- Email: [email protected]
Structural changes of Salix miyabeana cellulose fibres during dilute-acid steam explosion: Impact of reaction temperature and retention time
Accepted Manuscript Title: Structural changes of Salix miyabeana cellulose fibres during dilute-acid steam explosion: impact of reaction temperature and retention time Author: Ch´erif Ibrahima Khalil Diop Jean-Michel Lavoie Michel A. Huneault PII: DOI: Reference:
S0144-8617(14)01136-9 http://dx.doi.org/doi:10.1016/j.carbpol.2014.11.031 CARP 9459
To appear in: Received date: Revised date: Accepted date:
22-9-2014 11-11-2014 16-11-2014
Please cite this article as: Diop, C. I. K., Lavoie, J.-M., and Huneault, M. A.,Structural changes of Salix miyabeana cellulose fibres during dilute-acid steam explosion: impact of reaction temperature and retention time, Carbohydrate Polymers (2014), http://dx.doi.org/10.1016/j.carbpol.2014.11.031 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Structural changes of Salix miyabeana cellulose fibres during dilute-acid steam explosion:
2
impact of reaction temperature and retention time.
3
ip t
4
Chérif Ibrahima Khalil Diop, Jean-Michel Lavoie, Michel A. Huneault*
6
Department of Chemical and Biotechnological Engineering, Université de Sherbrooke, 2500
7
Boulevard de l’Université, J1K 2R1, Sherbrooke, Québec, Canada
us
cr
5
8
an
9 10
13
Corresponding author: Professor Michel A. Huneault
14
Department of chemical and biotechnological engineering, Université de Sherbrooke, 2500
15
Boulevard de l’Université, J1K 2R1, Sherbrooke, Québec, Canada
16
E-mail: [email protected]
17
Telephone: +1 819 821-8000 (Ext. 65580)
19
te
Ac ce p
18
d
12
M
11
20 21 22 23
1 Page 1 of 33
Abstract:
25
Dilute-acid steam explosion of Salix myabeana have been carried out to understand the effect of
26
processing conditions, expressed through a severity factors (SFT), on the changes in cellulose
27
fibre structures in a perspective of using these in polymer composites. This thermo-chemico-
28
mechanical extraction lead to the isolation of cellulose fibres as observed by SEM images. Fibre
29
length as well as length to diameter aspect ratios decreased with the severity of the treatment.
30
Likewise, fibre whiteness diminished with an increasing severity factor, which could be a
31
tangible effect of physical degradation. Variations in cristallinity seemed to be dependent upon
32
the reaction temperature, generally decreasing with regards to retention time. Above a severity
33
threshold, a structural disorganization was observed. Overall, dilute-acid steam explosion was
34
shown to be a valuable cellulose extraction process that can provide a variety of fibre structures.
M
an
us
cr
ip t
24
d
35
Key words: Salix miyabeana cellulose fibres; Dilute-acid steam explosion; Severity factor;
37
Structural changes; Physical characterization.
39 40 41 42
Ac ce p
38
te
36
43 44 45 46
2 Page 2 of 33
47
1. Introduction:
48 49
Cellulose fibres are generally produced from lignocellulosic feedstock (wood, straw, grass, etc.) and they can be extracted and/or purified depending on the final targeted application.
51
Beside its large and well-known use in the pulp and paper industries, cellulose is increasingly
52
being considered as a potential feedstock for biofuel production, implying its hydrolysis to
53
glucose and fermentation to “cellulosic” ethanol (Berberi et al., 2014). However, the crystalline
54
fraction of the cellulose is the most resistant to acid-hydrolysis and thus may be found as a post-
55
treatment residue. In the current work, in an effort to minimize chemical and energy
56
requirements, it is envisioned that fermentable sugars could be produced out of the easily
57
hydrolysable amorphous cellulose fractions. The crystalline residues could then be used in a
58
second step as natural fibre reinforcement for polymer composites.
cr
us
an
M
d
Various standard pulping methods are available at an industrial level including but not
te
59
ip t
50
limited to mechanical, thermo-mechanical, chemical and chemico-mechanical pulping. Recent
61
studies by Liu, Ju, Li and Jiang (2014) and Wang, Li, Cao, and Tang (2011) investigated the use
62
of ionic liquid extraction of cellulose from various sources and showed that up to 62% cellulose
63
extraction could be obtained using these green solvents, which in addition can be recycled.
Ac ce p
60
64
Among the less common techniques, steam processes as steam explosion, which were
65
first introduced in 1926 by Mason (Mason, 1926) is getting increasing attention in the field of
66
biofuels. The technique is more challenging to scale to an industrial level than classical pulping
67
process although it allows efficient separation of the different constitutive fraction of biomass
68
(Lavoie, Capek-Menard, Gauvin and Chornet, 2010). The process involves cooking water-
69
impregnated lignocellulosic biomass under pressure at temperatures in the range of 160-260˚C.
3 Page 3 of 33
70
After a few minutes of reaction, the content is rapidly purged out of the pressurized vessel and
71
undergoes a rapid decompression or “explosion”. The process is environmentally friendly due to
72
the low acid concentration and no organic solvents. The hydrolytic power of the steam processes is often assessed through a severity factor,
74
which is a valuable tool used to quantify the intensity of the hydrolytic conditions of the steam
75
treatments. When using an acid catalyst, the role of the pH in the reaction is combined to the
76
other hydrolytic parameters (temperature and residence time) generating a combined severity
77
factor (Chum, Johnson, Black and Overend, 1990). The versatility of the severity factor has
78
been demonstrated by adapting it to lignin removal for a wide array of biomass and conversion
79
processes including steam processes (Lee and Lavoie, 2013). Compared to other pre-treatment
80
technologies, steam-explosion presents several advantages particularly for the biorefineries.
81
These advantages are summarized in the work of Garotte, Domínguez, and Parajó (1999). They
82
noted the high hemicellulose yield with by-products showing little degradation and disruption of
83
the solid residues from bundles to individual fibres that occur during the explosion effect. Steam-
84
explosion is generally performed without addition of any chemicals. When a dilute-acid (i.e. with
85
acid concentration ranging between 1 to 5%) is used as a catalyst, it increase the hydrolysis rate
86
of hemicellulose and thus further improves the process.
cr
us
an
M
d
te
Ac ce p
87
ip t
73
Many investigations reporting on the steam processes focused on the extractability of
88
cellulose as a source for high yield fermentable sugars (Mosier et al., 2005) (Brownell, Yu and
89
Saddler, 1986). However, fewer researchers considered the impact of the steam-process on the
90
fibre structure and properties when used for others purposes as in pharmaceutical, biocomposites
91
and biomaterials. The use of steam processes on lignocellulosic materials can lead to significant
92
changes in their structure and properties, such as reduction of the fibre length, fibre degradation
4 Page 4 of 33
93
and modification of the crystalline structure. This process should be better optimized to better
94
understand the parameter for the production of high quality natural fibres.
95
The utilisation of natural fibres as reinforcing fillers for polymer composites has previously been investigated in the open literature. As an example, Petinakis, Yu, Simon and
97
Dean (2003) reviewed that using natural fibres compared with other reinforcing agents such as
98
glass fibres, talc or carbon fibres for improving the performance of biodegradable polymers
99
include the retention of the biodegradability of the composite, but generally also exhibit lower
cr
us
101
density, superior performance, and lower cost due to the abundance of cellulose fibres. Han and Choi (2010) explained the role of surface areas especially with regards to the
an
100
ip t
96
different mechanical properties of reinforced biocomposites. They showed as well that the
103
surface morphology of cellulose fibres is a significant factor to consider for the composite
104
interfacial adhesion. Other important fibre specifications include fibre strength, length to
105
diameter ratio and surface area.
te
d
M
102
In this work, the physical modifications induced by a dilute-acid steam explosion
107
treatment on Salix miyabeana (willow tree) cellulose fibres were investigated. In particular, the
108
objective of the work was to determine how dilute-acid steam-explosion process modifies the
109
fibre length, whiteness, thermal stability, crystallinity and degree of polymerization of cellulose.
110
These results were correlated with the steam process conditions via the combined severity factor.
111
The resulting equations generated from simple linear regressions allow an estimation of the
112
properties of the cellulose fibres at a known severity factor. In addition, the dimensions of
113
extracted fibres were compared with those of the existing commercial microcrystalline cellulose.
114
The potential use of dilute-acid steam-explosion as a new route for microcrystalline cellulose
115
isolation from biomass residues is discussed.
Ac ce p
106
5 Page 5 of 33
116 117
2) Materials and methods
118
2.1. Material Salix myabeana residues obtained from Le Saule Magique Inc. (Laval, Québec, Canada)
120
were used as feedstock. Biomass was processed using an 800g pilot-scale steam gun, which was
121
extensively described by Berberi et al. (2014); Lavoie et al. (2010) and Lavoie and Beauchet
122
(2012). Sulfuric acid 98% and sodium chlorite (technical grade nominally 80%) were bought
123
from Anachemia Canada Inc. Acetic acid glacial (99.8 %) was purchased from VWR
124
International LLC. All reagents were used as received.
an
us
cr
ip t
119
2.2. Method:
127
2.2.1. Steam explosion treatment:
d
126
M
125
Prior to steam processes, the feedstock was grinded and screened to an average size of
te
128
two (2) cm. The severity factor of the steam treatment was determined using an empirical
130
equation developed by Overend and Chornet (1987), which correlates temperature and residence
131
time:
132 133
Ac ce p
129
Log (R0) = log [t × e (T-Tref)/ 14.75]
1
Where Log (R0) is the severity factor, t is the residence time (min), T is the reaction
134
temperature (˚C), Tref is the reference temperature arbitrarily settled at 100˚C assuming that the
135
hydrolytic effect of water on cellulosic material is negligible below this temperature.
136
By including the effect of the pH on the reaction, equation 1 was further adapted to the
137
experimental conditions for the determination of the combined severity factor (SFT), under an
138
acidic catalysts:
6 Page 6 of 33
139 140 141
SFT= Log (R0)-pH
2
(Chum, et al., 1990)
With Log (R0) obtained from equation 1. Screened woodchips (200 g, dry basis) were initially impregnated in a solution of 3% (wt %) sulphuric acid (ratio solid to liquid 1:10, m/ v) for 10 minutes after which they were
143
dewatered with a filter press. The acid concentration was fixed to 3% (wt. %) following
144
preliminary investigation on the feedstock (not published). The acidic-wet biomass was then
145
cooked in the steam gun at three different temperatures (190, 205 and 220 ˚C), which were
146
maintained for a residence time arbitrarily fixed at 3, 5 and 10 minutes respectively. Following
147
steam treatment, the resulting lignocellulosic material was washed using hot water and press
148
filtered until neutral pH was attained. A laboratory filter press was used for dewatering the wet
149
biomass after treatment. By applying a controlled pressure of 0.7 MPa, the water content was
150
reduced to the minimum in order to facilitate the drying process. Fibres were then dried at room
151
temperature in a convection oven for 72 hours. The experimental uncertainty on the process
152
temperature was around 2oC. For the concentration measurement and residence time, the relative
153
uncertainty is below 0.5%. Based on the above, the overall uncertainty on the calculated SFT
154
was less than 1%.
156 157
cr
us
an
M
d
te
Ac ce p
155
ip t
142
2.2.2. Fibre delignification and bleaching Following the steam explosion treatment, Salix myabeana cellulose fibres were bleached
158
using a slight modification of the chlorine dioxide method employed by Kumar, Mago, Balan
159
and Wyman (2009). Bleaching was performed during six hours at 75˚C, using distilled water
160
(ratio solid to liquid 1:20 wt/ v) sodium chlorite (0.6g/ g of dry matter) and acetic acid (0.2 ml/ g
161
of dry matter). The reaction was done in three stages of two hours in which an equal amount of
7 Page 7 of 33
162
sodium chlorite and acetic acid were added to the medium. Fibres were then repeatedly washed
163
with hot and distilled water and dried at room temperature in an air-circulating oven for 72 h and
164
then kept in a sealed plastic bag until analysis.
