Enhancing the potential of enantioselective organocatalysis with light

Organocatalysis—catalysis mediated by small chiral organic molecules—is a powerful technology for enantioselective synthesis, and has extensive applications in traditional ionic, two-electron-pair reactivity domains. Recently, organocatalysis has been successfully combined with photochemical reactivity to unlock previously inaccessible reaction pathways, thereby creating new synthetic opportunities. Here we describe the historical context, scientific reasoning and landmark discoveries that were essential in expanding the functions of organocatalysis to include one-electron-mediated chemistry and excited-state reactivity. This Review discusses recent developments in the combination of organocatalysis and photochemistry for the activation of molecules, which has enabled previously inaccessible reaction pathways and influenced many fields of chemical research. Organocatalysis uses small, chiral organic molecules as catalysts and has been widely applied to asymmetric reactions in the past few decades. In this Review, Mattia Silvi and Paolo Melchiorre look at how, in combination with photochemical reactivity, the field has moved from traditional two-electron-pair reactivity to include one-electron and excited-state chemistry. The merger of these two fields has expanded the scope of asymmetric organocatalysis substantially, providing access to previously unavailable synthetic prospects, and this combination shows promise for developing new stereocontrolled catalytic processes.


Merging organo-and photoredox catalysis
It is of interest to consider why organocatalysis was combined with photochemical reactivity, and what motivated the exploration beyond the established boundaries of two-electron-pair reactivity. As is often the case in science, progress was spurred by a specific goal that could not be achieved with the available technologies. Here, that goal was the intermolecular enantioselective α -alkylation of carbonyl substrates with alkyl halides (Fig. 1a) using an enamine-mediated catalytic process. It is important to understand why this simple transformation attracted great interest in the enantioselective-catalysis community 18 . The α -alkylation of carbonyl compounds is among the most important of the classical synthetic reactions 19 . Generally, the process requires the preformation of stoichiometric metal-enolate nucleophiles that undergo an S N 2-type reaction with alkyl halides 20 . The development of an enantioselective catalytic version, however, has proven difficult, with the few reported methodologies being limited in scope 21,22 . Clearly, seeking to develop catalytic asymmetric methods that could directly functionalize unmodi fied carbonyl substrates was ambitious, and enamine-based chemistry was considered to be the most promising approach. This began with Gilbert Stork's fundamental studies 23 in the 1960s, which taught organic chemists that stoichiometric enamines could react with alkyl halides via S N 2 pathways. With the advent of enamine-mediated catalysis 7 , it was thought that implementing a direct stereoselective intermolecular α -alkylation of aldehydes would be not only feasible, but also straightforward. However, this synthetic target turned out to be much more difficult than expected 24 . The main reason was the modest reactivity of alkyl halides, which complicates the ionic alkylation step while favouring side reactions, for example N-alkylation of the Lewis-basic amine catalysts and self-aldol condensation.
In 2008, David MacMillan's group recognized that the main hurdle to overcome was intrinsic to the ionic S N 2 pathway 25 . Therefore, they used alkyl bromides not as electrophiles but as precursors for generating radicals. The underlying idea was to exploit the innate tendency of electron-deficient radicals to react rapidly with π -rich olefins, enabling the formation of carbon-carbon bonds that were otherwise difficult to make 26 . A ruthenium-based polypyridyl photocatalyst 5 ([Ru(bpy) 3 ] 2+ , in which bpy is 2,2′ -bipyridine) was used to generate open-shell species readily from α -bromo carbonyl compounds 2 (Fig. 1b). Before this point, photocatalyst 5 had a rich history as a single-electron transfer (SET) catalyst for inorganic applications 27 , but had limited use in synthetic chemistry 28 .
The reaction mechanism (Fig. 1c) is based on the integration of two independent catalytic cycles. On one side, the photoredox cycle proceeds Organocatalysis-catalysis mediated by small chiral organic molecules-is a powerful technology for enantioselective synthesis, and has extensive applications in traditional ionic, two-electron-pair reactivity domains. Recently, organocatalysis has been successfully combined with photochemical reactivity to unlock previously inaccessible reaction pathways, thereby creating new synthetic opportunities. Here we describe the historical context, scientific reasoning and landmark discoveries that were essential in expanding the functions of organocatalysis to include one-electronmediated chemistry and excited-state reactivity.

