THE ⁴Σ- STATES OF THE FO RADICAL

The development of solar energy conversion and storage is essential for alleviating the looming energy and environmental crisis triggered by the global population explosion and unprecedented economic growth [1-4]. However, the intermittent and diffuse nature of solar energy leave ample space to further developments. Increasingly, semiconductor photocatalysis can expand a new horizon for harnessing solar energy [5, 6]. Whereas visible light utilization rate in the solar spectrum still imposes restrictions on the available scope of transition metal oxides on Earth [7, 8]. For instance, TiO₂ can only absorb approximately 5% of natural sunlight owing to its wide bandgaps (3.0 eV for rutile, 3.2 eV for anatase) [9, 10]. Nevertheless, TiO₂ turns out to be a stepping stone to new photocatalysts that close the energy gap [11, 12].

Particularly, surface-modified TiO₂ with carboxyl, hydroxyl containing ligands such as ethylenediaminetetraacetic acid [13, 14], salicylic acid [15, 16], catechol [17, 18], glucose [19, 20], and cyclodextrin [21, 22] have aroused great interest on account of their high absorption of visible light and low recombination of electron-hole to achieve photocatalytic reactions efficiently. Besides, the properties of the TiO₂ complexes, such as photochemical stability, optical response, and binding mode, are closely associated with the structure of ligands [23-25]. Generally, aromatic organic acids containing other functional groups like hydroxyl can form more stable surface complexes than aromatic monocarboxylic acids [26]. Nevertheless, the π-conjugation of aromatic monocarboxylic acid ligands also strongly affects the complex's stability, surface binding, and charge transfer [27, 28]. Therefore, one can embrace the concept of molecular design to select redox-active ligand based on a previously disclosed strategy [29].

Hereby, with benzoic acid (BA) as the initial molecule, horizontally extending one or two benzene rings furnish 2-naphthoic acid (2-NA) and 2-anthracene acid (2-AA) which could be anchored onto TiO₂ for the photocatalytic selective oxidation of sulfides. It is worth noting that the selective oxidation of organic sulfides has recently attracted much interest [30, 31] because the resultant sulfoxides are vital intermediates for agrochemicals, pharmaceuticals, and valuable fine chemicals [32-36]. Meanwhile, TEA, an electron transfer mediator [37], maintains the integrity of these surface ligands on TiO₂. Additionally, TEA also stands out for its low boiling point in consideration of the subsequent separation and purification of sulfoxides [38]. Hence, cooperative photocatalysis of 2-AA-TiO₂ with TEA in solution is established via collisional electron transfer.

In general, the aromatic monocarboxylic acid-modified TiO₂ complex is implemented by the acidolysis and exchange of the surface hydroxyl groups of TiO₂ with carboxylate anions, forming the RCOO-Ti bond [26]. However, the π-conjugated system in aromatic monocarboxylic acids can greatly affect the electrostatic attraction of carboxylate anions to the titanol group, thus affecting the stability and photocatalytic activity of the ligand-modified TiO₂. With BA as the initial molecule, horizontally extending the π-conjugation affords two monocarboxylic acids ligands, namely 2-NA and 2-AA (Fig. 1a). In this work, Aeroxide P25 TiO₂ was adopted as the starting metal oxide to anchor BA, 2-NA, and 2-AA.

Figure 1

Figure 1.  (a) Horizontally extending the benzene ring of BA obtains 2-NA and 2-AA. (b) UV–vis DRS of TiO₂, BA-TiO₂, 2-NA-TiO₂, 2-AA-TiO₂. (c) Violet light-induced selective aerobic oxidation of phenyl methyl sulfide over BA-TiO₂, 2-NA-TiO₂, 2-AA-TiO₂. Reaction conditions: CH₃OH (1 mL), ligand (1.2 × 10⁻³ mmol), TiO₂ (40 mg), TEA (0.03 mmol), phenyl methyl sulfide (0.35 mmol), aerial O₂, violet LEDs (3 W × 4). (d) FTIR spectra of TiO₂, 2-AA, and 2-AA-TiO₂.

