Catalytic transfer hydrogenation of furfural for ... - Wiley Online Library

5 downloads 0 Views 2MB Size Report
ethyl levulinate: Interplay of Lewis and Brønsted acidities. Kan Tang, [a] Shaoqu Xie, ..... P. Panagiotopoulou, N. Martin, D. G. Vlachos, ChemSusChem. 2015, 8 ...
Energy Technology Generation, Conversion, Storage, Distribution

Accepted Article Title: Catalytic transfer hydrogenation of furfural for the production of ethyl levulinate: Interplay of Lewis and Brønsted acidities Authors: kan Tang, Shaoqu Xie, Gabriel R Cofield, Xiaokun yang, Emmy Tian, and Hongfei Lin This manuscript has been accepted after peer review and appears as an Accepted Article online prior to editing, proofing, and formal publication of the final Version of Record (VoR). This work is currently citable by using the Digital Object Identifier (DOI) given below. The VoR will be published online in Early View as soon as possible and may be different to this Accepted Article as a result of editing. Readers should obtain the VoR from the journal website shown below when it is published to ensure accuracy of information. The authors are responsible for the content of this Accepted Article. To be cited as: Energy Technol. 10.1002/ente.201700973 Link to VoR: http://dx.doi.org/10.1002/ente.201700973

www.entechnol.de

10.1002/ente.201700973

Energy Technology

FULL PAPER Catalytic transfer hydrogenation of furfural for the production of ethyl levulinate: Interplay of Lewis and Brønsted acidities Kan Tang, [a] Shaoqu Xie, [a] Gabriel Ross Cofield, [a] Xiaokun Yang, [b] Emmy Tian, [c] and Hongfei Lin*[a]

Introduction In recent years, alkyl levulinate as a bio-derived compound has attracted significant research interests due to its promising potential in various chemical applications.[1] First of all, it has been found to be a useful solvent since it has much lower vapor pressures compared to other organic solvents.[2–4] Alkyl levulinate also has some miscellaneous uses in food additives and latex coatings.[5] However, the possible large-quantity use of alkyl levulinate is as a fuel additive/precuror. Directly serving as a bio-based blender, it significantly improves the cold flow properties of biodiesel by the reduction of cloud point (CP), pour point (PP), and cold filter plugging points (CFPP).[6] Alternatively, alkyl levulinate can also be transformed into gamma-valerolactone (GVL), an important precursor for the downstream production of

[a]

[b]

[c]

K. Tang, Dr. S. Xie, G. R. Cofield, Prof. H.Lin The Gene and Linda Voiland School of Chemical Engineering and Bioengineering Washington State University, 1505 Stadium Way, Pullman, WA 99164, USA E-mail: [email protected] X. Yang The Department of Chemical and Materials Engineering University of Nevada, Reno, NV 89557, USA E. Tian The Department of Chemistry University of California, Berkeley CA 94720 USA Supporting information for this article is given via a link at the end of the document

liquid hydrocarbon fuels,[7–10] via hydrogenation subsequent intramolecular esterification.[11–13]

