Facile and high-yield synthesis of methyl levulinate from cellulose

Article abstract of DOI:10.1039/c7gc02883k

Efficient production of chemicals from cellulose provides sustainable routes for the utilization of natural renewable resources to meet the requirements of human society. Herein, we reported a highly efficient and simple metal salt catalyst, Al₂(SO₄)₃, for cellulose conversion to methyl levulinate (ML) under microwave conditions. A highest ML yield of 70.6% was obtained at 180 °C within a very short time of 40 min. The introduction of water could reduce humin/coke formation and solvent consumption, and could also switch the reaction pathway via the more reactive intermediate glucose. Kinetic and mechanistic studies of the subreactions showed that both cellulose hydrolysis and alcoholysis pathways were involved in the cellulose conversion to ML, with the former as the main route in the presence of water. The Lewis acid species [Al(OH)_x(H₂O)_y]ⁿ⁺ and the Br?nsted acid species H⁺, generated by in situ hydrolysis of Al₂(SO₄)₃, were responsible for the reaction conversions. The reaction with microwave heating showed accelerated reaction rates of 25 times the reaction with conventional oil heating, and even more times for the rates of glucose and methyl glucoside (MG) dehydration, resulting in higher reaction selectivity toward ML production. The catalyst was also successfully recycled and applied to the conversion of cellulose to other alkyl levulinates, as well as the conversion of raw biomass to ML with high yields. The homogeneous nature of Al₂(SO₄)₃, together with its high efficiency and excellent recyclability, make it a potential catalyst for the large-scale production of ML in industry.

Full text of DOI:10.1039/c7gc02883k

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: Y. Huang, T. Yang,
Y. Lin, Y. Zhu, H. Pan and L. li, Green Chem., 2018, DOI: 10.1039/C7GC02883K.
This is an Accepted Manuscript, which has been through the
Volume 18 Number
7 7 April 2016 Pages 1821–2242
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Accepted Manuscripts are published online shortly after
Cutting-edge research for a greener sustainable future
www.rsc.org/greenchem
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1463-9262
CRITICAL REVIEW
G. Chatel et al.
Heterogeneous catalytic oxidation for lignin valorization into valuable
chemicals: what results? What limitations? What trends?
rsc.li/green-chem

Please do not adjust margins
Green Chemistry
Page 1 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
Facile and high-yield synthesis of methyl levulinate from cellulose
Yao-Bing Huang* Tao Yang, Yu-Ting Lin, Ying-Zhi Zhu, li-Cheng Li, Hui Pan*
Received 00th January 20xx,
Accepted 00th January 20xx
Efficient production of chemicals from cellulose provides sustainable routes for the utilization of natural renewable
resources to meet the requirements of human society. Herein, we reported a highly efficient and simple metal salt catalyst
Al2(SO4)3 for the cellulose conversion to methyl levulinate (ML) under microwave condition. A highest ML yield of 70.6%
was obtained at 180 oC within a very short time of 40 min. The introduction of water could reduce humin/coke formation
and solvent consumption, and can also switch the reaction pathway via the more reactive intermediate glucose. Kinetic
and mechanistic study of the subreactions showed that both cellulose hydrolysis and alcoholysis pathways were existed in
the cellulose conversion to ML, with the former as the main route in the presence of water. The Lewis acid species
[Al(OH)x(H2O)y]n+ and Brønsted acid species H+, generated by in-situ hydrolysis of Al2(SO4)3, were responsible for the
reaction conversions. The reaction with microwave heating showed accelerated reaction rates of 25 times the reaction
with conventional oil heating, and even more times of the rates of glucose and methyl glucoside (MG) dehydration,
resulting in higher reaction selectivity toward ML production. The catalyst was also successfully recycled and applied to the
conversion of cellulose to other alkyl levulinates, as well as the conversion of raw biomass to ML with high yields. The
homogeneous nature of Al2(SO4)3, together with its high efficiency and excellent recyclability, make it a potential catalyst
for the large-scale production of ML in industry.
DOI: 10.1039/x0xx00000x
www.rsc.org/
further to a variety of useful chemicals such as furanics,8-10
acids,11,12 esters,13-15 polyols16,17 and alkanes18,19 which greatly
broaden the cellulosic compound family. One particular
Introduction
In recent years, there has been growing interest in the
utilization of renewable resources to produce energy or
chemicals to meet the demand of human society. One of the
attracting route, for chemical industry, is to convert biomass
(e.g. lignocellulose) to platform molecules that can serve as
starting materials for the production of other value-added
chemicals or liquid fuels.1-3 Of all the components of the
lignocellulose, cellulose attracts considerable attentions due to
its widely available, large annual production and highly
repetitive structural units that may generate simple
depolymerized products.4 Unlike the general surface
modification of cellulose to prepare materials, chemical
conversions of cellulose to valued-added compounds require
the hydrolysis of the glucosidic bonds of cellulose to release
glucose units before the subsequent downstream conversions.
However, the highly crystallinity and three-dimensional
hydrogen-binding networks of cellulose make it reluctant to be
hydrolysed, and thereby leaving it a challenging substrate for
chemical conversion.5 Albeit that, tremendous efforts have
been devoted to the conversions of cellulose to glucose,6,7 and
interest concerns the catalytic conversion of cellulose to alkyl
levulinates like (ML) and ethyl levulinate (EL). These esters are
a kind of short fatty acid esters which can be used as
plasticizing agents, solvents, or as a blending fuel additive.20,21
Apart from that, their reactive ester and carbonyl groups
enable them to be used for the synthesis of several
downstream chemicals or drugs.22
According to the literatures,23,24 both levulinic acid and
furfuryl alcohol are the most reported substrates, which
undergo the esterification and alcoholysis reaction,
respectively, to provide the alkyl levulinates products.
However, these two substrates are also the intermediates of
lignocellulose, which require extra reaction steps (e.g.
hydrolysis, hydrogenation) to be prepared from lignocellulose.
These processes obviously add complexity to the whole
reaction route and reduce the whole economy. Besides, the
associated separation and purify process are also costly.
Therefore, the establishment of new methods to produce alkyl
levulinates directly from easily available and cheaper natural
biopolymer cellulose is academically and economically more
attractive.
a.Jiangsu provincial key lab for the chemistry and utilization of agro-forest biomass,
College of Chemical Engineering, Nanjing Forestry University, Nanjing, China
Email: hyb123@mail.ustc.edu.cn;
Over the past few years, several catalytic systems have
been successfully employed for the conversion of cellulose to
alkyl levulinates. Both soluble and solid catalysts could be
found in the literatures. For example, Fu and Lu et al examined
a series of liquid acids for cellulose methanolysis and found
that H2SO4 and para-toluenesulfonic acid were the most
† Footnotes relating to the title and/or authors should appear here.
