Introduction

Organic photosensitizers, a representative class of non-metal photocatalysts, have received increasing research interest1,2,3. In fact, nature evolves pigments to utilize solar energy for important chemical transformations4. Compared with inorganic photocatalysts (PCs), the organic ones possess a relatively broader redox window, and thus a stronger capability to active inert molecules3. However, their disadvantages of a shorter excited-state lifetime and lower stability lead to an inferior catalytic performance in some cases. For example, diaryl dihydrophenazines are visible-light PCs with strongly reducing ability (E0(PC•+/3PC*) < − 2 V versus saturated calomel electrode) upon excitation. Although they efficiently catalyzed the atom transfer radical polymerization5, the carbon radical intermediate was recently found to attack the PCs (Fig. 1a)6,7. The susceptibility of dihydrophenazine PCs to the reaction with open-shell radical species may restrict their applications in photoredox catalysis.

Fig. 1: Diaryl dihydrophenazine catalysis.
figure 1

a Representative example of visible-light dihydrophenazine catalysis and its limitation. Ar aromatics, SCE saturated calomel electrode. b Enhancing the performance of dihydrophenazine EDA photoactivation by MOF platform. EDA electron donor–acceptor, MOF metal-organic framework; light blue: Zr, red: O, blue: N, gray: C.

Full size image

In addition to traditional photoredox catalysis enabled by photosensitizers, the recently developed electron donor–acceptor (EDA) photoactivation provides an alternative mode for photochemical synthesis8,9. The proper electron donor catalyst can interact with electron acceptor substrates to form a photoactive EDA complex. Upon light irradiation, intra-complex single electron transfer (SET) from donor to acceptor occurs to give active open-shell species for further transformations10,11,12,13. In most cases, the EDA interaction is “weak”, thus a relatively high concentration is preferred to promote the chances of generating EDA adducts. Despite fruitful achievements, the development of the corresponding heterogeneous version of EDA photoactivation remains inaccessible.

Metal-organic frameworks (MOFs)14,15,16,17,18,19,20,21,22,23,24,25,26, a class of 3D porous materials by bridging metal-containing second building units (SBUs) with organic linkers, have proved their superiority in engineering heterogeneous catalysis due to the wide synthetic tunability and unique porous structure27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43. Specific to MOF-based photocatalysis, factors including site isolation effect, confinement effect, high local concentration, and active-center protection endow MOF catalysts with superior catalytic performance compared with homogeneous counterparts44,45,46,47,48,49,50,51,52,53,54,55,56,57,58. A variety of metal-based PCs have been integrated into MOFs for photoinduced synthesis, while the merger of organic PCs remains less investigated because of their typically low symmetry and high steric demand59,60,61,62. Among the aforementioned factors, a high local concentration of active centers in MOF was surmised to promote the EDA interaction and related EDA photoactivation, which remains elusive in heterogeneous catalysis. Moreover, integrating diaryl dihydrophenazine into MOFs would reduce the risk of being attacked by radical intermediates and therefore expand their catalytic applications. Herein, we report the design of photoactive MOF Zr-PZDB (PZDB = 4,4’-(phenazine-5,10-diyl)dibenzoate), consisting of Zr6-SBUs and PZDB connecting ligands (Fig. 1b). Upon visible light irradiation, Zr-PZDB competently catalyzed the Minisci-type cross-coupling of N-heterocycles with ethers, alcohols, non-activated alkanes, amides, and aldehydes. Zr-PZDB also enabled the coupling reaction of aryl sulfonium salts with heterocycles. Furthermore, the late-stage functionalization of complex drug or bioactive molecules was attained by using Zr-PZDB as the catalyst63. In contrast, the homogeneous counterparts PZDB-H or PZDB-Me exhibited inferior catalytic efficiency. Our systematical spectroscopy study and analysis revealed the enhanced EDA interaction between Zr-PZDB and pyridinium salt or sulfonium salt. The subsequent photoinduced intra-complex SET allowed the generation of radicals and enabled the corresponding coupling reactions.

