Xiao Fan,Shoujie Ren,Baitang Jin,Shiguang Li,Miao Yu,Xinhua Liang,

1 Department of Chemical and Biochemical Engineering,Missouri University of Science and Technology,Rolla,MO 65409,United States

2 Gas Technology Institute,1700 South Mount Prospect Road,Des Plaines,IL 60018,United States

3 Department of Chemical and Biological Engineering,Rensselaer Polytechnic Institute,Troy,NY 12180,United States

Keywords:CO2 hydrogenation Cu-ZnO based catalyst Iron (Fe)Dimethyl ether (DME)Stability

ABSTRACT A series of iron (Fe) modified CuO-ZnO-ZrO2-Al2O3 (CZZA) catalysts,with various Fe loadings,were prepared using a co-precipitation method.A bifunctional catalyst,consisting of Fe-modified CZZA and HZSM-5,was studied for dimethyl ether(DME)synthesis via CO2 hydrogenation.The effects of Fe loading,reaction temperature,reaction pressure,space velocity,and concentrations of precursor for the synthesis of the Fe-modified CZZA catalyst on the catalytic activity of DME synthesis were investigated.Long-term stability tests showed that Fe modification of the CZZA catalyst improved the catalyst stability for DME synthesis via CO2 hydrogenation.The activity loss,in terms of DME yield,was significantly reduced from 4.2%to 1.4%in a 100 h run of reaction,when the Fe loading amount was 0.5(molar ratio of Fe to Cu).An analysis of hydrogen temperature programmed reduction revealed that the introduction of Fe improved the reducibility of the catalysts,due to assisted adsorption of H2 on iron oxide.The good stability of Femodified CZZA catalysts in the DME formation was most likely attributed to oxygen spillover that was introduced by the addition of iron oxide.This could have inhibited the oxidation of the Cu surface and enhanced the thermal stability of copper during long-term reactions.

1.Introduction

Dimethyl ether (DME) synthesis via CO2hydrogenation has gained considerable interest in recent years,since it is an important intermediate for synthesizing many chemicals.Also,it is a promising potential replacement for diesel fuel,because of its high cetane number,low NOxemission,and low soot production in the exhaust gas from an engine,due to no C—C bond structure [1–3].The synthesis of DME,through a direct route via catalytic CO2hydrogenation on a bifunctional catalyst,has several advantages over the typical two-step process (i.e.,synthesis of methanol and dehydration of methanol to DME).These include breaking the equilibrium constraints of methanol synthesis and,thus,obtaining a higher single-pass CO2conversion [4,5].Typically,DME synthesis,via CO2hydrogenation over bifunctional catalysts,is comprised of three independent reactions,as described by the following [6]:

Methanol synthesis reaction:

Currently,a variety of catalysts for the direct synthesis of DME,via CO2hydrogenation,have been studied,including Cu-ZnO based catalysts,Cu-Mn,CuO-TiO2-ZrO2,and Cu-Ni bimetallic catalysts[7–9].These exhibited good catalytic performance,with CO2conversion and DME selectivity being approximately 15%-29% and 40%-50%,respectively.Among these catalysts,CuO/ZnO/Al2O3(CZA) catalyst is the most investigated for methanol synthesis.Due to its high activity and easy preparation,the CZA catalyst has been used as the CO2hydrogenation catalyst,with solid acid catalysts (e.g.,HZSM-5 and γ-Al2O3) as the dehydration catalyst,to form bifunctional catalysts for DME synthesis via one-step CO2hydrogenation [5,10,11].However,challenges remain for direct DME synthesis over CZA based catalysts,due to catalyst deactivation and poor stability.The major factors limiting the catalytic performance of the CZA catalysts are the deposition of carbonaceous materials (coke) and sintering of copper [6,12–14].In addition,water generated through the reactions could have detrimental effects(e.g.,oxidation and sintering of Cu in the presence of water)on the activity of a hydrogenation catalyst and the acidity of a dehydration catalyst.Consequently,the stability of both catalysts could be significantly affected[15].Therefore,it is imperative that the catalytic performance of bifunctional catalysts should be enhanced by improving their resistance to copper oxidation and sintering,inhibiting coke deposition,extending the catalyst lifetime and,eventually,accomplishing a more efficient utilization of carbon dioxide.

