Hydrothermal production of furfural from xylose and xylan as model compounds for hemicelluloses

Article abstract of DOI:10.1039/c3ra43108h

The uncatalyzed microwave-assisted hydrothermal conversion of xylose and xylan as hemicellulose model compounds to furfural at temperatures of 160-240 °C and reaction times of 5-240 min is investigated. With acceptable reaction rates at 200 °C the conversion of xylose provides furfural with a maximum yield of 49%. The conversion of xylan, however, requires lower temperatures (160-180 °C) in order to promote the hydrolysis of the polymer, yielding xylose as a furfural precursor. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3ra43108h

View Article Online
View Journal
RSC Advances
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: M. Möller and U. Schröder, RSC Adv.,
2013, DOI: 10.1039/C3RA43108H.
This is an Accepted Manuscript, which has been through the RSC Publishing peer
review process and has been accepted for publication.
Accepted Manuscripts are published online shortly after acceptance, which is prior
to technical editing, formatting and proof reading. This free service from RSC
Publishing allows authors to make their results available to the community, in
citable form, before publication of the edited article. This Accepted Manuscript will
be replaced by the edited and formatted Advance Article as soon as this is available.
To cite this manuscript please use its permanent Digital Object Identifier (DOI®),
which is identical for all formats of publication.
More information about Accepted Manuscripts can be found in the
Information for Authors.
Please note that technical editing may introduce minor changes to the text and/or
graphics contained in the manuscript submitted by the author(s) which may alter
content, and that the standard Terms & Conditions and the ethical guidelines
that apply to the journal are still applicable. In no event shall the RSC be held
responsible for any errors or omissions in these Accepted Manuscript manuscripts or
any consequences arising from the use of any information contained in them.
www.rsc.org/advances
Registered Charity Number 207890

Page 1 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
Table of contents entry
The hydrothermal liquefaction of xylose and xylan as model compound for
hemicellulose is studied. The reaction rate constants of the decompositions are
determined for selected temperatures ranging from 180 – 220 °C. The production of
furfural is in the main focus: its yields and selectivities are evaluated.

RSC Advances
Page 2 of 20
View Article Online
DOI: 10.1039/C3RA43108H
Hydrothermal production of furfural from xylose
and xylan as model compounds for hemicelluloses
Maria Möller, Uwe Schröder*
Institute of Environmental and Sustainable Chemistry, Technische Universität
Braunschweig, Hagenring 30, 38106 Braunschweig, Germany.
*author of correspondence; phone: +49 5313918425; fax: 49 5313918424, email to:
uwe.schroeder@tuꢀbs.de
Abstract
The uncatalyzed microwaveꢀassisted hydrothermal conversion of xylose and xylan as
hemicellulose model compound to furfural at temperatures of 160 – 240 °C and
reaction times of 5 – 120 min is investigated. With acceptable reaction rates already
at 200 °C the conversion of xylose provides furfural with a maximum yield of 49%.
The conversion of xylan, however, requires lower temperatures (160 – 180 °C) in
order to promote the hydrolysis of the polymer, yielding xylose as furfural precursor.
Key words: hydrothermal liquefaction, furfural, xylose, xylan, microwave heating
ꢀ 1 ꢀ

Page 3 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
1. Introduction
Biomass is considered as a promising alternative to fossil feedstock for the
production of chemicals and energy. 1 The carbohydrate derived platform chemicals
5ꢀ(hydroxymethyl)furfural (5ꢀHMF), furfural, levulinic acid and lactic acid are
precursors for functionalized building blocks for fine chemicals, for monomer units of
diverse polymers, 2 and as precursors for potential biofuels or fuel additives. 3 Hence
the production of these substances gains increasing attention, see 4ꢀ6
.
The main constituents of plant biomass are cellulose, hemicelluloses and lignin. Their
relative fractions differ in dependence of the particular feedstock (hardꢀ and
softwoods, agriculture residues, etc.), yet an average ratio of 4:3:3 can be estimated
7,
8). As shown in Scheme 1 these biopolymers form the main skeleton of
lignocellulosic material and determine their mechanical properties. In plant cell walls
cellulose fibrils are embedded in hemicellulose jackets, which are crossꢀlinked to
lignin 9. Due to this jacket cellulose is protected and is thus inaccessible for biological
degradation processes 10. Hence, for the breakdown of lignocellulosic feedstock the
degradation behavior of hemicelluloses must be taken into account.
While lignin consists of phenylpropanoid fractions forming a highly branched
polymeric structure, 11 cellulose and hemicelluloses are based on carbohydrates. In
contrast to cellulose (an unbranched
β
ꢀ1,4ꢀglycosidic linked homopolymer consisting
ꢀ1,4ꢀlinked, partially branched,
of glucose units), hemicellulose refers to a group of
β
amorphous polysaccharides composed of pentoses (Dꢀxylose, Lꢀarabinose) and
hexoses (Dꢀglucose, Dꢀmannose, Dꢀgalactose). 10, 12 Among various hemicelluloses
xylan is a hemicellulose consisting of xylose as its major monomer. 7, 8 In this study
we use xylan as the homopolymer of xylose.
ꢀ 2 ꢀ

