Enantioselective photoredox dehalogenative protonation

We report an enantioselective photoredox dehalogenative protonation as a new type of asymmetric protonation.


Introduction
Enantioselective protonation is a fundamental method for synthesizing enantioenriched a-tertiary carbonyl compounds. 1 In the past decade, a number of efficient ground-state catalytic platforms have been established, allowing the precise delivery of a proton to prochiral carbanion intermediates in a highly enantioselective manner. [2][3][4][5][6][7][8] Nonetheless, the approach of constructing secondary a-carbon-halogen bonds for ketones still remains underdeveloped, as few examples with insufficient enantioselectivities have been described (Scheme 1a). 3b,4a,6a The stereoselective formation of C-F bonds is of critical importance in pharmaceutical chemistry given the ability of uorine atoms to act as the isostere of hydrogen atoms in decreasing the rate of metabolic degradation without inuencing the pharmacological effects. 9 Meanwhile, chiral C-Cl and C-Br bonds are applicable for introducing various important molecular architectures via simple and stereospecic transformations and are thus extensively adopted in the asymmetric synthesis of chiral complex molecules with signicant biological activities. 9a,b,10 In this context, the development of a new and generic catalytic asymmetric protonation approach to form valuable chiral tertiary a-haloketones 9,10 represents a highly desirable task that would be a complementary and more versatile method than direct catalytic enantioselective halogenation. Of note, a few direct asymmetric a-chlorination and a-uorination reactions of aliphatic ketones via enamine catalysis have been established (Scheme 1b); 11 however, the catalytic systems are not compatible with aromatic variants, likely because of the slow enamine formation and the approximately equimolar enamine rotational isomers. An approach through indirect enantioselective chlorination of b-ketocarboxylic acids 12 also failed to generate chiral a-secondary chloro-ketones in high enantioselectivities (Scheme 1c).
Due to the strong electron-withdrawing properties of ketones and the high electronegativity of halogens, the stereocenter of secondary a-haloketones should be particularly labile, especially for those aromatic variants. 6a,11b As such, a mild and neutral protonation strategy to avoid racemization of the products in the reaction process is inarguably crucial. Recently, we reported the asymmetric photoreduction of 1,2-diketones to chiral a-hydroxy ketones enabled by a synergistic Scheme 1 Previous studies on the enantioselective construction of secondary a-C-X bonds for ketones.
photosensitizer and H-bonding catalysis with weakly basic tertiary amines as the terminal reductant (Scheme 2a). 13 The excellent enantioselectivity clearly demonstrated the applicability of such an open-shell single-electron-transfer (SET) approach 14 for the formation of the labile tertiary carbon stereocenter of a-hydroxyl ketones via an asymmetric protonation procedure. Accordingly, we wondered whether the ability to combine this dual-catalysis 15 reduction system with competent oxidative feedstocks, from which halogenated carbon anions a to ketones could be produced via two SET reductions, might address this challenging task.
Alkyl halides (X) are an important class of readily accessible starting substrates in organic synthesis. 16 Since Tanaka and coworkers introduced the visible light-driven reductive dehalogenation of benzyl bromide to benzene in 1984, 17 the groups of Kellogg, 18a Fukuzumi,18b Stephenson, 18c-e Zeitler, 18f and Hisaeda 18g and so on 18h-o have accomplished the transformation of diverse alkyl halides to alkanes via photoredox catalysis with tertiary amines as the reductant. This conceptual strategy was developed by the Reiser group 19 by establishing an alternative reductive system, i.e., 1,5-dimethoxynaphthalene (DMN) with ascorbic acid, wherein the selective debrominative protonation of a,a-dibromoketones to the racemic monobromo ketones was described (Scheme 2b). Furthermore, Hyster and co-workers recently reported an enzyme catalytic debrominationenantioselective hydrogen atom transfer (HAT) to access enantioenriched a-tertiary esters from a-bromoesters. 20 Encouraged by these elegant studies, we hypothesized that the enantioselective platform of a,a-dihaloketones would be an expedient and modular strategy for accessing the desired chiral organohalides. Several signicant challenges should remain in this promising task, namely the weaker basicity of ketones relative to ketyls which would impair the stereocontrol of H-bonding catalysis, and both the competitive racemic background reaction and the possible racemization of the stereocenter which should diminish the enantioselectivity. More importantly, even employing a powerful enzyme catalysis and capturing a hydrogen atom from the bulky avin still could not provide a good enantiofacial result for the formation of a C-F bond. 20a Therefore, the direct delivery of the smallest proton to the prochiral intermediates through the distinctly weaker asymmetric H-bonding induction to provide stereocontrolling environments would be more formidable.

