Materials

The raw Mnt (purity >95%) used in this study was obtained from Inner Mongolia, China. The chemical composition of Mnt, as determined by X-ray fluorescence (XRF), was given as follows: SiO₂ (59.23%), Al₂O₃ (16.49%), Fe₂O₃ (4.39%), MgO (5.36%), CaO (2.81%), K₂O (0.07%) and Ti₂O (0.26%). Sodium chloride (NaCl), titanium tetrachloride (TiCl₄) and platinum(II) nitrate (Pt(NO₃)₂) were purchased from Shanghai Aladdin Chemical Reagent Co Ltd. Hydrochloric acid (HCl; 37 wt.%) was supplied by Guangzhou Chemical Reagent Factory, China.

Catalyst preparation

The scheme for material preparation is shown in Fig. S1.

Characterization methods

A Rigaku MiniFlex-600 X-ray diffractometer (Rigaku, Japan) was employed to obtain X-ray diffraction (XRD) traces using Cu-Kα radiation (λ = 0.154 nm), a Ni filter and a Hybrid pixel array detector at 40 kV and 15 mA. X-ray photoelectron spectroscopy (XPS) spectra were collected on a Thermo K-Alpha XPS instrument (Thermo Fisher Scientific, UK) featuring a monochromatic Al-Kα X-ray source (1468.6 eV), with pass energies set to 100 and 30 eV for survey and high-resolution scans, respectively. N₂ adsorption–desorption studies were conducted using an ASAP 2020 system (Micromeritics Instrument Corporation, USA) at 77 K. Following degassing at 120℃ for 12 h, the S _BET, micropore area and pore-size distribution (PSD) were calculated using the Brunauer–Emmett–Teller (BET) equation, t-plot method and the non-local density functional theory (NLDFT) model, respectively. Transmission electron microscopy (TEM) images and energy-dispersive X-ray spectroscopy (EDS) spectra were recorded using an FEI Talos F200S field emission TEM device (FEI Company, USA) at 200 kV, while scanning electron microscopy (SEM) images were captured via a SU-8010 cold-field emission SEM device (Hitachi, Japan). The Pt loading was determined using ICP-OES (Thermo Scientific, UK). In situ diffuse reflectance Fourier-transform infrared spectroscopy (DRIFTS; INVENIO R, Bruker, Germany) was conducted using a device equipped with a Harrick diffuse reflectance pool and a mercury–cadmium–telluride (MCT) detector. Sample were measured using a Bruker Vertex 70V infrared spectrometer (Bruker, Germany). The spectra were recorded in the wavenumber range of 4000–400 cm^–1 with a resolution of 4 cm^–1 across 64 scans.

Catalyst performance evaluation

The catalytic activity of as-prepared catalysts was evaluated in a fixed-bed quartz microreactor with an inner diameter of 6 mm. A total of 80 mg of the catalyst (sieved to 35–65 mesh) was loaded into the reactor. Toluene vapour was generated via the bubbling method and mixed with air to achieve a feed concentration of 1000 ppm. This represents a typical mid- to high industrial level, allowing for the rigorous assessment of catalyst efficiency. It is also a standard concentration used in laboratories for evaluating supported noble metal catalysts, enabling fair comparison with literature data.

The total flow rate was maintained at 80 mL min^–1, corresponding to a weight hourly space velocity (WHSV) of 60,000 mL g^–1 h^–1. The CO₂ concentration in the effluent gas was monitored quantitatively using an infrared gas analyser (GXH-3010E, Huayun, China). To investigate the light-off characteristics and complete oxidation temperature, relevant tests were conducted using the stepwise heating mode. The reactor was initially stabilized at 140℃ for 30 min. Subsequently, the temperature was increased stepwise, with each temperature point held for 30 min to ensure steady-state conditions were established, until complete toluene conversion was achieved. The catalytic performance was assessed based on the CO₂ yield. The temperatures corresponding to 10%, 50% and 90% CO₂ yields were denoted as T ₁₀, T ₅₀ and T ₉₀, respectively. The CO₂ yield was calculated using Equation 1:

(1)\begin{equation}\left( {C/{C_0}} \right){\text{ }} \times {\text{ }}100\% \end{equation}

where C is the CO₂ concentration in the outlet and C ₀ is the CO₂ concentration when toluene was completely oxidized.

