Coupling effect of bifunctional ZnCe@SBA-15 catalyst in 1,3-butadiene production from bioethanol

Zheng Wang, Sijia Li, Shengping Wang, Jiaxu Liu, Yujun Zhao,*, Xinbin Ma

1 Key Laboratory for Green Chemical Technology of Ministry of Education, School of Chemical Engineering and Technology, Collaborative Innovation Center of Chemical Science and Engineering, Tianjin University, Tianjin 300072, China

2 State Key Laboratory of Fine Chemicals, Department of Catalytic Chemistry and Engineering, Dalian University of Technology, Dalian 116024, China

Keywords:Biomass Catalyst Coupling effect Lewis acid 1,3-butadiene Optimization

ABSTRACT A series of bifunctional ZnCe@SBA-15 catalysts with different Zn/Ce ratios were prepared by a solid-state grinding strategy and used in the conversion of ethanol to 1,3-butadiene (ETB).For the supported metal oxides, ZnO serves as the active sites for the dehydrogenation of ethanol, and CeO2 promotes the aldolcondensation reaction.Based on the results of Py-FTIR and NH3-TPD, it suggests that the yield of 1,3-butadiene is positively correlated with the number of weak Lewis acid sites on the catalyst surface,given their benefit for aldol-condensation reactions.The catalyst with an optimal Zn/Ce ratio of about 1:5 has the highest concentration of weak Lewis acid.Coupling with the ZnO sites,it contributes to a 98.4%conversion of ethanol and a 45.2% selective of 1,3-butadiene under relatively mild reaction conditions(375°C, 101.325 kPa, and 0.54 h-1).

1.Introduction

Nowadays, climate change and ecosystem issues become increasingly severe, urging the exploration of alternative green chemical technologies.The use of biomass resources to replace the non-renewable fossil fuels such as coal and petroleum shows significant prospects in chemical production [1,2].The technologies required to produce ethanol from biomass resources is relatively mature [3-7].The large production capacity of bioethanol demands further upgrading to high value-added chemicals such as 1,3-butadiene, acetaldehyde, ethylene,and ethyl acetate, which has attracted widespread attention from various industries[8-11].As an essential raw feedstock, 1,3-butadiene (1,3-BD) is widely used by manufacturers of synthetic rubber, nylon, and resin.The primary source of 1,3-BD is the by-products in ethylene production.In recent years, affected by the continuous lightening of raw materials for ethylene production and the rapid growth of domestic synthetic rubber production capacity,the supply of 1,3-BD is in severe shortage [12,13].It is therefore important to develop new 1,3-BD production from bioethanol instead of the traditional cracking route.

The conversion reaction of ethanol to 1,3-BD (ETB) has two main production routes: the one-step method by Lebedev and the two-step one by Ostromislensky [14].The major difference between the two processes lies on the fact that the former only feeds with ethanol, while the latter partially converts ethanol to acetaldehyde first, which then reacts with extra ethanol to 1,3-BD.The widely accepted reaction mechanism of this ETB system is described in Fig.1 [15].The basic reactions include five steps:(a) ethanol is dehydrogenated to acetaldehyde; (b) acetaldehyde undergoes aldol-condensation to obtain acetaldol; (c) acetaldol is dehydrated to crotonaldehyde; (d) Meerwein-Ponndorf-Verley(MPV) reaction occurs between ethanol and crotonaldehyde to form crotyl alcohol; (e) crotyl alcohol is dehydrated to 1,3-BD.The major by-products of this two step process includ ethylene,ether, and acetaldehyde, which come from the dehydration and disproportionation of ethanol and some side reactions involving the intermediates [16].For such a multi-step reaction system, the catalyst should have multiple functions with the active sites for ethanol dehydrogenation and aldol-condensation (basic or acidic sites) essential.

Fig.1. Mechanism of the ETB reaction.

