Organophotocatalytic Late-stage N-CH3 Oxidation of Trialkylamines with O2 in Continuous Flow

We report an organophotocatalytic, N-CH3-selective oxidation of trialkylamines in continuous flow. Based on the 9,10dicyanoanthracene (DCA) core, a new catalyst (DCAS) was designed with solubilizing groups for processing in flow which allowed harnessing of O2 as a benign reagent for late-stage photocatalytic N-CH3 oxidation of natural products and active pharmaceutical ingredients. These substrates bear functional groups which are not tolerated by previous methods. The organophotocatalytic process benefited from the flow parameters, affording cleaner reactions in short residence time of 13.5 mins and productivities of up to 0.65 g / day. Mechanistic studies found that catalyst derivatization not only enhanced solubility of the new catalyst compared to DCA, it profoundly diverted the photocatalytic reaction mechanism from singlet electron transfer (SET) reductive quenching with amines to energy transfer (EnT) with O2. Introduction The quintessential theme of medicinal chemistry is probing structure activity relationships. While strategies such as de novo and diversity-oriented synthesis (DOS) are powerful tools to achieve this task, late-stage functionalization (LSF) is gaining traction over the past decade as it offers a quicker route to access libraries of complex bioactive molecules.1,2 Among the myriad of methods that could be applied for LSF reported thus far, C–H functionalization is undeniably an attractive and potent addition to a synthetic chemist’s arsenal.1–5 This umbrella term has stretched in scope from traditional transition metal catalysis to organocatalytic, and photocatalytically-enabled transformations with recent examples applied to C–H functionalization of simple and complex amides through ionic6–8 or oxidative9–13 mechanisms. Trialkylamines and their proximal C–H positions are attractive loci for transformations especially as their privileged representation in their alkaloid family. The proclivity towards studying this class of compounds, however, is beyond their ubiquity; their relevance crosses the borders of natural sciences (Figure 1).14–18 Moreover, rapid synthetic access to structurally-diverse trialkylamines is desired in pharmaceutical research as their minute structural variations carry substantial pharmacological effects, for instance, in the pharmacological activity of opiates.19–21 However, aside from the intrinsic basicity of amines, their C(sp3)–H positions are relatively inert. Thus, access to derivatives are typically carried out via a stepwise fashion usually requiring initial demethylations of trialkylamine N-CH3 groups to via free N–H for subsequent transformations.22–28 That is until the renaissance of single electron transfer (SET) redox methods, partly driven by photoredox catalysis, which revolutionized the practice of organic chemistry29 allowing direct C-C bond formations or nucleophilic additions to benzylic amines and a few examples on simple aliphatic amines.30–36 Still, reports on strategies for LSF of complex substrates especially trialkylamines are rather scarce.37 Figure 1. Bioactive trialkylamines and relevant target sites for NCH3 C–H functionalization. Direct transformation of a trialkylamine’s N-CH3 to an N-formyl group is another worthy endeavor as the structure of (and mechanisms to access) N-formyl groups is relevant to oxidative metabolite research,38–40 present in natural products,14,41–44 and could serve as a synthetic handle for further modifications including Barbier type amidation,45 C–C cross-coupling reaction,46 amino-carbonylation of alkenes or alkynes,47 and cross-coupling with phenols or amines affording carbamates48 or ureas49. Classically, N-formyls are accessed from trialkylamines using toxic Ru(IV) or Os(IV) oxidants (Figure 2A)50–54. Recently, Yamaguchi and co-workers circumvented this via an elegant Cu(II)/Cu(I) and Nitroxyl radical catalyst system.55 Song and co-workers reported a transition metal free deconstructive formylation reaction.56 The main drawbacks of such previous methods are i) the incompatibility of redox sensitive functionalities (commonly found on complex pharmaceuticals) hence limiting their application to relatively simple amines, ii) the expense of reagents (hindered nitroxyl radicals and excess difluorocarbene reagents) which are economically impractical for scale-up. These current challenges motivated us to develop a catalytic method that: i) utilizes the relatively mild conditions of visible light photocatalysis and abundant, benign reagents (O2), ii) is applicable to complex pharmaceutically-relevant molecules as an LSF strategy, iii) is amenable to continuous flow processing in a scalable, safe process (Figure 2b).