2.2.3. Fibres characterization.
167
2.2.3.1. Fibre Quality Analysis
Length of the cellulose fibres as well as diameter and length to diameter aspect ratios
us
168
cr
166
ip t
165
were determined using a Fibre Quality Analyzer (FQA) (Optest Hires model, Hawkesbury,
170
Canada).
171
M
173
2.2.3.2. X-Ray diffraction analysis (XRD)
The X-ray diffraction patterns were obtained using a Philips X’Pert diffractometer using
d
172
an
169
copper radiation Kα (λ = 1.5418 Ǻ), a voltage of 40 kV and an operation current of 50 mA. All
175
scans were performed from 2θ = 5° to 2θ = 40°. MDI Jade 10 software packages was used for
176
pattern processing as well as for calculation of the degree of crystallinity (DC). The degree of
177
crystallinity was calculated using the height ratios between the intensity of the crystalline peak
178
(I002 - IAm) and total intensity (I002) after subtraction of the background signal measured without
179
cellulose as also previously reported by Park, Baker, Himmel, Parilla and Johnson (2010).
180
The I002 intensity of the crystalline peak was taken at 2θ= 27˚ and IAm is the intensity of the
181
amorphous phase at 2θ = 18˚.
Ac ce p
te
174
182 183
2.2.3.3. Scanning Electron Microscopy (SEM)
8 Page 8 of 33
The microstructure of the fibres was observed using a Hitachi scanning electron
185
microscope model S-3000N (Hitachi High-Technologies Corporation, Japan) equipped with a
186
tungsten hairpin filament, under different magnifications. Prior to analysis, the surface of the
187
cellulose fibres was sputter-coated with gold/ platinum using a vacuum plasma spray and argon
188
as inert gas. The images were taken at an accelerating voltage of 10 kV.
cr
189
ip t
184
2.2.3.4. Thermogravimetric analysis (TGA)
191
Thermal properties of the different samples were measured using a SETSYS TG-DTA/DSC 24
192
Thermal Analyser with a mass spectrometer (Setaram Instrumentation, Canada). A known
193
weight of each sample that was previously dried at 60˚C for 24h was taken into silica crucible for
194
analysis. The mass loss analyses were performed using argon as a purge gas, with a furnace flow
195
of 20 ml/min and a temperature profile of 10˚C/ min. The normalized mass loss was plotted as
196
differential thermogravimetric curves from 30 to 600˚ C.
197
199
2.2.3.5. Optical properties
Ac ce p
198
te
d
M
an
us
190
The steam-treated bleached fibres were submitted to Technibrite Micro TB-1C
200
reflectometer for optical analysis. Difference in CIE L*, a*, b* color coordinates and whiteness
201
index, expressed in percentage were measured according to the TAPPI T 452 om-98 method.
202 203 204
2.2.3.6. Degree of polymerization Degree of polymerization was obtained following a method reported by Ibrahima,
205
Agblevor and El-Zawawy (2010). Bleached steam-exploded fibres (0.06 g) were wetted in 7.5
206
mL distilled water and stirred during 2 minutes after what 22.5 mL of 1M cupri-ethylenediamine
9 Page 9 of 33
(CED) (copper concentration 0.98-1.02 M, Anachemia Canada Co) solution was added (H2O:
208
CED ratio of 1: 3 (v: v)). The solution, protected from light and oxygen, was agitated during 2h
209
until complete dissolution of the cellulose was reached. A cannon-fenske viscometer N# 75 was
210
used to measure the flow time of the CED-cellulose solution that was used to calculate the
211
intrinsic viscosity [η] according to the equation of Solomon-Gatesman:
cr
212
ip t
207
[η] = (2× (ηsp – Ln (ηrel)) 0.5/ C
With: ηsp = Specific viscosity and ηrel = relative viscosity and C = solution concentration.
214
The degree of polymerization DP was calculated using the following equation proposed by
215
Evans and Wallis (1989):
an
DP0.9= [η] ×1.65
M
216
us
213
3
Finally, the average number of chain scissions (CS) per cellulose chain unit during the time
218
course of degradation was calculated using the equation:
219
CS = (DP0/DP)-1
te
d
217
4
5
Where DP0 is the degree of polymerization before scission. In the present work, the DP obtained
221
at lower severity factor (SFT = 2.88) was taken as the reference DP0 for CS calculation.
222
Ac ce p
220
223
3. Results and discussion
224
3.1 Crystallinity and Optical properties.
225 226
Table 1 presents a summary of the steam explosion process conditions and severity factor
227
as well as key structural features of the treated cellulose: optical properties, and degree of
228
crystallinity. The reaction temperature was varied between 190 and 220o C while the residence
229
time was varied from 3 to 10 min. The sulfuric acid content was kept constant at 3 wt%. These 10 Page 10 of 33
conditions allowed calculating severity factor ranging from 2.88 to 4.29 for the different
231
samples. The fibres, recovered after the steam explosion treatment, were highly crystalline with
232
degrees of crystallinity superior to 85% (determined by X-ray analysis). It is noteworthy that no
233
detectable amount of lignin or hemicellulose were present in the samples. The crystallinity was
234
relatively constant although a slight decrease was observed for higher severity factors. X-ray
235
diffraction patterns (not shown) showed that type I crystalline cellulose was present in all treated
236
samples showing characteristics peaks at 2θ of 14.9˚, 16.3˚, 22.5˚ and 34.6˚ which were assigned
237
to the 101,
us
cr
ip t
230
, 002 and 040 crystallographic planes respectively.
239
an
238
241
#
d
A B
245
C
246
D
247 248
E F G
190
te
244
Temperature Retention H2SO4 Severity (˚C) time (Wt. %)a factor b (min.)
3
Ac ce p
242 243
M
240
Table 1: Steam explosion conditions, severity factor, degree of crystallinity, and optical properties of bleached steam exploded Salix myabeana cellulose fibres Degree of crystallinity (%)
Optical properties L*
a*
b*
3
2.88
88.6 ± 0.7
92.7
-0.51
7.82
-1.11
7.75
190
5
3
3.1
90.6 ± 0.8
93.25
190
10
3
3.4
85.2 ± 0.7
90.05 -1.02
205
3
3
3.32
87.1 ± 0.6
91.94
-0.93
8.85
205
5
3
3.55
87.4 ± 0.7
86.75 -0.02
13.31
205
10
3
3.85
85.7 ± 0.6
69.86
2.18
22.82
220
3
3
3.77
88.2 ± 0.6
80.11
1.17
18.99
10.56
249
H
220
5
3
3.99
85.5 ± 0.7
73.23
1.8
17.95
250
I
220
10
3
4.29
85.3 ± 0.6
42.62
5.78
20.61
251 252
a = H2SO4 concentration corresponding to a pH of 0.24 b = Combined severity factor calculated according to equation (1)
11 Page 11 of 33
253
Similar observations were made by Wang, Jiang, Xu and Sun (2009), when working with
255
Lespedeza crytobotrya specie steam exploded for a fixed retention time of 4 minutes.
256
Nonetheless, other reports aiming at the cristallinity of cellulose after steam process operated on
257
longer periods mentioned that the cristallinity at one point seemed to be increasing
258
proportionally to the severity factor (Ibrahim et al., 2010) (Wang and Chen, 2013). Both
259
observations can be logically correlated to the chemical hydrolysis occurring during steam
260
processes. The severity of the latter being amplified by the dilute acid catalyst led to a more
261
crystalline material with the attack and the dissolution of part of the amorphous regions.
262
Kang, Bansal and Realff (2013) on the other hand, when using SO2 as catalyst observed the same
263
reduction of the fibre crystallinity when more severe treatment was employed. From this, they
264
managed to determine a severity threshold above which a decrease of the crystallinity was
265
perceived during steam treatments. This threshold although not localised in our work could
266
possibly explain the crystallinity reduction after a certain severity value was reached. In that
267
case, according to the reduction of the cristallinity noticed, it seem that the fibres degradation at
268
higher treatment severity was not limited to the cellulose wall, but also could in a lesser extent
269
affect the internal fibre crystalline organization.
cr
us
an
M
d
te
Ac ce p
270
ip t
254
On the other hand, when considering the percent cristallinity discussed in our work,
271
higher crystalline values were noticed compared to those of microcrystalline cellulose
272
investigated by other authors. As an example, Leppänen et al. (2009) when comparing micro
273
cellulose from different sources, revealed a crystallinity of 59% for commercial Avicel and 65%
274
for microcrystalline cellulose obtained from cotton linter.
12 Page 12 of 33
275
The optical properties were characterized using the CIE-LAB color coordinates data where L* (lightness), represents the reflectance in the mid-range of visible spectrum, a*
277
measures the redness (negative value) versus greenness (positive values) of the material and b*
278
corresponds to the measure of the color space going from yellowness (positive value) to blueness
279
(negative value). The variations of those optical parameters for each of the willow cellulose
280
samples are as well provided in Table 1. The latter shows that increasing the severity of the
281
steam treatment leads to a decrease of fibre lightness while the yellowness increased
282
accompanied by a color change from red to green. Figure 1 represents the changes of the
283
bleached cellulose fibres as a function of the temperature and residence time.
Ac ce p
te
d
M
an
us
cr
ip t
276
284 285
Figure 1
13 Page 13 of 33
286
Temperature, in the investigated range, had a stronger effect of fibre discoloration than residence time. As an example of the correlation between appearance of the pulp and severity
288
factor, samples C and D, which were treated at similar severity factor of respectively 3.4 and
289
3.32 shows similar discoloration. When the severity factor reached 3.55 and above, irreversible
290
visible discoloration took place, light at the beginning, but increasing proportionally with the
291
severity factor.
cr
A simple estimation of the optical properties of the material post-bleaching can as well be
us
292
ip t
287
provided by the percent whiteness since it is measured across the entire visible spectrum. Figure
294
2 presents the correlation between the whiteness of bleached fibres and the severity factor (fig.
295
2a) and the combined influence of the couple reaction temperature/ residence time on cellulose
296
whiteness (fig. 2b). The whiteness (W) decreased linearly with the increase of the severity
297
factor, a relation that can be described using the following equation:
d
M
an
293
W a SFT b
299
With: W = percent whiteness, a = - 40.487 ± 1.323 and b = 197.3 ± 5.2
( 6)
Ac ce p
te
298
300 301
Figure 2
14 Page 14 of 33
Hemicellulose present in the initial biomass could be responsible of the discoloration
303
observed in the fibres. However, it should be mostly hydrolyzed during the dilute-acid steam
304
explosion (Tucker, Kim, Newman and Nguyen, 2003). It is thus improbable that hemicelluloses
305
are to be taken accountable for the observed color changes. As well, the bleaching process used
306
in this work (chlorine dioxide) was proven to have no evident effect on cellulose fibres structure,
307
attacking only the aromatic rings of the remaining lignin on the lignocellulose (Lemeune, Jamee,
308
Chang and Kadla, 2004). Therefore, the observed browning of the fibres could be related to a
309
partial degradation of cellulose via successive degradation of carbohydrate structures to furans
310
(or aromatic rings) that would ultimately lead to the production of char at high severity.
