Ceo2 Based Catalysts for the Treatment of Propylene in Motorcycle’s Exhaust Gases

In this work, the catalytic activities of several single metallic oxides were studied for the treatment of propylene, a component in motorcycles’ exhaust gases, under oxygen deficient conditions. Amongst them, CeO2 is one of the materials that exhibit the highest activity for the oxidation of C3H6. Therefore, several mixtures of CeO2 with other oxides (SnO2, ZrO2, Co3O4) were tested to investigate the changes in catalytic activity (both propylene conversion and CO2 selectivity). Ce0.9Zr0.1O2, Ce0.8Zr0.2O2 solid solutions and the mixtures of CeO2 and Co3O4 was shown to exhibit the highest propylene conversion and CO2 selectivity. They also exhibited good activities when tested under oxygen sufficient and excess conditions and with the presence of co-existing gases (CO, H2O).


Introduction
The complete oxidation of propylene, one component in automobile exhaust gases, has been studied by many researchers. The most popular used catalysts are noble metals such as Pt, Pd... The catalysts OPEN ACCESS based on noble metals possess pollutant conversions higher than 90% [1][2][3][4]. Other catalysts such as perovskites, metallic oxides have also been investigated [5][6][7][8][9][10]. Amongst them, CeO2 has been widely used because it has a high oxygen storage/release capacity. However, with increasing temperature (>850 °C), CeO2 readily sinters, resulting in a deactivation of the catalyst [11]. Recently, some authors paid attention on the use of ceria-zirconia mixed oxides due to their good oxygen storage capacity (OSC) and their enhanced stability against thermal sintering [12][13][14]. A large number of investigations have shown that the addition of ZrO2 to CeO2 forms a CeO2-ZrO2 solid solution that could promote the bulk performance of CeO2 since it increases thermal stability, facilitates mobility and diffusion of bulk oxygen [15]. The addition of ZrO2 to CeO2, therefore, improves conversions of CO and hydrocarbon (HC) under reduction conditions or improves conversion of NO under oxidation conditions [16,17]. However, no agreement has been obtained about which Ce/Zr ratio could result in the highest activity of ceria-zirconia catalysts in the treatment of propylene.
Since metallic oxides are normally the cheapest and most convenient catalysts for the treatment of automobile exhaust gases, our research aims to find out some candidates with high ability to treat propylene, a main pollutant component in the motorcycle's exhaust gases. Therefore, several known metallic oxides (SnO2, TiO2, Al2O3, V2O5, Co3O4, NiO, CeO2, ZrO2, MnO2, ZnO, CuO) for the treatment of exhaust gases were investigated under same reaction conditions to select potential candidates. Since CeO2 possesses many remarkable properties as mentioned previously, the mixtures of CeO2 with some promising metal oxide candidates will be studied to see the influence of mixing on the catalytic activity. Different from other investigations in the literature, the screening of potential catalysts in this work was performed under oxygen deficient condition, which is close to the real high speed operating conditions of motorcycles. It was also expected that if a catalyst exhibits good activity for complete oxidation under oxygen deficient condition, it will exhibit even better activity under other oxygen sufficient conditions. Although the researches on the complete oxidation of hydrocarbons under oxygen deficient condition has not been adequately paid attention in the literature, in 1998 Lee et al. [27] investigated thoroughly the complete oxidation of CO and propylene under different oxygen deficient conditions. However, the authors studied Pd/Al2O3 catalysts, which is noble catalyst. Therefore, it will be interesting if one would compare activity of mixed metal oxide catalysts applied in this work with that of the noble catalysts at oxygen deficient conditions. It was also aimed that the optimal mixtures found from this research will be applied for the treatment of other components in the exhaust gases (other hydrocarbons, NOx, CO) in the followed papers.
Mixtures of CeO2 with ZrO2 or Co3O4 were also prepared by the above described sol-gel synthesis. During the gelation, if precipitation occurs, an appropriate amount of concentrated HNO3 solution was added.
X-ray powder diffraction (XRD) patterns of the catalysts were recorded with a D8 Bruker Advanced diffractometer (Bruker, Karlsruhe, Germany). The specific surface areas of the samples were measured at 77 K by the Brunauer-Emmett-Teller (BET) method using N2 adsorption/desorption on an ASAP 2010 and a Gemini VII Micromeritics apparatuses (Micromeritics, Norcross, GA, USA). The morphology of the catalysts was examined on a Hitachi X4800 scanning electron microscope. Temperature Programmed Reduction Hydrogen (TPR-H2) profiles of the catalysts were measured with a AutoChem 2920 II-Micromeritics device (Micromeritics, Norcross, GA, USA).