Introduction
The quintessential theme of medicinal chemistry is probing structure activity relationships. While strategies such as de novo and diversity-oriented synthesis (DOS) are powerful tools to achieve this task, late-stage functionalization (LSF) is gaining traction over the past decade as it offers a quicker route to access libraries of complex bioactive molecules. 1,2 Among the myriad of methods that could be applied for LSF reported thus far, C-H functionalization is undeniably an attractive and potent addition to a synthetic chemist's arsenal. [1][2][3][4][5] This umbrella term has stretched in scope from traditional transition metal catalysis to organocatalytic, and photocatalytically-enabled transformations with recent examples applied to C-H functionalization of simple and complex amides through ionic [6][7][8] or oxidative [9][10][11][12][13] mechanisms. Trialkylamines and their proximal C-H positions are attractive loci for transformations especially as their privileged representation in their alkaloid family. The proclivity towards studying this class of compounds, however, is beyond their ubiquity; their relevance crosses the borders of natural sciences ( Figure 1). [14][15][16][17][18] Moreover, rapid synthetic access to structurally-diverse trialkylamines is desired in pharmaceutical research as their minute structural variations carry substantial pharmacological effects, for instance, in the pharmacological activity of opiates. [19][20][21] However, aside from the intrinsic basicity of amines, their C(sp 3 )-H positions are relatively inert. Thus, access to derivatives are typically carried out via a stepwise fashion usually requiring initial demethylations of trialkylamine N-CH 3 groups to via free N-H for subsequent transformations. [22][23][24][25][26][27][28] That is until the renaissance of single electron transfer (SET) redox methods, partly driven by photoredox catalysis, which revolutionized the practice of organic chemistry 29 allowing direct C-C bond formations or nucleophilic additions to benzylic amines and a few examples on simple aliphatic amines. [30][31][32][33][34][35][36] Still, reports on strategies for LSF of complex substrates -especially trialkylamines -are rather scarce. 37 Direct transformation of a trialkylamine's N-CH 3 to an N-formyl group is another worthy endeavor as the structure of (and mechanisms to access) N-formyl groups is relevant to oxidative metabolite research, [38][39][40] present in natural products, 14,[41][42][43][44] and could serve as a synthetic handle for further modifications including Barbier type amidation, 45 C-C cross-coupling reaction, 46 amino-carbonylation of alkenes or alkynes, 47 and cross-coupling with phenols or amines affording carbamates 48 or ureas 49 . Classically, N-formyls are accessed from trialkylamines using toxic Ru(IV) or Os(IV) oxidants ( Figure 2A) [50][51][52][53][54] . Recently, Yamaguchi and co-workers circumvented this via an elegant Cu(II)/Cu(I) and Nitroxyl radical catalyst system. 55 Song and co-workers reported a transition metal free deconstructive formylation reaction. 56 The main drawbacks of such previous methods are i) the incompatibility of redox sensitive functionalities (commonly found on complex pharmaceuticals) hence limiting their application to relatively simple amines, ii) the expense of reagents (hindered nitroxyl radicals and excess difluorocarbene reagents) which are economically impractical for scale-up. These current challenges motivated us to develop a catalytic method that: i) utilizes the relatively mild conditions of visible light photocatalysis and abundant, benign reagents (O 2 ), ii) is applicable to complex pharmaceutically-relevant molecules as an LSF strategy, iii) is amenable to continuous flow processing in a scalable, safe process ( Figure 2b).

