Polymeric iridium catalysts for C–H borylation of arenes under batch and continuous flow conditions

Abstract

Iridium-catalyzed direct C–H borylation of arenes has been extensively studied as an efficient method for synthesizing arylboronates. However, this transformation predominantly relies on homogeneous iridium catalysts, which are typically air-sensitive and difficult to recover. Here we show that pyridine-containing polymeric iridium catalysts can promote C–H borylation under both batch and continuous flow conditions. A ternary copolymer synthesized from 4-vinylpyridine, 4-tert-butylstyrene, and N-isopropylacrylamide served as an effective support for immobilizing and stabilizing the iridium species. This catalyst enabled batch C–H borylation of 20 arenes with yields up to 87%, as well as continuous flow synthesis of borylated products under ambient air conditions for over 24 h. XAFS analysis revealed partial reduction of the iridium precursor during catalyst preparation and in situ during the borylation reaction. This study provides an instructive and practical approach for designing immobilized iridium catalysts, thereby advancing the diverse applications of C–H borylation reactions.

The iridium-catalyzed direct C–H borylation of arenes has been extensively studied as an efficient method for synthesizing arylboronates, however, homogeneous iridium catalysts are typically air-sensitive and difficult to recover. Here, the authors report a ternary copolymer that serves as an effective support for immobilizing and stabilizing iridium species in pyridine-containing polymer-supported iridium catalysts that enable batch C–H borylation of arenes and continuous flow synthesis of borylated products under air.

Introduction

Arylboronates are important synthetic building blocks for constructing value-added molecules, including functional materials, natural products, and pharmaceuticals^1–7. Compared with traditional methods that react aryl lithium or magnesium reagents with trialkyl borates, direct C–H bond borylation of readily available arenes is considered a more efficient one-step process for the synthesis of arylboronates^8–24. Since the pioneering work by Ishiyama, Miyaura, and Hartwig, direct C–H borylation has attracted significant attention in modern organic synthesis¹⁰. Most of these processes are facilitated by sensitive transition-metal complexes, such as iridium(I) complexes like [IrCl(cod)]₂ (cod = 1,5-cyclooctadiene) or [IrOMe(cod)]₂, which typically initiate the reaction through oxidative activation of boron reagents. However, these reactions require careful handling to maintain the activity of the catalytic species, often necessitating glovebox, Schlenk, or other inert gas protection techniques. Moreover, after the reaction, the iridium complex, as a homogeneous component, needs to be removed from the reaction mixture through additional procedures, and recovering or recycling the iridium is generally difficult. Although transition-metal-free approaches for the C–H borylation of arenes have been reported, they often require difficult-to-handle or corrosive reagents such as boron tribromide (BBr₃)^25,26.

Recently, replacing homogeneous metal catalysts with easily recoverable and reusable heterogeneous catalysts has garnered significant attention as a strategy to address the growing uncertainty metal resource supply^27–29, and accordingly, several research groups have sought to develop recyclable iridium catalysts. For example, Zhu et al. developed an iridium(I) salicylaldiminato-cyclooctadiene complex that facilitated arene C–H borylation in the presence of tetra-2-pyridylpyrazine and an ionic liquid, allowing catalyst reuse after product separation via distillation³⁰. Nishida et al. introduced an insoluble and reusable catalyst derived from [IrCl(cod)]₂ with 2,2-bipyridinedicarboxylic acid (BPDCA)³¹. The catalyst could be easily filtered from the reaction mixture for reuse under nitrogen in a glovebox, and later applied in continuous flow C–H borylation of arenes under the same inert conditions³². Immobilization of iridium(I) complexes onto solid supports such as bipyridine- or phosphine-modified organosilica^33–41, metal-organic frameworks⁴², covalent-organic frameworks⁴³, and porous organic polymers⁴⁴, has also been explored for developing heterogeneous iridium catalysts. For instance, Sawamura et al. immobilized [IrOMe(cod)]₂ on a silica-supported silicon-constrained monodentate trialkylphosphine (silica-SMAP), achieving efficient and selective C–H borylation of arenes with directing groups^37–41. However, the developed catalyst could not be reused. Although several systems have showed reusability for C–H borylation, they still require inert conditions due to the limited stability of iridium(I) complexes. Moreover, despite being more efficient than a traditional batch reaction, synthesis of arylboronates under continuous flow C–H borylation conditions using an immobilized iridium catalyst has not yet been achieved.