Introduction of inexpensive Fe into the Cu-based catalysts for the catalytic hydrogenation of CO2has been investigated [16–18],but there are no studies of Fe modified CZZA catalysts.The addition of Fe as a promoter in supported metal catalysts,in the form of isolated ions or nanosized oxide crystallites,is helpful for preventing Cu nanoparticles from sintering and inhibiting the oxidation of copper surfaces during a reaction [16,19,20].In this study,Femodified CuO/ZnO/ZrO2/Al2O3(CZZA) catalysts were prepared for DME synthesis.The CZZA catalyst was synthesized and investigated in our previous report [21].Fe-CZZA catalysts,with various Fe loadings (Fe/Cu molar ratio in a range of 0.1–0.7),were physically mixed with HZSM-5 for one-step DME synthesis,and were systematically characterized before and after reactions.The effects of the Fe loading amount,reaction temperature,reaction pressure,space velocity,and precursor concentration for Fe-CZZA synthesis on the catalytic activity of the developed bifunctional catalysts were investigated.

2.Experimental

2.1.Materials

All reagent chemicals,including copper nitrate(Cu(NO3)2.3H2O,99%),zinc nitrate (Zn(NO3)2.6H2O,98%),aluminum nitrate (Al(NO3)3.9H2O,98%),zirconium dinitrate oxide hydrate (ZrO(NO3)2.xH2O,99%),iron nitrate (Fe(NO3)3.9H2O,99.95%),and sodium carbonate (Na2CO3,99%),were purchased from Alfa Aesar.Ammonia-ZSM-5 (SiO2/Al2O3=23:1 molar ratio,surface area=425 m2.g-1) was purchased from Alfa Aesar.High purity H2and CO2gases with admixed molar ratio of 3:1 were purchased from Airgas Inc.,and used as feed gas for DME synthesis.

2.2.Catalyst preparation

To investigate Fe loading effects in this work,an optimal Cu/Zn/Zr/Al molar ratio was fixed at 4:2:1:0.5,as determined in our previous work[10,21].Various loadings of Fe-modified CZZA catalysts,with a Fe/Cu molar ratio in a range of 0.1–0.7,were prepared by a co-precipitation method.First,aqueous nitrate solutions of Cu(NO3)2,Zn(NO3)2,Al(NO3)3,ZrO(NO3)2,and Fe(NO3)3were mixed together in a glass beaker.The mixed solution and an aqueous solution of sodium carbonate were simultaneously added dropwise by two peristaltic tubing pumps into 400 ml of preheated deionized (DI) water (65 °C-70 °C),with vigorous stirring at a speed of 400 r.min-1.At the end of co-precipitation,the pH value was adjusted to 7.0 by an aqueous solution of sodium carbonate and,then,the solution was aged at 70 °C for 30 min.Finally,the obtained precipitates were filtered under vacuum and rinsed several times with warm DI water,dried in an oven at 110 °C overnight and,subsequently,calcined at 400 °C for 5 h,with a heating rate of 2 °C.min-1.The prepared Fe-CZZA catalysts were denoted as xFe-CZZA,in which x represents the molar loading of Fe (i.e.,0.1,0.3,0.5,and 0.7).The commercial ammonia-ZSM-5 was calcined in air at 500 °C for 5 h to obtain HZSM-5.

2.3.Catalyst characterizations

X-ray diffraction (XRD) measurements were made by a Philips X’Pert PRO PW3050 X-ray diffractometer,equipped with a CuKα radiation and a graphite generator.To obtain the crystal structure of fresh and used catalysts,the scan rate was set at 0.5(°).min-1from 5° to 90° at a tube voltage of 45 kV and a current of 40 mA.

The metallic copper surface area(MSA)and Cu dispersion(DCu)were determined by performing N2O-chemisorption measurements with a Micromeritics Autochem II 2920 instrument.Before analysis,the samples were degassed at 250 °C for 2 h in pure helium (He),then reduced in situ at 250 °C for 2 h in 10% H2/He,and further cooled down to 50°C.The N2O-chemisorption process was started by flushing 1% N2O/He gas.The calculation was based on the assumption of a Cu:N2O=2:1 molar ratio titration stoichiometry.

The reducibility of catalysts was analyzed by hydrogen temperature programmed reduction (H2-TPR) under the following conditions:25 mg sample,10% H2mixed with Ar at a flow rate of 20 ml.min-1,heating rate of 10 °C.min-1,using a Micrometritics Autochem II 2920 (Micrometritics instrument Corporation).

