Bioresource Technology 128 (2013) 297–304

Contents lists available at SciVerse ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Pilot-scale study on the acid-catalyzed steam explosion of rice straw using a

continuous pretreatment system
Wen-Hua Chen ⇑, Chia-Chin Tsai, Chih-Feng Lin, Pei-Yuan Tsai, Wen-Song Hwang
Chemistry Division, Institute of Nuclear Energy Research, Taoyuan, Taiwan, ROC

h i g h l i g h t s

" A pilot-scale continuous pretreatment system was successfully developed.

" The pretreatment system shows good operational stability and durability.
" 73% of the total saccharification yield was obtained.
" The system can be used for the production of bioethanol and bio-based chemicals.
" The total sugar yield of rice straw in different scale system was compared.

a r t i c l e i n f o a b s t r a c t

Article history: A continuous acid-catalyzed steam explosion pretreatment process and system to produce cellulosic eth-
Received 1 August 2012 anol was developed at the pilot-scale. The effects of the following parameters on the pretreatment effi-
Received in revised form 27 September ciency of rice straw feedstocks were investigated: the acid concentration, the reaction temperature, the
2012
residence time, the feedstock size, the explosion pressure and the screw speed. The optimal presteaming
Accepted 7 October 2012
Available online 2 November 2012
horizontal reactor conditions for the pretreatment process are as follows: 1.7 rpm and 100–110 °C with
an acid concentration of 1.3% (w/w). An acid-catalyzed steam explosion is then performed in the vertical
reactor at 185 °C for 2 min. Approximately 73% of the total saccharification yield was obtained after the
Keywords:
Pilot-scale
rice straw was pretreated under optimal conditions and subsequent enzymatic hydrolysis at a combined
Pretreatment severity factor of 0.4–0.7. Moreover, good long-term stability and durability of the pretreatment system
Rice straw under continuous operation was observed.
Acid-catalyzed steam explosion Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Various pretreatment strategies have been examined, including

Lignocellulosic materials, such as agricultural, hardwood and
softwood residues, are potentially viable sources of sugars for the
production of bioethanol, biobutanol, other biofuels or various
valuable chemicals. Lignocellulosic materials are particularly
attractive because they do not compete with food crops. In biomass
processing, the primary goal is to remove the hemicelluloses and
destroy the structure of the biomass so that cellulose is more
accessible to enzymatic hydrolysis. This process is generally re-
ferred to as pretreatment, and it is one of the most difficult pro-
cesses to optimize. Therefore, pretreatment is a central part of
the lignocellulose-to-ethanol process.
⇑ Corresponding author. Address: Chemistry Division, Institute of Nuclear Energy
Research, Atomic Energy Committee, Executive Yuan, No. 1000, Wenhua Rd., Jiaan
Village, Longtan Township, Taoyuan County 32546, Taiwan, ROC. Tel.: +886 3
4711400x5115.
E-mail address: wenhua@iner.gov.tw (W.-H. Chen).
dilute-acid hydrolysis, steam explosion, liquid hot water extraction,
alkaline hydrolysis, ammonia treatment and various biological
processes. Among these various types of pretreatment, dilute-acid
hydrolysis and steam explosion has been widely tested in pilot
scale equipment and is considered to more favorable method for
industrial applications (Wyman et al., 2005; Alvira et al., 2010;
Chandel et al., 2011; Modenbach and Nokes, 2012; Larsen et al.,
2012). Moreover, addition of a mineral acid results in more efficient
hemicelluloses hydrolysis and the subsequent enzymatic digestion
during steam pretreatment. Therefore, the acid-catalyzed steam
explosion process is considered to be close to commercialization
(Galbe and Zacchi, 2012).
An estimated 600–900 million tons of rice straw are produced
globally every year. Moreover, developing countries in Asia pro-
duce 90% of the world’s rice straw (Sarkar et al., 2012). Therefore,
rice straw is an attractive lignocellulosic material for biofuels and
other useful biomass chemicals. Nevertheless, various characteris-
tics of rice straw make it intractable. For example, rice straw is soft,

