Communications t o the Editor

Enhancement of Enzymatic Hydrolysis of

Sugar Cane Bagasse by Steam
Explosion Pretreatment
S. H. Kling, C. Carvalho Neto, M. A. Ferrara, J. C. R. Torres,
D. B. Magalhaes. and Dewey D. Y. Ryu*
Foundation for Industrial Technology, Lorena, Sao Paulo, Brazil
Accepted for publication May 14, 1986
In the Brazilian economic context, effective utili-
zation of sugarcane bagasse for ethanol production is
very important. A recent survey carried out by the
Foundation for Industrial Technology in Brazil esti-
mated that 6,750,000 tons of bagasse containing 50%
moisture could be harvested during the 1985-1986 sea-
son. Natural lignocellulosics are not so susceptible to
enzymatic hydrolysis due to their physical and chem-
ical proper tie^,'-^and a good pretreatment is required
if a practical enzymatic saccharification process were
to be developed.
Recently, a steam explosion technique has been in-
vestigated as a possible pretreatment process for such
lignocellulosics as wood chips and other bi oma~ses."~
Selected examples of steam explosion processes and
their process conditions are shown in Table I. In all
pretreatment processes, treatment times and temper-
atures ranged 0.3-20 min and 195-240C, respectively.
In this study, the possibility of applying a steam
explosion pretreatment process to sugarcane bagasse
was investigated, and the effectiveness of the pretreat-
ment in terms of hemicellulose solubilization and en-
hancement of enzymatic hydrolysis was determined.
The steam requirement for the pretreatment was also
investigated at the pilot-plant scale, and these results
are presented.
MATERIALS AND METHODS
Materials
The lignocellulosic samples used in this study were
sugarcane bagasse supplied by Sao Martinho sugar
and ethanol plant in Pradopolis, Sao Paulo, Brazil. The
sugarcane bagasse used in this study contained, on a
dry weight basis, 47% potential fermentable reducing
* To whom all correspondence should be addressed, at the Uni-
versity of California, Davis, CA 95616.
sugars, 26% nonfermentable reducing sugars, 23% lig-
nin, and ca. 4% ashes and other residues. The sugar-
cane bagasse used consists of shredded cane stalk with
sliver lengths ranging from 2.5 to 10 cm. The diameter
of the slivers was ca. 0.1-1 cm. The cellulase complex
used for saccharification tests was supplied by Bio-
ferm, S. A. , Brazil. It had 0.1 filter paper unit (FPU)
and 0.02 units cellobiase (P-g1ucosidase)lmg crude en-
zyme powder.
Pretreatment
Pretreatment by steam explosion was carried out by
heating the sugarcane bagasse with superheated steam
and exploding the heated bagasse by instantaneous de-
compression. The schematic design of the reactor sys-
tems used for steam explosion is shown in Figure 1 .
The reactor system consists of the reactor used for
heating and compression, the cyclone which collects
the exploded material after decompression, and other
peripherals and controls. In these studies two reactors
were used. The first one with nominal capacity of I .5
L was used in the bench scale experiments for hemi-
cellulose solubilization and enhancement of enzymatic
hydrolysis. The other with nominal capacity of 200 L
was used for the steam requirement tests.
The reactor system was first heated (in the range
190-220C), the lignocellulosics were charged into the
reactor and heated by superheated steam for a pre-
determined period, after which the treated lignocel-
lulosic material was exploded and released into the
cyclone by instantaneous decompression of the reac-
tor. The pretreatment time, which corresponds to the
mean residence time of sugarcane bagasse in the re-
actor system, was varied from 0 to 25 min while main-
taining the preset pretreatment temperature. It has been
noted that the time required to heat the bagasse in the
bench-scale reactor from the room temperature is usu-
ally ca. 10 s . The heating time for the bagasse in the
Biotechnology and Bioengineering, Vol. XXIX, Pp. 1035-1039 (1987)
0 1987 J ohn Wiley & Sons, Inc.
CCC 0006-3592/87/081035-05$04.00
Table I. Selected steam explosion processes and process conditions.
Mean Treatment
Process residence time temperature Steam pressure Mode of
(and reference) (min) ("C) (kdcm) operation
Iotech (4,s) 0.3-1.3 2 10-240 36.8-50.4 batch
Stake (6) 5-20 200-2 10 14.3-17.0 continuous
General Electric (7) 5-20 195 13.6 batch
pilot-scale unit was ca. 30 s. After the pretreatment,
the material was washed, and the effects of the pre-
treatment on hemicellulose solubilization and enhance-
ment of enzymatic hydrolysis were evaluated.
