems have not been shown to function with benzylic substrates or, for that matter, with substrates containing aromatic moieties.

1.3 Hydroxylation Reactions in Nature Awe-inspiring examples of selective oxygenation reactions can be found in Nature. Two major classes of oxygenase enzymes perform such reactions: metalloenzymes and flavin-dependent (organocatalytic) enzymes. The metalloenzymes

4 Towards the Selective Catalytic Hydroxylation of C–H Bonds

typically employ a copper- or iron-based active site; these include cytochrome P450s (Fe(II)), methane monooxygenase (Fe(II)), tyrosinase (Cu(I)), Rieske dioxygenases (Fe(II)), monocopper hydroxylases (Cu(I)), and pterin-dependent oxygenases (Fe(II)).4 Alternatively, there are the non-metallic enzymes, such as the flavin-containing monooxygenases; this class of enzymes features a polyaromatic organic scaffold that forms an organic peroxide. Both classes of oxygenases react with molecular oxygen to form reactive oxygenating species to effect oxygenation chemistry in metabolic processes. In terms of designing synthetic catalysts for C–H hydroxylation reactions, the cytochrome P450s and the flavin-containing monoxygenases enzymes are excellent archetypes. With these oxygenase enzymes, two alternative strategies to devising C–H hydroxylation methods are presented: transition metal-catalyzed and organocatalytic. To mimic the performance of these enzymes in the flask, there are two major characteristics of enzymatic behavior that must be understood: the enzyme active site and the enzyme binding site. While the concept of mimicking a binding site with molecular recognition motifs in a synthetic context is intriguing, this topic is beyond the scope of this review.5 This review will discuss how understanding the nature of enzyme-active sites has aided in the development of catalytic C–H hydroxylation methods.

1.4 Cytochrome P450 and Metalloenzyme Mimics

1.4.1 Cytochrome P450s Cytochrome P450s are a family of iron-containing heme enzymes that react with molecular oxygen to transfer a single oxygen-atom to substrates (Figure 1.3). Oxygen transfer by these enzymes occurs with remarkable site-selectivity and stereospecificity. The iron porphyrin cofactor is ligated by a cysteine at one of the axial positions of the metal, which attaches the heme cofactor to the protein backbone. While the structure of

4 Que, L.; Tolman, W. B. “Biologically inspired oxidation catalysis.” Nature 2008, 455, 333-340. 5 Several researchers have made considerable progress in this arena; see: (a) Breslow, R. “Biomimetic chemistry and artificial enzymes- catalysis by design.” Accounts Chem Res 1995, 28, 146-153. (b) Cacciapaglia, R.; Stefano, S. Mandolini, L. “Effective molarities in supramolecular catalysis of two- substrate reactions.” Accounts Chem Res 2004, 37, 113-122. (c) Breslow, R.; Dong, S. D. “Biomimetic reactions catalyzed by cyclodextrins and their derivatives.” Chem Rev 1998, 98, 1997-2011. (d) Das, S.; Incarvito, C. D.; Crabtree, R. H.; Brudvig, G. W. “Molecular recognition in the selective oxygenation of saturated C-H bonds by a dimanganese catalyst.” Science 2006, 312, 1941-1943.

5 Chapter 1

this enzyme active site is well defined, the intimate details of this mechanism of oxygen transfer by this complex remains an actively debated topic in the literature.6

O [O] N N N N Fe Fe N N N N

S S cys SO S cys O O HO O HO HO O HO

Figure 1.3. A simplified view of the active site of cytochrome P450.

It is known that upon coordination of dioxygen to the iron(II) heme complex and electron donation from a terminal reductant, an iron(III)-peroxo species is formed (Figure 1.4). Following two protonation steps, the O–O bond is cleaved to form an iron(V)oxo; however, due to the redox-active capabilities of the porphyrin ligand, this species is more accurately described as an iron(IV)oxo porphyrin radical cation.7 Kinetics studies show that the iron(IV)oxo radical cation oxidizes organic substrates via an oxo-transfer reaction; furthermore, hydrogen peroxide, alkyl hydroperoxides, peroxyacids, periodate, and iodosylbenzene are all functional with cytochrome P450s, indicating that an iron-oxo, rather than an iron-peroxo, is the active oxygenating species.8

6 Shaik, S.; de Visser, S. P.; Kumar, D. “One oxidant, many pathways: a theoretical perspective of monooxygenation mechanisms by cytochrome P450 enzymes.” J Biol Inorg Chem 2004, 9, 661-668. Que. The oxo/peroxo debate: a nonheme iron perspective. J Biol Inorg Chem 2004, 9, 684-690. 7 Que. The heme paradigm revisited: alternative reaction pathways considered - Introduction. J Biol Inorg Chem 2004, 9, 643-643. 8 Groves, J. T. In Cytochrome P450: Structure, Mechanism, and Biochemistry, 3rd ed.; Ortiz de Montellano, P. R., Ed.; Kluwer Academic/Plenum: New York, 2005; pp 1-44.

6 Towards the Selective Catalytic Hydroxylation of C–H Bonds

OH2 R–OH R–H Fe(III)

O S H Cys Fe(IV) Fe(III) R R–H S S Cys Cys e- H2O

2H+ O O Fe(III) Fe(II) R–H R–H S S Cys O Cys O O2 e- Fe(III) R–H S Cys

Figure 1.4. The proposed catalytic cycle for cytochrome P450-mediated hydroxylation of an aliphatic substrate.

The taxol biosynthesis represents one of the most inspiring examples of late stage introduction in complex molecule assembly.9 Taxol is an antimitotic agent that has been successfully employed as a cancer treatment.10 Due to the limited availability of this compound from its natural source, the yew tree, the synthesis of taxol represents a significant challenge for synthetic chemists. While many research groups have completed syntheses of taxol, no route has been sufficiently efficient to utilize on an industrial scale.11,12 To mimic the elegant and efficient biosynthesis of taxol in the lab, however, a chemist would need access to highly selective oxygenation catalysts like the CYP450s. A recent paper by Baran illustrates why many syntheses are not as efficient as

9 Taxol is the trade name for paclitaxel and is a registered trademark of Bristol-Meyers-Squibb. 10 Chau et al. Taxol biosynthesis: Molecular cloning and of a cytochrome p450 characterization taxoid 7 beta-hydroxylase. Chem Biol (2004) vol. 11 (5) pp. 663-672 11 a) R. A. Holton, C. Somoza, H.-B. Kim, F. Liang, R. J. Biediger, P. D. Boatman, M. Shindo, C. C. Smith, S. Kim. H. Nadizadeh, Y Suzuki, C. Tao, P. Vu, S. Tang, P. Zhang, K. K. Murthi, L. N. Gentile, J. H. Liu, J: Am. Chem. SO 1994, 116, 1597-1600; b) K. C. Nicolaou, Z Yang, J. J. Liu, H. Ueno, P. G Nantermet, R. K. Guy. C. F. Claiborne, J. Renaud, E. A. Couladouros, K Paulvannan, E. J. Sorensen, Nature 1994,367,630-634; c) S. J. Danishesfsky, J. J. Masters, W. B. Young, J. T. Link, L. B. Snyder, T V. Magee, D. K. Jung R. C. A. Isaacs, W. G. Bornmann, C. A. Alaimo, C. A. Coburn, M. J. Di Grandi, 1. Am. Chem. Soc. 1996, 118, 2843-2859; d) P. A. Wender. N. F Badham, S. P. Conway, P. E. Floreancig, T. E. Glass, J. B. Houze, N. E Krauss, D. Lae, D. G. Marquess, P. L. McGrane, W Meng. M. G. Natchus, A. J. Shuker, J. C. Sutton, R. E. Taylor, ibid. 1997, 119, 2757-2758. 12 Currently the synthesis of taxol is achieved through semi-synthesis starting from the advanced taxoid 10-deacetylbaccatin III,, which can be isolated in sufficient quantities from regenerable yew twigs.

7 Chapter 1

Nature’s; while chemists have sufficiently developed the methods necessary for the “cyclase” phase of a synthetic route, there is a dearth of methods available for the “oxidiase” phase (see Introduction).1 To illustrate this, Figure 1.5 depicts a step in the biosynthesis of taxol. The deoxygenated taxane skeleton is quickly assembled through a series of cyclization reactions to give 4,11-taxadiene (Figure 1.5). The first oxygenation step is catalyzed by an enzyme, taxadiene-5-hydroxylase, which can be characterized as a membrane-bound, NADPH-dependent, P450 oxygenase that employs molecular oxygen as its terminal oxidant.13 Stereoselective epoxidation of a single olefin is followed by rearrangement to give an enantiopure alcohol.

Me Me Me Me Me Me Me CYP450 Me Me Me Me Me H O OH H Me H Me Me H H Me CH2 H

Figure 1.5. The first oxygenation step in the biosynthesis of taxol.

Perhaps even more remarkable is the action of taxane 13α-hydroxylase, which installs an oxygen at a C–H bond despite the presence of more reactive olefin moieties (Figure 1.6).14 These examples highlight the impressive capabilities of the cytochrome P450s for achieving selective C–H hydroxylation reactions and one can envision how access to analogous methods would be highly beneficial to a synthetic chemist.

Me Me Me Me

Me CYP450 Me Me HO Me OH OH H H Me CH2 Me CH2

Figure 1.6. An example of the impressive site-selectivity exhibited by a CYP450 in the biosynthesis of taxol.

13 Jennewein, S.; Long, R. M.; Williams, R. M.; Croteau, R. “Cytochrome P450 taxadiene 5 alpha- hydroxylase, a mechanistically unusual monooxygenase catalyzing the first oxygenation step of taxol biosynthesis. Chem Biol 2004, 11, 379-387. 14 Jennewein et al. Taxol biosynthesis: Taxane 13 alpha-hydroxylase is a cytochrome P450-dependent monooxygenase. P Natl Acad Sci Usa 2001, 98, 13595-13600.

8 Towards the Selective Catalytic Hydroxylation of C–H Bonds

1.4.2 A Survey of “Outer-Sphere” Transition-Metal Catalyzed C–H Hydroxylation Methods15 Early work on developing synthetic oxygenation catalysts focused on mimicking the basic reactivity manifold of cytochrome P450 metalloenzymes. Heme, also known as iron protoporphyrin, is the prosthetic group for the cytochrome P450s. In an attempt to mimic the reactivity of the CYP450s, Groves and co-workers employed iron porphyrins with hypervalent iodine oxidants (Figure 1.7).16 Generally the synthetic utility of these systems was underwhelming: a large excess of substrate had to be employed to mitigate oxidative decomposition of catalyst and chemoselectivity, as shown in the example with cyclohexene, was only modest. Furthermore, attempts to employ these reaction conditions with saturated hydrocarbon substrates resulted in very low yields due to competing catalyst arrest and/or decomposition.

Cl 16 mol% catalyst 1 equiv PhIO Cl O N N CH2Cl2, rt OH Fe 33 equiv 55%a 15%a N N Cl aYield based on iodosylbenzene consumed Cl

catalyst

Figure 1.7. A synthetic CYP450 mimic developed by Groves and coworkers.

More recently, modifications to the porphyrin ligand have yielded impressive increases in product conversions. For instance, Dolphin et al introduced halogenation to the periphery of the porphyrin ligand; this modification had the effect of both preventing the oxidative degradation of the ligand, and thus extending the lifetime of the catalyst and increasing the reactivity by generating a more electrophilic oxidant.17 Mansuy and coworkers further exploited this concept by adding nitro groups to the β positions of the

15 “Outer-Sphere” in this case refers to a method where the substrate does not directly interact with the metal center, ie there is no oxidative addition of substrate. Instead, the oxidant and catalyst form a reactive species, such as a metal-oxo, that the substrate reacts with. For more information, see Dick and Sanford. Transition metal catalyzed oxidative functionalization of carbon-hydrogen bonds. Tetrahedron (2006) vol. 62 (11) pp. 2439-2463 16 Groves, J. T.; Nemo, T. E.; Myers, R. S. “Hydroxylation and epoxidation catalyzed by iron-porphine complexes – oxygen-transfer from iodosylbenzene.” J Am Chem Soc 1979, 101, 1032-1033. 17 Dolphin, D.; Traylor, T. G.; Xie, L. Y. “Polyhaloporphyrins: Unusual ligands for metals and metal- catalyzed oxidations.” Accounts Chem Res 1997, 30, 251-259.

9 Chapter 1

porphyrin ligand (Figure 1.8).18 Heptane was hydroxylated with 5 mol% catalyst in moderate yield to give a mixture of hydroxylation products. Impressively, cyclooctene was epoxidized in high yield using only 1 mol% catalyst loading and one equivalent of substrate.

5 mol% catalyst - NO2 Cl 1 equiv PhIO OH Cl Cl Me Me Me Me CH2Cl2, rt, 1h 40 equiv 66% yield Cl N Cl C1 : C2 : C3 : C4 1 : 49 : 33 : 17 O2N N Fe N NO2 O Cl N Cl 1 mol% catalyst 3 equiv H2O2

Cl Cl CH2Cl2/MeCN NO2 rt, 1h 1 equiv 97% yield catalyst

Figure 1.8. A synthetic CYP450 mimic developed by Mansuy and coworkers.

The results of Dolphin, Mansuy, and others provide a promising starting point for designing transition metal catalysts for C–H hydroxylation. There are several limitations to pursuing porphyrin-based catalysts, however, including the general requirement for a large excess of substrate relative to oxidant and the intrinsic difficulties associated with synthesizing substituted porphyrin-type ligands. Alternative catalysts for C–H hydroxylation employ non-heme systems in conjunction with simple terminal oxidants such as hydrogen peroxide. The combination of an iron salt and hydrogen peroxide as a reagent combination for oxidation dates back to the work of Fenton in the nineteenth century.19 These early findings show that iron catalyzes the homolytic cleavage of the peroxo O–O bond in hydrogen peroxide; thus, a major challenge has been to inhibit the production of non-selective oxygen-centered radicals when developing iron/hydrogen peroxide catalytic systems.20 The desire to understand and to model the enzymatic reactions of metalloenzymes have led the research groups of Barton, Nam, Que and others to produce synthetic, non-poryphrin iron

18 (a) Bartoli, J. F.; Battioni, P.; Defoor, W. R.; Mansuy, D. “Synthesis and remarkable properties of iron beta-polynitroporyphyrins as catalysts for monooxygenation reactions.” J Chem Soc Chem Comm 1994, 1, 23-24. (b) Mansuy, D. “Activation of alkanes- the biomimetic approach.” Coordination Chemistry Reviews 1993, 125, 129-141. 19 Fenton, H. J. H. Oxidation of tartaric acid in the presence of iron. J. Chem. Soc. 1894, 65, 889–910. 20 Que, L.; Tolman, W. B. “Biologically inspired oxidation catalysis.” Nature 2008, 455, 333-340.

10 Towards the Selective Catalytic Hydroxylation of C–H Bonds

catalysts.21,22,23 For instance, Que and Nam showed that the catalytic hydroxylation 24 activity could be achieved. Que showed that using [Fe(TPA)(CH3CN)2](ClO4)2 (TPA=trispyridylamine) in the presence of dilute hydrogen peroxide gave products of tertiary C–H hydroxylation (Figure 1.9).25 Subjecting cis-1,2-dimethylcyclohexane to the reaction conditions resulted in a single tertiary alcohol diastereomer; this experiment provided evidence for a stereospecific oxygen insertion event. Mechanistically, for a reaction of this type to occur sterospecifically, the reaction is either concerted of lifetime of any intermediate radical species must be considerably short-lived (vide infra).26

2+ Me 1 mol% [Fe(TPA)(CH3CN)2](ClO4)2 Me s 1 equiv H2O2 N OH N Fe N Me MeCN, rt Me s N 100 equiv 3.9 TON [Fe(TPA)(CH3CN)2](ClO4)2

Figure 1.9. C–H hydroxylation with 1 mol% [Fe(TPA)(CH3CN)2](ClO4)2.

While many of these studies on non-heme iron C–H hydroxylation catalysts have been pursued with an eye toward elucidating the mechanisms of metalloenzymes, several groups have also exploited these systems for chemical synthesis. Of note is the work of White and coworkers who have employed a non-heme iron catalyst for the purpose of selectively hydroxylating tertiary C–H bond substrates (Figure 1.1). Although the nature of the active catalyst is ambiguous, similar results to CYP450 enzymatic transformations are observed, suggesting that the reaction proceeds through an iron oxo species. As can be seen with the acylated menthol substrate in Figure 1.10, the reaction is selective for

21 Barton, D. H. R.; Csuhai, E.; Ozbalik, N.; Balavoine, G. “Mechanism of the selective functionalization of saturated-hydrocarbons by Gif systems- relationship with methane monooxygenase.” P Natl Acad Sci USA 1990, 87, 3401-3404. 22 Nam, W. “High-valent iron(IV)-oxo complexes of heme and non-heme ligands in oxygenation reactions.” Accounts Chem Res 2007, 40, 522-531. 23 Rohde, J. -U.; In, J. –H.; Lim, M. H.; Brennessel, W. W.; Bukowski, M. R.; Stubna, A.; Munck, E.; Nam, W.; Que, L. Jr. “Crystallographic and spectroscopic characterization of a nonheme Fe(IV)=O complex,” Science 2003, 100, 3665-3670. 24 Kaizer, J.; Klinker, E. J.; Oh, N. Y.; Rohde, J. U.; Song, W. J.; Subna, A.; Kim, J.; Nam, W.; Que, L. “Nonheme (FeO)-O-IV complexes that can oxidize the C-H bonds of cyclohexane at room temperature.” J Am Chem Soc 2004, 126, 472-473. 25 Chen, K., Costas, M., Kim, J., Tipton, A. K. & Que, L. Jr. “Olefin cis-dihydroxylation versus epoxidation by nonheme iron catalysts: two faces of an FeIII-OOH coin.” J. Am. Chem. Soc. 2002, 124, 3026–3035. 26 Chen, K.; Que, L. “Stereospecific alkane hydroxylation by non-heme iron catalysts: Mechanistic evidence for an Fe-V = O active species.” J Am Chem Soc 2001, 123, 6327-6337.

11 Chapter 1

tertiary C–H bonds; this reactivity pattern is analogous to what is observed for carbene and nitrene reactions.27

(SbF6)2 Me OAc cat (5 mol%) Me OAc Me OAc N HO H2O2 (1.2 equiv) x3 N MeCN Me AcOH (50 mol%) Me Me Fe Me N MeCN Me CH CN, rt Me N 3 OH

50% yield 11 : 1 Fe(S,S-PDP)

Figure 1.10. Selective tertiary C–H bond hydroxylation with Fe(S,S-PDP).

Despite the important gains made by Nam, Que and White, these reactions are far from achieving the performance of CYP450 transformations. These reactions are plagued by very low turnover numbers (2-6), high catalyst loadings (15 mol%), and the cumbersome procedure of adding catalyst and oxidant portionwise periodically over the course of the reaction. Recently Ribas and Costas have explored modified dipyridyldiamine ligand designs to enhance the catalytic efficiency.28 Bimolecular catalyst decomposition is ascribed to the poor performance of most transition-metal catalyzed hydroxylation reactions. Therefore catalysts that feature an active site that is embedded in an oxidatively robust coordination sphere were considered. By adding large pinene groups to pyridyl ligands, a highly efficient non-heme iron catalyst was designed that gives moderate to high product yields with 2 mol% catalyst loadings (Figure 1.11).

27 a) Du Bois, J. Chemtracts: Org. Chem. 2005, 18, 1; b) Fiori, K. W.; Espino, C. G.; Brodsky, B. H.; Du Bois, J. Tetrahedron 2009, 65, 3042. 28 Gómez, L.; Garcia-Bocsh, I.; Company, A.; Benet-Bucholz, B.; Polo, A.; Sala, X.; Ribas, X.; Costas, M.

“Stereospecific C–H Oxidation with H2O2 Catalyzed by a Chemically Robust Site-Isolated Iron Catalyst.” Angew. Chem. Int. Ed. 2009, 48, 5720-5723.

12 Towards the Selective Catalytic Hydroxylation of C–H Bonds

Me Me

N Me OAc cat (1 mol%) Me OAc Me OAc Me HO H2O2 (1.2 equiv) x2 N OSO2CF3 Me Me Me N Fe AcOH (50 mol%) OSO2CF3 Me Me Me Me N CH3CN OH

62% yield 17 : 1 Me Me

Fe(CF3SO3)2((S,S,R)-mcpp)

Figure 1.11. Improved efficiency for selective tertiary C–H bond hydroxylation with Fe((S,S,R)-mcpp).

The search for synthetic catalysts that mimic the performance of metalloenzymes for hydroxylation reactions has evolved beyond iron chemistry. Methods that utilize alternative transition metals, including rhenium, chromium and ruthenium, have yielded interesting and in some cases impressive results. The same set of guidelines apply for developing these non-iron catalyzed hydroxylation reactions; researchers have focused on exploring simple catalyst structures that are oxidatively robust and can yield highly reactive metallo-peroxo or –oxo intermediates in situ in the presence of inexpensive terminal oxidants. An example of a non-iron hydroxylation catalyst is methyltrioxorhenium (MTO). Treatment of MTO with hydrogen peroxide generates a metallodioxirane in situ. This oxidant is capable of hydroxylating a C–H bond in a stereospecific manner in high conversion (Figure 1.12).29,30 While the TON for this transformation is quite good, product over-oxidation and chemoselectivity is problematic, giving a number of products as the result of methylene oxidation and oxidative cleavage.

29 Murray, R. W.; Iyanar, K.; Chen, J. X.; Wearing, J. T. “Methyltrioxorhenim catalyzed C–H insertion reactions of hydrogen peroxide.” Tetrahedron Letters 1995, 36, 6415-6418. 30 Bianchini, G.; Crucianelli, M.; Canevail, C.; Crestini, C.; Morazzoni, F.; Saladino, R. “Efficient and selective oxidation of methyl substituted cycloalkanes by heterogeneous methyltrioxorhenium-hydrogen peroxide systems.” Tetrahedron 2006, 62, 12326-12333

13 Chapter 1

Me Me Me O OH Me O O 2 mol% MTO O O O Me 65% 18% Re or O Re 25 equiv H2O2 O O O O O Me Me Me t BuOH, 40 ºC Me OH Me active oxidant Me Me 7% 5%

Figure 1.12. Catalytic C–H hydroxylation with MTO.

Impressive results have been achieved by Fuchs and coworkers with their chromium-catalyzed hydroxylation reaction.31 While the mechanism of this transformation is still unknown, this reactionis postulated to occur through a chromoylperiodate, which is formed from CrO3 and HIO4. Decomposition of this intermediate is proposed to generate a highly reactive chromium dioxo, monoperoxo species (Figure 1.13).32 Subsequent reaction with an aliphatic C–H bond is believed to proceed in a concerted manner to afford a Cr(VI)-alkoxide. Particularly remarkable in this reaction is the chemoselectivity; in addition to the selectivity over numerous tertiary C–H bonds in a common substrate, a disubstituted, sterically unencumbered olefin is not epoxidized under these conditions. The reason for this unusual chemoselectivity, however, is not well understood and is unique to this catalytic system.

H OH O O CR3 O O O O O Cr CR CR OH O Cr I Cr O 3 Cr 3 O O O O O O O O

Me Me

Me CrO2(OAc)2 (5 mol%) Me OBz OBz Me Ac2O (3 equiv) Me H5IO6 (3 equiv) Me O Me O -40 ºC, 20 min OH

AcO AcO

Figure 1.13. C–H hydroxylation of a steroid substrate with catalytic chromium.

31 Lee, S.; Fuchs, P. L. “Cephalostatin support studies, part 23. Chemospecific chromium[VI] catalyzed oxidation of C-H bonds at -40 degrees C.” J Am Chem Soc 2002, 124, 13978-13979. 32 Lee, S.; Fuchs, P. L. “An efficient C-H oxidation protocol for alpha-hydroxylation of cyclic steroidal ethers”. Org Lett 2004, 6, 1437-1440.

14 Towards the Selective Catalytic Hydroxylation of C–H Bonds

Recently our laboratory and others have explored the use of a ruthenium-based 33 catalyst for C–H hydroxylation. RuCl3 reacts with a terminal oxidant such as KBrO3 to form a ruthenium tetraoxo species in situ (Figure 1.14). This powerful oxidant is believed to react with C–H bonds in a similar manner to the chromate oxidant described by Fuchs. This method has an advantage over others because it is operationally simple and displays reasonable compatibility with common functional groups.

O Me Me 5 mol% RuCl3 Me OH 10 mol% pyridine O Ru O O N O N Me 3 Me 3 H O 3 equiv KBrO3 Me O O O O Me R CH3CN/H2O, 60 ºC 73 % Me

Figure 1.14. Site-selective ruthenium-catalyzed C–H hydroxylation.

1.4.3 Summary of outer-sphere transition-metal catalyzed C–H hydroxylation methods Considerable progress has been made toward developing transition-metal catalyzed C–H hydroxylation methods. Metalloenzyme-like reactivity is often observed with these systems: in many cases, product analysis indicates that the hydroxylation event proceeds stereospecifically and site-selectivity for a particular C–H bond is fairly good. However, despite many years of research devoted to developing transition-metal catalyzed C–H hydroxylation methods, many common problems still plague these systems. For instance, with very few exceptions, the substrate scope is often limited to simple, unfunctionalized alkanes and these methods are generally incompatible with substrates containing polar functional groups. Additionally, in some cases the substrate concentration must exceed that of the terminal oxidant. Further, in almost every case of a transition-metal catalyzed C–H hydroxylation method, the nature of the active oxidant is not well understood and the fate of the catalyst is difficult to discern. While many researchers have looked to metalloenzymes for clues to developing catalytic C–H hydroxylation methods, there is an alternative strategy that is often overlooked. Flavin-containing monooxygenases (FMOs), like the cytochrome P450s, are known to catalyze the hydroxylation of aliphatic C-H bonds. FMOs, however, catalyze the hydroxylation event without employing a transition-metal cofactor. In terms of developing a synthetic hydroxylation method, this feature is quite appealing: eliminating the metal from the reaction mixture could mitigate problems with metal-catalyzed terminal oxidant

33 McNeill, E.; Du Bois, J. “Ruthenium-Catalyzed Hydroxylation of Unactivated 3º C–H Bonds” Manuscript in preparation.

15 Chapter 1

decomposition and ligand degradation. Furthermore, whereas transition-metal systems are often difficult to study, organic oxidants may be isolated, occasionally crystallized and are readily monitored by NMR. FMOs offer an alternative strategy for developing a bio- inspired C–H hydroxylation method.

1.5 Flavin-Containing Monooxygnease and Organocatalytic C–H Hydroxylation

1.5.1 Flavin-Containing Monooxygenase Microsomal flavin-containing monooxygenases (FMOs), also known as Zeigler’s enzyme, typically oxygenate nucleophilic heteroatoms in a wide range of substrates.34,35 These enzymes also have been implicated as catalysts for aliphatic C–H hydroxylation reactions.36 As opposed to the cytochrome P450s where the heme ligand is a requisite electron-reservoir, flavin adenine dinucleotide (FAD) is the prosthetic group responsible for managing the electrons for the FMOs.34 The flavin substructure is reduced by the terminal reductant NADPH; the reduced form (FADH2) is capable of donating electrons to molecular oxygen to form a 4a-hydroperoxy intermediate (Figure 1.15).37 This intermediate reacts with a substrate in a concerted, two-electron oxygen transfer reaction, distinct from the mechanism of oxygen transfer with the CYP450s. The resulting FAD-OH is protonated, releasing water and regenerating the FAD prosthetic group. The microsomal enzyme exists, for the most part, in an activated state as a peroxyflavin intermediate waiting to react with a substrate; this behavior is in contrast to the other microsomal oxygenase, CYP450, which requires the binding of the substrate to initiate the oxidation sequence.38

34 Eswaramoorthy, S.; Bonanno, J. B.; Burley, S. K.; Swaminathan, S. “Mechanism of action of a flavin- containing monooxygenase.” P Natl Acad Sci USA 2006, 103, 9832-9837. 35 Poulsen, L. L.; Ziegler, D. M. “Liver microsomal FAD-containing mono-oxygenase – spectral characterization and kinetic studies.” J Biol Chem 1979, 254, 6449-6455. 36 Li, L.; Liu, X.; Yang, W.; Xu, F.; Wang, W.; Feng, L.; Bartlam, M.; Wang, M.; Rao, Z. “Crystal structure of long-chain alkane monooxygenase (LadA) in complex with coenzyme FMN: Unveiling the long-chain alkane hydroxylase.” J Biol Chem 2008, 376, 453-465. 37 Ziegler, D. M. “An overview of the mechanism, substrate specificities, and structure of FMOs.” Drug Metab Rev 2002, 34, 503-511. 38 Mitchell, S. C. “Flavin mono-oxygenase (FMO) - The 'other' oxidase.” Curr Drug Metab 2008, 9, 280- 284.

16 Towards the Selective Catalytic Hydroxylation of C–H Bonds

R S SO R Me N N O Me N N O

FAD-OOH N N FAD-OH Me N R Me N R H O H O O O OH H

O2 H2O R H R Me N N O Me N N O

FADH2 FAD N N Me N R Me N R H NADPH O O

Figure 1.15. Proposed mechanism for C–H hydroxylation catalyzed by FMO.

An example of a C–H hydroxylation reaction catalyzed by an FMO is illustrated in Figure 1.16.36 LadA, a long-chain alkane monooxygenase, catalyzes the formation of a terminal alcohol. This enzyme targets long-chain alkanes, specifically carbon chains longer than fourteen carbons. The active site consists of a deep, hydrophobic cage pocket that seems to bind and position the terminal methyl group proximal to the flavin cofactor. The cavity serves two roles in addition to selecting for particular length aliphatic chains; it provides a proton donor (His138) that leads to the formation of the C4a- hydroperoxyflavin and it creates a solvent-free environment that prevents the rapid decomposition of the flavin-peroxide.

LadA OH R = CH (CH ) – where 12 ! n ! 33 R Me R 3 2 n O2

Figure 1.16. An example of a C–H hydroxylation reaction catalyzed by FMO.

The biosynthesis of resistoflavin provides another example of an FMO-catalyzed hydroxylation. Resistoflavin is a polyketide with a unique boat-like shape that is derived from the flat resistomycin. Recently the biosynthetic pathway relating the two structures has been elucidated, indicating that resitsomycin is a substrate for an FAD-dependent monooxygenase, RemO (Figure 1.17).39 This is a particularly unusual transformation in that most FMOs functionalize the periphery of polyphenols while RemO catalyzes the hydroxylation of a central carbon atom in a fused ring system.40

39 Ishida, K.; Maksimenka, K.; Fritzsche, K.; Scherlach, K.; Bringmann, G.; Hertweck, C. “The boat- shaped polyketide resistoflavin results from Re-facial central hydroxylation of the discoid metabolite resistomycin.” J Am Chem Soc 2006, 128, 14619-14624. 40 For examples of other natural products originating from a similar, but peripheral, oxygenation, see the references in reference (39).

17 Chapter 1

Me Me Me Me HO O O O

Me OH RemO Me OH OH

OH O OH OH O OH resistomycin resistoflavin

Figure 1.17. Hydroxylation of RemO with an FMO.

Examples of hydrocarbon oxidation by FMOs suggest that an alternative, biologically inspired solution to transition-metal catalyzed C–H hydroxylation exists. Organic molecules can catalyze C–H hydroxylation and may have unique advantages over transition-metal catalyzed processes (Section 1.4).

