Chemical Engineering Science - University of Tennessee

Page 1

Chemical Engineering Science 153 (2016) 10–20

Contents lists available at ScienceDirect

Chemical Engineering Science

http://d0009-25

n CorrGuangxtensifica

nn CorE-m

journal homepage: www.elsevier.com/locate/ces

CuO-Fe2O3-CeO2/HZSM-5 bifunctional catalyst hydrogenated CO2 forenhanced dimethyl ether synthesis

Xinhui Zhou a, Tongming Su a, Yuexiu Jiang a, Zuzeng Qin a,b,n, Hongbing Ji a,c,Zhanhu Guo b,nn

a School of Chemistry and Chemical Engineering, Guangxi Key Laboratory of Petrochemical Resource Processing and Process Intensification Technology,Guangxi University, Nanning 530004, Chinab Integrated Composites Laboratory (ICL), Department of Chemical and Biomolecular Engineering, University of Tennessee, Knoxville, TN 37966, USAc Department of Chemical Engineering, School of Chemistry & Chemical Engineering, Sun Yat-sen University, Guangzhou 510275, China

H I G H L I G H T S

G R A P H I C A L A

� CuO-Fe2O3 catalyst modified withdifferent amounts of CeO2 wereprepared.

� Addition of CeO2 led to the forma-tion of a stable Cu-O-Ce solid solu-tion.

� CeO2 can modify the amount of acidsites and acid type of CuO-Fe2O3

catalyst.� Cu-Fe-Ce/HZSM-5 catalyst with3.0 wt% CeO2 showed the optimalcatalytic activity.

x.doi.org/10.1016/j.ces.2016.07.00709/& 2016 Elsevier Ltd. All rights reserved.

esponding author at: School of Chemistryi Key Laboratory of Petrochemical Resourcetion Technology, Guangxi University, Nanningresponding author.ail addresses: [email protected] (Z. Qin),

B S T R A C T

a r t i c l e i n f o

Article history:Received 12 March 2016Received in revised form30 June 2016Accepted 4 July 2016Available online 5 July 2016

Keywords:Carbon dioxide hydrogenationDimethyl ether synthesisMixed oxideCu-Fe-Ce catalystsCerium oxide

a b s t r a c t

A series of CuO-Fe2O3-CeO2 catalysts with various CeO2 doping were prepared via the homogeneousprecipitation method, characterized and mechanically mixed with HZSM-5. Their feasibility and per-formance for the synthesis of dimethyl ether (DME) via CO2 hydrogenation in a one-step process wereevaluated. The formed stable solid solution after the CuO-Fe2O3 catalyst modified with CeO2 promotedthe CuO dispersion, reduced the CuO crystallite size, decreased the reduction temperature of highlydispersed CuO, modified the specific surface area of the CuO-Fe2O3-CeO2 catalyst, and improved thecatalytic activity of the CuO-Fe2O3-CeO2 catalyst. The addition of CeO2 to CuO-Fe2O3 catalyst increasedthe amount of Lewis acid sites and Brønsted acid sites, and enhanced the acid intensity of the weak acidsites, which in turn promoted the catalytic performance of CO2 hydrogenation to DME. The optimalintroduced amount of Ce in the catalyst was determined to be 3.0 wt%. The CO2 conversion and DMEselectivity were 20.9%, and 63.1%, respectively, when the CO2 hydrogenation to DME was carried out at260 °C, and 3.0 MPa with a gaseous hourly space velocity of 1500 mL gcat

�1 h�1.& 2016 Elsevier Ltd. All rights reserved.

and Chemical Engineering,Processing and Process In-530004, China.

[email protected] (Z. Guo).

1. Introduction

Dimethyl ether (DME) possesses a high cetane number andproduces less NOx and SOx than do fossil fuels when combusted,representing an environmental friendly renewable fuel (Frusteriet al., 2015; García-Trenco and Martínez, 2012; Rutkowska et al.,2015). DME can also be used as a chemical intermediate for

Page 2

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–20 11

dimethyl sulphate, methyl acetate and low carbon olefin synthesis(Ge et al., 1998). The fossil fuel usage has increased the amount ofCO2 in the atmosphere and the subsequent greenhouse effectposes a threat to the environment. In addition, the energy crisisfacing the whole world is becoming increasingly urgent. Therefore,converting CO2 into useful chemicals, such as DME, is an attractivemethod to reduce greenhouse-gas emissions and to recyclecarbon.

The hydrogenation of CO2 to form DME mainly includes threereactions, i.e., methanol synthesis reaction (Eq. (1)), methanoldehydration reaction to form DME (Eq. (2)) and a reversed water-gas shift reaction (Eq. (3)) (Bonura et al., 2014b; Yang Lim et al.,2016). DME can be produced from CO2 via two routes. The firstroute is a two-step process through two reactors (including me-thanol synthesis on a metallic catalyst and subsequent dehydra-tion of methanol on an acid catalyst). The second route is a one-step process using a bifunctional catalyst in the same reactor tosimultaneously perform the two steps. However, methanol dehy-dration in the same reactor disturbs the equilibrium of synthesisreaction. Therefore, one-step process is economically and ther-modynamically favored (Jia et al., 2006; Vakili et al., 2011).

CO2þ3H2⇋CH3OHþH2O; ΔH298¼�49.46 kJ mol�1 (1)

2CH3OH⇋CH3OCH3þH2O; ΔH298¼�23.5 kJ mol�1 (2)

CO2þH2⇋COþH2O; ΔH298¼41.17 kJ mol�1 (3)

The bifunctional catalysts for the one-step synthesis of DME viaCO2 hydrogenation include a Cu-base catalyst for the hydrogena-tion component and a solid acid, such as HZSM-5, γ-Al2O3 or HYzeolite for the dehydration component, resulting inCuO-ZnO-Al2O3/HZSM-5 (Liu et al., 2015), CuO-ZnO-Al2O3-ZrO2

/HZSM-5 (An et al., 2008), Cu-ZnO-ZrO2/HZSM-5 (Frusteri et al.,2015), CuO-TiO2-ZrO2/HZSM-5 (Wang et al., 2009) orCuO-Fe2O3-ZrO2/HZSM-5 (Liu et al., 2013; Qin et al., 2015). How-ever, the conversion of stable CO2 molecule requires high tem-perature and pressure and only 10–30% was reported with theDME selectivity of only 30–50% when the reaction was carried outat 4–10 MPa and 200–300 °C. Due to the conversion of CO2 tomethanol was limited by the thermodynamic equilibrium, it wasmore difficult to attain reasonable per-pass conversions for largeindustrial scale CO2 treatment.

CeO2 is an effective co-catalyst with a hollow structure, and canstabilize and disperse Cu and other precious metals (Bera et al.,2003). CeO2 and CuO can form a stable solid solution (Gamarraet al., 2007; Marbán and Fuertes, 2005) to improve both catalyticactivity and thermal stability of CuO-based catalysts. CeO2 hasbeen frequently selected as a component for use in complex oxidesor as a dopant to improve the catalyst performance. CeO2 has beenapplied in the field of preferential CO oxidation (Hornés et al.,2010; Jia et al., 2010) and methanol steam reforming (Tonelli et al.,2011). For the dimethyl ether synthesis from CO2, the acid prop-erties of the catalysts influenced the DME selectivity (García-Trenco and Martínez, 2012; Yoo et al., 2007). For example, theacidity of zeolite HY was enhanced by the addition of Ce, and itpossessed a high proportion of moderate strength acid sites andwere more stable for methanol dehydration to DME (Jin et al.,2007); the presence of Ce on the HZSM-5 could modify the acidicproperties of HZSM-5 including the amount of acid sites and acidtype (Wang et al., 2007); and the addition of Ce could enhance theacidity of MnOx/TiO2 (Wu et al., 2008). These indicate that Ceplayed an important role on the catalysts acid properties.

