Plasmonic O2 dissociation and spillover expedite selective oxidation of primary C–H bonds

Manipulating O2 activation via nanosynthetic chemistry is critical in many oxidation reactions central to environmental remediation and chemical synthesis. Based on a carefully designed plasmonic Ru/TiO2−x catalyst, we first report a room-temperature O2 dissociation and spillover mechanism that expedites the “dream reaction” of selective primary C–H bond activation. Under visible light, surface plasmons excited in the negatively charged Ru nanoparticles decay into hot electrons, triggering spontaneous O2 dissociation to reactive atomic ˙O. Acceptor-like oxygen vacancies confined at the Ru–TiO2 interface free Ru from oxygen-poisoning by kinetically boosting the spillover of ˙O from Ru to TiO2. Evidenced by an exclusive isotopic O-transfer from 18O2 to oxygenated products, ˙O displays a synergistic action with native ˙O2− on TiO2 that oxidizes toluene and related alkyl aromatics to aromatic acids with extremely high selectivity. We believe the intelligent catalyst design for desirable O2 activation will contribute viable routes for synthesizing industrially important organic compounds.

Finite-difference-time-domain (FDTD) simulation. FDTD simulation were performed using FDTD Solution 8.24 (Lumerical Solution). During the simulation, an electromagnetic pulse in the spectral range from 20 nm to 1100 nm was launched into a box containing the target nanostructure. A mesh size of 0.1 nm was employed in calculating the extinction spectra and charge distribution contours of the Ru nanosphere and Ru nanosphere-TiO 2 . The refractive index of the surrounding medium was 1.33 of water. The dielectric function of Ru and TiO 2 were extracted from the D. W. Lynch et al. and M. W. Ribarsky et al., respectively. 8,9 The diameter of the Ru sphere was set to be 2 nm. The size of the TiO 2 was set as 5×5×2 nm. The Ru sphere with diameter of 2 nm were set on the middle surface of TiO 2 without gap.   To directly characterize charge carrier dynamics, informative photoluminescence (PL) spectroscopy was adopted.
Under the excitation of a 300 nm laser, steady-state PL of TiO 2 displayed two emission peaks (Fig. S3a). The dominant peak at around 400 nm was the band-edge emission, while the other weaker one centered at 550 nm was from the radiative recombination of electrons trapped at OV states (~0.85 eV below the CB edge) with the holes at VB (inset of Fig. S3a). 10 Due to the high concentration of surface OVs on TiO 2-x , the defect emission peak became comparable to the band-edge emission. After Ru deposition, PL intensity of Ru/TiO 2 experienced no remarkable change, while both the band-edge and OV-induced emissions of Ru/TiO 2-x suffered from a substantial quenching (Fig. S3a). The PL quenching in Ru/TiO 2-x possibly stemmed from a combined effect between OVs and Schottky barrier on suppressing electron-hole recombination.
To validate this assumption, we performed time-resolved PL measurements on the OV emission peak. According to the decay curves, both TiO 2 and TiO 2-x could be fitted with a double-exponential function, including a fast interband recombination decay component (τ 1 ) and a slow OV-mediated recombination decay component (τ 2 ) (Fig. S3b). Due to the efficient trapping of photoelectrons by OVs, TiO 2-x displayed a much higher proportion of the long-living component (τ 2 , 68%), whose average time (2.99 ns) was therefore much longer than that of TiO 2 (1.42 ns) (Table S1). Ru/TiO 2 displayed an almost identical decay behavior to TiO 2 , suggesting the weak interactions between Ru and TiO 2 . As for Ru/TiO 2-x , its decay was the slowest among the as-prepared photocatalysts (Fig. S3b). After careful fitting, we found a third exponential constant (τ 3 ) with the longest lifetime of 19.05 ns was necessary to obtain a satisfactory fit, which further lengthened the carrier lifetime to 8.54 ns (Table S1). It is to be noted that the introduction of τ 3 was essential that signified the contribution of Schottky barrier. As schematically illustrated in Fig. S3c, the energetic electrons, either on the CB or trapped on the OV states, could be transferred to Ru nanoparticles. Whereas the presence of Schottky barrier at the interface would block the back-transfer of photoelectrons, thus suppressing their direct or indirect recombination with holes ( Fig. S3c).