and

The current synthesis methods of alkyl levulinate mostly involves levulinic acid as the reactant. Evidently, levulinic acid can be catalytically esterified with alcohols, by either homogenous or heterogeneous acid catalysts, to produce desired alkyl levulinates. Recently, an increasing effort has been developed to produce alkyl levulinate from (poly)saccharides. Hexose (C6) feedstocks such as glucose and fructose can be transformed into alkyl levulinates or their corresponding ether species, in alcohol via the formation of 5-hydroxymethylfurfural (HMF) as an intermediate, which can be subsequently rehydrated to the final ester upon the loss of formic acid.[5,14–16] However, the usage of C6 hexose monosaccharides or polysaccharides composed primarily of hexoses for fuel production might compete with the food supply. Meanwhile, alkyl levulinate production from inedible feedstocks, i.e. C5 compounds, is generally overlooked. Recently, Hu et al. reported the first direct synthesis of alkyl levulinates from xylose in the co-presence of Amberlyst-70 and Pd/Al2O3 using external pressurized hydrogenation, however the yield was rather low (~20%). [16] Ideally, it would be more sustainable to avoid high pressure external hydrogen. The pioneering work of Corma et al. [17–19] on Lewis acid catalyzed Meerwein–Ponndorf–Verley (MPV) transfer hydrogenation reactions, demonstrated that this mechanism has also been well adopted for producing biodiesel components via the cascade of transfer hydrogenation and etherification/acetalization of furfural/HMF.[20–25] Following the work of Corma et al. [17–19] on MPV reactions, Bui et al. [26] reported an elegantly integrated process for the production of GVL from furfural (FUR) through a domino reaction. The reaction was first catalyzed by the Lewis Acid in the MPV transfer hydrogenation of furfural and later by the Brønsted acid in the hydrolytic ring-opening and lactonization reaction. However, the sacrificial alcohol used as a hydrogen donor in that domino reaction was 2-butanol, which is apparently rather expensive. In this paper, we prove that the production of ethyl levulinate can also be realized by a tandem catalytic reaction combining transfer hydrogenation of furfural and subsequent ethanolysis of furfuryl alcohol into ethyl levulinate (EL) with a high yield, which was achieved by finely tuning the molar ratio of Brønsted to Lewis acidities. In this case, bio-ethanol, an abundant and eco-friendly alcohol, acted as the hydrogen donor in the MPV transfer hydrogenation of furfural into furfuryl alcohol. Consequently, EL was proven to be effectively producible, without requiring the use of high pressure hydrogen.

This article is protected by copyright. All rights reserved.

Accepted Manuscript

Abstract: alkyl levulinate is sought after due to its broad applications such as a fuel additive in biodiesel fuels. Currently it is industrially produced from levulinic acid using harmful, highly corrosive acids like H2SO4 and HCl as the catalysts. Herein, a new process is developed that produces ethyl levulinate from furfural and ethanol with the combination of both Brønsted and Lewis solid acids. The transfer hydrogenation of furfural by ethanol is catalyzed by the solid acids, leading to the alternative precursor for ethyl levulinate production, furfuryl alcohol. The subsequent ethanolysis of furfuryl alcohol, which produce ethyl levulinate, is also enhanced by using an optimal ratio of Brønsted to Lewis acids. Bio-ethanol is chosen as the sacrificial hydrogen donor for proton transfer hydrogenation since it is environmentally friendly, more abundant and affordable than other alcohols that are commonly used for large-scale production processes. The tandem catalytic reactions can be effectively maneuvered by tuning the Brønsted and Lewis acidities to produce ethyl levulinate in relatively fast kinetic rates.

10.1002/ente.201700973

Energy Technology

FULL PAPER Proposed reaction mechanisms and side reactions To start with, the proposed reaction pathway (Figure 1) is referred first to give an overview of the tandem reaction system. It has been previously mentioned that ethanol in the reaction system acted as sacrificial hydrogen donor and was converted into acetaldehyde. The reactive acetaldehyde was stoichiometrically converted into acetaldehyde diethyl acetal (DEA) via acetalization with two more parts of ethanol. Under the tested reaction conditions, acetaldehyde, as well as its derived products (other than DEA), was not detected, implying that acetaldehyde was completely converted to DEA with a 100% selectivity. Thus it might be sensible to regard the ratio of [DEA]/[FUR]0 as an indicator of the extent of MPV transfer hydrogenation. It should also be pointed out that DEA as a side product of the MPV transfer hydrogenation, can be regarded as a commercial fine chemical and recovered through hydrogenation over base metals or used for other chemical processes. [11,21] Thus it should not be viewed as waste or an undesired product. The intended reaction involves ethanol reacting with furfural as opposed to acetaldehyde, where furfural accepts the hydride transfer from ethanol and is reduced to furfuryl alcohol (FAL), the critical intermediate for the production of EL. It could also be validated later on that the co-presence of Lewis and Brønsted acids resulted in a synergistic effect on this crucial ring opening step to facilitate the production of EL. Meanwhile, in excess ethanol, furfural not only underwent ethanolysis ring opening into the final product, but was also partially acetalized by ethanol into 2-furaldehyde diethyl acetal (FAC). Furfuryl alcohol went through another side reaction, dimerization, to produce the reaction intermediate, difurfuryl ether (DFE). Presumably both FAC and DFE could

Figure 1 Postulated reaction pathway of the tandem catalytic conversion of furfural (FUR) into ethyl levulinate (EL) in the co-presence of Lewis (L) and Brønsted (B) acid.