2.4.Catalytic tests

A bifunctional catalyst,composed of 0.5 g Fe-CZZA (powder,particle size ≥125 μm,with various Fe loadings of 0.1,0.3,0.5,and 0.7) and 0.5 g HZSM-5 (powder,particle size ≥125 μm,SiO2/Al2O3=23:1 molar ratio),were physically mixed first and,then,loaded into a fixed-bed reactor (O.D.of 12.7 mm and I.D.of 9.525 mm) for DME synthesis.A reduction process was carried out at 250 °C for 10 h with 20 ml.min-1of pure H2under atmospheric pressure.Subsequently,the bifunctional catalyst was tested for DME synthesis by feeding a H2/CO2mixture with a molar ratio of 3:1(18 ml.min-1H2and 6 ml.min-1CO2,space velocity of 720 h-1).The effects of reaction temperature were first studied in a range of 220–280 °C while the pressure was kept at 2.8 MPa by a back pressure regulator.Then,the effects of reaction pressure were investigated with a pressure range of 2.1–4.2 MPa and a reaction temperature at 240 °C.Various space velocities were also tested in a range of 540–1,080 h-1with a reaction temperature of 240 °C and pressure at 2.8 MPa,as well as the effects of precursor concentration (0.1 mol.L-1to 0.6 mol.L-1).The reaction products were analyzed by an online gas chromatography (SRI 8610C),equipped with a TCD detector,a 6-foot HAYESEPD column,and a 6-foot molecular sieve 13X column.Long-term stability tests of CZZA and Fe-CZZA catalysts were conducted at 240 °C,2.8 MPa,18 ml.min-1H2and 6 ml.min-1CO2,with a space velocity of 720 h-1for 100 h.Methanol,DME,and CO were the main products observed.Based on GC data,the carbon balance was higher than 95%.

3.Results and Discussion

3.1.Characterizations of fresh catalyst

Typically,the active sites for methanol synthesis are recognized to be either CuZn alloy[22,23]or Cu-ZnO interfaces[24,25]in a Cu/ZnO/Al2O3catalyst.The ZnO component plays an important role in methanol synthesis and cannot be neglected.In this study,the Cu surface area was measured by N2O-chemisorption.As shown inTable 1,the Cu surface area of a fresh CZZA catalyst was 44.3 m2.g-1.With the introduction of Fe in the CZZA catalyst,there was an obvious decrease in the Cu surface area,compared to that of the fresh CZZA catalyst.This trend became more remarkable with a further increase of Fe loading.For example,with a Fe loading of 0.7,the Cu surface area showed a significant loss of from 44.3 m2.g-1to 20.3 m2.g-1.This could be ascribed to the introduction of Fe into the CZZA catalyst,resulting in the surfaces of Cu active sites being partially covered or blocked [16,18,26].This phenomenon of a decrease in Cu MSA,with the introduction of Fe or other promoters(e.g.,Pd,Ce),is in accordance with earlier reports [27–30].

Table 1N2O chemisorption results of fresh and used CZZA and Fe-modified CZZA catalysts with different Fe loadings

The reduction behavior of the fresh CZZA and Fe-CZZA catalysts,with 0.1,0.3,0.5,and 0.7 Fe loadings,and the effects of the presence of Fe on the reduction of the CuO species were analyzed by H2-TPR.The results are shown in Fig.1.Despite the fact that the CZZA catalyst was modified with different amounts of Fe,all of the Femodified CZZA and the fresh CZZA catalysts displayed a similar reduction profile,with a main reduction peak in a narrow range of 150–310 °C,which was attributed to the reduction process of CuO (Cu2+to Cu0) [4,30,31].However,the reduction peaks of Fe2O3and ZrO2were not detected because they would be typically reduced to a lower valence state of metal at a much higher temperature.It is noteworthy that a lower reduction temperature occurred for all of the Fe-modified CZZA catalysts,indicating that the reducibility of CuO was promoted by the introduction of a Fe element,enhancing the synergistic effects between CuO and Fe2O3[18,20,32].Besides,it was highly suggestive that iron oxides played an important role in improving the reducibility of the catalysts through adsorbing/activating H2[19,20,27].It is conceivable that a lower reduction temperature would be beneficial by inhibiting the growth of metallic Cu crystallites,yielded by H2-reduction,and favorable for keeping the adjacent Cu surface in a more reduced state and,thus,counteracting the oxidizing effect of CO2or H2O.

Fig.1.H2-TPR profiles of CZZA and xFe-CZZA catalysts prepared with different Fe loadings.(x=0.1,0.3,0.5,and 0.7,molar ratio of Fe to Cu).

