Tuning the selectivity of benzylamine photo-oxidation with different rhodium modes anchored on BiOIO₃

Abstract

Selective benzylamine (BA) oxidation to high value-added imines is of great significance in the chemical industry, but it still remains challenging. In this study, we demonstrate rhodium single atom and nanoparticle modes (Rh₁ and Rh_NP) anchored on BiOIO₃ nanosheets as two efficient photocatalysts (Rh₁/BOIO and Rh_NP/BOIO) to achieve the selective BA oxidation reaction. Notably, these two photocatalysts displayed completely opposite selectivity during BA photo-oxidation. Benzonitrile (BN) was the main product over Rh₁/BOIO, while Rh_NP/BOIO preferentially produced N-benzylidenebenzylamine (N-BBA) under the same reaction conditions. To reveal the reason for such differences, we combined density functional theory (DFT) simulation and experiments to investigate the BA oxidation reaction mechanism over these two photocatalysts. As a consequence, it was found that the different active intermediates ˙OH and ˙O₂⁻ generated on Rh₁ and Rh_NP would lead to the opposite selectivity of BA photo-oxidation over Rh₁/BOIO and Rh_NP/BOIO, respectively. This study offers a valuable strategy to tune the selectivity of photo-oxidation reactions as required.

---

Gang Wang

Gang Wang is currently as Associate Professor at College of Chemistry and Materials Science, Anhui Normal University. He received his PhD in 2015 from State Key Laboratory of Crystal Materials at Shandong University. During 2017–2019, he joined Kyoto University as a postdoc for studying Metal–organic frameworks (MOFs) compounds. His research interests focus on the development of single atom catalysts for solar energy conversion, including CO2 conversion, photosynthesis and overall water splitting.

---

1 Introduction

The selective oxidation of amines to imine, amide, and nitrile products is practically significant for feedstock manufacturing in the field of organic synthesis.^1–5 However, tuning the selectivity of oxidation products still remained a long-term goal and challenge due to the complicated reaction system and harsh reaction conditions (e.g., high temperature, high pressure, and the use of toxic cyanides) in current amine oxidation reaction systems.^6–9 Therefore, it is urgently desirable to develop economical and environmentally friendly catalytic systems toward realizing highly selective aniline oxidation reactions. In recent years, photocatalytic technology has been regarded as one of the promising approaches in organic synthesis owing to its green, cost-effective, and gentle reaction conditions.^10–12 Nevertheless, the application of photocatalysts in organic synthesis is still hindered by their inefficient catalytic capacity due to the low separation efficiency of photogenerated charge carriers.

Various efforts have been devoted to promoting the charge-separation efficiency of photocatalysts, including metal species loading, which is one of the most effective strategies.^13–16 Currently, most of the researches are focused on regulating the patterns of the loaded metal nanoparticles (NPs), such as the particle size, morphology, and their components in photocatalysts. For example, different sizes of Pt nanoparticles were successfully decorated on a TiO₂–SiO₂ support, leading to the size-dependent activity in photocatalytic CO₂ conversion. The charge-transfer efficiency can be promoted by decreasing the size of the Pt NPs.¹⁷ Liu and coworkers reported that the existence and absence of Pt NPs on a CdS/Fe₂O₃ Z-scheme heterojunction exhibited completely different selectivities for C=N and C–C coupling reactions, respectively. Pt NPs acted as cocatalysts for accelerating the charge separation to facilitate the formation of C=N over the CdS/Fe₂O₃ Z-scheme heterojunction.¹⁸ Recently, single-atom photocatalysts (SAPCs), involving the advantages of nearly 100% atom utilization, abundant active sites, and a well-defined coordination structure, have emerged as another encouraging method for promoting the charge-separation efficiency.^19–27 For example, a photocatalyst with Ag nanoparticles and in situ-generated Ag single atom (SAAg) on AgF was fabricated to realize selective hydrodehalogenation and dehalogenation–arylation reactions under visible-light irradiation.²⁸ The well-defined coordination structure of SAAg served as the catalytic active center to anchor activated halides for achieving the photocatalytic organic synthesis. Wen and coworkers reported that single-atom Ni on a TiO₂ support not only promoted the charge-carrier separation efficiency, but was also favorable to realize the selective sulfonation of enamides into amidosulfones.²⁹ However, there are few reports on the effects on the product selectivity in photocatalytic organic syntheses via anchoring different modes of metal.