Last but not least, the polymerization reaction of both furfural and furfuryl alcohol should also be taken into consideration since those side reactions dictate the final yield and selectivity of EL. Reactions with the Lewis acid catalyst The results of furfural conversion in subcritical ethanol with only the Lewis acid catalyst, Zr-SBA-15, are presented in Table 1. The conversion of FUR and the yield of EL after an 18 h reaction increased from 69.8% to 100% and from 0.9% to 41.8% , respectively, upon elevating the reaction temperature from 120oC to 180oC. Notably the conversion of furfural and the EL yield at 180oC in an 8 h reaction was even higher than those at lower temperatures in 18 h reactions.

Table 1 Product Distribution after the conversion of furfural over the single Lewis acid catalyst at different temperatures. Reaction condition: 1 wt% furfural, 100 mg Zr-SBA-15 catalyst, 30 mL ethanol, p(N2) = 400 psi.

FAL yield %

DFE yield%

[DEA]/[FUR]0 %

EL yield %

S.R. wt%

36.2

0

0

0

0

1.4

69.7

64.1

1.3

1.8

7.1

0

2.5

18

69.8

34.3

11.3

11.3

24.4

0.9

3.6

150

1

56.2

40.6

6.6

5.2

13.6

0

2.2

5

150

4

68.0

27.3

6.9

28.2

37.5

3.8

3.6

6

150

18

83.4

20.5

0

24.8

55.9

12.8

5.5

7

180

1

73.8

44.9

5.9

12.3

19.9

1.5

5.8

8

180

4

94.0

3.5

0

60.4

85.3

14.1

6.3

9

180

8

100

2.7

0

34.5

85.8

28.6

7.2

10

180

18

100

0

0

14.6

98.2

41.8

9.1

Entry

Temperature oC

Time h

FUR conversion %

1

120

1

39.4

2

120

4

3

120

4

FAC yield %

Denotation: FUR – furfural, FAL – furfuryl alcohol, FAC – 2-furaldehyde diethyl acetal, DFE – difurfuryl ether, DEA – acetaldehyde diethyl acetal, [DEA]/[FUR]0 – molar ratio of produced DEA to initial furfural, EL – ethyl levulinate, S.R. – solid residue.

This article is protected by copyright. All rights reserved.

Accepted Manuscript

be reversibly transformed back into furfural and furfuryl alcohol for the ultimate irreversible conversion into EL.

Results and Discussion

10.1002/ente.201700973

Energy Technology

However, the significant differences were observed in the rates of conversion of three key intermediates FAC, FAL and DFE upon temperature elevation. At 120 oC, FAL and DFE did not appear in a short time (1 h reaction) while FAC had been produced, which indicated that acetalization was faster than either transfer hydrogenation or dimerization at lower temperatures. After prolonging the reaction time till 18 hours at 120 oC, the yield of FAC increased initially and then decreased, while those of DFE and FAL continued to increase. At 150 oC or 180 oC, the EL yield increased significantly, while the yield of solid residue only slightly increased as the reaction time increased. Note that the diminishing rates of the yields of DFE and FAL was significant at higher temperatures. These phenomena suggested that both FAC and DFE could be reversibly transformed back into FUR and FAL, respectively, towards ultimate irreversible conversion to EL. FUR was unable to be converted into EL at 120 oC within 4 hours. Indeed, FAL, FAC and DFE were still the dominant products even after 18 hours. The formation of DFE occurred after 4 hours at 180 oC and it was further transformed into EL as the reaction proceeded. These observations implied that ethanolysis for ring opening of FAL could be the rate limiting step at 120oC. On the other hand, at 150oC or 180oC, the diminishing rates of the yields of three intermediates were in the order of FAL, FAC and DFE. Therefore, it is reasonable to conclude that at higher temperatures, the reversible acetalization and dimerization occur and the ethanolysis of FAL may no longer be the rate limiting step.

