Efficient synthesis of biofuel precursor with long carbon chains from fructose

Abstract

Long carbon biofuel precursors were efficiently synthesized via Aldol condensation of furans (furfural, 5-ethoxymethylfurfural, 5-hydroxymethylfural) and their derivatives (levulinic acid, ethyl levulinate) which were co-produced by fructose transformation in ethanol using acidic metal salt catalysts. High yields of furans and their derivatives were obtained and their mole ratios were adjusted. The effects of reaction time, temperature and water content on the mole ratios of furans and their derivatives were investigated in detail. The condensation reaction was conducted directly when the mole ratio of furans and their derivatives was located in the range of 1 to 2. The results showed that a total mole yield of furans and their derivatives as high as 91.6% could be obtained after fructose alcoholysis at 413 K for 2.0 h using Fe₂(SO₄)₃ as a catalyst. After that, Aldol condensations of the produced furans and their derivatives with mole ratios of 1.9 and 1.1 were directly conducted using NaOH as the catalyst and 91.2% and 89.1% yields of precursors were gained, respectively. This technical route indicates a simple and feasible method to produce renewable long carbon biofuel from biomass.

---

Introduction

Compositional changes to the atmosphere are embodied in the disequilibrium of our environment due to anthropogenic emissions.¹ Moreover, some conventional long carbon fuels are derived from sharply depleting fossil fuels. Therefore, it is urgent to exploit a new route for producing fuel from renewable biomass to substitute those originating from fossil fuels. Although short chain hydrocarbon fuels can be obtained from various platform chemicals derived directly from biomass, the formation of long carbon biofuels needs further C–C coupling between key platform chemicals over acid/base catalysts.^2–4 Meanwhile, the oxygen atoms incorporated in the precursors should be removed from the condensed compounds by hydrodeoxygenation when used as transportation fuels.^5–7

In recent years, the routes to long chain biofuels by using biomass include two methods. One is the conversion of biomass to syngas and then the use of the Fischer–Tropsch synthesis.⁸ However, the structures of these synthesized products are almost entirely restricted to straight chains. Another one is the conversion of biomass to platform chemicals (furans and derivatives) and then condensation by C–C coupling.⁹ Furfural (FA) and 5-hydroxymethylfural (5-HMF) are the most important furan compounds,^10,11 and they can be derived into 5-methylfurfual (5-MF), 2-methylfuran (2-MF), levulinic acid (LA) and angelica lactone (AL). In fact, all of these compounds could be selected as feedstocks to elongate carbon chains by C–C couplings. Fu and her team reported that the fuel precursors with C₁₀–C₁₄ could be easily obtained through direct self-coupling of FA and 5-MF under mild conditions in water.¹² After hydrodeoxygenation, straight or branched C₈–C₁₄ alkanes of diesel-range fuels were obtained in moderate to high yield. Corma also reported that high-quality diesel fuel could be achieved via 2-MF and 6-butylundecane condensation followed by hydrodeoxygenation, obtaining an overall yield of 87%.¹³ Another investigation found that a C–C bond could be formed between AL moieties through a free radical reaction under mild conditions without using a noble-metal catalyst and solvent, which gave carbon chain elongated products with 10 or 15 carbons with complete conversion and 100% selectivity.¹⁴

Actually, cross condensation is the most common method in the carbon chain elongation reaction. FA and 2-MF can be condensed efficiently over sulfonic acid catalyst under solvent-free conditions.¹⁵ C₈ and C₁₃ alkane precursors can be obtained from furfural and acetone via Aldol condensation over base catalyst in an aqueous system, and the selectivities for n-alkanes were higher than 50% through hydrodeoxygenation.¹⁶ The reaction of 2-MF with acetone and butanal was also investigated over a series of solid acid catalysts, and evidently higher carbon yields of diesel were obtained when the hydroxyalkylation/alkylation product of 2-MF with butanal was used as the feedstock.¹⁷

Although long chain hydrocarbons (C₈–C₁₅) could be obtained from the feedstocks via different condensation routes mentioned above, the feedstocks adopted in the present investigation still only depended on model chemicals (those chemicals like acetone that could not be obtained easily from biomass with high yields) and the main work was focused on hydrodeoxygenation with different catalysts. The directions on how to efficiently use biomass and obtain precursors with high yields are largely neglected. Therefore, it was high time to find new substrates and use a simple synthesis route to economically obtain precursors with long carbon chains for biofuel production.