0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biortech.2012.10.111
298 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304
hydrophobic and non-homogeneous. It also has a low bulk density
and tends to wind around the instrument and stick together during
processing. Hence, a detailed understanding of rice straw in a pilot-
scale continuous pretreatment system will be crucial in enabling
industrial applications for rice straw. In general, biomass is tradi-
tionally processed via a batch processing method; however, if eth-
anol production from rice straw is to be successful in the
marketplace, a platform for its continuous processing must be
developed.
Although many researchers have studied dilute-acid hydrolysis
or steam explosion pretreatment in a pilot-scale continuous pre-
treatment system (Schell et al., 2003; Thomsen et al., 2006; Weiss
et al., 2010; Fang et al., 2011a,b; Rocha et al.,2012a,b; Shekiro et al.,
2012), less attention has been paid to rice straw in a continuous
pretreatment system. INER has developed pretreatment systems
for different scales on the basis of previous dilute-acid hydrolysis
experiments that were performed at the lab-scale (400 g/batch),
the bench-scale (10 kg/batch) and at a pilot plant (1 ton/day) har-
boring a cellulosic ethanol-testing platform (Chen et al., 2011a,b;
Weng et al., 2011). Out of these in-house technologies, the pilot-
scale continuous pretreatment system uses the acid-catalyzed
steam explosion process to pretreat feedstocks such as rice
straw.
The goal of this study was to define the optimal conditions for
the acid-catalyzed steam explosion of rice straw in the pilot-scale
continuous pretreatment system. The effects of acid concentration,
reaction temperature, residence time, feedstock size, screw speed
and explosion pressure on the pretreatment efficiency of the rice
straw feedstock were evaluated. The pretreatment efficiency was
evaluated by estimating the xylose yield after pretreatment and
the glucose yield from the subsequent enzymatic hydrolysis. More-
over, the continuous operation stabilities and the corresponding
performance will also be reported.
2. Methods
2.1. Biomass materials
The rice straw was primarily collected from private farms in
Taiwan. The composition of the raw rice straw was as follows: glu-
can 32.9 ± 0.1%, xylan 18.7 ± 0.1%, arabinan 3.2 ± 0.1%, extractive
10.1 ± 0.2%, ash 11.3 ± 0.1% and lignin 19.1 ± 1.1%. The composition
was calculated on a dry weight basis.
2.2. Pretreatment system
The pilot-scale continuous pretreatment system consisted of a
feedstock receiving hopper with a weighmetric screw feeder, a
horizontal reactor (presteamer) with a long residence time for
the presteaming, and for the acid impregnation of the biomass, a
vertical high pressure steam explosion reactor, a receiving flash
tank and a solid liquid separator. The unique design of the pretreat-
ment system includes a piston pressure feeding device, special
mechanical equipment and a modified outlet device for steam
explosion. The horizontal and vertical high-pressure reactors are
heated to the desired temperature by the direct injection of 6
and 16 kg/cm2 saturated steam, respectively. The flow meters,
equipped with automatic temperature or pressure compensation,
maintained the desired temperature and pressure. The horizontal
presteamer is connected to a diaphragm pump that adds dilute
acid solution. The acid mixing and preheating times of rice straw
are controlled by the screw speed of the horizontal reactor. Addi-
tionally, the feeding rate of the screw feeder and diaphragm pump
can be used to regulate the solid–liquid ratio of rice straw and di-
lute sulfuric acid solution. An advanced Distributed Control System
(DCS) has been used in the pilot-scale continuous pretreatment
system. It provides continuous and automatic operation during
the pretreatment process. The throughput can vary from 100–
200 kg of dry material per hour, depending on the bulk density
of the raw material and processing conditions.
2.3. Pretreatment process
The acid-catalyzed steam explosion processes used in this pre-
treatment system are shown in Fig. 1. Air-dried rice straw with a
moisture content of 15% was chopped into smaller pieces
(62 cm) with a shredder and used as the raw material. These rice
straw pieces were immediately transferred to the hopper of the
pretreatment system by a pneumatic conveyor system designed
for continuously transporting lignocellulosic materials in the pilot
plant. Subsequently, the rice straw pieces were continuously fed
into the horizontal pressurized reactor with a piston pressure-
feeding device. The rice straw was presoaked with dilute sulfuric
acid and preheated under a selected temperature in the horizontal
reactor. Then, the presoaked rice straw was continuously trans-
ferred to the vertical high-pressure reactor that proceeded the
acid-catalyzed steam explosion. The resultant slurry is automati-
cally continuously transferred to the solid–liquid separator. The
xylose-rich hydrolysate and the cellulose-rich residues were
obtained.
In this study, the throughput was 100 ± 20 kg of rice straw per
hour, and the total feed amount was 1000 ± 50 kg for each test con-
dition. The rice straws were mixed with dilute sulfuric acid solu-
tion to an initial solid loading of about 40–50% (w/w) and
preheated as they passed through the horizontal pressurized reac-
tor at a screw speed of 1.7 rpm and temperature of 100–110 °C. The
test conditions of the acid-catalyzed steam explosion process var-
ied in temperature from 160 to 190 °C, residence times varied from
2 to 10 min and acid concentrations in the acid supply tank varied
from 1.0 to 8.4 wt.%. The effects of the feedstock size (61, 62 cm),
the screw speed of the horizontal reactor (1.7 rpm, 2.5 rpm,
3.4 rpm, 5.1 rpm) and the explosion pressure (10 kg/cm2,
13 kg/cm2, 16 kg/cm2) on the pretreatment efficiency were also
tested.
In general, the hydrolysate and the pretreated residues were
sampled at every 2 h. Samples were analyzed for the sugar concen-
trations of their hydrolysates, the composition of the pretreated
rice straw and the subsequent enzymatic digestibility. Four sam-
ples were taken for each test condition. The pH, concentration, per-
centage of oligomers in the hydrolysate, carbohydrate content of
the pretreated residues and the total enzymatic hydrolysis for each
test run were the average value of the four samples.
2.4. Enzymatic hydrolysis
The unwashed and washed pretreated residues (about 1 g dry
weight) were transferred to a 250 ml shake flask, and the pH was
adjusted to 4.8 by adding 50 mM acetate buffer at a 2% (w/v) solid
loading. The digestion of the cellulose was conducted at a cellulase
(Novozyme celtic CTec2) activity level of 15 FPU/g cellulose. The
enzymatic digestion was carried out at 50 °C and 100 rpm for 24 h.
In this study, the pretreatment efficiency is estimated by com-
bining total xylose and glucose conversion yields after both the
pretreatment and enzymatic hydrolysis processes. The definition
of the total saccharification yield (overall monosugar yield) is the
sum of the monomeric xylose and glucose saccharification yields.
The xylose/glucose saccharification yields are calculated by the fol-
lowing equations
 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304 299

Saturation Saturation
steam steam

Rice Presoaking & preheating Acid-catalyzed steam explosion

straw in the horizontal reactor in the vertical reactor

Dilute acid Liquid Solid

Solid/Liquid separation Enzymatic hydrolysis

Xylose-rich hydrolysate
Glucose yield (E)
Glucose yield (P)
Xylose yield (P) Xylose yield (E)

Total saccharification yield

Fig. 1. The flow charts of the acid-catalyzed steam explosion process for the pilot-scale continuous pretreatment system.