311
Coloration of the fibres was also observed by Ibrahim et al. (2010) who related this event to the
312
cell wall components, extractives and sugar degradation after steam pre-treatment at higher
313
severity factor.
cr
us
an
M
d
In this work, all steam-exploded (SE) lignocellulosic materials were bleached under the
te
314
ip t
302
same conditions. Chlorine dioxide bleaching is usually done in three stages using the same
316
amount of reagent for each stage under identical time and temperature conditions. Chlorination
317
of cellulose fibres was proven to produce white cellulose, free or with a relatively low lignin
318
content (Patrick and Martin, 1997) depending on the specie and the initial lignin content of the
319
material. The fact that white samples were obtained under the lower severity factors implies that
320
the browning cannot be related to a remaining lignin content. It seems that utilization of
321
sulphuric acid as catalyst, in addition to its fibres destructuration potential at high temperature
322
also induces a variety of secondary reactions. These reactions could involve formation of new
323
chemicals (such as furans) by the transformation carbohydrate (as cellulose or cellulose side-
324
chain). The darkening of steam-exploded fibres has been previously attributed to the production
Ac ce p
315
15 Page 15 of 33
of acetic, formic and other pyroligneous acids as well as the reactions of these organic acids with
326
lignocellulosic material (Morgan and Stewart, 1941). In addition, a possible polymerization of
327
furans and of other sugar degradation products is also an option (Wang et al., 2009). A simple
328
experiment with carbohydrate in an acid catalyzed mixture provides a pitch-black mixture after
329
reacting at 200 o C for a few minutes. Another possibility, although these combined reactions are
330
hard to monitor, could be the generation of “pseudo-lignin”, defined as aromatic compounds
331
yielding a positive klason lignin value not derived from native lignin during dilute-acid steam-
332
explosion pre-treatment (Hu, Jung, and Ragauskas, 2012). These compounds could be formed
333
following combination of carbohydrate macromolecules and by-products (such as furans) formed
334
during the acid pre-treatment of biomass at high temperature (Sannigrahi, Kim, Jung and
335
Ragauskas, 2011). Pseudo-lignin could be resistant to bleaching procedures and its formation
336
may become more likely at high severity.
d Acid-catalyzed steam processes leads to defibrillation of the lignocellulosic biomass
Ac ce p
339
3.2 Fibre dimensions
te
337 338
M
an
us
cr
ip t
325
340
matrix while the amorphous and peripheral fractions of cellulose are most likely hydrolyzed.
341
Figure 3a shows the evolution of average single fibres length (L), diameter (D) and length to
342
diameter (L/D) aspect ratio as a function of the severity factor (SFT). The fibre lengths were
343
drastically reduced following the steam treatments in a linear fashion with regards to the
344
increasing severity of the treatments. This tendency has been reported by various authors
345
including Datar et al. (2007) when working on non-catalyzed as well as dilute acid steam
346
exploded corn stover performed in the temperature range of 190 to 200 C˚ for a retention time of
347
3 to 5 minutes. Under high temperature and for long residence time (severe treatments), the
348
chemical hydrolysis, condensation, mechanical defiberization and structural disorganisation are 16 Page 16 of 33
amplified, thus leading to the production of shorter cellulose fibres. At lower SFT value of 2.88,
350
an average length of 0.23 mm was recorded, while shorter fibres with an average length of 0.14
351
mm were produced at higher severity. The liner relationship between fiber length and SFT can be
352
described by the following equation:
353
L c SFT d
ip t
349
7
With: L = fiber length, c = - 0.057 ± 0.013 and d = 0.387 ± 0.045
355
Figure 3b presents the evolution of the fine (defined here as fibres with length ranging from 0.07
356
to 0.2 mm) as a function of the severity factor.
Ac ce p
te
d
M
an
us
cr
354
357 358
Figure 3
359 17 Page 17 of 33
360
The fines (F%) accounted for 52% at a severity of SFT=2.8 and their fraction increased linearly
361
with the increasing SFT to reach value of 82% at the higher severity (SFT= 4.3). The correlation
362
between the two parameters followed the equation: F % e SFT g
364
With: F% = percentage of fine, e = 21.39 ± 2.31 and g = 8.51 ± 2.59
365
By contrast, the average fibre diameters value (D) was more stable with only a slight
8
cr
ip t
363
increase as function of the SFT leading to a decreasing L/D ratio with regards to severity. This
367
trend could be explained by agglomeration due to fibre degradation at higher temperature/
368
retention time as mentioned earlier. The curvature after the separation of fibre bundles as a result
369
of the steam processes could also explain the augmentation of the average fibre diameter
370
detected using FQA.
M
an
us
366
The products obtained after the dilute acid steam explosion could be related to
372
microcrystalline cellulose. In fact, Leppänen et al. (2009) reported that hydrolysis of cellulose
373
with dilute-acid cannot isolate nanosize particles of cellulose even after mechanical
374
disintegration. Strong acid must be used for this purpose. When dilute acid hydrolysis is
375
considered, the strong inter-crystalline contacts remain intact and the micron-size crystallites
376
aggregates leading to the formation of microcrystalline cellulose. The average fibre length for
377
commercial Avicel was found to be in the range of 0.02 mm to 0.25 mm, while it ranged from
378
0.02 to 0.125 mm for microcrystalline cellulose obtained from cotton linter (Leppänen et al.,
379
2009). This must be compared to average particle length ranging from 150 to 200 nm and the
380
average particle width ranging from 10 to 20 nm for nanocellulose (Ioelovich, 2012). On the
381
other hand, Ohwoavworhua and Adelakun (2010) found an average fibre diameter of 95 m for
382
microcrystalline cellulose produced by hydrochloric acid hydrolysis of α-cellulose extracted
383
from Sorghum caudatum. In term of fiber length, the average length of the steam exploded
Ac ce p
te
d
371
18 Page 18 of 33
cellulose fibres obtained in this work ranged from 0.15 to 0.23 mm, while their diameter varied
385
between 15.8 to 32.8 m. To some extent, these facts could corroborate the possibility to extract
386
microcrystalline cellulose directly by using dilute-acid steam explosion, which could be a new
387
process route to explore.
ip t
384
388
cr
390
3.5. Microstructural characterization
Figure 4 presents pictures of bleached steam-exploded cellulose fibres obtained with a
us
389
SEM system. The micrographs reveals separated elementary cellulose fibres with some irregular
392
filament bundles persisting in some cases (fig 4 A; B; C; D).
393
an
391
This observation is of interest for an application as polymer reinforcement since isolated fibres are usually targeted for such an application (Altman et al., 2003). In comparison to the
395
lower severity factor (SFT = 2.88), fibres obtained with higher severity (up to SFT= 3.77)
396
seemed to “unfold”. This observation, in addition to the fibre agglomeration observed at higher
397
SFT values (Fig 4 E; G; H; I), could explain the difference in fibre diameter generated from
398
steam processes as reported in Figure 3. The severity of the applied treatment impacted clearly
399
on the final cellulose microstructure with fibre length reduction and loss of fibrous structure
400
occurring progressively with increasing SFT.
d
te
Ac ce p
401
M
394
These structural changes, happening at higher severity factor, could be elucidated by the
402
extension of the hydrolytic chemical phases, subsequent to the lengthening of the residence time.
403
The pressure rise inside the reactor generated by the increasing reaction temperature as well led
404
to more intense depressurization and thus defibrillation. This phenomenon generates a
405
disordering of the biomass cell wall, as previously observed for wheat straw by Anupama and
406
Singh (2011).
19 Page 19 of 33
Ac ce p
te
d
M
an
us
cr
ip t
407
408 409
Figure 4
20 Page 20 of 33
On the other hand, two spherical globules with diameters in the 10 m range in the upper
410 411
left part of Figure 4I may be pseudo-lignin particles as observed by Hu, Jung and Ragauskas
412
(2012) and Sannigrahi et al. (2011) when severe treatments were applied.
414
ip t
413
3.6. Derivative thermo gravimetric analysis
Thermo gravimetric analyses were performed in order to determine the thermal
cr
415
stability of bleached cellulose fibres treated at different severities. Figure 5 presents the relative
417
weight loss and weight loss derivative (WLD) of the treated fibres. The weight loss occurred in
418
two stages for all the tested specimens. The main degradation phase occurred around 300o C
419
while a second minor phase occurs at temperatures above 400o C. The onset degradation
420
temperature (Tonset) as well as the maximum temperature of degradation (TMax) did not
421
significantly change with treatment severity.
422
The weight loss derivative curves (WLD) for all bleached cellulose fibres revealed that the
423
maximum degradation rate occurred between 298 ˚C and 308˚C. This main peak, with maximum
424
WLD appearing between 298˚ C and 308 ˚C depending on the applied conditions, was
425
accompanied with two secondary peaks between 70˚ and 100˚C and between 380˚ C and 550 ˚C
426
attributed respectively to H2O evaporation and remaining impurities (Morán, Alvarez, Cyras and
427
Vázquez, 2008). The last cited peak evidenced that these impurities degraded around
428
temperature range of 400˚C to 550˚C that may approximately correspond to the degradation
429
temperature of isolated lignin discussed by Ibrahim et al. (2010) for cotton stalk and banana
430
plant.
431 432
Ac ce p
te
d
M
an
us
416
However, due to the strong bleaching process applied and the absence of lignin, this peak could be assigned to the pseudo-lignin molecules that present a similar structure with lignin.
21 Page 21 of 33
433
The total carbonaceous residues (char) found at the end of the heating process was low,
434
between 2 and 4.2% with the highest value found for the highest severity factor. The same trend
435
was observed by Jacquet et al. (2011), for steam exploded cellulose fibres.
Ac ce p
te
d
M
an
us
cr
ip t
436
437 438 22 Page 22 of 33
439
Figure 5
440 441
3.7. Degree of polymerization The average chain length of the cellulose macromolecules in the bleached cellulose fibres
443
can be expressed as a degree of polymerization (DP) of the cellulose chains. The variation of the
444
degree of polymerization as a function of the severity factor (SFT) is presented in Figure 6a.
445
Figure 6b presents variation of DP as a function of the average fibre length and Figure 6c shows
446
the evolution of DP as a function of the actual treatment conditions (temperature, residence
447
time).
an
us
cr
ip t
442
Ac ce p
te
d
M
448
23 Page 23 of 33
ip t cr us an M d te 450 451
Ac ce p
449
Figure 6
As for the fiber length presented earlier, the degree of polymerization decreased as a
452
function of the severity factor. The linear relation between these two parameters can be described
453
by the equation:
454
DP h SFT i
9
455
With: DP = degree of polymerization. h = - 1326.3 ± 31.8 and i = 5645 ± 129
456
It is noteworthy that macromolecular chain length expressed as degree of polymerisation
457
and the macroscopic fiber length are closely related through a linear relationship. 24 Page 24 of 33
458
High value of DP of 1830 was observed for fibres treated at the lowest SFT (2.88). A
459
significant reduction of the degree of polymerization down to values around 100 were obtained
460
at higher SFT (4.29). This corresponds to an average chain scission value of 15.7. In Figure 6b, on the other hand, we can observe that the DP was reduced with increasing
ip t
461
temperature and residence time taken individually. Similar variations were also revealed by
463
Foston and Ragauskas (2010) for populus and switchgrass during dilute acid pre-treatment, with
464
the increase of the residence time.
us
cr
462
Terinte, Ibbett and Schuste (2011) observed that the leveling off degree of polymerization
466
(LODP) of microcrystalline cellulose ranges from 200 to 300. Actually, in this work, we noticed
467
that the degree of polymerization of cellulose fibres dropped to around 205 when a severity of
468
3.99 was applied, and reached value as low as 109 at higher severity factor of 4.29. For instance,
469
the average degree of polymerization calculated by Leppänen et al. (2009) for commercial
470
Avicel was found to be around 150 and was around 165 for microcrystalline cellulose obtained
471
from cotton linter. This statement could be in agreement with the above discussed possibility that
472
microcrystalline cellulose can be extracted directly by dilute-acid steam explosion.
