Zr-Based Metal−Organic Frameworks with Phosphoric Acids for the Photo-Oxidation of Sulfides

Heterogeneous Brønsted acidic catalysts such as phosphoric acids are the conventional activators for organic transformations. However, the photocatalytic performance of these catalysts is still rarely explored. Herein, a novel Zr-based metal−organic framework Zr-MOF-P with phosphoric acids as a heterogeneous photocatalyst has been fabricated, which shows high selectivity and reactivity towards the photo-oxidation of sulfides under white light illumination. A mechanism study indicates that the selective oxygenation of sulfides occurs with triplet oxygen rather than common reactive oxygen species (ROS). When Zr-MOF-P is irradiated, the hydroxyl group of phosphoric acid is converted into oxygen radical, which takes an electron from the sulfides, and then the activated substrates react with the triplet oxygen to form sulfoxides, avoiding the destruction of the catalysts and endowing the reaction with high substrate compatibility and fine recyclability.


Introduction
Porous solid Brønsted acids such as phosphoric acids are important heterogeneous catalysts for diverse chemical reactions [1][2][3]. An effective way to construct these catalysts is to introduce them into highly porous and stable metal−organic frameworks (MOFs) [4][5][6]. Phosphoric acid-based MOFs have received extensive attention in recent years due to their high reactivity and selectivity towards many types of reactions, especially asymmetric transformations [7][8][9][10][11]. Diverse phosphoric acid ligands are synthesized to construct different MOFs; among these, the coordination of carboxyl groups with Zr(IV) show the best stability [10,11]. However, in view of the large conjugate structure associated with phosphoric acid-based MOFs, their photocatalytic performance is rarely explored.
MOFs with tailorable structures and high porosities have been demonstrated to be efficient photocatalysts towards various types of reactions, such as H 2 O splitting [12][13][14][15], CO 2 reduction [16][17][18][19], organic pollutant degradation [20][21][22][23], and organic transformations [24][25][26][27][28]. Ligand structure in MOFs is essential to the regulation of their photocatalytic performance, including the photoresponsivity and active site. Compared to other ligands, phosphoric acids have versatile sites towards different substrates [29,30], and the moderate acidity of phosphoric acid makes it an ideal candidate for the construction of materials to undergo excited-state proton transfer (ESIPT) under photo-excitation, which is a common reaction in molecules with acidic hydrogen atoms and conjugate systems [31][32][33]. ESIPT products are reactive and the reaction is in general wholly reversible, so it would be an effective strategy to construct novel photocatalytic MOFs through introducing phosphoric acids into photo responsive ligands. Thereinto, binaphthol (BINOL)-based phosphoric acids with a large conjugated structure, ESIPT propriety, fine visible-light-response, and good modifiability could be functioned as effective photocatalysts [34][35][36][37][38][39]. In this context, BINOL 2 of 15 is chosen as a potential skeleton to fabricate photo responsive phosphoric acid ligands for MOF construction.
Photocatalyzed selective oxidation of sulfides using oxygen as the oxidizer is an environment-friendly way to produce sulfoxides, which are key intermediates for bioactive ingredients in the pharmaceutical industry [40][41][42]. Generally, oxygen is activated by the photocatalyst through energy transfer or electron transfer to generate reactive oxygen species (ROS) such as 1 O 2 and O 2 −• , followed by the reaction with sulfide to produce sulfoxide [43][44][45][46][47][48]. Due to the high activity of ROS, the generated sulfoxide and some substituent groups with low stability would also be oxidized, causing low selectivity and poor substrate compatibility. Moreover, the photocatalysts might also be affected by hyperactive ROS and the lifetime of catalysts is shortened, which are unfavorable for the application of photocatalysts, especially the recycling of heterogeneous photocatalysts. Avoiding the production of ROS, making the photocatalyst directly activate the sulfide and react with triplet oxygen, is an effective way to improve the selectivity and substrate compatibility [49]. Considering that the phosphoric acid-based materials may follow the ESIPT route, we believe phosphoric acids containing Zr-based MOFs would be a good candidate material for the selective photocatalytic oxygenation of sulfides to sulfoxides and may have better selectivity and substrate compatibility. In this view, a BINOL-based phosphoric acid ligand and corresponding Zr-MOFs were fabricated and employed in this reaction; their catalytic performance and reaction mechanism were also investigated.

