YAO Tai-zhen ，AN Yun-lei ，YU Hai-ling ，LIN Tie-jun ，YU Fei ，ZHONG Liang-shu,3,*

（1. CAS Key Laboratory of Low-Carbon Conversion Science and Engineering, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, China；2. University of Chinese Academy of Sciences, Beijing 100049, China；3. School of Physical Science and Technology, ShanghaiTech University, Shanghai 201210, China）

Abstract: The effects of supports (CeO2, ZrO2, MnO2, SiO2 and active carbon) on the structure and catalytic performance of Ru-based catalysts for Fischer-Tropsch synthesis to olefins (FTO) were investigated. It was found that the intrinsic characteristics of supports and the metal-support interaction (MSI) would greatly influence the catalytic performance. The catalytic activity followed the order: Ru/SiO2 > Ru/ZrO2 > Ru/MnO2 > Ru/AC > Ru/CeO2. As far as olefins selectivity was concerned, both Ru/SiO2 and Ru/MnO2 possessed high selectivity to olefins (>70%), while olefins selectivity for Ru/ZrO2 was the lowest (29.9%). Ru/SiO2 exhibited the appropriate Ru nanoparticles size ( ～ 5 nm) with highest activity due to the relatively low MSI between Ru and SiO2. Both Ru/AC and Ru/MnO2 presented low CO conversion with Ru nanoparticles size of 1-3 nm. Stronger olefins secondary hydrogenation capacity led to the significantly decreased olefins selectivity for Ru/AC and Ru/ZrO2. In addition, partial Ru species might be encapsulated by reducible CeO2 layer for Ru/CeO2 due to strong MSI effects, leading to the lowest activity.

Key words: Fisher-Tropsch synthesis；syngas；olefins；Ru nanoparticles；support effects

Olefins are value-added platform feedstock for preparation of rubber, polymer, pesticide, medicine,and so on[1-3]. Lower olefinsare always used as the main polymer monomers, while long-chain olefinsare important basic raw materials for the synthesis of high-grade lubricants and higher alcohols[1].Traditionally, olefins are generated from naphtha cracking or fluid catalytic cracking. With the rapid growth in petroleum consumption and the limited reserve of oil sources, alternative processes are developed for direct production of olefins from syngas[4]. Oxide-zeolite bifunctional catalysis and Fischer-Tropsch to olefins (FTO) process have been successfully developed for direct production of olefins from syngas[5]. For example, Jiao et al.[6]developed ZnCr composite oxides to activate CO molecule,combined with meso-SAPO-34 zeolite to offer the acid site for olefins formation. Selectivity to lower olefins for ZnCrOx/MSAPO catalyst could reach 80% with CO conversion of 17% under the condition of 400 °C,25 bar and a H2/CO ratio of 1.5. Cheng et al.[5]also reported a bifunctional catalyst that could realize direct conversion of syngas to lower olefins with an exceptionally high selectivity (74%) at CO conversion of 11%. It was suggested that the weak acid sites on the zeolites and suitable distance between oxides and the zeolites facilitated the production of lower olefins. FTO process is another promising route to produce olefins directly from syngas. FT catalysts could be modified to produce olefins with high selectivity by adding appropriate electronic and structural promoters[7,8].Traditionally, Fe-based catalysts are always studied for the FTO reaction at high-temperature (>300 °C)[2,9,10].Na and S promoted iron catalysts using α-alumina or carbon nanofiber were developed as weakly interacting supports for selective production of lower olefins(61%)[2]. Na promoter can benefit for suppressing methane selectivity and S promoter may selectively block the hydrogenation sites[2,10]. For Fe-based FTO catalysts, various mixed phases of iron carbide and iron oxide co-exist under reaction conditions[7,10], and large amount of CO2is produced during FTO reaction process due to the high activity for water-gas shift reaction (WGSR)[7]. Obviously, the high selectivity(>40%) to C1 products (CH4and CO2) greatly declines the carbon utilization efficiency and the technology competitiveness. Compared with Fe-based catalysts,Co-based catalysts always possess a strong hydrogenation ability with long-chain hydrocarbons as the dominant products and have rarely been reported for the FTO reaction[8]. However, our previous work discovered the Co2C nanoprisms with exposed (101)and (020) facets exhibited high selectivity to lower olefins ( ～ 60.8%) as well as low selectivity to methane( ～ 5.0%) at mild reaction conditions[11]. The products distribution deviated greatly from the classical ASF distribution with the ratio of olefin to paraffin (O/P) for the C2-4slate to be as high as 30. Similar to Fe-based catalyst, Co2C nanoprisms also suffer from high CO2selectivity (>40%) and low carbon efficiency. Very recently, some studies were focused on reducing the WGSR activity to decrease CO2selectivity during FTO process. Xu et al.[12]investigated hydrophobically modified Fe-FTO catalysts with 65% olefins selectivity at a CO conversion of 56.1%, and successfully reduced total selectivity of C1 products to 22.5%. It was proposed that the hydrophobic shell facilitated the desorption of water and reduced the side reaction related to water. Lin et al.[13]designed silica-coated CoMn-based catalyst for the FTO reaction, and CO2selectivity was largely suppressed from 44.7% to 15.1%, while enhancing olefins selectivity in total products from 39.7% to 58.8%. The coating of a hydrophilic silica layer over the Co2C active phase promoted the fast transfer of H2O away from internal Co2C to the outer layer SiO2shell, thus reducing the WGS reaction activity and improving the carbon efficiency of FTO process.

