Introduction

Organoboron compounds are now considered one of the most essential building blocks for synthesis1,2,3,4. Among these, organic molecules with C(sp2)–B bonds represent a large and important class of substrates that can be transformed into various derivatives through a multitude of straightforward approaches relying on the replacement of the tri-coordinate boron-based moiety5,6. As a general approach toward diverse boron-based functional groups from their simple pre-functionalized precursors, direct borylation of C(sp2)–H bonds utilizing transition metal catalysts7,8,9 or electrophilic aromatic substitution (EAS)10 have been previously studied with emphases on both reactivity and selectivity (Fig. 1a). For transition metal catalysis, various strategies to gain specific regioselectivity (ortho vs. meta vs. para) such as introduction of directing groups11,12,13 or utilization of steric14,15,16,17 or electronic effects18,19,20 resulting from the pre-installed functional groups have been extensively developed. In particular, alkylbenzenes can be borylated under sterically controlled catalytic systems which give meta selectivity to alkyl groups14,15,16,17. However, the steric control can also limit the scope of alkylbenzenes because the catalyst cannot access sterically hindered C–H bonds, notably those ortho to alkyl groups. Consequently, transition-metal catalyzed borylation of sterically hindered arenes such as mesitylene where all aromatic C–H bonds are ortho to alkyls has been limited21,22. An exceptional example was reported by Chatani et al. utilizing an NHC-ligated platinum catalyst, [(ICy)Pt(dvtms)] (ICy = 1,3-dicyclohexylimidazol-2-ylidene, dvtms = divinyltetramethyldisiloxane) where sterically encumbered alkylbenzenes such as mesitylene could be borylated to give corresponding aryl boronic esters21.

Fig. 1: Borylation of aromatic C–H bonds.
figure 1

a Conventional routes to aryl boronic esters by direct arene borylation. b Cage opening of [B10H10]2– to generate a non-classical boron cluster-based electrophile. c Synthetic approach through cage opening of [B10H10]2– followed by the deconstruction of the corresponding 6-Ar-nido-B10H13 species.

Full size image

As an alternative to transition metal-catalyzed chemistry, electrophilic arene borylation using boron-based electrophiles follows the general trend of reactivity and selectivity of EAS mechanism and thus shows selectivity in electron-rich C–H bonds (electronic control), often at the ortho- or para-positions of alkylbenzenes (Fig. 1a)10. In general, classical boron-based electrophiles consist of a single boron moiety connected to auxiliary mono- or polydentate ligands that control the electrophilicity of the boron center. However, the intrinsically weak electrophilicity of the currently known mononuclear boron-based electrophiles significantly confines the scope of their reactivity to highly electron-rich heteroarenes or anilines23,24,25,26,27. Some notable exceptions reported by Ingleson utilizing in situ generated [Y2BL]+ (Y2 = catecholato, Cl2; L = amine) as mononuclear boron-based electrophiles have shown activity toward alkylbenzene substrates28,29,30,31. However, these systems require high reaction temperatures (>70 °C) due to insufficient electrophilicity of the generated borenium intermediates. The high reaction temperature results in poor regioselectivity as well as undesirable alkyl shift or elimination reactions originating from the Wheland (arenium) intermediates32,33,34. Furthermore, these reactions normally necessitate rigorously inert conditions due to the high air and moisture sensitivity of the reagents employed. To overcome these limitations, stronger boron electrophiles that can be generated from bench-stable and simple precursors are highly desired.