Synthesis and Characterization
The ligand 3,3 ,6,6 -tetrakis(4-benzoic acid)-1,1 -binaphthyl phosphate (L 1 H 4 ) was synthesized with an optimized route according to the literature [7]; the new route was shortened to seven steps and the total yield was increased to 31%. To study the effect of the phosphate hydroxyl group, a phosphate-hydroxyl-protected ligand 3,3 ,6,6 -tetrakis(4methyl benzoate)-1,1 -binaphthyl methyl phosphate (L 2 Me 4 ) was also synthesized by an additional methylation reaction for the precursor of L 1 H 4 (L 1 Me 4 ). Zr-MOF-P was prepared through a solvothermal reaction with L 1 H 4 , ZrCl 4 , formic acid, and trifluoroacetic acid (TFA) in N,N-dimethylformamide (DMF) at 120 • C as light yellow octahedral crystals (Scheme 1). In this context, BINOL is chosen as a potential skeleton to fabricate photo responsive phosphoric acid ligands for MOF construction. Photocatalyzed selective oxidation of sulfides using oxygen as the oxidizer is an environment-friendly way to produce sulfoxides, which are key intermediates for bioactive ingredients in the pharmaceutical industry [40][41][42]. Generally, oxygen is activated by the photocatalyst through energy transfer or electron transfer to generate reactive oxygen species (ROS) such as 1 O2 and O2 −• , followed by the reaction with sulfide to produce sulfoxide [43][44][45][46][47][48]. Due to the high activity of ROS, the generated sulfoxide and some substituent groups with low stability would also be oxidized, causing low selectivity and poor substrate compatibility. Moreover, the photocatalysts might also be affected by hyperactive ROS and the lifetime of catalysts is shortened, which are unfavorable for the application of photocatalysts, especially the recycling of heterogeneous photocatalysts. Avoiding the production of ROS, making the photocatalyst directly activate the sulfide and react with triplet oxygen, is an effective way to improve the selectivity and substrate compatibility [49]. Considering that the phosphoric acid-based materials may follow the ESIPT route, we believe phosphoric acids containing Zr-based MOFs would be a good candidate material for the selective photocatalytic oxygenation of sulfides to sulfoxides and may have better selectivity and substrate compatibility. In this view, a BINOL-based phosphoric acid ligand and corresponding Zr-MOFs were fabricated and employed in this reaction; their catalytic performance and reaction mechanism were also investigated.

