Polymer Degradation and Stability 84 (2004) 331e339

www.elsevier.com/locate/polydegstab

Isolation and characterization of cellulose from

sugarcane bagasse
J.X. Suna, X.F. Suna, H. Zhaoa, R.C. Sunb,)
a
College of Forestry, The North-Western Sciences and Technology University of Agriculture and Forestry, Yangling 712100, China
b
State Key Laboratory of Pulp & Paper Engineering, South China University of Technology, Guangzhou 510640, China
Received 22 December 2003; received in revised form 22 December 2003; accepted 7 February 2004

Abstract

Three different procedures for isolation of cellulose from sugarcane bagasse (SCB) were comparatively studied. Sequential
extractions of dewaxed SCB with water with or without ultrasonic irradiation, various concentrations of alkali and alkaline peroxide
yielded 44.7 and 45.9% cellulose preparations, which contained 6.0 and 7.2% associated hemicelluloses and 3.4 and 3.9% bound
lignin, respectively. Delignification with acidic sodium chlorite followed by extraction with alkali (10% KOH and 10% NaOH) gave
cellulose yields of 44.7 and 44.2%, which contained 5.7 and 3.7% residual hemicelluloses and 1.6 and 1.5% remaining lignin,
respectively. Interestingly, a one-step treatment of SCB with an 80% acetic acide70% nitric acid mixture under the conditions given
yielded 43.0e43.6% of the more pure cellulose fractions, which contained minor amounts of bound hemicelluloses (3.2e4.3%) and
were relatively free of associated lignin (0.2e0.6%). The isolated six cellulose samples were comparatively studied by both degraded
methods such as acid hydrolysis and thermal analysis and non-degradation techniques such as FT-IR and CP/MAS 13C-NMR
spectroscopy, and the relative crystallinity was also comparatively estimated.
Ó 2004 Elsevier Ltd. All rights reserved.

Keywords: Cellulose; Bagasse; Isolation; Degradation; Thermal analysis; Crystallinity

1. Introduction of which is in a crystalline structure. Another 25e35% is

hemicelluloses, an amorphous polymer usually com-
In recent years, there has been an increasing trend posed of xylose, arabinose, galactose, glucose, and
towards more efficient utilization of agro-industrial re- mannose. The remainder is mostly lignin plus lesser
sidues, such as sugarcane bagasse (SCB), as raw amounts of mineral, wax, and other compounds [3,4].
materials for industrial applications. SCB is a residue Cellulose, the major constituent of all plant materials,
produced in large quantities by the sugar and alcohol forms about half to one-third of plant tissues and is
industries, and is mainly used as a fuel to power the constantly replenished by photosynthesis, with estimates
sugar mill [1]. However, the remaining bagasse still of annual world biosynthesis of 1011 tonnes [5]. In
continues to be a menace to the environment, a suitable particular, cellulose is the main constituent of higher
utilization of this residue being an important objective plants, including wood, cotton, flax, hemp, jute, SCB,
to be pursued. Several processes and products have been ramie, cereal straws, etc. This represents a vast potential
reported that utilize SCB as a raw material. These feedstock for a number of industries and has created
include electricity generation, pulp and paper produc- a great deal of research interest. Chemically, cellulose is
tion, and products based on fermentation [2]. About a linear natural polymer of anhydroglucose units linked
40e50% of SCB is the glucose polymer cellulose, much at the one and four carbon atoms by b-glycosidic bonds.
This is confirmed by the presence of three hydroxyl
) Corresponding author. Current address: The BioComposites
groups with different acidity/reactivity, secondary OH
Centre, University of Wales, Bangor, Gwynedd LL57 2UW, UK.
at the C-2, secondary OH at the C-3, and primary OH at
Tel.: C44-1248-370588; fax: C44-1248-370594. the C-6 position, and, accordingly, by the formation
E-mail address: [email protected] (R.C. Sun). of strong various intermolecular and intramolecular

0141-3910/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymdegradstab.2004.02.008
332 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339