In the study of one-step ETB reaction, silica-supported metal oxide catalysts are extensively used.Several support materials (e.g., silica gels, fumed, or mesoporous silica) with large surface area and suitable pore size have been applied in this reaction system[17,18].For the active species, Ag, Au, CuO, and ZnO are active for ethanol dehydrogenation.Magnesium, zirconium, tantalum,and cerium oxide are used as the active sites for aldolcondensation due to their appropriate acidity[19-24].These metal oxides serve as dehydrogenation sites to decide the total acid content, acid species [25,26], and acid-base properties of the catalyst[27].There is also synergetic effects between the two active sites so that it is vital to balance their content ratio.The selectivity and yield of 1,3-BD relies on several factors, and the proportion of acid sites[28]and the content of Lewis acid[29]are considered as the two main aspects.Doping alkali metal, such as Na [30]and Cs [31], and In2O3[32]can adjust the proportion of acid and base sites and promote dehydrogenation of ethanol as well.Besides silica materials,microporous zeolites are also considered as the support of metal oxides active phase.Kyriienkoet al.[33]found that ordered microporous Beta zeolite support can increase the content of Lewis acid sites and reduce the amount of Br?nsted acid sites,resulting in higher ethanol conversion and 1,3-BD selectivity.Qiet al.[34]constructed a series of dealuminated Beta zeolite catalysts grafted with isolated sites of Zn and Y,respectively,to clearly explain the specific functions of the two types of active sites in ETB reactions.The center of the Lewis acid site and the adjacent silanol groups can efficiently catalyze the formation of C—C bonds by promoting the direct conversion of co-adsorbed ethanol and acetaldehyde.The two-step ETB method exhibited superior reaction performance on Y-DeAlBEA catalysts with 1,3-BD selectivity of 66.4%.Hence, wisely choosing two appropriate types of functional sites and effectively tailoring their ratios is crucial to receive a highly efficient catalyst for ETB reactions.

SBA-15 has a large specific surface area,uniform mesopore distribution,tunable pore size,and high hydrothermal stability.When compared with microporous zeolites,it also has excellent diffusion performance.Therefore, when used in ETB reaction, the mass transfer inside SBA-15 supported catalyst can be significantly improved, so does the suppression of carbon deposition [35].In our previous work [36], a novel ZnCe@SBA-15 catalyst was prepared by solid-state grinding strategy using SBA-15 with organic templates.On the one hand, the confinement effect of the narrow space between the silica wall and templates can inhibit the agglomeration of CeO2.On the other hand,the template and metal have strong interactions, which can effectively disperse the metal oxide species.It is worthy to mention that adjusting the relative ratios of metal oxide sites will promote the effective coupling of dehydrogenation reaction and aldol-condensation reaction, which serve as one of the key steps in improvement of the catalyst performance.

In this work, we prepared a series of highly dispersed ZnCe@SBA-15 catalysts with distinct Zn/Ce mass ratios and used for ETB reactions.Based on the catalyst characterization and catalytic activity results, the correlation between the catalyst performance and the absolute quantity of Lewis acid was analyzed.It proved that weak Lewis acid is beneficial in the production of 1,3-BD.The effect of other reaction conditions such as temperature and residence time were also studied, together with the catalyst stability in ETB reaction.

2.Experimental

2.1.Catalyst preparation

The ordered mesoporous silica material was prepared by the hydrothermal method [37].About 4 g of P123 (organic template)was completely dissolved in deionized water (30 ml) and a hydrochloric acid solution(120 g,2 mol·L-1).The mixture was stirred in a flask fixed in a water bath at 35°C.After TEOS (silicon source, 8.5 g) was dropwise added, the solution was continuously stirred for 24 h.The received solution was then transferred to a crystallization kettle and hydrothermally treated (100°C, 24 h).After that,the crystallization kettle was cooled to ambient temperature, and the solid was filtrated and washed to neutral pH value.Finally,it was dried overnight(120°C)to obtain SBA-15 raw powder without removing the organic template.