especially at elevated temperatures. [31,32] The formed solid resin is dark coloured and insoluble in ethanol. At lower temperatures, the formation of the solid residue was rather limited. However, as the temperature increased to 180 oC, the formation of the solid residue was intensified and the weight percentage of the solid residue reached approximately 10 wt% after 18 h. Thus it implies that the resinification of furfural/furfuryl alcohol is sensitive to thermal activation, which is in agreement with previous reports. On the other hand, since the MPV reaction follows a redox mechanism, it is also worthwhile to consider the extent of hydrogen donation in the reaction. At a fixed reaction time, the ratio of [DEA]/[FUR]0 consistently increased with increasing temperature, indicating that elevated temperature promoted MPV transfer hydrogenation. The highest [DEA]/[FUR]0 is above 98% after an 18 h reaction at 180 oC. This unexpectedly high value indicated that FUR can be completely converted into FAL through the MPV transfer hydrogenation, implying that the undesired polymerization / resinification might be predominantly derived from FAL, but not directly from FUR itself. The amount of ethanol-soluble resin derived from FAL could originated from 25-35 wt% of the total FUR feedstock. Kim et al. suggested that the polymerization of FAL catalyzed by Brønsted acid resulted in highly conjugated diene structures. [33,34] However, based on our NMR characterization results (Figure 2), the ethanolsoluble resin showed no characteristic de-shielding peaks with large chemical shifts (i.e.> 6 ppm) of conjugated dienes. Thus we propose that the ethanol-soluble resin is more likely the Diels-Alders adduct of FAL.[31,35]

As for the side reactions, it is well known that FUR or FAL is very sensitive to resinification catalyzed by acids

(b)

Intensity a.u.

200

m/z=413.44

150

100

50

(a) 0 300

350

400

450

500

550

m/z

(c)

Figure 2 Panel (a) Liquid fraction after the conversion of furfural was removed by rotary evaporator to give out dark red resin component. Panel (b) MALDI mass

This article is protected by copyright. All rights reserved.

Accepted Manuscript

FULL PAPER

10.1002/ente.201700973

Energy Technology

FULL PAPER

Reactions with the combination of Brønsted acid and Lewis acid From the above discussion, we had illustrated that the tandem conversion of FUR could be achieved by the sole Lewis acid catalyst despite its lower yields and slower reaction kinetics. Herein, we demonstrate that inclusion of additional Brønsted acid will result in a synergistic effect on the production of EL by significantly improving both the EL yield and the reaction kinetics. The results of a combination of Brønsted acid (B) and Lewis acid (L) at various reaction temperatures are shown in Figure 3. Except for the reaction at 120 oC, all other reactions at higher temperatures achieved a 100% conversion of FUR in 18 h. The maximum yield of EL, 49.2%, was obtained at 180 oC. At the same time, the diminishing rates of the yields of three intermediates upon elevated temperatures were in the order of FAC, DFE and FAL. Especially, FAC and DFE were quickly consumed (the similar trend was also observed when the Lewis acid catalyst was solely used), while FAL was barely detected. FAC as the product of reversible acetalization seemed to be unstable at higher temperatures in the presence of both Brønsted acid and Lewis acid. The molar ratio of [DEA]/[FUR]0 increased as the temperature increased from 120 oC and reached its highest value of 69% at 150 oC before a drastic decline upon the further increase in temperature. Interestingly, the maximal yields of DEA and EL did not appear simultaneously. Whereas, the resin derived from FAL was formed largely at higher temperatures (>180 oC), which resulted in the rapid decrease of the EL yield.