1.5.2 A Survey of Potentially-Catalytic Organic Oxidation Methods To design a non-metallic C–H hydroxylation catalyst, a suitable organic mediator must be identified. Both ketones and imines, when treated with an appropriate terminal oxidant, can form reactive oxidizing reagents, although no examples of such groups functioning as catalysts for C–H hydroxylation have appeared.41 Therefore, the focus of this half of the review has been broadened to include both organic molecules that can catalyze O-atom transfer reactions to olefin and sulfide substrates as well as stoichiometric organic reagents that can affect C–H hydroxylation reactions. Three classes of organic catalysts will be discussed: synthetic flavins, ketones, and imines. 1.5.2a Synthetic Flavin Catalyzed Oxidations Nature’s example of a hydroxylating organocatalyst is the flavin-containing monooxygenase (FMO). The core of the flavin structure is 7,8-dimethylisoalloxazine and the active form of the catalyst is the enzyme bound 4a-hydroperoxyflavin (Figure 1.18).42 It is difficult to employ the hydroperoxyflavin as an oxidation catalyst because this molecule will spontaneously eliminate hydrogen peroxide when not protected by the enzyme cavity.42 Isolated riboflavin is ineffective as a catalyst for oxygen transfer reactions. Efforts to design more stable hydroperoxyflavin mimics involve substituting the 5-N position. For instance, by adding an ethyl group to the 5-N, Murahashi et al found a

41 Adam, W.; Saha-Moller, C. R.; Ganeshpure, P. A.; “Synthetic Applications of Nonmetal Catalysts for Homogeneous Oxidations.” Chem Rev 2001, 101, 3499-3548. 42 Gelalcha, F. G. “Heterocyclic hydroperoxides in selective oxidations.” Chem Rev 2007, 107, 3338- 3361.

18 Towards the Selective Catalytic Hydroxylation of C–H Bonds

catalyst for heteroatom oxidation employing hydrogen peroxide as the terminal oxidant.43 Substrates such as dialkylamines and sulfides are oxidized to nitrones and sulfoxides, respectively, using this system.

10 Me ClO - H 1 4 Me 8 N N O Me N N O 10 mol% cat. 2 equiv H2O2 3 N Ph2S Ph2SO 4a NH Me 7 N Me N Me MeOH, rt 5 96% O Et O

Figure 1.18. An example of a synthetic FMO.

Murahashi has also attempted to use molecular oxygen as the terminal oxidant with flavin-based catalysts.44 Knowing that molecular oxygen will react with a reduced form of the flavin catalyst (Figure 1.15), a mild reductant was sought that would take the cationic form of the catalyst to the reduced form. Hydrazine was identified as an excellent reductant for this purpose. Fluorinated solvents such as trifluoroethanol and hexafluoroisopropanol were chosen due to the high solubility of molecular oxygen in these media. The substrate scope is similar to their previous work but the use of molecular oxygen in an organocatalytic O-atom transfer reaction is truly remarkable.

Me S SO Me Me N N O Me N N O

N N Me N Me Me N Me O O Et O Et O OH H S = R2NH H+ R3N R2S O H2O2 2 H2O Z = H NNH Me Me 2 2 Me N N O Me N N O

N N Me N Me Me N Me Et O Z+ ZH Et O

Figure 1.19. Catalytic substrate oxidation with a synthetic FMO using molecular oxygen as the terminal oxidant.

1.5.2b Ketone Catalysts and C–H Hydroxylation by Dioxiranes Perhaps the most generally applied small molecule oxidation catalysts are ketones. Treatment of a ketone with an oxidizing agent can result in dioxirane formation

43 Murahashi, S. I.; Oda, T.; Masui, Y. “Flavin-catalyzed oxidation of amines and sulfur compounds with hydrogen peroxide.” J Am Chem Soc 1989, 111, 5002-5003. 44 Imada, Y.; Iida, H.; Ono, S.; Murahashi, S. I. “Flavin catalyzed oxidations of sulfides and amines with molecular oxygen.” J Am Chem Soc 2003, 125, 2868-2869.

19 Chapter 1

or a perhydrate, depending on the choice of terminal oxidant. Both dioxiranes and perhydrates are capable of O-atom transfer reactions and have enjoyed widespread use as reagents for alkene epoxidation. These reagents can be rendered catalytic with an appropriate terminal oxidant, as first shown by Curci.45 Sheldon and coworkers have found that perfluorinated, highly electron-deficient ketones such as hexafluoroacetone and perfluoroheptadeca-9-one will react with hydrogen peroxide to form a peroxyketal adduct (Figure 1.20).46,47 This species will transfer oxygen to olefins and regenerate the ketone catalyst, which can then be reoxidized under the reaction conditions.

5 mol% hexafluoroacetone Me Me Me Me O O 2 equiv. H2O2 HO O H Me rt, 1M hexafluoroisopropanol Me F3C CF3 99%

Figure 1.20. Catalytic olefin epoxidation with hexafluoroacetone and hydrogen peroxide.

If a persulfate such as Oxone® is employed as the terminal oxidant instead of hydrogen peroxide, a dioxirane is formed from the ketone (Figure 1.21).48 This strained heterocycle is capable of a wide-range of oxidation chemistry, including C–H hydroxylation. Furthermore, as with the previously described perhydrate chemistry, in certain cases only a catalytic amount of the ketone mediator is required. Denmark, Curci, Shi and Yang have done extensive work on developing a range of ketones as mediators for catalytic olefin epoxidation.

45 Curci, R.; Fiorentino, M.; Serio, M, R. “Asymmetric epoxidation of unfunctionalized alkenes by dioxirane intermediates generated from potassium peroxomonosulphate and chiral ketones.” J Chem Soc Chem Comm 1984, 3, 155-156. 46 Van Vliet, M.; Arends, I. W. C. E.; Sheldon, R. “Hexafluroacetone in heafluoro-2-propanol: A highly active medium for epoxidation with aqueous hydrogen peroxide.” SynLett 2001,1305. 47 Van Vliet, M.; Arends, I. W. C. E.; Sheldon, R. “Perfluoroheptadecan-9-one: a selective and reusable catalyst for epoxidations with hydrogen peroxide.” Chem. Commun. 1999, 263-264. 48 Murray, R. W.; Singh, M.; Jeyaraman, R. “Dioxiranes 20. Preparation and properties of some new dioxiranes.” J Am Chem Soc 1992, 114, 1346-1351.

20 Towards the Selective Catalytic Hydroxylation of C–H Bonds

R2SO

R N+–O- R2S 3 ROH R3N RH

O O R O R O KHSO O O 5 R R R R O O O SO - R R R R 3 R R R X R R R R R R O R X = OLi, O R O OTMS R R HO R

Figure 1.21. Dioxiranes can mediate a range of oxygenation reactions.

Denmark has thoroughly investigated the reaction conditions for ketone-catalyzed epoxidations.49 There are many points where the reaction can deviate from the desired catalytic cycle and it was determined that maintaining a precise pH was necessary to avoid some of these competing processes (Figure 1.22). For instance, at a pH greater than 8, Oxone self-decomposition is quite rapid relative to dioxirane formation (pathway b). After Oxone adds to the ketone, the Criegee intermediate can decompose in one of two ways: the intermediate can break down to form the desired dioxirane (pathway c) or an alkyl group can migrate in a Bayer-Villiger type reaction (pathway d), resulting in the irreversible consumption of ketone. Depending on the relative concentration of Oxone and other reaction conditions, the dioxirane can either go on to transfer oxygen to - substrate or be consumed by peroxomonosulfate HSO5 in an unproductive process (pathway e). Elucidating the causes for these unproductive processes allowed Denmark to devise ideal reaction conditions, such as slow addition of Oxone in a buffered solution using a biphasic system.

49 Denmark, S. E.; Forbes, D. C.; Hays, D. S.; Depue, J. S.; Wilde, R. G. “Catalytic epoxidation of alkenes with Oxone.” J. Org. Chem. 1995, 60, 1391-1407.

21 Chapter 1

O SO SO 2- - 5 2- R R HSO5 HSO4- + O2 + SO4 S b

a O - O O - HSO5 HSO4 + O2 + R R e R R O O O c R O - O R O SO3 R O SO - d R O R R 3

Figure 1.22. Catalytic cycle for dioxirane-mediated olefin epoxidation.

Denmark has systematically examined the ketone structure to identify steric and electronic factors that facilitate the formation of dioxirane. A summary of this work is presented in Table 1. Perhaps unsurprisingly, substituted ketones such as 2-butanone and 3-pentanone performed poorly under the reaction conditions compared to acetone (entries 1-3). The behavior of cyclic ketones was more difficult to predict and it is surprising that the performance is so highly dependent on ring size (entries 4-6). There are two competing trends at play with the cyclic ketones that determine the epoxidation efficiency: the propensity toward rehybridization of the carbonyl carbon from sp2 to sp3 and the tendency for ring-expansion via Bayer-Villiger oxidation. For instance, relief of ring strain would suggest that cyclobutanone should yield the most potent dioxirane for oxygen transfer. Under catalytic conditions, however, the Bayer-Villiger oxidation (pathway c in Figure 1.21) is very rapid for cyclobutanone and only a trace amount of product is observed.

Table 1.1. Effect of ketone structure on the efficiency of catalytic olefin epoxidation.

2 equiv ketone 10 equiv Oxone O Me OBn pH 7.8 DCM/H2O Me OBn 10% nBu4HSO4 24h, 0 ºC

entry ketone epoxide : olefin ratio

1 acetone 87 : 13 2 2-butanone 40 : 60 3 3-pentanone 5 : 95 4 cyclobutanone 2 : 98 5 cyclopentanone 3 : 97 6 cyclohexanone 67 : 33 7 1,1,1-trifluoroacetone 29 : 71 8 hexafluoroacetone 2 : 98

22 Towards the Selective Catalytic Hydroxylation of C–H Bonds

Despite the fact that trifluoromethyldioxirane (TFDO) has been shown to be a highly reactive reagent for stoichiometric olefin epoxidation, 1,1,1-trifluoroacetone and hexafluoroacetone performed very poorly as catalysts for alkene epoxidation.50 Catalyst inhibition is due to the highly electrophilic nature of these ketones; stable hydrates are formed under aqueous conditions and exchange with peroxymonosulfate is therefore slow (Figure 1.23).

O O – – O O O OH HSO5 H3C R O – H C CF OH O SO3 H3C CF3 3 3 F3C R

Figure 1.23. The formation of a stable hydrated ketone inhibits catalytic turnover.

In addition to mediating catalytic epoxidation reactions, dioxiranes are excellent reagents for effecting C–H hydroxylation. Curci and coworkers have spearheaded efforts to develop and refine these reagents for hydroxylation, particularly dimethyldioxirane (DMDO) and trifluoromethyldioxirane (TFDO).51,52 These reagents react with unactivated C–H bonds to transfer oxygen in a stereospecific manner (Figure 1.24).

O O DMDO DMDO 84% yield H3C CH3 17h, 25 ºC OH

Ph TFDO Ph O O TFDO Me H Me OH 90% yield H3C CF3 Et 1h, -23 ºC Et 72% ee 72% ee

Figure 1.24. Dioxiranes react with C–H bond in a stereospecific manner.

Considerable mechanistic and computational studies have been conducted on dioxirane reactions. As with metal-oxo reactions, there are at least two possible pathways for oxygen transfer: a concerted “oxene”-type mechanism and a fast, C–H abstraction/radical rebound mechanism. Experiments by the Curci group and others suggest that the reaction proceeds through a concerted asynchronous transition state. When a chiral, tertiary C–H bond substrate is treated with TFDO, the enantiopurity of the product is found to match that of the starting material (Figure 1.24). The absence of a

50 Mello, R.; Fiorentino, M.; Sciacovelli, O.; Curci, R. J. Org.Chem. 1988, 53, 389. 51 Curci, R.; D’accolit, L.; Fusco, C. “A Novel Approach to the Efficient Oxygenation of Hydrocarbons under Mild Conditions. Superior Oxo Transfer Selectivity Using Dioxiranes.” Accounts Chem Res, 2006, 39, 1-9. 52 Mello, R.; Fiorentino, M.; Sciacovelli, O.; Curci, R. “Oxidations by Methyl(trifluoromethyl)dioxirane .2. Oxyfunctionalization of saturated-hydrocarbons.” J. Am. Chem. Soc. 1989, 111, 6749-6757.

23 Chapter 1

radical intermediate under the reaction conditions is further indicated by the stereospecific hydroxylation of cis-1,2-dimethylcyclohexane. Radical clock experiments provide some of the most convincing experimental evidence for a concerted C–H hydroxylation event. Newcomb and coworkers developed a hypersensitive radical probe 1.1 (Figure 1.25); the radical formed by homolytic H-atom abstraction from the cyclopropylcarbinyl position rearranges to the ring-opened radical 1.2 with a rate constant of 1 x 1011 s-1 at ambient temperature.53 When 4 equiv of radical probe 1.1 was mixed with DMDO, the only products observed were alcohol A and ketone B. The lack of any ring opened-product suggests that if a radical was formed during the hydroxylation event, its maximum lifetime is only about 200 fs. These results strongly implicate a concerted mechanism for C–H hydroxylation with DMDO.

Me Me Me Ph Ph Ph 1.1 1.2

Me HO Me O Me O O acetone OH Me Me Me 22 ºC Ph Ph Ph Ph 4 equiv 1.1 A B not observed

Figure 1.25. Radical clock experiments suggest that dioxirane-mediated C–H hydroxylation occurs in a concerted fashion.

Houk and coworkers have investigated the transition state of both C–H hydroxylation and olefin epoxidation with dioxiranes through DFT calculation.54 For olefin epoxidation with dioxiranes, calculations confirm that the oxygen-atom transfer event proceeds through a concerted mechanism. There are two possible orientations for the transition structure: spiro and planar (Figure 1.26). By locating stationary points for both transition structures,55 Houk showed that there is a large preference (7.4 kcal/mol) for the spiro orientation. To put this number into context, the activation barrier for epoxidation of

53 Simakov, P, A.; Choi, S.-Y.; Newcomb, M. “Dimethyldioxirane Hydroxylation of a Hypersensitvie Radical Probe: Supporting Evidence for an Oxene Insertion Pathway” Tetrahedron Lett. 1998, 39, 8187- 8190. 54 Houk, K. N.; Liu, J.; Demello, N. C.; Condroski, K. R. “Transition States of Epoxidations: Diradical Character, Spiro Geometries, Transition State Flexibility, and the Origins of Stereoselectivity.” J Am Chem Soc 1997, 119, 10147-10152. 55 To make this comparison, a second order saddle point had to be identified for the planar transition structure. A true transition state was only identified for the spiro transition structure.

24 Towards the Selective Catalytic Hydroxylation of C–H Bonds

ethylene with dioxirane is calculated to be only 12.9 kcal/mol. Furthermore, a small but significant preference (2.4 kcal/mol) for a synchronous transition state was calculated. Singleton, however, has asserted that even very high-level calculations cannot unambiguously assign a synchronous or asynchronous transition state for olefin epoxidation with dioxiranes.56

Figure 1.26 Four different transition structures for the epoxidation of ethylene with dioxdirane.57

For C–H hydroxylation, calculations performed by Bach, Houk and Rauk suggest a similar concerted mechanism to olefin epoxidation is operative (Figure 1.27).58,59,60 A spiro, asynchronous transition structure characterizes the interaction between a C–H bond and dioxirane. In contrast with the olefin epoxidation transition structure, however, the transition state for C–H hydroxylation is highly polar with substantial carbocationic character at the carbon atom of the substrate. At the transition state, the C–H bond is partially broken, the O–H bond is almost entirely formed, and the O–O bond of the dioxirane is partially broken. Despite the more asynchronous nature of the C–H hydroxylation event, the mechanisms for olefin epoxidation and C–H hydroxylation share many common features.

56 Singleton, D. A.; Wang, Z. H. “Isotope effects and the nature of enantioselectivity in the Shi epoxidation. The importance of asynchronicity.” J. Am. Chem. Soc. 2005, 127, 6679-6685. 57 Figure is adapted from reference (54) 58 Glukhovtsev, M. N.; Canepa, C.; Bach, R. D. “The nature of the transition structure for the oxidation of alkanes with dioxiranes.” J. Am. Chem. Soc. 1998, 120, 10528-10533. 59 Du, X. H.; Houk, K. N. “Transition states for alkane oxidations by dioxiranes.” J. Org. Chem. 1998, 63, 6480-6483. 60 Shustov, G. V.; Rauk, A. “Mechanism of dioxirane oxidation of CH bonds: Application to homo- and heterosubstituted alkanes as a model of the oxidation of peptides.” J. Org. Chem. 1998, 63, 5413-5422.

25 Chapter 1

O O Me Me Me Me O O H H H H H H H H

Figure 1.27. The transition structures for olefin epoxidation and C–H hydroxylation with dioxiranes are similar.

Dioxiranes react with a strong preference for tertiary C–H bonds, consistent with the electrophilic nature of the oxidation event. Remarkable examples of chemoselective tertiary C–H bond hydroxylation with DMDO or TFDO can be found in the literature, 61 including the reaction of tetraepoxide of 3β-acetyl vitamin D2 (Figure 1.28). This reaction is particularly notable given the presence of five other tertiary C–H bonds. Other striking examples include Wender’s hydroxylation of a late-stage bryostatin intermediate with DMDO (Figure 1.2).

Me Me Me Me Me Me O Me O Me Me H Me OH 3 equiv TFDO O O CH2Cl2, 0 ºC, 1.7h O O

O O 61% AcO AcO

Figure 1.28. Hydroxylation of a single C–H bond with TFDO.

The site selectivity exhibited by dioxiranes is not entirely attributable to electronic effects within a molecule. The transition state for oxygen-atom transfer requires strict orbital alignment between the O–O σ* of the dioxirane and the C–H σ of the substrate. In many cases, electronically favorable C–H hydroxylation reactions will not occur due to steric impediments that block the necessary orbital alignment. An excellent example of this is the selective oxygenation of the Grundmann ketone by DMDO (Figure 1.29).62 The C25 hydroxylated product was isolated in 86% yield, although there are several other tertiary C–H bonds present in the molecule. As illustrated in the molecular model in

61 Curci, R.; Detomaso, A.; Lattanzio, M. E.; Carpenter, G. B. “Oxidation of natural targets by dioxiranes. 4. High stereo- and regioselective conversion of vitamin D2 to its (all-R) tetraepoxide and C-25 hydroxy derivative.” J. Am. Chem. Soc. 1996, 118, 11089-11092. 62 Bovicelli, P.; Lupatetelli, P.; Minicione, E.; Prencipe, T.; Curci, R. “Oxidation of natural products by dioxiranes .2. Direct hydroxylation at the side chain C25 of cholestane derivatives and of vitamin D3 Windaus-Grundmann ketone.” J. Org. Chem. 1992, 57, 5052-5054.

26 Towards the Selective Catalytic Hydroxylation of C–H Bonds

Figure 1.29, the site selectivity over other tertiary bonds is probably due to the optimal stereoalignment offered at C25.

Me Me Me 3 equiv DMDO Me CH2Cl2/acetone Me Me Me 20 ºC, 48h Me OH H H O O 86% yield

Figure 1.29. Selective C–H hydroxylation at C25 is attributed to optimal stereoalignment with the dioxirane. Figure taken from reference (39).

Dioxiranes should be ideal mediators for catalytic C–H hydroxylation reactions. They exhibit good selectivity towards electron-rich C–H bonds, react in a stereospecific fashion, and in principle, should be catalytic reagents, as exemplified by the numerous catalytic epoxidation reactions that they mediate. Unfortunately, as was shown in Figure 1.22, dioxiranes have a short half-life in the presence of peroxymonosulfate. With reactive substrates such as olefins, the rate of the substrate-dioxirane reaction is sufficiently fast to outcompete dioxirane decomposition. When the substrate is less reactive, such as an aliphatic C–H bond, oxygen transfer to substrate is much slower and dioxirane decomposition becomes the major reaction pathway. Yang and coworkers observed this phenomenon when attempting the catalytic hydroxylation of adamantane, arguably one of the easiest C–H hydroxylation substrates. Only 18% conversion to 1-

27 Chapter 1

adamantanol was achieved despite employing 100 mol% trifluoroacetone in a buffered Oxone solution (Figure 1.30).63

trifluoroacetone (100 mol%) OH Oxone/NaHCO3 18% conversion CH3CN/H2O, rt

Figure 1.30. The short half-life of dioxiranes in the presence of Oxone precludes the development of a catalytic C–H hydroxylation reaction mediated by dioxiranes.

An obvious way to increase the rate of the dioxirane-mediated C–H insertion is to take advantage of an intramolecular oxygen transfer event.64 By positioning the ketone proximal to the C–H bond, the rate of oxygen transfer to substrate is considerably faster and this outcompetes other decomposition pathways. As shown by Yang, reaction of substrate 1.3 with buffered Oxone gives tetrahydropyran hemiketal product 1.4 in high yield and with perfect site-selectivity (Figure 1.31). While this is a clever strategy for circumventing the problem of competitive dioxirane decomposition, it is not a solution to designing a general catalytic C–H hydroxylation reaction mediated by dioxiranes.

HO CF H H 3 CF3 Oxone/NaHCO3 CF3 O 78% O CH3CN/H2O, rt O O 1.3 1.4

Figure 1.31. Stoichiometric intramolecular oxygen-atom transfer with an in situ generated dioxirane.

1.5.2c Imine Catalysts and C–H Hydroxylation with Oxaziridines Oxaziridines are a class of strained three-membered ring heterocycles with many similarities to dioxiranes. The substrate scope for oxaziridine-mediated oxygen-atom transfer reactions is significantly limited compared to that of dioxirane-mediated oxygenation reactions. Generally oxaziridines are employed for the oxidation of strong nucleophiles such as metal enolates and sulfides. In certain instances, oxaziridines will

63 Yang, D.; Wong, M. K.; Wang, X. C.; Tang, Y. C. “Regioselective intramolecular oxidation of unactivated C–H bonds by dioxiranes generated in situ.” J. Am. Chem. Soc. 1998, 120, 6611-6612. 64 (a) Wong, M. K.; Chung, N. W.; He, L.; Yang, D. “Substituent effects on regioselective intramolecular oxidation of unactivated C–H bonds: Stereoselective synthesis of substituted tetrahydropyrans.” J. Am. Chem. Soc. 2003, 125, 158-162; (b) Yang, D; Wong, M. K.; Yan, Z. “Regioselective Intramolecular Oxidation of Phenols and Anisoles by Dioxiranes Generated in situ.” J. Org. Chem. 2000, 65, 4719-4184; (c) Wong, M.; Chung, N.; He, L.; Wang, X.; Yan, Z.; Tang, Y.; Yang, D. “Investigation of the Regioselectivities of Intramolecular Oxidation of Unactivated C–H Bonds by Dioxiranes Generated in situ.” J. Org. Chem. 2003, 68, 6321-6328.

28 Towards the Selective Catalytic Hydroxylation of C–H Bonds

mediate olefin epoxidation and in the case of highly electron-deficient oxaziridines, C–H hydroxylation is possible.41 Imines, when treated with a peroxidic oxidant like Oxone or a peracid, will form the corresponding oxaziridine. As with dioxiranes and ketones, a catalytic process could be effected where oxygen transfer from an oxaziridine generates an imine, which could be subsequently reoxidized back to the oxaziridine. As stoichiometric oxygen-transfer reagents, oxaziridines are generally less reactive than dioxiranes, as evidenced by their significantly limited substrate scope and higher calculated activation barriers for olefin epoxidation and other oxygen-transfer reactions.54 Given that dioxiranes and oxaziridines have similarities in ring strain and molecular geometry, the disparity in O-atom transfer capability between the two classes of oxidants is puzzling (Figure 1.32).65 One possible explanation is that there exists a stronger thermodynamic driving force for oxygen transfer from dioxiranes: N–O bonds are typically stronger than O–O bonds and the resulting N=C double bond is weaker than the C=O double bond formed after oxygen transfer from the dioxirane.66,67

O O O NH Bond BDE (kcal/mol) O–O 23 H H H H dioxirane O–Operacid ~50 17.3 18.8 C=O 90 C=N 77 G2(MP2) Strain Energies (kcal/mol)

Figure 1.32. Comparison of physical characteristics of dioxiranes and oxaziridines.

Houk and coworkers have compared calculated transition structures for dioxirane- and oxaziridine-mediated epoxidation and found that both transformations are characterized by spiro, concerted atom-transfer event.54 Analysis of the frontier molecular orbitals (FMOs), explains the nature of the spiro orientation (Figure 1.33). The interaction between the oxaziridine HOMO and the olefin LUMO is maximized in the spiro transition structure, as with dioxiranes. The transition state for oxaziridine-mediated oxygen-atom transfer, however, is considerably more asynchronous than with dioxiranes. Many 41,54,68 researchers liken oxygen transfer with these reagents to an SN2 process. Nucleophilic attack at oxygen implies significant negative charge development on the

65 Bach, R. D.; Dmitrenko, O. “Effect of Geminal Substitution of the Strain Energy of Dioxiranes. Origin of the Low Ring Strain of Dimethyldioxirane.” J. Org. Chem. 2002. 67, 3884-3896. 66 Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic Chemistry; 3rd ed.; Harper Collins Publishers, Inc.: New York, 1987. 67 Cremer, D.; Kraka, E.; Szalay; “Decomposition modes of dioxirane, methyldioxirane and dimethyldioxirane - a CCSD(T), MR-AQCC and DFT investigation.” Chem Phys Lett 1998, 292, 97-109. 68 Bach, R. D.; Wolber, G. J. J. Am. Chem. Soc. 1984, 106, 1410-1415.

29 Chapter 1

nitrogen of the oxaziridine and the other oxygen of the dioxirane. Because the nitrogen is less able to stabilize this negative charge, interaction between the filled n oxygen orbital and one side of the olefin π* becomes more pronounced, resulting in a highly asynchronous transition state.

Figure 1.33. Olefin epoxidation with an oxaziridine is characterized by a spiro, asynchronous transition structure.

There has not been a comprehensive theoretical study on oxaziridine-mediated C–H hydroxylation. In fact, with the exception of the work reported by this lab, there does not seem to be any literature that documents molecular modeling of a transition state involving the transfer of oxygen from an oxaziridine to a saturated aliphatic substrate.69 However, as shown previously in this chapter, DFT calculations of dioxirane-mediated epoxidation and C–H hydroxylation proceed through very similar transition structures. Furthermore, dioxirane- and oxaziridine-mediated olefin epoxidation have also been shown to proceed through similar transition states. Our calculations confirm that C–H hydroxylation with oxaziridines exhibit similar transition structures to those mediated by dioxiranes. Davis oxaziridines are the most widely applied oxaziridines in organic synthesis.70 There are two general structures for Davis oxaziridines (Figure 1.34) and they are both N- sulfonyl-derived.71 The placement of a sulfonyl group on the nitrogen of the oxaziridine greatly influences the intrinsic reactivity of these reagents; smaller, less electron- withdrawing groups on nitrogen give oxaziridines that transfer nitrogen in preference to

69 Brodsky, B. H.; Du Bois. J. “Oxaziridine-mediated catalytic hydroxylation of unactivated 3º C-H bonds using hydrogen peroxide.” J Am Chem Soc 2005, 127, 15391-15393. 70 Davis, F. A.; Sheppard, A. C. “Applications of oxaziridines in organic synthesis.” Tetrahedron 1989, 45, 5703-5742. 71 Davis, F. A.; Lamendola, J.; Nadir, U.; Kluger, E. W.; Sedergran, T. C.; Panunto, T. W.; Billmers, R.; Jenkins, R.; Turchi, I. J.; Watson, W. H.; Chen, J. S.; Kimura, M. J. Am. Chem. Soc. 1980, 102, 2000- 2005.

30 Towards the Selective Catalytic Hydroxylation of C–H Bonds

oxygen.72 Davis oxaziridines are uniquely useful for sulfur oxidation reactions because oxygen transfer to sulfides gives sulfoxide products with minimal amounts of over- oxidation.73 These reagents have also been employed for the stereospecific epoxidation of olefins; high conversions are achieved with simple olefin substrates.74 Perhaps the most common use of Davis oxaziridines is the α-hydroxylation of metal enolates. Treatment of a preformed enolate at -78 ºC with an N-sulfonyl oxaziridine results in an α- hydroxy carbonyl compound. Diastereoselective and enantioselective versions of this transformation have been achieved with chiral auxiliaries and optically active oxaziridines.75

R SO Davis Oxaziridines: 2

O O R N+–O- R2S 3 S ROH R3N O N Ph RH O O H S R N R3 R O R R R RLi/RMgX R = H, NO2 R 1 2 R X R ROH R C H R R 4 9 R O O R X = OLi, OTMS R N O S R R O O HO R

Figure 1.34. Oxaziridines can effect oxygenation reactions with nucleophilic substrates.

Almost all of the oxidation reactions that employ oxaziridines are stoichiometric transformations. The resulting imines can, in some cases, be isolated at the end of the reaction, suggesting that such processes could be rendered catalytic if an appropriate

72 Armstrong, A.; Edmonds, I. D.; Swarbrick, M. E.; Treweeke, N. R. “Electrophilic amination of enolates with oxaziridines: effects of oxaziridine structure and reaction conditions.” Tetrahedron 2005, 61, 8423- 8442. 73 Davis, F. A.; Reddy, R. T.; Han, W.; Carrol, P. J. “Chemistry of Oxaziridines .17. N- (phenylsulfonyl)(3,3-dichlorocamphoryl)oxaziridine – a highly efficient reagent for the asymmetric oxidation of sulfides to sulfoxides.” J. Am. Chem. Soc. 1992, 114, 1428-1437. 74 Davis, F. A.; Harakal, M. E.; Awad, S. B.; “Chemistry of Oxaziridines .4. Asymmetric epoxidation of unfunctionalized alkenes using chiral 2-sulfonyloxaziridines – evidence for a planar transition state geometry.” J. Am. Chem. Soc. 1983, 105, 3123-3126. 75 (a) Evans, D. A.; Morrissey, M. M.; Dorow, R. L. “Asymmetric oxygenation of chiral imide enolates – a general approach to the synthesis of enantiomerically pure alpha-hydroxy carboxylic-acid synthons.” J. Am. Chem. Soc. 1985, 107, 4346-4348. (b) Davis, F. A.; Weismiller, M. C.; Murphy, C. K.; Reddy, R. T.; Chen, B. C. “Chemistry of oxaziridines .18. Synthesis and enantioselective oxidations of the [(8,8-dihalocamphoryl)sulfonyl]oxaziridines.” J. Org. Chem. 1992, 57, 7274-7285.

31 Chapter 1

terminal oxidant were to be identified.76 The groups of Davis and Page have explored imine-catalyzed oxidations of sulfides, including asymmetric variants. Davis et al demonstrated the conversion of sulfides to sulfoxides with 20 mol% 1.5 (Figure 1.35).77 Page attempted an asymmetric version of this same reaction with Davis oxaziridine analogue 1.6, although the catalyst loading was quite high (100 mol%) and the enantiomeric excess was poor. These reactions represent the only examples of which we are aware of imine-catalyzed oxygen-atom transfer reaction.

O O O S S 20 mol% 1.3 N Ph Oxone S H CH2Cl2, 22 ºC O2N 90% yield 1.5

Me Me 100 mol% 1.4 O S Me 4 equiv aq. H2O2 S Me N Me 4 equiv DBU DCM, -15 ºC, 14h S Me O 61% yield O 25% ee 1.6

Figure 1.35. Rare examples of oxaziridine-mediated catalytic sulfide oxidation.

