Photocascade chemoselective controlling of ambident thio(seleno)cyanates with alkenes via catalyst modulation

Abstract

Controlling the ambident reactivity of thiocyanates in reaction manifolds has been a long-standing and formidable challenge. We report herein a photoredox strategy for installing thiocyanates and isothiocyanates in a controlled chemoselective fashion by manipulating the ambident-SCN through catalyst modulation. The methodology allows redox-, and pot-economical ‘on-demand’ direct access to both hydrothiophene and pyrrolidine heterocycles from the same feedstock alkenes and bifunctional thiocyanomalonates in a photocascade sequence. Its excellent chemoselectivity profile was further expanded to access Se- and N-heterocycles by harnessing selenonitriles. Redox capability of the catalysts, which dictates the substrates to participate in a single or cascade catalytic cycle, was proposed as the key to the present chemodivergency of this process. In addition, detailed mechanistic insights are provided by a conjugation of extensive control experiments and dispersion-corrected density functional theory (DFT) calculations.

Thiocyanates possess ambident reactivity, as the electron density can be localized on sulfur and nitrogen atoms of the moiety. Here, the authors present chemoselective, divergent protocols to install thiocyanates and isothiocyanates, modulated by choice of photocatalyst.

Introduction

Control of chemoselectivity in the reaction involving ambident species is a long-standing issue in chemistry, and continuous efforts are being made to address its subtle nature in organic synthesis¹. The thiocyanate ion is one such moiety which, if successfully tamed, can selectively lead to organic thiocyanates and isothiocyanates, both having a vast presence in bioactive molecules as well as serving as valuable synthetic intermediates (Fig. 1A)^2–6. However, its chemoselectivity is dependent upon numerous factors, making it a perennially tricky problem to solve^1,7. Though alkyl isothiocyanates are more stable than their thiocyanate counterparts in thermodynamic terms, the alkylation of SCN⁻ usually leads to bonding through the S-center due to its kinetic preference over N-attack in both S_N2 and S_N1 reactions (Fig. 1B)⁸. Traditionally, alkyl thiocyanates have been accessed by subjecting alkyl halides to nucleophilic substitution by SCN⁻. In recent years, installation of thiocyanates onto readily available feedstock alkenes is also quickly coming to the fore as a route of choice. This has commonly been achieved by three-component reactions, either involving the addition of thiocyanate radical (generated by oxidation of thiocyanate salts) onto an olefinic π–bond or trapping of thiocyanate anion by a carbocationic intermediate, generated in the radical reactions of alkenes (Fig. 1C)^3,9. Isothiocyanation of alkenes, on the other hand, is quite rare^10,11 making direct access to organic isothiocyanates limited to reactions of primary amines with electrophilic thiophosgene or carbon disulfide under harsh reaction conditions^12,13. Hence, the rational design of a chemodivergent reaction manifold for selective incorporation of thiocyanate and isothiocyanate functionalities from the same reagent under mild conditions is highly desirable.

Fig. 1. Ambident reactivity of thiocyanate and present work.

A Biologically relevant thiocyanates and isothiocyanates. B Ambident reactivity profile of thiocyanate anion. C Reported three-component thiocyanation of alkenes. D This work: Catalyst-regulated thiocyanation and isothiocyanation by bi-functional alkylthiocyanate reagents.

Chemodivergent strategies that allow chemists to access structurally diverse products from a common set of starting materials are highly valuable¹⁴. Recently, the advent of photoredox catalysis has further matured this strategy to access divergent products under mild conditions^15,16. Unlike others, photocatalysts have the unique ability to promote reactions via both single-electron-transfer (SET)^17,18 and energy-transfer (EnT)¹⁹ processes where redox potential and triplet energy level of the catalyst dictate their actions, respectively. Tactically merging the above two processes serially or by repeating either of the processes sequentially, a photo-cascade platform could be devised to possibly harness molecular complexity in a step- and redox-economic way. In contrast to commonly encountered single-cycle catalytic platforms, multicycle cascade photocatalysis allows different reaction outcomes by sequentially engaging the first cycle product as an intermediate for the subsequent cycle. To be operative, this necessarily requires that the intermediate responds in any of the catalyst activation modes. Seminal works from Gilmour²⁰, Glorius²¹, König²², and others^23–25 have powerfully demonstrated this elegant strategy for achieving unprecedented transformations. We envisioned exploring this elegant platform for the controlled chemoselective incorporation of thiocyanate and isothiocyanate, en route to 5-membered S- and N-heterocycles from alkenes and thiocyanomalonates (TCMs) by a tactical regulation of the photocatalyst (Fig. 1D). Highly substituted 5-membered heterocyclic scaffolds are usually accessed by Lewis acid catalyzed [3 + 2] cycloaddition of donor-acceptor cyclopropanes (DACs) with dipolarophiles²⁶. While this strategy is efficient, the requirement of using DACs has impeded their broad applications, particularly for late-stage functionalization of complex molecules. The use of olefins, which are widely accessible and inexpensive feedstock, in the place of DACs to afford 5-membered heterocycles in a step- and redox-economical way under mild conditions provides a challenging yet appealing solution. The proposed radical route involving TCMs would enable the use of unactivated olefins and thus have the potential to overcome the limitations of existing ionic methods.

