2.1 Catalyst preparation

Ce–Ti oxide (Ce_0.3Ti_0.7O₂) was synthesized using the sol–gel method [12,13]. Desired amount of an aqueous solution of cerium nitrate Ce(NO₃)₃·6H₂O [Rectapur, 99.5%] and ethanol CH₃CH₂OH were added under stirring to another solution of ethanol CH₃CH₂OH and titanium(IV) isopropoxide Ti(OC₃H₇)₄ [Across Organics; 98%] with molar ratio Ti(OC₃H₇)₄/CH₃CH₂OH = 1/2. The molar ratio between H₂O and titanium(IV) precursor is H₂O/Ti(OC₃H₇)₄ = 5. The solution was gelled after finishing the reaction between titanium(IV) isopropoxide Ti(OC₃H₇)₄ and water. The gel was dried at 80 °C during 24 h.

To investigate the effect of the calcination temperatures on the properties of Ce_0.3Ti_0.7O₂ oxide, the dried solid was calcined under air for 4 h at different temperatures: 400, 500 and 600 °C. These samples are denoted as Ce_0.3Ti_0.7O₂ (calc. T), where T indicates the calcination temperature.

The gold-based catalysts were prepared by the deposition precipitation (DP) method [14]. Aqueous solution of tetrachloroauric acid (10⁻³ mol L⁻¹) [Stremchemicals; 99.9%] was added under stirring to an aqueous suspension of oxide support Ce_0.3Ti_0.7O₂ and aqueous solution of urea in excess. The solution was heated at 80 °C to decompose urea and obtain pH equal to 6.7. The pH of solution was maintained at the value of 6.7 during 4 h to obtain high dispersion of fine gold particles on the oxide supports. The mixture was filtered and washed with deionised water at 60 °C several times in order to eliminate the chloride ions, dried during 24 h at 80 °C and finally calcined under air for 4 h at 400 °C. The catalyst thus obtained are denoted as Au/Ce_0.3Ti_0.7O₂ (calc. T), where T is the calcination temperature of the support.

2.2 Catalyst characterization

Elemental analyses were performed by inductively coupled plasma atomic emission spectroscopy at the CNRS Centre of Chemical Analysis (Vernaison, France).

BET surface area was measured by nitrogen adsorption at −196 °C in a Thermo-Electron QSurf M1 apparatus. Before analysis, the samples were treated for 30 min at 120 °C under a 30% N₂/He flow.

XRD analysis was performed on a BRUKER Advance D8 powder X-ray diffractometer using Cu Kα radiation. Diffraction patterns were recorded over a 2θ range of 15–80° and using a step size of 0.02° and a step time of 4 s. The mean crystallite sizes were estimated using the Scherrer equation.

Diffuse Reflectance UV–vis spectroscopy (DR/UV–vis) experiments were carried out on a VARIAN CARY 5000. The measurements were performed on air exposed samples at ambient temperature between 200 and 800 nm.

The temperature programmed reduction experiments were carried out in an Altamira AMI-200 apparatus. The TPR profiles were obtained by passing a 5% H₂/Ar flow (30 mL min⁻¹) through 50 mg of samples heated at 5 °C min⁻¹ from ambient temperature to 900 °C. The hydrogen concentration in the effluent was continuously monitored by a thermoconductivity detector (TCD).

2.3 Catalytic activity measurements

Catalytic tests were carried out at atmospheric pressure in a conventional fixed bed microreactor using 100 mg of fine catalyst powder. The reactive flow (100 mL min⁻¹) is composed of air and 6000 ppm of propene. The reactants and the reaction products were analysed by a VARIAN chromatograph equipped with FID and TCD detectors. The catalysts were first activated at 400 °C for 4 h under air (2 L h⁻¹) and the conversion measurement was studied on a slow heating ramp between 20 and 400 °C at 1 °C min⁻¹. The temperature ramp of 1 °C min⁻¹ was considered to be slow enough to reach a pseudo-steady state at every point.

3.1 Ce_0.3Ti_0.7O₂ oxide study

Fig. 1 presents the conversion of propene as a function of temperature in the presence of the catalytic support Ce_0.3Ti_0.7O₂ calcined at different temperatures. It can be observed that the catalytic activity of mixed oxides is quite similar whatever the calcination temperature of supports. Over none of these oxides 100% conversion was reached. However, regarding the conversion rate of propene at 400 °C in Table 1, a slightly better activity of the Ce_0.3Ti_0.7O₂ support calcined under air at 400 °C can be noticed. In order to identify the physico-chemical properties of these samples, these materials were characterized by several techniques.

