Extraction of Hemicellulose 253

Applied Biochemistry and Biotechnology Vol. 9193, 2001

Copyright 2001 by Humana Press Inc.
All rights of any nature whatsoever reserved.
0273-2289/01/9193/0253/$13.75
253
*Author to whom all correspondence and reprint requests should be addressed.
Continuous Countercurrent Extraction
of Hemicellulose
from Pretreated Wood Residues
KYOUNG HEON KIM, MELVIN P. TUCKER, FRED A. KELLER,
ANDY ADEN, AND QUANG A. NGUYEN*
National Renewable Energy Laboratory, 1617 Cole Boulevard,
Golden, CO, 80401-3393, E-mail: [email protected]
Abstract
Two-stage dilute acid pretreatment followed by enzymatic cellulose
hydrolysis is an effective method for obtaining high sugar yields from wood
residues such as softwood forest thinnings. In the first-stage hydrolysis step,
most of the hemicellulose is solubilized using relatively mild conditions.
The soluble hemicellulosic sugars are recovered from the hydrolysate slurry
by washing with water. The washed solids are then subjected to more severe
hydrolysis conditions to hydrolyze approx 50% of the cellulose to glucose.
The remaining cellulose can further be hydrolyzed with cellulase enzyme.
Our process simulation indicates that the amount of water used in the hemi-
cellulose recovery step has a significant impact on the cost of ethanol produc-
tion. It is important to keep water usage as low as possible while maintaining
relatively high recovery of soluble sugars. To achieve this objective, a proto-
type pilot-scale continuous countercurrent screw extractor was evaluated
for the recovery of hemicellulose from pretreated forest thinnings. Using the
274-cm (9-ft) long extractor, solubles recoveries of 98, 91, and 77% were
obtained with liquid-to-insoluble solids (L/IS) ratios of 5.6, 3.4, and 2.1,
respectively. An empirical equation was developed to predict the perfor-
mance of the screw extractor. This equation predicts that soluble sugar recov-
ery above 95% can be obtained with an L/IS ratio as low as 3.0.
Index Entries: Extraction; hemicellulose; softwood; pretreatment; acid
hydrolysis.
Introduction
In a previous study (1), we concluded that two-stage dilute sulfuric
acid pretreatment of softwood forest thinnings gave higher sugar yields
than single-stage pretreatment. Figure 1 shows a simplified block-flow
254 Kim et al.
Applied Biochemistry and Biotechnology Vol. 9193, 2001
diagram of a two-stage pretreatment process. In this process, high recovery
of soluble sugars in the first-stage pretreated material (or hydrolysate) is
essential because sugars remaining in the first-stage hydrolysate would be
destroyed in the second-stage hydrolysis. Furthermore, the amount of wash
water required to achieve high sugar recovery has a significant impact on
the process economics. Too much wash water would dilute the sugar stream
and increase the cost of fermentation and ethanol distillation. In general,
countercurrent washing of the pretreated biomass is required to achieve
adequate sugar recovery and high sugar concentration in the extract while
maintaining a reasonable water usage requirement. One common term
used to describe the amount of water used for extracting solubles from
material is the liquid-to-solids (L/S) ratio, which is the ratio of water used
in the extraction process over the dry wt of total solids (insoluble and
soluble) in the feed. Because the content of soluble solids varies with pre-
treated materials and diminishes as the solids are being washed, the liquid-
to-insoluble solids (L/IS) ratio is also used. The L/IS ratio is essentially
constant throughout a countercurrent extractor at steady state, whereas the
L/S ratio increases as the soluble solids are removed. We used the L/IS
ratio to ensure consistent comparison of extraction characteristics for dif-