While examples of sulfide and enolate oxidations with oxaziridines are can be found in the literature, there are only two known systems for C–H bond hydroxylation mediated by oxaziridines. Desmarteau has shown that highly electron-deficient perfluorinated oxaziridines are well suited for oxygen transfer chemistry. Not only are heteroatoms and olefins oxidized with these reagents but non-activated C–H bonds of aliphatic substrates can be hydroxylated.76 In one such example, a single tertiary C–H bond at the A/B ring fusion of a steroid substrate is stereospecifically hydroxylated in 70% yield (Figure 1.36).78 It is interesting to note that product analysis suggests that the reactivity of these perfluorinated oxaziridines is similar to dioxiranes; however, no calculated data or kinetics analysis to evaluate the transition state or activation barrier have appeared for these reagents.

76 Petrov, V. A.; Resnati, G. “Polyfluorinated oxaziridines: Synthesis and reactivity.” Chem. Rev. 1996, 96, 1809-1823. 77 Davis, F. A.; Lal, S. G. Durst, H. D. “Chemistry of oxaziridines. 10. Selective catalytic oxidation of sulfides to sulfoxides using N-sulfonyloxaziridines.” J. Org. Chem. 1988, 53, 5004-5007. 78 Arnone, A.; Cavicchioli, M.; Montarnari, V.; Resnati, G. “Direct oxyfunctionalization at unactivated sites – synthesis of 5-beta-hydroxysteroids by perfluorodialkyloxazirines.” J. Org. Chem. 1994, 59, 5511-5513.

32 Towards the Selective Catalytic Hydroxylation of C–H Bonds

Me Me Me O Me N n-C4H9 n-C3F7 Me CO2Me Me CO2Me F

CFCl3, 23 ºC AcO AcO H OH 70%

Figure 1.36. A polyfluorinated oxaziridine effects the hydroxylation of a steroidal C–H bond.

Brodsky and Du Bois reported the only other example of oxaziridine-mediated C– H hydroxylation (Figure 1.37).69 With a substituted 1,2,3-benzoxathiazine-2,2-dioxide- derived oxaziridine, adamantane is hydroxylated in 80% yield. Remarkably, through judicious choice of terminal oxidant and co-catalyst, this transformation is catalytic in the oxaziridine-precursor, making it one of only three reported examples of a catalytic oxaziridine-mediated oxygen-atom transfer reaction. Furthermore, this is the first and only known organocatalytic method for C–H hydroxylation. Although the substrate scope is limited to a very small number of simple aliphatic substrates, the Brodsky system is a promising lead for the development of a general method catalytic C–H hydroxylation.

20 mol% catalyst OH Cl O Cl O 1 mol% Ar2Se2 O O S S 4 equiv UHP N O N O DCE, 95h F3C F3C O 80% catalyst oxaziridine

Figure 1.37. The only known organocatalytic system for C–H hydroxylation.

1.6 Conclusion The selective hydroxylation of C–H bonds is a formidable challenge for synthetic organic chemists. Nature has provided much of the inspiration for the studies described in this chapter. Hydroxylation transformations in Nature are stereospecific in addition to being both chemo- and regioselective. This chapter outlined several strategies for pursuing C–H hydroxylation. Transition metal-catalyzed reactions attempt to mimic metalloenzymes such as the cytochrome P450s. Organocatalysts are inspired by the performance of flavin-containing monooxygenases. While many breakthroughs have been made in both classes of oxygenase-mimics, no general, catalytic system for C–H hydroxylation has been developed. The promising results of Brodsky notwithstanding, considerable work must be done to make this method generally applicable for organic synthesis.

33 Chapter 2

Chapter 2. Benzoxathiazines for Selective O-Atom Transfer

2.1 Introduction The invention of catalysts for the selective oxy-functionalization of unactivated C– H bonds remains a major challenge in synthetic chemical methods research. Catalytic reactions of this type could dramatically reshape the practice of modern chemical synthesis. The primary obstacle in the development of a catalytic C–H hydroxylation method lies not in the infamous inertness of aliphatic molecules, but rather in the difficulty in designing catalysts that can discriminate between particular C–H bonds in a complex molecule. A number of recent approaches, however, have shown progress towards this goal.79

2.1a Background As discussed in Chapter 1, oxaziridines represent a unique family of oxidants for C–H hydroxylation. These molecules react as stoichiometric reagents similarly to dioxiranes, which have been shown to promote selective C–H hydroxylation. The ability to functionalize the nitrogen of the oxaziridine offers a significant advantage over dioxiranes in the arena of catalyst development, where the ability to tune both the steric and electronic properties of the active oxidant is paramount. Additionally, the stability of oxaziridines is much greater than of dioxiranes; while greater stability implies a less reactive oxidant, this quality also renders oxaziridines easier to isolate and manipulate. Unlike dioxiranes, which can only be formed from ketones with select reagents (primarily peroxymonosulfate, Oxone), oxaziridines can be formed from imines with a variety of terminal oxidants (such as Oxone, peracids, perseleninic acids, etc). This feature imparts more flexibility in designing reaction conditions for a catalytic process. No such catalytic process involving dioxirane oxidants has been demonstrated for C–H hydroxylation; this fact may be due in part to the requirement to use Oxone, which decomposes the dioxirane faster than the rate of a typical reaction with a C–H bond (See Figure 1.22). For these reasons, oxaziridines were envisioned to be superior to dioxiranes as mediators for a catalytic C–H hydroxylation process. The pioneering work of Resnati and Davis provided motivation for pursuing catalytic oxaziridine-mediated C–H hydroxylation (Figure 2.1). Resnati showed that highly electron-deficient polyfluorinated oxaziridines are unusually reactive compared to

79 See references, Chapter 1.

34 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

other oxaziridines and can hydroxylate C–H bonds with a reaction scope similar to that of dioxiranes.80 Davis, in addition to exploring extensively the synthesis and reactivity of N- sulfonyl oxaziridines, demonstrated that oxaziridines can mediate catalytic sulfide oxidation, cycling between an imine and the corresponding oxaziridine with Oxone as the terminal oxidant.81 With these studies serving as a guide, developing a catalytic C–H hydroxylation reaction that would exploit a reactive oxaziridine intermediate appeared to be a feasible and worthwhile pursuit (Figure 2.1).

Resnati: Stoichiometric C–H Hydroxylation Mediated by Oxaziridines

O N OH n-C4H9 n-C3F7 F

CFCl3, 23 ºC 90%

Davis: Catalytic Sulfur Oxidation with N-Sulfonyloxaziridines

O O O 20 mol% oxaziridine S S S O N Ph Oxone H CH2Cl2, rt 90% yield O2N oxaziridine Brodsky and Du Bois: Oxaziridine-Mediated Catalytic C–H Hydroxylation

[O] O O 20 mol% 2.1 OH S R3 R3 1 mol% Ar Se O N N N 2 2 O CF R R R R 4 equiv UHP 3 1 2 1 2 DCE, 45-95h 80% Sox Sred Cl 2.1

Figure 2.1. Resnati and Davis established precedent for developing a catalytic oxaziridine-mediated C–H hydroxylation reaction.

Du Bois and Brodsky reported the first example of a catalytic oxaziridine-mediated olefin epoxidation and C–H hydroxylation reaction (Figure 2.1).82 The reaction employs a dual-catalytic system: an electron-deficient 1,2,3-benzoxathiazine-2,2-dioxide83 was designed to mediate oxygen transfer to a given hydrocarbon substrate; a diaryldiselenide

80 Sorochinsky, A. E.; Petrov, V. A.; Petrenko, A. A.; Soloshonok, V. A.; Resnati, G. “Regioselective Oxyfunctionalization of Bridgehead Adamantane Derivatives.” Tetrahedron 1997, 53, 5995-6000. 81 Davis, F. A. “Adventures in Sulfur−Nitrogen Chemistry.” J. Org. Chem. 2006, 71, 8993-9003. 82 Brodsky, B. H.; Du Bois. J. “Oxaziridine-mediated catalytic hydroxylation of unactivated 3º C-H bonds using hydrogen peroxide.” J Am Chem Soc 2005, 127, 15391-15393. 83 For simplicity, these heterocycles will be referred to as benzoxathiazines for the remainder of this document.

35 Chapter 2

was employed to mediate the transfer of oxygen from hydrogen peroxide to the benzoxathiazine heterocycle. Urea hydrogen peroxide (UHP) is a convenient, solid form of hydrogen peroxide and can oxidize a diaryldiselenide to a perseleninic acid. The combination of UHP, 1 mol% diaryldiselenide and 20 mol% benzoxathiazine 2.1 works well for epoxidation reactions and hydroxylation of simple, unfunctionalized substrates, as seen in Table 2.1. For instance, an electron-deficient trisubstituted olefin was epoxidized in nearly quantitative yield and adamantane was hydroxylated in 80% yield with 20 mol% catalyst 2.1. Although this system is much less effective with acyclic tertiary hydrocarbon substrates and the reaction duration is quite long, these early studies paved the way for the development of a general, synthetically useful C–H hydroxylation method. Importantly, this work established that catalytic turnover was possible and that C–H bonds could be hydroxylated using a heterocycle that was not polyfluorinated.

Table 2.1. Substrate scope for catalytic oxaziridine-mediated C–H hydroxylation.

Substrate Product time (h) temp ( ºC) Yield (%)

Me Me CO nBu 45 22 96 CO2nBu Me 2 Me O OH O O S 95 22 80 O N

CF3

Me Me OBz 96 50 37 OBz Cl Me OH Me 2.1 Me OBz Me OBz 73 50 13b Me Me OH

a Reactions conducted in DCE using 1 mol% diaryldiselenide, 20 mol% 2.1, and 4 equiv of UHP. b Conversion determined by 1H NMR of the unpurified reaction mixture.

Our goal at the outset of this project was to understand the factors that govern the success of this non-metal mediated hydroxylation reaction and to improve reaction performance with more recalcitrant substrates. We sought a more efficient method (higher turnover numbers, TON) that could be employed in a complex molecule setting. To achieve these goals, we chose to examine in detail the individual steps of the catalytic cycle and to identify the components that most significantly influence the reaction efficiency. A proposed catalytic cycle is illustrated in Figure 2.2. With several variables (ie the terminal oxidant, co-catalyst, etc.) playing a large role in the success of the catalytic process, there are many factors that can affect the reaction outcome; the

36 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

elements critical to catalysis will be discussed in detail in Chapter 3. Conversely, the work described in this chapter focuses solely on the structural characteristics of the catalyst that affect the rate of oxygen atom transfer to a substrate; as such, the stoichiometric reaction between an oxaziridine and substrate is discussed here.

10 mol% 2.1 1 mol% Ar2Se2 O OBz 4 equiv UHP OBz DCE, 36h 92% O O S O O N O F3C Se OH CF3 H2O2 OBz

CF3 Cl a b c

O O O H2O O S OBz F C Se OH O N 3 O CF3

CF3 2.1 Cl

Figure 2.2. A proposed catalytic cycle for oxaziridine-mediated C–H hydroxylation.

To begin our studies, we chose to focus first on step c, the transfer of oxygen from oxaziridine to substrate (Figure 2.2). Brodsky demonstrated that for this catalytic process, step c is rate-determining.84 Stoichiometric oxygen transfer to substrate is exceedingly slow, particularly with saturated aliphatic substrates. The longer the oxaziridine exists in solution, the more likely it will react in an unproductive manner, thus diminishing the overall performance of the reaction.85 Ideally, an improved catalyst would react rapidly with hydrocarbon substrates; thus, our attention focused on elucidating the factors that could increase the oxygen-transfer efficiency of benzoxathiazine-derived oxaziridines.

2.1b Project Goals and Methods In considering different oxaziridine designs, our attention focused exclusively on heterocyclic-based structures analogous to benzoxathiazine 2.1. For the purpose of

84 Brodsky, B. H. Ph. D. Thesis, Stanford University, 2007. 85 Denmark and coworkers observed such problems with the catalytic generation of dioxiranes (Figure 1.22); the catalytically generated dioxirane reacted with the terminal oxidant before it could react productively with the substrate.

37 Chapter 2

achieving our long-term goal of developing a catalytic reaction process, the stability of the “imine” catalyst was viewed as a critical design component. Acyclic imines are often highly susceptible to hydrolytic decomposition, and therefore were deemed inferior for our purposes. Consequently, cyclic structures were pursued over the more commonly documented acyclic imine-derived reagents, such as Davis’s stoichiometric oxidant (Figure 2.3). Hydration of an acyclic imine results in the formation of two molecules, an amine and an aldehyde. In the case of the cyclic imine, however, hydration should not necessarily result in a deleterious loss of catalytic function. Thus, in keeping with Brodsky’s original design, our initial pursuits focused on benzoxathiazines and related heterocyclic derivatives.

O O O O S S O N Ph N O O R H Ph

Davis Oxaziridine Du Bois Oxaziridine O O O O O O S S S H2O O O O O N H2O O NH Ph N S OH Ph – H2O NH2 H Ph R – H2O R H Ph

Figure 2.3. Comparison of the Davis oxaziridine with the Du Bois oxaziridine.

One of the principle reasons for choosing to study benzoxathiazine-derived catalysts stemmed in large part from our ability to make a number of structural modifications to the relatively simple heterocyclic core. We realized, however, that the ability to modify such catalysts at numerous positions would afford a prohibitively large number of structures to synthesize and evaluate. As such, we selected three general structural elements for investigation: 1) the nature of the heteroatoms within the heterocycle; 2) ring size; and 3) aromatic substituent groups (Figure 2.4).

Heterocycle: Ring-Size: Aromatic Substitution:

O X" O O O O O O O O O N X S O S N O S S X' N O S N O O N O N O O O O O

X X = S, P X = EWG, EDG X' = O, NR, CH2 X" = O, N

Figure 2.4. Structural elements to investigate.

38 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

For the purpose of evaluating the rate of oxygen transfer for the new oxaziridine structures, both theoretical and analytical methods were employed. Because we were interested in structural modifications that affected the rate of oxygen-atom transfer (step c, Figure 2.2), we turned initially to the study of stoichiometric alkene epoxidation with oxaziridines (Figure 2.5). As described in Chapter 1, oxaziridines react with olefins through a spiro-transition state that is in many ways analogous to the C–H hydroxylation event. Olefin epoxidation was used as a model reaction for kinetic studies due to the significantly faster rate of reaction as compared to C–H hydroxylation.86 The decision to explore stoichiometric alkene epoxidation also meant that the oxaziridines to be tested could be of a less reactive variety than those needed for C–H hydroxylation. Accordingly, oxaziridines derived from salicylaldehydes could be utilized instead of the more complex trifluoromethyl ketone derivatives previously employed in Brodsky’s original work (Figure 2.1, Table 2.1). With relatively simple structures to synthesize, we were able to synthesize a range of structurally varied oxaziridines (Figure 2.5), determine the observed rate constants for olefin epoxidation and use computational methods to guide the design of next-generation modified structures. This assessment format allowed us to quickly evaluate a large number of potential hydroxylation catalysts.

Computational A simplified Rate of stoichiometric methods were used to oxaziridine was olefin epoxidation was determine theoretical synthesized. determined. activation barriers for ethylene epoxidation.

O O O O O O S S S O N k O O N Me OBz obs O N Me OBz H O + + O H Me 25 ºC H Me H H H H 2.2 2.3 2.4 2.5

Figure 2.5. A strategy for evaluating potential catalysts: synthesis, kinetic analysis, and DFT calculations.

2.2 Oxaziridine Synthesis A variety of heterocycles were synthesized to probe the effect of heteroatom (ie S vs P vs C), ring size and aromatic substituents on the rate of stoichiometric oxygen transfer (Figures 2.4 and 2.5). The synthetic routes to these heterocycles are presented

86 Depending on substrate, olefin epoxidation with benxozathiazine-derived oxaziridines takes less than 6 hrs to go to completion while C–H hydroxylation will take > 72 hrs.

39 Chapter 2

here and the evaluation of these oxidants with kinetic studies and computational methods are presented in Section 2.4. Within the manifold of heterocyclic structures, there are several types that could be envisioned to support the formation of the corresponding oxaziridine. For instance, the heteroatom adjacent to the oxaziridine could be used to fine-tune the electronic properties of the scaffold. The size of the ring could also be adjusted to investigate the effect of ring strain on rate. We first chose to investigate the nature of the heterocycle—the heteroatom and the ring size—and then explored the introduction of substituents along the aromatic portion of the heterocycle.

2.2a Heterocycle synthesis The synthesis of two classes of heterocycles were pursued, N-sulfonyl and N- phosphinoyl oxaziridines. Sulfamoyl chloride, H2NSO2Cl, is a convenient reagent for synthesizing benxoxathiazines and related sulfonamide heterocycles (Figure 2.6).87 This reagent allowed for the efficient synthesis of a large range of N-sulfonyl imines in one step from salicylaldehyde starting materials. Buffered Oxone was then used to form the oxaziridine from the N-sulfonyl imine.88 Given the range of available salicylaldehydes, particularly with substituents at C6 and C8,89 this two-step sequence is quite useful.

O O O O OH O S 1 S 3 H2NSO2Cl O 2 N Oxone, NaHCO3 O N O H DMA, 0 ºC 8 4 H CH3CN, H2O H 7 5 6 2.2 2.4 Figure 2.6. A two-step process for the conversion of salicylaldehydes to benzoxathiazine-derived oxaziridines.

Unlike the benzoxathiazines whose syntheses are relatively straightforward, the synthesis of analogous benzoxaphosphorines proved challenging. In 1985, Boyd et al reported one of the first examples of an N-phosphinoyl oxaziridine synthesis.90 The oxidizing capabilities of these oxaziridines were shown to be similar to those of the N- sulfonyl oxaziridines reported by Davis; for instance cyclohexene is epoxidized in high

87 Okada, M.; Iwashita, S.; Koizumi, N. Tetrahedron Lett. 2000, 41, 7047-7051. Espino, C. G.; Wehn, P. M.; Chow, J; Du Bois, J. “Synthesis of 1,3-Difunctionalized Amine Derivatives Through Selective C–H Bond Oxidation.” J. Am. Chem. Soc. 2001, 123, 6935-6936. 88 Yang, D.; Wong, M. K.; Wang, X. C.; Tang, Y. C. “Regioselective intramolecular oxidation of unactivated C–H bonds by dioxiranes generated in situ.” J. Am. Chem. Soc. 1998, 120, 6611-6612. 89 For consistency, benzoxathiazine numbering is used throughout this document. 90 Boyd, D. R.; Jennings, W. B., McGuckin, R. M.; Rutherford, M.; Saket, B. M. “N-Phosphioyl oxaziridines – a new class of oxaziridines.” J Chem Soc Chem Comm 1985, 9, 582-583.

40 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

conversion with excess of a trifluoromethyl-substituted oxaziridine (Figure 2.7).91 A phosphorous heterocycle analogous to our N-sulfonyl oxaziridine, however, has never been reported.92 O P Ph N Ph 1.1 equiv O

Me CF3

O 90% conversion CDCl3, 35 ºC

Figure 2.7. An N-phosphinoyl oxaziridine is capable of effecting olefin epoxidation.

The syntheses of oxaziridines 2.6 and 2.7 were accomplished by adding a salicylaldehyde and triethylamine to a solution of POCl3 at 0 ºC (Figure 2.8). Subsequently triethylamine and diethylamine were added followed by aqueous ammonia. The phosphinoyl imine was then oxidized to the oxaziridine with an aqueous buffered Oxone solution. Interestingly, only one diastereomer of the product was isolated and a crystal structure was obtained that confirmed that the peroxymonosulfate had added trans to the diethylamine moiety (Figure 2.9). Comparison of the ground state energies for each diastereomer (B3LYP/6-31G*) indicates that the isomeric oxaziridine is less stable by 4.6 kcal/mol. Given the unconstrained nature of the diethylamine group, it is not immediately evident why the cis and trans oxaziridines should have such different ground state energies. Comparison of the geometrical parameters of 2.7 with known heterocyclic oxaziridines 2.893 and 2.94 shows that 2.7 displays an elongated N–O bond as compared to the other oxaziridines. It would be expected that the weaker N–O bond seen in this compound would correspond to a more potent oxygen transfer reagent.

91 Jennings, W. B.; Schweppe, A.; Testa, L. M.; Zallabos-Garcia, E.; Sepulveda-Arques, J. “Applications of N-phosphinoyloxaziridines in the conversion of alkenes to epoxides and esters to alpha- hydroxyesters.” Synlett 2003, 121-123. 92 A similar structure has been proposed and synthesized: Vandersteen, J.; Westra, J. G.; Benckhuysen, C.; Schulten, H. R. “A new oxaziridine derivative of cyclophosphamide obtaine from ozonolysis of O-3- butenyl, N,N-bis(2-chloroethyl)phosphorodiamidate.” J. Am. Chem. Soc. 1980, 102, 5691-5692. 93 Davis, F. A.; Reddy, R. T.; McCauley, J. P.; Przeslawski, R. M.; Harakal, M. E.; Carroll, P. J. “Chemistry of Oxaziridines. 15. Asymmetric Oxidations Using 3-substituted 1,2-Benzisothiazole 1,1- Dioxide Oxides.” J. Org. Chem. 1991, 56, 809-815.

41 Chapter 2

1) "salicylaldehyde", NEt , O NEt2 Oxone® O NEt2 3 P P DCM, 0 ºC O N NaHCO3 O N POCl O R = H 2.6 3 = Br 2.7 2) NEt2H, NEt3, DCM, 0 ºC H CH3CN/H2O H 3) NH4OH

R R

Figure 2.8. Synthesis of phosphorous oxaziridines 2.6 and 2.7.

Figure 2.9. Crystal structure of N-phosphinoyl oxaziridine 2.7.

2.2b Synthesis of expanded-ring benzoxathiazine-derived oxaziridines Although we knew that a cyclic structure would be preferable over an acyclic structure, it was unclear to us how the rigidity and the inherent ring strain of the various heterocycle rings would affect the rate of oxygen transfer. Furthermore, each ring size has potential advanages. For instance, the five- and six-membered rings might be obvious choices to pursue because their syntheses have been previously reported; however, the seven- and eight-membered rings, by the nature of their larger ring size, have the advantage of offering additional sites for modification. Because we wanted to

42 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

determine how the heterocycle ring size affected the rate of oxygen-atom transfer, we investigated benzoxathiazine-, benzoxathiazepine- and benzoxathiazecine-derived oxaziridines for oxygen-atom transfer. Davis has previously reported the synthesis of a five-membered ring heterocycle and the synthesis of the six-membered ring heterocycle is illustrated in Figure 2.6. The syntheses of both the benzoxathiazepine and benzoxathiazecine heterocycles were envisioned to proceed by sulfamoylation of a phenolic aldehyde substrate followed by oxidation with buffered Oxone. For the seven-membered ring, the necessary substrate was easily accessed by formylating napthalen-1-ol (Figure 2.10).94

O O O O O H OH OH S N S N O 1. tBuLi H2NSO2Cl O Oxone, NaHCO3 O

2. DMF DMA, 0 ºC CH3CN, H2O THP

2.10 2.11

Figure 2.10. Synthesis of seven-membered ring N-sulfonyl oxaziridine.

Synthesis of the analogous eight-membered ring heterocycle proved to be much more challenging (Figure 2.11). Efforts to synthesize phenolic aldehyde 2.13 were thwarted by spontaneous cyclization to for the hemi-acetal structure 2.14. Although synthetic routes that do not employ 2.13 were considered and briefly explored, the synthesis of this heterocycle was not achieved.16

O O OH O H S N O H2NSO2Cl O DMA, 0 ºC

OH 2.14 2.13 2.12

Figure 2.11. Attempted synthesis of a benzoxathiazecine heterocycle.

2.2c Incorporation of aromatic substituents For many of the aromatic substituents we hoped to investigate, the corresponding salicylaldehyde was commercially available; therefore the oxaziridine could be accessed by the sulfamoylation/oxidation protocol shown in Figure 2.6. On the occasion that the requisite salicylaldehyde was not commercially available, we often sought the most simple and direct transformations at the expense of yield.

94 This work was performed by Stanford undergraduate, Elisabeth Hennessey.

43 Chapter 2

For arene-halogenated heterocycles, commercial salicylaldehydes are generally only available with halogenation at the C6 and C8 positions. To incorporate halogenation at the C5 and C7 positions, the Riemer-Tiemann formylation reaction was employed with commercially available halogenated phenols (Figure 2.12).95 Generally this is a very poor reaction, giving regiochemical mixtures and low product yields. Because we sought to access a range of substituted salicylaldehydes quickly, the poor regioselectivity became an asset and often two desired salicylaldehydes could be isolated from one reaction.96 Other substituents, such as methyl and trifluoromethyl groups, could be incorporated into the heterocycle with this protocol as well. Salicylaldehyde-derivatives were then subjected to the sulfamoylation/oxidation protocol illustrated in Figure 2.6 (Table 2.2). A crystal structure was obtained for 5-bromo oxaziridine 2.30 (Figure 2.13). Surprisingly, despite the proximity of the bromine to the C4 hydrogen, no distortion of the oxaziridine bond lengths or bond angles was observed.

OH CHCl3 OH O O OH NaOH H + H 60 ºC Cl Cl Cl 1 : 1

Figure 2.12. The Riemer-Tiemann formylation reaction was employed to access salicylaldehydes from substituted phenols.

95 Hodgson, H. H.; Jenkinson, T. A. “The Reimer-Tiemann reaction with m-chlorophenol.” J Chem Soc 1927, 1740-1742. 96 When necessary, improved regioselectivity could be achieved employing an alternative protocol paraformaldehyde and MgCl2. See: Hansen and Skattebol. One-pot synthesis of substituted catechols from the corresponding phenols. Tetrahedron Letters 2005, 46, 3357-3358.

44 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

Table 2.2. Benzoxathiazines and benzoxathiazine-derived oxaziridines synthesized from substituted salicylaldehydes.

benzoxathiazine oxaziridine C5 C6 C7 C8

2.15 2.16 Cl H H H O O S 2.17 2.18 H Cl H H O N 2.19 2.20 H H Cl H 8 H 2.21 2.22 H Cl Cl H 7 5 2.23 2.24 Cl Cl H H 6 2.25 2.26 H Cl H Cl benzoxathiazine 2.27 2.28 Cl Cl H Cl O O 2.29 2.30 Br H H H S O N 2.31 2.32 H Br H H O CF H H 2.33 2.34 3 H H 2.35 2.36 H CF3 H H 2.37 2.38 F H H H oxaziridine CH H 2.39 2.40 3 H H

Figure 2.13. Crystal structure of 2.30.

With access to the brominated salicylaldehydes, simple cross-coupling reactions could be employed to introduce additional substituent groups. For instance, 6- bromosalicylaldehyde was used as a cross-coupling partner in various Suzuki reactions

45 Chapter 2

(Figure 2.14). This general procedure, along with methods to brominate other sites on the aromatic ring, allowed for rapid access to a large number of benzoxathiazine derivatives.

O O OMe O 3 equiv aryl boronic acid OMe O S Ar = Ph 2.41 O N 5 mol% Pd(PPh3)4 O pMePh 2.42 H H H oBrPh 2.43 2 equiv K2CO3 DMF, 80 ºC pNO2Ph 2.44 Br Ar 2,6-MePh 2.45 Ar 3,5-CF3Ph 2.46 2.X Figure 2.14. Cross-coupling route to substituted benzoxathiazine-derived oxaziridines.

2.3 Overview of Computational Studies and Kinetics Analysis of Oxaziridines for Stoichiometric Epoxidation

2.3a DFT Calculations Density functional theory (DFT) calculations are an excellent tool for predicting and visualizing transition states in organic transformations. As discussed in Chapter 1, many research groups, including ours, have employed these calculations to model transition states for dioxirane and oxaziridine oxidations of olefins and C–H bonds.97 Comparing calculated transition state activation energies for a range of oxaziridines helped us rationalize our experimental data as well as guide us to new oxaziridine designs. Because we were studying the rate of oxygen-atom transfer to an olefin, transition states for ethylene epoxidation were typically sought. DFT calculations were conducted using Gaussian ’03 with the B3LYP level of theory and the 6-31G* basis set. Previous theoretical studies have validated the use of the B3LYP functional and the 6-31G* basis set for calculations involving organic oxidants.19,98 Most calculations are gas-phase

97 (a) Du, X. H.; Houk, K. N. “Transition states for alkane oxidations by dioxiranes.” J. Org. Chem. 1998, 63, 6480-6483. (b) Houk, K. N.; Liu, J.; Demello, N. C.; Condroski, K. R. “Transition States of Epoxidations: Diradical Character, Spiro Geometries, Transition State Flexibility, and the Origins of Stereoselectivity.” J Am Chem Soc 1997, 119, 10147-10152. (c) Hirschi, J. S.; Takeya, T.; Hang, C.; Singleton, D. A. “Transition-State Geometry Measurements from 13C Isotope Effects. The Experimental Transition State for the Epoxidation of Alkenes with Oxaziridines.” J Am Chem Soc 2009, 131, 2397- 2403. (d) Bach, R. D.; Andres, J. L.; Su, M. D.; McDouall, J. J. W. “Theoretical model for electrophilic oxygen atom insertion into hydrocarbos.” J Am Chem Soc 1993, 115, 5768-5775. 98 Bach, R. D.; Gluskhovtsev, M. N.; Gonzalez, C.; Marquez, M., Estevez, C. M.; Baboul, A. G.; Schlegel, H. B. “Nature of the Transition Structure for Alkene Epoxidation by Peroxyformic Acid, Dioxirane, and Dimethyldioxirane” A Comparison of B3LYP Density Functional Theory with Higher Computational Levels.” J. Phys. Chem. A 1997, 101, 6092-6100.

46 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

unless otherwise noted. Transition states were identified by a stationary point with a single imaginary frequency and intrinsic reaction coordinate (IRC) analyses confirmed that the structure identified corresponded to a first-order saddle point consistent with a transition structure. Our calculations indicate that N-sulfonyl oxaziridines behave similarly to the oxaziridines and dioxiranes studied by Houk with a spiro, slightly asynchronous transition structure (See Chapter 1).99 If DFT calculations were to guide our studies, we needed to establish the accuracy of the chosen level of theory and basis set. To do so, we compared the geometrical parameters obtained from the crystal structure of catalyst 2.2 with the bond lengths and angles obtained from DFT calculations (Table 2.3). A plot of the experimental vs calculated values gives a line with an R2 value >0.999 and a slope of 1.02.100 From this analysis, it is clear that the predicted geometries are generally in accord with the metrical parameters of the crystal structure.

Table 2.3. Comparison of theoretical and experimental structural parameters for oxaziridine 2.30. (DFT calculations with B3LYP/6-31G*)

Bond Lengths (Å) Calc Exper Bond Angles (º) Calc Exper

N1-O4 1.468 1.489 N1-O4-C7 60.42 59.54

N1-C7 1.454 1.449 N1-C7-O4 61.39 62.32

O4-C7 1.421 1.428 S1-N1-C7 114.66 114.61

N1-S1 1.755 1.700 O1-S1-N1 102.94 104.84

O1-S1 1.669 1.584

Another useful metric for gauging the quality of our computed data is to compare IR spectra obtained from experiment with DFT-predicted vibrational data. Figure 2.15 shows an overlay of the predicted spectrum of 2.2 with the obtained spectrum. The predicted peaks match well with those recorded in the IR measurement. Due to the high level of similarity between the predicted and experimentally obtained crystal structure data and IR data as well as experimentally determined activation parameters (vide infra), we had confidence that DFT calculated transition state energies would be useful for guiding our studies.