Catalytic activities were measured on a micro reactor with an internal diameter of 0.4 cm. Prior to use, the catalyst were pressed, ground and sieved into 250-300 μm particles, 0.1 g of the obtained catalyst was used for each reaction. The total reactant gas flow was 80 mL/min, the gas hourly space velocities (GHSV) was 76,000 h −1 , the volume compositions of the reactant gas flow were 2.5% C3H6, 2.5% O2, 95% N2. In order to investigate the influences of co-existing gases and oxygen concentrations on the reactions, other compositions were also adjusted: 0.9% C3H6, 4.1% O2, N2 balance; 0.9% C3H6, 5% O2,, N2 balance; 0.9% C3H6, 0.3% CO, 5% O2, N2 balance. To evaluate the influence of H2O, the reactant gases were flowed through a water bubbler at 25 °C. The water concentration was calculated as 2% by simulation method with a Hysys program. The reaction temperatures ranged from 200 °C to 500 °C which were measured using a thermocouple attached at the position of the catalyst bed inside an electric furnace. Analysis of propylene, oxygen, CO2, CO and oxygenate products was performed using an on-line Focus-Thermo Scientific gas chromatograph with a thermal conductivity detector (TCD). The data were obtained when the reaction reached to the stable state, i.e., from 60 min after the starting of the reactant flows. The data were stable for at least 8 hours on stream with the continuous reactant flows and during at least three cycles of the same catalytic tests.
Pollutant concentrations were measured by a specific driving cycle ECE R40 (Economic Commission for Euro Regulation 40-Emission of gaseous pollutants of motorcycles, max speed: 50 km/h, average speed: 18.7 km/h, total time: 780 s, path length: 4.052 km). The compositions of motorcycles' exhaust gases were analyzed by a CEBII gas analyzer (Austria AVL-Germany Pierburg). CO, CO2 is analyzed by a non-dispersive infrared (NDIR) analyzer. NOx is analyzed by a chemiluminescence detector (CLD). The oxygen concentrations in the exhaust gases were measured under different operating conditions: idle, medium speed and high speed. Furthermore, the compositions of hydrocarbons in the exhaust gas were analyzed by a GC-MS 2010 Shimadzu (Kyoto, Japan) and a GC-FID Thermo Electron (San Jose, CA, USA).

Composition of Motorcycle Exhausts Gases
To establish a suitable composition of reactants which reflect the composition of motorcycles' exhaust gases, the compositions of the exhaust gases from different kinds of motorcycles in Hanoi, Vietnam were measured. The largest portions are CO2 and O2 which ranges from 4,000 to 120,000, 18,172 to 120,465 ppm (by volume), respectively. Due to the motorcycles in operation in Hanoi has been used for long, the compositions of HC (5000-45,000 ppm) and CO (5,000-80,000 ppm) are quite high. In the addition, the concentration of NOx is the lowest as 0-4000 ppm. Since ECE R40 driving cycle lasted for 780 s, the concentrations of each component were examined for every 5 s. The data presented the range of instantaneous concentrations at every measurement. These data were obtained when tested ten motorcycles (Dream, Wave, Nouvo, Jupiter), which were new and used for about 50,000 km. Therefore, large ranges of HC, CO, CO2 are due to different speed applied in the ECE R40 driving cycle, different situation of the engines (old or new), which may be representative for the compositions of general working motorcycles. Compared to the data of normal engine exhaust gases obtained from literature [33], the composition of toxic gases (CO, HC) in motorcycles' exhaust gases shown here are much higher since the motorcycles in operation in Hanoi has been used for long. Thus, the requirement for treatment is stringent. The investigation of oxygen concentration at different operating conditions (idle, medium and high speed conditions) showed that at high speed condition, the oxygen concentration is minimum (1.8 vol%). It is clear that under that operating conditions, oxygen in the exhaust gases is deficient for the complete oxidation of HC and CO. The investigation of HC in the different exhaust gases by GC-MS show that they include methyl-cyclopentane (1200-10,800 ppm), 2,3-dimethyl pentane (about 4500 ppm), toluene (500-7200 ppm), benzene (about 3000 ppm), methane (1,250-10,000 ppm), propylene (1000-8550 ppm), methanol (about 4500 ppm), ethanol (1000-17,000 ppm), acetaldehyde (2000-4500 ppm) and a few other minor gases. Amongst them, propylene was detected with high concentration. Therefore, in our work, C3H6 was chosen as the pollutant component for the treatment. Based on the real compositions of C3H6 and O2 in the exhaust gases, the concentration of the reactant gases was chosen as 2.5 vol% C3H6, 2.5 vol% O2 and 95 vol% N2, which is the oxygen deficient condition for the oxidation of HC. The selection of C3H6 as the object for the treatment is mainly to make a model to test the oxidation ability of the catalysts. Good catalysts found from this investigation will then be tested for the treatment of real exhaust gases, which include also other kinds of volatile organic compounds.

