Ce Du,Linet Gapu Chizema,Emmerson Hondo,Mingliang Tong,Qingxiang Ma,Xinhua Gao,Ruiqin Yang,Peng Lu,*,Noritatsu Tsubaki

1 Zhejiang Provincial Key Lab for Chem.&Bio.Processing Technology of Farm Product,School of Biological and Chemical Engineering,Zhejiang University of Science and Technology,Hangzhou 310023,China

2 Department of Applied Chemistry,School of Engineering,University of Toyama,Gofuku 3190,Toyama 930-8555,Japan

3 State Key Laboratory Cultivation Base of Natural Gas Conversion,Ningxia University,Yinchuan 750021,China

Keywords:Catalyst C--C coupling CO activation Hydrogenation Light olefins Syngas

ABSTRACT Light olefins(C2–C4)are fundamental building blocks for the manufacture of polymers,chemical intermediates,and solvents.In this work,we realized a composite catalyst,comprising MnxZry oxides and SAPO-34 zeolite,which can convert syngas(CO+H2)into light olefins.MnxZry oxide catalysts with different Mn/Zr molar ratios were facilely prepared using the coprecipitation method prior to physical mixing with SAPO-34 zeolite.The redox properties,surface morphology,electronic state,crystal structure,and chemical elemental composition of the catalysts were examined using H2-TPR,SEM,XPS,XRD,and EDS techniques,respectively.Tandem reactions involved activation of CO and subsequent hydrogenation over the metal oxide catalyst,producing methanol and dimethyl ether as the main reaction intermediates,which then migrated onto SAPO-34 zeolite for light olefins synthesis.Effects of temperature,pressure and reactant gas flow rate on CO conversion and light olefins selectivity were investigated in detail.The Mn1Zr2/SAPO-34 catalyst(Mn/Zr ratio of 1:2)attained a CO conversion of 10.8%and light olefins selectivity of 60.7%,at an optimized temperature,pressure and GHSV of 380°C,3 MPa and 3000 h?1 respectively.These findings open avenues to exploit other metal oxides with CO activation capabilities for a more efficient syngas conversion and product selectivity.

1.Introduction

Light olefins comprising ethylene,propylene and butylene,are fundamental elementary units for a wide range of formulated products in the modern chemical and energy industries.Traditionally,they are produced from thermal or catalytic cracking of petroleum-derived hydrocarbons like naphtha[1,2].With recent reports of drastic depletion of petroleum reservoirs[1–3],urgent need to explore sustainable alternatives for low cost petrochemical feed stocks has risen.The efficient catalytic conversion of syngas(CO/H2),obtained from supplementary sources such as biomass,natural gas(including shale gas and biogas),coal and waste material,to light olefins,promises to be an economical non-petroleum route to abate global energy demand and environmental concerns[2,4].

The Fischer Tropsch to Olefins(FTO)synthesis route transforms syngas to light olefins over F-T catalysts(mainly Fe or Co-based)via a single process.It has been attractive over the recent years due to its flexibility towards carbon feedstock which can be directly converted into various liquid fuels[2,5,6].Generally,light olefins synthesis via FT route involves the dissociation of CO,formation and coupling of CHx(x=1–3)species to form CnHmintermediates that undergo hydrogenation or dehydrogenation to CnHmparaffin or olefin products[5,7].However,the coupling process is uncontrollable,resulting in a broad range of hydrocarbon products[5,7].Furthermore,the target range of products,including light olefins and paraffins is restricted to ≤58%,basing on limitations by the Anderson-Schulz-Flory(ASF)distribution[2,8].

Alternatively,conversion of syngas to light olefins can be achieved indirectly by a two-step process:syngas to methanol and/dimethyl ether(DME)synthesis followed by the subsequent conversion of these intermediate products to olefins(MTO/DMTO)[4,9,10].At present,the technology for producing light olefins via DMTO process has been commercialized in Baotou,China[11].It has been considered as a breakthrough and crucial step for producing light olefins in coal or natural gas rich countries.However,the two-step process is energy intensive,therefore less cost effective.The one-step conversion of syngas to light olefins requires two catalytic active sites working sequentially to catalyze two reactions:CO activation to form C1 intermediate products and C--C coupling of the intermediates to form light olefins.Substantial efforts have been made to merge the two reactions by the combination of the active sites into one bifunctional catalyst.Earlier attempts using methanol synthesis metal oxide catalysts such as Cu-ZnO combined with MTO catalysts resulted in aromatics and paraffins being the major products [9,12].To date,some progress has been made and bifunctional catalysts have demonstrated higher performance potential and cost-effectiveness due to their reaction coupling ability.Nonetheless,little information on this topic can be obtained from previous literature.

