Acta Physico-Chimica Sinica  2017, Vol. 33 Issue (3): 602-610   (2238 KB)
 Article Options PDF (2238 KB) Full Text HTML Abstract Figures References History Received: August 26, 2016 Revised: November 25, 2016 Published on Web: November 25, 2016 Service Email this article to a colleague Add to my bookshell Add to citation manager Email Alert Feedback View Feedback Author GAO Xiao-Ping GUO Zhang-Long ZHOU Ya-Nan JING Fang-Li CHU Wei
Catalytic Performance and Characterization of Anatase TiO2 Supported Pd Catalysts for the Selective Hydrogenation of Acetylene
GAO Xiao-Ping1,2, GUO Zhang-Long1,2, ZHOU Ya-Nan1,2, JING Fang-Li1, CHU Wei1,2,*
1 School of Chemical Engineering, Sichuan University, Chengdu 610065, P. R. China;
2 Institute of New Energy and Low-Carbon Technology, Sichuan University, Chengdu 610207, P. R. China
Abstract: Anatase TiO2 nanospindles containing 89% exposed {101} facets (TiO2-101) and nanosheets with 77% exposed {001} facets (TiO2-001) were hydrothermally synthesized and used as supports for Pd catalysts. The effects of the TiO2 materials on the catalytic performance of Pd/TiO2-101 and Pd/TiO2-001 catalysts were investigated in the selective hydrogenation of acetylene to polymer-grade ethylene. The Pd/TiO2-101 catalyst exhibited enhanced performance in terms of acetylene conversion and ethylene yield. To understand these effects, the catalysts were characterized by H2 temperature-programmed desorption (H2-TPD), H2 temperatureprogrammed reduction (H2-TPR), transmission electron microscopy (TEM), pulse CO chemisorption, X-ray photoelectron spectroscopy (XPS), and thermogravimetric analysis (TGA). The TEM and CO chemisorption results confirmed that Pd nanoparticles (NPs) on the TiO2-101 support had a smaller average particle size (1.53 nm) and a higher dispersion (15.95%) than those on the TiO2-001 support (average particle size of 4.36 nm and dispersion of 9.06%). The smaller particle size and higher dispersion of Pd on the Pd/TiO2-101 catalyst provided more reaction active sites, which contributed to the improved catalytic activity of this supported catalyst.
Key words: Pd/TiO2 catalyst     Acetylene selective hydrogenation     Anatase TiO2     {101}plane     Structure characterization

1 Introduction

Pd-based catalyst is industrially used for the acetylene selective hydrogenation to remove trace amount of acetylene from ethylene feed stream in the commercial production of polymer-grade polyethylene1, 2. However, most of the supported Pd catalysts show poor selectivity and stability due to strong adsorption of reactant and product on contiguous Pd sites3, 4. Several attempts have been considered to improve its selectivity and stability such as (ⅰ) inducing the strong metal-support interaction (SMSI) effect to weaken the adsorption strength of ethylene on Pd surface5, 6, (ⅱ) adding a second metal (e.g. Ag 7-11, Zn 12, Ga 13, 14, In 15, 16) to form alloy with Pd or to suppress the multi-coordination sites of the Pd surface, (ⅲ) inducting inert material (e.g. carbonaceous deposits formed by pretreatment with feed gases17), (ⅳ) pretreating by plasma18-20. Previous studies have showed that Ti3+ species on the support surface could have contact with the Pd nanoparticle surfaces, hereby leading to the SMSI effect21, 22. Mobility of the Ti3+ from the lattice of TiO2 to the surface of Pd particles is usually facilitated by reduction at high temperature. According to reports in the literature, the specifically exposed planes of support play a crucial role in determining the metal-support interaction and catalytic behavior due to rather different atomic species, electron density and coordination environment of various facets23. Some publications have clarified that the exposed facets of the support nanocrystals could exert a profound influence on the catalytic performances24, 25.

Recent progress in the synthesis of anatase TiO2 nanomaterials enables to select the exposure of desirable crystal planes, and thus benefits more detailed studies on the catalytic behavior of supported metal nanoparticles on TiO2. For example, Ru nanoparticles loaded on {101} facets of TiO2 nanoparticles exhibited almost double higher turnover frequency in CO2 methanation than those over {001} facets of TiO2 nanosheets26. The {101} planes displayed a much stronger interaction with Ru nanoparticles than the {001} planes, which enhanced the adsorption and activation of CO2 and H2 molecules. TiO2 nanosheets and nanospindles were applied to disperse vanadia species as well. The {001} facets of TiO2 nanosheets benefited the deposition of octahedral vanadia species, whereas the {101} facets of TiO2 nanospindles resulted in the generation of tetrahedral vanadia species. Octahedrally coordinated vanadia species on TiO2 nanosheets showed a much higher activity in selective reduction of NO with NH3 mainly because of the existence of more V=O sites and V-O-V links27. The {100} facets of TiO2 promoted activation of O2 and the formation of Auδ+ and improved the catalytic activity for CO oxidation28. A theoretical study based on density functional theory calculations with a Hubbard U correction (DFT + U) reveals that the catalytic activity of selective hydrogenation of acetylene on oxygen defect surface is much higher than on the perfect one when Pd4 cluster supported on the anatase TiO2(101) surfaces29. However, the influence of the exposed crystal planes of TiO2 on the catalytic behavior of Pd nanoparticles for acetylene selective hydrogenation to ethylene, which is key to enhance the catalytic efficiency of noble mental from the viewpoint of electronic structure, have been not reported.