Results and Discussion
Photocatalyst and process design. At the onset, we attempted to apply our previous trialkylamine photocatalytic oxidation conditions (developed for the functionalization of N-alkyl THIQs 34 ) to dextromethorphan (1a). Despite the many reports proposing reductive quenching of [Ru(bpy) 3  We sought an alternative strategy and contemplated the use of an organophotocatalyst, in order to avoid toxic, precious, and unsustainable transition metal-based photocatalysts. 57 Simultaneously, we envisaged leveraging continuous flow conditions to efficiently and safely handle O 2 as a simple, abundant, benign terminal oxidant, motivated by numerous reports of gas-liquid organophotocatalytic processes. [58][59][60][61] The rapid uptake of continuous flow processing in synthesis of fine materials and pharmaceuticals is worth noting, as is its innovative marriage with visible light irradiation which has drastically enhanced the efficiency, sustainability and safety of photochemical processes. [62][63][64] While TPP and Rose Bengal are among popular organophotocatalyst choices for the activation of alkaloids, they are primarily optimized for N-demethylations of opiates and tropanoids, 65,66 endocyclic C-H cyanations -toN. 67 or oxidations of benzylic amines. 68 We were particularly drawn to 9,10-dicyanoanthracene (DCA) as utilized by Santamaria and coworkers. Using DCA as a potent photooxidant (E 1/2 [ 1 DCA*/DCA •− ] = +1.99 V vs SCE) 57 and air as terminal oxidant, variable amounts of N-formyl side product (2a) were obtained in the Ndemethylation of 1a to give 3a (Scheme 2). 69,70 Our attempts using modified reaction conditions reaction in batch (conditions: 30 W white LED floodlamp, air bubbling, 6 h) yielded complex reaction mixture (crm) with 2a as a major component (for HPLC profile and spectral output of LED used, see SI). In a tubular coil flow reactor, 2a was still the major product but the reaction profile was significantly cleaner. However, the very poor solubility of DCA in MeCN often led to flow channel blockages and longer reaction times. Thus, a catalyst with enhanced solubility was required.
Intuitively, introduction of polar substituents should improve the solubility of compounds in polar aprotic solvents. Nitro-and Sulfonic acid-groups are good choices for aromatic compounds as the synthetic process to access them is straightforward. Glöcklhofer and co-workers reported the synthesis of a dinitro derivative of DCA with improved solubility. 71 Sulfonic acids on the other hand carry the advantage of further derivatization of sulfonyl chlorides. Inspired by the intermediates reported in the synthesis of a water-soluble DCA analogue, we began our catalyst synthesis ( Figure 3). 72 Anthraquinone-2,6-disulfonic acid 4, commercial or easily synthesized from cheap anthraquinone 73,74 , was reduced by activated Zn in aq. (NH 4 ) 2 CO 3 to afford anthracene-2,6-disulfonic acid 5 in good (65%) yield after acidic workup and recrystallisation from aq. KCl. Electrophilic bromination of the central ring of 5 gave 6 in high (80%) yield. At this stage, our synthesis deviated from the literature cyanation which digested the crude product (containing CuCN) in conc. HNO 3 and liberated toxic HCN gas. However, both Rosenmund von-Braun and Pd-catalysed cyanations failed to cyanate 6 due to its poor solubility in organic solvents. Literature guided us to an alternative strategy; in an attempt to disrupt the  stacking properties of anthracene disulfonic acids (ADS), Tohnai and coworkers reported their derivatization as organic amine salts. 75,76 They found amines bearing long chains (i.e. n-heptyl 76 and npentyl 75 ) minimized or prevented-stacking interactions of ADS as observed in their crystal structures. Inspired by these reports, we derivatized 6 to increase its solubility in polar aprotic organic solvents, increasing the success of cyanation. Instead of ammonium salts that would hinder characterization and photocatalytic reaction work-up, we achieved this covalently via sulfonamides. Chlorination of 6 with POCl 3 and subsequent trapping of 7 with secondary amines of various chain lengths gave 9,10-dibromoanthracene-2,6-disulfonamides 8a, 8b and 8c (DBAS) in 87, 90 and 74% yields, respectively. Pleasingly, Rosenmund von-Braun cyanation of 8a, 8b and 8c under microwave-assisted (15 min) or thermal (see Supporting Informatios (SI)) heating afforded 9,10-dicyanoanthracene-2,6disulfonamides 9a, 9b and 9c (DCAS) as brilliant yellow solids in 66%, 89% and 26% yields, respectively. We note that the entire synthesis is carried out on gram scale, without chromatography, with straightforward purification via recrystallisation. Photocatalyst 9b (henceforth referred to as 'DCAS') was progressed to evaluation in reactions since it: i) displayed the highest solubility in MeCN and PhCN solvents consistent with its calculated physical property values 77 suggesting it was the least lipophilic and had the highest topological polar surface area and ii) was obtained in the highest overall yield (42% over 5 steps). DCA and DCAS gave similar UV-vis spectra (Figure 3, right). Both have absorption maxima ( max ) at 420 nm and 395 nm, suggesting sulfonamides at the 2,6-positions hardly affect the absorptive properties of the dicyanoanthracene core. Emission spectra were similar for 1 DCA* and 1 DCAS*. Further characterization is described subsequently (and see SI).
Studies using a homogeneous liquid flow photoreactor. Next, DCAS was tested under some initial photocatalytic flow conditions (Table 1) in a commercial tubular coil continuous flow photoreactor (Vapourtec Ltd R-series/UV-150). Using 1a (12 mM) as our substrate and 5 mol% of DCAS at rt, a maximum yield of 25% for 2a (with 4:1 of 2a:3a selectivity) was obtained under recycling conditions (90 min) no matter whether dry air, O 2 , or (1:1) N 2 /O 2 were used (entry 2). The absence of catalyst (entry 2) or O 2 led to no reaction. Single pass conditions in the absence of LiClO 4 gave a similar yield (25%) and improved selectivity for 2a (entry 4, 2c was not detected). When the temperature was increased to 40 o C, the yield improved to 40% (entry 5). Under similar conditions but employing DCA as catalyst afforded 2a in 15% yield, confirming superiority of DCAS under flow conditions.
We hypothesized the formation of 2a in only low yield was due to limited oxygen solubility, because the reaction under N 2 protection led to a purple coloration in the post-reactor reaction mixture (see SI), an observation consistent with the formation of DCAS •− . When the purple post-reactor reaction mixture was collected and exposed to air, immediate discoloration back to yellow was observed. We note that the related parent structure DCA •− is well-known to be purple in color. 78 [14][15][16]. While doubling concentration to 48 mM or using a residence time as short as R T = 6.8 min negatively impacted the yield of 2b (entries 14-15), we found that yield was preserved at a residence time of R T = 13.5 min (entry 16). This doubled productivity of 2b to 0.65 g /day. Next, we tested the scope of the reaction (Table 3). Since isolations of polar formamides were oftentimes challenging and involved a weak chromophore, the following discussion deems 1 H NMR yields more representative of reaction efficiency.   Compounds 2c (59%) and 2d (67%) were obtained from natural products Tropane and (free alcohol-bearing) Atropine. Even Scopolamine, which has a free alcohol, ester, and an epoxide, afforded 2e in 62% yield with no nor-scopolamine detected, albeit requiring 2 passes through the reactor (total R T = 27 min). This contrasts with Santamaria and co-workers' conditions using DCA and without LiClO 4 , which afforded a 1:1 mixture of 2e : norscopolamine. 69,70 Compared to 2b, the yield of 2f was lower (34%) presumably due to the presence of the Si protecting group (DPMS) known to stabilize radicals and quench excited photosensitizers via different pathways. 83 Benzoyl-containing compound 2g was afforded in good yield 73%. Electron poor (CF 3 ) and electron rich (-OMe) substituents on the benzoyl group were tolerated equally, affording 2h (56%) and 2i (58%) respectively. We note both 2g and 2i are natural products; novel tropanoid compound 2g was recently isolated from Pellacalyx saccardianus and this synthetic method corroborated the proposed structure. 84 Compound 2g (Confoline) was isolated from Convolvulus subhirsutus and our method accessed it from Convolvamine in a single step. (in the literature, semi-synthesis of 2i was achieved by formylation of demethylated Convolvamine). 44 Compounds 2j to 2o were obtained from piperazines as common API fragments (such as those present in Sildenafil and Danofloxaxin). 