On The Polymorphic and Morphological Changes of Cellulose Nanocrystals (CNC-I) Upon Mercerization and Conversion To CNC-II

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Carbohydrate Polymers 143 (2016) 327–335

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

On the polymorphic and morphological changes of cellulose


nanocrystals (CNC-I) upon mercerization and conversion to CNC-II
Ersuo Jin a , Jiaqi Guo b , Fang Yang a , Yangyang Zhu a , Junlong Song a,b,∗ , Yongcan Jin a ,
Orlando J. Rojas b,∗∗
a
Jiangsu Provincial Key Lab. of Pulp and Paper Science and Technology, Nanjing Forestry University, Nanjing 210037, Jiangsu, Peoples Republic of China
b
Bio-based Colloids and Materials, School of Chemical Technology, Aalto University, PO Box 16300, Aalto FIN-00076 Espoo, Finland

a r t i c l e i n f o a b s t r a c t

Article history: Polymorphic and morphological transformations of cellulosic materials are strongly associated to their
Received 8 November 2015 properties and applications, especially in the case of emerging nanocelluloses. Related changes that take
Received in revised form 9 January 2016 place upon treatment of cellulose nanocrystals (CNC) in alkaline conditions are studied here by XRD, TEM,
Accepted 22 January 2016
AFM, and other techniques. The results indicate polymorphic transformation of CNC proceeds gradually
Available online 25 January 2016
in a certain range of alkali concentrations, i.e. from about 8% to 12.5% NaOH. In such transition alkali
concentration, cellulose I and II allomorphs coexists. Such value and range of the transition concentration
Keywords:
is strongly interdependent with the crystallite size of CNCs. In addition, it is distinctively lower than that
Cellulose nanocrystals
CNC
for macroscopic fibers (12–15% NaOH). Transmission electron microscopy and particle sizing reveals that
Mercerization after mercerization CNCs tend to associate. Furthermore, TEMPO-oxidized mercerized CNC reveals the
Polymorphs morphology of individual nanocrystal of the cellulose II type, which is composed of some interconnected
Cellulose I and II granular structures. Overall, this work reveals how the polymorphism and morphology of individual CNC
Crystallinity change in alkali conditions and sheds light onto the polymorphic transition from cellulose I to II.
Crystallite size © 2016 Elsevier Ltd. All rights reserved.

1. Introduction (cellulose I−IV), the structure of which depends on the cellulose


source, method of extraction or treatment. Among them, cellulose
Cellulose, composed of ␤-1,4 linked glucopyranose units, is a polymorphs I and II are the most widely discussed. In nature, cel-
major resource for materials, (bio)chemicals and biofuels (Mosier lulose I is produced into crystalline forms with the Iˇ allomorph
et al., 2005; O’Sullivan, 1997; Wyman, Dale, Elander, Holtzapple, being the most common in plants, while I˛ in algae and bacteria
Ladisch, & Lee, 2005). In most cases, it is associated by hydro- (O’Sullivan, 1997). Cellulose II is usually obtained from cellulose I
gen bonding and other weak forces to form a semi-crystalline by mercerization (alkali treatment) or regeneration (solubilization
structure where highly ordered regions (the crystallites) are com- and recrystallization); there are a few reports suggesting the pro-
bined with less ordered (disorder or para-crystalline) domains, also duction of cellulose II from Gluconacetobacter xylinus (Kuga, Takagi,
known as the amorphous phase. The molecular orientation and the & Brown, 1993). Cellulose Iˇ has unit cell dimensions of a = 7.78 Å,
hydrogen-bond network that pack the cellulose chains within the b = 8.20 Å, c = 10.38 Å, ˛ = ˇ = 90◦ , and  = 96.5◦ (Nishiyama, Langan,
crystallites can vary widely, giving rise to different polymorphs & Chanzy, 2002). Cellulose II produced by the mercerization has
unit cell dimensions of a = 8.10 Å, b = 9.08 Å, c = 10.36 Å, ˛ = ˇ = 90◦ ,
and  = 117.3◦ (Gessler et al., 1995; Kobayashi, Kimura, Togawa,
& Wada, 2011; Kolpak & Blackwell, 1976; Langan, Nishiyama, &
Abbreviations: CNC, cellulose nanocrystal; CNC-I, cellulose nanocrystal, allo- Chanzy, 1999, 2001; Raymond, Kvick, & Chanzy, 1995; Sarko &
morph of cellulose I; CNC-II, cellulose nanocrystal, allomorph of cellulose II;
Muggli, 1974; Stipanovic & Sarko, 1976). There is a general con-
CNC-concentration%, cellulose nanocrystals treated under alkaline conditions at
given concentration (%) of NaOH solution. sensus for the parallel chain packing for cellulose I, while the chain
∗ Corresponding author at: Nanjing Forestry University, Jiangsu Provincial directionality of cellulose II is antiparallel (Lee et al., 2013). A pos-
Key Lab. of Pulp and Paper Science and Technology, Nanjing 210037, China. sible route how antiparallel chains of cellulose II produced from
Tel.: +86 25 85428163; fax: +86 25 85428689 (J.S.). cellulose I which has parallel chains was proposed by Yamane,
∗∗ Corresponding author at: Aalto University (O.J.R.). Tel.: +358 50 5124227;
Miyamoto, Hayakawa, and Ueda (2013) Cellulose I and cellulose
fax: +358 50 5124227.
E-mail addresses: junlong.song@njfu.edu.cn (J. Song), orlando.rojas@aalto.fi II possess different properties and advantages over the other. Com-
(O.J. Rojas). pared to cellulose II, cellulose I exhibits much better mechanical

http://dx.doi.org/10.1016/j.carbpol.2016.01.048
0144-8617/© 2016 Elsevier Ltd. All rights reserved.
328 E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335