Although it is difficult for H2to reduce Fe2O3from trivalence(III)to a zero(0)valence state at such low reduction temperatures[16,17],the reduction of trivalent Fe(III) to bivalent Fe(II) could occur at below 300 °C,with an initial temperature at about 230°C[33].The broad reduction peak at 350–750°C was probably associated with the consecutive reduction of Fe2O3to Fe3O4or FeO and,eventually,to Fe.Without an observation of a peak at 300–400 °C,this peak of Fe(III) to Fe(II) could be overlapped by the intense peak of Cu(II)to Cu(0).The slight reduction in Fe(III)would take place on the surface of catalyst and drive the surface oxygen away from bulk Fe2O3,instead of lattice oxygen,therefore forming an oxygen-deficient surface.With exposed defects,the oxygendeficient surface would probably be able to attract adatoms,due to the redox property Fe3+/Fe2+couple from the surrounding catalytic Cu sites,identified as oxygen spillover of Cu [20,27,34,35].This will be discussed in a later section.

The admixed calcined(without reaction)CZZA/HZSM-5 and Fe-CZZA/HZSM-5 bifunctional catalysts were analyzed by XRD.As shown in Fig.2,highly crystallized monoclinic CuO was detected in all tested samples (CZZA/HZSM-5 and Fe-CZZA/HZSM-5),and Cu was detected in both used CZZA and Fe-CZZA catalysts.Moreover,it was observed that the diffraction peaks at 43.3° corresponded to the (111) plane of copper,and the diffraction peaks at 31.8° and 56.6° corresponded to the (100) and (110) planes of zincite.These results are consistent with our previous report[21].We have reported that copper oxide species can be entirely reduced to metallic Cu,during H2reduction,before reaction.In addition,no peaks corresponded to iron(III) oxides and zirconium oxides,indicating that iron (III) oxides and zirconium oxides were either too highly dispersed to form very small particles or existed as an amorphous phase in CZZA or Fe-CZZA catalysts.

Fig.2.XRD patterns of calcined catalysts (a) 0.3Fe-CZZA/HZSM-5 and (c) CZZA/HZSM-5 catalysts,and used (b) 0.3Fe-CZZA/HZSM-5 and (d) CZZA/HZSM-5 catalysts,after 100 h of DME synthesis reaction.

3.2.Catalytic performance of Fe-CZZA/HZSM-5 catalysts for DME synthesis

3.2.1.Effects of Fe loading amount and reaction temperature on catalytic performance

The catalytic performance of Fe-modified CZZA catalysts,with different Fe loadings (from 0.1 to 0.7),was evaluated in the hydrogenation of CO2to DME at different reaction temperatures from 220 to 280 °C,a pressure of 2.8 MPa,and a space velocity of 720 h-1.Reaction temperature has significant effects on CO2conversion,DME selectivity,and DME yield thermodynamically and kinetically [31,36].As shown in Fig.3,the data collected to evaluate catalytic performance included the conversion of CO2,and the yield of CO,MeOH,and DME.The formation of CO during the reaction is considered as the exclusive carbonaceous byproduct via the reversed water gas shift (RWGS) reaction.The CO2conversion showed an increasing trend with the reaction temperature for all Fe-CZZA catalysts,suggesting that high temperature would favor a reaction thermodynamically.However,the RWGS reaction is endothermic,with an enthalpy change of 41.2 kJ.mol-1,while the reaction of CO2hydrogenation to DME is exothermic,with a total enthalpy change of -72.8 kJ.mol-1(-49.4 kJ.mol-1-23.4 kJ.mol-1) [37–39].This means that a higher reaction temperature would not be favorable for obtaining a higher DME selectivity,but lead to enhanced formation of CO in the final product(Fig.3b).Moreover,as shown in Fig.3d,all tested Fe-CZZA samples had the highest DME yield at 240 °C,which was close to the data reported in our previous study for a CZZA catalyst [21].The optimum reaction temperature for DME synthesis,via CZZA/HZSM-5 bifunctional catalyst,was at 240 °C with the highest DME yield(18.0%).