Structural characteristics of the samples

Figure 1a presents the XRD traces of Na-Mnt and TiO₂-Mnt. For Na-Mnt, the diffraction peak at 6.88°2θ corresponds to the ordered (001) reflection with a basal spacing (d ₀₀₁) of 1.27 nm, while the peak at 26.7°2θ is attributed to a quartz impurity phase. Considering that the thickness of a single Mnt layer is ∼0.96 nm, the calculated interlayer spacing is ∼0.29 nm, which is characteristic of interlayer hydrated sodium ions (Chmielarz et al., Reference Chmielarz, Piwowarska, Kuśtrowski, Węgrzyn, Gil and Kowalczyk2011; Zhu et al., Reference Zhu, Chen, Du, He, Liu, Xi and He2020). As shown in the inset of Fig. 1a, for the XRD trace of TiO₂-Mnt a diffraction peak corresponding to a d-spacing of 4.18 nm appears in the low-angle region (Saini et al., Reference Saini, Pinto and Pires2010). This indicates that the interlayer spacing significantly increased to 3.22 nm, confirming substantial interlayer expansion, which is consistent with previous reports on TiO₂-Mnt (Bergaya, Reference Bergaya1995). Upon calcination, the interlayer titanium polycations undergo dehydration and dehydroxylation, transforming into TiO₂. Furthermore, the peaks observed at 25.10°2θ, 48.12°2θ and 54.34°2θ correspond to the (101), (200) and (105) planes of anatase TiO₂, respectively, indicating that TiO₂ was successfully loaded into the Na-Mnt interlayers (Butman et al., Reference Butman, Ovchinnikov, Karasev, Kochkina, Agafonov and Vinogradov2018). For the TiO₂/SiO₂ composite obtained after subsequent acid leaching, a weak and broad diffraction peak is observed in the range of 20–23°2θ, which can be attributed to the formation of amorphous silica (Chen et al., Reference Chen, Zhu, Deng, Ma, He and Du2019). Additionally, the TiO₂ content was determined to be ∼5.37% by ICP-OES. To further rule out the potential influence of residual Fe₂O₃, the Fe content decreased significantly from 2.56 wt.% before acid leaching to 0.11 wt.% afterwards, as determined by EDS (Fig. S2), confirming that the acid treatment process effectively removed the metallic impurities. In the XRD traces of the resulting 0.17Pt–, 0.37Pt– and 0.69Pt–TiO₂/SiO₂ samples (Fig. 1b), no distinct diffraction peaks of platinum were detected, which can be ascribed to the low Pt loading or the high dispersion of Pt nanoparticles on the support surface (Xu et al., Reference Xu, Zhang, Gao, Zhu and Zhao2024).

Figure 1.

XRD traces of (a) Na-Mnt and TiO₂-Mnt and (b) TiO₂/SiO₂ and xPt–TiO₂/SiO₂ (x = 0.17, 0.37 or 0.69). (c) FTIR traces of Na-Mnt, TiO₂/SiO₂ and 0.37Pt–TiO₂/SiO₂. (d) N₂ adsorption–desorption isotherms and PSD traces (inset) of Mnt, TiO₂-Mnt and 0.37Pt–TiO₂/SiO₂.

Figure 1 Long description

The image A showing an X ray diffraction line plot labeled a that compares two samples labeled “TiO2-Mnt” and “Na-Mnt”. The y-axis label is “Intensity (a.u.)”, where a.u. is shown in the label. The x-axis label is “2theta”, with labeled x-axis values 10, 20, 30, 40, 50 and 60. A legend lists “Montmorillonite”, “Anatase” and “Quartz” with marker symbols. A numeric annotation “1.27 nm” is shown near a low-angle feature. A numeric annotation “26.7” is shown above a peak. An inset shows a low-angle zoom with labeled x-axis values 2.0, 2.5 and 3.0 and includes the annotation “4.18 nm”. Takeaway statement: the two traces show different low-angle annotated spacings (1.27 nm in the main plot and 4.18 nm in the inset) and multiple marked peak positions associated with the legend entries. The image B showing an X ray diffraction line plot labeled b that compares four samples labeled “TiO2/SiO2”, “0.17Pt-TiO2/SiO2”, “0.37Pt-TiO2/SiO2” and “0.69Pt-TiO2/SiO2”. The y-axis label is “Intensity (a.u.)”, where a.u. is shown in the label. The x-axis label is “2theta”, with labeled x-axis values 10, 20, 30, 40, 50 and 60. A legend lists “Anatase” and “Quartz” with marker symbols. Takeaway statement: the four traces show similar peak locations with marker symbols appearing near peaks and the traces are separated vertically to compare the different labeled compositions. The image C showing an infrared absorbance line plot labeled c that compares three samples labeled “Na-Mnt”, “TiO2/SiO2” and “0.37Pt-TiO2/SiO2”. The y-axis label is “Absorbance”. The x-axis label is “Wavenumber (cm-1)”, with labeled x-axis values 4000, 3500, 3000, 1000 and 500. Numeric peak labels shown on the traces are 3420, 3620, 1100, 1032, 960, 913, 842, 800, 620 and 520. Takeaway statement: the three spectra show multiple labeled band positions, including labels in the 3000 to 4000 region (3420 and 3620) and several labeled bands between 1100 and 520. The image D showing a nitrogen adsorption–desorption plot labeled d with multiple paired curves using point markers for three labeled samples: “Mnt”, “TiO2-Mnt” and “0.37Pt-TiO2/SiO2”. The y-axis label is “Quantity Adsorbed (cm3/g)”. The x-axis label is “Relative Pressure (P/P0)”, with labeled x-axis values 0.0, 0.2, 0.4, 0.6, 0.8 and 1.0. The curves rise gradually at lower relative pressure and rise more steeply near the highest relative pressure values. An inset shows pore-size distribution traces with y-axis label “Differential pore volume (cm3/g)” and x-axis label “Pore width (nm)” and the inset traces are labeled “Mnt”, “TiO2-Mnt” and “0.37Pt-TiO2/SiO2”. Takeaway statement: the three samples show different adsorption–desorption curve heights and different pore-width distribution trace shapes in the inset, enabling comparison across the labeled materials.