Characterization and Catalytic Activities of Several Single Metallic Oxides
Several single metallic oxides which were found pure from XRD investigation were tested under the same reaction conditions as mentioned previously. Since the reaction condition was oxygen deficient, oxygenate products and CO were also examined during the reaction. CO2 selectivity was calculated based on the reaction products detected by GC, which are CO2, CO, formic acid, acrolein, acetaldehyde, and ethanol. In fact, more oxidation products may still exist with low concentrations and the real selectivity may be lower. However, the CO2 selectivity calculated here is still suitable to compare catalytic activities of the catalysts working under the same conditions and being calculated by the same way.
Propylene conversion and CO2 selectivity of the investigated oxides are shown in Tables 1 and 2, respectively. Amongst these catalysts, NiO was only studied at temperatures below 350 °C due to its low thermal resistance. At higher temperatures, it was observed that NiO particles were broken-up, resulting in blocking of the reactor. The same observation was also seen with MnO2 at temperatures above 450 °C. However, the instability of these catalysts at high temperatures was only due to mechanical reason since TGA/DTA and XRD results indicated no change in phase compositions of the samples at the reaction temperatures. Therefore, their mechanically instability will not significantly influence their catalytic activities.
The results from Table 1 show that the propylene conversions of the catalysts almost reach to a maximum value at a certain temperature. MnO2 and Co3O4 exhibit high conversions from 250 °C on, whereas NiO shows a high conversion at 350 °C, V2O5 and CeO2 at 400 °C and SnO2, ZnO at 450 °C. For Al2O3, ZrO2 and CuO, propylene conversion still stays very low even at high temperatures up to 500 °C. Co3O4 and NiO catalysts result in the highest propylene conversion. MnO2 and CeO2 exhibited only a slightly lower conversion.  These catalysts also possess rather high CO2 selectivity as seen from Table 2. However, CO2 selectivity of Co3O4 decreases dramatically at high temperatures due to the formation of more CO (selectivity at 500 °C is 61%). Oxygenated products were observed in the oxidation of propylene on all investigated catalysts but especially found in the reactions with V2O5, SnO2, TiO2, ZnO since these catalysts are well known catalysts for partial oxidation of hydrocarbons. Oxygenated products may also be formed when using CuO and ZrO2 but because CuO and ZrO2 exhibited low conversions, the amount of formed oxygenated products may be too low to be detected. For V2O5, SnO2, TiO2, ZnO, CO2 selectivity was low at 350-400 °C because these temperatures are optimal for partial oxidation to form oxygenate products (selectivity of oxygenate products may reach about 30%). When increasing temperature, CO2 selectivity on these catalysts increased significantly as the formed oxygenated products were also completely oxidized to CO2. Amongst the investigated catalysts, CeO2, and especially MnO2 exhibit quite constant and highest CO2 selectivity at all examined temperatures. When BET surface areas (Table 1) were taken into account, it is clear that high surface area oxides such as Al2O3, TiO2, ZrO2 exhibited low activity, they are only suitable to be supports. Oppositely, low surface area oxides such as MnO2, Co3O4, NiO, CeO2 exhibited good activity and suitability to act as active phases. Therefore, if surface areas of highly active oxides are increased, the activities of the catalysts will be improved.
In general, MnO2, CeO2 and Co3O4 are the most promising catalysts to convert propylene under oxygen deficient conditions. CeO2 has a high ability to convert propylene with high CO2 selectivity at all investigated reaction temperatures due to a high OSC as discussed in literature [16]. Co3O4 has a high propylene conversion at low temperatures but also a low CO2 selectivity at high temperatures. MnO2 shows high activity for both propylene conversion and CO2 selectivity but it is mechanically unstable at high temperatures. Therefore, in the following investigation, catalyst mixtures containing CeO2 will be focused on.