RSC Advances
Page 4 of 20
View Article Online
DOI: 10.1039/C3RA43108H
Scheme 1 – Schematic illustration of lingocellulose
Hydrothermal liquefaction is considered as a promising technology for the conversion
of wet biomass to valueꢀadded products. 13 Therefore, the hydrothermal conversion
of lignocellulosic raw material (e.g. 8, 14) as well as of biomass model compounds,
e.g., cellulose, glucose and xylose in subcritical water has been investigated under
various experimental conditions. 15ꢀ18 Studies with the main focus on hemicelluloses
model compounds are rare, and to our knowledge there is only one study examining
the uncatalyzed hydrothermal conversion of the homopolymer xylan. 19 The use of a
hemicellulose model compound in a fundamental investigation provides advantages
since the impact of interactions between hemicellulose and cellulose and lignin can
be excluded. However, one has to be aware that for practical exploitations these
interactions have to be taken into account.
In this study, the hydrothermal decomposition of xylose and its polymer xylan as a
representative hemicellulose compound are investigated – with the emphasis on the
yield of the dehydration product furfural. Because of its advantages as an activation /
heating method for hydrothermal conversion processes (see, e.g., 20ꢀ22) the
ꢀ 3 ꢀ

Page 5 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
hydrothermal conversion experiments were performed using microwave heating. The
experiments were conducted in a temperature window between 160 and 240 °C and
at reaction times between 5 – 180 min – without the addition of any catalysts. The
suitability of microwave heating as well as the impact of the reaction variables
temperature, time and the extent of the substrate conversion was elucidated.
2. Materials and Methods
General Conditions
All experiments were performed without the addition of any catalyst. For this reason
we refer to the experiments as being uncatalysed, being aware though that catalytic
effects may occur during the reaction based on the formed products and
intermediates (e.g., organic acids) as well as on the greater extent of water
dissociation in subcritical water.
Materials
Xylan (from Beechwood, polyꢀ(βꢀDꢀxylopyranose[1→4]) 90%, SigmaꢀAldrich; CHNꢀ
analysis: C [%]: 40.77, H [%]: 6.04, N [%]: ꢀ), Dꢀ(+)ꢀxylose (for biochemistry, 99%,
Merck), were used as purchased, without further purification. Furfural (99%, Sigmaꢀ
Aldrich) was used for the preparation of reference sample for calibration curves.
Deionized water was used for the preparation of aqueous solutions.
Microwave assisted degradation of Xylose and Xylan
All hydrothermal reactions were performed using a Synthos 3000 microwave
synthesis system (Anton Paar, Austria) with a magnetron frequency of 2455 MHz and
a maximum microwave power output of 1400 W in unpulsed control mode. Two
80 mL quartz vessels (Anton Paar, XQ80, max. pressure limit: 8 MPa, temperature
limit: 300 °C, maximal filling: 60 mL), containing a magnetic stirrer were charged with
15 mL of an aqueous carbohydrate (xylose) solution (57 mM), closed with a lipꢀtype
seal (PTFEꢀTFM) and were put in position 1 and 5 of the 8ꢀpositionꢀrotor (Anton
Paar, S8). For xylan degradation reactions each of these vessels were charged with
ꢀ 4 ꢀ

