Volatile organic compounds (VOCs) have been considered the main contributor to air pollution^[1].The conventional approach for VOCs destruction is catalytic combustion. In that field of catalysis, noble metal/ TiO₂ catalysts have been extensively reported and stated to be highly efficient^[2-3]. The reason of the highly appreciated noble metal/ TiO₂ catalyst lies in two aspects. One is the unique semiconductive properties of TiO₂ material, for the ability of being particle reduced can help activate oxygen. The other is the strong metal-support interaction between palladium and TiO₂ which can inhibit agglomeration of metal particles. However, on the field of classical impregnation method to prepare Pd/TiO₂ (P25), Vera P. Santos have found that although the total oxidation of ethanol could be obtained at the temperature of 180 ℃, the total conversion into CO₂ can only be achieved at 260 ℃^[4]. Hence, even though significant improvement has been made towards developing noble metal catalysts, challenge still remains. It is well known that total oxidation into CO₂ of VOCs over noble metal catalysts should be occur at lower temperature.

Porous TiO₂ materials especially mesoporous TiO₂ now have been occupied an important position in the field of environmental technology, such as photocatalytic degradation of toxic organic contaminants, photocatalytic water pollutant and titania film might be used in Dye Sensitized Solar Cells^[5]. Ordered porous TiO₂ has attracted increasing interest for its high porosity and specific surface area, uniform aperture size and narrow pore size distribution^[6-9]. While little literature has been reported about the mesoporous powder TiO₂ used as support of noble metal catalyst. Actually the pore wall of orderly porous TiO₂ is usually composed of nanometer TiO₂ particles, which can combine the excellent characteristics of both ordered porous which is considered to be beneficial to the diffusion of the reactants and products and nanometer TiO₂ which is well known for its effectiveness nontoxicity and chemical stability. However, there are some obvious drawbacks in applying porous TiO₂, such as low thermal stability, low density of surface active site and weak surface acidity, which significantly limit the efficiency of catalysis^[10-11]. Efforts have been paid to improve properties of TiO₂ porous material. There are a number of strategies reported among which incorporation TiO₂ with another metal oxide has been highlighted^[12]. Cause composite materials can achieve the possession of high thermostability of anatase, orderly porous channel and defect TiO₂ nanoparticles with abundant chemical pro-perties. Alumina, due to its affordability, easy accessibility and high chemical stability, has been extensively studied. The literature has investigated the effect of preparation method on the properties of Al₂O₃-TiO₂ and found that the addition of Al₂O₃ to TiO₂ can produce novel feature because of their unique chemical and physical properties^[13-14].

Supported palladium catalyst is well-known for its high catalytic performance in the alcohol oxidation; however, development in catalytic selectivity has been in the quest and more efforts must be paid to enhance palladium based catalysts related to specific organic component reaction. An important factor concerning the VOCs oxidation over noble metal based catalysts is that oxidation of VOCs is considered to be structure sensitive reaction. At the base of fixed active components, turnover rates could be changed by the structure of catalysts. The influence of particle size on the catalytic combustion of different VOCs has been reported^[4]. While as to ethanol oxidation over palladium catalyst, more efforts have to be taken to understand mechanism of ethanol oxidation for the connection between catalytic performance and catalyst properties depends mainly on the type of VOC.

In our previous work, nano Al₂O₃-TiO₂ prepared by sol-gel method has been utilized as the support of Pd catalysts. And it found that Pd/Al₂O₃(5%)-TiO₂ catalyst showed a higher catalytic activity than those of either Pd/TiO₂ or Pd/Al₂O₃ for the total oxidation of ethanol. Nonetheless, only limited research attention has been devoted to coupling mesoporous TiO₂ with Al₂O₃, and conducted the resulting titania-composite materials as the carrier of supported catalysts^[15]. In this work, the template and sol-gel methods were employed for the preparation of mesoporous TiO₂-Al₂O₃. Palladium as active component was impregnated on the composite support, and the activity and yield of CO₂ of this catalyst produced by ethanol combustion was de-monstrated.