Optimization of reaction conditions
We began our study with 2-chloro-2-uoro-tetralone 1a as the model substrate (Table 1). A range of chiral H-bonding catalysts, photoredox catalysts, amine reductants, and inorganic bases as the acid-binding agents to remove the generated HCl were examined in conjunction with diverse reaction parameters (see Tables S1 and S2 in the ESI †). As a result, the desired product 2a was obtained in 92% yield with 96% ee when the reaction was performed in 1,2-dichloroethane at 5 C for 36 h in the presence of dicyanopyrazine-derived chromophore (DPZ, 13,16b,21 5 mol% L-tert-leucine-based squaramide-tertiary amine C1, 22 0.6 equiv. Amine-1 as the secondary amine, 1.0 equiv. H 2 O and 2.5 equiv. Na 2 CO 3 (entry 1). The structural features of the chiral catalyst, consisting of the aryl substituent of squaramide (C2), the structural skeleton of the tertiary amine (C3) and the type of H-bonding donor (C4 and C5), were modied to affect the enantiofacial selectivity (entries Scheme 2 Previous studies. (a) Photoredox catalytic asymmetric reduction of 1,2-diketones via enantioselective protonation to build the labile a-secondary alcohols of ketones. (b) Photoredox catalytic reductive dehalogenation of a,a-dibromoketones in a racemic manner.  [2][3][4][5]. When Amine-2 was the reductant, the yield of 2a decreased to 76% with a similar enantioselectivity (entry 6). Other plausible photoredox catalysts, including Ru(bpy) 3 , were tested (entries 7-9), but better comprehensive results were not achieved. The moderate yield obtained with Rose Bengal stemmed from an unclear reaction interruption, which resulted in unsatisfactory chemical conversion. The transformation in the absence of catalyst C1 offered rac-2a in 48% yield with >95% chemical conversion aer 36 h, indicating that a racemic background reaction occurred (entry 10). The lower yield was due to deterioration of the chemical selectivity, as a few unknown side products were observed. In the absence of H 2 O, the yield of 2a decreased to 74% with a similar enantioselectivity (entry 11), suggesting its important effect in increasing the solubility of inorganic salts. DPZ as the photoredox catalyst was found to accelerate the transformation (entry 12), although the type of EDA 23 is involved in the process (see ESI † for details). Control experiments also conrmed that reductants, inorganic bases and visible light were essential for the reaction (entries [13][14][15]. With the optimized conditions in hand, we next evaluated the substrate scope (Table 2). A range of 2-chloro-2-uorotetralones with diverse electron-withdrawing and electrondonating substituents on the aryl ring were rst examined. The reactions were nished within 24-36 h, and a series of chiral a-uorinated tetralones 2a-o were obtained in 68-96% yield and with 92 to >99% ee. The satisfactory results achieved with alkyne (2h), hydroxyls protected by tosyl (Ts, 2j), allyl (2k), mesyl (Ms, 2m) and tert-butyl-dimethylsilyl (TBS, 2n) groups and even the naked hydroxyl (2o) as the substituent underscore the high functional-group tolerance of this catalytic system. The reaction to furnish 2a was attempted on a 1.0 mmol scale, and a similar yield and enantioselectivity were observed with a slightly longer reaction time (48 h, see footnote b), indicating the promising synthetic utility of this method. Other signicant cyclic ketones, including 4-chromanone (2p), thiochroman-4one (2q), 1-indanones with distinct substituents on the aromatic ring (2r-2x) and 1-benzosuberone (2y), were subsequently tested, leading to a variety of corresponding enantioenriched a-uoroketones in 55 to 90% yield and 84 to 93% ee. Linear and nonaromatic ketones were also evaluated. To achieve the best enantioselectivity, catalyst C25 was used for the preparation of 2za and 2zc-e (82-90% ee), and C26 was used for 2zb (90% ee) (see ESI † for the structures of tertiary-squaramidebased C25 and C26). The generality of this method was thus illustrated. We also carried out enantioselective reductive debromination of 2-bromo-2-uoro-tetralone 3; 2a was obtained with a similar excellent enantioselectivity but in a lower yield than that achieved via the dechlorination of 1a. Note that these a-uoroketones are stable under the reaction conditions, as indicated by the ee value being maintained and no deuorinated parent ketone being observed when the reaction was performed for a longer time. Moreover, the determination of 0% ee of 2a before the reaction completed excluded the possibility of kinetic resolution of 2a.