Our group has focused on the development of insoluble polymeric metal catalysts for both batch and continuous flow reactions^45–50. Recently, we developed a durable polymeric palladium catalyst by complexing Pd species with a copolymer of 4-vinylpyridine (4VP) and 4-tert-butylstyrene (4tBS) using a molecular convolution method⁵⁰. The steric hindrance provided by the 4tBS component effectively suppressed palladium aggregation, enabling long-term Suzuki–Miyaura reactions under continuous flow conditions. Building on this success, we hypothesized that immobilizing iridium species on an appropriate polymeric framework could stabilize the active species, potentially enabling C–H borylation under both batch and continuous flow conditions.

As is well known, iridium(I) complex such as [IrCl(cod)]₂ is traditionally synthesized by reacting iridium(III) chloride with 1,5-cod in a mixed solvent of ethanol and water under reflux conditions. The Ir(I) species gradually forms through coordination with 1,5-cod and reduction by ethanol (Fig. 1a)⁵¹. Therefore, in this work, by replacing 1,5-cod with a nitrogen-containing polymer under similar reaction conditions, we developed an efficient and stable polymeric iridium catalyst containing Ir(I) species, thereby promoting C–H borylation of arenes under batch and continuous flow conditions (Fig. 1b). Notably, packing the optimal catalyst into a column reactor under ambient air conditions enabled continuous flow synthesis of multiple borylated products using a single reactor, demonstrating its practicality.

Fig. 1. Preparation and conceptual development of pyridine-containing polymeric iridium catalysts for C–H borylation.

Schematics of the a preparation of [IrCl(cod)]₂, and b development of pyridine-containing polymeric iridium catalysts for C–H borylation under batch and continuous flow conditions.

Results and discussion

Poly(4-vinylpyridine) (P4VP) was initially examined as a coordination site for iridium species. As described in Supplementary Information Section 3.1, IrCl₃·xH₂O and P4VP were mixed in a 3:1 molar ratio and reacted in a mixed solvent of ethanol and water at 60 °C for 24 h. During the reaction, the characteristic black color of IrCl₃·xH₂O gradually faded, and a yellow precipitate formed. After filtration and washing with methanol, a yellow powder was obtained in a 94% yield. The colorless filtrate indicated successful immobilization of iridium species on the polymer framework. The catalytic activity of the resulting P4VP-Ir catalyst was initially evaluated in the C–H borylation of anisole (1a, 2 mL, excess) using bis(pinacolato)diboron (B₂pin₂, 2a, 1 mol equiv) as a substrate at 108 °C for 22 h. The reaction afforded a mixture of borylated products (o-, m-, and p-isomers) in a 70% GC yield (Supplementary Table 1, entry 1). After the reaction, the catalyst retained its yellow color, similar to the fresh catalyst, and was easily recovered in a 95% yield. The recovered catalyst was reused under identical conditions, affording the borylated products in a 74% yield (Supplementary Table 1, entry 2). These results suggest that the active Ir(I) species for C–H borylation likely formed either during catalyst preparation in ethanol or in situ during the reaction. Encouraged by these findings, we further explored other catalysts for C–H borylation. We investigated the catalytic performance of poly(3-vinylpyridine) (P3VP) and a copolymer of 4-vinylpyridine and 4-methylstyrene (poly(4VP-co-4MeS)) as polymer ligands (polymer structures shown in Table 1). While the P3VP-Ir catalyst gave a 64% yield, the copolymer-supported catalyst, poly(4VP-co-4MeS)-Ir, significantly improved activity, affording the borylated anisole in a 95% yield (Supplementary Table 1, entries 3 and 4).

Table 1.