ferent pretreated biomass materials.
Figure 2 shows the effect of the L/IS ratio on soluble sugar recovery
from first-stage hydrolysate and cost of ethanol production for a 2000 dry
t/d softwood-to-ethanol plant using two-stage dilute acid hydrolysis (2).
The sugar recovery values were based on the results of three-stage,
stagewise, countercurrent extraction experiments (3). The process simula-
tion in Fig. 2 implies that reducing the amount of water used in extracting
sugars from first-stage hydrolysate from an L/IS ratio of 5.0 to 3.0 would
lower the cost of ethanol production from $0.34/L ($1.29/gal) to $0.31/L
Fig. 1. Two-stage dilute sulfuric acid pretreatment of softwood forest thinnings.
Extraction of Hemicellulose 255
Applied Biochemistry and Biotechnology Vol. 9193, 2001
($1.16/gal), even though the sugar recovery would be lowered from 97.9 to
93.8%. This reduction in production cost is owing mainly to the lower
energy requirements to recover ethanol resulting from fermentation of a
more concentrated sugar stream. Further reduction of L/IS to 2.8 raises the
cost of ethanol production because the negative impact of sugar loss is
greater than the savings from lower water usage. If a greater number of
extraction stages or a continuous countercurrent extraction device is used,
the L/IS ratio can probably be reduced to <3.0 while achieving high sugar
recovery and further reducing the cost of ethanol production.
Countercurrent extraction of solubles from biomass materials (such as
pulp, sugarcane, fruits, seeds, and pretreated lignocellulose) can be accom-
plished in a variety of commercial equipment (4,5). The main criteria for
selecting countercurrent washing of pretreated biomass to recover soluble
sugars include high sugar recovery, high sugar concentration in the extract
(i.e., low L/IS ratio), and low capital and operating costs. These criteria are
generally the same as those used in the food-processing industry for extrac-
tion of sugars and other soluble solids from a variety of feedstock. By com-
parison, most stagewise washers used in the pulp industry are designed
primarily for thorough washing of fibers and not necessarily for obtaining
a high concentration of solutes in the wash water. Therefore, our focus is on
continuous countercurrent extraction equipment used in the food-process-
ing industry because these systems are most effective in reducing water
requirements. Screw conveyors (single or twin) and screw towers are com-
monly used for extraction of sugar from sugarcane, sugar beet, and fruits
(4,6). Pilot-scale screw extractors were used to extract soluble components
from sweet sorghum silage (7) and steam-pretreated lignocellulosic ma-
Fig. 2. Effect of L/IS ratio on soluble sugar recovery and production cost of ethanol
(2000 dry t/d, $25/dry t softwood residues, two-stage dilute acid hydrolysis process,
without enzyme addition).
256 Kim et al.
Applied Biochemistry and Biotechnology Vol. 9193, 2001
terials (8). Twin-screw extractors provide better liquid/solid contact and,
thus, are generally more efficient than single-screw extractors. However,
twin-screw extractors are generally more expensive. The design of inter-
mittent-reversing of screw rotation direction was reported to overcome the
compaction problem and inefficient liquid/solid contact in single-screw
extractors (9,10). We installed mixing paddles on the auger of our single-
screw extractor to improve liquid/solid contact.
Good liquid/solid contact in screw extractors also depends on the
drainage characteristics of the pretreated biomass. The particle size of bio-
mass may be important in continuous countercurrent extraction because
very fine particles tend to compact and cause liquid to channel or block