Recently,Jiao et al.[8],and Cheng et al.[13],described the successful one-step catalytic conversion of syngas into C2–C4olefins.The former group reported an oxide-zeolite concept,where a combination of metal oxide(ZnCrOx)catalyst and a mesoporous SAPO(MSAPO)zeolite with a hierarchical pore texture,attained a light olefins selectivity of 80%at 17%CO conversion via ketene intermediate.The latter group reported that a bifunctional catalyst composed of metal oxide(Zr-Zn)and zeolite(SAPO-34)catalyzed the C2–C4olefins selectivity to 70%at 10%CO conversion via the MTO process.Both communications reported a mixed oxide-zeolite process which promoted the successful occurrence of tandem reactions.The respective roles of the active components,metal oxides and zeolites,in CO dissociative activation and selective C--C coupling for light olefins synthesis were clearly revealed.However,it is preferable to use other environmentally friendly metals in place of chromium.Furthermore,the detailed mechanism of CO dissociation over the above-mentioned metal oxide catalysts has not yet been completely understood.Zhu et al.[14],used partially reducible MnOxin combination with SAPO-34 and achieved light olefins selectivity and CO conversion of 80%and 8.5%,respectively.The molecular sieve used in the aforementioned investigations is noted for its unique shape selectivity for light olefins synthesis by virtue of its chabasite(CHA)topology and small pore size [10,11,15].Light olefins selectivity reported in these stated cases surpassed the 58%limit imposed on the traditional FT catalysts.

Over the years,zirconia and manganese oxides have received considerable attention in the field of heterogeneous catalysis as supports[16,17],promoters[18]and most recently as the main active catalysts for CO hydrogenation to light olefins [13,14].Mn oxides are cheap,environment-friendly and exhibit superior redox activities and oxygen storage properties[19].Zirconia is noted for its high thermal stability and it also possesses both acid and basic sites[20,21].As individual catalysts,their properties pose a considerable limitation to their CO hydrogenation capabilities[10,22].However,the properties of one pure metal oxide can be enriched by interacting with foreign metal cations such that when combined,the catalytic performance of these metal oxides can be improved.

The work presented here reports for the first time,a bifunctional catalyst composed of reducible binary metal(Mn-Zr)oxides and SAPO-34 for CO hydrogenation to light olefins.The synergy of the two reactive components resulted in high product selectivity.Generally,olefins selectivity is negatively correlated with the surface acidity strength of the catalyst.The binary metal oxide had an acidity reduction effect on the zeolite hence suppressing its strong hydrogenation ability.This in turn suppressed secondary hydrogenation reactions responsible for converting light olefins to paraffins,ultimately providing perfect conditions for light olefin formation.The effects of varied reaction parameters(temperature,pressure and flow rate),on the catalytic performance of the unique Mn-Zr/SAPO-34 bifunctional catalyst have also been discussed.

2.Experimental

2.1.Materials

Commercial SAPO-34(SiO2/Al2O3=0.5,BET=570,Nankai University Catalyst Co.,Ltd),Mn(NO3)2·6H2O,Zr(NO3)4·5H2O and Na2CO3(Shanghai Chemical Co.,Ltd).The above-mentioned chemicals were used without further purification.All the gases were purchased from Hangzhou Metal Working Gas Company.

2.1.1.Preparing Mn-Zr binary oxide catalysts

The traditional co-precipitation procedure adopted from Lopez et al.[19]and Bulavchenko et al.[16]was used to prepare Mn-Zr catalysts with different molar ratios(1:0,1:0.25,1:0.5,1:1,1:2,1:4 and 0:1)as shown in Fig.1.An aqueous metal nitrates salt solution and Na2CO3,as precipitating agent,were added simultaneously to 200 ml deionized water,under vigorous stirring.Temperature and pH values were maintained at 65°C and 10.0,respectively.The slurry was then aged at the same temperature for 24 h.The solid residue was filtered and washed repeatedly with deionized water enough to remove residual sodium cations.The obtained filter cakes were dried at 100°C for 12 h and subsequently calcined at 600°C for 4 h at a heat rate of 5°C·min?1in a muffle furnace.

2.1.2.Preparing the bifunctional catalyst

The precursors from the calcination process;ZrO2,MnOx,Mn-Zr oxide powders were mixed with SAPO-34 at a mass ratio of 1:1.5 each,and physically ground in an agate mortar for 10 min.For convenience,the catalysts were represented as MnxZry/SAPO-34,where subscripts x and y represents the molar ratios of Mn and Zr in each binary oxide catalyst respectively.Finally,the bifunctional catalysts were pelletized to standard 380–840 μm to mitigate internal mass transfer limitations.For comparative purposes a separate catalyst sample was prepared where the metal oxide and zeolite powders were pelletized separately for granule-stacking.

