Hydrothermal processing, as an alternative for upgrading agriculture
residues and marine biomass according to the biorefinery concept: A review
He
´
ctor A. Ruiz
n
, Rosa M. Rodrı
´
guez-Jasso, Bruno D. Fernandes, Anto
´
nio A. Vicente, Jose
´
A. Teixeira
IBB-Institute for Biotechnology and Bioengineering, Centre of Biological Engineering, University of Minho, Campus de Gualtar, 4710-057 Braga, Portugal
article info
Article history:
Received 5 August 2011
Received in revised form
22 November 2012
Accepted 26 November 2012
Available online 22 January 2013
Keywords:
Biorefinery
Autohydrolysis
Lignocellulosic material
Macroalgae
Microalgae
Biofuels
abstract
The concept of a biorefinery that integrates processes and technologies for biomass conversion
demands efficient utilization of all components. Hydrothermal processing is a potential clean
technology to convert raw materials such as lignocellulosic materials and aquatic biomass into
bioenergy and high added-value chemicals. In this technology, water at high temperat ures and
pressures is applied for hydrolysis, extraction and structural modification of mater ials. This review is
focused on providing an updated overview on the fundamentals, modelling, separation and applications
of the main components of lignocellulosic materials and conversion of aquatic biomass (macro- and
micro- algae) into value-added products.
& 2012 Elsevier Ltd. All rights reserved.
Contents
1. Introduction .......................................................................................................35
2. Hydrothermal processing of lignocellulosic material . . . ....................................................................36
2.1. Fundamentals and operating conditions of hydrothermal processing ....................................................36
2.2. Modeling of hydrothermal processing. ............................................................................38
2.3. Effect of hydrothermal processing on cellulose . . ...................................................................40
2.4. Effect of hydrothermal processing on hemicellulose .................................................................41
2.5. Effect of hydrothermal processing on lignin ........................................................................42
3. Hydrothermal processing of aquatic biomass .............................................................................43
3.1. Hydrothermal processing of macroalgae...........................................................................43
3.2. Hydrothermal processing in microalgae ...........................................................................45
4. Conclusions .......................................................................................................47
Acknowledgements .....................................................................................................47
References . . . .........................................................................................................47
1. Introduction
Considering the amount of biomass available, there is a clear
opportunity to develop commercial processes that could generate
products needed at very high volumes and low selling price. Most of
such products are now bein g made from non-renewable resources,
mainly through oi l refineries. T hese refineries, starting from a
complex mixture (petroleum), use a wide ra nge of u nit operations
to generate an impressive variety of products that are sold directly or
transformed into value-added products such plastics and fibers.
Approximately 17% of the volume of products deri ved fro m petro-
leum in the US is classified as chemicals [1]. If these chemicals could
be obtained from renewable resources (e.g., biomass in a biorefinery),
it would reduce petroleum dependence while also having a positive
environmental impact. In recent years the use of different renewable
Contents lists available at SciVerse ScienceDirect
journal homepage: www.elsevier.com/locate/rser
Renewable and Sustainable Energy Reviews
1364-0321/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.rser.2012.11.069
n
Corresponding author. Tel.: þ351 253 604 400; fax: þ351 253 604 429.
E-mail addresses: hector_ruiz@deb.uminho.pt,
domitilah@hotmail.com (H.A. Ruiz).