M
d
te
Ac ce p
473
an
465
474
4. Conclusions
475
The increased severity applied to salix myabeana during the Steam explosion process affects the
476
length, whiteness, length to diameter aspect ratio as well as the degree of polymerization of the
477
bleached cellulose fibres. The thermal stability of the fibres on the other hand was not a strong
478
function of the reaction parameters. Finally, the obtained fibre structures in term of fibre length,
479
diameter and degree of polymerization suggest that the dilute acid steam explosion process could
480
be used as a new route for direct extraction of microcrystalline cellulose from biomass residues. 25 Page 25 of 33
481 482
Acknowledgement: The authors are grateful to BiofuelNet Canada and to the Industrial Research Chair
484
on Cellulosic Ethanol and Biocommodities of the Université de Sherbrooke for their support of
485
the project. The Chair is co-funded by CRB Innovations, Enerkem and Ethanol Greenfield
486
Québec Inc., as well as the Ministère de l'énergie et des ressources naturelles du Québec. Special
487
thanks to the CRB team for using part of their facilities and expertise in steam treatments.
us
cr
ip t
483
.
488
an
489 490
M
491
d
492
te
493 494
Ac ce p
495 496 497
References:
498
Altman, G. H., Diaz. F., Jakuba, C., Calabro, T., Horan, R. L., Chen, J., Lu H., Richmond,
499 500
J., and Kaplan, D. L. (2003). Silk-based biomaterials. Biomaterials. 24, 401–416. Anupama, K., and Singh, M. (2011). Isolation of cellulose nanofibrils from wheat straw using
501
steam explosion coupled with high shear homogenization. Carbohydrate Research. 346, 76-
502
85.
503
Berberi, V., Fuente-Hernandez, A., Corcos, P-O, Lalonde, O., Almirall-Diaz, I. and Lavoie, J-M. 26 Page 26 of 33
504
(2014) Cellulosic Ethanol from Reed Canary Grass using Bench and Pilot Scale Steam
505
Treatments & Non-enzymatic Hydrolysis. Article submitted to Industrial Crops and
506
products. Brownell, H. H., Yu E. K. C and Saddler J. N. (1986). Steam-explosion pre-treatment of wood:
ip t
507
Effect of chip size, acid, moisture content and pressure drop. Biotechnology and
509
Bioengineering, 28 (6), 792-801.
Chum, H. L., Johnson, D. K., Black, S. K., and Overend, R. P. (1990). “Pre-treatment-
us
510
cr
508
catalyst effects and the combined severity parameter,” Applied Biochemistry and
512
Biotechnology. 24/25, 1-14.
513
an
511
Datar, R., Huang, J., Maness, P. C., Mohagheghi, A., Czernik, S., and Chornet, E. (2007). Hydrogen production from the fermentation of corn stover biomass pretreated
515
with a steam explosion process. International journal of hydrogen energy. 32, 932-939.
518
d
te
517
Evans, R., and Wallis, A. F. A. (1989). Cellulose Molecular Weights Determined by Viscometry. Journal of Applied Polymer Science, 37, 2331-2340. Foston, M., and Ragauskas, A. J. (2010). Changes in lignocellulosic supramolecular and
Ac ce p
516
M
514
519
ultrastructure during dilute acid pre-treatment of Populus and switchgrass. Biomass and
520
Bioenergy, 34, 1885-1895.
521 522
Garrot, G., Domínguez, H., and Parajó J. C. (1999). Hydrothermal processing of lignocellulosic materials. Holz als Roh- und Werkstoff, 57, 191- 202.
523
Han, S. O., and Choi, H. Y. (2010). "Morphology and surface properties of natural fiber treated
524
with electron beam," Microscopy: Science, Technology, Applications and Education 3(1),
525
1880-1887.
526
Hu, F., Jung. S., and Ragauskas, A. (2012). Pseudo lignin formation and its impact on enzymatic
27 Page 27 of 33
527 528
hydrolysis. Bioresource technology. 117, 7-12. Ibrahim, M. M., Agblevor, F. A., and El-Zawawy, W. K. (2010). Isolation and characterization of cellulose and lignin from steam-exploded biomass. BioResources. 5 (1), 397-418.
530
Ioelovich, M. (2012). Optimal conditions for isolation of nanocrystalline cellulose particles.
532
Nanoscience and Nanotechnology, 2 (2), 9-13.
Jacquet, N., Quiévy, N., Vanderghem, C., Janas, S., Blecker, C., Wathelet, B., Devaux,
cr
531
ip t
529
J., and Paquot, M. (2011). Influence of steam explosion on the thermal stability of cellulose
534
fibres. Polymer Degradation and stability. 96, 1582-1588
Kang, Y. Bansal, P. and Realff, M. J. (2013). SO2 catalyzed steam explosion: the effect
an
535
us
533
of different severity on digestibility, accessibility and crystallinity of lignocellulosic
537
biomass. Biotechnology progress. 29 (4), 909-916.
Kumar, R., Mago, G., Balan, V., and Wyman, C. E. (2009). Physical and chemical
d
538
M
536
characterization of corn stover and polar solid resulting from leading pre-treatment
540
technologies. 100, 3948-3962.
Lavoie, J. M., and Beauchet, R. (2012). Biorefinery of Cannabis sativa using one-and two-step
Ac ce p
541
te
539
542
steam treatments for the production of high quality fibres. Industrial Crops and Products,
543
37(1), 275-283.
544
Lavoie, J. M., Capek-Menard, E., Gauvin, H., and Chornet, E. (2010). Quality pulp from mixed
545
softwoods as an added value coproduct of a biorefinery. Industrial & Engineering Chemistry
546
Research, 49(5), 2503-2509.
547
Lee, R. A., and Lavoie, J. M. (2013). From first-to third-generation biofuels: Challenges of
548
producing a commodity from a biomass of increasing complexity. Animal Frontiers,
549
3(2), 6-11.
28 Page 28 of 33
550
Lemeune, S., Jamee, l H., Chang, H.-M., and Kadla, J. F. (2004). Effects of Ozone and Chlorine
551
Dioxide on the Chemical Properties of Cellulose Fibers. Journal of applied polymer sciences.
552
93, 1219-1223. Leppänen, K., Andersson, S., Torkkeli, M., Knaapila, M., Kotelnikova, N., and Serimaa, R.
554
(2009). Structure of cellulose and microcrystalline cellulose from various wood species,
555
cotton and flax studied by X-ray scattering. Cellulose, 16, 999– 1015.
561 562 563 564
cr
us
an
560
United States patent office. US1596159 A
Morán, J. I. Alvarez, V. A., Cyras, V. P., and Vázquez, A. (2008). Extraction of cellulose and
M
559
Mason, H. W. (1926). Low-temperature explosion process of disintegrating wood and the like.
preparation of nanocellulose from sisal fibers. Cellulose. 15, 149-159.
d
558
pretreated by alumina-doped MgO in AMIMCL. Carbohydrate Polymers, 113, 1–8.
Morgan, W. H., and Stewart, O. B. (1941). Process of making light-colored ligno-cellulose fiber.
te
557
Liu, L., Ju, M. T., Li, W. Z., and Jiang, Y. (2014). Cellulose extraction from Zoysia japonica
United States Patent Office. US 2234188 A. Mosier, N., Wyman, C., Dale, C., Elander, R., Lee. Y. Y., Holtzapple, M., and Ladish, M.
Ac ce p
556
ip t
553
565
(2005). Features of promising technologies for pre-treatment of lignocellulosic biomass.
566
BioResource technology, 96, 673-686.
567
Ovwoavworhua, F. O., and Adelakun, T. A. (2010). Non-wood fibre production of
568
microcrystalline cellulose from Sorghum caudatum: characterisation and tableting properties.
569
Indian Journal of Pharmaceutical Sciences. 72 (3), 295-301.
570 571 572
Overend, R.P., and Chornet, E. (1987). Fractionation of lignocellulosics by steam-aqueous pre-treatments. Philosophical Transaction of the Royal Society A. 321 (1561), 523-536. Park, S., Baker, J. O., Himmel, M. E., Parilla, P. A. and Johnson, D. K. (2010). Cellulose
29 Page 29 of 33
573
crystallinity index: measurement techniques and their impact on interpreting cellulose
574
performance. Biotechnology for biofuel. 3, 1-10.
576 577
Patrick, T., and Martin, S. (1997). Chlorine dioxide delignification kinetics and EOP extraction of softwood Kraft pulp. The Canadian Journal of Chemical Engineering. 75, 23-30.
ip t
575
Petinakis, E., Yu, L., Simon, G., and Dean, K. (2003). Natural Fibre Bio-Composites
Incorporating Poly (Lactic Acid). In: Alberto M (Ed) Fiber reinforced polymer: The
579
technology applied for concrete repairs. ISBN 978-953-51-0938-9, 41-59.
580
Sannigrahi, P., Kim, D. H., Jung, S., and Ragauskas, A. (2011).Pseudo-lignin and
582
us
Pretreatment chemistry Energy Environ. Sci.4, 1306-1310.
an
581
cr
578
Terinte, N., Ibbett, R., and Schuste, K. C. (2011). Overview on native cellulose and microcrystalline cellulose I structure studied by X-Ray diffraction (WAXD): comparison
584
between measurement techniques. Lenzinger Berichte, 89, 118-131.
d
Tucker, M. P., Kim, K. H., Newman, M. M., and Nguyen, Q. A. (2003). Effects of Temperature
te
585
M
583
and Moisture on Dilute-Acid Steam Explosion Pre-treatment of Corn Stover and Cellulase
587
Enzyme Digestibility. Applied Biochemistry and Biotechnology. 105–108, 165- 177.
Ac ce p
586
588
Wang, K., Jiang, J.X., Xu, F., and Sun, R.C. (2009). Influence of steaming pressure on the
589
steam explosion pre-treatment of stalks (lespedeza cryotobotra) Part1: characteristic of
590
degraded cellulose. Polymer Degradation and Stability. 94, 1379-1388
591
Wang, N., and, and Chen, H. Z. (2013). Manufacture of dissolving pulps from cornstalk by
592
novel method coupling steam explosion and mechanical carding fractionation. BioResource
593
Technology. 139, 59-65.
594 595
Wang, X. J., Li, H. Q., Cao, Y, and Tang, Q. (2011). Cellulose extraction from wood chip in an ionic liquid 1-allyl-3-methylimidazolium chloride (AmimCl). Bioresource Technology, 102,
30 Page 30 of 33
596
7959–7965.
597 598
ip t
599 600
cr
601
us
602 603
an
604 605
M
606
d
607
te
608
610 611 612 613
Ac ce p
609
Figures captions:
614
Figure 1: Aspects of bleached dilute acid steam-exploded cellulose fibres as a function of the
615
reaction temperature and retention time. (Conditions described in Table 1).
616
Figure 2: (a) Cellulose fibre whiteness (%) as a function of the severity factor and (b) evolution
617
of the whiteness with the retention time.
31 Page 31 of 33
Figure 3: (a) Fibre length, diameter, and Length /Diameter ratio as a function of the severity
619
factor and (b) fine fraction (fibres with length between 0.07 and 0.2 mm) as a function of the
620
severity factor.
621
Figure 4: SEM micrographs of bleached Salix myabeana fibres steam exploded at different
622
severity factors as described in Table 1.
623
Figure 5: Weight loss and weight loss derivative (WLD) of bleached Salix myabean fibres at
624
different steam explosion conditions as described in Table 1.
625
Figure 6: a) Degree of polymerization as a function of the severity factor (■), b) Degree of
626
polymerization as a function of the cellulose fibre length (●) and c) Degree of polymerization of
627
cellulose fibres as a function of the reaction temperature and residence time.
cr
us
an
M Ac ce p
te
d
628 629
ip t
618
32 Page 32 of 33
629
Research Highlights
630
● Extraction of cellulose fibres by dilute-acid steam explosion have been investigated.
632
● Increasing severity factor will signally affect the structure of cellulose fibres.