Electronic Supplementary Information (ESI) available: [details of any
supplementary information available should be included here]. See DOI:
10.1039/x0xx00000x
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 2 of 13
View Article Online
DOI: 10.1039/C7GC02883K
ARTICLE
Journal Name
efficient catalysts, offering the highest ML yield of 55%.25
Other weak acids such as HCOOH and H3PO4 were inferior for
this conversion. However, the undesired etherification of
methanol was unavoidable in presence of these liquid catalysts
at elevated reaction temperature, which consumed a lot of
solvent. Tominaga et al reported a mixed catalytic system
using 2-naphtalenesulfonic acid and In(OTf)3 for cellulose
conversion and a high yield of 75%, even the highest until now,
was obtained.26 The Lewis acid sites were proposed to
promote the rate-determining step of the intermediated MG
isomerization to generate ML. In consideration of the recycling
and product separation merits of solid catalysts, a few
heterogeneously catalysed systems were also reported over
the catalysts including sulphated metal oxides,13,27,28 Zeolite13
and heteropolyacid salts.29,30 However, the product yields
were relatively much lower as compared to the liquid acid
systems. Since both of cellulose and solid catalyst are insoluble
in alcohols, the mass transfer limitation may be primary factor
for lower efficiency of solid catalysts. Strikingly, Liu and Wang
et al reported the use of mesoporous niobium-based
phosphate (NbP) solid catalyst with proper Brønsted/Lewis
(B/L) ratio for cellulose conversion with high reaction
Scheme 1 Al2(SO4)3 catalyzed cellulose conversion into ML using microwave
heating.
Apart from that, microwave irradiation is a widely used
technology for chemical process. It is a volumetric and
selective dielectric heating, which could greatly accelerate the
reaction rate and reduce the requisite time to finish the
reaction.35 It may also save the energy consumption, and
sometimes reduce undesired side reactions by avoiding the
exposure of the reaction mixture to high temperature for long
time. This technology may just suitable for cellulose
conversion, due to the thermal instability of its downstream
carbohydrate derived products. Herein, we reported a facile
and economically feasible and high-yield synthesis of ML from
cellulose over metal salt catalyst aluminium sulfate under
microwave condition. Different metal salt catalysts were
investigated in the alcoholysis reaction. The influences of
reaction parameters on the product selectivity, as well as the
reaction kinetics, were studied. Critical role of additive water
in the reaction was discussed. Mechanism studies about the
reaction pathway were also proposed.
o
efficiency, offering 56% yield of ML at 180 C for 24h, which
represented a significant advance of heterogeneous system.31
Concerning all the above systems, soluble catalysts averagely
exhibited higher efficiency for cellulose conversion. Besides,
it’s not difficult to find that both Lewis and Brønsted acid sites,
as well as their ratios, are crucial to the final product yields.
These advances may provide fundamental principles for the
design of new catalytic systems for cellulose conversion to
alkyl levulinates.
Experimental
Materials
Microcrystalline cellulose (96%, 20 nm), ML (99%), EL (98%),
butyl levulinate (98%), glucose (98%), MG (98%) and fructose
(99%) were supplied by TCI Chemicals Co. Ltd (Shanghai
China). 5-hydroxymethylfurfural (HMF, 99%) and 5-
methoxymethylfurfural (MMF, 97%) was purchased from
Adamas-beta Inc. (Shanghai, China). All the metal salts were in
the form of crystalline hydrate and purchased from Sinopharm
Chemical Reagent Co., Ltd. (Shanghai, China). The raw biomass
materials (e.g. Bamboo, bagasse and polar) were collected
from local market.
Several reported literatures revealed that metal salts with
proper combination of cation and anion exhibited excellent
Brøntsd/Lewis acidity in alcoholic solutions, offering
inexpensive, simple and efficient methodologies for a variety
of biomass conversions.32-34 Strikingly, our previous studies
demonstrated that aluminium sulfate was such type of catalyst
which provided high reactivity for the alcoholysis of furfuryl
alcohol to alkyl levulinates.35 These results encouraged us to
further apply metal sulfate catalyst to the conversion of more
challenging substrate cellulose. It’s worth noting that Xu et al
had reported the use aluminium sulfate for glucose conversion
to alkyl levulinates.36 They also tested the conversion of
cellulose, but only a moderated ML yield was obtained. Due to
the big differences in of glucose and cellulose in chemical
property, the reaction system for glucose conversion cannot
be directly used for cellulose conversion. Besides, detailed
information about the influences of reaction parameters and
additives on reaction, kinetics of subreactions and the related
mechanism were still lack. Therefore, it’s attractive to establish
a comprehensive study on the metal salt catalysed alcoholysis
of cellulose.
Catalytic reactions
Conversion of cellulose under microwave heating: the
alcoholysis of cellulose was conducted in a 100mL Teflon
sealed microwave reactor. Typically, a mixture of 3 mmol
microcrystalline cellulose, 14 mL MeOH, a certain amount of
H2O and Al2(SO4)3 catalyst was charged into the reactor. Then,
the reactor was sealed and heated to the desired reaction
temperature in 2 min under microwave irradiation. The time
the reactor reached desired temperature was set as the zero
time of the reaction. After the reaction finished, the reaction
was quenched by placed the reactor in an ice cool water bath
to stop the reaction. All experiment was duplicated and the
average data was recorded.
Conversion of cellulose with conventional heating: The
reaction was carried out in a stainless steel pressurized reactor
with a total volume of 30 mL. Before each run, microcrystalline
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 3 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
cellulose (3 mmol), 14 mL MeOH, 0.6 mL H2O and a given was weighted (the solid coke and unreacted cellulose, W3).
amount of Al2(SO4)3 were added into the reactor. The reactor After that, the solid residue was soaked into 0.5 M H2SO4
was then sealed and heated to the set temperature in 30 min. aqueous solution and heated at 433K for 6 h to remove all
The zero time was taken when the medium in the reactor was unconverted cellulose.31 Finally, the left residue was washed,
heated to desired temperature. After reaction, the reactor was dried at 105 oC and weighted (W4). The cellulose conversion
quenched in an ice-water bath. And the reaction mixture was was calculated by the following formula:
filtered with a 0.22-mm membrane filter before analysis.
W −W
3×162×0.96
3
4
Cellulose conversion (%) = (1-
) × 100%
Recycling of the catalyst: after reaction, the solid particles
was firstly separated via centrifugation. The liquid reaction
o
solution was evaporated at 50 C to remove the solvents in a
After the removal of unconverted cellulose, the left residue
was coke, which was generated from the direct carbonization
of cellulose in reaction progress. The elemental composition of
rest residue (W4) was determined on a CHNS analyzer (Thermo
scientific Flash 2000). The carbon content was determined as
C1. The yield of coke was calculated using following equation:
round bottle. Next, the left viscous mixture was treated with
10 mL dichloromethane three times to extract the target
product ML. After that, the solid white residual in the bottle
was dried and reused for the next reaction. The collected
dichloromethane solution was then evaporated to separate
ML. All the extraction solvents were recycled after
evaporation.
W ×C1
3×6×12×0.96
2
Coke yield (%) = (
) × 100%
Product analysis
W4 (mg)= the weight of rest residue after the removal of
Analysis of the reaction products and intermediates (ML, 5- unconverted cellulose;
methoxymethyl furfural) were carried out by using Agilent C1 (%) = the amount of carbon element in rest residue.