Fructose is an important carbohydrate in the conversion of biomass to furan chemicals and its derivatives. Many investigations on the production of furan compounds and levulinate esters from fructose have been reported in recent years. The highest 5-ethoxymethylfurfural (EMF) yield of 71.2% was obtained directly from fructose through a one-pot reaction over AlCl₃.¹⁸ By controlling the reaction conditions, high 5-HMF and EL yields of up to 89% and 86%, respectively, were achieved from fructose.¹⁹

Herein, we used fructose as a raw material to produce furan compounds (FA, 5-HMF and EMF) and their derivatives (LA and EL) through one-pot conversion. According to the mechanism of Aldol condensation, there are at most two carbon active sites in furan derivatives formed under alkaline conditions. Therefore, only one or two aldehyde carbons of furans can be interlinked with active carbons via C–C coupling.^20,21 In this paper, a mixture furans and derivatives with appropriate mole ratios directly from fructose decomposition were then used as the feedstock to obtain long-chain precursors via Aldol condensation. When the mole ratio of furans and derivatives was at the range of 1 to 2, green Aldol condensation would occur with complete conversion of furans and derivatives. Due to the majority components of the alcoholysis products being EMF and EL (Fig. S1†), the main condensation scenario was proposed as in Fig. 1. Meanwhile, tiny amounts of FA and 5-HMF could also be condensed with LA and EL respectively over base catalyst. Finally, carbohydrate fuel precursors with different carbon chains were obtained via condensation.

Fig. 1 Proposed main reaction routes from fructose to long carbon fuel precursors.

Experimental section

Materials

Fructose (purity > 99%) was purchased from MYM Biological Technology Company Limited; H₂SO₄, Al₂(SO₄)₃, ZnSO₄, NaH₂PO₄, and Fe₂(SO₄)₃ (purity > 99%) were all bought from Damao chemical reagents company, China; C₂H₅OH (purity > 99%) and NaOH (purity > 96%) were bought from Guangdong Guanghua Sci-Tech Co. Ltd China.

General reaction procedure

The detailed experimental procedures are shown in Fig. 2. 2.0 g fructose, 0.2 g Fe₂(SO₄)₃ and 50 mL ethanol were placed into a slurry bed successively. Then, the reaction system was heated to desired temperature for a certain amount of time. After that, the reaction was halted and cooled down to room temperature with cold water. At this time, the main components of the products were furan compounds and ethyl levulinate (Fig. S1†). Sodium hydroxide was added directly into the mixed solution to induce the condensation reaction at 373 K for 4.0 hours. Afterwards, calcium oxide powder was introduced into the solution to remove Fe³⁺ and SO₄²⁻ followed by a filtration step. After that, the pH value of the filtrate was adjusted to 5.0 with hydrochloric acid. The final filtrate was concentrated by rotary evaporation and vacuum dried at 323 K overnight. Then, the residual solid was dissolved with ethanol and filtered again to remove sodium chloride. Finally, the pure ethanol solution of the precursors was achieved.

Fig. 2 Scheme of the acquisition of pure precursors from fructose.

Analytic methods

The analysis of the alcoholysis products was performed on an Agilent GC-7890A gas chromatograph (HP Innowax capillary column 19091N-133N, 30 m × 250 μm × 0.25 μm) with an external standard method. The column temperature was held at 323 K for 2 minutes and then heated via temperature programming at a rate of 283 K min⁻¹ to 523 K, holding for 3 minutes. At the same time, the conversion of substrates before and after condensation was also evaluated on the Agilent GC-7890A gas chromatograph. The conversion of fructose was detected using HPLC (Waters 2695) with a SH1011 column (8.0 × 300 mm, 6 μm particle size, Waters) held at 323 K, using 5 mM H₂SO₄ as the mobile phase (0.5 mL min⁻¹) through an external standard method. The composition of the precursor was analyzed using an elemental analyzer (vario EL III). Because of the viscidity of the condensed product at the mole ratio of 1.1, the substrate was first dissolved with methanol, and then the ¹³C chemical shifts were measured using a nuclear magnetic resonance scanner (Bruker Advance 400 III) with DMSO as the solvent (100 MHz). The relevant calculations of conversion, yield of products and mole ratio were as in the following formulae:

For the conversion calculation, n₀ and n_l are the initial and final mole contents of the feedstock before and after reaction. Meanwhile, n₀ in the mole yield calculation has the same meaning as in (1), and n_i was the mole content of each product respectively. When the mole ratios of the furans and their derivatives were adjusted in the optimized ranges, the ratios were calculated according to (3). n_i represents the mole content of FA, EMF and 5-HMF, and n_j represents the mole content of EL and LA, respectively. Finally, the yield of the precursor was evaluated based on formula (4), in which m_k is the mass of FA, EMF, 5-HMF, EL and LA.