We hypothesized that redox-neutral carbothiocyanation of alkene 1 to the difunctionalized product 3’ may be achieved by photo-reductive cleavage of thiocyanomalonate 2 followed by SCN-transfer radical addition (Fig. 2A)²⁷. Coupling this ATRA process with a cyclization event may further allow access to 2-imino-tetrahydrothiophene (2-ITHT) product 3. On the contrary, subjecting 3’ to a consecutive SET event rather than cyclization can possibly lead to an oxidative radical-polar crossover²⁸ in a subsequent step which may allow thiocyanate 3’ to isomerize to the thermodynamically stable isothiocyanate 4’^29–32. N-cyclized 2-thiopyrrolidone 4 may then be readily accessed by a similar cyclization step as before. Splitting the overall reaction draft into ATRA and isomerization realms, a strategic key step was the judicious choice of photocatalyst which would selectively allow the alkyl thiocyanates (2 and 3’) to participate in a single or cascade catalytic cycles by matching its redox potentials. It is worth mentioning that the radical reactivity of alkyl thiocyanates, particularly as a thiocyanate group transfer agent, has never been explored³³. However, such a road map for controlling ambident functionality would lead to alkylchalcogenonitriles as multifunctional reagents, thus amplifying their synthetic utility. Continuing our research efforts on photoredox catalysis^34,35, we describe here the implementation of this blueprint enabling the chemodivergent synthesis of 5-membered S-, Se- and N- heterocycles from alkenes and thio(seleno)-cyano malonates, which highlights both the potential as well as the challenges of ambident thio(seleno)cyanates in organic synthesis.

Fig. 2. Reaction design and catalyst evaluation.

A Working hypothesis. B Electrochemical scale (vs. SCE).

Results and Discussion

Optimization of the reaction

Pursuing this idea, we set out to introduce thiocyanomalonates (TCMs) as bifunctional reagents³⁶ since some of the malonate congeners (-SePh, -TEMPO, -halo) have previously been exploited in radical reactions^27,37–39. Based on the reduction potential of the model thiocyanates 2a (E_1/2 = − 1.32 V vs. SCE) and 3a’ (E_1/2 = − 1.81 V vs. SCE), we anticipated that metal-free phenothiazine catalysts (PTH = N-arylphenothiazine) could deliver our objectives, courtesy of their superior and tunable redox capacity in the excited state (Fig. 2B)^40,41. After screening the reaction parameters based on our working hypothesis (see Supplementary Table 1 for detail optimization), we found that the reaction between 4-methylstyrene 1a and diethyl thiocyanomalonate 2a occurred in the presence of PTH1 under LED irradiation (λ_max = 390 nm) in toluene to produce metastable ATRA product 3a’, which during silica gel column chromatography or upon treatment with AlCl₃ in the same pot (just after the photoredox reaction) afforded the desired hydro thiophene product 3a in 38% yield (entry 1, Table 1). The pursuit of higher yield led us to the synthesis of PTH2, which delightfully furnished 3a in 89% yield (entry 2, Table 1). Further attempts at optimization by changing the catalyst (entries 3–5, Table 1) and solvent did not improve the yield of 3a (entries 6–9, Table 1). The S-selective catalyst PTH2 features a 4-CN substituent on its N-phenyl ring while its replacement with 4-OPh (PTH4) results in N-selectivity. It is to be noted that while the S-cyclization sequence needs assistance from Lewis acid, the N-cyclization variant i.e., 2-thiopyrrolidone 4a (81%), is obtained as a direct product just under photoredox conditions through a spontaneous domino difunctionalization-isomerization-cyclization sequence (entry 10, Table 1). 3a was obtained as the major product using commercially available PTH3 in a short reaction time, which got converted exclusively to 4a in the long run (entry 3 vs 11, Table 1). Other S-selective catalysts did not exhibit the same kind of chemodivergent isomerization in long reaction time (entries 12-13, Table 1). Overall, catalyst screening indicated that the electron-deficient N-aryl ring of PTH favored the S-heterocycle product, whereas unsubstituted or electron-rich N-aryl of PTH preferred N-heterocycle formation. These observations can be rationalized with their increasing reduction profile in the exited state⁴⁰, as desired to initiate the 2^nd cycle for isomerization. Control studies revealed the necessity of catalyst and light for the reaction to occur (entries 14-15 and 16, Table 1)⁴². Aerial oxygen has a negative impact on the reaction yield (entry 17, Table 1). Additional control experiments ruled out the impact of thermal heating on the success of the reaction (entries 18-19, Table 1). Finally, catalysts PTH2 and PTH4 were selected for accessing S-heterocycles and N-heterocycles respectively.

