Article pubs.acs.org/accounts
Merging Visible Light Photoredox and Gold Catalysis Published as part of the Accounts of Chemical Research special issue “Photoredox Catalysis in Organic Chemistry”. Matthew N. Hopkinson,*,† Adrian Tlahuext-Aca, and Frank Glorius* Organisch-Chemisches Institut, NRW Graduate School of Chemistry, Westfälische Wilhelms-Universität Münster, Corrensstrasse 40, 48149 Münster, Germany CONSPECTUS: Since the beginning of this century, π-Lewis acidic gold complexes have become the catalysts of choice for a wide range of organic reactions, especially those involving nucleophilic addition to carbon−carbon multiple bonds. For the most part, however, the gold catalyst does not change oxidation state during the course of these processes and twoelectron redox cycles of the kind implicated in cross-coupling chemistry are not easily accessible. In order to address this limitation and expand the scope of gold catalysis beyond conventional hydrofunctionalization, extensive efforts have been made to develop new oxidative reactions using strong external oxidants capable of overcoming the high potential of the AuI/AuIII redox couple. However, these processes typically require superstoichiometric amounts of the oxidant and proceed under relatively harsh conditions. Moreover, to date, goldcatalyzed oxidative coupling reactions have remained somewhat limited in scope because, for many systems, the desired crosscoupling does not favorably compete with homodimerization or conventional hydrofunctionalization. In 2013, we disclosed a new concept for gold-catalyzed coupling reactions that, rather than involving external oxidants, employs aryl radicals that act as both the oxidant and the coupling partner in overall redox-neutral transformations. For this, we developed a dual catalytic system combining homogeneous gold catalysis with the emerging field of visible light photoredox catalysis. Using aryldiazonium salts, which are known to act as sources of aryl radicals upon activation with reducing photocatalysts, we could achieve intramolecular oxy- and aminoarylations of alkenes upon irradiating the reaction mixtures with visible light. Further studies on this transformation, in which nucleophilic addition onto a gold-activated alkene is followed by C(sp3)−C(sp2) bond formation, expanded the scope of the process to intermolecular, three-component oxyarylation, while inexpensive organic dyes and user-friendly diaryliodonium salts could be employed as alternative photocatalysts and aryl radical sources, respectively. The potential of dual gold/photoredox catalysis was quickly realized by several research groups and a range of diverse new coupling reactions involving nucleophilic addition to π-systems and even P−H and C(sp)−H functionalization have been developed. In addition to the ambient reaction conditions and the simple setup using household light sources or even sunlight, a key advantage of dual gold/photoredox catalysis results from the simultaneous oxidation of gold(I) and coordination of the coupling partner, which results in high levels of selectivity for the cross-coupled products over homodimers. Furthermore, when gold complexes that are not catalytically active prior to oxidation by the aryl radical are employed, background reactions not involving coupling can be suppressed. Notably, this feature has allowed for the successful use of allenes and alkynes, for which conventional hydrofunctionalization pathways are highly favored, opening the door to new transformations involving the most common substrate classes for gold catalysis. In this Account, we provide an overview of dual gold/photoredox catalysis and highlight the potential of this concept to greatly expand the scope of homogeneous gold catalysis and enable the efficient construction of complex organic molecules. Moreover, recent studies on the visible light-promoted synthesis of novel gold(III) complexes suggest that photoredox activation could yet find further applications in gold chemistry beyond coupling.
1. INTRODUCTION
hydrofunctionalized products. Based on this reactivity, a diverse array of transformations involving carbon, nitrogen, oxygen, or other nucleophiles and alkyne, allene, or, to a lesser extent, alkene substrates have been reported employing various gold(III) or, more commonly, gold(I) complexes as catalysts. In most cases, these processes are highly chemoselective and tolerant of many
Over the last two decades, homogeneous gold catalysis has become a powerful tool for the synthesis of complex organic molecules.1−6 In the majority of cases, the gold catalyst acts as a soft π-Lewis acid and selectively activates a carbon−carbon multiple bond in the starting material toward intra- or intermolecular nucleophilic addition. This step results in the formation of an organogold species (A for gold(I), Scheme 1), which then typically undergoes fast protodemetalation to afford © 2016 American Chemical Society
Received: July 5, 2016 Published: September 9, 2016 2261
DOI: 10.1021/acs.accounts.6b00351 Acc. Chem. Res. 2016, 49, 2261−2272
Article
Accounts of Chemical Research Scheme 1. Gold-Catalyzed Hydrofunctionalization of Carbon−Carbon Multiple Bonds
attractiveness with regards to cost, atom-economy, and functional group tolerance, while many reactions are still performed at elevated temperatures. Another significant drawback concerns the inherent difficulty in controlling the selectivity for the crosscoupling products over homodimers of either coupling partner. Unlike conventional redox-neutral cross-coupling, where the coordination of one coupling partner occurs during the oxidative addition step and the other via transmetalation, in oxidative coupling reactions, there is no such mechanistic distinction between the coordination steps, and selectivity for cross-coupling can be much harder to achieve.
