Conversion of carbohydrate biomass to methyl levulinate with Al 2(SO4)3 as a simple, cheap and efficient catalyst

Article abstract of DOI:10.1016/j.catcom.2014.02.021

Al₂(SO₄)₃ was developed as a simple and efficient catalyst for the synthesis of methyl levulinate (MLE) from carbohydrate biomass including fructose, glucose, mannose, sucrose, cellobiose, starch, and cellulose. The maximum MLE yield of 64% from glucose could be obtained at 160 C for 150 min. Important roles of Al^{3 +} and Br?nsted acid sites generated by the hydrolysis/methanolysis of Al ^{3 +} were elucidated. The reaction process was revealed using in situ real-time attenuated total reflection infrared spectroscopy. Al ₂(SO₄)₃ can be regenerated easily and reused at least four times without the loss of activity.

Full text of DOI:10.1016/j.catcom.2014.02.021

Catalysis Communications 50 (2014) 13–16
Contents lists available at ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Short Communication
Conversion of carbohydrate biomass to methyl levulinate with Al2(SO4)3
as a simple, cheap and efficient catalyst
a
b
b
Lipeng Zhou a, Haojie Zou a, Jiaxu Nan a, Lin Wu a, Xiaomei Yang a, , Yunlai Su , Tianliang Lu , Jie Xu
⁎
a
College of Chemistry and Molecular Engineering, Zhengzhou University, 100 Kexue Road, Zhengzhou 450001, PR China
b
State Key Laboratory of Catalysis, Dalian National Laboratory for Clean Energy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 30 January 2014
Received in revised form 20 February 2014
Accepted 21 February 2014
Available online 27 February 2014
Al2(SO4)3 was developed as a simple and efficient catalyst for the synthesis of methyl levulinate (MLE) from
carbohydrate biomass including fructose, glucose, mannose, sucrose, cellobiose, starch, and cellulose. The
maximum MLE yield of 64% from glucose could be obtained at 160 °C for 150 min. Important roles of Al3+ and
Brønsted acid sites generated by the hydrolysis/methanolysis of Al3+ were elucidated. The reaction process
was revealed using in situ real-time attenuated total reflection infrared spectroscopy. Al2(SO4)3 can be
regenerated easily and reused at least four times without the loss of activity.
Keywords:
Carbohydrate biomass
Glucose
© 2014 Elsevier B.V. All rights reserved.
Al2(SO4)3
Methyl levulinate
1. Introduction
from literature that Lewis acids are effective for isomerization of aldose
[23,24] while dehydration of fructose to LA or its esters is catalyzed by
With the gradual consumption of fossil resources, great effort is paid
to convert abundant and renewable biomass resources into chemicals
and fuels. As one of the most important chemicals derived from biomass
resources, levulinate esters showed numerous potential applications
either in flavoring and fragrance industries or as a blending component
in biodiesel [1,2]. Additionally, the ester and carbonyl functional groups
also enable levulinate esters to be used as substrates for various kinds
of condensation and addition reactions in organic chemistry [3]. Produc-
tion of levulinate esters from cheap, abundant, and renewable biomass
resources has attracted much attention in recent years. For example,
levulinate esters were obtained from direct conversion of cellulosic
biomass [4–8] or from platform compounds derived from cellulose
and hemicellulose, such as glucose [9,10], levulinic acid (LA) [11],
5-chloromethylfurfural [12,13], and furfuralcohol [14].
Various acid catalysts including mineral acids (HCl, H2SO4, etc.)