BET specific surface area of the samples calcined at 400, 500 and 600 °C are mentioned in Table 1. The surface area decreases with increasing calcination temperature. Regarding the activity measurement, it can be observed that the solid Ce_0.3Ti_0.7O₂ (calc. 400 °C) which present the best activity, possess the highest specific area (230 m² g⁻¹).

XRD patterns of the Ce_0.3Ti_0.7O₂ support at different calcination temperatures are reported in Fig. 2. This figure shows that all the samples are characterized by the absence of well resolved diffraction peaks of TiO₂ or CeO₂. This indicates that the solids are in a relatively amorphous state. The highly amorphous character of the samples can be deduced from the high level of noise detected in the patterns. Thus, for the sample Ce_0.3Ti_0.7O₂ calcined at 400 and 500 °C a broad peak at around 2θ = 28.5° is observed on their XRD profiles. However, when the solid was calcined at 600 °C, some patterns with bad resolution corresponding to anatase and cerianite phases are revealed. This result indicates that some CeO₂ and TiO₂ crystallites are present in the Ce_0.3Ti_0.7O₂ (calc. 600 °C) sample. This is in accordance with the BET measurement; which reveals the lowest value (86 m² g⁻¹) for this solid, due to its higher degree of crystallinity than the other samples. The same behavior has been observed by Rynkowski et al. [12] in the case of ceria–titania mixed oxides with atomic ratios Ti/Ce = 8/2. Lopez et al. [15] have evidenced the appearance of well resolved diffraction lines of cerianite phase in the mixed oxide TiO₂–CeO₂ (90 wt% of TiO₂) after calcination at 800 °C. Those authors have shown that the ceria is formed at low temperature in the form of small dispersed crystallites in the structure of amorphous TiO₂ and which are not detected by the XRD. When increasing the calcination temperature, these crystallites start growing and sintering to constitute larger ones, which allow their detection. Indeed the better crystallization state is in accordance with the decrease of the specific areas of the Ce_0.3Ti_0.7O₂ samples when increasing the calcination temperature to 600 °C.

DR/UV–vis optical spectra of Ce_0.3Ti_0.7O₂ samples are reported in Fig. 3. Several absorption bands characteristic of titanium are observed in the range 200–400 nm. According to the literature [16–18], the band at 215 nm should be due to isolated Ti(IV) in a tetra-coordinated environment, the band at 220 nm is attributed to isolated Ti atoms in octahedral coordination, the band at 260 nm is attributed to Ti⁴⁺ ions in octahedral coordination. The peak at 330 nm is ascribed to the O²⁻ → Ti⁴⁺ charge transfer transition, indicating the anatase TiO₂ species. The band observed at 300 nm is attributed to O²⁻ → Ce⁴⁺ charge transfer transition [19]. When increasing the calcination temperature, the adsorption edge at 380 nm assigned to the inter-band transitions which take place in CeO₂ shifts to the visible range. Several studies [20,21] have shown that increasing the size of semi-conducting micro-crystallites entails a shift of the absorbance edge of the inter-band transition. This result is in accordance with XRD analysis which presents the formation of crystallites of ceria with the increase of the calcination temperature of the solids.

Fig. 4 shows the profiles obtained after H₂-TPR measurements of Ce_0.3Ti_0.7O₂ calcined at various temperatures. For all oxides, a first broad peak starting at about 280 °C is detected. This reduction peak maximum shifts from 466 to 545 °C with increasing calcination temperature from 400 to 600 °C. In accordance with literature data [13,22], this reduction peak is generally attributed to reduction of oxygen species in the surface region of ceria whereas the very small peak observed at about 790 °C corresponds to the reduction of the oxygen species in the bulk of the ceria. The shift of the low temperature reduction peak to a higher temperature could be induced by the formation and agglomeration of nanoparticles of Ceria and Titania. Indeed, Mengfei et al. [13] have shown that with the increasing of the temperature, the H₂-TPR peak shifts to higher temperatures, probably, because of the sintering of the samples. The hydrogen consumption is also presented in Table 2. It could be observed that the H₂ consumption is relatively constant, whatever is the calcination temperature of the Ce–Ti oxides.

3.2 Au/Ce_0.3Ti_0.7O₂ catalysts' study

Gold is deposited on calcined oxides as described in Experimental section. The gold content (% wt) determined by elemental analysis is mentioned in Table 3.

Fig. 5 represents the conversion of propene as a function of temperature in the presence of different gold-based catalysts. Thus, adding gold on the oxide support improves the catalytic activity for the propene oxidation. The propene conversion is performed at lower temperature and presents 100% of selectivity towards formation of CO₂. It can be observed that the catalytic activity changes slightly according to the temperature of support calcination. It seems that the gold deposited on support calcined at 400 °C presents the best catalytic activity. Thus, the calcination temperature of the support seems to have an effect in the activity for gold-catalyzed propene oxidation.

