Arylboronic esters are important synthetic intermediates, which can be converted to multiple functional groups and further employed to prepare a wide range of fluorinated organic compounds [1-6]. Traditionally, the synthesis of arylboronic esters is based on the reaction of aryl Grignard or lithium reagents with boron electrophiles [7-10]. Direct C–H borylation of arenes over precious metals such as Rh, Re, Ru, and Ir has recently emerged as a powerful tool for the synthesis of arylboronic esters, due to their efficiency and functional group compatibility [2-5, 11-14]. In addition, highly selective transition metal-catalyzed borylation reactions of aryl halides, based on Pd, Fe, Co, Ni, Cu, Zn or even metal-free, have also been developed [15-27]. For example, Marder et al. [27] developed the Zn-catalyzed 1, 2-selective dual C–H/C–X borylation of aryl halides. Recently, they also reported an efficient catalytic procedure for the selective C–F borylation of polyfluoroaromatic compounds using the N-heterocyclic carbine (NHC) Ni(0) complex as a catalyst and bis(pinacolato)diboron (B₂pin₂) as the boron source [23]. Geetharani et al. [20] also achieved borylation of aryl halides including aryl chlorides based on a Co(Ⅱ)-NHC precursor, affording aryl boronates in good yields. Jiao et al. [28] reported the pyridine-catalyzed transition metal-free radical borylation of aryl halides. Photocatalysis has recently become particularly attractive in organic synthesis [29-34]. Significant recent progress has also been made in homogeneous photolysis-induced borylation. For example, Larionov et al. [35-37] showed that aryl halides can be photochemically borylated under UV irradiation. Photoredox catalysts such as eosin Y or fac-[Ir(ppy)₃] (ppy = phenylpyridine) can also achieve the borylation of aryl diazonium salts or aryl halides in the same manner, under visible light [38-40]. Glorius et al. [41] even developed the decarboxylative borylation of aryl N-hydroxyphthalimide esters with visible light in metal-free conditions. Very recently, Guo et al. [42] reported a highly selective and general photocatalytic C–F borylation protocol that employs a rhodium biphenyl complex as triplet sensitizer and the Ni(0)-NHC catalyst for the C–F bond activation and defluoroborylation through a triplet energy transfer process.

Although homogeneous catalysts show high activity and selectivity, heterogeneous systems have also attracted attention due to their easy recycling. Zhu et al. [43] achieved (aryl) C–H activation and borylation using Ir(0) nanoparticles stabilized with the trihexyltetradecylphosphonium methylsulfonate ionic liquid as catalyst, in the presence of tetra-2-pyridinylpyrazine. Wang et al. [44] reported the Fe₂O₃ nanoparticle-catalyzed borylation of arenes with tert-butyl peroxide as oxidant, affording the corresponding aromatic boronic esters in moderate yields and with selectivity controlled by the electronic effects of the substituents. Pucheault et al. [45] found that highly active Pd nanocrystals coated with 1, 10-bis(diphenylphosphino)ferrocene or tricyclohexylphosphane exhibited promising activity in the borylation of aryl bromides, with moderate to high yields at 100 ℃. Liu et al. [46] synthesized a molecular heterogeneous catalyst based on Ir by metalation of one-dimensional organosilica nanotubes containing 2, 2'-bipyridine ligands with an Ir complex; the catalyst showed high activity and durability in the C–H borylation of arenes because of the effective suppression of Ir-bpy complex aggregation and fast transport in the short nanotubes. Lin et al. [47] synthesized a terpyridine-based metal-organic layer (TPY-MOL) and metalated it with CoCl₂ to obtain a Co^Ⅱ·TPY-MOL catalysts for benzylic C–H borylation. They claimed that the formation of the MOL-stabilized Co^Ⅱ-(TPY^··)^2- species endowed the catalyst with unique and enhanced catalytic activities for C–H borylation.

To obtain a satisfactory product yield, most heterogeneous borylation processes are still conducted at elevated temperatures and with ligand additives. Because improving the catalytic efficiency by increasing the temperature is an energy-intensive process, enhancing the activity of a solid catalyst at lower temperatures by sunlight irradiation would represent a significant breakthrough. Moreover, electron-rich active sites in catalysts can accelerate the activation of C–halogen bonds; hence, there is considerable interest in employing light irradiation as an alternative approach to increase the electron density at active sites and promote the reactions involving aryl halides. Recently, Chandrashekar et al. [48] reported the photocatalytic transformation of diazonium salts to arylboronates with high turnover number in the presence of water-soluble 3-mercaptopropionic acid (MPA)-capped, CdSe-based quantum dots. The CdSe dots transfer excited-state electrons to the substrate to form aryl radicals, which react with the borylating agent to form the arylboronate. Our group previously employed a photoresponsive cubic β-SiC semiconductor with a band gap of 2.4 eV to absorb light and generate energetic electrons [49]. The transfer of these electrons to Pd can greatly enhance its activity in the photocatalytic Suzuki-Miyaura coupling [50], Sonogashira [51], Heck [52], and nitroarene hydrogenation [53] reactions. Herein, we report a catalyst consisting of Pd nanoparticles supported on β-SiC nanowires with hierarchical structure (Pd/SiC), designed for the photocatalytic borylation of aryl iodides and aryl bromides with B₂pin₂. The catalyst exhibits excellent activity under visible light irradiation at 30 ℃.