Renewable and Sustainable Energy Reviews 21 (2013) 35–51
raw materials (lignocellulose material (LCMs) and aquatic marine
materials) has been a growing trend for different applications and
products such as energy, fuels, chemical, cosmetics, medical applica-
tions, construction materials and high added-value products for food
or feed. The term ‘‘biorefinery’’ borrows its origin from the classical
petroleum refinery concept and refers to biomass conversion into
fuels and chemicals with high added-value through the integration of
clean processes [2,3]. Several technologi es have been developed
during the last decades that allow this conversion process to occur,
the clear objective being now to make this process cost-competitive
in today’s markets. Hydrothermal processing is an alternative for the
fractionation of these raw materials (LCMs and aquatic biomass). The
fractionation refers to the conversion into its main constituent: LCMs
(i.e., cellulose, hemicellulose and lignin); macroalgae (different poly-
saccharides depending of taxonomic groups); microalgae (oils, pro-
teins, carbohydrates) [4–6]. Bobleter et al. [7] pioneered in using
water for pretreatment to enhance the susceptibility of lignocellulosic
material to enzymatic hydrolysis. The processes with liquid water
under high temperature and pressure are also called autohydrolysis,
hydrothermal treatment, hot compressed water (HCW), hydrother-
molysis, liquid hot water (LHW), aquasolve process, aqueous proces-
sing and pressure-cooking in water [8–24]. The objective of this
review is to present research progresses in hydrothermal processing
of lignocellulosic materials and aquatic biomass for the fractionation
of their main components. The fundamentals, mathematical model-
ing, effects of hydrothermal processing on cellulose, hemicellulose
and lignin and their applications are reviewed. Additional information
on the application of this technology to aquatic biomass (macro- and
micro- algae) is also provided.
2. Hydrothermal processing of lignocellulosic material
Hydrothermal processing has been considered a cost-effective
pretreatment [25] and in general, the major advantages that this
process offers are: (1) the process does not require the addition
and recovery of chemicals different from water, (2) limited
equipment corrosion problems, (3) simple and economical opera-
tion [26–29]. For that reason, the hydrothermal processing can be
considered an environmentally friendly fractionation process
[30]. Lignocellulosic materials (LCMs) are the most abundant
renewable biomass and its annual production was approximately
estimated in 200 billion metric tons worldwide in 2007 [31].
LCMs are mainly composed of cellulose, hemicellulose and lignin
and have great potential as cheap and renewable feedstock for
different applications. In general, LCMs include agricultural resi-
dues, herbaceous, hardwood, softwood, cellulose wastes and
industry co-products. Table 1 shows the composition of different
lignocellulose materials. The fractionation of LCMs into products
derived from their structural components is an attractive possi-
bility leading to the biorefinery concept. However, the main
problem of fractionation is the recalcitrant nature of these
materials. Fractionation may be achieved through hydrothermal
processing, whose first step is hemicellulose solubilization. Fig. 1
shows the scheme of a biorefinery using hydrothermal proces-
sing. This process has been mainly used as a pretreatment for
bioethanol production; in recent researches, the use of a sequen-
tial process has been applied as an alternative of papermak-
ing production, also as a technology for converting agro-food
by-products into useful food ingredients [12, 65–71].
2.1. Fundamentals and operating conditions of hydrothermal
processing
In hydrothermal processing LCMs are exposed to water in the
liquid state, at elevated temperature and pressures, that penetrated
cell’s structures, hydrates cellulose, depolymerizes hemicellulose (to
oligomers and monomers) being between 40% and 60% of the total
biomass dissolved in the process (
Fig. 2). In water at high tempera-
tures (150–230 1C), the H-bonding starts weakening, allowing
Table 1
Composition of selected lignocellulosic materials (% dry matter).