The Fourier-transform infrared (FTIR) spectra of Na-Mnt, TiO₂/SiO₂ and 0.37Pt–TiO₂/SiO₂ are depicted in Fig. 1c & Table S1. In the case of Na-Mnt, peaks located at 3620 and 3420 cm^–1 are assigned to structural hydroxyls and interlayer water stretching vibrations, respectively. The disappearance of the peak at 3620 cm^–1 in both TiO₂/SiO₂ and 0.37Pt–TiO₂/SiO₂ is attributed to the elimination of surface hydroxyls via acid leaching and calcination (Tyagi et al., Reference Tyagi, Chudasama and Jasra2006). Significant structural evolution is evidenced by the blue shift of the Si–O–Si stretching peak from 1032 to 1100 cm^–1, reflecting the conversion of the Mnt tetrahedral framework into amorphous silica (Yang et al., Reference Yang, Zhu, Chen, Fu, He, Zhu and He2020b). This transformation is further supported by the disappearance of the Al–OH (913 cm^–1) and Mg–OH (842 cm^–1) bending modes (Xiang et al., Reference Xiang, Zhu, Chen, Lv and Yang2022), as well as the Si–O–Mg (620 cm^–1) and Si–O–Al (520 cm^–1) vibrations, indicating the effective leaching of octahedral cations (Al, Mg; Maleki & Karimi-Jashni, Reference Maleki and Karimi-Jashni2017). By contrast, the existence of peaks located at 800 cm^–1 (Si–O–Si) and 960 cm^–1 (Si–OH) is a confirmation of the formation of amorphous silica (Wang et al., Reference Wang, Wang, Yin and Wang2016; He et al., Reference He, Zhu, Chen, Tang, Ji and Wei2024).

Figure 1d displays the N₂ adsorption–desorption isotherms of Mnt, TiO₂-Mnt and 0.37Pt–TiO₂/SiO₂. The S _BET and pore volumes determined on these isotherms are summarized in Table 1. All three isotherms show typical Type-IV curves, indicating the presence of a mesoporous structure (Rouquerol et al., Reference Rouquerol, Rouquerol and Sing1999). Moreover, in the low-relative-pressure region, the isotherms of TiO₂-Mnt and 0.37Pt–TiO₂/SiO₂ show Langmuir-type (Type-I) characteristics. This behaviour stems from their high affinity for nitrogen molecules, confirming the existence of micropores. Additionally, all three isotherms present distinct H3-type hysteresis loops, corresponding to slit-shaped pores (Zuo et al., Reference Zuo, Ding, Tong, Feng and Qi2015). TiO₂-Mnt has a significantly lower N₂ adsorptive capacity at low relative pressures as compared to Mnt, indicative of a more developed microporous structure. This observation is consistent with the PSD shown in Fig. 1d. As indicated in Table 1, the S _BET and micropore surface area of TiO₂-Mnt increased to 162 and 116 m² g^–1, respectively. These findings also support the hypothesis that the titanium species were incorporated into the clay matrix successfully, developing a hierarchical pore network with micropores and mesopores. For the 0.37Pt–TiO₂/SiO₂ sample, substantial N₂ uptake is observed across the entire relative pressure range. Its S _BET further increased to 214 m² g^–1, with a significant rise in external surface area to 203 m² g^–1, and the pore volume expanded to 0.54 cm³ g^–1. Such a hierarchical pore structure would facilitate diffusion and mass transfer during catalytic reactions.

Table 1.

Textural characteristics of Mnt, TiO₂-Mnt and 0.37Pt–TiO₂/SiO₂.

Table 1 Long description

The table reports textural properties for three samples: Mnt, TiO2-Mnt, and 0.37Pt–TiO2/SiO2, including BET surface area, micropore surface area, external surface area, total pore volume, and micropore volume. Total surface area rises from 63 for Mnt to 162 for TiO2-Mnt and 214 for 0.37Pt–TiO2/SiO2. Micropore surface area is highest for TiO2-Mnt at 116, compared with 26 for Mnt and 11 for 0.37Pt–TiO2/SiO2. External surface area is largest for 0.37Pt–TiO2/SiO2 at 203, versus 37 for Mnt and 46 for TiO2-Mnt, indicating its surface area is dominated by non-microporous surface. Total pore volume is greatest for 0.37Pt–TiO2/SiO2 at 0.54, compared with 0.11 for Mnt and 0.15 for TiO2-Mnt. Micropore volume is highest for TiO2-Mnt at 0.05, lower for Mnt at 0.01, and very small for 0.37Pt–TiO2/SiO2 at 0.001. Values are derived from a t-plot approach and include an external surface estimate calculated from the difference between total and micropore surface areas.