Results and discussion

The PZDB-H linker was synthesized according to the reported procedure (see Supplementary Information for details)64. PZDB-H presents a rigid single-crystal structure (Supplementary Fig. S3), suggesting its potential to be bridging ligands for MOF synthesis. Solvothermal reaction of PZDB-H ligand with ZrCl4 in N,N-dimethylformamide (DMF) using trifluoroacetic acid as the modulator delivers octahedron single crystals of Zr-PZDB (Fig. 2a and Supplementary Fig. S4a; during the preparation of this manuscript, the Queen group reported a convenient synthesis of a series of MOFs with piperazine core65). The single-crystal structure suggests a formula of Zr6(μ3-O)4(μ3-OH)4(PZDB)6 for Zr-PZDB, which was further verified by TGA (thermogravimetric analysis, Supplementary Fig. S8) result with the residue weight of 23.7 wt% for ZrO2 close to the theoretical value of 23.1 wt% from the formula. Zr-PZDB exhibits a space group of Fm—3m (No. 225; see Supplementary Table S2) and fcu topology, wherein Zr6O4(OH)4 clusters are connected by 12 PZDB linkers in the face-centered-cubic array, in line with the iso-reticular structures of UiO-6866. The dihydrophenazine moiety in PZDB of Zr-PZDB retains the planar geometry. However, it is disordered due to the rotation around the PZDB backbone (Supplementary Fig. S4b). The shortest distance between two adjacent linkers is measured as ~2 Å, indicating the relatively steric demanding environment (Supplementary Fig. S5), which might account for the excellent protection of PZDB active centers by Zr-PZDB (vide infra).

Fig. 2: Synthesis and characterization of Zr-PZDB.
figure 2

a Schematic showing the synthesis of Zr-PZDB and the structure: X-ray single-crystal structure represented by an octahedron-like cage; disorder and H atoms are omitted for clarity; light blue: Zr, red: O, blue: N, gray: C. b PXRD patterns of Zr-PZDB; arb. u. = arbitrary units. c, d SEM (c) and HRTEM (d) of Zr-PZDB. e IR spectra of Zr-PZDB and PZDB-H. f Zoom-in 1H NMR of digested Zr-PZDB in DMSO-d6. g Normalized excitation and emission spectra of Zr-PZDB and PZDB-H in MeCN; arb. u. = arbitrary units. h UV-vis spectra of 1a, Zr-PZDB, and their mixture in MeCN (2.5 × 10−3 M); a. u. = absorbance units. i UV-vis spectra of 1a, PZDB-H, and their mixture in MeCN (2.5 × 10−3 M); a. u. = absorbance units. j UV-vis spectra of mixtures of 1a and Zr-PZDB with different ratios of 1a in MeCN (5 × 10−3 M); a. u. = absorbance units. k Job’s plot based on UV-vis data in (j).