Review ReSeARCH
through the reductive cleavage of 2, instigated by SET reduction from the Ru(i) intermediate 7, to afford electrophilic radical 8. Concurrently, in the organocatalytic pathway, enamine Ia is generated by the condensation of organocatalyst 4 with aldehyde 1. Then, the ground-state chiral enamine stereoselectively traps radical 8 to form the stereogenic centre within α -amino radical 9 with high fidelity. In the original study, it was proposed that this electron-rich intermediate 9 was finally oxidized by the excited state of the Ru(ii) photocatalyst 6, a SET event that closes the photoredox cycle while affording iminium ion 10. Hydrolysis of the latter species furnishes the α -alkylation product 3 while regenerating the catalyst 4. Luminescence quenching studies revealed that the reducing [Ru(bpy) 3 ] + species 7 was initially generated by the oxidation of a sacrificial amount of enamine Ia by the excited * Ru(ii) catalyst 6. Later, mechanistic investigations established a radical chain manifold as the main reaction pathway 29

Generic mechanisms of organocatalytic reactivity
Organic catalysts can exert their functions by following two different substrate activation patterns (see Box 1 Figure): Covalent-based modes of activation exploit the ability of an organic catalyst to covalently bind a substrate in a reversible fashion and form a reactive intermediate that can participate in many reaction types with consistently high enantioselectivity. Chiral primary and secondary amines belong to this class, activating carbonyl substrates via the formation of nucleophilic enamines I (refs 7, 87) (from enolizable aldehydes and ketones), electrophilic iminium ions II (refs 8, 88) (from unsaturated carbonyl compounds), and α -iminyl radical cation intermediates III (ref. 40) (upon single-electron oxidation of enamines by a chemical oxidant). N-heterocyclic carbene catalysts 89 offer an alternative activation mechanism for aldehydes, conferring an inverted (umpolung) reactivity to the normally electrophilic carbonyl carbon atom upon formation of Breslow intermediate IV (ref. 90), which acts as an acyl anion equivalent 91,92 . These activation modes, which rely on strong, directional interactions, enable the stereoselective functionalization of unmodified carbonyl compounds at the ipso, α , and β positions.
Non-covalent approaches are based on the cooperation of several weak attractive interactions between the catalyst and a basic functional group on the substrates 93 . Although the catalyst-substrate interactions are generally weaker and less directional than their covalent counterparts, they operate in concert to ensure a high level of transition state organization, resulting in a high degree of enantioselectivity. Hydrogen-bonding activation 74,94 , phase-transfer catalysis 6,70 , anion-binding activation 95  Review ReSeARCH (Fig. 1d). The photoredox catalyst therefore initiates a self-propagating radical process that is sustained by the ability of the α -amino radical 9 to regenerate radical 8 by directly reducing the organic bromide 2. The same reaction can be achieved by replacement of the ruthenium photocatalyst with organic dyes 30 or different metal-based polypyridyl complexes 31 .
This study has had many far-reaching implications. Synthetically, by combining enamine-mediated catalysis with the action of a photoredox catalyst, mechanistically related enantioselective α -alkylation reactions have been developed (Fig. 1e), including trifluoromethylation 32 , benzylation 33 , and cyanoalkylation 34 processes. The enantioselective α -alkylation of 1,3-dicarbonyl substrates to generate quaternary carbon stereocentres, which are synthetically useful yet difficult to form, has also been achieved 35 . However, the main impact of this study was the demonstration that radical intermediates could be generated at ambient temperatures, simply by using a photocatalyst activated by visible light. This meant that the tools and the mechanisms for stereocontrol of enantioselective organocatalysis, which require mild conditions for optimal efficiency, could be successfully applied within radical reactivity patterns. These studies, along with other investigations dealing with non-stereocontrolled transformations 36,37 , also laid the foundations for the development of the field of photoredox catalysis 17 . At present, synthetic chemists are exploring the benefits of integrating the activity of photoredox catalysts with other catalytic systems, including metal-based catalysis 38 and chiral Lewis acid catalysis 39 , although these aspects fall out of the scope of this Review.