The metal oxide was loaded on SBA-15 by using a solid-state grinding strategy [38].Appropriate amounts of Zn(NO3)2·6H2O(Shanghai Aladdin,99%purity),Ce(NO3)3·6H2O(Tianjin Jieerzheng,99.5% purity), and SBA-15 raw powder were placed in an agate mortar and grinded at ambient temperature for 12 h.The collected solid powder was then transferred to a tube furnace and calcined in air flow(500°C,5 h).The obtained ZnCe@SBA-15 catalysts were denoted as 10ZnxCey-AS(x:y=1:8,1:5,1:3,1:1,3:1),where 10 represents that the weight percentage of total Ce and Zn loading.The subscripts ofxandyin ZnxCeyindicates that the mass ratio of Zn and Ce respectively, and AS represents a catalyst prepared using SBA-15 raw powder containing organic templates.

2.2.Characterization

N2physical adsorption: The specific pore structure and surface area of the catalysts were measured using a Micromeritics Tristar II 3000 physical adsorption instrument at -196 °C.Before measurement, catalyst samples were pretreated in a nitrogen atmosphere(300°C, 3 h).

X-ray diffraction (XRD): The XRD powder diffraction results of catalyst samples were obtained by an X-ray diffractometer (C/Max-2500, Rigaku, Japan).The ray source was Cu Kα(λ=0.15406 nm), and working conditions were as follows:V=40 kV,I=200 mA, scanning range=10°-90° and scanning speed=8 (°)·min-1.

Ultraviolet-visible diffuse reflection spectrum(UV-Vis):UV-Vis DRS was measured by a Japanese Shimadzu UV-2550 type spectrophotometer (wavelength 200-800 nm).BaSO4was used as a reference in the experiment.

Transmission electron microscope (TEM): A field emission transmission electron microscope (JEM-2100F) was used to observe the catalyst morphology.Simultaneously, with an integrated energy dispersive spectrometer(EDS)module,the elements in catalyst samples were also measured qualitatively for their surface distribution.The following operation conditions were used:accelerating voltage=200 kV;point resolution ≤0.19 nm;line resolution ≤0.14 nm; tilt angle=25°; beam spot size ≤0.5 nm.The sample was first ground and a small amount of catalyst powder was suspended in ethanol and ultrasonicated for 30 min.The dispersed liquid was dropped on a copper microgrid prior to analysis.

Pyridine adsorption Fourier transform infrared spectrum (Py-FTIR): Py-FTIR was applied to identify the type of acid (Lewis and Br?nsted) on the materials.Py-FTIR spectra were collected from a Nicolet 6700 FT-IR spectrometer (4 cm-1optical resolution, one level of zero-filling).The sample was pressed into a selfsupporting wafer(13 mm in diameter)firstly.The sample was then pretreated in a cuvette under vacuum(10-3Pa)in situ(400°C,4 h).Pyridine was loaded into the cuvette with a flow of 30 ml·min-1(50°C, 30 min), followed by vacuum treatment (150°C, 1 h) for the wafer prior to collecting the FT-IR spectra data.

NH3programmed temperature desorption(NH3-TPD):A Micrometrics Auto-chem II 2910 chemisorption instrument was used to perform the programmed temperature desorption experiment.A catalyst sample was packed into a quartz tube and pretreated under He atmosphere(500°C,1 h),then cooled to 100°C and treated with NH3for 30 min.The sample was then purged in He until the TCD signal of the tail gas became constant.The catalysts were then heated to 500°C (10°C·min-1) and TCD signal was collected for the desorbed NH3.

2.3.Evaluation of the catalysts

The reaction performance of the prepared catalysts was evaluated using a fixed bed micro catalytic reaction system (WFSD-2017, Tianjin Xianquan Instrument Co., Ltd.).For each run, about 0.10 g of catalyst (size of about 0.25-0.4 mm) was placed into a quartz tube (inside diameter 6 mm, length 24 cm).The catalyst bed was heated in N2atmosphere at 400°C (heating rate of 10°C·min-1, 30 min), and then adjusted to reaction temperature(325, 350, 375, and 400°C, respectively).Nitrogen gas was then flown through a evaporation tank to carry saturated ethanol vapor into the reaction tube.Different feed rates (0.54, 1.08, 1.62, and 2.16 g ethanol·(g cat)-1·h-1) was made by changing the temperature of the saturated evaporation tank.The products were detected on line by an SP-7890A GC equipped with an HP-PLOT Q column and an FID detector.