Xylose saccharification yieldð%Þ ¼ ½ðMonomeric xylose ðgÞ released in pretreatment ðPÞ and enzymatic hydrolysis ðEÞÞ=
ðTotal monomeric xylose and glucose ðgÞ in the feedstockÞ  100

Glucose saccharification yieldð%Þ ¼ ½ðMonomeric glucose ðgÞ released in pretreatment ðPÞ and enzymatic hydrolysis ðEÞÞ=
ðTotal monomeric xylose and glucose ðgÞ in the feedstockÞ  100

Total saccharification yieldð%Þ ¼ ½ðMonomeric xylose and glucose ðgÞ released in pretreatment ðPÞ and enzymatic hydrolysis ðEÞÞ=
ðTotal monomeric xylose and glucose ðgÞ in the feedstockÞ  100

2.5. Analytical methods
The identified components are extractive, ash, cellulose, hemi-
cellulose and lignin of the untreated and treated rice straw were
determined according to Laboratory Analytical Procedures from
the National Renewable Energy Laboratory (NREL) (Sluiter et al.,
2005). The carbohydrate content was determined by measuring
the hemicellulose (xylan, arabinan) and cellulose (glucan) compo-
sition of derived sugars. The composition of the hydrolysates after
pretreatment and enzymatic hydrolysis was found to include
cellobiose, glucose, xylose, arabinose, acetic acid and furfural and
measured by high-performance liquid chromatography (HPLC)
using a packed column (Agilent 1200 series HPLC equipped with
a Coregrl-87H3 column from Transgenomic Technologies. The
column was run at 65 °C with 4 mM H2SO4 as the eluent. The flow
rate was 1.0 mL/min). The portion of oligomers in the hydrolysates
was determined by comparing the sugar monomer compositions
before and after a secondary full hydrolysis in sulfuric acid.
3. Results and discussion
3.1. Optimization of the acid-catalyzed steam explosion process
A total of 52 pretreatment runs were completed, including some
replicates. The monomeric xylose and glucose concentrations of
the hydrolysate were 13.1–37.0 g/L and 5.2–18.2 g/L, respectively.
The research addressed the optimization of xylan removal to max-
imize xylose yield during the acid-catalyzed steam explosion pro-
cess. It is also desirable to maintain a large portion of the glucan in
the pretreated residues for the subsequent enzymatic hydrolysis of
cellulosic ethanol production. The selectivity (the monomeric xy-
lose to glucose ratio in the hydrolysate) is an important factor in
optimizing the hydrolysis conditions. The selectivities ranged from
1.5 to 3.2 and increased with the pH of the hydrolysate. Xylan, glu-
can and lignin comprised 3.3–9.8%, 44.5–53.6% and 22.2–30.6% of
the pretreated rice straw, respectively.
Upon acidic pretreatment, xylan and glucan are degraded to
polysaccharides (oligosaccharides) and subsequently to monosac-
charides via autohydrolysis reactions. In general, during biomass
pretreatment with hot water or very dilute acid, a large fraction
of the solubilized hemicellulose is in the form of oligomers. There-
fore, the resultant hydrolysate after the acid-catalyzed steam
explosion pretreatment includes both monosaccharides and oligo-
saccharides. The percentage accounts for total xylose and glucose
as well as their conversion to either monosaccharides or oligosac-
charides. The percentage and conversion are presented in Fig. 2 as
a function of the hydrolysate pH. The oligosaccharide percentage of
xylose and glucose ranged from 7% to 46% and from 27% to 64% in
the hydrolysate, respectively. The oligosaccharide percentage in-
creased as the pH of the hydrolysate increased. (Fig. 2(a)). This
shows that the weaker acid concentration pretreatment (high pH
value) led to a higher concentration of oligosaccharides during
the acid-catalyzed steam explosion process. This can be observed
in Fig. 2(b). Conversions of oligomeric xylose and glucose were
found to increase with the pH of the hydrolysate. In contrast, the
conversion of monomeric xylose and glucose decreased with
increasing pH values. Kumar and Wyman (2008) reported a similar
trend: the selectivity of monomeric xylose formation from oligo-
mers increased with the acid concentration during dilute sulfuric
acid pretreatment. Moreover, xylooligomers degraded faster than
monomeric xylose at all the pH values tested, and the ratio of
the xylooligmer disappearance rate constant to the xylose degra-
dation rate constant increased with decreasing pH. The percentage
of xylan dissolved was 70–85%, whereas the total xylose recovery
was in the range of 30–61%. The total xylose recovery is similar
to the 60% and 12.8–53.7% xylose recovery of wheat straw by
Thomsen et al. (2006) and Fang et al. (2011b), who used a dilute
acid pretreatment at the 50 kg/h and 60 kg/h scale, respectively.
300 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304

(a) monosaccharides. For example, furfural and HMF (5-hydroxy-

80 methyl furfural) are categorized as furan-type inhibitors and are
also known to inhibit enzymes (Palmqvist and Hahn-Hägerdal,
Oligosaccharides percentage (%)

xylose
glucose 2000). Moreover, the presence of a b-glucosidase inhibitor, such
as lignin, is produced by a higher sulfuric acid concentration pre-
60
treatment and results in an altered ratio of glucose and cellobiose.
The previous study reported that glucose yields for the pretreated
rice straw after the acid catalyzed steam explosion were increased
40 from 25.0% to 85.2% after the removal of these inhibitors by wash-
ing (Chen et al., 2011a). In this context, a drastic increase in glucan
conversion was observed between unwashed and washed samples.
20 The unwashed pretreated rice straw was found to have a lower
glucan conversion (50–79%) during enzymatic hydrolysis than
those (77–89%) of washed pretreated rice straw. The glucan con-
version (77–89%) of washed and pretreated rice straw is similar
0
1.0 1.5 2.0 2.5 3.0 to the yield by Fang et al. (2011a) obtained from washed black
spruce. That study employed a continuous steam explosion at the
PH of hydrolysate
60 kg/h and 92 kg/h scale (45–82%). The glucan conversion is also
(b) similar to the level obtained by Weiss et al. (2010) for the washed
0.6
corn stover. This study employed a dilute acid pretreatment at the
(g/g glucan or xylan in raw material)