[4,10] and solid acids (sulfated metal oxides [9], heteropolyacids [2,5],
sulfonic resins [15,16], zeolites [9], etc.) were employed for conversion
of carbohydrates to LA or its esters. In general, carbohydrates are first
dehydrated to 5-hydroxymethylfurfural (HMF) or its derivatives,
which are then rehydrated to LA or its esters [9,17,18]. It is believed
that the generation of HMF or its derivatives from fructose is much
easier than from glucose [19–22]. With glucose as starting material, a
catalyst being able to isomerize glucose to fructose might show good
performance to obtain high yield of LA or its esters. It can be concluded
Brønsted acids [9,25–27]. Accordingly, a catalyst combined both Lewis
and Brønsted acidity should be appropriate for transformation of
glucose to LA or its esters. Similar strategy was utilized in conversion
of carbohydrates such as cellulose, sucrose, corn starch, or glucose. For
example, Ya'aini and coworkers used hybrid catalyst Cr/HY to catalyze
glucose conversion in water to LA; 62% of LA yield was obtained at
160 °C for 180 min [28]. Tominaga et al. employed mixed-acid systems
consisting In(OTf)3 and p-toluenesulfonic acid as Lewis and Brønsted
acids to synthesize methyl levulinate (MLE) from cellulose in methanol;
MLE yield reached 70% at 180 °C for 5 h, which was higher than In(OTf)3
(42%) or p-toluenesulfonic acid (20%) being solely used under the same
reaction conditions [6]. Recently, Mao and Wu et al. studied the co-
catalysis effect of Al salts and acid ionic liquid on the conversion of
sucrose or corn starch to MLE. It was found that the combination of Al
salts and acid ionic liquid was efficient for the formation of MLE [29,30].
In the present work, single component Al2(SO4)3 was developed as
catalyst for the conversion of various carbohydrates to MLE in methanol
at the absence of any co-catalyst. The catalytic mechanism of Al2(SO4)3
in the conversion of carbohydrate to MLE was studied.
2. Results and discussion
2.1. Glucose conversion to MLE catalyzed by various acid catalysts
Various catalysts were tested in the conversion of glucose to MLE in
methanol (Table 1). MLE yield was low when mineral acid such as
H2SO4 or HCl was used as catalyst although glucose was completely
⁎
Corresponding author. Tel.: +86 371 67781780; fax: +86 371 67766076.
E-mail address: yangxiaomei@zzu.edu.cn (X. Yang).
http://dx.doi.org/10.1016/j.catcom.2014.02.021
1566-7367/© 2014 Elsevier B.V. All rights reserved.

14
L. Zhou et al. / Catalysis Communications 50 (2014) 13–16
Table 1
Glucose conversion to methyl levulinate over various catalystsa.
Entry
Catalyst
Glucose conversion (%)
Methyl levulinate yield (%)
pH of methanol solution
1
No
H2SO4
H2SO4
HCl
H-USY
H-beta
H-MOR
H-ZSM-5
57
100
–
0
13
73
12
22
7
1
0.3
6
15
0
5
7
4
12
2
64
18
4
13
71
6.95
0.03
0.03
0.03
–
2
3b
4
100
100
100
100
100
100
100
74
5
6
7
8
–
–
–
9
Amberlyst 15
SnSO4
–
10
11
12
13
14
15
16
17
18
19
20
21
1.82
5.51
0.33
4.79
5.21
1.62
4.83
0.48
0.96
1.27
–
MnSO4·H2O
Fe2(SO4)3
CoSO4·7H2O
NiSO4·6H2O
CuSO4·5H2O
ZnSO4·7H2O
Al2(SO4)3·18H2O
AlCl3·6H2O
Al(NO3)3·9H2O
Al2(SO4)3·18H2O + NaOHc
99
100
99
98
100
100
100
100
100
100
c
Al2(SO4)3·18H2O + H2SO4
–
a
Glucose (1.72 mmol), catalyst (0.68 mmol, based on Men+ or H+; or 0.1 g of solid catalyst), methanol (15 mL), 160 °C, 150 min, 0.1 MPa N2.