Scheme 1. Synthesis of ligand and Zr-MOF-P.
Scanning electron microscope (SEM) images ( Figure 1) and powder X-ray diffraction (PXRD) patterns (Figure 2a) confirmed that Zr-MOF-P was successfully synthesized. Single crystal X-ray diffraction showed that Zr-MOF-P crystallizes in the monoclinic space Scanning electron microscope (SEM) images ( Figure 1) and powder X-ray diffraction (PXRD) patterns (Figure 2a) confirmed that Zr-MOF-P was successfully synthesized. Single crystal X-ray diffraction showed that Zr-MOF-P crystallizes in the monoclinic space group (a = 21.112 Å, b = 38.991 Å, c = 19.209 Å, and β = 120.902 • ) and is similar to most of the reported Zr-MOFs composed of carboxylate-based tetrahedral linkers [50][51][52][53]; it also exhibited the flu topology ( Figure 3, Table S1, and Figure S1). The phosphoric acid ligand presents a distorted tetrahedron structure, with a 55.70 • dihedral angle between the two naphthalene groups, constructing a cavity with a diameter of 12.2 Å. The total solvent accessible volume of Zr-MOF-P is estimated to be 73.6%, calculated by the PLATON routine [54]. The simulated PXRD pattern is similar to the experimental data, demonstrating the phase purity of Zr-MOF-P. Due to the large ligand and high porosity of Zr-MOF-P, its crystallinity was destroyed after removing the solvent molecules in the channels by vacuumdrying. Therefore, we tried the supercritical CO 2 -drying method to keep its structural integrity, and the results showed that the crystal structure of the dried MOF was still partly destroyed ( Figure 2a). Considering that the removal of the solvent would break the structure of Zr-MOF-P, it was directly used in the photocatalytic reactions after being washed with DMF several times through suction filtration. Unsurprisingly, Zr-MOF-P showed high thermal stability, thermogravimetric analysis (TGA) indicated that the guest molecules are removed before the temperature reaches 160 • C, and the frameworks remained stable below 300 • C ( Figure 2b). group (a = 21.112 Å, b = 38.991 Å, c = 19.209 Å, and β = 120.902°) and is similar to most of the reported Zr-MOFs composed of carboxylate-based tetrahedral linkers [50][51][52][53]; it also exhibited the flu topology ( Figure 3, Table S1, and Figure S1). The phosphoric acid ligand presents a distorted tetrahedron structure, with a 55.70° dihedral angle between the two naphthalene groups, constructing a cavity with a diameter of 12.2 Å. The total solvent accessible volume of Zr-MOF-P is estimated to be 73.6%, calculated by the PLATON routine [54]. The simulated PXRD pattern is similar to the experimental data, demonstrating the phase purity of Zr-MOF-P. Due to the large ligand and high porosity of Zr-MOF-P, its crystallinity was destroyed after removing the solvent molecules in the channels by vacuum-drying. Therefore, we tried the supercritical CO2-drying method to keep its structural integrity, and the results showed that the crystal structure of the dried MOF was still partly destroyed ( Figure 2a). Considering that the removal of the solvent would break the structure of Zr-MOF-P, it was directly used in the photocatalytic reactions after being washed with DMF several times through suction filtration. Unsurprisingly, Zr-MOF-P showed high thermal stability, thermogravimetric analysis (TGA) indicated that the guest molecules are removed before the temperature reaches 160 °C, and the frameworks remained stable below 300 °C ( Figure 2b).   