liquid flow completely. Water temperature may also have an effect on the
extraction of solubles from the pretreated biomass. The impact of these
parameters and the L/IS ratio on soluble sugar recovery from pretreated
softwood forest thinnings was explored in a series of experiments using
small column percolators and a pilot-scale single-screw extractor.
Materials and Methods
Pretreated Biomass
Whole-tree chips (passing through a 0.5-in. screen) from California
softwood forest thinnings were soaked in 0.66% (w/w) sulfuric acid
solution. The acid-impregnated chips were air-dried to 43% (w/w) solids
(the acid concentration of liquid in air-dried chips was 1.08% w/w), then
pretreated at 185C for 4 min using a 4-L steam explosion reactor described
previously (1). At these conditions, approx 85% of the hemicellulose was
solubilized. The water-insoluble fraction of the pretreated material was
72.9% on a dry wt basis. Table 1 gives the feedstock composition and the
theoretical component yields after pretreatment. As seen in Table 1, the
conversion yield of cellulose (i.e., glucan) was lower than that of hemicel-
lulose (i.e., mannan, galactan, xylan, and arabinan) owing to the mild pre-
treatment condition aimed at maximizing hemicellulose hydrolysis. The
total soluble solids concentration of the liquid fraction of the pretreated
material was 99.8 g/L, and the sugar composition of the liquid fraction is
given in Table 2.
Yellow poplar sawdust was pretreated at 0.3% (w/w) sulfuric acid
and 195C for 5 min using a Sunds Hydrolyzer installed at the Process
Development Unit of the National Renewable Energy Laboratory (NREL)
in Golden, CO. Pretreatment of yellow poplar sawdust using the Sunds
Hydrolyzer was reported previously (11). Yellow poplar chips (pulp chip
size) were pretreated at 0.55% sulfuric acid and 170C for 15 min using a
Sunds Hydrolyzer installed at the Tennessee Valley Authority pilot plant
(Muscel Shoals, AL). The water-insoluble fraction of the yellow poplar
chips was 71.8% on a dry wt basis, and the soluble solids concentration of
the liquid fraction of the pretreated material was 144.6 g/L.
Extraction of Hemicellulose 257
Applied Biochemistry and Biotechnology Vol. 9193, 2001
Effect of Water Temperature on Extraction of Pretreated Softwood
To determine the effect of water temperature on the extraction of
soluble solids from the pretreated softwood, stagewise batch extraction
was carried out in a glass beaker. For a single-batch extraction, 50 g (wet wt)
of pretreated softwood chips (68.3% moisture content) was mixed with
280 mL of deionized water at 25, 40, 60, and 80C. The slurry was stirred for
2 min, then filtered using a vacuum Buchner filter to separate the liquid
from insoluble solids. The extracted solids were dried in a 105C oven
overnight. The soluble solids recovery was determined by subtracting the
dry wt of the extracted solids from the dry wt of the starting material. Data
were also collected for multiple (as many as five), consecutive batch extrac-
Table 1
Feedstock Composition and Theoretical Component Yields
of Pretreated Softwood
Theoretical yield
Feedstock composition after pretreatment
Component (%) (%)
Glucan 43.2
Unconverted 90.5
To monomeric glucose 10.9
To oligomeric glucose 1.1
To HMF
a
0.3
Unaccounted for 2.7
Mannan 11.5
Unconverted 9.8
To monomeric mannose 74.0
To oligomeric mannose 12.3
Mannan to HMF
a
2.2
Unaccounted for +1.7
Galactan 4.3
Unconverted 29.3
To monomeric galactose 61.3
To oligomeric galactose 10.1
Unaccounted for 0.7
Xylan 7.7
Unconverted 15.2
To monomeric xylose 76.4
To oligomeric xylose 8.9
To furfural 5.3
Unaccounted for 5.9
Arabinan 2.2
Unconverted 9.4
To monomeric arabinose 96.1
To oligomeric arabinose 9.9
Unaccounted for 15.3
a
5-hydroxymethyl-2-furaldehyde.
2
5
8
K
i
m