Full size image

Powder X-ray diffraction (PXRD) analyses indicated the high phase purity of Zr-PZDB. The corresponding PXRD pattern matches well with the simulated one from the single crystal structure (Fig. 2b). In addition, Zr-PZDB retained the crystalline structure after being soaked in a series of solvents, as evidenced by PXRD analyses (Fig. 2b and Supplementary Fig. S10). Scanning electron microscopy (SEM) imaging showed the octahedron morphology of Zr-PZDB (Fig. 2c). High-resolution transmission electron microscopy (HRTEM) imaging of Zr-PZDB revealed a distance of ~2.4 nm between two Zr6-SBUs (Fig. 2d), matching well with the distance of ~2.3 nm in the single crystal structure (Supplementary Fig. S6). IR spectra of PZDB-H and Zr-PZDB were collected. A stretching band at 1690 cm−1 assignable to the carbonyl group was observed for PZDB-H, while the carbonyl groups in Zr-PZDB showed the stretching band at 1654 cm−1 (Fig. 2e). The difference originated from the bonding of carboxylic group with metals. Only one set of signals assigned to PZDB-H was detected by 1H NMR analysis of the digested Zr-PZDB, indicating the inertness of the bridging ligand under MOF preparation conditions and the high purity of MOF (Fig. 2f). The porosity of Zr-PZDB was investigated by gas absorption analyses (Supplementary Fig. S9). The Brunauer–Emmett–Teller surface area of Zr-PZDB was measured as 715 m2/g with the calculated main pore window of ~1.3 nm. As shown in Fig. 2g, upon excitation at 398 nm, PZDB-H exhibited two characteristic emission peaks at 450 and 550 nm, respectively. However, the intensity of the 450 nm peak was significantly decreased, with a main emission peak at 550 nm for Zr-PZDB. Meanwhile, PZDB-H and Zr-PZDB showed similar excitation signals for the emission at 550 nm.

Functionalization of N-heterocycles is of great importance from their potential bioactivity viewpoints67. In addition to traditional synthetic routes, photoinduced derivatization via the intermediacy of open-shell radicals represents an alternative strategy. Taking this into consideration, the Minisci-type cross-coupling of pyridinium salt68,69,70 with tetrahydrofuran was chosen as the model reaction to evaluate the catalytic performance of Zr-PZDB. As shown in Table 1, Zr-PZDB competently catalyzed the cross-coupling reaction upon blue LED irradiation (Kessil PR160L-427, 390–470 nm, Supplementary Fig. S30) to give the target product 3a in 80% isolated yield. Screening other solvents proved that CH3CN was the optimal choice (entries 2–4, Table 1). Base NaHCO3 outperformed other inorganic bases, due probably to its relatively better solubility (entries 5,6, and 8, Table 1). The use of NEt3 led to a reduced yield of 12% for 3a (entries 5-8, Table 1), because of its potential to quench methoxy radical via hydrogen atom transfer reaction. The linker PZDB-H or PZDB-Me instead of Zr-PZDB as the catalyst delivered 3a in only 6% and 13% yields, respectively (entries 9 and 10, Table 1). Loading PZDB-H into activated carbons as the catalyst provided a yield of 15% for 3a (entry 17, Table 1 and Supplementary Fig. S36). These results indicated the important role of the MOF platform in the catalytic performance. Noteworthily, the use of several representative molecular photosensitizers, including Ir(ppy)3, Ru(bpy)32+, and Eosin Y, afforded inferior results (entries 11–13, Table 1). The control experiments suggested that both light and MOF catalyst were indispensable for the efficient transformation (entries 14 and 15, Table 1). A combination of ZrCl4 and PZDB-H instead of Zr-PZDB led to trace amounts of 3a (entry 16, Table 1).