The aforementioned success encouraged us to further evaluate this protocol in constructing chiral tertiary a-chloroketones from a,a-dichloroketones. The preliminary studies with 2,2-dichloro-tetralone 4a as the starting substrate showed that the slightly modied reaction conditions, wherein the reductant was Amine-2 and the temperature was À30 C, afforded chiral 2-chloro-tetralone 5a in 87% yield and with 92% ee within 12 h (eqn (1)). 25 However, the ee value of this product Table 2 Dehalogenative protonation for the synthesis of chiral secondary a-fluoroketones a a Reaction performed on a 0.10 mmol scale. Yields were determined from the isolated material aer chromatographic purication. Enantiomeric excesses were determined by HPLC analysis on a chiral stationary phase. b When the reaction was performed on a 1.0 mmol scale, t ¼ 48 h, 86% yield, 97% ee. c Catalyst C25 was used instead of C1. d Catalyst C26 was used instead of C1. e The modied reaction parameters: Amine-2 instead of Amine-1, no H 2 O in THF at 40 C for 60 h. was observed to deteriorate when ash column chromatography was performed, revealing that such a tertiary stereocenter is considerably labile.
Accordingly, we planned to explore a sequential strategy involving subsequent one-pot reduction of ketone to chlorohydrin, thus facilitating purication. Aer careful examination, we found that the desired product 6a was obtained in 79% yield over two steps with a maintained ee of 92% and >20 : 1 dr when the reaction mixture from the rst reductive dechlorination step was directly treated with 3.0 equiv. of diisobutylaluminum hydride (DIBAL-H) at À40 C under an argon atmosphere for 3 h (Table 3). In the presence of the established tandem protocol, a series of 2,2-dichloro-tetralones with different substituents on the aryl ring were examined, leading to the important chlorohydrins 24 6b-k in 60-82% yield with 84-94% ee and >20 : 1 dr. With respect to 2,2-dichloro-1-indanone, the reaction conditions were further modied for achieving higher enantioselectivity (see footnote c). As a result, product 6l was obtained in 65% yield, 87% ee and >20 : 1 dr. Inarguably, these results demonstrated that a secondary C-Cl bond was successfully introduced at the a-position of these cyclic ketones with high enantioselectivities, indicating the compatibility of this catalytic system even though these a-secondary C-Cl bonds of ketones are unstable. The attempt with linear ketones showed that C25 was a suitable H-bonding catalyst, leading to a series of a-chloroketones 5m-o in 73-83% yield with 80-82% ee. Because the products were stable during ash column chromatography, the further reduction of ketone to alcohol was unnecessary. Notably, some amount of doubly dechlorinated products was found before the reactions were nished. Hence, we tested the ee of 5a at different reaction stages and aer prolonging the reaction beyond completion and found that the amount of tetralone continued to increase. The results demonstrated that all the ee values were consistent, indicating that kinetic resolution is not possible in the dechlorination of monochlorinated ketones.
Given the almost identical utility of chiral C-Cl and C-Br bonds in organic synthesis, 2,2-dibromo-tetralone 7 was selected as a representative to evaluate the viability of enantioselective photoredox debrominative protonation. The investigation revealed that the chiral a-bromo-tetralone stereocenter exhibited lability similar to that of the chloro variant. Accordingly, the cascade reaction was performed, and chiral bromohydrin 8 was successfully achieved in 61% yield, 90% ee and >20 : 1 dr (eqn (2)).