The 0.1Fe-CZZA/HZSM-5 catalyst showed the best performance,among all of the Fe-CZZA/HZSM-5 catalysts.The optimal catalytic activities of 0.3Fe-CZZA,0.5Fe-CZZA,and 0.7Fe-CZZA were close to each other,which were achieved at around 11% of DME yield,with the reaction temperature at 240 °C.This demonstrated that the activities of the Fe-CZZA catalysts were correlated with the copper surface area obtained from N2O-chemisorption (Table 1).This could probably be due to the introduction of Fe,with a low loading amount (Fe/Cu=0.1),that covered/blocked a small portion of Cu active sites and,consequently,resulted in a smaller drop in catalytic activity.However,by increasing the Fe loading from a low to a high amount (Fe/Cu=0.3,0.5,or 0.7),more Cu active sites were blocked and fewer of them were accessible to catalyze reaction.This was in accordance with the change in the copper surface areas of the fresh Fe-modified CZZA catalysts.In this work,the 0.3Fe-CZZA/HZSM-5 catalyst was chosen for investigation in the rest of this study to determine the effects of reaction conditions(i.e.,reaction pressure,space velocity,and precursor concentration)on catalytic performance.This is because the 0.3Fe-CZZA/HZSM-5 catalyst(DME loss in an absolute value of 1.5%)showed much better stability than that of the 0.1Fe-CZZA/HZSM-5 catalyst (DME loss in an absolute value of 3.7%)in 100 h DME synthesis reaction.This will be discussed in detail in the later part.

3.2.2.Effects of reaction pressure on catalytic performance

To study the effects of pressure on catalytic performance,a bifunctional catalyst composed of 0.3Fe-CZZA and HZSM-5 was used.From the results (shown in Fig.4),a higher catalytic activity was obtained at higher pressures(3.5 MPa and 4.2 MPa);however,their losses (absolute value) of DME yield after 24 h were 1.10%and 2.05%,respectively.Although the catalytic activity at lower pressures (2.1 MPa and 2.8 MPa) was lower than that at higher pressures (3.5 and 4.2 MPa),their DME yield losses for them in a 24 h run of reaction were much less (0.44% for 2.1 MPa,and 0.50% for 2.8 MPa).Importantly,the aim of the introduction of Fe into a CZZA catalyst was to protect the copper active sites and,thus,further enhance catalytic stability in a long time reaction[16,40].Due to the low conversion of CO2obtained when the reaction pressure was 2.1 MPa and the high loss of activity at 3.5 MPa and 4.2 MPa,2.8 MPa was used for the following catalytic tests in this work.Specifically,with higher reactant concentrations and favored equilibrium conversion under higher pressures,enhanced activity was observed,thereby exhibiting a higher conversion of CO2.However,as shown in Fig.4b,too high or too low reaction pressures would obtain a higher yield of CO than that at 2.8 MPa,which negatively affected the overall catalytic activities of bifunctional catalysts.Besides,a high reaction pressure was typically needed for higher requirements for material in the reaction system and for standards of safety[40].Applications of low reaction pressure will become a promising trend in the future production of DME in industry,from the viewpoints of reaction safety,cost consumption,and sustainable production.Based on these results,reaction conditions at a temperature of 240 °C and a pressure of 2.8 MPa were applied in the rest of all experiments.

Fig.3.Effects of reaction temperature and Fe loading amount on(a)CO2 conversion,(b)CO yield,(c)MeOH yield,and(d)DME yield for CZZA/HZSM-5 and xFe-CZZA/HZSM-5 bifunctional catalysts(x varied from 0.1 to 0.7).Reaction conditions:temperature 220–280°C,pressure at 2.8 MPa,H2/CO2 molar ratio at 3:1,and space velocity at 720 h-1.

Fig.4.Effects of reaction pressure on(a)CO2 conversion,(b)CO yield,(c)MeOH yield,and(d)DME yield for 0.3Fe-CZZA/HZSM-5 bifunctional catalyst.Reaction conditions:pressure 2.1–4.2 MPa,temperature at 240 °C,H2/CO2 molar ratio at 3:1,and space velocity at 720 h-1.

Fig.5.Effects of space velocity on CO2 conversion,yield of CO,MeOH,and DME for 0.3Fe-CZZA/HZSM-5 bifunctional catalyst.Reaction conditions:pressure at 2.8 MPa,temperature at 240 °C,H2/CO2 molar ratio at 3:1,and space velocity 540 h-1 to 1,080 h-1.