Table 1.

Optimization of the photochemical chemodivergent reaction^a

Entry	Catalyst	Solvent	Time	3a (%)c	4a (%)c
1	PTH1	Toluene	1 h	38	2
2	PTH2	Toluene	30 min	89 (86)d	7
3	PTH3	Toluene	20 min	68	21
4	PTH4	Toluene	20 min	70	22
5e	fac-Ir(ppy)3	Toluene	6 h	67	8
6	PTH2	CH3CN	45 min	73	6
7	PTH2	DMF	1 h	38	2
8	PTH2	DMSO	1 h	35	2
9	PTH2	1,2-DCE	30 min	76	5
10	PTH4	Toluene	24 h	3	81 (79)d
11	PTH3	Toluene	24 h	4	76
12	PTH2	Toluene	24 h	82	11
13e	fac-Ir(ppy)3	Toluene	24 h	61	12
14	-	Toluene	30 min	2	0
15	-	Toluene	24 h	4	0
16 f	PTH2	Toluene	30 min	0	0
17 g	PTH2	Toluene	30 min	78	5
18 h	PTH2	Toluene	24 h	3	0
19 h	-	Toluene	24 h	3	0

^aConditions: 1a (0.4 mmol), 2a (0.2 mmol), photo-catalyst (5 mol%), solvent (2 mL), degassed condition, irradiation with LEDs light (λ_max = 390 nm) with 100% intensity at 30–35 °C; after completion of photo-reaction, 1 equivalent AlCl₃ (0.2 mmol) was added at 0 °C and stirred for 1 h; ^bAlCl₃ is necessary only facilitate S-cyclization; ^cCrude ¹H NMR yield (%) using 1,1,2,2-tetrachloroethane as internal standard; ^dIsolated yield; ^e1 mol% catalyst loading, irradiation with LEDs light (λ_max = 450 nm); ^fReactions performed in dark; ^gReactions performed in open air; ^hReactions performed at 60 °C.

Substrate scope

We examined the generality of hydrothiophene construction by exploring a variety of alkenes under conditions A (Fig. 3). Aromatic alkenes with different electronic properties or steric hindrance performed well, leading to 2-ITHT derivatives (3a-k) in good to excellent yields (58%–88%). Vinyl heteroarenes bearing furan, thiophene, indole, and thiazole rings were suitable substrates delivering corresponding products (3l-o), although the yield of the latter two is poor (35%–38%) potentially due to the decomposition of these alkenes in our reaction conditions. Substituted-styrenes (both α- and β-) also survived to provide the desired products (3p-r) satisfactorily (58%–84% yield). Unfortunately, triphenylethylene and stilbene were unreactive, likely because of steric hindrance. Similarly, alkynes were found to be incompatible to this reaction, forming complex and insoluble mixtures upon reaction with the TCMs (see the list of incompatible substrates, Supplementary Table 6). Next, we turned our attention to the more challenging unactivated alkenes, which have proven to be difficult substrates under the radical carbothiocyanation process^35,43. Remarkably, this method can be extended to aliphatic alkenes. A range of terminal substituted aliphatic alkenes bearing various functionalities such as ester, ketone, alcohol, bromide, and sulfonate were well tolerated (3s-x) with good to excellent yield (63%–84%). Both α- (3y-aa) and β- (3ab,ac) substituted alkenes were also suitable substrates in this transformation (53%–72% yield). Surprisingly, electron-deficient acrylates (unsubstituted and α-substituted) provided the desired products (3ad-ag) as well, albeit obtained in modest yields (39%–47%), which has interesting mechanistic implications⁴⁴. The ethyl group of 3a could be substituted by other alkyl groups by choosing the appropriate TCM reagents (3ah-ak). Finally, the suitability of this mild photocatalytic method to pharmaceutically relevant molecules was demonstrated by derivatization of estrone (3al), naproxen (3am), oxaprozin (3an) as well as ribose-derived sugar substrate (3ao) in good yields (66%–78%), thus opening up the possibility of late-stage modification of biologically active molecules using this toolkit.

