1 Introduction

Natural gas (> 90% CH4 content) is an attractive energy source due to its abundance and high energy density (high H/C ratio). However, the release of unburned CH4 during flame combustion can pose serious environmental problems because of the strong greenhouse effect of CH4 that is twenty times greater than that of CO2 at the same molar amount [1]. The catalytic combustion of methane is regarded as the most promising method for reducing the emissions of CH4 in industrial processes at relatively low temperatures (~ 500 °C), with the advantages of high combustion efficiency and low emission of other toxic products, such as NOx [2]. Nevertheless, this method requires the development and screening of high-performance catalysts.

Noble metal catalysts (such as Pt, Pd and Rh) are widely used for methane combustion due to their high catalytic activity at low temperatures [3–7]. These catalytic activities are mainly dependent on the dispersion and electronic structure of the noble metal and the properties of the support. Although encouraging progress has been made in improving the catalytic performance of CH4 oxidation, the reliability of CH4 oxidation at low temperatures is still unclear. Recently, bimetallic catalysts have been exploited in CH4 oxidation because of their unique synergistic and electronic effects, especially Pd–Pt catalysts. Chetyrin et al. [8] reported that during CH4–O2 catalysis, single Pt atoms and small Pt clusters dispersed on an Al2O3 support are largely inactive. In contrast, large Pt–Pd alloy clusters (> 5 nm) with a thin PdO shell covering a Pt-rich core dispersed on an Al2O3 support are highly reactive. On these cluster surfaces, the O2− anions are highly nucleophilic, whereas the Pd2+ cations are highly electrophilic because they are in contact with the underlying Pt-rich core. These form Pd2+–O2− site pairs that catalyze the kinetically relevant C–H bond cleavage of CH4 via the formation of a highly stabilized four-center transition state (H3Cδ−–Pd2+–Hδ+–O2−) more effectively than monometallic O*-covered Pt or PdO clusters [9, 10]. In contrast, Ru-based catalysts are utilized less for CH4 combustion, mainly because of their poor high-temperature stability [11]. However, Ru–Pt bimetallic catalysts have been studied in the field of C1 catalysis, particularly for methanol electrooxidation reactions (MOR). For example, porous carbon-supported Ru–Pt catalysts prepared via strong electrostatic adsorption and electroless deposition are more active than pure Pt [12]. Moreover, the interaction between Ru and Pt can enhance methanol conversion [13]. There could be a “bifunctional mechanism” in Ru–Pt bimetallic catalysts. In particular, the C–H bond of methanol could be cleaved and CO adsorbed at Pt sites; however, Ru atoms can provide more O species to consume the intermediate poisoning on Pt sites [14, 15]. Although Ru–Pt bimetallic catalysts have been widely used for MOR, limited reports exist on Ru–Pt bimetallic catalysts for CH4 combustion. However, the properties of support materials generally affect the activity of the supported catalysts. TiO2 has been widely used as catalyst support, and the crystalline phase of rutile-type TiO2 matches that of RuO2, enhancing the high dispersion and stability of the RuO2 active phase [16, 17]. Herein, a bimetallic RuPt/TiO2 catalyst precursor was synthesized by an in situ reduction method [14], and then the precursor was oxidized to obtain a catalyst (RuPt-O/TiO2) with a stable structure. The various characterizations demonstrated that both PtOx particles (4 nm) and twinned noble metal structures were present on the RuPt-O/TiO2 catalyst, that is, RuO2 was epitaxially grown on the TiO2 substrate and PtOx adhered to the RuO2 surface. As a result, the Pt species in the bimetallic RuPt-O/TiO2 catalyst consisted of Pt2+ and Pt4+, while Pt0 was present in the Pt-O/TiO2 catalyst. The synergistic effect of RuO2 and PtOx species afforded superior CH4 oxidation performance on RuPt-O/TiO2 catalyst. PtOx can prompt CH4 activation and oxidation, and RuO2 species can provide additional oxygen species to PtOx to furnish the redox cycle of PtOx species. Consequently, the performance of RuPt-O/TiO2 catalyst in CH4 oxidation was significantly enhanced compared with that of the corresponding monometallic catalysts.