Identical weight of fructose as substrate.
b
c
Both n(H+)/n(Al3+) and n(NaOH)/n(Al3+) are equal to 1.0.
converted (Table 1, entries 2 and 4). Zeolites, such as H-β, H-MOR, or H-
MLE. Al2(SO4)3 can afford Lewis acid sites from Al3+ ions and Brønsted
acid sites (protons) generated from the hydrolysis/methanolysis of Al3+
ions, so the conversion of glucose to MLE can proceed smoothly. Metal
salts with Lewis and Brønsted acidity acting as catalysts for conversion
of carbohydrates to levulinate ester were also observed by other
researchers [6,8,33,34]. For this reaction, another concern is the
formation of undesired dimethyl ether due to the acid-catalyzed inter-
molecular dehydration of methanol. The acidity of the system is not
very strong, so very small quantity of methanol (less than 3%) was
converted to dimethyl ether.
Formation of MLE from glucose catalyzed by Al2(SO4)3 is sensitive
to reaction conditions such as the amount of catalyst and substrate,
reaction temperature, and reaction time (Supporting information,
Table S1). The maximum MLE yield of 64% was obtained with
glucose/Al2(SO4)3 of 5 in mole at 160 °C for 150 min.
ZSM-5, and sulfonic resin of Amberlyst 15 exhibited lower MLE yield
than mineral acids (Table 1, entries 6–9). Only H-USY gave higher
MLE yield than mineral acids (Table 1, entry 5). When fructose was
used as substrate, however, 73% of MLE yield was obtained with
H2SO4 as catalyst (Table 1, entry 3). These results indicate that strong
Brønsted acid can efficiently catalyze fructose to MLE, but it is inactive
for the conversion of glucose to MLE due to its lower catalytic ability
to isomerize glucose to fructose [31]. For H-USY zeolite, some extra-
framework Al species were generated during preparation process
through dealumination of Y with high-temperature vapor, which
could act as Lewis acid sites for the isomerization of glucose to fructose
[32]. So, H-USY zeolite gave relatively higher MLE yield than the other
Brønsted acids.
As mentioned above, metal ions could act as Lewis acid and show
good catalytic performance in the isomerization of glucose to fructose.
Several metal sulfates were employed in the present work to catalyze
the conversion of glucose to MLE. The results are listed in Table 1. MLE
yields were low with most of the metal sulfates except Al2(SO4)3, and
64% of MLE yield could be obtained with Al2(SO4)3 as catalyst
(Table 1, entry 17). pH measurement indicated that pH value of the
methanol solution of Al2(SO4)3 was lower than the other metal sulfates
except Fe2(SO4)3. The above results suggest that both Lewis and
Brønsted acidity of catalyst might affect the formation of MLE. Further-
more, other Al3+ salts such as AlCl3 or Al(NO3)3 gave much lower MLE
yields than Al2(SO4)3 (Table 1, entries 18 and 19). Considerable amount
of etherification and acetalization products from HMF and methanol
was observed with AlCl3 or Al(NO3)3 as catalyst. The difference in
catalytic performance for these Al3+ salts might be caused by the differ-
ent Brønsted acidity of their methanol solution. The pH value of
Al2(SO4)3 methanol solution is the lowest among these Al3+ salts,
which might enable Al2(SO4)3 to promote the dehydration of fructose
to HMF and its subsequent conversion to MLE more efficiently than
the others. The importance of Brønsted acidity was also proved by the
following experiments. When NaOH was added to the reaction system
containing Al2(SO4)3 (Table 1, entry 20), the yield of MLE was reduced
drastically for the neutralization of the Brønsted acid sites; whereas
the yield of MLE was increased when H2SO4 was introduced due to
the presence of more Brønsted acid sites (Table 1, entry 21). On the
other hand, anions probably also play some role for the formation of
MLE. Combination of Al3+ with SO24− showed peculiar selectivity for
In situ real-time attenuated total reflection infrared spectroscopy
was used to reveal the reaction process of glucose to MLE catalyzed by
Al2(SO4)3 (Supporting information, Fig. S1). According to the results,
the reaction pathway of glucose to MLE was proposed (Fig. 1). First,
glucose was isomerized to fructose under the catalysis of Lewis acid
sites (Al3+). Then, fructose was dehydrated to HMF catalyzed by
protons generated from hydrolysis/methanolysis of Al3+. Finally, HMF
was decomposed to MLE in the presence of methanol catalyzed by
protons.