Fig. 3. Scope of the photochemical synthesis of 2-imino-tetrahydrothiophenes.

^aReaction conditions A: 1 (0.4 mmol), 2 (0.2 mmol), PTH2 (5 mol%), and degassed toluene (2 mL) under argon with irradiation of LEDs (λ_max = 390 nm) with 100% intensity at 30–35 °C for 30 mins. Then add AlCl₃ (1 equiv.) at 0 ^oC for 1 h. ^bisolated yield. ^cfor gram scale reaction. ^daliphatic alkene used (4 equiv.) for all cases, reaction time 1 h. ^eacrylate used (1 equiv.) for all cases, reaction time 1 h.

We further surveyed the applicability of this catalyst-controlled chemodivergent reaction by assessing the substrate scope of 2-thiopyrrolidinone synthesis using conditions B (Fig. 4)^45,46. For aromatic alkenes, all the substrates reacted well, leading to the corresponding thiopyrrolidinone scaffolds (4a-r) in good to excellent yields (51%–79%), barring 4m,n providing modest yields of 36% and 32% respectively possibly due to the deterioration of the alkene under the reaction conditions as mentioned above. Expectedly, unactivated aliphatic alkenes (1-dodecene, methylenecyclohexane) and acrylates proved to be unsuccessful, freezing at the carbothiocyanation stage (i.e., 1^st cycle of ATRA only operative) due to the difficulty in photoredox-isomerization in 2^nd cycle through oxidative radical-polar-crossover (vide infra)⁴⁴. Guided by the above failure, we postulated that electron-rich aliphatic alkenes might facilitate the above isomerization through SET-oxidation, to provide the expected N-heterocycles. This was indeed the case with a range of electron-rich alkenes, such as vinyl ethers (4s-u) and vinyl amine (4v), leading to the desired thiopyrrolidone products in good yields (64%–78%). In the case of 4s, a considerable amount of aldehyde 4s’ is formed possibly due to the generation of α-ethoxy carbocation during the oxidative polar crossover step⁴⁷. Ene-carbamate (4w), and enamides (4x-z) were also suitable in the current method, although the yields were moderate (40%–63%) due to polymerization of such alkenes under the reaction conditions. Variations of malonate coupling partners were readily implementable to their desired products (4aa–ad) in good yields (56%–73%). Moreover, the expedient transformation of the intricate olefins derived from estrone (4ae), vitamin E (4af), L-valine (4ag), and D-glucose (4ah) further showcased both the mild nature of this protocol along with the unique ‘on-demand’ selective diversification of medicinally interesting molecules. The formation of N-heterocycles was unambiguously proved by X-ray crystallography of 4k and 4z.

Fig. 4. Scope of the photochemical synthesis of 2-thiopyrrolidinone.

^aReaction conditions B: 1 (0.4 mmol), 2 (0.2 mmol), PTH4 (5 mol%), and degassed toluene (2 mL) under argon with irradiation of LEDs (λ_max = 390 nm) with 100% intensity at 30–35 °C for 24 h. ^bisolated yield. ^cfor gram scale reaction.

This chemoselective catalytic manifold involving TCMs was also extended to a range of other C-centered radicals (Fig. 5). A variety of thiocyanomethyl reagents (2f-j) substituted with carbonyl (-keto, -ester) or non-carbonyl (-nitrile, -benzyl) functionalities were effectively engaged with styrene as a bifunctional reagent, providing versatile carbo-thiocyanates (5a-e) and carbo-isothiocyanates (6a-e) with high synthetic value^3,4,6. Further investigations on tertiary alkyl thiocyanate reagents allowed the incorporation of quaternary center γ- to thiocyanate (5f-h) and isothiocyanate products (6f-h) in synthetically useful yields (49%–84%). Moreover, Fluorinated alkyl-thiocyanate 2n also successfully led to chemo-divergent products 5i and 6i with good yields (69%–78%). A few thiocyanates such as 2-thiocyanatopropane and (thiocyanatomethylene)dibenzene (See incompatible substrates list, Supplementary Table 6), however, could not be harnessed successfully into this manifold. While the former is an aliphatic thiocyanate having a reduction potential which is well beyond the scope of the catalysts in discussion, the later is easily oxidized to a highly stabilized benzhydryl carbocation and rather undergoes self-isomerization to isothiocyanate.