Raw material Cellulose (%) Hemicellulose (%) Lignin (%) References
Agricultural residues
Corn cobs 38.8–44 33–36.4 13.1–18 Liu et al. [32]; Wang et al. [33]
Corn stover 34.32–36.5 20.11–31.3 11.9–13.55 Weiss et al. [34]; Liu and Cheng [35]
Wheat straw 33–40 20–33.8 15–26.8 Ruiz et al. [12]; Talebnia et al. [36]
Rice straw 35–36.6 16.1–22 12–14.9 Hsu et al. [37]; Yadav et al. [38]
Sugar cane bagasse 34.1–49 15.79–29.6 19.4–27.2 Mesa et al. [39]; Maeda et al. [40]
Barley straw 37.5 25.1–37.1 15.8–16.9 Sun et al. [41]; Garcı
´
a-Aparicio et al. [42]
Rice husk 33.43 20.99 18.25 Garrote et al [43]; Abbas and Ansumali [44]
Rye straw 41.1–42.1 23.8–24.4 19.5–22.9 Ingram et al. [21]; Gullo
´
n et al. [45]
Rapessed straw 36.59–37 19.6–24.22 15.55–18 Dı
´
az et al. [13]; Lu et al. [46]
Sunflower stalks 33.8 20.2–24.27 14.6–19.9 Ruiz et al. [47]; Caparro
´
s et al. [48]
Sweet sorghum bagasse 41.33–45.3 22.01–26.3 15.2–16.47 Zhang et al. [49]; Goshadrou et al. [50]
Herbaceous
Switchgrass 41.2–32.97 25.95–31.1 17.34–19.1 Keshwani and Cheng [51]; Hu et al. [52]
Alfalfa stems 24.7 14.7 14.9 Ai and Tschirner [53]
Coastal Bermuda grass 25.59 19.29 19.33 Wang et al. [54]
Hardwood
Aspen 43.8 18 20.8 Tian et al. [55]
Hybrid Poplar 48.95 21.73 23.25 Pan et al. [56]
Eucalyptus 44.6 21.4 30.1 Gonzalez et al. [57]
Eucalyptus globulus 44.4 21.8 27.7 Romanı
´
et al. [9]
Softwood
Pinus radiata 45.3 22.5 26.8 Araque et al. [58]
Spruce 43.8 20.8 28.83 Shafiei et al. [59]
Cellulose wastes
Newspapers 60.3 16.4 12.4 Lee et al. [60]
Recycled paper sludge 60.8 14.2 8.4 Peng and Chen [61]
Industry co-products
Distiller’s grains 12.63 16.9 – Kim et al. [62]
Brewer’s spent grain 18.8–20.97 15.18–32.8 21.7–25.62 Carvalheiro et al. [63]; Pires et al. [64]
H.A. Ruiz et al. / Renewable and Sustainable Energy Reviews 21 (2013) 35–5136
autoionization of water into acidic hydronium ions (H
3
O
þ
)thatactas
catalysts and basic hydroxide ions (OH
). In the subcritical region
(100–374 1C) the ionization constant (K
w
) of water increases with
temperature. However, when exceeding its critical point (374 1Cand
22.1 MPa), the values of dielectric constant, ionization constant (K
w
)
and ionic product of water drop drastically. Moreover, hydronium
ions are generated from organic acids, mainly acetic acid from acetyl
groups and uronic acid [28,72–75]. Acetyl groups are present in LCMs
and as they are associated with hemicellulose, the hydration of the
acetyl groups leads to the acidificatio n of the liquor and thus,
formation of hydrogen ions. A number of hypotheses have been
suggested to explain this phenomenon. According to these considera-
tions, in a recent work, Liu [76] presented the following model for
hemicellulose solubilization in hydrothermal processing.
H
2
O2H
þ
þOH
ð1Þ
H
þ
þH
2
O2H
3
O
þ
ð2Þ
RPOA
C
þH
þ
2RPOA
C
n
H
þ
ð3Þ
RPOA
C
n
H
þ
þH
2
O2RPOH
n
H
þ
þHOA
C
ð4Þ
RPOA
C
n
H
þ
þH
2
O2ROH
n
H
þ
þHPOA
C
ð5Þ
HOA
C
2H
þ
þOA
C
ð6Þ
Fig. 2. Batch reactor systems for hemicellulose depolymerization in hydrothermal processing.
Fig. 1. Scheme of a biorefinery using hydrothermal processing and LCMs as raw material.