For practical application, the reusability and stability of the catalyst
are crucial besides the activity and selectivity. Therefore, the reusability
of Al2(SO4)3 was investigated and the results are shown in Fig. 2. After
five reaction runs, the yield of MLE was still comparable to that over
fresh catalyst. It suggests that Al2(SO4)3 is a stable, reusable, and
efficient catalyst for the transformation of glucose to MLE in methanol.
In addition, MLE in the reaction mixture can be separated easily through
distillation [9]. The detail isolation procedure for MLE was given in the
Supporting information (Scheme S1).
2.2. Production of MLE from other saccharides catalyzed by Al2(SO4)3
Formation of MLE from methyl α-D-glucopyranoside and other
saccharides catalyzed by Al2(SO4)3 is presented in Table 2. Under the
same reaction conditions, methyl α-D-glucopyranoside gave 45% of
MLE yield (Table 2, entry 2), indicating that methyl glucopyranoside
as an intermediate can be easily converted to MLE. 49% of MLE yield

L. Zhou et al. / Catalysis Communications 50 (2014) 13–16
15
CH OH
2
O
O
O
H
H
OR
H
H
OH
H
OR
+
OH
OH
O
H
OH
H+
ROH
Lewis acid site Al3+
O
CH OH
2
OH
H+
HOH C
CHO
2
O
H
OH
H
CH OH
2
H
OH
Fig. 1. Reaction pathway of glucose to levulinate ester.
was obtained with fructose as substrate, which is lower than that with
glucose as substrate (Table 2, entry 3). According to the results reported
in literatures [9,17,18], fructose was more easily transformed to MLE
than glucose. So, we speculated whether the reaction time of 150 min
was too long for fructose conversion, which might cause the further
conversion of MLE and decrease the yield of MLE. When the reaction
time was shortened to 90 min, the MLE yield became much lower
(Table 2, entry 4). Thus, other reason should account for the low yield
of MLE for fructose. During glucose conversion, fructose was generated
gradually from the isomerization of glucose and the concentration of
fructose was kept at low level all the time. When fructose was used
directly as substrate, however, the high concentration of fructose
might induce the formation of byproducts such as humins [15], which
led to the decrease of MLE yield. Nevertheless, MLE yield was much
higher over sole Brønsted acid catalyst at identical fructose concentra-
tion. As shown in Table 1, entry 3, as high as 73% of MLE yield was
obtained over H2SO4. So, it can be deduced that Lewis acid sites primar-
ily induce the formation of byproducts. This conclusion agrees with that
reported by Weingarten et al. who proposed that Lewis acid sites were
principally accountable for the formation of humin byproducts during
dehydration of xylose [35]. MLE can also be obtained from mannose
with the yield of 46% (Table 2, entry 5). The above results suggest that
Al2(SO4)3 can catalyze not only the aldose–ketose isomerization
between glucose and fructose but also the epimerization between
glucose and mannose. In this case, Al2(SO4)3 behaves like Sn-β zeolite
which can catalyze the isomerization among glucose, fructose, and
mannose, because equilibrium mixture of glucose–fructose–mannose
can be obtained at sufficiently long reaction times for whatever the
starting sugar [36]. For sucrose that contains one glucose and one
fructose units per molecule, the yield of MLE fell between those from
glucose and fructose. MLE yield from cellobiose was slightly lower
than from sucrose due to the difficult hydrolysis/methanolysis of β-
glycosidic bonds. Starch as a polysaccharide containing α-glycosidic
bonds gave higher MLE yield than cellobiose and sucrose, indicating
that Al2(SO4)3 in methanol can effectively depolymerize α-glycosidic
bonds. However, cellulose gave only 19% of MLE yield under the same
reaction conditions (Table 2, entry 9). The yield of MLE was still
low even with prolonging the reaction time to 300 min. Considerable
amount of cellulose remained unconverted. The poor reactivity
of cellulose is likely due to the ineffective depolymerization and
decrystallization in methanol at 160 °C [33]. Higher temperature was
often required for cellulose conversion [4–7,9]. The MLE yield was im-
proved to 44% when the reaction temperature was elevated to 180 °C
(Table 2, entry 11).