properties, while the first offer benefits in terms of functionality


(Diddens, Murphy, Krisch, & Muller, 2008; Wang, Li, Yano & Abe,
2014).
Cellulose II, as indicated before, can be produced by alkali treat-
ment of cellulose I or by regeneration from cellulose solution;
however, cellulose II can be obtained from native cellulose through
other approaches as well. For example, Ago, Endo, and Hirotsu
(2004) demonstrated that a mechano-chemical treatment of native
cellulose with a given amount of water (∼30 wt% as present in the
cellulose solid state) via ball milling for several hours, caused the
crystalline transformation to cellulose II polymorph, even though
the crystallinity index of cellulose II was quite low, ranging from 30
to 37%. Sèbe, Ham-Pichavant, Ibarboure, Koffi, and Tingaut (2012)
used a combination of extended treatment time (>50 min) and high
final acid concentration (>64% H2 SO4 ) to obtain cellulose nanocrys-
tals with some cellulose II. They also found that pure cellulose II
nanowhiskers are obtained only in a narrow range of conditions
(66% H2 SO4 /60 min addition). Martins et al. (2015) also produced
cellulose II crystals through sulfuric hydrolysis (under a concentra-
tion of acid not reported). In principle, sulfuric acid is a swelling Fig. 1. Schematic diagram illustrating the general process to obtaining individual
cellulose nanocrystals with dominant polymorphs I or II.
agent for cellulose and therefore recrystallization of cellulose may
have occurred upon subsequent removal of the swelling agent or
during the hydrolysis process. (Sèbe et al., 2012). that CNCs obtained by sulfuric acid hydrolysis undergo complete
Cellulose nanocrystals (CNC) display a high specific strength, polymorphic transformation by reaction with 1% NaOH solution in
modulus and aspect ratio which are main reasons for their effect in 3 s. Liu and Hu (2008) treated bamboo nanofibers with 12% NaOH
improving the mechanical properties of synthetic polymers at low solution for 20 min to generate cellulose II with some residual cel-
loading levels (Habibi, Lucia, & Rojas, 2010). Benefits derived from lulose I while full conversion to cellulose II was achieved at 16%
the utilization of CNCs are being reported in fields of material sci- NaOH concentration. Abe and Yano (2011) treated wood cellulose
ence, electronics, catalysis and biomedicine (Gatenholm & Klemm, nanofibers and full conversion of cellulose I into II was accom-
2010). In most cases, CNCs are obtained by sulfuric acid hydrolysis plished in 15% NaOH solution. Interestingly, the reported values
of native cellulose, which results in purified CNCs. The properties of of NaOH concentration for mercerization of nanocellulose are very
the nanocrystals (e.g., shape, length, and diameter) depend on the close to those needed for polymorphic transformation of macro-
source of the cellulose and the degradation process (e.g., time and scopic fibers.
temperature for acid hydrolysis, high-pressure homogenization So far, investigations dealing with mercerization of cellulose
conditions, etc.). Thus, the properties of CNC-based materials are have only been focused on whole cellulose fibers, i.e., organized
closely related to CNC pretreatment method and resultant nanopar- assemblies of microfibrils. Very few investigations have dealt with
ticle morphology (Liu, Song, Anderson, Chang, & Hua, 2012; Lu & the properties of cellulose I and II nanocrystals (Yue et al., 2012).
Hsieh, 2010). Thus, it is necessary to conduct a systematic study on the trans-
Despite the fact that Mercer introduced the “mercerization” formation of individual CNCs, especially their polymorphic and
transformation in 1850, the unraveling of its mechanism remains morphological changes. Our results show that there are effects
a question of intense debate. As for the macroscopic state of cel- induced by the nano-scale size of the CNCs, which do not hold
lulose, i.e. cellulosic fibers, the general agreement indicates that in the case of macroscopic fibers. The polymorphic transforma-
at room temperature the polymorphic transformation starts at an tion of CNCs upon mercerization proceeds gradually, starting below
alkali concentration of 12.5% and is completed at a concentration 8% NaOH concentration and is complete above 12.5% NaOH. This
of 17.5% (Halonen, Larsson, & Iversen, 2013; Purz & Fink, 2003; transition range also depends on the initial cellulose source. CNCs
Revol, Dietrich, & Goring, 2011). However, if the temperature or produced by sulfuric acid hydrolysis contain negatively-charged
other conditions are changed, the polymorphic transformation may groups (anionic sulfate half-ester groups) that can be removed
occur at different alkali concentrations. For instance, the transi- during mercerization under alkaline conditions. Additionally, indi-
tion occurs at 14–15% NaOH concentration for the mercerization vidual CNCs are difficult to observe (TEM, AFM) but appear to form
of jute fibers at 85 ◦ C and 4 h (Yu et al., 2014). Polymorphic trans- aggregates. Thus, moderate TEMPO-oxidation of the mercerized
formation from cellulose I to cellulose II for cotton linter by alkali cellulose crystals (schematic illustration in Fig. 1) is required for
pretreatment and urea additive was investigated by Gupta, Uniyal, better dispersion and observation.
and Naithani (2013). Their results showed a sudden polymorphic
transformation at a concentration of 15 wt% NaOH, regardless of 2. Experimental
the presence of urea. In two-phase solvent systems, such as the
mercerization experiments conducted with four different solution 2.1. Materials
systems (ethanol/water, acetone, dimethyl sulphoxide and xylene),
structural changes of cellulose crystallites depended primarily on Two microcrystalline cellulose grades were used in this study,
the distribution and solubility of sodium hydroxide in the sol- from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China),
vent. However, there was a critical concentration in the hydrophilic thereafter referred to as MCC1 and, AvicelTM PH-101, provided
phase that must exceed 7–8% NaOH before cellulose I undergoes by Sigma-Aldrich Co., Ltd (Shanghai, China), referred to as MCC2.
complete crystal change to cellulose II (Mansikkamaki, Lahtinen 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO), 99% (GC) purity
and Rissanen, 2005). was purchased from Jiana Chemical Co., Ltd (Changzhou, China).
The understanding of the polymorphic transformation of cel- Sodium hypochlorite, sodium hydroxide, sulfuric acid and other
lulose in the nanoscale, i.e. nanocellulose crystals, has found some reagents were analytical grade and purchased from Nanjing Chem-
controversies in the literature. Li, Ding, Li, and Jiang (2002) reported ical Reagent Co., Ltd (Nanjing, China).
E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335 329