hydrogen bonds [6]. It is organised into fibrils, which sunlight and then cut into small pieces (1e3 cm). The
are surrounded by a matrix of lignin and hemicelluloses. cut bagasse was ground to pass a 1.0 mm size screen.
At present, it seems generally accepted that cellulose I The dried powder was first extracted with toluenee
has a parallel chain orientation, while in cellulose II, the ethanol (2:1, v/v) in a Soxhlet apparatus for 6 h, and the
chains are anti-parallel [7]. Furthermore, by the use of dewaxed meal was allowed to dry in an oven at 60 (C
X-ray diffraction patterns and solid-state cross polar- for 16 h. The chemical composition (%, w/w) of the
isation magic-angle spinning carbon-13 nuclear magnetic SCB is cellulose 43.6%, hemicelluloses 33.5%, lignin
resonance (CP/MAS 13C-NMR) spectroscopy, four 18.1%, ash 2.3%, and wax 0.8% on a dry weight basis.
major polymorphs of cellulose have been reported,
namely celluloses I, II, III, and IV. Cellulose I is the
native and predominate crystalline structure of alga, bac- 2.2. Isolation of cellulose
terial, some animal and most higher plants, and can be
converted into the other polymorphs through a variety of The procedure for isolation of cellulose using alkaline
treatments [8]. The cellulose fibril is partly crystalline, peroxide with or without ultrasonic treatment is illus-
with two different crystal forms, cellulose Ia and cellulose trated in Fig. 1. The dewaxed SCB (40 g) was sequen-
Ib. Cellulose Ia has one-chain triclinic structure and tially treated with 300 ml H2O at 55 (C for 2 h with or
cellulose Ib has two-chain monoclinic structure [9], and without first ultrasonic irradiation for 40 min, 0.5 M
they differ in hydrogen bonding [10]. In addition, cellu- NaOH, 0.5%, 1.0%, 1.5%, 2.0%, and 3.0% H2O2 in
lose Ia is reported as the dominant polymorph in 200 ml 0.5 M NaOH, and 200 ml 2 M NaOH at 55 (C
bacterial and alga celluloses, while cellulose Ib is pre- for 2 h. The insoluble residue was collected by filtration,
dominant in higher plants such as cotton, wood and washed with distilled water until the pH of the filtrate
bagasse. It is also known that cellulose Ia can be irrevers- was neutral, then dried at 60 (C, and named as cellulose
ibly converted to cellulose Ib by the application of heat preparation C1 and C2, respectively.
[11]. It can be seen from the NMR spectra of cellulose I The scheme for isolation of cellulose by delignifica-
that the broad single at 110 ppm has two shoulders, tion with acidified sodium chlorite is shown in Fig. 2.
which are due to a mixture of two sub-polymorphs Ia and The extractive free bagasse (40 g) was first treated
Ib. This has also been noticed to happen during pulp- with distilled water for 2 h at 70 and 80 (C, respectively.
ing [12]. Non-crystalline cellulose forms are also present Further the two water-soluble free samples were delig-
in the fibril: paracrystalline cellulose and cellulose at nified with 1.3% sodium chlorite at pH 3.5e4.0, ad-
inaccessible and accessible fibril surfaces [13,14]. justed with 10% acetic acid, at 75 (C for 2 h. Finally the
A protocol originally described by Green [15] using holocellulose was extracted with 10% potassium hy-
acidified sodium chlorite is frequently used to delignify droxide and 10% sodium hydroxide for 10 h at 20 (C,
wood as an initial step in the isolation of cellulose. respectively. After filtration, the two residues were
However, environmental concerns associated with chlo- washed thoroughly with distilled water and 95% ethanol
rinating agents (e.g. NaClO2) have led to the increased use and dried in an oven for 16 h at 60 (C. They are
of more environmentally benign agents for delignification considered to be cellulosic fractions C3 and C4.
such as hydrogen peroxide or an acetic acidenitric acid The method for isolation of celluloses with an 80%
mixture in both elemental chlorine-free (ECF) and totally acetic acide70% nitric acid mixture (10:1, v/v) follows
chlorine-free (TCF) isolation sequences [16,17]. In this the chemistry described by Crampton and Maynard [18]
study we compare these three different procedures for and Brebdel et al. [19]. SCB sample (5.0 g) was weighed
cellulose isolation by using alkaline peroxide, acidified
sodium chlorite, and an acetic acidenitric acid mixture.
The cellulosic preparations obtained were characterized
by their yield, content of lignin and hemicelluloses,
viscosity, molecular weight, and thermal stability using
both degraded methods and non-degraded techniques. In
particular, CP/MAS 13C-NMR was used to investigate
the structural changes of the cellulosic polymers.

2. Experimental

2.1. Materials

Sugarcane bagasse was obtained from a local sugar Fig. 1. Scheme for isolation of cellulose from sugarcane bagasse with
factory (Guangzhou, China). It was first dried in or without ultrasonic irradiation.
 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339 333