Compared with Fe- and Co-based catalysts, Rubased catalysts exhibit higher catalytic activity with lower CH4and CO2selectivity, which are usually applied to produce long-chain saturated alkanes in FTS due to the strong hydrogenation ability[14]. However,few studies are concerned on the production of olefins via Ru-based FT catalysts. Recently, it was reported that Na-Ru/SiO2catalyst exhibited ultrahigh carbon efficiency that realized 80.1% olefins selectivity at CO conversion of 45.8%[15]. The undesired C1 products were limited under 5%. The reaction pathway was tailored to promote the formation of olefins due to the change in the local electronic structure and the low reactivity of chemisorbed H species on Ru surfaces.Supported catalysts are usually used for Fischer-Tropsch reaction to enhance the mechanical stability,thermal conductivity and improve the dispersion of the active phase, and the nature of the support can significantly influence the catalytic performance.Although Na-promoted Ru-based catalysts exhibited high carbon efficiency during FTO reaction, the effects of support on the Ru-based FTO reaction were still unclear. Herein, MnO2, CeO2, ZrO2, SiO2and active carbon (AC) were selected as supports of Ru-based FTO catalysts and their effects on catalytic activity and product selectivity were investigated in detail.

1 Experimental

1.1 Material

Ruthenium nitrosyl nitrate (14.14%, aqueous solution) was purchased from Heraeus Precious Metal Technology Co., Ltd, AR. Sodium nitrate (NaNO3,AR), ammonium carbonate ((NH4)2CO3, AR),zirconium nitrate (Zr(NO3)4·5H2O; AR) and manganese nitrate (50% Mn(NO3)2, aqueous solution) were purchased from Sinopharm Chemical Reagent Co., Ltd.Cerium nitrate (Ce(NO3)2·6H2O, AR) was purchased from Shanghai Aladdin Biochemical Technology Co.,Ltd. (Shanghai, China). Aerosol silica (SiO2, AEROSIL 380) was purchased from Evonik Degussa China Co.,Ltd. Active carbon (AC) was purchased from Chengdu Institute of Organic Chemistry, Chinese Academy of Sciences. All materials were used as received without further purification.