In this work, we now report a strategy for C(sp2)–H borylation relying on the use of non-classical boron cluster-based electrophiles which are significantly more reactive than their mononuclear congeners (see above). Hawthorne has previously shown an acid-induced cage-opening reaction of dicesium closo-decaborate (Cs2[B10H10], 1) that forms an intermediate capable of reacting with several arenes and generating 6-Ar-nido-B10H13 as the products, which represents a unique example of a transition-metal-free cluster-based borenium species (Fig. 1b)35,36,37. Importantly, this polyhedral borane 1 can be obtained in one step from thermolysis of sodium borohydride and is air and moisture stable. We hypothesized that an acid-generated non-classical, strongly electrophilic boron cluster intermediate, B10H13+, would be suitable for a broad scope of electrophilic aromatic borylation of various densely substituted arenes. We further hypothesized that the resulting arylated nido-decaborane compounds could be converted into simple Bpin-containing products via a selective cage deconstruction which allows the breakage of B–B bonds while preserving B–C bonds. While large boron clusters such as [B10H10]2– are generally thought to be very chemically inert entities38,39,40,41, we recently discovered that a selective cage deconstruction is feasible in the context of [B6H5R]2– (R = alkyl) species, providing a potential pathway for converting multinuclear boron clusters into mononuclear congeners42. Here, we show that combining these two synthetic steps provides a precious metal- and ligand-free strategy for the preparation of aryl boronic esters from 1 and alkyl- and haloarenes by sequential cage opening of 1 and the deconstruction of the resulting boron cages (Fig. 1c).

Results and discussion

In order to evaluate the feasibility of our proposed strategy, we have prepared 6-Tol-nido-B10H13 (2) (Tol, tolyl) as a model substrate by reacting 1 and 5 equivalents of HOTf in toluene36. The resulting product was purified by silica gel column chromatography and isolated in 89% yield as a white, air-stable solid characterized by 1H, 13C and 11B NMR spectroscopy. To evaluate whether 2 can undergo cage deconstruction, we commenced with an NMR spectroscopic study to examine the potential conversion of 2 into a mono-boron species in the presence of pinacol (Fig. 2). We hypothesized that these conditions could be feasible, considering our previous study which showed that nido-B5R5 (R = alkyl) species could undergo deconstruction forming RBpin products42. First, in a J. Young NMR tube, a CD3CN solution of 2 (p:m = 72:28) was prepared under a static vacuum (Fig. 2b, c, red line). The solution was heated at 65 °C for 24 h, but no conversion of 2 was observed, showing high thermal stability of 2 in acetonitrile (Fig. 2b, c, olive line). Next, 5 equivalents of pinacol were added to the CD3CN solution of 2. When the solution was heated at 65 °C for 24 h, formation of TolBpin (3) and gaseous H2 was observed by 1H NMR spectroscopy along with a decreased concentration of 2 (Fig. 2b, green line). Also, 11B NMR spectrum of this solution revealed the formation of HBpin as the byproduct (Fig. 2c, green line). Further heating of the solution for a total of 72 h achieved a nearly quantitative conversion of 2 and HBpin, successfully yielding 3 as the only observable major organic species by NMR spectroscopy (Fig. 2b, c, purple line). The ratio of para- and meta-isomers of 3 were 73:27 almost identical to those of the starting 2, which indicates that rearrangement of the C(Tol)-B bond does not take place during the reaction. Although the detailed reaction mechanism is currently elusive due to a lack of observable core intermediates other than HBpin, the formation of H2 indicates similarity to the reactions of nido-B10H14 with small nucleophiles, where H2 is formed as the major byproduct and thus implies coordination of pinacol to the boron cage is likely to occur during the reaction43,44,45.

Fig. 2: Monitoring of deconstruction of 2 by NMR spectroscopy.
figure 2

a Formation of 3, H2, HBpin and borates by the deconstruction of 2 in the presence of pinacol in CD3CN. b Stacked 1H NMR spectra of the reaction mixture. c Stacked 11B NMR spectra of the reaction mixture.