cellulose samples was determined according to Tappi

method T 249 cm-85.
Viscosity of the cellulosic preparations was measured
by British Standard Methods for determination of
limiting viscosity number of cellulose in dilute solutions,
Part 1. Cupriethylenediamine (CED) method (BS 6306:
Part 1: 1982). The viscosity average DP (degree of
polymerisation) of the cellulose samples was determined
from their intrinsic viscosity [h] in cupriethylenediamine
hydroxide (cuene) solution, P0:90 ¼ 1:65½h
=ml g1 ,
where P is an indeterminate average DP [22]. Molecular
weight of the cellulosic preparations was then calculated
from their P multiplied by 162, molecular weight of an
anhydroglucose.
Fourier transform infrared (FT-IR) spectra of the
cellulose fractions were recorded on a Nicolet 750
Fig. 2. Scheme for isolation of cellulose from delignified sugarcane spectrophotometer in the range 4000e400 cm1 using
bagasse.
a KBr disc containing 1% finely ground sample. CP/
MAS 13C-NMR spectra were recorded using a Bruker
into 200 ml Pyrex tubes. Subsequently, 100 ml 80% MSI400 spectrometer at 25 (C and 62.9 MHz for car-
aqueous acetic acid (w/w) and 10 ml 70% nitric acid bons. The MAS rate was 3 kHz. Each spectrum was
(w/w) was added. The tubes were sealed using screw-caps obtained with an accumulation of 5000 scans. The delay
fitted with Teflon liners and placed into an oil bath set at time was 60 s, the proton 90( pulse width was 9 mm, and
the required temperature at 110 and 120 (C for 20 min, the contact time for cross polarisation was 2 ms.
respectively. When the tube was removed from the oil Thermal analysis of the cellulose samples was per-
bath and cooled, 60 ml of distilled water was added, and formed using thermogravimetric analysis (TGA) and
the reagent was decanted off. The residue was then differential scanning calorimetry (DSC) on a simulta-
thoroughly washed with distilled water and 95% ethanol neous thermal analyzer (STA 625). The apparatus was
to remove the nitric acid and extraction breakdown continually flushed with nitrogen. The sample weighed
products. The residue was then dried in an oven at 60 (C between 9 and 12 mg. Each sample was heated from
for 16 h and was labelled as cellulosic preparations C5 room temperature to 600 (C at a rate of 10 (C min1.
and C6. To reduce errors and confirm the results, each
experiment was repeated in triplicate under the same
conditions, and the yield of cellulose was given as the 3. Results and discussion
average of these three replicates. The standard errors or
deviation were observed to be lower than 5.8%. 3.1. Yield of cellulose

Hydrogen peroxide is widely used for bleaching

of mechanical pulps under alkaline conditions [23].
2.3. Characterisation of cellulose Currently, with the development of totally chlorine-free
bleaching technologies, there is also a growing interest in
The neutral sugar composition of the six cellulosic the use of hydrogen peroxide as one of the oxidants
preparations was determined by gas chromatography replacing chlorine-based reagents [24]. It is generally
(GC) analysis of the corresponding alditol acetates. The accepted that perhydroxyl anion HOO is the most
sample (10 mg) was treated with 72% H2SO4 (0.125 ml) important active species involved in the suppression of
for 45 min at room temperature by agitation on a vortex chromophores in lignin macromolecules [25]. On the
mixture. The solution was then diluted to 1.475 ml, other hand, the radicals such as HOc and Oc 2 produced
heated at 100 (C for 2.5 h, cooled, and neutralised with at high pH levels can participate in lignin degradation
0.32 ml 15 M ammonia. After reduction, the resulting and hemicellulose dissolution [21,26]. As can be seen in
alditols were acetylated for GC analysis as described by Table 1, sequential treatment of dewaxed SCB with
Blakeney et al. [20]. Method for the determination of distilled water at 55 (C for 2 h with or without first
phenolic acids and aldehydes in nitrobenzene oxidation ultrasonic irradiation for 40 min, 0.5 M NaOH, 0.5%,
mixtures of lignins associated in the cellulosic prepara- 1.0%, 1.5%, 2.0%, and 3.0% H2O2 under alkaline
tions with high performance liquid chromatography condition, and 2 M NaOH at 55 (C for 2 h resulted in
(HPLC) has been described in a previous paper [21]. dissolution or degradation of 95.5 and 94.7% of the
Klason lignin content in the unbleached and bleached original hemicelluloses, and 91.7 and 90.2% of the
334 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339

Table 1
Isolation conditions and yield of cellulose from sugarcane bagasse
Fraction no. Isolation conditions Yield
(% dry SCB)
C1 Sequential treatments of dewaxed SCB (9.92 g) with 300 ml H2O at 55 (C for 2 h with 44.7
ultrasonic irradiation for 40 min, 0.5 M NaOH, 0.5%, 1.0%, 1.5%, 2.0%,
and 3.0% H2O2 in 200 ml 0.5 M NaOH, and 200 ml 2 M NaOH at 55 (C for 2 h
C2 Sequential treatments of dewaxed SCB (9.92 g) with 300 ml H2O at 55 (C for 2 h without 45.9
ultrasonic irradiation, 0.5 M NaOH, 0.5%, 1.0%, 1.5%, 2.0%, and
3.0% H2O2 in 200 ml 0.5 M NaOH, and 200 ml 2 M NaOH at 55 (C for 2 h
C3 Treatment of the delignified SCB (30.7 g) with 10% KOH (600 ml) for 10 h at 20 (C 44.7
C4 Treatment of the delignified SCB (30.5 g) with 10% NaOH (600 ml) for 10 h at 20 (C 44.2
C5 Treatment of 5.0 g SCB with 100 ml 80% acetic acide70% nitric acid (10/1, v/v) at 110 (C for 20 min 43.6
C6 Treatment of 5.0 g SCB with 100 ml 80% acetic acide70% nitric acid (10/1, v/v) at 120 (C for 20 min 43.0