99 See ref 19 (b). 100 For an example of a similar analysis of calculated data, see: Stephens, P. J.; Pan, J. J.; Devlins, F. J.; Krohn, K.; Kurtan, T. “Determination of the absolute configurations of natural products via density functional theory calculations of vibrational circular dichroism, electronic circular dichroism, and optical rotation: The iridoids plumericin and isoplumericin.” J Org Chem 2007, 72, 3521-3536.

47 Chapter 2

!" !"

!" !" !" !"

Figure 2.15. Predicted (left) and experimentally obtained (right) IR spectra of 2.2.

2.3b Kinetics method To gain a sense of how changes in the oxaziridine structure affect the rate of oxygen-atom transfer, we sought a kinetics method to determine the rate constant for oxaziridine-mediated epoxidation (Figure 2.16). An HPLC method was created and pseudo-first order kinetic studies were conducted using a 10-fold excess of olefin substrate 2.3 in the presence of an internal standard, N,N-dimethyltoluenesulfonamide. To monitor the reaction rate, vials of a triphenylphosphine solution were prepared and, at predetermined time points, aliquots of the reaction mixture were extracted by auto-pipet and added to these vials. Any remaining oxaziridine was immediately reduced and the ratio of olefin to epoxide could be determined by HPLC analysis of these samples. Although HPLC analysis was typically performed almost immediately after sample acquisition, these samples could sit for several days without any observable deterioration of the materials.

48 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

Ratio of Product : I.S. vs Time (min) 0.7

0.6

0.5

0.4

0.3 matlab fit y=A-Ae-kt 0.2 k=5.332(30)x10-4 A=5.653(11)x10-1 0.1

0 0 50 100 150 200 250 300

O O O O S S O O N Me OBz Me2NTs (I.S.) O N Me OBz O + + H Me DCE, 25 ºC H Me 2.3 2.5

Figure 2.16. Raw kinetics data obtained for oxaziridine 2.18 following a pseudo-first order kinetics scheme.

Raw data from our kinetic studies was analyzed using non-linear regression with Matlab. The following relationship derived from the Arrhenius equation was used to obtain kobs under our pseudo first- order conditions (Eq. 1).

-kt A(t) = A(t∞) – A(t∞)e (1)

-4 -1 101 For the parent oxaziridine structure 2.2, a kavg of 2.6 x 10 s was obtained. This number will be used in the following discussion as a basis for comparison and this structure will be referred to as the “parent structure.” A kobs was obtained for almost every oxaziridine synthesized. This comprehensive set of data allowed us to make comparisons and formulate hypotheses about trends in oxaziridine structure and reactivity. Using our kinetic method, Eyring analysis was conducted to compare experimental activation barriers with those calculated (Figure 2.17).102 An Eyring plot of the data (ln(kobs) vs. 1/T (K)) gave an Arrhenius activation energy of Ea = 10.9 kcal/mol. DFT calculations overestimate the activation energy for the reaction of oxaziridine 2.2 and olefin 2.3 by approximately 2.2 kcal/mol. Based on this result, density functional theory

101 The experimental rate constants were assessed two to three times and averaged to give kavg. 102 Eyring analysis was conducted by Dr. Benjamin Brodsky

49 Chapter 2

appears to provide reasonable estimates of experimental activation barriers for olefin epoxidation with benzoxathiazine-derived oxaziridines.

O O O O S S O N Me OBz DCE O N O O Me OBz + + Ea = 10.9 kcal/mol -35 – 5 ºC CF3 Me CF3 Me

2.47 2.3 2.48 2.5

Figure 2.17. Eyring analysis.

2.4 Evaluation of benzoxathiazine-derived oxaziridines as stoichiometric oxidants With synthetic routes to a wide range of oxaziridines and experimental and theoretical methods for assessing the rate of oxygen-atom transfer from these oxaziridines, we set out to determine the most effective heterocycle for oxygen transfer as well as the structural factors most relevant to creating a highly reactive oxygen-transfer reagent.

2.4a Heteroatom and the Heterocycle While it is known that electron-withdrawn oxaziridines are more reactive as oxygen-atom transfer agents, it is unclear how the nature of the heterocycle will influence this. We therefore sought to explore and compare a variety of known and unknown oxaziridine-containing heterocycles (Figure 2.18). For instance, both sulfur- and phosphorous-substituted oxaziridines are capable of O-atom transfer but the analogous carbamates have not been explored in this context. Furthermore, there are no reports of sulfamide-derived oxaziridines performing oxygen transfer reactions although oxygen- transfer reactions using both cyclic and acyclic N-sulfonyl oxaziridines are well documented by Davis and others.

O O O O O O O O NEt2 S S Me S P O N N N N O N O N O O O O O

2.2 2.49 2.50 2.51 2.6

Figure 2.18. Oxaziridine reagents related to 2.2.

50 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

Heterocycles depicted in Figure 2.18 were initially analyzed by DFT calculations. A comparison of oxaziridines 2.2 and 2.49 is illustrated in Figure 2.19.103 The presence of the phenolic oxygen in 2.2 greatly lowers the calculated solution-phase activation barrier for oxygen transfer as compared to 2.49 (B3LYP/6-31G*/CPCM, DCE). Because 2.2 and

2.49 are nearly isosteric, the large difference in Ea is attributed to the inductive power of the phenolic oxygen atom. This result guided us to explore more benzoxathiazine catalysts that feature an electron-withdrawing group para to the phenolic oxygen in order to further lower the activation barrier.

X = CH2 ‡ X = O O O H OAc S X N O H Me Me

X = O 2.2

X = CH2 2.49

2.4 !Eact(TS ) = 16.0 kcal/mol 2.49 !Eact(TS ) = 19.4

Figure 2.19. DFT calculations suggest that the phenolic oxygen plays a large role delocalizing negative charge in the transition structure.

Theoretical evaluation of heterocycles 2.50 and 2.51 against the benzoxathiazine- derived oxaziridine 2.2 suggested that neither of these two would effect O-atom transfer with an activation barrier comparable to that of the parent structure. Moreover, in considering the difficulties associated with the synthesis of these structures as compared to 2.2, we turned our attention towards other heterocyclic systems. We were particularly interested in a phosphorous-based analogue because the possibility it would afford to tune the ancillary ligand. Furthermore, the stereogenicity of the phosphorous center could perhaps be exploited in future designs for asymmetric O- atom transfer. As discussed previously, the crystal structure of 2.6 displays an elongated N–O bond as compared to the other oxaziridines, suggesting that it may indeed be a more potent oxygen transfer reagent. When phosphorous heterocycle 2.6 was monitored with pseudo-first order kinetics, the observe rate constant was found to be two orders of magnitude slower than the benzoxathiazine parent (Table 2.5). We hypothesized that the electron-donating

103 These calculations were carried out by Dr. Benjamin Brodsky.

51 Chapter 2

nature of the diethyl amine in 2.6 was contributing to the slow reaction rate. Therefore, we chose to evaluate the influence of different substituent groups on the calculated activation barriers for this reaction. Not surprisingly, replacing the diethyl amine with a trifluoromethyl group lowered the activation energy by 5.1 kcal/mol in relation to 2.6 (2.53, Table 2.5). Unfortunately, synthesis of this molecule proved to be quite difficult and there are no known routes to similar molecules reported in the literature. We thus chose to identify a phosphorous heterocycle with a nitrogen-containing ligand that could be prepared following the same general synthetic route as used for 2.6 (Figure 2.8). Calculations showed that morpholine analogue 2.54 had a lower activation energy than the diethylamine analogue 2.6 by 2.1 kcal/mol, and lower than the parent sulfur heterocycle 2.2 by ~1 kcal/mol. DFT calculations suggest that 2.54 should perform as well or better than our parent structure; such findings would provide validation of our computational analysis and set precedence for employing N-phosphoryl oxaziridines for C–H hydroxylation reactions.

Table 2.4. Comparison of activation barriers for ethylene epoxidation.

O O O O S S k O O N Me OBz obs O N Me OBz O + + H Me 25 ºC H Me

2.3 2.5

O O O CF O N O O NEt2 NEt2 3 S P P P P O N O N O N O N O N O O O O O H H H H

2.2 2.6 2.52 2.53 2.54

2 3 krel 1 0.01 nd nd nd

1 Ea 24.2 25.2 27.2 20.1 23.1

1Activation energy calculated for ethylene epoxidation using Gaussian. 2 This diastereomer was not obtained. 3 This oxaziridine could not successfully prepared.

2.4b Heterocycle ring size We sought to determine the effect of heterocyclic ring size on the rate of oxygen- atom transfer based on the assumption that different ring-strain energies would affect this process. In a related study, Davis and coworkers compared the rate of oxygen transfer

52 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

for acyclic oxaziridine 2.55 with the five-membered ring analogue 2.56 (Table 2.6). These results, however, were inconsistent and varied as a function of the substrate tested.104 With sulfide oxidation, the five-membered ring heterocycle 2.56 was twice as fast as the acyclic version 2.55. Olefin epoxidation, however, gave opposite results; the acyclic oxazirdine 2.55 was three times faster at epoxidizing methylcyclohexene than the cyclic analogue 2.56. The authors ascribe these disparate data to the different degrees of steric interaction between substrate and oxidant; this important insight suggests that the rate of oxygen transfer is a function of many variables, not just the electrophilicity of the oxaziridine.

Table 2.5. Comparison of performance of Davis’s cyclic and acyclic oxaziridines.

reaction relative rate

2.55 2.56 O oxaziridine O O O O O O S S 1.00 1.98 S S Ph N O N CDCl3, 25 ºC Me Me Me Me O H Ph Me Me oxaziridine Me O 1.00 0.33 2.55 2.56 CDCl3, 25 ºC

To expand upon and perhaps clarify the results of Davis, we compared the rates of stoichiometric epoxidation between oxaziridines 2.55, 2.56, 2.57, 2.2, 2.11, and 2.58. To our surprise, calculated activation barriers suggested that the six-, seven-, and eight- membered ring heterocycles should react with comparable rates and considerably faster than the acyclic and five-membered ring analogues (structures 2.55-2.57) (Table 2.7). These predicted activation barriers correspond well with the observed kinetics results for alkene 2.3 epoxidation; the six- and seven-membered ring heterocycles performed similarly and both displayed a rate of oxygen transfer two orders of magnitude faster than acyclic oxaziridine 2.55.

104 Davis, F. A.; Billmers, J. M.; Gosciniak, D. J.; Towson, J. C.; Bach, R. D. “Chemistry of oxaziridines .7. Kinetics and mechanism of the oxidation of sulfoxides and alkenes by 2-sulfonyloxaziridines – relationship to the oxygen-transfer reactions of metal peroxides.” J. Org. Chem. 1986, 51, 4240-4245.

53 Chapter 2

Table 2.6. Comparison of observed rate constants and calculated activation barriers for a range of heterocycle ring sizes.

O O O O S S k O R N Me OBz obs R N Me OBz O + + R H Me 25 ºC R H Me 2.3 2.5

O O O O O O O O O O O S S N O S N O S O O N S N S N O O Ph N O O O O

Ph H Me H

structure 2.55 2.56 2.57 2.2 2.11 2.58

a b krel <.04 ~0.01 -- 1 1 --

1 Ea 26.2 29.1 26.7 24.2 23.3 24.8

a Value estimated based on known performance with other tri-substituted olefins. b The synthesis of 2.58 was not successfully accomplished.

It seems that regardless of the size of the ring, heterocycles with the same heteroatom connectivity (2.2, 2.11, 2.58) perform similarly, despite the high variability in the rigidity of these heterocycles. We hypothesized that the higher activation barriers calculated for 2.55-2.57 could be accounted for by the lack of a phenolic oxygen; we have previously shown that this oxygen plays a significant role in stabilizing charge build-up during the transition state (Figure 2.19). Further, the steric impediment introduced by the methyl group on structure 2.56 seems to add 2.4 kcal/mol to the activation barrier. Ultimately, the six-membered ring benzoxathiazine heterocycles were chosen over the larger ring heterocycles due to their convenient synthesis from commercially-available salicylaldehydes (Figure 2.6)

2.4c Halogenation of aromatic ring A decided advantage of the benzoxathiazine scaffold is that electronic manipulation of the aromatic ring is relatively straightforward and it is expected that such modifications should greatly influence oxaziridine reactivity. Davis conducted Hammett analysis for olefin and sulfide oxidation with a variety of substituted analogues of 2.55. By varying substituents on both phenyl groups (ArXSO2N-O-CHArY), ρ values were obtained for both sites (ρ(ArX) = 0.51 – 1.06; (ρ(ArY) = 0.91 – 1.05). The small positive ρ values led Davis to hypothesize that in the transition structure for olefin epoxidation, a small amount of negative charge is distributed onto the carbon and the nitrogen of the oxaziridine.

54 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

Accordingly, electron-withdrawing groups positioned to stabilize this polarization of electron density should have a beneficial effect on the activation energy.26 Kinetics analysis of chloride-substituted oxaziridines 2.16-2.28 (Table 2.8) revealed that the kobs for C7 = Cl and C6 = Cl were approximately the same and two times that of parent oxaziridine 2.2, which indicates that substituents at both positions contribute to charge delocalization in the transition state. Furthermore, these effects seem to be somewhat additive: an oxaziridine 2.22 gives a rate constant that is approximately the sum of both 2.18 and 2.20. Chlorination at C5 and C8 offer a more substantial increase in the rate that at C6 and C7. The rate increase resulting from substitution at C5 could be due to steric distortion of the neighboring oxaziridine; this topic is discussed in more detail in Section 2.4e. Overall, however, these chlorination studies indicate that the most electron deficient oxaziridines give the fastest rate of oxygen transfer. Furthermore, DFT calculations for measuring the effect of chlorination on the previously discussed ring- expanded heterocycles (2.56-2.60) show a similar, albeit small, effect on the rate constant (Table 2.8); the small benefit chlorination offers these heterocycles did not warrant their syntheses.

Table 2.7. Comparison of rate of oxygen transfer for a variety of chlorinated oxaziridines.

O O O O S S k O O N Me OBz obs O N Me OBz O + + H Me 25 ºC H Me

X X 2.3 2.5

O O O O O O O O O O O O S S S S S O N O N S O N O N O N O O O N O O O O Cl

Cl Cl Cl Cl Cl Cl Cl 2.2 2.20 2.18 2.22 2.16 2.28

krel 1 1.9 2.1 3.6 14.0 28.3

1 Ea 24.2 23.6 23.4 23.0 22.6 21.4

1Activation energy calculated for ethylene epoxidation using Gaussian (B3LYP/6-31G*).

55 Chapter 2

Table 2.8. Comparison of halogenated heterocycles of different ring sizes.

O O O O O O O S O O O O S N O S N S N O N O O S O O Ph N O Me ClpPh H Cl Cl Cl Cl

56 57 58 59 60

1 27.3 28.5 23.4 Ea 22.6 24.0

1Activation energy calculated for ethylene epoxidation using Gaussian (B3LYP/6-31G*).

2.4d Hammett analysis Given that halogenation at C7 and C6 gave similar results, Hammett analysis was pursued with modified C6 analogues (Table 2.9, Figure 2.20). A small, positive ρ value (0.83) indicating a build-up of negative charge in the transition state was measured, confirming the proposed electrophilic oxidation mechanism and in accord with the small, 26 positive ρ value observed by Davis. Interestingly, fitting the correlation data to σp gave a slightly better fit than fitting to σm, further suggesting that the phenolic oxygen plays a more substantial role in charge stabilization.

Table 2.9. Hammett analysis.

O O O O S S k O O N Me OBz obs O N Me OBz O + + H Me 25 ºC H Me

R R -1 Entry R Group !p kobs (s ) krel 1 OMe -0.27 2.2 x 10-4 0.85 2 H -0.01 2.6 x 10-4 1.0 3 Ph 0.00 2.8 x 10-4 1.1 4 Cl 0.23 5.4 x 10-4 2.1 -4 5 C6F5 0.27 6.1 x 10 2.3 -4 6 CF3 0.54 7.4 x 10 2.8 -3 7 NO2 0.78 1.6 x 10 6.0

56 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

Hammett Analysis 0.9

0.6

0.3 kobs 0 y = 0.83x + 0.08 R2 = 0.95

-0.3 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 σ value

Figure 2.20. Hammett plot.

2.4e C5 modification We were surprised to find that halogenation at C5 had such a dramatic effect on the rate of oxygen transfer (section 2.4c). Chlorination at C5 gave a kobs that was notably larger than the C7 and C6 chlorinated analogues. As halogenation at C5 should impart approximately the same inductive stabilization of the phenol as at C7, we rationalized that this rate enhancement could be due to the release of steric strain upon oxygen transfer. Consequently, we synthesized a variety of oxaziridines that were modified at the C5 position (Table 2.11). A methyl group at C5 gives a slower kobs than the parent, suggesting that even if steric factors were contributing to the rate enhancement, the electronic stabilization imparted by the substituents is a more significant contributor.

Table 2.10. Effect of C5 modification of the rate of expoxidation.

O O O O S S k O O N Me OBz obs O N Me OBz O + + H Me 25 ºC H Me

R R

a Calculated Ea (kcal/mol) -1 Entry R Group !p/!m k (s ) krel ethylene epoxidation obs 1 Me -0.17/-0.07 24.5 1.6 x 10-4 0.6 2 Ph -0.01/0.06 24.1 1.6 x 10-4 0.6 3 H 0/0 24.2 2.6 x 10-4 1.0 4 F 0.06/0.34 23.1 9.5 x 10-4 3.7 5 Cl 0.23/0.37 23.4 1.4 x 10-3 5.4 6 Br 0.23/0.39 21.5 1.3 x 10-3 5.0

57 Chapter 2

The crystal structure of 2.30 sheds some light on the uniquely fast rate of oxygen transfer of the C5 substituted benzoxathiazines (Table 2.8). The dihedral angle for C5- C6-C7-H105 is much smaller than in the other crystal structures; in other words, the C–N bond of the oxaziridine lies closer to the same plane as the phenyl group than with the other oxaziridines. Upon oxygen transfer, the resulting imine needs fall into planarity with the phenyl group to benefit from electron delocalization through the conjugated π system. The presence of the C5 substituent could somehow be enforcing a more planar oxaziridine, perhaps through hydrogen bonding, thus reducing the kinetic barrier for oxygen transfer. This hypothesis corresponds with the unusually fast rate of epoxidation observed for oxaziridines halogenated or oxygenated at C5.

2.4 Conclusions The design elements for benzoxathiazine-derived oxaziridines have been thoroughly explored. The following observations regarding the reactivity of the oxaziridines for oxygen-atom transfer are noted: 1) N-sulfonyl oxaziridines are the best heterocycles tested for stoichiometric oxygen transfer; N-phosphoryl oxaziridines, however, could prove useful in this context if a suitable phosphorous ligand could be identified and appended. 2) Ring size of the heterocyclic “imine” from which the oxaziridine is derived appears to have minimal impact on the absolute rate of oxygen-atom transfer. 3) As expected, the most electron-deficient oxaziridines are the most reactive for oxygen transfer. A positive ρ value in the Hammett plot confirms the electrophilic nature of the transformation. 4) The position of the electron-withdrawing group on the aromatic ring affects the rate of oxygen transfer. Substituents at both C7 and C6 seem to impart the similar degree of transition state stabilization. This result is in accord with charge build-up occurring at both the carbon and nitrogen on the oxaziridine during oxygen transfer. Adding substitutents to C5 and C8, however, increases the rate of stoichiometric oxygen transfer even more than the C7 and C6 positions. This effect cannot be explained by electronic stabilization alone. We hoped that the knowledge gained by exploring and understanding the factors that contribute to a more reactive oxaziridine O-atom transfer reagent could be applied to the design of a next-generation benzoxathiazine catalyst for selective C–H bond

105 Numbering assigned by the crystallographer.

58 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

oxidation. Our efforts towards developing a catalytic process for C–H hydroxylation are discussed in Chapter 3.

59 Chapter 2

Experimental Section General. All reagents were obtained commercially unless otherwise noted. Moisture-sensitive reactions were performed using flame-dried glassware under an atmosphere of dry nitrogen. Air- and moisture-sensitive liquids and solutions were transferred via syringe or stainless steel cannula. Organic solutions were concentrated under reduced pressure (ca. 15 Torr) by rotary evaporation. Dichloroethane (DCE) was distilled from CaH2 immediately prior to use. Dichloromethane, tetrahydrofuran (THF), acetonitrile, and N,N–Dimethylformamide (DMF) were dried by passage under 12 psi N2 through columns containing activated alumina. N,N–Dimethyl-acetamide (DMA) was dried over activated 4Å molecular sieves. Chromatographic purification of products was accomplished using forced-flow chromatography on Silicycle Ultra Pure Silica Gel Silia-P (40-63 µm). Compounds purified by chromatography on silica gel were typically applied to the adsorbent bed using the indicated solvent conditions with a minimum amount of added chloroform as needed for solubility. Thin layer chromatography was performed on

EM Science silica gel 60 F254 plates (250 µm). Visualization of the developed chromatogram was accomplished by fluorescence quenching and by staining with ethanolic anisaldehyde, aqueous potassium permanganate, or aqueous ceric ammonium molybdate (CAM) solution. HPLC analyses were conducted using a Hewlett Packard Series 1100 Chemstation. Samples were analyzed using a Zorbax CN normal phase analytical column (4 x 250 mm) with hexanes/i-PrOH (0.5% to 50% gradient elution) as eluent. Optical absorption was determined at 254 nm. NMR spectra were acquired on a Varian Mercury-400 operating at 400 and 100 MHz or a Varian Inova-500 operating at 500 and 125 MHz for 1H and 13C, respectively, and are referenced internally according to residual solvent signals. Data for 1H NMR are recorded as follows: chemical shift (δ, ppm), multiplicity (s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet), integration, coupling constant (Hz). Data for 13C are reported in terms of chemical shift (δ, ppm). Infrared spectra were recorded as thin films using NaCl salt plates on a Perkin-Elmer Paragon 500 FTIR spectrometer or a Thermo-Nicolet IR300 spectrometer and are reported in frequency of absorption. High-resolution mass spectra were obtained the Vincent Coates Foundation Mass Spectrometry Laboratory, Stanford University.

60 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O NEt2 P O N

2.6red Triethylamine (0.77 mL, 5.5 mmol, 1.1 equiv) was added to a solution of salicylaldehyde (0.51 mL, 4.8 mmol, 0.96 equiv) in 6.0 mL of CH2Cl2. This solution was added dropwise over 10 minutes to a precooled (0 ºC) solution of POCl3 (distilled from

CaH, 0.46 mL, 5 mmol, 1 equiv) in 9.0 mL of CH2Cl2. The resulting opaque, pale yellow solution was allowed to stir at 0 ºC for ~1h. Diethylamine (0.52 mL, 5 mmol, 1 equiv) was then added to the ice-cold solution immediately followed by triethylamine (0.77 mL, 5.5 mmol, 1.1 equiv). The reaction mixture was allowed to warm to 22 ºC over 3 hr. 25%

NH4OH in water (2.0 mL, 12.5 mmol, 2.5 equiv) was added to the reaction mixture at 0 ºC and the reaction was subsequently diluted with 30 mL of H2O and 30 mL CH2Cl2. The mixture was transferred to a separatory funnel, the organic layer was collected, and the aqueous layer was extracted with 3 x 20 mL CH2Cl2. The combined organic extracts were dried over anhydrous Na2SO4 and concentrated under reduced pressure to afford a yellow oil. Purification by chromatography on silica gel using gradient elution, 2:1→1:1 hexanes/EtOAc furnished the product as a clear oil (350 mg, 30%). TLC Rf = 0.11 (33% 1 EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 8.81 (1H, d, J = 53 Hz), 7.68 (1 H, m), 7.66 (1 H, m), 7.08 (1 H, d, J = 8.4 Hz), 3.08 (4 H, m), 1.16 (6H, t, J = 7.0 Hz) ppm; 13C

NMR (CDCl3, 100 MHz) δ 171.3, 171.2, 154.0, 135.1, 130.1, 123.4, 118.5, 118.5, 117.5, 117.3, 39.3, 13.9 ppm.

Et2N O P O N

Br 2.7red Prepared according to the procedure outlined above using 5- 1 bromosalicylaldehyde. TLC Rf = 0.14 (33% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 8.81 (1H, d, J = 53 Hz), 7.68 (1 H, m), 7.66 (1 H, m), 7.08 (1 H, d, J = 8.4 Hz), 3.08 (4 13 H, m), 1.16 (6H, t, J = 7.0 Hz) ppm; C NMR (CDCl3, 100 MHz) δ 169.9, 169.8, 153.3,

61 Chapter 2

137.8, 132.5, 120.8, 120.7, 119.0, 118.7, 115.6, 29.6, 39.6, 14.1, 14.1 ppm; IR (thin film) ν 3230, 2975, 2934, 2876, 1625, 1556, 1470, 1383, 1261, 1209, 1038, 960, 848, 758 cm- 1.

General procedure for benzoxathiazine preparation. To a solution of salicyclaldehyde (1.00 mmol) in 2.0 mL of DMA at 0 °C was quickly transferred solid

H2NSO2Cl (3.00 mmol, 3.0 equiv). Caution: A mild exotherm is generally noted upon combination of these reagents. The mixture was allowed to warm to ambient temperature and was stirred until TLC analysis indicated completion of the reaction (2–12 h). The reaction was quenched with 5 mL of a pH 7 NaH2PO4/NaOH aqueous buffer solution and transferred to a separatory funnel with 10 mL of Et2O. The organic layer was separated, and the aqueous layer was extracted with 2 x 5 mL of Et2O. The combined organic layers were washed successively with 2 x 3 mL of H2O and 1 x 5 mL of saturated aqueous NaCl, dried over MgSO4, and concentrated under reduced pressure. Purification by chromatography on silica gel (conditions given below) afforded the desired benzoxathiazine.

O O S N O

2.10 Prepared according to the procedure outlined above using 8-hydroxy-1- 1 naphthaldehyde. TLC Rf = 0.1 (20% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 8.76 (1H, s), 8.40 (1H, d, J = 8.4 Hz), 7.92 (1H, d, J = 8.2 Hz), 7.75 (3H, m), 7.56 (1H, d, J = 13 8.5Hz) ppm; C NMR (CDCl3, 100 MHz) δ 168.2, 153.5, 131.7, 128.4, 128.2, 125.8, 123.6, 123.1, 122.7,110.6 ppm; IR (thin film) ν 2361, 1595, 1382, 1195 cm-1.

62 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N

Cl 2.17 Prepared according to the procedure outlined above using 5- chlorosalicylaldehyde. Purified by chromatography on silica gel using 20% 1 EtOAc/hexanes (white solid, 93%). TLC Rf = 0.23 (15% EtOAc/hexanes); H NMR

(CDCl3, 400 MHz) δ 8.64 (s, 1H), 7.71 (dd, 1H, J = 8.7, 2.5 Hz), 7.68 (d, 1H, J = 2.3 Hz), 13 7.27 (d, 1H, J = 8.8 Hz) ppm; C NMR (CDCl3, 100 MHz) δ 166.7, 152.8, 137.6, 131.9, 130.2, 120.4, 116.3 ppm; IR (thin film) ν 1604, 1560, 1469, 1384, 1347, 1178, 832 cm-1; + + HRMS (ES ) calcd for C7H4ClNO3S 216.9600 found 239.9492 (MNa ).

O O S O N Cl

Cl Cl 2.27 Prepared according to the procedure outlined above using 3,5,6- trichlorosalicylaldehyde. Purified by chromatography on silica gel using 20% 1 EtOAc/hexanes (white solid, 35%). TLC Rf = 0.21 (20 % EtOAc/hexanes); H NMR 13 (CDCl3, 400 MHz) δ 9.11 (s, 1H), 7.90 (s, 1H) ppm; C NMR (CDCl3, 100 MHz) δ 164.4, 149.8. 137.6, 132.7, 131.0, 123.2, 115.4 ppm; IR (thin film) ν 1602.6, 1401.9, 1196.6, -1 + + 815.4 cm ; HMRS (ES ) calculated for C7H2Cl3NO3S 284.8821 found (MNa ).

O O S O N

F 2.37 Prepared according to the procedure outlined above using 6- fluorosalicylaldehyde. Used without further purification (white solid, 72%). TLC Rf = 0.15

63 Chapter 2

1 (20% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 8.95 (1H, s), 7.76 (1H, td, J = 6.15, 13 8.50 Hz), 7.13 (2 H, m) ppm; C NMR (CDCl3, 100 MHz) δ 162.2, 161.9 (d, JCF = 5.7 Hz),

159.6, 154.3 (d, JCF = 3.6 Hz), 139.0 (d, JCF = 10.7 Hz), 114.3 (d, JCF = 3.8 Hz), 112.7 (d,

JCF = 19.6 Hz), 105.6 (d, JCF = 16.3 Hz) ppm; IR (thin film) ν 3097, 1608, 1475, 1391, 1197, 1030 cm-1.

O O S O N

2.41red Prepared according to the procedure outlined above using 4-hydroxybiphenyl-3- carbaldehyde. The product was judged to be sufficiently pure for use without 1 chromatography (yellow crystalline solid, 55%). TLC Rf = 0.27 (4:1 hexanes/EtOAc); H

NMR (CDCl3, 500 MHz) δ 8.78 (s, 1H), 7.96 (dd, 1H, J = 8.6, 2.3 Hz), 7.85 (d, 1H, J = 2.2 Hz), 7.58-7.55 (m, 2H), 7.53-7.50 (m, 2H), 7.47-7.44 (m, 1H), 7.39 (d, 1H, J = 8.6 Hz) 13 ppm; C NMR (CDCl3, 125 MHz) δ 167.7, 153.4, 139.9, 137.9, 136.2, 129.3, 128.8, 128.6, 127.0, 119.0, 115.5 ppm; IR (thin film) ν 1615, 1570, 1373, 1185, 762 cm-1; HRMS + + (ES ) calcd C13H9NO3S 259.0303 found 282.0207 (MNa ).

O O S O N

Me 2.42red Prepared according to the procedure outlined above using 4-hydroxy-4'- methylbiphenyl-3-carbaldehyde. Purified by chromatography on silica gel using 14% 1 EtOAc/hexanes (yellow solid, 41%). TLC Rf = 0.17 (9% EtOAc/hexanes); H NMR

(CDCl3, 500 MHz) δ 8.73 (s, 1H), 7.93 (ddd, 1H, J = 1.1, 2.2, 8.56 Hz), 7.83 (dd, 1H, J = 0.7, 2.2 Hz), 7.46 (d, 2H, J = 7.7 Hz), 7.32 (d, 1H, J = 7.91 Hz), 7.34 (m, 2H) 2.44 (s, 3H)

64 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

13 ppm; C NMR (CDCl3, 125 MHz) δ 167.9, 152.9, 139.6, 138.6, 135.8, 134.8, 129.9, 128.6, 126.7, 118.7, 115.4, 21.1 ppm; IR (thin film) ν 1613, 1569, 1380, 1182, 1126, 809 cm-1.

O O S O N

NO2

2.44red Prepared according to the procedure outlined above using 4-hydroxy-4'- nitrobiphenyl-3-carbaldehyde. Purified by chromatography on silica gel using gradient elution, 4:1→1:1 hexanes/EtOAc (yellow solid, 42%). TLC Rf = 0.13 (4:1 1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.79 (s, 1H), 8.38 (m, 2H), 8.00 (dd, 1H, J = 8.5 Hz, 2.3 Hz), 7.92 (d, 1 H, J = 2.3 Hz), 7.75 (m, 2H), 7.47 (d, 1H, J = 8.5 Hz) ppm; 13 C NMR (CDCl3, 125 MHz) δ 167.2, 154.4, 147.9, 144.1, 137.3, 136.2, 129.2, 127.9, 124.6, 119.6, 115.7 ppm; IR (thin film) ν 1609, 1569, 1517, 1388, 1347, 1190, 835 cm-1; + + HRMS (ES ) calcd for C13H8N2O5S 304.0154 found 327.0038 (MNa ).