85,86 Despite having 3 possible sites for functionalization (one exocyclic N-CH 3 and two endocyclic N-CH 2 -R sites), selective N-CH 3 oxidation afforded N-formyl compounds in respectable yields. We were surprised by the tolerance of halogen-bearing substrates, affording 2l (55%), 2m (21%), and 2n (30%) (as well as 2h) under the reaction conditions. Formation of DCA •− is well-known via the reductive quenching of 1 DCA* by trialkylamines 87,88 together with our aforementioned detection of the characteristic purple coloration of DCA •− under the reaction conditions. Photoexcited radical anions are known to facilitate reductive cleavages of aryl halides and other strong bonds, 87-92 while C-F bonds and N-Ts groups are prone to cleavage under reductive photocatalysis 93,94 or by photoexcited super electron donors. 95 Simple piperidine 2p (39%) was also tolerated. Our success with 2b, 2d, 2e, and 2p whose substrates bear free alcohol groups encouraged us to explore more complex molecules. Gratifyingly, conditions were successfully applied to macrolide antibiotics with dense functionalities (free alcohols, oxime ether, ketone). Erythromycin, clarithromycin and roxithromycin afforded 2r, 2s, and 2t in 61%, 44%, and 24% yields, respectively. However, compounds bearing benzylic amines, benzylic -hydroxy amines (separated by conformationally free rotating -bond), free carboxylic acids, and olefins such as 1u, 1v and 1w were unsuccessful. Benzaldehyde formation (C-N cleavage, possibly via endocyclic iminium ion formation and hydrolysis) and complex reaction mixtures were observed for these substrates. Our initial hypothesis was thus based on the SET mechanism proposed by Santamaria ( Figure 5). 69,70 In this premise, 1 DCAS* underwent reductive quenching by trialkylamine 1 and DCAS was regenerated from oxidation of DCAS •− via O 2 (initial SET reactions of trialkylamines were also reported as the main pathway for photocatalytic oxidations to N-oxides and demethylations using thiazine and fluoresceine dyes). 99 Trialkylamine radical cation 1' would undergo transformation to iminium ion 10, either via hydrogen atom transfer (HAT) with superoxide or by deprotonation to -amino radical 1" followed by oxidation of 1" by peroxyl radical. In the presence of H 2 O, 10 forms 11a which undergoes either hydrolysis to secondary amine or further oxidation to product 2. Alternatively, peroxide anion is more nucleophilic than H 2 O and reacts with 10 to form product 2 via the collapse of 11b. Intermediate 11b could also form via radical combination of 1" and peroxyl radical.   In support of this initial hypothesis, control batch experiments employing 2.0 equiv. of DCA and DCAS under strict N 2 protection in MeCN afforded clean conversion of 1a in both cases to a 1 : 1 mixture of 1a : 3a (Figure 7, top). Upon exposure to light, the reaction mixtures darkened from pale yellow to purple ( Figure  6, bottom). Upon removal of light and exposure to air, the reaction mixtures lightened to yellow again. These observations are hallmarks of cyanoanthracene radical anions and the Ndemethylation reaction confirms the SET oxidation of 1 to 1' by the organophotocatalyst. Here, PhCN was needed instead of MeCN as solvent to improve solubility in the case of the stoichiometric quantities of cyanoanthracenes employed. In the absence of additional base, 1a deprotonates 1a •+ to afford 1" (10 after a second SET oxidation), meaning the reaction fundamentally could never exceed 50% conversion. Although both reactions were slow (days), the reaction employing DCAS was markedly faster, presumably either due to its higher excited state oxidation potential or, more likely, its higher solubility (DCA was not fully partially soluble in this batch reaction). In light of the preceding discussion in support of the mechanism in Figure 5, one would expect 1 DCAS* to undergo more rapid fluorescence quenching than 1 DCA* by amines. Very surprisingly, the opposite was true. The Stern-Volmer quenching rate constant for 1 DCAS* by 1a (k q = 1.44 × 10 9 M -1 s -1 ) was two orders of magnitude smaller than for 1 DCA* (k q = 1.69 × 10 11 M -1 s -1 ). 100,101 Presumably, either i) the 2-methoxyethyl groups of DCAS inhibit bimolecular quenching events by sterically obstructing approach of the trialkylamine, or ii) aggregation of DCA accelerates its reductive quenching by trialkylamines which is broken up in the case of DCAS.