633
● Correlations show fibre structures changing proportionally with severities factors
634
● The steam process could be applicable to the extraction of microcrystalline cellulose
ip t
631
cr
635
us
636 637 638
an
639 640
M
641
645 646 647 648 649 650 651 652
te
644
Ac ce p
643
d
642
653 654 655 656 657 33 Page 33 of 33
S0144-8617(14)01136-9 http://dx.doi.org/doi:10.1016/j.carbpol.2014.11.031 CARP 9459
To appear in: Received date: Revised date: Accepted date:
22-9-2014 11-11-2014 16-11-2014
Please cite this article as: Diop, C. I. K., Lavoie, J.-M., and Huneault, M. A.,Structural changes of Salix miyabeana cellulose fibres during dilute-acid steam explosion: impact of reaction temperature and retention time, Carbohydrate Polymers (2014), http://dx.doi.org/10.1016/j.carbpol.2014.11.031 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Structural changes of Salix miyabeana cellulose fibres during dilute-acid steam explosion:
2
impact of reaction temperature and retention time.
3
ip t
4
Chérif Ibrahima Khalil Diop, Jean-Michel Lavoie, Michel A. Huneault*
6
Department of Chemical and Biotechnological Engineering, Université de Sherbrooke, 2500
7
Boulevard de l’Université, J1K 2R1, Sherbrooke, Québec, Canada
us
cr
5
8
an
9 10
13
Corresponding author: Professor Michel A. Huneault
14
Department of chemical and biotechnological engineering, Université de Sherbrooke, 2500
15
Boulevard de l’Université, J1K 2R1, Sherbrooke, Québec, Canada
16
E-mail: [email protected]
17
Telephone: +1 819 821-8000 (Ext. 65580)
19
te
Ac ce p
18
d
12
M
11
20 21 22 23
1 Page 1 of 33
Abstract:
25
Dilute-acid steam explosion of Salix myabeana have been carried out to understand the effect of
26
processing conditions, expressed through a severity factors (SFT), on the changes in cellulose
27
fibre structures in a perspective of using these in polymer composites. This thermo-chemico-
28
mechanical extraction lead to the isolation of cellulose fibres as observed by SEM images. Fibre
29
length as well as length to diameter aspect ratios decreased with the severity of the treatment.
30
Likewise, fibre whiteness diminished with an increasing severity factor, which could be a
31
tangible effect of physical degradation. Variations in cristallinity seemed to be dependent upon
32
the reaction temperature, generally decreasing with regards to retention time. Above a severity
33
threshold, a structural disorganization was observed. Overall, dilute-acid steam explosion was
34
shown to be a valuable cellulose extraction process that can provide a variety of fibre structures.
M
an
us
cr
ip t
24
d
35
Key words: Salix miyabeana cellulose fibres; Dilute-acid steam explosion; Severity factor;
37
Structural changes; Physical characterization.
39 40 41 42
Ac ce p
38
te
36
43 44 45 46
2 Page 2 of 33
47
1. Introduction:
48 49
Cellulose fibres are generally produced from lignocellulosic feedstock (wood, straw, grass, etc.) and they can be extracted and/or purified depending on the final targeted application.
51
Beside its large and well-known use in the pulp and paper industries, cellulose is increasingly
52
being considered as a potential feedstock for biofuel production, implying its hydrolysis to
53
glucose and fermentation to “cellulosic” ethanol (Berberi et al., 2014). However, the crystalline
54
fraction of the cellulose is the most resistant to acid-hydrolysis and thus may be found as a post-
55
treatment residue. In the current work, in an effort to minimize chemical and energy
56
requirements, it is envisioned that fermentable sugars could be produced out of the easily
57
hydrolysable amorphous cellulose fractions. The crystalline residues could then be used in a
58
second step as natural fibre reinforcement for polymer composites.
cr
us
an
M
d
Various standard pulping methods are available at an industrial level including but not
te
59
ip t
50
limited to mechanical, thermo-mechanical, chemical and chemico-mechanical pulping. Recent
61
studies by Liu, Ju, Li and Jiang (2014) and Wang, Li, Cao, and Tang (2011) investigated the use
62
of ionic liquid extraction of cellulose from various sources and showed that up to 62% cellulose
63
extraction could be obtained using these green solvents, which in addition can be recycled.
Ac ce p
60
64
Among the less common techniques, steam processes as steam explosion, which were
65
first introduced in 1926 by Mason (Mason, 1926) is getting increasing attention in the field of
66
biofuels. The technique is more challenging to scale to an industrial level than classical pulping
67
process although it allows efficient separation of the different constitutive fraction of biomass
68
(Lavoie, Capek-Menard, Gauvin and Chornet, 2010). The process involves cooking water-
69
impregnated lignocellulosic biomass under pressure at temperatures in the range of 160-260˚C.
3 Page 3 of 33
70
After a few minutes of reaction, the content is rapidly purged out of the pressurized vessel and
71
undergoes a rapid decompression or “explosion”. The process is environmentally friendly due to
72
the low acid concentration and no organic solvents. The hydrolytic power of the steam processes is often assessed through a severity factor,
74
which is a valuable tool used to quantify the intensity of the hydrolytic conditions of the steam
75
treatments. When using an acid catalyst, the role of the pH in the reaction is combined to the
76
other hydrolytic parameters (temperature and residence time) generating a combined severity
77
factor (Chum, Johnson, Black and Overend, 1990). The versatility of the severity factor has
78
been demonstrated by adapting it to lignin removal for a wide array of biomass and conversion
79
processes including steam processes (Lee and Lavoie, 2013). Compared to other pre-treatment
80
technologies, steam-explosion presents several advantages particularly for the biorefineries.
81
These advantages are summarized in the work of Garotte, Domínguez, and Parajó (1999). They
82
noted the high hemicellulose yield with by-products showing little degradation and disruption of
83
the solid residues from bundles to individual fibres that occur during the explosion effect. Steam-
84
explosion is generally performed without addition of any chemicals. When a dilute-acid (i.e. with
85
acid concentration ranging between 1 to 5%) is used as a catalyst, it increase the hydrolysis rate
86
of hemicellulose and thus further improves the process.
cr
us
an
M
d
te
Ac ce p
87
ip t
73
Many investigations reporting on the steam processes focused on the extractability of
88
cellulose as a source for high yield fermentable sugars (Mosier et al., 2005) (Brownell, Yu and
89
Saddler, 1986). However, fewer researchers considered the impact of the steam-process on the
90
fibre structure and properties when used for others purposes as in pharmaceutical, biocomposites
91
and biomaterials. The use of steam processes on lignocellulosic materials can lead to significant
92
changes in their structure and properties, such as reduction of the fibre length, fibre degradation
4 Page 4 of 33
93
and modification of the crystalline structure. This process should be better optimized to better
94
understand the parameter for the production of high quality natural fibres.
95
The utilisation of natural fibres as reinforcing fillers for polymer composites has previously been investigated in the open literature. As an example, Petinakis, Yu, Simon and
97
Dean (2003) reviewed that using natural fibres compared with other reinforcing agents such as
98
glass fibres, talc or carbon fibres for improving the performance of biodegradable polymers
99
include the retention of the biodegradability of the composite, but generally also exhibit lower
cr
us
101
density, superior performance, and lower cost due to the abundance of cellulose fibres. Han and Choi (2010) explained the role of surface areas especially with regards to the
an
100
ip t
96
different mechanical properties of reinforced biocomposites. They showed as well that the
103
surface morphology of cellulose fibres is a significant factor to consider for the composite
104
interfacial adhesion. Other important fibre specifications include fibre strength, length to
105
diameter ratio and surface area.
te
d
M
102
In this work, the physical modifications induced by a dilute-acid steam explosion
107
treatment on Salix miyabeana (willow tree) cellulose fibres were investigated. In particular, the
108
objective of the work was to determine how dilute-acid steam-explosion process modifies the
109
fibre length, whiteness, thermal stability, crystallinity and degree of polymerization of cellulose.
110
These results were correlated with the steam process conditions via the combined severity factor.
111
The resulting equations generated from simple linear regressions allow an estimation of the
112
properties of the cellulose fibres at a known severity factor. In addition, the dimensions of
113
extracted fibres were compared with those of the existing commercial microcrystalline cellulose.
114
The potential use of dilute-acid steam-explosion as a new route for microcrystalline cellulose
115
isolation from biomass residues is discussed.
Ac ce p
106
5 Page 5 of 33
116 117
2) Materials and methods
118
2.1. Material Salix myabeana residues obtained from Le Saule Magique Inc. (Laval, Québec, Canada)
120
were used as feedstock. Biomass was processed using an 800g pilot-scale steam gun, which was
121
extensively described by Berberi et al. (2014); Lavoie et al. (2010) and Lavoie and Beauchet
122
(2012). Sulfuric acid 98% and sodium chlorite (technical grade nominally 80%) were bought
123
from Anachemia Canada Inc. Acetic acid glacial (99.8 %) was purchased from VWR
124
International LLC. All reagents were used as received.
an
us
cr
ip t
119
2.2. Method:
127
2.2.1. Steam explosion treatment:
d
126
M
125
Prior to steam processes, the feedstock was grinded and screened to an average size of
te
128
two (2) cm. The severity factor of the steam treatment was determined using an empirical
130
equation developed by Overend and Chornet (1987), which correlates temperature and residence
131
time:
132 133
Ac ce p
129
Log (R0) = log [t × e (T-Tref)/ 14.75]
1
Where Log (R0) is the severity factor, t is the residence time (min), T is the reaction
134
temperature (˚C), Tref is the reference temperature arbitrarily settled at 100˚C assuming that the
135
hydrolytic effect of water on cellulosic material is negligible below this temperature.
136
By including the effect of the pH on the reaction, equation 1 was further adapted to the
137
experimental conditions for the determination of the combined severity factor (SFT), under an
138
acidic catalysts:
6 Page 6 of 33
139 140 141
SFT= Log (R0)-pH
2
(Chum, et al., 1990)
With Log (R0) obtained from equation 1. Screened woodchips (200 g, dry basis) were initially impregnated in a solution of 3% (wt %) sulphuric acid (ratio solid to liquid 1:10, m/ v) for 10 minutes after which they were
143
dewatered with a filter press. The acid concentration was fixed to 3% (wt. %) following
144
preliminary investigation on the feedstock (not published). The acidic-wet biomass was then
145
cooked in the steam gun at three different temperatures (190, 205 and 220 ˚C), which were
146
maintained for a residence time arbitrarily fixed at 3, 5 and 10 minutes respectively. Following
147
steam treatment, the resulting lignocellulosic material was washed using hot water and press
148
filtered until neutral pH was attained. A laboratory filter press was used for dewatering the wet
149
biomass after treatment. By applying a controlled pressure of 0.7 MPa, the water content was
150
reduced to the minimum in order to facilitate the drying process. Fibres were then dried at room
151
temperature in a convection oven for 72 hours. The experimental uncertainty on the process
152
temperature was around 2oC. For the concentration measurement and residence time, the relative
153
uncertainty is below 0.5%. Based on the above, the overall uncertainty on the calculated SFT
154
was less than 1%.
156 157
cr
us
an
M
d
te
Ac ce p
155
ip t
142
2.2.2. Fibre delignification and bleaching Following the steam explosion treatment, Salix myabeana cellulose fibres were bleached
158
using a slight modification of the chlorine dioxide method employed by Kumar, Mago, Balan
159
and Wyman (2009). Bleaching was performed during six hours at 75˚C, using distilled water
160
(ratio solid to liquid 1:20 wt/ v) sodium chlorite (0.6g/ g of dry matter) and acetic acid (0.2 ml/ g
161
of dry matter). The reaction was done in three stages of two hours in which an equal amount of
7 Page 7 of 33
162
sodium chlorite and acetic acid were added to the medium. Fibres were then repeatedly washed
163
with hot and distilled water and dried at room temperature in an air-circulating oven for 72 h and
164
then kept in a sealed plastic bag until analysis.