Mechanism studies
The mechanism of these reactions via a photoredox dehalogenation-protonation process is established. 18,19 Our Stern-Volmer experiments demonstrated the occurrence of the same catalytic cycle in which the transformation was triggered by reductive quenching of the photoredox catalyst (see the ESI †). Notably, all chiral catalysts (C1, C25 and C26) contain a squaramide group, and this diketone moiety (e.g., E red 1/2 of C1 ¼ À0. 22,À0.79 V vs. SCE in CH 3 CN) should be reduced by DPZc À more readily than a,a-dihaloketones (e.g., E red 1/2 of 1a ¼ À0.80, À1.23 V vs. SCE in CH 3 CN). In this context, the structure of C1 aer the transformation of 1a to 2a was analyzed. As shown in Scheme 3a, C1 was recovered in 90% yield and no reduced product was observed. In the absence of 1a, a high recovery yield of C1 was still obtained (Scheme 3b). These Table 3 Dechlorinative protonation for the synthesis of chiral secondary a-chlorohydrins or a-chloroketones a a Reaction performed on a 0.10 mmol scale. The reaction time stated below the product structure refers to the reductive dechlorination process. Yields were determined from the isolated material aer chromatographic purication. Enantiomeric excesses were determined by HPLC analysis on a chiral stationary phase. The dr was determined by the crude 1 H NMR analysis. b 2.0 mL DCE was used. c The modied reaction parameters: 20 mol% C1, 1.5 equiv. Na 2 CO 3 in DCM at À60 C. d Catalyst C25 was used instead of C1.
Scheme 3 a Determining the structure of C1 after the reaction completed. b Determining the structure of C1 when without the starting substrate 1a. c The reaction of 1a to 2a was performed under standard reaction conditions but with C28 instead of C1 as the chiral catalyst.
results suggest a perfect tolerance of C1 under the reaction conditions, likely because of its evidently lower concentration than the substrates in the reaction system. For the formation of the a-secondary stereocenter, the use of structurally different Amine-1 and Amine-2 as the reductants offering similar enantioselectivities (entries 1 and 6, Table 1) could roughly prove a protonation process rather than HAT. The deuterium labelling studies could further demonstrate this protonation process (see ESI †).
To probe the role of C1's bifunctional groups, i.e., tertiary amine and squaramide, the transformation of 1a was performed under the standard reaction conditions, as shown in Table 1, but by using C1-derived quaternary ammonium salt C28 as the catalyst. Product 2a was obtained in 78% yield with 84% ee (Scheme 3c). The negligible change in enantioselectivity and the maintained absolute conguration suggest that the tertiary amine moiety of C1 might not participate in proton transfer but instead provides a steric factor only for establishing stereocontrol. From the similar enantioselective results in the construction of different C-X bonds with the same chiral catalyst C1 (see the results for 2a, 5a and 8 in Table 2, eqn (1) and (2), respectively), the hydrogenbonding interaction between the squaramide of C1 and halogen of ketones might be excluded. Meanwhile, the remarkably different enantioselectivity with C1 and C3 catalysts (entries 1 and 3, Table 1) indicates a slight possibility of the strongly acidic squaramide participating in proton transfer interchange to present H + to carbanion intermediates in the protonation step. Accordingly, a plausible transition state involving H-bonding interactions between the squaramide of C1 and ketone for facilitating stereocontrol and improving the chemoselectivity was proposed and is shown in the ESI. †

Conclusions
In summary, we have developed a novel asymmetric protonation strategy that is the visible-light-driven photoredoxcatalyzed dehalogenation-enantioselective protonation. Through the establishment of a dual organocatalytic system with a secondary amine reductant, a range of important cyclic and acyclic ketones with labile chiral secondary C-F, C-Cl and C-Br bonds at the a-position were collectively obtained in high yields and with high ee values. Although the stereocenters of these products are liable to proceed racemization in the reaction process and a competitive racemic background reaction exists in the reaction system, the satisfactory results suggest the amazing feasibility and compatibility of this strategy. Given the convenient preparation of alkyl halides, we expect this method to be extensively utilized in the construction of diverse signicant and, especially, complex chiral a-tertiary carbonyls and their variants, thus leveraging the broad applicability of enantioselective protonation in organic synthesis.

Conflicts of interest
There are no conicts to declare.