Characterization and Catalytic Activities of Mixtures of CeO2
The catalytic activity of mixtures of CeO2 and SnO2 with different compositions was investigated since it was expected that the addition of the high conductive semiconductor SnO2 on the highly active catalyst CeO2 will improve the reaction due to the increase of the available lattice oxygen, which may act as an oxidizing agent at high temperatures. However, no development of catalytic activity has been observed with the mixtures of CeO2 and SnO2.
Although catalytic activity of ZrO2 was low as seen from the previous section, the catalytic activities of CeO2-ZrO2 chemical mixtures were also studied since the literature reports that ZrO2 is able to modify the sub-lattice oxygen in the CeO2-ZrO2 mixed oxides, generating defective structures and highly mobile oxygen atoms in the lattice which can be released even at moderate temperatures [12,13,34]. Therefore, the activities of these chemical mixtures are expected to be increased. Moreover, ZrO2 exhibited higher surface area than CeO2 as shown in Table 1, thus, the addition of ZrO2 may help to increase surface area of the catalysts to improve their activities.
XRD patterns of CeO2-ZrO2 chemical mixtures are shown in Figure 1. CeO2 exhibits a cubic structure (a, b, c parameter is 5.4 nm) represented by 2θ = 28.6° and 33.1°, ZrO2 calcined at 550 °C (temperature lower than 1170 °C) exhibits a monoclinic structure represented by 2θ = 28.2° and 31.6°. The evidence of solid solution formation is given by the shift of ceria reflections to higher values for Ce0.8Zr0.2O2 and Ce0.5Zr0.5O2 samples, where Zr (atomic radius is 160 pm) replaced for Ce (atomic radius is 181.8 pm) in the cubic structure of CeO2. In the case of sample with higher Zr ratio, it has been reported in the literature [35] that a compound Ce0.25Zr0.75O2 shows the strongest XRD reflection at 2θ = 30°. In our work, the sample Ce0.1Zr0.9O2 also showed a cubic structure with the strongest XRD reflection at 2θ = 30°. Besides, single ZrO2 monoclinic phase still existed, which represented by small peaks at 2θ = 28.2° and 31.6° as observed in XRD pattern of this sample. Here, it should be noticed that ZrO2 also exist as cubic structure (a, b, c parameter is 5.1 nm) when synthesized at high temperature (more than 2370 °C). Thus, cubic structure of ZrO2 is very close to that of CeO2 with almost the same parameters. The strongest XRD reflection of this cubic ZrO2 is at 2θ = 30°. Therefore, in the presence of CeO2, ZrO2 cubic structure may already be formed at low temperature because a few percentage of other oxide may stabilize cubic ZrO2. Since the structure of cubic ZrO2 and cubic CeO2 are very similar (very close a, b, c parameters and strongest XRD reflections) and the content of CeO2 in Ce0.1Zr0.9O2 sample was small, XRD patterns of Ce0.1Zr0.9O2 sample looked like that of pure cubic CeO2 with 2θ shifted to 30° but the existed phase could be mainly assigned for cubic ZrO2.   All investigated CeO2-ZrO2 chemical mixtures possess surface areas around 50 m 2 /g, which are almost equal to those of pure ZrO2 (52 m 2 /g). Pure CeO2 possesses a little lower surface area (33 m 2 /g).