3.2.3.Effects of space velocity on catalytic performance

Fig.5 illustrates the effects of space velocity of admixed feed gas(CO2+H2) on the catalytic performance of a 0.3Fe-CZZA/HZSM-5 bifunctional catalyst for one-step synthesis of DME.Space velocity is an important parameter that represents the efficiency of a reaction taking place in the reactor [41–43].Various space velocities,including 540,720,900,and 1,080 h-1were tested for direct DME synthesis reaction with a H2/CO2molar ratio of 3,at a pressure of 2.8 MPa and temperature of 240 °C.As shown in Fig.5,CO2conversion and DME yield dropped continuously from 21.7%to 16.7% and from 12.4% to 7.4%,respectively,with a certain augmentation of space velocity from 540 to 1,080 h-1.This might be ascribed to the fact that increased space velocity would shorten the residence time between a reactant gas (CO2/H2) and a Fe-CZZA/HZSM-5 bifunctional catalyst,leading to insufficient reaction time between the CO2/H2mixing gases and the active centers of the catalyst [31,36,39].Compared with the effects of temperature and pressure shown in Figs.3 and 4,the effects of the space velocity of feed gas on the catalytic activity of Fe-CZZA/HZSM-5 bifunctional catalysts were the least in a range of 540–1,080 h-1.The highest DME yield was achieved at a space velocity of 540 h-1with a reaction temperature of 240 °C and a pressure of 2.8 MPa.The yield of CO was constantly kept lower than 10%,which implied that the change of space velocity had little effect on the formation of CO under the reaction conditions in this work.Low space velocity would enhance the overall reaction conversion by increasing the contact time of reactant gases with catalysts,as well as a longer contact time between the byproduct H2O and the catalysts,which may deactivate the catalyst via Cu oxidation.Therefore,we simply chose the space velocity of 720 h-1for our experiments.

3.2.4.Effects of precursor concentration on catalytic performance

After investigating the reaction parameters,including Fe loading amounts,reaction temperatures,pressures,and space velocities,we further studied the effects of precursor concentration(0.1 mol.L-1,0.2 mol.L-1,0.4 mol.L-1,and 0.6 mol.L-1) to prepare a 0.3Fe-CZZA catalyst for catalytic performance of DME synthesis,via CO2hydrogenation.Fig.6 clearly indicates that the precursor concentration (less than 0.4 mol.L-1) had very little effect on CO2conversion and DME yield under the reaction conditions in this study.With an increase in precursor concentration,from 0.1 mol.L-1to 0.6 mol.L-1in a 100 h run of reaction,the CO2conversion decreased by an absolute value of 8.5% (from 20.3% to 11.8%),and DME yield dropped by 7.6% (from 11.0% to 3.4%).The yields of CO were very close to each other with different precursor concentrations.The changes in catalytic activity might have been caused by a significant decrease in the Cu surface area,when the concentration of precursor increased.The same phenomenon was observed and reported in our previous study [21],which was attributed to the fact that higher precursor concentrations could reduce the dispersion of copper particles and enhance precipitate aggregation during the co-precipitation process.In the following stability test,a 0.1 mol.L-1solution of precursor was applied in all Fe-modified CZZA catalysts with various Fe loading amounts.

3.2.5.Catalyst stability

The development of catalysts,possessing a good stability and long life,has a significant value for production of DME via CO2hydrogenation in industrial application.In our previous studies[10,21],the CZZA catalyst showed good activity in a synthesis of MeOH,but the stability of the CZZA/HZSM-5 bifunctional catalyst for direct synthesis of DME was still undesirable,especially in the first 24 h of reaction.

A long-term stability test (100 h) of direct DME synthesis reaction was performed for 0.1 mol.L-1Fe-modified CZZA catalysts(with various Fe loadings) mixed with HZSM-5.The CZZA/HZSM-5 bifunctional catalyst was also tested under the same reaction conditions for comparison.Fig.7 shows the experimental results of CO2conversion,and yield of CO,MeOH,and DME at a reaction temperature of 240 °C,pressure of 2.8 MPa,and space velocity at 720 h-1for 100 h of reaction with various Fe loading amounts(0.1–0.7 molar ratio of Fe/Cu)of Fe-CZZA/HZSM-5 bifunctional catalysts.For a fresh CZZA/HZSM-5 catalyst,after 100 h of reaction,the CO2conversion dropped from an initial 26.1% to 24.1%,and DME yield dropped from 18.5% to 14.1% (loss in an absolute value of 4.4%).After the introduction of Fe into a CZZA catalyst,the 0.1Fe-CZZA/HZSM-5 catalyst showed a less deactivation trend of catalytic activity than that of a CZZA/HZSM-5 catalyst in 100 h.The CO2conversion dropped from an initial 22.8% to 19.9%,and DME yield dropped from 14.2% to 10.5% (loss in an absolute value of 3.7%),which showed a slightly enhanced stability,as compared to that of the CZZA/HZSM-5 catalyst.This could be ascribed to the effect of Fe modification at a very low loading,which resulted in a relatively weak interaction between Cu and Fe.