group (a = 21.112 Å, b = 38.991 Å, c = 19.209 Å, and β = 120.902°) and is similar to most the reported Zr-MOFs composed of carboxylate-based tetrahedral linkers [50][51][52][53]; it a exhibited the flu topology ( Figure 3, Table S1, and Figure S1). The phosphoric acid liga presents a distorted tetrahedron structure, with a 55.70° dihedral angle between the t naphthalene groups, constructing a cavity with a diameter of 12.2 Å. The total solv accessible volume of Zr-MOF-P is estimated to be 73.6%, calculated by the PLATON ro tine [54]. The simulated PXRD pattern is similar to the experimental data, demonstrati the phase purity of Zr-MOF-P. Due to the large ligand and high porosity of Zr-MOF its crystallinity was destroyed after removing the solvent molecules in the channels vacuum-drying. Therefore, we tried the supercritical CO2-drying method to keep its str tural integrity, and the results showed that the crystal structure of the dried MOF was s partly destroyed ( Figure 2a). Considering that the removal of the solvent would break structure of Zr-MOF-P, it was directly used in the photocatalytic reactions after bei washed with DMF several times through suction filtration. Unsurprisingly, Zr-MOF showed high thermal stability, thermogravimetric analysis (TGA) indicated that the gu molecules are removed before the temperature reaches 160 °C, and the frameworks mained stable below 300 °C ( Figure 2b).     To identify the possible photocatalytic application of Zr-MOF-P, its photo-electrochemical properties were tested. As shown in Figure 4a, UV-Vis spectra indicated that this MOF could be excited by visible light for its obvious absorption below 600 nm, and the band gap was estimated to be 2.83 eV. Mott-Schottky measurements were performed at the frequency of 1000, 1500, and 2000 Hz to identify the semiconductor characteristics of Zr-MOF-P, the flat band position determined from the same intersection is about −1.26 V vs. Ag/AgCl (−1.04 V vs. NHE), and the positive slope of the C -2 values indicates the character of n-type semiconductors [55][56][57]. Thus, the conduction band (CB) is −1.04 V vs. NHE, and the valence band (VB) is 1.79 V vs. NHE (Figure 4b). The VB of Zr-MOF-P is higher than the oxidation potentials of sulfides, but lower than that of sulfoxides [49], indicating that it could be used in the photo-oxidation of sulfides to sulfoxides. Photo-electrochemical measurements showed that Zr-MOF-P had an obvious photocurrent response, illustrating that the hole−electron pair could be separated under visible light irradiation ( Figure 4c). The weak fluorescence emission indicated the low electron−hole recombination rate in Zr-MOF-P, which favored the electron transfer between photocatalysts and substrates ( Figure 4d). All the photo-electrochemical measurements clearly demonstrated that Zr-MOF-P would be an ideal photocatalyst for the photo-oxidation of sulfides to sulfoxides. To identify the possible photocatalytic application of Zr-MOF-P, its photo-electrochemical properties were tested. As shown in Figure 4a, UV-Vis spectra indicated that this MOF could be excited by visible light for its obvious absorption below 600 nm, and the band gap was estimated to be 2.83 eV. Mott-Schottky measurements were performed at the frequency of 1000, 1500, and 2000 Hz to identify the semiconductor characteristics of Zr-MOF-P, the flat band position determined from the same intersection is about −1.26 V vs. Ag/AgCl (−1.04 V vs. NHE), and the positive slope of the C -2 values indicates the character of n-type semiconductors [55][56][57]. Thus, the conduction band (CB) is −1.04 V vs. NHE, and the valence band (VB) is 1.79 V vs. NHE (Figure 4b). The VB of Zr-MOF-P is higher than the oxidation potentials of sulfides, but lower than that of sulfoxides [49], indicating that it could be used in the photo-oxidation of sulfides to sulfoxides. Photoelectrochemical measurements showed that Zr-MOF-P had an obvious photocurrent response, illustrating that the hole−electron pair could be separated under visible light irradiation ( Figure 4c). The weak fluorescence emission indicated the low electron−hole recombination rate in Zr-MOF-P, which favored the electron transfer between photocatalysts and substrates ( Figure 4d). All the photo-electrochemical measurements clearly demonstrated that Zr-MOF-P would be an ideal photocatalyst for the photo-oxidation of sulfides to sulfoxides.

