Excitation of acyclic trisubstituted alkenes via triplet energy transfer for diastereo- and enantioselective [2 + 2] photocycloaddition Authors: Jinguo Long, Jiayun Wen, Zhenzhen Mo, Xianhui Chen, Ting Wang, Xianjie Fang Published: 2026 ABSTRACT AbstractPhotochemistry offers a powerful platform for enabling unique bond cleavage and formation. However, achieving precise stereochemical control while suppressing persistent racemic background reactions remains a fundamental challenge in this field. Herein, we report a highly enantio- and diastereoselective [2 + 2] photocycloaddition of 2-cyanoacrylamides—acyclic trisubstituted alkenes—using a dual catalytic system. Despite non-selective triplet energy transfer to both catalyst-bound and unbound substrates, high stereoselectivity is achieved through a dual Curtin–Hammett kinetic control strategy. This method features broad substrate scope, accommodating diverse arenes and alkenes, and enables the construction of α-chiral quaternary nitriles as well as contiguous quaternary stereocenters. Notably, stereoconvergent transformations are observed from mixtures of varied E/Z isomers and rotamers, and dearomative cycloadditions proceed efficiently across a range of aromatic systems. Mechanistic studies support a pathway involving fast ring-closure kinetics of matched diradical intermediates, with the cyano group playing a critical role.Subject terms: Synthetic chemistry methodology, Photocatalysis Introduction The persistent competition from racemic background reactions poses a formidable challenge in asymmetric photocatalysis, as it impedes precise stereoselective control1–3. Since Bach’s pioneering work4, a key strategy for high stereoselectivity in asymmetric [2 + 2] photocycloadditions centered on direct control of substrate excitation4–47. This is typically achieved through wavelength-selective excitation that distinguishes between catalyst-bound substrates (“on-catalyst”) and unbound substrates (“off-catalyst”), or through precise modulation of triplet-state energy (Fig. 1a, left)48. However, when the excitation profiles of unbound substrates overlap with those of the chiral complexes, non-selective sensitization of both species can occur, leading to loss of stereocontrol48,49. Additionally, the substrate scope in asymmetric [2 + 2] photocycloadditions remains largely confined to aromatic rings, cyclic alkenes, and a limited set of acyclic 1,2-disubstituted alkenes3. Excitation of acyclic trisubstituted alkenes, by contrast, has been scarcely explored23,26. The highly enantioselective transformation of acyclic trisubstituted alkenes bearing a substituent at the α-position adjacent to the coordinating group remains unreported. Working hypothesis. a Enantioselective [2 + 2] cycloaddition. b Enantioselective [2 + 2] cycloaddition under Curtin-Hammett kinetic control. c Selected compounds with a nitrile containing an α-quaternary center. d This work: Enantioselective [2 + 2] cycloadditions of 2-cyanoacrylamides via Lewis acid catalysis. Notably, successful examples of asymmetric [2 + 2] photocycloadditions under non-selective sensitization conditions remain scarce (Fig. 1a, right)48–51. Although the existing breakthroughs have revealed promising new strategies, considerable challenges persist. For instance, the Miller system circumvents achiral background reactions by leveraging the intrinsic inertness of unbound substrates toward the desired photocycloaddition upon excitation48. In contrast, Zhong’s enzyme compartmentalization strategy physically isolates the chiral process within protein cavities to afford protection, albeit at the expense of requiring stoichiometric quenchers in the bulk phase49. Yoon and Jiang independently achieved high enantioselectivity through a post-sensitization differential reaction kinetics strategy50,51; however, the diastereoselectivity obtained was only moderate. Moreover, this kinetically controlled approach is highly sensitive to external variables, such as substrate feed ratio and structural variations. Consequently, achieving effective suppression of racemic background pathways under uncontrolled substrate excitation, while maintaining stereochemical fidelity, remains a fundamental and long-standing challenge in asymmetric photocatalysis. Herein, we propose a dual Curtin–Hammett kinetic control strategy to suppress background