Dearomatization Through Photoredox Hydroarylation: Discovery of Radical-Polar Crossover Strategy

Abstract

Indole dearomatization has been achieved via radical hydroarylation. Under mild photoredox conditions, a range of indole derivatives undergo hydroarylation to form 2-arylindoline products. Mechanistically, radical termination occurs primarily via stepwise reduction/protonation, with a small contribution from concerted hydrogen atom transfer. This mechanistic understanding prompted the extension of this reactivity to benzenoid dearomatization. This work formed the foundation of our program, which utilizes reductive radical-polar crossover to drive highly-selective dearomatization pathways.

Graphical Abstract

Dearomatization methods are attractive for their abilities to transform simple, planar arenes into highly complex, three-dimensional structures.¹ We aimed to develop a general catalytic strategy for dearomatization that operates through a radical mechanism. More specifically, we sought to leverage the highly-reactive nature of radical species to overcome aromatic stabilization energy.² As shown in Figure 1, addition of aryl radicals to aromatics would give rise to dearomatized cyclohexadiene intermediates. While, typically, these intermediates undergo rapid oxidation (E_1/2° = – 0.34 V vs SCE)³ and deprotonation (e.g. Minisci pathway),⁴ we recognized that formal interception of the key dienyl radical species with a hydrogen atom would deliver significant value. As a mechanistically distinct approach, this would complement existing oxidative,⁵ enzymatic,⁶ and photochemical⁷ dearomatization methods. Drawing from our experience in olefin hydroarylation,⁸ we postulated that the key to realizing this dearomatization goal would lie within the mechanism of radical termination. Here, the net addition of a hydrogen atom could reasonably be accomplished via concerted hydrogen atom transfer (HAT) or through a stepwise radical-polar crossover process (sequential addition of an electron and proton). We recently reported that radical-polar crossover effectively enables the dearomative spirocyclization of aryl⁹ and alkyl¹⁰ radical species. In this report, we share the initial insights that led to this program.

Figure 1.

Key lesson learned in development of radical dearomative indole hydroarylation.

We first chose to interrogate this blueprint for dearomatization through the reactivity of indole substrate 1. We understood that single electron reduction of the bromopyridine group would deliver the corresponding aryl radical species. Because radical cyclization onto the indole would be facile, this would allow us to focus on trapping the cyclized radical intermediate. Successful radical hydroarylation according to this plan would yield 2-arylindoline products in a complementary manner to methods developed by Jia,¹¹ Lautens,¹² and others,¹³ which utilize Heck/anionic capture chemistry and require the formation of a quaternary center at C2. Early radical approaches to this scaffold are limited due to the use of harsh conditions (Bu₃SnH or (Me₃Si)₃SiH, AIBN).¹⁴ In this work, we detail the development of a photoredox-catalyzed system for indole hydroarylation that functions under mild conditions (organic photocatalyst and amine reductant powered by visible light) and is highly selective for the desired dearomative pathway. Critical to this selectivity is the reductive radical-polar crossover mechanism, which, under the conditions that we describe, occurs in preference to the classical radical arylation pathway.

Our study began with reacting indole 1 under conditions we previously developed for hydroarylation of electron-rich olefins.^8a After a series of minor adjustments to these conditions (see Supporting Information for details) we obtained dearomatized product 2 in 77% yield (Table 1, entry 1). However, these conditions were not ideal because they require extremely low concentrations (0.02 M), employ a precious metal photocatalyst, and involve laborious product purification (i.e. removal of the Hantzsch ester oxidation product). Consequently, we explored the use of trialkylamine reductants, finding tributylamine as a suitable replacement with a yield of 59% (Table 1, entries 2–4). Increasing the reaction concentration and removing the thiol additive improved the yield to 84% (Table 1, entry 5). Finally, we found that Ir(ppy)₂dtbbpy⁺ can be replaced with the organic donor-acceptor cyanoarene catalysts 4CzIPN (P2) and 3DPAFIPN (P3) to give 2 in comparable yield (Table 1, entries 6–7).¹⁵ It is notable that these conditions result in efficient dearomatization without catalytic thiophenol (PhSH), even though we have previously seen thiol play a crucial role in the HAT termination of carbon-centered radicals.^8a

Table 1.

Optimization of conditions for dearomative indole hydroarylation^a

entry	reductant	additive	photocatalyst	yield of 2d (%)
1	HEH	AcOH, PhSH	P1	77
2	Et3H	PhSH	P1	24
3	i-Pr2Net	PhSH	P1	43
4	Bu3N	PhSH	P1	59
5b	Bu3N	–	P1	84
6b,c	Bu3N	–	P2	80
7b,c	Bu3N	–	P3	83

^aConditions: aryl radical precursor (1 equiv), P1 (1 mol %), AcOH (20 equiv), reductant (3 equiv), PhSH (5 mol %), 2,2,2-trifluoroethanol (0.02 M), blue LEDs, 23 °C, 16h.

^b0.2 M.

^c5 mol % photocatalyst.

^dYields determined by GC.