Conversion of ethanol, selectivity, and yield of 1,3-butadiene was calculated based on the following equations (1), (2), (3),respectively:

3.Results and Discussion

3.1.Texture properties of catalysts

All catalysts exhibit Langmuir Type IV isotherms with H2hysteresis loops(Fig.2(a)),revealing that all those samples have a typical hexagonal mesoporous structure.When the relative pressure(P/P0) is within a specific range of 0.44-0.8, the adsorption amounts of these samples increased rapidly with the elevation ofP/P0,proving that all catalysts have a relatively consistent pore distribution and a similar pore size.It can also be verified by the results in Fig.2(b).The pore size is mainly distributed within the range of 3-6 nm for all catalysts.Meanwhile, the specific surface area of these samples(Table 1)is very close when the ZnO content is low enough (Zn/Ce ratio below 1:5).When further increase the amount of ZnO by raising the Zn/Ce ratio to 1:3 or higher, the specific surface area of the catalyst gradually decreases.This suggests that CeO2can be dispersed much easier than ZnO in the given conditions.Yinet al.[39]have adopted the solid-state grinding method to prepare similar catalyst (12% (mass), CeO2-ASS).Compared with SBA-15, the surface area of their catalysts is only slightly reduced.A similar catalyst(10%,Ce/SBA-15)was also synthesized by Puet al.[40]with an impregnation method.This catalyst has the similar surface area and a lower pore volume in comparison with SBA-15.As we found before[36],compared with SBA-15, the hysteresis loop of 10Zn1Ce5-AS shifts to a lowerP/P0value, indicating that a smaller pore size of the support after the metal oxide loading [41].This can be inferred that ZnO and CeO2species are dispersed in the mesopores to some extent.Moreover,CeO2can contribute to more specific surface area than ZnO[42,43].Therefore, 10Zn1Ce5-AS and 10Zn1Ce8-AS have a similar specific surface area as SBA-15, but a lower pore volume due to a higher loading of CeO2in catalyst.

Table 1The textural structure of 10ZnxCey-AS catalysts and SBA-15

3.2.Metal oxidecrystals and their dispersion in catalyst

The crystal structure of the series of 10ZnxCey-AS catalysts and the dispersed state of supported oxides were analyzed by X-ray diffraction(XRD)spectrum characterization in a wide-angle range.In Fig.3, all catalyst samples show distinct broad peaks (2θ=23°)corresponding to amorphous silica, and none of the typical ZnO characteristic peaks appears.This suggests that the ZnO species is highly dispersed on catalyst, whose crystal size is too small to effectively diffract X-ray photons with sufficient signal/noise ratio beyond the detection limit of the used XRD instrument.Another possibility lies on the potential amorphous states of ZnO in catalyst.Notably, at a low content of cerium (Zn/Ce is higher than 1:5), no characteristic peaks or only very weak peaks of CeO2are found on those catalyst samples.It implies that the CeO2is also well dispersed on SBA-15 at low loading contents.However, further increasing the cerium species, the XRD spextrum of catalyst 10Zn1Ce8-AS clearly shows multiple characteristic peaks of CeO2.It illustrates that excess ceria species will cause their aggregation and worsen their well-dispersed states.For the series of 10ZnxCey-AS catalysts, the dispersion of ceria is more sensitive to the mass ratio of Zn/Ce than zinc oxide, which can be used to effectively adjust the dispersion of ceria on catalyst.

3.3.Electronic states and transition of metal oxides on catalyst

Fig.2. (a) The N2 adsorption-desorption isotherms and (b) pore size distribution of 10ZnxCey-AS catalysts.

Fig.3. XRD patterns of 10ZnxCey-AS catalysts.

Fig.4. The DR UV-Vis spectra of 10ZnxCey-AS catalysts.