RSC Advances
Page 6 of 20
View Article Online
DOI: 10.1039/C3RA43108H
125 mg xylan (correspondent to 60 mM xyloseꢀunits, assuming no acetylation),
15 mL deionized water was added. The comparably low concentration levels of
xylose and xylan were chosen to focus the study on the fundamental degradation
processes, without significant interference with secondary effects like the mentioned
polymerization. Empty vessels were put in position 3 and 8 for rotor symmetry
reasons. The vessel in position 1 was equipped with a temperature/pressureꢀsensor
(immersed gasꢀbulb thermometer (max. 300 °C)) to control the reaction by
temperature vs. time mode. 23 A typical microwave heating profile is shown in Fig. 1.
For this study, reaction temperatures of 180 °C, 200 °C, 220 °C and 240 °C and
reaction times ranging from 5 min to 120 min were chosen.
250
reaction temperature
200
150
100
50
7.8 min
preheating
time
xylose
xylan
reaction time
total time
heat dissipation time (no power supply)
0
0
10
20
30
40
time [min]
Fig. 1 – Typical temperature profiles of a microwave heated hydrothermal experiment. The
preheating time was 5 min, reaction time 10 min, the target temperature was 220 °C. The
Fig. shows temperature profiles of aqueous xylose solution (57 mM, 2 x 15 mL, blue) and
xylan suspension (60 mM xyloseꢀunits, 2 x 15 mL, green) during microwave heating.
A preheating or ramp time of 5 min was selected for all runs. During the final coolingꢀ
down phase no extra power (apart from a low air stream in the device) was applied to
the microwave device to accelerate the cooling process. After the completed
microwave run, the product solutions were transferred into sample vials. A loss of
ꢀ 5 ꢀ

Page 7 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
volume during the hydrothermal reaction was not observed. Each time 2 mL solution
were sampled for analysis and were filtered using a syringe filter (PhenexꢀRC
membrane, 15 mm, 0.2 ꢁm) in order to remove solid particles.
The data of each experimental run including the measured temperature, pressure
and the power input were recorded in time segments of one second and were
downloaded from the microwave system for subsequent analysis.
All experiments were started at ambient pressure. The pressure buildup in the
reaction vessels during the experiments at temperatures between 180 – 240 °C was
between 1.2 MPa at 180 °C and about 4.3 MPa at 240 °C.
As illustrated before, 22 the microwave heating of water and aqueous solutions can
follow the programmed temperature ramp up to target temperatures of about 180ꢀ
200 °C. Although the presence of sugars in aqueous solution improves the
absorption of microwave energy, still a prolonged heating time is necessary to reach
target temperatures above 200 °C. 22 This is illustrated in figure 1 for an aqueous
xylose solution: almost three minutes extra are required to reach the programmed
reaction temperature of 220 °C. However, the suspension of xylan in water provides
a good microwave absorbing medium – the programmed temperature profile fits very
well.
Analysis
The determination of the xylose and furfural concentration in the liquid phase was
performed by a HPLC system (Spectrasystem P4000, Finnigan Surveyor RI Plus
Detector, Fisher Scientific) equipped with a Rezex ROAꢀOrganic Acid H+ (8%)
column (300 x 7.8 mm). Sulfuric acid (2.5 mM, flow rate 0.5 mL minꢀ1) served as the
eluent. The column was kept at room temperature; the detector was set at 40 °C.
Product concentrations were determined using calibration curves obtained from
reference samples.
Qualitative Analysis was carried out by GC/MS (Trace GC Ultra, DSQ II, Thermo
Scientific, TRꢀWaxMS (30 m x 0.25 mm ID x 0.25 ꢁm film GC column)). The total
carbon content (TC) of the liquid phase of selected samples was determined by the
use of a TOC cuvette test (Hach Lange (LCK381), Germany).
ꢀ 6 ꢀ