Table 1 Optimization of reaction conditionsa
Full size table

The individual solution of 1a and suspension of Zr-PZDB were almost colorless; however, their mixture exhibited a light-yellow color (Fig. 2h). This phenomenon suggested the potential involvement of EDA interaction. Noteworthily, the Lakhdar lab pioneered the identification of an EDA interaction between the organic dye Eosin Y and N-ethoxy-2-methylpyridinium tetrafluoroborate71. To further verify the feasibility, a detailed spectroscopy study was then carried out. UV-vis spectrum of the mixture presented a significantly enhanced peak from 330 to 430 nm with respect to the absorption profiles of the individual 1a and Zr-PZDB (Fig. 2h), indicating the formation of an EDA complex. In sharp contrast, the UV-vis spectrum of a mixture of 1a and PZDB-H was almost identical to the sum of individual spectra of 1a and PZDB-H (Fig. 2i). The obvious difference might be attributed to the high local concentration of dihydrophenazine active centers in Zr-PZDB, which are believed to benefit the formation of EDA adducts kinetically. UV-vis spectra of a series of mixtures with different Zr-PZDB/1a ratios were subsequently collected (the ratio of Zr-PZDB is based on the PZDB linker, Fig. 2j). The maximum absorption enhancement was observed for a Zr-PZDB/1a ratio of 1/1 (green line in Fig. 2j). The corresponding Job plot further verified the optimal ratio of 1/1 for the plausible EDA complex (Fig. 2k and Supplementary Table S3). In addition, no absorption enhancement was detected by mixing 1a with NaHCO3 (Supplementary Fig. S13). To further illustrate the role of EDA interaction in this photoinduced Minisci-type transformation, a 420 nm band filter, which blocks lights with a wavelength smaller than 420 nm, was used with a Kessil PR160L-427 lamp (390-470 nm). No product 3a was detected with 1a being intact (Supplementary Figs. S32 and S33). Furthermore, the utilization of the Kessil PR160L-390 lamp (370-420 nm) provided a yield of ~50% for 3a (Supplementary Fig. S32), with the detection of ~50% 4-methylpyrdine (Supplementary Fig. S34). The addition of a 380 nm band filter increased the yield to ~80% (Supplementary Figs. S32 and S35). These results indicate that (1) light with a wavelength larger than 420 nm cannot induce the reaction; (2) light less than 380 nm in wavelength might decrease the selectivity for 3a; (3) the likely involvement of EDA interaction in this photoinduced transformation72. To demonstrate the considerable interaction between Zr-PZDB and pyridinium salt 1a, an adsorption control experiment was designed and performed (Supplementary Fig. S38). After being stirred with 1 equiv of Zr-PZDB overnight, the CH3CN solution of 1a only retained 5% of pyridinium salt. In other words, Zr-PZDB competently adsorbed 95% of 1a, indicating their strong interaction.

The scope of dehydrogenative cross-coupling was then explored, and the results were compiled in Fig. 3. With Zr-PZDB as the catalyst, five different kinds of C-H coupling partners, including ethers, alcohols, unactivated alkyls, amides, and aldehydes, reacted with pyridinium salts efficiently, evidencing the versatility of MOF catalyst. Both cyclic and linear ethers worked well to give products 3a-3d in moderate to very good yields. Pyridinium salts bearing various substituents such as alkyl, phenyl, ester, cyanide, methoxy, and trifluoromethyl groups were compatible (3e-3k and 3m-3n), and no obvious electronic effects were observed. The substrate 1-methoxylepidinium methyl sulfate afforded the corresponding 3 l in 70% yield. Pyridinium salts without a substituent on the C4 position delivered ortho- and para-functionalized products with C2 to C4 ratios of 1.2 to 2.9 (3o-3s). Noteworthily, relatively bulky substituents on the C3 position sterically protect the adjacent C2 and C4 positions, resulting in very good C6 selectivity (3aj and 3al). Methanol and ethanol also served as good coupling partners, and the corresponding pyridinyl alcohols 3t-3v were obtained. Without C4 protection, both ortho- and para-substituted products were isolated with ratios from 1.4 to 5.4 (3w-3y).

Fig. 3: Zr-PZDB catalyzed Minisci-type cross-coupling reactions.
figure 3

Reactions were conducted at 0.05 mmol scale; yields of isolated products; red: the newly formed bond; blue: the reaction sites.

Full size image

Direct replacement of ether by amine, for example, triethylamine, did not yield the coupling product. The easy oxidation of the corresponding α-C radical of triethylamine might account for the failed attempt. However, the dimethylacetamide (DMA) was tolerated to produce 3ad-3ag efficiently. Furthermore, the non-activated, cyclic alkanes worked as appropriate feedstocks, giving the resultant 3z-3ac in good yields. The ring size had little effect on the reaction efficiency. In addition to nucleophilic C(sp3)-H coupling partners, substrates containing carbonyl C(sp2)-H also underwent the cross-coupling reaction. Treatment of dimethylformamide (DMF) or benzaldehyde with 1-methoxylepidinium methyl sulfate under standard reaction conditions afforded 3ad and 3ae, respectively. No target product was observed when 1a was used as the substrate. Noteworthily, Zr-PZDB effectively catalyzed the late-stage functionalization of complicated drug or bioactive molecules containing Nikethamide (3aj), Admiral (3ak), and Myristyl Nicotinate (3al).