1.2 Support preparation

All metal oxide supports were prepared by coprecipitation method and (NH4)2CO3was used as precipitant. Appropriate amount of Zr(NO3)4·5H2O,Ce(NO3)2·5H2O or Mn(NO3)2was dissolved in deionized water. The salt aqueous solution and(NH4)2CO3solution were simultaneously added into a beaker which contained 100 mL deionized water under continuous stirring. The pH value was fixed at about 8.After aging for 2 h at 70 °C, the obtained suspension was centrifuged and washed with deionized water until pH value of 7.0, then dried at 80 °C for 10 h. Then the samples were calcined in a muffle furnace at 400 °C for 4 h under static air. Finally, three metal oxide supports including ZrO2, MnO2and CeO2could be obtained. AC was previously treated with 50% nitric acid at 80 °C for 2 h and then washed with deionized water until the pH value reached about 7.0, followed vacuum drying at 120 °C for 12 h to serve as a support.

1.3 Catalyst preparation

The supported Ru catalysts with Ru loading amount of about 2.5% were prepared by impregnation method. The molar ratio of alkali metal Na to Ru was kept at 1∶2 to determine the addition content of NaNO3. The catalysts were stirred for 4 h at 30 °C.After being dried in an oven at 80 °C for 10 h, the catalyst precursors were calcined at 400 °C for 4 h at a heating rate of 2 °C/min. Then Ru/MnO2, Ru/ZrO2,Ru/CeO2, Ru/SiO2and Ru/AC catalysts could be obtained.

1.4 Catalyst characterization

Power X-ray diffraction (XRD) data were acquired using a Rigaku Ultima IV X-ray diffractometer (40 kV, 40 mA) equipped with CuKα radiation (λ= 1.54056 ?) with scanning angle from 5°to 90° at a scanning speed of 2(°)/min. Structure phases were identified by JCPDS standard card.

Transmission electron microscopy (TEM) and high-solution transmission electron microscopy(HRTEM) images were obtained on a FEI Tecnai G2 F20 S-TWIN equipment with 200 kV accelerating voltage. The samples for TEM were pretreated under a pure H2flow at 400 °C for 4 h and then dispersed into ethanol. After ultrasonication for 10 min, the suspension was deposited on the copper grids for the measurement. The average particle size was calculated upon more than 200 particles.

Nitrogen adsorption measurements were performed on a Micromeritics ASAP 2420 instrument.100 mg sample was degassed at 200 °C in vacuum for 10 h and then transferred to the analysis port for testing. The specific surface area was calculated by the Brunauer-Emmett-Teller (BET) equation and the pore size was determined by the Barrett-Joyner-Halenda(BJH) method.

Metal content in the reduced catalyst samples was determined using an Optima 8000 DV PerkinElmer Inductively Coupled Plasma-Optical Emission Spectrometer (ICP).

Hydrogen temperature-programmed reduction(H2-TPR) was tested on a Micromeritics ChemSorb2920 with TCD. The samples (50 mg) were loaded into a U-shaped quartz tube and then treated in He (30 mL/min) at 120 °C for 1 h. TCD started to record when the temperature rose from 50 to 800 °C in 5% H2/Ar (30 mL/min).

C3H6pulse transient hydrogenation experiments were tested on VDSorb-9Xi instruments with an MKS Cirrus 2 mass spectrometer. The reduced samples(50 mg) were loaded into a U-shaped quartz tube and then treated in H2(30 mL/min) at 400 °C for 1 h and followed syngas (H2/CO = 2, 30 mL/min) at 270 °C for 0.5 h. Subsequently, C3H6was pulsed into the system and the effluent was detected by mass spectrometer.The effluent for C3H6(m/z= 42) and C3H8(m/z= 44)was monitored using an MKS Cirrus 2 mass spectrometer. The integrated peak area ratio of C3H6/C3H8detected by mass spectrometer was calculated and denoted asR, which displays the hydrogenation capacity of different catalysts.

X-ray photoelectron spectroscopy (XPS) scans of the materials were obtained by a Thermo Fisher ScientificK-Alpha spectrometer. An achromatic AlKα source (1486.6 eV) was used and operated while maintaining a pressure of 1 × 10-6mbar in the chamber.The standard C 1speak as a reference was at a binding energy (BE) of 284.8 eV.