H.A. Ruiz et al. / Renewable and Sustainable Energy Reviews 21 (2013) 35–51 37
RPOH
n
H
þ
2RPOH þH
þ
ð7Þ
ROH
n
H
þ
2ROH þ H
þ
ð8Þ
Reaction steps to solubilize hemicellulose
RX
n
OHþH
þ
2RX
n
OH
n
H
þ
ð9Þ
RX
n
OH
n
H
þ
þH
2
O2RX
m
OH
n
H
þ
þHX
S
OH ð10Þ
Reaction steps to reduce the chain length inside the hydrothermal
process liquor
HX
n
OHþH
þ
2HX
n
OH
n
H
þ
ð11Þ
HX
n
OH
n
H
þ
þH
2
O2HX
m
OH
n
H
þ
þHX
S
OH ð12Þ
where X
n
represents an n-xylooligomer middle group, mþs¼n,
R- denotes the cellulose or lignin bonded to LCMs, P represents a
segment of hemicellulose, HX
n
OH is an n-xylooligomer, HOA
C
represents the acetic acid molecule and A
C
is CH
3
CO [76]. On the
other hand, the most important operational variables of hydro-
thermal processing include temperature, residence time, particle
size, moisture content (ratio liquid/solid) and pH influence on the
fractionation of LCMs and must be taken into consideration to
maximize the product yield (i.e., hemicellulose sugar production,
accessible surface area for enzymatic saccharification, etc.).
The relationship between temperature-time strongly influences
the physical–chemical characteristics of LCMs in hydrothermal
processing. Ballesteros et al. [77] reported an increase of
hemicellulose-sugar degradation at higher temperatures and
residence times, concluding that at more severe operational
conditions there are more losses of hemicellulosic sugar. For this
reason, a strict control is required for high temperature reactions
due to thermal degradation. Several works showed that the
products (pentose and oil yield) from hydrothermal processing
are favored at lower reaction temperatures and longer residence
times [78–80]. Normally, when larger particle sizes are used, heat
transfer problems lead to overcooking of the exterior (with
consequent formation of inhibitors) and incomplete autohydro-
lysis of the interior. This problem can be overcome by reducing
particle size as the first pretreatment step. This size reduction
process not only changes the particle size and shape, but also
increases bulk density, improves flow properties, increases por-
osity, increases surface area and is usually required to make
material handling easier before hydrothermal processing. The
higher surface area increases the number of contact points for
chemical reaction [12,81]. Mosier et al. [82] reported that size
reduction is not needed since the lignocellulose particles break
apart when cooked in water. Ballesteros et al. [83] showed that
the utilization of very small chips of softwood in hydrothermal
processing would not be desirable to optimize the effectiveness of
the process and improve economy, due to the significant energy
requirements of particle reduction process. However, in recent
work, Hosseini and Shah [84] reported that it is possible to
improve in 50% the energy efficiency of pretreatment by the
optimization of particle size properties. According to Ruiz et al.
[12], the use of blends with different particle size distributions
has a selective influence over the sugar extraction: thus, the use
of a blend with defined percentages of the various particle sizes is
recommended before carrying out a hydrothermal processing.
Moisture content and ratio liquid/solid may also greatly influence
the ability of heat and chemicals (H
3
O
þ
) to penetrate LCMs,
causing an uneven treatment of material. An uneven treatment
can potentially result in the selective degradation of the outer
portion of the LCMs, while at the same time the interior is less
affected by the treatment [81]. Cullis et al. [85] reported that the
moisture content has a dramatic effect on the efficacy of the
hydrothermal processing as a substantial decrease in the amount
of hemicellulose-derived carbohydrates recovered in the water-
soluble fraction was observed when increasing the starting
moisture content from 12 to 30%, Rodrı
´
guez et al. [86] showed
that it is possible to obtain high glucose, xylose, arabinose and
acetic acid concentrations by combining high temperatures with a
medium-low treatment time and liquid/solid ratio. On the other
hand, the formation of hydronium ions from water and from
organic acids is an important factor during hydrothermal proces-
sing, since the LCMs and water mixture will reach high tempera-
tures and pressures during the process. These high temperatures
and pressures will accelerate the acid-catalyzed hydrolysis of
cellulose and hemicellulose as well as the acid-catalyzed degra-
dation of glucose and xylose. Monitoring and control of the pH in
hydrothermal processing will maximize the solubilization of the
hemicellulose fraction as liquid soluble oligosaccharides while
minimizing hydronium ions concentration and, more impor-
tantly, the degradation of these oligosaccharides and monosac-
charides to degradation products [29,87]. Mosier et al. [88]
pretreated corn fiber using pH controlled liquid hot water at
160 1C and a pH value above 4.0 and found that 50% of the fiber
was dissolved in 20 min. The carbohydrates dissolved by the
pretreatment were 80% soluble oligosaccharides and 20% mono-
saccharides with o 1% of the carbohydrates lost to degradation
products. Cara et al. [89] reported a slight pH decrease of
hydrothermal processing hydrolyzates, in the range of 3.8 to
3.3, and an increase of degradation product concentrations
(furfural) from 0.4 to 1.7 g/L, respectively.