2.2.Characterization

Powder XRD was performed by a RINT 2400 diffractometer using monochromatic CuKαradiation(λ=0.15418 nm)source operated at 40 kV and 20 mA in the scanning 2θ range from 5°to 80°with a step of 0.02°.N2physisorption was performed on an automatic gas adsorption system(QuantachromeAutosorb-iQ)and the specific surface area,pore volume and average pore diameter of the samples were obtained using the Brunauer Emmett Teller (BET) method.The samples were degassed under vacuum at 250 °C for 8 h prior to the measurement.Scanning electron microscopy(JEOL JSM-6360LV) equipped with an energy-diffusive X-ray spectroscopy(EDX)was used to observe the surface morphology and the elemental composition of samples respectively.X-ray photoelectron spectroscopy (XPS) tests were studied using a Thermo Scientific EscaLab 250 X-ray photoelectron spectrometer.The C 1s,Mn 2p,Zr 3d and O 1s spectra of prepared Mn1Zr2were recorded with an energy step size of 0.1 eV,using monochromatic Al Kα radiation.The binding energies were regulated with respect to the internal standard C 1s peak binding energy of adventitious carbon at 285 eV.Optimum deconvolution of Zr 3d and Mn 2p peaks was conducted using the Gaussian–Lorentzian function.Temperature programmed reduction(H2-TPR)and temperature programmed desorption(NH3-TPD)were conducted on a Belcat-B3 automatic chemical adsorption instrument(BEL Co.Japan).In each run,30 mg of zeolite or metal oxide was loaded at the center of a U-shaped quartz reactor and pretreated for 1 h at 200°C,followed by cooling to a temperature of 100°C using Ar flowing at 20 ml·min?1.For the bifunctional catalyst,50 mg (MnxZryto zeolite=1:1.5) was used.10% H2/Ar (30 ml·min?1) or 5% He/NH3flow was then passed over the sample and the temperature was linearly raised to 750°C(H2-TPR)and 650°C(NH3-TPD)with a heating rate of 10°C·min?1.The consumption of H2and NH3was detected by a thermal conductivity detector(TCD).

2.3.Catalytic activity measurements

Fig.1.Illustration of bifunctional catalyst MnxZry/SAPO-34 preparation method.

The catalytic hydrogenation of CO to light olefins tests were performed in the same continuous-flow fixed-bed stainless steel tubular reactor (6.8 mm internal diameter).A control experiment as blank was conducted prior to actual catalyst loading to investigate the effects,if any,of an empty reactor on experimental proceedings.The tubular reactor is designed with a low diameter-to-height ratio which radically promotes homogeneous distribution of heat to ensure isothermallity across its contents.1.0 g (380–840 μm) of composite catalyst was used for every test.This mass was mixed with a small amount of inert SiO2as diluent and placed at the center of the stainless steel reactor sandwiched by quartz wool for each test.All fresh catalysts were reduced at 500°C for 4 h in pure H2.After reduction,the reactor temperature was reduced to 300°C and then purged by feed gas(H2/CO=2,Ar=4%)before temperature and pressure were raised to a desired set point.Various measurements to observe the effects of temperature between 290 to 410°C and reactant pressure(1.0 to 3.5 MPa)were performed on catalyst with best metal oxide ratios.On the effect of reactant flow,a steady purge of syngas,carefully supplied at varied flow,30 to 50 ml·min?1was employed for the investigation on the best performing catalyst.Online gas chromatography equipment consisting of a thermal conductivity detector(TCD)equipped with a TDX-01 column and flame ionization detector(FID)equipped with an Agilent HP-Plot/Q column were used to quantitatively analyze the gaseous products.Internal normalization was employed to compute CO conversion.Hydrocarbon distribution for CO hydrogenation was evaluated on a molar carbon basis(without considering CO2).The catalytic performance was evaluated after 6 h for all reactions.Each experimental test was conducted at least three times to test and ascertain the reproducibility of the results.Life and stability of the best catalyst(Mn1Zr2/SAPO-34)under optimized parameters was investigated for 50 h of reaction(Fig.S4).

3.Results and Discussion

3.1.XRD analysis

The crystalline structures of the as-prepared binary metal oxides prior mixing with SAPO-34 zeolite were presented in Fig.2.According to the JCPDS card(PDF number 80-2155),the four prominent diffraction peaks(Fig.2(g))at 2? degree values of 30.3°,35.3°,50.4°and 60.2°correspond to(011),(110),(112)and(121)planes revealing the tetragonal polymorph of zirconium oxide.For the pure manganese sample(Fig.2 (a)),characteristic peaks of Mn2O3(JCPDS card-PDF number 41-1442) and hausmannite (Mn3O4,JCPDS card-PDF number 80-0382)were observed,the latter being more prevalent.On introducing zirconium oxide(Fig.2(b?g)),main characteristic peaks of hausmannite started to decay,indicating a phase transition into amorphous or microcrystalline state[17,19].In contrast,diffraction peaks corresponding to Mn2O3gradually become more sharp until the Mn:Zr=1,indicating improved crystallinity.The Mn2O3peaks shifted slightly towards higher angles with increase in ZrO2composition(Mn1Zr0to Mn1Zr1),which was perhaps related to the interaction between ZrO2and Mn oxides.At Mn:Zr=1 M ratio(Fig.2(d)),the diffraction peaks for ZrO2became more dominant.The peaks indexed to all manganese oxide species gradually disappeared,as conspicuously translated in the magnified extract(Fig.3).This behavior resulted in the progressive stabilization of the tetragonal phase of ZrO2(t-ZrO2).For activation and hydrogenation of CO bonds,tetragonal phase (t-ZrO2)exhibits exceptional activity[13,20].