Full size image

With the initial observation of the formation of aryl boronic ester in hand, optimization experiments were conducted using in situ generated 6-Ph-nido-B10H13 as a model substrate to generate PhBpin (4) as a target product (Table 1). The model substrate, 6-Ph-nido-B10H13 was generated from 1 and benzene at room temperature for 3 h under aerobic conditions and then used in situ after simple filtration through a silica plug without thorough isolation. For the deconstruction step, a reaction mixture of in situ generated 6-Ph-nido-B10H13 with 10 equivalents of pinacol and 12 equivalents of MgSO4 as drying reagent in MeCN was stirred at 65 °C for 24 h under air conditions. Without any deviations from these conditions, 4 was produced in 24% NMR spectroscopic yield along with 40% of the remaining 6-Ph-nido-B10H13 determined by 1H NMR spectroscopy using mesitylene as an internal standard (entry 1). The addition of 10 mol% p-BQ (1,4-benzoquinone) as an oxidant did not affect the reaction efficiency, unlike the previous hexaborate degradation where superstoichiometric amounts of oxidants were needed (entry 2)42. However, the use of 10 mol% of a stronger oxidant, CAN (cerium(IV) ammonium nitrate), increased the reaction yield to 48% and only 9% of 2 remained after 24 h (entry 3). To complete the screening, increased loading (20 mol%) of CAN was examined, and it produced 4 in 71% NMR spectroscopic yield with the full conversion of 6-Ph-nido-B10H13 (entry 4). When the reaction was performed under an N2 atmosphere, a slightly decreased yield of 4 was obtained, but the difference was not significant (entry 5). Because the reaction can proceed without an oxidant, we postulate that the role of CAN is to facilitate the oxidation of boranes followed by the coordination of pinacol46.

Table 1 Screening experiments
Full size table

With the optimized conditions, various alkyl- and halobenzenes were examined (Fig. 3a). The target compound of the screening experiment, 4, was isolated in 67% yield after purification of the reaction mixture by silica gel flash column chromatography. Mono-substituted benzene compounds gave up to 83% yields of the corresponding aryl boronic esters (3,5–9) with high para selectivity following the electronic control commonly observed in EAS reactions. Sterically bulkier alkyl groups showed higher para selectivity in general. Next, alkylbenzenes that do not have available para C–H bonds were investigated. For 1,2-disubstituted benzene derivatives, substitutions at the 4-positions were obtained like the previously reported electrophilic borylations (10–12)28. However, very unusual selectivity was observed for 1,3-disubstituted benzenes such as m-xylene. Unlike the previous electrophilic borylation examples which showed selectivity for 4-positions (ortho to alkyl groups)28, substitution at 5-positions (meta to alkyl groups) was obtained with very high regioselectivity (13–15). As all the previously reported meta-selective borylation of alkylbenzenes were proposed to follow the steric control7,8,9,14,15,16,17, this unusual meta selectivity also could result from the steric effect caused by the bulkiness of the non-classical boron cluster electrophile. Next, alkylbenzene compounds that have only ortho C–H bonds were examined. Previous electrophilic borylation reactions which were typically conducted at high temperature (>70 °C) usually did not work with these substrates and even a few known examples exclusively showed undesirable side reactions such as alkyl migration and elimination28,29,30,31. However, our reaction conditions showed good reactivity with p-xylene resulting in a 79% yield of 16 without any side reactions. To further investigate the steric effect in this system, 1,4-dialkylbenzenes with two different alkyl groups were tested. Consequently, the less sterically hindered C–H bonds ortho to methyl groups were exclusively borylated instead of the sterically encumbered C–H bonds ortho to isopropyl or tert-butyl groups (17, 18), supporting the high steric effect of this system. Next, tri- or even more alkylated benzene derivatives such as mesitylene, 1,3,5-triethyl-, 1,2,4,5-tetramethyl- and pentamethylbenzene were examined. These substrates also showed good reactivity and the reactions did not give any undesirable side products (20–23), likely due to the low reaction temperature insufficient to promote alkyl shifts. It should be noted that direct C–H borylation of these sterically hindered multi-alkylbenzenes is challenging and hardly be obtained at room temperature by other methods21.