2.2. Preparation of cellulose nanocrystals (CNCs) signal. X-ray diffraction data were collected from 2 = 5–55◦ in steps
of 0.02◦ ; and were further analyzed using the MDI Jade 6.5 software
The procedure for preparation of CNCs is described elsewhere (Materials Data, Inc.). XRD patterns were smoothed and baseline-
(Beck-Candanedo, Roman, & Gray, 2005; Fu, He, Jin, Cheng, & Song, corrected by a software-generated parabolic curve. Curve-fitting
2012; Song, Fu, Cheng, & Jin, 2014; Viet, Beck-Candanedo, & Gray, was performed to identify individual peaks and to calculate the
2007). Briefly, MCC powder (10 g) was hydrolyzed with sulfuric crystallinity and crystallite size. The Segal method (Creely, Segal, &
acid (250 mL, 64 wt%) at 45 ◦ C for 1 h under continuous stirring. Ziifle, 1956) uses the ratio of the peak height at the 2 = 18◦ position,
The hydrolysis was quenched by adding a large amount of water which is assumed to result from the contribution of amorphous
(∼10 times the volume of the acid solution used). The system was regions of cellulose, to that of plane 200. This method is usually
centrifuged (12,000 rpm) at 4 ◦ C for 10 min and the supernatant applied to cellulose I but not for cellulose II. Here, we used the
was decanted. This procedure for acid removal was repeated three areas of the crystalline and amorphous regions in the XRD spectra.
times until a white suspension was obtained. Then the suspen- Therefore, the crystallinity based on the ratio of crystalline region’s
sion was dialyzed with cellulose acetate membrane tubes (MWCO area to the total area was employed for convenient comparison.
14000) against deionized water, for about 5–7 days, until a constant The Scherrer’s equation was used for estimating the crystallite size
pH was reached. A portion of the CNC aqueous dispersion (there- (Mittal, Katahira, Himmel, & Johnson, 2011):
after referred to as CNC-I) was stored in a bottle and the other was
freeze-dried. k
ˇ=
 cos 
2.3. Mercerization of CNC-I where  is the wavelength of the incident X-ray (0.15406 nm),  is
the Bragg angles corresponding to the planes, ˇ is the full-width
10 mL of NaOH solutions (1.0, 2.0, 5.0, 8.0, 10, 12.5, 15 and 17.5% at half maximum (FWHM) of the X-ray peak corresponding to the
concentration) were prepared initially. Then 100 mg precisely planes;  is the X-ray crystallite size, and k is a constant of 0.89.
weighted CNC was added into the alkali solution under constant
stirring for 30 min at ambient temperature. Upon completion of the
2.6. CNC imaging and sizing
reaction, the dispersion was neutralized with 5% sulfuric acid solu-
tion. The respective mercerized CNC dispersion was dialyzed with
A drop of aqueous CNC dispersion in water (∼0.01%w/v) was
cellulose acetate membrane tubes (MWCO14000) against deion-
deposited on a carbon-coated electron microscope grid and then
ized water for 3–5 days, until a constant conductivity was reached.
negatively stained with phosphotungstic acid. A piece of tissue
The obtained samples, herein indicated by a numeral representing
paper was used to absorb excess water and then allowed to dry
the NaOH concentration used during mercerization (e.g. CNC-10%),
in air. The sample on the grids was imaged with a JEM-2100
were freeze-dried and stored in a desiccator. For comparison, the
UHR (JEOL, Japan) transmission electron microscope operated at
other MCC sample, MCC1, was mercerized with 17.5% NaOH for 24 h
an accelerating voltage of 200 kV.
at first, and then subjected to sulfuric acid hydrolysis as described in
AFM observation was performed using Multimode Nanoscope
the previous section (a shorter time was used, 20 min). This sample
V controller (Bruker Corporation, USA) in air at room temperature
is thereafter denoted as CNC-II.
(23 ◦ C). A drop with respective CNC dispersion was placed on a sil-
icon wafer and spin-dried at a speed of 3000 rev/min using a spin
2.4. TEMPO-mediated oxidation of mercerized CNCs
coater (WS-650-23, Laurell Technology Co. Ltd., USA). Images from
AFM were recorded in tapping mode by using silicon cantilevers.
Negative charges on CNC-I are reduced to some degree after
The commercial AFM probes had a spring constant of 20–80 N/m
treatment with alkali, due to hydrolysis of ether bonds (Lokanathan,
and a resonance frequency of 300–340 kHz.
Uddin, Rojas, & Laine, 2014). As a consequence, CNCs tend to
The relative size of mercerized CNC aggregates as well as CNC-
agglomerate after mercerization, limiting the possibility of detailed
II was determined by dynamic light scattering using a Malvern
studies, for example, via imaging. Therefore mercerized CNCs were
Zetasizer NanoS90 at a 90◦ scattering angle.
subjected to TEMPO-mediated oxidation, which adds negative
charges to the surface and improves colloidal stabilization. The
TEMPO-mediated oxidation followed the method described else- 3. Results and discussion
where (Habibi, Chanzy, & Vignon, 2006) except that more moderate
conditions were used. Briefly, the CNC-17.5% sample (50 mL dis- 3.1. Untreated cellulose nanocrystals
persion) was first sonicated for 5 min. Then TEMPO (14.75 mg,
0.094 mmol) and NaBr (162 mg, 1.57 mmol) were added to the dis- Two untreated cellulose nanocrystal samples (CNC-I1 and
persion. NaClO solution (0.4 mL, 1.24 M) was added slowly and the CNC-I2) were obtained by sulfuric acid hydrolysis from two micro-
pH of the mixture was maintained at 10 by adding 0.5 M NaOH. crystalline cellulose sources, MCC1 and MCC2, respectively. In
The oxidation was terminated after 30 min by adding methanol Fig. 2a, both XRD patterns of CNC-I1 and CNC-I2 exhibit a sharp
(1 mL) and the pH was adjusted to 7 using 0.5 M HCl solution. The high peak at 2 = 22.6◦ and two overlapping, weaker diffraction
aqueous dispersion was then introduced in cellulose acetate mem- peaks at 2 = 14.7◦ and 2 = 16.4◦ , which are assigned to the typ-
brane tubes (MWCO14000) for dialysis against deionized water. ical pattern of cellulose I (French, 2014; French & Cintrón, 2013).
The obtained sample, after freeze–drying, is referred thereafter as The d-spacings of these planes calculated by Bragg equation are
CNC-17.5%-TEMPO. 0.39 nm, 0.60 nm and 0.54 nm, respectively. These parameters are
consistent with those reported in the literature (Ishikawa, Okano,
2.5. X-ray diffraction Sugiyama, Ishikawa, & Okano, 1997; Sèbe et al., 2012). However,
compared to the diffraction bands of CNC-I1, the pattern for CNC-
X-ray diffraction measurements were carried out by using a I2 is smoother and overlaps more markedly. Two diffraction peaks
Goniometer Ultima IV (Rigaku Co. Ltd., Japan) X-ray diffractome- at 14.7◦ and 16.4◦ merge into one broad band and the small peak at
ter equipped with CuK␣ radiation (0.15406 nm) generated at 40 kV 2 = 20.1 (plane 012) is integrated into the most pronounced peak
and 30 mA at room temperature. Vacuum dried cellulose powder of plane 200. The crystallinity index (CrI) of CNC-I1 and CNC-I2 are
samples were placed on quartz substrates that have no background quite similar, 77.9% and 77.5%, respectively.
330 E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335

Fig. 2. X-ray diffraction patterns of CNC-I1 and CNC-I2 (a) and a representative TEM image of CNC-I1 (b) (the images for CNC-I2 are very similar, not included here).