The objective of this report was composed in three aspects: firstly to synthesize TiO2 nanosheets mainly with the {001} facets or nanospindles mainly with the {101} facets in the presence of F- or Cl- as morphology-directing agents, respectively; secondly to load Pd NPs on the different shapes of TiO2 carriers and characterization of the as-prepared supports and catalysts; thirdly to investigate the effects on acetylene selective hydrogenation to ethylene on these catalysts, and supporting characterization evidences for explanation of the better performance of Pd/TiO2-101 sample. In detail, the samples were characterized by X-ray diffraction (XRD), Raman spectra, electron spin resonance (ESR), high-resolution transmission electron microscope (HRTEM), Xray photoelectron spectroscopy (XPS), N2 adsorption-desorption, and hydrogen temperature-programmed desorption (H2-TPD), hydrogen temperature-programmed reduction (H2-TPR). In addition, the connection between surface properties of catalysts and catalytic behavior was also investigated.

2 Experimental
2.1 Catalysts preparation

The chemicals (Chengdu Kelong Chemical Reagent Co., Ltd.) are analytical grade and are utilized without further purification. The dominated {001} facets of anatase TiO2 nanosheets were fabricated by hydrothermal method as reported in literature30. Typically, 25 mL of Ti (OBu)4 and 4 mL of HF solution (40% (w, mass fraction)) were mixed in a dried Teflon-lined autoclave with a volume of 100 mL and then kept at 180 ℃ for 24 h. After being cooled down to ambient temperature, the hydrothermal product was washed with 0.1 mol∙L-1 NaOH aqueous solution to remove fluorine. Then, the white powder was filtrated, washed with ethanol as well as deionized water several times, and dried at 80 ℃ for 6 h. Finally, the obtained sample was denoted as TiO2-001.

Anatase TiO2 nanoparticles with exposed {101} facets were prepared by a method reported by Liu et al.28. TiCl4 (6.6 mL) was added dropwise into a 0.43 mol∙L-1 aqueous solution of HCl in an ice bath with magnetic stirring. The resulting clear TiCl4 solution was then added into NH3 ∙H2O solution (5.5% (w)) to generate the white precipitate Ti (OH)4. This system was kept pH around 6-7 by the addition of 10 mL aqueous NH3∙H2O (4% (w)). This precipitate was recovered by filtration and washing, after aging for 2 h at ambient temperature. The resulting Ti (OH)4 precursor and 0.4 g NH4Cl dissolved in a mixture containing 30 mL of water and 30 mL of isopropyl alcohol. A suspension was gained after treating with stirring and ultrasonic. Then, the suspension was transferred to a 100 mL Teflon-lined autoclave and kept for 24 h at 180 ℃. Finally, the powers were filtrated and washed by water as well as ethanol until there was no Cl- in the mother solution determined by aqueous AgNO3 (0.05 mol∙L-1). The obtained sample was named as TiO2-101.

Pd nanoparticles were loaded onto TiO2 nanocrystals (TiO2-101 and TiO2-001) through impregnation. The loading of Pd is 1% (w). TiO2 nanocrystals (1 g) were put into an aqueous of PdCl2 (0.0229 mol∙L-1, 4.15 mL) at ambient temperature for 4 h with stirring. After drying at 60 ℃ for 12 h, the derived powers were treated at 400 ℃ for 3 h under N2 atmosphere. The gained catalysts were labeled to as Pd/TiO2-001 and Pd/TiO2-101, respectively. Elemental analysis by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) reveals that the content of Pd is 0.56% (w) for Pd/TiO2-001 sample and 0.57% (w) for Pd/TiO2-101 sample, which are lower than the original theoretical content.

2.2 Characterization of catalysts

The power X-ray diffraction patterns were recorded on DX-2700 diffractometer (Haoyuan, China) using Cu radiation at 40 kV and 30 mA. The 2θ scanning range was from 10° to 80° with a scan step of 0.03 (°)∙s-1 in a continuous mode.

Morphologies were analyzed on a Tecnai G2F20 transmission electron microscope (TEM). The lattice fringes of the catalysts were characterized through a high-resolution transmission electron microscope (HRTEM). The samples were crushed and dispersed in ethanol, and the resulting suspensions were allowed to dry on carbon film supported on copper grids.

The N2 adsorption-desorption isotherms were measured at -196 ℃ using an automated surface area & pore size analyzer (Quantachrome NOVA1000e apparatus). The specific surface area was calculated by the Brunauer-Emmett-Teller (BET) equation.

The TiO2 samples were also investigated by Raman spectroscopy. Radiation of 532 nm from an argon-ion laser was used for excitation. The instrument version is LabRAM HR800.

Electron spin resonance spectroscopy was conducted under vacuum at -150 ℃ using a JES-FA200 electron spin resonance spectrometer. It was performed to qualitatively monitor the Ti3+ species on the surface of the TiO2.

The hydrogen temperature-programmed reduction measurement was performed in a fixed-bed reactor at atmospheric pressure. 50 mg sample was loaded in the middle of reactor tube, and the reductive gas of 5% H2/N2 with a total gas flow rate of 30 mL∙min-1 was introduced. The system was kept at 30 ℃ for 1 h until the baseline was stable, and then it was heated linearly from 30 to 600 ℃ at a heating rate of 10 ℃∙min-1. The H2 uptake amount during the reduction was recorded by gas chromatograph (SC-200) equipped with a thermal conductivity detector (TCD). Prior to hydrogen temperature-programmed desorption (H2-TPD), 100 mg of catalysts were heated at 120 ℃ for 2 h in nitrogen and then placed in H2 with a flow rate of 23 mL∙min-1 for 0.5 h at 30 ℃. TPD was carried out in a stream of nitrogen with a flow rate of 40 mL∙min-1 and a temperature ramp of 10 ℃∙min-1.