reactions arising from non-selective sensitization and to achieve effective stereocontrol of the photocycloaddition (Fig. 1b). The strategy hinges on the precise modulation of diradical intermediate pathways through differential reaction rates, thereby delivering high stereoselectivity. Diradical intermediates derived from unbound substrates undergo rapid C–C bond rotation to revert to alkenes, effectively suppressing racemic product formation (Fig. 1b, left). Under chiral catalyst induction, stereochemically matched diradical intermediates within the chiral complex undergo rapid cyclization, whereas cyclization of mismatched intermediates—even those generated via rotation—is suppressed, thus securing high stereoselectivity (Fig. 1b, right). Nevertheless, this imposes a stringent demand on the chiral catalyst: under conditions where the background reaction is effectively suppressed, the cycloaddition rate of the chiral complex must not only remain uncompromised but also substantially outpace that of the background process, while simultaneously achieving precise discrimination among different stereoisomers. While nitriles are common pharmacophores in natural products and drugs, their incorporation with quaternary carbon centers can significantly enhance the efficacy of bioactive molecules (Fig. 1c)52–54, however, synthetic approaches to α-chiral quaternary nitriles via asymmetric catalytic [2 + 2] photocycloaddition remain elusive4,55–59. Herein, we report an asymmetric [2 + 2] photocycloaddition of acyclic trisubstituted alkenes—specifically 2-cyanoacrylamides, readily accessible via Knoevenagel condensation—enabled by a dual catalytic system (Fig. 1d). Although both catalyst-bound and free substrates undergo unselective activation through triplet energy transfer, the employment of an appropriate chiral Lewis acid effectively enforces high stereoselectivity in this transformation. Notably, stereoconvergent conversion is achieved from mixtures of substrates. Furthermore, the reaction smoothly enables dearomative photocycloadditions across a diverse range of aromatic systems. This strategy not only provides access to α-chiral quaternary nitriles but also allows for the construction of contiguous quaternary stereocenters (CQS) and a variety of chiral fused polycyclic architectures. Results Optimization of the reaction conditions Our investigation commenced with the choice of 2-cyanoacrylamide E−1a as the model substrate and [Ir(dF(CF₃)ppy)₂(dtbbpy)]PF₆ as the photocatalyst (Fig. 2). Upon irradiation with visible light (455 nm) at room temperature, full conversion of E−1a gave rise to the cycloaddition product rac−2a with 95% yield and 5.7/1 dr in the absence of a chiral catalyst (Fig. 2A). Lowering the temperature to –60 °C under otherwise identical conditions reduced the yield from 95% to 14% while increasing the diastereoselectivity to 8:1 dr. The recovered 1a was found to be a ca. 1:1 mixture of E/Z isomers (Fig. 2A), suggesting that excitation of unbound substrates still occurred at this temperature, albeit with a slower cycloaddition rate. Based on these findings, oxazaborolidine catalysts 3, which are recognized as highly effective and readily available chiral Lewis acids, were introduced at –60 °C to control the stereoselectivity (Fig. 2B). Notably, this class of catalysts had not been previously applied to asymmetric [2 + 2] photocycloadditions via acyclic trisubstituted alkene excitation4–12,18. Gratifyingly, employing catalyst 3a (30 mol%) delivered the desired product 2a in excellent yield, high dr, and 58% ee. Subsequent optimization focused on structural modifications of the catalyst 3a. Introduction of ortho-substituents on the B-aryl group (3b–3h) had a pronounced effect on enantioselectivity. An ortho-methyl group (3e) enhanced the ee to 92%, and replacing the phenyl group of the proline-derived amino alcohol with a naphthyl group (3j) further increased the ee to 94%. Solvent screening did not improve the results (Fig. 2C, left); presumably, acetone coordinates with the catalyst, leading to a sharp drop in reaction efficiency (60% yield, 26% ee). Reducing the amount of photosensitizer (1.8 mol% to 0.4 mol%) can also suppress the background reaction, thereby improving the ee value from 94% to 96%, whereas lowering the 3j loading adversely affected the reaction outcome (Fig. 