CeO2
(b) Figure 2b shows CO2 selectivity of CeO2-ZrO2 chemical mixtures at different reaction temperatures. It can be seen that CO2 selectivity of all CeO2-ZrO2 chemical mixtures is much higher than that of the least active pure component (ZrO2). However, only samples with low ZrO2 content (Ce0.9Zr0.1O2 and Ce0.8Zr0.2O2) exhibit comparable and quite constant high CO2 selectivity. The high ZrO2 content samples possess low CO2 selectivity at temperatures ranging from 350-400 °C since these temperatures were favorable for partial oxidation to form oxygenated products. Combining propylene conversion and CO2 selectivity, it is clear that Ce0.9Zr0.1O2 and Ce0.8Zr0.2O2 samples exhibit the highest activity for the complete oxidation of propylene into the nontoxic product CO2. The activity of these catalysts even increased significantly when the reaction was performed under the oxygen sufficient condition (molar ratio of C3H6/O2 was 1/4). The propylene conversions under the oxygen sufficient condition at 500 °C on Ce0.9Zr0.1O2 and Ce0.8Zr0.2O2 catalysts were 77.10% and 85.42%, respectively. The CO2 selectivity under this condition was above 97%. The catalytic activity of CexZr1−xO2 for the oxidation of propylene under oxygen excess conditions have been reported by D. Homsi et al. [36], who found no enhancement of activity of the Ce0.75Zr0.25O2 solid solution compared to that of pure CeO2. That result is a bit different with the result of this work as Ce0.8Zr0.2O2 exhibited higher activity than that of pure CeO2. The reason may be the different synthesis methods, which may result in different properties of the final products. The catalyst in our work was prepared by solgel synthesis, which may allow better access of zirconium into the structure of ceria, leading to the increase of mobile oxygen as seen from TPR-H2 results (Table 3). Oppositely, the catalysts in D. Homsi's work were prepared by precipitation, the method allows worse mixing of components than sol-gel method. Moreover, surface area of our Ce0.8Zr0.2O2 was higher than that of our CeO2 while it is opposite in D. Homsi's work. Therefore, both catalytic activity and reduction ability of Ce0.75Zr0.25O2 was not higher than that of CeO2 in D. Homsi's work. Besides, different reaction conditions (oxygen deficient in our work and oxygen excess in D. Homsi's work) may also influenced since the advance properties of a solid solution will be more presented under oxygen deficient conditions. The reason for the enhancement in catalytic activity of Ce0.9Zr0.1O2 and Ce0.8Zr0.2O2 samples, thus, may be the formation of the solid solution Ce1−xZrxO2 in the chemical mixtures of CeO2-ZrO2. To prove this assumption, the CeO2-ZrO2 mechanical mixture containing 80% mol CeO2 was also tested for the reaction (Figure 3). This mechanical mixture showed the presence of only single CeO2 and ZrO2 phases but not solid solution. It can be observed that both propylene conversion and CO2 selectivity of the mechanical sample are much lower than that of the sol-gel prepared sample, where the formation of a solid solution Ce0.8Zr0.2O2 was detected.  However, the sol-gel samples with high content of ZrO2 did not show improvement of activity. Thus, it may be assumed only a little change in the structure of Ce1−xZrxO2 solid solution compared to that of CeO2 helps to increase catalytic activity. For other Ce1−xZrxO2 solid solution (x > 0.5), where the structure shows a big shift of ceria reflections to higher values, the increase of catalytic activity will not happen. The formation of a solid solution with a little replacement of Zr ions to Ce ions may increase the catalytic activity since this replacement may result in appropriate vacancies inside the bulk structures of the catalysts, which increases mobility of oxygen transported inside the bulk structures, i.e., also increases the OSC of the catalyst. An evidence of the increase of OSC of the Ce1−xZrxO2 catalyst with low ZrO2 content compared to pure oxides was estimated based on TPR-H2 profiles of the pure oxides and chemically mixed samples ( Table 3). The consumed H2 quantities on the solid solutions of CeO2 and ZrO2 (Ce0.9Zr0.1O2 and Ce0.8Zr0.2O2) were much higher than those on pure oxides (CeO2 and especially ZrO2). At the same time, the temperature of hydrogen reduction decreased significantly on the Ce0.9Zr0.1O2 catalyst (the lowest reduction temperature of Ce0.9Zr0.1O2 sample is only 316 °C while that of CeO2 is 474 °C and of ZrO2 is 623 °C), therefore, the catalyst reached to high activity at lower temperature. However, the fact that the catalytic activity of Ce0.9Zr0.1O2 was a little lower than that of Ce0.8Zr0.2O2 although the consumed hydrogen amount of Ce0.9Zr0.1O2 was a little higher than that of Ce0.8Zr0.2O2 is a bit non