Pretreatment and Fractionation of Corn Stover by Ammonia Recycle
Percolation (ARP) Process
Tae Hyun Kim and Yoon Y. Lee
Department of Chemical Engineering, Auburn University, AL 36849, U.S.A.
Corresponding author: Y.Y. Lee
Department of Chemical Engineering, Auburn University, AL 36849
yylee@eng.auburn.edu, Phone) 334-844-2019,
Fax) 334-844-2063
INTRODUCTION
A pretreatment method based on aqueous ammonia is being developed in our
laboratory. It utilizes aqueous ammonia in a recycle mode, thus termed as Ammonia
Recycled Percolation (ARP). Its process flow diagram is shown in Fig. 1. Ammonia is an
efficient delignification reagent. The ARP is a pretreatment method that raises the
enzymatic digestibility and also achieves high degree of delignification.
The primary factors influencing the reactions occurring in the ARP are reaction
time, temperature, ammonia concentrations, and the amount of liquid throughput. Of
these, the liquid throughput and the reaction temperature were identified as the factors
most sensitively affecting the processing cost. The main interest in early phase of this
work was to minimize the liquid input and adjust other conditions so as to attain
acceptable levels of digestibility and delignification. During the ARP, a significant
fraction of hemicellulose is also removed along with lignin. This partial separation of
1
hemicellulose makes the process of utilizing hemicellulose sugars complicated. To
resolve this problem, a two-stage pretreatment was investigated. This process scheme is
designed to separate hemicellulose sugars in the first stage by hot water, and separate
lignin in the second stage by the ARP. The remaining solid thus contains mostly cellulose.
Upon completion of this process, fractionation of biomass is to be achieved. Fractionation
of biomass would improve the overall biomass conversion process since each of the
biomass constituents are utilized more efficiently.
This paper describes the process conditions of the ARP and the two-stage process,
and assesses their performance as a pretreatment method. The samples of these processes
were analyzed for composition and the enzymatic digestibilities were determined.
Samples were further examined and analyzed by SEM, X-Ray crystallography, and FTIR.
The results of the various solid analyses are reviewed in connection with the enzymatic
digestibility.
2
MATERIALS AND METHODS
Material
Air-dried ground corn stover was supplied by the National Renewable Energy
Laboratory (NREL, Golden, CO). The corn stover was screened to a nominal size of
935 mesh. The initial composition of the corn stover, as determined by NREL, was: 36.1
wt.% glucan, 21.4 wt.% xylan, 3.5 wt.% arabinan, 1.8 wt.% mannan, 2.5 wt.% galactan,
17.2 wt.% Klason lignin, 7.1 wt.% ash, 3.2 wt.% acetyl group, 4.0 wt.% protein, and 3.6
wt.% uronic acid. -Cellulose was purchased from Sigma (Cat. No. C-8200, Lot No.
11K0246). Cellulase enzyme, Spezyme CP (Genencor, Lot No. 301-00348-257), was
obtained from NREL. The average activity of the enzyme, as determined by NREL, was:
31.2 filter paper unit (FPU)/mL. Activity of -glucosidase (Novozyme 188 from Novo
Inc., Lot No. 11K1088) was 750 CBU/mL.
Simultaneous saccharification and fermentation (SSF)
The microorganism used in the SSF was Saccharomyces cerevisiae ATCC®
200062 (NREL-D5A). The growth media was YP medium, which contained 1% yeast
extract (Sigma Cat. No. Y-0500) and 2% peptone (Sigma Cat. No. P-6588).
Experimental setup and operation
Figure 2 shows the overall layout of the ARP apparatus. The system consists of a
stock solution reservoir, pump, temperature-programmable oven, SS-316 column reactor
(9/10 in. ID  10 in. L, internal volume of 101.9 cm3), and liquid holding tank. The
reactor was operated in a flow-through mode, in which the liquid flows through a reactor
column packed with biomass. The reactor system was pressurized by nitrogen at 2.3 MPa
3
to prevent flash evaporation. In a typical ARP experiment, 15 g of biomass were packed
into the reactor.