7890A GC instrument fitted with a DB-5 capillary column (30 m
The humins generated in the reaction mixture were also
×
0.25 mm × 0.25 μm, Agilent) and
a
FID detector. calculated by carbon yield according to previous literatures.38
Naphthalene was used as internal standard. The other Typically, insoluble products after reaction were firstly filtered
intermediates glucose, MG, fructose and levulinic acid were and washed with 30 mL MeOH. All filtrate was collected. The
analyzed on an Agilent 1200 series HPLC using a Bio-Rad solvent methanol and some volatile small molecules were then
o
o
Aminex HPX-87H column operating at 60 C with a refractive removed by evaporation at 100 C for 12 h. The dried residue
index (RI) detector. 5 mM H2SO4 aqueous solution was used as including humins, (side)products and catalysts was weighted
the mobile phase with a flow rate of 0.6 mL/min. The amount (W5) and characterized by element analysis to obtain carbon
of HMF was analyzed on a Waters 1525 series HPLC equipped content C2. Then, the total weights of all the carbons in
with an UV diode array detector at 280 nm and a C18 column (side)products were calculated as w6. The carbon weight of
at 35 oC (4.6 × 150 mm, 5μm, Agilent). A methanol-water humins can then be determined as W5*C2-w6. And the mole
(50/50, v/v) was used as eluent with a volumetric flow rate of yield of humins formation was calculated by using the
0.4 mL/min. The content of these reaction intermediates in the following formula:
samples was calculated by interpolation from calibration
W ×C2 −W
3×6×12×0.96
curves
5
6
Humin yield (%) = (
) × 100%
The amount of dimethyl ether was determined by
calculating the weight loss of the reactor before and after
reaction, as reported in the previous literature.37 Before the
reaction started, all the reagents and catalyst were placed in
the reactor and weighted. After the reaction was finished and
cooled down to room temperature, the reactor was opened to
exclude the gaseous dimethyl ether and then weighted
(dimethyl ether has a lower boiling point of -34 oC which is
gaseous at room temperature.). The weight loss before and
after the reaction (W1) was approximately equal to the
amount of dimethyl ether formed during the reaction, and the
yield of DEE was calculated as follows:
W5 (mg) = the weight of dried residue;
W6 (mg) = the total weights of all the carbons of reaction
intermediates and products including glucose, MG, fructose,
HMF, MMF and ML.
C2 (%) = the content of carbon element in dried residue.
Results and discussion
Catalytic performances of varied catalysts
A variety of metal salt catalysts were selected to evaluate
their performances in cellulose conversion under microwave
condition. The reaction parameter was firstly set to 180 oC and
40 min with a fixed amount of metal cation loading (Table 1).
To facilitate cellulose conversion, a small amount of water was
also added to promote ML yield according to the previously
reported literatures.31,39 A blank experiment was firstly tested
and no desired ML was detected after reaction, indicating a
catalysing process of cellulose conversion (Table 1, entry 1).
W
W
1
×64
×46
DEE yield (%) =
× 100%
2
W1 (mg) = the weight loss before and after the reaction;
W2 (mg) = the initial methanol weight
Cellulose conversion
：after reaction, the solid residue was
separated by filtration, washed with 30 mL MeOH and dried in
o
vacuum at 105 C for 1h. Then the obtained dry solid residue
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 4 of 13
View Article Online
DOI: 10.1039/C7GC02883K
ARTICLE
Journal Name
Then, a commonly used Fe2(SO4)3 catalyst was investigated unconverted (Table 1, entry 3). This may be originated from
and the reaction provided a moderate ML yield of 48.2% at full the fact that the acidic strengthen of FeCl3 solution was not
cellulose conversion (Table 1, entry 2). A small amount of the enough to convert cellulose under current reaction condition.
side products was detected including MMF, glucose and MG. Besides, the varied performances of the two metal salt
Considerable amount of humins and coke were obtained. By catalysts also highlight the critical role of anions in determining
changing the catalyst to FeCl3, a drastic decrease in the ML the acidity of
yield was observed (6%) and much of cellulose remained
Table 1 Direct conversion of cellulose with various catalysts
Products yield (%)
Carbon balance
(%)
Entry Catalyst
Conv. (%)
ML
0
MMF
0
Glucose
MG
0
Fructose Humins
Coke
3.7
10.4
4.3
3.9
4.5
3.1
3.5
4.5
3.8
8.2
6.1
11.1
3.7
5.7
9.7
1
Blank
4.3
0.2
1.2
2.7
0.1
1.7
0.3
0.6
2.1
1.4
5.5
7.4
0
0
0
0.1
31.6
1.7
0.5
1.8
1.4
1.1
1.3
0.9
4.3
2.4
14.5
3.5
4.8
55.1
-
2
Fe2(SO4)3
FeCl3
100
18.4
5.1
48.2
6.0
0
1.9
0.7
0
0.8
1.3
0
94.1
92.4
88.2
93.2
89.6
87.7
92.2
95.0
92.7
91.9
96.4
93.8
91.8
95.5
3
0.3
0
4
Cr2(SO4)3
CrCl3
5
19.3
6.7
6.1
0.8
0.1
4.4
3.1
10
3.2
0.2
0
0.4
0.2
0.4
0.9
1.5
2.1
4.3
0.5
1.9
5.8
0
0.3
0
6
ZnSO4
ZnCl2
7
6.5
0
8
CuSO4
CuCl2
15.4
11.9
35.7
38.5
100
24.3
41.6
100
0.8
0.5
2.7
6.4
0.1
1.5
1.3
0.3
0.2
0.1
0.3
2.5
0
9
10
11
12
13
14
15a
SnSO4
SnCl4
6.3
70.6
8.2
12.6
30.4
Al2(SO4)3
AlCl3
3.7
7.5
0
0.3
0.5
0
Al(NO3)3
H2SO4
Reaction condition: microcrystalline cellulose 3 mmol, metal salt (Mn+ 1.2 mmol), 0.6 mL H2O, 14 mL MeOH, 800 W, 180 oC, 40 min; a: H2SO4 1.2 mmol.
metal salt catalyst in methanol solution. Encouraged by this respectively (Table 1, entries 13-14). The different reactivities
promising result, other metal salts including those of different of the three Al salt further confirmed the critical role of anion
anions were also evaluated. For Cr and Zn salts, they only in determining the acid strengthen of metal salt in methanol
provided trace amount of ML under the identical condition solution. Based on the above experimental results, only
(Table 1, entries 4-7). Noteworthy, the low solubility of Fe2(SO4)3 and Al2(SO4)3 catalyst offered full cellulose
Cr2(SO4)3 and ZnSO4 in methanol solution was most likely be conversions under the identical reaction condition, suggesting
responsible for their poor reactivity in the reaction.40 For Cu stronger acid strengthen of these two metal salts in alcoholic
salts, the results were still unsatisfying (Table 1, entries 8-9). In solution. But Al2(SO4)3 was the more suitable one with
case of Sn salts, the reactions gave no more than 10% yields of appropriate B/L ratio to control the product selectivity. Besides,
ML with lower cellulose conversion (Table 1, entries 10-11). we also carried out the conventional oil heating for cellulose
This was in accordance to our previous research that that the conversion and found that a longer reaction time of 12h was
use of single Lewis salt SnCl4 was unable to catalyse cellulose needed to obtain the highest ML yield of 68.8% (Table S9). This
conversion in high yield, except the co-existence of Brønsted result suggested that the Al2(SO4)3 catalyst could efficiently
acid H2SO4.41 This result enlightened that both Brønsted and catalyze cellulose conversion in conventional oil heating, but in
Lewis acidic species were essential to the reaction conversion a much lower reaction rate. Finally, single H2SO4 was also
and product yield. To our surprise, Al2(SO4)3 exhibited superior tested in the reaction. Although cellulose was fully converted,
catalytic performances for cellulose transformation, providing only 30.4% yield of ML was obtained, associated with much
70.6% yield of ML at full cellulose conversion (Table 1, entry humins and coke formation (Table 1, entry 15). This may be
12). This yield was close to the highest yield of the mixed-acid caused by the lack of Lewis acid sites of the system, that
system reported by Tominaga,26 but significantly reduced the cannot efficiently convert sugar intermediates (i.e. glucose and
reaction complexity. The good result can be originated from MG from cellulose hydrolysis/alcoholysis) into the downstream
the synergism of Brønsted and Lewis acid species generated by products through the isomerization step. These hydroxyl-rich
the hydrolysis of Al2(SO4)3 in the reaction solution. Similarly, sugars intermediates may easily decompose when they were
-
when the anion was changed to Cl- or NO3 , the reaction exposed in the thermal reaction mixture for longer time, and
provided much lower ML yields, only 8.2% and 12.6%, generate humins and coke. This was also verified by the
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 5 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
MeOH, 0.6 mL H2O, 800 W; a: 0.6 mmol Al2(SO4)3 (20 mol% based on
cellulose content); b:180 oC, 40 min. (orange squares: cellulose conversion,
others referred to fructose, HMF and MMF, see Table S2 and S3 for detail).