logical. Here, the influence of surface area might be a reason as surface area of Ce0.8Zr0.2O2 was higher than that of Ce0.9Zr0.1O2, which may help to expose more active sites. Figure 4 shows the morphology change of the Ce0.8Zr0.2O2 sol-gel sample before and after reaction (24 h on stream). Before the reaction, the sample possesses many pores. After reaction, almost all pores are encapsulated. The reason may be the exothermicity of complete oxidation of C3H6 and the formation of coke. Chemical mixtures of CeO2 and Co3O4 were also studied since the results in Section 3.2 shows that Co3O4 exhibits high propylene conversion at low temperature although it did not exhibit high CO2 selectivity at high temperatures. Meanwhile, CeO2 exhibits high CO2 selectivity at high temperatures although its propylene conversion is not as high as that of Co3O4 at low temperatures. Therefore, when CeO2 and Co3O4 are mixed together, the obtained catalysts may exhibit high conversion of propylene at low temperatures and high CO2 selectivity at high temperatures. XRD patterns of some CeO2-Co3O4 chemical mixtures (20% and 50% mol of CeO2) in the comparison with XRD patterns of pure CeO2 and Co3O4 are presented in Figure 5. Pure Co3O4 synthesized at 550 °C exhibited a strong amorphous nature with a high and rough baseline but the strongest XRD reflections of a cubic Co3O4 structure (a, b, c parameters of 8.1 nm) were still detected at 2θ = 31°, 37°, 45°, 59° and 65°. Meanwhile, XRD patterns of CeO2-Co3O4 chemical mixtures (even up to 80% Co3O4) show the presence of mainly CeO2-like structure. However, there are shifts of ceria reflections to higher 2θ values and rougher baselines than that of pure CeO2. Thus, like the chemical mixtures of CeO2 and ZrO2, there may be also a formation of a solid solution in CeO2-Co3O4 chemical mixtures with the replacement of Co (atomic radius is 125 pm) for Ce (atomic radius is 181.8 pm) in the structure of CeO2. Because the content of Co3O4 was higher, Co3O4 phase may still be existed but as small or amorphous particles surround the solid solution of CeO2-Co3O4, leading to the rougher baselines than that of pure CeO2. The CeO2-Co3O4 chemical mixtures possess surface areas around 45 m 2 /g, which are higher than those of pure CeO2 (33 m 2 /g) and pure Co3O4 (11 m 2 /g). These may be reasons for the higher activity of the mixtures compared to pure components as described below.   68  64  60  56  52  48  44  40  36  32  28  24  20 2 theta, degrees Propylene conversions and CO2 selectivity of CeO2-Co3O4 chemical mixtures are presented in Figure 6. Compared to the single oxides, CeO2-Co3O4 chemically mixed catalysts show high propylene conversions at lower temperature (200 °C). These chemical mixtures also possess as high CO2 selectivity as that of pure CeO2 at all reaction temperatures up to 400 °C. Under the oxygen sufficient condition (molar ratio of C3H6/O2 was 1/4), CeO2-Co3O4 chemical mixtures converted about 87% propylene since 250 °C with CO2 selectivity of about 98%. Under oxygen excess condition (molar ratio of C3H6/O2 was 1/5.5), CeO2-Co3O4 chemical mixtures converted 100% propylene since 250 °C with CO2 selectivity of 100%. Thus, the combination of Co3O4 with CeO2 improves propylene conversion and CO2 selectivity. Especially, this combination lowers the temperature of the maximum activity to 200 °C, which is very important for the treatment of hydrocarbon during the starting operation of the engines. However, the mechanical stability of Co3O4-CeO2 chemical mixtures is low. The materials were broken-up at temperatures higher than 400 °C. Therefore, Co3O4-CeO2 catalysts should be supported on high thermal resistant supports. The reason for the enhancement in catalytic activity of CeO2-Co3O4 chemical mixture samples may be the formation of the solid solution in the chemical mixtures of CeO2-Co3O4 resulted by the replacement of Co atoms for Ce atoms as seen from XRD patterns ( Figure 5). To prove this assumption, the CeO2-Co3O4 mechanical mixture containing 50 mol% CeO2 was also tested for the reaction (Figure 7). It can be observed that both propylene conversion and CO2 selectivity of the mechanical sample are lower than that of the sol-gel prepared sample, where the formation of a solid solution was detected. Especially, like pure CeO2 and Co3O4, the mechanical mixture exhibited very low propylene conversion at low temperature (200 °C) while much higher propylene conversion had already been obtained on the chemical mixture at this temperature. TPR-H2 data of pure CeO2, Co3O4 and the chemical mixture of 20% CeO2-80% Co3O4 in Table 3 helped to explain the activity of the mixtures and the pure oxides. TPR-H2 shows that Co3O4 exhibited an excellent mobility of oxygen as its consumed H2 quantity was