1 Experimental 1.1 Preparation of Al₂O₃ modified porous TiO₂

Titania and alumina sols were prepared separately, and mixed together to obtain porous Al₂O₃-TiO₂ (95%). The preparation details of is alumina sol was described in our previous work^[16]. Titania sol was synthesized by a traditional template method^[17]. Ti (OC₄H₉)₄ was added to a HNO₃ aqueous solution slowly under vigorous stirring at 313 K to get a transparent solution. Alumina sol was added into the system, and kept stirring for another 1 h. The transparent sol was separated from hydrolysis products and aged at ambient temperature for 3 h. Ethanol solution of dodecylamine was added into the transparent sol under vi-gorous stirring and then obtained a transparent sol which was treated with ultrasonic for 1 h, aged at 353 K for 24 h in a closed flask. The resulting white gel was transferred to a Petri dish and dried at 353 K for another 24 h to give a yellow powder. The yellow powder was rinsed with absolute ethanol and subsequently dried at 353 K overnight. The products were calcined from 523 to 673 K. Pd was loaded on the carriers by a wet-impregnation method, and all samples contained 1% Pd theoretically. The samples prepared in different conditions were designed as Pd/AT-n, Pd/T-n, in which n denotes different calcination temperatures, AT denotes the resulting titania-composite Al₂O₃-TiO₂, and T denotes single TiO₂.

1.2 Catalyst evaluation

The catalytic properties of the catalysts were eva-luated using a fixed-bed reactor connected with an online gas chromatograph. The compositions of ethanol and products were analyzed using a GC-950 apparatus with two FID detectors, one of which incorporated a methanator. Approximately 0.30 g of catalyst was used (0.450~0.280 mm), and the feed gas consisted of 0.5% C₂H₅OH, 5% air, and a balance of N₂. The flow rate was 100 mL/min of total gas and the gas hourly space velocity (GHSV) was 24 000 h^-1. Ethanol conversion (X_i) and the yields (Y_j) of by-pro-ducts and carbon dioxide from ethanol combustion were calculated according to the formulas:

${X_i} = \frac{{{C_{{\rm{i, in}}}}-{C_{{\rm{i, out}}}}}}{{{C_{{\rm{i, in}}}}}}\;\;\;\;{Y_{\rm{j}}} = \frac{{{C_{j, {\rm{out}}}}}}{{{C_{{\rm{i, in}}}}}} \times \frac{1}{n} \times 100\% $

Where C_i and C_j are the concentrations of ethanol and products from ethanol combustion, respectively, n is the products and reactants carbon atom ratios. And

${C_{\rm{i}}} = {A_{\rm{i}}} \times {f_{\rm{i}}}\;\;\;{C_{\rm{j}}} = {A_{\rm{j}}} \times {f_{\rm{j}}}$

where A and f are the peak areas and the relative correction factors of reactants (i) or products (j), respectively.

1.3 Characterizations

Low-angle X-ray diffraction (XRD) patterns of all samples were collected in θ-2θ mode using Rigaku D/MAX-2550 apparatus (Cu K_α1 radiation, λ=1.540 6 Å), operated at 40 kV and 200 mA. Wide-angle XRD patterns were collected in the same mode, but operated at 100 mA. Intensities of the diffraction peaks were recorded in the 2θ range 5°~80° with a step size of 0.01°, and the scanning speed was 8 °/min.

Transmission electron micrographs (TEM) were carried out on a Tecnai G²F20, operating at 200 kV. The sample was diluted in ethanol and dropped onto a copper grid coated with a thin film of carbon.

The porous texture of the samples was analyzed by N₂-adsorption at 77 K. By using a Micromeritics ASAP 2000 system, the BET and BJH methods were applied for the determination of the specific surface area, and the mean mesoporous equivalent diameter, respectively. Before measurement, samples were evacuated at 473 K.

FT-IR analysis was carried out using NICOLET360 type spectrometer. The spectra of solids were obtained using KBr pellets. Before measurement, the catalyst and potassium bromide were mixed together in accordance with the quality ratio 1:100. The vibrational transition frequencies are reported in wave numbers (cm^-1).

XPS analysis was carried out on a Kratos Axis Ultra DLD multifunctional surface analysis photoelectron spectrometer with X-ray source at 450 W and Al-Kα radiation (1 486.6 eV). The element binding energies were referenced to the C 1s line at 284.6 eV and measurements were restricted an estimated error of ± 0.1 eV.