Polymeric Ir catalyzed C–H borylation under batch conditions

Entry	Polymeric Ir catalyst	Temp. (°C)	Decane (mL)	Yield (%)a (o-, m-, p-)	Selectivity (%) (o-3a/m-3a/p-3a)
1	poly(4VP-co-4MeS)-Ir (4VP:4MeS = 1:1)	118	2	29	4/73/23
2	poly(4VP-co-4MeS)-Ir (4VP:4MeS = 1:4)	118	2	21	4/74/22
3	poly(4VP-co-4FS)-Ir (4VP:4FS = 1:1)	118	2	6	9/66/25
4	poly(4VP-co-4tBS)-Ir (4VP:4tBS = 1:1)	118	2	21	7/69/24
5	poly(4VP-co-4tBS-co-4FS)-Ir (4VP:4tBS:4FS = 1:1:1)	118	2	21	5/72/23
6	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4)	118	2	45	4/74/22
7	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4)	118	1	50	5/73/22
8	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4)	125	1	61	5/72/23
9b	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4)	125	0.5	55	6/71/23
10b	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4)	125	0.5	60	4/70/26
11b	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4)	125	none	56	8/68/24
12c	poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4)	125	0.5	74	5/72/23

^aCalculated via GC using tridecane as an internal standard.

^b1 mol% Ir was used.

^c1.2 mol% Ir was used.

The effect of copolymer structure on the performance of polymeric iridium catalysts in the C–H borylation of anisole was next examined using Na₃IrCl₆·xH₂O as the iridium precursor due to its superior solubility in the EtOH/H₂O mixture compared to IrCl₃·xH₂O. The preparation procedures for these catalysts are described in Supplementary Section 3.2. As shown in Table 1, anisole (1a) (1.0 mmol, 1 mol equiv) reacted with B₂pin₂ (1.5 mmol) in decane at 118 °C for 24 h. Here, we selected 1a as substrate to demonstrate both synthetic utility and process practicality. Under these conditions, the previously described poly(4VP-co-4MeS)-Ir (4VP:4MeS = 1:1, 3 mol% Ir) promoted the reaction but afforded only 29% yield of borylated products (Table 1, entry 1). Increasing the proportion of 4-methylstyrene (4MeS) in the copolymer to a 4:1 ratio relative to 4VP further decreased the catalytic activity, affording only 21% yield (Table 1, entry 2). Other pyridine-containing copolymers were also screened, including poly(4VP-co-4FS) (4-fluorostyrene as the comonomer) and poly(4VP-co-4tBS) (4-tert-butylstyrene as the comonomer). However, the corresponding iridium catalysts exhibited poor activity, affording the borylated product 3a in 6% and 21% yields, respectively (Table 1, entries 3 and 4). A ternary copolymer, poly(4VP-co-4tBS-co-4FS) (4VP:4tBS:4FS = 1:1:1), was evaluated as a polymer ligand, but the resulting catalyst also showed unsatisfactory performance (Table 1, entry 5). Interestingly, incorporation of N-isopropylacrylamide (NIPAAm) into the polymer framework afforded a 45% yield of 3a when poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4) was used as the catalyst. The monomer ratio in the copolymer was confirmed via ¹H NMR spectroscopy (Supplementary Fig. 1). Using this catalyst, optimization of solvent amount and reaction temperature further increased the yield to 61% (Table 1, entries 7 and 8). Furthermore, increasing the NIPAAm content to 4 mol equiv produced poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4), which exhibited even higher catalytic activity for the C–H borylation of anisole, delivering improved yields at 1 mol% Ir, compared to the catalyst containing 2.4 mol equiv of NIPAAm (Table 1, entries 9 vs. 10). Attempts to further improve the yield by removing the solvent resulted in a decreased yield (Table 1, entry 11), whereas increasing the iridium content to 1.2 mol% improved the yield to 74% (Table 1, entry 12). As a result, this ternary copolymer-immobilized iridium catalyst, poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4), was selected as the optimized catalyst, with the Ir loading confirmed to be 5.6 wt% via inductively coupled plasma mass spectrometry (ICP-MS) analysis. We also conducted the C–H borylation of anisole using unsupported Na₃IrCl₆·xH₂O and [Ir(cod)OMe]₂ as catalysts, respectively. However, only low yields were obtained (Supplementary Table 2, entries 1 and 2). Moreover, when Na₃IrCl₆·xH₂O was added together with the optimal copolymer under the identical reaction conditions, nearly no product was detected (Supplementary Table 2, entry 3). These results indicate that the immobilization of iridium species on the polymer support is crucial.