xylose glucose 30 kg/h scale (52–94%). Additionally, the glucan conversion is

mono- mono- approximately the same level as the conversion Fang et al.
oligo- oligo-
(2011b) obtained for washed wheat straw. This study used a con-
0.4
tinuous steam explosion at the 60 kg/h scale (76–97%). However,
Conversion

as illustrated in Fig. 3, washing pretreated rice straw or including

an inhibitor removal step could improve the enzymatic digestibil-
ity efficiency from 7% up to 27%. The results show that the effect of
0.2 enhancement of enzymatic digestibility by washing process was
increased with an increase in acid concentration (low the hydroly-
sate pH value).
In general, the total saccharification yield (after combining the
0.0 total xylose and glucose yields) following both the pretreatment
and enzymatic hydrolysis processes is an index of the pretreatment
1.0 1.5 2.0 2.5 3.0 efficiency for ethanol bioconversion. The monomeric xylose,
PH of hydrolysate glucose and total saccharification yields for the acid-catalyzed
steam explosion pretreatment (P), and the subsequent enzymatic
Fig. 2. (a) The oligosaccharide percentage in the total xylose and glucose and (b)
digestion of pretreated residues (E) are shown in Fig. 4 as a
conversions of monosaccharides and oligosaccharides versus the pH of the
hydrolysate under different pretreatment conditions. function of the combined severity factor (CSF). The normalized
combined severity factor (CSF = log R0-pH; where R0 ¼
t  exp½ðT r  100Þ=14:75) is a combination of the reaction temper-
100 30
ature (Tr), the residence time (t) of the vertical high pressure steam
explosion reactor and the pH of the obtained hydrolysate during
the acid-catalyzed steam explosion pretreatment. Xylose was
25
Glucan conversion (%)
(enzymatic hydrolysis)

80 mainly obtained from the pretreatment process, whereas glucose

Enhancement (%)

20
60
(g/g glucan and xylan in raw material)

15 1.0 Total sugar:

100 kg/h
Xylose(P) Glucose(P) 100 kg/h;
40 Xylose(E) Glucose(E) 1 kg/batch
10 0.8
0.1 kg/batch
Saccharification yield

20 unwashed
washed 5
0.6

0 0
1.0 1.5 2.0 2.5
0.4
PH of hydrolysate

Fig. 3. The glucose yield (g/g glucan) in terms of enzyme digestibility for unwashed 0.2
and washed pretreated rice straw and the enhancement versus the pH of the
obtained hydrolysate.
0.0

0.0 0.5 1.0 1.5 2.0

As shown in Fig. 3, the glucose yield (g/g glucan) in terms of
enzymatic hydrolysis for the unwashed pretreated rice straw was CSF
between 50% and 79% after 24 h at 15 FPU/g of cellulose, which in- Fig. 4. The total saccharification yield of the acid-catalyzed steam explosion of rice
creased with increasing pH in the hydrolysate. The acid pretreat- straw in the different scale steam-explosion system as a function of the combined
ment is also known to irreversibly produce an inhibitor from severity factor (CSF).
 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304 301

was mainly obtained from enzymatic hydrolysis. The maximum
xylose yield (P) (22%), corresponding to 61% of xylose recovery,
occurred at a combined severity factor range of 1.3. A similar result
was obtained from the study of Schell et al. (2003). They reported
that the optimum xylose recovery (70%) occurred for dilute-acid
pretreatment conditions with a CSF in the range of 1.4–1.7 using a
pilot-scale system (32 kg/h). Moreover, the result suggests the glu-
cose yield (32–58%) from enzymatic hydrolysis (glucose yield (E))
contributes more significantly to the total saccharification yield
than to the xylose yield (6–22%) from pretreatment (xylose yield
(P)). As shown in Fig. 4 and 73% of the maximum total saccharifica-
tion yield was obtained using the acid-catalyzed steam explosion
pretreatment process (as part of the pilot-scale continuous pre-
treatment system) at a combined severity factor range of 0.4–0.7.
A comparison of the total saccharification yield after acid-
catalyzed steam explosion on different scales is also presented in
Fig. 4. The acid-catalyzed steam explosion experiments were car-
ried out on a self-designed bench-scale steam explosion system
at INER. Approximately 1 kg of rice straw presoaked in an appro-
priate amount of sulfuric acid solution (0.5–2% (w/w)) at room
temperature was introduced into a 20 L reactor at the bench-scale
pretreatment system that was subsequently heated. The reactor
heat-up time varied from 10 to 30 s, depending on the pretreat-
ment temperature (160–200 °C). After the desired steaming time
(2–20 min), a ball valve at the bottom of the reactor was used to
rapidly reduce the vessel to atmospheric pressure. As a result,
the optimal operating conditions for the acid-catalyzed steam
explosion pretreatment were carried out at 200 °C for 1 min with
1% H2SO4. Under these conditions, the xylose conversion and enzy-
matic hydrolysis efficiency reached 70% and 93%, respectively.
Moreover, 80% of the total saccharification yield was obtained at
a combined severity factor of 1.3–1.5. In comparison, the optimal
combined severity factor of the pilot-scale system (100 kg/h) was
milder than (CSF = 1.3–1.5) the bench scale system (1 kg/batch)
and the lab scale system (100 g/batch). This can be explained by
the behavior of rice straw under dilute-acid hydrolysis conditions
on different scales. The behavior of the rice straw is altered due
to its capacity, the presoaking process, the reactor design and the
operational (batch or continuous) module.
To summarize, the result suggests that the maximum total
sugar saccharification yield was obtained at a combined severity
factor of 0.4–0.7 after samples were mixed with a 1.3 wt.% acid
solution and preheated to 100–110 °C at a screw speed of
1.7 rpm. Subsequently, samples were subjected to an acid-cata-
lyzed steam explosion process at 185 °C for 2 min.
3.2. Effect of operational parameters on pretreatment efficiency
The pretreatment efficiency is dependent not only on opera-
tional conditions, such as temperature, residence time and acid
concentration, but also on operational parameters, such as feed-
stock size, presoaking time and explosion pressure during the
acid-catalyzed steam explosion. In the report, the effects of the rice
straw size and screw speed of the horizontal reactor were studied.
This study also investigated the effect of the steam explosion pres-
sure from the vertical reactor to the flash tank on the concentra-
tions of the hydrolysate. Additionally, glucan conversion via
enzymatic hydrolysis was also studied.
Table 1 summarizes the operational parameters and results.
Excluding the screw speed at 5.1 rpm (AC-SE-2-5.1-13), the mono-
meric sugar concentrations in the hydrolysate were in the range of
7.6 ± 0.1 g/L to 9.5 ± 0.6 g/L glucose and 20.3 ± 1.1 g/L to
23.4 ± 1.5 g/L xylose. Xylan and glucan contents obtained in the
pretreated residues were measured in the range of 3.6 ± 0.1% to
6.0 ± 1.1% and 45.1 ± 0.3% to 49.3 ± 0.8%, respectively. Normally,
the screw speed is inversely proportional to the residence time of
the rice straw in the horizontal reactor. In this study, the rice straw
was mixed with 1.3% (w/w) dilute sulfuric acid solution at 100–
110 °C as it passed through the horizontal reactor. Therefore, the
accelerated screw speed with a corresponding decrease in resi-
dence time in the horizontal reactor might allow insufficient time
to soak the rice straw in the acid solution. As shown in Table 1, a
screw speed of 5.1 rpm (AC-SE-2-5.1-13) yielded the lowest mono-
meric xylose concentrations (17.5 ± 1.0 g/L) and the highest xylan
contents in the pretreated residue (12.1 ± 0.4%). With regard to fer-
mentation inhibitors, the concentrations of HMF and furfural have
been found to decrease with increasing screw speed. Increasing the
screw speed decreases the presoaking time accordingly. This im-
plies that the presoaking time is an important factor in xylan
degradation.
The influence of rice straw size, steam explosion pressure and
screw speed on the subsequent enzymatic digestibility of the pre-
treated residues is also shown in Table 1. The enzymatic hydrolysis
conversions were found to slightly decrease from 68.8 ± 1.9% to
64.5 ± 1.5% as the rice straw size reduced, whereas a drastic in-
crease in enzymatic hydrolysis conversion was observed from