In this study, we fabricated two photocatalysts with Rh single atom and nanoparticle modes, respectively, anchored on BiOIO₃ nanosheets (Rh₁/BOIO and Rh_NP/BOIO), both of which exhibited highly efficient BA photo-oxidation performances. The difference was that BN was the main product over Rh₁/BOIO, while the N-BBA product dominated for Rh_NP/BOIO under the same reaction conditions. More importantly, the similar phenomena of opposite selectivity also occurred over Rh₁/BOIO and Rh_NP/BOIO with other BA-based compounds as substrates. By combining theoretical calculations and experiments, we could reveal that the Rh–O₂ mode in Rh₁/BOIO favored combining with ˙OH active intermediates, which were derived from the decomposition of tert-butyl hydroperoxide (t-BuOOH), and then oxidized the PhCHNH intermediate to BN. In the case of Rh_NP/BOIO, O₂ molecules could be generated due to the easy disproportionation of ˙OH radicals on the Rh_NP mode. After capturing the photogenerated electrons, the O₂ molecules became ˙O₂⁻ active species and thus could oxidize PhCHNH to PhCHO. Another BA molecule then combined with PhCHO to form the final product of N-BBA.

2 Experimental

2.1. Chemicals

Bismuth nitrate pentahydrate (Bi(NO₃)₃·5H₂O), potassium iodate (KIO₃), rhodium(III) chloride (RhCl₃·xH₂O), ethylene glycol (EG), polyvinylpyrrolidone (PVP, K-30), sodium borohydride (NaBH₄), methanol, ethanol, acetone, tert-butyl hydroperoxide (t-BuOOH, 70 wt% in H₂O), benzylamine (BA), 4-methylbenzylamine, 4-methoxybenzylamine, 4-fluorobenzylamine and 4-chlorobenzylamine were all obtained from Sinopharm. All the chemicals were analytical grade and used without further treatment for purification. Ultrapure water (18.2 MΩ cm) was purified with a Milli-Q instrument.

2.2. Preparation of photocatalysts

Preparation of BiOIO₃ (BOIO): BiOIO₃ was prepared via a hydrothermal method. First, 8 mmol Bi(NO₃)₃·5H₂O and 8 mmol KIO₃ were dissolved in 80 mL H₂O. After stirring for 30 min, the mixed suspension was transferred into a 100 mL sealed Teflon-lined autoclave and heated at 150 °C for 5 h. After cooling to room temperature, the obtained sample was washed with water and ethanol several times and dried at 60 °C for 12 h.

Preparation of Rh single-atom-loaded BiOIO₃ (Rh₁/BOIO): first, 0.3 g of the above BOIO was dispersed in 40 mL H₂O and 10 mL ethanol mixed solution. Then, 122 μL RhCl₃·xH₂O solution (0.1 g mL⁻¹) was added into the suspension. After 0.5 h constant stirring, the mixed suspension was irradiated under full light (PLS-SEX300D, 300 W Xenon lamp, Beijing Perfectlight Technology Co., Ltd.) for about 0.5 h. Rh₁/BOIO was obtained after washing with water and ethanol several times, and dried at 60 °C for 12 h. Different amounts of Rh₁ were also loaded on BOIO surface by controlling the added RhCl₃·xH₂O. In detail, 61 and 183 μL RhCl₃·xH₂O solution (0.1 g mL⁻¹) were respectively added into 0.3 g BOIO suspensions to obtain two contrastive samples after light irradiation. The mass fractions of Rh species in these two samples were about 0.34 and 1.04 wt% according to inductively coupled plasma atomic emission spectroscopy (ICP-AES), and denoted as Rh₁-0.34 wt% and Rh₁-1.04 wt%.

Preparation of Rh nanoparticles-loaded BiOIO₃ (Rh_NP/BOIO): the Rh nanoparticles were synthesized via a chemical reduction method. In detail, 20 mL EG, 418 μL RhCl₃·xH₂O solution (0.1 g mL⁻¹), and 100 mg PVP were mixed together by constant stirring at 85 °C. After forming a dark mixed solution, the reaction temperature was increased to 115 °C and maintained for 2 h via an oil bath method. Then, NaBH₄ was added after the solution cooled to room temperature. The Rh_NP was obtained after stirring the reaction for 12 h. Before use, Rh_NP was washed with ethanol and acetone several times. For Rh_NP/BOIO, 0.3 g BOIO and 2.1 mg of the as-prepared Rh_NP were added into a mixed solution of 20 mL methanol and 20 mL water. After stirring the reaction for 12 h, Rh_NP/BOIO was obtained after centrifugation and treatment at 100 °C for 12 h in a vacuum. Different amounts of Rh_NP were also loaded on the BOIO surface by controlling the added as-prepared Rh_NP. On the basis of the ICP-AES results, the samples with different Rh_NP amounts were obtained, and were denoted as Rh_NP-0.35 wt% and Rh_NP-1.04 wt%.