In addition, the used poly(4VP-co-4tBS-co-NIPAAm)-Ir catalyst (Table 1, entry 12) could be readily recovered from the reaction mixture by simple centrifugation. The recovered catalyst promoted the reaction under the same conditions, affording the borylated products in 65% yield. The recovered catalyst changed in color from yellow to gray and became progressively darker. Moreover, iridium leaching was evaluated by ICP-MS analysis of both the isolated product and the reaction mixture. The results showed 3.13 ppm of Ir (~2.7% leached from the fresh catalyst) was detected in the reaction mixture, and only 0.45 ppm of Ir was detected in the product after silica gel column chromatography purification. These results indicate that gradual aggregation and minor leaching of iridium species probably reduce the catalytic performance under the current batch reaction conditions.

To investigate the oxidation state of Ir in the catalyst, fresh and used poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4) was analyzed using Ir L₃-edge X-ray absorption near-edge structure (XANES) spectroscopy. As shown in Fig. 2a, the spectral profile of the fresh catalyst (red line) differed from that of the precursor (Na₃IrCl₆·xH₂O, green line), indicating that the Ir(III) species interact with the pyridine-containing copolymer, and that the oxidation state of Ir changes during catalyst preparation. Moreover, comparison with reference materials—iridium powder for Ir(0) (black line), [Ir(OMe)(cod)]₂ for Ir(I) (orange line), and IrO₂ for Ir(IV) (purple line)—showed that the spectral profile of the catalyst was distinct from all references, suggesting the presence of a mixture of iridium species with different oxidation states in the prepared catalyst⁵². The recovered catalyst from the C–H borylation of anisole (Table 1, entry 6) exhibited a spectral profile (blue line) that closely resembled that of the fresh catalyst (red line, Fig. 2b). The absorption edges of both samples were positioned between the Ir(I) and Ir(III) reference samples. To further clarify the oxidation states, linear combination (LC) fitting of the XANES spectra was performed for both the fresh and used catalysts from 11.20 to 11.25 keV (50 eV) using Na₃IrCl₆·xH₂O and [Ir(OMe)(cod)]₂ as Ir(III) and Ir(I) references, respectively (Supplementary Fig. 7a). The results indicated that ~52% of the fresh catalyst consisted of Ir(I) component, which increased to 66.1% after the reaction. Although this approach provides semi-quantitative information, the trend observed in the oxidation state changes of the same catalyst before and after use suggests that the Ir(III) component was partially reduced during catalyst preparation and further reduced in situ during the C–H borylation of anisole. Similar analyses were performed for poly(4VP-co-4MeS)-Ir (4VP:4MeS = 1:1) and poly(4VP-co-4FS)-Ir (4VP:4FS = 1:1). As shown in Supplementary Fig. 6, the spectral profiles of the fresh and used catalysts were similar. LC fitting revealed that the Ir(I) component in poly(4VP-co-4MeS)-Ir (4VP:4MeS = 1:1) was 41.6%, while in poly(4VP-co-4FS)-Ir (4VP:4FS = 1:1), it was 39% (Supplementary Fig. 7b, c).

Fig. 2. XAFS measurements of polymeric Ir catalysts.

a L₃-edge XANES spectra of poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:2.4) (red) compared with reference materials: Ir powder (black), IrO₂ (purple), Na₃IrCl₆·xH₂O (green), [Ir(OMe)(cod)]₂ (orange); b direct comparison of L₃-edge XANES spectra for fresh (red) and used (blue) Ir catalysts; c Fourier transform of k³-weighted Ir L₃-edge EXAFS oscillations of fresh and used Ir catalysts and references.