Hierarchical SiC nanowires were prepared by a sol-gel process followed by carbothermal reduction [54, 55]. The Pd/SiC catalyst was prepared by a one-step impregnation-reduction method. Briefly, a 485 mg amount of SiC nanowires was dispersed in 14.1 mL of an aqueous solution of Pd(NO₃)₂ (0.01 M). After 30 min of stirring, 10 mL of an aqueous solution of lysine (0.53 M) was added dropwise to the above mixture. After another 30 min of stirring, 5 mL of an aqueous NaBH₄ solution (0.35 M) was added dropwise, and 5 mL of 0.3 M HCl was then added to the above suspension. Finally, the mixture was left in air for 24 h. The solid product was then separated, washed, and dried to obtain the Pd/SiC catalyst with a metal loading of 3 wt%. According to inductively coupled plasma mass spectrometry (ICP-MS, Perkin-Elmer ELAN 5000) measurements, the actual Pd content was 2.87 wt%. Pd/TiO₂, Pd/Al₂O₃, and Pd/SiO₂ catalysts with the same Pd loading, as well as Pd/SiC catalysts with different Pd loadings, were also prepared by the same method but with different supports or different amounts of Pd(NO₃)₂.

The microstructures of the catalysts were investigated by high-resolution transmission electron microscopy (HRTEM) and scanning transmission electron microcopy (STEM). X-ray photoelectron spectroscopy (XPS) measurements were performed on a Kratos XSAM800 spectrometer, using an Al K_α (hν = 1486.6 eV) X-ray excitation source. The crystalline phases were characterized by X-ray diffraction (XRD, Rigaku D-Max/RB). Diffuse reflectance UV-Vis absorption spectra were measured with Al₂O₃ as the reference, using a UV-3600 spectrophotometer (Shimadzu). Photoluminescence (PL) spectra were recorded using a Hitachi F-4500 fluorescence spectrophotometer. We ensured that the quality and tablet thickness of the different samples used in the UV-Vis absorption and PL measurements were consistent.

Unless specified otherwise, all reactions were conducted in ambient Ar atmosphere at 30 ℃. A mixture of aryl halides (1 mmol), B₂pin₂ (1.2 mmol, 305 mg), base (1 mmol, 98 mg), and Pd/SiC catalyst (40 mg) was suspended in 10 mL of N, N-dimethylformamide (DMF) in an oven-dried 25 mL quartz vial equipped with a magnetic stirring bar. During the reaction, the mixture was stirred at 500 rpm using a magnetic stirrer and then exposed to a xenon lamp with a wavelength range of 375 to 800 nm (the spectral output is shown in Fig. S1). A low-pass optical filter was employed to block wavelengths below 400 nm. The light intensity was maintained at 0.65 W/cm². The effect of the light wavelength on the catalytic performance was investigated using various light-emitting diode (LED) lamps with different wavelengths.

After reaction, the mixture was diluted with dichloromethane (DCM, 10 mL). The organic phase was extracted and filtered through a millipore filter (pore size: 0.22 μm). Then, 0.5 mmol of n-dodecane was added as an internal standard. The product yield was determined by gas chromatography-mass spectrometry (GC-MS, Bruker Scion SQ 456) using n-dodecane as the internal calibration standard. The reported values are the average of two experiments. The yields were calculated based on the amount of aryl halide. The residue was purified by column chromatography on silica gel (200–300 mesh; eluant: hexane/ethyl acetate) to isolate the desired product.

All NMR spectra were recorded at ambient temperature using a Bruker Avance III 400 (¹H, 400 MHz; ¹³C{¹H}, 101 MHz; ¹⁹F, 376 MHz). The reported ¹H NMR chemical shifts are relative to tetramethylsilane (TMS) and referenced via the residual proton resonances of the corresponding deuterated solvent (CDCl₃: 7.26 δ). The reported ¹³C{¹H} NMR spectra are referenced to TMS via the carbon signals of the deuterated solvent (CDCl₃: 77.16 δ), whereas the quoted ¹⁹F NMR chemical shifts are relative to CFCl₃ as external standard. All ¹³C NMR spectra were broad-band ¹H decoupled. High-resolution mass spectrometry (HRMS) measurements were carried out on a Thermo Scientific Exactive Plus instrument equipped with an Orbitrap analyzer. Electrospray ionization (ESI) measurements were conducted using a heated ESI (HESI) source with an auxiliary gas temperature of 50 ℃. Atmospheric pressure chemical ionization (APCI) measurements were conducted using an APCI source with a corona needle; the auxiliary gas temperature was 400 ℃.