Table 1
a
The operational parameters of the acid-catalyzed steam explosion and the concentration of the hydrolysate, composition of the pretreated rice straw and the subsequent
enzymatic digestibility.

Sample Concentration in hydrolysate (g/L) Composition content (%)c EH conversion (%)d

b b
Glucose Xylose HMF Furfural Glucan Xylan ASL AIL
Rice straw size
AC-SE-1-3.4-10e 7.6 ± 0.8 23.4 ± 1.5 1.6 ± 0.4 1.7 ± 0.2 47.7 ± 1.0 5.1 ± 0.6 5.0 ± 0.4 20.1 ± 3.5 68.8 ± 1.9
AC-SE-2-3.4-10 8.5 ± 0.3 22.1 ± 1.0 0.4 ± 0.1 1.8 ± 0.2 48.8 ± 0.8 5.3 ± 0.2 4.6 ± 0.2 20.4 ± 1.1 64.5 ± 1.5
Explosion pressure
AC-SE-2-1.7-10 7.6 ± 0.1 20.3 ± 1.1 1.7 ± 1.1 2.1 ± 0.1 45.1 ± 0.3 4.4 ± 2.1 4.7 ± 0.2 23.4 ± 1.3 66.4 ± 2.3
AC-SE-2-1.7-13 9.0 ± 0.5 21.9 ± 0.6 1.5 ± 0.3 2.4 ± 0.4 48.7 ± 0.5 5.6 ± 0.9 5.0 ± 0.3 19.8 ± 1.6 73.9 ± 1.2
AC-SE-2-1.7-16 9.5 ± 0.6 23.4 ± 1.2 2.6 ± 0.2 2.5 ± 0.1 48.0 ± 0.6 3.6 ± 0.1 4.8 ± 0.1 21.1 ± 1.3 76.0 ± 1.4
Presoaking time
AC-SE-2-1.7-13 9.0 ± 0.5 21.9 ± 0.6 1.5 ± 0.3 2.4 ± 0.4 48.7 ± 0.5 5.6 ± 0.9 5.0 ± 0.3 19.8 ± 1.6 73.9 ± 1.2
AC-SE-2-2.5-13 9.4 ± 0.6 20.6 ± 1.1 1.6 ± 0.5 2.2 ± 0.3 49.3 ± 0.8 6.0 ± 1.1 4.9 ± 0.4 19.5 ± 0.8 70.3 ± 1.5
AC-SE-2-3.4-13 8.6 ± 0.5 21.9 ± 0.6 0.7 ± 0.1 1.4 ± 0.1 45.6 ± 0.7 5.3 ± 0.3 5.4 ± 0.1 17.5 ± 1.0 68.7 ± 1.3
AC-SE-2-5.1-13 7.0 ± 0.2 17.5 ± 1.0 0.5 ± 0.1 0.8 ± 0.1 48.5 ± 1.2 12.1 ± 0.4 5.2 ± 0.1 18.3 ± 0.7 64.6 ± 1.2
a
The reaction using a mixed with a 1.3 wt.% acid solution and preheated at 100–110 °C before an acid-catalyzed steam explosion process at 185 °C for 2 min.
b
The concentration represents monomeric glucose and xylose.
c
Each value is the average of three replicates and calculated on the basis of dry weight.
d
Cellulose to glucose conversion after 24 h of enzymatic hydrolysis. Each value is the average of three replicates.
e
AC-SE-1-3.4-10; 1 represents rice straw size (cm); 3.4 represents screw speed of the horizontal reactor (rpm); 10 represents steam explosion pressure (kg/cm2).
302 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304