To further evaluate the ability of Zr-PZDB donor catalyst, aryl sulfonium salt was examined as the acceptor and substrate (Fig. 4). Upon getting one electron, aryl sulfonium salt would undergo decomposition to generate aryl radical, which is more reactive and sometimes incompatible with homogenous catalytic systems13,73. To our delight, Zr-PZDB effectively catalyzed the coupling reaction between aryl sulfonium salt 4a and 1-methylpyrrole upon light (PR160L-427) irradiation, giving 6a in 87% isolated yield (entry 1 in Supplementary Table S4). Replacement of Zr-PZDB by PZDB-H or PZDB-Me led to decreased yields of 39% and 12% for 6a, respectively (entries 2 and 3 in Supplementary Table S4). PZDB-H@C as the catalyst provided a reduced yield of 36% (entry 7 in Supplementary Table S4 and Supplementary Fig. S37). The combination of ZrCl4 and PZDB-H instead of Zr-PZDB afforded 6a in 45% yield (entry 4 in Supplementary Table S4). The absence of Zr-PZDB or light resulted in almost no target product (entries 5 and 6 in Supplementary Table S4), suggesting their importance in the coupling reaction.

Fig. 4: Zr-PZDB catalyzed cross-coupling of aryl sulfonium salts and heterocycles.
figure 4

Reactions were conducted at 0.2 mmol scale; yields of isolated products; red: the newly formed bond.

Full size image

A series of aryl sulfonium salts and heterocycles were then evaluated (Fig. 4). Functional groups, including alkyls, ethers, esters, halides, and triflate, were well tolerated, among which esters, halides, and triflate could be conveniently transformed to other substituents (6a-6h). Pyrrole, indole, thiophene, and furan underwent the coupling reaction smoothly to deliver 6i-6r in 52-91% isolated yields. No obvious electronic effects were observed. In view of the prevalence of heterocycle derivatives in drug or bioactive molecules, this methodology might find potential applications in biochemistry.

Investigation on the EDA interaction between Zr-PZDB and 4a was also performed. UV-vis spectrum of the corresponding mixture showed the absorption enhancement from 350 nm to 430 nm compared with the absorption profiles of the individual 4a and Zr-PZDB (Fig. 5a). In contrast, no obvious absorption enhancement was observed from the UV-vis spectrum of the mixture of PZDB-H and 4a (Fig. 5b). Moreover, at a high concentration (2.5 × 10−3 M), absorption enhancement was more obvious for the mixture of Zr-PZDB/4a with little changes for the mixture of PZDB-H/4a (Supplementary Fig. S12). These results further evidenced the enhancement effect of the MOF platform on the EDA interaction. To verify the involvement of radical species, radical capture and clock experiments were carried out (Fig. 5c). The addition of a radical scavenger, TEMPO, to the reaction mixture of 1a/THF completely inhibited the formation of product 3a, with the detection of radical capture species 7 by high-resolution mass spectrometry (HRMS). On the other hand, radical scavenger 1,1-diphenylethene successfully captured the more reactive aryl radical as proved by the formation of 8. In addition, the results of radical clock experiments further verified the involvement of alkyl and aryl radicals in the coupling reactions, respectively.