e
t

a
l
.
A
p
p
l
i
e
d

B
i
o
c
h
e
m
i
s
t
r
y

a
n
d

B
i
o
t
e
c
h
n
o
l
o
g
y
V
o
l
.

9
1

9
3
,

2
0
0
1
Table 2
Sugar Composition of Liquid Fraction
of Starting Pretreated Softwood and Liquid Extract from Continuous Countercurrent Extraction (g/L)
Liquid Cellobiose
a
Glucose Xylose Galactose Arabinose Mannose
Liquid fraction 1.9 17.4 22.6 9.4 7.7 32.4
of pretreated softwood
Extract from L/IS = 2.1 ND 13.3 16.4 7.8 6.2 27.8
Extract from L/IS = 3.4 ND 7.6 10.9 5.2 4.3 17.1
Extract from L/IS = 5.6 ND 5.4 7.3 4.2 2.0 10.2
a
ND, not detected by HPLC.
Extraction of Hemicellulose 259
Applied Biochemistry and Biotechnology Vol. 9193, 2001
tions. For a two-batch extraction, the filtered solids obtained from the first
extraction were reslurried with 280 mL of deionized water, stirred, and
filtered. This was repeated as many times as the number of extraction
batches before the extracted solids were dried in the 105C oven. The weight
ratio of added water to dry insoluble solids (L/IS) was 24 for the single-
batch extraction, 47 for the two-batch extraction, and 118 for the five-batch
extraction. At 80C, the resulting soluble solids concentrations were 1.2,
0.25, 0.11, 0.05, and 0.02% (w/w) corresponding to the number of extraction
of 1, 2, 3, 4, and 5, respectively.
Drainage Rate of Pretreated Biomass
The continuous countercurrent screw extractor used in the present
study relies on percolation of water by gravity through the pretreated
biomass. If the pretreated biomass has poor water drainage proper-
ties (i.e., very slow drainage rate), channeling or blockage may occur inside
the extractor, which can result in low sugar recovery or low throughput.
Therefore, bench-scale percolation tests were performed using silicone
columns to compare the water drainage rates for three pretreated materi-
als: softwood chips, yellow poplar sawdust, and yellow poplar chips.
Pretreated wood residues (15.6 g on a dry wt basis) were placed in a
2.5-cm (1-in.) diameter 30.5-cm (12-in.) high silicone column, which was
fitted with a filter at the bottom. Hot water at 60C was added to the top of
the column. To keep the total slurry concentration in the column in the first
percolation batch (15.2% on a dry wt basis) constant per pretreated mate-
rial, the weight ratio of added water to total dry solids (L/S) was varied
depending on pretreated materials. The L/S ratios in the first batch were
3.4, 4.4, and 3.9 for the pretreated softwood, yellow poplar sawdust, and
yellow poplar chips, respectively. Seven consecutive percolations, each
with the same amount of water, were performed on the same column to
determine whether the drainage rate changed as the amount of water used
increased. The time period required for the liquid to completely drain from
the column was recorded. The average drainage rate was calculated by
dividing the mass of liquid collected by the draining time.
Countercurrent Extraction of Pretreated Biomass
Figure 3 shows a schematic diagram of the pilot-scale continuous coun-
tercurrent extractor designed by NREL. A 10-cm (4-in.) diameter 305-cm
(10-ft) long, U-trough screw conveyor, driven by a Link-Belt

drive
(Rexnord, Philadelphia, PA), was purchased from FMC (Tupelo, MS) and
modified for the purpose of this process. The helical screw has 36 short-
pitch flights. Half-inch holes were drilled into the flights and mixing
paddles were installed between the flights to improve liquid/solid contact.
Because the solids discharge opening was 31 cm (1 ft) from the top of the
conveyor, the effective length of the extraction zone was only 274 cm (9 ft).
The total working volume of the extractor was 28 L. To minimize channel-
260 Kim et al.
Applied Biochemistry and Biotechnology Vol. 9193, 2001
ing of water along the bottom of the trough, the screw extractor was
mounted with an inclined angle of 50 from horizontal.
At the beginning of a run, a batch of pretreated softwood was loaded
into a constant volumetric feeder (Acrison