original lignin, yielding 44.7 and 45.9% cellulose (Table the lignin, non-cellulose polysaccharides and other com-
1), respectively. In comparison, treatment with sonica- ponents, yielding 43.6 and 43.0% pure cellulose, which
tion time for 40 min solubilized 0.8 and 1.5% higher of are equal or rather close to the value of a-cellulose in
the original hemicelluloses and lignin, respectively. This SCB (43.6%). The improved yield reported here might
slightly higher efficiency of the ultrasound-assisted be a consequence of using a one-step protocol to mini-
extraction can be explained by the mechanical action mize cellulose loss, since the method for a two-step
of the ultrasound on the cell walls resulting in an in- protocol to isolate cellulose involves a sodium chlorite
creased accessibility and extractability of the hemi- delignification as the first stage followed by extraction
cellulosic and lignin components [27,28]. with alkali to remove hemicelluloses.
Traditionally, acidified sodium chlorite is used to
delignify wood as an initial step in the isolation of pure
3.2. Content of hemicelluloses and lignin
cellulose, and chlorine is widely used as a bleaching
agent in the pulp or cellulose industry [29]. In aqueous
It is well known that cellulose and hemicelluloses are
media their reactions with lignocellulosic materials
the major components of the secondary layers of the cell
result in aromatic substitution, in some cases accompa-
wall in lignocellulosic fibres [32]. These hemicelluloses
nied by displacement of side chains, and in oxidation
are present in the surface layer, i.e., the outer layer of the
reactions leading to quinonoid structures [30]. Its re-
fibres, where these polymers can act as adhesives to
action with lignin is fast compared with its reaction with
strengthen the bonds between individual fibres in the
cellulose. However, its rate of reaction with the cellulose
three-dimensional network of a sheet of paper [33].
is still happened and the reaction damages the fibres
Hydrolysis of the cellulose preparations using sulphuric
during the usual conditions of chlorine bleaching [31]. In
acid and analysed using GC revealed that glucose was
the present study, delignification of water treated SCB
a predominant monosaccharide, comprising 92.8e
with acidified sodium chlorite and following removal of
96.8% of the total neutral sugars (Table 2). Xylose
hemicelluloses using 10% KOH and 10% NaOH yielded
(1.7e4.7%) and galactose (0.3e1.6%) appeared as small
44.7 and 44.2% cellulose (C3, C4), in which 92.5 and
amounts. Arabinose and mannose exhibited minor
95.5% of the original hemicelluloses were released. This
quantities. Thus, SCB cell wall material processed in
significant solubility of hemicelluloses was probably that
these three procedures can provide cellulose with high
the hemicelluloses are present mainly on outer fibre
surface, from where they dissolve easily in the alkaline
solution. Contrastly, cellulose is located in the inner Table 2
parts of the fibres and therefore is not easily dissolved. The content of neutral sugars (relative % cellulosic sample, w/w) in six
cellulosic preparations obtained from SCB
In recent years, the environmental risks associated
with the traditional delignification and bleaching using Sugars (%) Cellulosic preparationsa
elemental chlorine fostered the development of new C1 C2 C3 C4 C5 C6
systems free from elemental chlorine or totally chlorine- Rhamnose Trb Tr NDc ND ND ND
free. In this experiment we also use a totally chlorine- Arabinose 0.70 0.98 0.68 0.56 Tr Tr
free system for cellulose isolation and purification based Xylose 3.39 4.69 2.32 1.73 3.15 2.48
Mannose 0.38 0.98 0.43 0.33 0.61 0.41
on the simultaneous delignification and removal of non- Galactose 1.50 1.56 1.24 1.05 0.52 0.34
cellulose polysaccharides using an acetic acidenitric Glucose 94.03 92.78 94.28 96.33 95.72 96.85
acid mixture. As shown in Table 1, treatment of SCB a
Corresponding to the cellulosic preparations in Table 1.
with an 80% acetic acide70% nitric acid (10/1,v/v) b
Tr, trace.
c
mixture at 110 and 120 (C for 20 min removed most of ND, not detected.
 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339 335