We next sought to understand the substrate scope of this indole hydroarylation system and generated a series of electronically-varied substrates. On a preparative scale, we found that a toluene cosolvent (25% v/v) improved photocatalyst solubility and reaction reproducibility. Under these conditions, hydroarylation of a variety of indole substrates was efficacious (Table 2). Both bromopyridines and aryl iodides served as effective aryl radical precursors, cyclizing to give hydroarylation products in moderate to good yields (2–3, 47–86%). Reduction of an electron-rich aryl iodide bearing a methoxy substituent was possible under these conditions (4, 62%). Further halogenated aryl iodide substrates demonstrated selective iodide cleavage to give arylindoline products in good yields (5–6, 66–70%). Indole substitution is well-tolerated at the 4, 5, and 6 positions. Electron-rich indoles containing methyl, silyl ether, and methoxy substituents undergo hydroarylation in great yields (7–10, 76–86%). Bromination of indole is also tolerated and results in the formation of arylindoline products in good yields (11–13, 69–83%), demonstrating the regioselective reduction of the most electron-poor ring in the system. These aryl bromides, which would not typically survive Pd-catalyzed arylation conditions,^{11b, 12b, 13a} serve as handles for further reactivity. A key limitation of this system is seen when reacting electron-poor indoles. The instability of the indole amide bond in concert with the basic reaction conditions results in solvolysis of the amide prior to aryl radical formation.

Table 2.

Scope of dearomative indole hydroarylation^a

^aConditions: aryl radical precursor (1.0 equiv), P3 (5 mol %), tributylamine (3.0 equiv), 25% (v/v) toluene/2,2,2-trifluoroethanol (0.1 M), blue LEDs, 23 °C, 16 h, isolated yields.

^bDeviations: P2 (5 mol %) catalyst, 2,2,2-trifluoroethanol (0.1 M) solvent, 40 °C.

Depicted in Figure 2 are two reasonable mechanisms that we envisioned. Here, electron transfer from amine reductant to P3* (E_1/2(P3*/P3^•−) = +1.09 V vs SCE)^15a results in formation of the corresponding cyanoarene radical anion (P3^•−), and amine radical cation. Single electron reduction (SET) of 1 (E_1/2(P3/P3^•−) = −1.59 V vs SCE)^15a would, after loss of bromide, give pyridyl radical 14. Cyclization through the 5-exo mode would give rise to benzylic radical 15. Radical termination via HAT from amine radical cation (BDE (C–H) = ~42 kcal/mol)¹⁶ would deliver desired indoline 2. Alternatively, reduction of 15 (E_1/2° = – 1.24 V vs SCE; see Supporting Information for details) to the corresponding anion 16 would precede a protonation event, also arriving at 2.

Figure 2.

Mechanistic hypothesis for indole hydroarylation and deuterium labeling experiments to determine the termination event.

To evaluate the relative contributions of each of the outlined pathways to the observed results, we conducted two deuterium-labeling experiments (Figure 2, A and B). To probe the first hypothetical pathway (path a), we replaced tributylamine with triethyl-d₁₅-amine, where HAT from the oxidized amine would result in deuterium incorporation in the product 2 (experiment A). However, under these conditions, we saw only 5% deuterium incorporation at the benzylic position of 2. To consider the radical-polar crossover pathway (b), we replaced TFE with 2,2,2-trifluoroethanol-OD under otherwise standard reaction conditions (B). We observed 86% deuterium incorporation at the 3-position of indoline 2. These implicate reductive radical-polar crossover as the primary operative mechanism in this system, though it is likely that both scenarios contribute to the observed reaction outcome.

We recognized that this mechanistic approach, where radical termination occurs through a stepwise reduction/protonation sequence, would be useful in a number of other settings. One direction that we were particularly interested in exploring involved the dearomative cyclization of aryl radicals with unactivated aromatics (i.e. benzenoids). To evaluate this idea, we constructed 2-bromo-3-benzyloxypyridine 17. Through the presumed mechanistic pathway that is outlined in Figure 3, we obtained spirocyclic diene 20 in meaningful yield (26%). Key to this transformation is exo-cyclization of pyridyl radical 18, and reduction of the obtained intermediate to afford dienyl anion 19. This Birch-like intermediate would be selectively reacted with electrophiles (proton in this case) to deliver the desired spirocyclic system.¹⁷ Here, deuterium-labeling studies demonstrated >95% D incorporation from solvent, in support of the outlined mechanistic design. This initial example of radical benzene dearomatization led to a broad method for the dearomatization of unactivated arenes through hydroarylation.⁹ Additional work has been done in our lab to extend this method to the dearomative hydroalkylation of unactivated arenes via α-acyl radical intermediates,¹⁰ and Yu detailed that anion trapping can be performed effectively with CO₂.¹⁸

Figure 3.

Extension of dearomative hydroarylation to benzene derivatives.

In summary, we have developed a system for the intramolecular dearomatization of indoles that functions under mild conditions and acts without the requirement of C2 substitution. Mechanistically, we discovered that this process terminates predominantly through reductive radical-polar crossover. Understanding the utility of this stepwise reduction/protonation mechanism, we demonstrated the extension of this reactivity to benzenoid dearomatization through aryl radical spirocyclization.