Open Access Article
Xiao-Fan
Cui
ab,
Le-Cheng
Wang
ab and
Xiao-Feng
Wu
*ab
aDalian National Laboratory for Clean Energy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China. E-mail: xwu2020@dicp.ac.cn
bLeibniz-Institut für Katalyse e.V, Rostock, Germany. E-mail: Xiao-Feng.Wu@catalysis.de
First published on 17th November 2025
Organoboron compounds are fundamental building blocks in organic synthesis, and recent advances in boron-migration reactions have attracted significant attention. However, carbonylative transformations featuring a 1,2-boron shift remain largely unexplored, with limited strategies available for the incorporation of exogenous carbonyl groups. We report an unprecedented 1,2-boron migratory carbonylation enabled by visible-light photoredox catalysis, in which carbon monoxide (CO) serves as an abundant and convenient C1 source to trap translocated alkyl radicals generated after 1,2-boron migration. The methodology provides an efficient and streamlined approach to synthesize structurally complex and diversely functionalized β-boryl thioesters under mild conditions. Notably, this reaction enables a functional-group translocation within a single molecule, allowing for the positional exchange between a carbonyl and a boryl group via sequential CO2 extrusion, 1,2-boron migration, and CO insertion.
Migration reactions have long been recognized as a powerful strategy for constructing boron-containing compounds, ever since 1960s.26–34 While extensive efforts have been devoted to 1,2-migrations of tetracoordinated borate species, radical-driven 1,2-boron migrations remain comparatively less explored. The earliest study dates back to 1999, when Batey and Smil disclosed that boron-tethered radical cyclizations could undergo a 1,2-boryl radical shift.35 After nearly two decades of inactivity, Studer and co-workers realized a notable intermolecular radical 1,3-difunctionalization by employing simple allylboron compounds as boryl-migrating substrates.36–38 Independently, Aggarwal and co-workers developed a photocatalytic single-electron oxidation of 1,2-diborates to efficiently generate β-boryl radicals.39–41 Very recently, Song's work advanced radical 1,2-boron migration by designing specialized boronic ester precursors to overcome limitations in β-boron radical generation.42 Subsequently, Li group developed a photocatalysis/NHC co-catalysis strategy to achieve 1,2-boron migratory acylation reaction for the efficient synthesis of β-boryl ketones.43 However, despite boron's versatile reactivity profile, current borylation-mediated carbonylation transformation face inherent regiochemical limitations in incorporating exogenous carbonyl groups (Fig. 1).
![]() | ||
| Fig. 1 a, Traditional 1,2-boron migration via tetracoordinate borons. b, Radical induced 1,2-boron migration. c, This work: metal-free 1,2-boron migrative carbonylation. | ||
Carbon monoxide (CO) serves as an inexpensive and readily available C1 source, making it a cornerstone reagent in classical carbonylation chemistry.44–46 Nevertheless, traditional metal-catalyzed carbonylations frequently rely on elevated temperatures and toxic noble-metal catalysts (Pd, Rh, Ir), which inevitably narrows the substrate scope and raises both capital and environmental costs.47–57 Radical-mediated carbonylation has attracted increasing interest owing to its broad radical-generation strategies and relatively mild reaction conditions. In these processes, a transient carbon-centered radical is intercepted by CO to give an acyl radical, which can subsequently be transformed into esters,58–66 amides67–77 or ketones.78–81 The key advantage is that the radical generation step can be decoupled from the metal-mediated bond-forming events; light, peroxides or SET reagents can be used to trigger the radical chain, after which the acyl radical is quenched by a nucleophile or engaged in cross-coupling.82–86 Inspired by the powerful ability of visible-light photocatalysis to generate radical species amenable to carbonylative transformations, as demonstrated in our previous work on radical carbonylations, we envisioned carbon monoxide (CO) as a key radical trapping agent. This strategy would leverage the generated alkyl radicals not only for migration but crucially for coupling with CO, thereby installing a carbonyl group concurrently with the 1,2-boron shift, unlocking new synthetic pathways for structurally complex boronic esters. In addition, achieving functional-group translocation within a single molecule to enable positional exchange remains highly challenging, as it demands exceptional regioselectivity and chemoselectivity. Such transformations often require precise bond activation without perturbing other functionalities and are frequently complicated by the inherent stability of the original groups and potential side reactions. Therefore, efficient and selective intramolecular functional group translocation remains a formidable synthetic challenge. Moreover, this strategy also enables a functional-group translocation within a single molecule, allowing for the positional exchange between a carbonyl and a boryl group via sequential CO2 extrusion, 1,2-boron migration, and CO insertion.