Dual-catalyst systems-covalent organocatalysis
The use of redox-active photocatalysts in conjunction with wellestablished chiral organocatalytic intermediates other than enamines Box 2

Light in organocatalysis
Two main strategies, the dual-and the single-catalyst approach, have been used to successfully combine organocatalysis and photochemical reactivity (see Box 2 Figure). Dual-catalyst approach. In this strategy, the activity of a photoredox catalyst 17 synergistically combines with the generic mechanisms of activation that define the ground-state reactivity of chiral organocatalytic intermediates. This approach involves the use of metal or organic photocatalysts that absorb light in the visible region; upon excitation, these photocatalysts either remove an electron from or donate an electron to simple organic substrates. This single-electron transfer mechanism facilitates access to radical species under mild conditions 100 . The unique reactivity of such photocatalytically generated open-shell intermediates enables the expansion of the organocatalytic functions from a polar to a radical reactivity domain. The field of photoredox catalysis, which is a fast-moving area of modern synthetic chemistry 17 , has led to the development of many novel synthetic methodologies. Single-catalyst approach. This strategy exploits the ability of organocatalytic intermediates to reach an excited state directly upon light absorption, and to participate in the activation of substrates, without external photocatalysts. At the same time, the chiral organocatalyst ensures effective stereochemical control. This approach demonstrates that the synthetic potential of organocatalytic intermediates is not limited to the ground-state domain, but can be expanded by exploiting their photochemical activity. By bringing an organocatalytic intermediate to an electronically excited state, light excitation unlocks reaction manifolds that are unavailable to conventional ground-state organocatalysis.   25 . c, The originally proposed closed catalytic cycle 25 . d, The key propagation step of the radical chain mechanism 29 . e, Further synthetic applications of this dual catalytic strategy for the direct stereocontrolled α -alkylation of aldehydes: trifluoromethylation 32 , benzylation 33 , cyanoalkylation 34 , and the formation of quaternary carbon stereocentres 35 . CFL, compact fluorescence lamp.

Review ReSeARCH
has been successively explored. Two examples of such strategies are given below.

Merging SOMO activation and photoredox catalysis
The example of singly-occupied molecular orbital (SOMO) activation illustrates how the combination of established modes of organocatalytic reactivity with photoredox catalysis could lead to unconventional transformations. SOMO activation, first introduced in 2007 (ref. 40), exploits the SET oxidation of enamines I by a chemical oxidant, which generates an electrophilic α -iminyl radical cation III amenable to a range of open-shell reactions. Because III can be stereoselectively intercepted by electron-rich functionalized olefins (for example, allyl silanes), the subsequent α -alkylation products result from umpolung reactivity. The main drawback of this strategy is that it requires an excess of stoichiometric oxidant. This issue was solved using a lightactivated catalyst to trigger the key SET oxidation of enamines to generate the intermediate III (ref. 41) (Fig. 2a). The milder radical-generation conditions offered the possibility of intercepting III with unactivated olefins, such as simple styrenes, in a stereocontrolled fashion 42 (path i in Fig. 2a). By avoiding the use of organic halides, this approach further expanded the potential of the organocatalytic intermolecular α -alkylation technology. The chemistry required the combination of organocatalysis with both an iridium photoredox catalyst 43 , which generated intermediate III, and a hydrogen atom transfer 44 thiol catalyst, which reduced the intermediate V resulting from the radical addition to the styrene.
The chemistry of α -iminyl radical cation III, generated under photoredox conditions, is not limited to radical addition manifolds. It can be expanded to realize unconventional and difficult-to-achieve transformations, such as the direct β -arylation of saturated carbonyl substrates 45 (path ii in Fig. 2a). The allylic C-H bonds in intermediate III are sufficiently weakened to enable proton abstraction by a suitable base, such as DABCO (1,4-diazabicyclo[2.2.2]octane), giving the β -enaminyl radical intermediate VI. This species can undergo radical coupling with the long-lived radical anion 13, generated by SET reduction of 1,4-dicyanobenzene 12 from an iridium(iii) photocatalyst. This bond-forming event, which is governed by the persistent radical effect 46 , forms a new carbon-carbon bond at the β -position of the original carbonyl. The strategy is synthetically appealing, given the lack of alternative methods for the direct β -functionalization of carbonyl substrates bearing saturated alkyl chains; however, only a single enantioselective example has been reported so far. Nonetheless, this study provided an initial demonstration that classical organocatalytic tools, such as the chiral amine catalyst 14, could serve to control the stereochemical outcome of a radical coupling event, which is greatly complicated by its intrinsic high rate 47 .
Overall, the studies detailed in Fig. 2a indicate that the native reactivity of an established organocatalytic intermediate (that is, enamines) can be switched from a closed-shell to an open-shell manifold with a lightactivated photoredox catalyst. These studies also highlight the ability of traditional organic catalysts, generally used in enantioselective ionic processes, to control the geometry of the ensuing radical intermediates (such as III and VI) while creating a suitable chiral environment for stereocontrolled bond formation.