purity. This is particularly true when an acetic acide dehyde, acetosyringone, vanillin, p-hydroxybenzalde-
nitric acid mixture was used to isolate cellulose. The hyde, acetovanillone, and syringic acid. Small amounts
data in Table 2 also indicated that treatment of SCB of p-hydroxybenzoic and ferulic acids, and traces of
with ultrasonic irradiation resulted in increase in the p-coumaric and vanillic acids were also found to be pre-
purity of cellulose, since the glucose content increased sent in the nitrobenzene oxidation mixtures. This rela-
from 92.8% in C2 (without ultrasonic treatment) to tively high amount of syringaldehyde and acetosyringone
94.0% in C1 (with ultrasonic irradiation). Similarly, an suggested that the majority of the hemicelluloses in the
increase of pre-treatment temperature from 70 (C3) to cell walls of SCB are bonded to lignin via syringyl units. In
80 (C (C4) combining increase of alkali strength from addition, the data in Table 3 showed that the two cellulose
10% KOH to 10% NaOH led to an increment in fractions C1 and C2 isolated from the lignified SCB,
cellulose purity from 94.3% in C3 to 96.3% in C4. In contained a higher amount of bound lignin (3.4e3.9%)
addition, as Table 1 shows, an increasing cellulose purity than that of the two cellulose preparations extracted from
can be observed between C5 ( glucose, 95.7%) and C6 the delignified SCB, 1.5e1.6%. More interestingly, trace
( glucose, 96.9%) as the temperature increased from 110 of contaminated lignin in C5 and C6 preparations stated
(C5) to 120 (C (C6). that using an acetic acidenitric acid mixture was a rapid
Evidence for the existence of chemical bonds between and simple method to isolate pure a-cellulose, which is
lignin and polysaccharide components in lignocellulosic relatively free of bound lignin.
materials, i.e., ligninecarbohydrate complexes, has been
accumulated over a period of several years and the 3.3. Viscosity and molecular weight
concept is now largely accepted [34]. It has been
suggested that the major part of the linkages occur Viscosity is a measure of molecular weight and the
between lignin and the side groups of hemicelluloses higher the molecular weight, the higher is its viscosity
[35]. It was found that the most likely linkage types were [37]. Therefore, in this study the molecular weight of the
benzyl ether, benzyl ester and glycosidic [36]. To verify six cellulose preparations was determined with the
the residual lignin linked to the residual hemicelluloses knowledge of their intrinsic viscosity. The intrinsic
in the cellulose preparations, the six cellulose samples viscosity, a measure of the hydrodynamic volume occu-
were subjected to oxidize by alkaline nitrobenzene pied by a molecule, is a measure of the capacity of a
oxidation, which provided an estimate of the amounts polymer molecule to enhance the viscosity [38]. The
of associated lignin and indication of its composition. decrease in intrinsic viscosity means a decrease of the
Table 3 gives the yield (% cellulosic sample, w/w) of hydrodynamic volume of the macromolecular chain
phenolic acids and aldehydes from alkaline nitro- [39]. In other words, intrinsic viscosity is a characteristic
benzene oxidation of the associated lignin in the six of macromolecules that is related directly to their ability
isolated cellulosic preparations. As can be seen in the to disturb flow and indirectly to their size and shape.
table, the major products were identified to be syringal- For molecules that can exist with a variety of molecular
weights, the relation between intrinsic viscosity and
molecular weight is one of the most important pro-
Table 3
The content of lignin and the yield (% cellulosic sample, w/w) of
perties [40]. Based on an extensive study of cellulose
phenolic acids and aldehydes from alkaline nitrobenzene oxidation of molecular weights determined by viscosity, Evans and
the associated lignin in the six isolated cellulosic preparations obtained Wallis [22] demonstrated that the degree of polymerisa-
from SCB tion (DP) of a cellulose sample from its intrinsic
Phenolic acids Cellulosic preparationsa viscosity h in cupriethylenediamine hydroxide (cuene)
and aldehydes C1 C2 C3 C4 C5 C6 solution can be calculated by using the equation of
p-Hydroxybenzoic acid 0.056 0.058 0.011 0.012 0.013 0.010 P0:90 ¼ 1:65½h
, where P is an indeterminate average DP.
p-Hydroxybenzaldehyde 0.14 0.16 0.10 0.10 0.010 0.010 Table 4 lists the intrinsic viscosity (h), viscosity average
Vanillic acid 0.002 0.004 0.006 0.004 0.007 Trb DP (degree of polymerisation) (P), and molecular
Vanillin 0.19 0.20 0.076 0.055 0.043 0.015 weight (Mw) of the six isolated cellulosic preparations.
Syringic acid 0.10 0.12 0.023 0.022 Tr NDc
Clearly, the data showed that the first four cellulose
Syringaldehyde 0.37 0.41 0.20 0.18 0.038 0.012
Acetovanillin 0.12 0.11 0.082 0.055 0.019 0.008 preparations isolated from both lignified and delignified
Acetosyringone 0.25 0.27 0.056 0.061 0.022 0.009 SCB had higher intrinsic viscosity (356.2e415.6 ml/g)
p-Coumaric acid 0.007 0.010 0.008 0.010 0.010 0.006 and molecular weight (192 000e227 850 g/mol) than
Ferulic acid 0.012 0.013 0.009 0.011 0.005 ND those of the last two fractions (h, 256.3e303.2 ml/g;
Total 1.25 1.36 0.57 0.46 0.17 0.070 Mw, 133 250e160 570 g/mol), indicating that treatment
Content of Klason lignin 3.35 3.86 1.62 1.45 0.58 0.18 of SCB with an acetic acidenitric acid mixture under the
a
Corresponding to the cellulosic preparations in Table 1. condition used does degrade the macromolecule of
b
Tr, trace. cellulose. In addition, a lower h and Mw in C1 fraction
c
ND, not detected. than those in C2 preparation stated that a noticeable
336 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339