reaction using sugar intermediates as reaction substrates in
the presence/absence of Lewis acid catalyst as presented in
Table S8. It’s not difficult to find that both glucose and MG
were fully converted under the standard reaction condition,
but with more humins and coke formation in the case of
lacking Lewis acid, further highlighting the importance of Lewis
acid sites in the cellulose-to-ML conversion.
for 2 h (Table S1). By increasing the temperature to 170 oC, the
reaction could reach a full conversion at 80 min with a
maximum ML of 68.4%, which was close to the highest yield
reported in this work (Figure 1a). The other carbons were
mainly converted to humins and coke. A higher reaction
Overall, the simple metal salt catalyst Al2(SO4)3 exhibited
the highest reactivity in converting cellulose to ML under
microwave condition, providing a very competitive high ML
yield as compared to the literature reported ones. The
hydrolysis of the salt generated the key active Lewis and
Brønsted species that enabled the successive reaction steps
involved in cellulose conversion. In addition, detailed
experiments about the influence of reaction parameters on
the reaction efficiency were also provided in the following part.
o
temperature of 180 C could further reduce the reaction time
to 40 min and increase the ML yield to 70.6%. Obvious
intermediate conversions (e.g. MG, glucose and others) to the
products or humins/coke were observed as the reaction
proceeded. Although further rising the reaction temperature
o
to 190 C could further increase the reaction rate and reduce
the reaction time, the maximum ML yield slightly decreased to
66.1% at 30 min of reaction time. Therefore, the suitable
reaction temperature was set to 180 oC. The reaction time was
40 min, which has been significantly reduced as compared to
the conventional oil heating (e.g. 12-24 h). Besides, a short
reaction time was also benefit for preventing the undesired
side reactions to form humins or coke, representing the great
advantage of microwave heating over oil heating in cellulose
conversion.
Influence of reaction temperature and catalyst loading
The reaction temperature has a profound effect on the
reaction conversion and product distribution. Normally, the
reported optimal temperature for cellulose conversion ranges
from 170-220 oC,22 otherwise may easily lead to either
incomplete cellulose conversion or large production of
humins/coke. Since microwave could generate hot spots in the
reaction media,42 the reaction temperature need to be
carefully evaluated to avoid the above situations. Lower
reaction temperature of 150 or 160 oC only led to lower
cellulose conversion even under the microwave condition,
with a maximum 16.2% ML yield at 49.5% cellulose conversion
The catalyst loading was also found to affect the reaction
conversion and product distribution (Figure 1b). At lower
Al2(SO4)3 loading of 6.7 mol%, the cellulose conversion was
only 59% under the identical reaction condition, accompanying
by considerable amount of the intermediate products.
Continuously increasing the catalyst loading to 20 mol%
resulted in full cellulose conversion and higher ML yield, with
much of the intermediate products converted. These results
indicated that sufficient catalyst concentration was necessary
for maximizing ML yield by promoting the reaction rates.
However, excessive catalyst loading may cause adverse
decrease in ML yield, as the formed ML may undergo
decomposition to humins over stronger acidic solution at
elevated reaction temperature, which was also verified by the
ML decomposition test as presented in Figure S4. Thus, 20
mol% Al2(SO4)3 would be the optimal catalyst loading for
cellulose conversion.
Role of water on cellulose conversion
During our experiments, we noticed that the addition of a
certain amount of water in methanol was important to the
final ML yield and solvent consumption by the etherification of
two methanol molecule. This phenomenon was also found in
several reported literatures such as using 95% ethanol or
adding water to the cellulose alcoholysis system.39 However,
few studies carried in-depth research on the role of water in
cellulose conversion. According to our experience on metal
salt catalyzed system and the reported literature,32,39 water
may affect the catalytic sites, reaction rate or the reaction
pathway. To further reveal these aspects,
a series of
experiments were designed and presented in the following
parts.
Figure 1 Conversion of cellulose with Al2(SO4)3 with different temperature
(a) and catalyst loading (b). Reaction conditions: 3 mmol cellulose, 14 mL
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 6 of 13
View Article Online
DOI: 10.1039/C7GC02883K
ARTICLE
Firstly, the introduction of water may facilitate the
Journal Name
undergoing thermal decomposition to form coke. At the same
time, the yield of humins was firstly reduced and then
increased as the water content increased, with a relatively
lower level of 14.5% yield around 0.6 mL water content. Apart
from that, the addition of water could significantly inhibit the
etherification reaction of methanol itself as shown in the
figure, which was unavoidable in the presence of acidic
catalysts at elevated temperature. When no water was added,
over 20% solvent was consumed to generate DME. As the
water content increased, the DME yield was quickly decreased
to an acceptable level of <5% as the water content exceeded
0.6 mL. The comparison of the etherification of methanol
between Al2(SO4)3 catalyst and other commonly used acid
catalysts that typically used for cellulose conversion were also
investigated, as listed in Table S7. Al2(SO4)3 system provided
the minimum DME production and methanol loss as compared
to that with several other equivalent acid catalysts including
H2SO4, amberlyst-15 and heteroployacids. This may be resulted
from the dynamical equilibrium between the salt and its
hydrolysed acidic species that could prevent the system from
too acidity. However, further evidence about this effect is still
under research. From this point, Al2(SO4)3 may have better
economic effectiveness in cellulose conversion in methanol at
elevated reaction temperature. Based on the above results, it
can be concluded that proper addition of water could
effectively change the product distribution by inhibiting the
undesired side reactions.
hydrolysis of metal salt to form the active catalytic species.
According to work from our and other groups,32,43,44 metal salt
(e.g. Al2(SO4)3) tended to hydrolyze with water to generate
[Al(OH)x(H2O)y]n+ and H+, which were the actual Lewis and
Brøntsd acid sites that participated in the subreactions
involved in cellulose conversion. The water amount in
methanol could affect the equilibrium of metal salt hydrolysis,
as well as the acid density of the reaction mixture, which
would then affect their reactivity in converting cellulose. Thus,
a certain amount of water in methanol to form a mixed solvent
was beneficial for the formation of active catalytic site, and
thereby facilitating cellulose conversion.