The highly pronounced characteristic peaks of tetragonal zirconia(Fig.2(e?f)),suggest exceptional crystallization and incorporation of Mn into ZrO2forming a solid solution based on ZrO2structure.This observation is consistent with other reports where it was suggested that Mn2+ions(ionic radius ≤67 pm)would disperse into Zr4+(ionic radius=86 pm)lattice structure by direct substitution[16,19,23].In Fig.4,we observe typical diffraction lines of the parent SAPO-34 zeolite before and after physical mixing with the metal oxides.The zeolite main diffractions remained sturdy after mixing with all binary (or single)oxide catalysts ratios,indicating that the interaction did not compromise its sound morphology.However,the slight reduction in peak height for all bifunctional catalysts was a result of its lower mass proportion.Additionally,the characteristic peaks corresponding to Mn1Zr0,Mn0Zr1and Mn1Zr2were clearly visible in XRD pattern for the respective bifunctional catalyst.

3.2.H2-TPR measurements

Fig.2.XRD patterns of the samples after calcination at 673 K:(a)Mn1Zr0;(b)Mn1Zr0.25;(c)Mn1Zr0.5;(d)Mn1Zr1;(e)Mn1Zr2;(f)Mn1Zr4;(g)Mn0Zr1.

Fig.3.The enlarged patterns of peaks around 28°‐40°for(a)Mn1Zr0;(b)Mn1Zr0.25;(c)Mn1Zr0.5;(d)Mn1Zr1;(e)Mn1Zr2;(f)Mn1Zr4;(g)Mn0Zr1.

Temperature programmed reduction was performed to determine reduction behavior of the calcined catalysts.Fig.5 shows two well resolved peaks at 250 °C and 446 °C ascribed to reduction of Mn4+→Mn3+and Mn3+→Mn2+,respectively,as a result of the partial removal of the abundant surface oxygen species[24,25].Cogent evidence that the generation of these oxygen vacancies is vital for CO activation and that the unpaired electrons entrapped in the defective sites weaken the C--O bond thereby inducing C--O disproportionation reaction,has been gathered[14,26].Comparing TPR patterns(a)and(b),reduction temperature of the Mn oxides species was lower than that of zirconia corresponding to the behavior of their diffractions in XRD profiles.Detailed investigations have revealed that the reducibility of Mn oxides by H2is mainly determined by the concentration and crystallinity[17,24].H2consumption of ZrO2(curve b)in the 600–700°C temperature range can be attributed to the reduction of ZrO2surface hydroxyl groups as proposed by other researchers[16,27].Additionally,the position and intensity of ZrO2TPR profiles have been proven to depend on the phase composition and synthesis method used [27,28].For the more crystalline binary metal oxide(Mn1Zr2)(see XRD)there was a notable increase in the reduction temperatures as indicated by the shift from low to higher reduction temperatures(curve a–c).The TPR profile(c)showed significant H2consumption at 385°C,which was attributed to the easy reduction of the bulk composition of the manganese phase,mainly composed of Mn4+to Mn3+.In contrast,a low and broad peak was observed at higher temperatures,between 450 and 490°C,which was assigned to H2consumed for reduction of Mn3+and a small concentration of Mn incorporated in the Zr lattice[28].However,the ultimate incorporation of SAPO-34 zeolite resulted in broadening and backward shift of the reduction peaks to lower temperatures(Fig.5(d)).This phenomenon suggests that despite physical mixing,a robust interaction between the molecular sieve and metal oxide species established a relative ease for the reduction of composite catalyst.

3.3.XPS measurements

XPS analysis was conducted to examine the surface concentration and oxidation states of elements.Fig.S3 reports the XPS spectra of Zr 3d and Mn 2p of the prepared Mn1Zr2oxide catalyst.The Zr 3d5/2binding energy(BE)ranging from 182.2 to 182.5 eV corresponds to the Zr4+state of zirconium[16].However,the BE of the prepared catalyst was located at a lower 181.9 eV,indicating the intimate interaction between Zr and Mn species[25].Manganese cations incorporated into the lattice of ZrO2by direct substitution result in the establishment of solid solutions Zr1?xMnxO2,[19,23]in which lattice oxygen has adequately high mobility and consequently high reactivity[16].We propose that the partial reduction on catalyst surface would then generate oxygen vacancy sites which are vital to the activation of CO.Mn 2p XPS spectra,at the 640.0–650.0 eV and 650.0–660.0 eV BE regions were attributed to Mn 2p3/2and Mn 2p1/2respectively.The spectrum of Mn 2p3/2was deconvoluted into three peaks at 640.7–641.7,641.7–641.9,and 641.9–642.9 eV which indicated the co-presence of Mn2+,Mn3+and Mn4+species,respectively [16].These results are in fair agreement with the XRD and TPR analysis.The XPS spectrum of O 1s in Fig.S3 c has been fitted with two peak contributions,representing surface labile oxygen(Olab)and lattice oxygen(Olat)at 531.8 eV and 529.6 eV respectively[25].Table S1 presents the relative concentration ratios of Olab/O and Olat/(Mn+Zr).