RESULTS
ARP Pretreatment
Corn stover was pretreated with aqueous ammonia in a flow-through column
reactor, a process termed as Ammonia Recycle Percolation (ARP). The ARP
pretreatment gives a high and adjustable degree of delignification. The most significant
composition change after pretreatment is in the lignin. The delignification reaction is
rapid; 70% of the lignin was removed within 10 minutes of treatment. The ARP removes
75–85% of the total lignin and solubilizes 50–60% of hemicellulose, but retains more
than 92% of the cellulose content (Table 1 and Fig. 3). Decomposition of carbohydrates
during the pretreatment is insignificant. ARP treatment of corn stover for 90 min renders
near quantitative enzymatic digestibility with 60 FPU/g-glucan and above 93%
digestibility with 10 FPU/g-glucan (Table 1),
Since a large quantity of xylan and lignin was removed, it became of interest to
examine the physical changes of the biomass. The SEM pictures of the treated samples
(Fig. 4) indicate that the micro-fibrils are separated from the initial connected structure
and fully exposed. It would certainly increase the external surface area and the porosity
of the biomass. We also find that the wet treated biomass is much softer than the wet
untreated biomass.
The crystallinity of biomass has been frequently brought up as a factor
influencing the enzymatic hydrolysis. We measured the X-ray diffraction pattern of
4
treated, untreated corn stover, and -cellulose (Fig. 5) from which the crystallinity
indexes were determined. The main effect on the composition of the ARP treatment is
removal of xylan and lignin that are of amorphous structure. The crystallinity index is
influenced by the composition of the biomass (Fig.3). The increase of CrI is caused by
the removal of amorphous substances in the biomass (lignin + xylan), which raises the
enzymatic digestibility. However, there is no indication that the crystalline structure of
the glucan content of the biomass is changed as the result of ARP treatment. We therefore
find no direct relationship between the digestibility and the CrI.
Infrared spectroscopy is an analytical method frequently used for investigation of
the structure of constituents and its chemical changes taking place in lignocellulosic
materials. In this work, diffuse reflectance infrared (DRIFT) spectra were measured for
study of the change of chemical structure brought upon pretreatment, particularly the
lignin content. In Fig.6, the bands of lignin in the ARP treated sample are found to be
lower than that of the untreated sample, a positive proof of the delignification effect of
ARP.
Low-liquid ARP
Amount of liquid throughput is one of the major cost factors in the ARP. A
modification of the original ARP was made in which the amount of liquid input and the
residence time is minimized. We find that the liquid input and residence time can be
reduced to 3.3 mL/g-biomass and 10–12 min without adversely affecting the overall
performance of the ARP. Low-liquid ARP shows that 0.47 g of NH3 and 2.7 g of water
are required for pretreatment of one g of corn stover. The low-liquid ARP at 170 oC was
5
as effective as the conventional ARP in that it achieves 73.4% delignification and 88.5%
digestibility with enzyme loading of 15 FPU/g glucan (Fig.7-a). The data of Fig. 7-b also
indicate that cellulase enzyme (Spezyme CP) has xylanase activity to give the xylan-toxylose yield only slightly lower than that of glucan.
Simultaneous saccharification and fermentation (SSF) of ARP-treated corn
stover and α-cellulose was performed with Saccharomyces cerevisiae ATCC® 200062
(NREL-D5A). Low-liquid ARP-treated solid samples were prepared under the optimum
conditions of 170 ºC, 3.3 mL of 15 wt.% ammonia per gram of corn stover liquid input, 5
mL/min flow rate, and 10 min reaction time. The glucan content of the pretreated corn
stover was 62%. The ethanol yields from glucose for Low-liquid ARP-treated corn stover
and α-cellulose are presented in Fig. 8. With an initial loading of 3% w/v glucan, the
ethanol yield of treated corn stover reached 84% of the theoretical maximum at 96 h (Fig.
8-a). The ethanol yield using 6% w/v glucan loading is presented in Fig. 8-(b) and gave
about the same level of ethanol yield, 84% at 120 h, as the 3% w/v glucan loading case.
The glucose level in the broth decreased to and stayed at near zero level after 12 h for 3%
w/v glucan loading and 24 h for 6% w/v glucan loading. The SSF therefore proceeded
mostly under glucose limited condition.
Two-stage percolation process
To prevent hemicellulose loss during the ARP, we have devised a two-stage
process in which hot-water percolation process and the ARP are operated in succession.
This process scheme is designed to separate hemicellulose sugars in the first stage and
lignin in the second stage. Two-stage processing of corn stover (hot water treatment
6
followed by ARP) can effectively fractionate corn stover into three main constituents.