RSC Advances
Page 8 of 20
View Article Online
DOI: 10.1039/C3RA43108H
To determine the soluble xylan residue after the hydrothermal experiment, a
24
subsequent hydrolysis was carried out (procedure adapted from
and modified):
5 mL of the crude reaction product were transferred into a test tube, 284 mg (174 ꢁL)
sulfuric acid (72%) were added, the tube was sealed and heated at 100 °C for 1 h.
For reference a second test tube was prepared similarly but was not heated. After
removal of solid particles, the samples were taken for xylose analysis.
Assuming the absence of any reaction product at
t = 0, the product yields (Yi) of the
analysed liquid products ( ), the selectivity ( i) and the degree of educt conversion (X)
i
S
were calculated by the subsequent equations (with
number):
c
: concentration, ν: stoichiometric
(1)
(2)
(3)
νEduct
ci [mM]
Y [%]=
⋅⋅
νi cEduct,0 [mM]
⋅100[%]
i
νEduct ci [mM]
Si [%]=
X [%]=
⋅⋅
⋅100[%]
νi ꢀcEduct [mM]
ꢀcEduct [mM]
⋅100[%]
cEduct,0 [mM]
Beside solid products, gaseous products were presumably formed during the
experiments, however, the analysis of these products was not within the scope of this
study.
All reactions were carried out in duplicate, reported are the average values, the
deviations did not exceed 5%.
The rate constants for the xylose and the xylan decomposition were calculated
assuming first order kinetics (fitting not shown). The calculation was based on the
pure reaction time and did not take into account the preheating and cooling down
phase (see Fig. 1). This may produce a certain systematic error, which, however may
be tolerable within a comparative study.
3. Results and discussion
Temperature dependence of Xylose and Xylan Decomposition
ꢀ 7 ꢀ

Page 9 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
Xylose and xylan are effectively decomposed in subcritical water. As illustrated in
figure 2 (A, B), increasing reaction temperatures and increasing reaction time lead to
an enhancement of the xylose conversion. Whereas at 200 °C a reaction time of
60 min is required for the xylose conversion, at 240 °C 90% xylose is converted
within 10 min. The conversion rate of xylose is comparable to those of the hexoses
glucose and fructose (Table 1, represented by the reaction rate constant k [sꢀ1] under
the assumption of a first order kinetic), showing the following gradation:
fructose > xylose > glucose. The estimated reaction rate constants of this study for
the xylose decomposition are consistent with literature data. 25
Table 1. – Reaction rate constant k for uncatalyzed hydrothermal conversion of
carbohydrates. A first order reaction is assumed for the calculation.
T [°C]
k [sꢀ1]
Glucose
Fructose
6.5*10ꢀ4 a
1.8*10ꢀ3 a
3.6*10ꢀ3 a
Xylose
7.5*10ꢀ5 c
180
200
220
ꢀ
ꢀ
ꢀ
4.5*10ꢀ4 b
1.2*10ꢀ3 b
4.0*10ꢀ4 c
1.1*10ꢀ3 c
5.1*10ꢀ4 a
a: 19, c (hexose) = 55 mM, b: calculation base on experimental data of this study c: 21, c
(xylose) = 72 mM.
For the decomposition of the xylose homopolymer xylan lower temperatures and
shorter reaction times are sufficient, 30 min at 200 °C or 10 min at 220 °C yield a
complete conversion (figure 2 (C, D)). This is reflected by the reaction rate constant
of the xylan decomposition for which a value of 4.2*10ꢀ4 sꢀ1 is determined for a
reaction temperature of 180 °C. This rate constant can be considered as a mixed rate
constant comprising the hydrolytic cleavage of xylan to xylose as well as side
reactions like the (intramolecular) dehydration of xylan. Assuming that the hydrolytic
cleavage is the dominating progress an excess of xylose is expected since the
hydrolysis of xylan to xylose is estimated to be five times faster than the subsequent
dehydration of xylose to furfural (see table 1, 180 °C reaction temperature). At 220 °C
a reaction rate constant of 1.3*10ꢀ3 sꢀ1 for the xylan hydrolysis has been determined
by Pinkowska, 19 correlated with the present data, this indicates that the xylan and
xylose degradation proceed at the same rate. However, as shown in figure 2 (B, D)
this is inconsistent with our data. Possible explanations for the deviation may
comprise experimental differences e.g. in the heating procedure (in heatingꢀup rates
ꢀ 8 ꢀ