Pulse CO chemisorption was employed to determine Pd dispersion on ChemStar TPx chemisorption analyzer. Prior to CO adsorption measurements, the Pd/TiO2 samples were purged in helium at room temperature for 30 min. The system was switched to H2 (30 mL∙min-1) and heated to 400 ℃ with a heating rate of 5 ℃∙min-1. CO pulses were injected (50 μL of 10% of CO in helium) from a calibrated loop over the Pd/TiO2 catalysts at 30 ℃ and repeated until the desorption peaks were constant. The number of exposed Pd atoms on the Pd/TiO2 catalyst surface was calculated by the total amount of CO adsorption. In this paper, the CO/ Pd stoichiometry of 1 is used for calculation.

The X-ray photoelectron spectroscopy data were collected on XSAM800 spectrometer with an Al (=1486.6 eV) X-ray source. The charging effectswere corrected by adjusting the binding energy of C 1s peak from carbon contamination to 284.6 eV.

The amount of carbonaceous deposited on used catalysts was determined by running thermo-gravimetric analysis (on TGA Q500) in air atmosphere (40 mL∙min-1). The temperature was first kept at 30 ℃ for 30 min, and then increased to 700 ℃ with a ramp of 10 ℃∙min-1. Results of derivative thermogravimetry (DTG) were obtained from the thermo-gravimetric analysis (TGA) curves by differentiating the latter with respect to temperature.

2.3 Selective hydrogenation of acetylene

Acetylene selective hydrogenation reaction was performed in a fixed-bed reactor in the temperature range from 40 to 80 ℃ at the atmospheric pressure with a gas hourly space velocity (GHSV) of 30000 mL∙g-1∙h-1. The gaseous feed with a total flow of 50 mL∙min-1 contained 1% C2H2, 2% H2 and the balance N2. 0.1 g catalyst was diluted by 0.4 g quartz sand for the sake of avoiding temperature and concentration gradients. Prior to the reaction, the catalysts were pretreated in hydrogen (20 mL∙min-1) at 400 ℃ for 2 h. Before sampling, the reaction temperature was kept constant for 0.5 h before being raised to the next one. The reactants and products were detected by gas chromatography (GC) equipped with a flame ionization detector (FID). For the sake of reproducible data, five tests were carried out. Conversion of acetylene, selectivity to ethylene, selectivity to ethane, yield towards ethylene, and carbon balance (Bc) are calculated as follows31:

 $\begin{array}{*{20}{l}} {{\rm{Conversion}} = \frac{{[{{\rm{C}}_2}{{\rm{H}}_2}]({\rm{inlet}}) - [{{\rm{C}}_2}{{\rm{H}}_2}]({\rm{outlet}})}}{{[{{\rm{C}}_2}{{\rm{H}}_2}]({\rm{inlet}})}} \times 100\% }\\ {{{\rm{C}}_2}{{\rm{H}}_4}\;{\rm{Selectivity = }}\frac{{[{{\rm{C}}_2}{{\rm{H}}_2}]({\rm{outlet}})}}{{[{{\rm{C}}_2}{{\rm{H}}_2}]({\rm{inlet}}) - [{{\rm{C}}_2}{{\rm{H}}_2}]({\rm{inlet}})}} \times 100\% }\\ {{{\rm{C}}_2}{{\rm{H}}_4}\;{\rm{Selectivity = }}\frac{{[{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{6}}}]\left( {{\rm{outlet}}} \right)}}{{[{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{2}}}]({\rm{inlet}}) - [{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{2}}}]\left( {{\rm{outlet}}} \right)}} \times 100\% }\\ {{\rm{Yield = Conversion}} \times {\rm{Selectivity}}}\\ {{B_{\rm{c}}} = \frac{{[{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{2}}}]({\rm{outlet}}){\rm{ + }}[{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{4}}}]\left( {{\rm{outlet}}} \right){\rm{ + }}[{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{6}}}]\left( {{\rm{outlet}}} \right)}}{{[{{\rm{C}}_{\rm{2}}}{{\rm{H}}_{\rm{2}}}]({\rm{inlet}})}} \times {\rm{100}}\% } \end{array}$

where [C2H2](inlet) and [C2Hx](outlet) (x=2, 4, 6) in the formulas represent the mole concentrations.

3 Results and discussion
3.1 Crystalline structure of TiO2 nanoparticles

To demonstrate the crystal structure of the TiO2 materials and the Pd/TiO2 catalysts, XRD analysis was carried out and the corresponding results are exhibited in Fig. 1. All samples displayed several typical characteristic peaks attributed to the anatase TiO2 phase (JCPDS # 21-1272, space group: I41/amd (141))32, 33. Obviously, the sample TiO2-001 (curve C) exhibits relatively stronger diffraction peak at (200) than that at (004) reflection, indicating a predominant exposure of the {001} facets30, 34.Whereas the TiO2- 101 (curve A) displays a decrease in the (200) reflection and an improvement on intensity of diffraction peak at (004), implying the occupancy of the {101} planes27. Moreover, no obvious change in reflections of Pd/TiO2-101 (curve B) and Pd/TiO2-001 catalysts (curve D) is observed when compared with the corresponded pristine supports, implying the TiO2 nanoparticles remained in the original structure and morphology26. It should be noted that there is no X-ray diffractions of Pd species, which is because of the comparatively low loading amount (1% (w), below the detection limit of XRD) of Pd. The TiO2 materials are further analyzed by Raman spectra (Fig.S1 (in Supporting Information)) and the results were consistent with the results from XRD analysis.