^a Based on t-plot method.

^b S _external = S _BET – S _micropore.

^c Obtained at a relative pressure of 0.97.

Figure 2a–d shows SEM images of Mnt and TiO₂/SiO₂. Mnt exhibited a typical layered structure, whereas the TiO₂/SiO₂ obtained after pillaring and acid treatment displayed a characteristic loose and porous morphology. This structural evolution is attributed to the synergistic effect of Ti pillaring and the acid etching process: the former constructed a porous framework, while the latter dissolved the aluminium oxide octahedral sheets on the Mnt surface, leading to the in situ generation of amorphous silica nanoparticles and the further formation of a rich porous structure.

Figure 2.

SEM images of (a,b) Mnt and (c,d) TiO₂/SiO₂; (e–g) TEM images and (h) SAED image of 0.37Pt–TiO₂/SiO₂; and (i) EDS mapping of 0.37Pt–TiO₂/SiO₂.

Figure 2 Long description

The image A showing a grayscale micrograph with a rough, cracked surface texture. A scale bar at the lower right reads 1 micrometer. The image B showing a grayscale micrograph with folded, ridge-like surface texture. A scale bar at the lower right reads 500 nanometers. The image C showing a grayscale micrograph with granular surface texture. A scale bar at the lower right reads 1 micrometer. The image D showing a grayscale micrograph with granular surface texture and scattered brighter spots. A scale bar at the lower right reads 500 nanometers. The image E showing a grayscale micrograph of an irregular particle cluster on a dark background. Three regions are marked with dashed outlines and labeled (i), (ii) and (iii). A scale bar at the lower right reads 500 nanometers. The image F showing a grayscale lattice image with a dashed yellow rectangle and a dashed red rectangle. Text inside the image includes “Pt(111) 0.23 nm” near the lower left and “TiO2(101) 0.35 nm” near the lower center. A scale bar at the lower right reads 20 nanometers. The image G showing a grayscale lattice image with two large, rounded darker areas near the top and a dashed blue rectangle. Text includes “TiO2(001)” near the top and “Pt(111)” near the lower right. A scale bar at the lower right reads 20 nanometers. The image H showing a circular diffraction pattern on a dark background with multiple labeled spots and rings. Labels include “(004)”, “(101)”, “(200)”, “(105)” and “Pt(111)”. A scale bar at the lower right reads 5 nanometers to the minus one. The image I showing a grayscale micrograph at left with a bright irregular region on a dark background and a dashed outline labeled (iii). A scale bar reads 50 nanometers. To the right are five separate color maps on black backgrounds, each with a scale bar reading 50 nanometers: a green map labeled “O”, a cyan map labeled “Si”, a blue map labeled “Ti” and a yellow map labeled “Pt”.

The composition and morphology of 0.37Pt–TiO₂/SiO₂ were examined in more detail using TEM images (Fig. 2e–h). The selected area in Fig. 2e was magnified for closer observation (Fig. 2f,g). The (101) crystal planes of anatase TiO₂ were clearly observed, with lattice spacings of 3.52 Å. However, lattice fringes at 2.23 Å were found and attributed to the (111) plane of face-centred cubic (fcc) Pt, which meant that the Pt nanoparticles were effectively immobilized on the TiO₂/SiO₂ surface. These TEM data confirmed the coexistence of TiO₂ and Pt nanoparticles within one microregion and their wide dispersion. Notably, overlapping regions of TiO₂ and Pt lattice fringes were observed in Fig. 2g; such intimate interfacial contact provides a structural basis for potential electron transfer between TiO₂ and Pt. The electron diffraction pattern (selected area electron diffraction (SAED); Fig. 2h) showed distinct diffraction rings associated only with the (101), (004) and (220) planes of anatase TiO₂ and the (111) plane of fcc Pt. Additionally, energy-dispersive X-ray spectroscopy (EDX) elemental mapping (Fig. 2i) showed equal distributions of Ti, Pt and Si elements over the support surface, which indicated that the TiO₂ and Pt components were well dispersed.

XPS analysis

The XPS spectra of the catalysts with Pt loadings of 0.17%, 0.37% and 0.69% are shown in Fig. 3a, and the corresponding distribution of surface Pt valence states is summarized in Table 2. The Pt 4f core-level spectra were deconvoluted into metallic Pt⁰ and oxidized Pt²⁺ species. Specifically, the characteristic peaks located at ∼71.6 and ∼74.6 eV were attributed to Pt⁰, whereas the peaks at ∼72.7 and ∼75.6 eV were assigned to Pt²⁺. The ratio of surface Pt⁰/(Pt⁰ + Pt²⁺) showed an increasing trend with Pt loading; notably, the 0.37Pt–TiO₂/SiO₂ catalyst displayed the highest fraction of Pt⁰ species. Generally, metallic Pt⁰ is considered the more active phase for catalytic oxidation reactions.