Secondly, the introduction of water may also affect the
product distribution. This was verified by conducting the
reaction with varied water addition to methanol under the
identical reaction condition. Figure 2 shows that the final coke
yield was gradually reduced with increasing water amount in
Lastly, the addition of water could also change the reaction
rates and pathway of cellulose conversion. This was verified by
the comparing the product distributions of the reactions with
varied water addition at the earlier stages (Figure 3). The
conversions of cellulose increased as the water content
increased at all the three stages, which indicated the higher
cellulose conversion rates as water was gradually added to the
reaction. Another important phenomenon related to reaction
pathway should be noticed that as the water content
increased, the ratio of the cellulose alcoholysis intermediate
MG to the hydrolysis intermediate glucose (R) was also varied.
When no water was added, MG was presented as the major
intermediate and R were 1.80, 1.95 and 2.65 at 5, 10 and 20
min, respectively.
Figure 2. Influence of Water content on side products. Reaction conditions:
cellulose 3 mmol, 0.6 mmol Al2(SO4)3, 14 mL MeOH, 800 W, 40 min (see Table S4
for detail).
These results suggested that cellulose was more readily to
be converted to the intermediate MG in the absence of
sufficient water. When 0.2 ml water was added to the
reaction, the above R values decreased to 1.11, 1.09 and 1.08,
respectively, and further to 0.43, 0.43 and 0.65 as 0.6mL water
was added. These changes in R values clearly showed that the
introduction of water had important influence on the
pathways, by switching the main route from cellulose
Figure
3 Products distribution of cellulose conversion with different water
content (Orange squares: cellulose conversion). Reaction conditions: cellulose 3
mmol, 0.6 mmol Al2(SO4)3, 14 mL MeOH, 800 W, 180 oC (see Table S5 for detail).
the reaction mixture until a relative constant level was
obtained from 0.6 mL water content. According to the
previous work,45 coke was mainly resulted from the
carbonization of cellulose at high reaction temperature for
longer time. By adding more water, cellulose was more easily
to be attacked by water molecules, moving the reaction
towards to the glycosidic C-O bond hydrolysis rather than
6 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 7 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
reaction proceed mainly through the intermediate glucose
instead of MG, accompanying with higher reaction rates and (3)
change the product distribution, resulting in higher ML yield,
less humin/coke formation and less solvent consumption.
These results further highlighted the importance of water in
cellulose alcoholysis, and also well explained why water was
necessary in the conversion of cellulose.
Reaction kinetic and mechanism
During our experiments, a series of intermediates and side
products were detected as listed in Table 1. This information
provided direct evidences related to the reaction mechanism
of cellulose conversion. To further reveal the reaction pathway,
the product evolution of cellulose conversion under the
Figure
4 ML production from glucose (solid) and MG (hollow). Reaction
conditions: substrate 3 mmol, 0.6 mmol Al2(SO4)3, 0.6 mL H2O, 14 mL MeOH, 800
W, 180 oC.
alcoholysis-to-MG to hydrolysis-to-glucolse, which was
believed to had a big impact on the reaction rate, and even the
product distribution.
To verify this point, we carried out the comparing
experiments by using MG and glucose as substrates for ML
production under the optimized reaction condition (Figure 4).
As shown, it required less time for glucose to reach a full
substrate conversion as compared to MG, which indicated that
glucose reacted faster (higher conversion rate) than MG under
the same reaction condition. At the same time, higher ML yield Figure 5 Product evolution of cellulose conversion. Cellulose 3 mmol, 0.6 mmol
Al2(SO4)3, 14 mL MeOH, 0.6 mL H2O, 180 oC, 800 W.
(82.3% vs 68.1%), as well as its formation rate, was also
obtained in glucose conversion. These results all suggested
that glucose was more readily to be converted to ML with
higher reactivity and reaction rate.46 Also, the mutual
transformations between these two substrates were observed
as both of their intermediates were detected in the reaction,
which was in consistent with the previously reported work.14
Combing these results, it’s not difficult to find that the addition
of water may help to alter the reaction pathway via the more
reactive intermediate glucose, which provided higher reaction
rate and product yield.
Based on the above experiments, it can be concluded that
the water has three main functions in the alcoholysis of
cellulose: 1) provide a reaction media for metal salt hydrolysis
to generate active Brøntsd and Lewis acid species; 2) make the
Scheme 2 Reaction pathway and kinetic model for cellulose conversion.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 8 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
Table 2 Estimated kinetic parameters for cellulose conversion under microwave condition.
Entry
add H2O
(mL)
Rate constants (×10-3 min-1)
T (oC)
K1
K2
K3
K4
K5
K-5
K6
K7
K8
K9
1
2
3
0
0.2
0.6
0.6
0.6
0.6
180
180
180
170
190
180
1.49
1.08
0.81
0.20
3.56
0.02
3.22
6.74
8.21
2.55
17.78
0.33
6.42
4.28
3.65
1.05
9.01
0.15
15.68
15.79
15.94
7.68
29.21
0.48
6.24
6.02
5.81
3.95
8.38
0.24
1.38
2.34
3.95
2.65
6.19
0.15
72.3
1.10
1.07
1.02
0.43
2.55
0.056
151.7c
9.01
9.07
9.10
3.64
18.11
0.30
2.71
2.67
2.69
1.24
5.40
0.13
8.04
7.95
7.80
5.76
10.49
0.31
51.1
4
5
6a
Ea (KJ/mol)
add H2O
(mL)
0
245.4 165.8a 183.5 114.0b 64.1
136.9 125.4d
Entry
Rate constants (×10-3 min-1)
T (oC)
K-9
K10
K11
K12
K13
K14
K-14
1.03
2.40
3.55
2.58
4.76
0.14
52.1
K15
2.58
2.63
2.67
1.54
4.61
0.11
93.5h
K16
2.51
2.45
2.38
1.31
4.30
0.10
101.3
1
2
3
180
180
180
170
190
180
1.61
3.75
5.11
3.67
7.15
0.20
56.8
15.81
28.74
35.45
22.72
53.32
1.45
2.52
3.05
3.26
1.47
7.11
0.13
32.90
32.69
32.55
20.80
49.98
1.37
2.35
2.87
3.04
1.48
6.77
0.12
40.35
40.76
40.85
32.92
51.87
1.71
0.2
0.6
0.6
0.6
4
5
6a
0.6
Ea (KJ/mol)
72.7e 134.4f
74.7
129.6 38.7g
Reaction conditions: cellulose (3 mmol), 0.6 mmol Al2(SO4)3, 14 mL MeOH, 800 W. The apparent activation energy (Ea) of each reaction were calculated based on the
rate constants in entry 3-5. a: reaction conducted with conventional oil heating (Table S6). Apparent activation energies in previous literatures: a: 105.6-177.6
kJ/mol;47,48 b: 86-160 kJ/mol;49-51 c: 70.97-167 kJ/mol;49,52 d: 104.2 kJ/mol;27 e: 57-107 kJ/mol;49 f: 70-142 kJ/mol;49,50,53 g: 29.4-54.3 kJ/mol;49,53 h: 61.7-150 kJ/mol.49,53
identical reaction condition was monitored (Figure 5). added. The rate of cellulose-to-glucose (k2) was relatively
Cellulose was gradually converted as reaction proceeded. MG lower than cellulose-to-MG (k3) without water addition (3.22
and glucose were firstly increased and the decreased to a vs 6.42), revealing that the alcoholysis route 2 was the
negligible level. Other side products such as fructose, HMF and prevalent pathway for cellulose conversion. However, k2 was
MMF were all detected during the reaction. Based on the continuously increased while k3 decreased with increasing
product distribution, a skeleton reaction network related to water content, which indicated that the addition of water
cellulose conversion over Al2(SO4)3 was constructed (Scheme could gradually switch the reaction pathway via cellulose
2).