Photo-Oxidation of Thioanisol
Considering the excellent photo-electric performance of Zr-MOF-P, we investigated its photocatalytic activity towards the photo-oxidation of sulfide into sulfoxide, and thioanisole was selected as the model substrate. The reaction was initially carried out in ac-

Photo-Oxidation of Thioanisol
Considering the excellent photo-electric performance of Zr-MOF-P, we investigated its photocatalytic activity towards the photo-oxidation of sulfide into sulfoxide, and thioanisole was selected as the model substrate. The reaction was initially carried out in acetonitrile with Zr-MOF-P under white light irradiation and an O 2 atmosphere at room temperature. As shown in Table 1, after 9 h, 19% of sulfide was oxidized into sulfoxide and no overoxidized product (sulfone) was produced. Based on the reported works, the yield is obviously affected by the type of solvents. Therefore, different solvents were explored, and the protic solvent trifluoroethanol (TFEA) was found to be the optimal solvent with a yield of 97% (

Photocatalytic Mechanism
Most of the research reported that ROS such as 1 O 2 , O 2 −• , or ·OH originating from oxygen under the activation of a photocatalyst were important active species in the photocatalytic oxidation of sulfide. Therefore, we carried out a series of experiments to confirm whether ROS participate in the Zr-MOF-P-catalyzed reaction. Quenching experiments through adding different scavengers of ROS was firstly performed. Diazabicyclo[2.2.2]octane (DABCO, TCI, Tokyo, Japan) is a scavenger for 1 O 2 ; its addition showed no effect on the yield of sulfoxide, excluding the participation of 1 O 2 in the reaction (  entries 6 and 7). Moreover, the addition of the sulfide radical cation scavenger 1,4-dimethoxybenzene (DMB, Aladdin, Shanghai, China) also repressed the reaction with a decreased yield of 78% (Table 2, entry 8). Considering that the optimal solvent TFEA is conducive to maintaining the stability of cations [58,59], the sulfide radical cation should be a critical intermediate in the catalytic process. Quenching experiments using L 1 Me 4 as the catalyst also showed similar results (Table 2, entries 9-15), which indicated that the ligand in Zr-MOF-P was the active component. To further exclude the participation of ROS, electron paramagnetic resonance (EPR) tests were performed by adopting 2,2,6,6-tetramethylpiperidine (TEMP) and 5,5-dimethyl-1-pryyoline-Noxide (DMPO) as trappers. As shown in Figure 5a, no EPR signal was detected under white light irradiation, which means that no 1 O 2 , O 2 −• , or ·OH was produced by the photocatalyst. Moreover, the probe molecules Singlet Oxygen Sensor Green, nitrotetrazolium blue chloride, and coumarin-3-carboxylic acid for 1 O 2 , O 2 −• , and ·OH were added to the suspension of Zr-MOF-P under white light irradiation; the results still showed that no ROS was produced (Figure 5b-d). Based on the above mechanism research experiments, the possibility of ROS participating in the reaction was ruled out; sulfide was directly activated by Zr-MOF-P, and then reacted with 3 O 2 .
The active site of Zr-MOF-P was found through another controlled experiment. L 1 Me 4 showed fairly good activity towards the reaction with a yield of 99% (Table 2, entry 9), while the phosphate-hydroxyl-protected L 2 Me 4 almost had no catalytic activity with a yield as low as 2% ( Table 2, entry 16). EPR spectra showed a single and unstructured signal of Zr-MOF-P with a g value of 2.0033 after illumination, and the solid L 1 Me 4 also had the same signal while L 2 Me 4 did not, demonstrating the existence of photo-induced oxygen radicals in Zr-MOF-P and L 1 Me 4 [60]. Moreover, the EPR signal of Zr-MOF-P after illumination was significantly decreased after the addition of thioanisole, which indicated an electron transfer process between thioanisole and the photo-induced oxygen radical ( Figure 6). Therefore, a proposed mechanism was shown in Scheme 2. The ligand was firstly excited by visible light, followed with an ESIPT process, producing the oxygen radical A. The photo-induced oxygen radical takes an electron from sulfide, generating the reduced ligand B and sulfide radical cation C. Then C reacts with 3 O 2 , which converts into the persulfoxide radical D. The reduced ligand B donates an electron to D, which is recovered, and D is transformed into persulfoxide E. Finally, E reacts with another sulfide molecule and two molecules of sulfoxide are produced.
were added to the suspension of Zr-MOF-P under white light irradiation; the showed that no ROS was produced (Figure 5b-d). Based on the above mec search experiments, the possibility of ROS participating in the reaction was ru fide was directly activated by Zr-MOF-P, and then reacted with 3 O2. The active site of Zr-MOF-P was found through another controlled e L1Me4 showed fairly good activity towards the reaction with a yield of 99% (Ta 9), while the phosphate-hydroxyl-protected L2Me4 almost had no catalytic act yield as low as 2% ( Table 2, entry 16). EPR spectra showed a single and unstru nal of Zr-MOF-P with a g value of 2.0033 after illumination, and the solid L1M the same signal while L2Me4 did not, demonstrating the existence of photo-in gen radicals in Zr-MOF-P and L1Me4 [60]. Moreover, the EPR signal of Zr-M illumination was significantly decreased after the addition of thioanisole, whic an electron transfer process between thioanisole and the photo-induced oxy ( Figure 6). Therefore, a proposed mechanism was shown in Scheme 2. The

6121
firstly excited by visible light, followed with an ESIPT process, producing ical A. The photo-induced oxygen radical takes an electron from sulfid reduced ligand B and sulfide radical cation C. Then C reacts with 3 O2, wh the persulfoxide radical D. The reduced ligand B donates an electron t covered, and D is transformed into persulfoxide E. Finally, E reacts wit molecule and two molecules of sulfoxide are produced.

Substrate Compatibility and Recyclability
Encouraged by the unusual photocatalytic mechanism of Zr-MOF-P with different substituents were employed in the reaction (Scheme 3). M fide derivatives with halogen at the ortho-or para-positions of phenyl rin pletely transformed into corresponding sulfoxides, and the conversion tuted sulfide were slower than the para-substituted sulfide due to the ste 7). Other substituents such as nitro (8), methyl (9), and methoxy (10, 11) w in the reaction. Without the participation of ROS, amino (12)-and hyd tuted sulfides could be oxidized to sulfoxides and no side reaction wa photosensitive iodine was also well tolerated, and almost quantitativel sulfoxide was obtained (14). Moreover, even the diphenyl sulfide which be oxidized by most photocatalysts could be successfully transformed in with excellent yields. All of the above results indicated that the avoidan ROS would endow Zr-MOF-P with high substrate compatibility and sel