Fig. 5. Scope of the bifunctional chalcogenonitriles in photochemical chemodivergent reaction.

^aReaction conditions A: 1 (0.4 mmol), 2 (0.2 mmol), PTH2 (5 mol%), and degassed toluene (2 mL) under argon with irradiation of LEDs (λ_max = 390 nm) with 100% intensity at 30–35 °C for 30 mins. ^bReaction conditions B: 1 (0.4 mmol), 2 (0.2 mmol), PTH4 (5 mol%), and degassed toluene (2 mL) under argon with irradiation of LEDs (λ_max = 390 nm) with 100% intensity at 30–35 °C for 24 h. ^cisolated yield. ^dfac-Ir(ppy)₃ (1 mol%) as catalyst. ^eafter completion of photoredox reaction, AlCl₃ (1 equiv.) was added at 0 ^oC for 1 h.

We were also curious to explore other heavier chalcogenonitriles, like hitherto unexplored ambident reactivity of selenonitrile which could possibly be tamed in a similar way by our strategy. We were delighted to observe the selenocyanatomalonate 2o undergo chemoselective photoredox annulation with various vinyl arenes leading to Se- (7a-e) and N- (8a-e) heterocycles with moderate to good yields (44%–72%) under conditions A and B, respectively (Fig. 5). Selenocyanatomalonate 2p containing the bulky isopropyl ester groups was also used successfully as a bifunctional reagent in this process. Moreover, the photocatalytic switch was also effective for phenacyl selenocyanate 2q to access γ-keto–selenocyanate (7g) and –isoselenocyanate (8g), classes of useful building blocks for the assembly of selenium-containing heterocycles and peptidomimetics^48,49. Overall, these studies certified the generality and broad applicability of our current strategy in controlling ambident reactivity of chalcogenonitriles.

The synthesized N-heterocyclic scaffolds (both thiopyrrolidones 4 and selenopyrrolidones 8) are normally stable at room temperature. However, their substituted hydrothiophene 3 and hydroselenophene 7 counterparts are not bench stable and require cooler storage temperatures.

The efficiency of this photo-annulation reaction could be further highlighted by its scalability, ease, and simplicity. The reaction between diethyl thiocyanomalonate 2a and 4-methylstyrene 1a was accomplished on a 6 mmol scale affording 3a (1.63 g, 81%, Fig. 3) and 4a (1.45 g, 72%, Fig. 4) under conditions A and B respectively, demonstrating the robustness of the process. Interestingly, the reaction of 4-methylstyrene 1a with diethyl bromomalonate and ammonium thiocyanate under slightly modified conditions also delivered the desired chemo-divergent products 3a (63%) and 4a (57%), albeit less yield compared to the corresponding bifunctional thiocyanate 2a. This modular three-component recipe further enriches this developed method to access hydrothiophenes and pyrrolidine heterocycles from ready-stock materials by just regulating the photocatalyst (Fig. 6A). To demonstrate the utility of this method, we manipulated the 2-imino-hydrothiophene and 2-thiopyrrolidone products (Fig. 6B). For example, acid hydrolysis of 3b resulted in thiolactone 9, a useful monomeric skeleton for polymerization⁵⁰. Treatment of geminal diester 3b with catalytic Lewis acid leads to 2-amino-4,5-dihydrothiophenes 10 through a decarboalkoxylation-isomerization sequence. Notably, highly intricate spiro thia-oxazete⁵¹ 11 and bridged this bicycle 12 were synthesized via inter- and intra-molecular nucleophilic addition-annulation at the imine center of 3 v and 3aa, separately. Oxidation of 2-thiopyrrolidone 4b to the corresponding pyrrolidone 13 went cleanly with mCPBA. In addition, the alkylation of 4b with dimethyl sulfate gave way to thioimidate 14 in good yield. The synthetic utility of free thiocyanates and isothiocyanates was also investigated. Nucleophilic substitution of diphenylphosphine oxide on thiocyanate compound 5f produced phosphonothioate 15, which is an essential component of therapeutic oligonucleotides⁵². The thiocyanate group of 5f was also derivatized to trifluoromethyl thioethers 16 using TMSCF₃, which has been known to show high lipophilicity⁵³. On the other hand, treatment of 6f with benzamidine hydrochloride afforded 1,2,4-thiadiazoles 17 that can act as dual 5-lipoxygenase and cyclooxygenase inhibitors⁵⁴. Finally, thiourea derivative 18 was prepared from the reaction of 6 f with morpholine which has a wide range of pharmacological effects⁵⁵.