100 FUR conversion FAC yield DFE yield [DEA]/[FUR]0

60

EL yield

%

80

noted that FAC and FAL as two of the key reaction intermediates had been completely converted after 18 h at any B/L ratios and thus their yields were not shown in the figure. Under the complete conversion of FUR at all B/L ratios, even the most stable intermediate, DFE, proved to be completely converted except in a few cases with very low B/L ratios (i.e.0.15) obviously triggered more side reactions, such as the resinification of FAL which resulted in the sharp deterioration of the EL yields. As seen in Figure 4, on the hydrogen donor side, a drastic decline of [DEA/FUR] 0 occurred (a reduction from nearly 100% to just above 30%) with increasing the B/L ratio, implying that the inclusion of Brønsted acid might significantly promote the ethanolysis of acetals. This observation is well coincided with the interpretations on the hydrogen acceptor side since the ethanolysis of FAC, FAL and DEA were also promoted by the mild Brønsted acidities in our previous discussion.

40

100

20

80

0

60

58

EL yield

56 54 52 50 %

48 46 44 42 40 38

140

160

180

200

220

240

0

%

120

30

40

50

60

70

80

FUR conversion DFE yield [DEA]/[FUR]0

o

20

In Figure 4, the results of finely tuning the molar ratio between Brønsted and Lewis acids are presented. It is

20

[B] /[L] %

Temperature C

Figure 3 Distribution of the products after the conversion of furfural at different temperatures with a combination of Lewis and Brønsted acid catalysts. Reaction condition: 1wt% furfural, 100mg Zr-SBA-15 catalyst (Lewis acid), 35mg ZSM-5 catalyst (Brønsted acid), 30ml 100wt% ethanol as solvent, p(N2) = 400psi, 18 h, FAL was not detected except at 120 °C. The acidities of Zr-SBA-15 and ZSM-5 were determined by NH3-TPD to be 0.74mmol/g and 0.54mmol/g, the molar ratio of B/L= 0.258

10

40

EL yield 0 0

10

20

30

40

50

60

70

[B] / [L] %

Figure 4 Distribution of the products of after the conversions of furfural over a series of the combination of ZSM-5 and Zr-SBA-15. Reaction conditions: 0.23 g furfural (1 wt%), 100 mg Zr-SBA-15 catalyst, 30 mL ethanol, p(N2) = 400 psi, 180oC, 18 h. FAL and FAC were not detected.

This article is protected by copyright. All rights reserved.

Accepted Manuscript

spectrum of the ethanol soluble resin, 2,5-dihydroxybenzoic acid (2-DHB) as matrix. Panel (c) 1H NMR spectrum of the ethanol soluble resin, CDCl3 as solvent (δ=7.26ppm). Reaction conditions: 100mg Zr-SBA-15 catalyst, 30ml ethanol, p(N2) = 400psi, 180oC, 18h.

10.1002/ente.201700973

Energy Technology

FULL PAPER Figure 6 Comparison of the ethyl levulinate yields in the two-step approach and the one-pot tandem approach over a combination of Zr-SBA-15 with various Brønsted acids. Reaction condition: 100 mg Zr-SBA-15 catalyst, 0.054 mmol Brønsted acid (equivalent to 10 mg of ZSM-5 catalyst), 30 mL ethanol, p(N2) = 400 psi, 180 oC, 18 h. EL-ethyl levulinate.

60 B/L=0 B/L=0.25 B/L=0.074

50

Conclusions

30 20 10 0 0

2

4

6

8

10

12

14

16

18

Reaction time / h

Figure 5 Plots of ethyl levulinate yield v.s. reaction time at different B/L molar ratios. Condition: 0.23 g furfural (1 wt%), 100 mg Zr-SBA-15 catalyst, 30 mL ethanol, p(N2) = 400 psi. The acidities of Zr-SBA-15 and ZSM-5 were determined to be 0.74 mmol/g and 0.54 mmol/g.

Since the synergistic effect had been proven in the cases of ZSM-5, it would be intuitive to verify whether or not other types of Brønsted acids could play a similar role. After a screening on the three different Brønsted acids, ZSM-5, WZrO2 and phosphotungstic acid (PTA), it was determined that only ZSM-5 contributed to the synergistic effect on EL production, while the non-porous heterogeneous catalyst (W-ZrO2) and the homogenous catalyst (PTA) did not exhibit the synergistic effect under the tested reaction conditions (Figure 6). However, the probability cannot be ruled out that other Brønsted acids may still display a synergistic effect under different conditions. Indeed, as suggested by the work of Dumesic’s group,[36] the unique pore structure and the hydrophobic properties of ZSM-5 might give rise to the promotion of the production of EL by significantly suppressing the polymerization side reactions. Thus the future direction of this tandem conversion would be the development of bi-functional solid Brønsted/Lewis acid catalysts that have a similar structure to that of ZSM-5.