The Pd dispersion was determined by pluse H₂ chemisorption on a Micrometrics AutoChem Ⅱ 2920 instrument with a TCD. Before measurement, 100 mg sample was pretreated by helium flow for 1h at 200 ℃ to remove gaseous species adsorbed, and then cooling down to room temperature. The H₂/Pd average stoichiometry of 1 has been assumed for the calculation of dispersion^[18].

NH₃-TPD-MS experiment was performed on a TP-5080. Prior to experiment, 100 mg catalyst was pretreated at 300 ℃ for 30 minutes in He atmosphere to remove the adsorbed matter and then decreased to 50 ℃ to adsorb ammonia. Physically adsorbed ammonia was removed by purging with He. And then the temperature was increased linearly at 10 ℃/min from 50 to 810 ℃ to desorbed chemisorbed NH₃ and the gas coming from the reactor be analyzed by QIC mass spectrometry detector.

2 Results and discussion 2.1 Phase and morphology analysis

A series of composite Al₂O₃-TiO₂ porous materials and palladium based catalysts were synthesized by a template associated wet-impregnation method. Evidence for template removal of the as-prepared samples has been confirmed by FT-IR (Fig. 1).

Their low-angle XRD patterns are shown in Fig. 2a. The low-angle XRD patterns of Pd/T-150, Pd/AT-150, Pd/T-250, Pd/AT-250 appear only one broad peak respectively. The peaks at the lower angle suggest the existence of a large lattice plane distance (D value), which correspond to the mesoporous structure with not well-ordered pores in these materials. As only one peak is observed in the low-angle range, the mesoporous structure of the samples is typically of long-range order in nature^[19]. In low-angle XRD patterns of mesoporous silica HMS with wormhole framework, only one broad peak could be observed, which was similar to that of our sample^[20]. So the samples were not layered structures but wormhole framework structures and that has been proven by further characteristic method TEM (Fig. 2a insert). The XRD pattern of catalysts calcined at 400 ℃ does not show any peak, indicating no mesoporous structure in the samples^[21]. This means that higher calcination temperature will lead to the collapse of the pore structure.

In order to confirm crystal structure of the synthesized mesoporous carriers and catalyst calcined at 250 ℃, XRD analysis was performed in a wide angle range, as shown in Fig. 2b. As seen in the patterns, the characteristic peaks of anatase phase appear in all samples, while the brookite phase appears in the single component carrier T-250 and the composite oxide carrier AT-250. It also can be seen that the intensity of the diffraction peak of brookite becomes weaker when Al₂O₃ was introduced into mesoporous TiO₂, which means that the addition of Al₂O₃ would restrain the generation of brookite TiO₂. No peak correlated to γ-Al₂O₃ or PdO was observed, probably due to the relatively low contents and/or high dispersion of γ-Al₂O₃ and PdO over the catalysts^[22].

2.2 N₂ adsorption analysis

In order to further examine the porous structure of supports and catalysts calcined at 250 ℃, N₂ adsorption-desorption isotherms were employed. According to the IUPAC classification^[23], the adsorption isotherms of all samples as shown in Fig. 3 exhibit the typical Ⅳ isotherm, which indicates the existence of mesoporous structure. Moreover, a clear absorption hysteresis loop located at 0.40 < P/P⁰ < 1.0 and a flat "S"-shape is observed in the isotherms of T-250 and AT-250, which belongs to the H1 type hysteresis loop, suggesting the presence of uniform pores in the samples. The curves of Pd/T-250 and Pd/AT-250 belongs to type H2, suggesting that the presence of slit-like pores in the samples. Fig. 3 insert shows the pore size distribution curve calculated by the BJH method from the desorption branch of a nitrogen isotherm. It can be seen that the diameter distribution of samples at 250 ℃ mainly concentrated in the ranges of mesopores from 3 to 20 nm, which can be concluded that the samples have good mesoporous structure.