Extended X-ray absorption fine structure (EXAFS) spectra of the fresh and used poly(4VP-co-4tBS-co-NIPAAm)-Ir catalyst (Fig. 2c, red and blue lines) showed a decrease in the intensity of the peak at 2.0 Å, corresponding to the Ir–Cl interaction⁵². Additionally, a peak at ~2.6 Å, assignable to the Ir–Ir interaction, exhibited almost no change in the catalyst after use. These results suggest that although the Ir–Cl interaction in the fresh catalyst was reduced during the reaction, the activated iridium species were stabilized by the polymer framework, thereby preventing severe aggregation. This feature contributes to the potential of the polymeric Ir catalyst, with its suitable structure, as a stable and active species for long-term use.

The batch C–H borylation of several arene compounds was subsequently investigated. As shown in Fig. 3, under the optimized reaction conditions using 1.2 mol% poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4), product 3a, consisting of m- and p-isomers, was isolated in a 52% yield. Other methyl- and methoxy-substituted arenes also underwent the reaction, yielding the corresponding products 3b, 3c, and 3d in 50–61% yields. Notably, product 3d has been utilized as an intermediate in the synthesis of the natural products pestaphthalide A and B⁵³. Additionally, mono- and di-chloro-substituted arenes reacted smoothly with B₂pin₂, providing the corresponding borated products 3e, 3f, and 3g in 62–65% yields. Arenes containing tertiary amine and ester group were also compatible with the catalyst, affording 3h and 3i, both containing m- and p-isomers, in 55% and 71% yields, respectively. Dimethyl isophthalate reacted with B₂pin₂, yielding product 3j in a 66% yield. Furthermore, the C–H borylation of several heteroatom-containing arenes was investigated. Pyridine, indole, benzofuran, and benzothiophene derivatives reacted with B₂pin₂, yielding borylated products 3k–3n in isolated yields of 63–81%. Similarly, 2-butylfuran and 2-hexylthiophene afforded the borylated products 3o and 3p in 75% and 87% yields, respectively. Given the importance of thiophene derivatives in electronic functional materials, we further tested the C–H borylation reaction with several thiophene derivatives. For 3-hexylthiophene, the total yield of the product 3q was 70%, with 26% of the product corresponding to the 2,5-diborylated form, as determined by ¹H NMR spectroscopy. The C–H borylation of 2-chlorothiophene gave product 3r in a 57% yield. Increasing the amount of B₂pin₂ from 1.5 to 2.5 mol equiv enabled the diborylation of both 2,2’-bithiophene and thieno[3,2-b]thiophene, affording products 3s and 3t in yields of 83% and 72%, respectively. However, during the investigation, we found several arenes with strong coordinating ability, such as nitrobenzene, benzoxazole, and N,N-dimethylbenzamide; showed poor reactivity under the standard reaction conditions. In addition, the arenes bearing highly reactive functional groups, such as bromo and amino groups, are also incompatible with our catalytic system because of byproduct formation (Supplementary Table 3).

Fig. 3. Substrate scope for batch C–H borylation using the poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4) catalyst.

^aIsolated yields are shown in this table. ^bDetermined via GC analysis. ^cDetermined via ¹H NMR analysis. ^dContaining 26% diborylated product, 2-[3-hexyl-5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)thiophen-2-yl]-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (3qa). ^e2.5 mol equiv B₂pin₂ was used.