Fig. S4
Simulated extinction spectra of a Ru nanosphere of 2 nm in diameter supported on a TiO 2 nanosheet of 2 nm in thickness. Inset shows the electric field intensity enhancement contours of the Ru/TiO 2 hybrid system. The finite-difference time-domain (FDTD) simulations reveal that a Ru sphere (diameter of 2 nm) on TiO 2 displays a surface plasmon resonance (SPR) absorption ranging from 100 nm to 600 nm centered at 215 nm ( Fig. S4a). 11 It is to be noted that our experimentally determined UV-vis absorption spectrum displays a flat and extended absorption tail in the visible light region. 12 This enhanced and widened absorption tail could be possibly due to the hybridization effect between the defective substrate (TiO 2-x ) and plasmonic metal (Ru), and also the hybridization among Ru nanoparticles. [13][14][15] Under visible light (λ = 420 nm), Ru can generate electric field intensity enhancement several times of the incident field (|E/E 0 | 2 ) at the Ru-TiO 2 interface (Fig. S4b).      XRD patterns of the Ru/TiO 2-x before and after multicycle photocatalytic toluene oxidation were the same, indicating Ru/TiO 2-x possessed good chemical stability for photocatalytic toluene oxidation (Fig. S11a). Meanwhile, after multicycle test, TEM image showed Ru nanoparticles on TiO 2-x were still in the metallic state based on the Ru(101) lattice fringes (Fig. S11b). Thus, the possible decomposition of TiO 2-x support and oxidation of Ru nanoparticles were ruled out.
It is highly possible that OVs in Ru/TiO 2-x were gradually oxidized during photocatalytic toluene oxidation. Based on the EPR spectra, OVs in Ru/TiO 2-x after multicycle test were slightly decreased (Fig. S11c). There are two mechanisms associated with the excitation of OVs in TiO 2 . First, photoelectrons can be excited from the valence band of TiO 2 to the OVs states, and then to the conduction band (Fig S11d: Mechanism I). Second, the localized electrons on the OVs states can be directly excited to the conduction band of TiO 2 (Fig S11d: Mechanism II). After the localized electrons are trapped by adsorbed O 2 , the OVs are quenched through the second mechanism. In our case, we believe both mechanisms co-exist.
However, the dominant excitation pathway is inferred to be the first one, based on the slightly decreased OVs in Ru/TiO 2-x after multiple photocatalytic toluene oxidation. We prepared Ru nanoparticles with larger diameters from 3.5 nm to 6 nm, and from 5.0 nm to 10.0 nm by increasing the thermal reduction time of Ru 3 (CO) 12 during the Ru/TiO 2-x synthesis ( Fig. S12a and S12b). Increasing the Ru nanoparticle size led to the decrease of photocatalytic toluene oxidation activity (Fig. S12c). We reckon that larger Ru nanoparticle size might occupy the OVs, inhibit reactants adsorption, and decreases the interfacial area, all of which are disadvantageous for photocatalytic toluene oxidation.    To investigate the influence of OVs in Ru/TiO 2-x for photocatalytic toluene oxidation, we adjusted the annealing temperature for TiO 2-x synthesis from 200 o C to 350 o C with NaBH 4 as a reducing reagent. In our study, the default TiO 2-x used to synthesize Ru/TiO 2-x was prepared at 300 o C. According to the EPR spectra, the symmetrical signal of OV was gradually enhanced along with the annealing temperature increase (Fig. S17a). Subsequently, Ru nanoparticles were deposited onto TiO 2-x with different concentrations of OVs through the same method. It was interesting to find along with OVs increase, Ru/TiO 2-x displayed both enhanced conversion efficiency and selectivity for toluene oxidation (Fig. S17b).
As discussed in our manuscript, the OVs on TiO 2 play an important role in facilitating O 2 activation into •O 2 − that S14 selectively oxidizes toluene into benzaldehyde. Together with the •O 2 dissociated from the Ru nanoparticles, benzaldehyde can be further oxidized into benzoic acid (Fig. 5d).  relative concentration in the multiexponential decay and is expressed as . After 6 hours of reaction, the products distributions and concentrations were determined by GC-MS. S18  under a 300-W Xe lamp with a 400 nm cutoff filter. After 6 hours of reaction, the products distributions and concentrations were determined by GC-MS.