Fig. 5: Mechanistic investigation.
figure 5

a UV-vis spectra of 4a, Zr-PZDB, and their mixture in DMSO (5 × 10−4 M); a. u. = absorbance units. b UV-vis spectra of 4a, PZDB-H, and their mixture in DMSO (5 × 10−4 M); a. u. = absorbance units. c Radical capture and clock experiments. d, e Light on/off experiments for coupling reactions of 1a/THF (d) and 4a/pyrrole (e). f HRMS spectrum of crude reaction mixture catalyzed by PZDB-Me. g Zoom-in 1H NMR of digested Zr-PZDB as prepared (up) and after catalytic reaction (down). h, i Proposed reaction mechanisms for Minisci-type reaction (h) and pyrrole arylation (i).

Full size image

Light on/off experiments for the coupling between 1a and THF indicated the continuation of the reaction in the dark (Fig. 5d), suggesting a radical chain pathway. In sharp contrast, almost no reaction was detected for coupling between 4a and pyrrole in the dark (Fig. 5e). In addition, its corresponding quantum yield was measured as 0.003 (Supplementary Information “Determination of quantum yield” section), negating the radical progression pathway74. A similar phenomenon was observed in hot-filtration control experiments. The removal of the Zr-PZDB catalyst from the reaction mixture of 1a and THF after 12-hour irradiation did not totally shut down the reaction. The yield for 3a was increased from 36% to 52% without MOF catalyst under light irradiation (Supplementary Fig. S26), albeit a higher yield of 80% in the presence of MOF catalyst. However, little yield increase for 6a was observed after removing the Zr-PZDB catalyst from the reaction mixture of 4a and pyrrole (Supplementary Fig. S27). In addition, the leaching of Zr was detected as <0.3% for both coupling reactions by ICP-MS, demonstrating the stability of Zr-PZDB under the coupling reaction conditions. These results combined with the light on/off experiments indicated: (1) the coupling reaction between 1a and THF would involve the radical chain pathway, whereas the reaction of 4a and pyrrole did not undergo radical progression; and (2) the heterogeneous nature of Zr-PZDB catalysis in both coupling reactions.

The Zr-PZDB catalyst was recovered and used in three runs of cross-coupling reactions of 1a/THF and 4a/pyrrole with slightly decreased catalytic performance (Supplementary Figs. S16 and S17). Zr-PZDB retained the crystalline structure, as evidenced by the matching PXRD pattern after the reaction (Fig. 2b). To compare the stability of PZDB active center in Zr-PZDB and homogeneous counterpart, attempts to recover and characterize the catalysts after the reaction were carried out. 1H NMR spectrum of the digested Zr-PZDB-AR (AR = after reaction) was almost identical to that of the as-prepared Zr-PZDB (Fig. 5g and Supplementary Fig. S18). In contrast, thin layer chromatography (TLC) and 1H NMR analyses of the homogeneous reaction mixture indicated the disappearance of the PZDB-Me catalyst (Supplementary Figs. S19 and S20). HRMS analysis further suggested the formation of mono-, di-, and tri-substituted derivatives of PZDB-Me probably originating from the attack of the corresponding aryl radicals (Fig. 5f and Supplementary Fig. S21). These results verified the reported susceptibility of dihydrophenazine molecular catalyst against active radical species and highlighted the protection ability of the MOF platform.

On the basis of the above experimental results and seminal literature precedents75,76,77,78,79,80, plausible reaction mechanisms are proposed (Fig. 5h, i). For Minisci-type reaction, integration of Zr-PZDB with pyridinium salt gives an EDA complex. Upon visible light irradiation, the intra-complex SET from Zr-PZDB to the pyridinium salt occurs to afford neutral pyridine and methoxy radical. The latter competently abstracts hydrogen from C-H bonds to generate carbon-centered radicals, which then attack the pyridinium salt to give the radical adduct intermediate B. Deprotonation of B by neutral pyridine or NaHCO3 delivers the intermediate C. The intermediate C then undergoes decomposition to form the final product and methoxy radical. The oxygen radical can further go through HAT with C-H coupling partners to start another cycle. Alternatively, intermediate C may reduce the MOF radical cation to regenerate the MOF catalyst and afford the substituted pyridinium salt D, which then forms an EDA adduct with Zr-PZDB followed by light excitation and decomposition to give the final product. For coupling reaction between aryl sulfonium salt and pyrrole, Zr-PZDB interacts with sulfonium salt to form the EDA adduct, which undergoes SET upon light irradiation to give MOF radical cation and aryl radical. The latter attacked pyrrole to deliver intermediate E, which is then oxidized by MOF radical cation to regenerate the Zr-PZDB catalyst and give intermediate F. Deprotonation of F finally delivers the coupling products.