The BET surface areas, pore sizes and pore volumes of samples are summarized in Table 1. Without Al₂O₃, the BET surface areas of T-250 and Pd/T-250 were 187 and 127 m²·g^-1, respectively. The surface areas of samples AT-250 and Pd/AT-250 were increased to 256 and 206 m²·g^-1 respectively when Al₂O₃ was added. This may due to Al₂O₃ itself has a larger surface area. Meanwhile, the surface area of Pd/AT calcined at 400 ℃ sharply decreased to 69 m²·g^-1, that could be ascribed to the higher calcination temperature, resulting in the collapse of the pore structure.

2.3 Pd dispersion

Pd dispersion of different catalysts was shown in Fig. 4. It can be seen that the dispersion of palladium on the Pd/AT-250 was higher than that on Pd/AT-150 and Pd/AT-400, and the change feature is consistent with their surface area. It is known that higher dispersion of active components resulted in higher thermal stability and catalytic performance.

2.4 XPS

We carried out XPS to determine chemical states of elements Ti, O and Pd in pure TiO₂, Al₂O₃-TiO₂, catalysts Pd/TiO₂ and Pd/Al₂O₃-TiO₂ calcined at 250 ℃ and shown in Fig. 5. When compared Pd/Al₂O₃-TiO₂ with Pd/TiO₂ calcined at 250 ℃, there is no significant difference of chemical state of palladium species. And the increased catalytic performance of ethanol oxidation over palladium catalysts could be ascribed to Pd dispersion.

As shown in Ti 2p, the main peak of Ti 2p_3/2 on the T-250 is located at 458.4 eV, which can be ascribed to Ti⁴⁺ of TiO₂. In comparison, the corresponding peak showed a negative shift of about 0.4 eV over the AT-250. The shift of Ti 2p_3/2 peak to a lower position suggests that particle Ti⁴⁺ might be existed in lower chemical state like Ti³⁺ for the introduction of Al₂O₃. It has been proven that at the presence of oxygen vacancies assistant with TiO_x (x < 2), content of chemisorption oxygen can be enhanced^[2].

According to asymmetry of O 1s spectra, we use peak fitting to further figure out chemical state of oxygen species at the surface of palladium catalysts. The main peak at around 530.0 eV could be attributed to the lattice oxygen of TiO₂. And the O_Ⅱ could be attributed to chemisorption oxygen. It can be seen in the O1s of catalyst Pd/T-250 and Pd/AT-250, the peak position of O_Ⅰ shifts to lower energy by about 0.4 eV compared with Pd/T-250, and this is consistent with Ti 2p curves. And the content of chemisorption oxygen over Pd/AT-250 was significantly larger than that of Pd/T-250. The well-dispersed palladium species and rich chemisorbed oxygen is closely related to the activity of Pd catalysts on ethanol oxidation.

2.5 NH₃-TPD measurement

To analyze acidity in sample, NH₃-TPD profiles corresponding to the sample calcined at 150 and 250 ℃ is shown in Fig. 6. It is obviously to see a desorbed peak at low temperature which was related to weak acidity^[24]. And the peak area is almost unchanged between different catalysts indicating the weak acidity is consistent with each other. While samples calcined at 150 ℃ shows a broad desorption peak at temperature range 400~600 ℃ which is attributed to strong acidity resulted from the residual template under low calcination temperature. And the peak area was a little different between Pd/TiO₂ and Pd/AT which was for the different acidity of Al₂O₃ and TiO₂. Reference has shown that catalyst acidity was introduced by support than metal. Additionally, for supported Pd catalyst, the larger amount of strong acidity would lead to difficulty in Pd oxidation and restrain ethanol oxidation^[25]. And this could be an aspect for lower catalytic performance of catalysts calcined at 150 ℃.

2.6 Catalytic activity evaluate

The ethanol catalytic oxidation activity over different Pd catalysts was evaluated. Evidence of by-product yields has shown the route of ethanol oxidation among various catalysts (Fig. 7c-f). The first step is ethanol dehydrogenized into acetaldehyde, which was then at the presence of oxygen further transformed into CO and CO₂. It is found that the difference between various catalysts lies in the yield of CO. CO yield on Pd/AT-250 is apparent lower than that on any other catalysts and that could be due to the high specific area. The catalytic performance of different catalysts was investigated by detecting the oxidation proportions of ethanol and yield to CO₂ and the results are shown in Fig. 7. It is found that ethanol conversion and yield to CO₂ on Pd/TiO₂-Al₂O₃ were higher than that on the others, and that might be due to the higher surface area and the relatively higher Pd dispersion. From the profiles in Fig. 7(a), the significant decrease of ethanol conversion and CO₂ yield on Pd/TiO₂ when reaction temperature was higher than 150 ℃(calcined temperature) could also be of another proof for the destroyed structure and decreased surface area.