Fig.4.XRD patterns of SAPO-34 and composite samples with mass ratio of oxide/zeolite equal to 1:1.5.

Fig.5.H2-TPR profiles of(a)Mn1Zr0;(b)Mn0Zr1;(c)Mn1Zr2;(d)Mn1Zr2/SAPO-34.

3.4.NH3-TPD measurements

Fig.6 displays four distinct NH3desorption curves for Mn1Zr2,SAPO-34,Mn1Zr2/SAPO-34 and for Mn1Zr2/SAPO-34g.For SAPO-34,two desorption peaks observed at low temperature(168°C)and medium temperature (372 °C),were ascribed to weak and strong acid sites[12]which are its Lewis and Br?nsted acid sites respectively[21].Basing on the position of the metal oxides(Mn1Zr2)peak(about 200°C),it could be deduced that the acidity was characteristically Lewis in nature [23].There was a slight decrease in the intensity of both the low and high temperature peaks for Mn1Zr2/SAPO-34 indicating a decrease in the density of acid sites.It has been deduced that the Br?nsted acidity of SAPO-34 is equivalent to the negative charge resulting from the exchange mechanism of H+ions or Si by Al or P[29,30].The grinding process employed during catalyst preparation may possibly have altered the coordination environments of H,Si,Al and P,thereby reducing its acidity.As can be seen,the high temperature peak shifted slightly backwards to a lower temperature,representing an obvious reduction in acid strength of the zeolite.These results indicate that the intimacy introduced by physically mixing metal oxide and SAPO-34 had significant effect of reducing the surface acidity of SAPO-34 zeolite.On the other hand,NH3-TPD signals for granule staked oxides and SAPO-34 did not show any alteration on the zeolite acid strength and density.

3.5.SEM-EDS measurements

As observed in Fig.7(a),the morphology of Mn1Zr2particles has a rosette like stature with an overall spherical orientation,emanating from a successful fusion of the metal oxides.The presence of a uniform particle size distribution,consistent particle shape and minimum particle agglomeration could be observed from the Mn1Zr2catalyst.Excess Zr normally exhibits block like crystal morphology(Fig.S1 b),therefore,the variation observed may be due to Mn structural promotion effects(Fig.S1 c).The ultra-small particles(2–4 nm)formation was probably due to the slow precipitation process and longer aging time during catalyst preparation(Fig.S1 a).The bifunctional catalyst(Mn1Zr2/SAPO-34)prepared by physical mixing retained its individual metal oxide and zeolite structures as illustrated in Fig.7(b).SAPO-34 particles exhibited a refined uniform cubic structure.Images obtained by the EDS elemental color mappings(Fig.S2 a,b)indicate that Mn particles were dispersed over the ZrO2on the overall Mn1Zr2oxide catalyst.This is in sync with XRD profiles already discussed.The EDS spectrum in Fig.7(c)presents peaks corresponding to manganese and zirconium.The results confirm the absence of impurities in the Mn1Zr2oxide catalyst.

3.6.Textural property measurements

As can be observed from Table S2,the BET surface area of the Mn1Zr2binary oxide was substantially larger than that of the individual Mn or Zr oxides.BET surface area of the bifunctional catalysts increased from 251 m2·g?1to 298 m2·g?1with an increase in Zr indicating a significant dependence on the Zr content.However,it has been reported that increasing Mn content in Mn-Zr catalysts has similar but reverse effects,meaning it results in smaller specific surface areas[16,17].The results indicate that variations in metal oxide composition contributed differently to the properties of the catalyst.Generally,the surface area of each bifunctional catalyst was smaller than that of SAPO-34 alone.Dang et al.[21],reported similar results suggesting the plugging of zeolite pores with oxide crystallites.The bifunctional catalyst Mn1Zr2/SAPO-34 had the highest specific surface area effectively providing more catalytic active sites and more contact between active components,as compared to others listed in Table S2.

As shown in Fig.8,N2adsorption and desorption for the bifunctional catalysts follows type-IV isotherms,while SAPO-34 alone follows type-I isotherms according to the IUPAC classification.The metal oxides,predominantly comprising mesopores,as shown by the hysteresis loops observed between 0.4–1.0 P/P0range,remarkably enhanced the diffusion and consequently suppressed mass transfer limitations of reactants and intermediates in consecutive reactions,improving catalytic performance.