Analysis
Immediately after the pretreatment of sugarcane ba-
gasse by steam explosion, the samples were washed
and subjected to standard enzymatic saccharification
tests. The conditions used in this test were: tempera-
ture of 50C, pH 4.8 (with sodium citrate buffer of
0.05M), agitation of 120 rpm, cellulose concentration
of 35 mg/mL, ratio of cellulase to substrate of 14 FPU/g
cellulose, and reaction time of 24 h. The total reducing
sugar obtained after enzymatic hydrolysis were deter-
mined by the method of Nelson8 using glucose as stan-
dard. The saccharification yield evaluated was based
on the cellulose content of pretreated bagasse. The
fermentable reducing sugars and the nonfermentable
reducing sugars were determined by the method of
Saeman and co-w~rkers.~ Analyses of total carbohy-
drate content in the pretreated and untreated bagasse
were evaluated after hydrolyzing them with 72% sul-
furic acid by using Dunning's method.'O In this work,
the potential fermentable reducing sugar and nonfer-
mentable reducing sugar are closely related to the con-
tents of cellulose and hemicellulose, respectively.
The lignin content was gravimetrically estimated from
Figure 1. Schematic design of steam explosion reactor system: (1)
reactor, (2) cyclone, (3) pneumatic valve, (4) thermocouple, (5) bi-
metal thermometer, (6) manometer, (7) steam distribution ring, (8)
vent.
the insoluble residue by the method of Moore after the
acid hydrolysis. Hemicellulose recovery in water ex-
tract was measured by determining the content of non-
fermentable reducing sugar before and after hydrolysis
with 4% H2S04 (1 atm steam pressure, 121"C, and 15
min). The objective of this acid hydrolysis was to con-
vert the oligosaccharides present in the water extract
to monosaccharides.
Steam Requirements
The weighed sugarcane bagasse samples were charged
into the reactor and heated from 25 to 190"C, and then
maintained at this temperature for 15 min by injecting
saturated steam (15 kg/cm2) directly into the reactor.
After the heating, the steam valve was closed and the
reactor was kept completely closed until it was cooled
to environmental temperature. After cooling the re-
actor for ca. 48 h, the reactor was opened and all ma-
terial inside was weighed. The difference between final
and initial weights was determined as the steam con-
densed during thermal treatment of sugarcane bagasse
samples. From this the steam requirement was esti-
mated, although the operational efficiency could be
improved considerably.
RESULTS AND DISCUSSION
Hemicellulose Solubilization
The hydrolysis of the hemicellulose fraction during
steam pretreatment involves solubilization of hemi-
cellulose and partial destruction of the reducing sugar
produced. As a consequence, the amount of reducing
sugar recovered from the pretreated bagasse is a func-
tion of the pretreatment time.
Figure 2 shows the reducing sugar recovery from
pretreated hemicellulose as a function of pretreatment
time and temperature. The differences between the
nonfermentable reducing sugar recovered from the
hemicellulose before and after the acid hydrolysis with
4% H2S04 represent an incomplete hydrolysis of the
hemicellulose fraction to oligomers. The oligomers
produced by the steam pretreatment can be converted
mainly to xylose in the acid hydrolysis step.
Using the best conditions of time and temperature,
60-65% hemicellulose recovery would be possible us-
1036 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 29, JUNE 1987
g80-
9
3
w
J
260 E
I
e 5 40..
>a
w
>0
:: 20..
a d a
a
W
3 In
Figure 2. Recovery of nonfermentable sugars in water extracts
(-) before and (----) after hydrolysis with 4% H2S04 (1.05
atm/l5 min) following the pretreatment of bagasse by steam explo-
sion at varying temperatures: (0) 190C (D) 200C (A) 210"C, and
(V) 220C.
els of 190 and 210C (up to 10 min at 190C, 6 rnin at
200"C, and 8 min at 210C), and it decreased with in-
creasing pretreatment time and temperature to ca. 87%. ,+.\,, ~ ,- , ~-
.... - Q* >
'. 4 P p. ,_, ,", a,\
d'\ b,,,,
Enzymatic Hydrolysis
The results of enzymatic saccharification of bagasse
pretreated by steam explosion at four different tem-
perature levels and varying times are shown in Figure
4. The sugar yields showed a maximum at 1-min treat-
h .