2.3. Characterizations

The crystal phases of the samples were analyzed by X-ray diffraction (XRD, D/MAX-2550). The morphologies of the samples were determined by transmission electron microscopy (TEM, JEOL 200CX) and AC HAADF-STEM (JEOL JEM-ARM 200F) with an energy-dispersive X-ray spectrometry system (EDS). Raman spectra and Fourier transform infrared (FT-IR) were performed on a Renshaw in Via Raman spectrometer and Bruker Tensor 22 instrument, respectively. Micromeritics ASAP 2460 adsorption apparatus was used to measure the N₂ adsorptions of the samples. UV/vis diffuse reflectance spectra (DRS) of the photocatalysts were recorded on a Hitachi U-3010 spectrophotometer. A Thermo Scientific Plasma Quad inductively coupled plasma atomic emission spectroscopy system (ICP-AES) was used to determine the content of Rh in the samples. The photoluminescence (PL) properties and time-resolved photoluminescence (TRPL) spectra of the samples were measured on fls980 and Horiba Fluomax-4, respectively. X-Ray photoelectron spectroscopy (XPS) tests were performed on a PHI ESCA-5000C electron spectrometer. A gas chromatography-mass spectrometry system (GC-MS, Thermo Scientific Trace 1300) was employed to detect the products. Free radical trapping experiments were conducted using an electron spin resonance (ESR) instrument. For the ˙OH active species, 20 mg sample, 1.0 mL H₂O, 50 μL t-BuOOH, and 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO, 0.5 mmol L⁻¹) were mixed together in an argon atmosphere after ultrasonic dispersion. The ESR signals of the samples were obtained after 10 min light irradiation. As for the ˙O₂⁻ active species, the test process was similar to that for the ˙OH active species, except for changing the 1.0 mL H₂O to 1.0 mL methanol in the mixed suspension. X-Ray absorption fine structure (XAFS) measurements of Rh₁/BOIO were conducted at the BL14W1 station in the Shanghai Synchrotron Radiation Facility (SSRF).

2.4. Photocatalytic reaction

The photo-oxidation performances over the samples were conducted under full light irradiation. In detail, 20 mg photocatalyst, 1.0 mL t-BuOOH (70 wt%), and 100 μL BA were mixed together after 2 min stirring in a 15 mL thick-walled pressure bottle. Before sealing, Ar gas was injected in to the bottle to remove air. 300 W Xenon lamp (PLS-SEX300D, Beijing Perfectlight Technology Co., Ltd.) was used as the light source for the photocatalytic reactions. After reacting for 6 h, the products were detected via GC-MS. In the scope of the photo-oxidation reactions, the illumination time was increased to 12 h. The conversion rate and selectivity were calculated via the following equations:

Conversion rate (%) = [(C₀ − C_r)/C₀] × 100%

Selectivity rate (%) = [C_p/(C₀ − C_r)] × 100%

C₀: original concentration of substrates, C_r: concentration of substrates after reaction, C_p: concentration of products after reaction.

2.5. Photoelectrochemical performances

The photoelectrochemical performances of the samples were tested using a three-electrode system in 0.5 M Na₂SO₄ electrolyte solution. In detail, the sample-modified FTO glasses, platinum foil (1.0 × 1.0 cm²), and Ag/AgCl (saturated KCl) were employed as the working, counter, and reference electrode, respectively. A 300 W Xe lamp was used as the light source and the data were collected on a CHI660E (China) electrochemical station under light illumination and in the dark. During the preparation of the working electrode, 5.0 mg sample, 20 μL of Nafion solution (5.0 wt%), and 1.0 mL H₂O were mixed together via ultrasonic dispersion for 1.0 h to form a photocatalyst ink. After that, the photocatalyst-modified FTO glasses were obtained by coating 100 μL ink on FTO glass (1.0 × 1.0 cm²), and dried in the air at room temperature. The same three-electrode system was also used for the Mott–Schottky (M–S) measurements.

2.6. Theoretical calculations

Density functional theory (DFT) calculations were performed using the Vienna ab initio Simulation Package (VASP) with the projector augmented-wave potentials.^30,31 The generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) functional was employed to describe the exchange-correlation.^32,33 The cutoff energy for the plane-wave basis was set to 520 eV for all the calculations. The energy convergence criterion for the self-consistent-field iteration was 10⁻⁵ eV, and the atomic positions were fully optimized until all the residual forces were smaller than 0.01 eV Å⁻¹. First, the structure optimization of bulk BiOIO₃ was performed with a 7 × 7 × 3 Monkhorst–Pack mesh for the Brillouin zone integration. The optimized bulk BiOIO₃ had a cell size of a = 5.71 Å, b = 5.78 Å, c = 11.25 Å with 4 Bi atoms, 4 I atoms, and 16 O atoms in total. Then, the slab model of BiOIO₃ was constructed with the IO-terminated (001) surface exposed, with a 15 Å vacuum space along the z-direction included to avoid interlayer interactions. After the optimization of the BiOIO₃ slab, four initial configurations of the adsorption structure of a single Rh atom on the BiOIO₃ surface were created and optimized. As a result, the four configurations eventually tended to a similar adsorption structure with Rh bonded to two terminal O atoms. Finally, the free energy differences of the t-BuOOH decomposition into O₂ were calculated on both the Rh single atom and the Rh nanoparticle (the (111) surface of bulk Rh was created to mimic the Rh nanoparticle). For all the simulations, the DFT-D3 Grimme strategy for dispersion correction of the total energy was used to include the vdW interactions.³⁴