Elsewhere, 1 DCA* is known as an efficient singlet oxygen sensitizer (k q = 4.3 x 10 9 M -1 s -1 ) via a photosensitized energy transfer (E n T) mechanism. 102 The high quantum yield (reaching almost 2.0) supports the generation of two molecules of 1 O 2 per 1 DCA*. Yet, comparison of quenching rate constants revealed that 1a and 1x (N-methyl tetrahydroisoquinoline) as surrogate trialkylamines quench 1 DCA* more efficiently than O 2 does ( Figure 7). Furthermore, our enhanced yields with increasing [O 2 ] in solution are inconsistent with an initial SET between 1 DCAS* and amine, as O 2 should decrease the population of 1 DCAS* available for "initial SET." 99 This was reflected in the relative intensity change of light transmitted through the coil of the aforementioned tubular flow reactor as detected by an online fibre-optic transmission spectrometer probe. When an aerated reaction mixture of DCA (5 mol%) + 1a (12 mM) was compared with an aerated solution of DCA only, the latter led to strong absorption (Figure 8, A), indicating a larger steady-state concentration of DCA directly afforded via the rapid E n T quenching of 1 DCA* by O 2 . On the other hand, the reaction mixture (Table 1, entry 6) gave minimal absorption of light (Figure 8, B). Reductive quenching of 1 DCA* by 1a, even faster than quenching by O 2 , does not directly afford DCA but affords DCA •− whose absorption 92 is red-shifted far into the visible (≥580 nm). In contrast, k q for quenching of 1 DCAS* by O 2 was comparable, if slightly higher than that of 1 DCA*, and was markedly (100x) higher than k q for quenching of 1 DCAS* by 1a.