Table 4
The intrinsic viscosity (h), viscosity average DP (degree of polymerisation) (P), and molecular weight (Mw) of the isolated cellulosic preparations
Cellulosic preparationsa
C1 C2 C3 C4 C5 C6
Intrinsic viscosity (h, ml/g)b 356.2 400.8 415.6 412.8 303.2 256.3
Viscosity average DP (P)c 1185.2 1351.0 1406.5 1396.0 991.2 822.5
Molecular weight (Mw)d 192 000 218 860 227 850 226 150 160 570 133 250
a
Corresponding to the cellulosic preparations in Table 1.
b
Determined by British Standard Methods for determination of limiting viscosity number of cellulose in dilute solutions, Part 1.
Cupriethylenediamine (CED) method.
c
Calculated by P0:90 ¼ 1:65½h
, P represents the viscosity average DP (degree of polymerisation).
d
Calculated by P!162.

degradation of cellulose occurred during the ultrasonic sponds to the bending mode of the absorbed water [36].
irradiation. Furthermore, as compared with the fraction Each spectrum gives a peak at 1428 cm1 which is
C5, lower values of h and Mw in C6 preparation implied attributed to the CH2 bending and this at 1374 cm1 to
that cellulose favoured degradation at higher tempera- the OeH bending. The absorbance at 1328 cm1 arises
ture during the isolation of pure cellulose with acetic from the CeC and CeO skeletal vibrations [42]. The
acidenitric acid mixture. peak at 1261 cm1 is indicative of the OH in-plane
bending cellulose. The absorption band at 1169 cm1
relates to CeO antisymmetric bridge stretching. The
3.4. Spectroscopic characterization CeOeC pyranose ring skeletal vibration occurs in the
region 1076e1023 cm1. The peak at 903 cm1 is origi-
FT-IR spectroscopy has been extensively used in nated from b-glycosidic linkages between glucose units
cellulose research, since it presents a relatively easy in cellulose [43]. In comparison, the spectral profiles and
method of obtaining direct information on chemical relative intensities of the bands in spectra 1 and 2 are
changes that occur during various chemical treatments rather similar; indicating same structures of the two
[41]. Fig. 3 illustrates FT-IR spectra of cellulosic celluloses isolated from either lignified SCB with
preparations C1 (spectrum 1), C4 (spectrum 2) and C6 alkaline peroxide or delignified bagasse with alkali. On
(spectrum 3) isolated from dewaxed sugarcane bagasse. the other hand, it should be noted that the spectrum 3
Obviously, all the extraction procedures removed most provides evidence of slight acetylation by showing the
of the lignin polymers because of the disappearance presence of the three acetyl ester bands at 1745 (CaO
of the lignin-associated absorbances at 1600 and ester), 1374 (eCeCH3), and eCeOe stretching band at
1510 cm1. The absorption of 3420 cm1 is due to 1261 cm1 [44], indicating that acetylation did occur
stretching of CeH groups and that one at 2910 cm1 to during the treatment with an 80% acetic acide70%
the CeH stretching. The band at 1639 cm1 corre- nitric acid mixture under the condition used.
Solid-state 13C-NMR is one of the promising new
70 experimental methods that can be applied to the study
1 of cellulose crystal structure and morphology. Such 13C
65
spectra, acquired with the combined techniques of
60 protonecarbon cross polarisation and magic-angle
2
55
sample spinning, not only contain unique signatures
corresponding to the various recognized crystal forms of
50 3
cellulose but also distinguish different forms within
45 a grouping such as cellulose I. Some of these latter
1328
1639 903 distinctions arise due to morphological features, such as
40 1745
1261
varying degrees of disorder [45]. The CP/MAS 13C-
35 1374 671 NMR spectra of the cellulosic preparations C2 (spec-
1428
30 1169 trum a), C3 (spectrum b), and C6 (spectrum c) are shown
1023 in Fig. 4. Starting on the upfield side of the spectrum,
25 1076
the two peaks at 62.4 and 64.8 ppm is assigned to C-6 of
4000 3000 2000 1000 cellulose (64.8 ppm, crystalline cellulose, 62.4 ppm,
amorphous cellulose). The cluster of resonances from
Fig. 3. FT-IR spectra of cellulosic preparations C1 (spectrum 1), C4
72.5 to 74.8 ppm represent C-2, C-3, and C-5. The next
(spectrum 2) and C6 (spectrum 3) isolated from dewaxed sugarcane two peaks at 82.3e83.7 and 88.8 ppm relate to C-4
bagasse. (88.8 ppm, crystal-interior cellulose, and 82.3e83.7 ppm,
 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339 337

cellulose preparations, extracted with alkali from de-

lignified bagasse and obtained directly by treatment of
SCB with an acetic acidenitric acid mixture, since the
intensity of peaks at 83.7 and 62.4 ppm assigned to non-
crystalline cellulose in the spectra b and c are larger than
that of the signals at 88.8 and 64.8 ppm, which cor-
respond to crystalline cellulose C-4 and C-6, although
the residual hemicellulose content of three samples is
in a decreasing order, 7.2% in C2, 5.7% in C3 and 3.2%
in C6. The combined intensities of the two signals at
83.7 and 62.4 ppm previously assigned to accessible and
inaccessible surface were the same in the cellulose pre-
paration C3 and C6. The current results revealed that
the crystallinity of cellulose decreased by the treatment
of SCB with acidified sodium chlorite followed by alkali
extraction or with an acetic acidenitric acid mixture
under the condition given.