feeder BDFM; Acrison,

Moonachie, NJ). Process water was heated in the water heater to 60C, then
circulated through the screw conveyor jacket. When the return water tem-
perature reached a steady-state value of about 57C, the feeder was then
switched on and set at a predetermined feed rate to begin to introduce
pretreated wood into the bottom of the extractor. The conveyor drive was
then activated and set at a predetermined forward speed such that the
flights were less than 50% filled with pretreated wood. At about the same
time, a split stream of hot water was metered through a rotameter and
sprayed on top of the pretreated material through a spray nozzle
installed approx 31 cm (1 ft) upstream of the solids discharge opening.
The extracted liquid passed through coarse filters at the bottom of the
extractor and was collected every 5 min into a flask connected to a vacuum
pump. The extracted solids were discharged into a barrel via the solids
discharge chute at the top of the extractor. The Acrison feeder, extract
receiving flask, and barrel for receiving discharged solids were placed on
electronic balances, and their weight changes were recorded every 5 min.
Each extraction run lasted 110120 min. Steady state (i.e., no appreciable
change in pH of the extracted solids) was obtained after approx 60 min.
At the end of each run, samples of solids inside the extractor were taken at
2-ft intervals and analyzed for insoluble solids and percentage of soluble
solids extracted.
Fig. 3. Schematic of the pilot-scale continuous countercurrent extractor.
Extraction of Hemicellulose 261
Applied Biochemistry and Biotechnology Vol. 9193, 2001
Analysis of Extracted Solids and Liquid
The concentration of monomeric sugars in the extract was analyzed by
high-performance liquid chromatography (HPLC) using the same method
for hydrolysate liquor analysis as described previously (12). The fraction of
insoluble solids (FIS) of feed materials and extracted solids were deter-
mined to find the amount of solubles extracted from pretreated wood by
the screw extractor. The FIS is defined as the dry weight ratio of water-
insoluble solids over the unwashed solids.
To determine the FIS, approx 7 L of tap water at 40C was added to
solids with known total weight and solid content (approx 180 g of total
dry solids or 120 g of insoluble solids for pretreated softwood) in a con-
tainer. The resulting slurry was mixed vigorously with a portable mixer
for 5 min and left standing for 5 min. The approximate L/S and L/IS
ratios of the slurry were 44 and 58, respectively. The slurry was fil-
tered with a 24-cm diameter glass-fiber filter (1.5-m particle retention,
Whatman grade 934-AH; Whatman, Maidstone, England) under a
vacuum and resuspended with tap water for the next washing. These
procedures were repeated at least three more times or until the pH of
the slurry was higher than 6.0, and the solids were then washed once
more with 40C deionized water. The total L/S and L/IS ratios used in
determining the FIS of pretreated softwood were 220 and 290, respec-
tively. All the washed solids were recovered, mixed, and weighed. Three
representative samples were dried overnight in a 105C oven to deter-
mine the solid content of the washed solids. The FIS values for pretreated
biomass are generally in the 0.650.80 range. We define 100% soluble
solids (or solubles) recovery as 100 (1 FIS). The percentage of solubles
recovery yields of partially extracted solids were calculated according to
the following equation:
Percentage of solubles recovery = [(1 FIS
i
)/(1 FIS
0
)] 100 (1)
in which FIS
i
is the FIS of the partially extracted material, and FIS
0
is the FIS
of the starting material.
Fitting Empirical Equation to Experimental Data
The following empirical equation was fitted to the experimental data
to predict the recovery yield of solubles against the extractor length and the
L/IS ratio:
R = 100 a exp(bL) (2)
in which R is the recovery yield of solubles (%), a and b are constants, and
L is either the length of extractor or L/IS ratio. Nonlinear regression was
carried out using the scientific graphing software SigmaPlot