/f&-.- dAb --:<~ _ _ _
D
'h
4 n
ing the acid hydrolysis. The conditions at which the
maximum yield of hemicellulose hydrolysis was achieved
were 220C and 2 min; 210C and 3 min; 200C and 6
min; and 190C and 10-15 min. Although the prolonged
reaction time improved the hydrolysis of hemicellu-
lose, a higher degree of degradation of sugar with an
increased reaction time was observed.
Composition of Pretreated Sugarcane Bagasse
Figure 3 summarizes the data related to the com-
position of steam pretreated bagasse as a function of
time and temperature of pretreatment. It shows a slight
increase in the relative amount of cellulose and lignin
content due to the solubilization of hemicellulose. The
final cellulose content reached ca. 65% for all samples
pretreated at varying temperature levels tested. The
2c
I
5 10 15 2 0 25
ment temperature levels 200 and 210"C, the yield pro-
files go through a maximum at ca. 5-7 and 3-4 rnin
treatment times, respectively. When bagasse was pre-
treated at 190"C, the sugar yields seem to stay constant
at or near its maximum level for a wide range or pre-
treatment time (10-20 min).
The sugar yield of bagasse pretreated at 190C did
not decrease as much as those pretreated at higher
temperature levels with the increasing treatment time.
This observation was considered very important, es-
pecially when the steam explosion process is to be
scaled up to a larger reactor where a temperature gra-
dient within the reactor exists and treatment time has
to be prolonged somewhat. Under the saccharification
test conditions used in this study, the maximum sugar
yield obtained was 36% at 190, 200, and 210"C, and
32% at 220C. When untreated bagasse was hydrolyzed
enzymatically, only 10% sugar yield was achieved. The
pretreatment by steam explosion improved the enzy-
matic hydrolysis and yielded an additional 0.13 kg fer-
mentable sugar based on 24-h hydrolysis assay and 1
kg steam consumed for the pretreatment. When the
pretreated bagasse was hydrolyzed for a prolonged pe-
riod up to 100 h, the additional sugar yielded was as
L L I
I
V c,
fn
a 5 10 I5 20 2 5
PRE-TREATMENT TIME (M!N)
PRE - TREATMENT TIME (MINI
Figure 4. Saccharification yield of bagasse pretreated by steam
explosion based on the cellulose content of pretreated bagasse. Pre-
treatment temperature levels were: (0) l WC, (0) 2WC, (A) 210"C,
and (V) 220C.
Figure 3. Relative content of (-) cellulose, (----) hemicel-
lulose, and (---) lignin of sugarcane bagasse after pretreatment
by steam explosion: (0) l WC, (0) 20O0C, (A) 210"C, and (V) 220C.
COMMUNICATIONS TO THE EDITOR 1037
Table 11. Steam requirements for steam explosion of various lignocellulosic materials.
Experimental Theoretical
steam consumption Steam Consumption
Weight Steam
Initial of moist explosion Moist raw Dried raw Moist raw Dried
moisture bagasse pretreatment material material m a t e r i a I material
Raw content sample temperature (kg steadkg (kg steadkg (kg steadkg (kg steadkg
material (W (kg) (0 raw material) raw material) raw material) raw material) Reference
~~ ~
Sugarcane 38.6 50.4 190 0.56 0.91 0.30 0.49
65.5 43.9 190 0.65 1.90 0.37 1.06
bagasse 57.8 58.2 190 0.55 1.30 0.33 0.78 this work
Biomass 50 - 150-200 - - - 0.45-1.30* 13
- - 1.28 12
Biomass - - - -
20 - 200 - -
33 - 200
43 - 200
50 - 200
- 0.21
- 0.38
- 0.49
- 0.58
- -
Wood 240 - 0.88-1.12 - 0.64 14
- -
- -
* This assumed 50% of heat recovery.
high as 0.37 kg/kg steam consumed for pretreatment.
Investigations are continuing to improve the total sugar
yield by using cellulase with optimal ratios of cellulase
components and prolonged saccharification time.
Based on the considerations described above and of
the lower temperature required for the energy econ-
omy, we have selected 190C and 10 min as the optimal
condition for pretreatment of bagasse. These condi-
tions also improve the hemicellulose recovery, as pre-
viously mentioned.