hydrolysis route 1 to generate glucose. And this reaction route
K1 is the rate constant of cellulose-to-coke conversion; k2 became prevalent (k2=8.21 vs k3=3.65) under our optimized
and k3 are the rate constants of cellulose conversion to glucose reaction condition with 0.6 mL water addition, which was in
and MG, respectively; k5 and k-5 are the rate constants of accordance to our above discussions. The rate constants of
mutual transformation between glucose and MG; K4 and k7 are mutual transformation between glucose and MG (k5 and k-5)
rate constants of glucose-to-HMF and MG-to-MMF exhibited opposite changes as water content increased, with a
dehydration, respectively; k10 and k12 are rate constants of slight increase in k-5 and decrease in k5. This result indicated
HMF-to-LA and MMF-to-ML, respectively; k9 and k-9 are the that more MG was likely to be converted to glucose as water
rate constants of mutual transformation between HMF and content increased, which would also contribute to the reaction
MMF, respectively; k6, k11 and K15 are the rate constants of proceeded via glucose intermediate. The above results clearly
humins formation from glucose, HMF and LA, respectively; k8, suggested that both of the two reaction routes (route 1 and 2)
k13 and k16 are rate constants of humins formation from MG, were parallelly existed in cellulose conversion and the water
MMF and ML, respectively. All the elemental reactions are content played a dominated role in determining which route
considered as pseudo-first-order reactions.
With the increasing of water content in the cellulose ≥0.6mL), the more prevalent of the hydrolysis route it will be.
hydrolysis reaction, the rate of coke formation (k1) decreased As for the successive subreactions in route 1 through the
the reaction would proceed. The more addition of water (e.g.
while that of cellulose conversion (k2+k3) increased. This intermediate glucose, the rate of glucose-to-HMF (k4) was
indicated that the coke formation was suppressed while the almost invariable with varied water contents, which was due
cellulose conversion was promoted as water was gradually to the independence of the dehydration process that didn’t
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 8
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 9 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
require the participation of water molecule. This was also the
In our work, microwave condition is another important
case for LA-to-ML esterification (k14). The rate of HMF-to-LA factor for the high yield of ML production in a shorter reaction
rehydration (k10) increased significantly as water content time as compared to the conventional oil heating systems. To
increased and maintained at a high level when water content further reveal the acceleration effect of microwave on
was 0.6 mL. This was also the case for the ML-to-LA hydrolysis cellulose conversion, oil heating for cellulose conversion was
reaction (k-14), but still at much lower level as compared to carried out and the related reaction rates of the subreactions
that of the LA esterification (K14). As for route 2, the rates of were also investigated as presented in table 2, entry 6. The
MG-to-MMF dehydration (k7) and MMF-to-ML alcoholysis (k12) reaction rates were significantly reduced as compared to the
were invariable with water addition. This suggested that the reaction conducted under microwave condition in table 2,
subreactions in route 2 wouldn’t affect by the water. Apart entry 3. As known, oil heating provides a homogeneously
from that, the glucose conversion rate k4 was larger than MG heating in the reaction media which allows all the subreactions
conversion rate k7, demonstrating that glucose was more proceed under the same reaction temperature. However,
reactive than MG in the conversion. The rate constants of microwave irradiation is a volumetric and dielectric heating
mutual transformation between HMF and MMF showed resource which is significantly affected by the dielectric
similar changes to that of the mutual transformation between properties of the substrates, namely dielectric constant ε` and
glucose and MG, further confirming the role of water that dielectric loss ε``. They reflect the abilities of a substrate in
effectively switched the reaction proceed via route 1. The side absorbing the electromagnetic energy and turning the energy into
reaction rates of humins formation from the two routes (k6, k11, heat, respectively.54 Substrate that possess high dipole moments, as
k15) and (k8, k13, k15) were comparable and remained almost well as the polar chemical bonds or functional groups, can easily
unchanged under the current reaction conditions.
absorb the microwave energy and may generate hot spots (local
Finally, by comparing all the reaction rates of the higher temperature or energy), which would greatly accelerate the
subreactions, the hydrolysis of cellulose and the alcoholysis of reaction rates.42
cellulose had the lowest reaction rates in the two routes,
respectively, thereby being the rate-determining steps of the
two reaction routes. The alcoholysis of cellulose to MG was
prevalent when no water was added to the reaction (k3>k2),
while the hydrolysis of cellulose to glucose became prevalent
after the introduction of 0.6 ml water(k2>k3). Besides, the total
rate of cellulose conversion (k2+k3) with 0.6 mL water content
was also higher than that without water addition, further
suggesting the acceleration effect of water on cellulose
conversion.
To further reveal the apparent activation energies (Ea) of
each subreactions involved in cellulose conversion, two other
o
reaction temperatures (170 and 190 C) were used to obtain
the rates of the subreactions. The Arrhenius rate data plot
were recorded and presented in Figure S2a. All R-squared
values for these reaction steps were higher than 0.98. The
Figure 6 Comparison of the rates of the subreactions with microwave and oil
heating. R represents the ratio of rates constants of microwave heating to that of
oil heating.
apparent activation energy of cellulose hydrolysis (Ea2) and
alcoholysis (Ea3) were 165.8 and 183.5 kJ/mol, respectively,
which were relatively higher than the other subreactions
involved in the main route toward ML formation. This result
further indicated that these two steps were difficult to
proceed and more sensitive to reaction temperature. The
apparent activation energies of the other subreactions of the
two main routes were very similar, which further indicated
that both of these two routes were parallel reactions in
cellulose conversions.
To investigate which subreation in cellulose conversion are more
sensitive to the microwave magnetic field. The reaction rates of the
subreactions under microwave and oil heating were compared. R
represents the ratio of the rates with microwave heating to oil
heating. As indicated in Figure 6, most of the rates of the
subreaction with microwave heating increased 24~25 times of that
with conventional oil heating. A few exceptional subreactions were
found to exhibit varied R values. The rate of cellulose-to-coke
conversion was more accelerated with about 41 times, which may
be caused by the hot spots that lead to local higher temperature
and make cellulose easier to be degraded to coke due to the poor
solubility of cellulose crystal. Albeit that, K1 was still relatively
smaller as compared to the rates of the other subreactions. It’s
worth noting that the rates of glucose-to-HMF (k4) and MG-to-MMF
(k7) conversions were also more accelerated while the rates of
glucose-to-humins (k6) and MG-to-Humins (k8) conversions were
Finally, the accuracy of the kinetic model was examined by
comparing the experimental data to the kinetic model
(predicted ones) in the parity plot in Figure S2b. As seen, a
good fit between the kinetic predicted data for cellulose
conversion and the experimental data was observed, which
further demonstrated the reliability of the current reaction
kinetic models.