RSC Advances
Page 10 of 20
View Article Online
DOI: 10.1039/C3RA43108H
and cooling down times) as well as the applied kinetic model for xylan degradation
and resulting calculations.
100
80
100
80
A
B
60
40
60
20
200 °C
220 °C
40
0
180
190
200
210
220
230
240
0
20
40
60
80
100
120
100
100
80
C
xylan conversion
xylan conversion
D
80
60
180 °C
200 °C
220 °C
240 °C
60
40
6
40
12
10
8
xylose yield
4
2
0
6
xylose yields
4
2
0
180
190
200
210
220
230
240
0
20
40
60
80
100
120
temperature [°C]
time [min]
Fig. 2 A, B: temperature and time dependence of the hydrothermal xylose conversion
(conditions: aqueous xylose solution (56 mM, 2 x 15 mL), C, D: temperature and time
dependence of the hydrothermal xylan conversion (conditions: xylan suspension
(2 x (125 mg xylan in 15 mL)). The temperature dependence was determined for a reaction
time of 10 minutes.
Various kinetic models for the (acidꢀcatalyzed) hydrolysis of the hemicellulose
8
components of lignocellulosic material have been reported in literature . A widely
applied approach is a pseudoꢀhomogeneous, firstꢀorder mechanism, which is also
based on the decomposition of lignocellulosic material yielding carbohydrates via
irreversible reactions 14 shown in Figure 3. The formation of intermediate (xyloꢀ)
oligomers as well as xylose degradation reaction is considered. In addition it is
distinguished between fast and slow hydrolyzed xylan components. The fast reaction
reflects the hydrolysis of “free” xylan, whereas the slow reaction takes into account
the decreased accessibility of the polymer due to interactions to lignin and
cellulose. 26 A general scheme for this approach is presented in figure 3, providing
the basis for further extension (e.g. considering fast and slow hydrolyzed oligomers
ꢀ 9 ꢀ

Page 11 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
or the furfural formation and its decomposition as well as condensation of furfural with
reaction intermediates 8, 27).
Fig. 3 – pseudohomogeneous, firstꢀorder mechanism of xylan hydrolysis (here, xylan refers
to hemicellulosesꢀconsisting part of lignocellulosic material) 14
.
Overall, it seems acknowledged that the rateꢀlimiting process during the uncatalysed
hydrothermal conversion of xylan are the hydrolysis steps of the polymer to the
monomer xylose, whereas under acid catalysis, the decomposition of xylose to
furfural becomes rateꢀlimiting. 10,27
Since in this study isolated xylan is used as a hemicellulose model compound, the
slow kinetic path (k2) can be neglected. The above assumptions may be
oversimplified; however, the quantification of all degradation products required for the
thorough kinetic modeling was not within the scope of this study.
Formation of Furfural
As depicted in Figure 4, during the decomposition of xylose the same types of
reactions known from the decomposition of hexoses takes place, with analogue
products. Beside furfural – the product of the acidꢀcatalyzed cyclodehydration of
10
xylose
– glyceraldehyde and glycolaldehyde, dihydroxyacetone, pyruvaldehyde,
lactic acid, hydroxyacetone are detected products. 28 Acids (e.g., acetic acid, derived
from acetyl derivative xylose units, pyruvic acid) are known as further decomposition
products. 29 The identified side products in this study were dihydroxyacetone,
hydroxyacetone and formic acid.
Furfural is identified as main decomposition product of the uncatalyzed hydrothermal
conversion of xylose yielding a maximum of 49% furfural (55% selectivity at 200 °C
and 60 minutes reaction time). Increasing reaction temperatures and increasing
reaction times do not lead to further furfural production.
ꢀ 10 ꢀ

RSC Advances
Page 12 of 20
View Article Online
DOI: 10.1039/C3RA43108H
Fig. 4 – Known reaction pathways of the degradation of xylan (simplified structure, since
xylan may contain a low fraction of partly acetylated xylose units) and xylose respectively. 10,
13, 28 LBET: Lobry de Bruyn–Alberda van Ekensteinꢀtransformation.
Furfural itself is an intermediate product, thus it is sensitive against acidic compounds
30
(which refers to acid catalyzed cleavage of the furan structure) and autoꢀoxidation
leading to decomposition to formic acid and to polymeric products. 31 Therefore the
actual furfural yield from xylose is limited by these processes. As shown in Figure 5A,
at 200 °C the selectivity of the furfural formation shows only a minor decrease for
increased reaction times, since the complete xylose conversion is reached not until
120 min reaction time and furfural is formatted continuously upon xylose dehydration.
In contrast, at 220 °C the xylose conversion is completed within 60 min. Thus, the
furfural decomposition becomes prevailing, indicated by a significant decrease of the
furfural selectivity at prolonged reaction times.
In comparison to 5ꢀHMF – the hydrothermal dehydration product derived from
hexoses – furfural is less affected by decomposition processes under the applied
ꢀ 11 ꢀ