To clarify the oxidation state and electronic transition behavior of Ce in the catalyst samples,the 10ZnxCey-AS catalysts were char-Note: Calculated by ①BET method.②BJH method.acterized by diffuse reflection UV-visible (DR UV-Vis) spectrum.The bands (Fig.4) at 210 nm and 260 nm are attributed to the charge transfer from O2-to Ce3+.At 294 nm, the band represents the metal charge transfer of the CeO2clusters (O2-to Ce4+).When decreasing the ratio of Zn/Ce from 3:1 to 1:3, the strength of the O2-to Ce3+charge transfer band is more apparent, while the O2-to Ce4+charge transfer band that represents the CeO2clusters is very weak.This indicates that the Ce on the catalyst surface is well dispersed, and no apparent CeO2clusters are formed.However,when the ratio of Zn/Ce becomes 1:8, the charge transfer band from O2-to Ce4+is significantly enhanced, indicating that the charge transfer band moves toward the bulk CeO2cluster.This means the CeO2species in the 10Zn1Ce8-AS catalyst agglomerate to certain degrees already.For ZnO species, the bands at 265 and 365 nm correspond to nanoclusters and large zinc oxide particles,respectively.The nano-size ZnO bands are apparent in all catalyst samples, and the bands from large-size ZnO are almost negligible.With all these obversations, we infer that ZnO species is well dispersed, which are also in consistent with the XRD results.

3.4.SBA-15 structure and element distribution on catalyst

Fig.5. (a) EDS mapping, elemental distribution of (b) Ce, (c) Zn, and (d) TEM image of the 10Zn1Ce5-AS catalyst.

The element distribution and morphology of the catalyst can be observed from the TEM images.Through EDS mapping (Fig.5(a),(b), (c)), it can be found that both ZnO and CeO2species are uniformly distributed on the catalyst.The catalyst has ordered mesopores of SBA-15, and no large aggregates of oxide additives are found (Fig.5 (d)).It illustrates the successful synthesis of SBA-15 and proves the excellent dispersion state of active sites.

3.5.Acid properties of catalysts

Py-FTIR characterization method was applied to distinguish Lewis and Br?nsted acid on those catalysts.As shown in Fig.6(a),all 10ZnxCey-AS catalysts have only Lewis acid sites (1450 cm-1),but the band for the Br?nsted acid site(1540 cm-1)is hardly found.To further understand the strength and amount of the acid sites on those catalysts, NH3-TPD tests were performed.CeO2is generally believed to provide Lewis acid sites [44,45].In Fig.6(b), it can be found that all catalysts have two desorption peaks,which are considered as the weak acid sites (near 180°C) and the medium-tostrong acid sites (around 290°C), respectively.Nevertheless, the integrated peak areas of the two acid sites show totally different trends.As the Zn/Ce ratio changes from 3:1 to 1:8, the peak area of medium-to-strong acid sites decreases.But for weak acid sites,the area of the diffraction peaks increases first and then falls.The number of weak acid sites reaches the largest for catalyst with a Zn/Ce ratio of 1:5.It indicates that the ceria on the surface of catalyst will agglomerate when increasing its loading as proved by the XRD results.Such agglomeration of CeO2becomes especially severe for the catalyst with a Zn/Ce ratio of 1:8,with a dramatic reduction in both useful weak acid sites and the total acidity.When the Zn/Ce ratio on 10ZnxCey-AS catalysts is 1:5, some medium-tostrong acidic sites are inhibited, and more weak acidic sites are formed.This could appropriately balance the dehydrogenation active sites and aldol-condensation active sites to maximize the conversion of ethanol and the selectivity of 1,3-BD.

Fig.6. (a) The Py-FTIR spectra and (b) the NH3-TPD profiles for 10ZnxCey-AS catalysts.