The reaction mixture (Table 1, Figure 9A). Firstly, when -terpinine was employed as a substrate, ascaridole was formed in 63% yield as quantified by 1 H NMR. Endoperoxide formation is a hallmark reporter for 1 O 2 via its Diels-Alder [4+2]-cycloaddition with dienes. [103][104][105] Secondly, the presence of DABCO as an additive inhibited conversion in 1b's reaction. We confirmed this was not due to it competing for 1 DCAS* as a reductive SET quencher, since the rate constant (k q = 7.28 x 10 8 M -1 s -1 ) showed it was an even less efficient quencher than O 2 or 1a/1b. Rather, DABCO is a wellknown physical quencher of 1 O 2 , 107,108 as demonstrated by the linear correlation between the reciprocal relative rate and [DABCO]. 107,108 In summary, increased efficiency of DCAS over DCA in the reaction is not only attributed to enhanced solubility in flow. Sulfonamide substitution at the 2,6-positions of the cyanoanthracene markedly inhibited reductive quenching of 1 DCAS* by trialkylamines, diverting the mechanism to 1 O 2 sensitization (a similar "steric-bulk" strategy was employed the literature with tert-butyl substituents preventing EDA complexation between the catalyst and substrate which are unproductive to their reactions). 109   Precious metal Ru-and Ir-based polypyridyl complexes are well known to participate in both photosensitization (E n T) and photoredox catalysis (SET), where structural tuning of ligands can effect switching between divergent pathways. To our knowledge, such a concept has rarely been exploited in organophotocatalysis on the same core, with privileged organophotocatalyst structures developed either for SET or E n T pathways. Consistent with the lack of -stacking aggregation in the XRD of DCAS (distance between -planes of anthracene = 13.60 Å, Figure 9, B), we tentatively propose that the bulky, freely-rotating sulfonamide substituents hinder bimolecular (or unimolecular) 101 quenching events with trialkylamines. Notably smaller O 2 outcompetes trialkylamines to reach the cyanoanthracene core. In lieu of i) the high oxidation regioselectivity for N-CH 3 over N-CH 2 -R groups, ii) the failure of simpler/less constrained trialkylamine substrates in favor of more constrained substrates, and iii) redox potentials indicating endergonic SET between trialkylamines (E p ox >+0.5 V vs SCE) and 1 O 2 (E p red >+0.1 V vs SCE), 110 an E n T followed by HAT mechanism is proposed ( Figure 10). Photoexcitation of DCAS affords 1 DCAS* which undergoes E n T with 3 O 2 The generated 1 O 2 interacts with the trialkylamine via a well-studied exciplex, 102,107,110,111 which can undergo one of two pathways. Firstly, HAT forms 1" and liberates a peroxyl radical. Further oxidation of 1", followed by combination with O 2 •− (and subsequent HAT, such as with solvent) affords 11b, which could also be formed via radical combination of 1" with proximally-generated peroxyl radical. Liberation of H 2 O affords 2. 3 DCAS* is annihilated by a second molecule of 3 O 2 to regenerate DCAS.
In a recent study of Rovis and co-workers' photocatalytic functionalizations of cyclic amines, 112 they proposed that a reversible and fast HAT is responsible for the endocyclic selectivity. In our case, we deem that the HAT between the singlet oxygen and amine substrate is irreversible, thus sterics govern the selectivity (i.e. at the N-methyl position). Further discussion on the selectivity of singlet oxygen reactions with related amines is ongoing. [113][114][115] In the case of less-constrained trialkylamines, the 1 O 2 -bound exciplex can react promiscuously with endocyclic / non N-CH 3 positions and other functional groups of substrates (e.g. benzylic groups, free alcohols) leading overall to degradation.

Conclusion
Herein we demonstrate the use of DCAS as a new organophotocatalyst for late-stage oxidation of pharmaceutical agents with trialkylamine moieties using molecular oxygen. Redox sensitive functionalities were tolerated allowing N-formyl functionalization of alkaloids and macrolide antibiotics in good yields with excellent selectivity. Succinct synthesis of N-Formyl tropanoids were achieved in continuous flow. The small reaction volume at a given time allowed safe handling of O 2 under back pressure, shortening reaction (residence) times to several minutes and unleashing synthetically useful productivities (0.65 g / day). Mechanistic insights demonstrate how seemingly minor structural variations in an organophotocatalyst can not only increase solubility, but profoundly divert the excited state mechanism from photoredox catalysis to photosensitization. With the generation of 1 O 2 revealed, our study provides one of a few examples of natural product synthesis using 1 O 2 as a reagent. 116 Ongoing, deeper studies probe the nature of interactions between DCAS, O 2 and trialkylamine quenchers.

ASSOCIATED CONTENT Supporting Information
Experimental procedures, optimization studies, 1 H and 13 C spectra of all novel compounds, XRD data, photophysical spectroscopic investigations and LC-MS/NMR data from which conclusions were drawn.