3 Results and discussion

3.1. Characterization of Rh₁/BOIO and Rh_NP/BOIO

The atomically dispersed Rh was directly anchored on BOIO (Rh₁/BOIO) via photodeposition. Rh_NP/BOIO was obtained after activating the mixed Rh_NP and BOIO at high temperature in a vacuum (see details in the Experimental part). Rh₁/BOIO and Rh_NP/BOIO were collected as powders for X-ray diffraction (XRD) technique measurements, and both of them could be well matched with the pure BOIO sample (ICSD#262019, Fig. S1†).^35–38 Only slightly changes could be detected in the Fourier transform infrared (FT-IR) spectrometry, Raman spectra, and Brunauer–Emmett–Teller (BET) results for the samples, indicating the maintenance of the BOIO crystal structure after the Rh₁ and Rh_NP modes, respectively, were decorated on it (Fig. S2–S4†). Compared with pure BOIO, the slight changes in the Raman, FT-IR, and BET results over Rh₁/BOIO and Rh_NP/BOIO might be attributed to the decreased crystallinity of BOIO after the loading procedure. Next, transmission electron microscopy (TEM) was employed to investigate the morphologies of Rh₁/BOIO, Rh_NP/BOIO and pure BOIO, all of which displayed a random nanosheets morphology (Fig. S5–S7†). Fig. 1a shows that the Rh_NP were sized about 10 nm and distributed on the surface of BOIO nanosheets. High-resolution TEM (HRTEM) was implemented to obtain more microstructure information of Rh_NP/BOIO. As presented in Fig. 1b, lattice fringes with inter-fringe distances of 0.282 and 0.225 nm, assigned to the BOIO (200) and Rh (111) lattice planes, respectively, could be observed.^39,40 The corresponding energy-dispersive X-ray spectroscopy (EDS) mapping images in Fig. 1c further demonstrated the Rh_NP mode anchored on the surface of BOIO nanosheets. In the case of Rh₁/BOIO, no nanoparticles or clusters could be detected in Fig. 1d. The atomic resolution high-angle annular dark-field scanning TEM (AC HAADF-STEM) images combined with the corresponding EDS mapping images together further confirmed the uniform dispersion of Rh atoms on the BOIO surface (Fig. 1e and f). The content of Rh species was about 0.69 wt% for both Rh₁/BOIO and Rh_NP/BOIO, according to the ICP-AES results.

Fig. 1 (a) TEM image of Rh_NP/BOIO. (b) HRTEM image of Rh_NP/BOIO. (c) The EDS mapping images of Rh_NP/BOIO. (d) HRTEM image of Rh₁/BOIO. (e) Representative AC HAAFD-STEM image of Rh₁/BOIO. The inset shows the line intensity profiles taken along the atoms. (f) The EDS mapping of Rh₁/BOIO. (g) Rh K-edge XANES spectra of Rh₁/BOIO and the reference samples. The inset is the magnified figure. (h) FT-EXAFS spectra of Rh₁/BOIO and the reference samples. (i) R-space fitting results. The inset is the optimized structure of Rh₁/BOIO, purple ball: Bi, green ball: I, red ball: O, blue ball: Rh.

Moreover, X-ray absorption near-edge structure (XANES) measurements were conducted to investigate the chemical and coordination environment of Rh species in Rh₁/BOIO. As shown in Fig. 1g, the near-edge of Rh₁/BOIO at the Rh K-edge was between Rh foil and Rh₂O₃ references, indicating a partial oxidation state of Rh had been formed via the photodeposition method. The Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS, Fig. 1h) spectrum of Rh₁/BOIO displayed a prominent peak at about 1.5 Å, which was the first coordination shell of Rh and agreed well with the Rh–O scattering. Obviously, no peak for Rh–Rh coordination could be observed for Rh₁/BOIO. In addition, the wavelet transform (WT) EXAFS of Rh₁/BOIO presented a maximum peak at about 4.2 Å⁻¹, similar to that for the Rh–O structure in Rh₂O₃ (Fig. S8†). All the above results suggested the Rh single atom coordinated with an O atom on the surface of Rh₁/BOIO. On the basis of the EXAFS fitting results and theoretical simulations (Fig. 1i, S9, S10, and Table S1†), we could confirm the coordination number of Rh was two, indicating the Rh atom anchored on BOIO through a Rh–O₂ coordination mode in Rh₁/BOIO. By comparing the XPS results of Rh species, a more positive binding energy was also detected in Rh₁/BOIO, which was in accordance with the XANES results above. Besides, the metal Rh mode of Rh_NP/BOIO was confirmed by the XPS measurements (Fig. S11 and S12†).