Steam Requirement
The steam requirement for the pretreatment by steam
explosion was examined in terms of the amount of
steam actually consumed per unit amount of bagasse
during the pretreatment. The effect of initial moisture
content of sugarcane bagasse on the steam requirement
for pretreatment was determined. These results, shown
in Table 11, are average values of triplicate samples in
each case. l2-l4 It shows that the steam requirement for
pretreatment of bagasse ranges from 0.55 to 0.65 kg
steadkg moist bagasse on moist bagasse weight basis
and from 0.91 to 1.90 kg stearnkg on dry bagasse weight
basis for the range of moisture content (38.6-65.5%)
tested.
Table I1 also shows theoretical and experimental val-
ues of steam requirement which we have estimated
and/or measured. To determine its theoretical values,
wehave considered steam requirements for heating of
bagasse samples (cellulosic material plus moisture) and
the reactor, and heat loss to environment. For this
calculation it was assumed that a specific heat for ba-
gasse (dry matter) of 0.42 kcallkg C.
Table I1 reveals also that the steam consumption
data obtained are in good agreement with those re-
ported by Datta13 and Spano et a1.,I2 and our data
appear to be better than those values presented by
Brownell and Sadler.I4 A significant part of this dif-
ference could be due to the heat loss during reactor
heating between the pretreatment cycles, which was
not considered by Brownell and Sadler but was con-
sidered by us. Our calculations showed that reactor
heating consumes ca. 15% of the total heat required to
treat sugarcane bagasse samples with 39% moisture
content.
Theoretical values of the steam requirement that we
calculated are considerably lower than the experimen-
tal values. It is certain that we will be able to reduce
the heat loss significantly and improve the energy ef-
ficiency further, and at the same time more accurate
measurement of the parameters involved in the cal-
culation of the steam consumption could be achieved.
Specific heat of bagasse and better control of the pre-
treatment reactor system during the experiments are
such examples.
CONCLUSIONS
Steam explosion as a pretreatment method for sug-
arcane bagasse appears to have some promise for prac-
tical application. The results obtained from this study
showed that a significant improvement of sugar yield
can be made when the bagasse is pretreated by steam
explosion and the susceptibility of bagasse to the en-
zymatic saccharification is improved. Further im-
provement in the energy efficiency for the steam ex-
plosion pretreatment of bagasse and the effective
recovery and utilization of hemicellulose hydrolysate
as the by-product of pretreatment should enable us to
bring this pretreatment technique to an early realiza-
tion of practical application.
1038 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 29, JUNE 1987
References 2nd Annual Symposium on Fuels from Biomass, Troy, NY,
1.
2.
3.
4.
5.
6.
7.
D. D. Y. Ryu, S. B. Lee, T. Tassinari, and C. Macy, Biotechnol.
Bioeng. , 24, 1047 (1982).
M. A. Millet, A. J . Baker, and L. D. Satter, Biotechnol. Bioeng.
Syrnp., 6, 125 (1976).
L. T. Fan, M. M. Gharpuray, and Y. H. Lee, Biotechnol. Bioeng.
Syrnp., 11, 29 (1981).
L. J urasek, Dev. Znd. Microbiol., 20, 177 (1979).
P. Foody, Optimization of steam explosion pre-treatment,
Final Report, Iotech Corporation, Ottawa, Ontario, Canada,
1980.
J . D. Taylor and W. J . Esdalle, Increased utilization of crop
residues as animal feed through autohydrolysis, Stake Tech-
nology Report. Ottawa, Canada.
R. E. Brooks, W. D. Bellamy, and T. Su, Proceedings of the
1978, pp. 511-531.
8. N. Nelson, J . Bi d . Ckem., 153, 375 (1944).
9. J . F. Saeman, E. E. Hams, and A. A. Kline, Znd. Eng. Chern.,
10. J . W. Dunning and D. E. Dallas, Anal. Chern., 21, 727 (1949).
11. W. E. Moore and D. B. J ohnson, Procedures for Chemical
Analysis of Wood and Wood Products, Forest Product Lab-
oratory, U.S. Department of Agriculture, Madison, WI, 1967.
12. L. Spano, A. Allen, T. Tasinari, M. Mandels, and D. D. Y.
Ryu, Proceedings of 2nd Annual Symposium on Fuels from
Biomass, Troy, NY, 1978. pp. 671-684.
13. R. Datta, Process Biochem. 42, 16 (1981).
14. H. H. Brownell, J . N. Sadler, Biorechnol. Bioeng. Symp., 14,
17, 95 (1945).
55 (1984).
COMMUNI CATI ONS TO THE EDITOR 1039