2.2.3. Fibres characterization.
167
2.2.3.1. Fibre Quality Analysis
Length of the cellulose fibres as well as diameter and length to diameter aspect ratios
us
168
cr
166
ip t
165
were determined using a Fibre Quality Analyzer (FQA) (Optest Hires model, Hawkesbury,
170
Canada).
171
M
173
2.2.3.2. X-Ray diffraction analysis (XRD)
The X-ray diffraction patterns were obtained using a Philips X’Pert diffractometer using
d
172
an
169
copper radiation Kα (λ = 1.5418 Ǻ), a voltage of 40 kV and an operation current of 50 mA. All
175
scans were performed from 2θ = 5° to 2θ = 40°. MDI Jade 10 software packages was used for
176
pattern processing as well as for calculation of the degree of crystallinity (DC). The degree of
177
crystallinity was calculated using the height ratios between the intensity of the crystalline peak
178
(I002 - IAm) and total intensity (I002) after subtraction of the background signal measured without
179
cellulose as also previously reported by Park, Baker, Himmel, Parilla and Johnson (2010).
180
The I002 intensity of the crystalline peak was taken at 2θ= 27˚ and IAm is the intensity of the
181
amorphous phase at 2θ = 18˚.
Ac ce p
te
174
182 183
2.2.3.3. Scanning Electron Microscopy (SEM)
8 Page 8 of 33
The microstructure of the fibres was observed using a Hitachi scanning electron
185
microscope model S-3000N (Hitachi High-Technologies Corporation, Japan) equipped with a
186
tungsten hairpin filament, under different magnifications. Prior to analysis, the surface of the
187
cellulose fibres was sputter-coated with gold/ platinum using a vacuum plasma spray and argon
188
as inert gas. The images were taken at an accelerating voltage of 10 kV.
cr
189
ip t
184
2.2.3.4. Thermogravimetric analysis (TGA)
191
Thermal properties of the different samples were measured using a SETSYS TG-DTA/DSC 24
192
Thermal Analyser with a mass spectrometer (Setaram Instrumentation, Canada). A known
193
weight of each sample that was previously dried at 60˚C for 24h was taken into silica crucible for
194
analysis. The mass loss analyses were performed using argon as a purge gas, with a furnace flow
195
of 20 ml/min and a temperature profile of 10˚C/ min. The normalized mass loss was plotted as
196
differential thermogravimetric curves from 30 to 600˚ C.
197
199
2.2.3.5. Optical properties
Ac ce p
198
te
d
M
an
us
190
The steam-treated bleached fibres were submitted to Technibrite Micro TB-1C
200
reflectometer for optical analysis. Difference in CIE L*, a*, b* color coordinates and whiteness
201
index, expressed in percentage were measured according to the TAPPI T 452 om-98 method.
202 203 204
2.2.3.6. Degree of polymerization Degree of polymerization was obtained following a method reported by Ibrahima,
205
Agblevor and El-Zawawy (2010). Bleached steam-exploded fibres (0.06 g) were wetted in 7.5
206
mL distilled water and stirred during 2 minutes after what 22.5 mL of 1M cupri-ethylenediamine
9 Page 9 of 33
(CED) (copper concentration 0.98-1.02 M, Anachemia Canada Co) solution was added (H2O:
208
CED ratio of 1: 3 (v: v)). The solution, protected from light and oxygen, was agitated during 2h
209
until complete dissolution of the cellulose was reached. A cannon-fenske viscometer N# 75 was
210
used to measure the flow time of the CED-cellulose solution that was used to calculate the
211
intrinsic viscosity [η] according to the equation of Solomon-Gatesman:
cr
212
ip t
207
[η] = (2× (ηsp – Ln (ηrel)) 0.5/ C
With: ηsp = Specific viscosity and ηrel = relative viscosity and C = solution concentration.
214
The degree of polymerization DP was calculated using the following equation proposed by
215
Evans and Wallis (1989):
an
DP0.9= [η] ×1.65
M
216
us
213
3
Finally, the average number of chain scissions (CS) per cellulose chain unit during the time
218
course of degradation was calculated using the equation:
219
CS = (DP0/DP)-1
te
d
217
4
5
Where DP0 is the degree of polymerization before scission. In the present work, the DP obtained
221
at lower severity factor (SFT = 2.88) was taken as the reference DP0 for CS calculation.
222
Ac ce p
220
223
3. Results and discussion
224
3.1 Crystallinity and Optical properties.
225 226
Table 1 presents a summary of the steam explosion process conditions and severity factor
227
as well as key structural features of the treated cellulose: optical properties, and degree of
228
crystallinity. The reaction temperature was varied between 190 and 220o C while the residence
229
time was varied from 3 to 10 min. The sulfuric acid content was kept constant at 3 wt%. These 10 Page 10 of 33
conditions allowed calculating severity factor ranging from 2.88 to 4.29 for the different
231
samples. The fibres, recovered after the steam explosion treatment, were highly crystalline with
232
degrees of crystallinity superior to 85% (determined by X-ray analysis). It is noteworthy that no
233
detectable amount of lignin or hemicellulose were present in the samples. The crystallinity was
234
relatively constant although a slight decrease was observed for higher severity factors. X-ray
235
diffraction patterns (not shown) showed that type I crystalline cellulose was present in all treated
236
samples showing characteristics peaks at 2θ of 14.9˚, 16.3˚, 22.5˚ and 34.6˚ which were assigned
237
to the 101,
us
cr
ip t
230
, 002 and 040 crystallographic planes respectively.
239
an
238
241
#
d
A B
245
C
246
D
247 248
E F G
190
te
244
Temperature Retention H2SO4 Severity (˚C) time (Wt. %)a factor b (min.)
3
Ac ce p
242 243
M
240
Table 1: Steam explosion conditions, severity factor, degree of crystallinity, and optical properties of bleached steam exploded Salix myabeana cellulose fibres Degree of crystallinity (%)
Optical properties L*
a*
b*
3
2.88
88.6 ± 0.7
92.7
-0.51
7.82
-1.11
7.75
190
5
3
3.1
90.6 ± 0.8
93.25
190
10
3
3.4
85.2 ± 0.7
90.05 -1.02
205
3
3
3.32
87.1 ± 0.6
91.94
-0.93
8.85
205
5
3
3.55
87.4 ± 0.7
86.75 -0.02
13.31
205
10
3
3.85
85.7 ± 0.6
69.86
2.18
22.82
220
3
3
3.77
88.2 ± 0.6
80.11
1.17
18.99
10.56
249
H
220
5
3
3.99
85.5 ± 0.7
73.23
1.8
17.95
250
I
220
10
3
4.29
85.3 ± 0.6
42.62
5.78
20.61
251 252
a = H2SO4 concentration corresponding to a pH of 0.24 b = Combined severity factor calculated according to equation (1)
11 Page 11 of 33
253
Similar observations were made by Wang, Jiang, Xu and Sun (2009), when working with
255
Lespedeza crytobotrya specie steam exploded for a fixed retention time of 4 minutes.
256
Nonetheless, other reports aiming at the cristallinity of cellulose after steam process operated on
257
longer periods mentioned that the cristallinity at one point seemed to be increasing
258
proportionally to the severity factor (Ibrahim et al., 2010) (Wang and Chen, 2013). Both
259
observations can be logically correlated to the chemical hydrolysis occurring during steam
260
processes. The severity of the latter being amplified by the dilute acid catalyst led to a more
261
crystalline material with the attack and the dissolution of part of the amorphous regions.
262
Kang, Bansal and Realff (2013) on the other hand, when using SO2 as catalyst observed the same
263
reduction of the fibre crystallinity when more severe treatment was employed. From this, they
264
managed to determine a severity threshold above which a decrease of the crystallinity was
265
perceived during steam treatments. This threshold although not localised in our work could
266
possibly explain the crystallinity reduction after a certain severity value was reached. In that
267
case, according to the reduction of the cristallinity noticed, it seem that the fibres degradation at
268
higher treatment severity was not limited to the cellulose wall, but also could in a lesser extent
269
affect the internal fibre crystalline organization.
cr
us
an
M
d
te
Ac ce p
270
ip t
254
On the other hand, when considering the percent cristallinity discussed in our work,
271
higher crystalline values were noticed compared to those of microcrystalline cellulose
272
investigated by other authors. As an example, Leppänen et al. (2009) when comparing micro
273
cellulose from different sources, revealed a crystallinity of 59% for commercial Avicel and 65%
274
for microcrystalline cellulose obtained from cotton linter.
12 Page 12 of 33
275
The optical properties were characterized using the CIE-LAB color coordinates data where L* (lightness), represents the reflectance in the mid-range of visible spectrum, a*
277
measures the redness (negative value) versus greenness (positive values) of the material and b*
278
corresponds to the measure of the color space going from yellowness (positive value) to blueness
279
(negative value). The variations of those optical parameters for each of the willow cellulose
280
samples are as well provided in Table 1. The latter shows that increasing the severity of the
281
steam treatment leads to a decrease of fibre lightness while the yellowness increased
282
accompanied by a color change from red to green. Figure 1 represents the changes of the
283
bleached cellulose fibres as a function of the temperature and residence time.
Ac ce p
te
d
M
an
us
cr
ip t
276
284 285
Figure 1
13 Page 13 of 33
286
Temperature, in the investigated range, had a stronger effect of fibre discoloration than residence time. As an example of the correlation between appearance of the pulp and severity
288
factor, samples C and D, which were treated at similar severity factor of respectively 3.4 and
289
3.32 shows similar discoloration. When the severity factor reached 3.55 and above, irreversible
290
visible discoloration took place, light at the beginning, but increasing proportionally with the
291
severity factor.
cr
A simple estimation of the optical properties of the material post-bleaching can as well be
us
292
ip t
287
provided by the percent whiteness since it is measured across the entire visible spectrum. Figure
294
2 presents the correlation between the whiteness of bleached fibres and the severity factor (fig.
295
2a) and the combined influence of the couple reaction temperature/ residence time on cellulose
296
whiteness (fig. 2b). The whiteness (W) decreased linearly with the increase of the severity
297
factor, a relation that can be described using the following equation:
d
M
an
293
W a SFT b
299
With: W = percent whiteness, a = - 40.487 ± 1.323 and b = 197.3 ± 5.2
( 6)
Ac ce p
te
298
300 301
Figure 2
14 Page 14 of 33
Hemicellulose present in the initial biomass could be responsible of the discoloration
303
observed in the fibres. However, it should be mostly hydrolyzed during the dilute-acid steam
304
explosion (Tucker, Kim, Newman and Nguyen, 2003). It is thus improbable that hemicelluloses
305
are to be taken accountable for the observed color changes. As well, the bleaching process used
306
in this work (chlorine dioxide) was proven to have no evident effect on cellulose fibres structure,
307
attacking only the aromatic rings of the remaining lignin on the lignocellulose (Lemeune, Jamee,
308
Chang and Kadla, 2004). Therefore, the observed browning of the fibres could be related to a
309
partial degradation of cellulose via successive degradation of carbohydrate structures to furans
310
(or aromatic rings) that would ultimately lead to the production of char at high severity.