Substrate Compatibility and Recyclability
Encouraged by the unusual photocatalytic mechanism of Zr-MOF-P, various sulfides with different substituents were employed in the reaction (Scheme 3). Methylphenyl sulfide derivatives with halogen at the orthoor parapositions of phenyl rings were all completely transformed into corresponding sulfoxides, and the conversion of ortho-substituted sulfide were slower than the para-substituted sulfide due to the steric hindrance (2-7). Other substituents such as nitro (8), methyl (9), and methoxy (10,11) were all tolerated in the reaction. Without the participation of ROS, amino (12)-and hydroxy (13)-substituted sulfides could be oxidized to sulfoxides and no side reaction was observed. The photosensitive iodine was also well tolerated, and almost quantitatively corresponding sulfoxide was obtained (14). Moreover, even the diphenyl sulfide which was difficult to be oxidized by most photocatalysts could be successfully transformed into sulfoxide (15) with excellent yields. All of the above results indicated that the avoidance of producing ROS would endow Zr-MOF-P with high substrate compatibility and selectivity. As a heterogenous photocatalyst, recyclability is an advantage compared w geneous catalysts. Zr-MOF-P can be easily separated through centrifugation w action finished, and it can be directly used for subsequent runs without addi cessing. The photocatalytic activity of Zr-MOF-P shows no noticeable chang cycles of the photo-oxidation of thioanisole (Figure 7a). PXRD spectra and SEM the recycled photocatalyst also show little change compared with the pristine As a heterogenous photocatalyst, recyclability is an advantage compared with homogeneous catalysts. Zr-MOF-P can be easily separated through centrifugation when the reaction finished, and it can be directly used for subsequent runs without additional processing. The photocatalytic activity of Zr-MOF-P shows no noticeable change after five cycles of the photo-oxidation of thioanisole (Figure 7a). PXRD spectra and SEM im-ages of the recycled photocatalyst also show little change compared with the pristine Zr-MOF-P (Figure 7b,c). Therefore, a photocatalyst with high stability and recyclability for the photo-oxidation of sulfides has been constructed.
As a heterogenous photocatalyst, recyclability is an advantage compared with ho geneous catalysts. Zr-MOF-P can be easily separated through centrifugation when th action finished, and it can be directly used for subsequent runs without additional cessing. The photocatalytic activity of Zr-MOF-P shows no noticeable change after cycles of the photo-oxidation of thioanisole (Figure 7a). PXRD spectra and SEM imag the recycled photocatalyst also show little change compared with the pristine Zr-MO (Figure 7b,c). Therefore, a photocatalyst with high stability and recyclability for the ph oxidation of sulfides has been constructed.

Instruments
Powder X-ray diffraction (PXRD) was carried out with a PANalytical X'Pert3 P der-17005730 X-ray Powder Diffractometer equipped with two Cu anodes (λ1 = 1.54 Å, λ2 = 1.544426 Å, ratio K-α2/K-α1 = 0.5) at 40 kV and 40 mA. Thermogravimetric ana (TGA) was performed using a TA Discovery SDT 650 heated from room temperatu 800 °C under N2 atmosphere at the heating rate of 10 °C•min −1 . Scanning electron mi copy (SEM) images were obtained using a Hitachi SU-8010 microscope (Tokyo, Jap UV-Vis diffuse reflectance spectra were obtained on a Shimadzu UV-2600i (Kyoto, Ja spectrophotometer equipped with an integrated sphere and a white standard of B