 Fig. 1 XRD patterns of TiO2-101 (A), Pd/TiO2-101 (B), TiO2-001 (C) and Pd/TiO2-001 (D)

The morphology of the TiO2 nanoparticles is studied by TEM. As shown in Fig. 2, TiO2-001 sample exhibits uniform sheet-like shape while the sample TiO2-101 shows spindle-like shape. The average thickness and side length for the sample TiO2-001 are approximately 5 and 40 nm (Fig. 2(a)), respectively. Fig. 2(b) shows that a well-defined sheet structure was observed, which had an interplane spacing of 0.235 nm. All of these features implied that the exposed planes of anatase TiO2 is the {001} facets. In contrast, TiO2-101 sample has a spindle-like shape with an average size of 15.7 nm long and 10 nm wide (Fig. 2(c)); from the side view, the interplanar spacing of 0.35 nm is corresponded with the {101} planes of anatase TiO2 (Fig. 2(d)). On the basis of Wulff construction (Fig.S2(a) and Fig.S2(b) (in Supporting Information)), the percentage of each crystalline facet in the applied samples is calculated. In signal-crystalline TiO2 nanospindles, the {101} facets are the dominant facets with the ratio of 89% and the other 11% is the {001} facets, while the proportion of {001} planes as well as {101} planes in TiO2 nanosheets are 77% and 23%, respectively. Hence, in our case, the percentage of each crystalline facet is beyond 75%, demonstrating that the applied TiO2 nanomaterials can serve as model supports28.

 Fig. 2 TEM, HRTEM images of TiO2-001 (a, b) and TiO2-101 (c, d)

3.2 Catalytic performances in selective hydrogenation of acetylene

The catalytic behavior of the Pd/TiO2-101 and Pd/TiO2-001 catalysts was evaluated by using partial hydrogenation of acetylene to ethylene as probe reaction under the employed reaction conditions. The catalytic performances of both samples are represented in Fig. 3, showing globally an increase of acetylene conversion with the increasing reaction temperature from 40 to 80 ℃ (Fig. 3(a)). The Pd/TiO2-101 catalyst exhibits higher acetylene conversion until the reaction temperature reached at 70 ℃ (100% for Pd/TiO2-101 vs 94% for Pd/TiO2-001) after which full conversion was obtained over the two samples. It is worth noting that the conversion of acetylene is as high as 92% over Pd/TiO2- 101 while only 50% is gotten over Pd/TiO2-001. Owing to its preferred catalytic performance, Pd/TiO2-101 catalyst surely displays the excellent yield in ethylene of 57% at 60 ℃, which is 1.9 times higher than those over Pd/TiO2-001 catalyst (Fig. 3(b)). The preferable catalytic activity of the Pd/TiO2-101 catalyst might be assigned to its large specific surface area35, high Pd dispersion36, 37 leading to the formation of more active centers. The selectivity toward ethylene on both catalysts decreased with the increasing conversion (Fig.S3 (in Supporting Information)), which is due to the fact that the ethylene is produced as an intermediate in acetylene semi-hydrogenation reaction. Furthermore, Pd/TiO2- 101 catalyst shows higher selectivity in ethylene when compared with that of the Pd/TiO2- 001 catalyst on the basis of equal conversion. The selectivity to ethane (Fig.S4(a) (in Supporting Information)) increases with the increasing temperature, especially after 60 ℃. As shown in Fig.S4(b) (in Supporting Information), the carbon balance in both cases is close to 100%.

 Fig. 3 Acetylene conversion as a function of temperature (a) and ethylene yield versus reaction temperature (b) over Pd/TiO2-001 and Pd/TiO2-101 catalysts at a total flow rate of 50 mL∙min-1 with varying reaction temperatures from 40 to 80 ℃

3.3 Texture properties of the catalysts

In order to explain the difference in the catalytic activity of the two Pd/TiO2 catalysts, the structure and size of the Pd nanoparticles were characterized by TEM and HRTEM, and the results are shown in Fig. 4. The HRTEM image of Pd/TiO2-101 (Fig. 4(d)) shows that Pd NPs distribute homogeneously without apparent accumulation, while some Pd NPs on the Pd/TiO2-001 catalyst accumulate after reduction (Fig. 4(a)). More than 100 Pd nanoparticles are obtained from different regions, randomly selected for the sake of getting the average size of Pd and the results are displayed in the histograms (Fig. 4(c, f)).As we can see, the Pd/ TiO2-101 catalyst has a narrow size distribution in the range from 1.00 to 2.20 nm and the average size of Pd particles is 1.53 nm which is rather small than that of Pd/TiO2-001 catalyst (4.36 nm). It is obviously observed that the dispersion of the Pd/TiO2-101 catalyst is higher than that of the Pd/TiO2-001 catalyst. This result is well agreement with the CO chemisorption (Table 1). Remarkably, the narrower size distribution, smaller average particle size and higher dispersion of the Pd/TiO2-101 catalyst can be ascribed to two factors. The specific surface areas of TiO2-101 were significantly larger than that of TiO2-001. The BET specific surface areas, nitrogen adsorption-desorption isotherms and BJH pore-size distributions of the samples are given in Fig.S5 and Table S1 (in Supporting Information). Generally, the higher surface area of the TiO2-101 could contribute to higher dispersion of Pd NPs, which facilitates the enhancement of the catalytic activity38, 39. This result was accordance with the tendency of catalytic behavior in Fig. 3. Besides, as reported in the literature, high dispersion of Pd NPs can also be attributed to the strong metal-support interaction effect40, 41. The following section will demonstrate this effect of the Pd/TiO2-101 catalyst.