Fig. 6. Synthetic transformations.

A Three-component reaction. B Postsynthetic modifications.

Mechanistic studies

To gain mechanistic insights into the plausible mechanism, a series of control experiments were conducted (Fig. 7). Firstly, the radical nature of the reaction was indicated by inhibition of the reaction in the presence of TEMPO, with concomitant detection of TEMPO adducts 19, 20, and 21 (Fig. 7A). The same reactive intermediates (19’ - 21’) were also traced by reaction with 2-phenylimidazo[1,2-a]pyridine⁵⁶. When α-cyclopropyl-4-chlorostyrene 22 was reacted with 2a under standard conditions, ring-opened product 23 was isolated (Fig. 7B). Moreover, exposing diethyl 2,2-diallylmalonate 24 with 2k to carbo-thiocyanation conditions resulted in the 5-exo-trig cyclized product 25. Both these radical probe experiments confirm the initial formation of malonyl radicals in this process. To understand the source of the N-H proton in 4a, deuterated-TCM 2a-d (73% D) was synthesized and reacted with 1a under standard conditions with PTH4. Product 4a-d was detected in crude ¹H NMR with 58% deuterium incorporation, indicating the methine hydrogen of 2a is mainly supplying the N-H proton of product 4a (Fig. 7C). The necessity of continuous photo-irradiation for the reaction was confirmed by a light-on/off experiment (Fig. 7D). UV/Vis measurements of individual reaction component and their combination do not hint towards the formation of an electron donor-acceptor complex between 1a and 2k (spectra I, Fig. 7E)⁵⁷. In addition to the control results presented in Table 1 (entries 2, 14–19), the ineffective overlay between absorption spectra of starting thiocyanates (2k and 5f) with emission spectra of LEDs (λ_max = 390 nm) used, eliminates the possibility of C-S bond cleavage by direct light excitation of 2k or 5f (spectra II, Fig. 7E)^58,59. In isothiocyanation event, control studies indicated that isomerization of 5f to 6f only took place in the presence of both catalyst and light (Fig. 7F). Moreover, exposure of benzyl thiocyanate 5f (the first cycle product) to various nucleophiles (ethanol, 1,2,4-trimethoxybenzene) under conditions B afforded nucleophile-trapped 26a,b along with the isomerized isothiocyanation product 6f (Fig. 7G). This result indicated the intermediacy of a carbocation 5f’, generated by an oxidative radical-polar crossover during the isomerization process with catalyst PTH4 in the reaction. The non-isomerization of 1-Dodecene derived alkylthiocyanates 27 (Fig. 7G) is tentatively attributed to their unfavorable reduction potential (E^red = −2.39 V vs. SCE, see Supplementary Fig. 15) compared to PTH4 (E_1/2^* = − 2.16 V vs. SCE), which cannot initiate the second photoredox cycle. To understand the trajectory of the cascade catalysis with PTH4, the reaction of 1a with 2k was monitored by ¹H NMR spectroscopy (Fig. 7H). In the first 20 minutes, complete conversion of the starting thiocyanomalonate 2k took place, majorly to benzylthiocyanate 5f (70%), along with a minor amount of isothiocyanate 6f (22%) (Graph on first 1 h reaction, Fig. 7H). Subsequently, 5f gradually converted to 6f in 24 h time, suggesting that isomerization involving cascade cycle is the rate-limiting step (Graph on 24 h reaction, Fig. 7H). Stern-Volmer fluorescence quenching studies also indicated that both 2k and 5f are effective quenchers for excited state photocatalyst PTH4 (Fig. 7I), while 2k is the sole quencher for catalyst PTH2 (Supplementary Fig. 26). To substantiate the potential mode of activation of the photocatalyst (via SET or EnT), the possibilities of both the modes were entertained by attempting each of the steps of our cascade reaction (thicyanation and isomerization) individually with a series of established energy transfer catalysts along with our initial SET capable catalysts (Fig. 7J). Computation of the triplet energies of the starting thiocyanates 2k (44.9 kcal/mol) and 5f (43.6 kcal/mol) revealed that all the photocatalysts listed in Fig. 7J (all having E_T > 45 kcal/mol) rationally should provide both the desired products 5f and 6f (albeit to different extents), if an EnT pathway was at play at all. However as observed, the only productive catalysts were those whose reduction potentials exceeded that of the starting thiocyanates 2k (E^red = −1.36 V vs SCE, entries 2, 5–8) and 5f (E^red = −1.83 V vs SCE, entries 5 and 8, see Supplementary Figs. 12 and 14 for electrochemical measurements), irrespective of their triplet energy. This bolstered our initial hypothesis of SET processes being responsible while eliminating any possibility of energy transfer.

