In order to conserve finite fossil fuels and reduce green house gas emission, the development of sustainable energy is an inevitable challenge in the 21st century. Among the alternative energy sources, cellulosic biomass (e.g. energy grasses such as miscanthus or agricultural waste such as corn stover, bagasse, cereal straws, and wood chips) presents several unique advantages: (1) it can produce liquid fuels or chemicals that substitute the existing petroleum-derived ones without requiring significant modification of infrastructure, (2) it is renewable and potentially carbon-neutral in overall life cycle, and (3) it is inexpensive, ubiquitous and faces less land use competition with food compared to edible biomass (sugar, starch and oil). Aiming to contribute to the development of biomass conversion technology, my Ph.D. study was conducted at the Energy Biosciences Institute, a multidisciplinary joint initiative for biomass research between UC Berkeley, Lawrence Berkeley National Laboratory, University of Illinois at Urbana-Champaign, and BP. Specifically, my focus has been the downstream conversion of sugar/sugar derivatives into chemicals, as described below.

Chapter 1: Polyethylene terephthalate (PET) is a plastic in large demand worldwide, due to its utility in the manufacturing of a range of products from beverage containers to synthetic fibers. The key monomer of PET, terephthalic acid, is currently produced by oxidizing p-xylene from crude oil. In an effort to reduce our dependence on fossil fuel, we demonstrated a route to synthesize solely biomass-derived p-xylene as a drop-in replacement. Namely, we explored the feasibility of converting 2,5-dimethylfuran (derived from 5-(hydroxymethyl)furfural, a typical by-product of cellulose hydrolysis) and acrolein (produced from glycerol, a side-product of fatty acid methyl ester biodiesel production) into p-xylene. This synthesis consisted of a sequential Diels-Alder reaction, oxidation, dehydration, and decarboxylation. In particular, the pivotal first step, the Diels-Alder reaction to construct 7-oxabicyclo[2,2,1]hept-2-ene core structure, was studied in detail to provide useful kinetic and thermodynamic data. The concept was realized and the bio-derived p-xylene was obtained in 34 % overall yield over four steps.

Chapter 2: The deoxygenation reaction of sugar moieties is essential for the conversion of cellulosic biomass to chemicals and fuels. While numerous reports focus on the pyrolysis, hydrogenolysis and acid-catalyzed dehydration reaction of biomass, one much less developed deoxygenation pathway is the deoxydehydration (DODH) reaction, which removes two adjacent hydroxyl groups from vicinal diols to afford alkenes. We have developed an oxorhenium-catalyzed DODH reaction using a sacrificial alcohol (e.g. 1-butanol, 3-pentanol) as a recyclable and environmentally friendly solvent/reductant, and successfully applied it to sugars, sugar acids and sugar alcohols. When combined with the alcohol reductant, oxorhenium compounds, namely methyltrioxorhenium (MTO) and perrhenic acid (HReO₄), showed much higher activity than other precedented DODH systems and enabled the unstable polyol substrates to undergo clean DODH reactions. Linear polyene products and aromatic compounds were obtained with remarkable selectivity. Mechanistic insights were acquired by studying the isolated Re(V) species as well as by examining the unprecedented modes of DODH on 2-ene-1,4-diol and 2,4-diene-1,6-diol moieties.

In recent years, gold(I) complexes have seen increased use as catalysts for the activation of alkynes toward nucleophilic addition. Now, studies show that carbene–like intermediates are accessible, often depending on the substrates and catalyst ligands utilized. While metal carbenoid behavior for gold(I) has been proposed in other transformations, our research has expanded the reactivity of gold beyond simple electrophilic activation, based on evidence that gold can function as both a π–acid and as an electron donor. For instance, electron donation from gold d–orbitals can serve to facilitate the formation of gold carbenoid intermediates after initial cycloisomerization. The gold carbenoid species is then capable of undergoing subsequent reactivity to generate synthetically useful products that can also be further functionalized. Chapter 1 will discuss the development of new gold(I) carbenoid reactivity whereby an oxidation even results in oxygen atom transfer to the carbenoid center to generate carbonyl functional groups. The broad range of reactions for which this methodology can be applied will be discussed as will catalyst and ligand effects for the transformations.