Effect of microwave heating
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 9
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 10 of 13
View Article Online
DOI: 10.1039/C7GC02883K
ARTICLE
Journal Name
less accelerated. These results suggested that microwave heating lower yields may be resulted from the low reactivity and steric
enabled the reaction proceed with higher selectivity toward ML hindrance of these alcohols.35,40 However, by extending the
production and suppress the undesired side reactions toward reaction time, the reactions could also reach satisfying product
humin. This may be originated from the high polarity of glucose and yields. In addition, different raw biomass resource was also
MG that rich in -hydroxyl groups, which would then easily absorb tested for ML production. When sucrose and inulin were used,
more energy to create a high-energy local reaction environment 83.4% and 64.7% yields of ML were obtained in 15 min,
and increased the reaction rates substantially.55 The lower R values respectively. Another commonly used α-Cellulose could also
of k6 and k8 were just as the result of the higher R values of k4 and effectively be converted to ML in 68.9% yield under the
k7, respectively. This is because that the increased times of the rates optimized condition. Other raw materials such as Bagasse,
of glucose and MG conversion should match that of their formation Poplar powder and Bamboo powder were also successfully
rates k2 and k3, as these two intermediates (glucose and MG) converted to ML with 65.1, 62.7 and 64.2% yields, respectively,
exhibited similar concentration regularity over the entire reaction based on the cellulose contents of these raw materials. These
time with both microwave and oil heating. Overall, microwave results further demonstrated the great compatibility of
heating did have changed the reaction selectivity toward higher Al2(SO4)3 catalytic system for the production of ML from
selectivity to target product by suppressing the side reaction to different biomass resources.
humins, as compared to the conventional oil heating, which was
also the advantage of microwave heating over conventional heating.
Table 3 Production of alkyl levulinates (AL) with different
This finding may provide preliminary insight into the selective
alcohols or from different biomass resources.
heating of microwave on a molecular level, which may generate
more unique chemistry selectivity.
Substrate
Cellulose
Cellulose
Cellulose
Sucrose
solvent Time (min) AL yield (%)
1
2
3
4
5
6
7
8
9
EtOH
40/55
40/75
40/75
15
53.9/70.1
26.6/53.7
36.6/63.1
83.4
Catalyst reusability
iPrOH
nBuOH
MeOH
MeOH
MeOH
MeOH
MeOH
For the application of a catalyst in biomass conversion, its
recyclability is very important. Although the current reaction
was a homogeneous system, the catalyst could also be
efficiently recovered by simple procedures. The recycled
catalyst was directly subjected to the next reaction. After five
recycling, the yield of ML decreased by ca 5.3%, from 70.6% to
65.3% in the sixth run. The slight decrease in the product yield
may be attributed to the loss of acid species during the
recycling process. Albeit that, the catalyst still possessed high
reactivity for cellulose conversion. Concerning the high
reaction efficiency and good recyclability of Al2(SO4)3, it would
be an ideal catalyst for the production of ML from cellulosic
biomass, in spite of its homogeneous nature in the reaction.
Inulin
15
64.7
α-Cellulose
Bagasse
40
68.9
35
65.1
Poplar powder
35
62.7
Bamboo powder MeOH
40
64.2
Reaction condition: substrate 500 mg, 0.6 mmol Al2(SO4)3, 14 mL MeOH, 0.6 mL
H2O, 800 W, 180 oC. ML yields from bagasse, poplar powder and bamboo powder
were calculated based on the cellulose content in dry feedstocks.
Conclusions
In conclusion, we have presented a simple and efficient
metal salt Al2(SO4)3 catalyst system for the high yield
production of ML from cellulose in methanol solvent under
microwave condition. A high yield of 70.6% of ML, comparable
to the reported highest ML yield (75%), was obtained at 180 oC
for 40 min, with the addition of a certain amount of water. The
addition of water can not only effectively increase the rate of
cellulose conversions, but also reduce the etherification of
solvents and the coke/humin formation during the reaction.
Mechanistic and kinetic studies revealed that the reaction
proceed mainly through cellulose hydrolysis intermediate
glucose and alcoholysis intermediate MG, with the former as
the prevalent route for ML production in the presence of the
optimized amount of water. Besides, the route through
glucose intermediate showed higher reaction rates as
compared that through MG intermediate. Microwave was
demonstrated to effectively accelerate the reaction rate ~25
times of that with conventional oil heating, with more times in
the rate of glucose/MG conversion and less in their side
reactions to humins, resulting in higher selectivity toward the
target ML product. Finally, the catalyst could be successfully
recycled with high reactivity and also showed good
Figure 7 Recycle study of Al2(SO4)3 catalyst. Reaction conditions: 3 mmol
cellulose, 14 mL MeOH, 0.6 mL H2O, 800 W, 180 oC,40 min.
Applications for biomass conversions
Finally, the established method was applied to the
conversion of other carbohydrates and raw biomass. Before
that, different alcohols were also tested for cellulose
conversion to produce different alkyl levulinates. The reactions
with EtOH, iPrOH and nBuOH as solvents provided 53.9, 26.6
and 36.6% yields of alkyl levulinate product, respectively. The
10 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 11 of 13
View Article Online
DOI: 10.1039/C7GC02883K
Journal Name
ARTICLE
19 Q. N. Xia, Z. J. Chen, Y. Shao, X. Q. Gong, H. F. Wang, X. H. Liu,
S. F. Paker, X. Han, S. H. Yang and Y. Q. Wang, Nat. Commun.,
2016, 7, 11162.
20 E. Christensen, A. Williams, S. Paul, S. Burton and R. L.
compatibility for the conversion of other carbohydrates and
raw biomass materials to ML with slightly varied reaction time,
making Al2(SO4)3 a green, practical and potential catalyst for
industrial ML production from biomass resources.
McCormick, Energy Fuels, 2011, 25, 5422.
21 M. J. Climent, A. Corma and S. Iborra, Green Chem., 2014, 16
516.
,
22 A. Démolis, N. Essayem and F. Rataboul, ACS Sustainable
Chem. Eng., 2014, , 1338.
23 M. Mascal and E. B. Nikitin, Green Chem., 2010, 12, 370.
Acknowledgements
2
This work was supported by the by Natural Science Foundation
of Jiangsu Province (BK20150872), National Natural Science 24 F. Su, Q. Y. Wu, D. Y. Song, X. H. Zhang, M. Wang and Y. H.
Guo, J. Mater. Chem. A, 2013, 1, 13209.
Foundation of China (NSFC 21502095), Jiangsu student’s
platform for innovation training program (SPITP
201710298008, 2016NFUSPITP028), Top-notch Academic
Programs Project of Jiangsu Higher Education Institutions
25 X. Y. Wu, J. Fu and X. Y. Lu, Carbohydr. Res., 2012, 358, 37.
26 K. Tominaga, A. Mori, Y. Fukushima, S. Shimada and K. Sato,
Green Chem., 2011, 13, 810.
27 L. C. Peng, L. Lin and H. Li, Ind. Crops Prod., 2012, 40, 136.
(TAPP) and Priority Academic Program Development of Jiangsu 28 C. H. Luo, A. S. Poyraz, L. Jin, Y. Meng, L. Pahalagedara, S. Y.
Chen, D. A. Kriz, C. Guild, A. Gudz and S. L. Suib, Green
Chem., 2014, 16, 785.
Higher Education Institutions (PAPD)
29 W. P. Deng, M. Liu, Q. H. Zhang and Y. Wang, Catal. Today,
2011, 164, 461.
30 Y. Liu, C. L. Liu, H. Z. Wu and W. S. Dong, Catal. Lett., 2013,
143, 1346.
31 D. Q. Ding, J. X. Xi, J. J. Wang, X. H. Liu, G. Z. Lu and Y. Q.
Wang, Green Chem., 2015, 17, 4037.