 Fig. 4 TEM images and the corresponding Pd particles size distributions of Pd/TiO2 catalysts (a, b, c) Pd/TiO2-001 and (d, e, f) Pd/TiO2-101

Table 1 Properties of the Pd/TiO2-001 and Pd/TiO2-101 catalysts

Based on the reported results in literature42, we can conjecture that Ti3+ species combined with Pd surface in the interface might play the part of a new reaction site which can greatly improve hydrogen activation as well as its dissociation. Although H2-TPD is unable to give direct information about the hydrogen activation/ dissociation ability of the catalysts, it can supply evidence on the recombination of atomic hydrogen, quantity and which kind of hydrogen desorption. Therefore, H2-TPD tests over the two cat-alysts were performed and the results are delivered in Fig. 5. Both catalysts show two main peaks (α, β) of desorbed H2, indicating that at least two types of active centers exist on the catalysts surface. The β-peak at higher temperature is associated with the desorption of hydrogen adsorbed strongly, while the α-peak at low temperature that is more weakly bounded to catalysts surface arises from the desorption of physically adsorbed hydrogen43-45. With regard to Pd/TiO2-101 catalyst, the intensity of β-peak remarkably increases comparing with Pd/TiO2-001 sample, suggesting that H2 dissociation/activation occurs more easily on the Pd/TiO2-101 catalyst.

 Fig. 5 H2-TPD profiles of the Pd/TiO2-001 and Pd/TiO2-101 catalysts

As reported in the literature46, the Ti3+ species in TiO2 materials are produced by trapping of electrons at defective sites of TiO2 and the quantity of gathered electrons might reflect the amount of defect sites. Nakaoka and Nosaka47 reported six signals of ESR technology occurring on the surface of TiO2: (ⅰ) Ti4+O-Ti4+OH-, (ⅱ) surface Ti3+, (ⅲ) adsorbed oxygen (O2-), (ⅳ) Ti4+O2-Ti4+O2-, (v) inner Ti3+, and (vi) adsorbed water. Therefore, ESR experiments were carried out over TiO2-001 and TiO2-101 catalysts for qualitatively studying the Ti3+ defects, as shown in Fig. 6. In our study, it is obviously seen that both TiO2-001 and TiO2-101 supports show only one strong signal at g values of 1.997 (less than 2), which can be ascribed to Ti3+ (3d1) on the surface48. Moreover, the relatively higher intensity of the Ti3+ signal over the TiO2-101 support than that of the TiO2-001 implied significant amount of surface Ti3 + defects on TiO2-101 support. In view of the catalytic results, the Pd NPs loaded on TiO2- 101 with more Ti3 + defective sites presented much higher activity than those on the TiO2-001 support which indicated that the presence of Ti3+ defects may contribute to the catalytic performance.

 Fig. 6 ESR spectra of TiO2-001 and TiO2-101 catalysts

H2-TPR experiments were performed to investigate the influence of the support materials on the reducibility of Pd species and the results are depicted in Fig. 7. A negative peak is observed at about 80 ℃ on both catalysts, which was attributed to the decomposition of the palladium hydride formed by exposure to hydrogen at ambient temperature49. Noticeably, on Pd/TiO2-101 catalyst, this negative peak with slightly lower peak area occurs at slightly lower temperature and a small H2 consumption peaks located at 105 ℃ was observed. This result suggests that there is higher dispersion of Pd nanoparticles on the TiO2-101 support and implied that the interaction between Pd nanoparticles and TiO2- 101 support might be stronger than that between Pd nanoparticles and TiO2-001 support50, 51. As well-known, palladium hydride is related to the particle size of Pd, and palladium hydride decreases with increasing in dispersion of Pd (decreasing Pd crystallite size)52-54. The previous study indicated that the improvement in dispersion of Pd may be correlated with the presence of abundant Ti3+ species on TiO2-101 support55. Considering the reduction peak of TiO2 at high temperature, the broad peak between 300 and 450 ℃ is because of the reduction of Ti4+ (nearby or interacting with the Pd nanoparticles) to Ti3+. As discussed in the references56, the dissociative hydrogen chemisorbed on palladium may transfer from Pd nanoparticle surface to TiO2 support and thus reduce Ti4+ to Ti3+.

 Fig. 7 H2-TPR profiles of catalysts (a) Pd/TiO2-001 and (b) Pd/TiO2-101

The surface atomic compositions of Pd/TiO2-001 and Pd/TiO2- 101 catalysts were analyzed by XPS measurements. The results of Pd/TiO2-001 and Pd/TiO2-101 samples were given in Table 1. It can be seen that Pd/TiO2-101 catalyst gives a lower atomic ratio of Pd/Ti (0.018) than that of Pd/TiO2-001 (0.025), which could be due to decoration of Pd0 metal surface by the mobile reducible TiO2. Reducible TiO2 can be reduced at high temperature and consequently migrate onto the Pd surface, which enhanced the Pd electron density and then weakened the ethylene adsorption, thus the selectivity of ethylene is improved19, 50, 57, 58.