(Jandel,
Chicago, IL) to determine constants a and b with different operating
parameters.
262 Kim et al.
Applied Biochemistry and Biotechnology Vol. 9193, 2001
Results and Discussion
Effect of Water Temperature on Extraction of Pretreated Softwood
As shown in Fig. 4, the recovery of solubles from the pretreated soft-
wood forest thinnings increased significantly as the wash water tempera-
ture was raised from 25 to 80C. This effect of water temperature was
especially pronounced when low amounts of water were used such as in
single- and double-batch extractions, which were carried out with an L/IS
ratio of 24 and 47, respectively. Therefore, it is important to use hot water
in countercurrent extraction in which a low L/IS ratio is maintained to
obtain high extraction efficiency. In the pilot-scale continuous extraction
experiments, 57C water was sprayed on the pretreated biomass and 60C
water was circulated through the screw conveyor jacket.
Comparison of Drainage Rates
Pretreated biomass has different particle sizes depending on the par-
ticle size of the starting materials and the pretreatment conditions. If the
particle size is too small to give an adequate drainage rate, channeling and
blockage of liquid flow may become serious problems for countercurrent
screw extractors. In the present study, drainage tests were performed on
three different pretreated wood residues; Figure 5 shows the results. The
drainage rate through pretreated softwood chips was significantly higher
than those obtained with pretreated yellow poplar sawdust and chips,
which contain a large amount of fines. This higher drainage rate of the
pretreated softwood was most likely attributable to its large particle sizes
in comparison to the pretreated yellow poplar materials. The increase in
Fig. 4. Effect of wash water temperature on the extraction of solubles from pre-
treated softwood.
Extraction of Hemicellulose 263
Applied Biochemistry and Biotechnology Vol. 9193, 2001
drainage resistance probably was caused by the compaction of the bed of
washed materials. The bed heights of both the pretreated yellow poplar
residues shrank approx 40% after four consecutive extraction batches,
whereas that of the pretreated softwood shrank about 25%. A similar pack-
ing problem in a column extractor was also reported in oil extraction from
fine soybean flour (13). The drainage rates of pretreated wood residues
generally decreased as the number of percolation batches increased and
later on remained constant. This was probably because water-absorbed
wood matrix hampered the water drain in the packed column. Basing our
selection on the results of the drainage tests, we chose pretreated softwood
for running extraction experiments with the countercurrent extractor.
The two pretreated yellow poplar materials with lower water drainage
rates will be considered in future studies.
Countercurrent Extraction of Pretreated Softwood Forest Thinnings
Countercurrent extraction of pretreated softwood forest thinnings
was performed at three different L/IS ratios and a fixed solid feed rate of
approx 220 g/min (228, 209, and 234 g/min for L/IS = 2.1, 3.4, and 5.6,
respectively). The L/IS ratio was determined by dividing the amount of
water in the extract by the amount of insoluble solids in solids feed. The
wash water flow rate was varied for different L/IS ratios (175, 240, and
400 g/min for L/IS = 2.1, 3.4, and 5.6, respectively) and the solids feed rate
was kept constant for all runs. Because the screw rotation speed was fixed
at 20 rpm for the entire extraction runs, the average solids residence times
in the extractor were essentially the same (approx 20 min) for all runs.
Figure 6 shows the concentrations of solubles and pH of extracts
recovered during the steady-state operation of the continuous countercur-
rent extractor at various L/IS ratios. The concentration of solubles in the
Fig. 5. Comparison of water drainage rates for different pretreated materials using
a 1-in. (2.5-cm) diameter 12-in. (30.5-cm) high column.
264 Kim et al.
Applied Biochemistry and Biotechnology Vol. 9193, 2001
extract decreased as the L/IS increased because of dilution of the sugar
solution contained in the pretreated wood residues at higher water flow
rates. Therefore, it is necessary to keep the L/IS ratio low to achieve high
solute concentrations in the extract as long as the extraction efficiency is
maintained at an adequate level. The increase in extract pH with the L/IS
ratio also indicates the dilution effect at higher water flow rates. Table 2 lists
the sugar composition of the liquid fraction of the starting pretreated
material (i.e., before extraction) and extracts collected from the countercur-
rent extraction runs at the different L/IS ratios. As expected, the sugar
concentration decreased as the L/IS ratio was raised.
Fig. 6. Extract pH and concentration of solubles in extract recovered from the bot-
tom of the extractor in continuous countercurrent extraction of pretreated softwood at
different L/IS ratios.
Fig. 7. Percentage of insolubles in extracted solids discharged from the top of the
extractor in continuous countercurrent extraction of pretreated softwood at different
L/IS ratios.
Extraction of Hemicellulose 265
Applied Biochemistry and Biotechnology Vol. 9193, 2001
Figure 7 illustrates the percentage of insolubles in extracted solids