2 Experimental

2.1 Catalysts preparation

The rutile-type TiO2 purchased from Shanghai Macklin Biochemical Co., Ltd. was used as support. The bimetallic catalyst was synthesized by the produces described in Ref. [14]. First, a certain amount of TiO2 was added into 20 ml H2O and then ultrasonicated for 30 min. Second, a certain amount of H2PtCl6·6H2O and RuCl3 were added into the suspension liquid and ultrasonicated for 30 min again. Then, the suspension liquid was placed in a water bath of 50 °C and kept continuously stirring until the solvents were evaporated. The sample was dried at 60 °C overnight in a vacuum oven, and then was reduced under 10 vol% H2/Ar at 300 °C for 3 h. Afterward, the as-prepared sample was washed and centrifuged repeatedly with deionized water to wash away residual ions and then dried at 60 °C overnight in a vacuum oven. Finally, the sample was reduced to under 10 vol% H2/Ar at 700 °C for 3 h, which was labeled as RuPt/TiO2.

On this basis, the RuPt/TiO2 was calcined at 500 °C in air for 1 h, which was labeled as RuPt-O/TiO2. The monometallic Ru-O/TiO2 and Pt-O/TiO2 catalysts were synthesized by the same produces. The theoretical total metal loading of all catalysts was 1 wt%, and the atomic ratio of Ru/Pt in the bimetallic catalyst was 1.

2.2 Catalysts characterization

The loading of metal was detected by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Agilent 725ES). X-ray diffraction (XRD) patterns were detected on a Bruker D8 Focus diffractometer with Cu Kα radiation (λ = 0.154056 nm, 40 kV and 40 mA), scanning from 10° to 80° at a speed of 6 (°)·min−1. The surface areas of catalysts were obtained on the Quantachrome Nova Touch LX3 instrument at − 196 °C and were calculated by the Brunauer–Emmett–Teller (BET) method. Laser Raman spectra of catalysts were obtained on a Renishaw Raman spectrometer at ambient condition, and the 532 nm line of a spectra physics Ar+ laser was used as an excitation. Electron paramagnetic resonance (EPR) measurements were performed on a Bruker EMX-8/2.7 EPR Spectrometer. Aberration-corrected scanning transmission electron microscopy (AC-STEM) characterization was performed using a ThermoFisher Themis Z electron microscope. High-angle annular dark-field (HAADF)-STEM images were recorded using a convergence semi-angle of 0.6303°, and inner- and outer collection angles of 3.381° and 11.46°, respectively. Energy-dispersive X-ray spectroscopy (EDS) was carried out using 4-in-column Super-X detectors. X-ray photoelectron spectra (XPS) of all the catalysts were detected on a Thermo ESCALAB 250Xi spectrometer equipped with Al Kα (1486.6 eV) radiation as the excitation source. All binding energies (BE) were determined with respect to the C 1s line (284.8 eV) originating from adventitious carbon.

In situ diffuse reflectance and infrared Fourier transform spectra (DRIFTS) of CH4 adsorption on the catalysts were measured on a Nicolet Nexus 6700 spectrometer with an MCT detector, and the sample cell was equipped with ZnSe windows. DRIFT spectra were collected with a resolution of 4 cm−1 and 64 scans in Kubelka–Munk units. The catalysts were pretreated at 500 °C for 1 h in Ar, and then were exposed to the feed gas consisting of 20 vol% CH4/N2 (0.5 ml·min−1) balanced by Ar (49.5 ml·min−1) at 350 °C. DRIFT spectra were recorded for 0, 1, 3, 5, 10 and 20 min under the continued feed gas.

CH4/O2 pulse experiments were carried out on a Micromecitics Auto II 2920 instrument with a mass spectrometer (HPR-20 QIC). The specific test process was similar to that described in Ref. [3]. The catalysts (50 mg) were pretreated in a 3 vol% O2/He with a flow rate of 40 ml·min−1 at 400 °C for 1 h. After cooling down to 350 °C, the catalyst was purged with He at 350 °C for 30 min. Then a stream of 5 vol% CH4/Ar was injected into the catalyst for ten pulses with a precise analytical syringe (loop volume 0.5173 ml) until the baseline remained stable. CH4 (mass-to-charge ratio of m/e = 15), CO2 (m/e = 44), CO (m/e = 29), H2 (m/e = 2) and O2 (m/e = 32) were continuously recorded by a mass spectrometer. After ten pulses were injected, the catalysts were purged with He for 20 min, and then a stream of 3 vol% O2/He was injected into the catalyst for ten pulses. Similarly, the catalysts were purged with He for 20 min. The above is a cycle. Therefore, CH4/O2 pulse measurements were conducted by repeatedly dosing 5 vol% CH4/Ar or 3 vol% O2/He into the catalyst bed at 350 °C.