DownLoad: Full-Size Img PowerPoint

Based on the ultraviolet (UV)–visible absorption spectra (Fig. S1a in Supporting information), it can be found that compared with BA, there is no absorption response to visible light completely in 2-NA, while 2-AA can harvest a band of visible light at approximately 400–420 nm. When they were assigned as redox-active ligands for the surface of TiO₂, UV–visible diffuse reflectance spectra (UV–vis DRS) (Fig. 1b) show that the difference between BA-TiO₂, 2-NA-TiO₂, and pristine TiO₂ is relatively small. Conversely, the difference between 2-AA-TiO₂ and pristine TiO₂ is rather obvious. Compared with TiO_2, which can only absorb UV light, the response range of 2-AA-TiO₂ has an obvious redshift. In addition, combining with Fig. S1b (Supporting information), Fig. 1b reveals that the light-emitting spectrum of the violet LED overlaps with the absorbance of 2-AA-TiO₂. Despite the loading of these ligands, there was no change in the powder X-ray diffraction (PXRD) peaks of rutile and anatase TiO₂ (Fig. S1c in Supporting information).

Subsequently, these complexes were applied to the violet light-induced selective oxidation of phenyl methyl sulfide (Fig. 1c). It is intriguing to find a correlation between the conversions of phenyl methyl sulfide and the structure of these monocarboxylic acid ligands. With only one benzene ring, the photocatalytic activity of BA-TiO₂ was almost zero under the standard conditions. With the augment of π-conjugation, the violet light-induced conversion of phenyl methyl sulfide increased from 26% over 2-NA-TiO₂ to 86% over 2-AA-TiO₂ for 1 h, which was attributed to the more stabilized binding of carboxyl from 2-AA with the surface hydroxyl group of TiO₂. Compared with anatase TiO₂ (ST-01) or rutile TiO₂ with much higher specific surface areas, Aeroxide P25 TiO₂ with 2-AA was the best complex photocatalyst (Table S1 in Supporting information).

Next, the chemical states of Ti, O, and C in 2-AA-TiO₂ were measured by X-ray photoelectron spectroscopy (XPS) (Fig. S2 in Supporting information). The Ti 2p_1/2 (464.4 eV) and Ti 2p_3/2 (458.5 eV) peaks conform to the results of Ti 2p (Fig. S2a), which are consistent with a previous investigation [37]. Meanwhile, in Fig. S2b, the XPS data of O 1s can be split into O−Ti at about 529.6 eV, O=C at about 530.9 eV, O−C at about 531.8 eV, which might be due to the presence of excessive 2-AA on the surface of the 2-AA-TiO₂. According to the C 1s XPS (Fig. S2c), the carboxylate group from 2-AA successfully combined with the titanol group of TiO₂ to generate the 2-AA-TiO₂ complex.

On the other hand, the formation of 2-AA-TiO₂ complex was explored by Fourier-transform infrared (FTIR) spectroscopy. In Fig. 1d, the two broad peaks at 1423 cm⁻¹ and 1403 cm⁻¹ were attributed to the stretching vibrations of –COOTi– group. The peak (1423 cm⁻¹) in the FTIR spectrum of 2-AA corresponded to the symmetric stretching of the hydroxy group in a carboxyl group. Therefore, it is reasonable to accept that 2-AA is successfully adsorbed on the surface of TiO₂. Nevertheless, a more profound analysis is required to further determine whether the binding mode of 2-AA-TiO₂ is through bidentate or monodentate via the carboxylate ligand. The FTIR spectra of pristine TiO₂ and 2-AA were also presented in Fig. 1d. The primary signal peak of the carboxylic group in 2-AA are analyzed as following: 1683 cm⁻¹ is the stretching vibration peak of C=O; the signal peaks at 1582 cm⁻¹, 1481 cm⁻¹, and 1461 cm⁻¹ are assigned to C=C skeleton stretching mode; the signal peak at 1423 cm⁻¹ is the vibration peak of O−H in carboxylate group; the strong peak at 1290 cm⁻¹ is the stretching vibration peak of C−O [39, 40]. If the 2-AA-TiO₂ possesses the monodentate mode, the FTIR spectrum of 2-AA-TiO₂ should show the vibration peak of C=O. However, the actuality that the peak of C=O was vacant in the curve, revealed that 2-AA-TiO₂ was more tend to the bidentate binding mode. Meanwhile, the peak of −C−O−Ti− presented in the FTIR spectrum of 2-AA-TiO₂ and the slightly shifted peak position further manifested that the two oxygen atoms in the −COO− connected to the adjacent surface Ti sites respectively and formed the bridging bidentate mode. Furthermore, the FTIR spectra of BA-TiO₂ and 2-NA-TiO₂ are also shown in Fig. S3a (Supporting information).