| Entry | PC (x mol%) | HE (y equiv.) | DIPEA (z equiv.) | Solvent | Yieldb |
|---|---|---|---|---|---|
| a Reaction conditions: 1a (0.1 mmol), 2a (1.5 equiv., 0.15 mmol), photocatalyst (2 mol%, 0.002 mmol), HE (1.5 equiv., 0.3 mmol), DIPEA (2 equiv., 0.2 mmol), solvent (1 mL) and CO (60 bar), 40 W blue LEDs, rt, 18 h. b Isolated yield. c EA (2 ml). d 1a (0.2 mmol), 2a (0.1 mmol). e 1a (0.15 mmol), 2a (0.1 mmol). f 1a (0.1 mmol), 2a (0.1 mmol). HE (diethyl 1,4-dihydro-2,6-dimethyl-3,5-pyridinedicarboxylate), DIPEA (N,N-diisopropylethylamine), EA (ethyl acetate), DCE (1,2-dichloroethane). Eosin Y (CAS: 17372-87-1). Acid red 94 (CAS: 632-69-9). | |||||
| 1 | Ir(ppy)3 (2 mol%) | 1.5 | 2 | DCE | 46% |
| 2 | Ir(ppy)3 (2 mol%) | 1.5 | 2 | EA | 54% |
| 3 | Ir(ppy)3 (2 mol%) | 1.5 | 2 | 1,4-Dioxane | 49% |
| 4 | Eosin Y (2 mol%) | 1.5 | 2 | EA | 57% |
| 5 | Ru(bpy)3Cl2 (2 mol%) | 1.5 | 2 | EA | 54% |
| 6 | Ir(ppy)3 (2 mol%) | 1.5 | 2 | EA | 57% |
| 7 | Acid red 94 (2 mol%) | 1.5 | 2 | EA | 58% |
| 8 | Acid red 94 (3 mol%) | 1.5 | 2 | EA | 59% |
| 9 | Acid red 94 (5 mol%) | 1.5 | 1 | EA | 55% |
| 10 | Acid red 94 (3 mol%) | 1.5 | 1.5 | EA | 57% |
| 11 | Acid red 94 (3 mol%) | 1.5 | 0 | EA | Trace |
| 12 | Acid red 94 (3 mol%) | 2 | 1 | EA | 34% |
| 13 | Acid red 94 (3 mol%) | 1 | 1 | EA | 62% |
| 14c | Acid red 94 (3 mol%) | 1.5 | 1 | EA | 62% |
| 15d | Acid red 94 (3 mol%) | 1.5 | 1 | EA | 69% |
| 16e | Acid red 94 (3 mol%) | 1.5 | 1 | EA | 69% |
| 17f | Acid red 94 (3 mol%) | 1.5 | 1 | EA | 69% |
With the optimized conditions established, we set out to investigate the generality and limitation of the photo-catalyzed 1,2-boron migrative carbonylation accelerated (Table 2). For β-boryl NHPI esters, when acyclic alkyl groups were used to replace the cyclohexyl group in the substrate, the carbonylated products were obtained in good yields with little variation in yields (3b–3f). To understand the impact of the stability of radicals on the reactivity and selectivity, substrates containing methylene tethers with different lengths were exposed to the reaction as well, it was found that the length of the chain does not affect the formation of the desired products (3g–3i). The generality with respect to the thiosulfonate coupling partner proved equally broad. A series of thiosulfonates with electron-donating groups such as –Me and –OMe could generate the desired products (4a–4d) in good yields. Furthermore, different functional groups, such as naphthalene ring, primary alkyl chains and trimethylsilyl group (4e–4i) could be tolerated. Meanwhile, thiosulfonates with trimethylsilyl group could also generate the products 4g in 63% yield. Remarkably, Alkenyl sulfones participated via radical addition yielding vinyl-functionalized ketone 4j in 64% yield, which provides a unique modification pathway for functional molecules containing olefinic fragments and also supports the radical nature of this transformation. The universality of thiosulfonates bearing substituents is further demonstrated in the synergistic reaction with β-boryl NHPI esters. When the thiosulfonate containing different substituents (e.g., electron-withdrawing group and electron-donating group) was combined with the acyclic alkyl-containing β-boronyl NHPI ester, the reaction system still maintained its high efficiency, and structurally diversified bifunctionalized products (5a–5g, 68–82%) were successfully constructed. For example, thiosulfonates containing p-fluorophenyl (5b), o-methoxyphenyl (5d), and tert-butyl (5f) were compatible with boronic acid ester precursors containing primary/secondary carbons at the α-position, demonstrating that the free radical migration and thiosulfonation processes are insensitive to both electronic and site-resistance effects.