Merging iminium-ion and photoredox catalysis
Iminium-ion activation has found many applications in ionic domains, facilitating the conjugate additions of soft nucleophiles to the β -carbon atom of unsaturated carbonyl compounds. However, the development of a stereoselective trap of nucleophilic radicals has not been trivial. This is because the addition of radicals to a cationic iminium ion II creates a reactive α -iminyl radical cation VII (Fig. 2b), an unstable intermediate with a high tendency to undergo β -scission 48 and reform the more stable iminium ion II. Recently, a strategy was reported that enabled enantioselective radical conjugate additions to β-substituted cyclic enones 15, forming quaternary carbon stereocentres with high fidelity 49

Review ReSeARCH
imine VIII, thus avoiding a possible competitive back-electron transfer. Finally, the long-lived carbazole radical cation in VIII, arising from the intramolecular SET, undergoes single-electron reduction from the reduced photoredox catalyst (PC red in Fig. 2b). This restores the neutral carbazole moiety while yielding the quaternary product 16. Notably, a photocatalyst (PC) both creates the nucleophilic radical and promotes the final redox process, which was identified as the turnover-limiting step of the overall reaction 50 .
This process provides a route to the generation of quaternary carbon stereocentres in an enantioselective manner, and exploits the tendency of radicals to connect structurally congested carbons because their reactivity is only marginally affected by steric factors 51 . However, radical-based catalytic enantioselective strategies had previously found limited application in the formation of quaternary carbon sterecentres 35 , and were not mentioned in a recent comprehensive review of available methods 52 . It appears that organocatalysis, in combination with photoredox catalysis, may offer effective tools to better exploit the intrinsic merits of radical reactivity.
N-heterocyclic carbene catalysis can also be used in conjunction with photoredox catalysts 53 . Although this approach has not yet been used to stereoselectively trap photochemically generated radicals, this target appears feasible.