Results and discussion

Effect of different catalysts on fructose alcoholysis

Fructose is a more efficient and selective substrate than other carbohydrates in the synthesis of furan compounds over acid catalysts.²² Thus, we adopted several simple mineral acids to convert fructose into furan compounds and their derivatives, and the results are listed in Table 1. According to the results, Fe₂(SO₄)₃ had an obvious effect on the conversion of fructose into furan compounds and their derivatives. EMF and EL were the main products and the total mole yield was up to 71.2% after the feedstock was completely converted in ethanol. Meanwhile, other furans and their derivatives, for example, FA, 5-HMF and LA were accompanied by alcoholysis with lower mole yields. The catalytic behavior of Al₂(SO₄)₃ was similar to that of Fe₂(SO₄)₃.

Table 1 Conversion of fructose into key platform chemicals over different catalysts^a

Entry	Catalysts	Conversion/%	Mole yield/%					pH	Mole ratio
			FA	EMF	5-HMF	EL	LA
a Reaction conditions: 5% fructose; 413 K; 50 mL ethanol; 1.0 h; 0.2 g catalyst.
1	Fe2(SO4)3	100	0.5	49.9	0.9	21.3	1.0	3.5	2.3
2	Al2(SO4)3	100	0.6	47.6	0.5	19.8	0.6	4.0	2.4
3	H2SO4	100	0.4	5.3	—	53.6	1.0	3.0	0.1
4	ZnSO4	82.3	0.24	7.8	25.3	1.6	—	5.5	20.6
5	NaH2PO4	8.6	0.36	—	8.2	—	0.5	6.0	16.4

---

Usually, fructose is dehydrated to 5-HMF, catalyzed by protons generated from hydrolysis/alcoholysis of Al³⁺ or Fe³⁺. Then, 5-HMF decomposes into EL in the presence of ethanol, catalyzed by the protons.²³ Although the lowest pH value of H₂SO₄ was detected as 3.0 in ethanol, a lower total mole yield of 63.4% of furans and the derivatives were achieved. Moreover, the majority products in the final mixed system were EL and LA (86.1%). This result indicated that a stronger proton acid could be propitious for the dehydration of fructose, forming EL and LA. However, humins were inevitably produced under stronger proton acid conditions, resulting in lower yields of furans and their derivatives. According to the pH value sequence of the different catalysts in ethanol solution, weaker proton acids were less able to dehydrate 5-HMF to LA. Meanwhile, undesired humins were difficult to produce under such reaction conditions. For example, 49.9% and 47.6% mole yields of EMF were achieved respectively over Fe₂(SO₄)₃ (pH = 3.5) and Al₂(SO₄)₃ (pH = 4.0). A lower pH value of the Fe₂(SO₄)₃ ethanol solution was found compared with that of Al₂(SO₄)₃, which might enable Fe₂(SO₄)₃ to alcoholyze more protons to promote the dehydration of fructose to 5-HMF, and its subsequent conversion to EL, a little more efficiently than that of Al₂(SO₄)₃. Therefore, Fe₂(SO₄)₃ was adopted in our following investigation to convert fructose into EMF mixed with EL in ethanol solution.

Effect of temperature on fructose alcoholysis

The reaction was carried out at different temperatures from 373 to 453 K. The results for the conversion of fructose are summarized in Fig. 3. A significant effect of reaction temperature on product distribution has been observed. An increasement in the reaction temperature improved the EL yield as a result of decreasing the EMF yield. This is because the increasement in temperature could also promote the decomposition of EMF to EL. At 373 K, much more 5-HMF and EMF were detected, and this temperature was in favor of 5-HMF etherification to EMF. However, the total mole yields of the furan compounds and their derivatives was only up to 44.6%, not to mention the adjustment of the mole ratios. At 413 K, EMF and EL were observed as the main components of the mixed products, and a very high total mole yield for the furan compounds and their derivatives (73.6%) was obtained. Meanwhile, based on the ideal mole ratios for Aldol condensation in Fig. 1, the mole ratio of furans and their derivatives at 413 K was about 2.3, which was only one step away from 2 : 1. Therefore, the reaction time was optimized next to achieve this idea. Hence, the optimum temperature was set to 413 K for the highest yield of furan compounds and their derivatives.