3. Conclusions
Al2(SO4)3 was found to be an efficient and recyclable catalyst for the
conversion of glucose to MLE in methanol under comparably mild
conditions (160 °C, 150 min). Al2(SO4)3 could provide Lewis and
Brønsted acid site through Al3+ and the hydrolysis/methanolysis of
Al3+ at the same time, which related to the isomerization of glucose
to fructose and dehydration of fructose to MLE, respectively. Fructose,
mannose, sucrose, cellobiose, and starch can also be converted to MLE
effectively at 160 °C for 150 min, while higher reaction temperature
80
70
60
50
40
30
20
10
0
Table 2
Yields of methyl levulinate obtained from methyl α-D-glucopyranoside and other
saccharidesa.
Entry
Substrate
Reaction
time (min)
Yield of methyl
levulinate (%)
1
2
3
4
5
6
7
8
Glucose
Methyl α-D-glucopyranoside
Fructose
Fructose
Mannose
Sucrose
Cellobiose
Starch
α-Cellulose
α-Cellulose
α-Cellulose
150
150
150
90
150
150
150
150
150
300
300
64
45
49
39
46
55
50
57
19
23
44
9
1
2
3
4
5
10
11b
Runs
a
Reaction conditions: substrate (0.34 g), Al2(SO4)3 (0.34 mmol), methanol (15 mL),
160 °C, 0.1 MPa N2.
Fig. 2. Reusability of Al2(SO4)3 in the transformation of glucose to methyl levulinate in
methanol (see Supporting information for experimental details).
b
Reaction temperature is 180 °C.

16
L. Zhou et al. / Catalysis Communications 50 (2014) 13–16
[14] J. Lange, W.D. van de Graaf, R.J. Haan, ChemSusChem 2 (2009) 437–441.
[15] P.A. Son, S. Nishimura, K. Ebitani, React. Kinet. Mech. Catal. 106 (2012) 185–192.
[16] J. Zhang, E. Weitz, ACS Catal. 2 (2012) 1211–1218.
[17] R. Weingarten, W.C. Conner, G.W. Huber, Energy Environ. Sci. 5 (2012) 7559–7574.
[18] S. Van de Vyver, J. Thomas, J. Geboers, S. Keyzer, M. Smet, W. Dehaen, P.A. Jacobs, B.F.
Sels, Energy Environ. Sci. 4 (2011) 3601–3610.
(180 °C) is needed for cellulose to obtain considerable amount of MLE
due to the difficult depolymerization of its glycosidic bonds.
Acknowledgments
[19] C.M. Lew, N. Rajabbeigi, M. Tsapatsis, Ind. Eng. Chem. Res. 51 (2012) 5364–5366.
[20] T. Ståhlberg, S. Rodriguez-Rodriguez, P. Fristrup, A. Riisager, Chem. Eur. J. 17 (2011)
1456–1464.
[21] J. Wang, J. Ren, X. Liu, J. Xi, Q. Xia, Y. Zu, G. Lu, Y. Wang, Green Chem. 14 (2012)
2506–2512.
[22] G. Yong, Y. Zhang, J.Y. Ying, Angew. Chem. Int. Ed. 47 (2008) 9345–9348.
[23] Y. Román-Leshkov, M. Moliner, J.A. Labinger, M.E. Davis, Angew. Chem. Int. Ed. 49
(2010) 8954–8957.