Using iodobenzene 1a (Table 1) as the substrate, we screened different bases, solvents, and Pd loading amounts in Pd/SiC to assess the scope and limitations of the present photocatalytic borylation reaction (see Tables S1–S3 in the Supporting Information for details). The borylated product 1b (phenylpinacolborane) was obtained in the highest yield of 95% using potassium acetate (KOAc) as base and 3 wt% Pd/SiC as photocatalyst in DMF. Compared with those of other supported catalysts for heterogeneous borylation of aryl halides, reported in previous studies (Table S4), the photocatalytic performances of Pd/SiC exhibit some improvements both in reaction conditions and catalyst repeatability. The reaction did not occur under light irradiation without any catalyst, and no product was detected using only pure SiC as catalyst. Control experiments showed that the 1b yield decreased to 5% in the dark under identical conditions. Moreover, the 1b yield decreased linearly when the light intensity was decreased while keeping all other experimental conditions unchanged (Fig. 1(A)), indicating that the reaction was driven by light irradiation. The reduced catalytic activity is likely due to the decrease in the number of activated electrons generated at low irradiation intensity. The obtained linear relationship indicates that the process should be first order in photon, suggesting that the reaction is dominated by a single photon absorption event. Fig. 1(B) shows the dependence of the catalytic activity on the irradiation wavelength, determined using procedures similar to those reported in literature [50-53]. The 1b yields in irradiation wavelength ranges of 400–800, 450–800, 500–800, 550–800, and 600–800 nm were 95%, 61%, 39%, 23%, and 10%, respectively. As the yield of 1b without light irradiation was only 5%, the light-induced yields within each wavelength range were about 34% (400–450 nm), 22% (450–500 nm), 16% (500–550 nm), 13% (550–600 nm), and 5% (600–800 nm). When the reaction was conducted under UV irradiation (200–300 nm, 0.3 W/cm²) with a light intensity of 0.3 W/cm², a nearly 100% yield of 1b was achieved. Usually, the higher energy associated with short-wavelength light results in the bound energy of Pd going to higher energy levels, facilitating chemical reactions on the Pd nanoparticles [56]. The above values are consistent with the UV-Vis absorption data of the Pd/SiC catalyst.

We further examined the aryl halide scope in the present photocatalytic borylation under optimized conditions, and the results are summarized in Table 1. Aryl iodides with electron-withdrawing (2a–7a) or electron-donating (7a–11a) substituents showed high to moderate reactivity. Besides the aryl iodides, aryl bromides (12a–18a) were also converted to their corresponding boronate esters over Pd/SiC under visible light irradiation. However, the present catalyst could not convert aryl chlorides to the target products.

Different microscopic and spectroscopic methods were used to characterize the structural features of the Pd/SiC catalyst. Hierarchical SiC nanowires (Fig. S2) were employed to support Pd nanoparticles, because the hierarchical structure promotes an increase in light absorption and specific surface area, as well as a reduction in the recombination rate of photoelectron and holes, thus improving the photocatalytic activity [57, 58]. The transmission electron microscopy (TEM) images of Pd/SiC (Fig. 2(A) and S3) show that the Pd nanoparticles are uniformly dispersed on the support and have an average diameter of 3.7 nm (Fig. S4). Based on the HRTEM image in Fig. 2(B), the interplanar crystal spacing of the Pd nanoparticles is 0.23 nm, corresponding to the Pd (111) crystal faces. In the UV-Vis absorption spectra (Fig. 1(B) and S5(A)), the strong absorption peak at ca. 375 nm originates from SiC, while the absorption above 500 nm mainly derives from the near-infrared absorption of SiC and the reflection from the sample. However, Pd/SiC shows stronger absorption peaks than pure SiC in both the UV and visible ranges. Generally, the binding energy (BE) values of metallic Pd are in the ranges of 334.7–335.5 eV for Pd 3d_5/2 and 340.3–340.8 eV for Pd 3d_3/2. Therefore, the Pd particles in Pd/SiC are metallic (Fig. 2(C)). However, the lower values of the Pd 3d BE (334.7 and 340.1 eV) in the Pd/SiC sample suggest an electron transfer to Pd. The work functions of SiC and Pd are 4.0 and 5.12 eV, respectively; when they are in contact with each other, a built-in potential is formed, which forces the electrons to transfer from SiC to Pd, resulting in electron enrichment of Pd [18].