O O S O N

F3C CF3 2.46red Prepared according to the procedure outlined above using 4-hydroxy-3',5'- bis(trifluoromethyl)biphenyl-3-carbaldehyde. Purified by chromatography on silica gel using 1 10:1 hexanes/EtOAc (white solid, 49%). TLC Rf = 0.13 (9% EtOAc/hexanes); H NMR

(CDCl3, 500 MHz) δ 8.03 (s, 1H), 7.96 (d, 1H, J = 2.2 Hz), 7.86 (dd, 1H, J = 8.5, 2.3 Hz), 7.39 (d, 1H, J = 8.5 Hz), 5.54 (s, 1H) ppm; IR (thin film) ν 1619, 1421, 1383, 1280, 1185, 1135, 1060, 860 cm-1.

65 Chapter 2

O O S O N

F F

F F F Prepared according to the procedure outlined above using 2',3',4',5',6'-pentafluoro- 4-hydroxybiphenyl-3-carbaldehyde. Purified by chromatography on silica gel using 4:1 1 hexanes/EtOAc (white solid, 61%). TLC Rf = 0.41 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.74 (s, 1H), 7.84 (d, 1H, J = 8.5 Hz), 7.80 (s, 1H), 7.46 (d, 1H, J = 8.6 Hz) 13 ppm; C NMR (CDCl3, 125 MHz) δ 167.2, 154.3, 144.0 (dm, JCF = 248 Hz), 141.2 (dm,

JCF = 257 Hz), 139.1, 138.1 (dm, JCF = 253 Hz), 132.5, 124.8, 119.2, 115.4, 112.7 (m) ppm; IR (thin film) ν 1617, 1574, 1524, 1500, 1397, 1192, 1066, 992 cm-1.

General procedure for oxaziridine preparation. To a solution of benzoxathiazine (2.73 mmol) in 10.9 mL of CH3CN and 8.2 mL of H2O was added a mixture of solid Oxone (4.09 mmol, 1.5 equiv) and NaHCO3 (12.29 mmol, 4.5 equiv) over a ~1 min period, during which time gas evolution was observed. After stirring this mixture for 15 min, the slurry was diluted with 40 mL of EtOAc and 25 mL of H2O and transferred to a separatory funnel. The organic layer was collected and the aqueous layer was extracted with 2 x 10 mL of EtOAc. The combined organic fractions were washed with 1 x

10 mL of saturated aqueous NaCl, dried over MgSO4, and concentrated under reduced pressure. Purification to obtain the desired oxaziridine was performed as described below.

O O S O N O H

2.2 The product was judged to be sufficiently pure for use without chromatography 1 (colorless oil that solidifies on standing, 97%). TLC Rf = 0.47 (30% EtOAc/hexanes); H

NMR (CDCl3, 400 MHz) δ 7.70 (dd, 1H, J = 7.6, 1.6 Hz), 7.64-7.60 (m, 1H), 7.43 (dt, 1H, J 13 = 7.6, 1.1 Hz), 7.22 (d, 1H, J = 8.2 Hz), 5.41 (s, 1H) ppm; C NMR (CDCl3, 100 MHz) δ

66 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

150.2, 133.9, 132.0, 127.5, 119.7, 116.0, 79.1 ppm; IR (thin film) ν 1621, 1491, 1467, -1 + 1414, 1254, 1192, 1104, 906, 751 cm ; HRMS (ES ) calcd for C7H5NO4S 198.9939 found 221.9835 (MNa+).

O NEt2 P O N O

2.6

Prepared according to the procedure outlined above using benzoxathiazine 2.6red.

Purified by chromatography on silica gel using 2:1 hexanes/EtOAc. TLC Rf = 0.21 (33% 1 EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.56 (d, 1H, J = 7.7 Hz), 7.44 (t, 1H, J = 8.7, Hz), 7.19 (t, 1H, J = 7.7 Hz), 7.00 (d, 1H, J = 8.0 Hz), 5.41 (d, 1H, J = 11.9 Hz), 2.99 13 (4H, m), 1.03 (t, 6H, J = 7.3 Hz) ppm; C NMR (CDCl3, 100 MHz) δ 150.9, 150.8, 132.8, 131.0, 124.4, 118.3, 118.2, 116.9, 116.8, 79.7, 79.6, 40.1, 40.1, 14.4 ppm; IR (thin film) ν 2979, 2936, 1492, 1466, 1384, 1282, 1033, 933, 759 cm-1.

Et2N O P O N O H

Br 2.7

Prepared according to the procedure outlined above using benzoxathiazine 2.7red. Purified by chromatography on silica gel using 2:1 hexanes/EtOAc (white solid, 39%). 1 TLC Rf = 0.25 (33% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.72 (d, 1H, J = 2.4 Hz), 7.59 (ddd, 1H, J = 8.7, 2.4, 2.4 Hz), 6.97 (d, 1H, J = 8.7 Hz), 5.39 (d, 1H, J = 11.7 Hz), 3.07 (q, 2H, J = 7.1 Hz), 3.03 (q, 1H, J = 7.1 Hz), 1.12 (t, 6H, J = 7.1 Hz) ppm; 13C

NMR (CDCl3, 100 MHz) δ 150.1, 150.0, 135.7, 133.6, 120.3, 120.2, 119.1, 119.0, 116.7, 79.0, 79.0, 40.3, 40.3, 14.5 ppm; IR (thin film) ν 2976, 1485, 1283, 1209, 1034, 960, 858 cm-1.

67 Chapter 2

O O S N O O

2.11 Prepared according to the procedure outlined above using benzoxathiazine 2.10. 1 TLC Rf = 0.46 (20% EtOAc/hexanes); H NMR (CDCl3, 500 MHz) δ 8.18 (m, 1H), 7.93 (m, 1H), 7.89 (d, 1H, J = 9.2), 7.68 (m, 2H), 7.64 (d, 1H, J = 8.4), 5.54 (s, 1H) ppm; 13C NMR

(CDCl3, 125 MHz) δ 136.1, 129.5, 128.5, 128.1, 128.1, 127.1, 125.3, 124.1, 121.0, 111.0, 79.5 ppm; IR (thin film) ν 1413, 1269, 1204, 1159, 1068, 892 cm-1.

O O S O N O

Cl 2.20 Prepared according to the procedure outlined above using benzoxathiazine 2.19. Purified by chromatography on silica gel using 2:1 hexanes/EtOAc (white solid, 80%). 1 TLC Rf = 0.31 (20% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.66 (d, 1H, J = 8.2 Hz), 7.43 (d, 1H, J = 1.9, 8.2 Hz), 7.25 (d, 1H, J = 1.9 Hz), 5.42 (s, 1H) ppm; 13C NMR

(CDCl3, 100 MHz) δ 150.1, 140.0, 132.6, 127.7, 120.1, 114.2, 78.4 ppm.

O O S O N O

Cl Cl 2.24 Prepared according to the procedure outlined above using benzoxathiazine 2.23.

Purified by chromatography on silica gel using 15:1 hexanes/EtOAc. TLC Rf = 0.3 (1:10 1 EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.70 (1H, d, J = 8.9 Hz), 7.13(1H, d, J = 13 8.9 Hz), 6.00 (1H, s) ppm; C NMR (CDCl3, 100 MHz) δ 149.1, 135.6, 134.4, 132.6, 118.8, 115.8, 77.3 ppm.

68 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O Cl

Cl Cl 2.28 Prepared according to the procedure outlined above using benzoxathiazine 2.23. 1 TLC Rf = 0.27 (20% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.80 (s, 1H), 6.00 (s, 13 1H) ppm; C NMR (CDCl3, 100 MHz) δ145.2, 134.5, 133.9, 132.6, 124.1, 117.0, 76.8 ppm.

O O S O N O

CF3 2.36 Prepared according to the procedure outlined above using benzoxathiazine 2.25. Purified by chromatography on silica gel using 10% EtOAc/hexanes (white solid, 80%). 1 TLC Rf = 0.34 (10% EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.99 (d, 1H, J = 2.0 Hz), 7.90 (dd, 1H, J = 8.6, 1.6 Hz), 7.37 (d, 1H, J = 8.7 Hz), 5.48 (s, 1H) ppm; 13C NMR

(CDCl3, 100 MHz) δ 152.3, 131.2 (q, JCF = 4 Hz), 130.3 (q, JCF = 34 Hz), 129.4 (q, JCF = 4 19 Hz), 122.9 (q, JCF = 273 Hz), 120.6, 117.1, 78.4 ppm; F NMR (CDCl3, 376 MHz) δ –63.6 (s, 3F) ppm; IR (thin film) ν 1631, 1505, 1425, 1332, 1238, 1204, 1135, 1073, 923, 834 -1 + + 734 cm ; HRMS (ES ) calcd for C8H4F3NO4S 266.9813 found 289.9709 (MNa ).

O O S O N O

F 2.38 Prepared according to the procedure outlined above using benzoxathiazine 2.37.

Purified by chromatography on silica gel using 10:1 hexanes/EtOAc. TLC Rf = 0.42 (20% 1 EtOAc/hexanes); H NMR (CDCl3, 400 MHz) δ 7.60 (td, 1H, J = 6.0, 8.5 Hz), 7.19 (td, 1H, 13 J = 0.9, 8.8 Hz), 7.05 (d, 1H, J = 8.4), 5.77 (s, 1H) ppm; C NMR (CDCl3, 100 MHz)

69 Chapter 2

δ164.2, 161.7, 150.4 (d, JCF = 4 Hz), 134.6 (d, JCF = 10 Hz), 115.2 (d, JCF = 3.9 Hz),

114.8, 104.7 (d, JCF = 15.5 Hz), 73.7 (d, JCF = 6.4 Hz) ppm.

O O S O N O

2.40 Prepared according to the procedure outlined above using benzoxathiazine 2.39. 1 TLC Rf = 0.67 (33% EtOAc/hexanes); H NMR (CDCl3, 500 MHz) δ 7.45 (t, 1H, J = 8.0 Hz), 7.21 (d, 1H, J = 7.8 Hz), 7.03 (d, 1H, J = 7.2 Hz), 5.64 (s, 1H), 2.60 (s, 3H) ppm; 13C

NMR (CDCl3, 125 MHz) δ 150.8, 141.6, 133.2, 129.6, 117.4, 113.7, 76.6, 18.2 ppm; IR (thin film) ν 1621, 1471, 1414, 1203, 1025, 962, 851 cm-1.

O O S O N O

2.41 Prepared according to the procedure outlined above using benzoxathiazine

2.41red. Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (yellow solid, 1 85%). TLC Rf = 0.46 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 7.86 (d, 1H, J = 2.2 Hz), 7.79 (dd, 1H, J = 8.5, 2.2 Hz), 7.58-7.55 (m, 2H), 7.53-7.49 (m, 2H), 7.57-7.43 13 (m, 1H), 7.29 (d, 1H, J = 8.5 Hz) ppm; C NMR (CDCl3, 125 MHz) δ 149.0, 141.0, 138.2, 132.1, 130.3, 129.2, 128.5, 127.1, 119.8, 116.0, 79.0 ppm; IR (thin film) ν 3037, 2925, -1 + 1619, 1485, 1416, 1196, 1117, 917, 844, 762 cm ; HRMS (ES ) calcd for C13H9NO4S 275.0252 found 298.0105 (MNa+).

70 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

Me 2.42 Prepared according to the procedure outlined above using benzoxathiazine

2.42red. Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (off-white 1 solid, 76%). TLC Rf = 0.3 (9% EtOAc/hexanes); H NMR (CDCl3, 500 MHz) δ 7.84 (d, 1H, J = 2.2), 7.76 (dd, 1H, J = 2.2, 8.5 Hz), 7.46 (2H, m), 7.31 (dd, 2H, J = 0.5, 8.5 Hz), 7.27 13 (d, 1H, J = 8.51), 5.46 (s, 1H), 2.44 (s, 3H) ppm; C NMR (CDCl3, 125 MHz) δ 148.5, 140.6, 138.5, 135.0, 131.6, 130.0, 129.8, 126.7, 119.4, 115.6, 79.0, 21.0 ppm; IR (thin film) ν 3030, 2923, 1618, 1492, 1420, 1194, 1116, 916, 836 cm-1.

O O S O N O

2.43 Prepared according to the procedure outlined above using benzoxathiazine

2.43red. Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (off-white 1 solid, 52%). TLC Rf = 0.32 (9% EtOAc/hexanes); H NMR (CDCl3, 500 MHz) δ 7.75 (d, 1H, J = 2.1 Hz), 7.72 (dd, 1H, J = 1.2, 9.0 Hz), 7.65 (dd, 1H, J = 2.2, 8.4 Hz), 7.43 (m, 13 1H), 7.33 (dd, 1H, J = 1.7, 7.5 Hz), 7.29 (m, 2H), 5.45 (s, 1H) ppm; C NMR (CDCl3, 125 MHz) δ 149.2, 140.4, 139.5, 134.6, 133.4, 132.7, 131.0, 129.9, 127.8, 122.2, 119.2, 115.3, 78.9 ppm; IR (thin film) ν 1619, 1502, 1418, 1195, 1117, 918, 843 cm-1.

71 Chapter 2

O O S O N O

NO2 2.44 Prepared according to the procedure outlined above using benzoxathiazine

2.44red. Purified by chromatography on silica gel using 4:1 hexanes/EtOAc (off-white 1 solid, 30%). TLC Rf = 0.33 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.38 (m, 2H), 7.94 (d, 1H, J = 2.2 Hz), 7.85 (dd, 2H, J = 8.5 Hz, 2.2 Hz), 7.75 (m, 2H), 7.38 (d, 1H, J = 8.5 Hz), 5.51 (s, 1H) ppm; IR (thin film) ν 2923, 1514, 1410, 1351, 1186, 1112, 917, 842 cm-1.

O O S O N O

F3C CF3 2.46 Prepared according to the procedure outlined above using benzoxathiazine

2.46red. Purified by chromatography on silica gel using 10:1 hexanes/EtOAc. TLC Rf = 1 0.13 (9% EtOAc/hexanes); H NMR (CDCl3, 500 MHz) δ 8.81 (s, 1H), 7.98 (m, 5H), 7.48 (d, 1H, J = 8.7 Hz) ppm; IR (thin film) ν 1616, 1574, 1383, 1280, 1188, 1134, 1061 cm-1.

O O S O N O

F F

F F F Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (white solid, 1 75%). TLC Rf = 0.53 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 7.79 (m, 1H),

72 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

13 7.70 (d, 1H, J = 8.5 Hz), 7.37 (d, 1H, J = 8.6 Hz), 5.47 (s, 1H) ppm; C NMR (CDCl3, 125 13 MHz) δ C NMR (CDCl3, 125 MHz) δ 150.5, 144.3 (dm, JCF = 250 Hz), 141.5 (dm, JCF =

256 Hz), 138.3 (dm, JCF = 256 Hz), 135.9, 133.8, 126.3, 120.4, 116.8, 113.3, 78.9 (m) ppm; IR (thin film) ν 1526, 1500, 1421, 1207, 1190, 1066, 991, 869 cm-1.

General procedure for pseudo-first order rate constant determination

A 10 mL Schlenk flask fitted on the sidearm with a rubber septum and a N2 gas inlet was charged with Me2NTs (19.9 mg, 0.10 mmol, used as an internal standard). The flask was sealed with a lightly greased glass stopper and a 1.00 M dichloroethane solution of olefin 3 (1.00 mL, 1.00 mmol, 10 equiv) was added via syringe through the side-arm portal. The stopcock was closed and the reaction flask was placed in a temperature-regulated water bath pre-heated to 25 °C using an IKA RCT basic hot plate/stirrer with attached ETS-D4 thermocouple. The solution was allowed to equilibrate with gentle stirring for 15 min. The glass stopper was removed and a single portion of the finely powdered oxaziridine (1.00 mmol) was quickly added. The flask was re-sealed and the solution stirred at the set temperature. At 15 pre-determined time points, the flask was quickly unstoppered and, with the aid of a 200 µL Gilson brand pipetter, a 25 µL aliquot of the reaction solution was removed and immediately quenched into a mixture containing PPh3 (50 µL of a 0.1 M CH2Cl2 solution) and 0.50 mL of HPLC-grade hexanes. Analysis of each sample was performed by analytical HPLC using iPrOH/hexanes as the eluent (gradient elution, 0.5%→50% iPrOH/hexanes). The ratio of integrated absorbances A(t) = A(t)epoxide/A(t)internal standard measured at 254 nm was plotted as a function of time (t, min). The observed rate constant, kobs, was determined by fitting the -kt data according to A(t) = A(t∞) – A(t∞)e using a graphing program (Matlab, version 7.6.0). Kinetic runs were performed a minimum of two times for each oxaziridine. From this data, an average kobs and standard deviation was determined.

73 Chapter 2

O O O O S S k O O N Me OBz obs O N Me OBz O + + H Me 25 ºC H Me

R 2.3 R 2.5

Average kobs Entry Oxaziridine (x 10–4 s-1) 1 R = OMe 2.21(4) 2 H 2.6(1) 3 Ph 3.2(6) 4 Cl 5.45(4) 5 p-C6H4NO2 5.2(7) 6 C6F5 7(1)

7 CF3 7.39(2) 8 NO2 15.7(1)

74 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O NEt2 P O N O

2.6

75 Chapter 2

O NEt2 P O N

2.6red

76 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

Et2N O P O N O H

Br 2.7

77 Chapter 2

Et2N O P O N

Br 2.7red

78 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S N O

2.10

79 Chapter 2

O O S N O O

2.11

80 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

2.20

81 Chapter 2

O O S O N O

Cl Cl 2.24

82 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N Cl

Cl Cl 2.27

83 Chapter 2

O O S O N O Cl

Cl Cl 2.28

84 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O H

Br 2.32

85 Chapter 2

O O S O N

2.37

86 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

2.38

87 Chapter 2

O O S O N O

2.40

88 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N

2.41red

89 Chapter 2

O O S O N O

2.41

90 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N

2.42red

91 Chapter 2

O O S O N O

Me 2.42

92 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

2.43

93 Chapter 2

O O S O N

NO2

2.44red

94 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

F F

F F F

95 Chapter 2

O O S O N

F F

F F F

96 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

NO2 2.44

97 Chapter 2

O O S O N

F3C CF3

2.46red

98 Benzoxathiazine-Derived Oxaziridines for Selective O-Atom Transfer

O O S O N O

F3C CF3 2.46

99 Chapter 3

Chapter 3. A Catalytic Method for the Hydroxylation of C–H Bonds

3.1 Introduction The benzoxathiazine heterocycle functions as a unique platform for the design of C–H hydroxylation catalysts (Figure 3.1). The design, synthesis and analysis of a large number of the corresponding oxaziridines were presented in Chapter 2. While N-sulfonyl oxaziridines have proven to be excellent reagents for stoichiometric O-atom transfer, the goal of this work was to develop a catalytic oxaziridine-mediated system for C–H hydroxylation. The work presented in this chapter documents our efforts towards this end.

O O O O S S O N O N H [O] O OH CF3 R' R' Me CF3 Me Me Me

R R

Figure 3.1. Benzoxathiazines react with a terminal oxidant to form a N-sulfonyl oxaziridine capable of transferring oxygen to a C–H bond.

The proposed catalytic cycle, originally illustrated in Chapter 2, is presented in a revised form in Figure 3.2. Stoichiometric oxygen transfer from oxaziridine to substrate (step c) was analyzed both by theory and experiment for a large number of oxaziridines (Chapter 2). Herein is described the application of the corresponding benzoxathiazines in a catalytic reaction. Two distinct reaction protocols were developed that employ different terminal oxidants and co-catalysts and ultimately result in different substrate scopes.

O O S O N O O H O X 2 2 R OH R S O O a b c R S O N O H O SO 2 X R R OOH

Figure 3.2. A catalytic cycle for C–H hydroxylation mediated by a benoxathiazine-derived oxaziridine.

100 A Catalytic Method for the Hydroxylation of C–H Bonds

3.2 Protocol A: Diaryldiselenide and Urea Hydrogen Peroxide

3.2a Diaryldiselenides: background and synthesis The choice of terminal oxidant for the benzoxathiazine catalyzed C–H hydroxylation is an important design element; the terminal oxidant should be able to oxidize the heterocycle to an oxaziridine without undesirable background oxidation. While reagents such as m-CPBA and Oxone are capable of oxidizing a benzoxathiazine to an oxaziridine, these are not desirable terminal oxidants because they exhibit poor atom economy and, particularly in the case of m-CPBA, can require cumbersome purifications to remove by-products. Conversely, hydrogen peroxide would be an ideal oxidant because it is the most atom-economical source of an oxygen atom besides molecular oxygen. With water as its only by-product, it is environmentally benign and the lack of organic waste obviates the need to separate by-products from the reaction mixture. Hydrogen peroxide is incapable of oxidizing a benzoxathiazine to the corresponding oxaziridine at a discernable rate and a suitable co-catalyst is required if it is to be used successfully for this purpose. A promising candidate for this task is a diaryldiselenide. A diaryldiselenide will react with hydrogen peroxide to form seleninic acid (Figure 3.3).106 Further reaction with hydrogen peroxide yields the perseleninic acid, which, upon oxygen transfer to substrate, regenerates the seleninic acid, enabling catalytic turnover. The oxidizing reactivity of the perseleninic acid can be tuned through the judicious choice of the aryl electron-withdrawing group. We envisioned that the combination of a diaryldiselenide and urea hydrogen peroxide (UHP), a crystalline, anhydrous form of hydrogen peroxide, would be an operationally simple procedure.

H2O2 O O R Se H O R Se R Se 2 2 OH OOH Se R

SO S

Figure 3.3. Diaryldiselenides react with hydrogen peroxide to form a perseleninic acid.

The research groups of Sheldon and Syper have explored diaryldiselenides as precatalysts for olefin epoxidation and Bayer-Villger oxidation reactions.106,107 Figure 3.4

106 Syper, L.; Mlochowski, J. “Benzeneperoxyseleninic Acids- Synthesis and Properties.” Tetrahedron 1987, 43, 207-213. 107 (a) ten Brink, G. J.; Vis, J. M.; Arends, I. W. C. E.; Sheldon, R. A. “Selenium-catalyzed oxidations with aqueous hydrogen peroxide. 2. Baeyer-Villiger reactions in homogeneous solution.” J Org Chem 2000,

101 Chapter 3

illustrates a variety of diaryldiselenides used as precatalysts in the epoxidation of olefins. The efficiency of this transformation is dependent on the electrophilicity of the resulting perseleninic acid. Accordingly, the most electron-deficient diaryldiselenides, bis(3,5- bis(trifluoromethyl)phenyl)diselenide, gives the highest yield of epoxide following a 1 hr reaction.

0.5 mol% catalyst 0.2 mol% NaOAc 2 equiv H O 2 2 O 2,2,2-trifluoroethanol 20 ºC, 1 h

Entry 1 2 3 4

F3C Se Se Se 2 none Catalyst 2 2 F Cl CF3 Yield (%) 64 70 98 <1

Figure 3.4. Electron-deficient diaryldiselenides are excellent precatalysts for simple olefin epoxidation reactions.

Because arylseleninic acids are known to catalyze the epoxidation of olefins, we first performed control experiments with saturated hydrocarbon substrates. When cis- dimethylcyclohexane was subjected to 20 mol% diaryldiselenide (Ar2Se2) and 8 equivalents UHP in acetonitrile, ~5% conversion to product was observed by GCMS (Figure 3.5). We were curious to see if more electron-deficient diaryldiselenides would give higher conversions of product and thus diaryldiselenides were investigated as potential pre-catalysts for C–H hydroxylation systems. A variety of diaryldiselenides were synthesized following the general protocol illustrated in Figure 3.5.108 Substrates such as indane, cis-decalin, and cis-dimethylcyclohexane were treated with 20 mol% of a diaryldiselenide and eight equivalents of UHP in acetonitrile at room temperature. Surprisingly, the highly electron-deficient bis(trifluoromethyl)phenyl)diselenide gave only trace conversion to product. All of the diaryldiselenides tested gave less than 10% conversion to product (Figure 3.5).

66, 2429-2433; (b) ten Brink, G. J.; Fernandes, B. C. M.; van Viliet, M. C. A.; Arends, I. W. C. E.; Sheldon, R. A. “Selenium catalysed oxidations with aqueous hydrogen peroxide. Part I: epoxidation reactions in homogeneous solution.” J Chem Soc Perk Transactions 1 2001, 3, 224-228. 108 Ph2Se2 is commercially available from Aldrich. Bis(3,5-bis(trifluoromethyl)phenyl)diselenide is courtesy of Dr. Benjamin Brodsky. For Pyr2Se2 see: Toshimitsu, A.; Owada, H.; Terao, K.; Uemura, S.; Okano, M. J. Org. Chem. 1984, 49, 3796. For NMeImid2Se2 see: Roy, G.; Nethaji, M.; Mugesh, G. J. Am. Chem. Soc. 2004, 126, 2712.

102 A Catalytic Method for the Hydroxylation of C–H Bonds

Test Reaction: Ar Se = Ph Se 20 mol% Ar Se 2 2 2 2 Me 2 2 Me ~5% conversion 8 equiv UHP OH = 3,5-CF -Ph Se Me CH CN, rt Me 3 2 2 3 <2% conversion Diaryldiselenide Synthesis: 0 Br 1) Mg , Et2O 2) Se0 R Se R = -CF , -Cl, -F Se R 3 3) HCl, H2O R 4) air, MeOH

F3C Se Me Se Se Se Se 2 2 N Se 2 2 2 N 2 F Cl N CF3 Figure 3.5. C–H hydroxylation was briefly examined with a range of diaryldiselenides.

Although the diaryldiselenides did not prove to be effective as precatalysts for C– H hydroxylation, they were quite useful as co-catalysts for the oxaziridine-mediated catalytic C–H hydroxylation reaction. High conversion of benzoxathiazine 3.1 to oxaziridine 3.2 was achieved employing 2 mol% of bis(3,5- bis(trifluoromethyl)phenyl)diselenide (3.3) (Figure 3.6).109,110

O O O O 2 mol% Ar2Se2 (3.3) S 2 equiv UHP S O N O N O 92% conversion DCE, 22 ºC CF3 CF3

3.1 3.2

Figure 3.6. Diaryldiselenides are useful catalytic reagents for oxidizing a benzoxathiazine to a oxaziridine.

With the diselenide pre-catalyst established, the proposed catalytic reaction was explored. The perseleninic acid co-catalyst proved successful at mediating oxygen transfer between UHP and the benzoxathiazine. A dual catalytic system was developed that employed 20 mol% benzoxathiazine catalyst 3.4, 1 mol% diaryldiselenide 3.3, and 4 equivalents of UHP in DCE. Adamantane was hydroxylated in good yield; however, when

109 Brodsky, B. H. Ph. D. Thesis, Stanford University, 2007. 110 Although it would be preferable to employ the commercially available Ph2Se2, upon oxidation to the perseleninic acid this reagent is prone to irreversible rearrangement to give a catalytically inactive Se(VI) species. Incorporating electron-withdrawing groups to the perseleninic acid suppresses the rearrangement to give the selenate. For more information, see reference (1).

103 Chapter 3

this protocol was employed with acyclic tertiary C–H bond substrates, the product conversion dropped significantly (Table 3.1).111

Table 3.1. Substrate scope for C–H hydroxylation with a first generation benzoxathiazine catalyst.

Substrate Product time (h) temp ( ºC) Yield (%)

95 22 80 O O S O N

Me Me CF3 OBz OBz 96 50 37 Me Me OH Cl Me OBz Me OBz 3.4 73 50 13b Me Me OH

a Reactions conducted in DCE using 1 mol% diaryldiselenide 3.3, 20 mol% catalyst 3.4, and 4 equiv of UHP. b Conversion determined by 1H NMR of the unpurified reaction mixture.

3.2b Improved catalyst design With conditions for a catalytic reaction established, we set out to improve the product conversions and reaction scope by exploring the effect of oxaziridine structure on reaction efficiency. Based on the SAR studies outlined in Chapter 2, our hypothesis was that electron-deficient benzoxathiazines, which afford the most reactive oxaziridines for stoichiometric oxygen transfer, would furnish the most effective catalysts (see Conclusions, Chapter 2). Further, previous results by Brodsky suggested that installation of an electron-withdrawing group at C4 had a profound effect on product conversion; benzoxathiazines with C4 = H gave trace conversion to hydroxylated product whereas benzoxathiazines with C4 = CF3 were suitable catalysts for C–H hydroxylation (Table 3.1).109 With these considerations in mind, we explored the effect of adding various electron-withdrawing groups to C4.112

DFT calculations show that oxazirdine 3.1 (where C4 = CF3) gave one of the lowest activation barriers tested (Entry 3, Table 3.2). Because these C4-substituted oxaziridines are considerably more reactive than the salicylaldehyde analogues described in Chapter 2, the less reactive terminal olefin substrate 3.5 was prepared for kinetic analysis (Table 3.2). These studies confirm the results of the DFT calculations with the

111 Brodsky, B. H.; Du Bois. J. “Oxaziridine-mediated catalytic hydroxylation of unactivated 3º C-H bonds using hydrogen peroxide.” J Am Chem Soc 2005, 127, 15391-15393. 112 These studies were performed in conjunction with B. H. Brodsky.

104 A Catalytic Method for the Hydroxylation of C–H Bonds

CF3-substituted benzoxathiazine 3.1 giving rates of oxygen transfer 86 times faster than the parent oxaziridine 3.7.

Table 3.2. The effect of C4 substitution of the rate of olefin epoxdiation.

O O O O S S O N kobs O N O O + OBz + OBz 8 4 R 25 ºC R 7 5 6 3.5 3.6

a -1 k Entry Oxaziridine R Group Calculated Ea (kcal/mol) kobs (s ) rel

1 3.7 H 24.9 0.58 x 10-5 1.0 -5 2 3.8 CO2Et -- 6.3 x 10 11 -4 3 3.1 CF3 22.4 5.0 x 10 86 -4 4 3.9 C3F7 22.3 3.6 x 10 62 5 3.10 Ph 28.9 --b -- b 6 3.11 C6F5 ------

a DFT calculated activation barriers for ethylene epoxidation using B3LYP/6-31G* b Rate was too slow to get an accurate rate constant.

When larger groups were appended to C4, such as C6H5 and C6F5, the rate of stoichiometric oxygen transfer for these oxaziridines was too slow to get an accurate rate constant (Table 3.2). The calculated activation barrier for ethylene epoxidation with oxaziridine 3.10 is 4 kcal/mol greater than the value calculated for the parent 3.7. Although electron-deficient oxaxiridine 3.11 would be expected to be a more electrophilic oxaziridine and thus a more potent C–H oxidant, adding a substituent so close to the reactive site inhibited activity by hindering the approach of the nucleophile. Ultimately, the trifluoromethyl group was chosen as the C4 substituent. In order to continue to take advantage of the diversity of starting materials offered by salicylaldehydes (Chapter 2), a short reaction sequence was employed to introduce the trifluoromethyl group (Figure 3.7). Ruppert’s reagent (CF3·TMS) was used with a catalytic amount of tetrabutylammonium fluoride to generate a trifluoromethyl nucleophile.113 The addition of this reagent to an aldehyde results in an alcohol that was oxidized with bis(acetoxy)iodobenzene and catalytic TEMPO to give the α-trifluoromethyl ketone. After deprotection of the phenolic methyl group, sulfamoylation gave the desired benzoxathiazine. Due to the more electron-deficient nature of benzoxathiazine 3.1 as compared to the benzoxathiazine 3.7 (C4 = H), buffered Oxone conditions generally

113 Wayman, K. A.; Sammakia, T. “O-nucleophilic amino alcohol acyl-transfer catalysts: the effect of acidity of the hydroxyl group on the activity of the catalyst.” Org Lett 2003, 5, 4105-4108.