The morphology of CNC-I1 and CNC-I2 as observed by TEM two noticeable peaks at 2 = 14.7 and 16.4, assigned to cellulose I,
is quite similar (Fig. 2b). The crystals show a needle-like shape, still remain. Samples subjected to NaOH treatment at concentra-
100–300 nm in length and ca. 10 nm in width, consistent with other tions above 12.5% (i.e. 15 and 17.5%) indicate similar XRD patterns
values reported in the literature for MCC-derived CNCs (Habibi and display only the characteristic profile of cellulose II, with three
et al., 2010). peaks positioned at 2 = 12.0, 20.0 and 21.8o .
The alkali-treated CNC-I2 samples produce similar trends as
3.2. CNC polymorphic transformation from cellulose I into those discussed before for CNC-I1, except that the transition con-
cellulose II and size-induced polymorphism centration is shifted from 10% to 8% (CNC-I1).
Crystallinity index, d-spacing and crystallize size for each
A systematic evaluation of the process by which cellulose I plane were determined via Jade code. The d-spacings of the
nanocrystals evolve into cellulose II crystals was undertaken and (1 −1 0), (1 1 0) and (2 0 0) planes for cellulose I are 6.01 ± 0.06 Å,
the role of size-induced polymorphic transformation upon mercer- 5.35 ± 0.05 Å and 3.92 ± 0.02 Å, respectively. While these values are
ization was assessed. To this end, the two CNC samples (CNC-I1 and shifted to 7.41 ± 0.06 Å, 4.43 ± 0.03 Å and 4.08 ± 0.04 Å for cellulose
CNC-I2) were treated with NaOH solutions of given concentrations, II. The crystallite size for the (1 −1 0) and (1 1 0) planes of cellulose I,
from 1% to 17.5%, for 30 min and at room temperature. The X-ray and (1 1 0) and (2 0 0) for cellulose II varied extensively due to over-
diffraction patterns of CNC-I1 and CNC-I2 after alkali treatment are lap. However the values of the average crystallize size of the most
shown in Fig. 3a and b, respectively. prominent peak(s), i.e., 2 = 22.6◦ for cellulose I and 2 = 20.0 and
The XRD patterns of CNC-I1 upon treatment with dilute alkali, 21.8◦ for cellulose II, can be considered accurate. The crystallinity
e.g. 1, 2, 5, and 8% NaOH solutions display similar peaks as those indices and average crystallite sizes of CNC-I1 and CNC-I2 after
of untreated CNC and indicate the main contribution of cellulose treatment with alkali solutions of given concentrations are listed
I (French, 2014). However, for the sample treated by 8% NaOH, in Tables 1 and 2, respectively.
a small bump at the position of 2 = 12.0 starts emerging. When In the case of CNC-I1, CrI continuously decreases with the
the alkali concentration is increased to 10%, two additional distinc- increase of alkali concentration used in the treatment. The frac-
tive diffraction peaks appear at 2 = 12.0 and 2 = 20.0, which are tion of cellulose I and cellulose II can be estimated by the area of
assigned to cellulose II (French, 2014). This observation hints to fitted curves. At 10% NaOH concentration, the fraction of Cellulose
the fact that at concentration of ca. 10% NaOH, both polymorphs, I allomorph is 72.8%, indicating the presence of 27.2% cellulose II
cellulose I and cellulose II, coexist. When the alkali concentration allomorph. In the case of 8% NaOH, cellulose I dominates (84.0%)
is increased to 12.5%, three peaks from cellulose II dominate but and only a small portion (16.0%) of cellulose II allomorph is

Fig. 3. XRD patterns after mercerization (30 min, room temperature) of CNC-I1 (a) and CNC-I2 (b) under given concentrations of NaOH solution, from 1 to 17.5%.
E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335 331

Table 1
Crystallinity indices and crystallite size of CNC-I1 after alkali treatment for 0.5 h.

NaOH Conc. (%) CrI (%) Crystallite size (nm)

Cellulose I Cellulose II

MCC1 83.7 6.9 –


CNC-I1 77.9 6.6 –
1 76.5 6.8 –
2 75.9 6.8 –
5 73.8 6.8 –
8 71.4 6.5 5.2
10 66.6 6.3 5.3
12.5 68.2 5.1 5.5
15 68.2 – 5.5
17.5 56.7 – 5.5

Table 2
Crystallinity indices and crystallite size of CNC-I2 after alkali treatment for 0.5 h.
Fig. 4. XRD patterns of CNC-I1 treated with alkali at the transition concentration
NaOH Conc. (%) CrI (%) Average crystallite size (nm) (10%) for given periods of time.
Cellulose I Cellulose II