The catalysts were characterized using CO pulse chemisorption to determine the palladium dispersion (Table 1). The dispersion of Pd/TiO2-101 is 15.95%, which is higher than that of Pd/TiO2- 001 (9.06%). This better dispersion of Pd/TiO2-101 catalyst implied a larger number of active sites than that on the Pd/TiO2-001 sample, which is considered as one of the key reasons for the catalytic performance enhancement.

3.4 Catalytic stability measurement

The catalyst stability was tested on both Pd/TiO2-001 and Pd/ TiO2-101 samples at 70 ℃. From the results in Fig. 8, the acetylene conversion decreased from 100% to 97.1% for Pd/TiO2-101 catalysts with reaction time on stream of 900 min. While that was from 95.5% to 92.4% for the Pd/TiO2-001 sample after reaction for 900 min.

 Fig. 8 Durability test for Pd/TiO2-001 and Pd/TiO2-101 catalysts at 70 ℃

The deactivation of Pd-based catalysts for selective hydrogenation of acetylene was mainly caused by accumulation of hydrocarbon species which hindered not only the pathway of H2 and C2H2 to the active sites but also the release of C2H4 59. Thus TGA was performed to study the carbon deposition on the spent catalysts after the stability experiment, as shown in Fig. 9. For the Pd/ TiO2-001 sample, the weight loss below 165 ℃ can be assigned to the loss of water, while the weight loss after 165 ℃ can be attributed to the oxidation or decomposition of carbonaceous deposits formed on the catalysts. The quantitative calculation is done based on the weight loss between 165 and 440 ° C. The amount of deposited carbonaceous species for the used Pd/TiO2- 001 was 13.6 g∙g-1 (gram carbonaceous per gram catalyst), and that was 11.55 g∙g-1 (gram carbonaceous per gram catalyst) for the Pd/TiO2-101 catalysts.

 Fig. 9 TGA/DTG results of (a) Pd/TiO2-001 and (b) Pd/TiO2-101 after the durability tests at 70 ℃ for 900 min

Moreover, the corresponding DTG results of Pd/TiO2-001 catalyst comprises two peaks at 299 and 355 ℃, which imply two types of carbonaceous species. However, the DTG profiles of the spent Pd/TiO2-101 catalyst is similar but less resolved after reaction of 900 min. According to the literature60-62, the peak at 281 ℃ is associated with the combustion of trapped hydrocarbons absorbed in catalyst pores or on the catalyst surface. The peak centered at 343 ℃ can be assigned to the combustion of amorphous carbon located on or in the vicinity of Pd NPs, a precursor of graphitic carbon that has a structure of oligomeric hydrocarbon, CxHy, which decreases the utilizability of H2 and/or C2H2. What′s more, the peak area of the Pd/TiO2-101 catalyst at high temperature is significantly smaller than that of Pd/TiO2-001, implying higher resistance against carbonaceous compound deposition and hence possessing better stability.

4 Conclusions

We have synthesized two types of Pd/TiO2 catalysts with different crystal-plane of TiO2 supports ({101} and {001} facets) and investigated the effects on catalytic properties in the selective hydrogenation of acetylene to ethylene reaction. The characterization results showed that a smaller particles and higher Pd NPs dispersion on Pd/TiO2-101 catalyst than that on the Pd/TiO2-001 catalyst. Pd/TiO2-101 catalyst present significantly higher catalytic activity than that of Pd/TiO2-001 catalyst. The catalytic behavior is dependent on the exposed facets of TiO2 supports. These results not only manifested that the structure and catalytic properties of Pd/TiO2 catalysts can be tuned by controlling the crystal-plane of the TiO2 support, but also greatly deepened the understanding of the selective hydrogenation of acetylene reaction by Pd/TiO2 catalysts.

Supporting Information: available free of charge via the internet at http://www.whxb.pku.edu.cn.