According to the EXAFS results, the polymer framework appears to stabilize the iridium species, potentially enabling the long-term use of the catalyst. To this regard, we examined the catalytic performance of P4VP-Ir under continuous flow conditions. As described in Supplementary Fig. 9, 100 mg (0.163 mmol Ir) of P4VP-Ir was mixed with ~2 g of sea sand and packed into a column reactor (5 mm ID, 10 cm length). A solution containing B₂pin₂ (10 mmol) as a substrate in 20 mL anisole, with 500 μL of tridecane as an internal standard, under the neat conditions was preheated and introduced into the reactor using a syringe pump. The reaction mixture was sampled hourly from the outlet and analyzed via GC. Initially, at 120 °C and a flow rate of 50 μL min⁻¹, the product yield was only 7%. After 2 h, the flow rate was reduced to 25 μL min⁻¹, and the reaction temperature was increased to 130 °C, resulting in a 22% product yield after 6 h, with similar o-, m-, and p-isomer selectivity to that observed in batch reactions. After a 6-h operation, the reactor was washed with anisole and stored under air. On the second day, the reactor was tested again and retained similar activity, affording a 22% yield. Increasing the temperature to 140 °C improved the product yield to 50%. These results indicate that higher temperatures are required under flow conditions to achieve comparable performance to batch reactions. Notably, even though the reaction was conducted at 140 °C, the collected solution remained clear and colorless, and the recovered P4VP-Ir catalyst remained yellow, similar to that of the fresh catalyst (Supplementary Fig. 9).

Building on these initial results, the poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4) catalyst was tested under continuous flow conditions. The catalyst (555 mg, 0.163 mmol Ir) was mixed with sea sand and packed into a column reactor (10 mm ID, 10 cm length). A clear solution containing tridecane (internal standard), B₂pin₂, and 1,2-dichlorobenzene (1f) was injected into the reactor at a flow rate of 25 μL min⁻¹, with the reactor maintained at 140 °C. The residence time was directly measured to be 116 min 40 s. The reaction was conducted for 6 h each day; after each 6-h run, the flow device was cooled to room temperature, and the reactor was stored under air. On the following day, the flow reaction was resumed under identical conditions. In this manner, the reaction was carried out for a total operation time of 24 h over 4 days (Fig. 4, entry 1). As a result, although an induction period was observed on day 1, the reaction proceeded smoothly thereafter, affording the borylated product 3f in an average yield of 181% (calculated based on B₂pin₂ as 1 mol equiv) with a TOF of 8.3 h⁻¹. The space-time yield (STY) was calculated to be 1.13 kg/(L.day). These results indicate that this continuous-flow method has potential as an effective approach for C–H borylation using a fixed amount of Ir catalyst.

Fig. 4. C–H borylation of 1,2-dichlorobenzene (entry 1) and multiple arenes using a single column reactor (entry 2) under continuous flow conditions.

The yields were calculated based on B₂pin₂ as 1 mol equiv via GC analysis, using tridecane as an internal standard.

To further assess the practicality of this catalytic system, three different arenes were subjected to C–H borylation using the same reactor (Fig. 4, entry 2). The experimental procedures were similar to those described in entry 1, and 1,2-dichlorobenzene was again investigated as the first substrate. Although the yield differed from that in entry 1 due to the catalyst being prepared at a different time, an induction period was also observed. The yield of 3f exceeded 100% after 2 h and reached 171% after 5 h (Fig. 4, entry 2, reaction 1). On the following day, the substrate was changed to methyl benzoate. A fresh starting solution was pumped into the reactor for 2 h at room temperature (~22 °C) prior to heating. After this period, the heater was turned on to raise the temperature to 140 °C. The starting solution was then continuously pumped into the reactor for an additional 2 h, after which collection of the reaction mixture was initiated. In this reaction, although the yields of 3i (including the m- and p-isomers) did not exceed 100%, the polymeric Ir catalyst still demonstrated good performance, affording yields ranging from 86% to 95% over a 4-h reaction (Fig. 4, entry 2, reaction 2). These results demonstrate that the same reactor can be used sequentially to synthesize multiple products in a continuous flow system. On the third day, anisole was employed as the substrate. Although only 19% of borylated anisole (including the o-, m-, and p-isomers) was detected in the first hour, the yield gradually increased to 66% after 4 h (Fig. 4, entry 2, reaction 3), after which the reaction was stopped. The same reaction using anisole was restarted on the fourth day, giving a yield of 57% for 3a, which stabilized at ~65% from the second hour onward. Extending the reaction to 11 h resulted in stable and slightly increased product yields. To further evaluate the catalytic stability, the reaction was continued on days 5 and 6 for 9 h each day. Although the yield decreased to ~55%, the system remained stable without significant catalyst deactivation. On the seventh day, we attempted to restart the reaction under the same conditions. However, the collected solution from the outlet initially did not match the calculated amount, and the system soon became clogged. We therefore suspect that the polymer material may have swollen after prolonged contact with the organic reagents. In total, the continuous-flow reactions were performed for 42 h.