1.5 Catalytic evaluation

FTO catalytic performance was evaluated in a continuous flow fixed-bed reactor with 10 mm inner.1 g of catalyst sieved into 40-60 mesh was diluted with quartz sand (2 g, 40-60 mesh) and then loaded into the constant temperature zone of the reactor. Prior to the catalytic reaction, the catalysts were reduced with pure H2(100 mL/min) at 400 °C for 4 h. After the reactor was cooled down to 180 °C, syngas (H2∶CO∶N2=64.7∶32.3∶3) was fed in and the pressure was improved to 1 MPa, where N2was used as internal standard. The gaseous product flowed through a hot trap (120 °C) and a cold trap (1 °C), and then was analyzed online by Agilent chromatograph (GC) equipped with flame ionization detector (FID) and thermal conductivity detector (TCD). H2, N2, CO, CH4and CO2were detected by TCD, the hydrocarbons with carbon number in the range of 1-7 (C1-C7) were detected by FID. The other products were collected by the hot trap and cold trap which were analyzed by off-line Shimadzu GC. The carbon balance, mass balance and oxygen balance were calculated and maintained at(100 ± 5)%. All experiments were repeated for more than twice to keep the results convinced.

CO conversion (xCO) and product selectivity (si)are calculated on a carbon atom basis.

where COinletand COoutletdenotes moles of CO at the inlet and the outlet, respectively,sirepresents the selectivity of the producti,Niis the molar fraction of producti, andniis the carbon number of producti.

Chain growth probability (α) is calculated by Anderson-Schulz-Flory distribution:

wherewnandndenotes mass fraction of the hydrocarbon and the carbon number of hydrocarbon.(α) is chain growth probability which could be obtained by calculation the slope as mentioned before.

2 Results and discussion

2.1 Catalytic performance

The catalytic performance for different supported Ru-based catalysts were investigated as shown in Figure 1 and Table 1. For all supported catalysts, the catalytic activity was in the following order: Ru/SiO2>Ru/ZrO2> Ru/MnO2> Ru/AC > Ru/CeO2. Among all of the supported catalysts tested at 270 °C, SiO2supported catalyst showed the highest CO conversion of 33.6%. CO conversion over Ru/ZrO2, Ru/MnO2and Ru/AC catalyst decreased to 17.6%, 12.0% and 7.8%,respectively. Under the same reaction conditions, CeO2supported catalysts exhibited the lowest CO conversion of 2% and only increased to 8.3% even the reaction temperature was increased to 300 °C.

For olefins selectivity, it was found that olefins were generated easily for SiO2, MnO2and CeO2supported catalysts with olefins selectivity higher than 49%. For Ru/AC and Ru/ZrO2catalysts, only about 30% for olefins selectivity were obtained. In addition,the O/P ratio for SiO2, MnO2and CeO2supported catalysts was all higher than 5, while ZrO2and AC supported catalysts presented much lower O/P ratio(＜ 3), which further confirmed that olefins were preferentially formed over SiO2, MnO2and CeO2supported catalysts. Except for olefins, only a small amount of oxygenates (＜ 10%) were produced over various supported catalysts. In addition, CH4selectivity was rather low for Ru/SiO2(2.5%) and Ru/MnO2(4.4%) catalysts. However, methanation reaction was violent and nearly half of CO was converted to CH4on Ru/AC catalyst with CH4selectivity of 46.5%. For ZrO2and CeO2supported catalysts, CH4selectivity was 18.0% and 12.4%, respectively. The results suggested that SiO2and MnO2supported catalysts could suppress CH4formation effectively, while AC favored methanation reaction. Ru-based FTO process often possesses low activity toward WGSR with high carbon efficiency[14,15]. In our study, the WGSR activity seemed to be influenced by different supports. CeO2supported catalyst exhibited the highest WGS reaction activity with CO2selectivity of 21.5%, which greatly reduced the carbon efficiency. And CO2selectivity of Ru/MnO2,Ru/AC, Ru/SiO2and Ru/ZrO2catalysts were all lower than 10%, suggesting high carbon efficiency for olefins formation.