2.3 Catalytic performances measurements

The catalytic activity of all the catalysts for CH4 combustion was tested in a fixed bed quartz reactor at atmospheric pressure. 200 mg catalysts (40–60 mesh) was used. The feed gas consisted of 1 vol% CH4, 20 vol% O2 and N2 balanced with a flow rate of 50 ml·min−1, and the gas hourly space velocity (GHSV) was 15,000 ml·g−1·h−1. The catalyst bed was heated from 150 to 500 °C at a rate of 3 °C·min−1, and the CH4 concentrations were detected by an online gas chromatograph (Agilent GC 7890A) equipped with thermal conductivity detectors (TCD). The selectivity of CO2 was close to 100% over the catalysts, and the other possible products were under the detective limit of the GC.

The reaction rate of the catalysts in CH4 oxidation was also measured in the feed gas consisted of 1 vol% CH4, 20 vol% O2, and N2 balanced. The GHSV was 15,000 and 60,000 ml·g−1·h−1 for the monometallic and bimetallic catalysts, respectively, to achieve the conversion of CH4 lower than 5%–15%. The reaction rates (\({r}_{{\mathrm{CH}}_{4}}\)) were calculated by the following equation:

$${r}_{{\mathrm{CH}}_{4}}=\frac{{C}_{{\mathrm{CH}}_{4}}{X}_{{\mathrm{CH}}_{4}}V{P}_{\mathrm{atm}}}{{m}_{\text{cat}}{w}_{\text{M}}RT}$$
(1)

where \({C}_{{\mathrm{CH}}_{4}}\) is the concentration of CH4 in the feed gas; \({X}_{{\mathrm{CH}}_{4}}\) is the conversion of CH4; mcat is the mass of the catalyst used; wM is the loading of metal; Patm is the atmosphere pressure, equaled to 101.3 kPa; R is the molar gas constant, equaled to 8.314 Pa m3·mol−1·K−1; T is the room temperature, equaled to 25 °C.

3 Results and discussion

3.1 Catalyst structure

The structures of the RuPt/TiO2 catalysts were first investigated. Figure 1 shows the aberration-corrected HAADF–STEM images and corresponding EDS elemental maps of the RuPt/TiO2 sample reduced at 700 °C. Two types of particles of distinct sizes were identified in the catalyst (Fig. 1a). The sizes of most small particles were ~ 4.0 nm, and the d-spacing of {111} was 0.243 nm, as shown in Fig. 1a, b. Compared with the standard lattice parameter of Pt, it can be determined that this phase should be Pt, which is further demonstrated by EDS result in Fig. 1f, g. In contrast, the size of large particles was ~ 10.0 nm, and the structure is face-centered cubic (fcc) with the d-spacing around 0.216 nm for {111} planes, e.g., the measured values of 0.216, 0.217 and 0.214 nm in Fig. 1c–e. Based on the contracted lattice constant and the uniform distribution of Pt and Ru in the EDS maps (Fig. 1f–i), it can be concluded that the large nanoparticles are Pt–Ru alloys. Owing to a large difference in the reduction potential and crystal lattice between Pt and Ru [18, 19], Ru–Pt alloys are difficult to obtain through facile synthesis.