In summary, we have integrated dihydrophenazine into MOF to realize a series of cross-coupling reactions. The Zr-PZDB MOF served as the heterogeneous donor catalyst to interact with the pyridinium and sulfonium acceptors for generating the photoactive EDA complexes, respectively. Photoactivation of the EDA adducts triggered the intra-complex SET and the coupling reactions. The high local concentration of PZDB active centers in Zr-PZDB is believed to promote the EDA interaction kinetically, accounting for the superior catalytic performance of Zr-PZDB over homogeneous counterparts. Zr-PZDB competently catalyzed the dehydrocoupling between pyridinium salts and ethers, alcohols, non-activated alkanes, amides, and aldehydes, leading to a wide variety of N-heteroarene derivatives. In addition, Zr-PZDB enabled the synthesis of various aryl heteroarenes from aryl sulfonium salts and heterocycles. Some of the resultant products may have potential bioactivity. Moreover, the developed Zr-PZDB catalysis can be used for attaining the late-stage functionalization of drug and/or bioactive molecules, including Nikethamide, Admiral, and Myristyl Nicotinate. The MOF scaffold effectively protected the dihydrophenazine active center, resulting in relatively low catalyst loading and good durability. This research not only highlights the potential of MOF engineering to address the limitations of organic catalysis but also paves an innovative avenue to green and sustainable MOF-based EDA photoactivation.

Methods

Preparation of Zr-PZDB: PZDB-H (4.2 mg, 0.01 mmol), ZrCl4 (8.4 mg, 0.036 mmol), and CF3COOH (28.5 mg, 0.25 mmol) were mixed in DMF (0.4 mL) in a closable flash. The mixture was then heated at 120 °C for 48 h. After cooling to room temperature naturally, the yellow crystalline solid was obtained by centrifugation and then sequentially washed with DMF three times. Solvent exchange with benzene (3 × 5 mL, replaced by fresh benzene every 8 h) was then conducted. Then, the resultant crystalline solid was dried by vacuum to afford Zr-PZDB (4.4 mg, 82% yield).

General procedure for Minisci-type reaction: N-Methoxy pyridinium methylsulfate (0.05 mmol), CH coupling partners (3.8 mmol), NaHCO3 (8.4 mg, 0.10 mmol), and Zr-PZDB (1.3 mg, 2.5 μmol, 5 mol% based on linker) were mixed in acetonitrile (0.5 mL) in a sealed test tube. The resulting mixture was stirred under blue LED irradiation (PR160L-427, 390-470 nm) at room temperature in a N2 atmosphere for 24 h. After that, the solvent was removed under vacuum, and the residue was subjected to column chromatography on silica gel to give products 3.

General procedure for heterocycle arylation: Dibenzothiophenium salt (0.2 mmol), heterocycle (8.0 mmol), and Zr-PZDB (5.3 mg, 10.0 μmol, 5.0 mol% based on the linker) were mixed in DMSO (1.0 mL) in a sealed test tube. The resulting mixture was stirred under blue LED irradiation (PR160L-427, 390–470 nm) at room temperature in a N2 atmosphere for 24 h. After that, the reaction was quenched with aqueous saturated NaHCO3 and diluted with EtOAc. The organic layer was washed with brine, dried with Na2SO4, filtered, and concentrated in vacuo. The residue was subjected to column chromatography on silica gel to give products 6.