MCC2 69.9 4.4 – is slightly higher than that of the produced CNCs. This is explained
CNC-I2 77.5 4.1 – by the fact that in the production of CNC some molecules located
1 78.9 3.9 –
in the periphery of the crystals react with sulfuric acid and become
2 81.2 3.9 –
5 73.1 3.9 – dissolved. The reason why two different MCC sources have differ-
8 72.1 2.8 4.5 ent crystallite sizes may be due to the difference in fiber source and
10 70.8 – 5.0 treatments applied in their production. This issue needs further
12.5 67.7 – 4.8
elucidation.
15 59.2 – 5.0
17.5 55.3 – 4.9
3.3. Time for CNC polymorphic transformation

identified. For 12.5% NaOH, cellulose II dominates (82.4%) and a The XRD pattern of CNC pretreated at the transition NaOH con-
small portion (17.6%) of cellulose I allomorph is identified. These centration includes all the peaks of cellulose I and cellulose II. This
observations hint that the polymorphic transformation proceeds facilitates determination of the extent of completion of the poly-
gradually, starting from below 8% NaOH and ending at a NaOH morphic transformation, i.e., via monitoring the change in XRD
concentration above 12.5%. patterns with the pretreatment time. As such, Fig. 4 includes XRD
This gradual transition is also evidenced in the average crys- patterns of CNC-I treated with 10% NaOH during given times, up to
tallize size. The average crystallize size for MCC1 is 6.9 nm. This 24 h. In all cases, from 0.5 h and up to 24 h treatment, the same dis-
size is reduced to 6.6 nm for CNC-I1. When CNC-I1 is treated in tinguishable features in the XRD profiles are observed, indicating
dilute alkali, the size remains similar. When the alkali concentration that an extensive treatment is not conducive of further polymor-
increased to 8%, the average crystallize size for cellulose I crys- phic transformation. The polymorphic transformation occurs quite
tallites are reduced slightly, to 6.5 nm. When the concentrations rapidly, in less than 30 min, regardless of the concentration of alkali
increased to 10% and 12.5%, the average crystallize size for cellu- used. This result is in agreement with the conclusion of Liu and Hu
lose I crystallites is reduced further to 6.3 and 5.1 nm, respectively. (2008).
In the meantime, the average size of cellulose II crystallites grows
from 5.2 nm to 5.5 nm. At an alkali concentration above 12.5%, the 3.4. Morphological changes of cellulose nanocrystals during
average size of cellulose II crystallites doesn’t change further. polymorphic transformation
The CrI variation upon pretreatment of CNC-I2 with alkali fol-
lows the same trend as that for CNC-I1, i.e. CrI continuously An AFM image of CNC-I1 after treatment with 10% NaOH is
decreases with the increase in alkali concentration used in the shown in Fig. 5a. Needle-like shapes are observed, with ca. 10 nm
treatment. At 8% NaOH, the CNC mixture comprises 68.0% Cellu- in width and 100–200 nm in length. Negligible differences in size
lose I and 32.0% Cellulose II crystallites. The fraction of cellulose II and morphology are observed if compared to the untreated CNC-I.
is higher than that in the case of CNC-I1 for treatment with the same CNC-I1 pretreated with 17.5% NaOH resulted in important mor-
alkali concentration. Furthermore, negligible amounts of cellulose phological changes as observed by AFM and TEM (Fig. 5b and c),
I or II are found at 5% and 10% NaOH. respectively. Round-shaped agglomerates, about 200 nm in size,
In terms of the average crystallite size of CNC-I2, it also fol- composed of granules (size of around 30 nm), can be observed in
lows the same trend as that of CNC-I1, i.e. the average crystallite Fig. 5b; no individual crystals are identified, even if the dispersion
size for cellulose I gradually reduces at the transition alkali con- was diluted before drying and imaging. Such agglomerates can be
centration, while the average crystallite size for cellulose II grows. clearly seen under TEM (Fig. 5c) as uniform spherical aggregates
A very interesting is that the average crystallite size of cellulose of ca. 150–200 nm. The size of granules for mercerized CNC is also
I in CNC-I2, 4.0 nm, is much smaller than that of CNC-I1, 6.8 nm. consistent with that of regenerated cellulose (Yamane et al., 2013).
The difference in crystallize size may be one of the reasons why the Untreated CNCs produced by sulfuric acid hydrolysis carry neg-
polymorphic transformation for CNC-I2 takes place at a lower alkali atively charged groups and are dispersed easily in aqueous medium
concentration and is completed within a narrow range. The differ- (Habibi et al., 2010). The ester bonds of negatively charged sulfate-
ence in crystallize size for both CNC samples can be traced back to half ester groups break down, at least partially, upon treatment in
the MCC sources. The crystallite size in MCC1 is 6.9 nm, while that an alkaline solution (Lokanathan et al., 2014). Therefore, the surface
in MCC2 is only 4.4 nm. The crystallite size for both MCC samples charge of CNCs is reduced and facilitates their agglomeration. This
332 E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335

Fig. 5. AFM image of CNC-I1 upon treatment at the critical concentration of alkali, 10% (a) and corresponding images (AFM, b and TEM, C) upon CNC-I1 treatment at higher
concentration (17.5% NaOH).

solution is somewhat similar. For CNC-17.5%, the size shown


indicates large aggregates. As a comparison, the size of CNC-II
(hydrolysis product of mercerized MCC), is shown to be 20-30 nm
(see Fig. 7a), which is in agreement with the granule size in AFM
observations for CNC-17.5% (Fig. 5b).
Since CNCs loss their charges after mercerization and tend to
agglomerate, TEMPO-mediated oxidation was carried out in order
to facilitate dispersion and to observe the morphology of individual
particles. Oxidation strength was carefully adjusted by the molar
ratio of NaClO/equivalent anhydroglucose unit (AGU). In a conven-
tional TEMPO-mediated oxidation, this value is about 0.5, which is
too severe for the intended purpose. After several tests, it was found
that NaClO/AGU = 0.15 is the minimum necessary to promote col-
loidal stabilization of the CNCs, i.e., by oxidation of some of its C6
hydroxyl groups. A TEM image of CNC-17.5%-TEMPO is shown in
Fig. 7b. The morphology of individual mercerized cellulose crystal
Fig. 6. Size of CNC and corresponding agglomerates as determined by light scatter- indicates interconnected spherical features and bended kidney-like
ing in aqueous dispersions of nanocrystals (CNC-I1) before and after treatment at
granules. The size of these granules is consistent with that observed
given alkali concentrations. The image of CNC-II is also included.
in AFM (Figs. 5b and 7a), ca. 30–40 nm.
To the best of our knowledge, this is the first report on the
tendency is clearly demonstrated in Fig. 6 for size measurements morphology of individual CNC with cellulose II polymorph, espe-
that were carried out in aqueous dispersion via light scattering. cially obtained from the mercerization. Sèbe et al. (2012) have
Bearing in mind that the size recorded can only be taken as a imaged individual cellulose nanocrystals with cellulose II allo-
relative value and cannot be compared with that determined by morph obtained from acid regeneration. Compared with that of
TEM, it is still interesting to note similar observations: the diame- cellulose I nanowhiskers, their morphology (AFM images) showed
ter of untreated CNC-I1 and that after treatment with 10% NaOH a ribbon-like shape with rounded tips, with a smaller size but
E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335 333