In hydrothermal processing there are different types of reactor
configurations. (1) Batch reactor: LCMs solid particles are mixed
with water in the reactor (Fig. 3A). The residence time of the
reacting solid is long [8,16,90–92]. In a recent work, Gullo
´
n et al.
[45] reported a conversion of 69.2% from initial xylan into
xylooligosaccharides using a batch reactor configuration at
208 1C and rye straw as raw material; (2) Semi-continuous reactor
of (flow-through partial flow-through): hot water is passed over a
stationary bed of LCMs and dissolves lignocellulose components
while the liquid products are rapidly swept out (Fig. 3B). The
residence time of liquid products is short, compared to a batch
reactor [16,82,93–98]. Liu and Wyman [99] reported that in this
type of reactors the fluid velocity in flow through has a significant
impact on hydrothermal processing. Increasing fluid velocity
significantly accelerated solubilization of total mass, hemicellu-
lose and lignin even at the similar liquid residence times; (3)
Continuous reactor (co-current, counter-current): the LCMs are
passed in one direction while water is passed in the same or
opposite direction (Fig. 3C and D). A continuous reactor system is
also typically required to operate at high temperatures and
pressures to achieve a high conversion of the feedstock within a
short residence time [16,82,100–102]. Makishima et al. [103]
reported a 82% conversion of xylan in xylose and xylooligosac-
charides using a continuous flow type reactor. Yu and Wu [16]
suggested that the characteristics of liquid products are strongly
influenced by the reactor configuration.
2.2. Modeling of hydrothermal processing
Modeling in hydrothermal processing provides a way to
compare results from experiments carried out at different condi-
tions. Table 2 shows the main mathematical models used in both
isothermal or non-isothermal hydrothermal processing (Fig. 4A
and B) [69,104]. An often used option to modeling the effects of
the main operational variables by pseudo first order kinetics is the
severity factor (R
0
) proposed by Overend and Chornet [105] and
Chornet and Overend [106]. This empirical model has been
H.A. Ruiz et al. / Renewable and Sustainable Energy Reviews 21 (2013) 35–5138
Fig. 3. Representation of different reactor configurations for hydrothermal processing. (A) batch; (B) semi-continuous (flow-through reactor); (C) continuous (co-current);
(D) continuous (counter-current)(adapted from Mosier et al. [82]; Yu and Wu [16]).
Table 2
Models used in hydrothermal processing.
Effect Model Variables Reference
Severity factor R
0
, easy way for comparing
results among experiments carried out under
different conditions of temperature and time.
R
0
¼
R
t
0
exp
T100
14:75
dt
t is the reaction time (min), T is temperature (1C),
100 is the temperature of reference and 14.74 is
an empirical parameter related with activation
energy, assuming pseudo first order kinetics. The
results are usually represented as a function of log
(R
0
).
Overend and
Chornet [105];
Chornet and
Overend [106]
H-factor (HF), is a relationship between time and
temperature, which is only an approximation
of the reaction due to the fact that the proton
concentration changes with time and
activation energy. The concept of HF was
development for Kraft/chemical pulping.