3.6.Balance of acid sites and ZnO towards yield of 1,3-BD via ETB process

The catalytic performances(Fig.7)of the 10ZnxCey-AS catalysts with different Zn/Ce ratios were evaluated in the ETB reaction system at mild reaction conditions.With the decreasing Zn/Ce ratio,the ethanol conversion rate declines from 90.8% to 50.2% with a drastical conversion drop occurred on catalyst with a Zn/Ce ratio of 1:8.It is believed that the dehydrogenation ability of the 10Znx-Cey-AS catalysts gradually fades away as the amount of ZnO species decreases, which in turn low the ethanol conversion.When the Zn/Ce ratio changing from 3:1 to 1:5, the selectivity of 1,3-BD gradually increases and the highest level (45.3%) reaches on catalyst 10Zn1Ce5-AS.As shown in Fig.7, the major by-products of the ETB reaction are ethylene, acetaldehyde, and ether.The selectivity of ethylene and ether remains at around 20% and 1.5%,respectively, while the selectivity of acetaldehyde decreases from 52.0% to 21.6% significantly with the decreasing Zn/Ce ratio.The weak acid sites are considered the favorable active sites for aldol-condensation [46], the most extensive content of weak acid on the catalyst with Zn/Ce of 1:5, therefore, contributes to the superior selectivity to 1,3-BD.On the other hand, higher content of cerium benefits the formation of the weak acid sites, which effectively promotes the conversion of acetaldehyde to 1,3-BD.Thereby, it can further enhance the aldol-condensation reaction eventually.However,when the Zn/Ce ratio reaches 1:8,agglomeration of the cerium species becomes severe and the exposed acidic sites, especially those weak acid sites, decreases.As the consequence,the aldol-condensation reaction is adversely affected,leading to a lower selectivity of 1,3-BD.Among these catalyst samples,the one with an appropriate Zn/Ce ratio(1:5)has a balanced metal oxide species of ZnO and CeO2,providing more weak acid sites contributing to the highest yield of 1,3-BD.

Fig.7. Catalytic performances of the 10ZnxCey-AS catalysts.Conditions:P=101.325 kPa, T=375°C, TOS=5 h, WHSV=1.62 h-1.

Gaoet al.[47]found that the Zr—O—Si band on ZrO2/nano-SiO2catalyst provided moderate acid and basic sites, which showed excellent catalytic activity (selectivity 93.18% and conversion 58.52%).Zhuet al.[48]also confirmed that Lewis acid sites significantly improved 1,3-BD selectivity.To reflect the relationship between acid properties of catalysts and their reaction performances,the number of surface acid sites is summarized in Table 2.The yield of 1,3-BD is positively correlated to the acidity of the weak acid,indicating that more weak acid sites promote the selective conversion to 1,3-BD.On the contrary, the yield of 1,3-BD is negatively impacted by the medium-to-strong acid sites.It indicates that higher CeO2content contributes to the formation of more weak acid sites.However, over loading of CeO2(e.g., 10Zn1-Ce8-AS catalyst) leads to a lower yield of 1,3-BD.The agglomeration of Ce reduces the content of weak Lewis acid, and a lower ZnO content gives fewer active sites for dehydrogenation,both lead to the low conversion of ethanol and yield of 1,3-BD.

Table 2The amount of surface Lewis acid of 10ZnxCey-AS catalysts and yield of 1,3-BD

Based on the above discussion,we propose a possible ETB reaction mechanism on the 10Zn1Ce5-AS catalyst(Fig.8).The two main reaction steps (dehydrogenation and aldol-condensation) occur sequentially under the balance of the two types of active sites.Generally speaking,the dissociative adsorption of hydroxyl groups occurs on metal oxides [34,49].For 10ZnxCey-AS catalysts, ethanol dehydrogenation mainly occurs on the ZnO sites to form acetaldehyde.Later, acetaldehyde molecules transfer to acetaldol under aldol-condensation and then dehydrate to crotonaldehyde on the CeO2sites.Following that, 1,3-BD is generated through the MPV reduction and dehydration reaction.In brief,on this optimized catalyst, the two active components are effectively integrated to perform a coupling effect.The dehydrogenation and the aldolcondensation activity reach a suitable balance in the ETB system.