The catalyst Pd/AT-250 exhibited the highest activity and achieved complete oxidation of ethanol at 275 ℃. While at reaction temperature below 150 ℃, it can be seen that the conversion of ethanol over catalysts calcined at 150 ℃ (Pd/T-150) is higher than that at higher calcination temperature. Calcination can help remove organic components and also do harm to porous structure. The more ordered porous might be beneficial for reactant and surface active sites interaction. Additionally, the introduction of aluminium could restrain crystalline growth. And efforts have been paid to prepare anatase at low temperature for the stable phase^[26]. In general, anatase is more active than rutile and brookite for catalytic activity. So it can be proposed that the Pd/T-150 with relatively higher catalytic performance than that on the others is due to the stable anatase and the relation between anatase and surface active sites. While it can be seen the extremely lower performance when reaction temperature is higher than 200 ℃ over Pd/T-150. High temperature treatment would cause residual organic species diffusion to form compound resulting in sintering of active sites and deteriorating catalytic activity^[26]. Fig. 7(b) presents the yield of CO₂ over Pd-based catalysts calcined at diffe-rent temperatures, which was constant with catalytic activity of ethanol conversion. The catalyst Pd/AT-250 had the best yield to CO₂, and its yield to CO₂ reached 80% at 275 ℃. The porous TiO₂ supported Pd catalyst with high BET surface area can improve the dispersion of surface Pd active sites. The higher dispersed active sites can improve catalytic efficiency towards ethanol convert to CO₂. Moreover, the abundant chemisorbed oxygen atoms can enhance the relation between oxethyl produced at the first step of ethanol oxidation reaction and oxygen to form CO₂.

2.7 Stability test

The stability test under different temperature and different time of ethanol oxidation over Pd/AT-250 and Pd/T-150 was shown in Fig. 8. The flow rate was 100 mL/min of total gas and the gas hourly space velocity (GHSV) was 24 000 h^-1.

The test temperature range is 75~500 ℃. For catalyst used to remove emission of vehicle, it should maintain high catalytic performance at high temperature and long work time. From Fig. 8a, we can see the temperature of total oxidation of ethanol is 275 ℃ over Pd/AT-250 while that over Pd/T-150 is 325 ℃ and it decreased obviously as the temperature increased continuously. It is noting that at temperature range of 275~400 ℃, the higher catalytic performance towards ethanol conversion and CO₂ yield could be maintained over Pd/AT-250 which means that the relatively higher thermostability of Pd/AT-250, while as the temperature is higher than 400 ℃, the ethanol conversion and CO₂ yield is decreased. The decreased thermostability of Pd based catalyst could be attributed to the relatively lower support structure, and higher reaction temperature leads to carbon deposition and active sites sintering. From Fig. 8b, we carried out stability test at different time. It could be found that at reaction time ethanol conversion decreased from 98.7% to 96.2% and the average decreased rate is 0.25%/h^-1 which revealed the higher stability.

While the ethanol conversion and CO₂ yield over Pd/T-150 is decreased from 84.5% to 8.29% which means the lower stability, revealing that the increased calcination temperature and the introduction of alumi-nium could improve the stability of Pd/TiO₂ catalyst.

3 Conclusions

A series of wormhole-like porous material Al₂O₃-TiO₂ and Pd/Al₂O₃-TiO₂ catalysts have been synthesized by template-impregnation technique. The catalyst Pd/Al₂O₃-TiO₂ calcined at 250 ℃ exhibited the highest activity and yield of CO₂ and that can be attributed to the high surface area, larger pore size, higher-dispersed palladium species and the abundant chemisorbed oxygen.

Acknowledgements: Authors acknowledge the financial support from Project supported by the State Key Program of National Natural Science of China (Grant No. 21336006), and Shanxi Province Scientific and Technological Project (20140313002-2).