discharged during the steady-state operation of the countercurrent extrac-
tor. Higher FIS (i.e., insoluble fraction in extracted solids) values were
obtained at higher L/IS ratios. This result indicates that higher amounts of
soluble solids were extracted at higher L/IS ratios. The lower value of
percentage of insolubles in extracted solids at 70-min extraction time from
L/IS = 3.4 can be attributed to a disruption in steady-state operation when
the clogged filters were replaced. The gradual decline in FIS value for the
run of L/IS = 3.4 was most likely caused by plugging of the filters at the
bottom of the extractor. After the filter was replaced at 70 min, the extrac-
tion efficiency improved, as indicated by the rising FIS trend. We installed
a different type of filter for the other two runs and did not observe severe
plugging problems.
Operating Line
To establish the operating line of the recovery of solubles with respect
to location in the extractor, extracted solids remaining in the extractor were
recovered at the end of each extraction run and analyzed for percentage of
solubles extracted. The experimental data are indicated by closed symbols
in Fig. 8. The solid lines represent the best-fitted operating lines to the
experimental data by the empirical equation (Eq. 2).
The percentage of solubles recovery increased as the distance from the
bottom of the extractor increased. This was confirmed by an increase in pH
of the liquid in the extracted solids. At 274 cm (9 ft) from the bottom of the
extractor, where solids were discharged, the solubles recoveries for L/IS
ratios of 5.6, 3.4, and 2.1 were 98, 91, and 77%, respectively. In plant opera-
Fig. 8. Solubles recovery and pH of liquid in extracted solids at different locations
in the continuous countercurrent extractor operated at steady state after 120 min.
Closed symbols indicate experimental data and lines represent the predicted curves.
266 Kim et al.
Applied Biochemistry and Biotechnology Vol. 9193, 2001
tion, assuming that the extracted solids would be pressed to approx 45%
(w/w) solid content to recover the entrained solubles and the obtained
liquid returned to the extractor, one would expect a slight enhancement in
the solubles recovery, as represented by the open symbols in Fig. 8. The
soluble recovery was higher at a higher L/IS ratio for the same extractor
length. The values of constants a and b for the empirical equation used
(Eq. 2) for plotting the operating lines in Fig. 8 were estimated to be 99.9120
and 0.1910, 99.9938 and 0.2327, and 99.9710 and 0.6251 for L/IS = 2.1, 3.4,
and 5.6, respectively.
Figure 9 presents the prediction of soluble recoveries when the length
of the extractor is extended to increase the extraction stages. The estimated
values of constants a and b for Eq. 1 used in plotting the predicted curves
in Fig. 9 were 99.9950 and 0.8937, 99.9969 and 0.9926, 99.9981 and 1.0920,
and 99.9988 and 1.1916 for extractor lengths of 274.3, 304.8, 335.3, and
365.8 cm, respectively (9, 10, 11, and 12 ft, respectively). A 12-ft extractor
with the same configuration as the current extractor used in this work is
expected to achieve >95% recovery of solubles at an L/IS ratio of 3.0.
This predicted performance is better than the predicted 93.8% soluble
recovery for the three-stage stagewise countercurrent washer mentioned
earlier (Fig. 2).
Conclusion
We have demonstrated that continuous countercurrent extraction of
hemicellulosic sugars from pretreated softwood residues using a pilot-
scale screw extractor can be effectively achieved. The soluble recovery yield
decreased as L/IS ratio was reduced. The empirical equation predicts that
Fig. 9. Prediction of soluble recovery with respect to L/IS ratio for continuous
countercurrent extraction of the pretreated softwood with different extractor lengths.
Extraction of Hemicellulose 267
Applied Biochemistry and Biotechnology Vol. 9193, 2001
adequate recovery of soluble sugars can be obtained in the low-range L/IS
ratio of 2.53.0, if the length of the extractor is extended to about 366 cm (12 ft).
Acknowledgment
This work was funded by the U.S. Department of Energy, Office of
Fuels Development.
References
1. Nguyen, Q. A., Tucker, M. P., Keller, F. A., and Eddy, F. P. (2000), Appl. Biochem.
Biotechnol. 8486, 561576.
2. Nguyen, Q. A. and Aden, A. (1999), Report no. 4083, National Renewable Energy
Laboratory, Golden, CO.
3. Schell, D. (1997), Report no. 4115, National Renewable Energy Laboratory, Golden, CO.
4. Schwartzberg, H. G. (1980), Chem. Eng. Prog. 76, 6785.
5. Smook, G. A. (1992), Handbook for Pulp and Paper Technologies, 2nd ed., Angus Wilde,
Vancouver, BC.
6. Rundle, K. W. (1989), US patent 4,873,095.
7. Noah, K. S. and Linden, J. C. (1989), Trans. ASAE 32, 14191425.
8. Nguyen, Q. A. (1993), Canadian patent 1,322,366.
9. Brinkley, C. R. and Wiley, R. C. (1978), J. Food Sci. 43, 10191023.
10. Gunasekaran, S., Fisher, R. J., and Casmir, D. J. (1989), J. Food Sci. 54, 12611265.
11. Tucker, M. P., Farmer, J. D., Keller, F. A., Schell, D. J., and Nguyen, Q. A. (1998),
Appl. Biochem. Biotechnol. 7072, 2535.
12. Nguyen, Q. A., Tucker, M. P., Boynton, B. L., Keller, F. A., and Schell, D. J. (1998),
Appl. Biochem. Biotechnol. 7072, 7787.
13. Nieh, C. D. and Snyder, H. E. (1991), J. Am. Oil Chem. Soc. 68, 246249.
Continuous Countercurrent Extraction of Hemicellulose from Pretreated
Wood Residues