Fig. 3 Effect of reaction temperature on the product distribution and mole ratio. Reaction conditions: 5% fructose; 50 mL ethanol; 0.2 g Fe₂(SO₄)₃; 1.0 h.

Effect of time on fructose alcoholysis

Based on the optimization of the catalyst and reaction temperature, a typical time profile for the alcoholysis of fructose is depicted in Fig. 4. High fructose conversion of 100% was achieved within only 0.25 h in ethanol solution over Fe₂(SO₄)₃ catalyst. It could be seen that the reaction time had little effect on the total mole yields of furans and their derivatives. Moreover, EMF and EL were the predominant products when reaction time was prolonged from 0.25 h to 6.0 h. Water could be produced during dehydration of fructose, conversion of ethanol to diethyl ether, and EL production by EMF dehydration, but it was proposed that tiny amounts of water might not play an essential role in hindering further hydrolysis of furans.²⁴ Thus, the highest total mole yield is achieved after fructose is alcoholyzed for 2.0 h, and the value could reach 91.6%. Simultaneously, the value of the mole ratio (1.9) in this case was located between 1 and 2, which met the initial objective for synthesis of long carbon chain precursors. Therefore, 2.0 h was the optimal reaction time. However, undesired humins and other byproducts appeared during a prolonged alcoholysis process, resulting in the decrease of the total mole yield of furans and their derivatives.²⁵

Fig. 4 One-pot alcoholysis of fructose at different times. Reaction conditions: 5% fructose; 413 K; 50 mL ethanol; 0.2 g Fe₂(SO₄)₃.

Effect of water percentage on fructose alcoholysis

Water is not only a solvent in this system, but also a product of the alcoholysis procedure. Previous investigations had suggested that the presence of enough water would also enhance the rehydration of 5-HMF to LA and formic acid.²⁶ Thus, the effect of different water contents in this system was investigated and the results are shown in Fig. 5.

Fig. 5 Effect of water on the distribution of alcoholysis products. Reaction conditions: 5% fructose; 413 K; 2.0 h; 0.2 g Fe₂(SO₄)₃.

Obviously, a great deal of furans were detected in the ethanol–water system when the water content was only 10%, and this result was in accord with the previous investigation.²⁴ 5-HMF was esterified rapidly after dehydration of fructose, resulting in the large amount of EMF. However, esterification was inhibited sharply when the water content increased up to 40%, leading to the increase of 5-HMF in system. Interestingly, an increase of LA occurred only when the water content was at a certain percentage (60%). Moreover, fructose conversion decreased gradually when the water content was increased. According to these results, the conclusion is that water not only shifts the equilibrium to products but also affects the rate of the hydrolysis reaction.

Because of the disproportionality of the furans and their derivatives, the mole ratio was not located in the ideal range (1–2) due to the presence of plenty of furans. However, the large amount of LA produced after adding 60% water was beneficial to reduce the value of the mole ratio. Although the mole ratios could be adjusted by changing the percentage of water, the total mole yields of furan and its derivatives were obviously lower than in the pure ethanol system, as shown by the results of alcoholysis in ethanol (91.6%) and water (37.7%).

Condensation reactions in the ethanol system

According to the mechanism of condensation and a previous investigation,¹⁶ NaOH was adopted to catalyze the conversion of furans and their derivatives to form long carbon chain compounds. Here, we presented the condensation results of the furans and their derivatives with mole ratios of 1.9 and 1.1, which were achieved after fructose was alcoholyzed in pure ethanol for 2.0 and 3.0 h over Fe₂(SO₄)₃. As reported in ref. 20, the condensed products at the mole ratio of 1 : 1 could be further condensed via self-Michael additions, as well as formation of other Aldol adducts, resulting in a furanic-keto acid polymer over strong base catalyst for a long time (48 h).²⁰ We thought that the mole ratio had something to do with the self-Michael additions and the unnecessarily long reaction time. That is, not only reaction time but also mole ratio effected the final condensed product components. After that, condensation was conducted in the ethanol system directly after fructose was alcoholyzed completely. In these two systems, the mole yields of FA, 5-HMF, EMF, LA and EL were 0.7%, 0.3%, 45.9%, 0.9% and 41.9% when the mole ratio was 1.1, and when the mole ratio was 1.9, they were 0.6%, 0.8%, 58.4%, 0.6% and 31.2%, respectively.