Dual-catalyst systems-non-covalent organocatalysis
Non-covalent modes of organocatalytic reactivity have also been used with photoredox catalysis. So far there have been only a few reports, but these have offered solutions to important synthetic problems. The initial approaches used photochemical strategies to generate, in situ, reactive closed-shell species (for example, iminium ions 54 and singlet oxygen 55 ), which were successively intercepted by chiral organocatalytic intermediates. The first application of non-covalent organocatalysis in light-mediated radical chemistry provided a strategy for an asymmetric aza-pinacol cyclization 56 (Fig. 3a). The combination of the chiral phosphoric acid catalyst 20 and an iridium photoredox catalyst promoted the intramolecular reductive coupling between the ketone and hydrazone moieties within substrate 18 to furnish the syn 1,2-amino alcohol derivatives 19 with high enantioselectivity. The process is triggered by the formation of the ketyl radical 21, which is generated by a concerted proton-coupled electron transfer (PCET) process 57 driven by the cooperation of the photoredox and organic catalysts. PCET involves the simultaneous transfer of a proton and an electron in a single elementary step to enable processes that would be precluded by sequential, discrete proton and electron transfer steps. In this specific case, the direct SET reduction of the aryl ketone in 18 by the iridium photocatalyst alone would not be feasible. The ketyl radical 21, generated by PCET, is primed to cyclize into the hydrazone. Subsequent hydrogen atom transfer from a terminal reductant (Hantzsch dihydropyridine) to the generated hydrazyl radical leads to the final product 19. The high level of enantiocontrol indicates that the neutral ketyl radical 21 could maintain a considerable association, via tight hydrogen-bonding interactions, with the coordinating phosphate anion of the chiral Brønsted acid 20 during the course of the stereodefining cyclization. This study established the possibility of using concerted PCET to realize enantioselective radical processes by streamlining the preparation of radicals that are otherwise difficult to achieve. It also suggested the somewhat unexpected finding that the weak interactions inherent to non-covalent organocatalysis are appropriate for the selective binding of radical intermediates while channelling the resulting processes towards stereoselective manifolds.
Recently, Takashi Ooi expanded on this concept by using chiral P-spiro tetraaminophosphonium ion 25, which could selectively bind the anion-radical 26 via ion-pairing interactions 58 (Fig. 3b). This approach required the concomitant action of an iridium photoredox catalyst to reduce N-sulfonyl aldimines 22 and oxidize N,N-arylaminomethanes 23. The radical coupling of 26 and 27, governed by the chiral ion pair, gave the amine product 24 in high enantioselectivity. This study further demonstrated that organocatalysis can provide effective approaches to address issues in enantioselective radical chemistry that were previously considered unattainable, such as the precise stereocontrol of radical coupling processes 47 .

Organocatalysis in the excited state
The great potential of combining photoredox catalysis and organocatalysis lies mainly in the possibility of accessing open-shell species, the unique reactivity of which enables transformations not accessible through polar pathways. A different strategy has recently emerged, which offers possibilities to expand the field of organocatalysis. Researchers are exploring the potential of some chiral organocatalytic intermediates to reach an excited state directly upon the absorption of visible light, to enable new catalytic functions. The chemical reactivity of molecules differs fundamentally between the ground and electronically excited states 59 . For example, a molecule in an excited state is both a better electron donor (that is, a better reductant) and a better electron acceptor (that is, a better oxidant) than it is in the ground state 60 . This explains why, upon excitation, some organocatalytic intermediates can activate substrates via SET manifolds without the need for an external photocatalyst. At the same time, the chiral intermediate can provide effective stereochemical control over the ensuing bond-forming process. In this strategy, stereoinduction and photoactivation are combined in a single chiral organocatalyst.

Photochemistry of enamines
The reaction in Fig. 1b was also instrumental to the discovery that the synthetic potential of chiral enamines is not limited to the ground-state domain, and can be expanded by exploiting their photochemical activity. During investigations into the direct α -alkylation of aldehydes with electron-deficient alkyl bromides 28 using the organocatalyst 11 (Fig. 4a), a control experiment revealed that the reaction could be efficiently conducted in a stereoselective fashion under light illumination but without the need for any external photoredox catalyst 61 . Mechanistic studies revealed the ability of enamines Ib to trigger the photochemical formation Review ReSeARCH of radicals from alkyl bromides using two different photochemical mechanisms, depending on the substrate. The first mechanism 61 relies on the formation of electron donor-acceptor (EDA) complexes that absorb visible light 62,63 , which are generated in the ground state upon association of the electron-rich enamine Ib with the electron-deficient dinitrobenzyl bromide 28a (Fig. 4a, path i). Irradiation of the coloured EDA complex IX induces a SET event, enabling access to the radical intermediate 30a. A second radical generation mechanism 64 (Fig. 4a, path ii) exploits the ability of the enamine Ib to directly reach an electronically excited state (Ib*) upon absorption of light in the near-UV region and then act as a potent single-electron reductant. SET reduction of bromomalonate 28b induces the formation of radical 30b. Mechanistic studies 65 established that both enamine-mediated photochemical alkylations proceeded through a self-propagating radical chain mechanism 66 (Fig. 4a, right), in analogy to the reactions in the presence of a photoredox catalyst (Fig. 1d).
These studies demonstrated that enamines, which behave as nucleophiles in the ground state, can become reductants upon light excitation and trigger the formation of radicals. At the same time, ground-state chiral enamines control the stereochemical course of the radical-trapping event. This strategy was expanded to develop mechanistically related enantioselective α -functionalization reactions, including phenacyl alkylation 61 , amination 67 , and arylsulfonyl alkylation 68 of aldehydes and the alkylation of cyclic ketones 69 .