Instruments
Powder X-ray diffraction (PXRD) was carried out with a PANalytical X'Pert3 Powder-17005730 X-ray Powder Diffractometer equipped with two Cu anodes (λ 1 = 1.540598 Å, λ 2 = 1.544426 Å, ratio K -α2 /K -α1 = 0.5) at 40 kV and 40 mA. Thermogravimetric analysis (TGA) was performed using a TA Discovery SDT 650 heated from room temperature to 800 • C under N 2 atmosphere at the heating rate of 10 • C·min −1 . Scanning electron microscopy (SEM) images were obtained using a Hitachi SU-8010 microscope (Tokyo, Japan). UV-Vis diffuse reflectance spectra were obtained on a Shimadzu UV-2600i (Kyoto, Japan) spectrophotometer equipped with an integrated sphere and a white standard of BaSO 4 was used as a reference. UV-Vis spectra were obtained on a Shimadzu UV-2600i spectrophotometer. Fluorescence spectra and quantum yield were obtained on an Edinburgh Instruments FLS1000 fluorescence spectrophotometer (Livingston, UK). Nuclear magnetic resonance (NMR) data were collected on a Bruker Avance III 500 spectrometer (Berlin, Germany). HRMS was recorded on an Agilent G6545 Q-TOF (Santa Clara, CA, USA). Electrochemical characterizations were carried out with a CH Instruments CHI660E workstation (Shanghai, China). The photocatalytic reactions were performed in a PerfectLight PCX50C photoreactor (Beijing, China) with 5 W white light LED. Gas chromatographic (GC) analyses were performed using a Shimadzu 2010 gas chromatograph (Kyoto, Japan) equipped with an HP-5MS capillary column (30 m × 0.25 mm × 0.25 µm) and a flame ionization detector. Gas chromatography-mass spectrometry (GC-MS) was recorded on a Waters GCT Premier mass spectrometer (Milford, MA, USA). Electron paramagnetic resonance (EPR) measurements were carried out on a Bruker model A300 spectrometer (Berlin, Germany).
The synthesis of I is as follows: 6,6 -dibromo-3,3 -diiodo-1,1 -binaphthyl-2,2 -diol (2.00 g, 2.87 mmol), 4-(methoxycarbonyl)benzeneboronic acid (5.17 g, 28.74 mmol), Pd(OAc) 2 (129 mg, 0.57 mmol), Na 2 CO 3 (2.13 g, 20.09 mmol), DMF (32 mL), and H 2 O (32 mL) were added into a 350 mL Schlenck tube under Ar atmosphere. The reaction was stirred for 24 h in a 60 • C oil bath. After the reaction finished, it was cooled to room temperature, the mixture was extracted with CH 2 Cl 2 , and the organic phase was washed with H 2 O three times. Then, the organic phase was dried over anhydrous Na 2 SO 4 , and the solvent was filtered and concentrated. Crude product was purified by column chromatography on silica gel (2/1 petroleum ether/ethyl acetate, R f = 0.55) to afford 1.  Figure S2). 13  The synthesis of Zr-MOF-P is as follows: L 1 H 4 (200 mg, 0.242 mmol), ZrCl 4 (169 mg, 0.0.725 mmol), anhydrous formic acid (10 mL), and trifluoroacetic acid (2 mL) were added in DMF (40 mL). After 10 min of ultrasonic vibration, the mixture was heated in a 100 mL Teflon-sealed autoclave at 120 • C for 3 days. Then, the mixture was cooled to room temperature, light yellow powders (310 mg) were collected through centrifugation, and washed with DMF. Because the removing of solvent molecules from MOF channels will distort the framework, Zr-MOF-P was dipped in DMF and was collected through suction filtration before use.

Electrochemical Characterization
Electrochemical characterizations were carried out using a CH Instruments CHI660E workstation through a three-electrode system in 0.2 M Na 2 SO 4 aqueous solution.
Mott-Schottky plots of Zr-MOF-P were measured using the photocatalyst-coated glassy carbon as working electrode, Ag/AgCl as reference electrode, and Pt plate as counter electrode at frequencies of 1000, 1500, and 2000 Hz, respectively. Preparation of the working electrode is as follows: 5 mg Zr-MOF-P was dispersed in 1 mL ethanol, and 10 µL 5 wt% Nafion was added as binder. Then, 20 µL of the solution was coated on the surface of the glassy carbon electrode and dried at room temperature. This process was repeated until the electrode was completely covered. Photocurrent measurements of Zr-MOF-P were measured using the photocatalystcoated Pt plate as working electrode, Ag/AgCl as reference electrode, and Pt plate as counter electrode, and a 40 W White light LED was used as light source. Preparation of the working electrode is as follows: 5 mg Zr-MOF-P was dispersed in 1 mL ethanol, and 10 µL 5 wt% Nafion was added as binder. Then, 50 µL of the solution was coated on the Pt plate and dried at room temperature. This process was repeated until 1 cm 2 of the Pt plate was completely covered.