Figure 3.

XPS spectra of (a) Pt 4f, (b) Ti 2p and (c) O 1s of xPt–TiO₂/SiO₂ (x = 0.17, 0.37 or 0.69).

Figure 3 Long description

The image A showing a plot labeled “Pt 4f” with y-axis label “Intensity (a.u.)” and x-axis label “Binding energy (eV)”. The x-axis runs from 80 to 68 with tick labels at 80, 78, 76, 74, 72, 70 and 68. Three stacked spectra are shown, each with a main line trace and multiple fitted peak components. The top spectrum is labeled “0.17Pt–TiO2/SiO2” and has labeled peak positions “74.99 eV” and “71.88 eV”. The middle spectrum is labeled “0.37Pt–TiO2/SiO2” and has labeled peak positions “74.89 eV” and “71.73 eV”. The bottom spectrum is labeled “0.69Pt–TiO2/SiO2” and has labeled peak positions “74.74 eV” and “71.53 eV”. The image B showing a plot labeled “Ti 2p” with y-axis label “Intensity (a.u.)” and x-axis label “Binding energy (eV)”. The x-axis runs from 470 to 450 with tick labels at 470, 465, 460, 455 and 450. Three stacked spectra are shown, each with a main line trace and fitted peak components. The top spectrum is labeled “0.17Pt–TiO2/SiO2” and has labeled peak positions “463.65 eV” and “457.91 eV”. The middle spectrum is labeled “0.37Pt–TiO2/SiO2” and has labeled peak positions “463.81 eV” and “458.03 eV”. The bottom spectrum is labeled “0.69Pt–TiO2/SiO2” and has labeled peak positions “463.96 eV” and “458.09 eV”. The image C showing a plot labeled “O 1s” with y-axis label “Intensity (a.u.)” and x-axis label “Binding energy (eV)”. The x-axis runs from 536 to 528 with tick labels at 536, 534, 532, 530 and 528. Three stacked spectra are shown, each with a main line trace and fitted peak components. The top spectrum is labeled “0.17Pt–TiO2/SiO2” and includes labeled peak positions “532.28 eV” and “532.12 eV”. The middle spectrum is labeled “0.37Pt–TiO2/SiO2” and includes a labeled peak position “532.12 eV”. The bottom spectrum is labeled “0.69Pt–TiO2/SiO2” and includes a labeled peak position “531.94 eV”.

Table 2.

XPS result of xPt–TiO₂/SiO₂ (x = 0.17, 0.37 or 0.69).

Table 2 Long description

The table reports XPS binding energies for platinum zero 4f seven halves and 4f five halves peaks, the percent of platinum present as metallic platinum, and titanium four plus 2p one half and 2p three halves peaks for three Pt–TiO2/SiO2 samples. For 0.17Pt–TiO2/SiO2, Pt zero 4f seven halves is 71.88 eV and 4f five halves is 74.99 eV, with 66.34 percent Pt zero; Ti four plus 2p one half is 463.66 eV and 2p three halves is 457.92 eV. For 0.37Pt–TiO2/SiO2, Pt zero peaks are 71.73 and 74.89 eV and Pt zero is 66.80 percent; Ti peaks are 463.80 and 458.02 eV. For 0.69Pt–TiO2/SiO2, Pt zero peaks are 71.53 and 74.74 eV and Pt zero is 66.71 percent; Ti peaks are 463.88 and 458.07 eV. As Pt loading increases, both Pt zero binding energies decrease slightly, while the Ti four plus binding energies change only minimally. The metallic Pt fraction remains essentially stable around two thirds across all samples. Small shifts are within the range where instrument calibration and fitting choices can influence reported peak positions, so differences should be interpreted cautiously.

As the Pt loading increases from 0.17 to 0.69 wt.%, the Pt 4f binding energy in the XPS spectra exhibits a distinct negative shift, indicating that Pt species acquire electrons via interfacial charge transfer. Correspondingly, the Ti 2p binding energy shows a positive shift, confirming the electron loss from the Ti⁴⁺ species (Fig. 3b). These findings substantiate the directional electron transfer from Ti to Pt, which elevates the electron cloud density on the Pt surface, maintaining it in an electron-rich state. This MSI not only facilitates the adsorption and activation of gas-phase O₂ into highly reactive nucleophilic oxygen species but also induces the generation of abundant oxygen vacancies (O_v) at the Pt–TiO₂ interface (Xiao et al., Reference Xiao, Yu, Guo and Ge2022; Zhang et al., Reference Zhang, Li, Sun, Chen, Zhu and Zhao2022).