Page 13 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
reaction conditions. Thus, the furfural selectivity decreases much lesser within a
comparable reaction times (see figures 5 A and B). The maximal selectivities of
furfural and 5ꢀHMF formation are similar: The maximum furfural selectivity is 49%
(obtained at 220 °C, 30 minutes reaction time), the value for 5ꢀHMF is 48%
(10 minutes reaction time 22).
60
50
40
30
20
10
0
60
50
40
30
20
10
0
X = 98%
X = 100%
X = 100%
X = 99%
200 °C
220 °C
A
furfural
B
5-HMF
10
0
20
40
60
80
100
120
0
20
30
40
50
60
time [min]
time [min]
Fig. 5 – Time dependence of the selectivity of the furfural and 5ꢀHMF formation: A.
Selectivity of the furfural formation from xylose (aqueous xylose solution (56 mM, 2 x 15 mL))
B. Selectivity of 5ꢀHMF formation from fructose; data adapted from 22 (aqueous fructose
solution (56 mM, 2 x 15 mL)).
The hydrothermal decomposition of xylan provides significantly lower furfural yields
than the decomposition of xylose. At 200 °C only 16% of the starting material is
converted to furfural – as compared to 49% from xylose. This corresponds to a
difference of 33% based on the initial number of xyloseꢀunits. Figure 6A shows the
difference in the furfural yields in dependence of the reaction time at two exemplary
temperatures.
200 °C
xylose
B
A
50
40
30
20
10
0
50
40
30
20
10
0
220 °C
xylose
xylan
200 °C
xylan
220 °C
180
200
220
240
0
20
40
60
80 100 120
temperature [°C]
time [min]
Fig. 6 – Furfural yield from different starting materials (xylose, xylan): A. Time dependence.
ꢀ 12 ꢀ

RSC Advances
Page 14 of 20
View Article Online
DOI: 10.1039/C3RA43108H
B. Temperature dependence at 10 min reaction time (conditions: aqueous xylose solution
(56 mM, 2 x 15 mL), xylan suspension (2 x (125 mg xylan in 15 mL))).
Here it is remarkable, that the maximal furfural yield from xylan is reached at shorter
reaction times than from xylose. Since the hydrolytic cleavage of xylan via the
xylooligomers to xylose is required as preꢀstage for the formation of furfural, a shift to
a longer reaction time was expected. However, at 220 °C, a maximal xylose yield of
only 7% is already detected at 5 min reaction time, whereas xylan, which has to
undergo a hydrolytic depolymerization, is converted within 10 min (previous section,
figure 2 (C, D)). Since only a low amount of xylose is available for furfural formation,
the furfural decomposition becomes more pronounced at an early reaction stage
during xylan conversion.
Furthermore, figure 6B shows that the maximum furfural yield from xylan at 220 °C is
clearly lower than at 200 °C, while the xylose decomposition results in similar
maximal furfural yields in the investigated temperature range (at different reaction
times). Although at a constant reaction time of 10 minutes the furfural yield from
xylose is still increasing with a temperature rise from 220 °C to 240 °C (Figure 6B),
the xylan derived furfural yield shows a maximum at 220 °C. Consistently, the
temperature dependence of xylose yield (from xylan) shows a maximum at 200 °C
(10 min; data not shown). These observations are reflected by the course of the
furfural yield over time at different reaction temperatures (Figure 7). Lower
temperatures and extended reaction times favor the furfural production as a result of
preferred xylose formation by xylan hydrolysis.
ꢀ 13 ꢀ

Page 15 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
20
18
16
14
12
10
8
6
160 °C
180 °C
200 °C
220 °C
240 °C
4
2
0
0
20
40
60
80 100 120 140 160 180 200 220 240
time [min]
Fig. 7 – Time dependence of the furfural yield from xylan (xylan suspension: 2 x 125 mg
xylan in 15 mL).
In order to elucidate the reaction pathway of the xylan conversion, figures 8 AꢀC
depict the extent of the furfural and the xylose generation as function of the degree of
the xylan conversion. Whereas the maximal xylose yield (18%, at 160 °C) is obtained
at a conversion of 75%, the maximum furfural yield and that of the integrated
xylose/furfural yield is achieved at conversions of 94ꢀ96%. Lower reaction
temperatures provide higher xylose yields already at a medium degree of conversion
(Figure 8A). The only slow decrease of the xylose yield at these temperatures is
caused by an only low xylose decomposition rate, indicated by the respective slow
increase of the furfural yield (Figure 8B).
For the xylan decomposition the impact of the reaction temperature becomes
apparent from the dependence of the furfural yield from the degree of the xylan
conversion (Figure 8B). An increasing conversion of the starting material causes an
enhanced furfural yield, but the total yield is decreasing with the applied temperature.
At temperatures of 200 °C and above, short reaction times are sufficient to convert at
least 90% of the starting material, and extended reaction times lead to reduced
yields.
ꢀ 14 ꢀ