[24] C.M. Lew, N. Rajabbeigi, M. Tsapatsis, Microporous Mesoporous Mater. 153 (2012)
55–58.
We are grateful to the National Natural Science Foundation of China
(Grant: U1304209; J1210060) and the Undergraduate Innovation
Education Project of Zhengzhou University for the financial support
(Grant: 2013sjxm008).
Appendix A. Supplementary data
[25] L. Zhou, L. Wu, H. Li, X. Yang, Y. Su, T. Lu, J. Xu, J. Mol. Catal. A (2014), http://dx.doi.org/
10.1016/j.molcata.2014.01.017.
[26] S. Saravanamurugan, A. Riisager, Catal. Commun. 17 (2012) 71–75.
[27] G. Yang, E.A. Pidko, E.J.M. Hensen, J. Catal. 295 (2012) 122–132.
[28] N. Ya'aini, N.A.S. Amin, S. Endud, Microporous Mesoporous Mater. 171 (2013)
14–23.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.catcom.2014.02.021.
References
[1] F. Su, L. Ma, D. Song, X. Zhang, Y. Guo, Green Chem. 15 (2013) 885–890.
[2] J. Zhang, S. Wu, B. Li, H. Zhang, ChemCatChem 4 (2012) 1230–1237.
[3] E.S. Olson, M.R. Kjelden, A.J. Schlag, R.K. Sharma, ACS Symp. Ser. 784 (2001) 51–63.
[4] C. Chang, G. Xu, X. Jiang, Bioresour. Technol. 121 (2012) 93–99.
[5] F. Rataboul, N. Essayem, Ind. Eng. Chem. Res. 50 (2011) 799–805.
[6] K. Tominaga, A. Mori, Y. Fukushima, S. Shimada, K. Sato, Green Chem. 13 (2011)
810–812.
[7] X. Wu, J. Fu, X. Lu, Carbohydr. Res. 358 (2012) 37–39.
[8] L. Peng, L. Lin, J. Zhang, J. Zhuang, B. Zhang, Y. Gong, Molecules 15 (2010)
5258–5272.
[29] H. Mao, H. Huang, Chem. Ind. Eng. Prog. 31 (2012) 1816–1819.
[30] M. Wu, H. Mao, J. Zhu, Mod. Chem. Ind. 32 (2012) 56–58.
[31] M. Moliner, Y. Román-Leshkov, M.E. Davis, Proc. Natl. Acad. Sci. U. S. A. 107 (2010)
6164–6168.
[32] S. Saravanamurugan, M. Paniagua, J.A. Melero, A. Riisager, J. Am. Chem. Soc. 135
(2013) 5246–5249.
[33] Y. Yang, C. Hu, M.M. Abu-Omar, Bioresour. Technol. 116 (2012) 190–194.
[34] C.B. Rasrendra, I.G.B.N. Makertihartha, S. Adisasmito, H.J. Heeres, Top. Catal. 53
(2010) 1241–1247.
[35] R. Weingarten, G.A. Tompsett, W.C. Conner, G.W. Huber, J. Catal. 279 (2011)
174–182.
[9] L. Peng, L. Lin, H. Li, Q. Yang, Appl. Energy 88 (2011) 4590–4596.
[10] L. Peng, L. Lin, H. Li, Ind. Crop. Prod. 40 (2012) 136–144.
[11] D.R. Fernandes, A.S. Rocha, E.F. Mai, C.J.A. Mota, V.T. da Silva, Appl. Catal. A 425–426
(2012) 199–204.
[36] M.S. Holm, Y.J. Pagán-Torres, S. Saravanamurugan, A. Riisager, J.A. Dumesic, E.
Taarning, Green Chem. 14 (2012) 702–706.
[12] M. Mascal, E.B. Nikitin, Angew. Chem. Int. Ed. 47 (2008) 7924–7926.
[13] M. Mascal, E.B. Nikitin, Green Chem. 12 (2010) 370–373.