32 J. J. Tang, X. W. Guo, L. F. Zhu and C. W. Hu, ACS Catal., 2015,
Notes and references
‡ Footnotes relating to the main text should appear here. These
might include comments relevant to but not central to the
matter under discussion, limited experimental and spectral data,
and crystallographic data.
5
, 5097.
§
33 Y. L. Wang, W. P. Deng, B. J. Wang, Q. H. Zhang, X. Y. Wan, Z.
C. Tang, Y. Wang, C. Zhu, Z. X. Cao, G. C. Wang and H. L. Wan,
Nat. Commun., 2013, 4, 2141.
1
G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106,
4044.
34 Y. Yang, C. Hu and M. M. Abu-Omar, Bioresour. Technol.,
2012, 116, 190.
35 Y. B. Huang, T. Yang, M. C. Zhou, H. Pan and Y. Fu, Green
Chem., 2016, 18, 1516.
36 L. P. Zhou, H. J. Zou, J. X. Nan, L. Wu, X. M. Yang, Y. L. Su, T. L.
Lu and J. Xu, Catal. Commun., 2014, 50, 13.
37 L. C. Peng, L. Lin, H. Li and K. L. Chen, BioResources, 2013, 8,
2
3
J. O. Metzger, Angew. Chem., Int. Ed., 2006, 45, 696.
S. N. Naik, V. V. Goud, P. K. Rout and A. K. Dalai, Renewable
Sustainable Energy Rev., 2010, 14, 578.
S. Dora, T. Bhaskar, R. Singh, D. V. Naik and D. K. Adhikari,
Bioresour. Technol., 2012, 120, 318.
4
5
T. Komanoya, H. Kobayashi, K. Hara, W. J. Chun and A.
Fukuoka, Appl. Catal., A, 2011, 407, 188.
5895.
38 T. M. C. Hoang, E. R. H. van Eck, W. P. Bula, J. G. E.
6
7
Y. B. Huang and Y. Fu, Green Chem., 2013, 15, 1095.
J. S. Luterbacher, J. M. Rand, D. M. Alonso, J. Han, J. T.
Youngquist, C. T. Maravelias, B. F. Pfleger and J. A. Dumesic,
Science, 2014, 343, 277.
R. Carrasquillo-Flores, M. Käldström, F. Schüth, J. A. Dumesic
and R. Rinaldi, ACS Catal., 2013, 3, 993.
Gardeniers, L. Lefferts and K. Seshan, Green Chem., 2015, 17
,
959.
39 A. S. Amarasekara and B. Wiredu, BioEnergy Res., 2014,
1237.
7
,
8
9
40 Y. B. Huang, T. Yang, B. Cai, X. Chang and H. Pan, RSC Adv.,
2016, , 2106.
41 Y. B. Huang, T. Yang, A. F. Liu, X. C. Zhou and H. Pan, Chin. J.
Org. Chem, 2016, 3 , 1483.
C. Chiappe, M. J. R. Douton, A. Mezzetta, C. S. Pomelli,
G.Assanelli and A. R. de Angelis, ACS Sustainable Chem. Eng.,
6
2017,
10 P. Zhou and Z. H. Zhang, Catal. Sci. Technol., 2016,
5, 5529.
6
6, 3694.
42 C. O. Kappe, Angew. Chem. Int. Ed., 2004, 43, 6250.
11 J. L. Xu, H. Y. Zhang, Y. F. Zhao, Z. Z. Yang, B. Yu, H. J. Xu and Z.
M. Liu, Green Chem., 2014, 16, 4931.
12 J. L. Song, H. L. Fan, J. Ma and B. X. Han, Green Chem., 2013,
15, 2619.
43 For details about the characterization and discussion of Al
catalysts, see our recent work: (1) T. Yang, B. Cai, X. Chang, H.
Pan and Y. B. Huang, ChemistrySelect, 2017, 2, 3775; (2) T
Yang, Y.-H. Zhou, S. Z. Zhu, H. Pan, and Y. B. Huang,
ChemSusChem 2017, 10, 4066.
44 V. Choudhary, S. I. Sandler and D. G. Vlachos, ACS Catal.,
13 L. C. Peng, L. Lin, H. Li and Q. L. Yang, Appl. Energy, 2011, 88
,
4590.
14 S. Saravanamurugan and A. Riisager, ChemCatChem, 2013, 5,
1754.
15 F. de Clippel, M. Dusselier, R. Van Rompaey, P. Vanelderen, J.
Dijkmans, E. Makshina, L. Giebeler, S. Oswald, G. V. Baron, J.
F. M. Denayer, P. P. Pescarmona, P. A. Jacobs and B. F. Sels, J.
Am. Chem. Soc., 2012, 134, 10089.
2012, 2, 2022.
45 S. C. Du, J. A. Valla and G. M. Bollas, Green Chem., 2013, 15
,
3214.
46 E. Ahmad, M. I. Alam, K. K. Pant and M. A. Haider, Green
Chem., 2016, 18, 4804.
47 A. O. Converse, I. K. Kwarteng, H. E. Grethlein and H.
Ooshima, Appl. Biochem. Biotech., 1989, 20, 63.
48 Y. Sun, J. Zhuang, L. Lin and P. O. Yang, Biotechnol. Adv.,
2009, 27, 625.
16 M. Y. Zheng, J. F. Peng, R. Y. Sun, A. Q. Wang and T. Zhang,
ACS Catal., 2017, 7, 1939.
17 K. Fabičovicová, M. Lucas and P. Claus, Green Chem., 2015,
17, 3075.
49 N. A. S. Ramli and N. A. S. Amin, Chem. Eng. J., 2016, 283
150.
50 B. Girisuta, K. Dussan, D. Haverty, J. J. Leahy and M. H. B.
Hayes, Chem. Eng. J., 2013, 217, 61.
,
18 B. O. de Beeck, M. Dusselier, J. Geboers, J. Holsbeek,
E.Morré, S. Oswald, L. Giebeler and B. F. Sels, Energy Environ.
Sci., 2015, 8, 230.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 11
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 12 of 13
View Article Online
DOI: 10.1039/C7GC02883K
ARTICLE
Journal Name
51 L. Atanda, M. Konarova, Q. Ma, S. Mukundan, A. Shrotrib and
J. Beltramini, Catal. Sci. Technol., 2016, , 6257.
6
52 W. N. Zhu, C. Chang, C. Ma and F. G. Du, Chin. J. Chem. Eng.,
2014, 22, 238.
53 A. Mukherjee, M. J. Dumont, Ind. Eng. Chem. Res., 2016, 55
8941.
,
54 M. Tripathia, J. N. Sahub,c, P. Ganesana, P. Monashc and T. K.
Dey, J. Anal. Appl. Pypol., 2015, 112, 306.
55 J. J. Fan, M. De bruyn, V. L. Budarin, M. J. Gronnow, P. S.
Shuttleworth, S. Breeden, D. J. Macquarrie and J. H. Clark, J.
Am. Chem. Soc., 2013, 135, 11728.
12 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 13 of 13
Green Chemistry
View Article Online
DOI: 10.1039/C7GC02883K
The production of alkyl levulinates from cellulose was successfully achieved by using simple
and efficient metal salt catalyst Al2(SO4)3, offering a high yield up to 70.6% under microwave
condition within a short reaction time of 40 minutes. Kinetic and mechanistic studies revealed
that reaction proceeded mainly via cellulose hydrolysis route in the presence of extra water,
and the cellulose alcoholysis route was also parallelly involved.