Fig. 7. AFM image of CNC-II (a) and TEM image of CNC-17.5%-TEMPO (b).

expanded width. The morphology shown in AFM images demon- of mercerized CNCs, 5.5 nm. While the average crystallite size of
strate that such ribbons were composed of several granules. Hence TEMPO-oxidized sample (CNC-17.5%-TEMPO), is only 4.2 nm. This
the morphology includes several interconnected granules shared observation indicates that the crystallite size of CNC-II and CNC-
by both individual nanocellulose samples, regardless the treat- 17.5%-TEMPO is reduced in their production.
ment process used, either by concentrated acid or mercerization; All the above observations provide some evidence and shed
they undergo swelling and recrystallization. Yue et al. (2012) have some light to better understanding how the polymorphic trans-
reported rod-like crystals for cellulose II having similar widths but formation of cellulose takes place. For a regular cellulose I crystal,
reduced length, compared with pristine cellulose I crystals. How- with a width of several nanometers and a length of hundreds
ever these authors obtained cellulose II crystals from mercerized nanometers, and at a NaOH concentration above the critical value,
macroscopic fibers, via sulfuric acid hydrolysis and followed by hydroxide ions start to penetrate into the crystals. As hydroxide
high-pressure homogenization. This means that the polymorphic ions diffuse within the crystallite and swells it, the cellulose chains
transformation occurred at the macroscopic state and the mor- rearrange their orientation as hydroxide ions are removed from the
phological changes may have been restricted by the surrounding lattice of the cellulose matrix. This process takes place as suggested
crystals and were not able to relax, which is in contrast to the by Yamane et al. (2013) who indicated that some six ring conform-
present observations. ers of glucose form hair-pin turns and therefore cellulose chains
The XRD patterns of CNC-17.5%-TEMPO, and CNC-II are shown fold and pack antiparallel forming a more thermodynamically sta-
in Fig. 8. The XRD pattern for CNC-17.5% is added for comparison. ble state, allomorph cellulose II. As such, the cellulose II crystallites
They all include the characteristic peaks of cellulose II but with formed have much shorter length, ca. 30–40 nm (see TEM images),
different intensity. After TEMPO-mediated oxidation, mercerized if compared with their original state, cellulose I, 100–300 nm. Our
CNCs still retain their crystalline structure. The crystallinity indices results also demonstrate that one cellulose chain may be involved
and crystallite size of CNC-17.5%-TEMPO and CNC-II are given in in the formation of several cellulose II crystallites. The “bridge”
Table 3. Since CNC-II contains mainly the crystalline component connecting granules provides this evidence. While images of indi-
(upon removal of amorphous regions), its crystallinity, 75.4%, is vidual CNC composed of the cellulose II polymorph can be obtained,
close to the original CNC samples, and higher than that of mercer- the process by which cellulose I evolves into cellulose II, and the
ized CNC (CNC-17.5%), 56.7%. This value is much higher than that variables that influence this process, remain unresolved subjects.
of TEMPO-oxidized sample (CNC-17.5%-TEMPO), 37.6%. The aver- Computational simulation and related approaches may shed more
age crystallite size of CNC-II, 5.1 nm, is slightly smaller than those light on these issues.
Overall, the results obtained suggest that: (1) cellulose I allo-
morph cannot be converted under dilute alkali treatment (<5%
NaOH); (2) polymorphic transformation proceeds gradually, in the
alkali transition concentration range. In the transition state, both
allomorphs, cellulose I and II, coexist in the cellulose nanocrystals.
The ratio of each fraction depends only on the alkali concentra-
tion used in the treatment. Below this transition, cellulose I clearly
dominates the system while above the transition concentration,
cellulose II dominates. (3) If compared to the polymorphic trans-
formation of macroscopic fibers, where the polymorph conversion
only occurs when alkali concentration is above 12.5%, the critical

Table 3
Crystallinity indices and crystallite size of CNC-17.5%-TEMPO and CNC-II.

Sample Polymorph CrI (%) Average crystallite size (nm)

CNC-17.5%-TEMPO II 37.6 4.2


CNC-II II 75.4 5.1
Fig. 8. XRD patterns for cellulose nanocrystals with polymorph of cellulose II.
334 E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335