As shown in Fig. 3c, deconvolution of the O 1s spectra reveals three characteristic peaks at ∼529.5, 532.1 and 533.4 eV, assigned to lattice oxygen (O_latt), adsorbed oxygen species (O_ads) and hydrated oxygen (O_wat), respectively (Zhang et al., Reference Zhang, Chen, Fu, Liu, Xie and Geng2025a). Note that the O 1s peaks in 0.17Pt– and 0.37Pt–TiO₂/SiO₂ shift towards lower binding energies, implying the formation of O_v. XPS characterization reveals that toluene oxidation over 0.37Pt–TiO₂/SiO₂ probably proceeds via a synergistic mechanism combining Langmuir–Hinshelwood (L–H) and Mars–van Krevelen (MvK) pathways. Specifically, the Pt sites serve as ‘electronic pumps’ that promote the activation of gas-phase O₂ into highly reactive O_ads species (O^– or O₂^–) via electron transfer. Simultaneously, the Pt–TiO₂ interface acts as the catalytic site where the mobility of O_latt is significantly boosted. O_latt participates in the oxidation reaction through the MvK mechanism. The transfer of electrons from Ti to Pt reduces the formation energy of oxygen vacancies, thereby accelerating the redox process (Zhang et al., Reference Zhang, Ruan, Gao, Dai, Wei and Li2025b). This enhanced redox cycle at the interface, characterized by efficient interfacial electron transfer and O_v turnover, is the fundamental driver of the superior catalytic performance of 0.37Pt–TiO₂/SiO₂ (Sun et al., Reference Sun, Fan, Cheng, Mo, Ke and Ren2021; Jiang et al., Reference Jiang, Zou, Li, Zhao, Zhao, Liu and Zhu2026; Ma et al., Reference Ma, Zhang, Shen, He, Fang, Xie and Gong2026).

Catalytic performance

The toluene oxidation performance of xPt–TiO₂/SiO₂, TiO₂/SiO₂ and Pt–SiO₂ catalysts is presented in Fig. 4a, with the corresponding T ₁₀, T ₅₀ and T ₉₀ values (temperatures required for 10%, 50% and 90% CO₂ yields, respectively) listed in Fig. 4b & Table 3. Pt–SiO₂ and TiO₂/SiO₂, synthesized under identical conditions, served as contrastive samples. According to the T ₉₀ values, the catalysts exhibit a clear activity hierarchy, decreasing in the order of 0.37Pt–TiO₂/SiO₂ (190℃), 0.69Pt–TiO₂/SiO₂ (189℃), Pt–SiO₂ (194℃), 0.17Pt–TiO₂/SiO₂ (219℃) and TiO₂/SiO₂ (280℃). The inferior performance of 0.17Pt–TiO₂/SiO₂ is attributed to a deficiency in active Pt sites, while TiO₂/SiO₂ showed poor activity due to the limited intrinsic oxidative ability of TiO₂. Specifically, the Pt–SiO₂ catalyst exhibited a T ₉₀ of 194℃ for toluene oxidation. The marked improvement in performance following the introduction of TiO₂ underscores its critical contribution to the reaction. This suggests that the interplay between Pt and TiO₂ facilitates a more efficient catalytic pathway. Furthermore, a control sample prepared by physically mixing TiO₂ and SiO₂ followed by Pt loading (Pt–MTiO₂/SiO₂) showed a significantly T ₉₀ of 206℃ (Fig. S3). Notably, the Pt loading from 0.37 to 0.69 wt.% only reduced T ₉₀ from 190°C to 189°C, indicating a limited improvement in performance. Based on careful consideration of economic feasibility and catalytic efficiency, the activity of 0.37Pt–TiO₂/SiO₂ is superior to that of all 0.69Pt–TiO₂/SiO₂, Pt–SiO₂ and Pt–MTiO₂/SiO₂. This result suggests a trade-off effect: although an increased Pt loading can promote the establishment of MSIs and facilitate favourable electron transfer from Ti to Pt, excessively high loadings inevitably induce Pt nanoparticle agglomeration, thereby diminishing the density of accessible Pt active sites. To further evaluate the industrial application potential of the Pt–TiO₂/SiO₂ catalyst, the toluene oxidation performance of typical supported Pt-based catalysts reported in the recent literature is summarized in Table S2. Among the catalysts prepared by impregnation methods, 0.37Pt–TiO₂/SiO₂ achieves a lower T ₉₀ (190℃) under a higher toluene concentration while using a lower Pt loading, demonstrating notable performance advantages.

Figure 4.

(a) CO₂ generation curves of toluene oxidation over xPt–TiO₂/SiO₂ (x = 0.17, 0.37 or 0.69), Pt–SiO₂ and TiO₂. (b) T ₅₀ and T ₉₀ of xPt–TiO₂/SiO₂ (x = 0.17, 0.37 or 0.69), Pt–SiO₂ and TiO₂. (c) Effect of different water vapour regimes on toluene oxidation for 0.37Pt–TiO₂/SiO₂. (d) Catalytic oxidation of 0.37Pt–TiO₂/SiO₂ over three cycles. (e) Catalyst stability during 48 h of continuous testing at 190℃ of 0.37Pt–TiO₂/SiO₂ under the conditions of toluene concentration = 1000 ppm, total flow rate = 80 mL min^–1 and gas hourly space velocity = 60,000 cm³ g^–1 h^–1.