3.2. Photo-oxidation performances

Based on the well-defined structures of Rh₁/BOIO and Rh_NP/BOIO, we assessed their photo-oxidation activities under full light irradiation. In the model BA photo-oxidation reaction (Fig. 2a and b), Rh₁/BOIO displayed a selectivity rate of 86.7% to BN, and completed the reaction without N-BBA product formation after 6 h light irradiation. In sharp contrast, N-BBA was the main product over Rh_NP/BOIO with a selectivity rate of 82.8% and only 4.6% selectivity to BN. More than a 99% BA conversion rate could be achieved in the photo-oxidation reaction over these two photocatalysts. The as-prepared Rh₁/BOIO and Rh_NP/BOIO exhibited better BA photo-oxidation performances than those in previous reports (Tables S2 and S3†). Comparatively, the selectivity rates of BN and N-BBA products over pure BOIO were much lower than those of Rh₁/BOIO and Rh_NP/BOIO, respectively. In addition, no BN or N-BBA product could be detected in the absence of a photocatalyst or light irradiation, suggesting that the photocatalyst and light irradiation were indispensable for BA photo-oxidation reactions. All these results extensively demonstrated that the different Rh modes on the BOIO surface were able to achieve obviously opposite selectivities during the BA photo-oxidation reaction. To optimize the Rh in Rh₁/BOIO and Rh_NP/BOIO, different amounts of Rh species were loaded on the BOIO surface. As a result, 0.69 wt% was determined to be the optimized Rh loading amount for Rh₁/BOIO and Rh_NP/BOIO (Fig. S13–S15†). Further, cyclic experiments were conducted to evaluate the stability of the Rh₁/BOIO and Rh_NP/BOIO photocatalysts. Fig. 2c and d demonstrate that both of the selectivity rates for BN over Rh₁/BOIO and N-BBA over Rh_NP/BOIO were almost unchanged during the cyclic tests. Next, crystal structure characterizations, including XRD, TEM, and EDS, were carried out to confirm the structural maintenance of the used Rh₁/BOIO and Rh_NP/BOIO (Fig. S16–S18†). In addition, the scope of the photo-oxidation reactions of the BA-based substrates over Rh₁/BOIO and Rh_NP/BOIO, e.g., 4-methylbenzylamine, 4-methoxybenzylamine, 4-fluorobenzylamine, and 4-chlorobenzylamine, also displayed the opposite selectivities. As shown in Fig. 2e and f, R–Bn≡N was the main product over Rh₁/BOIO, while the R–Bn=N–Bn–R product was obtained over Rh_NP/BOIO (R is CH₃, OCH₃, F, or Cl), indicating the opposite selectivity of the BA oxidation reaction is applicable to a wider range of substrates over these two photocatalysts.

Fig. 2 (a) Reaction scheme for the BA photo-oxidation reaction. (b) Selectivity rates of BN and N-BBA after 6 h photo-oxidation reaction over the samples. (c) Stability of Rh₁/BOIO. (d) Stability of Rh_NP/BOIO. (e) Scope of the photo-oxidation reaction over Rh₁/BOIO. (f) Scope of the photo-oxidation reaction over Rh_NP/BOIO.

3.3. Charge-carrier dynamics

The light-harvesting abilities of the samples were characterized by UV/Vis diffuse reflectance spectroscopy (DRS). As displayed in Fig. 3a, Rh₁/BOIO displayed stronger visible-light absorption than that of pure BOIO and Rh_NP/BOIO in the range of 400–800 nm. A slight red-shift of the absorption edges could be observed over Rh₁/BOIO and Rh_NP/BOIO compared with that of BOIO. According to the Kubelka–Munk function and Mott–Schottky results, the bandgaps of pure BOIO and Rh₁/BOIO were determined as 3.38 and 3.34 eV, respectively (Fig. 3b, S19 and S20†). The Rh–O₂ coordination mode in Rh₁/BOIO was able to enhance the light harvesting via the narrower bandgap of BOIO. As for Rh_NP/BOIO, the promotion of its light-absorption capacity could be mainly attributed to the black color of Rh_NP. To investigate the dynamic behaviors of the photogenerated carriers, several spectroscopic characterizations were employed. Photoelectrochemical tests were conducted to evaluate the separation efficiency of charge carriers. First, all the samples were evenly coated on the FTO glasses before testing.⁴¹ As presented in Fig. 3c, the photocurrent densities of Rh₁/BOIO and Rh_NP/BOIO were much higher than that of pure BOIO. Compared with pure BOIO, Rh₁/BOIO possessed the smallest semicircle diameter in the electrochemical impedance spectra (EIS) (Fig. S21†).⁴² These results suggest that both Rh₁ and Rh_NP were able to effectively facilitate the photogenerated charge transfer. The steady-state PL spectra results of the samples are also given (Fig. S22†). The PL intensity was significantly decreased after Rh₁ and Rh_NP, respectively, were anchored on BOIO, implying that Rh₁/BOIO and Rh_NP/BOIO exhibited lower charge-carrier recombination efficiencies than that of BOIO under light excitation.^43,44 Based on the TRPL and fitting results, the lifetimes of BOIO, Rh_NP/BOIO, and Rh₁/BOIO were determined as 1.16, 1.22, and 1.35 ns, respectively (Fig. 3d and Table S4†), which suggests that the lifetimes of the charge carriers were improved after anchoring Rh₁ and Rh_NP on the BOIO surface,⁴⁵ which corresponds to the steady-state PL spectra results. Taken together, the separation and transmission efficiency of the photogenerated carriers could be improved after respectively anchoring Rh₁ and Rh_NP on the BOIO surface. Despite the separation and transfer efficiency of charge carriers being promoted, the mechanism for the opposite selectivity that occurred over Rh₁/BOIO and Rh_NP/BOIO in the BA photo-oxidation reaction remained unclear.