In this work, a series of CuO-Fe2O3 catalysts modified usingdifferent amounts of CeO2 were prepared via a homogeneousprecipitation method; characterized using X-ray diffractometer

(XRD) to determine the crystal structure of catalysts, Ramanspectroscopy for obtaining the vibrational and rotation mode ofthe lattice and the molecules of catalysts, high-resolution trans-mission electron microscopy (HRTEM) to obtain the informationon the structure and composition of the particles, N2 adsorption-desorption to determine the pore structures and the textureproperties of the catalysts, H2-temperature programmed reduction(H2-TPR) to obtain the quantitative reduction information, thermalanalysis (TG-DTA) for determining the weight changes in the cal-cination process of the catalysts, Fourier transform infrared spec-troscopy (FTIR) of adsorbed pyridine and temperature-pro-grammed desorption of ammonia (NH3-TPD) for obtaining thesurface acidities of the catalysts, and temperature programmedoxidation (TPO) to detect the coke formed during the reaction. Thecatalysts were mixed with HZSM-5 to synthesize dimethyl ethervia CO2 hydrogenation in a one-step process.

2. Experimental

2.1. Catalyst preparation

The precursor for producing the CuO-Fe2O3-CeO2 catalyst wasprepared via a homogeneous precipitation method.Cu(NO3)2 �3H2O and Fe(NO3)3 �9H2O (Sinopharm Chemical Re-agent Co., Ltd.) were weighed at Cu/Fe mole ratio of 3:2 (Liu et al.,2013; Qin et al., 2015; Qin et al., 2016). Then, the amount of theCe(NO3)3 �6H2O (Sinopharm Chemical Reagent Co., Ltd) was de-termined by the desired CeO2 content of 1.0, 2.0, 3.0 and 4.0 wt% inCuO-Fe2O3-CeO2 in the final catalysts and was added to obtain ametal nitrate solution (0.5 mol L�1). A urea (China GuangdongGuanghua Sci-Tech Co., Ltd.) solution (5 mol L�1) was then addedto the metal nitrate solution, in which the urea to nitrates molarratio was 20. The mixture was stirred and reacted for 24 h at 90 °C.The obtained precipitates were filtered and dried at 110 °C for 12 h,ground through a 20–40 mesh, calcined at 400 °C for 4 h, and thusthe catalyst CuO-Fe2O3-CeO2 was obtained for methanol synthesis.HZSM-5, using as the methanol dehydration component, with asilica-alumina ratio of 300:1 (China Shanghai Novel ChemicalTechnology Co., Ltd.) was mechanically mixed with theCuO-Fe2O3-CeO2 oxide composites at a mass ratio of 1:1. Forcomparison, a CuO-Fe2O3/HZSM-5 catalyst without CeO2 wasprepared via the same method.

2.2. Catalyst characterization

The XRD, N2 adsorption-desorption and H2-TPR catalyst char-acterizations can be referenced in previous studies (Liu et al.,2013). Briefly, the X-ray diffraction (XRD) analysis was performedusing a Bruker D8 Advance X-ray diffractometer. The isotherm ofnitrogen adsorption and desorption was measured by an ASAP2000 physical adsorption instrument (Micromeritics InstrumentCorporation). The catalyst surface area was calculated via Bru-nauer-Emmett-Teller (BET) method, and the pore size distributioncurve was determined using the Barrett-Joyner-Halenda (BJH)model, which was based on the isotherm of desorption side. TheH2-TPR, NH3-TPD, and TPO experiments of catalyst samples weretaken in a PX200 multifunction catalyst analysis system (ChinaTianjin Golden Eagle Technology Co., Ltd.).

The Raman spectra were obtained using Lab RAMHR800 LaserConfocal Micro-Raman Spectroscopy (Horiba, Ltd.), including aYAG laser with a 532 nm excitation wavelength. High resolutiontransmission electron microscope (HRTEM) images of the catalystswere obtained using a FEI tecnai G20 instrument with an accel-eration voltage of 200 kV. A thermal analysis was performed usinga ZRY-1P thermal analyser (Techcomp Jingke Scientific

Page 3

Fig. 2. Raman spectra of the CuO-Fe2O3-CeO2 catalysts modified by CeO2 content of(a) 0, (b) 1.0, (c) 2.0, (d) 3.0, and (e) 4.0 wt% CeO2 calcined at 400 °C without H2

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–2012

Instruments (Shanghai) Co., Ltd.). About 8 mg catalyst was used forthe analysis, at a heating rate of 10 °C min�1 from 50 to 500 °C.The FTIR of adsorbed pyridine was conducted using a Magna-IR750 FTIR spectrometer (Thermo Nicolet Corporation). Self-sup-porting catalysts wafers (ca. 10 mg cm�2) were produced andloaded into an IR cell. The wafers were evacuated at 150 °C for 2 hto record the background spectrum and subsequently saturatedwith pyridine and evacuated at 150 °C for 2 h. The Py-IR spectrawere recorded at a spectra resolution of 4 cm�1 after subtractingthe sample background.

2.3. Catalytic hydrogenation of CO2 to DME

The catalytic activity was investigated using a continuous fixed-bed stainless steel reactor, which is detailed in the literature (Liuet al., 2013). Briefly, 1.0 g hybrid catalyst was added to the fixedbed reactor and reduced at 300 °C for 4 h with 30 mL min�1 of H2

(99.999%) at atmospheric pressure. A gas mixture composed of20% CO2 and 80% H2 was then introduced and reacted at 240–280 °C, 2.0–4.0 MPa with a gaseous hourly space velocity (GHSV)of 1500–3500 mL gcat�1 h�1.

reduction.

3. Results and discussion

3.1. XRD analysis

Fig. 1 illustrates the XRD patterns of the CuO-Fe2O3 catalystsmodified by different CeO2 amounts before H2 reduction. Thediffraction peaks at 2θ¼35.49°, 38.69°, and 58.26° are associatedwith monoclinic CuO (JCPDS No. 48-1548). The peaks at2θ¼30.24°, 35.63°, and 62.93° can be attributed to cubic crystalFe2O3 (JCPDS No. 39-1346). The CuO-Fe2O3 peaks are sharper thanthose of CuO-Fe2O3 catalysts modified by CeO2, suggesting thatCuO-Fe2O3 possesses an enhanced crystallinity (Wang et al., 2016).The CuO diffraction peaks become weakened and broadened asthe amount of CeO2 was increased, indicating that CuO dispersionwas improved by the addition of CeO2. The crystallite sizes of CuO(111) for CuO-Fe2O3-CeO2 catalysts modified with 1.0, 2.0, 3.0 and4.0 wt% CeO2 calculated by the Scherrer equation are 17.6, 14.7, 14.3and 15.9 nm, respectively. However, the crystallite size of CuO(111) for CuO-Fe2O3 catalyst is 22.5 nm, indicating that the mod-ification with CeO2 can decrease the crystallite size of CuO andpromote the dispersion of catalysts. The XRD patterns suggest that

Fig. 1. The XRD patterns of the CuO-Fe2O3-CeO2 catalysts modified by (a) 0, (b) 1.0,(c) 2.0, (d) 3.0, and (e) 4.0 wt% CeO2 calcined at 400 °C without H2 reduction.