66.4 ± 2.3% to 76.0 ± 1.4% as the steam explosion pressure in-
creased from 10 kg/cm2 to 16 kg/cm2. Notably, the mechanical ef-
fects are caused because the pressure is suddenly reduced and
fibers are separated owing to the explosive decompression during
the steam explosion process. Therefore, the higher pressure differ-
ence resulted in higher glucan conversion in terms of enzymatic
hydrolysis. With respect to the effect of presoaking time, the glu-
can conversion decreased from 73.9 ± 1.2% to 64.6 ± 1.2% when
the screw speed increased from 1.7 rpm to 5.1 rpm.
The experimental results suggest that the rice straw size and
steam explosion pressure do not significantly affect either the su-
gar concentrations of the hydrolysate or the glucan, xylan and lig-
nin compositions in the pretreated rice straw. Notably, small rice
straw, high explosion pressure and slow screw speed (long presoa-
king time) can increase the subsequent enzymatic digestibility of
the pretreated residues.
3.3. Operational stability of the continuous steam explosion system
A total of 10 replicate pretreatment runs were completed using
the pilot-scale continuous steam explosion system operating at a
feed rate of approximately 100 kg/h. The total feed amount was
969–1080 kg for each run. Normally, the average productivity
was 2432 ± 238 kg of hydrolysate and 718 ± 20 kg of pretreated
dry solid per dry ton of rice straw. Table 2 lists the following
information: the average concentration of the hydrolysate, the
composition of the pretreated rice straw, the enzymatic digestibil-
ity after using a mixed 1.3 wt.% acid solution preheated to
100–110 °C at a screw speed of 1.7 rpm and the subsequent acid-
catalyzed steam explosion process at 185 °C for 2 min. The
acid-catalyzed steam explosion process reduced the xylan content
from 18.7 ± 0.1% to 4.3 ± 0.4% as the result of degradation and
solubilization, which increased the monomeric/oligomeric xylose
concentration (37.8 ± 2.6 g/L). Consequently, the glucan content
of the pretreated residues was increased from 32.9 ± 0.1% to
46.2 ± 1.3%. This helps the enzymes to contact the glucan and xylan
in the pretreated residues and convert the fermented glucose and
xylose during the enzymatic hydrolysis process. The subsequent
enzymatic hydrolysis of the pretreated residues after the acid-
catalyzed steam explosion pretreatment yielded glucose levels of
70.4–76.0% for 24 h at 15 FPU/g of cellulose.
The average mass balance of xylose, glucose and lignin of 10
replicate pretreatment runs is presented in Table 3. The recovery
of xylose in the hydrolysate was 43.4 ± 6.1%, while 14.7 ± 1.7% of
xylan retained in the pretreated residues. The xylose mass closure
value for the acid-catalyzed steam was 58.1 ± 7.0%. Thomsen et al.
(2006) and Fang et al. (2011b) also reported the similar xylose
recovery of wheat straw used a continuous dilute acid pretreat-
ment at 50 kg/h and 60 kg/h scale, respectively. The loss of 41.9%
xylan may be due to high temperature degradation of xylose into

Table 2
The average values a for concentration of the hydrolysate, composition of the pretreated rice straw and the subsequent enzymatic digestibility after using a mixed acid solution
(1.3 wt.%) and preheating at the screw speed of 1.7 rpm to a temperature of 100–110 °C. This process was followed by an acid-catalyzed steam explosion process at 185 °C for
2 min.

Monomeric Oligomeric Fermentation inhibitor

Glucose Xylose Glucose Xylose Acetic acid HMF Furfural
Concentration (g/L) in hydrolysate in the xylose-rich hydrolysate
6.8 ± 1.1 18.1 ± 2.4 10.0 ± 1.1 19.7 ± 2.7 6.0 ± 0.7 1.5 ± 0.4 2.2 ± 0.3
Solid recoveryb(%) Composition in residuesc (%) Enzymatic Hydrolysisd (%)
Glucan Xylan ASL AIL Glucose Xylose
The composition and enzymatic digestibility of the cellulose-rich residues
71.8 ± 2.0 46.1 ± 1.3 3.8 ± 0.4 4.7 ± 0.3 20.3 ± 2.7 73.2 ± 2.8 88.5 ± 5.1
a
Each value is the average of ten replicates pretreatment runs.
b
Solid recovery represents the percentage of obtained residues after solid–liquid separation from the total raw rice straw and calculated on the basis of dry weight.
c
Compositions are calculated on the basis of dry weight.
d
Cellulose to glucose and xylan to xylose conversion after 24 h of enzymatic hydrolysis.

Table 3
a
The xylose, glucose and lignin mass balance for the acid-catalyzed steam explosion process.