Fig. 3: Substrate scopes.
figure 3

a Substrate scopes of the boron cluster-mediated arene C–H borylation. Longer reaction times in the deconstruction step are shown in parentheses. b Alternative strategy to borylate a highly electrophilic 1,3-dibromobenzene using [B10H10]2– as the boron source.

Full size image

While borenium chemistry is fundamentally incompatible with molecules containing coordinating groups (e.g., N, O) due to a strong thermodynamic propensity for the formation of strong B–N and B–O bonds, we hypothesized that borylation of haloarenes should nevertheless be feasible47. To probe this, several fluoro- and chlorobenzene compounds were investigated (Fig. 3a). Fluoro- and chlorobenzene gave the desired boronic ester products with para selectivity (24, 25), but the overall yields were decreased compared to alkylbenzene substrates due to the deactivating effect of halogens in EAS36. When both fluoro and alkyl groups were present, the strong para-directing effect of fluorine overrode the electronic and steric effects of alkyl groups in general (26, 27, 31). However, in the case where the para-position to fluorine is sterically hindered by two methyl groups, relatively poorer selectivity was obtained due to the steric hindrance by the methyl groups (28). When a C–H bond para to fluorine is not available, selectivity was observed ortho to fluorine (31, 33). In the case where all para and ortho C–H bonds to halogens are not present, borylation meta to halogen C–H bonds was obtained without undesirable side products (34, 35). Chlorobenzenes generally showed lower selectivity and reactivity compared to fluorobenzenes when substrates have the same substitution patterns (30, 32, 35). Overall, these studies suggest that B10H13+ intermediate can be compatible with aromatic compounds containing C–Cl and C–F bonds.

Although pinacol boronic ester (Bpin) is the most popular and desired boronic ester group in organic synthesis, other alcohol-derived boronic esters were also prepared to highlight the versatility of our borylation reaction (Fig. 3a)4,48. Trimethylene glycol, neopentyl glycol, hexylene glycol, 2,4-dimethyl-2,4-pentanediol, and (1S,2S,3R,5S)-(+)−2,3-pinanediol yielded the corresponding phenyl boronic esters under the standard reaction conditions (36–40). This result shows that various types of boronic esters can be prepared in a divergent manner using bare alcohols, which can be advantageous over the current borylation reactions that require multi-step syntheses to prepare pre-functionalized boron reagents such as diboron compounds49.

A gram scale synthesis of 17 was carried out to demonstrate the scalability of this process (Fig. 3a). From 10.0 mmol of 1 in 20 mL p-cymene, 1.27 g of 17 was produced in a 49% isolated yield over two steps. Overall, the borylation reaction could be performed under fully aerobic conditions on a gram scale without the use of pre-functionalized boron reagents or precious metal catalysts along with unique regioselectivity and reactivity.

Despite successful C–H borylation of sterically congested alkyl-, fluoro- and chloroarenes, other types of substrates such as heteroarenes, aryl bromides or CF3-substituted arenes were not suitable for our reaction conditions for two reasons. First, the use of the superacid, HOTf, resulted in rapid salt formation with most heteroarenes instead of inducing the cage opening of 1. Also, a deleterious side reaction, formation of nido-B10H13OTf 36, was dominant when electron-deficient arenes such as aryl bromides and CF3-substituted arenes were used. Therefore, an alternative approach was sought after to supplement these limitations and to enable 1 as a more general boron reagent for arene C–H borylation. As a result, we adopted a strategy of catalyst-free thermal arylation of closo-dodecaborate ([B12H12]2–) where various arenes including 1,3-dibromobenzene were used as substrates to generate polyarylated dodecaborates by direct C–H bond functionalization and authors suggested that these processes occur via EAS-type mechanism50,51. In our attempt to apply this reaction to closo-decaborate functionalization52, heating of a [TBA]2[B10H10] solution (41, TBA = tetrabutylammonium) in 1,3-dibromobenzene (42) at 200 °C for 72 h resulted in full conversion of 41 giving aryl-substituted decaborates, [TBA]2[B10ArnH10-n] (Ar = 1,3-Br2Ph, n = 6–7) (Fig. 3b). The crude reaction mixture was simply washed with hexane to remove excess amount of remaining 42 and further heated in the presence of 2 equivalents of CAN, 10 equivalents of pinacol and 12 equivalents of magnesium sulfate in acetonitrile at 70 °C for 72 h. Consequently, 3,5-Br2PhBpin (43) was obtained as the final product in 41% isolated yield based on the starting 41 with exclusive regioselectivity. This example showcases that a thermal route could potentially circumvent the scope of limitations of the strong acid-mediated electrophilic borylation with 1 and further shows the generality of cage deconstruction of arylated B10-based species. This deconstruction is reminiscent of the similar processes occurring with alkylated B6-based clusters previously reported by our group36.