During the course of our investigations into homogeneous gold catalysis, several relevant questions consistently arise. Among them are whether the catalyst can be recovered and recycled, what the cost is of the catalyst, and how load the catalyst loading can be. In order to address these issues, we initiated a research program focused on studying the translation of homogeneous catalytic processes to heterogeneous ones, all in the solution phase. In doing so, we could obtain catalytic activity with heterogeneous catalysts for reactions that were previously only obtained homogeneously. Chapter 2 details our efforts in pursuit of this goal and successes in developing oxidatively modified, electrophilic platinum nanoparticles that display activity for a range of π–bond activation/cyclization reactions. Our characterization studies for the designed nanoparticle system and examination of catalyst activity are presented, as are catalyst leaching investigations. Initial efforts into flow reactors and new dendrimer capping agents are also discussed. The discovery of treatments for nanoparticles that induce the desired homogeneous catalytic activity should lead to the further development of reactions previously inaccessible in heterogeneous catalysis.

Making and Breaking C‒F Bonds via Palladium-Catalysis

By

Richard Thornbury

Doctor of Philosophy in Chemistry

University of California, Berkeley

Professor F. Dean Toste, Chair

Chapter 1 – A ligand controlled, palladium-catalyzed, enantioselective 1,3-arylfluorination of [2H]-chromenes was developed. The products were obtained in high enantioselectivity and with a syn- relationship of the introduced substituents. The pyranyl fluoride products were further derivatized to demonstrate the utility of the products. A ligand dependent divergent formation of 1,3- and 2,1- alkene difunctionalization products was also observed. This bifurcation in reactivity was investigated with a combination of experimental, computational, and statistical analysis tools. Ultimately, the site selectivity was found to be dependent on the ligand denticity and metal electrophilicity, the electronics of the boronic acid, and the donor ability of the directing group in the substrate.

Chapter 2 – A palladium-catalyzed defluorinative coupling of 1-aryl-2,2-difluoroalkenes with boronic acids was developed. Broad functional tolerance arises from a redox-neutral process in which a palladium(II) active species which is proposed to undergo a β-fluoride elimination to afford the products. The monofluorostilbene products were formed with excellent diastereoselectivity (≥50:1) in all cases. As a demonstration of this method’s unique combination of reactivity and functional group tolerance, a Gleevec analogue, using a monofluorostilbene as an amide isostere, was synthesized in 4 steps from commercially available materials.

The ability to control absolute stereochemistry is a powerful tool in organic synthesis. Given the utility of enantiopure molecules in a number of industries, the pharmaceutical perhaps most prominent, many research efforts have focused on asymmetric catalysis. Recently, the Toste Group has developed and implemented a chiral anion phase-transfer strategy to achieve high enantioselectivity in a number of transformations. This thesis describes new reactivity discovered within this manifold and also discusses the design and synthesis of new phase-transfer catalysts to improve enantioinduction.

Chapter 1 details an asymmetric aminocyclization reaction of tryptamine derivatives enabled by chiral anion phase-transfer of aryldiazonium salts. The immediate products, C3-diazenated pyrroloindolines, can be converted directly to cyclotryptamine natural product derivatives, providing an efficient entryway into a complex scaffold that has been previously accessed by much longer synthetic sequences.

Chapter 2 is a natural extension of the work presented in Chapter 1 and describes an enantioselective α-amination of activated ketone derivatives, again employing aryldiazonium salts as aminating reagents. Obtaining high enantioselectivity initially proves to be very difficult, as use of our traditional library of BINOL-derived chiral phosphoric acids leads to poor enantioinduction. Eventually, we overcome this challenge through the design and synthesis of a new library of chiral phosphoric acids, specifically those derived from a 1,1′-binaphthyl-2,2′-diamine (BINAM) backbone. Chapter 2 also presents potential advantages of our methodology over more commonly used α-amination reactions that utilize dialkyl azodicarboxylates.

Chapter 3 discusses the use of aryldiazonium salts in a more traditional role as cross coupling electrophiles. By merging chiral anion phase-transfer and palladium catalysis, we report an enantioselective Heck-Matsuda arylation of cyclic olefins. Through the process of reaction optimization, we observe counterion dependent reactivity and are ultimately able to avoid the formation of undesired olefin isomers by judicious choice of phosphate. While chiral anions have almost exclusively been used to optimize enantioselectivity, this result indicates that counterions can also be used to modulate chemical reactivity in transitional metal catalyzed processes.

The formation of saturated carbon-carbon bonds in a precise and

controlled manner is arguably the principal objective of organic synthesis.

Carbocyclic ring systems comprise the underlying structure for the

preponderance of natural products and pharmaceutical agents. Therefore,

synthetic methods capable of selectively initiating polycyclization processes in

the presence of spectating functionality are of significant value, particularly so

when multiple stereocenters are formed enantioselectively.

The emergence of phosphine gold(I) catalysis over the past decade has

opened up new avenues to carbocycle formation via π activation of alkynes

occurring under exceedingly mild conditions and with excellent chemoselectivity.