transition concentration for CNCs is reduced, to about 8% NaOH. The nano-size induced polymorphic transformation plays a role in
This indicates a size-induced polymorphic transformation, which reducing the critical concentration required for allomorph conver-
is strongly interconnected with the CNC crystallite size. However, sion. Finally, morphological changes (TEM and AFM) of individual
this effect is not as pronounced as that claimed by Li et al. (2002) cellulose crystal during polymorphic transformation illustrate that
who reported complete conversion upon treatment with 1% NaOH CNC comprising the Cellulose II polymorph exhibits interconnected
solution for three seconds. This may be attributed to the con- granules. This indicates that single cellulose chains may be involved
centrated alkali that remains after mercerization during drying, in Cellulose II the formation of Cellulose II crystals.
which facilitates the polymorphic transformation. At the transition
NaOH solution concentration of 8–12.5%, and for the polymorphic
transformation of the studied CNCs, it is hypothesized that the Acknowledgements
hydrated hydroxide ions start to penetrate the cellulose crystals
and undergo partial reaction with cellulose. At alkali concentra- This project was financially supported by Special Fund for
tions above the transition concentration, hydroxide ions penetrate Forestry Scientific Research in the Public Interest (201404510),
extensively within the crystals, producing CNC swelling. Upon ion National Natural Science Foundation of China (Nos. 31270613,
removal during dialysis, the cellulose chains recrystallize into cellu- 31200454), Qing-Lan Project, Jiangsu Co-Innovation Center for Effi-
lose II, which is more stable thermodynamically. As a consequence, cient Processing and Utilization of Forest Resource (NJFU) and the
the respective XRD patterns display the contribution of cellulose II Priority Academic Program Development of Jiangsu Higher Edu-
allomorph while the characteristic peaks of native cellulose I are cation Institutions. JS acknowledges funding support from China
absent. Treatment with alkali at a concentration below the transi- Scholarship Council provided for his research mission in Aalto Uni-
tion value provides conditions that are not sufficient for hydroxide versity. This work was also partially supported by the Academy of
ions to penetrate into the cellulose I crystals and therefore no poly- Finland Centres of Excellence Programme (2014–2019) “Molecular
morphic transformation takes place, even if the size of the particle Engineering of Biosynthetic Hybrid Materials Research” (HYBER).
is in the nano-scale (Abe & Yano, 2011; Dinand, Vignon, Chanzy,
& Heux, 2002; Fengel, Jakob, & Strobel, 1995). Based on the dis-
cussion of Liu and Hu (2008), the size of the hydroxide ions vary References
with the alkali concentrations. At a low concentration, hydrox- Abe, K., & Yano, H. (2011). Formation of hydrogels from cellulose nanofibers.
ide ions are fully hydrated and their size may prevent them from Carbohydrate Polymers, 85(4), 733–737.
penetrating and disrupting the cellulose lattice. However, as the Ago, M., Endo, T., & Hirotsu, T. (2004). Crystalline transformation of native cellulose
from cellulose I to cellulose II polymorph by a ball-milling method with a
NaOH concentration increases, the amount of free water available
specific amount of water. Cellulose, 11(2), 163–167.
for hydrating the hydroxide ions is limited. As such, the charac- Beck-Candanedo, S., Roman, M., & Gray, D. G. (2005). Effect of reaction conditions
teristic size of hydrated hydroxide ions are relatively smaller and on the properties and behavior of wood cellulose nanocrystal suspensions.
more easily penetrate into the cellulose lattices. When their size is Biomacromolecules, 6(2), 1048–1054.
Creely, J. J., Segal, L., & Ziifle, H. M. (1956). Determination of the degree of
reduced to dimensions small enough to penetrate into the cellulose crystallite orientation in cotton fibers by means of the recording X-ray
lattices, the critical alkali concentration required for polymorphic diffraction spectrometer. Textile Research Journal, 26(10), 789–795.
transformation is identified. At the NaOH transition concentration, Diddens, I., Murphy, B., Krisch, M., & Muller, M. (2008). Anisotropic elastic
properties of cellulose measured using inelastic X-ray scattering.
some crystals of cellulose I are converted into cellulose II. The ratio Macromolecules, 41(24), 9755–9759.
of cellulose I to II is a function of alkali concentration used. There Dinand, E., Vignon, M., Chanzy, H., & Heux, L. (2002). Mercerization of primary wall
is indication that the polymorphic transformation of CNCs is actu- cellulose and its implication for the conversion of cellulose I → cellulose II.
Cellulose, 9(1), 7–18.
ally a dynamic process. Above the NaOH transition concentration, Fengel, D., Jakob, H., & Strobel, C. (1995). Influence of the alkali concentration on
hydroxide ions penetrate easily into all pristine crystals and convert the formation of cellulose II. Study by X-ray diffraction and FTIR spectroscopy.
them into crystals of cellulose II polymorph. Holzforschung-International Journal of the Biology, Chemistry, Physics and
Technology of Wood, 49(6), 505–511.
In terms of the penetration of hydrated hydroxide ions, as for
French, A. D. (2014). Idealized powder diffraction patterns for cellulose
CNC particles of inherently large surface area, the penetration is polymorphs. Cellulose, 21(2), 885–896.
facilitated. This the reason for size-induced polymorphic transfor- French, A. D., & Cintrón, M. S. (2013). Cellulose polymorphy, crystallite size, and
the segal crystallinity index. Cellulose, 20(1), 583–588.
mation. In regards to the CNC sample with a small crystallite size,
Fu, G., He, A., Jin, Y., Cheng, Q., & Song, J. (2012). Fabrication of hollow silica
hydroxide ions penetrate easily into the matrix and therefore shift nanorods using nanocrystalline cellulose as templates. Bioresources, 7(2),
the concentration required for polymorphic transformation to a 2319–2329.
lower value and within a narrower range. Gatenholm, P., & Klemm, D. (2010). Bacterial nanocellulose as a renewable
material for biomedical applications. Mrs Bulletin, 35(3), 208–213.
Gessler, K., Krauss, N., Steiner, T., Betzel, C., Sarko, A., & Saenger, W. (1995).
Beta.-d-cellotetraose hemihydrate as a structural model for cellulose II. An
4. Conclusions X-ray diffraction study. Journal of the American Chemical Society, 117(46),
11397–11406.
Gupta, P. K., Uniyal, V., & Naithani, S. (2013). Polymorphic transformation of
The polymorphic and morphological changes of cellulose cellulose I to cellulose II by alkali pretreatment and urea as an additive.
nanocrystals were systematically investigated by treating CNCs Carbohydrate Polymers, 94(2), 843–849.
Habibi, Y., Chanzy, H., & Vignon, M. R. (2006). TEMPO-mediated surface oxidation
under various concentrations of NaOH (1, 2, 5, 8, 10, 12.5, 15 and of cellulose whiskers. Cellulose, 13(6), 679–687.
17.5%). It is found that a transition range of alkali concentration Habibi, Y., Lucia, L. A., & Rojas, O. J. (2010). Cellulose nanocrystals: Chemistry,
exist for polymorphic transformation of cellulose and the mercer- self-assembly, and applications. Chemical Reviews, 110(6), 3479–3500.
Halonen, H., Larsson, P. T., & Iversen, T. (2013). Mercerized cellulose biocomposites:
ization process is completed in less than 30 min.
A study of influence of mercerization on cellulose supramolecular structure,
The transition NaOH concentration is found to be in the range water retention value and tensile properties. Cellulose, 20(1), 57–65.
of 8–12.5%, which strongly correlates with the CNC crystallite size. Ishikawa, A., Okano, T., Sugiyama, J., Ishikawa, A., & Okano, T. (1997). Fine structure
and tensile properties of ramie fibres in the crystalline form of cellulose I, II, IIII
Within this transition concentration range, both allomorphs of cel-
and IVI. Polymer, 38(2), 463–468.
lulose I and cellulose II coexists while below the polymorph of Kobayashi, K., Kimura, S., Togawa, E., & Wada, M. (2011). Crystal transition from
cellulose I dominates. Above this concentration range, cellulose I cellulose II hydrate to cellulose II. Carbohydrate Polymers, 86(2), 975–981.
converts to cellulose II completely. The polymorphic transforma- Kolpak, F. J., & Blackwell, J. (1976). Determination of the structure of cellulose II.
Macromolecules, 9(2), 273–278.
tion proceeds gradually, which is evidenced by the gradual decrease Kuga, S., Takagi, S., & Brown, R. M. (1993). Native folded-chain cellulose II. Polymer,
in crystallite size of cellulose I and the growth of cellulose II. 34(15), 3293–3297.
E. Jin et al. / Carbohydrate Polymers 143 (2016) 327–335 335