3.5. Thermal analysis

The thermal decomposing pattern of the cellulosic

preparations C1 (Fig. 5a), C3 (Fig. 5b), and C6 (Fig. 5c),
obtained by sequential extraction with alkali and
alkaline peroxide under ultrasonic assistance for
40 min, by delignification with acidified sodium chlorite
and followed alkali extraction, and with an acetic acide
nitric acid mixture gave additional evidence to the
relatively higher stability of the more pure cellulosic
preparation. As illustrated in Fig. 5, the TGA curves of
three cellulose fractions C1, C3, and C6 started to
decompose at 205 (Fig. 5a), 240 (Fig. 5b), and 260 (C
(Fig. 5c), respectively. Similarly, at 50% weight loss the
decomposition temperature of the three cellulose pre-
parations occurred at 305 (C1), 318 (C3) and 320 (C
(C6), respectively. This indicated that the thermal
stability of the cellulose increased with its purity. The
reason for this relatively higher thermal stability of the
pure cellulose is probably due to the substantial removal
of less stable hemicellulosic polymers from SCB during
the treatment with 80% acetic acide70% nitric acid
Fig. 4. The CP MAS 13C-NMR spectra of the cellulosic preparations
C2 (spectrum a), C3 (spectrum b), and C6 (spectrum c). under the conditions given. In addition, the DSC
thermogram of the cellulosic sample C1 gave two
smaller exothermic peaks centred at 305 and 430 (C,
crystal-surface cellulose), and finally a peak at 104.7 ppm while the cellulosic preparations C3 and C6 exhibited
is due to C-1 of cellulose. It is apparent that there are a larger asymmetrical exothermic peak at 320 and
differences in supermolecular structure among the three 340 (C, respectively. These exothermic peaks are due to
samples. In the spectrum a of cellulose fraction C2 two the disintegration of intramolecular interaction and the
sharp signals from crystalline cellulose are dominant decomposition of the polymer.
at 88.8 (C-4) and 64.8 (C-6) ppm, while the signals In short, the crystallinity index of the last four
attributed to non-crystalline cellulose forms occur by cellulose preparations (C3eC6), isolated with alkali from
broad peaks at 83.7 ppm (C-4, accessible fibril surface) delignified SCB or with an 80% acetic acide70% nitric
and 82.3 ppm (C-4, inaccessible fibril surface and hemi- acid mixture, was detected to be lower than that of the
celluloses) and at 62.4 ppm (C-6) [46]. This indicated first two fractions (C1 and C2), extracted sequentially
that the cellulose fraction C2, isolated successively with with alkali and alkaline peroxide from lignified SCB.
alkali and alkaline peroxide from lignified SCB, was This can be considered as an indication of cellulose
affected to a much smaller extent as compared to the crystallinity being lower or the size of crystallites smaller
338 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339

110 45 relatively free of contaminated lignin and hemicellu-

100 (a) loses. However, it should be noted that this method also
40
90 resulted in some degradation and partial acetylation of
35
80
30 the macromolecular structure of cellulose.
70
25
Weight loss

mCal/s
60
20
50
15
40
10
Acknowledgements
30
20 5
10 0
The authors are grateful for the financial support of
0 -5
this research from the major program of Ministry of
0 100 200 300 400 500 600 700
Education China, National Natural Science Foundation
Temperature (°C) of China (No. 30025036), Guangdong Natural Science
Foundation (No. 36567), and Shaanxi Foundation of
110 25 Science and Technology.
100 (b)
90 20

80
15
70 mCal/s References
Weight loss

60
10
50 [1] Simkovic I, Mlynar J, Alfoldi J. Holzforschung 1990;44:113.
40 5 [2] Pandey A, Soccol CR, Nigam P, Soccol VT. Bioresource Technol
30 2000;74:69.
20 0 [3] Jacobsen SE, Wyman CE. Ind Eng Chem Res 2002;41:1454.
10 [4] Wyman CE. Annu Rev Energy Environ 1999;24:189.
0 -5 [5] Goodger EM. Hydrocarbon fuels, production, properties and
0 100 200 300 400 500 600 700
performance of liquids and gases. London: Macmillan; 1976.
[6] Kadla JF, Gilbert RD. Cellulose Chem Technol 2000;34:197.
Temperature (°C)
[7] Lennholm H, Iversen T. Nord Pulp Pap Res J 1995;10:104.
[8] Hayashi J, Kon H, Takai M, Hatano M, Nozawa T. ACS Symp
110 16
Ser 1987;340:134.
100 (c)
14 [9] Sugiyama J, Vuong R, Chanzy H. Macromolecules 1991;24:4168.
90 [10] Sugiyama J, Persson J, Chanzy H. Macromolecules 1991;24:2461.
12
80 [11] Atalla RH, VanderHart DL. Science 1984;223:283.
10
70 [12] Newman RH, Hemmingson JA. Eighth International Symposium
mCal/s

8
Weight loss

60 on Wood Pulp Chem, Helsinki, Finland; 1995. p. 519.