Acknowledgment: We would like to thank LIU Ming (Sichuan University), LIU Yue-Feng (Institute of Metal Research, CAS), and ZHENG Jian (Southwest University of Science and Technology) for their assistances on TEM and Raman measurement; meanwhile, we also thank CHEN Min (Sichuan University), LIAO Xue-Mei (Xihua University), DENG Jie (Chengdu University), and ZHENG Jian for useful discussion and helps. GAO Xiao-Ping thanks the China ChengDa Engineering Co., Ltd for scholarship.
Reference
 (1) Kuhn M.; Lucas M.; Claus P Ind. Eng. Chem. Res. 2015, 54, 6683. doi: 10.1021/acs.iecr.5b01682 (2) Studt F.; Abild-Pedersen F.; Bligaard T.; Sørensen R. Z.; Christensen C. H.; Nørskov J. K Science 2008, 320, 1320. doi: 10.1126/science.1156660 (3) Kim S. K.; Kim C.; Lee J. H.; Kim J.; Lee H.; Moon S. H J. Catal. 2013, 306, 146. doi: 10.1016/j.jcat.2013.06.018 (4) Crespo-Quesada M.; Yarulin A.; Jin M.; Xia Y.; Kiwi-Minsker L J. Am. Chem. Soc. 2011, 133, 12787. doi: 10.1021/ja204557m (5) Hong J.; Chu W.; Chen M.; Wang X.; Zhang T Catal.Commun. 2007, 8, 593. doi: 10.1016/j.catcom.2006.08.010 (6) Kim W. J.; Moon S. H Catal. Today 2012, 185, 2. doi: 10.1016/j.cattod.2011.09.037 (7) Pei G. X.; L iu; X. Y.; Wang A.; Lee A. F.; Isaacs M. A.; Li L.; Pan X.; Yang X.; Wang X.; Tai Z.; Wilson K.; Zhang T ACS Catal. 2015, 5, 3717. doi: 10.1021/acscatal.5b00700 (8) Lee J. H.; Kim S. K.; Ahn I. Y.; Kim W. J.; Moon S. H Catal. Commun. 2011, 12, 1251. doi: 10.1016/j.catcom.2011.04.015 (9) Wang Z. Q.; Zhou Z. M.; Zhang R.; Li L.; Cheng Z. M Acta Phys.-Chim. Sin. 2014, 30, 2315. [王沾祺, 周志明, 张锐, 李莉, 程振民. 物理化学学报, 2014, 30, 2315.] (10) Gu H.; Xu B. L.; Zhou J.; Li Y. Z.; Fan Y. N Acta Phys.-Chim. Sin. 2006, 22, 712. [顾虹, 许波连, 周静, 李远志, 范以宁. 物理化学学报, 2006, 22, 712.] (11) Guo Z. L.; Huang L. Q.; Chu W.; Luo S. Z Acta Phys.-Chim.Sin. 2014, 30, 723. [郭章龙, 黄丽琼, 储伟, 罗仕忠. 物理化学学报, 2014, 30, 723.] (12) Kontapakdee K.; Panpranot J.; Praserthdam P Catal.Commun. 2007, 8, 2166. doi: 10.1016/j.catcom.2007.03.003 (13) He Y.; Liang L.; Liu Y.; Feng J.; Ma C.; Li D J. Catal. 2014, 309, 166. doi: 10.1016/j.jcat.2013.09.017 (14) Osswald J.; Giedigkeit R.; Jentoft R.; Armbruster M.; Girgsdies F.; Kovnir K.; Ressler T.; Grin Y.; Schlogl R J. Catal. 2008, 258, 210. doi: 10.1016/j.jcat.2008.06.013 (15) Neumann M.; Teschner D.; Knop-Gericke A.; Reschetilowski W.; Armbrüster M J. Catal. 2016, 340, 49. doi: 10.1016/j.jcat.2016.05.006 (16) Gao Z.; Zhang Y.; Li D.; Werth C. J.; Zhang Y.; Zhou X J. Hazard. Mater. 2015, 286, 425. doi: 10.1016/j.jhazmat.2015.01.005 (17) Teschner D.; Borsodi J.; Wootsch A.; Révay Z.; Hävecker M.; Knop-Gericke A.; Jackson S. D.; Schlögl R Science 2008, 320, 86. doi: 10.1126/science.1155200 (18) Chen M. H.; Chu W.; Dai X. Y.; Zhang X W. Catal. Today 2004, 89, 201. doi: 10.1016/j.cattod.2003.11.027 (19) Li Y.; Jang B.W. L Appl. Catal. A 2011, 392, 173. doi: 10.1016/j.apcata.2010.11.008 (20) Chu W.; Xu J.; Hong J.; Lin T.; Khodakov A Catal. Today 2015, 256, 41. doi: 10.1016/j.cattod.2015.05.024 (21) Panpranot J.; Nakkararuang L.; Ngamsom B.; Praserthdam P Catal. Lett. 2005, 103, 53. doi: 10.1007/s10562-005-6502-x (22) Panpranot J.; Kontapakdee K.; Praserthdam P Appl. Catal. A 2006, 314, 128. doi: 10.1016/j.apcata.2006.08.024 (23) Wang N.; Qian W.; Chu W.; Wei F Catal. Sci. Technol. 2016, 6, 3594. doi: 10.1039/C5CY01790D (24) Si R.; Flytzani-Stephanopoulos M Angew. Chem. Int. Ed. 2008, 47, 2884. doi: 10.1002/(ISSN)1521-3773 (25) Liu L.; Yao Z.; Deng Y.; Gao F.; Liu B.; Dong L ChemCatChem 2011, 3, 978. doi: 10.1002/cctc.v3.6 (26) Wang F.; Zhang S.; Li C.; Liu J.; He S.; Zhao Y.; Yan H.; Wei M.; Evans D. G.; Duan X RSC Adv. 2014, 4, 10834. doi: 10.1039/c3ra47076h (27) Shi Q.; Li Y.; Zhou Y.; Miao S.; Ta N.; Zhan E.; Liu J.; Shen W J. Mater. Chem. A 2015, 3, 14409. doi: 10.1039/C5TA02897C (28) Liu L.; Gu X.; Cao Y.; Yao X.; Zhang L.; Tang C.; Gao F.; Dong L ACS Catal. 2013, 3, 2768. doi: 10.1021/cs400492w (29) Yang J.; Cao L. X.; Wang G. C J. Mol. Model. 