Conclusion

In this study, a series of insoluble pyridine-containing polymeric iridium catalysts were developed. A ternary copolymer, synthesized from the polymerization of 4-vinylpyridine, 4-tert-butylstyrene, and N-isopropylacrylamide, was identified as an effective support for immobilizing and stabilizing the iridium species. Using this optimized catalyst, C–H borylation of various arenes was achieved under both batch and continuous flow conditions. Notably, multiple borylated products were synthesized under air in a single reactor simply by switching the substrate solutions. XAFS analysis of both the fresh and used catalysts revealed that the iridium precursor was partially reduced during catalyst preparation and further reduced in situ during the C–H borylation reaction. Therefore, although the recovered catalyst still faces a decline in catalytic activity, this novel strategy for the design and synthesis of immobilized iridium catalysts provides valuable guidance for advancing the development of practical catalytic systems for C–H borylation reactions.

Methods

Preparation of copolymers

The stabilizer (hydroquinone, HQ) in 4VP was removed via column chromatography with activated Al₂O₃ (~300 mesh, FUJIFILM Wako Pure Chemical Corporation). The stabilizer (4-tert-butylcatechol, TBC) in 4tBS was removed by washing with aq. NaOH (0.9 M) three times in a separating funnel. The resulting organic layer was washed with brine and dried over anhydrous MgSO₄ prior to use. N-Isopropylacrylamide (NIPAAm) was used as received. Polymerization was carried out in a 50-mL two-necked round-bottomed flask equipped with a stopcock. The flask was charged with 10 mL of 1-dodecanol and 2 mL of toluene. The atmosphere was exchanged with argon gas three times, followed by the addition of 1.051 g (10 mmol) of 4VP, 1.603 g (10 mmol) of 4tBS, 3.395 g (30 mmol) or 5.658 g (50 mmol) of NIPAAm, and 1 wt% of AIBN relative to the total amount of monomers. The mixture was degassed with argon gas five times, sealed, and stirred at 70 °C for 40 h. The resulting precipitate was transferred into a 200-mL beaker, washed with hexane and H₂O, and dried under reduced pressure overnight to obtain the corresponding copolymers, which were then ground in a mortar. For copolymer F (using 30 mmol NIPAAm), 4.895 g (81% yield) was obtained as white powder. For copolymer G (using 50 mmol NIPAAm), 5.755 g (69% yield) was obtained as white powder.

Preparation of polymeric Ir catalyst

Poly(4VP-co-4tBS-co-NIPAAm) (4VP:4tBS:NIPAAm = 1:1:4, 997.4 mg) was dispersed in 6 mL of EtOH and stirred at 23 °C in a 20-mL flask. Na₃IrCl₆·xH₂O (189.6 mg, 0.4 mmol, anhydrous basis) was dissolved in 2 mL of H₂O and added dropwise to the copolymer solution. The mixture was stirred at 23 °C under air for 1 h, forming a gray precipitate. A mixed solvent (EtOH/H₂O = 3:1, 8 mL) was then added, the flask was purged with argon three times, sealed, and heated at 60 °C for 24 h upon stirring. The resulting yellow precipitate was collected via filtration, washed with H₂O and MeOH, and dried under reduced pressure overnight to afford 1043.2 mg of the poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4) catalyst (yield: 93%). The filtrate was colorless and transparent. ICP-MS analysis confirmed an Ir content of 5.6 wt% in the catalyst.