3.7.Reaction performance for CO hydrogenation

The aim of this work is to investigate the effect of varying the molar ratios of Mn:Zr in the bifunctional catalysts and reaction conditions on their catalytic performance for the direct synthesis of light olefins from syngas.Mn-Zr catalysts played the critical role of synthesizing reaction intermediates(mainly methanol and DME),while the SAPO-34 molecular sieve was chosen for the shape selective C--C coupling to generate light olefins.Reported kinetic and thermodynamic conditions for light olefins direct synthesis over bifunctional catalysts have suggested feasibility at temperatures around 380–500°C mainly due to associated activation energy requirements for C--C coupling[13].Thus,in this work,optimum parameters for light olefin synthesis were investigated and discussed with detail.

Fig.6.NH3-TPD profiles of SAPO-34 and Mn1Zr2/SAPO-34 bifunctional catalyst,g represents granule staking.

Individual evaluation of each bifunctional catalyst component activity assisted in revealing roles played towards olefin synthesis.ZrO2and MnOxsurface oxygen vacancies are well-known to be beneficial for CO activation and generation of surface methoxide via formate and/or ketene formation in the presence of H2[14,31].Reported IR studies revealed that ZrO2could also be used as a hydrogenation catalyst because of its unique capacity to adsorb H2by heterolytic or homolytic splitting[32].Because of the reaction intermediates and product distribution observed in our work,it is obvious that the detailed reaction mechanism could be similar to previous reports[13,33].A simplified reaction pathway has been presented(Fig.9).

3.7.1.Effect of different molar ratios of Mn-Zr/SAPO-34 on reaction performance

Fig.10 shows the syngas conversion and reaction products of the different MnxZry/SAPO-34 bifunctional catalysts.Our results suggest that different molar compositions had different effects on product selectivity.Although Mn0Zr1/SAPO-34(ZrO2/SAPO-34)and Mn1Zr0/SAPO-34(MnOx/SAPO-34) produced light olefins as clearly shown in Fig.10,the CO conversion was very low (2.1% and 6.9%,respectively,Table S3).Mn1Zr2acatalyst(excluding SAPO-34)mainly converted syngas to CH4(16.2%) and CO2(36.5%),shown in Table S4.Methanol,light olefins and paraffins selectivity was only 31.1%,20.7%and 8.3%of the total hydrocarbons produced,respectively(Table S3).These results indicated that Mn-Zr oxide surfaces possessed the ability to catalyze CO cleavage,yet exhibiting limited ability to hydrogenate olefins to paraffins probably due to the presence of Mn.The presence of MnOxhas an electronic effect that prohibits the secondary hydrogenation of light olefins[34,35].Zhu et al.,have already proven that MnOxspecies rely on their strong capacity for CO activation rather than H2dissociation in olefin synthesis[14].Upon SAPO-34 incorporation,CH4selectivity declined to only 3.5%,CO conversion and olefins selectivity increased significantly with no trace of residual methanol(Table S3).This is an indication that the transformation of intermediates(CH3OH/DME)to C2–C4olefins can thermodynamically drive syngas conversion.SAPO-34 revealed excellent catalytic performance with ca.100%CH3OH conversion followed by C--C coupling on its Br?nsted acid sites.The microporous structure of SAPO-34 significantly contributed to the inhibition of C5+hydrocarbons synthesis.Consequently,the C5+hydrocarbons selectivity was very low (<8.5%) for composite MnxZry/SAPO-34 catalysts resulting in C2–C4olefins being the chief products in the current study.CO conversion for the as-prepared bifunctional catalyst was higher than that of individual metal oxides combined with SAPO-34,as already noted in previous reports[14].We attributed this to the synergistic effect introduced by the interaction between the two metal oxides.

Fig.7.(a)SEM image of Mn1Zr2 catalyst(b)SEM image of Mn1Zr2/SAPO-34(c)EDS spectrum of the Mn1Zr2 catalyst.

Fig.8.N2-adsorption–desorption isothermal curves of SAPO-34 and bifunctional catalysts.

CO2selectivity was between 38 and 50%over the bifunctional catalysts and a linear increase with CO conversion,possibly because of the water-gas shift(WGS)reaction(CO+H2O →CO2+H2)was observed[33].A similar trend was observed for CH4production on increasing Zr content,which could have been a result of the direct hydrogenation of methanol over the Mn-Zr active sites(Table S4).The highest C2–C4olefins selectivity of 59.6%,constituting mainly propylene,was obtained over Mn1Zr2/SAPO-34 at a CO conversion of 10.6%.This conversion increased significantly because of the accessibility of more active sites for CO activation over the Mn1Zr2oxide catalyst (see BET results),followed by the fast diffusion and consumption of reaction intermediates to light olefins by SAPO-34.