Figure 4 Long description

The image contains multiple graphs comparing CO2 generation during toluene oxidation. The first graph (A) is a line graph showing CO2 generation (%) on the y-axis and temperature (°C) on the x-axis, ranging from 120 to 300°C. It compares catalysts: 0.17Pt-TiO2/SiO2, 0.37Pt-TiO2/SiO2, 0.69Pt-TiO2/SiO2, TiO2/SiO2 and Pt-SiO2. The 0.37Pt-TiO2/SiO2 catalyst reaches high CO2 generation at the lowest temperature, indicating superior performance. The second graph (B) is a bar chart showing temperature (°C) on the y-axis and catalyst types on the x-axis. It includes T50 and T90 values for each catalyst. The 0.37Pt-TiO2/SiO2 catalyst has the lowest T50 and T90 values, demonstrating efficient CO2 generation. The third graph (C) is a line graph showing CO2 generation (%) on the y-axis and time (min) on the x-axis, ranging from 0 to 600 minutes. It examines the effect of different water vapor regimes (H2O-off, H2O-on 3 vol.%, 5 vol.%, 10 vol.%) on CO2 generation, which remains nearly constant across all regimes. The fourth graph (D) is a line graph showing CO2 generation (%) on the y-axis and temperature (°C) on the x-axis, ranging from 140 to 210°C. It tests catalytic oxidation over three cycles, showing consistent CO2 generation across cycles. The fifth graph (E) is a line graph showing CO2 generation (%) on the y-axis and time (h) on the x-axis, ranging from 0 to 48 hours. It tests catalyst stability, with CO2 generation remaining near 90–100 percent throughout the duration. Visual differentiation includes different markers for each catalyst in graph A and different colored bars for T50 and T90 in graph B. Graphs A and B both compare catalysts via light-off behavior and T50/T90 values, while graph C tests water effect over time, graph D tests repeatability across cycles and graph E tests long-term stability.

Table 3.

Catalytic performance of xPt–TiO₂/SiO₂ (x = 0.17, 0.37 or 0.69), Pt–SiO₂ and TiO₂/SiO₂ for toluene oxidation.

Table 3 Long description

The table reports CO2 generation temperatures in degrees Celsius for toluene oxidation, listed as T10, T50, and T90 for each catalyst. Lower temperatures indicate earlier CO2 generation and better low temperature activity. The best overall performer is 0.37Pt–TiO2/SiO2, with T10 175 C, T50 185 C, and T90 190 C. Pt–SiO2 is close, at T10 185 C, T50 191 C, and T90 194 C. 0.69Pt–TiO2/SiO2 is similar to Pt–SiO2, with T10 180 C, T50 187 C, and T90 189 C, while 0.17Pt–TiO2/SiO2 is higher at T10 207 C, T50 217 C, and T90 219 C. TiO2/SiO2 without Pt is the least active, requiring much higher temperatures: T10 240 C, T50 272 C, and T90 280 C. Across all catalysts, T10 is lowest and T90 is highest, and adding Pt generally lowers the required temperatures compared with TiO2/SiO2 alone.

Water vapour is ubiquitous in industrial exhaust and is known to inhibit the catalytic oxidation of VOCs via competitive adsorption on active sites. This phenomenon is particularly observed on hydrophilic supports that may become ‘cloaked’ with water molecules, thereby preventing contact for the reaction. To investigate the effect of water vapour on the catalytic process under actual conditions, the performance of the catalyst in the presence of water was evaluated. To this end, water vapour was introduced into the reaction system at concentrations of 3, 5 and 10 vol.% at 190°C to examine its effect on the 0.37Pt–TiO₂/SiO₂ catalyst (Fig. 4c). The findings indicate that the catalyst shows outstanding toluene oxidation efficiency without evidence of activity decay at water concentrations of 3, 5 and 10 vol.%. This reflects the exceptional water resistance of the catalyst. This robust performance may be attributed to the partial removal of surface hydroxyl groups from the acid-treated TiO₂/SiO₂ support during thermal treatment, which reduces the strong adsorption of water molecules on the active sites (Capel-Sanchez et al., Reference Capel-Sanchez, Barrio, Campos-Martin and Fierro2004).

The reusability of the catalyst was assessed based on three successive reaction cycles. As shown in Fig. 4d, the 0.37Pt–TiO₂/SiO₂ catalyst maintained a high toluene conversion rate after three cycles without apparent activity decay, indicating excellent cyclic stability and reusability. Additionally, a long-term on-stream stability test was conducted on the 0.37Pt–TiO₂/SiO₂ catalyst at 190°C for 48 h (Fig. 4e). The toluene conversion remained stable throughout the test period with no significant degradation, suggesting superior long-term stability and resistance to sintering. This stability can probably be ascribed to the high dispersion of Pt nanoparticles and the strong MSIs between Pt and the TiO₂/SiO₂ support, which effectively inhibit the migration and agglomeration of Pt particles during the reaction.