2012, 18, 3329. doi: 10.1007/s00894-011-1337-4 (30) Han X.; Kuang Q.; Jin M.; Xie Z.; Zheng L J. Am. Chem.Soc. 2009, 131, 3152. doi: 10.1021/ja8092373 (31) He Y.; Fan J.; Feng J.; Luo C.; Yang P.; Li D J. Catal. 2015, 331, 118. doi: 10.1016/j.jcat.2015.08.012 (32) Tan Z.; Sato K.; Takami S.; Numako C.; Umetsu M.; Soga K.; Nakayama M.; Sasaki R.; Tanaka T.; Ogino C.; Kondo A.; Yamamoto K.; Hashishin T.; Ohara S RSC Adv. 2013, 3, 19268. doi: 10.1039/c3ra43383h (33) Zheng J.; Liu Z.; Liu X.; Yan X.; Li D.; Chu W J. Alloy.Compd. 2011, 509, 3771. doi: 10.1016/j.jallcom.2010.12.152 (34) Tian F.; Zhang Y.; Zhang J.; Pan C J. Phys. Chem. C 2012, 116, 7515. (35) Komhom S.; Mekasuwandumrong O.; Praserthdam P.; Panpranot J Catal. Commun. 2008, 10, 86. doi: 10.1016/j.catcom.2008.07.039 (36) Sárkány A.; Schay Z.; Frey K.; Széles É.; Sajó I Appl. Catal.A 2010, 380, 133. doi: 10.1016/j.apcata.2010.03.042 (37) Menezes W. G.; Altmann L.; Zielasek V.; Thiel K.; Bäumer M J. Catal. 2013, 300, 125. doi: 10.1016/j.jcat.2012.12.023 (38) Vincent M. J.; Gonzalez R. D Appl. Catal. A 2001, 217, 143. doi: 10.1016/S0926-860X(01)00586-5 (39) Wang N.; Xu Z.; Deng J.; Shen K.; Yu X.; Qian W.; Chu W.; Wei F ChemCatChem 2014, 6, 1470. (40) Douidah A.; Marécot P.; Szabo S.; Barbier J Appl. Catal. A 2002, 225, 21. doi: 10.1016/S0926-860X(01)00627-5 (41) Dole H. A. E.; Safady L. F.; Ntais S.; Couillard M.; Baranova E. A J. Catal. 2014, 318, 85. doi: 10.1016/j.jcat.2014.07.003 (42) Panagiotopoulou P.; Kondarides D. I J. Catal. 2009, 267, 57. doi: 10.1016/j.jcat.2009.07.014 (43) Yu W. Y.; Mullen G. M.; Mullins C. B J. Phys. Chem. C 2013, 117, 19535. doi: 10.1021/jp406736b (44) Huang L.; Chu W.; Zhang T.; Yin Y.; Tao X J. Nat. Gas Chem. 2009, 18, 35. doi: 10.1016/S1003-9953(08)60082-1 (45) Han X.; Chu W.; Ni P.; Luo S. Z.; Zhang T J. Fuel Chem.Technol. 2007, 35, 691. doi: 10.1016/S1872-5813(08)60004-3 (46) Ikeda S.; Sugiyama N.; Murakami S. Y.; Kominami H.; Kera Y.; Noguchi H.; Uosaki K.; Torimoto T.; Ohtani B Phys.Chem. Chem. Phys. 2003, 5, 778. doi: 10.1039/b206594k (47) Nakaoka Y.; Nosaka Y J. Photochem. Photobiol. A 1997, 110, 299. doi: 10.1016/S1010-6030(97)00208-6 (48) Salama T. M.; Hattori H.; Kita H.; Ebitani K.; Tanaka T J. Chem. Soc. Faraday Trans. 1993, 89, 2067. doi: 10.1039/FT9938902067 (49) McCue A. J.; McKenna F. M.; Anderson J. A Catal. Sci.Technol. 2015, 5, 2449. doi: 10.1039/C5CY00065C (50) Riyapan S.; Boonyongmaneerat Y.; Mekasuwandumrong O.; Praserthdam P.; Panpranot J Catal. Today 2015, 245, 134. doi: 10.1016/j.cattod.2014.07.017 (51) Neyertz C.; Volpe M Colloids Surf. A 1998, 136, 63. doi: 10.1016/S0927-7757(97)00249-5 (52) Ziemecki S. B.; Michel J. B.; Jones G. A Reac. Solids 1986, 2, 187. doi: 10.1016/0168-7336(86)80082-1 (53) Gómez-Quero S.; Cárdenas-Lizana F.; Keane M. A Ind. Eng.Chem. Res. 2008, 47, 6841. doi: 10.1021/ie0716565 (54) Aytam H. P.; Akula V.; Janmanchi K.; Kamaraju S. R. R.; Panja K. R.; Gurram K.; Niemantsverdriet J W. J. Phys.Chem. B 2002, 106, 1024. doi: 10.1021/jp012357a (55) Panpranot J.; Kontapakdee K.; Praserthdam P J. Phys. Chem.B 2006, 110, 8019. doi: 10.1021/jp057395z (56) Xu J.; Sun K.; Zhang L.; Ren Y.; Xu X Catal. Commun. 2005, 6, 462. doi: 10.1016/j.catcom.2005.04.006 (57) Liu Y. N.; Feng J. T.; He Y. F.; Sun J. H.; Li D. Q Catal. Sci.Technol. 2015, 5, 1231. doi: 10.1039/C4CY01160K (58) Kim E.; Shin E.W.; Bark C.W.; Chang I.; Yoon W. J.; Kim W. J Appl. Catal. A 2014, 471, 80. doi: 10.1016/j.apcata.2013.11.036 (59) Zhang S.; Chen C. Y.; Jang B.W. L.; Zhu A. M Catal. Today 2015, 256, 161. doi: 10.1016/j.cattod.2015.04.002 (60) Pachulski A.; Schödel R.; Claus P Appl. Catal. A 2011, 400, 14. doi: 10.1016/j.apcata.2011.03.019 (61) Lopez E.; Ordonez S.; Diez F. V Appl. Catal. B 2006, 62, 57. doi: 10.1016/j.apcatb.2005.06.014 (62) Azizi Y.; Petit C.; Pitchon V J. Catal. 2008, 256, 338. doi: 10.1016/j.jcat.2008.04.003