functional groups, while the stability of gold toward air and moisture allows for mild and convenient reaction conditions. In addition to exploring the range of π-systems and nucleophiles suitable for these transformations, many research efforts have focused on the second bond-forming step in the catalytic cycle (step II, Scheme 1), during which the active gold catalyst is liberated from the organogold species A. Replacing the protonation step typically observed with alternative demetalation pathways would dramatically expand the scope of this chemistry by opening the door to powerful reactions involving the formation of two new C−C or C−heteroatom bonds across the π-system. One attractive option in this regard concerns engaging organogold intermediates of type A in cross-coupling reactions. In such processes, gold acts not only in its traditional role as a catalyst for the nucleophilic addition step but also as a catalyst for the subsequent coupling process. These kinds of redox reactions require the catalyst to undergo two-electron oxidation state changes during the catalytic cycle, and while gold(I) and gold(III) are isoelectronic with palladium(0) and palladium(II), respectively, the redox potential of the AuI/AuIII couple is significantly higher (E0 = 1.41 V) than that of the Pd0/PdII couple (E0 = 0.92 V).7 As such, gold(I) does not generally undergo oxidative addition with standard organic electrophiles used in cross-coupling reactions (e.g., aryl halides) and successful goldcatalyzed coupling processes have instead largely been oxidative transformations, requiring the use of a strong external oxidant. Using this strategy, a number of impressive gold-catalyzed tandem nucleophilic addition/oxidative cross-coupling reactions have been developed typically using either hypervalent iodine species or the electrophilic fluorinating reagent Selectfluor as the oxidant (Scheme 2).8,9 There remain, however, a number of disadvantages associated with gold-catalyzed oxidative coupling. For example, superstoichiometric amounts of the strong external oxidant are normally employed in these processes, limiting their
2. CONCEPT AND INITIAL SUCCESS Motivated by the desire to address some of these limitations and more generally to provide an alternative approach for goldcatalyzed coupling, we considered whether overall redox-neutral processes more akin to conventional cross-coupling reactions could ever be accessible with gold. As stated above, the major stumbling block in these reactions is the challenging oxidative addition step, where a gold(I) catalyst is converted to an organogold(III) intermediate. Rather than performing this twoelectron oxidation process as a single step, we designed an alternative mechanistic sequence featuring two single electron oxidation processes. The first step would constitute a single electron oxidation of gold(I) involving addition of an organic radical generated in situ during the reaction (Scheme 3a). The resulting organogold(II) species (B) would then be expected to undergo a second single electron oxidation to afford the desired organogold(III) intermediate (C). Evidence for the oxidation of stoichiometric gold(I) by organic radicals was provided by the pioneering work of Puddephat in the 1970s and more recently by Toste using trifluoromethyl iodide.10,11 Moreover, in 2005, Corma and co-workers demonstrated that in situ generated benzyl and aryl radicals were capable of reacting with either gold(I) or even colloidal gold to afford organogold(III) species.12 Aryl radicals have also been previously engaged as coupling partners in palladium-catalyzed cross-coupling reactions. In 2009, Chan and co-workers developed a C−H arylation process employing benzoyl peroxide as a source of aryl radicals.13 Two years later, however, the Sanford group introduced a new protocol that combined palladium-catalyzed cross-coupling with the emerging field of visible light photoredox catalysis.14−19 Using this system, aryl radicals could be generated at room temperature from aryldiazonium salts upon irradiation with visible light from a standard household desk lamp. Inspired by these studies, we designed a dual gold/photoredox catalytic system that, we hoped, would not only be able to mediate redox-neutral cross-coupling reactions, but also operate under similarly mild conditions. A general mechanistic outline of our designed dual gold/photoredox catalyzed nucleophilic addition/cross-coupling process is shown in Scheme 3b. In the first step, a coordinatively unsaturated gold(I) complex activates a π-bond in the substrate toward nucleophilic attack, affording an organogold(I) intermediate A. Concomitantly, irradiation of a
Scheme 2. Gold-Catalyzed Oxidative Difunctionalization Reactions Using External Oxidants
2262
DOI: 10.1021/acs.accounts.6b00351 Acc. Chem. Res. 2016, 49, 2261−2272
Article
Accounts of Chemical Research Scheme 3. Initial Mechanistic Design of Dual Gold/Photoredox-Catalyzed Difunctionalization Reactions
Scheme 4. Dual Gold/Photoredox-Catalyzed Intramolecular Oxyarylation of 1a (NMR Yields)
photoredox catalyst such as [Ru(bpy)3]2+ results in the formation of a reducing excited state species that is capable of donating an electron to the aryldiazonium salt (E1/2(RuIII/*RuII) = −0.81 V) to deliver [Ru(bpy)3]3+ and, after extrusion of dinitrogen, the key aryl radical.20 This reactive open-shell species can then add to the organogold(I) intermediate A to give an organogold(II) species B bearing both coupling partners. Single electron oxidation of B could then occur with the oxidized photocatalyst [Ru(bpy)3]3+, regenerating the [Ru(bpy)3]2+ complex and delivering the organogold(III) species C. Reductive elimination at this stage would complete the catalytic gold cycle, liberating the cross-coupled product and regenerating gold(I). For our initial studies, we selected the intramolecular oxyarylation of alkenes as a model reaction. This process is conceptually related to the Mizoroki−Heck coupling but, rather than delivering styrene products, involves the formation of a C(sp3)−O and a C(sp3)−C(sp2) bond across the alkene. Gold catalysts are well-suited to such transformations because they do not generally undergo β-hydride elimination and so do not lead to undesired Mizoroki−Heck side-products. Indeed, oxidative oxy- and aminoarylation processes using arylboronic acids or arylsilanes as the aryl coupling partners and Selectfluor as an external oxidant have been previously reported.21−25 In a first experiment, 1-penten-5-ol (1a) was reacted with benzenediazonium tetrafluoroborate (2a, 4 equiv) in the presence of Ph3PAuCl (10 mol %) and [Ru(bpy)3](PF6)2 (5 mol %) as photocatalyst in degassed methanol. After irradiation with visible light for 6 h from a 23 W household compact fluorescent lamp (CFL), we were delighted to observe the nucleophilic addition/ cross-coupling product 3aa in 51% (Scheme 4). Control reactions conducted in the absence of the gold catalyst led to no product formation, while 3aa was delivered in only trace
amounts (