The results from two-stage process are summarized in Table 2. The conditions of the
ARP treatment were fixed: 170oC, 15 wt% NH3, 30 min or 60 min and five different
temperatures covering 170210oC were applied in hot-water treatment stage. The end
product of two-stage pretreatment contains about 82% glucan, a product equivalent to a
filler-fiber used in papermaking. Under the optimum condition, on the basis of
fractionation and digestibility (190oC, 30 min of water treatment and 170oC, 60 min of
ARP at 2.3 MPa, 5 mL/min of flow rate), corn stover treated by two-stage processing
gives digestibility slightly lower than that of ARP alone (93.6% at 60 FPU/g-glucan and
84.8% at 15 FPU/g-glucan). Hot-water treatment alone at 210–220oC gives exceptionally
high digestibility. Two-stage processing above 200oC increases the residual “Klason
lignin”, an indication that lignin recondensation and/or lignin-carbohydrate complex may
have occurred (Table 2). It was also observed that the Klason lignin increases above 200–
210oC from both water only pretreatment and water-ARP pretreatment. High temperature
would induce faster lignin dissolution. It appears that the solublilized lignin re-condenses
onto the biomass during the latter phase of the treatment, thus increasing the net lignin
content in biomass. This becomes more discernible in the water-ARP pretreatment
because most lignin is removed from the second stage (ARP). From Table 2, it is
noticeable that the lignin content is directly correlated to the digestibility, especially the
digestibility with 15 FPU enzyme loading.
7
CONCLUSIONS
The ARP is highly effective in delignifying biomass and increasing the enzymatic
digestibility. Lignin removal by ARP is 7085%. Digestibility of the ARP treated corn
stover is near quantitative with 60 FPU/g-glucan and above 93% with 10 FPU/g-glucan.
The SEM pictures indicate that the biomass structure is deformed and its fibers are
exposed by the pretreatment. ARP pretreatment increases the crystallinity index as the
amorphous portion of biomass is removed. The crystalline structure of the biomass
cellulose, however, is not altered by the ARP treatment. Low-liquid ammonia treatment
reduces the liquid throughput to the level of 3.3 mL of liquid per gram of corn stover,
leading to a shorter residence time and lower energy requirements. A higher fraction of
xylan is retained than in conventional ARP. A high degree of delignification is not
necessary to attain high enzymatic digestibility or high ethanol yield.
The overall ethanol yield of 85% of the theoretical maximum was achieved in
SSF tests using corn stover treated by Low-liquid ARP. The two-stage process combines
hot-water and ARP treatments. The hot water processing removes mostly hemicellulose,
and the subsequent ARP removes lignin. With two-stage treatment (190oC, 30 min of
water treatment followed by 170oC, 60 min of ARP at 2.3 MPa, 5 mL/min of flow rate), a
virtual fractionation of the biomass was achieved: 9295% of xylan hydrolysis with
8386% recovery yields, and 7581% of lignin removal. The solid residue after twostage treatment contained 7885% cellulose.
8
Table 1
Effect of reaction time on the composition in ARP treatment. 1
Time
Solid
S.R.2
Liquid
K-Lignin3 A.S.L.4 Glucan
[min]
[%]
[%]
Untreated
100
10
20
Xylan
Total
Glucan Xylan
Delign.
Glucan
Xylan
Digestibility 5
60 FPU 10 FPU
[%]
[%]
[%]
[%]
[%]
[%]
[%]
[%]
[%]
[%]
17.2
2.3
36.1
21.4
0.0
0.0
36.1
21.4
0.0
21.2
14.3
 0.4
 0.2
 0.3
 0.2
 0.3
 0.2
61.4
5.5
1.1
35.4
12.8
0.9
8.8
36.3
21.6
 2.5
 0.4
 0.2
 0.4
 0.2
 0.1
 0.3
 0.5
 1.5
 1.2
67.9
92.2
83.9
 0.5
 2.2
 2.4
 2.9
88.3
58.5
3.4
1.0
34.7
11.0
1.1
11.0
35.8
22.0
80.5
92.0
 1.5
 0.2
 0.2
 0.4
 0.4
 0.2
 0.1
 0.2
 0.5
 1.1
 1.1
 2.1
40
56.4
3.1
0.9
34.6
10.5
1.3
11.2
35.9
21.7
82.2
95.1
89.2
 2.1
 0.2
 0.2
 0.4
 0.1
 0.2
 0.1
 0.2
 0.0
 1.0
 2.1
 0.4
60
55.2
2.8
0.8
34.0
9.7
1.6
11.7
35.6
21.4
83.9
94.4
88.3
 1.0
 0.1
 0.2
 0.5
 0.3
 0.2
 0.2
 0.3
 0.5
 0.6
 2.3
 1.8
92.5
90
-Cellulose
53.6
2.6
0.8
33.1
9.2
1.8
12.2
34.9
21.5
84.7
99.6
 1.1
 0.2
 0.2
 0.6
 0.3
 0.2
 0.2
 0.8
 0.5
 1.1
 2.5
 2.0
100
0.0
0.0
91.4
4.0
0.0
0.0
91.4
4.0
0.0
93.4
71.7
 0.8
 0.3
 0.8
 0.3
 1.5
 1.4
Notes. 1. All sugar and lignin content based on the oven-dry untreated biomass. Values are expressed as mean and standard deviation. Pretreatment conditions:
15 wt% of ammonia, 170C, 5 mL/min of flow rate, 2.3 MPa
2. S.R.: solid remaining after reaction
3. Klason lignin
4. A.S.L. stands for acid soluble lignin during analysis
5. Conditions of Enzymatic hydrolysis: 72 h, 60 or 10 FPU/g of glucan, pH 4.8, 50C, 150 rpm.
9
Table 2
Effect of temperature on composition of biomass in two-stage (hot water-ARP) treatment 1
Temp.
[°C]
Untreated
Solid
S.R.2 K-Lignin3 Glucan Xylan
[%]
[%]
[%]
[%]
100
17.2
Hot water only
65.4
13.5
170
58.1
11.8
180
55.0
11.3
190
53.0
10.3
200
51.0
8.8
210
50.5
10.1
220
Hot water-ARP
47.7
3.7
170
47.0
3.8
180
44.9
4.4
190
44.7
5.7
200
43.2
5.4
210
Liquid
Glucan Xylan
[%]
[%]
Total
Glucan Xylan
[%]
[%]
Yield in liquid
Glucan Xylan
[%]
[%]
Digestibility4
60 FPU
15 FPU
[%]
[%]
37.5
20.8
-
-
36.1
21.4
-
-
21.2
16.1
35.4
34.6
35.0
34.9
34.4
33.4
8.2
4.4
2.7
1.4
1.1
0.1
1.0
1.3
1.4
1.8
1.9
2.7
12.4
16.1
18.4
18.2
18.1
13.5
35.4
35.9
36.4
36.7
36.3
36.1
20.6
20.5
21.1
19.6
19.2
13.5
2.7
3.4
4.1
5.1
5.3
7.5
57.9
75.2
86.0
85.1
84.6
63.1
59.1
73.9
86.8
90.9
93.6
95.0
47.5
62.7
71.6
76.6
88.9
93.3
34.4
33.5
33.7
32.0
32.0
3.8
2.8
1.7
1.2
0.7
1.5
1.8
1.6
2.1
2.3
14.9
15.9
17.9
18.0
15.8
35.9
35.3
35.3
34.1
34.3
18.7
18.7
19.6
19.2
16.5
3.8
4.7
4.3
5.5
5.9
69.6
74.3
83.6
84.1
73.8
96.3
95.9
93.6
94.5
94.7
87.0
87.1
84.8
79.6
82.7
Note. 1. Data in the table based on the oven dry untreated biomass; Pretreatment conditions: 5.0 mL/min, 30 min, 2.5 MPa; ARP-15 wt.% of ammonia, 5.0
mL/min, 60 min, 2.5 MPa; •All reactions are carried out in a Bed-Shrinking Flow-Through (BSFT) Reactor.2. S.R. stands for solid remaining after
reaction.; 3. Klason lignin;
4. Digestibility at 72 h, Enzymatic hydrolysis conditions: 60 or 15FPU/g of glucan, pH 4.8, 50°C, 150 rpm
5. The data in the table show the mean value (n=2; SE 0.1% for K-lignin, SE<0.8% for S.R., SE<0.2% for Glucan and Xylan in solid and liquid,
SE<2.0% for digestibilities, SE : standard error).
10
Ammonia recycling
Make-up water
Biomass
(Corn stover)
Ammonia
Continuous Reactor
Steam
Liquid
Steam
Solid
Evaporator
Water
Pump
Washing
Crystallizer
Lignin & Other
sugar
Steam
Ethanol
(Bio-fuel)
Fermentor (SSF)
Soluble sugar
Washing
Fig. 1. Projected ARP (Ammonia Recycled Percolation) Process Diagram
11
Ethanol Yield [% of Theoretical] or Glucose Conc. [g/L]
(b)
Lignin
100
(Fuel)
80
60
40
20
0
0
24
48
Water
Aqueous Ammonia
Temp.
monitoring
system (DAS)
N2 Gas
C.W.