Composition Feedstock Pretreatment products Recoveryc

Rice straw (kg) Xylose-rich hydrolysate Cellulose-rich residues Liquid (%) Solid (%)
Liquid (kg) Solid (kg) Total (kg) Dry weight (kg)
Xylose
Monomeric – 25.4 ± 6.5 13.2 ± 1.6 38.6 ± 7.8 – 21.0 ± 4.3 –
Oligomeric – 27.1 ± 4.7b 14.2 ± 2.3 41.3 ± 5.9 – 22.4 ± 3.2 –
Total 184.1 ± 5.8b 52.5 ± 9.9 27.4 ± 2.7 79.9 ± 10.9 26.9 ± 2.5b 43.4 ± 6.1 14.7 ± 1.7
Glucose
Monomeric – 9.5 ± 2.8 4.8 ± 0.9 14.3 ± 3.5 – 4.4 ± 1.1 –
Oligomeric – 13.7 ± 2.2b 7.2 ± 1.0 20.9 ± 2.8 – 6.4 ± 0.9 –
Total 327.0 ± 10.4b 23.2 ± 4.4 12.0 ± 1.0 35.2 ± 5.1 318.8 ± 15.3b 10.8 ± 1.6 97.5 ± 4.1
Lignin
ASL 35.8 ± 1.3 – – – 29.1 ± 2.1 – 17.5 ± 1.4
AIL 131.0 ± 4.4 – – – 126.3 ± 16.3 – 75.9 ± 10.9
Total 166.8 ± 5.7 – – – 155.4 ± 18.0 – 93.4 ± 12.2
a
Each value is the average of ten replicates pretreatment runs.
b
Reported in monometric equivalents.
c
Recovery represents the percentage of obtatined xylose or glucose or lignin (ASL/AIL) in hydrolysate or residues divided by total xylose or glucose or lignin (ASL + AIL) in
the raw rice straw.
 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304 303

(a)
25 100
Glucose
Xylose
HMF
Furfural
20 80

Glucan conversion (%)

(enzymatic hydrolysis)
Concentration (g/L) 15 60

10 40

5 20

0 0
3 4 5 6 7 8 9
Operational Time (h)

(b)
25 100
Glucose
Xylose
HMF
Furfural
20 80

Glucan conversion (%)

(enzymatic hydrolysis)
Concentration (g/L)

15 60

10 40

5 20

0 0
1 2 3 4
Operational Day (d)

Fig. 5. The variations in the concentration of the obtained hydrolysate and enzymatic hydrolysis efficiency during the pretreatment process for (a) one batch and (b)
continuous operation for 4 days.