The research described herein describes the use of homogenous gold(I)

complexes to initiate electrophilic cyclization cascades. Through rational

substrate design, carbocationic centers may be generated in a predictable

manner and employed in subsequent intramolecular cyclization processes.

Chapter 1 introduces the unique reactivity observed in complexes of gold

imparted by its relativistically accelerated valence electrons. One consequence of

this perturbation is the linear geometry maintained by gold(I) complexes,

minimizing the influence of ligand-based chirality on reactions occurring at

coordinated alkynes. In spite of this challenge, moderate levels of

enantioselectivity were achieved in the desymmetrization of dienynes by

cycloisomerization using chiral bisphosphite gold(I) catalysts.

Ultimately, we were able to achieve selectivities up to 98% ee using

hindered chiral bisphosphine gold(I) catalysts during the evaluation of another

enyne cycloisomerization reaction, described in chapter 2. In this process, an

initial regioselective cyclization was used to generate a carbocationic species

poised to undergo intramolecular trapping. Consistently high enantioselectivity

was maintained using various pendant oxygen, carbon and nitrogen nucleophiles.

The diastereomerically pure bi- and tricyclization products obtained provided

support for a concerted polyene cyclization mechanism as predicted by the Stork-

Eschenmoser postulate.

Chapter 3 describes another tandem process exploiting the transient

cationic species arising from gold(I)-promoted enyne cycloisomerization. In this

case, a gold(I)-initiated tandem cyclopentannulation reaction was employed in

the total synthesis of the novel triquinane ventricosene. A cyclopropanol unit

embedded in the enyne substrate underwent a semipinacol rearrangement in

response to the carbocation, leading cleanly to bicyclo[3.2.0]heptan-6-one

products. For cyclopentenyl substrates, the hindered all-carbon quaternary center

and all of the ring fusions of the angular triquinane ring system were formed at

once. The choice of a hydrocarbon target highlighted the utility of gold(I) catalysts

as selective activators of carbon unsaturation; throughout the synthesis only a

single heteroatom was present.

This work concludes by extending the scope gold(I)-catalyzed

carbocyclization reactions which generate useful cationic intermediates. The

gold(I)-catalyzed Rautenstrauch rearrangement forms a cyclopentene-based

cationic species which was shown to undergo efficient trapping by pendant

arenes to give a saturated 5,6-ring fusion comprising a chiral benzylic quaternary

center. The chirality transfer observed in the parent process was found to be

conserved in the tandem process. Interestingly, cyclization of racemic substrates

by chiral bisphosphine digold catalysts was found to proceed with moderate

enantioselectivity, suggesting a competing mechanism is in effect which

proceeds through an achiral intermediate.

Throughout my doctoral studies, I have endeavored to address the question of how to rationally design chiral catalysts to control enantioselectivity in a predictable fashion. To approach this problem, I have focused on the strategic implementation of attractive non-covalent interactions between catalysts and their substrates at the enantiodetermining transition state. A significant portion of this work involved the development of an approach to elucidating the structural origins of selectivity in complex data sets, with an eye towards rational catalyst design. The insights gained from these studies constitute the subject matter of this dissertation.

Chapter 1 provides an overview of a subset of attractive non-covalent interactions involving aromatic rings. Examples are highlighted in which these interactions have been characterized in the ground state with the explicit intention of quantifying their strengths and geometric requirements. Using the lessons learned from these reports, case studies are presented in which the interactions under discussion have been leveraged to catalyze chemical reactions via the paradigm of transition state stabilization. The insight gleaned from this literature survey is intended to provide a framework for the discussion of the catalyst design concepts outlined in the following chapters.

Chapter 2 describes the development of a novel class of axially chiral triazole containing phosphoric acid catalysts designed to impart enantioselectivity via attractive non-covalent interactions with a substrate at the enantiodetermining TS. This strategy stands in contrast to that typically discussed in the literature in which catalyst steric bulk is frequently invoked to rationalize asymmetric induction. Using this newly prepared catalyst library, an enantioselective cross dehydrogenative coupling was developed within the conceptual framework of the Toste group’s strategy of chiral anion phase transfer catalysis. Preliminary experiments suggested that the triazole substituents played a role in the determination of asymmetric induction beyond steric bulk, as initially hypothesized.

The third chapter describes our efforts to elucidate the structural origins of enantioselectivity in the oxidative coupling reaction presented in Chapter 2. By obtaining an enantioselectivity data set based on strategically modulated catalysts and substrates, we applied linear regression techniques to correlate our experimental data with molecular descriptors describing the structural variation throughout the data set. The resulting models, along with the overall enantioselectivity trends, were used to develop intuitively sensible mechanistic hypotheses that were subsequently tested through the application of new catalysts specifically tailored to address them. Based on the hypothesis that the triazole catalysts were engaged in enantiodetermining  interactions with the substrates at the TS, we rationally designed catalysts that afforded the products with the highest enantiomeric excesses reported to date.