RSC Advances
Page 16 of 20
View Article Online
DOI: 10.1039/C3RA43108H
As shown in figure 8C, the cumulative yield of xylose and furfural at 160 °C exceeds
the yields that are obtained at higher temperatures. Thus, at the lower temperatures
the hydrolysis of xylan to xylose (with the subsequent reaction to furfural) is the
preferred reaction path.
20
A
15
10
5
0
40
50
60
60
60
70
70
70
80
80
80
90
90
90
100
100
100
20
15
10
5
B
160 °C
180 °C
200 °C
220 °C
240 °C
0
40
50
30
C
20
10
0
40
50
conversion [%]
Fig. 8
– A. Xylose selectivity as function of the degree of xylan conversion at different
reaction temperatures. B. Furfural selectivity at different reaction temperatures. C. Sum of
furfuralꢀ and xyloseꢀselectivity at different reaction temperatures.
ꢀ 15 ꢀ

Page 17 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
Especially at increasing temperatures other reaction pathways prevail. Fragmentation
reactions at the end group of the carbohydrate polymer leading to lowꢀmolecular
degradation products, as reported for cellulose, 32 may also occur at the xylan chain.
Similar to the decomposition of cellulose, 17, 33 a dominant side reaction appears to
be intramolecular dehydration of xylan, leading to a solid reaction residue. In the
experiments in this study this was visible by means of a clouding of the reaction
solution caused by suspended particles. As a result, the total organic carbon (TOC)
of the supernatant is observed. Thus, reactions conducted at 240 °C compared to
220 °C (10 min reaction time) provoke a decrease of 15% TOC. In addition
increasing reaction temperatures accelerate the formation of gaseous products.
Pinkowska observed an increase of gaseous products in the temperature range of
180 – 240 °C up to 40wt% (based on weighꢀout quantity of xylan). 19
4. Conclusion & Outlook
Good furfural yields (max. 49%) are achieved by the uncatalyzed hydrothermal
conversion of xylose within reaction times of 10 – 120 min, and in a temperature
range of 200 – 220 °C. For these conditions, reaction rate constants of 4.5*10ꢀ4 sꢀ1
(200 °C) and 1.2*10ꢀ3 sꢀ1 (200 °C) are determined. A decline of the furfural yield
caused by decomposition or condensation/polymerization reactions becomes
prevailing after full xylose conversion is reached.
In comparison to the xylose conversion the decomposition of its polymer xylan
provides low furfural yields, since the hydrolysis of xylan is reduced at increasing
temperature. Therefore a catalytic (e.g. acidic or enzymatic) hemicellulose
pretreatment at low reaction temperatures is advisable in order to trigger the
hydrolysis to xylose – prerequisite for an enhanced furfural yield. Further, for real
biomass the complex interplay of hemicellulose with lignin and cellulose has to be
taken into consideration.
ꢀ 16 ꢀ

RSC Advances
Page 18 of 20
View Article Online
DOI: 10.1039/C3RA43108H
Acknowledgements
U.S. acknowledges the foundation of the professorship Sustainable Chemistry and
Energy Research by the Volkswagen AG and the Verband der Deutschen
Biokraftstoffindustrie e.V.
ꢀ 17 ꢀ