2C, right). Optimization of the reaction conditions for [2 + 2] cycloaddition. A Model reaction and initial condition optimization. B The effect of Lewis acids. C The effect of reaction parameters (with 3j). a Reactions conducted on a 0.05 mmol scale under a N2 atmosphere. Yields and dr were determined by GC analysis by using tetradecane as an internal standard. The ratio of E/Z was determined by 1H NMR analysis. b [Ir] (1.8 mol%), 3 (30 mol%), CHCl3 (0.6 mL), –60 °C, 455 nm (30 W), 18 h. Determined by HPLC over a chiral stationary phase. c 3j (30 mol%), solvent (0.6 mL). d [Ir] (0.4 mol%), CHCl3 (0.6 mL). Substrate scope Under the optimized reaction conditions, the model substrate reached complete conversion within 4 h, affording the product in 91% yield, 95% ee, and >20/1 dr. Given that certain substrates are deeply colored and may not reach full conversion within a short period, we set the reaction time to 18 h and increased the photosensitizer loading for several reactions during the substrate-scope investigation (Fig. 3). Notably, substrate mixtures containing various isomers—including E/Z isomers and rotamers—all underwent the stereoconvergent transformation smoothly (Fig. 3). Our investigation began with the scope of 2- cyanoacrylamides containing alkene functionality (Fig. 3, top). We first explored substrates with various aryl substituents. The reaction tolerated a broad range of electronic properties, accommodating both electron-donating (e.g., para-OMe, 1b) and electron-withdrawing groups (e.g., para-CF₃, 1c). Halogenated substrates (including Br and I, 1d–1h) were also compatible, and the substituent position had little effect on reaction efficiency. Substrates bearing heteroaryl groups such as furyl (1i), thienyl (1j) and indolyl (1k) reacted smoothly under the standard conditions. When the aryl group was replaced with an alkenyl group (1l), the desired product 2l was obtained in 63% yield and high stereoselectivity (95% ee, dr > 20/1). However, such transformation cannot be achieved when the aryl group is replaced with an alkyl group (1a’), in intermolecular reactions (1t). Notably, N-alkyl substituents on the substrates (1m–1p) exerted only a moderate effect on enantioselectivity, affording 84–92% ee. This effect may arise from the proximity of the coordinating group to the catalyst center. Furthermore, variation of the alkene component—replacing the methyl group with ethyl or phenyl—was also well tolerated. Although increased steric bulk only slightly eroded stereoselectivity (for 2r: 92% ee, 13:1 dr), the isolated yield was moderate (49%), likely owing to poor solubility of the product. Cycloalkene substrates (1s) were likewise compatible, enabling the synthesis of polycyclic products in high yield and selectivity (89% yield, 96% ee, dr > 20/1). The absolute configurations of products 2m and 2r were unambiguously assigned by X-ray crystallography. Scope of alkene-alkene photocycloaddition. 1 (0.1 mmol, 1.0 equiv.), [Ir] (0.4 mol%), 3j (30 mol%), CHCl3 (0.6 mL), –60 °C, 455 nm (30 W), 18 h. Yield of isolated product. Determined by HPLC over a chiral stationary phase. dr was determined by GC analysis or 1H NMR analysis. a 4 h. b [Ir] (1.8 mol%). c 1H NMR yield. d w/o 3j. We next sought to extend this catalytic system to dearomative [2 + 2] photocycloadditions (Fig. 3, bottom). The reaction scope was investigated using 2-cyanoacrylamides paired with a range of arene partners. Notably, although enantioselectivity varied considerably (up to 95% ee), diverse arenes—including furan (4a, 4l), benzofuran (4b, 4c), thiophene (4d–4g), benzothiophene (4h), thieno[3,2-b]thiophene (4i), and phenanthrene (4j)—all underwent the desired dearomative transformation. Except for phenanthrene, for which the dr could not be determined, all products were obtained with high diastereoselectivity. The nature and position of substituents on the arene ring were found to exert a significant influence on the reaction outcome (4d–4g). Among these, a methyl group at the 2-position (4g) proved particularly detrimental, leading to lower yield and enantioselectivity compared with the other analogues. Although the catalyst enables high-yield dearomatization of indoles (5k), the product was obtained as a racemate, possibly owing to competing background reactions and product racemization. Unfortunately, replacement of the aryl group