Next, we conducted a deeper investigation into the borylation of a challenging substrate, 1,3,5-tri-isopropylbenzene (44), where all C–H bonds are sterically hindered by two adjacent, sterically bulky isopropyl groups (Fig. 4). Although C–H borylation ortho to one isopropyl group has been reported21, borylation of a C–H bond between two adjacent isopropyl groups has never been achieved before. When 44 was introduced to the standard reaction conditions, some degree of migration and elimination of an isopropyl group was observed unlike methyl- or ethyl-substituted alkylbenzene compounds, probably due to the relative stability of the resulting secondary carbocation that can form from the Wheland intermediate (Fig. 4a). Nevertheless, the major product was still the desired product, 2,4,6-tri-isopropylphenyl pinacol boronic ester (45) with 90% selectivity determined by 1H NMR spectroscopy while the alkyl migration (46) and elimination (14) products were also observed in 4% and 6% ratio, respectively. When a lower reaction temperature (0 °C) was introduced to get 45 as an exclusive product, a slightly lower yield was obtained (14%), but as expected, 94% selectivity of 45 was achieved. On the other hand, a higher reaction temperature at 60 °C decreased the selectivity of 45 to 76% along with a higher ratio of 14 in 16%. Next, the reaction of HOTf with 44 at room temperature was tested because protonation of 44 by HOTf could promote alkyl shifts to make 1,2,4-tri-isopropylbenzene or 1,3-di-isopropylbenzene33,53, followed by borylation to generate 46 or 14 (Fig. 4b). However, it turned out that HOTf does not produce the alkyl shift products. This result suggests that the alkyl migration and elimination should proceed through the formation of a Wheland intermediate formed by electrophilic addition of the boron cluster electrophile to an ipso position of 44 (Int_B in Fig. 4c). Between two possible pathways depicted in Fig. 4c, path a generating Int_A as a Wheland intermediate is likely preferred.

Fig. 4: Borylation of 44 and DFT computational studies.
figure 4

a Borylation of 44 giving three isomers (45, 46 and 14) as the products. b Reaction of HOTf and 44. c Two pathways in the regioselective step in borylation of 44. d DFT-optimized geometries of B10H13+ and 44 and visualized LUMO of B10H13+ and HOMO of 44. e DFT-calculated reaction profile for the formation of 47 from B10H13+ and 44. All DFT calculations were conducted at the ⍵B97X-D/6-311+G(d,p), CPCM(n-hexane) level of theory.