5, 5-Dimethyl-1-pyrroline N-oxide (DMPO) is an electron-trapping agent that can capture electrons from Pd nanoparticles [59]. When 0.5 mL of DMPO was added in the photocatalytic borylation of iodobenzene, the 1b yield decreased to 6%. This yield is almost the same as that observed in the dark reaction, suggesting that the light-driven reaction is completely quenched. In addition, triethanolamine (TEA) was employed as a scavenger to trap the photogenerated holes on the surface of SiC. No product was detected when 0.8 mL of TEA was added to the reaction system. The above results indicate that the coupling reaction cannot proceed without the reduction of electrons or oxidation of holes.

Thus, a visible light photoinduced catalytic process could contribute to the enhanced borylation of aryl halides over the Pd/SiC catalyst. When Pd nanoparticles are dispersed on the surface of SiC, they can form a Schottky junction. As the work function of SiC (4.0 eV) is lower than that of Pd (5.12 eV), a built-in potential of 1.12 eV can form between them. This potential forces electrons to transfer from SiC to Pd nanoparticles and results in electron-rich Pd nanoparticles. In the room-temperature PL spectra of pure SiC and Pd/SiC under excitation at 325 nm (Fig. S5(B)), the PL peaks of both samples range from 400 to 550 nm, in agreement with the literature. However, the PL intensity of Pd/SiC shows an obvious decrease compared with that of pure SiC, indicating that the recombination of the photogenerated electrons and holes has been effectively suppressed. This further confirms the transfer of electrons from SiC to the Pd particles. The electron transfer results in a positively charged region in SiC and negatively charged Pd nanoparticles. The C–X oxidative addition is often the rate-determining step in the catalytic cross-coupling reactions of aryl halides, while transmetalation is generally a facile process. The highly negatively charged Pd particles facilitate the carbon-halogen bond cleavage, and thus enhance the catalytic activity.

The performances of Pd/TiO₂, Pd/Al₂O₃, and Pd/SiO₂ with 3 wt% catalyst loading in the 1a borylation were also studied. Without irradiation, the 1b yield was 8% for Pd/TiO₂, 3% for Pd/Al₂O₃, and 5% for Pd/SiO₂, while under irradiation the yields increased to 24%, 15%, and 18%, respectively. The slightly increased activity of these catalysts under visible light is attributed to the Pd nanoparticles absorbing light through interband electronic transitions to produce photoexcited electrons, which can promote the breaking of C–I bonds. These results further confirm that the intrinsic catalytic activity of Pd is significantly enhanced by visible light irradiation when SiC is used as the support.

We also determined the thermodynamic parameters for the borylation of iodobenzene and bromobenzene with B₂pin₂ in the temperature range 30–100 ℃, under irradiation and in the dark (Fig. 3). The apparent activation energy for the borylation of iodobenzene was calculated to be 47.3 kJ/mol under irradiation (400–800 nm) and 73.1 kJ/mol in the dark, based on the Arrhenius equation. Using bromobenzene as substrate, the apparent activation energies with and without irradiation were 59.8 and 92.9 kJ/mol, respectively. The large difference in activation energy suggests that the reaction mechanisms with and without light irradiation are very different. The light-activated process is mainly driven by charge carriers. The energetic electrons concentrated in the Pd particles activate the adsorbates to form excited states and move to a different potential energy surface to take part in the reaction. The process without light irradiation is mainly driven by phonon effects. The thermal energy results in the adsorbed reactants to move to the ground-state potential energy surface of the product.

The main advantages of a heterogeneous catalyst lie in its good stability, recyclability, and ease of separation. To investigate the recyclability of the Pd/SiC catalyst, the used catalyst was regenerated and tested for five cycles under the same reaction conditions, and did not show any measurable loss in catalytic activity (Fig. S6(A)), which indicates its excellent stability. The TEM images of the catalyst after five runs show no obvious change in the size and morphology of the Pd nanoparticles (Fig. S6(B)).

In summary, we have shown that Pd nanoparticles supported on hierarchical SiC nanowires can effectively utilize visible light to catalyze the C–X borylation of aryl halides to the corresponding boronate esters under mild conditions. The Mott-Schottky heterojunction between Pd and SiC can continuously transfer photogenerated electrons to the Pd nanoparticles. The highly negatively charged Pd can facilitate the cleavage of C–I or C–Br bonds, which is the rate-determining step in catalytic cross-coupling reactions of aryl halides. This work will inspire further applications of transition metal-decorated, semiconductor-supported metal nanoparticles as photocatalysts for a wide range of organic transformations driven by light.

Acknowledgments

We are grateful to the Alexander von Humboldt Foundation for providing a postdoctoral fellowship to X. N. G.