CeO2 can significantly improve the CuO dispersion based on a3.0 wt% CeO2. Further increasing the CeO2 amount did not improvethe CuO dispersion. The results reveal that appropriate CeO2 ad-dition can inhibit the CuO crystals growth, resulting in small CuOcrystallites and increasing the catalysts dispersion. In addition, theCeO2�x peaks at 2θ of 26.3°, 43.9° and 55.7° (JCPDS No. 49-1415)were found in the CuO-Fe2O3-CeO2 catalysts, including that aportion of smaller Cu2þ ions (0.69 Å) and Fe3þ ions (0.64 Å)(Wang et al., 2014) entered to the CeO2 lattice (the radius of Ce4þ

ionic is 0.97 Å) to form a solid solution (Li et al., 2011, 2014a;Marbán and Fuertes, 2005).

3.2. Raman spectra

Fig. 2 shows the Raman spectra of the CuO-Fe2O3-CeO2 cata-lysts. The peaks of the CuO-Fe2O3 catalyst observed at 210, 276,373, and 574 cm�1 are related to typical A1g and 2Eg Raman he-matite modes, whereas the peak at 1288 cm�1 is associated withhematite two-magnon scattering (Li et al., 2014b; Wang et al.,2014). The peak at 276 cm�1 arises from CuO (Tsoncheva et al.,2013). The peaks of the modified CuO-Fe2O3-CeO2 catalyst at468 cm�1 can be attributed to the strong F2g mode of CeO2 (Qiet al., 2012). The broad band between 550 and 700 cm�1 is relatedto the oxygen vacancies after the introduction of cerium, sug-gesting the formation of a Cu-O-Ce or Fe-O-Ce solid solution (Chenet al., 2010; Li et al., 2011; Wang et al., 2014). The substitution ofCu2þ , Fe3þ for Ce4þ can generate oxygen vacancies aroundCu2þ-O-Ce4þ , Fe3þ-O-Ce4þ to maintain charge neutrality (Beraet al., 2002), respectively, which is consistent with the XRD result.

3.3. Transmission electron microscopy

The atomic level morphology and lattice structure of theCuO-Fe2O3 catalyst and CeO2 catalyst were investigated via HRTEMmeasurements. The low-magnification TEM images in Figs. 3(a)and 4(a) illustrate that the catalyst particles form disk-like shapes,and the addition of Ce decreases the particle size. A smaller particlesize generally resulted in the catalysts with more edges, corners,defects and oxygen vacancies, all of these could benefit the reactionperformance during the structure-sensitive reactions (Liu et al.,2005). The HRTEM CuO-Fe2O3 images in Fig. 3(b,c) clearly illustratethe presence of 0.232 and 0.252 nm interplanar spacing, corre-sponding to the characteristic (111) lattice plane of monoclinic CuO

Page 4

Fig. 3. Typical TEM images (a), HRTEM images (b and c) and EDX (d) of the CuO-Fe2O3 catalyst calcined at 400 °C without H2 reduction, while the inset (b) represents a SAEDpattern.

Fig. 4. Typical TEM images (a), HRTEM images (b and c) and EDX (d) of the CuO-Fe2O3-CeO2 catalyst with 3.0 wt% CeO2 calcined at 400 °C without H2 reduction, while theinset (b) represents a SAED pattern.

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–20 13

and characteristic (311) lattice plane of cubic crystal Fe2O3, respec-tively. The interplanar spacing of CuO (11�1) (0.2524 nm) andFe2O3 (311) (0.2518 nm) are nearly equal. Therefore, the 0.252 nminterplanar spacing may also correspond to CuO (11�1). TheCuO-Fe2O3-CeO2 catalyst with 3.0 wt% CeO2 HRTEM image (Fig. 4

(b)) displays 0.252 and 0.295 nm interplanar spacing, correspondingto the (311) and (220) planes of cubic crystal Fe2O3, respectively. Themagnified catalyst image in Fig. 4(c) is based on the square frame inFig. 4(b). This magnified image displays 0.187 and 0.271 nm inter-planar spacing, corresponding to the (20�2) plane of monoclinic

Page 5

Fig. 5. Nitrogen adsorption/desorption isotherms (A) and pore size distribution profiles (B) of CuO-Fe2O3-CeO2 with CeO2 content of (a) 0, (b) 1.0, (c) 2.0, (d) 3.0, and(e) 4.0 wt%.

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–2014

CuO and the (200) plane of CeO2, respectively. However, the (111)plane of monoclinic CuO was not observed in any of theCuO-Fe2O3-CeO2 (CeO2¼3.0 wt%) catalysts, suggesting that CuOwas relatively well-dispersed in the catalyst (Gamarra et al., 2007).The selected area electron diffraction (SAED) patterns are given inthe inset of Figs. 3(b) and 4(b). The SAED patterns include diffractionrings, indicating that the two catalysts are polycrystalline. In addi-tion, the energy-dispersive X-ray spectrum (EDX) displays the Fe, Ceand O signals, further proving that the mixed metal-oxide catalystshave been fabricated.

3.4. Nitrogen adsorption/desorption of catalysts

The specific surface areas and pore size distribution curves ofthe CuO-Fe2O3-CeO2 catalysts with different CeO2 amounts wereinvestigated using nitrogen adsorption-desorption isotherms,Fig. 5. The isotherms of the catalysts are similar, Fig. 5A, exhibitingIV type isotherms according to the IUPAC classification, suggestingthat all the catalysts possess the mesoporous features (Huanget al., 2016; Zhang et al., 2014). Furthermore, the pore size dis-tribution profiles (Fig. 5B) of the catalysts exhibited a narrow poresize distribution, with a maximum of �3.0 nm. The BET methodwas used to calculate the N2 adsorption-desorption isotherms andobtain the specific surface areas of the catalysts, which are listed inTable 1.

The specific surface areas of the CuO-Fe2O3-CeO2 catalysts were50.32, 56.41, 62.26, 67.34 and 56.60 m2 g�1 for the CeO2 contentsof 0, 1.0, 2.0, 3.0 and 4.0 wt%, respectively, suggesting that theaddition of proper amount of CeO2 can increase the specific sur-face areas of the CuO-Fe2O3 catalysts. In addition, CeO2 can pro-duce larger BET surfaces area than CuO (Tsoncheva et al., 2013; Xuet al., 2016). Furthermore, the CeO2 particle size of theCuO-Fe2O3-CeO2 catalyst will be reduced due to the 4f orbit and

Table 1Textural properties of the CuO-Fe2O3-CeO2 catalysts with different CeO2 contents.

CeO2 content (wt%) BET surface area (m2 g�1) Average pore diameter (nm)

0 50.32 25.541.0 56.41 20.622.0 62.26 19.443.0 67.34 18.964.0 56.60 20.86

the Ce structural relaxation, which consequently increases thesurface area and the concentration of defects, such as oxygen va-cancies (Zhan et al., 2012). The largest surface area (67.34 m2 g�1)was observed for the catalysts with 3.0 wt% CeO2. The largerspecific surface area can provide more absorption and/or reactionsites for CO2 and H2, resulting in a higher catalytic activity (Huanget al., 2016). However, the specific surface area was decreasedwhen the CeO2 loading was increased from 3.0 to 4.0 wt%, whichmay be associated with the crystallite growth as the amount ofCeO2 increases above 3.0 wt%.