Fig. 1
figure 1

a HAADF–STEM image; b–e enlarged images of B–E regions marked by yellow boxes in a; f–i EDS elemental maps of RuPt/TiO2 sample corresponding to region shown in a

Full size image

Figure 2 shows HAADF–STEM images and corresponding EDS elemental maps (Fig. 2b–e) of RuPt-O/TiO2 catalyst oxidized in air at 500 °C. Small Pt particles still existed and the size did not change significantly from ~ 4.0 nm (Fig. 2a). However, the size and structure of the Ru–Pt bimetallic nanoparticles (NPs) changed after calcination in air for 1 h. Ru was oxidized to RuO2 and epitaxially grown on TiO2 substrate due to lattice matching between them, and PtOx adheres to RuO2 surface. A fast Fourier transform (FFT) image of the TiO2 matrix is shown in the inset of Fig. 2a, demonstrating the tetragonal structure of TiO2 crystal along [111] zone axis. Compared with the standard structure of RuO2, it can be determined that the region marked by the red circle in Fig. 2f was a single crystal of RuO2 with a tetragonal structure in the [111] direction (the FFT of this region is shown in Fig. 2g), whose orientation and lattice parameters were almost the same as those of the TiO2 matrix. Notably, EDS elemental maps showed phase separation of Pt and Ru in the RuPt-O/TiO2 catalyst. The atomic-scale HAADF–STEM image also shows a varied Z-contrast, indicating the different regions of PtOx and RuO2, which was different from that of Pt–Ru alloy. This was also verified by the measured lattice distances in Fig. 2f, which showed Pt-rich region (0.358 nm) and RuO2 region (0.349 nm).

Fig. 2
figure 2

a, f HAADF–STEM images, b–e corresponding EDS elemental maps and g FFT patterns of RuPt-O/TiO2 catalyst

Full size image

Figures S1 and S2 show HAADF–STEM images and corresponding EDS elemental maps of monometallic Ru-O/TiO2 and Pt-O/TiO2, respectively. The particle size of Ru and Pt was ~ 11.3 and 5.2 nm with a broad size distribution, respectively. The size of the Pt NPs was similar to that of the Pt in RuPt-O/TiO2 catalyst.

Briefly, both Pt and bimetallic Ru–Pt NPs were present in the bimetallic RuPt-O/TiO2 catalyst after calcination at 500 °C for 1 h in air. Although the relative proportions of both NPs could not be identified, the special structure of the bimetallic RuPt-O/TiO2 catalyst could provide unique catalytic activity compared with the monometallic Pt and Ru catalysts [20–22].

XRD patterns of Ru-O/TiO2, Pt-O/TiO2 and RuPt-O/TiO2 catalysts are shown in Fig. 3. All catalysts exhibited diffraction peaks corresponding to rutile-type TiO2 (JCPDS No. 21-1276) [23]. In addition, XRD pattern of Ru-O/TiO2 catalyst showed a diffraction peak at 2θ = 35.1°, which was assigned to the (101) crystalline planes of RuO2 (JCPDS No. 70-2662). In contrast, XRD pattern of Pt-O/TiO2 catalyst showed no diffraction peaks assigned to Pt and PtOx due to its small particle size (5.2 nm), as observed by transmission electron microscopy (TEM). No diffraction peaks attributed to Pt and Ru were observed in XRD pattern of the bimetallic RuPt-O/TiO2 catalyst due to the relatively small particle sizes of both PtOx and Ru–Pt bimetallic NPs (Fig. 2).

Fig. 3
figure 3

XRD patterns and enlarged view of the corresponding pattern of TiO2, Ru-O/TiO2, Pt-O/TiO2 and RuPt-O/TiO2 catalysts

Full size image

Figure 4a shows Raman spectra of Ru-O/TiO2, Pt-O/TiO2 and RuPt-O/TiO2 catalysts. The vibration bands of the rutile-type TiO2 were located at 136, 229, 442 and 605 cm−1, which correspond to the classical B1g, two-photon scattering process, Eg (planar O–O vibration), and A1g (Ti–O stretch) Raman modes, respectively [24]. All catalysts exhibited similar spectra and an obvious red-shift of the Eg mode compared to the support TiO2, indicating that a large number of defects were present in TiO2 and decreased after NP loading. Moreover, the amounts of defects in Ru-O/TiO2 and RuPt-O/TiO2 catalysts were lower than those in Pt-O/TiO2 catalyst, due to the epitaxial growth of RuO2 on the TiO2 substrate that consumed more surface defects. This conclusion was further confirmed by EPR spectra. As shown in Fig. 4b, all the catalysts and the support TiO2 showed a signal at g = 2.006 (g is a constant of proportionality, whose value is the property of the electron in a certain environment), which was assigned to surface oxygen vacancies [25]. The signal intensity for the support TiO2 was stronger than that for the catalysts, indicating that a large number of surface oxygen vacancies were present on the support TiO2, and they were consumed by NP loading to a certain degree. Moreover, the decrease was more pronounced for Ru-O/TiO2 and RuPt-O/TiO2 catalysts, which is consistent with Raman results. In summary, metal NPs tended to be located in oxygen vacancies on the support TiO2, and Ru and Pt–Ru loading consumed more surface defects due to the epitaxial growth of RuO2 on TiO2 substrate.