with a naphthalene ring or a terminal alkyne did not lead to the desired [2 + 2] photocycloaddition. In cases where competitive cycloaddition pathways existed (4l, furyl vs. alkenyl), the dearomative pathway was preferentially favored (54% vs. 27% yield). Although steric effects cannot be entirely ruled out, the observed reactivity preference suggests that the dearomative cycloaddition is intrinsically more facile, which renders suppression of the background reaction challenging. This likely accounts for the generally modest erosion in enantioselectivity observed across these substrates. To demonstrate the synthetic utility of this [2 + 2] photocycloaddition, we conducted large-scale reactions and further product derivatizations (Fig. 4). For example, scaling up the model reaction to a 2 mmol scale (Fig. 4a) afforded the target product 2a in quantitative yield (98%) with retained stereoselectivity (96% ee, dr > 20/1). The PMB protecting group on 2p was efficiently cleaved using cerium(IV) ammonium nitrate60. In addition, 2a could undergo amidation to furnish corresponding products 2a-1 and 2a-2 (Fig. 4c, left)61. Furthermore, subjecting 2a to DIBAL-H reduced both the amide and cyano functionalities (Fig. 4c, right), generating the primary amine 2a-3. This amine was then condensed with carboxylic acids to furnish the corresponding products 2a-4 and 2a-562. Scale up reaction and product derivatizations. a Large-scale reaction. b Synthetic application of 2p. c Synthetic applications of 2a. Mechanistic studies To gain deeper insight into the reaction mechanism, a series of control experiments were carried out (Fig. 5). A time-course study using (E)-1a (Fig. 5a) revealed partial isomerization to (Z)-1a, with the mixture ultimately reaching an E/Z ratio of approximately 1:1. Over time, product 2a was formed steadily with high stereoselectivity. In a parallel experiment conducted in the absence of chiral catalyst 3j, olefin isomerization was still observed, but the formation of rac−2a proceeded at a considerably slower rate (Fig. 5b). Taken together with the UV–vis absorption spectra (SI, Fig. S2), these results indicate that although substrate excitation is catalyst-independent, 3j enhances reaction conversion and enforces stereoselectivity through coordination with 1a. This enables rapid, stereoconvergent product formation irrespective of the starting olefin geometry (SI, 7.4). Further experiments revealed that neither individual diastereomers nor rac-2a underwent deracemization under the standard conditions (SI, 7.5), indicating that high stereoselectivity does not originate from stereoisomer interconversion. Upon adding a triplet quencher to the reaction49, the yield of 2a decreased significantly (Fig. 5c). In addition, Stern−Volmer fluorescence quenching experiments indicate that the excited-state quenching of photocatalyst [Ir] is significantly more pronounced in the presence of both 1a and Lewis acid than with 1a alone (Fig. 5d). The reaction requires an external photosensitizer to proceed (Fig. 5e). Furthermore, a correlation was found between the yield and the triplet energy of the photocatalyst (Fig. 5e), but not with the redox potential, indicating that the reaction likely proceeds via triplet energy transfer (EnT) rather than a photoredox process (SI, Fig. S4)63,64. Notably, excellent yield and high enantioselectivity were still attained with the achiral photosensitizer 4CzIPN (Fig. 5e), suggesting that this efficient conversion may not rely on cooperative stereoinduction20. Elevating the reaction temperature to room temperature resulted in 2a with 77% ee (Fig. 5f), demonstrating that the chiral catalyst maintains significant stereoselectivity control despite substantial background reactions. This experimental evidence (Fig. 5f), which includes the detection of the minor chiral diastereomer of 2a (5% yield, 38% ee), unambiguously establishes competitive ring-closure between matched and mismatched diradical species within the chiral catalytic complex, with the mismatched pathway being kinetically disfavored at reduced temperatures. Notably, replacing the cyano group with other substituents (X = H, Me, CO₂Et, CF₃ or Ph) consistently yielded suboptimal results (Fig. 5g). When X = H, Me or Ph, photoexcitation facilitated alkene isomerization without productive cycloaddition. Substrates bearing CO₂Et