Concerning the BET measurement (Table 3), a slight decrease in surface area was observed when gold is deposited on the support. Moreover, it can be noticed that the catalyst with the support calcined at 400 °C exhibits the highest surface area and the best catalytic activity towards propene oxidation.

The XRD patterns for the gold-based catalysts are given in Fig. 6. No change concerning the structure of the Ce_0.3Ti_0.7O₂ support due to the gold deposition has been evidenced. However, a small diffraction peak at 2θ = 38.3° appears for all samples. This peak corresponds to nanoparticles of metallic gold. The size of the gold particles, deduced from the broadening of this reflexion peak is evaluated. Their values are 9, 12, and 7 nm for catalysts with the support calcined at 400, 500 and 600 °C, respectively.

In order to identify the nature of the active gold particles, the catalysts were analysed by DR/UV–vis spectrophotometry. The obtained spectra are represented in Fig. 7. Besides, the support contribution, the UV–vis spectra of the gold-based catalyst show a new band at around 560 nm. The observation of this absorption peak is corresponding to metallic gold nanoparticules. This phenomenon has also been revealed in previous work [9]. This band is characteristic of surface plasmon resonance of metallic gold particles which arises from the collective oscillations of the free conduction band electrons that are induced by the incident electromagnetic radiation. Such resonances are observed when the wavelength of the incident light far exceeds the particle diameter. The small variations of wavelength of this type of bands are dependent on particle size, shape, and surrounding environment [14,23–25]. Moreover, it is interesting to note that another absorption band at around 630 nm appeared which belongs to the longitudinal plasmon resonance band of aggregates or deviation from spherical geometry [26–28] of gold nanoparticles.

The H₂-TPR profiles of the gold deposited Ce_0.3Ti_0.7O₂ supports represented in Fig. 8 show a reduction peak at low temperature. The TPR profile of the Ce_0.3Ti_0.7O₂ represented in Fig. 4 shows a massif peak of reduction in the range of 466–545 °C, this one is attributed to the reduction of surface oxygen species from ceria. When gold is added to the support, modification in the TPR profiles was observed. Indeed a reduction peak appears at lower temperature. This peak corresponds to the reduction of the oxygen species on the gold particles and on ceria. Andreeva et al. [29,30] have shown that adding noble metals (like gold) on CeO₂ permits to lower the reduction temperature of oxygen species on ceria surface. Table 4 resumes the H₂-TPR experimental and calculated consumption of catalyst samples for peaks at low temperature (<400 °C). For the H₂ calculated consumption, we consider that gold species are in the ionic state Au³⁺. Thus, knowing the gold content in solids and following the reduction reaction of Au³⁺ into Au⁰, the amount of H₂ over consumption by the catalysts could be estimated. The H₂ over consumption exists for all gold catalysts. This result shows the influence of gold to facilitate the reduction of oxygen species from Ce_0.3Ti_0.7O₂ support surface on the border with gold. These results could be correlated to the results of the catalytic tests. Indeed, it can be noticed that the solid named Au/Ce_0.3Ti_0.7O₂ (Calc. 400 °C) shows the best activity towards the total oxidation of propene, showing also higher hydrogen over consumption. This observation could be explained by the interaction between gold nanoparticles and cerium oxide. Ceria is very well known for its oxygen storage capacity (OSC), defined as the capacity to release and accept oxygen. Centeno et al. [6] have studied the catalytic combustion of VOCs on Au/CeO_x/Al₂O₃, and they concluded that the presence of cerium ions has a positive influence on the fixation and the final dispersion of gold on alumina support. Moreover, ceria stabilizes the gold particles with low crystallite size. In addition to these beneficial features, ceria, because of its redox properties, may improve the catalytic behavior of the catalyst by increasing the supply of active oxygen. Similar conclusions were proposed by Sciré et al. [22] in their study of VOC oxidation over Au/CeO₂. The mechanism of hydrocarbon oxidation over ceria and other reducible metal oxides is usually considered to be of the redox or Mars and van Krevelen type [6,22]. The key steps of this mechanism are believed to be the supply of active oxygen by the readily reducible oxide and its reoxidation by oxygen. Gluhoi et al. [7] have shown that gold-based catalysts that contain a transition metal oxide or ceria are highly active in propene oxidation. The role of the oxide is twofold: it stabilizes the gold particles against sintering (ceria) and provides active oxygen for the reaction.