Photocatalytic Reaction
The photocatalytic reactions were performed on a PerfectLight PCX50C photoreactor (Beijing, China) equipped with 5 W white LEDs. In addition, the reaction was carried out at 25 • C by circulating refrigeration equipment. For the photo-oxidation of sulfides to sulfoxides, 4 mg photocatalyst, 0.1 mmol substrate, and 2 mL solvent were added into a 10 mL Schlenck tube under O 2 atmosphere. The reaction mixture was magnetically stirred at 150 rpm and illuminated with 5 W white LEDs. After the reaction finished, 20 µL of anisole was added as the internal standard and stirred for 10 min. Then, the photocatalyst was separated through centrifugation and washed with solvent. The products were analyzed by GC and GC-MS.
For gram-scale reaction, thioanisol (8.86 mmol, 1.10 g), TFEA (100 mL), and Zr-MOF-P (20 mg) were stirred at room temperature for 7 days in oxygen atmosphere (1 atm) under the irradiation of white LEDs. After the reaction finished, photocatalyst was separated through centrifugation and washed with ethyl acetate several times. The combined organic phase was concentrated over a rotary evaporator, and 1 (1.18 g, 95%) was obtained through column chromatography as a colorless oil.

EPR Measurements
EPR spectra were obtained on a Bruker model A300 spectrometer (Berlin, Germany) at room temperature. The spectrometer parameters are shown as follows: sweep width, 100 G; center field, 3510.890 G; microwave bridge frequency, 9.839 GHz; power, 20.37 mW; modulation frequency, 100 kHz; modulation amplitude, 1 G; conversion time, 42.00 s; sweep time 42.00 s; receiver gain, 2.00 × 10 4 . The preparation of the liquid samples was similar to the photocatalyst reaction. The signal after irradiation was measured after 5 min of irradiation with a 50 W Xe lamp with stirring, and the mixture was transferred to 3 mm diameter glass tubes as soon as possible to record the signals. Furthermore, for solid samples, about 2 mg of target compound was put into a 3 mm diameter glass tube, and the signal after irradiation was also measured after 5 min of irradiation with a 50 W Xe lamp.

ROS Detection with Probe Molecules
1 O 2 detection: Zr-MOF-P (2 mg) and TFEA (1.5 mL) containing Singlet Oxygen Sensor Green (10 µM) were added into a 10 mL Schlenck tube under air atmosphere. The reaction mixture was magnetically stirred for 30 min in the dark before illuminated with white LEDs for 1 h. After the reaction finished, photocatalyst was separated through centrifugation and the supernatant was examined with fluorescence spectrophotometer. The result was compared with the blank group and unilluminated control group, showing that no 1 O 2 produced. O 2 −• detection: Zr-MOF-P (2 mg) and TFEA (1.5 mL) containing nitrotetrazolium blue chloride (0.1 mM) were added into a 10 mL Schlenck tube under air atmosphere. The reaction mixture was magnetically stirred for 30 min in the dark before illuminated with white LEDs for 2 h. After the reaction finished, photocatalyst was separated through centrifugation and the supernatant was examined with UV-Vis spectrophotometer. The result was compared with the unilluminated control group, showing that no O 2 −• produced. ·OH detection: Zr-MOF-P (2 mg) and TFEA (1.5 mL) containing coumarin-3-carboxylic acid (0.1 mM) were added into a 10 mL Schlenck tube under air atmosphere. The reaction mixture was magnetically stirred for 30 min in the dark before illuminated with white LEDs for 2 h. After the reaction finished, photocatalyst was separated through centrifugation and the supernatant was examined with UV-Vis spectrophotometer. The result was compared with the unilluminated control group, showing that no ·OH produced.

Conclusions
A photocatalyst Zr-MOF-P based on a BINOL-derived phosphoric acid ligand for the selective oxidation of sulfides under white light irradiation was prepared. Comprehensive mechanistic studies indicated that Zr-MOF-P had appropriate photo-electrochemical properties for this reaction, and the ESIPT process produced the reactive oxygen radical, which would take an electron from the sulfides. Thus, the sulfides were activated and, subsequently, react with ground state oxygen, producing sulfoxides. The unique mechanism without the participation of ROS ensured the high selectivity and substrate compatibility of the reaction. Moreover, as a heterogeneous photocatalyst, Zr-MOF-P had sufficient stability, as it can be easily separated and re-used at least five times without any noticeable change in reactivity. This study demonstrates that phosphoric acids with a large conjugate structure can be used as photocatalysts, and they might have potential applications in more kinds of photocatalytic reactions. Further applications for Zr-MOF-P are under study in our group.