or CF₃ groups generated cycloaddition products but exhibited diminished diastereoselectivity. The observation that non-cyano derivatives (X = CO₂Et/CF₃) gave progressively lower stereoselectivity, while simpler systems (X = H/Me/Ph) were inert, highlights that the cyano group is essential for enantioselective cycloaddition. Mechanistic studies. a Product distribution. b Alkene E/Z isomerization. c Triplet quenching experiment. d Stern–Volmer quenching experiments. e Photosensitizer-controlled studies. f Effect of temperature. g Substituent effect. To elucidate the origin of the observed diastereoselectivity, density functional theory (DFT) calculations were performed (Fig. 6)65. Upon photoexcitation and intersystem crossing (ISC), E-configured 1E-1a generates a triplet diradical intermediate that bifurcates into two competing pathways: direct intramolecular 5-exo-trig radical cyclization followed by radical–radical C–C bond formation to give one diastereomer of the [2 + 2] product, or conformational rotation to a Z-configured intermediate followed by cyclization, which furnishes the alternative diastereomer. In the absence of a chiral Lewis acid (Fig. 6a), the two pathways exhibit comparable energy barriers, with the direct cyclization transition state (3E-TSa) being only 0.4 kcal/mol lower in free energy than the E → Z conformational rotation (3TSar), indicative of poor diastereoselectivity. Upon introduction of the simplified model catalyst ent-3a, four possible binding modes between the catalyst and substrate 1E-1a were evaluated (Fig. S13). Among them, the B–N–Al and B–O–Al binding modes were found to be relatively stable, exhibiting similarly low Gibbs free energies (ΔG = 3.6 and 3.4 kcal/mol, respectively). Both adopt bidentate chelation motifs, in which the substrate simultaneously coordinates to the boron center and the AlBr₃ moiety of the catalyst through either the cyano group (B–N–Al, 1E-1a_N)66–68 or the carbonyl oxygen atom (B–O–Al, 1E-1a_O)4,11,12. Computational studies. a Competing diastereomeric pathways without the oxazaborolidine catalyst. b Competing diastereomeric pathways in the presence of the oxazaborolidine catalyst via the B-N-Al binding mode. c 3D structures and noncovalent interactions of 3E-TSa_N and 3Z-TSa_N. Frontier molecular orbital analysis reveals a reduced HOMO–LUMO gap for the B–N–Al mode relative to the B–O–Al mode (4.039 vs 4.327 eV, Fig. S14)69. Time-dependent DFT calculations further show a lower S₁ excitation energy and a reduced S₁–T₁ gap (Fig. S15)70, collectively indicating more facile access to the reactive triplet state of the B–N–Al complex (3E-1a_N). Moreover, the entire reaction pathway on the triplet potential energy surface is energetically more favorable for the B–N–Al coordination mode (Figs. 6b and S16). Together, these results establish cyano coordination as the key interaction governing formation of the operative B–N–Al catalyst–substrate complex. This binding mode promotes triplet-state reactivity through electronic modulation and positions the substrate for stereochemical differentiation within the chiral environment. Cyano coordination also enables rapid E/Z interconversion, with diastereoselectivity determined by the competing E−5-exo-trig and Z−5-exo-trig cyclization transition states (3E-TSa_N and 3Z-TSa_N), under Curtin–Hammett control. The computed energy difference between these competing transition states (ΔΔG‡ = 1.7 kcal/mol) predicts a dr of ~18:1 at 298 K, in excellent agreement with experiment. To further rationalize the origin of this selectivity, an independent gradient model based on Hirshfeld partition (IGMH) was performed (Fig. 6c)71–74. 3E-TSa_N displays continuous regions of stabilizing H–π and dispersion interactions, whereas 3Z-TSa_N is dominated by a misaligned aromatic arrangement with weakened π–π interactions and pronounced steric repulsion (dH···H = 2.25 Å). These combined geometric and noncovalent effects are responsible for the observed high diastereoselectivity. Discussion In summary, we have developed a highly enantioselective [2 + 2] photocycloaddition of 2‑cyanoacrylamides using a dual catalytic strategy that effectively overcomes persistent racemic background reactions even under non-selective excitation. By combining chiral Lewis acid control with triplet energy transfer, the system achieves excellent