Full size image

Next, density functional theory (DFT) calculations were conducted to further investigate the selectivity of C–H versus C–C bond activation in borylation of 44. All DFT calculations were carried out at the ⍵B97X-D/6-311+G(d,p), CPCM(n-hexane) level of theory. First, optimized geometries of B10H13+ and 44 and the lowest unoccupied molecular orbital (LUMO) of B10H13+ and highest occupied molecular orbital (HOMO) of 44 were calculated (Fig. 4d). As previously suggested by Hawthorne35,36, the LUMO of B10H13+ showed high surface density located in the B6 position of the cluster, consistent with electrophilic substitution proceeding by the interaction of B6 of B10H13+ with the π-orbital of the aryl ring (HOMO of 44). As discussed above, this reaction forms two possible intermediates, Int_A (B6 ipso to H) and Int_B (B6 ipso to isopropyl). The computed energies of Int_A and Int_B (Fig. 4e) indicate that Int_A is 11.0 kcal/mol more stable than Int_B, consistent with the expectation for stabilization of the intermediate by o and p alkyl groups. The overall reaction from B10H13+ and 44 to the decaborane product, 6-(2,4,6-iPr3Ph)-nido-B10H13 (47) is highly exergonic by 108.0 kcal/mol, consistent with the extraordinarily high reactivity of the non-classical boron cluster B10H13+. Ion pairing and coordination of water do decrease the extent of exergonicity, but they do not affect the calculated regioselectivity (see Supplementary Figs. 149 and 150).

To examine the steric effect between alkyl groups and the boron cluster in 6-Ar-nido-decaborane, 6-(2,4,6-iPr3Ph)-nido-B10H13 (47) was independently synthesized and characterized by NMR spectroscopy and X-ray diffraction (Fig. 5). Addition of 5 equivalents of HOTf to a mixture of 1 and 44 at 0 °C produced the desired decaborane 47 in 16% isolated yield after 3 h of stirring (Fig. 5a). Slow evaporation of a saturated benzene solution of 47 afforded a single crystal suitable for X-ray diffraction (Fig. 5b). Next, the solid-state structures of 47 and 236 obtained by single crystal X-ray diffraction were compared to investigate the effect of two isopropyl groups versus hydrogen atoms ortho to the boron cluster (Fig. 5b, c). A bond distance between the boron cluster and the aryl ring (B6-C1) in 47 is elongated (1.582(2) Å) compared to the corresponding bond distance in 2 (1.555(3) Å). Also, a torsional angle between the B6-B2 plane in the cluster and the aryl plane in 47 (B2-B6-C1-C2 157.3(1)°) is closer to 180° rather than 90°, which means the aryl ring is vertical to the boron cluster, while the corresponding torsional angle in 2 (B2-B6-C1-C7 116.6(2)°) is closer to 90° meaning the phenyl ring is rather planar to the boron cage. The observed significant geometric difference in two decaborane compounds, 44 and 2, can be attributed to the steric effect of the isopropyl groups in 47. To reduce the steric hindrance by the hydrogen atoms of the boron cluster, one isopropyl group is located above the boron cage and another isopropyl group is situated aside the boron cluster with a slight distortion due to the steric hindrance between the two hydrogen atoms (H2 and H13). This observation clearly shows that the steric hindrance should affect the geometry and stability of Wheland intermediates as well as 6-Ar-nido-decaborane compounds, and it supports the existence of the steric effect imposed by the non-classical boron cluster electrophile.

Fig. 5: Comparison of solid-state structures of 47 and 2.
figure 5

a Synthesis of 47. b Solid-state structure of 47. Selected bond distance (Å) and torsional angle (°): B6-C1 1.5818(15); B2-B6-C1-C2 157.27(11). c Solid-state structure of 2 (ref. 36) for comparison. Selected bond distance (Å) and torsional angle (°): B6-C1 1.555(3); B2-B6-C1-C7 116.6(2). The insets are the side views projecting from the boron cages to the aryl rings.