However, also has been applied for
hydrothermal processing Griebl et al. [64].
HF ¼
R
t
0
exp 43:186
16115K
T
dt
t is the time (min), T is the temperature in (1C) and
the constants are related with the activation
energy.
Liu et al. [76]
Severity factor R
0
, in a non-isothermal
hydrothermal processing, which includes the
combination of temperature and reaction time
along heating and cooling.
logR
0
¼ log R
0 Heating
þR
0 Cooling
logR
0
¼
Z
t
MAX
0
TðtÞ100
o
dt þ
Z
t
F
t
MAX
T
0
ðtÞ100
o
dt
t
MAX
(min) is the time needed to achieve
maximum autohydrolysis temperature, t
F
(min) is
the time needed for the whole heating-cooling
period, T(t) and T
0
(t) stand for the temperature
profiles in heating and cooling, respectively and
o
is an empirical parameter.
Romanı
´
et al.
[9]
Model that explains the severity factor in
function of chip size and processing time
taking into account the diffusion of liquid into
LCMs.
R
0
¼
t
rr
2
ðð10:5ln jÞ=2MjD
D
CÞ
n
e
ðT100ðÞ=14:75Þ10
pH
t is the time of reaction (min), T is the temperature
(1C), D, is the diffusion coefficient,
r
is the density
of the fluid,
f
is thevoid fraction (porosity), r is the
particle radius (mm), M is the molecular weight
and
D
C is the concentration gradient.
Hosseini et al.
[84]
Relationship between the severity factor and the
viscosity of slurries made from sewage sludge
during hydrothermal processing.
m
¼ 2:755 10
5
R
0:8250
0
m
is viscosity (Pa s/s) and R
0
is the severity
parameter.
Yanagida et al.
[107]
Model that explains the time needed for the
chips to reach the desired temperature of
wood with round or square cross section 1 and
rectangular cross section 2 in hydrothermal
processing.
ð1Þ t ¼ aT
ht
ðÞ
b
T
ctr
ðÞ
c
T
init
ðÞ
d
D
e
M
f
G
g
ð2Þ t ¼ aT
ht
ðÞ
b
T
ctr
ðÞ
c
T
init
ðÞ
d
ðTHÞ
d
W
f
M
g
G
h
t is the time estimated (min) for the center reach
target temperature, T
ht
is the heating temperature
(1F), T
ctr
is the target center temperature, T
init
is the
initial wood temperature (1F), D is de diameter of
round cross section (in), TH is the thickness of
rectangular board (in), W is the width of
rectangular board (in), M is the moisture content
(%), G is the specific gravity, a–h are the regression
coefficients.
Simpson [108]
Model for calculating the time needed for water
diffusion into the LCMs as a function of the
process and LCMs characteristics (assuming
that LCMs have a porous structure) in
hydrothermal process.
t
w
¼
rr
2
10:5ln j
ðÞ
2MjD
D
C
Where
r
is the density, ris the particle,
f
is the
porosity, D is the diffusion coefficient, M is the
molecular weight and
D
C is the concentration
gradient.
Hosseini et al.
[109]
Model for calculating the temperature needed
for different particle sizes. Temperature as a
function of severity factor and radius in
hydrothermal processing.
T
2
¼ T
1
14:7ln
R
2
R
1
r
1
r
2
2
T
2
is the temperature (1C) needed for compensate
the particle size increase, T
1
is the initial LCMs
temperature which can be assumed as 20 1CC r
1
and r
2
are the radius (cm), R
1
and R
2
are de
severity factor.
Hosseini et al.
[109]
H.A. Ruiz et al. / Renewable and Sustainable Energy Reviews 21 (2013) 35–51 39
![Fig. 4. Heating and cooling temperature profiles corresponding to: (A)non-isothermal and (B) isothermal regimen in hydrothermal processing(adapted from Romanı́ et al. [104]; Ruiz et al. [69]).](/figures/fig-4-heating-and-cooling-temperature-profiles-corresponding-134f6rp2.webp)