50 6 [13] Newman RH. Holzforschung 1998;52:157.
40 [14] Duchesne I, Hult EL, Molin U, Daniel G, Iversen T, Lennholm
4
30 H. Cellulose 2001;8:103.
2
20 [15] Green JW. In: Whistler RL, editor. Methods of carbohydrate
10 0 chemistry, III. New York: Academic Press; 1963. p. 9.
0 -2 [16] Crampton EW, Maynard LA. J Nut 1938;15:383.
[17] Brendel O, Iannetta PPM, Stewart D. Phytochem Anal 2000;11:7.
0 100 200 300 400 500 600 700
[18] Crampton EW, Maynard LA. J Nut 1938;15:383.
Temperature (°C) [19] Brebdel O, Iannetta PPM, Stewart D. Phytochem Anal 2000;11:7.
Fig. 5. TGA/DSC curves of the cellulosic preparations C1 (a), C3 (b), [20] Blakeney AB, Harris PJ, Henry RJ, Stone BA. Carbohydr Res
and C6 (c). 1983;113:291.
[21] Sun RC, Fang JM, Tomkinson J. Ind Crops Prod 2000;12:71.
[22] Evans R, Wallis AFA. J Appl Polym Sci 1989;37:2331.
on the surface of the last four cellulose preparations. [23] Allison RW, Graham KL. J Pulp Pap Sci 1989;15:145.
This is mainly caused by the delignification with [24] De Choudents C, Agelier R, Chirat C. Tappi Int Pulp Bleaching
acidified sodium chlorite or treatment with organic Conference, Proceedings; 1996. p. 547.
acids at high temperature, while the sequential treatment [25] Rukowski J. Przeglad Papierniczy 1994;11:524.
with alkali and alkaline peroxide prior to with or [26] Wojciak A, Sikorski M, Moreno RG, Bourdelande JL, Wilkinson
F. Wood Sci Technol 2002;36:187.
without ultrasonic irradiation was not severe enough to [27] Hromadkova Z, Kovacikova J, Ebringerova A. Ind Crops Prod
cause decrystallization of cellulose. These results con- 1999;9:101.
firmed the validity of the approach considered in this [28] Sun RC, Tomkinson J. Carbohydr Polym 2002;50:263.
study for obtaining high purity of cellulose by treatment [29] Kempf AW, Dence CW. Tappi 1970;53:864.
with 80% acetic acide70% nitric acid mixture at 120 (C [30] Simson B, Ayers J, Schwab G, Galley M, Dence CW. Tappi
1978;61:41.
for 20 min. This method described here may be also [31] Singh A. J Pulp Pap Sci 1990;16:J48.
considered as an improvement upon other published [32] Timell TE. Wood Sci Technol 1967;1:45.
methods for isolation of high pure cellulose that is [33] Dahlman O, Jacobs A, Sjoberg J. Cellulose 2003;10:325.
 J.X. Sun et al. / Polymer Degradation and Stability 84 (2004) 331e339 339

[34] Gubiz GM, Stebbing DW, Johansson CI, Saddler JN. Appl [41] Ristolainen M, Alen R, Malkavaara P, Pere J. Holzforschung
Microbiol Biotechnol 1998;50:390. 2002;56:513.
[35] Tenanen M, Tamminen T, Hortling B. Appl Microbiol Biotechnol [42] Pastarova I, Botto RR, Arisz PW, Boon JJ. Carbohydr Res
1999;51:241. 1994;262:27.
[36] Sun RC, Sun XF, Tomkinson J. ACS Symp Ser 2003;864:2. [43] Pappas C, Tarantilis PA, Daliani I, Mavromoustakos T, Polissiou
[37] Togrul H, Arslan N. Carbohydr Polym 2003;54:63. M. Ultrason Sonochem 1999;5:163.
[38] Bradrup J, Immergut EH. Polymer handbook. New York: Wiley; [44] Sun RC, Fang JM, Tomkinson J, Jones GL. Ind Crops Prod
1975. 1999;10:209.
[39] Hirrien M, Desbrieres J, Rinaudo M. Carbohydr Polym [45] Vanderhart DL, Atalla RH. Macromolecules 1984;17:1465.
1996;31:243. [46] Wickholm K, Larsson PT, Iversen T. Carbohydr Res 1998;
[40] Johnston HK, Sourirajan S. J Appl Polym Sci 1973;17:3717. 312:123.