Figure 2 plotted the hydrocarbon distribution and chain growth probability for various supported Rubased catalysts. For all supported catalysts, ln(Wn/n)andnfollowed well with the classical ASF distribution.Carbon chain growth factor (α) of hydrocarbons could be derived from the linear part of the plot and obtained by fitting the catalytic results using the ASF model.The carbon chain growth probability was the weakest for the Ru/AC catalyst and the α value was only 0.52,which suggested that AC support hindered the formation of high carbon products. However, the α values for Ru/SiO2, Ru/MnO2, Ru/CeO2and Ru/ZrO2catalysts were all higher than 0.69.

Figure 2 Hydrocarbon distribution of various supported catalysts with the chain growth probability (α) obtained by the fitting the catalytic results using the ASF model

2.2 Characterization

BET surface area and pore size distribution of various supports and supported samples were measured by N2physisorption as shown in Table 2. The specific surface area for SiO2, MnO2, CeO2, AC and ZrO2was 279, 110, 88, 1438 and 83 m2/g, respectively. For pore size distribution, mesoporous structure was presented for various supports and supported catalysts. However,after Ru impregnation over various supports, both specific surface area and pore size decreased. In addition, according to ICP test, the amount of Ru loading content was 1.94%, 2.28%, 1.82%, 2.03% and 1.79% for SiO2, MnO2, CeO2, AC and ZrO2supported catalysts, respectively.

Table 2 Texture property and Ru loading content of various catalysts

Figure 3 showed the XRD patterns of various supports and calcined, reduced and spent catalysts.

Figure 3 Powder X-ray diffraction patterns of (a) supports, (b) calcined catalysts, (c) reduced catalysts and (d) spent catalysts

Figure 4 (HR)TEM images of the reduced catalysts: (a)Ru/SiO2, (b) Ru/MnO2, (c) Ru/CeO2, (d) Ru/AC, (e) Ru/ZrO2

There were no obvious characteristic peaks for the AC and aerogel SiO2, indicating that AC and SiO2supports were all in amorphous form. For CeO2, ZrO2and MnO2supports, clear diffraction peaks of CeO2, ZrO2and MnO2could be detected. After impregnation of Ru on various supports, diffraction peaks of RuO2were observed for Ru/SiO2and Ru/ZrO2catalysts. However,there were no clear peaks corresponded to Ru species for the AC, MnO2and CeO2supported catalysts,suggesting the high dispersion of Ru species[16]. For the reduced and spent samples, metallic Ru phase could be detected only for Ru/SiO2and Ru/ZrO2catalysts. In addition, the crystallite size of metallic Ru nanoparticles was calculated based on XRD characterization. The size metallic Ru nanoparticles for the reduced Ru/SiO2and Ru/ZrO2catalyst was 4.9 and 7.1 nm, respectively. In addition, there was no crystal structure change for the bulk phase of ZrO2and CeO2,while MnO2underwent phase transformation to MnO.

Figures 4 and 5 showed the (HR)TEM images and the corresponding distributions of the particle sizes of Ru species for the reduced and spent catalysts.

For all reduced samples, metallic Ru nanoparticles could be observed except Ru/CeO2. For ZrO2and SiO2supported catalysts, the average size of Ru0nanoparticles reached 7.5 and 4.5 nm, respectively.However, the size of Ru nanoparticles supported on AC and MnO2decreased to 1-3 nm. Similar (HR)TEM images were obtained for the spent catalysts and only metallic Ru nanoparticles with average size of 5.0, 1.8,2.2 and 7.6 nm were presented in the spent Ru/SiO2,Ru/MnO2, Ru/AC and Ru/ZrO2catalysts, respectively(Figure 5). Ru metallic phase was still hard to be observed in the spent Ru/CeO2catalyst due to the high dispersion of Ru.