Fig.7.Long-term stability test of CZZA/HZSM-5 and xFe-CZZA/HZSM-5(x varied from 0.1 to 0.7)bifunctional catalysts,with various Fe loadings,for 100 h of DME synthesis:(a) CO2 conversion,(b) CO yield,(c) MeOH yield,and (d) DME yield.Reaction conditions:temperature at 240 °C,pressure at 2.8 MPa,H2/CO2 molar ratio at 3:1,and space velocity at 720 h-1.

With an increment of Fe loadings from 0.1 to 0.3 and from 0.5,and 0.7,respectively,the catalytic activities of Fe-modified CZZA with HZSM-5 catalysts was reduced by some extent.This is in accordance with the loss of Cu active sites and a decrease in Cu dispersion of Fe-modified CZZA catalysts,mentioned in Table 1.Though catalytic activity was a bit lower,the catalyst stability of the CZZA/HZSM-5 bifunctional catalyst was dramatically improved by the introduction of Fe,with various Fe loading amounts(i.e.,0.3,0.5,and 0.7).Specifically,after a 100 h run of reaction,the CO2conversion of the 0.3Fe-CZZA/HZSM-5 catalyst decreased from 20.5%to 18.5%,DME yield decreased from 11.0%to 9.5%(loss in an absolute value of 1.5%);the CO2conversion of the 0.5Fe-CZZA/HZSM-5 catalyst decreased from 20.5%to 19.3%,DME yield decreased from 11.0% to 9.7% (loss in an absolute value of 1.3%);the CO2conversion of 0.7Fe-CZZA/HZSM-5 catalyst decreased from 20.7% to 19.5%,and DME yield decreased from 11.5% to 9.8% (loss in an absolute value of 1.7%).Regardless of the CO2conversion,the production of CO for Fe-CZZA/HZSM-5 bifunctional catalysts was mostly constant throughout the whole reactions,and slightly higher than that of the CZZA/HZSM-5 catalyst.Hence,this series of results suggested that the addition of Fe might enhance the RGWS reaction to form more CO,with a decrease in DME selectivity.As a result,the loss of DME yield for the 0.5Fe-CZZA/HZSM-5 catalyst was only 1.3%,with reference to the initial yield in 100 h,which possessed the best stability and the lowest activity loss among Fe-CZZA/HZSM-5 bifunctional catalysts,with different Fe loadings.Therefore,the results proved that the addition of Fe into the CZZA catalyst system could be helpful for preventing active centers from sintering and inhibiting the deactivation of Cu surfaces during long-term reaction and,thus,significantly improve the stability of CZZA catalysts.Although the CZZA/HZSM-5 bifunctional catalyst showed better catalytic performance at the beginning of DME synthesis reaction,the rapid deactivation rate was remarkable and demanded a prompt solution.From the perspective of economic benefit and environmental protection,enhancing catalyst stability for a long time in industrial production is worthy of considerably more importance,in spite of sacrificing a portion of active sites after catalyst modification by promoters.

The enhanced stability of Fe-CZZA/HZSM-5 could be ascribed to the migration of oxygen adatoms on active Cu sites towards oxygen-deficient iron oxide.Our previous study also indicated that the presence of water in the reaction system could assist Cu sintering and oxidation.This was identified as the main cause of deactivation in this work,especially for bifunctional catalysts with intimate surface contact [6,21].Oxygen atoms (i.e.,CO2and H2O)participated in the whole reaction process for the Fe-CZZA catalyst,which resulted in intensive oxidation of copper and has been considered as the main contribution to the deactivation of the Cu/Zn/Al catalyst [12,37,38,44].With the introduction of iron into the CZZA catalyst,the oxygen on copper surface would be more inclined to migrate to the iron surface through an oxygen spillover process [20,27,34].This phenomenon was typically observed on metal or metal oxides.It has been reported that there is a correlation between oxygen mobility and intrinsic oxide properties,particularly their surface basicity and metal-oxygen bond strengths.Oxygen mobility increases with the increase of oxide basicity and a decrease in the strength of the metal-oxygen bond[45,46].With higher oxygen mobility,the iron oxide can also enhance basicity.For example,Chen et al.[27] studied the temperature programed desorption analysis of CO2on Cu/SiO2,Fe/SiO2,and Cu-Fe/SiO2catalysts,and they found that Fe showed a higher adsorption ability for CO2than Cu did.Therefore,in the Fe-modified CZZA catalyst,Fe2O3could have provided a higher adsorption ability for CO2than Cu did.