3.7.The reaction conditions effect and catalyst stability

The influence of the reaction temperature and weight hour space velocity (WHSV) on the ETB performance of those catalysts is further investigated.As shown in Fig.9, the ethanol conversion jumps from 22.5% to 83.9% as the reaction temperature rises from 325°C to 400°C at a WHSV of 1.62 h-1.Because dehydrogenation and dehydration reactions of ethanol are both endothermic,elevating temperature is favorable for these reactions to promote the ethanol conversion.Similarly, the selectivity of ethylene also increases, while the selectivity of acetaldehyde falls when raising the reaction temperature.However,the selectivity of 1,3-BD shows only some minor changes, with the highest yield achieved at 375°C (35.7%).Given a low reaction temperature makes both the conversion of ethanol and the catalyst activity unfavorable while an excessively high temperature will add adverse penalty on energy consumption and catalyst stability, the reaction temperature needs to be chosen within some appropriate ranges to balance the ethanol conversion, 1,3-BD selectivity, and energy expense.

Fig.8. The predicted mechanism of ETB reaction.

Fig.9. Catalytic performances of 10Zn1Ce5-AS catalyst at different temperatures.

The effect of WHSV on the catalytic performances was further evaluated on the 10Zn1Ce5-AS catalyst at the desired temperature(375°C), as shown in Fig.10(a).As the increase of WHSV shortens the contact time between reactants (including the intermediates)and catalysts, a high ethanol WHSV impacts the conversion of ethanol as expected,making it decrease from 98.4%to 65.9%when WHSV varies from 0.54 h-1to 2.16 h-1.Given the selectivity of 1,3-BD is relatively stable at various WHSVs, the impact of WHSV on selectivity lies mainly for byproduct: the production of ethylene decreases while that of acetaldehyde rises along with the increase of WHSV.It seems that the shortened contact time is conducive to the dehydrogenation rate,but unfavorable to the dehydration rate.Although the overall selectivity of 1,3-BD is high, its space-time yield is insufficient(Fig.10(b)).Hence,the choice of WHSV of ethanol should be decided based on both the selectivity and space-time yield of 1,3-BD.Based on our results,a suitable WHSV is suggested to be around 1.62 h-1for the 10Zn1Ce5-AS catalyst.

Fig.10. (a) Catalytic performances of the 10Zn1Ce5-AS catalyst at different WHSV, (b) the correlation of WHSV to STY of 1,3-BD.

Fig.11. Stability of the 10Zn1Ce5-AS catalyst within 10 h.Conditions:P=101.325 kPa, T=375°C, WHSV=1.62 h-1.

The stability of the 10Zn1Ce5-AS catalyst was finally evaluated at the optimized reaction conditions.As shown in Fig.11,the ethanol conversion is maintained between 70%-78%, and the selectivity of 1,3-BD at about 40%within 8 h.With the prolonged reaction time,the selectivity of 1,3-BD and ethylene decreases while that of acetaldehyde shifts upward.It can be ascribed to the gradual coke deposition with the ETB reaction ongoing,leading to a slight deactivation of those weak Lewis acid sites.

4.Conclusions

In this study,we prepared a series of 10ZnxCey-AS catalysts with different Zn/Ce ratios and evaluated their dehydrogenation and aldol-condensation activity.The 10Zn1Ce5-AS catalyst was found to have the most suitable metal oxide ratio with an optimal balance of their roles on the two determining reactions.The ZnO species promote the dehydrogenation functions,which allows the ETB reaction to start quickly so that the ethanol conversion is enhanced.The well-dispersed ceria species posses appropriate strength and number of Lewis acid sites on catalyst.More weak Lewis acids and fewer medium-to-strong acid sites was found beneficial for the formation of 1,3-BD.By coupling these two types of active sites, the catalyst gives a selectivity of 1,3-BD as higher as 45.3%and low reaction temperature is favorable for its production.High temperature will cause not only more energy consumption,but also the ethanol dehydration to form excessive ethylene.A large WHSV of ethanol will reduce the ethanol conversion due to the shorter residence time, but barely alters the selectivity of 1,3-BD.With the aspect of high space-time yield of 1,3-BD, the optimal reaction temperature is found to be 375°C,and a suitable WHSV at 1.62 h-1.Long reaction time will lead to slight declination of the conversion of ethanol and selectivity of 1,3-BD ascribed to the gradual coke deposition.These results provide valuable insights in designing new catalyst for efficient valorization of bioethanol.

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

We appreciate the National Natural Science Foundation of China for financial support (21878227).