Chapter 4 builds on the strategy outlined in Chapter 3 in the context of a boronic acid-directed phosphoric acid catalyzed enantioselective fluorination of allylic alcohols. Specifically, the logic developed in this chapter allows for the generalization of our data intensive strategy to scenarios in which the mechanism for selectivity determination changes throughout a data set. Through the careful organization of enantioselectivity data and strategically implemented mechanistic experiments, we identified a region of structural space in which we propose that a lone pair- interaction governs enantioselectivity. Based on this hypothesis, we rationally designed a catalyst system capable of producing either enantiomer of a chiral fluorinated building block in high enantiomeric excess using the same chiral catalyst.

The total synthesis and synthetic efforts toward norditerpene natural products isolated from Pseudopterogorgia elisabethae and macrocyclic polyketides from Archangium gephyra is described. A highly convergent synthesis of archazolid B, uniting three linear subunits, is accomplished using various modern transition-metal catalyzed reactions. A relay ring-closing metathesis is featured for closure of the 24-membered macrocyclic ring. Construction of the carbon skeleton of archazolid A is also described, featuring two robust, scaleable building blocks and their unification. New strategies for the construction and elaboration of a 1,4 skipped diene moiety are established in this synthesis.

The total synthesis of norditerpenes amphilectolide and sandresolide B are also described. A palladium mediated carbonylative coupling is featured en route to the preparation of a key furan building block that can be used to access several norditerpene natural products produced from Pseudoterogorgia elisabethae. This key building block was then elaborated to amphilectolide utilizing a lanthanide catalyzed ring closure and furan oxidation with singlet oxygen. The ring-closure product may also be used to access caribenol B.

The furan building block was also used to construct sandresolide B, where a Myers alkylation overcame steric constraints in this system. An intramolecular Friedel Crafts acylation is used to close a 7-membered ring. These strategies also provide insight to a next generation route toward caribenol A, and the progress toward the initial strategy via Stork-Danheiser ring closure is described.

Chapter 1 – The gold(I)-catalyzed, stereoselective hydrofluorination of electron-deficient alkynes with triethylamine trihydrogen fluoride (Et3N·3HF) is described. Fluorinated α,β-unsaturated aldehydes, amides, esters, ketones, and nitriles were isolated in moderate to good yields as single diastereomers. In all but four cases, the (Z)-vinyl fluorides were initially formed in ≥97% diastereoselectivity. This work constitutes the first catalytic example of the diastereoselective preparation of a variety of β-alkyl, β-fluoro Michael acceptors from alkynes. Additionally, the described work expands access to β-aryl, β-fluoro Michael acceptors to the synthesis of β-fluoro- α,β-unsaturated amides and nitriles. The monofluoroalkenes formed through this strategy were readily transformed into other fluorine-containing compounds, and the developed method was applied to the synthesis of a fluorinated analogue of Exoderil, a topical antimycotic.

Chapter 2 – The copper-catalyzed silylation of propargylic difluorides to generate axially chiral, tetrasubstituted, monofluoroallenes in both good yields (24 examples >80%) and enantioselectivities (82–98% ee). Compared to previously reported synthetic routes to axially chiral allenes (ACAs) from prochiral substrates, a mechanistically distinct reaction has been developed: the enantiodiscrimination between enantiotopic fluorides to set an axial stereocenter. DFT calculations and vibrational circular dichroism (VCD) suggest that β-fluoride elimination from an alkenyl copper intermediate likely proceeds through a syn-β-fluoride elimination pathway rather than an anti- elimination pathway. The effects of the C1-symmetric Josiphos-derived ligand on reactivity and enantioselectivity were investigated. Not only does this report showcase that alkenyl copper species—like their alkyl counterparts— can undergo β-fluoride elimination, but that this elimination can be achieved in an enantioselective fashion. The resulting monofluoroallenes were shown to undergo a copper-catalyzed desilylative acylation. A kinetic resolution was developed, generating propargyl fluorine-containing quaternary stereocenters. Current efforts are underway to better understand the system that aim to achieve dynamic kinetic resolution.

Understanding the interaction between gold and its ligands is essential in exploiting this versatile metal in many areas of chemistry. Organogold chemistry and the reactivity of gold-carbon bonds, particularly in the context of catalysis, have been extensively investigated over the past two decades. Markedly less research has been dedicated to gold-heteroatom bonds. A desire to understand these bonds has resulted in four studies with vastly different goals and conceptual underpinnings, but that are unified by utilizing gold's interaction with X-type heteroatom ligands.