Fig. 7. Mechanistic studies.

A Radical trapping. B Radical probe. C Deuterium labeling experiment. D Light ON/OFF experiment. E UV/Vis absorption. F Control experiments on isomerization. G Isomerization intermediate evaluation. H Monitoring of cascade reaction (between 1a and 2k) by ¹H NMR. I Stern-Volmer quenching with PTH4. J Investigation into catalyst activation modes.

From the detailed mechanistic studies and related literature reports^60–62, we proposed a mechanism to execute the formation of divergent products in Fig. 8A. Initially, a single-electron transfer from photoexcited catalyst PC* to thiocyanomalonate 2 creates thiocyanate anion and malonyl radical A which undergoes addition to the alkene 1 generating radical B. This proceeds to form 3’ via two possible pathways namely radical-radical cross-coupling (combination of B with thiocyanate radical, generated by oxidation of thiocyanate anion through SET oxidation during catalytic turnover) and radical chain transfer (kinetically feasible thiocyanate group transfer from thiocyanomalonate 2). Exploration of these pathways by DFT calculations (Fig. 8B) revealed that the addition of malonyl radical A either onto 4-methylstyrene (aromatic conjugated olefin) or 1-dodecene (aliphatic variant) 1 (via TS1) proceeds irreversibly with a low energy barrier (8.8 and 12.3 kcal/mol, respectively) to deliver the intermediate B, downhill in energy by 16–21 kcal/mol. For the subsequent conversion of B to 3’ by selective radical-radical cross-coupling between B and SCN radical, though kinetically feasible based on the Ingold−Fischer “persistent radical effect”⁶³, would heavily compete with a significant number of highly probable unproductive pathways^64,65 thus diminishing its synthetic viability. A more plausible route to explain the high efficiency and selectivity observed would be envisioning intermediate B as a radical chain mediator, engaging with a second molecule of thiocyanatomalonate 2a via TS2 (with an energy barrier of 21.4 kcal/mol (R = p-Tol) and 15.7 kcal/mol (R = n-decyl). This leads to the metastable compound 3’ while the resulting second malonyl radical is easily sequestered by the excess olefin in the reaction. The metastable 3’ can subsequently undergo Lewis acid assisted cyclization to generate the S-heterocycles 3. For styrene-derived alkenes, the possibility of an additional radical-polar crossover mechanism via a stable benzylic carbocation was also explored (see, additional reaction mechanism in SI). However, the high quantum yield (φ) of the reaction for both aromatic (36.9, R = p-Tol) as well as aliphatic (25.2, R = n-decyl) alkenes validates radical chain pathway being the major contributor⁶⁶.

Fig. 8. Proposed mechanism supported by computational studies.

A Proposed mechanism. B Computational study. Free energy profile calculated at the [SMD(toluene) B3LYP-D3/def2-TZVPP] level of theory. Bond lengths are reported in Ångstroms.

When the reductively more potent PTH4 is used, the previously formed thiocyanate 3’ undergoes a second SET event which splits it into radical B and thiocyanate anion. Hereon, if radical B is susceptible to being oxidized to a stable carbocation (as in case of styrenes, vinyl ethers, vinyl amines, etc.) by an oxidative polar-crossover event, the generated carbocation can be trapped by isothiocyanate anion to form the thermodynamically stable isothiocyanate 4’, which subsequently cyclizes to afford the 2-thiopyrrolidone products 4. DFT studies for the 2^nd cycle suggest that the generation of thiocyanate anion and alkyl radical B from the photocatalytic reductive cleavage of alkyl thiocyanate 3’ takes place by a process that is exergonic by up to ~ 1.7 kcal/mol (R = p-Tol). This is followed by the endergonic polar crossover of radical B to carbocation C (by ~ 2.6 kcal/mol) which combines with isothiocyanate anion and undergoes spontaneous cyclization to N-Heterocycles 4 via intermediate 4’. The minute amounts of 3’ produced by the competitive action of thiocyanate anion are also converted to 4 as it further takes part in the isomerization step.