Figure 5 (HR)TEM images of the spent catalysts: (a) Ru/SiO2,(b) Ru/MnO2, (c) Ru/CeO2, (d) Ru/AC, (e) Ru/ZrO2

H2-TPR profiles (Figure 6(a)) reflected the reduction behavior of various supported Ru catalysts.Only one peak centered at 90-170 °C appeared over Ru/SiO2and Ru/ZrO2catalysts, which could be ascribed to the reduction of bulk RuO2to metallic Ru.Two obvious peaks occurred in Ru/CeO2catalyst, and the first peak at 95 °C could be assigned to the reduction of well-dispersed Ru species on the surface,while the other peak located at 127 °C was corresponded to the reduction of the Ru species incorporated into the surface of ceria and the surface oxygen of CeO2around Ru species[16-18]. For MnO2supported Ru based catalysts, the first peak at 145 °C was corresponded to the reduction of Ru4+to Ru0, and the second peak at 240 °C could be assigned to the reduction of MnO2to MnO[19]. For Ru/AC catalysts,three obvious wide peaks at 94, 300 and 431 °C were accompanied by the reduction of RuO2, the formation of CH4and the partial gasification of the AC support around Ru nanoparticles, respectively[20,21].

Figure 6 (a) H2-TPR profiles of Ru/CeO2, Ru/MnO2, Ru/ZrO2, Ru/AC, Ru/SiO2; (b) Ru 3p regions of Ru/CeO2,Ru/MnO2, Ru/ZrO2, Ru/AC and Ru/SiO2

XPS characterization is widely used to study the chemical valence and elemental composition of catalyst surfaces. The Ru 3pregion was chosen to investigate the valence state of Ru (Figure 6(b)). The peaks with binding energy (BE) of about 462.4 and 484.6 eV were attributed to metallic Ru0[22]. The binding energy of Ru 3pof various supported catalysts were shown in Table 2. For Ru/ZrO2, Ru/AC and Ru/SiO2catalysts,the binding energy for metallic Ru was 461.92, 461.78 and 461.71 eV, respectively. The binding energy for metallic Ru increased to 462.28 and 462.48 eV for Ru/MnO2and Ru/CeO2, respectively. It seemed that the binding energy for Ru/MnO2and Ru/CeO2catalysts were slightly higher than that for other catalysts, which could be attributed to the strong MSI effects on the support of MnO2and CeO2.

2.3 Effects of the catalyst support

The hydrogenation capacity is one of the key factors affecting the olefins proportion in the hydrocarbon products. In order to further investigating the support influence on the hydrogenation capacity,propylene hydrogenation reaction was carried out in pulse experiments[23]. The peak area ratio (R) of propylene and propane in the tailor gas was calculated by mass spectrometry (Figure 7). For AC and ZrO2supported catalysts, a considerable amount of propylene was hydrogenated to propane, and itsRvalues were only 1.2 and 0.5, respectively. For MnO2,CeO2and SiO2supported catalysts, the propane production was significantly interrupted with highRvalues of 43.4, 37.8 and 26.5, respectively. It seemed that SiO2, MnO2and CeO2supported catalysts could significantly inhibit the secondary hydrogenation of olefin, which was also consistent with the high olefins selectivity for syngas conversion over SiO2, MnO2and CeO2supported catalysts.