Page 19 of 20
RSC Advances
View Article Online
DOI: 10.1039/C3RA43108H
5. References
1.
2.
3.
B. Kamm, Angew. Chem., Int. Ed., 2007, 46, 5056ꢀ5058.
A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411ꢀ2502.
J.ꢀP. Lange, E. van der Heide, J. Van Buijtenen and R. Price, ChemSusChem
,
2012, 5, 150ꢀ166.
4.
5.
R.ꢀJ. van Putten, J. C. van der Waal, E. de Jong, C. B. Rasrendra, H. J.
Heeres and J. G. de Vries, Chem. Rev., 2013, 113, 1499ꢀ1597.
B. F. Sels, M. Dusselier, P. Van Wouwe, A. Dewaele and E. Makshina, Energy
Environ. Sci., 2013 , , 1415ꢀ1442.
W. de Jong and G. Marcotullio, Int. J. Chem. Reactor Eng., 2010,
K. B. Bastawde, World J. Microbiol. Biotechnol., 1992, , 353ꢀ368.
6
6.
7.
8.
8.
8
G. Garrote, H. Dominguez and J. C. Parajo, Holz Roh Werkst., 1999, 57, 191ꢀ
202.
9.
K. Wang, H. Yang, X. Yao, F. Xu and R.ꢀc. Sun, Bioresour. Technol., 2012,
116, 99ꢀ106.
10.
A. S. Mamman, J.ꢀM. Lee, Y.ꢀC. Kim, I. T. Hwang, N.ꢀJ. Park, Y. K. Hwang, J.ꢀ
S. Chang and J.ꢀS. Hwang, Biofuels, Bioprod. Biorefin., 2008, 2, 438ꢀ454.
11.
12.
13.
M. Kleinert and T. Barth, Chem. Eng. Technol., 2008, 31, 736ꢀ745.
H. V. Scheller and P. Ulvskov, Annu. Rev. Plant Biol., 2010, 61, 263ꢀ289.
M. Möller, P. Nilges, F. Harnisch and U. Schröder, ChemSusChem, 2011,
566ꢀ579.
4,
14.
15.
P. M.ꢀA. MäkiꢀArvela, T. Salmi, B. Holmbom, S. Willför and D. Y. Murzin,
Chem. Rev., 2011, 111, 5638ꢀ5666.
K. Lamminpää, J. Ahola and J. Tanskanen, Ind. Eng. Chem. Res., 2012, 51
,
6297ꢀ6303.
16.
17.
18.
G. Marcotullio and W. de Jong, Green Chem., 2010, 12, 1739ꢀ1746.
M. Möller, F. Harnisch and U. Schröder, RSC Adv., 2013.
X. Qi, M. Watanabe, T. M. Aida and R. L. Smith Jr, Catal. Commun., 2008, 9,
2244ꢀ2249.
19.
20.
21.
22.
23.
24.
H. Pinkowska, P. Wolak and A. Zlocinska, Biomass Bioenergy, 2011, 35,
3902ꢀ3912.
C. O. Kappe, D. Dallinger and S. Murphree, Practical Microwave Synthesis for
Organic Chemists, WileyꢀVCH Weinheim, Germany, 2009.
A. Sobhy and J. Chaouki, Chemical Engineering Transactions, 2010, 19, 25ꢀ
30.
M. Möller, F. Harnisch and U. Schröder, Biomass Bioenergy, 2012, 39, 389ꢀ
398.
J. M. Kremsner and C. O. Kappe, Eur. J. Org. Chem., 2005, 2005, 3672 ꢀ
3679.
A. Sluiter, B. Hames, R. Ruiz, C. Scarlata, J. Sluiter and D. Templeton,
Technical Report 2006, NREL/TP-510-42623
25.
26.
Q. Jing and X. LÜ, Chinese J. Chem. Eng., 2007, 15, 666ꢀ669.
A. Mittal, G. Chatterjee Siddharth, M. Scott Gary and E. Amidon Thomas,
Holzforschung, 2009, 63, 307.
27.
28.
J. Shen and C. E. Wyman, Bioresour. Technol., 2011, 102, 9111ꢀ9120.
M. J. Antal, T. Leesomboon, W. S. Mok and G. N. Richards, Carbohydr. Res.
1991, 217, 71ꢀ85.
,
29.
30.
P. J. Oefner, A. H. Lanziner, G. Bonn and O. Bobleter, Monatsh. Chem., 1992,
123, 547ꢀ556.
A. P. Dunlop, Ind. Eng. Chem., 1948, 40, 204ꢀ209.
ꢀ 18 ꢀ

RSC Advances
Page 20 of 20
View Article Online
DOI: 10.1039/C3RA43108H
31.
G. Marcotullio, M. A. T. Cardoso, W. De Jong and A. H. M. Verkooijen, Int. J.
Chem. Reactor Eng., 2009,
K. Ehara and S. Saka, Cellulose, 2002,
7.
32.
33.
9, 301ꢀ311.
C. Falco, N. Baccile and M.ꢀM. Titirici, Green Chem., 2011, 13, 3273ꢀ3281.
ꢀ 19 ꢀ