| a Reaction conditions: 1a (0.1 mmol), 2a (0.1 mmol), acid red (5 mol%), HE (1.5 equiv., 0.3 mmol), DIPEA (1 equiv., 0.1 mmol), solvent (2 mL) and CO (60 bar), 40 W blue LEDs, rt, 18 h. All yields were isolated yields. |
|---|
|
To gain deeper mechanistic insights into the origin of the observed reactivity and to delineate the elementary steps responsible for product formation, we carried out a comprehensive suite of control experiments that systematically probed each putative mechanism. The complete suppression of the reaction in the presence of radical scavengers such as TEMPO, coupled with the detection of corresponding radical trapping adducts by HRMS, revealed the involvement of radical intermediates. High-resolution mass-spectrometric (HRMS) analysis of the crude reaction mixture revealed a prominent peak at m/z = 422.3436, corresponding to the TEMPO-trapped carbon-centered radical adduct [M + H]+ (calcd for C30H49BO4, 422.3436). Having confirmed radical involvement, we next evaluated the individual contributions of each component to the overall catalytic cycle. In a stepwise fashion, we omitted the photocatalyst, base, HE and visible-light irradiation (λ = 456 nm, 40 W blue LEDs). In every case, the reaction was completely arrested, with no detectable product formation (≤1% by 1H NMR). Notably, extending the reaction time did not significantly increase the reaction yield. The strict requirement for every component (photocatalyst, base, Hantzsch ester, and light) demonstrates that synergistic coordination of multiple catalytic cycles is essential to achieve this transformation.
Based on the control experiments, a possible reaction mechanism is proposed as outlined in Fig. 2. One possible reaction pathway is that the photocatalyst was excited upon irradiation of visible light. Then, single-electron transfer (SET) from visible light-excited photocatalyst PC* to β-boryl NHPI esters 1a generates a radical anion A and PC˙+, the latter is then reduced by Hantzsch ester to PC. Next, radical anion A undergoes a decarboxylation reaction, releasing CO2, phthaloyl anion (Phth−) and generating an alkyl radical B, which transform into a more stable radical C through a facile 1,2-boron migration.42,43 Trapping of this alkyl radical by CO forms an acyl radical D, which subsequently engages in PhS-group transfer with 2a. This critical step delivers the observed product 3a while generating thiyl radical E. With the proton transfer from the Hantzsch ester radical cation, culminating in the formation of pyridine F and p-toluene sulfinic acid.87,88
To further underscore the preparative utility and broad synthetic reach of the newly developed 1,2-boron-migrative acylation protocol, we carried out an extensive set of scale-up experiment and synthetic transformations. First, the reaction was first translated from the standard 0.1 mmol scale to a 1.0 mmol scale under standard conditions. Gratifyingly, the 1,2-boron shift proceeded smoothly, delivering 3a in 60% isolated yield, only a modest erosion compared to the 69% obtained on 0.1 mmol scale. Then, to examine the range of various S-phenyl sources, a range of control experiments were performed. Neither simple thiophenol (PhSH) nor diphenyl disulfide (PhSSPh) furnished detectable product 4b under the standard photoredox conditions (<5% by LC-MS), consistent with their inability to generate the requisite sulfur radical or sulfonyl radical. In contrast, methyl benzenesulfonate (CH3SO2Ph) and sodium benzenesulfonate (PhSO2Na) both participated productively, providing 4b in 46% and 48% isolated yield, respectively. To showcase the versatility of the boryl-carbonyl motif as a synthetic linchpin, we subjected representative adduct 3a to a concise sequence of chemoselective manipulations.
The positional exchange of two functional groups within a single molecule presents significant challenges, primarily due to the need for high regioselectivity and chemoselectivity. Such transformations require the precise cleavage and formation of specific bonds without affecting other functionalities, often necessitating finely tuned reaction conditions and catalysts. Moreover, the inherent stability of the original functional groups and the potential for undesired side reactions further complicate the process. Achieving an efficient and selective intramolecular functional group transposition remains a formidable synthetic challenge in organic chemistry. We successfully achieved the functional group exchange between the carboxyl and boryl groups within a single molecule through a sequence involving NHPI esterification, 1,2-boron migratory carbonylation, and subsequent hydrolysis. As shown in Fig. 3c, LiOH-mediated hydrolysis of compound 3a proceeded rapidly and afforded the corresponding acid 6 in excellent yield.
![]() | ||
| Fig. 3 a, Scale-up experiments. b, Evaluation of different S-phenyl sources. c, Mutual positional exchange of carboxyl and boryl groups. d, Synthetic transformations. | ||
To further illustrate the synthetic value of this methodology, we explored a series of diverse and practical derivatizations of the Bpin moiety. First, the Bpin group was smoothly converted into the more stable trifluoroborate salt 7 in 87% yield (path I), enabling improved stability and handling for subsequent transformations. In addition, treatment of compound 3a with vinylmagnesium bromide and I2 efficiently delivered the vinylated product 8 in 72% yield (path II), thereby expanding the molecular complexity via a facile C–C bond-forming process. Furthermore, the β-boryl ketone intermediate underwent facile oxidation to furnish the corresponding alcohol 9 in 68% yield (path III), showcasing the compatibility of the boryl group with post-functionalization strategies. Collectively, these successful transformations highlight the broad applicability and robustness of our protocol for late-stage diversification of structurally complex molecules, underscoring its potential utility in synthetic and medicinal chemistry.
Supplementary information: general comments, general procedure, analytic data, and NMR spectra. See DOI: https://doi.org/10.1039/d5cy01255d.
| This journal is © The Royal Society of Chemistry 2026 |