Investigation of the catalytic mechanism by DRIFTS

To elucidate the dynamic evolution of reaction intermediates and reveal the specific catalytic pathway of the 0.37Pt–TiO₂/SiO₂ catalyst, in situ DRIFTS measurements were conducted. Before the experiment, the catalyst was pretreated in the infrared spectroscopy cell at 300°C under N₂ flow for 1 h to remove adsorbed impurities. Then, a 1000 ppm (by volume) toluene/air mixture was used to treat the pretreated catalyst, and the spectra were measured at various temperatures. The assignment of the corresponding DRIFTS absorbance bands is listed in Table S2.

As displayed in Fig. 5 & Table S3, the peaks located at 2877, 2924 and 3029 cm^–1 are assigned to the symmetric and asymmetric C–H stretching vibrations of the methyl group in toluene molecules (Sun et al., Reference Sun, Liu, Chen and Quan2015), implying that the toluene molecules are adsorbed onto the surface of the catalyst. The two peaks at 1495 and 1603 cm^–1 are associated the skeletal vibrations of the aromatic ring (Einaga & Futamura, Reference Einaga and Futamura2006; Liu et al., Reference Liu, Zeng, Shi, Wang, Zhu and Chen2017); the peak intensities tended to decrease gradually with the progress of the reaction, which suggested a step-by-step cleavage of the aromatic ring structure. Two characteristic peaks at 1016 and 1123 cm^–1 can be ascribed to alkoxide species (PhCH₂OH) formed by the cleavage of the methyl C–H bonds of adsorbed toluene followed by oxidation to generate benzyl alcohol (Keresszegi et al., Reference Keresszegi, Ferri, Mallat and Baiker2005; Augugliaro et al., Reference Augugliaro, Kisch, Loddo, López-Muñoz, Márquez-Álvarez and Palmisano2008). Additionally, characteristic bands of benzaldehyde species were detected at 1180, 1448 and 1589 cm^–1 (Busca, Reference Busca1996). C–O stretching vibrations assigned to benzoates were observed at 1396 and 1520 cm^–1, which indicated the presence of intermediates similar to benzoic acid (Zhao et al., Reference Zhao, Li, Jiang and Li2016). Furthermore, an absorption band for maleic anhydride was observed at 1306 cm^–1 (Zhong et al., Reference Zhong, Zeng, Zhang, Feng, Xiao and Wu2020).

Figure 5.

In situ FTIR spectra of 0.37Pt-TiO₂/SiO₂ in the ranges of (a) 1000–1700 cm^–1 and (b) 1700–3100 cm^–1 at various temperatures.

Figure 5 Long description

The image A showing FTIR spectra of 0.37Pt-TiO/SiO. Six overlaid line spectra are shown for 80 degrees Celsius, 120 degrees Celsius, 160 degrees Celsius, 180 degrees Celsius, 200 degrees Celsius and 220 degrees Celsius. The x-axis label is Wavenumber (centimeter superscript minus 1), with tick labels from 1000 to 1700. The y-axis label is Absorbance (a.u.). Labeled peak positions on the spectra are 1016, 1123, 1180, 1306, 1396, 1448, 1495, 1520, 1589, 1603 and 1700. The spectra show multiple peaks in the 1300 to 1603 range, with the tallest labeled peak at 1306. The image B showing FTIR spectra of 0.37Pt-TiO/SiO. Six overlaid line spectra are shown for 80 degrees Celsius, 120 degrees Celsius, 160 degrees Celsius, 180 degrees Celsius, 200 degrees Celsius and 220 degrees Celsius. The x-axis label is Wavenumber (centimeter superscript minus 1) and includes an axis break, with one section spanning 1700 to 1800 (tick labels at 1700 and 1800) and another section spanning 2900 to 3000 (tick labels at 2900 and 3000). The y-axis label is Absorbance (a.u.). Labeled peak positions are 1716 and 1733 in the 1700 to 1800 section and 2877, 2924 and 3029 in the 2900 to 3000 section. Text near the 2900 to 3000 section reads minus C H subscript 3 group. The spectra show three labeled peaks in the 2877 to 3029 range and two labeled peaks in the 1716 to 1733 range.

As discussed in the detailed spectral analysis earlier, it is possible to develop a full mechanistic description of an oxidation reaction of toluene using a Pt–TiO₂/SiO₂ catalyst. The reaction initiates with the chemisorption of toluene, followed by the sequential oxidation of the methyl group to form benzyl alcohol and benzaldehyde. These intermediates are quickly transformed into surface benzoates, and they build up to become the major surface intermediates. Next, reactive oxygen species attack the aromatic ring of the benzoate, which results in ring-opening and the production of maleic anhydride. Ultimately, all such intermediates are fully oxidized to CO₂ and H₂O.