Fig. 4
figure 4

a Laser Raman and b EPR spectra of Ru-O/TiO2, Pt-O/TiO2 and RuPt-O/TiO2 catalysts

Full size image

3.2 CH4 oxidation performance of catalyst

After determining the catalyst structure, its catalytic performance in methane oxidation was determined. Figure 5 shows the methane conversion as a function of temperature on the support TiO2, monometallic Pt-O/TiO2, Ru-O/TiO2, and bimetallic RuPt-O/TiO2 catalysts. The support TiO2 was inert for methane oxidation. After loading Pt and Ru, the catalytic performance in methane oxidation was significantly enhanced, and methane was completely converted on the Pt-O/TiO2 and Ru-O/TiO2 catalysts at 450 and 490 °C, respectively. The bimetallic RuPt-O/TiO2 catalyst with a constant total loading of 1 wt% of Pt and Ru showed excellent catalytic activity for methane oxidation. The temperature for complete methane conversion declined to 357 °C, which was much lower than that for both Pt-O/TiO2 and Ru-O/TiO2 catalysts. More importantly, after the addition of 5 vol% H2O to the feed gas, the bimetallic RuPt-O/TiO2 catalyst could preserve the activity of methane oxidation, but the Ru-O/TiO2 catalyst showed a low ability to H2O resistance.

Fig. 5
figure 5

Catalytic activity of Ru-O/TiO2, Pt-O/TiO2 and RuPt-O/TiO2 catalysts for CH4 combustion in dry condition (solid line) and wet condition containing 5 vol% H2O (dot line), where feed gas consisted of 1 vol% CH4, 20 vol% O2, 5 vol% H2O (when used), and N2 balance, and GHSV was 15,000 ml·g−1·h−1

Full size image

To further confirm the catalytic activities of the catalysts, the reaction rates were determined, and the results are listed in Table S1. The reaction rate of the bimetallic RuPt-O/TiO2 catalyst was estimated based on the fact that both Ru and Pt species were involved in methane oxidation. The reaction rate for CH4 oxidation was 13.9 × 10–5 over the RuPt-O/TiO2 catalyst at 303 °C, and it was 1.55 × 10–5 and 1.59 × 10–5 \({\mathrm{mol}}_{{\mathrm{CH}}_{4}}^{-1}\cdot {\mathrm{g}}_{(\mathrm{Ru}+\mathrm{Pt})}^{-1}\cdot {\mathrm{s}}^{-1}\) over Ru-O/TiO2 and Pt-O/TiO2 catalysts, respectively, confirming the superior activity of the bimetallic RuPt-O/TiO2 catalyst for methane oxidation compared with the monometallic catalysts.

The stability of the catalysts for CH4 combustion was also evaluated. As shown in Fig. 6a, both Ru-O/TiO2 and RuPt-O/TiO2 catalysts maintained their catalytic activities during the three reaction cycles, indicating reasonable stability. However, the catalytic activity of Pt-O/TiO2 catalyst improved slightly during the three reaction cycles, which can be attributed to the activation process during the first cycle as it is typical for Pt-based catalysts [26]. The stability of the catalysts during long-term oxidation reactions was further evaluated. As shown in Fig. 6b, both Ru-O/TiO2 and RuPt-O/TiO2 catalysts exhibited high stability, and methane conversion was maintained at 16% and 85% within 30 h, respectively. Curiously, the methane conversion on Pt-O/TiO2 catalyst gradually increased from 32% to 55% within 5 h and then remained constant, possibly because of the formation of an active layer of strongly adsorbed oxygen species during the reaction [26], which was represented in three reaction cycles on the Pt-O/TiO2 catalyst.