Figure 7 Transient response curves obtained for propylene-pulse experiment: (a) Ru/SiO2, (b) Ru/MnO2, (c) Ru/CeO2, (d) Ru/AC,(e) Ru/ZrO2, (f) the relationship between R value and olefins selectivity over various catalysts

Obviously, support had great effects on the Rubased catalyst structure and catalytic performance. It is well known that the particle size of active phase is a major factor which greatly influences the final catalytic performance for FT reaction. The CO dissociation would be hampered on the active phase with very small particle size[24,25], leading to a relatively low catalytic activity. Due to the enhanced Ru dispersion by AC and MnO2supports, Ru nanoparticles with very small size( ～ 2 nm) were generated, which exhibited low activity.The low intrinsic activity for small Ru nanoparticles may be related to the stronger CO adsorption and concomitant partial blocking of active sites[24,25]. For ZrO2and SiO2supported catalysts, Ru nanoparticles with moderate size (4-7 nm) were obtained, which benefited for CO chemisorption, leading to a high CO conversion. MSI plays an important role in the design of highly efficient FTS catalysts. It is suggested that MSI would influence the dispersion, reduction, and activation behavior of the active metal phase. In this work, H2-TPR, XRD and TEM results showed that Ru species were highly dispersed on CeO2and it might be encapsulated by CeO2supports due to the strong MSI.FTO process is recognized as a surface-catalyzed structure-sensitive reaction, the exposed active metals can directly contact syngas and catalyze CO/H2to produce olefins. However, as the active phase was encapsulated by supports, the contact frequency between Ru and syngas would be decreased and the CO adsorption and dissociation ability was also substantial hindered. In previous work, Zhang et al.[14]proved active phase was over-encapsulated by supporters, which caused a severe hindrance in the adsorption of CO and resulted in a lower conversion of CO. Friedrich et al. reported that Ru nanoparticles were covered by partially reduced supports, which also led to a decrease in CO chemisorption[26].

In addition, olefins were easily generated with high selectivity on SiO2, MnO2and CeO2supported catalysts than Ru/AC and Ru/ZrO2catalysts.Propylene-pulse experiments verified that SiO2, MnO2and CeO2supported catalysts would depress olefin hydrogenation ability effectively. Previous study confirmed that the olefins hydrogenation reaction would be encouraged by high ambient hydrogen concentrations on the catalyst surface, leading to a low olefin selectivity[23]. In syngas conversion, Na promoter was homogenously distributed on both support and Ru nanoparticles, and it might benefit the strong electronic interaction between Ru nanoparticles and Na promoter,which suppressed the olefins hydrogenation process[15].In this study, SiO2, MnO2and CeO2supported catalysts might enhance CO adsorption but weaken H2adsorption effectively. As a result, H/C ratio on the surface of catalysts was decreased and the olefins formation was promoted, resulting in a high selectivity towards to olefins[27,28]. Besides, It was noteworthy that WGSR activity over CeO2supported catalyst was higher than the other supported catalysts. In previous study, reducible oxide supports were reported to provide oxygen vacancies, and the oxygen vacancies may facilitate the dissociation of H2O and thus accelerate the overall reaction rate for WGSR[29-31]. As CeO2contained large amount of oxygen vacancies, CO2might be formed easily on Ru/CeO2catalyst. However,CO2selectivity on Ru/ZrO2catalyst was bellow than 5%. Jackson et al.[32]found the ZrO2surface could not be reduced to generate enough oxygen vacancies under 270 °C. Therefore, a low CO2selectivity was observed for Ru/ZrO2catalyst.

3 Conclusions

The effects of various supports (CeO2, ZrO2,MnO2, SiO2and AC) on Ru-based FTO catalytic performance were investigated. Ru/SiO2showed the highest activity and olefins selectivity for FTO reaction. MnO2, AC and CeO2supported catalysts showed much lower activity. Strong olefins second hydrogenation was found over ZrO2and AC supported catalysts, which led to low olefins selectivity. It was suggested that the intrinsic characteristics of supports and the metal-support interaction (MSI) would greatly influence the catalytic performance. Support may influence the size of Ru nanoparticles. Much smaller size (1-3 nm) for Ru nanoparticles was found for the catalysts supported on AC and MnO2, leading to a relatively low CO conversion. In addition, Ru species might be encapsulated by CeO2layer for Ru/CeO2due to strong MSI effects, leading to the lowest activity compared with other supported catalysts.