The results after condensation showed that fructose was highly efficiently converted into furans mixed with their derivatives. Traces of furans and EL were detected in the ethanol system (Fig. 6 and 7) after reaction. That is to say, the condensation could be more efficiently conducted in the alcohol system compared to previous investigations.¹⁶ High yield values of the final precursors of 89.1% and 91.2% were achieved with mole ratios of 1.1 and 1.9. In fact, the two main ideal reaction processes are as shown in Fig. 1. Precursor I, precursor II and precursor III were selectively produced, respectively, by changing the mole ratio of EMF and EL in the same system. Although other furan chemicals and derivatives like FA, 5-HMF and LA are co-produced simultaneously, all of them were also the feedstock of condensation.^12,16 Moreover, our results suggested that the two reaction routes occurred at the same time in spite of the mole ratio being 1.1. Under this condition, the final precursor was a mixture, and this conclusion was supported by ¹³C NMR. Comparing the chemical shifts of the synthesized precursors with theoretical precursors (Fig. S2–S7†) revealed that cross condensation and different feedstocks are the factors having an effect on the precursor components.²⁷ Meanwhile, further condensed products formed by self-Michael additions were not observed. This suggested that a shorter condensation time could not induce self-Michael additions. The carbon elongation process was only Aldol condensation at the mole ratio of 1.1 with a shorter reaction time (4.0 h).

Fig. 6 Conversion of furans and its derivatives, and the yield of precursor after condensation. Reaction conditions: 373 K; 4.0 h; 0.4 g NaOH; furans and their derivatives are alcoholyzed from fructose in pure ethanol for 3.0 h over Fe₂(SO₄)₃ and the mole ratio was 1.1.

Fig. 7 Conversion of furans and its derivatives, and the yield of precursor after condensation. Reaction conditions: 373 K; 4.0 h; 0.4 g NaOH; furans and their derivatives are alcoholyzed from fructose in pure ethanol for 2.0 h over Fe₂(SO₄)₃ and the mole ratio was 1.9.

However, the condensation routes based on the mole ratio of 1.9 were composed of Aldol condensation and self-Michael additions. The ¹³C NMR results showed that mixed precursors were achieved when the mole ratio of furans and derivatives was 1.9 (Fig. S8†). A polymer was formed under appropriate higher mole ratio conditions even with a shorter reaction time. Therefore, we could conclude that the mole ratio of furans and their derivatives affected the condensation method and controlled the length of the carbon chains of the precursors.

Based on the results of the element analysis of the precursors (Table 2), the final precursor^a was composed of only C, H, O and the molecular formula was C_5.1H_6.8O_2.0 in short form. In fact, the theoretical molecular formulas of precursors I, II and III were C₁₅H₂₁O₅, C₁₅H₂₁O₅ and C₂₃H₃₀O₈, respectively. In order to compare the differences in components of actual and theoretical products, the theoretical molecular formulas were changed to C_6.0H_8.5O_2.0, C_6.0H_8.5O_2.0 and C_5.8H_7.5O_2.0 according to the oxygen content in the actual condensed product. This result indicated that many more –CH₂CH₃ functional groups in the synthesized esters might be dissociated from the precursors during condensation, resulting in the decrease of carbon and hydrogen number. The elemental content results of precursor^b showed that the molecular formula was C_5.7H_5.8O_2.0 in short form. The relative content of C in precursor^b was higher than that in precursor^a (C_5.1H_6.8O_2.0). Meanwhile, the lower relative content of H in precursor^b indicated that the polymer was formed by self-Michael additions, resulting in longer carbon chains (Fig. S8†).²⁰ Compared with the theoretical molecular formulas of precursors I, II (C_6.0H_8.4O_2.0) and III (C_5.8H_7.5O_2.0), it also suggested that some –CH₂CH₃ functional groups could be dissociated from the precursors during condensation under strong base conditions.

Table 2 Elemental composition analysis of precursors^a

Element content/%
Precursor^a	C	H	O	Precursor^b	C	H	O
a Precursor^a and precursor^b were produced when the mole ratio of furans and their derivatives was 1.1 and 1.9, respectively.
	60.7	6.8	32.3		64.6	5.4	30.0

---

Conclusions

Long carbon biofuel precursors were efficiently synthesized via condensation from fructose. Furans and their derivatives could be co-produced at a high yield of 91.6% after fructose alcoholysis. Reaction time, water content and temperature had obvious effects on the yields and mole ratio adjustment of furans and derivatives. Condensation could occur when the mole ratio was adjusted in the range of 1 to 2. Meanwhile, –CH₂CH₃ functional groups in the ester dissociated during condensation under strong base conditions. Two Aldol condensation processes occurred simultaneously, even if the mole ratio was 1.1, resulting in the complex composition of the final precursors. However, the condensation route based on the mole ratio of 1.9 was composed of Aldol condensation and self-Michael additions, resulting in polymer formation with different longer carbon chains.