3.5. TG-DTA analysis of catalysts

Fig. 6 shows the precursor TG-DTA curves for different CeO2

amounts. Fig. 6A shows the TGA curves of the precursor. Theweight loss is observed to occur in three steps. The weight loss inthe range of 50–200 °C can be attributed to the desorption ofphysically adsorbed surface water and the bound water fromprecursors (Cao et al., 2008; Liu et al., 2015). The precursor weightlosses were 10.1%, 6.2%, 7.1%, 8.0% and 8.4% for the CeO2 amountsof 0, 1.0, 2.0, 3.0 and 4.0 wt%, respectively. The weight loss from200 to 400 °C corresponds to the thermal decomposition ofCu(OH)2 (Eq. (4)) and Fe(OH)3 (Eq. (5)), forming CuO and Fe2O3,respectively (Bonura et al., 2014a). The DTA curve (Fig. 6B) of thecorresponding precursors represents an endothermic peak. Theprecursor weight losses in this step were 12.2%, 10.2%, 9.9%, 8.0%and 11.3% for the CeO2 amount of 0, 1.0, 2.0, 3.0 and 4.0 wt%,respectively.

Cu(OH)2-CuOþH2O (4)

2Fe(OH)3-Fe2O3þ3H2O (5)

The Ce2O3 conversion was unstable in air and a portion of theCe2O3 was oxidized to form CeO2 during the Ce(OH)3 to formCe2O3 step (Eq. (6)) (Li et al., 2014a; Mo et al., 2005). CeO2 and CuOform a solid solution during the oxidation process. However, theweight was increased from Ce2O3 to CeO2 (Eq. (7)). Therefore, theweight loss of the precursors modified with CeO2 is less than thatof the CuO-Fe2O3 catalyst precursor.

2Ce(OH)3-Ce2O3þ3H2O (6)

2Ce2O3þO2-4CeO2 (7)

No significant weight change was observed at temperatures

Page 6

Fig. 6. (A) TGA and (B) DTA profiles for precursors of CuO-Fe2O3-CeO2 catalysts with CeO2 content of (a) 0, (b) 1.0, (c) 2.0, (d) 3.0, and (e) 4.0 wt%.

Fig. 7. H2-TPR profiles for CuO-Fe2O3-CeO2 with CeO2 content of (a) 0, (b) 1.0,(c) 2.0, (d) 3.0, and (e) 4.0 wt%. The solid curves are experimental curves andbroken curves are Gaussian multi-peak fitting curves.

Table 2Temperatures and area distributions of the CuO-Fe2O3-CeO2 reduction peaks fordifferent CeO2 contents.a

CeO2 content (wt%) Peak α Peak β Peak γ Total area

T (°C) Areab T (°C) Area T (°C) Area

0 172 0.89 219 7.30 247 3.26 11.451.0 156 2.31 178 2.39 195 3.34 8.042.0 167 1.51 178 2.30 198 1.47 5.283.0 155 3.02 177 2.67 187 3.05 8.744.0 170 2.23 182 0.93 203 7.73 10.89

a The results were measured based on the H2-TPR profiles and the area dis-tributions were calculated by integrating of the areas under the peaks.

b The peak area unit is �104 units.

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–20 15

above 400 °C (Mo et al., 2005). Therefore, a 400 °C calcinationtemperature was adopted to ensure a complete decomposition ofthe catalyst precursors and minimize sintering phenomena.

3.6. H2-TPR analysis of catalysts

The CuO-Fe2O3-CeO2 catalysts with different CeO2 amountswere investigated via H2-TPR to study the redox abilities of thecatalysts and the effects of CeO2 on the reduction of Cu species.The results of these analyses are displayed in Fig. 7. The H2-TPRprofiles peaks of the CeO2 modified CuO-Fe2O3 catalysts are si-milar but differ from those of the CuO-Fe2O3 catalysts. ThreeGaussian fitting peaks (α, β and γ) are shown for all of the catalystsin the H2-TPR profiles. When the temperature was below 300 °C,CuO and Fe2O3 had been reduced to Cu and Fe3O4, respectively.The reductions of CeO2 often occur at high temperatures(4300 °C), resulting in the formation of peaks (Maciel et al.,2011). The three reduction peaks (α, β and γ) correspond to thereductions of highly dispersed CuO, bulk CuO (Xu et al., 2016; Zenget al., 2013) and Fe2O3 (Cao et al., 2008), respectively (Avgour-opoulos and Ioannides, 2003). The temperatures and areas of eachreduction peak are summarized in Table 2.

According to Table 2, the α hydrogen reduction peaks of theCuO-Fe2O3-CeO2 catalysts are centered at 156 °C, 167 °C 155 °C and170 °C for the catalyst with a CeO2 content of 1.0, 2.0, 3.0 and4.0 wt%, respectively. These values are lower than the tempera-tures associated with the CuO-Fe2O3 catalyst (172 °C), indicatingthat the CeO2 modification can decrease the reduction tempera-ture and increase the reducibility of highly dispersed CuO to in-crease the reducibility of the catalysts. The observed lowest tem-perature of peak α at 155 °C, corresponded to the CuO-Fe2O3-CeO2

catalyst with 3.0 wt% CeO2, suggesting that this catalyst exhibitedan optimum reducing behavior. The ratios of the α reduction peakarea to the total area of the CuO-Fe2O3-CeO2 reduction peaks were7.77%, 28.7%, 28.6%, 34.6% and 20.5%, for the CeO2 amount of 0, 1.0,2.0, 3.0 and 4.0 wt%, respectively. These results suggest that theproportion of highly dispersed CuO was increased due to the CeO2

modification. Moreover, the CuO-Fe2O3-CeO2 catalyst with 3.0 wt%CeO2 possessed the largest proportion of highly dispersed CuO.The total hydrogen consumption of CeO2 modified CuO-Fe2O3

catalysts was decreased compared to the CuO-Fe2O3 catalyst. TheXRD results suggest that a portion of the copper and iron em-bedded in ceria form a solid solution, and the solid solution be-comes more difficult to be reduced (Han et al., 2013; Wang et al.,2005). Therefore, the total hydrogen consumption of CeO2 mod-ified CuO-Fe2O3 catalysts was decreased.

Page 7

Fig. 8. FT-IR spectra of pyridine adsorbed on CuO-Fe2O3 (a) and CuO-Fe2O3-CeO2

catalysts with 3.0% wt% CeO2 (b) at a desorption temperature of 150 °C.

Table 3Effect of the CeO2 amount on the catalytic hydrogenation of CO2 to DME onCuO-Fe2O3-CeO2/HZSM-5 catalysts.a

CeO2 con-tent (wt%)

CO2 conversion(mol%)

Products selectivity (mol%) DME yield(mol%)

DME CH3OH CO CH4

0 12.3 18.3 0.9 30.5 50.3 2.31.0 18.1 52.0 2.1 25.4 20.5 9.42.0 18.8 56.9 3.5 23.8 15.8 10.73.0 20.9 63.1 5.2 24.8 6.8 13.24.0 17.1 50.4 4.6 39.5 5.6 8.6

a Reaction conditions: T¼260 °C, P¼3.0 MPa, GHSV¼1500 mL gcat�1 h�1, andV(H2)/V(CO2)¼4.