highest amongst the investigated catalysts. Co3O4 was also reduced at lower temperatures than CeO2, which explains for the fact that Co3O4 exhibited good activity at lower temperature than CeO2. The chemical mixture of 20% CeO2-80% Co3O4 did not possess a larger quantity of mobile oxygen than pure Co3O4 but was reduced at lower temperature (since 279 °C); therefore, the chemical mixture of 20% CeO2-80% Co3O4 was able to convert propylene at lower temperature than Co3O4. For the chemical mixture of 20% CeO2-80% Co3O4, a solid solution was also formed as seen from XRD pattern of the sample, but the amount of consumed H2 of the mixture was not higher than that of pure Co3O4. Therefore, this mixture did not result in higher conversion of propylene than those of pure oxides; the advantage of this catalyst was combining well activity of both CeO2 and Co3O4 resulting in a catalyst with good activity at a wider temperature range. To explain for the fact that although the chemical mixture of 20% CeO2-80% Co3O4 possesses lower quantity of mobile oxygen than pure Co3O4 but there was no decrease in catalytic activity, a careful look on both TPR-H2 results and catalytic activities of good chemical mixtures of CeO2 with ZrO2 and Co3O4 also means something. It was seen that the chemical mixtures of CeO2 with ZrO2 (Ce0.8Zr0.2O2, Ce0.9Zr0.1O2) also exhibited good activity but the amount of consumed H2 was only about 30 mL/g while that of CeO2-Co3O4 chemical mixture was higher than 100 mL/g. Thus, it may be assumed that to ensure good oxidation of propylene, amount of mobile oxygen may be a certain value. If a catalyst possesses the amount of mobile oxygen higher than that necessary value, the activity may not increase significantly any more.
In order to explore the influences of co-existing gases and oxygen concentrations on catalytic performances, potential chemical mixtures of CeO2 with ZrO2 and Co3O4 (Ce0.8Zr0.2O2 and 20% CeO2 + 80% Co3O4) were studied in more details (Table 4). It is clear from Table 4 that these mixtures exhibited good activity for the oxidation of propylene not only under oxygen deficient condition but also under oxygen sufficient condition. Especially, under oxygen excess condition (condition 3 and 4), the chemical mixture of 20% CeO2 and 80% Co3O4 was able to convert 100% propylene and CO since 200 °C. A presence of 2% H2O did not influence significantly on the activity of this catalyst except that the minimum active temperature increases from 200 to 250 °C. The influence of CO and H2O was also not significant for the catalyst Ce0.8Zr0.2O2, proving that Ce0.8Zr0.2O2 and 20% CeO2-80% Co3O4 catalysts are stable catalyst for the oxidation of propylene under different reaction conditions. Compare to 20%CeO2-80% Co3O4 catalyst, Ce0.2Zr0.2O2 catalyst exhibited less activity under oxygen sufficient and excess conditions as conversion of propylene was less than and the temperature of the maximum activity was higher than those of 20% CeO2-80% Co3O4 catalyst. Catalytic activities of these catalysts are comparable to those of noble catalysts under the same deficient and sufficient conditions (at the same air to fuel ratios of 10 and 14) [27]. The catalyst mixtures of CeO2 and Co3O4 in this work even exhibited an advantage of having maximum activity at lower reaction temperature. The catalysts were stable during different catalytic cycles; the conversion and selectivity were almost unchanged during at least three catalytic cycles. Mixtures of CeO2 and MnO2 were also expected to exhibit good activities since MnO2 was one of the most active catalysts as investigated in Section 3.2. However, this catalyst system requires a lot of detailed investigations in order to explain thoroughly their properties. Therefore, theses mixtures will be studied in a separate paper.

Conclusions
The results of this work show that MnO2, CeO2 and Co3O4 are the most promising catalysts to convert propylene under oxygen deficient conditions. CeO2 not only shows a high propylene conversion but also a high CO2 selectivity at high temperatures. Therefore, CeO2 should be chosen as one component of the mixed catalysts for the treatment of propylene.
The catalytic properties of CeO2 could be improved by the combination with other components such as ZrO2 and Co3O4. Ce0.9Zr0.1O2 and Ce0.8Zr0.2O2 solid solutions exhibit the highest propylene conversions and CO2 selectivities. The chemical mixture of CeO2 and Co3O4 not only exhibits the highest propylene conversion and CO2 selectivity but also converts the largest amount of propylene at much lower temperatures compared to other catalysts. These catalysts also exhibited good activities when tested under oxygen sufficient and excess conditions and with the presence of co-existing gases (CO, H2O).