Photochemistry of other organocatalytic intermediates
The discovery that the catalytic functions of enamines could be expanded by exploiting their excited-state reactivity 61 motivated the quest for other chiral organocatalytic intermediates that could use similar photochemical mechanisms. Owing to their electronic similarities with enamines, enolates of type XI-generated in situ under phase-transfer catalysis conditions 70 (see Box 1) by deprotonation of cyclic β -ketoesters 31were suggested as suitable donors to facilitate the formation of photoactive ground-state EDA complexes 71 (Fig. 4b). Perfluoroalkyl iodides (R F I, 32) served as electron-accepting substrates, leading to the formation of the coloured EDA complex XII. A single-electron transfer, promoted by visible light, triggered the formation of the perfluoralkyl radical 34 (R F ·) through the reductive cleavage of the C-I bond. The electrophilic nature of R F · enabled it to be stereoselectively trapped by the chiral enolate XI, generated using a cinchonine-derived phase-transfer catalyst. This strategy provided access to enantio-enriched ketoester products 33 bearing a perfluoroalkyl-containing quaternary stereocentre.
Recently, it was also established that iminium ions can participate in photochemistry 72 (Fig. 4c). Condensation of the chiral amine catalyst 38 with aromatic enals 35 converts an achromatic substrate into a coloured iminium ion II. Selective excitation with a violet light-emitting diode forms an electronically excited state (II*), converting an electrophilic species into a strong oxidant 73   the β -enaminyl radical intermediate XIII along with the radical 39, which is generated upon irreversible fragmentation of the carbon-silicon bond. A stereocontrolled intermolecular coupling of the chiral β -enaminyl radical XIII and 39 then forms the stereogenic centre in the β -functionalized aldehyde product 37. The silane reagents 36 are non-nucleophilic substrates, which are recalcitrant to classical conjugate addition manifolds. Thus, in contrast to other examples of excited-state organocatalytic intermediates, the excitation of chiral iminium ions enables transformations that could not be realized by conventional catalytic asymmetric methodologies. A further difference is that the stereoselectivity is dictated by the chiral radical intermediate XIII, which governs the radical coupling event, and not by the ground-state iminium ion.

Non-covalent activation in asymmetric photochemistry
The photochemical organocatalytic strategies discussed so far all relied on the stereoselective interception of photogenerated radicals or radical ions in their ground states. But organocatalysis can also provide effective tools for catalytic stereocontrol in reactions of electronically excited intermediates. This is a difficult target because it requires the control of a photochemical process in a high-energy hypersurface, in which the action of a catalyst is greatly complicated by the absence of considerable activation barriers. Hydrogen-bonding catalysis 74 , which relies on several weak interactions to activate the substrates, has provided effective solutions. Chiral ketones, appropriately adorned with hydrogenbonding motifs 75 , were used to catalyse light-triggered stereocontrolled cyclizations 76 . The ketone-based organic catalysts effectively bind the substrate through a directional double-hydrogen-bond interaction, enabling the selective photoexcitation of a chiral catalyst-substrate complex. This ensures that the substrate resides in a suitable chiral environment when reaching an excited state. This strategy has been used successfully in both photo-induced redox processes and energy-transferinduced photochemical reactions. In an example of the latter 77 (Fig. 5a), a visible-light-absorbing thioxanthone moiety was incorporated within the catalyst 42. The lactam functionality of 42 was essential for binding the substrate 40 via a double-hydrogen-bond interaction. Meanwhile, the thioxanthone, upon excitation with light, activated the substrate via a proximity-driven Dexter energy transfer mechanism 59 and directed the [2 + 2] cyclization in the triplet energy hypersurface. The final product 41 was obtained with excellent enantioselectivity. Other strategies for the enantioselective catalysis of photochemical processes 15,78 have been successively developed. For example, it has been demonstrated that an intramolecular [2 + 2] photocycloaddition is promoted with high stereoselectivity by chiral thiourea catalysts 79 , which are traditional ground-state hydrogen-bonding organocatalysts 80 .