Fig. 6
figure 6

Stability of Ru-O/TiO2, Pt-O/TiO2, and RuPt-O/TiO2 catalysts during a recycling experiments and b prolonged oxidation at 350 °C, where feed gas consisted of 1 vol% CH4, 20 vol% O2, and N2 balance, and GHSV was 15,000 ml·g−1·h−1

Full size image

3.3 Nature of activity difference between catalysts

XPS spectra were obtained to explore the chemical states of Pt and Ru in the catalyst as crucial parameters for methane oxidation. As shown in Fig. 7a, the Ti 2p spectra of all catalysts consisted of two individual peaks at BE = 464.4 and 458.7 eV, and the peaks of Ti 2p3/2 and Ti 2p1/2 showed a difference of ~ 5.7 eV, which was consistent with the state of Ti4+, indicating the presence of Ti4+ in all catalysts [27, 28]. Ti 2p peaks of Ru-O/TiO2, Pt-O/TiO2, and RuPt-O/TiO2 shifted to a higher binding energy than those of TiO2, indicating that TiO2 possessed a lower valence state, which resulted from more oxygen vacancies available on the support TiO2. The O 1s spectra of all the catalysts can be divided into two peaks (Fig. 7b). The two peaks at 529–530 and 532 eV were assigned to lattice oxygen (Olat) and surface-adsorbed oxygen (Oads), respectively [29–31]. The ratio of Oads/Olat was calculated using the peak area ratio of Oads to Olat. The ratios of the four catalysts were very similar: Pt-O/TiO2 (0.13) ≈ RuPt-O/TiO2 (0.14) ≈ Ru-O/TiO2 (0.15) ≈ TiO2 (0.18).

Fig. 7
figure 7

a Ti 2p, b O 1s, c Pt 4f, and d Ru 3d XPS spectra of Ru-O/TiO2, Pt-O/TiO2, and RuPt-O/TiO2 catalysts

Full size image

Pt 4f XPS spectra of Pt-O/TiO2 and RuPt-O/TiO2 catalysts were deconvoluted (Fig. 7c). XPS spectrum the Pt-O/TiO2 catalyst included three components of Pt0, Pt2+ and Pt4+, and the metallic Pt (Pt0) at BE = 70.9 and 74.2 eV was a bare majority [32]. Pt species in the bimetallic RuPt-O/TiO2 catalyst also consisted of Pt0, Pt2+ and Pt4+. However, the ratio of Pt0/Pt on RuPt-O/TiO2 catalyst (5.1%) was significantly lower than those on Pt-O/TiO2 catalyst (42.3%), as shown in Table S1. The significant difference in Pt valence between Pt-O/TiO2 and RuPt-O/TiO2 catalysts can be attributed to the hybrid bimetallic NP in RuPt-O/TiO2 catalyst. The presence of a RuO2 transition layer between TiO2 and PtOx in RuPt-O/TiO2 catalyst can significantly inhibit the decomposition of PtOx to Pt during the calcination process, leading to a small amount of metallic Pt. Regarding Ru 3d spectrum (Fig. 7d), the bands at ~ BE = 280.5 and 284.6 eV were assigned to Ru4+, and those at ~ BE = 282.2 and 286.3 eV were assigned to the satellite peaks of RuO2 [33–35]. Therefore, both the monometallic Ru-O/TiO2 and bimetallic RuPt-O/TiO2 catalysts exhibited Ru4+, indicating no difference in Ru valence between them. Summarily, there was only a difference in the valence state of Pt among the three catalysts.

Regarding CH4 combustion, the adsorption of CH4 molecules and dissociation of C–H bonds on the surface of precious metal oxides (PtOx and RuOx) is the first rate-controlling step, followed by the gradual oxidation of CH3 by O from precious metal oxides to form CO2 and H2O, accompanied by the reduction of metal oxides, such as PtO to Ptδ+ (0 < δ < 2). Subsequently, the metal oxides are directly re-oxidized to their initial states by both active O atoms on the surface of the support and gas-phase O2 [36–38]. Therefore, CH4 adsorption and activation, as well as the redox cycle of precious metal oxides, were determined by DRIFT spectra of CH4 adsorption and CH4/O2 pulse experiments, respectively.