Fig. 3 (a) UV/Vis DRS spectra. (b) Band structures of the samples. (c) Photocurrent densities of Rh₁/BOIO, Rh_NP/BOIO, and BOIO. (d) TRPL spectra and fitting results of the samples.

3.4. Mechanism analysis

To elucidate the opposite selectivity mechanism over Rh₁/BOIO and Rh_NP/BOIO, we conducted free radical trapping experiments to explore the active species during the photo-oxidation reaction. As shown in Fig. 3a and b, the characteristic peak intensities of DMPO-˙OH and DMPO-˙O₂⁻ species among these samples were totally different. In Rh₁/BOIO, DMPO-˙OH were the dominant species under light irradiation, much higher than with Rh_NP/BOIO and pure BOIO. By contrast, Rh_NP/BOIO displayed the highest capacity for DMPO-˙O₂⁻ species generation among the three samples. Consequently, ˙OH active species could be easily produced on Rh₁/BOIO, while ˙O₂⁻ were the active species for Rh_NP/BOIO. Such entirely different active species formation performances were more likely the cause of the opposite selectivity during the BA photo-oxidation reaction over Rh₁/BOIO and Rh_NP/BOIO. Density functional theory (DFT) calculations were performed to further reveal the reaction mechanism. By comparing the density of states between pure BOIO and Rh₁/BOIO, an obvious contribution from Rh species appeared near the conduction band minimum of BOIO (Fig. 4c and S23†). Namely, the bandgap of BOIO was narrowed after Rh₁ species were anchored on, which was consistent with the band structural analysis in Fig. 3b. Considering the thermodynamic of the active species generation, ˙OH and O₂ have been recognized as the main intermediates for decomposing t-BuOOH/OOH⁻.^46,47 Compared to Rh₁/BOIO, the free energy changes over Rh_NP/BOIO were more conductive to generating O₂ molecules (Fig. 4d), which implied that the ˙OH could easily undergo coupling with Rh₁ and this impedes the disproportionation reaction of two ˙OH. These observations reveal that Rh₁ and Rh_NP in BOIO facilitated the generation of ˙OH and ˙O₂⁻ active species, respectively, as well as achieving the opposite selectivity in the BA photo-oxidation reaction.

Fig. 4 Electron spin resonance (ESR) signals of (a) DMPO-˙OH and (b) DMPO-˙O₂⁻ species after 10 min light irradiation over pure BOIO, Rh₁/BOIO, and Rh_NP/BOIO. (c) PDOS results of Rh₁/BOIO. (d) Calculated free energy diagram of all the steps for OOH⁻ decomposition to different intermediates on Rh₁/BOIO and Rh_NP/BOIO. Schematic illustration of the possible reaction mechanism for the BA photo-oxidation reaction over (e) Rh₁/BOIO and (f) Rh_NP/BOIO.

On the basis of the above experiments and our theoretical analysis, we propose the possible reaction mechanism over Rh₁/BOIO and Rh_NP/BOIO.^48–51 As shown in Fig. 4e, Rh₁/BOIO is excited to generate electron–hole pairs during light irradiation. The photogenerated electrons are able to decompose t-BuOOH to t-BuO˙ and ˙OH (Step 1). Besides, BA can be oxidized to PhCHNH intermediate by the photogenerated holes and t-BuO˙ species (Steps 2 and 3). The Rh₁ mode of Rh₁/BOIO fixes one ˙OH to form ˙OH–Rh₁, which not only prevents the disproportionation reaction of two ˙OH to produce O₂ and H₂O, but also further oxidizes the PhCHNH intermediate to BN product (Step 4). In the case of Rh_NP/BOIO (Fig. 4f), the Rh_NP on the BOIO surface prefers to catalyze the disproportionation reaction of ˙OH to generate O₂ and H₂O, due to the multi-˙OH species adsorbed on Rh_NP (Step 4). O₂ molecules react with photogenerated electrons to generate ˙O₂⁻ active species, which then oxidize the PhCHNH to the PhCHO intermediate (Step 6). Another BA molecule couples with PhCHO in the last step to produce N-BBA (Step 7).