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–2016

3.7. Surface acidity analysis

Fig. 8 shows the FTIR spectra of the adsorbed pyridine on theCuO-Fe2O3 and CuO-Fe2O3-CeO2 (CeO2¼3.0 wt%) catalysts. Thespectra were normalized and recorded under identical operatingconditions. The two characteristic bands at ca. 1450 and 1540 cm�1

are associated with pyridine adsorption on Lewis acid sites andBrønsted acid sites, respectively (Wang et al., 2007). The weak bandat 1442 cm�1 for CuO-Fe2O3 and the intense band at 1438 cm�1 forCuO-Fe2O3-CeO2 (CeO2¼3.0 wt%) indicated that the modified ofCuO-Fe2O3 catalyst with Ce increased the amount of Lewis acid sites.The Ce metal cations possess empty f orbits. The empty f orbit canprovide potential locations for Lewis acid sites (Wang et al., 2007).Therefore, the empty f orbits modify the acidity of the CuO-Fe2O3

catalysts. Moreover, the shift towards lower wavenumbers can betaken as an indicative of a lower acid strength of the Lewis acid sites(García-Trenco and Martínez, 2012). The CuO-Fe2O3-CeO2 (CeO2

¼3.0 wt%) band at 1546 cm�1 demonstrates that the Brønsted acidsites were formed after the Ce modification. This result indicatesthat the acid is altered by the CeO2 modification. The observed bandat 1476 cm�1 for both catalysts has also been identified in the FTIRspectra of the adsorbed pyridine on HY zeolites (Boréave et al.,

Fig. 9. NH3-TPD profiles of the CuO-Fe2O3-CeO2 catalysts modified by (a) 0, (b) 1.0,(c) 2.0, (d) 3.0, and (e) 4.0 wt% CeO2 calcined at 400 °C without H2 reduction.

1997), and possibly results from the interactions between the ad-sorbed pyridine and metal oxides (Tang et al., 2010). The 1608 cm�1

band is associated with pyridine adsorption vibrations at Brønstedacid sites (Stevens, 2003).

Fig. 9 shows the NH3-TPD profiles of the CuO-Fe2O3 catalystsmodified by different CeO2 amounts before H2 reduction. FromFig. 9, three NH3 desorption peaks were detected at 50–200 °C(peak α), 200-350 °C (peak β), and 350–670 °C (peak γ), indicatingthe presence of three strength acid sites on the catalysts (Gaoet al., 2008). The α, β, and γ peaks corresponded to the desorptionof the adsorbed NH3 on weak, medium, and strong acid sites, re-spectively (Jin et al., 2007). It can be found that after different CeO2

amounts were added into the CuO-Fe2O3 catalyst, the intensities ofweak acid sites were enhanced, indicating increased weak acidsites of the CuO-Fe2O3 catalysts modified by different CeO2

amounts and consistent with the Py-IR result. When the content ofCeO2 was 3.0 wt%, the acid intensity of weak acid site became thestrongest. However, the γ peak of the strong acid sites becamebroadened after the Ce modification, implying that CeO2 decreasedthe catalyst strong acidity.

3.8. Catalytic hydrogenation of CO2 to DME on CuO-Fe2O3-CeO2

/HZSM-5

3.8.1. Effect of the CeO2 contentThe catalytic performances of the CuO-Fe2O3-CeO2/HZSM-5

catalysts with different amounts of CeO2 during DME synthesisfrom CO2 hydrogenation are presented in Table 3. According toTable 3, the CeO2 modifier can regulate the product distribution forDME synthesis via CO2 hydrogenation. The CO2 conversion andDME selectivity are significantly improved compared with thenon-modified CuO-Fe2O3/HZSM-5, and inhibited the production ofCO and CH4. The CeO2 addition caused the CH4 selectivity mono-tonously decrease from 50.3% to 5.6%, with a minimum value of5.6% observed for the CuO-Fe2O3-CeO2/HZSM-5 catalyst with4.0 wt% CeO2. Moreover, the CO2 conversion and the DME se-lectivity were increased when the CeO2 amount was increasedfrom 0 to 3.0 wt% and decreased when the CeO2 content was in-creased above 3.0 wt%. The CuO-Fe2O3-CeO2 catalyst with 3.0 wt%CeO2 exhibited the maximum CO2 conversion and DME selectivityvalues of 20.9% and 63.1%, respectively, indicating that specificCeO2 additions can promote the catalytic performance.

The XRD analysis and Raman spectra showed that the CeO2

modification can increase the CuO dispersion and lead to the for-mation of a solid solution, which may affect the catalytic hydro-genation. The specific surface area was increased as the CeO2

amount was increased from 0 to 3.0 wt% during the nitrogen ad-sorption-desorption process, but declined when the CeO2 amountwas increased from 3.0 to 4.0 wt%. This indicates that the CeO2

modification can change the specific surface areas of theCuO-Fe2O3-CeO2 catalysts. The larger specific surface areas can

Page 8

Fig. 10. Effect of temperature on catalytic CO2 hydrogenation to DME forCuO-Fe2O3-CeO2/HZSM-5. Reaction conditions: T¼240–280 °C, P¼3.0 MPa,GHSV¼1500 mL gcat�1 h�1, and V(H2)/V(CO2)¼4.

Fig. 11. Effects of the reaction pressure on the catalytic CO2 hydrogenation to formDME for CuO-Fe2O3-CeO2/HZSM-5. Reaction conditions: T¼260 °C, P¼2.0–4.0 MPa,GHSV¼1500 mL gcat�1 h�1, and V(H2)/V(CO2)¼4.

Fig. 12. Effect of the space velocity on the catalytic CO2 hydrogenation to form DMEfor CuO-Fe2O3-CeO2/HZSM-5. Reaction conditions: T¼260 °C, P¼3.0 MPa,GHSV¼1500–3500 mL gcat�1 h�1, and V(H2)/V(CO2)¼4.

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–20 17

increase the absorption and/or the number of CO2 and H2 reactionsites, which may be partially responsible for improving the cata-lytic hydrogenation process. Combined the H2-TPR results, thereduction temperature of the highly dispersed CuO was decreasedafter modified by CeO2, in addition, the proportion of highly dis-persed CuO was increased, improving the activity of the catalyst.The results of pyridine FTIR spectra and NH3-TPD suggest thatmodifying the CuO-Fe2O3 catalyst with CeO2 can alter the acidtype, increase the amount of total acid sites (Lewis acid sites andBrønsted acid sites), and enhance the acid intensity of weak acidsite, which can improve the CO2 hydrogenation to DME catalyticperformance.

3.8.2. Reaction temperature effectFig. 10 shows the hydrogenation of CO2 to form DME on the

CuO-Fe2O3-CeO2/HZSM-5 bifunctional catalysts with 3.0 wt% CeO2

for temperatures ranging from 240 to 280 °C. Fig. 10 illustrates thatthe CH3OH selectivity was increased with increasing the tem-perature. In addition, the CO2 conversion and DME selectivity wereincreased when the reaction temperature rose from 240 to 260 °C,reaching maximum values of 20.9% and 63.1% at 260 °C, respec-tively. However, further increasing the reaction temperaturecaused the CO2 conversion and the DME selectivity to decrease.According to the Arrhenius law, increasing the temperature duringthe CO2 hydrogenation can enhance the reaction rate constant andimprove the reaction rate of the DME synthesis. However, the DMEsynthesis from the CO2 hydrogenation is a reversible exothermicreaction. Increasing the temperature will cause the reaction tofavor the reverse process, causing the DME selectivity to decrease.Moreover, a higher temperature can cause the Cu species to sinterand crystallize, decreasing the catalyst activity and coke formation,provoking partial deactivation of the HZSM-5 catalyst (Zha et al.,2013). Therefore, the CO2 conversion and DME selectivity initiallyincrease and subsequently decrease with increasing the tem-perature. As a result, the CO2 hydrogenation to form DME synth-esis was investigated on the CuO-Fe2O3-CeO2/HZSM-5 bifunctionalcatalysts at 260 °C.