stereoselectivity through a dual Curtin–Hammett kinetic regime, which suppresses racemic pathways while promoting rapid, stereocontrolled cyclization within the catalyst complex. This method enables the synthesis of α-chiral quaternary nitriles and fused polycyclic architectures bearing contiguous stereocenters. The cyano group plays a crucial structural and electronic role in securing high selectivity. The approach exhibits broad substrate scope, stereoconvergent behavior, and good scalability. Methods General procedure for asymmetric [2 + 2] photocycloaddition In a N2-filled glovebox, a fresh solution of 3j (30 mol%) in CHCl3 (0.4 mL) and the solution obtained by rinsing with CHCl3 (0.2 mL) were added to an oven-dried 10 mL sealed pressure tube with 1 or 4 (0.1 mmol, 1.0 equiv.) and [Ir] (0.4 mg, 0.4 mol%). The mixture was stirred for 15 min at –60 °C. The reaction tube was irradiated with blue LEDs (30 W, λmax = 455 nm) and stirred at –60 °C for approximately 18 h. Subsequently, the diastereomeric ratio (dr) was determined by GC or 1H NMR spectroscopy analysis, and the mixture was purified by column chromatography or flash column chromatography on silica gel (eluent: petroleum ether/ethyl acetate) to afford the desired product 2 or 5. Supplementary information Supplementary Information Transparent Peer Review file Source data Source data Acknowledgments We gratefully acknowledge the financial support from the National Natural Science Foundation of China (Nos. 22278265, U22B20137, 22302063), the Natural Science Foundation of Zhejiang Province (LR24B020002), the start-up funding from Hangzhou Normal University (Nos. 4095C50222204165, 4095C50225204020, 4095C50223204074), the Hangzhou Leading Innovation and Entrepreneurship Team Project of Hangzhou Science and Technology Bureau (TD2025004), and the Organized Research Program of Hangzhou Normal University. Author contributions X.F. and J.L. conceived the project and designed the experiments. J.L., J.W., and X.C. performed the experiments and analyzed the experimental data. Z.M. and T.W. performed the theoretical calculations. J.L., T.W., and X.F. co-wrote the paper. Peer review Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available. Funding X.F., J.L., and T.W. disclose support for the research of this work from the National Natural Science Foundation of China (Nos. 22278265, U22B20137, 22302063), the Natural Science Foundation of Zhejiang Province (LR24B020002), the start-up funding from Hangzhou Normal University (Nos. 4095C50222204165, 4095C50225204020, 4095C50223204074), the Hangzhou Leading Innovation and Entrepreneurship Team Project of Hangzhou Science and Technology Bureau (TD2025004), and the Organized Research Program of Hangzhou Normal University. J.W., X.C., and Z.M. declare no relevant funding. Data availability The X-ray crystallographic data reported in this article have been deposited at the Cambridge Crystallographic Data Center (CCDC), under deposition numbers 2494665 (2 m), 2494668 (2r) and 2552781 (diastereomer of 5d). These data can be obtained free of charge from the CCDC via www.ccdc.cam.ac.uk. The authors declare that the data supporting the findings of this study are available within the article, Supplementary Information, and Source Data. Data is also available from the corresponding author upon request. Source data are provided with this paper. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Jinguo Long, Jiayun Wen, Zhenzhen Mo. Contributor Information Jinguo Long, Email: jglong@hznu.edu.cn. Ting Wang, Email: wangting@hznu.edu.cn. Xianjie Fang, Email: fangxj@hznu.edu.cn. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77403-5. References Associated Data Supplementary Materials Supplementary Information Transparent Peer Review file Source data Data Availability Statement The X-ray crystallographic data reported in this article have been deposited at the Cambridge Crystallographic Data Center (CCDC), under deposition numbers 2494665 (2 m), 2494668 (2r) and 2552781 (diastereomer of 5d). These data can be obtained free of charge from the CCDC via www.ccdc.cam.ac.uk. The authors declare that the data supporting the findings of this study are available within the article, Supplementary Information, and Source Data. Data is also available from the corresponding author upon request. Source data are provided with this paper.