60 EL yield 50

%

40

This study proved that Zr-SBA-15 as a Lewis acid plays a critical role by catalyzing the Meerwein–Ponndorf–Verley reduction reaction, converting furfural into furfural alcohol using hydrogen from the sacrificial donor, ethanol. At the same time, intermediates difurfuryl ether and furfural acetal were reversibly formed via dimerization of furfural alcohol and acetalization of furfural, respectively. After introducing a finely tuned amount of ZSM-5 with Brønsted acidity (the molar ratio of B/L=0.074), an optimized ethyl levulinate yield of 55% was achieved after an 8hour reaction at 180 ºC, which proved the significant synergistic effect of the combination of Brønsted and Lewis acidities.

Experimental Section Materials The following reagents and products were used for the experiments: furfural (99%), furfuryl alcohol (98%), 2-furaldehyde diethyl acetal (>97%), acetaldehyde diethyl acetal (99%), triblock copolymer Pluronic P123, hydrochloric acid (36.5-38.0%, BioReagent), tetraethyl orthosilicate (>99.0%) and zirconyl chloride octahydrate precursor (98%) were purchased from Sigma Aldrich. Ethyl levulinate (99%), acetaldehyde diethyl acetal (99%), ZSM-5 (Si:Al = 80:1) were purchased from Alfa Aesar. Tungsten doped zirconium hydroxide (W/ZrO2) was purchased from Melchemicals (UK). The solid acid catalysts were calcined at 550 oC for 6 h to be activated before use. Catalyst preparation Zr-SBA-15 was synthesized according to the procedure described in reference[27] and had been well implemented in our previous articles.[28][29] Briefly, 2 g of Pluronic P123 was added to 75 ml of 1.6 M HCl solution. The mixture was stirred at 40°C for 3 hours until all P123 was dissolved. Next, 4.25 g of TEOS and the appropriate amount of zirconia precursor (the ratio of Si/Zr is 20) were added into the solution and the mixture was stirred for another 24 hours at 40°C. The resulting gel was placed in a Teflon-lined autoclave and heated at 100 °C for 24 hours. The solid product was filtered and rinsed with de-ionized water, dried at 100 °C overnight, and calcined in flowing air at 550 °C for 6 hours. Catalyst characterization

30

In our previous study,[28][29] a detailed characterization of Zr-SBA-15 had already been carried out. Based on the pyridine-DRIFTS results, it had been proved that Zr-SBA-15 was predominantly a Lewis acid. When Zr-SBA-15 was further analyzed, the results from N2 adsorption, TEM and SAXS techniques mutually concluded the pore size of ZrSBA-15 to be approximately 9 nm.

20

10

0 Zr-SBA-15

Zr-SBA-15 Zr-SBA-15 Zr-SBA-15 W-ZrO2 ZSM-5 ZSM-5 Two steps Tandem Tandem

Zr-SBA-15 PTA Tandem

The critical variable in this paper was the molar ratio between Brønsted and Lewis acid (B/L). Thus the acidity of each solid acid used was determined by NH3-TPD to ensure comparisons on the same acidity basis. In a typical NH 3-TPD experiment using a

This article is protected by copyright. All rights reserved.

Accepted Manuscript

EL yield %

40

10.1002/ente.201700973

Energy Technology

Micromeritics AutoChem ӀӀ 2920 Chemisorption Analyzer, the catalyst was first degassed in helium flow at 250 °C for 1 hour, then the temperature was decreased to 100 °C. After that, 10% ammonia in helium was exposed to the sample at 100 °C for 60 min. 50 mL/min helium was then flowed over to remove physically adsorbed ammonia. Temperature programmed desorption was carried out from 100 °C to 550 °C with a temperature ramp at 10 °C/min. The results of NH 3-TPD were shown in Figure S1.