3.8.3. Reaction pressure effectSimulations were conducted for the pressures ranging from

2.0 to 4.0 MPa while all other variables were maintained constant.As shown in Fig. 11, the CH3OH selectivity was increased as thereaction pressure was increased from 2.0 to 4.0 MPa. In addition,the CO2 conversion was increased when the pressure was

increased from 2.0 to 3.0 MPa, but decreased when the reactionpressure was increased from 3.0 to 4.0 MPa. The DME selectivityalso exhibited similar trends. This results imply that the directCO2-to-DME hydrogenation reaction was a volume decreasingconsecutive reaction, suggesting that high reaction pressure mayincrease the DME selectivity. However, the CO2 conversion (from20.9% to 17.2%) and DME selectivity (from 63.1% to 55.5%) weredecreased when the pressure was increased from 3.0 to 4.0 MPa.This trend was most likely caused by the accumulation of water,which cannot be removed from the fixed bed reactor during theCO2 hydrogenation to form DME process and may decrease thecatalyst activity (Bonura et al., 2014b; Liu et al., 2015). The resultssuggest that the optimal DME synthesis pressure is 3.0 MPa.

3.8.4. Space velocity effectsThe CuO-Fe2O3-CeO2/HZSM-5 catalyst with 3.0 wt% CeO2 was

used to synthesize DME from CO2 and H2, and the space velocityeffect was investigated over a range of 1500–3500 mL gcat�1 h�1,Fig. 12. The CO2 conversion and DME selectivity were decreased asthe space velocity was increased from 1500 to 3500 mL gcat�1 h�1.However, the CH3OH selectivity was increased from 5.2% to 11.3%as the space velocity was increased from 1500 to

Page 9

Fig. 13. CO2 conversion (A) and DME selectivity (B) effects of CuO-Fe2O3/HZSM-5 and CuO-Fe2O3-CeO2/HZSM-5 ( 3.0 wt%) based on the time on stream stability. The TGA (C),and the TPO profiles (D) of the used catalysts.

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–2018

3500 mL gcat�1 h�1. The contact time between the CO2/H2 mixtureand the catalyst surface was decreased by increasing the spacevelocity, leading to an insufficient contact between the CO2/H2

mixture and the active catalysts sites. Methanol could notpromptly dehydrate to DME on the HZSM-5 catalyst surface andrapidly left the fixed bed reactor (An et al., 2008). Therefore, in-creasing the space velocity caused the DME selectivity to decreasefrom 63.1% to 47.5% and the methanol selectivity to increase. Thesefindings suggest that the optimal space velocity for DME synthesisis 1500 mL gcat�1 h�1.

3.8.5. Stabilities and coke analysisThe stabilities of the CuO-Fe2O3/HZSM-5 and 3.0 wt% CeO2

modified CuO-Fe2O3-CeO2/HZSM-5 catalysts were studied in thecontext of DME synthesis via CO2/H2 at 260 °C and 3.0 MPa, withGHSV¼1500 mL gcat�1 h�1 and V(H2)/V(CO2)¼4. The results wererecorded every 0.5 h, Fig. 13. The CO2 conversion and DME se-lectivity of the CuO-Fe2O3-CeO2/HZSM-5 catalyst with 3.0 wt%CeO2 remained almost constant throughout the 15-h reaction at19.9% and 58.9%, respectively. However, the CO2 conversion andDME selectivity of the CuO-Fe2O3/HZSM-5 catalysts sharply de-clined with the time-on-stream, indicating that the CeO2 mod-ification can improve the CuO-Fe2O3/HZSM-5 catalyst stability.Modifying the CuO-Fe2O3 catalyst with CeO2 improved the CuOdispersion, resulting in less amalgamated fine Cu crystal (Tan et al.,2005) and increases the catalyst activity and stability.

Fig. 13C shows the TGA curves of the used catalysts. Accordingto the TGA curve of the used CuO-Fe2O3-CeO2/HZSM-5 (3.0 wt%)

catalyst, the weight loss at 400 °C is attributed to the coke com-bustion (Tonelli et al., 2015), indicating that slight carbon de-position occurred on the catalyst surface after the 15-h reaction.Thus, DME selectivity and CO2 conversion slightly decreased astime went on. Nevertheless, the used CuO-Fe2O3/HZSM-5 catalystexhibited a large weight loss above 400 °C after the 6-h reaction,indicating that modifying the CuO-Fe2O3 catalyst with CeO2 caninhibit the coke generation. Fig. 13D shows the TPO profiles of theused catalysts, and three peaks (α, β and γ) are shown for the usedCuO-Fe2O3/HZSM-5 and CuO-Fe2O3-CeO2/HZSM-5 (3.0 wt%) cata-lyst. The α peak corresponded to the oxidation of Cu to CuO andFe3O4 to Fe2O3, and the β and γ peaks were the oxidation pro-cesses of the monatomic carbon and whisker carbon, respectively(Hou et al., 2003). The peak β of the used CuO-Fe2O3-CeO2/HZSM-5 (3.0 wt%) catalyst significantly became weakened compared tothe used CuO-Fe2O3/HZSM-5 catalyst, indicating that carbon de-position was suppressed because of the formation of solid solution(Hou et al., 2003; Sierra et al., 2011) and then exhibited a highactivity and stability.

4. Conclusions

The CuO-Fe2O3-CeO2 catalysts with various CeO2 amounts weresuccessfully prepared using a homogeneous precipitation method.The CuO-Fe2O3-CeO2 was mixed with HZSM-5 zeolite and used inthe CO2 hydrogenation to produce DME. The results showed thatthe CeO2 modified CuO-Fe2O3 catalysts formed a stable solid

Page 10

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–20 19

solution, which promoted CuO dispersion, decreased the CuOcrystallite size, decreased the reduction temperature of highly dis-persed CuO, altered the specific surface areas of CuO-Fe2O3-CeO2

catalysts, and improved the catalytic activity of the CuO-Fe2O3-CeO2

catalyst. The addition of CeO2 to the CuO-Fe2O3 catalyst increasedthe amount of Lewis acid sites and Brønsted acid sites, and en-hanced the acid intensity of the weak acid sites, both of which in-creased the catalytic performance of the CO2 hydrogenation to formDME. Reactions were conducted using a CuO-Fe2O3-CeO2/HZSM-5catalyst with 3.0 wt% CeO2 at 260 °C and 3.0 MPa with a gaseoushourly space velocity of 1500 m gcat�1 h�1, resulting in a 20.9% CO2

conversion and 63.1% DME selectivity.

Acknowledgement

This work was supported by National Natural Science Foun-dation of China, China (21366004, 21425627), Guangxi ZhuangAutonomous Region special funding of distinguished experts, andOpen Project of Guangxi Key Laboratory of Petrochemical ResourceProcessing and Process Intensification Technology (2015K004).

References

An, X., Zuo, Y.Z., Zhang, Q., Wang, D.Z., Wang, J.F., 2008. Dimethyl ether synthesisfrom CO2 hydrogenation on a CuO-ZnO-Al2O3-ZrO2/HZSM-5 bifunctional cata-lyst. Ind. Eng. Chem. Res. 47, 6547–6554.