After confirming 2-AA-TiO₂ complex, various control experiments were conducted for the selective aerobic oxidation of sulfide into sulfoxide. It can be found that photocatalyst synthesized in situ is more effective than the pre-prepared 2-AA-TiO₂ for the selective aerobic oxidation of phenyl methyl sulfide to sulfoxide (Table S2 in Supporting information). Meanwhile, kinetic studies have shown that there is no induction time when aromatic acids modified TiO₂ was in situ formed (Fig. S3b in Supporting information). Hereafter, the exploration of the effect of different wavelengths of LEDs is stated in Fig. S4a (Supporting information), proposing that the corresponding conversion is commensurate to the UV–vis DRS of 2-AA-TiO₂. In Fig. S4b (Supporting information), it was found that very few sulfide were transformed into the corresponding sulfoxide with 2-AA. When 2-AA was loaded onto the surface of TiO₂ to form a complex photocatalyst, the conversion of phenyl methyl sulfide was significantly improved (Fig. S4b). Synchronously, the recycling tests were implemented to testify the durability of the 2-AA-TiO₂ photocatalyst (Fig. S4c in Supporting information). The recovered 2-AA-TiO₂ was also reused twice with the temperate refurbishing of 2-AA, implying its superior endurance.

Importantly, all the components of the photocatalytic system were indispensable based on the thorough control experiments. Further, the effect of the amounts of aromatic acids was explored (Fig. S5a in Supporting information). The conversion of phenyl methyl sulfide sharply increased once 2-AA was added. However, with the further increase of the dosage of 2-AA, the conversion of phenyl methyl sulfide increased slightly. With the amount of TiO₂ enhanced, the yield of phenyl methyl sulfoxide advanced gradually and the selectivity of the product declined slightly (Fig. S5b in Supporting information). TEA, the electron transfer mediator, exhibited a considerable effect on the violet light-induced selective aerobic oxidation of sulfide over the 2-AA-TiO₂ photocatalyst (Fig. S5c in Supporting information). Furthermore, when TEA was replaced by trimethylamine, the conversion of phenyl methyl sulfoxide decreased slightly (Table S3 in Supporting information). CH₃OH, as a redox-active solvent, plays an important role in the formation of phenyl methyl sulfoxide (Table S4 in Supporting information). With the increasing power of violet LEDs, the conversion of phenyl methyl sulfide presented an increasing trend (Fig. S5d in Supporting information).

Next, different scavengers were introduced into the system to determine the reactive oxygen species (ROS) (Fig. 2a). First, a singlet oxygen (¹O₂) scavenger, 1, 4-diazabicyclo[2.2.2]octane (DABCO) was added to the reaction system. The slightly reduced conversion indicated that ¹O₂ did not produce a decisive impact on the reaction. In contrast, when p-benzoquinone (p-BQ) was introduced to capture O₂^•–, the selective oxidation of sulfide was almost completely restrained, suggesting that O₂^•– was the pivotal ROS. Nevertheless, no reaction occurred in an atmosphere of N₂, suggesting O₂ is the terminal oxidant. Besides, AgNO₃ completely restrained the reaction by capturing e_CB⁻. Furthermore, according to Fig. S6a (Supporting information), the flat band potential that is closely related to the conduction band (CB) potential could be linearly fitted to be −0.85 V vs. Ag/AgCl from the intercept of the x-axis, which is negative than the potential of O₂/O₂^•– (−0.48 V vs. Ag/AgCl). According to the value of the bandgap and the CB potential, the valence band (VB) potential was calculated as +1.9 V vs. Ag/AgCl (Fig. S6b in Supporting information).

Figure 2

Figure 2.  (a) Quenching experiments to identify the ROS for the violet light-induced selective aerobic oxidation of phenyl methyl sulfide. Standard reaction conditions: CH₃OH (1 mL), 2-AA (1.2 × 10⁻³ mmol), TiO₂ (40 mg), TEA (0.03 mmol), phenyl methyl sulfide (0.35 mmol), aerial O₂, violet LEDs (3 W × 4), 40 min. The EPR signals recoded during the violet light-induced selective aerobic oxidation of phenyl methyl sulfide over the 2-AA-TiO₂ photocatalyst, (b) e_CB⁻ of 2-AA-TiO₂ and (c) O₂^•– captured by DMPO.