105 Chapter 3

resulted in starting material decomposition. Peracids, such as m-CPBA, were sufficiently mild and reactions to generate the corresponding oxaziridine could be performed under non-aqueous conditions.

OMe O CF3TMS MeO HO CF PhI(OAc)2 OMe O cat. TBAF 3 cat. TEMPO H H CF3 THF, 0 ºC to rt CH2Cl2 3.12 3.13 3.14

BBr3 CH2Cl2 -78 ºC to rt O O O O S S OH O O N m-CPBA O N H2NSO2Cl O CF3 CF3 CH2Cl2 CF3 DMA, 0 ºC

3.2 3.1 3.15

Figure 3.7. A synthetic route to trifluoromethylated benzoxathiazines and related oxaziridines.

Generally, the Suzuki cross-coupling reactions described in Chapter 2 were carried out prior to the installation of the trifluoromethyl ketone because the propensity for the electron-deficient ketone to hydrate made subsequent manipulations difficult. In one exception, the precursor to benzoxathiazine 3.16 was synthesized after the installation of the trifluoromethylketone (Figure 3.8). When the salicylaldehyde containing the pendant pentafluorophenyl group was treated with CF3TMS, as in Figure 3.7 (3.12 → 3.13), the reaction failed. An SNAr reaction at an aryl-fluorine bond could be responsible for the falilure of this reaction, although this hypothesis was not confirmed. The pentafluorophenyl group was installed by a novel C–H activation method developed by Fagnou and coworkers.114 This method is operationally simple because the catalyst and phosphine ligand are both air stable. The yield for this reaction was consistently low to moderate (40-60%), although starting material could often be recovered.

114 Lafrance, M.; Rowley, C. N.; Woo, T. K.; Fagnou, K. “Catalytic intermolecular direct arylation of perfluorobenzenes.” J Am Chem Soc 2006, 128, 8754-8756.

106 A Catalytic Method for the Hydroxylation of C–H Bonds

O O OMe O S O N OMe O 1.5 equiv pentafluorobenzene CF3 5 mol% Pd(OAc)2 CF3 CF3 10 mol% PtBu3Me·HBF4 F F 1.1 equiv K2CO3 F F Br DMA, 120 ºC F F F F F F 3.16 Figure 3.8. A novel cross-coupling method developed by Fagnou and coworkers was used to install a pentafluorophenyl group at C6.

3.2c Results with diaryldiselenide cocatalyst Several benzoxathiazines that were tested as C–H hydroxylation catalysts are depicted in Figure 3.9. These benzoxathiazines proved to be competent catalysts under our reaction conditions (4 equiv urea•H2O2, 1 mol% diaryldiselenide 3.3). The production of 3° alcohol, however, did not differ substantially between experiments and, surprisingly, no other heterocycle outperformed non-substituted benzoxathiazine 3.1.115 Furthermore, benzoxathiazine 3.19, for which the corresponding oxaziridine gave one of the highest rates of stoichiometric oxygen transfer, performed quite poorly in the catalytic reaction. It was clear from this data that factors other than the oxidizing ability of the intermediate oxaziridine conspired to limit catalyst turnover in these reactions.

O O O O O O O O O O S S S S S O N O N O N O N O N Cl CF CF3 3 CF3 CF3 CF3

H Ph Cl CF3 Cl 3.1 3.17 3.4 3.18 3.19

O O Catalyst Conversion Me 20 mol% catalyst Me OH S O N R' = H R = H 3.1 55% 1 mol% Ar Se Me 2 2 Me R' Ph 3.17 40% CF 3 Cl 3.4 45% urea•H2O2 CF3 3.18 40% OBz OBz DCE, 50 °C, 96 h 20% R catalyst R' = Cl R = Cl 3.19

Figure 3.9. Electron-deficient catalysts give diminished conversions in the catalytic reaction.

115 A catalytic reaction with the C6–NO2 derived benzoxathiazinane afforded < 10% of the desired product.

107 Chapter 3

One possibility for the surprisingly poor catalytic performance of the electron- deficient benzoxathiazines is that the more electron-deficient benzoxathiazines favor to a greater extent hydration at C4, a factor that could retard the rate at which the oxaziridine is regenerated (Figure 3.10). It is also possible that other nucleophilic species (e.g.,

H2O2, urea, tertiary alcohol product) could add at C4. Evidence for the hydrated form of the benzoxathiazine can be obtained by 1H NMR (Figure 3.11). Two sets of signals corresponding to the benzoxathiazine (3.16) and the hydrated benzoxathiazine are observed after purification; when the compound is concentrated for >24 h in vacuo, only a single set of signals for 3.16 is observed.

O O O O O O S S S O N O O N O NH O OH H O product ArSeOOH 2 CF3 4 CF3 CF3

catalyst R R R arrest

Figure 3.10. A stable hydrate formed at C4 could inhibit oxaziridine formation under catalytic conditions.

* * *

Figure 3.11. 1H NMR spectra of benzoxathiazine 3.16 (C6=pentafluorophenyl). Asterisks are used to demarcate the signals corresponding to the hydrated benzoxathiazine. After >24 h at <1 torr, the catalyst is entirely dehydrated (spectrum on the right).

Perhaps these findings are not surprising given that other groups have obtained similar results with dioxiranes. For instance, while investigating dioxirane-mediated epoxidation chemistry, Denmark and coworkers noted that highly electron-deficient ketones gave poor product conversions under catalytic conditions.116 Although the

116 Denmark, S. E.; Forbes, D. C.; Hays, D. S.; Depue, J. S.; Wilde, R. G. J Org Chem 1995, 60 1391- 1407.

108 A Catalytic Method for the Hydroxylation of C–H Bonds

corresponding dioxiranes are known to be highly reactive oxygenation reagents, these ketones reacted with Oxone very slowly to form the dioxirane.117 This phenomenon is attributed to the fact that such electron-deficient ketones exist as stable hydrates in the presence of water; thus exchange with persulfate occurs much more slowly than with unsubstituted ketones such as acetone (Figure 3.12).

O O O O O OH- HSO - 5 O H3C H3C - H C CF OH O SO3 H3C CF3 3 3 F3C F3C

Figure 3.12. Electron-deficient ketones form stable hydrates and the necessary exchange with monoperoxysulfate is subsequently slow.

The fact that highly electron-deficient benzoxathiazines performed poorly as C–H hydroxylation catalysts, particularly in light of our Hammett data, presented something of an impasse. Factors that contributed to a more reactive O-atom transfer reagent—i.e., more electron-deficient oxaziridines—did not correlate to a more effective catalyst. Attempts to sequester water from the reaction mixture with molecular sieves were tried but not rigorously pursued because they are known to catalyze the decomposition of 118 hydrogen peroxide. Other standard dehydrating methods, i.e. MgSO4 and Na2SO4, did not give improved results, suggesting that perhaps hydrogen peroxide addition to the benzoxathiazine might be responsible for catalyst inhibition. It became evident that we needed to pursue alternative strategies in order to improve the effectiveness of this family of hydroxylation catalysts.

3.3 Protocol B: Acetic Acid and Aqueous Hydrogen Peroxide Concomitant with investigations of the benzoxathiazine electronic structure on catalytic function, the influence of alternative reaction conditions on catalytic performance was also explored. These studies revealed that catalytic C–H hydroxylation with a benzoxathiazine could be achieved using a combination of aqueous H2O2, acetic acid, and H2O (Figure 3.13). Under such conditions, peracetic acid is formed. There is no need to add a cocatalyst because the peracetic acid is likely the oxidant responsible for converting the benzoxathiazine to the corresponding oxaziridine. Control reactions show that AcOH is required for benzoxathiazine oxidation with H2O2. Furthermore, although the

117 Curci, R.; D’Accolti, L.; Fusco, C. “A Novel Approach to the Efficient Oxygenation of Hydrocarbons under Mild Conditions. Superior Oxo Transfer Selectivity Using Dioxiranes.” Accounts Chem. Res. 2006, 39, 1-9. 118 Nishimura, T.; Onoue, T.; Ohe, K.; Uemura, S. “Palladium(II)-catalyzed oxidation of alcohols to aldehydes and ketones by molecular oxygen.” J. Org. Chem. 1999, 64, 6750-6755.

109 Chapter 3

benzoxathiazine is entirely hydrated under these aqueous reaction conditions, the acidic environment helps to facilitate the exchange between the hydrated benzoxathiazine and the peracid adduct of the benzoxathiazine. Acetic acid also acts as a co-solvent to help solubilize apolar substrates.

O O S O N O Me H2O2 AcOH R OBz 8 equiv Me O O a b c R S O N AcOOH Me OH H2O R OBz Me

R 20 mol%

Figure 3.13. New catalytic conditions feature peracetic acid that is formed from hydrogen peroxide and acetic acid in situ.

Reaction conditions utilizing acetic acid and aqueous hydrogen peroxide offered a small but significant boost in catalytic performance when benzoxathiazine 3.4 was employed. For instance, conversion of substrate 3.20 more than doubled with this new protocol (Figure 3.14).

O O S 20 mol% 3.4 Conditions %Conv. O N Me OBz 50 ºC, 72 h Me OBz 1 mol% Ar Se 3.3, UHP HO 2 2 13 CF3 Me Me "conditions" 1:1 AcOH:H2O, H2O2 30 3.20 3.21 3.4 Cl

Figure 3.14. A marked improvement in product conversion is observed with aqueous reaction conditions.

3.3a Hydrophobic Effects in Catalysis With the knowledge that the hydroxylation reaction performs well under aqueous conditions, we envisioned exploiting hydrophobic effects to accelerate the rate of the hydroxylation event. Breslow demonstrated that hydrophobic effects can be operative in bimolecular reactions, particularly in the case of the Diels-Alder reaction, which is significantly accelerated when performed in water (Figure 3.15).119 The differences in rate are not simply due to polar solvent effects, as can be seen when comparing the rates of reaction in isooctane, methanol, and water. While methanol gives an approximately 10-

119 Rideout, D.; Breslow, R. J. Am. Chem. Soc. 1980, 102, 7816-7817.

110 A Catalytic Method for the Hydroxylation of C–H Bonds

fold increase in the rate of reaction compared to isooctane, the rate increase for the reaction performed in water is almost two orders of magnitude faster. This rate increase is ascribed to the hydrophobic effect and results from the favorable aggregation of two hydrocarbon surfaces in the transition state of the reaction.120

solvent relative rate

Me solvent isooctane 1 20 ºC O methanol 12 O Me water 730

Figure 3.15. A large increase in the rate of a Diels-Alder reaction is attributed to the hydrophobic effect.

Although the hydrophobic effect plays a prominent role in enzymatic catalysis, it is rarely exploited in catalysis. Breslow and coworkers recently showed that in D2O aromatic olefin substrates could be selectively epoxidized in preference to aliphatic olefin substrates by aromatic oxaziridinium salts (Figure 3.16).121 Selectivity was achieved by using an oxaziridinium oxidant that can π-stack with an aromatic group on the substrate. Because dioxirane epoxidation is thought to display a very similar transition-state geometry to that of oxaziridinium epoxidation, DMDO was employed in control experiments. A competition reaction between substrates 3.22 and 3.23 was conducted for each oxidant; a notable difference in selectivity was observed between catalysts. Remarkably, the hydrophobic oxaziridinium salt favors epoxidation of olefin 3.22 in a ratio of 97 to 3.

120 Breslow, R. “Hydrophobic effects on simple organic reactions in water.” Acc. Chem. Res., 1991, 24, 159-164. 121 Biscoe, M. R.; Breslow, R. “Oxaziridinium salts as hydrophobic epoxidation reagents: Remarkable hydrophobically-directed selectivity in olefin epoxdiation.” J. Am. Chem. Soc., 2005, 127, 10812-10813.

111 Chapter 3

O O O Oxidant O O O O O Me O D2O Me O

3.22 3.23 A B

entry Oxidant A : B transition state

1 MMPP 44:56 aromatic groups perpendicular

2 DMDO 61:39 no hydrophobic region

3 Oxazirdinium salt 96.5:3.5 aromatic groups parallel

O Me BF - OH N 4 O O O O 2+ O Mg Me Me O 2 MMPP DMDO Oxaziridinium salt

Figure 3.16. Competition studies show that oxidants that can π-stack with a substrate give high selectivity for that substrate.

Interestingly, when a 1:1 mixture of D2O and iPrOD is employed, little or no change in A:B selectivity is observed for reactions employing DMDO. With the oxaziridinium salt, however, the presence of the alcohol mitigates the hydrophobic effect, thus reducing the favorable contact between the substrate 3.22 and reagent and the overall selectivity (A:B 88:12 with 1:1 D2O:iPrOD), corresponding to a ~1 kcal/mol difference in ΔΔG‡. These results suggest that the selectivity observed in water is a consequence of the increased rate of epoxidation between substrate 3.22 and the oxaziridinium salt, although one should be cautious of over-interpreting these results given the observed selectivity in water/isopropanol.

3.3b Results Contemporaneous with the work of Breslow, we envisioned the successful combination of our newly developed aqueous reaction conditions with a highly hydrophobic benzoxathiazine catalyst. Because the transition state of the hydroxylation event involves a one to one complex of the alkane and oxaziridine, increasing the effective molarity between alkane and oxidant should make the hydroxylation reaction faster. Accordingly, the aqueous acetic acid medium of our reaction could help drive the aggregation of substrate with an appropriately configured catalyst. While initial exploratory reactions employing aqueous acetic acid reaction conditions with catalyst 3.1 gave modest returns, discernible improvement in product

112 A Catalytic Method for the Hydroxylation of C–H Bonds

conversion with catalysts 3.17 (C6 = Ph), 3.4 (C6 = Cl), and 3.18 (C6 = CF3) was observed (Table 3.3, entries 2–4). Optimal performance under these new conditions seemed to peak with the Cl-substituted catalyst 3.4. These results are similar to what was observed by us and Denmark previously: increasing the electron deficiency of the catalyst is beneficial until the point where hydration of the catalyst becomes a competitive reaction and thus diminishing the catalyst performance. Also intriguing was the fact that the C6–Ph catalyst 3.17 was more effective than 3.1 despite the apparent similarity in the electronic structures of these two systems. This latter result, albeit small, caused us to question the influence of catalyst shape (i.e., polar surface area, molecular volume) and solubility on catalytic function. To examine such effects, two additional benzoxathiazines were designed with C6-aryl substituent groups chosen to match closely the σp parameter of Cl (entries 5, 6). 122

Table 3.3 Catalytic C–H hydroxylation with aqueous H2O2.

R O O S Me 20 mol% O Me OBz N OBz Me F3C Me OH

8 equiv. H2O2 1:1 AcOH:H2O, 50 ºC, 96h

Entry R Catalyst !p Conversion (%)

1 H 3.1 0 40 2 Ph 3.17 0 50 3 Cl 3.4 0.23 70 4 CF3 3.18 0.54 60 5 C6F5 3.16 0.27 95 6 p-NO2C6H4 3.24 0.26 20

aConversion determined by 1H NMR integration of the unpurified reaction mixture. Reaction conditions: 20 mol% catalyst, 8 equiv 50% H2O2, 0.25 M 1:1 AcOH/H2O, 50 °C, 96 h.

Oxaziridines generated from catalysts 3.16 and 3.4 were found to possess nearly identical reactivity. The Hammett plot illustrated in Figure 3.17 shows that for the stoichiometric oxygen transfer reaction, these oxaziridines react with alkene with nearly -4 -1 -4 -1 identical rates (kobs(3.16) = 7.0 x 10 s , kobs(3.4) = 5.5 x 10 s ) Accordingly, the function of both catalysts should be comparable if this parameter were the principal determinant of reaction efficiency under the aqueous conditions. The results from hydroxylation reactions with 3.16 and 3.4 are markedly disparate, however, with the

122 Hansch, C.; Leo, A.; Taft, W. “A survey of Hammett substituent constants and resonance and field parameters.” Chem. Rev. 1991, 91, 165.

113 Chapter 3

pentafluorophenyl (PFP) catalyst 3.16 clearly outshining 3.4 and all others (Table 3.1). It is evident that C–H hydroxylation using benzoxathiazine-based catalysts in aqueous media is responsive to factors other than oxidant reactivity. Moreover, it appears that the mechanistic details underlying catalytic turnover are considerably more complex than initially envisioned. At present, we believe that the fluorocarbon moiety in 3.16 may present a large hydrophobic surface area that promotes catalyst-substrate aggregation and thus elevates the effective molarity between these two reacting partners. The improved performance of 3.16 is not simply due to higher catalyst stability under reaction conditions. In the case of C–H hydroxylation reactions that employ either 3.16 or 3.1 as the catalyst, >90% of the benzoxathiazine catalysts are recovered following a reaction.

Hammett Analysis

0.9 NO2 R= 0.6 O O S pNO2Ph O N CF3 0.3 Cl H C6F5 log(k/kH) H 0 OMe y = 0.83x + 0.08 Ph R2 = 0.95 R -0.3 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 ! p

Figure 3.17. Hammett analysis.

Results depicted in Table 3.4 define the substrate profile for reactions catalyzed by the PFP-substituted benzoxathiazine 3.16 under aqueous acetic acid-H2O2 reaction conditions. The active oxidant is exquisitely selective for 3° C–H bonds, the products of 2° hydroxylation having never been observed for any of the substrates shown. As predicted based on prior art, the catalyst is sensitive to the electronic environment of the C–H bond undergoing reaction.123 In substrates possessing two 3° C–H bonds, the center furthest removed from the electron-withdrawing moiety reacts preferentially (entries 3, 5, 6). Not surprisingly, based on the concerted nature of the C–H insertion event, hydroxylation of a starting material possessing a stereogenic C–H center (entries 5–7) typically gives a single stereoisomeric product.124 While reaction times are somewhat

123 a) Du Bois, J. Chemtracts: Org. Chem. 2005, 18, 1; b) Fiori, K. W.; Espino, C. G.; Brodsky, B. H.; Du Bois, J. “A mechanistic analysis of the Rh-catalyzed intramolecular C–H amination reaction.” Tetrahedron 2009, 65, 3042. 124 Only a single diastereomer is observed within the limits of 1H NMR.

114 A Catalytic Method for the Hydroxylation of C–H Bonds

prolonged, it is notable that the reaction can proceed for an extended period without observable catalyst or substrate decomposition. A reaction conducted for 48 h with the substrate appearing in entry 1 of Table 3.4 returned >90% of the benzoxathiazine catalyst and >80% of starting material is recovered. Stereospecific C–H hydroxylation of mono- and bicyclic ring systems to afford tetrasubstituted carbinol centers is a desirable transformation for complex molecule synthesis. Through judicious choice of protecting groups, positional selectivity in substrates such as menthol (entries 7 and 8, Table 3.4) can be controlled. Although the total conversion to product is efficient for the O-benzyl ester derivative, hydroxylation occurs indiscriminately at C3 and C7. In the former example, the steric demands of the t larger BuPh2Si group are presumably enough to steer the reaction towards C3. Oxidation at fused ring junctures is highlighted for both cis-fused 5,5 and 5,6 ring systems (entries 5, 6). Applications for this type of transformation are potentially high, as angular hydroxyl groups appear with some regularity as structural motifs in natural products. Finally, we note that the reaction conditions are tolerant to a number of common functional groups including esters, silyl ethers, sulfonylated amines, and carboxylic acids.

115 Chapter 3

Table 3.4. Substrate profile for reactions catalyzed by 3.16.

Entry Substrate Product Yield

Me OBz Me OBz 1 3.25 3.26 44 Me Me OH Me Me 2 OBz 3.27 OBz 3.28 75 Me Me OH

BzO 3 Me BzO Me 3 3.29 3.30 70b Me 7 Me Me OH Me t t O2C Bu O2C Bu 4 3.31 3.32 61 Me Me HO OTroc H H OTroc 5 3.33 3.34 33

H OH

H OBz H OBz 6 3.35 3.36 82

H OH

3 t Me OSi BuPh2 HO t OSi BuPh2 Me 7 7 Me 3.37 3.38 34 Me Me Me 3 Me OBz R1 OBz 1 2 c 8 7 Me 3.39 Me R = H, R = OH 3.40 47 R2 R1 = OH, R2 = H Me Me Me Me NHSO Ar 9 2 3.41 Me 3.42 66 Me NHSO2Ar Me OH Ar = o-C6H4NO2 O O O 10 OH 3.43 3.44 (50)

a Optimized reaction conditions: 20 mol% catalyst 3.16, 8 equiv 50% H2O2, 0.25 M 1:1 AcOH/H2O, 50 °C, 96 h. bIsolated yield after 48 h. cThe ratio of C3/C7 hydroxylation products is ~1:1. An additional 10–15% of the product resulting from benzoate migration to the C7–OH is also obtained. dProduct volatility accounts for some diminution in yield.

116 A Catalytic Method for the Hydroxylation of C–H Bonds

Aromatic substrates are conspicuously absent from Table 3.4. While electron- deficient aromatic substitutents are tolerated under the reaction conditions, electron- neutral and electron-rich aromatic moieties are not. Furthermore, no product was obtained in catalytic reactions with tertiary benzylic substrates, which, based on analogous results from nitrene and carbene literature,125 would be expected to be excellent substrates. To probe these surprising results, the stoichiometric reactions between oxaziridine 3.4 and s-butylbenzene and ethylanisole in DCE were conducted (Figure 3.18). The reaction was monitored periodically by 1H NMR over the course of 72 hours; not only was no product observed, but the oxaziridine had been completely reduced to benzoxathiazine. We believed there were two possible reasons for this: either the oxaziridine was catalyzing its own decomposition or the presence of the aromatic substrate was somehow initiating the reduction of the oxaziridine.

Me Me O O O O S S Me O N Me O N O DCE or + CF or + CF 3 22 ºC or 50 ºC 3 Me Me MeO Cl MeO Cl

Figure 3.18. Reaction of aromatic substrates with an oxaziridine.

To determine what was causing the decomposition of the oxaziridine, a simple

NMR experiment was conducted. A 0.2 M solution of oxaziridine 3.4 in CD2Cl2 was 1 prepared and a H NMR spectrum of 3.4 was obtained. After two hours in CD2Cl2, a new spectrum was acquired and its signals were identical to the previously acquired spectrum; this result indicates that the oxaziridine is not likely facilitating its own decomposition in solution. Following the acquisition of the second spectrum, 1 equiv of ethyl anisole was added to the solution, the NMR tube was vigorously shaken, and an arrayed NMR experiment was set up to acquire a NMR spectrum approximately every two minutes over the course of 20 minutes. These spectra depicted the rapid decomposition of oxaziridine to the corresponding benzoxathiazine as well as many other intractable products. When the reaction solution was removed from the NMR after 20 minutes, the solution that had started clear had become translucent orange. Although we still do not understand the

125 For general reviews, see: (a) Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic Methods for Organic Synthesis with DiazoCompounds: FromCyclopropanes to Ylides; Wiley and Sons: New York, NY,1997; (b) Davies, H. M. L.; Beckwith, E. J. Chem. Rev. 2003, 103, 2861–2904; (c) Doyle, M. P. J. Org. Chem. 2006, 71, 9253–9260. See also reference (18).

117 Chapter 3

nature of the decomposition, it is apparent that only small amounts of electron-rich aromatic substrate are necessary to catalyze the decomposition of oxaziridine.

3.3c Mechanistic Insights Based on previous work with dioxiranes117 and DFT calculations by us and others,126 the hydroxylation of enantiopure tertiary C–H bond substrates was expected to occur with complete stereoretention. What was observed, however, in the case of substrate 3.45, was that as the reaction proceeded, the enantiopurity of the product decreased (as monitored by HPLC). Furthermore, when enantiopure product 3.46 was resubmitted to the reaction conditions, the ee eroded to 80%, indicating that the tertiary alcohol is not stable to the acidic aqueous reaction conditions for long durations (Figure 3.19). Checking the enantiopurity of the product at early time points (< 24 h, ~20% conversion), however, confirms that the hydroxylation event itself is stereospecific as the product is isolated in >99% ee. Ionization of the alcohol product is substrate dependent and it can be expected that for substrates that cannot easily form a stable carbocation erosion of the stereochemistry should not be as problematic.

20 mol% 3.16 Me OBz Me OBz

Me 8 equiv. H2O2 Me OH >99% ee 1:1 AcOH:H2O, 50 ºC, 96h 90% ee 3.45

Me OBz 20 mol% 3.16 Me OBz Me OH Me OH 8 equiv. H2O2 >99% ee 1:1 AcOH:H2O, 50 ºC, 24h 80% ee 3.46

Figure 3.19. The enantiopurity of the product diminishes when resubjected to the reaction conditions.

We wished to understand the reasons for the improved performance of the pentafluorophenyl catalyst 3.16 over other catalysts tested. We briefly investigated competition reactions similar to those performed by Breslow; however, because aromatic substrates are not tolerated in the reaction conditions, it was very difficult to find a substrate that would benefit from π-stacking with the catalyst. Furthermore, while the results of competition experiments could be informative, the most useful data could be acquired from kinetics experiments. Specifically, we wanted to know if the rate of oxygen

126 See Chapter 1 for a discussion of DFT calculations regarding the concerted nature of the oxygen insertion event.

118 A Catalytic Method for the Hydroxylation of C–H Bonds

atom transfer with the oxaziridine derived from catalyst 3.16 is faster than that of 3.4 in aqueous acidic media. The σ values for the substituents on these two heterocycles are identical and the corresponding oxaziridines have been shown to give similar rates of oxygen transfer in DCE. If the rate of oxygen transfer in water was much faster for the oxaziridine derived from 3.16 as compared to 3.4, it could be concluded that hydrophobic effects were responsible for accelerating the oxidation even. This experiment would be analogous to Breslow’s original study of the Diels-Alder reaction in water (Figure 3.15). Unfortunately, carrying out this seemingly simple experiment proved challenging. Adapting the previously-established HPLC kinetics method to aqueous conditions did not work: the immiscibility of the reagents in aqueous solvent made the composition of the sampled aliquots unreliable. Furthermore, within a half hour, epoxide-opening by water became a significant problem. Efforts to identify a C–H hydroxylation substrate to replace the olefin substrate was thwarted by the need for a much higher reaction temperature and the inability to design a substrate that was both easily hydroxylated and soluble in aqueous conditions. Although these complications proved frustrating, we are optimistic that an experiment could be designed in the future that would allow for the measurement of the desired rate constants. Regardless, the data presented in this chapter suggest the hydrophobic effect is responsible for the enhanced performance of catalyst 3.16.

3.4 Conclusions A unique catalytic process for C–H hydroxylation has been advanced from a proof-of-principle to that of an effective tool for chemical synthesis. In doing so, aqueous

H2O2 conditions have been identified that greatly simplify the experimental protocol and facilitate the large-scale application of this method. The stereospecificity and the predictable site-selectivity exhibited by this reaction make it one of the more generally applicable catalytic C–H hydroxylation methods reported to date. Our findings suggest a possible strategy for promoting kinetically slow C–H hydroxylation events through the hydrophobic aggregation of catalyst and substrate. Considerable experimentation needs to be done in order to discern the true nature of the greater conversions displayed by catalyst 3.16. Results intimating that this reaction benefits from hydrophobic aggregation should encourage future researchers to exploit secondary catalyst-substrate interactions to optimize reaction performance; this strategy is inspired by reactions that in occur in Nature that are catalyzed in the presence of an enzyme binding pocket. While enzyme active sites have long been a source of inspiration

119 Chapter 3

for the design of catalysts for C–H hydroxylation, perhaps enzyme binding pockets will guide the design of more effective catalytic reactions.

120 A Catalytic Method for the Hydroxylation of C–H Bonds

Experimental Section

General. All reagents were obtained commercially unless otherwise noted. Moisture-sensitive reactions were performed using flame-dried glassware under an atmosphere of dry nitrogen. Air- and moisture-sensitive liquids and solutions were transferred via syringe or stainless steel cannula. Organic solutions were concentrated under reduced pressure (ca. 15 Torr) by rotary evaporation. Dichloroethane (DCE) was distilled from CaH2 immediately prior to use. Dichloromethane, tetrahydrofuran (THF), acetonitrile, and N,N–Dimethylformamide (DMF) were dried by passage under 12 psi N2 through columns containing activated alumina. N,N–Dimethyl-acetamide (DMA) was dried over activated 4Å molecular sieves. Chromatographic purification of products was accomplished using forced-flow chromatography on Silicycle Ultra Pure Silica Gel Silia-P (40-63 mm). Compounds purified by chromatography on silica gel were typically applied to the adsorbent bed using the indicated solvent conditions with a minimum amount of added chloroform as needed for solubility. Thin layer chromatography was performed on

EM Science silica gel 60 F254 plates (250 mm). Visualization of the developed chromatogram was accomplished by fluorescence quenching and by staining with ethanolic anisaldehyde, aqueous potassium permanganate, or aqueous ceric ammonium molybdate (CAM) solution. HPLC analyses were conducted using a Hewlett Packard Series 1100 Chemstation. Samples were analyzed using a Zorbax CN normal phase analytical column (4 x 250 mm) with hexanes/i-PrOH (0.5% to 50% gradient elution) as eluent. Optical absorption was determined at 254 nm. NMR spectra were acquired on a Varian Mercury-400 operating at 400 and 100 MHz or a Varian Inova-500 operating at 500 and 125 MHz for 1H and 13C, respectively, and are referenced internally according to residual solvent signals. Data for 1H NMR are recorded as follows: chemical shift (d, ppm), multiplicity (s, singlet; d, doublet, t, triplet; q, quartet; m, multiplet), integration, coupling constant (Hz). Data for 13C are reported in terms of chemical shift (d, ppm). Infrared spectra were recorded as thin films using NaCl salt plates on a Perkin-Elmer Paragon 500 FTIR spectrometer or a Thermo-Nicolet IR300 spectrometer and are reported in frequency of absorption. High-resolution mass spectra were obtained the Vincent Coates Foundation Mass Spectrometry Laboratory, Stanford University.

121 Chapter 3

Experimental protocols for benzoxathiazine catalyst preparation

Benzoxathiazines and the corresponding oxaziridines were prepared using the same protocols described in Chapter 2 unless otherwise noted.

O O S O N

3.10red To facilitate ring-closure, the ring-open precursor was stirred over night with Amberlyst-15 in toluene at 80 ºC. Purified by chromatography on silica gel using gradient 1 elution, 4:1→2:1 hexanes/EtOAc (white solid). TLC Rf = 0.19 (20% EtOAc/hexanes); H

NMR (CDCl3, 500 MHz) δ 7.80-7.73 (m, 3H), 7.70-7.63 (m, 2H), 7.59-7.55 (m, 2H), 7.42- 13 7.37 (m, 2H) ppm; C NMR (CDCl3, 125 MHz) δ 176.4, 154.5, 136.9, 133.6, 133.6, 133.2, 131.7, 130.5, 128.8, 125.7, 119.4, 116.5 ppm; IR (thin film) ν 1412, 1211, 1179, 763, 609 cm-1.