[6]

H. Joshi, B. R. Moser, J. Toler, W. F. Smith, T. Walker, Biomass and Bioenergy 2011, 35, 3262–3266.

[7]

D. M. Alonso, J. Q. Bond, J. C. Serrano-Ruiz, J. a. Dumesic, Green Chem. 2010, 12, 992–999.

[8]

S. M. Sen, E. I. Gürbüz, S. G. Wettstein, D. M. Alonso, J. a. Dumesic, C. T. Maravelias, Green Chem. 2012, 7, 3289–3294.

[9]

J. Q. Bond, D. Wang, D. M. Alonso, J. a. Dumesic, J. Catal. 2011, 281, 290–299.

All reactions were carried out in a 160 mL stirred Parr reactor, whereby the catalyst was suspended in a solution of furfural in ethanol (30 ml) and the reactor was charged with 400 psi N 2 initially and then heated at a ramp rate of 10 °C/min until the desired set temperature was reached. During the reaction, mixing was achieved through an overhead propeller operating at 600 RPM. Once the set temperature was attained, the reactor was held for the set reaction time, and then was quenched quickly in an ice bath to stop the reaction. The reactor was cooled to approximately 25 °C before being vented after the gas pressure was recorded. The reactor was then immediately broken down and the solid residue remaining in the reactor was recovered and dried. The aqueous and solid fractions were separated by centrifugation. The solid fractions were calcined at 550 oC for 12 h to remove the organic residue on the catalyst for calculating the solid residue (S.R.) yields.

[10]

J. Q. Bond, D. M. Alonso, D. Wang, R. M. West, J. a Dumesic, Science (80-. ). 2010, 327, 1110–1114.

[11]

M. Chia, J. a. Dumesic, Chem. Commun. 2011, 47, 12233–12235.

[12]

a. M. Hengne, N. S. Biradar, C. V. Rode, Catal. Letters 2012, 142, 779–787.

[13]

W. R. H. Wright, R. Palkovits, ChemSusChem 2012, 5, 1657–1667.

[14]

R. Liu, J. Chen, X. Huang, L. Chen, L. Ma, X. Li, Green Chem. 2013, 15, 2895–2903.

[15]

S. Saravanamurugan, A. Riisager, ChemCatChem 2013, 5, 1754– 1757.

[16]

X. Hu, C.-Z. Li, Green Chem. 2011, 13, 1676–1679.

Chemical analysis

[17]

A. Corma, M. E. Domine, S. Valencia, J. Catal. 2003, 215, 294–304.

The products in the liquid ethanol were qualitatively and quantitatively analyzed by GC-MS and GC-FID, respectively. Briefly, the liquid phase solution after the reaction was filtered through a 0.45 micron syringe filter, and then diluted 15 times with ethanol. The samples were injected into an Agilent 6890 series GC/MS equipped with an Agilent DB5-MS column (30 m x 0.25 mm ID, 0.25 um film thickness) and an Agilent 5973 Mass Selective Detector for qualitative analysis of the products. The same procedure was also applied in preparing the samples for quantitative analysis, which were injected into a Shimadzu GC-2010 equipped with a SHRXI-5MS column (30m x 0.25 mm ID, 0.25 um film thickness) and a FID detector. The GC-FID responses of furfural, furfuryl alcohol, 2-furaldehyde diethyl acetal and acetaldehyde diethyl acetal were determined by the calibration curves of the chemical standards, as shown in Figure S2. The response factor of difurfuryl ether was estimated from furfuryl alcohol by implementing a wellestablished estimation method.[30] The analysis was repeated three times for each sample to obtain the average value. The yield of the product was calculated as follows:

[18]

A. Corma, M. E. Domine, L. Nemeth, S. Valencia, J. Am. Chem. Soc. 2002, 124, 3194–3195.

[19]

A. Corma, M. Renz, Angew. Chemie - Int. Ed. 2007, 46, 298–300.

[20]

P. Panagiotopoulou, N. Martin, D. G. Vlachos, ChemSusChem 2015, 8, 2046–2054.

[21]

J. Jae, E. Mahmoud, R. F. Lobo, D. G. Vlachos, ChemCatChem 2014, 6, 508–513.

[22]

W. Hao, W. Li, X. Tang, X. Zeng, Y. Sun, S. Liu, L. Lin, Green Chem. 2015, 18, 1080–1088.

[23]

J. D. Lewis, S. Van De Vyver, A. J. Crisci, W. R. Gunther, V. K. Michaelis, R. G. Griffin, Y. Rom??n-Leshkov, ChemSusChem 2014, 7, 2255–2265.