Acknowledgements

This work was supported by grants from the National High-Tech Research and Development Program (No. 2012AA101806), the National Key Basic Research and Development Plan (No. 2012CB215304), the National Natural Science Foundation of China (No. 51376185) and the Key Laboratory of Renewable Energy Open Foundation (No. y407j51001).

References

1. C. J. Chuck and J. Donnelly, Appl. Energy, 2014, 118, 83–91.
2. Q. Zhang, T. Jiang, B. Li, T. J. Wang, X. H. Zhang, Q. Zhang and L. L. Ma, ChemCatChem, 2012, 4, 1084–1087.
3. K. Y. Chen, M. Tamura, Z. L. Yuan, Y. Nakagawa and K. Tomishige, ChemSusChem, 2013, 6, 613–621.
4. N. Li and G. W. Huber, J. Catal., 2010, 270, 48–59.
5. G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106, 4044–4098.
6. M. J. Climent, A. Corma and S. Iborra, Green Chem., 2014, 16, 516–547.
7. D. J. Liu and E. Y. X. Chen, ACS Catal., 2014, 4, 1302–1310.
8. E. F. Iliopoulou, Curr. Org. Synth., 2010, 7, 587–598.
9. O. O. James, S. Maity, L. A. Usman, K. O. Ajanaku, O. O. Ajani, T. O. Siyanbola, S. Sahu and R. Chaubey, Energy Environ. Sci., 2010, 3, 1833–1850.
10. J. X. Long, B. Guo, X. H. Li, Y. B. Jiang, F. R. Wang, S. C. Tsang, L. F. Wang and K. M. K. Yu, Green Chem., 2011, 13, 2334–2338.
11. J. X. Long, G. Bin, J. J. Teng, Y. H. Yu, L. F. Wang and X. H. Li, Bioresour. Technol., 2011, 102, 10114–10123.
12. Y. B. Huang, Z. Yang, J. J. Dai, Q. X. Guo and Y. Fu, RSC Adv., 2012, 2, 11211–11214.
13. A. Corma, O. de la Torre, M. Renz and N. Villandier, Angew. Chem., Int. Ed., 2011, 50, 2375–2378.
14. J. Xin, S. Zhang, D. Yan, O. Ayodele, X. Lu and J. Wang, Green Chem., 2014, 16, 3589.
15. M. Balakrishnan, E. R. Sacia and A. T. Bell, ChemSusChem, 2014, 7, 1078–1085.
16. L. Faba, E. Diaz and S. Ordonez, ChemSusChem, 2014, 7, 2816–2820.
17. G. Li, N. Li, J. Yang, A. Wang, X. Wang, Y. Cong and T. Zhang, Bioresour. Technol., 2013, 134, 66–72.
18. B. Liu, Z. Zhang, K. Huang and Z. Fang, Fuel, 2013, 113, 625–631.
19. R. Liu, J. Chen, X. Huang, L. Chen, L. Ma and X. Li, Green Chem., 2013, 15, 2895–2903.
20. A. S. Amarasekara, T. B. Singh, E. Larkin, M. A. Hasan and H. J. Fan, Ind. Crops Prod., 2015, 65, 546–549.
21. J. P. Lange, E. van der Heide, J. van Buijtenen and R. Price, ChemSusChem, 2012, 5, 150–166.
22. X. Hu, L. Wu, Y. Wang, Y. Song, D. Mourant, R. Gunawan, M. Gholizadeh and C. Z. Li, Bioresour. Technol., 2013, 133, 469–474.
23. F. Yang, J. Fu, J. Mo and X. Lu, Energy Fuels, 2013, 27, 6973–6978.
24. L. Peng, L. Lin, J. Zhang, J. Shi and S. Liu, Appl. Catal., A, 2011, 397, 259–265.
25. H. Wang, T. Deng, Y. Wang, Y. Qi, X. Hou and Y. Zhu, Bioresour. Technol., 2013, 136, 394–400.
26. A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411–2502.
27. H. E. Gottlieb, V. Kotlyar and A. Nudelman, J. Org. Chem., 1997, 62, 7512–7515.

---

Footnote

† Electronic supplementary information (ESI) available: See DOI: 10.1039/c5ra07448g

---