Avgouropoulos, G., Ioannides, T., 2003. Selective CO oxidation over CuO-CeO2 cat-alysts prepared via the urea–nitrate combustion method. Appl. Catal. A: Gen.244, 155–167.

Bera, P., Gayen, A., Hegde, M.S., Lalla, N.P., Spadaro, L., Frusteri, F., Arena, F., 2003.Promoting effect of CeO2 in combustion synthesized Pt/CeO2 catalyst for COoxidation. J. Phys. Chem. B 107, 6122–6130.

Bera, P., Priolkar, K.R., Sarode, P.R., Hegde, M.S., Emura, S., Kumashiro, R., Lalla, N.P.,2002. Structural investigation of combustion synthesized Cu/CeO2 catalysts byEXAFS and other physical techniques: formation of a Ce1�xCuxO2-delta solidsolution. Chem. Mater. 14, 3591–3601.

Bonura, G., Cordaro, M., Cannilla, C., Arena, F., Frusteri, F., 2014a. The changingnature of the active site of Cu-Zn-Zr catalysts for the CO2 hydrogenation re-action to methanol. Appl. Catal. B: Environ. 152–153, 152–161.

Bonura, G., Cordaro, M., Cannilla, C., Mezzapica, A., Spadaro, L., Arena, F., Frusteri, F.,2014b. Catalytic behaviour of a bifunctional system for the one step synthesis ofDME by CO2 hydrogenation. Catal. Today 228, 51–57.

Boréave, A., Auroux, A., Guimon, C., 1997. Nature and strength of acid sites in HYzeolites: a multitechnical approach. Microporous Mater. 11, 275–291.

Cao, J.-L., Wang, Y., Yu, X.-L., Wang, S.-R., Wu, S.-H., Yuan, Z.-Y., 2008. MesoporousCuO–Fe2O3 composite catalysts for low-temperature carbon monoxide oxida-tion. Appl. Catal. B: Environ. 79, 26–34.

Chen, J., Zhan, Y., Zhu, J., Chen, C., Lin, X., Zheng, Q., 2010. The synergetic mechanismbetween copper species and ceria in NO abatement over Cu/CeO2 catalysts.Appl. Catal. A: Gen. 377, 121–127.

Frusteri, F., Cordaro, M., Cannilla, C., Bonura, G., 2015. Multifunctionality ofCu–ZnO–ZrO2/H-ZSM5 catalysts for the one-step CO2-to-DME hydrogenationreaction. Appl. Catal. B: Environ. 162, 57–65.

Gamarra, D., Munuera, G., Hungria, A.B., Fernandez-Garcia, M., Conesa, J.C., Midgley,P.A., Wang, X.Q., Hanson, J.C., Rodriguez, J.A., Martinez-Arias, A., 2007. Struc-ture-activity relationship in nanostructured copper-ceria-based preferential COoxidation catalysts. J. Phys. Chem. C 111, 11026–11038.

Gao, J., Guo, J., Liang, D., Hou, Z., Fei, J., Zheng, X., 2008. Production of syngas viaautothermal reforming of methane in a fluidized-bed reactor over the com-bined CeO2–ZrO2/SiO2 supported Ni catalysts. Int. J. Hydrog. Energy 33,5493–5500.

García-Trenco, A., Martínez, A., 2012. Direct synthesis of DME from syngas on hy-brid CuZnAl/ZSM-5 catalysts: new insights into the role of zeolite acidity. Appl.Catal. A: Gen. 411–412, 170–179.

Ge, Q., Huang, Y., Qiu, F., Li, S., 1998. Bifunctional catalysts for conversion ofsynthesis gas to dimethyl ether. Appl. Catal. A: Gen. 167, 23–30.

Han, X., Yu, Y., He, H., Zhao, J., 2013. Low CO content hydrogen production fromoxidative steam reforming of ethanol over CuO-CeO2 catalysts at low-tem-perature. J. Energy Chem. 22, 861–868.

Hornés, A., Hungría, A.B., Bera, P., Cámara, A.L., Fernández-García, M., Martínez-Arias, A., Barrio, L., Estrella, M., Zhou, G., Fonseca, J.J., Hanson, J.C., Rodriguez, J.A., 2010. Inverse CeO2/CuO catalyst as an alternative to classical direct config-urations for preferential oxidation of CO in hydrogen-rich stream. J. Am. Chem.Soc. 132, 34–35.

Hou, Z., Yokota, O., Tanaka, T., Yashima, T., 2003. Investigation of CH4 reformingwith CO2 on meso-porous Al2O3-supported Ni catalyst. Catal. Lett. 89, 121–127.

Huang, Y., Long, B., Tang, M., Rui, Z., Balogun, M.-S., Tong, Y., Ji, H., 2016. Bifunctionalcatalytic material: An ultrastable and high-performance surface defect CeO2

nanosheets for formaldehyde thermal oxidation and photocatalytic oxidation.Appl. Catal. B: Environ. 181, 779–787.

Jia, A.P., Jiang, S.Y., Lu, J.Q., Luo, M.F., 2010. Study of catalytic activity at theCuO-CeO2 interface for CO oxidation. J. Phys. Chem. C 114, 21605–21610.

Jia, G.X., Tan, Y.S., Han, Y.Z., 2006. A comparative study on the thermodynamics ofdimethyl ether synthesis from CO hydrogenation and CO2 hydrogenation. Ind.Eng. Chem. Res. 45, 1152–1159.

Jin, D., Zhu, B., Hou, Z., Fei, J., Lou, H., Zheng, X., 2007. Dimethyl ether synthesis viamethanol and syngas over rare earth metals modified zeolite Y and dual Cu–Mn–Zn catalysts. Fuel 86, 2707–2713.

Li, J., Han, Y., Zhu, Y., Zhou, R., 2011. Purification of hydrogen from carbon monoxidefor fuel cell application over modified mesoporous CuO–CeO2 catalysts. Appl.Catal. B: Environ. 108–109, 72–80.

Li, L., Song, L., Chen, C., Zhang, Y., Zhan, Y., Lin, X., Zheng, Q., Wang, H., Ma, H., Ding,L., Zhu, W., 2014a. Modified precipitation processes and optimized coppercontent of CuO–CeO2 catalysts for water–gas shift reaction. Int. J. Hydrog. En-ergy 39, 19570–19582.

Li, X., Lu, A., Yang, H., 2014b. Structure of ZnO–Fe2O3–P2O5 glasses probed by Ramanand IR spectroscopy. J. Non-Cryst. Solids 389, 21–27.

Liu, R.-W., Qin, Z.-Z., Ji, H.-B., Su, T.-M., 2013. Synthesis of dimethyl ether from CO2

and H2 using a Cu–Fe–Zr/HZSM-5 catalyst system. Ind. Eng. Chem. Res. 52,16648–16655.

Liu, R., Tian, H., Yang, A., Zha, F., Ding, J., Chang, Y., 2015. Preparation of HZSM-5membrane packed CuO–ZnO–Al2O3 nanoparticles for catalysing carbon dioxidehydrogenation to dimethyl ether. Appl. Surf. Sci. 345, 1–9.

Liu, X.-M., Lu, G.Q., Yan, Z.-F., 2005. Nanocrystalline zirconia as catalyst support inmethanol synthesis. Appl. Catal. A: Gen. 279, 241–245.

Maciel, C.G., Profeti, L.P.R., Assaf, E.M., Assaf, J.M., 2011. Hydrogen purification forfuel cell using CuO/CeO2–Al2O3 catalyst. J. Power Sources 196, 747–753.