Previous studies have suggested that the intimacy between the two active components and density of Br?nsted acid sites on SAPO-34 substantially influence product distribution[13,33].To confirm this idea,we also prepared and tested Mn1Zr2oxide and SAPO-34 packed in granule stacking arrangement.When the two active components are in a close physical proximity,there is reason to believe that a thermodynamic driving force arises from the swift removal of methanol/DME in forming C2–C4olefins on zeolite site,which facilitates CO conversion on metal site,hence increased product selectivity[30,33].Lower CO conversion(7.9%)and C2–C4olefin selectivity(35.4%)characterized the mixture of Mn1Zr2oxide and SAPO-34 granules,probably because there was limited thermodynamic driving force owing to the lower proximity assumed between the metal and zeolite granules in this arrangement (Table S3).CH4(31.1%) selectivity was high most probably because of repeated interaction of intermediates with each other instead of a rapid transfer to SAPO-34 which was not in close contact,and similarly C2–C4paraffins(30.8%)selectivity was high due to some secondary hydrogenation reactions of the C2–C4olefins on SAPO-34,as expected.After increasing the proximity of the two components by physical grinding to powder form,a substantial increase in both CO conversion and C2–C4olefin selectivity to 10.6%and 59.6%respectively,was observed.From these results we were able to determine that a closer proximity between the two components results in higher CO conversion and higher C2–C4olefin selectivity.

Furthermore,by physically mixing the metal oxides with the zeolite catalyst,the interaction of active sites was enhanced,at the same time reducing the density of Br?nsted acid sites(see NH3-TPD)for improved product distribution.The reduced acidity of zeolite catalyst can effectively decelerate rapid coking(Fig.S5) and deactivation resulting in the prolonged lifetime(Fig.S4).The density of Br?nsted acid sites was an essential factor in determining light olefins selectivity from intermediates.

3.7.2.Effect of reaction temperature on products distribution

In order to optimize the reaction conditions for the catalytic conversion of syngas(H2/CO=2)to target products over Mn1Zr2/SAPO-34 bifunctional catalyst,the most appropriate reaction temperature was investigated in the range 290–410°C.Conversion of CO,as shown in Fig.11,increased directly(6.9%–11.3%)with the increase in reaction temperature.Adversely,a similar linear correlation between temperature increase and CH4selectivity was observed by Raveendra et al.[18].This could be assigned to DME decomposition and product cracking as previously investigated[36].

Generally,the product spectrum mainly comprised ethylene and propylene,which accounted for more than 60%of total hydrocarbons produced at various temperatures.Selectivity of dimethyl ether and methanol decreased with increasing temperature confirming that zeolite catalyzed C--C coupling requires high temperatures[12,33].C2–C4olefins yield was maximum at 380°C at CO conversion of 10.6%with C2–C4olefins/C2–C4paraffins ratio of 1.9.Intermediate product synthesis and subsequent conversion over SAPO-34 zeolite was also very high at this temperature range.

Fig.9.Proposed reaction pathway for syngas conversion over MnxZry/SAPO-34 bifunctional catalyst.

Fig.10.Catalytic performance of bifunctional catalysts for CO hydrogenation to light olefins(CO2 free);Reaction conditions:T=380°C,P=2.5 MPa,mass(MnxZry/SAPO-34)=1 g,reaction gas(H2/CO=2),GHSV=3000 h?1,reaction time=6 h.=for olefins,o for paraffins,C5+for long chain hydrocarbons.

Further increase in reaction temperature to 410°C resulted in a steep decline in product selectivity.The tendency of coke accumulation,which causes channel blockages of zeolite at this relatively high temperature,may have influenced product selectivity.C2–C4paraffins selectivity increased with increasing temperature most likely due to increased hydrogenation of light olefins over catalyst surface.The selectivity of C5+hydrocarbons varied with reaction temperature although it was generally very low.This could have been a function of diffusion limitations of SAPO-34 small 8 membered ring pore openings which controlled carbon chain length.Therefore,in this work,the optimized temperature for C2–C4olefins synthesis over Mn1Zr2/SAPO-34 bifunctional catalyst was 380°C.

3.7.3.Effect of reaction pressure on products distribution

Fig.12 shows effect of syngas pressure on catalytic behaviors of Mn1Zr2/SAPO-34 catalyst.As the CO conversion increased,product distribution varied consistently with increase in pressure.At a reaction pressure of 1 MPa,the conversion and selectivity of C2–C4olefins were 6.3% and 53.6%,respectively.Both the selectivity and conversion increased at 2 MPa.It has been reported that increase in reaction pressure leads to elevated activation of surface carbon species due to improved encounter probability of reactants and catalyst active surface so that the reaction rate is increased consequently increasing product selectivity[30,37].The optimized reaction pressure was 3 MPa,where the direct conversion of syngas was 10.8%at C2–C4selectivity of 60.7%.