O O S O N O Ph

3.10 Purified by chromatography on silica gel using 4:1 hexanes/EtOAc (white solid).

Note: purified material decomposes upon standing. TLC Rf = 0.39 (20% EtOAc/hexanes); 1 13 H NMR (CDCl3, 400 MHz) δ 7.60-7.51 (m, 6H), 7.29-7.19 (m, 3H) ppm; C NMR (CDCl3, 100 MHz) δ 150.0, 133.4, 133.2, 131.0, 130.7, 128.8, 127.2, 126.7, 119.4, 118.6, 87.6 ppm.

O O S O N

C6F5

3.11red To facilitate ring-closure, the ring-open precursor was stirred over night with Amberlyst-15 in toluene at 80 ºC. Purified by chromatography on silica gel using 4:1

122 A Catalytic Method for the Hydroxylation of C–H Bonds

1 hexanes/EtOAc (white solid, 78%). TLC Rf = 0.28 (20% EtOAc/hexanes); H NMR 13 (CDCl3, 500 MHz) δ 7.86-7.81 (m, 1H), 7.44-7.38 (m, 1H) ppm; C NMR (CDCl3, 125 MHz) δ 165.7, 153.9, 144.5 (dm, J = 251 Hz), 143.5 (dm, J = 251 Hz), 138.4, 137.9 (dm, J = 251 Hz), 129.6, 126.5, 119.4, 116.0, 108.8 (m) ppm; IR (thin film) ν 1595, 1502, 1400, 1346, 1191, 995 cm-1.

O O S O N O

C6F5

3.11 Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (white solid, 1 43%). TLC Rf = 0.39 (20% EtOAc/hexanes); H NMR (CDCl3, 500 MHz) δ 7.69 (t, 1H, J = 7.6 Hz), 7.39 (t, 1H, J = 7.7 Hz), 7.31 (d, 1H, J = 8.3 Hz), 7.22 (d, 1H, J = 7.7 Hz) ppm; 13 C NMR (CDCl3, 125 MHz) δ 149.1, 145.0 (dm, J = 254 Hz), 144.4 (dm, J = 254 Hz), 143.2 (dm, J = 254 Hz), 137.7 (dm, J = 254 Hz), 134.6, 131.0, 127.7, 120.0, 116.2, 106.9 (m), 81.2 ppm; IR (thin film) ν 1511, 1425, 1341, 1212, 1179, 1100, 998, 885 cm-1.

OMe OH

CF3

Br 1-(5-Bromo-2-methoxyphenyl)-2,2,2-trifluoroethanol. Prepared according to the general procedure of Sammakia.127 To an ice cold solution of 5-bromo-2- methoxybenzaldehyde (3.16-S1) (12.0 g, 55.8 mmol) and Me3SiCF3 (9.9 mL, 67.0 mmol, n 1.2 equiv) in 140 mL of THF was added 22.9 mL of Bu4NF•H2O (1.0 M in THF, 22.9 mmol, 0.41 equiv). The golden solution was warmed to ambient temperature and stirred for 10 h. Following this time, a 4.4 M aqueous solution of HCl (25.8 mL, 111.6 mmol, 2 equiv) was slowly added. The mixture was stirred for 1 h then diluted with 30 mL of

EtOAc, and solid Na2CO3 was cautiously added. Once effervescence had ceased, the solution was dried over MgSO4, filtered and concentrated under reduced pressure to an orange oily residue. Purification of this material by chromatography on silica gel (gradient elution, 10:1→4:1 hexanes/EtOAc) yielded the desire product as a golden crystalline solid

127 Wayman, K. A.; Sammakia, T. “O-nucleophilic amino alcohol acyl-transfer catalysts: the effect of acidity of the hydroxyl group on the activity of the catalyst.” Org. Lett. 2003, 5, 4105-4108.

123 Chapter 3

1 (11.0 g, 70%). TLC Rf = 0.14 (10:1 hexanes/EtOAc); H NMR (CDCl3, 400 MHz) δ 7.57 (d, 1H, J = 2.3 Hz), 7.45 (dd, 1H, J = 8.8, 2.5 Hz), 6.81 (d, 1H, J = 8.8 Hz), 5.33 (dq, 1H, J 13 = 6.7, 6.7 Hz), 3.85 (s, 3H), 3.76 (d, 1H, J = 6.0 Hz) ppm; C NMR (CDCl3, 125 MHz) δ

156.4, 133.2, 131.7, 124.3 (q, JCF = 285 Hz), 124.3, 113.2, 112.8, 67.9 (q, JCF = 33 Hz), 55.9 ppm; IR (thin film) ν 3455, 2945, 3844, 1490, 1257, 1175, 1140, 1063, 810 cm-1.

OMe O

CF3

3.16-S2 1-(5-Bromo-2-methoxyphenyl)-2,2,2-trifluoroethanone. To a solution of 1-(5- bromo-2-methoxyphenyl)-2,2,2-trifluoroethanol (11.0 g, 38.6 mmol) in 390 mL of CH2Cl2 was added 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO, 1.93 mmol, 301 mg, 0.05 equiv) followed by PhI(OAc)2 (24.8 g, 77.2 mmol, 2.0 equiv). The mixture was stirred for

13 h and quenched by the addition of 250 mL of 1.0 M aqueous Na2S2O3. The mixture was transferred to a separatory funnel and the organic layer was collected. The aqueous layer was extracted with 3 x 200 mL of CH2Cl2. The organic fractions were combined, dried over Na2SO4, filtered, and concentrated under reduced pressure to yield an oily, peach-colored residue that solidifies upon standing. Purification of this material by chromatography on silica gel using 7:1 hexanes/EtOAc as eluent afforded the desired 1 product as a yellow crystalline solid (9.0 g, 82%). TLC Rf = 0.39 (4:1 hexanes/EtOAc); H

NMR (CDCl3, 500 MHz) δ 7.75 (d, 1H, J = 2.5 Hz), 7.69 (dd, 1H, J = 8.9, 2.5 Hz), 6.94 (d, 13 1H, J = 8.8 Hz), 3.91 (s, 3H) ppm; C NMR (CDCl3, 125 MHz) δ 182.0 (q, JCF = 37 Hz),

158.6, 138.2, 133.3, 123.1, 113.9, 115.8 (q, JCF = 290 Hz), 112.7, 56.0 ppm; IR (thin film) ν 3545, 2946, 3849, 1718, 1593, 1281, 1169, 1022, 965, 816 cm-1.

124 A Catalytic Method for the Hydroxylation of C–H Bonds

OMe O

CF3

F F

F F F 3.16-S3 2,2,2-Trifluoro-1-(2',3',4',5',6'-pentafluoro-4-methoxybiphenyl-3-yl)ethanone. Prepared according to the general procedure of Fagnou.128 A 10 mL round bottom flask was charged with solid K2CO3 (1.51 g, 10.93 mmol, 1.1 equiv), t-Bu2MeP•HBF4 (246 mg,

0.99 mmol, 0.1 equiv), Pd(OAc)2 (112 mg, 0.50 mmol, 0.05 equiv) and 3.16-S2 (2.81 g, 9.94 mmol). The vessel was equipped with a reflux condenser, and evacuated and backfilled with N2 three times. Neat pentafluorobenzene (1.7 mL, 14.9 mmol, 1.5 equiv) and 1.7 mL of deoxygenated DMA (sparged with N2 for 10–15 min prior to use) were added successively and the reaction was stirred at 120 °C for 12 h. The reaction was cooled to ambient temperature, diluted with 20 mL of EtOAc, and quenched by the addition of 5 mL of saturated aqueous NH4Cl and 20 mL of H2O. The contents were transferred to a separatory funnel, the organic layer was collected, and the aqueous fraction was extracted with 2 x 20 mL of EtOAc. The combined organic fractions were washed with 3 x 50 mL of H2O and 1 x 50 mL of saturated aqueous NaCl, dried over

MgSO4, filtered, and concentrated under reduced pressure to a brown amorphous solid. Purification of this material by chromatography on silica gel using 10:1 hexanes/EtOAc affored the desired product as a golden crystalline solid (1.95 g, 53%). TLC Rf = 0.33 (4:1 1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 7.75 (s, 1H) 7.68 (d, 1H, J = 8.7 Hz), 7.18 13 (d, 1H, J = 8.8 Hz), 4.00 (s, 3H) ppm; C NMR (CDCl3, 125 MHz) δ 182.3 (q, JCF = 37

Hz), 160.2, 144.1 (dm, JCF = 251 Hz), 140.6 (dm, JCF = 254 Hz), 137.9 (dm, JCF = 249

Hz), 137.3, 133.1, 122.1, 118.9, 116.0 (q, JCF = 290 Hz), 113.9 (m), 112.6, 56.1 ppm; IR (thin film) ν 2962, 1691, 1496, 1288, 1211, 1165, 1001, 984, 834 cm-1.

128 LaFrance, M.; Rowley, C. N.; Woo, T. K.; Fagnou, K. “Catalytic intermolecular direct arylation of perfluorobenzenes.” J. Am. Chem. Soc., 2006, 128, 8754-8756.

125 Chapter 3

OH O

CF3

F F

F F F

3.16-S4 2,2,2-Trifluoro-1-(2',3',4',5',6'-pentafluoro-4-hydroxybiphenyl-3-yl)ethanone.

To a solution of S3 (810 mg, 2.19 mmol) in 15 mL of CH2Cl2 cooled to –78 °C was added dropwise a 1.0 M CH2Cl2 solution of BBr3 (5.5 mL, 5.47 mmol, 2.5 equiv). The contents were warmed slowly warm to ambient temperature and stirred for 10 h. The mixture was then cooled to 0 ºC and carefully quenched by the slow addition of 15 mL of H2O. The biphasic mixture was transferred to a separatory funnel and extracted with 3 x 20 mL of

CH2Cl2. The combined organic extracts were dried over Na2SO4, filtered, and concentrated under reduced pressure to a brown oil. Purification of this material by chromatography on silica gel using 10:1 hexanes/EtOAc afforded the desired product as 1 an off-white solid (764 mg, 98%). TLC Rf = 0.53 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 11.20 (s, 1H), 7.91 (s, 1H) 7.70 (d, 1H, J = 8.9 Hz), 7.25 (d, 1H, J = 8.9 Hz) 13 ppm; C NMR (CDCl3, 125 MHz) δ 184.3 (q, JCF = 36 Hz), 164.9, 144.2 (dm, JCF = 245

Hz), 138.0 (dm, JCF = 256 Hz), 140.8 (dm, JCF = 253 Hz), 140.2, 132.6, 119.9, 118.2,

116.2 (q, JCF = 286 Hz), 115.0, 114.0, 113.8 (m) ppm; IR (thin film) ν 3203, 1669, 1524, 1501, 1202, 1156, 1072, 993 cm-1.

O O S O N

CF3

F F

F F F 3.16 6-(2’,3’,4’,5’,6’-Pentafluorophenyl)-4-trifluoromethyl-[1,2,3]-benzoxathiazine- 2,2-dioxide. Sulfamoylation of 2,2,2-trifluoro-1-(2',3',4',5',6'-pentafluoro-4- hydroxybiphenyl-3-yl)ethanone (570 mg, 1.60 mmol) was performed as described above (see, General procedure for benzoxathiazine preparation, Chapter 2). Purified by silica gel chromatography using gradient elution, 4:1→1:1 hexanes/EtOAc (pale yellow 1 crystalline solid, 430 mg, 64%); TLC Rf = 0.10 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500

126 A Catalytic Method for the Hydroxylation of C–H Bonds

MHz) δ 8.01 (s, 1H), 7.94 (d, 1H, J = 8.7 Hz), 7.59 (d, 1H, J = 8.7 Hz) ppm; 13C NMR

(CDCl3, 125 MHz) δ 161.6 (q, JCF = 38 Hz), 155.1, 144.1 (dm, JCF = 245 Hz), 141.5 (dm,

JCF = 257 Hz), 140.2, 138.1 (dm, JCF = 253 Hz), 130.2, 125.5, 120.3, 118.1 (q, JCF = 283 Hz), 112.0 (m), 111.8 ppm; IR (thin film) ν 2926, 1620, 1524, 1502, 1417, 1210, 1168, 1070, 996 cm-1.

O O S O N

CF3

Ph 3.17 6-Phenyl-4-trifluoromethyl-[1,2,3]-benzoxathiazine-2,2-dioxide. Purified by chromatography on silica gel using 2:1 hexanes/EtOAc (yellow solid, 67%). TLC Rf =0.19 1 (2:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.09-8.08 (m, 1H), 8.08-8.06 (m, 1H), 13 7.59-7.49 (m, 6H) ppm; C NMR (CDCl3, 125 MHz) δ 162.2 (q, JCF = 37 Hz), 154.0,

140.5, 137.6, 137.5, 129.4, 128.9, 127.1, 126.5, 119.9, 118.3 (q, JCF = 281 Hz), 111.7 ppm; IR (thin film) ν 1617, 1569, 1414, 1211, 1161, 980 cm-1.

O O S O N

CF3

CF3 3.18 6-Trifluoromethyl-4-trifluoromethyl-[1,2,3]-benzoxathiazine-2,2-dioxide. (Purified by chromatography on silica gel using 1:1 hexanes/EtOAc (white solid, 79%). 1 TLC Rf = 0.12 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.19 (s, 1H), 8.13 (d, 13 1H, J = 9.0 Hz), 7.60 (d, 1H, J = 8.8 Hz) ppm; C NMR (CDCl3, 125 MHz) δ 161.3 (q, JCF

= 38 Hz), 156.9, 135.7, 129.5 (q, JCF = 34 Hz), 125.8, 122.3 (q, JCF = 272 Hz), 120.8,

118.0 (q, JCF = 282 Hz), 111.5; IR (thin film) ν 3090, 1629, 1583, 1399, 1173, 1132, 987, 845 cm-1.

127 Chapter 3

O O S O N

CF3

NO2 3.24 6-(4’-Nitrophenyl)-4-trifluoromethyl-[1,2,3]-benzoxathiazine-2,2-dioxide. (Purified by chromatography on silica gel using 2:1 hexanes/EtOAc (pale yellow solid, 1 65%). TLC Rf = 0.17 (2:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.41-8.39 (m, 13 2H), 8.10-8.08 (m, 2H), 7.75-7.73 (m, 2H), 7.60-7.58 (m, 2H) ppm; C NMR (CDCl3, 125

MHz) δ 161.9 (q, JCF = 38 Hz), 155.0, 148.1, 143.8, 138.0, 137.4, 128.2, 127.0, 124.6,

120.5, 118.3 (q, JCF = 278 Hz), 112.0 ppm; IR (thin film) ν 2926, 1616, 1522, 1414, 1348, 1211, 1162, 981, 837, 803 cm-1.

General procedure for C–H hydroxylation. To a solution of substrate (0.40 mmol), catalyst 3.16 (33 mg, 80.0 µmol, 0.20 equiv), and Na2EDTA (1.4 mg, 4.0 µmol,

0.01 equiv) in 1.6 mL of a 1:1 acetic acid/H2O mix was added 50% aqueous H2O2 (181 µL, 3.20 mmol, 8.0 equiv). The reaction vessel was fitted with a reflux condenser and placed in an oil bath pre-heated to 50 °C. The contents were stirred vigorously at this temperature for 96 h, then cooled to ambient temperature and transferred to a separatory funnel containing 7 mL of 1.0 M aqueous Na2S2O3 and 10 mL of EtOAc. The organic layer was collected and the aqueous layer was extracted with 4 x 4 mL of EtOAc. The combined organic extracts were washed with 1 x 10 mL of saturated aqueous NaCl, dried over MgSO4, filtered, and concentrated under reduced pressure. Purification by chromatography on silica gel (conditions given below) furnished the desired product.

Me OBz Me OH

3.28 4-Hydroxy-4-methylpentyl benzoate. Purified by chromatography on silica gel using 4:1 hexanes/EtOAc (clear oil, 75%); TLC Rf = 0.17 (4:1 hexanes/EtOAc); H NMR

(CDCl3, 400 MHz) δ 8.05-8.01 (m, 2H), 7.56-7.52 (m, 1H), 7.45-7.40 (m, 2H), 4.33 (t, 2H, J = 2.8 Hz), 1.90-1.82 (m, 2H), 1.70-1.58 (br s, 1H), 1.62-1.58 (m, 2H) 1.25 (s, 6H); 13C

NMR (CDCl3, 125 MHz) δ 166.8, 133.1, 130.5, 129.7, 128.5, 70.8, 65.6, 40.1, 29.5, 24.0

128 A Catalytic Method for the Hydroxylation of C–H Bonds

ppm; IR (thin film) ν 3418, 2969, 1718, 1452, 1279, 1116, 713 cm-1; HRMS (ES+) calcd for C13H18O3 222.1256 found 245.1144 (MNa+).

Me OBz Me OH

3.21 3-Hydroxy-3-methylbutyl benzoate. Purified by chromatography on silica gel 1 using 4:1 hexanes/EtOAc (clear oil, 44%); TLC Rf = 0.17 (4:1 hexanes/EtOAc); H NMR

(CDCl3, 400 MHz) δ 8.05-8.02 (m, 2H), 7.57-7.55 (m, 1H), 7.47-7.43 (m, 2H), 4.51 (t, 2H, 13 J = 6.7 Hz), 2.00 (t, 2H, J = 6.7), 1.70 (br s, 1H), 1.34 (s, 3H) ppm; C NMR (CDCl3, 125 MHz) δ 166.6, 132.9, 130.2, 129.5, 128.4, 70.1, 61.9, 41.7, 29.7 ppm; IR (thin film) ν -1 + 3447, 2971, 1719, 1278, 1116, 711 cm ; HRMS (ES ) calcd for C12H16O3 208.1099 found 231.1002 (MNa+).

BzO Me

Me Me OH

3.30

(±)-7-Hydroxy-3,7-dimethyloctyl benzoate. Purified by chromatography on 1 silica gel using 5:1 hexanes/EtOAc (clear oil, 70%); H NMR (CDCl3, 400 MHz) δ8.04- 8.00 (m, 2H), 7.55-7.50 (m, 2H), 7.44-7.39 (m, 2H), 4.39-4.29 (m, 2H), 1.84-1.76 (m, 2H), 1.69-1.52 (m, 3H), 1.46-1.30 (m, 5H), 1.19 (s, 6H), 0.95 (d, 3H, J = 6.4 Hz) ppm; 13C NMR

(CDCl3, 125 MHz) δ 166.8, 133.0, 130.6, 129.7, 128.5, 71.0, 63.6, 44.2, 37.5, 35.7, 30.1, 29.4, 21.7, 19.7 ppm; IR (thin film) ν 3418, 2965, 2936, 1720, 1453, 1279, 1315, 1277, -1 + 1176, 1115, 1027, 939 cm ; HRMS (ES ) calcd for C17H26O3 278.1882 found 301.1775 (MNa+).

t O2C Bu Me HO 3.32 (±)-4-Hydroxy-4-methylcyclohexyl pivalate. Purified by chromatography on silica gel using 3:1 hexanes/EtOAc (white solid, 61%); TLC Rf = 0.24 (3:1 1 hexanes/EtOAc); H NMR (CDCl3, 400 MHz) δ 4.94-4.89 (m, 1H), 1.92-1.84 (m, 2H), 13 1.70-1.60 (m, 4H), 1.52-1.46 (m, 2H), 1.26 (s, 3H), 1.18 (s, 9H) ppm; C NMR (CDCl3,

129 Chapter 3

125 MHz) δ 178.1, 69.5, 69.4, 39.1, 34.6, 30.5, 27.4, 26.5 ppm; IR (thin film) ν 3426, 2961, 2933, 1727, 1481, 1398, 1371, 1286, 1174, 1136, 1011, 968, 907 cm-1; HRMS + + (ES ) calcd for C12H22O3 214.1569 found 237.1472 (MNa ).

H OTroc

OH 3.34 (±)-3a-Hydroxyoctahydropentalen-1-yl 2,2,2-trichloroethyl carbonate. Purified by chromatography on silica gel using 4:1 hexanes/EtOAc (colorless oil, 33%); 1 TLC Rf = 0.12 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 5.19 (ddd, 1H, J = 7.0, 7.0, 7.0 Hz), 4.83 (d, 1H J = 11.9 Hz), 4.73 (d, 1H J = 11.9 Hz), 2.52 (ddd, 1H, J = 7.8, 7.8, 7.8 Hz), 2.21-2.05 (m, 1H), 1.95-1.85 (m, 2H), 1.83-1.69 (m, 5H), 1.61-1.54 (m, 2H) 13 ppm; C NMR (CDCl3, 125 MHz) δ 153.6, 90.2, 81.2, 76.6, 54.9, 42.4, 37.4, 31.5, 26.7, 26.3 ppm; IR (thin film) ν 3402, 2958, 1756, 1250, 991 cm-1; HRMS (ES+) calcd for + C11H15Cl3O4 316.0036 found 338.9917 (MNa ).

H OBz

OH 3.36

(±)-3a-Hydroxyoctahydro-1H-inden-1-yl benzoate. Purified by chromatography on silica gel using 7:1 hexanes/EtOAc (colorless oil, 82%); TLC Rf = 0.20 (4:1 1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz, single diastereomer) δ 8.04-8.02 (m, 2H), 7.57-7.53 (m, 1H), 7.45-7.42 (m, 2H), 5.72 (dt, 1H, J = 9.2, 6.4, 6.4 Hz), 2.55-2.47 (m, 13 1H), 2.19-1.16 (m, 12H) ppm; C NMR (CDCl3, 125 MHz) δ 166.3, 132.8, 130.4, 129.5, 128.3, 80.8, 78.2, 49.5, 36.5, 32.0, 27.4, 24.2, 24.1, 23.6 ppm; IR (thin film) ν 3469, 2933, -1 + 2860, 1716, 1451, 1279, 1120 cm ; HRMS (ES ) calcd for C16H20O3 260.3282 found 283.1298 (MNa+).

HO t OSi BuPh2 Me Me

Me 3.38

(±)-3-(tert-Butyldiphenylsilyloxy)-4-isopropyl-1-methylcyclohexanol. Purified by chromatography on silica gel using 15:1 hexanes/EtOAc (colorless solid, 34%); TLC Rf

130 A Catalytic Method for the Hydroxylation of C–H Bonds

1 = 0.15 (15:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 7.74-7.68 (m, 4H). 7.44-7.37 (m, 6H), 3.78 (ddd, 1H, J = 10.0, 10.0, 4.3 Hz), 2.40-2.34 (m, 1H), 1.70-1.66 (m, 1H), 1.55-1.43 (m, 2H), 1.38 (dd, 1H, J = 13.2, 10.6 Hz), 1.29-1.19 (m, 4H), 1.07 (s, 3H), 1.06 13 (s, 9H), 0.89 (d, 3H, J = 7.1 Hz), 0.57 (d, 3H, J = 6.8 Hz) ppm; C NMR (CDCl3, 125 MHz) δ 135.98, 135.97, 135.2, 134.0, 129.6, 129.3, 127.5, 127.3, 71.3, 70.7, 50.2, 48.3, 37.7, 31.4, 27.0, 25.0, 21.3, 19.4, 18.5, 15.8 ppm; IR (thin film) ν 3435, 3071, 2958, 2931, -1 + 2857, 1428, 1110, 1072, 703 cm ; HRMS (ES ) calcd for C26H38O2Si 410.2641 found 433.2545 (MNa+).

Me OBz

Me OH Me

3.40A

(±)-2-(2-Hydroxypropan-2-yl)-5-methylcyclohexyl benzoate. Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (colorless oil, 21%); TLC Rf = 1 0.35 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.04-8.02 (m, 2H), 7.60-7.56 (m, 1H), 7.47-7.43 (m, 2H), 5.11 (ddd, 1H, J = 10.8, 10.8, 4.4 Hz), 2.76 (br s, 1H), 2.19-2.16 (m, 1H), 2.00-1.96 (m, 1H), 1.90-1.84 (m, 1H), 1.78-1.75 (m, 1H), 1.68-1.57 (m, 1H), 1.20 13 (s, 3H), 1.18 (s, 3H), 1.26-0.98 (m, 3H), 0.94 (d, 3 H, J = 6.5 Hz) ppm; C NMR (CDCl3, 125 MHz) δ 165.3, 133.2, 130.2, 129.4, 128.6, 76.6, 73.0, 51.7, 41.0, 34.3, 31.4, 28.4, 27.1, 26.1, 21.7 ppm; IR (thin film) ν 3443(br), 2960, 2873, 1715, 1697, 1275, 1116, 711 -1 + + cm ; HRMS (ES ) calcd for C17H24O3 276.1725 found 299.1621 (MNa ).

HO OBz Me Me

3.40B (±)-5-Hydroxy-2-isopropyl-5-methylcyclohexyl benzoate. Purified by chromatography on silica gel using 10:1 hexanes/EtOAc (colorless oil, 26%); TLC Rf = 1 0.29 (4:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.07-8.04 (m, 2H), 7.58-7.54 (m, 1H), 7.47-7.43 (m, 2H), 5.28 (ddd, 1H, J = 10.3, 10.3, 4.6 Hz), 2.23-2.17 (m, 1H), 2.04- 1.97 (m, 1H), 1.76-1.72 (m, 1H), 1.64-1.57 (m, 2H), 1.52 (dd, 1H, J = 12.9, 11.1 Hz), 1.44- 1.34 (m, 1H), 1.3 (s, 3H), 0.96 (d, 3H, J = 7.0 Hz), 085 (d, 3H, J = 7.0 Hz) ppm; 13C NMR

(CDCl3, 125 MHz) δ 166.0, 132.7, 129.5, 128.4, 128.3, 72.6, 71.2, 47.2, 44.7, 37.8, 31.4,

131 Chapter 3

26.5, 20.8, 19.4, 16.6 ppm; ; IR (thin film) ν 3449(br), 2960, 2873, 1715, 1697, 1275, -1 + + 1116, 711 cm ; HRMS (ES ) calcd for C17H24O3 276.1725 found 299.1623 (MNa ).

Me NHSO2Ar Me OH

3.42 N-(4-Hydroxy-4-methylpentyl)-2-nitrobenzenesulfonamide. Purified by chromatography on silica gel using 1:1 hexanes/EtOAc (colorless oil, 66%); TLC Rf = 0.17 1 (1:1 hexanes/EtOAc); H NMR (CDCl3, 500 MHz) δ 8.14-8.12 (m, 2H), 7.86-7.84 (m, 1H), 7.61-7.72 (m, 2H), 5.63 (t, 2H, J = 5.9 Hz), 3.13 (dt, 2H, J = 7.2, 6.4 Hz), 1.67-1.61 (m, 13 2H), 1.50-1.46 (m, 2H), 1.18 (s, 6H) ppm; C NMR (CDCl3, 125 MHz) δ 148.1, 133.7, 133.5, 132.7, 131.0, 125.3, 70.5, 44.2, 40.1, 29.3, 24.5 ppm; IR (thin film) ν 3536, 3349, -1 + 3099, 2971, 1542, 1365, 1339, 1166 cm ; HRMS (ES ) calcd for C12H18N2O5S 302.0936 found 325.0825 (MNa+).

132 A Catalytic Method for the Hydroxylation of C–H Bonds

O O S O N

3.10red

133 Chapter 3

O O S O N O Ph

3.10

134 A Catalytic Method for the Hydroxylation of C–H Bonds

O O S O N

C6F5

3.11red

135 Chapter 3

O O S O N O

C6F5

3.11

136 A Catalytic Method for the Hydroxylation of C–H Bonds

MeO H O CF3 H

3.13

137 Chapter 3

OH O

CF3

3.15

138 A Catalytic Method for the Hydroxylation of C–H Bonds

OMe O H

Br 3.16-S1

139 Chapter 3

OMe O

CF3

3.16-S2

140 A Catalytic Method for the Hydroxylation of C–H Bonds

OMe O

CF3

F F

F F F 3.16-S3

141 Chapter 3

OH O

CF3

F F

F F F

3.16-S4

142 A Catalytic Method for the Hydroxylation of C–H Bonds

O O S O N

CF3

F F

F F F 3.16

143 Chapter 3

O O S O N

CF3

NO2

144 A Catalytic Method for the Hydroxylation of C–H Bonds

Me OBz Me OH

3.21

145 Chapter 3

H OTroc

OH 3.34

146 A Catalytic Method for the Hydroxylation of C–H Bonds

H OBz

OH 3.36

147 Chapter 3

Me OBz Me OH Me

3.40A

148 A Catalytic Method for the Hydroxylation of C–H Bonds

HO OBz Me Me

3.40B

149 Chapter 3

Me OBz Me

3.45

150 A Catalytic Method for the Hydroxylation of C–H Bonds

151 X-Ray Crystal Structure Data

Appendix A. X-ray Crystallographic Data for Chapter 2

X-ray Structure Report

for

2.7

Dept. of Chemistry, Stanford University

650 724 4558

[email protected]

Solved by Allen Oliver

Wednesday, June 03, 2009

152 Appendix B

DISCUSSION

The compound crystallizes as colorless block-like crystals. There are eight molecules of the compound in the unit cell of the primitive, centrosymmetric space group Pbca.

The structure of the compound is as expected. Due to the centrosymmetric nature of the space group the compound is a racemic mixture. The bond distances and angles are all as would be expected. The bond angles about the epoxy oxygen are by necessity very constrained (C7-O1-N1 angle = 58.50(18)°).

CRYSTAL SUMMARY

Crystal data for C11H14BrN2O3P; Mr = 333.12; orthorhombic; space group Pbca; a = 14.785(2) Å; b = 8.7316(13) Å; c = 20.268(3) Å; α = 90°; β = 90°; γ = 90°; V = 2616.6(7) 3 -1 - Å ; Z = 8; T = 150(2) K; λ(Mo-Kα) = 0.71073 Å; µ(Mo-Kα) = 3.266 mm ; dcalc = 1.691g.cm 3 ; 18743 reflections collected; 2695 unique (Rint = 0.0656); giving R1 = 0.0348, wR2 = 0.0694 for 1792 data with [I>2σ(I)] and R1 = 0.0691, wR2 = 0.0809 for all 2695 data. Residual electron density (e–.Å-3) max/min: 0.394/-0.437.

An arbitrary sphere of data were collected on a colorless block-like crystal, having approximate dimensions of 0.19 × 0.10 × 0.05 mm, on a Bruker Kappa X8-APEX-II diffractometer using a combination of ω- and φ-scans of 0.3°. Data were corrected for absorption and polarization effects and analyzed for space group determination. The structure was solved by direct methods and expanded routinely. The model was refined by full-matrix least-squares analysis of F2 against all reflections. All non-hydrogen atoms were refined with anisotropic thermal displacement parameters. Unless otherwise noted, hydrogen atoms were included in calculated positions. Thermal parameters for the hydrogens were tied to the isotropic thermal parameter of the atom to which they are bonded (1.5 × for methyl, 1.2 × for all others).

153 X-Ray Crystal Structure Data

Table 1. Crystal data and structure refinement for 2.7.