In conclusion, we have successfully developed a photochemical chemodivergent route to incorporate thiocyanate and isothiocyanate functionalities onto olefins by controlling the ambident-SCN through catalyst modulation. The successful introduction of thio- and seleno-cyanomalonates as a bifunctional group transfer reagent in the photoredox process has enabled the redox-, and step-economic synthesis of 5-membered S-, Se- and N-heterocycles from feedstock alkenes through a cascade process. Due to the fundamental reactivity issues that are being solved here, this study on photocatalyst regulated tuning of ambident chalcogenonitriles will stimulate its wider application in radical chemistry research and as a synthetic tool in general. In addition, evidence from detailed control experiments along with density functional theory (DFT) calculations also provides a solid mechanistic backbone to the developed photo cascade strategy.

Methods

General procedure for photocatalytic 2-Imino-tetrahydrothiophenes synthesis

An oven-dried culture tube equipped with a magnetic stir bar was charged with PTH2 (3 mg, 0.01 mmol, 5 mol%), thiocyanatomalonate 2 (0.2 mmol), and dry toluene (2 mL). The tube was sealed with a Teflon screw cap before olefin 1 (0.4 mmol aromatic olefin/ 0.8 mmol aliphatic olefin/ 0.2 mmol acrylate) was added to it. Then, the reaction mixture was degassed by Freeze-Pump-Thaw cycles with argon and irradiated at 30–35 °C with 390 nm LEDs (100% intensity) at a distance of ~ 5 cm for 30 min (1 h for aliphatic olefin and acrylate). A high-speed fan was used to maintain the temperature. After the completion of the ATRA reaction (confirmed by TLC), Aluminum chloride (27 mg, 0.2 mmol) was added to the ice-cold reaction mixture. After 1 h, 2 mL of ethyl acetate was added and quenched with saturated ammonium chloride solution (2 mL). The crude reaction mixture was extracted with ethyl acetate (2 × 2 mL), washed with brine (3 mL), and dried over anhydrous Na₂SO₄. The organic portion was concentrated, and the residue was purified by silica gel column chromatography using EtOAc/petroleum ether as eluent to afford the corresponding 2-imino-tetrahydrothiophenes product 3.

General procedure for photocatalytic thiopyrrolidinones synthesis

An oven-dried culture tube equipped with a magnetic stir bar was charged with PTH4 (3.7 mg, 0.01 mmol, 5 mol%), thiocyanatomalonate 2 (0.2 mmol), and dry toluene (2 mL). The tube was sealed with a Teflon screw cap before olefin 1 (0.4 mmol) was added to it. Then, the reaction mixture was degassed by Freeze-Pump-Thaw cycles with argon and irradiated at 30–35 °C with 390 nm LEDs (100% intensity) at a distance of approximately 5 cm for 24 h. A high-speed fan was used to maintain the temperature. After the completion of the reaction (confirmed by TLC), reaction crude was concentrated and purified by silica gel column chromatography using EtOAc/petroleum ether as eluent to afford the corresponding thiopyrrolidinones product 4.

Author contributions

S. M. conceived the concept, supervised the project, and drafted the manuscript. I. U. H. performed most of the experiments and mechanistic study with the help of A.S. S.P., and S. R. C. took part in the preparation of some of the starting materials. R.L. performed the computational studies. All authors discussed the results and contributed to the final paper.

Peer review

Peer review information

Nature Communications thanks Jeh-Jeng Wang, and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The data supporting the findings of this study are available within the paper and its Supplementary Information. The X-ray crystallographic coordinates for the structures reported (4k and 4z) have been deposited at the Cambridge Crystallographic Data Center (CCDC), under deposition numbers CCDC 2164590 and 2182605, respectively. The data can be obtained free of charge from the Cambridge Crystallographic Data Center via www.ccdc.cam.ac.uk. Cartesian coordinates of computationally optimized geometries are available in Supplementary Data 1. Further relevant data are available from the corresponding author upon request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-024-49279-w.