However,pressure above 3.0 MPa shifted the product selectivity towards C2–C4paraffins while C2–C4olefins selectivity decreased to 52.3%.These results suggest that the hydrogenation ability of H2species over the Br?nsted acid sites of SAPO-34 might have been enhanced with the increase in pressure which is in agreement with recent reports[22].According to Liu et al.[6],increase in reaction pressure results in the accumulation of tail gas which causes the readsorption and secondary reactions of ethylene.These secondary reactions decrease the overall yield of light olefins and increase paraffins in the product gas.However,reaction pressure had lesser effect on C4olefins and C5+hydrocarbons selectivity due to minimum secondary reactions thus the results showed slight changes.CO2selectivity also increased with increase in pressure which could be a result of increase in water-gas shift reaction.

3.7.4.Effect of reactant gas flow on products distribution

Fig.11.Effect of reaction temperature on products distribution(CO2 free);Reaction conditions:P=2.5 MPa,mass(Mn1Zr2/SAPO-34)=1 g,reaction gas(H2/CO=2),GHSV=3000 h?1,reaction time=6 h.=for olefins,o for paraffins,C5+for long chain hydrocarbons,C1int for intermediates including methanol and dimethyl ether.

Fig.12.Effect of reaction pressure on products distribution(CO2 free);Reaction conditions:T=380°C,mass(Mn1Zr2/SAPO-34=1 g,reaction gas(H2/CO=2),GHSV=3000 h?1,reaction time=6 h.=for olefins,o for paraffins,C5+for long chain hydrocarbons.

In Fig.13,it is evident that light olefins selectivity increased significantly with the increase in gas hourly space velocity(GHSV).At GHSV of 2250 h?1,the results were characterized by high CO conversion(up to 12.3%)and low light olefins selectivity.This could have been a result of increased residence time in the reactor which encouraged more contact of reactants with the catalyst,consequently manifesting secondary reactions which favored paraffins and long chain hydrocarbons selectivity.The highest light olefins selectivity of 60.7%(CO conversion=10.8%)was obtained at GHSV of 3000 h?1,at a steady reactant flow of 40 ml·min?1.Thus,at the aforementioned GHSV,there was adequate supply of reactants,swift regeneration of active sites by removal of product from catalyst surface hence alleviation of mass transfer limitations.An increase in GHSV from 3000 to 3375 h?1was characterized by a slight increase in product selectivity from 60.7%to 63.4%,possibly because the formed methanol could quickly be converted to light olefins.At the highest GHSV of 3750 h?1,a steep decrease in CO conversion was observed.This could be attributed to reduced reactant-catalyst contact time as a result of vigorous flow.The results clearly revealed that space velocity has significant influence on the reaction efficacy.

4.Conclusions

Our studies strongly suggest that the C2–C4olefins were synthesized from syngas over Mn1Zr2/SAPO-34 bifunctional catalyst.The binary metal oxide catalysts had a relevantly higher activity,with respect to that of the pure oxides(MnOxand ZrO2)which was found to be related to the strong synergetic effect between Mn and Zr sites.The bifunctional catalyst yielded remarkable product selectivity,achievable through CO and H2activation over Mn1Zr2oxide surface and C--C bond formation on the Br?nsted acid sites within the constricted cages of SAPO-34.The intimacy of the two active components influenced the density of Br?nsted acid sites and also played a crucial role in the product distribution.Reaction intermediates mainly methanol and dimethyl ether were imperative for the generation of light olefins.Catalytic activity was significantly influenced by reaction temperature and pressure which were varied to determine the best reaction conditions.There was linear correlation between temperature increase and product selectivity.However,our studies also showed a steady increase in CH4and CO2with increasing reaction temperature and pressure.The highest yield of light olefins was achieved under the reaction conditions of Mn/Zr ratio 1:2,380°C,3.0 MPa and 3000 h?1.CO conversion and product selectivity attained 10.8%and 60.7%respectively,breaking the ASF limit.Therefore,optimum reaction conditions give optimal catalytic performance of the bifunctional catalyst for optimal synthesis of light olefins in a single step process which can assist to alleviate growing product demand.Nonetheless,further investigations to improve the CO conversion and product selectivity,while reducing methane and CO2production are still needed.

Declaration of Competing Interest

There are no conflicts to declare.

Fig.13.Effect of reactant gas flow on products distribution(CO2 free);Reaction conditions:T=380°C,P=3 MPa,mass(Mn1Zr2/SAPO-34)=1 g,reaction gas(H2/CO=2),reaction time=6 h.=for olefins,o for paraffins,C5+for long chain hydrocarbons.

Acknowledgements

This research is financially supported by Youth Foundation of ZUST,China (2019QN23)and Foundation of State Key Laboratory of Highefficiency Utilization of Coal and Green Chemical Engineering(2019-KF-21).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2020.09.004.