[24]

M. M. Antunes, S. Lima, P. Neves, A. L. Magalhães, E. Fazio, A. Fernandes, F. Neri, C. M. Silva, S. M. Rocha, M. F. Ribeiro, et al., J. Catal. 2015, 329, 522–537.

[25]

M. Koehle, R. F. Lobo, Catal. Sci. Technol. 2016, 6, 3018–3026.

[26]

L. Bui, H. Luo, W. R. Gunther, Y. Román-Leshkov, Angew. Chemie - Int. Ed. 2013, 52, 8022–8025.

[27]

D. M. Alonso, J. Q. Bond, J. C. Serrano-Ruiz, J. a. Dumesic, M. M. Antunes, S. Lima, P. Neves, A. L. Magalhães, E. Fazio, A. Fernandes, et al., Chem. Mater. 2004, 16, 4174–4180.

[28]

L. Yang, X. Yang, E. Tian, H. Lin, ChemSusChem 2016, 9, 36–41.

[29]

L. Yang, X. Yang, E. Tian, V. Vattipalli, W. Fan, H. Lin, J. Catal. 2016, 333, 207–216.

[30]

J. T. Scanlon, D. E. Willis, J. Chromatogr. Sci. 1985, 23, 333–340.

[31]

A. Gandini, M. N. Belgacem, Prog. Polym. Sci. 1997, 22, 1203– 1379.

Catalytic reactions

Yield of product % = Mole of product/Mole of initial furfural

Keywords: Ethyl levulinate • furfural • ethanol • tandem catalytic process • Lewis acid [1]

M. J. Climent, A. Corma, S. Iborra, Green Chem. 2014, 16, 516– 547.

[2]

L. Lomba, B. Giner, I. Bandres, C. Lafuente, M. R. Pino, Green Chem. 2011, 13, 2062–2070.

[3]

H. Guerrero, C. Lafuente, F. Royo, L. Lomba, B. Giner, Energy and Fuels 2011, 25, 3009–3013.

[4]

L. Lomba, C. Lafuente, M. Garcia-Mardones, I. Gascon, B. Giner, J. Chem. Thermodyn. 2013, 65, 34–41.

[5]

A. Demolis, N. Essayem, F. Rataboul, ACS Sustain. Chem. Eng. 2014, 2, 1338–1352.

This article is protected by copyright. All rights reserved.

Accepted Manuscript

FULL PAPER

10.1002/ente.201700973

Energy Technology

FULL PAPER R. Mariscal, P. Maireles-Torres, M. Ojeda, I. Sádaba, M. López Granados, Energy Environ. Sci. 2016, 9, 1144–1189.

[33]

T. Kim, R. S. Assary, C. L. Marshall, D. J. Gosztola, L. a. Curtiss, P. C. Stair, ChemCatChem 2011, 3, 1451–1458.

[34]

T. Kim, R. S. Assary, H. Kim, C. L. Marshall, D. J. Gosztola, L. a. Curtiss, P. C. Stair, Catal. Today 2013, 205, 60–66.

[35]

R. Gheneim, C. Perez-Berumen, A. Gandini, Macromolecules 2002, 35, 7246–7253.

[36]

M. A. Mellmer, J. M. R. Gallo, D. Martin Alonso, J. A. Dumesic, ACS Catal. 2015, 5, 3354–3359.

Accepted Manuscript

[32]

This article is protected by copyright. All rights reserved.

10.1002/ente.201700973

Energy Technology

Accepted Manuscript

FULL PAPER

This article is protected by copyright. All rights reserved.