311
Coloration of the fibres was also observed by Ibrahim et al. (2010) who related this event to the
312
cell wall components, extractives and sugar degradation after steam pre-treatment at higher
313
severity factor.
cr
us
an
M
d
In this work, all steam-exploded (SE) lignocellulosic materials were bleached under the
te
314
ip t
302
same conditions. Chlorine dioxide bleaching is usually done in three stages using the same
316
amount of reagent for each stage under identical time and temperature conditions. Chlorination
317
of cellulose fibres was proven to produce white cellulose, free or with a relatively low lignin
318
content (Patrick and Martin, 1997) depending on the specie and the initial lignin content of the
319
material. The fact that white samples were obtained under the lower severity factors implies that
320
the browning cannot be related to a remaining lignin content. It seems that utilization of
321
sulphuric acid as catalyst, in addition to its fibres destructuration potential at high temperature
322
also induces a variety of secondary reactions. These reactions could involve formation of new
323
chemicals (such as furans) by the transformation carbohydrate (as cellulose or cellulose side-
324
chain). The darkening of steam-exploded fibres has been previously attributed to the production
Ac ce p
315
15 Page 15 of 33
of acetic, formic and other pyroligneous acids as well as the reactions of these organic acids with
326
lignocellulosic material (Morgan and Stewart, 1941). In addition, a possible polymerization of
327
furans and of other sugar degradation products is also an option (Wang et al., 2009). A simple
328
experiment with carbohydrate in an acid catalyzed mixture provides a pitch-black mixture after
329
reacting at 200 o C for a few minutes. Another possibility, although these combined reactions are
330
hard to monitor, could be the generation of “pseudo-lignin”, defined as aromatic compounds
331
yielding a positive klason lignin value not derived from native lignin during dilute-acid steam-
332
explosion pre-treatment (Hu, Jung, and Ragauskas, 2012). These compounds could be formed
333
following combination of carbohydrate macromolecules and by-products (such as furans) formed
334
during the acid pre-treatment of biomass at high temperature (Sannigrahi, Kim, Jung and
335
Ragauskas, 2011). Pseudo-lignin could be resistant to bleaching procedures and its formation
336
may become more likely at high severity.
d Acid-catalyzed steam processes leads to defibrillation of the lignocellulosic biomass
Ac ce p
339
3.2 Fibre dimensions
te
337 338
M
an
us
cr
ip t
325
340
matrix while the amorphous and peripheral fractions of cellulose are most likely hydrolyzed.
341
Figure 3a shows the evolution of average single fibres length (L), diameter (D) and length to
342
diameter (L/D) aspect ratio as a function of the severity factor (SFT). The fibre lengths were
343
drastically reduced following the steam treatments in a linear fashion with regards to the
344
increasing severity of the treatments. This tendency has been reported by various authors
345
including Datar et al. (2007) when working on non-catalyzed as well as dilute acid steam
346
exploded corn stover performed in the temperature range of 190 to 200 C˚ for a retention time of
347
3 to 5 minutes. Under high temperature and for long residence time (severe treatments), the
348
chemical hydrolysis, condensation, mechanical defiberization and structural disorganisation are 16 Page 16 of 33
amplified, thus leading to the production of shorter cellulose fibres. At lower SFT value of 2.88,
350
an average length of 0.23 mm was recorded, while shorter fibres with an average length of 0.14
351
mm were produced at higher severity. The liner relationship between fiber length and SFT can be
352
described by the following equation:
353
L c SFT d
ip t
349
7
With: L = fiber length, c = - 0.057 ± 0.013 and d = 0.387 ± 0.045
355
Figure 3b presents the evolution of the fine (defined here as fibres with length ranging from 0.07
356
to 0.2 mm) as a function of the severity factor.
Ac ce p
te
d
M
an
us
cr
354
357 358
Figure 3
359 17 Page 17 of 33
360
The fines (F%) accounted for 52% at a severity of SFT=2.8 and their fraction increased linearly
361
with the increasing SFT to reach value of 82% at the higher severity (SFT= 4.3). The correlation
362
between the two parameters followed the equation: F % e SFT g
364
With: F% = percentage of fine, e = 21.39 ± 2.31 and g = 8.51 ± 2.59
365
By contrast, the average fibre diameters value (D) was more stable with only a slight
8
cr
ip t
363
increase as function of the SFT leading to a decreasing L/D ratio with regards to severity. This
367
trend could be explained by agglomeration due to fibre degradation at higher temperature/
368
retention time as mentioned earlier. The curvature after the separation of fibre bundles as a result
369
of the steam processes could also explain the augmentation of the average fibre diameter
370
detected using FQA.
M
an
us
366
The products obtained after the dilute acid steam explosion could be related to
372
microcrystalline cellulose. In fact, Leppänen et al. (2009) reported that hydrolysis of cellulose
373
with dilute-acid cannot isolate nanosize particles of cellulose even after mechanical
374
disintegration. Strong acid must be used for this purpose. When dilute acid hydrolysis is
375
considered, the strong inter-crystalline contacts remain intact and the micron-size crystallites
376
aggregates leading to the formation of microcrystalline cellulose. The average fibre length for
377
commercial Avicel was found to be in the range of 0.02 mm to 0.25 mm, while it ranged from
378
0.02 to 0.125 mm for microcrystalline cellulose obtained from cotton linter (Leppänen et al.,
379
2009). This must be compared to average particle length ranging from 150 to 200 nm and the
380
average particle width ranging from 10 to 20 nm for nanocellulose (Ioelovich, 2012). On the
381
other hand, Ohwoavworhua and Adelakun (2010) found an average fibre diameter of 95 m for
382
microcrystalline cellulose produced by hydrochloric acid hydrolysis of α-cellulose extracted
383
from Sorghum caudatum. In term of fiber length, the average length of the steam exploded
Ac ce p
te
d
371
18 Page 18 of 33
cellulose fibres obtained in this work ranged from 0.15 to 0.23 mm, while their diameter varied
385
between 15.8 to 32.8 m. To some extent, these facts could corroborate the possibility to extract
386
microcrystalline cellulose directly by using dilute-acid steam explosion, which could be a new
387
process route to explore.
ip t
384
388
cr
390
3.5. Microstructural characterization
Figure 4 presents pictures of bleached steam-exploded cellulose fibres obtained with a
us
389
SEM system. The micrographs reveals separated elementary cellulose fibres with some irregular
392
filament bundles persisting in some cases (fig 4 A; B; C; D).
393
an
391
This observation is of interest for an application as polymer reinforcement since isolated fibres are usually targeted for such an application (Altman et al., 2003). In comparison to the
395
lower severity factor (SFT = 2.88), fibres obtained with higher severity (up to SFT= 3.77)
396
seemed to “unfold”. This observation, in addition to the fibre agglomeration observed at higher
397
SFT values (Fig 4 E; G; H; I), could explain the difference in fibre diameter generated from
398
steam processes as reported in Figure 3. The severity of the applied treatment impacted clearly
399
on the final cellulose microstructure with fibre length reduction and loss of fibrous structure
400
occurring progressively with increasing SFT.
d
te
Ac ce p
401
M
394
These structural changes, happening at higher severity factor, could be elucidated by the
402
extension of the hydrolytic chemical phases, subsequent to the lengthening of the residence time.
403
The pressure rise inside the reactor generated by the increasing reaction temperature as well led
404
to more intense depressurization and thus defibrillation. This phenomenon generates a
405
disordering of the biomass cell wall, as previously observed for wheat straw by Anupama and
406
Singh (2011).
19 Page 19 of 33
Ac ce p
te
d
M
an
us
cr
ip t
407
408 409
Figure 4
20 Page 20 of 33
On the other hand, two spherical globules with diameters in the 10 m range in the upper
410 411
left part of Figure 4I may be pseudo-lignin particles as observed by Hu, Jung and Ragauskas
412
(2012) and Sannigrahi et al. (2011) when severe treatments were applied.
414
ip t
413
3.6. Derivative thermo gravimetric analysis
Thermo gravimetric analyses were performed in order to determine the thermal
cr
415
stability of bleached cellulose fibres treated at different severities. Figure 5 presents the relative
417
weight loss and weight loss derivative (WLD) of the treated fibres. The weight loss occurred in
418
two stages for all the tested specimens. The main degradation phase occurred around 300o C
419
while a second minor phase occurs at temperatures above 400o C. The onset degradation
420
temperature (Tonset) as well as the maximum temperature of degradation (TMax) did not
421
significantly change with treatment severity.
422
The weight loss derivative curves (WLD) for all bleached cellulose fibres revealed that the
423
maximum degradation rate occurred between 298 ˚C and 308˚C. This main peak, with maximum
424
WLD appearing between 298˚ C and 308 ˚C depending on the applied conditions, was
425
accompanied with two secondary peaks between 70˚ and 100˚C and between 380˚ C and 550 ˚C
426
attributed respectively to H2O evaporation and remaining impurities (Morán, Alvarez, Cyras and
427
Vázquez, 2008). The last cited peak evidenced that these impurities degraded around
428
temperature range of 400˚C to 550˚C that may approximately correspond to the degradation
429
temperature of isolated lignin discussed by Ibrahim et al. (2010) for cotton stalk and banana
430
plant.
431 432
Ac ce p
te
d
M
an
us
416
However, due to the strong bleaching process applied and the absence of lignin, this peak could be assigned to the pseudo-lignin molecules that present a similar structure with lignin.
21 Page 21 of 33
433
The total carbonaceous residues (char) found at the end of the heating process was low,
434
between 2 and 4.2% with the highest value found for the highest severity factor. The same trend
435
was observed by Jacquet et al. (2011), for steam exploded cellulose fibres.
Ac ce p
te
d
M
an
us
cr
ip t
436
437 438 22 Page 22 of 33
439
Figure 5
440 441
3.7. Degree of polymerization The average chain length of the cellulose macromolecules in the bleached cellulose fibres
443
can be expressed as a degree of polymerization (DP) of the cellulose chains. The variation of the
444
degree of polymerization as a function of the severity factor (SFT) is presented in Figure 6a.
445
Figure 6b presents variation of DP as a function of the average fibre length and Figure 6c shows
446
the evolution of DP as a function of the actual treatment conditions (temperature, residence
447
time).
an
us
cr
ip t
442
Ac ce p
te
d
M
448
23 Page 23 of 33
ip t cr us an M d te 450 451
Ac ce p
449
Figure 6
As for the fiber length presented earlier, the degree of polymerization decreased as a
452
function of the severity factor. The linear relation between these two parameters can be described
453
by the equation:
454
DP h SFT i
9
455
With: DP = degree of polymerization. h = - 1326.3 ± 31.8 and i = 5645 ± 129
456
It is noteworthy that macromolecular chain length expressed as degree of polymerisation
457
and the macroscopic fiber length are closely related through a linear relationship. 24 Page 24 of 33
458
High value of DP of 1830 was observed for fibres treated at the lowest SFT (2.88). A
459
significant reduction of the degree of polymerization down to values around 100 were obtained
460
at higher SFT (4.29). This corresponds to an average chain scission value of 15.7. In Figure 6b, on the other hand, we can observe that the DP was reduced with increasing
ip t
461
temperature and residence time taken individually. Similar variations were also revealed by
463
Foston and Ragauskas (2010) for populus and switchgrass during dilute acid pre-treatment, with
464
the increase of the residence time.
us
cr
462
Terinte, Ibbett and Schuste (2011) observed that the leveling off degree of polymerization
466
(LODP) of microcrystalline cellulose ranges from 200 to 300. Actually, in this work, we noticed
467
that the degree of polymerization of cellulose fibres dropped to around 205 when a severity of
468
3.99 was applied, and reached value as low as 109 at higher severity factor of 4.29. For instance,
469
the average degree of polymerization calculated by Leppänen et al. (2009) for commercial
470
Avicel was found to be around 150 and was around 165 for microcrystalline cellulose obtained
471
from cotton linter. This statement could be in agreement with the above discussed possibility that
472
microcrystalline cellulose can be extracted directly by dilute-acid steam explosion.