DownLoad: Full-Size Img PowerPoint

To better comprehend the reaction process, electron paramagnetic resonance (EPR) experiments were then carried out. The e_CB⁻ signal of 2-AA-TiO₂ gradually increased with the prolongation of exposure time in the absence of O₂ (Fig. 2b). Once O₂ was poured, the e_CB⁻ signal reduced to the initial state, indicating the relevance between O₂ and e_CB⁻. The EPR signal of O₂^•– was then tested with 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) as the capturing agent (Fig. 2c). The EPR signal of O₂^•– dramatically increased under violet light irradiation from 0 min to 2 min and continued to slightly increase at 4 min, suggesting that O₂^•– captured by DMPO was accumulated and maintained after switching off the light.

Tentatively, a mechanism of violet light-induced selective aerobic oxidation of sulfide over 2-AA-TiO₂ photocatalyst with TEA is proposed in Scheme 1. Firstly, the carboxyl group of 2-AA is in combination with the surface hydroxyl group of TiO₂, constituting the 2-AA-TiO₂ photocatalyst, which enables the reaction to take place at the violet light range. Next, the charge separation is generated over the surface complex by the stimulation of violet light. 2-AA would straightly infuse electrons into the CB of TiO₂ to form 2-AA^•+, without running into an excited state of 2-AA*. Meanwhile, O₂^•– is generated by the reduction of O₂ at the CB of TiO₂. Thirdly, the transition between TEA and TEA^•+ is driven by collisional electron transfer with 2-AA^•+, and the resulted TEA^•+ interacts with phenyl methyl sulfide to deliver a sulfur-centered radical cation, which is pivotal to connect the TEA redox cycle and photocatalytic cycle. This connection can effectively prevent the 2-AA-TiO₂ photocatalyst from being undermined by ROS. Fourthly, the sulfur-centered radical cation combines with O₂^•– to generate phenyl methyl persulfoxide. Ultimately, protons from CH₃OH terminate phenyl methyl persulfoxide to afford phenyl methyl sulfoxide.

Scheme 1

Scheme 1.  A proposed mechanism of violet light-induced selective aerobic oxidation of phenyl methyl sulfide by cooperative photocatalysis of 2-AA-TiO₂ with TEA.

DownLoad: Full-Size Img PowerPoint

Then, the range of organic sulfides subjected to the photocatalytic protocol was investigated (Table 1). Phenyl methyl sulfide was oxidized to corresponding phenyl methyl sulfoxide in 88% selectivity and 85% conversion for 50 min (Table 1, entry 1). With a comparable reaction time, the sulfides containing electron-donating groups generally afforded higher conversions than that of phenyl methyl sulfide (Table 1, entries 2–4). On the contrary, conversions of sulfides containing electron-withdrawing groups were lower than that of phenyl methyl sulfide (Table 1, entries 7–12). The sulfides with an m‑methoxy group or an o‑methoxy group holding the fairly strong steric hindrance effect desired an extended reaction time (Table 1, entries 5 and 6). A similar trend was observed in sulfides containing electron-withdrawing groups (Table 1, entries 11 and 12). Therefore, affected by the steric effect, the required reaction time of the regioisomer augmented in the order of para < meta < ortho isomer. In addition, when the methyl group of phenyl methyl sulfide was replaced by an ethyl group or a phenyl group, longer reaction time was acquired (Table 1, entries 13 and 14). The selective oxidation of 2-naphthyl methyl sulfide took a longer time to reach a similar conversion (Table 1, entry 15). Moreover, the selective oxidation of aliphatic sulfides required less reaction time to achieve the considerable conversion (Table 1, entry 16) or achieved a higher conversion and selectivity with the standard reaction time (Table 1, entry 17).

Table 1

Table 1.  Violet light-induced selective aerobic oxidation of sulfides into sulfoxides with air by cooperative photocatalysis of 2-AA-TiO₂ with TEA.^a

DownLoad: CSV

In summary, extending the π-conjugation of aromatic acids has been adopted to construct TiO₂ complex visible light photocatalysts. Intriguingly, with BA as the initial molecule, horizontally extending one or two benzene rings acquires 2-NA and 2-AA. Among three aromatic acids, 2-AA has delivered the best result over TiO₂ due to the most extended π-conjugation. Importantly, TEA maintains the integrity of anchored aromatic acids and accelerates electron transfer. Conveniently, violet light-induced selective aerobic oxidation of sulfides to sulfoxides has been realized by cooperative photocatalysis of 2-AA-TiO₂ with 10 mol% of TEA. This work extends the library of next-generation ligands for semiconductors to expand visible light-induced selective organic reactions.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment

This work was funded by the National Natural Science Foundation of China (Nos. 22072108 and 21773173).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2021.10.068.