4 Conclusions

In summary, two different Rh modes (Rh₁ and Rh_NP) were respectively anchored on BOIO surface to realize a completely opposite selectivity of BA photo-oxidation reaction. The as-prepared Rh₁/BOIO exhibited an 86.7% selectivity rate for BN product, while it was 4.6% over Rh_NP/BOIO. N-BBA was the main product over Rh_NP/BOIO with an 82.8% selectivity rate, but this could not be detected when using Rh₁/BOIO as the photocatalyst under the same reaction conditions. The theoretical simulations and experimental results revealed that the opposite selectivity performances arose from the different active species of ˙OH and ˙O₂⁻ generated over Rh₁/BOIO and Rh_NP/BOIO, respectively, during the photocatalytic reaction. Meanwhile, this phenomenon is applicable equally to other types of BA-based substrates, which further confirms that the selective tuning ability can be achieved via anchoring different Rh modes on BOIO. Our work highlights a convenient approach to fabricate catalysts with an ability to tune the product selectivity during photocatalytic reactions.

Author contributions

J. L and Y. W. designed, performed the experiments and wrote the paper. G. W. and Z. Lou. supervised the project and revised the manuscript. K. S., T. J. and Z. Li. performed the theoretical calculations. Q. C. and R. Y. analyzed the XAFS results data.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (22202011, 21971002), the Natural Science Research Project of Higher education in Anhui Province (2022AH020020), and Natural Science Foundation of Shandong Province (ZR2022QB056). The authors thank the BL14W1 station in the Shanghai Synchrotron Radiation Facility (SSRF) for help with characterization.