M
d
te
Ac ce p
473
an
465
474
4. Conclusions
475
The increased severity applied to salix myabeana during the Steam explosion process affects the
476
length, whiteness, length to diameter aspect ratio as well as the degree of polymerization of the
477
bleached cellulose fibres. The thermal stability of the fibres on the other hand was not a strong
478
function of the reaction parameters. Finally, the obtained fibre structures in term of fibre length,
479
diameter and degree of polymerization suggest that the dilute acid steam explosion process could
480
be used as a new route for direct extraction of microcrystalline cellulose from biomass residues. 25 Page 25 of 33
481 482
Acknowledgement: The authors are grateful to BiofuelNet Canada and to the Industrial Research Chair
484
on Cellulosic Ethanol and Biocommodities of the Université de Sherbrooke for their support of
485
the project. The Chair is co-funded by CRB Innovations, Enerkem and Ethanol Greenfield
486
Québec Inc., as well as the Ministère de l'énergie et des ressources naturelles du Québec. Special
487
thanks to the CRB team for using part of their facilities and expertise in steam treatments.
us
cr
ip t
483
.
488
an
489 490
M
491
d
492
te
493 494
Ac ce p
495 496 497
References:
498
Altman, G. H., Diaz. F., Jakuba, C., Calabro, T., Horan, R. L., Chen, J., Lu H., Richmond,
499 500
J., and Kaplan, D. L. (2003). Silk-based biomaterials. Biomaterials. 24, 401–416. Anupama, K., and Singh, M. (2011). Isolation of cellulose nanofibrils from wheat straw using
501
steam explosion coupled with high shear homogenization. Carbohydrate Research. 346, 76-
502
85.
503
Berberi, V., Fuente-Hernandez, A., Corcos, P-O, Lalonde, O., Almirall-Diaz, I. and Lavoie, J-M. 26 Page 26 of 33
504
(2014) Cellulosic Ethanol from Reed Canary Grass using Bench and Pilot Scale Steam
505
Treatments & Non-enzymatic Hydrolysis. Article submitted to Industrial Crops and
506
products. Brownell, H. H., Yu E. K. C and Saddler J. N. (1986). Steam-explosion pre-treatment of wood:
ip t
507
Effect of chip size, acid, moisture content and pressure drop. Biotechnology and
509
Bioengineering, 28 (6), 792-801.
Chum, H. L., Johnson, D. K., Black, S. K., and Overend, R. P. (1990). “Pre-treatment-
us
510
cr
508
catalyst effects and the combined severity parameter,” Applied Biochemistry and
512
Biotechnology. 24/25, 1-14.
513
an
511
Datar, R., Huang, J., Maness, P. C., Mohagheghi, A., Czernik, S., and Chornet, E. (2007). Hydrogen production from the fermentation of corn stover biomass pretreated
515
with a steam explosion process. International journal of hydrogen energy. 32, 932-939.
518
d
te
517
Evans, R., and Wallis, A. F. A. (1989). Cellulose Molecular Weights Determined by Viscometry. Journal of Applied Polymer Science, 37, 2331-2340. Foston, M., and Ragauskas, A. J. (2010). Changes in lignocellulosic supramolecular and
Ac ce p
516
M
514
519
ultrastructure during dilute acid pre-treatment of Populus and switchgrass. Biomass and
520
Bioenergy, 34, 1885-1895.
521 522
Garrot, G., Domínguez, H., and Parajó J. C. (1999). Hydrothermal processing of lignocellulosic materials. Holz als Roh- und Werkstoff, 57, 191- 202.
523
Han, S. O., and Choi, H. Y. (2010). "Morphology and surface properties of natural fiber treated
524
with electron beam," Microscopy: Science, Technology, Applications and Education 3(1),
525
1880-1887.
526
Hu, F., Jung. S., and Ragauskas, A. (2012). Pseudo lignin formation and its impact on enzymatic
27 Page 27 of 33
527 528
hydrolysis. Bioresource technology. 117, 7-12. Ibrahim, M. M., Agblevor, F. A., and El-Zawawy, W. K. (2010). Isolation and characterization of cellulose and lignin from steam-exploded biomass. BioResources. 5 (1), 397-418.
530
Ioelovich, M. (2012). Optimal conditions for isolation of nanocrystalline cellulose particles.
532
Nanoscience and Nanotechnology, 2 (2), 9-13.
Jacquet, N., Quiévy, N., Vanderghem, C., Janas, S., Blecker, C., Wathelet, B., Devaux,
cr
531
ip t
529
J., and Paquot, M. (2011). Influence of steam explosion on the thermal stability of cellulose
534
fibres. Polymer Degradation and stability. 96, 1582-1588
Kang, Y. Bansal, P. and Realff, M. J. (2013). SO2 catalyzed steam explosion: the effect
an
535
us
533
of different severity on digestibility, accessibility and crystallinity of lignocellulosic
537
biomass. Biotechnology progress. 29 (4), 909-916.
Kumar, R., Mago, G., Balan, V., and Wyman, C. E. (2009). Physical and chemical
d
538
M
536
characterization of corn stover and polar solid resulting from leading pre-treatment
540
technologies. 100, 3948-3962.
Lavoie, J. M., and Beauchet, R. (2012). Biorefinery of Cannabis sativa using one-and two-step
Ac ce p
541
te
539
542
steam treatments for the production of high quality fibres. Industrial Crops and Products,
543
37(1), 275-283.
544
Lavoie, J. M., Capek-Menard, E., Gauvin, H., and Chornet, E. (2010). Quality pulp from mixed
545
softwoods as an added value coproduct of a biorefinery. Industrial & Engineering Chemistry
546
Research, 49(5), 2503-2509.
547
Lee, R. A., and Lavoie, J. M. (2013). From first-to third-generation biofuels: Challenges of
548
producing a commodity from a biomass of increasing complexity. Animal Frontiers,
549
3(2), 6-11.
28 Page 28 of 33
550
Lemeune, S., Jamee, l H., Chang, H.-M., and Kadla, J. F. (2004). Effects of Ozone and Chlorine
551
Dioxide on the Chemical Properties of Cellulose Fibers. Journal of applied polymer sciences.
552
93, 1219-1223. Leppänen, K., Andersson, S., Torkkeli, M., Knaapila, M., Kotelnikova, N., and Serimaa, R.
554
(2009). Structure of cellulose and microcrystalline cellulose from various wood species,
555
cotton and flax studied by X-ray scattering. Cellulose, 16, 999– 1015.
561 562 563 564
cr
us
an
560
United States patent office. US1596159 A
Morán, J. I. Alvarez, V. A., Cyras, V. P., and Vázquez, A. (2008). Extraction of cellulose and
M
559
Mason, H. W. (1926). Low-temperature explosion process of disintegrating wood and the like.
preparation of nanocellulose from sisal fibers. Cellulose. 15, 149-159.
d
558
pretreated by alumina-doped MgO in AMIMCL. Carbohydrate Polymers, 113, 1–8.
Morgan, W. H., and Stewart, O. B. (1941). Process of making light-colored ligno-cellulose fiber.
te
557
Liu, L., Ju, M. T., Li, W. Z., and Jiang, Y. (2014). Cellulose extraction from Zoysia japonica
United States Patent Office. US 2234188 A. Mosier, N., Wyman, C., Dale, C., Elander, R., Lee. Y. Y., Holtzapple, M., and Ladish, M.
Ac ce p
556
ip t
553
565
(2005). Features of promising technologies for pre-treatment of lignocellulosic biomass.
566
BioResource technology, 96, 673-686.
567
Ovwoavworhua, F. O., and Adelakun, T. A. (2010). Non-wood fibre production of
568
microcrystalline cellulose from Sorghum caudatum: characterisation and tableting properties.
569
Indian Journal of Pharmaceutical Sciences. 72 (3), 295-301.
570 571 572
Overend, R.P., and Chornet, E. (1987). Fractionation of lignocellulosics by steam-aqueous pre-treatments. Philosophical Transaction of the Royal Society A. 321 (1561), 523-536. Park, S., Baker, J. O., Himmel, M. E., Parilla, P. A. and Johnson, D. K. (2010). Cellulose
29 Page 29 of 33
573
crystallinity index: measurement techniques and their impact on interpreting cellulose
574
performance. Biotechnology for biofuel. 3, 1-10.
576 577
Patrick, T., and Martin, S. (1997). Chlorine dioxide delignification kinetics and EOP extraction of softwood Kraft pulp. The Canadian Journal of Chemical Engineering. 75, 23-30.
ip t
575
Petinakis, E., Yu, L., Simon, G., and Dean, K. (2003). Natural Fibre Bio-Composites
Incorporating Poly (Lactic Acid). In: Alberto M (Ed) Fiber reinforced polymer: The
579
technology applied for concrete repairs. ISBN 978-953-51-0938-9, 41-59.
580
Sannigrahi, P., Kim, D. H., Jung, S., and Ragauskas, A. (2011).Pseudo-lignin and
582
us
Pretreatment chemistry Energy Environ. Sci.4, 1306-1310.
an
581
cr
578
Terinte, N., Ibbett, R., and Schuste, K. C. (2011). Overview on native cellulose and microcrystalline cellulose I structure studied by X-Ray diffraction (WAXD): comparison
584
between measurement techniques. Lenzinger Berichte, 89, 118-131.
d
Tucker, M. P., Kim, K. H., Newman, M. M., and Nguyen, Q. A. (2003). Effects of Temperature
te
585
M
583
and Moisture on Dilute-Acid Steam Explosion Pre-treatment of Corn Stover and Cellulase
587
Enzyme Digestibility. Applied Biochemistry and Biotechnology. 105–108, 165- 177.
Ac ce p
586
588
Wang, K., Jiang, J.X., Xu, F., and Sun, R.C. (2009). Influence of steaming pressure on the
589
steam explosion pre-treatment of stalks (lespedeza cryotobotra) Part1: characteristic of
590
degraded cellulose. Polymer Degradation and Stability. 94, 1379-1388
591
Wang, N., and, and Chen, H. Z. (2013). Manufacture of dissolving pulps from cornstalk by
592
novel method coupling steam explosion and mechanical carding fractionation. BioResource
593
Technology. 139, 59-65.
594 595
Wang, X. J., Li, H. Q., Cao, Y, and Tang, Q. (2011). Cellulose extraction from wood chip in an ionic liquid 1-allyl-3-methylimidazolium chloride (AmimCl). Bioresource Technology, 102,
30 Page 30 of 33
596
7959–7965.
597 598
ip t
599 600
cr
601
us
602 603
an
604 605
M
606
d
607
te
608
610 611 612 613
Ac ce p
609
Figures captions:
614
Figure 1: Aspects of bleached dilute acid steam-exploded cellulose fibres as a function of the
615
reaction temperature and retention time. (Conditions described in Table 1).
616
Figure 2: (a) Cellulose fibre whiteness (%) as a function of the severity factor and (b) evolution
617
of the whiteness with the retention time.
31 Page 31 of 33
Figure 3: (a) Fibre length, diameter, and Length /Diameter ratio as a function of the severity
619
factor and (b) fine fraction (fibres with length between 0.07 and 0.2 mm) as a function of the
620
severity factor.
621
Figure 4: SEM micrographs of bleached Salix myabeana fibres steam exploded at different
622
severity factors as described in Table 1.
623
Figure 5: Weight loss and weight loss derivative (WLD) of bleached Salix myabean fibres at
624
different steam explosion conditions as described in Table 1.
625
Figure 6: a) Degree of polymerization as a function of the severity factor (■), b) Degree of
626
polymerization as a function of the cellulose fibre length (●) and c) Degree of polymerization of
627
cellulose fibres as a function of the reaction temperature and residence time.
cr
us
an
M Ac ce p
te
d
628 629
ip t
618
32 Page 32 of 33
629
Research Highlights
630
● Extraction of cellulose fibres by dilute-acid steam explosion have been investigated.
632
● Increasing severity factor will signally affect the structure of cellulose fibres.
633
● Correlations show fibre structures changing proportionally with severities factors
634
● The steam process could be applicable to the extraction of microcrystalline cellulose
ip t
631
cr
635
us
636 637 638
an
639 640
M
641
645 646 647 648 649 650 651 652
te
644
Ac ce p
643
d
642
653 654 655 656 657 33 Page 33 of 33