Langan, P., Nishiyama, Y., & Chanzy, H. (1999). A revised structure and O’Sullivan, A. C. (1997). Cellulose: The structure slowly unravels. Cellulose, 4(3),
hydrogen-bonding system in cellulose II from a neutron fiber diffraction 173–207.
analysis. Journal of the American Chemical Society, 121(43), 9940–9946. Purz, H. J., & Fink, H.-P. (2003). Zur Morphologie und Gitterumwandlung bei der
Langan, P., Nishiyama, Y., & Chanzy, H. (2001). X-ray structure of mercerized Alkalisierung von Cellulose. Acta Polymerica, 34(9), 546–558.
cellulose II at 1 a resolution. Biomacromolecules, 2(2), 410–416. Raymond, S., Kvick, A., & Chanzy, H. (1995). The structure of cellulose II: A revisit.
Lee, C. M., Mittal, A., Barnette, A. L., Kafle, K., Park, Y. B., Shin, H., et al. (2013). Macromolecules, 28(24), 8422–8425.
Cellulose polymorphism study with sum-frequency-generation (SFG) vibration Revol, J. F., Dietrich, A., & Goring, D. A. I. (2011). Effect of mercerization on the
spectroscopy: Identification of exocyclic CH2 OH conformation and chain crystallite size and crystallinity index in cellulose from different sources.
orientation. Cellulose, 20(3), 991–1000. Canadian Journal of Chemistry, 65(8), 1724–1725.
Li, G., Ding, E., Li, X., & Jiang, Y. (2002). Studies on the preparation and Sèbe, G., Ham-Pichavant, F., Ibarboure, E., Koffi, A. L. C., & Tingaut, P. (2012).
characterization of nano-crystal cellulose II. Journal of Cellulose Science and Supramolecular structure characterization of cellulose II nanowhiskers
Technology, 10(2), 12–19. produced by acid hydrolysis of cellulose I substrates. Biomacromolecules, 13(2),
Liu, D., Song, J., Anderson, D. P., Chang, P. R., & Hua, Y. (2012). Bamboo fiber and its 570–578.
reinforced composites: Structure and properties. Cellulose, 19(5), 1449–1480. Sarko, A., & Muggli, R. (1974). Packing analysis of carbohydrates and
Liu, Y. P., & Hu, H. (2008). X-ray diffraction study of bamboo fibers treated with polysaccharides. III. Valonia cellulose and cellulose II. Macromolecules, 7(4),
NaOH. Fibers and Polymers, 9(6), 735–739. 486–494.
Lokanathan, A. R., Uddin, K. M. A., Rojas, O. J., & Laine, J. (2014). Cellulose Song, J. L., Fu, G. S., Cheng, Q., & Jin, Y. C. (2014). Bimodal mesoporous silica
nanocrystal-mediated synthesis of silver nanoparticles: Role of sulfate groups nanotubes fabricated by dual templates of CTAB and bare nanocrystalline
in nucleation phenomena. Biomacromolecules, 15(1), 373–379. cellulose. Industrial & Engineering Chemistry Research, 53(2), 708–714.
Lu, P., & Hsieh, Y. L. (2010). Preparation and properties of cellulose nanocrystals: Stipanovic, A. J., & Sarko, A. (1976). Packing analysis of carbohydrates and
Rods, spheres, and network. Carbohydrate Polymers, 82(2), 329–336. polysaccharides. 6. Molecular and crystal structure of regenerated cellulose II.
Mansikkamaki, P., Lahtinen, M., & Rissanen, K. (2005). Structural changes of Macromolecules, (5), 851–857.
cellulose crystallites induced by mercerisation in different solvent systems; Viet, D., Beck-Candanedo, S., & Gray, D. G. (2007). Dispersion of cellulose
determined by powder X-ray diffraction method. Cellulose, 12(3), 233–242. nanocrystals in polar organic solvents. Cellulose, 14(2), 109–113.
Martins, D. F., de Souza, A. B., Henrique, M. A., Silverio, H. A., Neto, W. P. F., & Wang, H. Y., Li, D. G., Yano, H., & Abe, K. (2014). Preparation of tough cellulose II
Pasquini, D. (2015). The influence of the cellulose hydrolysis process on the nanofibers with high thermal stability from wood. Cellulose, 21(3), 1505–1515.
structure of cellulose nanocrystals extracted from capim mombaca (Panicum Wyman, C. E., Dale, B. E., Elander, R. T., Holtzapple, M., Ladisch, M. R., & Lee, Y. Y.
maximum). Industrial Crops and Products, 65, 496–505. (2005). Coordinated development of leading biomass pretreatment
Mittal, A., Katahira, R., Himmel, M. E., & Johnson, D. K. (2011). Effects of alkaline or technologies. Bioresource Technology, 96(18), 1959–1966.
liquid-ammonia treatment on crystalline cellulose: Changes in crystalline Yamane, C., Miyamoto, H., Hayakawa, D., & Ueda, K. (2013). Folded-chain structure
structure and effects on enzymatic digestibility. Biotechnology for Biofuels, 4, 41. of cellulose II suggested by molecular dynamics simulation. Carbohydrate
Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y. Y., Holtzapple, M., et al. (2005). Research, 379, 30–37.
Features of promising technologies for pretreatment of lignocellulosic Yu, L. B., Lin, J. Y., Tian, F., Li, X. H., Bian, F. G., & Wang, J. (2014). Cellulose
biomass. Bioresource Technology, 96(6), 673–686. nanofibrils generated from jute fibers with tunable polymorphs and
Nishiyama, Y., Langan, P., & Chanzy, H. (2002). Crystal structure and crystallinity. Journal of Materials Chemistry A, 2(18), 6402–6411.
hydrogen-bonding system in cellulose Ibeta from synchrotron X-ray and Yue, Y. Y., Zhou, C. J., French, A. D., Xia, G., Han, G. P., Wang, Q. W., et al. (2012).
neutron fiber diffraction. Journal of the American Chemical Society, 124(31), Comparative properties of cellulose nano-crystals from native and mercerized
9074–9082. cotton fibers. Cellulose, 19(4), 1173–1187.

You might also like