Full size image

Conventionally, many of the polynuclear boron-rich molecules are thought of as exceedingly stable54 and thus little attention has been paid toward their potential selective degradation pathways into mononuclear species besides notable examples in partial carborane deboronations55. This work showcases how a large polynuclear B10-based cluster species can be selectively deconstructed into well-defined mononuclear boron-based small molecule organic compounds. Importantly, this deconstruction chemistry can be done with substituted B10-based clusters containing exopolyhedral B–C bonds, leading to the selective destruction of only cluster-based B–B bonds. The selective deconstruction coupled with unique non-classical electrophilic properties of the parent B10H13+ borenium intermediate can be thus leveraged into a straightforward, open flask two-step sequence for arene borylation. This discovery opens up a route towards introducing functional mononuclear boron-based sites into arene molecules via direct C–H functionalization and further bridges the chemistry of boron clusters and single-site boron-containing small molecules.

Methods

Caution: (1) triflic acid is a superacid and thus is very reactive, so special caution is required when it is added to any potentially reactive or flammable chemicals, for example, nitroarenes and so on. (2) The parent unfunctionalized decaborane, nido-B10H14, is known to be neurotoxic.

Preparation of Cs2[closo-B10H10] (1) from [HNEt3]2[B10H10]36

In a typical experiment, a 20 mL scintillation vial was charged with a magnetic stir bar, 3.22 g (10 mmol) of [HNEt3]2[B10H10] and 5 mL distilled H2O under N2 or aerobic conditions. 3.36 g (20 mmol) of CsOH•H2O was dissolved in a minimum amount of distilled H2O, and the aqueous CsOH solution was slowly added to the solution of [HNEt3]2[B10H10] with vigorous stirring. After 24 h, precipitates from the aqueous solution were filtered using a glass frit, and the collected solid on the glass frit was washed with 2 × 10 mL of EtOH and 3 × 10 mL of Et2O then dried under vacuum to give 3.2 g (8.3 mmol, 83%) of a white solid identified as 1.

Preparation of 6-Tol-nido-B10H13 (2) from 136

In a typical experiment, a 20 mL scintillation vial was charged with a magnetic stir bar, 770 mg (2.0 mmol) of 1 and 5 mL toluene under aerobic conditions. Then, 0.90 mL of HOTf (10.0 mmol, 5.0 equiv) was added to the reaction mixture and the mixture was stirred for 3 h at 23 °C. After 3 h, the volatiles were removed under vacuum and the resulting residue was purified by silica gel flash column chromatography using hexane as an eluent. Fractions were collected and the volatiles were removed under vacuum to give 378 mg (1.78 mmol, 89%) of a white solid identified as 2.

General procedure for borylation of alkyl- and halobenzene

In a typical experiment, a 4-mL dram vial was charged with a magnetic stir bar, 154 mg of 1 (0.40 mmol, 1.0 equiv), and 0.5 mL liquid substrates or 5.0 equiv of solid substrates (2.0 mmol) in 0.5 mL hexanes under aerobic conditions. Then, 180 µL of HOTf (2.0 mmol, 5.0 equiv) was added to the reaction mixture and the mixture was stirred for 3 h at 23 °C. The reaction mixture was filtered through a silica plug prepared using a Pasteur pipette packed with ~2 cm length of silica gel. After the filtration, the filter cake was washed with 10 mL hexanes, and the filtrate was collected. Volatiles of the filtrate were removed under vacuum, and the resulting crude mixture was transferred to a 25 mL Schlenk flask charged with a magnetic stir bar, 44 mg of CAN (20 mol%, 0.04 mmol), 4.0 mmol of diol, 580 mg of magnesium sulfate (4.8 mmol, 12.0 equiv) and 10 mL MeCN. The reaction mixture was stirred at 65 °C for 24 h unless otherwise noted. After 24 h, the reaction mixture was filtered through a pad of Celite, and the filtrate was dried under a vacuum to remove volatiles. The crude mixture was purified by silica gel flash column chromatography using hexanes/EtOAc as eluents. Finally, the collected eluents were concentrated under a vacuum to give the desired aryl boronic ester as the product, which was analyzed by NMR spectroscopy.