3-way
v/v
TG
PG
TG
#1
#2
N2 Cylinder
Pump
PG : Press. Gauge
TG : Temp. Gauge
C.W.: Cooling
Vent
PG
Oven
(Preheating Coil and Reactor)
Water
Fig. 2. Experimental set-up of laboratory ARP
12
Holding Tank
#1 : For Hot-water
#2 : For ARP
100
Digestibility at 60 FPU/g of glucan
90
Solid composition [%]
20
80
CrI
70
15
60
50
10
40
30
5
20
10
0
Enzymatic digestibility [%] or Crystallinity index [ ]
25
0
0
10
20
40
60
90
Reaction time [min]
Xylan remaining
Lignin remaining
Digestibility (60 FPU/g of glucan)
CrI (Crystallinity index)
Fig. 3. Effect of reaction time on solid composition, crystallinity index (CrI) and
enzymatic digestibility in ARP 1
Note. 1. All sugar and lignin content based on the oven-dry untreated biomass. Pretreatment
conditions: 15 wt% of ammonia, 170C, 5 mL/min of flow rate, 2.3 MPa
Enzymatic hydrolysis conditions: 72 h, 60 FPU/g of glucan, pH 4.8, 50C, 150 rpm.
13
(a) Untreated (X50)1
(b) ARP-90min (X50) 1
(c) Untreated (X300) 1
(d) ARP-90min (X300) 1
Fig. 4. Scanning electron micrographs (SEM) of treated and untreated corn stover.
Note 1. For example; ARP 90 min (X50) means 90 min of reaction time by ARP and X50 of manification
14
7000
Untreated
ARP-10
ARP-20
ARP-40
ARP-60
ARP-90
α-Cellulose
6000
Intensity
5000
4000
-cellulose
3000
2000
Untreated
1000
0
10
15
20
25
30
2
Fig. 5. XRD diagram of ARP-treated samples 1
Note 1. Number in legend indicates reaction time (e.g. ARP-10 indicating 10 min. of ARP treatment).
Pretreatment conditions: 15 wt% of ammonia, 170C, 5 mL/min of flow rate, 2.3 MPa
15
30
Untreated
ARP 10-min
ARP 20-min
ARP 30-min
Intensity (Kubelka Munk)
25
20
(a)
(b) (c)
15
10
Untreated
5
0
1000
2000
3000
4000
Wavenumber [cm-1]
Fig. 6. FTIR spectra of various ARP-treated samples
Note. Number in legend indicates reaction time (e.g. ARP-10min is 10 min ARP treatment)
Pretreatment conditions: 15 wt% of ammonia, 170C, 5 mL/min of flow rate, 2.3 MPa
a. IR band of C-O in guaiacyl or syringyl ring
b. IR band of aromatic skeletal vibration + C=O stretching
c. IR band of aromatic skeletal vibration
16
(a) Glucan digestibility at 15 FPU/g-glucan
100
Enzymatic digestibility [%]
90
80
170°C
70
150°C
60
130°C
50
110°C
40
α-Cellulose
30
Untreated
20
10
0
0
48
24
72
Time [h]
(b) Xylan digestibility at 15 FPU/g-glucan
100
Enzymatic digestibility [%]
90
80
70
170?
60
150?
50
130?
40
110?
30
20
10
0
0
24
48
72
Time [h]
Fig. 7. Enzymatic digestibility of Low-liquid ARP-treated samples
Notes. Pretreatment conditions: 110–170C, 3.3 mL of 15 wt.% NH3 throughput per gram of corn stover,
5.0 mL/min of flow rate; Enzymatic hydrolysis conditions: 72 h, 15 FPU/g-glucan, pH 4.8, 50C,
150 rpm.
17
Ethanol Yield [% of Theoretical] or Glucose Conc. [g/L]
(a) 3% w/v of glucan Loading
100
80
60
40
Low-liquid ARP-Treated, EtOH Yield
Cellulose, EtOH Yield
Low-liquid ARP-Treated, Glucose
Cellulose, Glucose
20
0
0
24
48
72
96
120
144
168
Time [h]
Ethanol Yield [% of Theoretical] or Glucose Conc. [g/L]
(b) 6% w/v of glucan Loading
100
80
60
40
Low-liquid ARP-Treated, EtOH Yield
Cellulose, EtOH Yield
Low-liquid ARP-Treated, Glucose
Cellulose, Glucose
20
0
0
24
48
72
96
120
144
168
Time [h]
Fig. 8. Ethanol yield in SSF of Low-liquid ARP treated samples
Note. Pretreatment conditions: 170C, 3.3 mL of 15 wt.% NH3 throughput per gram of corn stover, 5.0
mL/min of flow rate; SSF test conditions: 15 FPU/g-glucan enzyme loading, D5A yeast in YP medium,
pH 5.0, 38C, 150 rpm; (a) n=2 for Low-liquid ARP, n=4 for α-cellulose (b) n=2 for Low-liquid ARP,
n=2 for α-cellulose
18