Photoexcitation of enzyme cofactors
Recently, a strategy has been reported that exploits the excited-state reactivity of common biological cofactors to enable enzymes to catalyse completely different processes than those for which they evolved. The natural reactivity of nicotinamide-dependent ketoreductases (KREDs) can be altered upon light excitation of the NADH/NADPH cofactor, which is bound to the enzyme active site 81 . KREDs have found extensive use in the preparation of chiral alcohols via the reduction of ketones 82 . This native polar reactivity arises from the ability of such enzymes to simultaneously bind, through non-covalent weak interactions, a carbonyl compound and the cofactor, and the tendency of NADH (or NADPH) to serve as a hydride source. Visible-light excitation, however, switches on a completely different reactivity in which the NAD(P)H becomes a strong reducing agent 83 , enabling access to radical manifolds (Fig. 5b). This photochemical behaviour was used in the enantioselective dehalogenation of racemic α -bromo lactones 43. Once the NAD(P)H and the substrate 43 are brought into close proximity in the active site of the enzyme, they can form a visible-light-absorbing EDA complex XIV, which triggers the formation of the prochiral radical intermediate 45 upon reductive cleavage of the substrate C-Br bond. The cofactor radical cation 46 drives the formation of the reduced chiral product 44.
This brief detour into enzymatic catalysis 84 highlights how the power of photochemistry to unlock unconventional reactivity is influencing other established fields of catalytic enantioselective synthesis, including metal catalysis 85 .

Conclusions and outlook
Over the last decade, combination with light has created exciting opportunities for expanding the scope of organocatalysis beyond conventional two-electron reactivity. Groundbreaking developments have taught synthetic chemists how to translate the generic mechanisms of activation, which govern the success of enantioselective polar organocatalysis, into the realm of excited-state reactivity and radical chemistry. The resulting light-driven methodologies are greatly expanding the way in which chemists consider the sustainable preparation of chiral molecules.
Major developments are probably still to come. This prediction is encouraged by the rapidly growing stream of innovation in photoredox catalysis, which is continuously offering new ways to generate radicals, and by the fact that the potential of excited-state organocatalytic reactivity is far from being fully revealed. Novel synthetic developments are expected to arise from the combination of photoredox catalysis and the activation mechanisms of ground-state organocatalysis. Considering the powerful photoredox methods available for the generation of radicals upon selective C-H activation of unactivated substrates (that is, PCET and hydrogen-atom-transfer mechanisms), the development of challenging enantioselective C(sp 3 )-C(sp 3 ) coupling strategies is likely to become an ambitious target. Efforts will also be devoted to the use of continuous-flow photoreactors, which may enable the scale-up of photochemical organocatalytic asymmetric methods 86

Review ReSeARCH
for the continued expansion of organocatalysis will be to fully explore the unique modes of reactivity revealed by the excitation of organo catalytic intermediates. Along these lines, traditional photosensitizers could provide reliable support by facilitating-by means of energy-transfer mechanisms-the generation of excited-state chiral intermediates that cannot be accessed by the direct absorption of light. This approach will require a deep understanding of the photophysical properties of the organo catalytic intermediates. It is expected that the combination of conventional physical organic chemistry tools and photophysical investigations will play an increasingly relevant role. Another force for innovation may arise from the integration of the photochemical activity of chiral organocatalytic intermediates within metal-mediated catalytic cycles, which could enable unconventional mechanisms for stereocontrolled bond formation. Finally, we expect great advances in the development of photochemical radical cascade processes, in which the unique excitedstate organocatalytic reactivities can be combined to provide powerful transformations for the one-step synthesis of complex chiral molecules 10 .