References

1. R. D. Patil and M. K. Gupta, Adv. Synth. Catal., 2020, 362, 3987–4009.
2. B. Xu, E. M. Hartigan, G. Feula, Z. Huang, J.-P. Lumb and B. A. Arndtsen, Angew. Chem., Int. Ed., 2016, 128, 16034–16038.
3. M. Lv, F. Tong, Z. Wang, Y. Liu, P. Wang, H. Cheng, Y. Dai, Z. Zheng and B. Huang, J. Mater. Chem. A, 2022, 10, 19699–19709.
4. M. T. Bender and K. S. Choi, JACS Au, 2022, 2, 1169–1180.
5. J. Qin, Y. Long, F. Sun, P. P. Zhou, W. D. Wang, N. Luo and J. Ma, Angew. Chem., Int. Ed., 2022, 61, e202112907.
6. S. Wang, C. Lai, Y. Zhang, S. Bao, K. Lv and L. Wen, J. Mater. Chem. A, 2022, 10, 20975–20983.
7. R. Ray, A. S. Hazari, G. K. Lahiri and D. Maiti, Chem. Asian. J., 2018, 13, 2138–2148.
8. R. V. Jagadeesh, H. Junge and M. Beller, Nat. Commun., 2014, 5, 4123.
9. Z. Yu, E. R. Waclawik, Z. Wang, X. Gu, Y. Yuan and Z. Zheng, J. Mater. Chem. A, 2017, 5, 4607–4615.
10. Z. Tian, C. Han, Y. Zhao, W. Dai, X. Lian, Y. Wang, Y. Zheng, Y. Shi, X. Pan, Z. Huang, H. Li and W. Chen, Nat. Commun., 2021, 12, 2039.
11. G. Wang, Y. Liu, J. Zhang, Q. Chen, K. Fang and J. Mao, J. Mater. Chem. A, 2022, 10, 21349–21355.
12. F. Wang, X. Han, Z. Jia, Y. Li, T. Zhang, A. Han and J. Liu, Catal. Sci. Technol., 2021, 11, 6947–6951.
13. S. Peiris, J. McMurtrie and H.-Y. Zhu, Catal. Sci. Technol., 2016, 6, 320–338.
14. Y. Li, J. Li and Y. Xu, EnergyChem, 2021, 3, 100047.
15. Z. Xue, D. Luan, H. Zhang and X. Lou, Joule, 2022, 6, 92–133.
16. L. Tong, L. Ren, A. Fu, D. Wang, L. Liu and J. Ye, Chem. Commun., 2019, 55, 12900–12903.
17. C. Dong, C. Lian, S. Hu, Z. Deng, J. Gong, M. Li, H. Liu, M. Xing and J. Zhang, Nat. Commun., 2018, 9, 1252.
18. X. Liu, D. Dai, Z. Cui, Q. Zhang, X. Gong, Z. Wang, Y. Liu, Z. Zheng, H. Cheng, Y. Dai, B. Huang and P. Wang, ACS Catal., 2022, 12, 12386–12397.
19. G. Wang, R. Huang, J. Zhang, J. Mao, D. Wang and Y. Li, Adv. Mater., 2021, 33, 2105904.
20. F. Zhang, J. Ma, Y. Tan, G. Yu, H. Qin, L. Zheng, H. Liu and R. Li, ACS Catal., 2022, 12, 5827–5833.
21. Y. Li, Y. Qu and G. Wang, J. Mater. Chem. A, 2022, 10, 5990–5997.
22. G. Wang, Z. Chen, T. Wang, D. Wang and J. Mao, Angew. Chem., Int. Ed., 2022, 61, e202210789.
23. E. Zhao, M. Li, B. Xu, X. L. Wang, Y. Jing, D. Ma, S. Mitchell, J. Perez-Ramirez and Z. Chen, Angew. Chem., Int. Ed., 2022, 61, e202207410.
24. G. Wang, C. He, R. Huang, J. Mao, D. Wang and Y. Li, J. Am. Chem. Soc., 2020, 142, 19339–19345.
25. X. Shi, Y. Huang, Y. Bo, D. Duan, Z. Wang, J. Cao, G. Zhu, W. Ho, L. Wang, T. Huang and Y. Xiong, Angew. Chem., Int. Ed., 2022, 61, e202203063.
26. G. Wang, Y. Wu, Z. Li, Z. Lou, Q. Chen, Y. Li, D. Wang and J. Mao, Angew. Chem., Int. Ed., 2023, 62, e202218460.
27. J. Cai, A. Cao, Z. Wang, S. Lu, Z. Jiang, X. Dong, X. Li and S. Zang, J. Mater. Chem. A, 2021, 9, 13890–13897.
28. W. Wu, E. Cui, Y. Zhang, C. Zhang, F. Zhu, C.-H. Tung and Y. Wang, ACS Catal., 2019, 9, 6335–6341.
29. J. Yang, Z. Sun, K. Yan, H. Dong, H. Dong, J. Cui, X. Gong, S. Han, L. Huang and J. Wen, Green Chem., 2021, 23, 2756–2762.
30. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
31. P. E. Blöchl, Phys. Rev. B, 1994, 50, 17953–17979.
32. J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh and C. Fiolhais, Phys. Rev. B, 1992, 46, 6671–6687.
33. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
34. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104.
35. F. Chen, H. Huang, L. Ye, T. Zhang, Y. Zhang, X. Han and T. Ma, Adv. Funct. Mater., 2018, 28, 1804284.
36. W. Wang, B. Huang, X. Ma, Z. Wang, X. Qin, X. Zhang, Y. Dai and M. H. Whangbo, Chem. - Eur. J., 2013, 19, 14777–14780.
37. F. Chen, H. Huang, C. Zeng, X. Du and Y. Zhang, ACS. Sustain. Chem. Eng., 2017, 5, 7777–7791.
38. S. Yu, H. Huang, F. Dong, M. Li, N. Tian, T. Zhang and Y. Zhang, ACS Appl. Mater. Interfaces, 2015, 7, 27925–27933.
39. F. Chen, Z. Ma, L. Ye, T. Ma, T. Zhang, Y. Zhang and H. Huang, Adv. Mater., 2020, 32, e1908350.
40. G. Lin, Z. Zhang, Q. Ju, T. Wu, C. U. Segre, W. Chen, H. Peng, H. Zhang, Q. Liu, Z. Liu, Y. Zhang, S. Kong, Y. Mao, W. Zhao, K. Suenaga, F. Huang and J. Wang, Nat. Commun., 2023, 14, 280.
41. M. Ni, Y. Zhu, C. Guo, D. Chen, J. Ning, Y. Zhong and Y. Hu, ACS Catal., 2023, 13, 2502–2512.
42. A. Etogo, R. Liu, J. Ren, L. Qi, C. Zheng, J. Ning, Y. Zhong and Y. Hu, J. Mater. Chem. A, 2016, 4, 13242–13250.
43. C. Li, H. Wu, D. Zhu, T. Zhou, M. Yan, G. Chen, J. Sun, G. Dai, F. Ge and H. Dong, Appl. Catal., B, 2021, 297, 120433.
44. L. Li, X. Dai, D. Chen, Y. Zeng, Y. Hu and X. Lou, Angew. Chem., Int. Ed., 2022, 61, e202205839.
45. X. Yan, M. Xia, H. Liu, B. Zhang, C. Chang, L. Wang and G. Yang, Nat. Commun., 2023, 14, 1741.
46. J. Van Der Zee, D. P. Barr and R. P. Mason, Free Radicals, 1996, 20, 199–206.
47. M. Zhao, K. Zhang, J. Xu and J. Li, Org. Chem. Front., 2022, 9, 4063–4069.
48. S. Q. Gomes and A. G. Salles, Synth. Commun., 2019, 49, 3389–3399.
49. Y. R. Girish, R. Biswas and M. De, Chemistry, 2018, 24, 13871–13878.
50. Y. Fu, M. Zheng, Q. Li, L. Zhang, S. Wang, V. V. Kondratiev and B. Jiang, RSC Adv., 2020, 10, 28059–28065.
51. D. Tantraviwat, A. Nattestad, J. Chen and B. Inceesungvorn, J. Colloid Interface Sci., 2023, 629, 854–863.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3ta01804k

‡ These authors contributed equally to this work.

---