Batch C–H borylation reaction

A 10-mL round-bottom screw-top vial equipped with a stir bar was charged with 5.6 wt% poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4; 41.75 mg, 1.2 mol% Ir), anisole (108.1 mg, 1 mmol), B₂pin₂ (380.91 mg, 1.5 mmol), and decane (0.5 mL). The vial was purged with argon gas, sealed, and stirred at 125 °C in a ChemiStation device for 24 h. After cooling to room temperature, the mixture was diluted with ethyl acetate (8 mL), and 50 μL of tridecane was added as an internal standard. The solution was then filtered through Celite, yielding a clear, colorless solution for GC analysis. The yield was determined using a GC calibration curve with tridecane as the internal standard.

Continuous flow C–H borylation of 1,2-dichlorobenzene

Poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4, 555 mg, 0.163 mmol Ir) was mixed in a vial with sea sand (~8.5 g) using a vortex mixer (KENIS Ltd., Vortex-Genie 2) for 5 min at 1000 rpm, and then packed into a glass column (10 cm L × Φ10 mm) with two PTFE filters installed at both ends. The packed column was placed in a flow device (purchased from TOKYO RIKAKIKAI Co, Ltd., MCR-1000) heated at 140 °C. A substrate solution containing B₂pin₂ dissolved in 1,2-dichlorobenzene with tridecane as an internal standard was preheated at 140 °C using a hot plate and injected into the reactor using a syringe pump at a flow rate of 25 μL min⁻¹. After the effluent was collected hourly from the outlet and diluted with EtOAc prior to analysis via GC. The yield was calculated using tridecane as an internal standard. After the reaction, the system was cooled to room temperature and the reactor was stored under air. On the following day, the starting material solution was injected into the reactor under the same conditions.

TOF calculation

[flow rate × 60 (min) × concentration × yield]/n (Ir).

Space-time yield (STY) calculation

STY = m (product)/time×V (catalyst).

Continuous flow C–H borylation of multiple arenes

Poly(4VP-co-4tBS-co-NIPAAm)-Ir (4VP:4tBS:NIPAAm = 1:1:4, 555 mg, 0.163 mmol Ir) was mixed in a vial with sea sand (~8.0 g) using a vortex mixer (KENIS Ltd., Vortex-Genie 2) for 5 min at 1000 rpm, and then packed into a glass column (10 cm L × Φ10 mm) with two PTFE filters installed at both ends. The packed column was placed in a flow device heated at 140 °C. A substrate solution containing B₂pin₂ dissolved in arene with tridecane as an internal standard, was preheated at 140 °C using a hot plate and injected into the reactor using a syringe pump at a flow rate of 25 μL min⁻¹. The effluent was collected hourly from the outlet and diluted with EtOAc prior to analysis via GC. The yield was calculated using tridecane as an internal standard. After the reaction, the system was cooled to room temperature, and the reactor was stored under air. On the following day, the substrate was changed to methyl benzoate (reaction 2), and the starting material solution was prepared using the same procedure. For this reaction, the fresh starting solution was pumped into the reactor for 2 h at room temperature before heating. After this period, the heater was turned on to raise the temperature to 140 °C. The starting solution was then pumped into the reactor for an additional 2 h, after which collection of the reaction mixture was initiated. On day 3, the substrate was changed to anisole (reaction 3), following the same procedure described for reaction 2. On days 4–6, the flow device heater was set directly to 140 °C. The starting solution was pumped into the reactor for 1 h, after which collection of the reaction mixture was initiated.

Author contributions

Y.M.A.Y. directed the project. Z.Z. and Y.M.A.Y. designed the experiments. Z.Z. conducted the experiments, A.O. performed ICP-MS measurements. Z.Z., A.O., and Y.M.A.Y. wrote and revised the manuscript. All authors have approved the final version of the manuscript.

Peer review

Peer review information

Communications Chemistry thanks Malek Y. S. Ibrahim and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Data availability

All data generated during this study (experimental methods, compound characterizations, ¹H NMR, ¹³C NMR, GC-MS, HRMS) are provided in the Supplementary Information file (PDF), the NMR spectra are shown in Supplementary Data 1 file (PDF).

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42004-026-02044-0.