other compounds and mass loss during steam explosion process.
As shown in Table 3, the amount of glucose released into hydroly-
zate from pretreatment was less than 10.8 ± 1.6%. Glucose mass
closure for the acid-catalyzed was 108.3 ± 5.5%, which was laid in
the range of 91–108% for different pretreatment process in the lit-
erature. (Garlock et al., 2011) With regard to lignin, 17.5 ± 1.4% of
ASL (acid-soluble lignin) and 75.9 ± 10.9% of AIL (acid-insoluble lig-
nin) were retained in the pretreated residues. This indicated that
the removal of lignin was insignificant and most was retained in
the pretreated residues during the acid-catalyzed steam explosion.
The steam consumption for the acid-catalyzed steam explosion
was also measured by the saturated steam flow meter. The steam
requirement for pretreatment of rice straw was ca. 2.3 kg steam/kg
dry rice straw. The experimental value is considerably higher than
theoretical value (1.2 kg steam/kg dry rice straw). Kling et al.
(1987) reported that the experimental steam requirement for
steam explosion pretreatment of bagasse range from 0.91 to
1.9 kg steam/kg dry bagasse. They also found that theoretical
values (0.49–1.06 kg steam/kg dry bagasse) are lower than the
experiment results. Further reduction of heat loss and improve-
ment in the energy efficiency for the acid-catalyzed steam explo-
sion is necessary.
Long-term operational stability under continuous operation is a
key issue for commercialization that usually goes neglected. Con-
tinuous pretreatment is more challenging than batch pretreatment
because process conditions such as heat and mass transfer and res-
idence time cannot be as tightly controlled. To respond to this chal-
lenge, the long-term stability under continuous operation was
studied over 4 days. The feed rate was approximately 100 kg/h,
the pretreatment process was operated for 10 h each batch. Fig. 5
shows variations in the concentration of the obtained hydrolysate
and the enzymatic hydrolysis efficiency during the pretreatment
process for one batch (Fig. 5(a)) and the continuous operation for
4 days (Fig. 5(b)). As shown in Fig. 5, the hydrolysate concentra-
tions of monomeric xylose, glucose, HMF, furfural, and the glucan
conversion in the enzymatic hydrolysis were 19.1 ± 0.5 g/L,
304 W.-H. Chen et al. / Bioresource Technology 128 (2013) 297–304
8.9 ± 0.4 g/L, 2.4 ± 0.1 g/L, 2.5 ± 0.3 g/L, and 72.8 ± 1.7%, respec-
tively. In the case of the continuous operation for 4 days, the con-
centrations were 18.5 ± 0.4 g/L for xylose, 9.0 ± 0.2 g/L for glucose,
2.3 ± 0.1 g/L for HMF and 2.6 ± 0.2 g/L for furfural. The glucan con-
version in terms of the enzymatic hydrolysis was 73.0 ± 2.1%. The
small standard deviations clearly demonstrate that the pilot-scale
continuous pretreatment system possesses good operational sta-
bility and durability.
4. Conclusions
A pilot-scale continuous pretreatment system on the basis of an
acid-catalyzed catalyzed steam explosion pretreatment process
was successfully developed. It shows good operational stability
and durability. A total saccharification yield of 73% can be achieved
under optimal conditions. The behavior of rice straw under dilute-
acid hydrolysis is different due to its capacity, the presoaking pro-
cess, the reactor design and the operational module. Future work
will examine the performance of feedstock diversity, such as ba-
gasse, bamboo and sorghum. Investigations are continuing to esti-
mate energy consumption and economic analysis of the cellulosic
ethanol production process.
Acknowledgements
The authors thank Mr. Ching-Jui Lo, Chien-Hsien Huang, Sheng-
Shieh Chen and Chung-Chih Hsu for assistance with the pretreat-
ment test. The authors also thank Mr. Yi-Feng Jiang for providing
technical assistance. This study was supported by the National Sci-
ence and Technology Program (Energy) from the government of
Taiwan.
References
Alvira, P., Tomás-Pejó, E., Ballesteros, M., Negro, M.J., 2010. Pretreatment
technologies for an efficient bioethanol production process based on
enzymatic hydrolysis: a review. Bioresour. Technol. 101, 4851–4861.
Chandel, A.K., Chandrasekhar, G., Radhika, K., Ravinder, R., Ravindra, P., 2011.
Bioconversion of pentose sugars into ethanol: a review and future directions.
Biotechnol. Mol. Biol. Rev. 6 (1), 8–20.
Chen, W.H., Pen, B.L., Yu, C.T., Hwang, W.S., 2011a. Pretreatment efficiency and
structural characterization of rice straw by an integrated process of dilute-acid
and steam explosion for bioethanol production. Bioresour. Technol. 102, 2916–
2924.
Chen, W.H., Xu, Y.Y., Hwang, W.S., Wang, J.B., 2011b. Pretreatment of rice straw
using an extrusion/extraction process at bench-scale for producing cellulosic
ethanol. Bioresour. Technol. 102, 10451–10458.
Fang, H., Deng, J., Zhang, T., 2011a. Dilute acid pretreatment of black spruce using
continuous steam explosion system. Appl. Biochem. Biotech. 163, 447–547.
Fang, H., Deng, J., Zhang, X., 2011b. Continuous steam explosion of wheat straw by
high pressure mechanical refining system to produce sugars for bioconversion.
Bioresource 6 (4), 4468–4480.
Galbe, M., Zacchi, G., 2012. Pretretment: the key to efficient utilization of
lignicellulosic materials. Biomass Bioenergy. http://dx.doi.org/10:1016/
j.biombioe.2012.03.026.
Garlock, R.J., Balan, V., Dale, B.E., Pallapolu, V.R., Lee, Y.Y., Kim, Y., Mosier, N.S.,
Ladisch, M.R., Holtzapple, M.T., Falls, M., Sierra-Ramirez, R., Shi, J., Ebrik, M.A.,
Redmond, T., Yang, B., Wyman, C.E., Donohoe, B.S., Vinzant, T.B., Elander, R.T.,
Hames, B., Thomas, S., Warner, R.E., 2011. Comparative material balance around
pretreatment technologies for conversion of switchgrass to soluble sugars.
Bioresour. Technol. 102, 11063–11071.
Kling, S.H., Neto, C.C., Ferrara, M.A., Torres, J.C.R., Magalhaes, D.B., Ryu, D.D.Y., 1987.
Enhancement of enzymatic hydrolysis of sugar cane bagasse by steam
explosion pretreatment. Biotechnol. Bioeng. 29, 1035–1039.
Kumar, R., Wyman, C.E., 2008. The impact of dilute sulfuric acid on the selectivity of
xylooligomer depolymerization to monomers. Carbohydr. Res. 343, 290–300.
Larsen, J., Haven, M.Ø., Thitup, L., 2012. Inbicon makes lignocellulosic ethanol a
commercial reality. Biomass Bioenergy. http://dx.doi.org/10:1016/
j.biombioe.2012.03.033.
Modenbach, A.A., Nokes, S.E., 2012. The use of high-solids loading in biomass
pretreatment-a review. Biotechnol. Bioeng. 109, 1430–1442.
Palmqvist, E., Hahn-Hägerdal, B., 2000. Fermentation of lignocellulosic
hydrolysates. II: inhibitors mechanisms of inhibition. Bioresour. Technol. 74,
25–33.
Rocha, G.J.M., Martin, C., Siliva, V.F.N., Gómez, E.O., Goncalves, A.R., 2012a. Mass
balance of pilot-scale pretreatment of sugarcane bagasse by steam explosion
followed by alkaline delignification. Bioresour. Technol. 111, 447–452.
Rocha, G.J.M., Goncalves, A.R., Oliveira, B.R., Olivares, E.G., Rossell, C.E.V., 2012b.
Steam explosion pretreatment reproduction and alkaline delignification
reactions performed on a pilot scale with sugarcane bagasse for bioethanol
production. Ind. Crop. Prod. 35, 274–279.
Sarkar, N., Ghosh, S.K., Bannerjee, S., Aikat, K., 2012. Bioethanol production from
agricultural wastes: an overview. Renew. Energy 37, 19–27.
Schell, D.J., Farmer, J., Newman, M., McMillan, J.D., 2003. Dilute-sulphuric acid
pretreatment of corn stover in pilot-scale reactor. Appl. Biochem. Biotechnol.
105–108, 69–85.
Shekiro, J., Kuhn, E.M., Selig, M.J., Nagle, N.J., Decker, S.R., Elander, R.T., 2012.
Enzymatic conversion of xylan residues from dilute acid pretreated corn stover.
Appl. Biochem. Biotechnol.. http://dx.doi.org/10.1007/s12010-012-9786-5.
Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., 2005.
Determination of structural carbohydrates and lignin in biomass. Laboratory
Analytical Procedure from the National Renewable Energy Laboratory Biomass
Analysis Technology (NREL BAT) Team.
Thomsen, M.H., Thygesen, A., Christensen, B.H., Larsen, J., Jørgensen, H., Thomsen,
A.B., 2006. Preliminary results on optimising hydrothermal pre-treatment used
in co-production of biofuels. Appl. Biochem. Biotechnol. 129–132, 448–460.
Weiss, N.D., Farmer, J.D., Schell, D.J., 2010. Impact of corn stover composition on
hemicellulose conversion during dilute acid pretreatment and enzymatic
cellulose digestibility of the pretreated solids. Bioresour. Technol. 101, 647–
678.
Weng, Y.H., Huang, H.C., Guo, G.L., Hwang, W.S., 2011. Technology and production
of cellulosic ethanol in a pilot plant in Taiwan. J. Energy Power Eng. 5, 928–933.
Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005.
Coordinated development of leading biomass pretreatment technologies.
Bioresour. Technol. 96, 1959–1966.