Figure 8 shows DRIFT spectra of CH4 adsorption on Ru-O/TiO2, Pt-O/TiO2, and RuPt-O/TiO2 catalysts. Regarding the spectra on Ru-O/TiO2 catalyst, the absorption bands assigned to gaseous methane appeared at 1300, 2953, 3019 and 3095 cm−1 after exposure for 1 min. Meanwhile, there were two weak bands at 2330 and 2360 cm−1 assigned to gaseous CO2 [39]. Increasing the exposure time of CH4 to 5 min, the intensity of both weak bands increased significantly, due to CH4 adsorption and oxidation on the catalyst. In addition, the new band at 2009 and 2078 cm−1 appeared for 5 min, assigned to CO linear adsorption (Ru-(CO)) and twin adsorption (Ru-(CO)3) on Ru species [40, 41]. With a further increase in the exposure time to 20 min, the intensity of all bands slightly increased. Overall, CH4 was adsorbed and oxidized on RuO2 sites to produce CO2 on the Ru-O/TiO2 catalyst, accompanied by the reduction of RuO2 to Ruδ+ (2 < δ < 4) under O free conditions. Subsequently, the oxidation ability decreased, and an incomplete oxidation product of CO was generated and adsorbed on Ruδ+ species.

Fig. 8
figure 8

In situ DRIFT spectra of CH4 adsorption over a Ru-O/TiO2, b Pt-O/TiO2, and c RuPt-O/TiO2

Full size image

DRIFT spectra for CH4 adsorption on Pt-O/TiO2 catalysts were significantly different from those on Ru-O/TiO2 catalysts. As shown in Fig. 8b, after exposure for 1 min, there were two intense absorption bands at 3019 and 2069 cm−1, which were assigned to gaseous CH4 and CO linearly adsorbed on metallic Pt, respectively [22]. CO is produced only from the incomplete oxidation of CH4. In addition, two weak absorption bands assigned to gaseous CO2 appeared at 2330 and 2360 cm−1. With an increase in the exposure time of CH4 to 20 min, the intensities of all the bands did not change significantly. Combined with the spectra of Ru-O/TiO2 catalyst, it can be seen that Pt-O/TiO2 catalyst has a higher ability to activate and oxidize CH4. However, a significantly higher amount of CO than CO2 was produced on Pt-O/TiO2 catalyst due to the low valence state of Pt, as shown in XPS results.

As shown in Fig. 8c, for the spectra of RuPt-O/TiO2 catalyst, all bands assigned to gaseous CH4, gaseous CO2, and CO linearly adsorbed on metallic Pt and Ru appeared for an exposure time of 1 min. Compared with the spectra of both monometallic catalysts for an exposure time of 1 min, the amount of CO2 produced increased significantly, accompanied by a marked decrease in the amount of CO linearly adsorbed on metallic Pt. Furthermore, with an increase in the exposure time of CH4 to 20 min, the intensities of these bands increased slightly. This superior CH4 oxidation performance was ascribed to the synergistic effect of the RuO2 and PtOx species based on the special structures of PtOx adhered to the RuO2 surface. Firstly, compared with Pt-O/TiO2 catalyst, the higher valence state of Pt can prompt CH4 activation. Secondly, RuO2 species could provide additional oxygen species to oxidize CO adsorbed on metallic Pt to CO2, furnishing the redox cycle of PtOx species.