Marbán, G., Fuertes, A.B., 2005. Highly active and selective CuOx/CeO2 catalystprepared by a single-step citrate method for preferential oxidation of carbonmonoxide. Appl. Catal. B: Environ. 57, 43–53.

Mo, Z., Sun, Y., Chen, H., Zhang, P., Zuo, D., Liu, Y., Li, H., 2005. Preparation andcharacterization of a PMMA/Ce(OH)3, Pr2O3/graphite nanosheet composite.Polymer 46, 12670–12676.

Qi, L., Yu, Q., Dai, Y., Tang, C., Liu, L., Zhang, H., Gao, F., Dong, L., Chen, Y., 2012.Influence of cerium precursors on the structure and reducibility of mesoporousCuO-CeO2 catalysts for CO oxidation. Appl. Catal. B: Environ. 119–120, 308–320.

Qin, Z.-Z., Su, T.-M., Ji, H.-B., Jiang, Y.-X., Liu, R.-W., Chen, J.-H., 2015. Experimentaland theoretical study of the intrinsic kinetics for dimethyl ether synthesis fromCO2 over Cu-Fe-Zr/HZSM-5. AICHE J. 61, 1613–1627.

Qin, Z.-Z., Zhou, X.-H., Su, T.-M., Jiang, Y.-X., Ji, H.-B., 2016. Hydrogenation of CO2 todimethyl ether on La-, Ce-modified Cu-Fe/HZSM-5 catalysts. Catal. Commun.75, 78–82.

Rutkowska, M., Macina, D., Mirocha-Kubień, N., Piwowarska, Z., Chmielarz, L., 2015.Hierarchically structured ZSM-5 obtained by desilication as new catalyst forDME synthesis from methanol. Appl. Catal. B: Environ. 174–175, 336–343.

Sierra, I., Ereña, J., Aguayo, A.T., Arandes, J.M., Olazar, M., Bilbao, J., 2011. Co-feedingwater to attenuate deactivation of the catalyst metallic function(CuO–ZnO–Al2O3) by coke in the direct synthesis of dimethyl ether. Appl. Catal.B: Environ. 106, 167–173.

Stevens, R., 2003. In situ infrared study of pyridine adsorption/desorption dynamicsover sulfated zirconia and Pt-promoted sulfated zirconia. Appl. Catal. A: Gen.252, 57–74.

Tan, Y., Xie, H., Cui, H., Han, Y., Zhong, B., 2005. Modification of Cu-based methanolsynthesis catalyst for dimethyl ether synthesis from syngas in slurry phase.Catal. Today 104, 25–29.

Tang, X., Li, J., Sun, L., Hao, J., 2010. Origination of N2O from NO reduction by NH3

over β-MnO2 and α-Mn2O3. Appl. Catal. B: Environ. 99, 156–162.Tonelli, F., Gorriz, O., Arrúa, L., Abello, M.C., 2011. Methanol steam reforming over

Cu/CeO2 catalysts: influence of zinc addition. Quim. Nova 34, 1334–1338.Tonelli, F., Gorriz, O., Tarditi, A., Cornaglia, L., Arrúa, L., Cristina Abello, M., 2015.

Activity and stability of a CuO/CeO2 catalyst for methanol steam reforming. Int.J. Hydrog. Energy 40, 13379–13387.

Tsoncheva, T., Issa, G., Blasco, T., Concepcion, P., Dimitrov, M., Hernandez, S., Ko-vacheva, D., Atanasova, G., Lopez Nieto, J.M., 2013. Silica supported copper andcerium oxide catalysts for ethyl acetate oxidation. J. Colloid Interface Sci. 404,155–160.

Vakili, R., Rahmanifard, H., Maroufi, P., Eslamloueyan, R., Rahimpour, M.R., 2011. Theeffect of flow type patterns in a novel thermally coupled reactor for simulta-neous direct dimethyl ether (DME) and hydrogen production. Int. J. Hydrog.Energy 36, 4354–4365.

Wang, S., Mao, D., Guo, X., Wu, G., Lu, G., 2009. Dimethyl ether synthesis via CO2

hydrogenation over CuO–TiO2–ZrO2/HZSM-5 bifunctional catalysts. Catal.Commun. 10, 1367–1370.

Wang, W., Zhang, Y., Wang, Z., Yan, J.-M., Ge, Q., Liu, C.-J., 2016. Reverse water gasshift over In2O3–CeO2 catalysts. Catal. Today 259, 402–408.

Wang, X., Rodriguez, J.A., Hanson, J.C., Gamarra, D., Martínez-Arias, A., Fernández-García, M., 2005. Unusual physical and chemical properties of Cu in Ce1�xCuxO2

oxides. J. Phys. Chem. B 109, 19595–19603.Wang, X., Zhao, Z., Xu, C., Duan, A., Zhang, L., Jiang, G., 2007. Effects of light rare

earth on acidity and catalytic performance of HZSM-5 zeolite for catalyticcracking of butane to light olefins. J. Rare Earths 25, 321–328.

Wang, Y., Wang, F., Chen, Y., Zhang, D., Li, B., Kang, S., Li, X., Cui, L., 2014. Enhanced

Page 11

X. Zhou et al. / Chemical Engineering Science 153 (2016) 10–2020

photocatalytic performance of ordered mesoporous Fe-doped CeO2 catalysts forthe reduction of CO2 with H2O under simulated solar irradiation. Appl. Catal. B:Environ. 147, 602–609.

Wu, Z., Jin, R., Liu, Y., Wang, H., 2008. Ceria modified MnOx/TiO2 as a superiorcatalyst for NO reduction with NH3 at low-temperature. Catal. Commun. 9,2217–2220.

Xu, C., Hao, X., Gao, M., Su, H., Zeng, S., 2016. Important properties associated withcatalytic performance over three-dimensionally ordered macroporous CeO2–

CuO catalysts. Catal. Commun. 73, 113–117.Yang Lim, J., McGregor, J., Sederman, A.J., Dennis, J.S., 2016. Kinetic studies of CO2

methanation over a Ni/γ-Al2O3 catalyst using a batch reactor. Chem. Eng. Sci.141, 28–45.

Yoo, K.S., Kim, J.-H., Park, M.-J., Kim, S.-J., Joo, O.-S., Jung, K.-D., 2007. Influence ofsolid acid catalyst on DME production directly from synthesis gas over the

admixed catalyst of Cu/ZnO/Al2O3 and various SAPO catalysts. Appl. Catal. A:Gen. 330, 57–62.

Zeng, S., Liu, K., Chen, T., Su, H., 2013. Influence of crystallite size and interface onthe catalytic performance over the CeO2/CuO catalysts. Int. J. Hydrog. Energy38, 14542–14549.

Zha, F., Tian, H., Yan, J., Chang, Y., 2013. Multi-walled carbon nanotubes as catalystpromoter for dimethyl ether synthesis from CO2 hydrogenation. Appl. Surf. Sci.285, 945–951.

Zhan, W., Guo, Y., Gong, X., Guo, Y., Wang, Y., Lu, G., 2012. Surface oxygen activationon CeO2 and its catalytic performances for oxidation reactions. Sci. Sin. Chim.42, 433–445.

Zhang, L., Shi, L., Huang, L., Zhang, J., Gao, R., Zhang, D., 2014. Rational design ofhigh-performance DeNOx catalysts based on MnxCo3–xO4 nanocages derivedfrom metal–organic frameworks. ACS Catal. 4, 1753–1763.