Identification code 2.7

Empirical formula C11H14BrN2O3P

Formula weight 333.12

Temperature 150(2) K

Wavelength 0.71073 Å

Crystal system orthorhombic

Space group Pbca

Unit cell dimensions a = 14.785(2) Å α = 90°

b = 8.7316(13) Å β = 90°

c = 20.268(3) Å γ = 90°

Volume 2616.6(7) Å3

Z 8

Density (calculated) 1.691 g.cm-3

Absorption coefficient (µ) 3.266 mm-1

F(000) 1344

Crystal size 0.19 × 0.10 × 0.05 mm3

ω range for data collection 2.01 to 26.46°

Index ranges -18 ≤ h ≤ 18, -10 ≤ k ≤ 10, -25 ≤ l ≤ 25

Reflections collected 18743

Independent reflections 2695 [Rint = 0.0656]

Completeness to θ = 26.46° 99.7 %

Absorption correction numerical

Max. and min. transmission 0.9682 and 0.7353

Refinement method Full-matrix least-squares on F2

Data / restraints / parameters 2695 / 0 / 165

Goodness-of-fit on F2 1.000

Final R indices [I>2σ(I)] R1 = 0.0348, wR2 = 0.0694

154 Appendix B

R indices (all data) R1 = 0.0691, wR2 = 0.0809

Largest diff. peak and hole 0.394 and -0.437 e–.Å-3

155 X-Ray Crystal Structure Data

Table 2. Atomic coordinates and equivalent isotropic displacement parameters (Å2)

for 2.7. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

x y z U(eq)

Br(1) 0.23459(2) 0.99517(4) 0.43260(2) 0.032(1)

P(1) 0.39506(6) 0.58697(10) 0.15349(4) 0.022(1)

O(1) 0.48569(14) 0.5633(3) 0.26344(10) 0.025(1)

O(2) 0.41865(15) 0.4333(3) 0.13090(11) 0.027(1)

O(3) 0.30648(13) 0.5890(2) 0.19884(10) 0.023(1)

N(1) 0.48060(17) 0.6607(3) 0.20091(13) 0.023(1)

N(2) 0.37854(18) 0.7108(3) 0.09554(13) 0.026(1)

C(1) 0.2926(2) 0.6864(4) 0.25194(15) 0.019(1)

C(2) 0.2045(2) 0.7137(4) 0.27051(16) 0.023(1)

C(3) 0.1875(2) 0.8077(4) 0.32377(16) 0.026(1)

C(4) 0.2587(2) 0.8730(4) 0.35736(15) 0.024(1)

C(5) 0.3474(2) 0.8472(4) 0.33856(16) 0.023(1)

C(6) 0.3650(2) 0.7515(3) 0.28531(16) 0.020(1)

C(7) 0.4590(2) 0.7205(4) 0.26564(16) 0.023(1)

C(8) 0.4232(2) 0.6877(4) 0.03063(16) 0.031(1)

C(9) 0.5230(2) 0.7234(4) 0.03017(19) 0.039(1)

C(10) 0.3249(3) 0.8523(4) 0.10294(19) 0.040(1)

156 Appendix B

C(11) 0.3795(3) 0.9966(4) 0.10026(19) 0.044(1)

H(2A) 0.1560 0.6682 0.2469 0.027

H(3A) 0.1270 0.8273 0.3372 0.031

H(5A) 0.3957 0.8945 0.3618 0.027

H(7A) 0.5050 0.7920 0.2846 0.027

H(8A) 0.3928 0.7533 -0.0025 0.037

H(8B) 0.4146 0.5798 0.0170 0.037

H(9A) 0.5471 0.7081 -0.0144 0.059

H(9B) 0.5543 0.6553 0.0610 0.059

H(9C) 0.5323 0.8301 0.0436 0.059

H(10A) 0.2926 0.8485 0.1457 0.048

H(10B) 0.2788 0.8555 0.0675 0.048

H(11A) 0.3397 1.0849 0.1068 0.066

H(11B) 0.4092 1.0045 0.0571 0.066

H(11C) 0.4255 0.9946 0.1351 0.066

157 X-Ray Crystal Structure Data

Table 3. Anisotropic displacement parameters (Å)2 for 2.7. The anisotropic

2 2 2 displacement factor exponent takes the form: -2π [ h a* U11 + ... + 2 h k a* b* U12 ]

U11 U22 U33 U23 U13 U12

Br(1) 0.0320(2) 0.0334(2) 0.0313(2) -0.0050(2) 0.0061(2) 0.0036(2)

P(1) 0.0187(4) 0.0229(5) 0.0247(5) 0.0011(4) -0.0006(4) 0.0020(4)

O(1) 0.0269(12) 0.0231(12) 0.0256(13) -0.0005(10) -0.0030(10) 0.0065(11)

O(2) 0.0321(13) 0.0213(12) 0.0287(13) -0.0018(11) -0.0022(10) 0.0043(11)

O(3) 0.0160(11) 0.0258(13) 0.0266(13) -0.0010(11) 0.0000(9) -0.0036(10)

N(1) 0.0207(14) 0.0217(15) 0.0269(16) 0.0017(13) -0.0024(12) 0.0009(12)

N(2) 0.0269(16) 0.0243(16) 0.0256(16) 0.0050(13) 0.0033(13) 0.0081(13)

C(1) 0.0198(16) 0.0159(17) 0.0202(18) 0.0003(14) -0.0039(14) -0.0015(14)

C(2) 0.0149(16) 0.0251(19) 0.0281(19) 0.0041(16) -0.0041(14) -0.0023(15)

C(3) 0.0152(17) 0.0311(19) 0.032(2) 0.0085(17) 0.0034(14) 0.0004(16)

C(4) 0.0284(18) 0.0213(18) 0.0215(18) 0.0031(14) 0.0026(14) 0.0026(15)

C(5) 0.0200(17) 0.0193(17) 0.029(2) 0.0020(15) -0.0020(15) -0.0011(15)

C(6) 0.0174(17) 0.0175(17) 0.0254(19) 0.0027(15) 0.0004(14) 0.0015(13)

C(7) 0.0171(17) 0.0231(19) 0.0272(19) 0.0009(15) -0.0005(15) 0.0006(14)

C(8) 0.038(2) 0.033(2) 0.020(2) -0.0015(17) -0.0019(16) 0.0015(18)

C(9) 0.041(2) 0.041(2) 0.035(2) -0.0021(19) 0.0111(18) -0.001(2)

C(10) 0.044(2) 0.037(2) 0.038(2) 0.0167(19) 0.0130(19) 0.019(2)

158 Appendix B

C(11) 0.067(3) 0.030(2) 0.035(2) -0.0014(19) 0.004(2) 0.010(2)

159 X-Ray Crystal Structure Data

Table 4. Bond lengths [Å] for 2.7.

atom-atom distance atom-atom distance

Br(1)-C(4) 1.895(3) P(1)-O(2) 1.460(2)

P(1)-O(3) 1.600(2) P(1)-N(2) 1.615(3)

P(1)-N(1) 1.714(3) O(1)-C(7) 1.429(4)

O(1)-N(1) 1.528(3) O(3)-C(1) 1.387(4)

N(1)-C(7) 1.448(4) N(2)-C(10) 1.476(4)

N(2)-C(8) 1.486(4) C(1)-C(2) 1.377(4)

C(1)-C(6) 1.387(4) C(2)-C(3) 1.379(5)

C(3)-C(4) 1.377(4) C(4)-C(5) 1.384(4)

C(5)-C(6) 1.390(4) C(6)-C(7) 1.471(4)

C(8)-C(9) 1.508(5) C(10)-C(11) 1.498(5)

C(2)-H(2A) 0.9500 C(3)-H(3A) 0.9500

C(5)-H(5A) 0.9500 C(7)-H(7A) 1.0000

C(8)-H(8A) 0.9900 C(8)-H(8B) 0.9900

C(9)-H(9A) 0.9800 C(9)-H(9B) 0.9800

C(9)-H(9C) 0.9800 C(10)-H(10A) 0.9900

C(10)-H(10B) 0.9900 C(11)-H(11A) 0.9800

C(11)-H(11B) 0.9800 C(11)-H(11C) 0.9800

160 Appendix B

Table 5. Bond angles [°] for 2.7.

atom-atom-atom angle atom-atom-atom angle

O(2)-P(1)-O(3) 112.70(13) O(2)-P(1)-N(2) 115.04(14)

O(3)-P(1)-N(2) 106.66(13) O(2)-P(1)-N(1) 110.17(13)

O(3)-P(1)-N(1) 106.13(12) N(2)-P(1)-N(1) 105.54(14)

C(7)-O(1)-N(1) 58.50(18) C(1)-O(3)-P(1) 124.97(19)

C(7)-N(1)-O(1) 57.33(18) C(7)-N(1)-P(1) 118.7(2)

O(1)-N(1)-P(1) 106.99(17) C(10)-N(2)-C(8) 116.3(3)

C(10)-N(2)-P(1) 124.6(2) C(8)-N(2)-P(1) 119.1(2)

C(2)-C(1)-O(3) 117.2(3) C(2)-C(1)-C(6) 121.7(3)

O(3)-C(1)-C(6) 121.1(3) C(1)-C(2)-C(3) 119.3(3)

C(4)-C(3)-C(2) 119.6(3) C(3)-C(4)-C(5) 121.4(3)

C(3)-C(4)-Br(1) 119.2(2) C(5)-C(4)-Br(1) 119.4(2)

C(4)-C(5)-C(6) 119.3(3) C(1)-C(6)-C(5) 118.7(3)

C(1)-C(6)-C(7) 121.4(3) C(5)-C(6)-C(7) 119.9(3)

O(1)-C(7)-N(1) 64.17(19) O(1)-C(7)-C(6) 116.5(3)

N(1)-C(7)-C(6) 121.4(3) N(2)-C(8)-C(9) 114.3(3)

N(2)-C(10)-C(11) 114.3(3) C(1)-C(2)-H(2A) 120.3

C(3)-C(2)-H(2A) 120.3 C(4)-C(3)-H(3A) 120.2

C(2)-C(3)-H(3A) 120.2 C(4)-C(5)-H(5A) 120.4

C(6)-C(5)-H(5A) 120.4 O(1)-C(7)-H(7A) 115.1

N(1)-C(7)-H(7A) 115.1 C(6)-C(7)-H(7A) 115.1

N(2)-C(8)-H(8A) 108.7 C(9)-C(8)-H(8A) 108.7

N(2)-C(8)-H(8B) 108.7 C(9)-C(8)-H(8B) 108.7

H(8A)-C(8)-H(8B) 107.6 C(8)-C(9)-H(9A) 109.5

C(8)-C(9)-H(9B) 109.5 H(9A)-C(9)-H(9B) 109.5

161 X-Ray Crystal Structure Data

C(8)-C(9)-H(9C) 109.5 H(9A)-C(9)-H(9C) 109.5

H(9B)-C(9)-H(9C) 109.5 N(2)-C(10)-H(10A) 108.7

C(11)-C(10)-H(10A) 108.7 N(2)-C(10)-H(10B) 108.7

C(11)-C(10)-H(10B) 108.7 H(10A)-C(10)-H(10B) 107.6

C(10)-C(11)-H(11A) 109.5 C(10)-C(11)-H(11B) 109.5

H(11A)-C(11)-H(11B) 109.5 C(10)-C(11)-H(11C) 109.5

H(11A)-C(11)-H(11C) 109.5 H(11B)-C(11)-H(11C) 109.5

162 Appendix B

Table 6. Torsion angles [°] for 2.7.

atom-atom-atom-atom angle atom-atom-atom-atom angle

O(2)-P(1)-O(3)-C(1) 145.2(2) N(2)-P(1)-O(3)-C(1) -87.6(3)

N(1)-P(1)-O(3)-C(1) 24.6(3) C(7)-O(1)-N(1)-P(1) -113.7(2)

O(2)-P(1)-N(1)-C(7) -127.6(2) O(3)-P(1)-N(1)-C(7) -5.4(3)

N(2)-P(1)-N(1)-C(7) 107.6(2) O(2)-P(1)-N(1)-O(1) -66.1(2)

O(3)-P(1)-N(1)-O(1) 56.20(19) N(2)-P(1)-N(1)-O(1) 169.16(17)

O(2)-P(1)-N(2)-C(10) 157.6(3) O(3)-P(1)-N(2)-C(10) 31.8(3)

N(1)-P(1)-N(2)-C(10) -80.8(3) O(2)-P(1)-N(2)-C(8) -25.6(3)

O(3)-P(1)-N(2)-C(8) -151.3(2) N(1)-P(1)-N(2)-C(8) 96.1(3)

P(1)-O(3)-C(1)-C(2) 157.1(2) P(1)-O(3)-C(1)-C(6) -23.8(4)

O(3)-C(1)-C(2)-C(3) 178.9(3) C(6)-C(1)-C(2)-C(3) -0.2(5)

C(1)-C(2)-C(3)-C(4) 0.2(5) C(2)-C(3)-C(4)-C(5) 0.5(5)

C(2)-C(3)-C(4)-Br(1) -177.4(2) C(3)-C(4)-C(5)-C(6) -1.1(5)

Br(1)-C(4)-C(5)-C(6) 176.8(2) C(2)-C(1)-C(6)-C(5) -0.4(5)

O(3)-C(1)-C(6)-C(5) -179.5(3) C(2)-C(1)-C(6)-C(7) 179.3(3)

O(3)-C(1)-C(6)-C(7) 0.3(5) C(4)-C(5)-C(6)-C(1) 1.0(5)

C(4)-C(5)-C(6)-C(7) -178.7(3) N(1)-O(1)-C(7)-C(6) 113.9(3)

P(1)-N(1)-C(7)-O(1) 92.5(2) O(1)-N(1)-C(7)-C(6) -106.6(3)

P(1)-N(1)-C(7)-C(6) -14.0(4) C(1)-C(6)-C(7)-O(1) -55.9(4)

C(5)-C(6)-C(7)-O(1) 123.8(3) C(1)-C(6)-C(7)-N(1) 18.6(5)

C(5)-C(6)-C(7)-N(1) -161.6(3) C(10)-N(2)-C(8)-C(9) 102.1(4)

P(1)-N(2)-C(8)-C(9) -75.1(4) C(8)-N(2)-C(10)-C(11) -65.5(4)

P(1)-N(2)-C(10)-C(11) 111.5(3)

163 X-Ray Crystal Structure Data

Appendix B. X-ray Crystallographic Data for Chapter 3

X-ray Structure Report

for

3.11

Department of Chemistry, Stanford University

650 724 4558

[email protected]

Solved by Allen Oliver

Thursday May 21, 2009

164 Appendix B

DISCUSSION

The compound crystallizes as colorless block-like prisma from a dichloromethane solution. There are four molecules of the compound in the centrosymmetric, monoclinic space group P21/n.

The structure of the compound is as expected. There are no unusual bond distances or angles within the molecule. Because the molecule crystallizes in a centrosymmetric space group, the crystals consist of a racemate of the compound.

CRYSTAL SUMMARY

Crystal data for C13H4F5NO4S; Mr = 365.23; monoclinic; space group P21/n; a = 7.1946(5) Å; b = 11.6848(8) Å; c = 15.4871(10) Å; α = 90°; β = 99.041(3)°; γ = 90°; V = 1285.79(15) 3 -1 - Å ; Z = 4; T = 150(2) K; λ(Mo-Kα) = 0.71073 Å; µ(Mo-Kα) = 0.339 mm ; dcalc = 1.887g.cm 3 ; 22738 reflections collected; 3234 unique (Rint = 0.0260); giving R1 = 0.0317, wR2 = 0.0866 for 2824 data with [I>2σ(I)] and R1 = 0.0371, wR2 = 0.0906 for all 3234 data. Residual electron density (e–.Å-3) max/min: 0.485/-0.401.

An arbitrary sphere of data were collected on a colorless block-like crystal, having approximate dimensions of 0.31 × 0.23 × 0.18 mm, on a Bruker Kappa X8-APEX-II diffractometer using a combination of ω- and φ-scans of 0.3°. Data were corrected for absorption and polarization effects and analyzed for space group determination. The structure was solved by direct methods and expanded routinely. The model was refined by full-matrix least-squares analysis of F2 against all reflections. All non-hydrogen atoms were refined with anisotropic thermal displacement parameters. Unless otherwise noted, hydrogen atoms were included in calculated positions. Thermal parameters for the hydrogens were tied to the isotropic thermal parameter of the atom to which they are bonded (1.5 × for methyl, 1.2 × for all others).

165 X-Ray Crystal Structure Data

Table 1. Crystal data and structure refinement for 3.11.

Identification code 3.11

Empirical formula C13H4F5NO4S

Formula weight 365.23

Temperature 150(2) K

Wavelength 0.71073 Å

Crystal system monoclinic

Space group P21/n

Unit cell dimensions a = 7.1946(5) Å α = 90°

b = 11.6848(8) Å β = 99.041(3)°

c = 15.4871(10) Å γ = 90°

Volume 1285.79(15) Å3

Z 4

Density (calculated) 1.887 g.cm-3

Absorption coefficient (µ) 0.339 mm-1

F(000) 728

Crystal size 0.31 × 0.23 × 0.18 mm3

ω range for data collection 2.19 to 28.42°

Index ranges -9 ≤ h ≤ 9, -15 ≤ k ≤ 15, -20 ≤ l ≤ 20

Reflections collected 22738

Independent reflections 3234 [Rint = 0.0260]

Completeness to θ = 28.42° 99.5 %

Absorption correction Numerical

Max. and min. transmission 1.0000 and 0.9201

Refinement method Full-matrix least-squares on F2

Data / restraints / parameters 3234 / 0 / 217

Goodness-of-fit on F2 1.033

Final R indices [I>2σ(I)] R1 = 0.0317, wR2 = 0.0866

166 Appendix B

R indices (all data) R1 = 0.0371, wR2 = 0.0906

Largest diff. peak and hole 0.485 and -0.401 e–.Å-3

167 X-Ray Crystal Structure Data

Table 2. Atomic coordinates and equivalent isotropic displacement parameters (Å2)

for 3.11. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

x y z U(eq)

S(1) 0.17584(6) 0.73404(3) 0.30267(2) 0.027(1)

F(1) -0.11854(12) 0.82490(8) 0.58829(6) 0.031(1)

F(2) -0.02045(13) 0.97463(8) 0.71963(6) 0.029(1)

F(3) 0.32716(14) 1.06859(8) 0.74530(6) 0.031(1)

F(4) 0.58409(12) 1.00483(8) 0.64386(6) 0.031(1)

F(5) 0.48605(12) 0.85628(8) 0.51132(6) 0.028(1)

O(1) 0.0576(2) 0.76802(10) 0.22561(7) 0.039(1)

O(2) 0.37346(18) 0.73659(10) 0.31015(8) 0.036(1)

O(3) 0.10345(15) 0.61161(9) 0.32409(6) 0.026(1)

O(4) -0.04727(14) 0.78136(9) 0.41654(7) 0.026(1)

N(1) 0.12929(17) 0.81833(10) 0.38595(8) 0.023(1)

C(1) 0.17696(19) 0.56689(12) 0.40718(9) 0.021(1)

C(2) 0.2257(2) 0.45368(12) 0.41221(10) 0.024(1)

C(3) 0.2913(2) 0.40864(12) 0.49425(10) 0.025(1)

C(4) 0.3075(2) 0.47669(13) 0.56791(10) 0.025(1)

C(5) 0.25881(19) 0.59188(12) 0.56085(9) 0.022(1)

C(6) 0.19012(19) 0.63827(12) 0.47977(9) 0.019(1)

168 Appendix B

C(7) 0.13017(19) 0.75990(12) 0.46911(9) 0.020(1)

C(8) 0.18029(19) 0.83916(11) 0.54499(9) 0.021(1)

C(9) 0.0545(2) 0.86998(12) 0.59989(9) 0.023(1)

C(10) 0.1026(2) 0.94641(12) 0.66726(9) 0.022(1)

C(11) 0.2805(2) 0.99385(12) 0.68068(9) 0.023(1)

C(12) 0.4103(2) 0.96300(12) 0.62851(9) 0.023(1)

C(13) 0.3592(2) 0.88638(12) 0.56157(9) 0.021(1)

H(2A) 0.2150 0.4074 0.3612 0.029

H(3A) 0.3254 0.3301 0.4998 0.030

H(4A) 0.3521 0.4447 0.6237 0.030

H(5A) 0.2727 0.6387 0.6116 0.026

169 X-Ray Crystal Structure Data

Table 3. Anisotropic displacement parameters (Å)2 for 3.11. The anisotropic

2 2 2 displacement factor exponent takes the form: -2π [ h a* U11 + ... + 2 h k a* b* U12 ]

U11 U22 U33 U23 U13 U12

S(1) 0.0419(2) 0.0204(2) 0.0192(2) -0.0013(1) 0.0043(2) -0.0029(1)

F(1) 0.0243(4) 0.0350(5) 0.0355(5) -0.0077(4) 0.0068(4) -0.0066(4)

F(2) 0.0331(5) 0.0295(5) 0.0266(4) -0.0025(4) 0.0104(4) 0.0049(4)

F(3) 0.0407(5) 0.0262(5) 0.0252(4) -0.0100(4) 0.0027(4) -0.0031(4)

F(4) 0.0258(5) 0.0333(5) 0.0309(5) -0.0037(4) 0.0002(4) -0.0088(4)

F(5) 0.0251(4) 0.0324(5) 0.0268(4) -0.0039(4) 0.0073(3) 0.0012(4)

O(1) 0.0669(9) 0.0277(6) 0.0197(5) 0.0008(4) -0.0040(5) -0.0006(6)

O(2) 0.0443(7) 0.0300(6) 0.0380(6) -0.0056(5) 0.0183(5) -0.0064(5)

O(3) 0.0369(6) 0.0198(5) 0.0202(5) -0.0018(4) -0.0022(4) -0.0037(4)

O(4) 0.0238(5) 0.0265(5) 0.0257(5) 0.0008(4) -0.0013(4) 0.0021(4)

N(1) 0.0294(6) 0.0193(6) 0.0205(5) -0.0002(5) 0.0017(5) -0.0007(5)

C(1) 0.0209(6) 0.0204(6) 0.0212(6) -0.0003(5) 0.0018(5) -0.0025(5)

C(2) 0.0244(7) 0.0194(7) 0.0294(7) -0.0043(5) 0.0062(6) -0.0028(5)

C(3) 0.0220(7) 0.0175(6) 0.0365(8) 0.0025(6) 0.0056(6) -0.0003(5)

C(4) 0.0232(7) 0.0233(7) 0.0278(7) 0.0065(5) 0.0031(6) -0.0007(5)

C(5) 0.0213(6) 0.0218(7) 0.0219(6) 0.0006(5) 0.0027(5) -0.0021(5)

C(6) 0.0189(6) 0.0168(6) 0.0224(6) -0.0007(5) 0.0028(5) -0.0010(5)

170 Appendix B

C(7) 0.0211(6) 0.0195(6) 0.0196(6) -0.0011(5) 0.0019(5) -0.0001(5)

C(8) 0.0252(7) 0.0160(6) 0.0199(6) -0.0002(5) 0.0005(5) 0.0010(5)

C(9) 0.0223(7) 0.0202(6) 0.0245(7) 0.0009(5) 0.0024(5) -0.0006(5)

C(10) 0.0267(7) 0.0193(6) 0.0216(6) 0.0018(5) 0.0057(5) 0.0045(5)

C(11) 0.0315(7) 0.0176(6) 0.0186(6) -0.0016(5) -0.0006(5) 0.0000(5)

C(12) 0.0233(7) 0.0208(7) 0.0222(6) 0.0021(5) -0.0008(5) -0.0028(5)

C(13) 0.0239(7) 0.0198(6) 0.0201(6) 0.0013(5) 0.0034(5) 0.0030(5)

171 X-Ray Crystal Structure Data

Table 4. Bond lengths [Å] for 3.11.

atom-atom distance atom-atom distance

S(1)-O(2) 1.4086(13) S(1)-O(1) 1.4094(12)

S(1)-O(3) 1.5755(11) S(1)-N(1) 1.6974(12)

F(1)-C(9) 1.3377(16) F(2)-C(10) 1.3328(16)

F(3)-C(11) 1.3310(16) F(4)-C(12) 1.3284(16)

F(5)-C(13) 1.3361(16) O(3)-C(1) 1.4124(17)

O(4)-C(7) 1.4244(17) O(4)-N(1) 1.4878(16)

N(1)-C(7) 1.4569(18) C(1)-C(2) 1.368(2)

C(1)-C(6) 1.3910(19) C(2)-C(3) 1.388(2)

C(2)-H(2A) 0.9500 C(3)-C(4) 1.380(2)

C(3)-H(3A) 0.9500 C(4)-C(5) 1.391(2)

C(4)-H(4A) 0.9500 C(5)-C(6) 1.3856(19)

C(5)-H(5A) 0.9500 C(6)-C(7) 1.4869(19)

C(7)-C(8) 1.4952(18) C(8)-C(9) 1.383(2)

C(8)-C(13) 1.3869(19) C(9)-C(10) 1.376(2)

C(10)-C(11) 1.380(2) C(11)-C(12) 1.376(2)

C(12)-C(13) 1.375(2)

172 Appendix B

Table 5. Bond angles [°] for 3.11.

atom-atom-atom angle atom-atom-atom angle

O(2)-S(1)-O(1) 122.06(8) O(2)-S(1)-O(3) 111.39(7)

O(1)-S(1)-O(3) 104.98(7) O(2)-S(1)-N(1) 104.03(7)

O(1)-S(1)-N(1) 108.37(7) O(3)-S(1)-N(1) 104.81(6)

C(1)-O(3)-S(1) 116.03(9) C(7)-O(4)-N(1) 59.99(8)

C(7)-N(1)-O(4) 57.85(8) C(7)-N(1)-S(1) 115.21(9)

O(4)-N(1)-S(1) 110.72(8) C(2)-C(1)-C(6) 123.34(13)

C(2)-C(1)-O(3) 117.75(12) C(6)-C(1)-O(3) 118.87(12)

C(1)-C(2)-C(3) 117.79(14) C(1)-C(2)-H(2A) 121.1

C(3)-C(2)-H(2A) 121.1 C(4)-C(3)-C(2) 120.71(14)

C(4)-C(3)-H(3A) 119.6 C(2)-C(3)-H(3A) 119.6

C(3)-C(4)-C(5) 120.34(13) C(3)-C(4)-H(4A) 119.8

C(5)-C(4)-H(4A) 119.8 C(6)-C(5)-C(4) 119.96(13)

C(6)-C(5)-H(5A) 120.0 C(4)-C(5)-H(5A) 120.0

C(5)-C(6)-C(1) 117.85(13) C(5)-C(6)-C(7) 122.04(12)

C(1)-C(6)-C(7) 120.11(12) O(4)-C(7)-N(1) 62.17(9)

O(4)-C(7)-C(6) 116.74(11) N(1)-C(7)-C(6) 120.46(12)

O(4)-C(7)-C(8) 115.09(12) N(1)-C(7)-C(8) 111.82(11)

C(6)-C(7)-C(8) 118.35(12) C(9)-C(8)-C(13) 117.42(13)

C(9)-C(8)-C(7) 122.96(13) C(13)-C(8)-C(7) 119.61(12)

F(1)-C(9)-C(10) 118.38(13) F(1)-C(9)-C(8) 120.00(13)

C(10)-C(9)-C(8) 121.62(13) F(2)-C(10)-C(9) 120.68(13)

F(2)-C(10)-C(11) 119.86(13) C(9)-C(10)-C(11) 119.46(13)

F(3)-C(11)-C(12) 119.93(13) F(3)-C(11)-C(10) 119.70(13)

C(12)-C(11)-C(10) 120.36(13) F(4)-C(12)-C(13) 120.47(13)

173 X-Ray Crystal Structure Data

F(4)-C(12)-C(11) 120.35(13) C(13)-C(12)-C(11) 119.16(13)

F(5)-C(13)-C(12) 118.95(13) F(5)-C(13)-C(8) 119.11(12)

C(12)-C(13)-C(8) 121.94(13)

174 Appendix B

Table 6. Torsion angles [°] for 3.11.

atom-atom-atom-atom angle atom-atom-atom-atom angle

O(2)-S(1)-O(3)-C(1) 57.00(12) O(1)-S(1)-O(3)-C(1) -168.98(11)

N(1)-S(1)-O(3)-C(1) -54.89(11) C(7)-O(4)-N(1)-S(1) 107.46(10)

O(2)-S(1)-N(1)-C(7) -88.03(11) O(1)-S(1)-N(1)-C(7) 140.71(11)

O(3)-S(1)-N(1)-C(7) 29.02(11) O(2)-S(1)-N(1)-O(4) -151.23(9)

O(1)-S(1)-N(1)-O(4) 77.51(10) O(3)-S(1)-N(1)-O(4) -34.18(10)

S(1)-O(3)-C(1)-C(2) -138.50(12) S(1)-O(3)-C(1)-C(6) 43.69(16)

C(6)-C(1)-C(2)-C(3) 0.1(2) O(3)-C(1)-C(2)-C(3) -177.63(12)

C(1)-C(2)-C(3)-C(4) -0.4(2) C(2)-C(3)-C(4)-C(5) -0.3(2)

C(3)-C(4)-C(5)-C(6) 1.3(2) C(4)-C(5)-C(6)-C(1) -1.6(2)

C(4)-C(5)-C(6)-C(7) 177.93(13) C(2)-C(1)-C(6)-C(5) 0.9(2)

O(3)-C(1)-C(6)-C(5) 178.62(12) C(2)-C(1)-C(6)-C(7) -178.60(13)

O(3)-C(1)-C(6)-C(7) -0.93(19) N(1)-O(4)-C(7)-C(6) -112.08(13)

N(1)-O(4)-C(7)-C(8) 102.52(13) S(1)-N(1)-C(7)-O(4) -99.55(10)

O(4)-N(1)-C(7)-C(6) 106.25(14) S(1)-N(1)-C(7)-C(6) 6.70(17)

O(4)-N(1)-C(7)-C(8) -107.77(13) S(1)-N(1)-C(7)-C(8) 152.68(10)

C(5)-C(6)-C(7)-O(4) -132.17(14) C(1)-C(6)-C(7)-O(4) 47.36(18)

C(5)-C(6)-C(7)-N(1) 155.90(13) C(1)-C(6)-C(7)-N(1) -24.57(19)

C(5)-C(6)-C(7)-C(8) 12.1(2) C(1)-C(6)-C(7)-C(8) -168.40(13)

O(4)-C(7)-C(8)-C(9) 44.78(19) N(1)-C(7)-C(8)-C(9) 113.20(15)

C(6)-C(7)-C(8)-C(9) -100.03(16) O(4)-C(7)-C(8)-C(13) -134.17(13)

N(1)-C(7)-C(8)-C(13) -65.76(16) C(6)-C(7)-C(8)-C(13) 81.01(17)

C(13)-C(8)-C(9)-F(1) -178.37(12) C(7)-C(8)-C(9)-F(1) 2.7(2)

C(13)-C(8)-C(9)-C(10) 1.4(2) C(7)-C(8)-C(9)-C(10) -177.56(13)

F(1)-C(9)-C(10)-F(2) 0.1(2) C(8)-C(9)-C(10)-F(2) -179.74(12)

175 X-Ray Crystal Structure Data

F(1)-C(9)-C(10)-C(11) 179.83(12) C(8)-C(9)-C(10)-C(11) 0.0(2)

F(2)-C(10)-C(11)-F(3) -0.8(2) C(9)-C(10)-C(11)-F(3) 179.42(12)

F(2)-C(10)-C(11)-C(12) 178.24(12) C(9)-C(10)-C(11)-C(12) -1.5(2)

F(3)-C(11)-C(12)-F(4) 2.0(2) C(10)-C(11)-C(12)-F(4) -177.02(13)

F(3)-C(11)-C(12)-C(13) -179.46(12) C(10)-C(11)-C(12)-C(13) 1.5(2)

F(4)-C(12)-C(13)-F(5) -1.2(2) C(11)-C(12)-C(13)-F(5) -179.70(12)

F(4)-C(12)-C(13)-C(8) 178.55(12) C(11)-C(12)-C(13)-C(8) 0.0(2)

C(9)-C(8)-C(13)-F(5) 178.26(12) C(7)-C(8)-C(13)-F(5) -2.73(19)

C(9)-C(8)-C(13)-C(12) -1.5(2) C(7)-C(8)-C(13)-C(12) 177.56(13)

176