To further confirm the redox cycle on the catalyst surface, a CH4 or O2 pulse was applied at 350 °C on Ru-O/TiO2, Pt-O/TiO2, and RuPt-O/TiO2 catalysts. After ten pulses of CH4 or O2 during every cycle were injected, the total amount of CH4 and O2 consumed was calculated and is represented by a bar in Fig. 9. The amount of CH4 consumed over Ru-O/TiO2 (5.7 μmol·g−1) and Pt-O/TiO2 (15.7 μmol·g−1) catalysts in the first cycle of CH4 pulses was much lower than that observed for RuPt-O/TiO2 catalyst (50.0 μmol·g−1), demonstrating that RuPt-O/TiO2 catalyst possessed more active oxygen species to convert CH4. Both CO and CO2 were detected as products formed during the CH4 pulse, being consistent with in situ DRIFT spectra of CH4 adsorption. Subsequently, ten O2 pulses were injected to replenish the oxygen species consumed in the CH4 pulses, leading to re-oxidation of Pt and Ru species. The amount of O2 consumed over RuPt-O/TiO2 catalyst was 4.4 μmol·g−1, which was higher than that consumed over Ru-O/TiO2 (0.5 μmol·g−1) and Pt-O/TiO2 (1.4 μmol·g−1) catalysts, due to the low amount of CH4 consumed in CH4 pulses for the monometallic catalysts. In the second cycle of CH4 pulses, the amount of CH4 consumed over RuPt-O/TiO2 catalyst slightly decreased to 40.8 μmol·g−1 and then remained unchanged in the third cycle of CH4 pulses. These results indicate that the O species consumed in the cycle of CH4 pulses can be recovered by O2 pulses on RuPt-O/TiO2 catalyst, that is, the re-oxidation of Pt and Ru species, leading to the reservation of a high amount of CH4 consumption for CH4 pulses in the subsequent cycles. Similarly, the amount of CH4 consumed in the second cycle of CH4 pulses over Pt-O/TiO2 catalyst decreased and then remained unchanged in the third cycle of CH4 pulses. However, the amount of CH4 consumed was significantly lower than that consumed by RuPt-O/TiO2 catalyst. Regarding Ru-O/TiO2 catalyst, the amounts of CH4 consumed in three cycles of CH4 pulses were quite close to each other and were significantly lower than those consumed by both RuPt-O/TiO2 and Pt-O/TiO2 catalysts.

Fig. 9
figure 9

CH4 and O2 pulse experiments on Ru-O/TiO2, Pt-O/TiO2, and RuPt-O/TiO2 catalysts, where specific columns represent total amount of CH4 or O2 consumed for ten pulses during each cycle

Full size image

In summary, RuPt-O/TiO2 catalyst possessed a high amount of active oxygen species on the catalyst surface, and a relatively stable redox cycle of RuOx and PtOx species was achieved. In contrast, Ru-O/TiO2 catalyst had a significantly lower amount of active O species on the catalyst surface, although it exhibited the most stable redox cycle of the catalyst. The superior redox cycle of RuPt-O/TiO2 catalyst was attributed to the synergistic effect of RuO2 and PtOx species based on the special structures of PtOx adhered onto RuO2 surface, as in the case of CH4 adsorption and activation on RuPt-O/TiO2 catalyst.

4 Conclusion

The bimetallic RuPt-O/TiO2 catalyst was prepared by in situ reduction, followed by calcination in air. Specific Ru and Pt structures were constructed on the RuPt-O/TiO2 catalyst. RuO2 was epitaxially grown on a TiO2 substrate, and part of the PtOx adhered to the RuO2 surface, in addition to a single PtOx nanoparticle with a diameter of 4 nm. Owing to the synergistic effect of Ru and Pt, the RuPt-O/TiO2 catalyst exhibited superior catalytic activity compared with monometallic Ru-O/TiO2 and Pt-O/TiO2 catalysts for CH4 oxidation. The oxidation reaction rate of the RuPt-O/TiO2 catalyst at 303 °C was approximately an order of magnitude higher than that of the monometallic catalysts. Moreover, the RuPt-O/TiO2 catalyst exhibited excellent stability and water resistance. No difference was observed in the valence state of Ru between RuPt-O/TiO2 and Ru-O/TiO2 catalysts, but a significant difference was observed in the valence state of Pt between RuPt-O/TiO2 and Pt-O/TiO2 catalysts. Pt2+ and Pt4+ mainly existed on RuPt-O/TiO2 catalyst and Pt0 largely presented on Pt-O/TiO2 catalyst, in addition to Pt2+ and Pt4+. Consequently, the adsorption and activation of CH4 were strengthened on RuPt-O/TiO2 catalyst. Meanwhile, the RuO2 species could provide additional oxygen species to facilitate the redox cycle of PtOx species. Therefore, the bimetallic catalyst was more effective for methane oxidation than monometallic catalysts.