3.3.Characterization of used catalysts

The Cu surface areas of all used Fe-CZZA catalysts,with various Fe loadings,were evaluated by N2O-chemisorption in this work(Table 1).The percentage of Cu surface area loss was calculated based on the Cu surface areas of Fe-CZZA catalysts,with various Fe loading amounts,before and after 100 h run of DME synthesis reaction.As shown in Table 2,the loss of the Cu surface area of the CZZA catalyst in a 100 h reaction was 68.2%.Though CZZA showed a much better catalyst stability than that of the CZA catalyst (e.g.,nearly 60% loss of Cu surface area after only 25 h run of reaction,reported by Sun et al.[12]),it was not comparable to Fe modified CZZA catalysts.With the introduction of Fe into a CZZA catalyst,the loss of Cu surface area in a 100 h stability test for 0.1Fe-CZZA,0.3Fe-CZZA,0.5Fe-CZZA,and 0.7Fe-CZZA was 52.1%,18.9%,20.9%,and 14.8%,respectively.These results indicated a significant enhancement of catalyst stability,which made copper active sites energetic for a longer time in the reaction.The prepared catalysts,in the current work,were also compared to other Cu-based catalysts in the literature [12,44,47] and showed more superior stability and much less loss of copper surface area than those of others.It eventually turns out that adding Fe to a Cubased methanol synthesis catalyst was beneficial for preventing copper from thermal sintering and for keeping the copper in a reduced state for a long-term run of reaction.The loss of catalytic activity of all Fe-CZZA/HZSM-5 bifunctional catalysts was correlated with Cu surface area,in almost a linear relationship,when the Fe loadings were less than 0.3.However,for Fe loadings higher than 0.3,there was only a slight change for a copper surface area and catalytic activity.For example,a CZZA/HZSM-5 catalyst showed 68.2% loss of its copper surface area,corresponding to a 4.4% loss of DME yield in an absolute value;0.1Fe-CZZA/HZSM-5 catalyst showed a 52.1%loss of its copper surface area,corresponding to a 3.7%loss of DME yield.Besides,0.3Fe-CZZA/HZSM-5,0.5Fe-CZZA/HZSM-5,and 0.7Fe-CZZA/HZSM-5 catalysts showed an aver-age loss of DME yields in an absolute value of about 1.5%,corresponding to losses in copper surface areas within a range of 14.8% to 20.9%.This result demonstrated the importance of preventing the oxidation of Cu via promoter introduction for the long-term stability of catalyst in a practical industrial application.However,since the mechanism of such bifunctional catalysts with multi-components is complicated and still unclear,more work has to be done in the future,in order to determine the optimal conditions (e.g.,pH value,calcination temperature,and atomic ratio of different promoters) for preparation of catalysts with multicomponents,as well as how each promoter interacts with the Cu active centers,and affects each other in the reaction of CO2hydrogenation to DME.

Table 2Percentage of Cu surface area loss for various xFe-CZZA catalysts

4.Conclusions

A bifunctional catalyst,consisting of Fe-modified CZZA and HZSM-5,was developed for DME synthesis via one-step CO2hydrogenation.Our study indicated that iron provided several advantages in the prevention of sintering and oxidation of copper.The reducibility of CuO benefited from the introduction of Fe element through adsorbing/activating H2and,subsequently,keeping the adjacent Cu surface in a more reduced state,thereby counteracting the oxidizing effect of CO2or H2O.The catalytic performance of all Fe-modified CZZA catalysts,with different Fe loadings,was slightly lower due to the coverage of the Cu surface area,but the stability of Fe-CZZA catalysts was greatly enhanced due to a much lower loss of Cu surface area,as compared to that of the CZZA catalyst.Nearly a 1.3%(absolute value)loss of DME yield,with reference to the initial yield for 0.5Fe-CZZA/HZSM-5 bifunctional catalysts in a 100 h run of reaction,was obtained.This showed a much better stability than other Cu-based methanol synthesis catalysts (e.g.,CZA or CZZA).Via the oxygen spillover between deficient iron oxide and Cu,the oxidation of Cu surface by CO2or H2O and sintering of Cu active centers during the long-term reaction could be effectively mitigated.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the U.S.Department of Energy through contract DE-AR0000806.We thank DOE ARPA-E Project Director,Dr.Grigorii Soloveichik,for his assistance and support.