You are on page 1of 10

Carbohydrate Polymers 224 (2019) 115202

Contents lists available at ScienceDirect

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

Transparent and strong polymer nanocomposites generated from Pickering T


emulsion gels stabilized by cellulose nanofibrils

Xinyue Liua, Xiaojun Qia, Yupeng Guana, Yingying Hea, Shuai Lia, Hongxia Liua, , Li Zhoua,

Chun Weia, Chuanbai Yua, Yunhua Chenb,
a
College of Material Science & Engineering, Guilin University of Technology, Guilin 541004, China
b
School of Materials Science and Engineering, South China University of Technology, Guangzhou 510640, China

A R T I C LE I N FO A B S T R A C T

Keywords: We report here the development of transparent and strong polymer composites reinforced by unmodified cel-
Cellulose nanofibrils lulose nanofibrils (CNFs) with a Pickering emulsion gelation strategy. The CNFs entangle and firmly stabilize on
Pickering emulsion the surface of emulsion droplets containing polymethyl methacrylate (PMMA) solution, leading to the gelation of
Polymer nanocomposite the emulsions. CNFs/PMMA composites were generated via vacuum filtration and solvent washing of the gel and
PMMA
a subsequent two-step hot pressing. The composites contained a unique self-assembled two-tier hierarchy of
Three-dimensional network
CNFs networks and demonstrate promising transparency, tensile strength, flexibility, and an extremely low
thermal expansion. Remarkably, these properties are highly tunable with varying the concentration of CNFs and
the volume ratio of the water to oil phase. This work offers a facile route to realize the well dispersion of
unmodified CNFs in hydrophobic polymer matrix and achieve high performance of polymeric materials re-
inforced by CNFs.

1. Introduction dispersed in most polar or hydrophilic polymers (Qing, Sabo, Cai, &
Wu, 2013; Xiao, Gao, Gao, & Li, 2016;Nordqvist et al., 2007). How-
Cellulose nanofibrils (CNFs) are typically a fibrous component of ever, at low concentrations of CNFs suspension, CNFs have a gel-like
cellulose fibers that have nanoscale (< 100 nm) diameter and lengths structure due to their high agglomeration, which leads to poor disper-
of up to several micrometers (Kiziltas, Kiziltas, Bollin, & Gardner, 2015; sion in non-polar or hydrophobic polymer matrix (Ahmadi et al., 2017).
Stelte & Sanadi, 2009). CNFs, which provide structural support to plant Incompatibility with hydrophobic polymers represents a critical issue
bodies in nature, are characterized by biodegradability, high surface during the development of these composites. Many methods have been
area, high aspect ratio, low density (˜1.5 g/cm3) (Ahmadi, Behzad, attempted in an effort to increase the compatibility and interaction
Bagheri, Ghiaci, & Sain, 2017), high strength (2–6 GPa) (Saito, between CNFs and the hydrophobic polymer matrix. These include
Kuramae, Wohlert, Berglund, & Isogai, 2013; Wu, Moon, & Martini, solvent exchange (Kiziltas et al., 2015), surface modification of CNFs
2014), high elastic moduli (130–150 GPa) (Iwamoto, Kai, Isogai, & (Soeta, Fujisawa, Saito, Berglund, & Isogai, 2015; Soman, Chacko, &
Iwata, 2009; Sakurada, Nukushina, & Ito, 1962; Sturcova, Davies, & Prasad, 2017), addition of compatibilizing agents (Volk, He, & Magniez,
Eichhorn, 2005), and low thermal expansion coefficients (4–6 ppm K-1) 2015), and in situ polymerization (Sain, Ray, & Mukhopadhyay, 2014).
(Diaz, Wu, Martini, Youngblood, & Moon, 2013; Horii & Wada, 2005). Although these methods can improve the dispersion of CNFs in the
These unique properties have been exploited for the development of polymeric matrix and the subsequent mechanical properties of the
CNFs as a reinforcing filler in polymer nanocomposite materials (Azizi composites, complex chemical reactions and expensive reagents or
Samir, Alloin, & Dufresne, 2005; Eichhorn et al., 2010; Sakakibara, additional latexes are required. Meanwhile, the transparency of the
Moriki, Yano, & Tsujii, 2017; Alidadi-Shamsabadi, Behzad, Bagheri, & nanocomposite (especially when transparent polymers such as PMMA
Nari-Nasrabadi, 2014; Sato et al., 2016; Cai et al., 2016; Ballner et al., are used) could be significantly affected by this process (Kiziltas et al.,
2016; Stone & Korley, 2010; Jahed, Khaledabad, Almasi, & Hasanzadeh, 2015). In addition, mixing some waterborne latexes, such as polyvinyl
2017; Ljungberg et al., 2005). acetate (PVAc), polylactide (PLA), nature rubber (NR) and so on, with
Due to the hydrophilic nature of cellulose, CNFs can be generally to prepare nanocomposites was also used to solve the aggregation of the


Corresponding authors.
E-mail addresses: aozihx@foxmail.com (H. Liu), msyhchen@scut.edu.cn (Y. Chen).

https://doi.org/10.1016/j.carbpol.2019.115202
Received 2 June 2019; Received in revised form 30 July 2019; Accepted 12 August 2019
Available online 13 August 2019
0144-8617/ © 2019 Elsevier Ltd. All rights reserved.
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

hydrophilic CNFs in hydrophobic polymer matrix (LópLez-Suevos, as the polymeric matrix and effectively reinforced by a multiple scale
Eyholzer, Bordeanu, & Richter, 2010; Larsson, Berglund, Ankerfors, & network of CNFs. Highly transparent composites with tensile strength
Lindström, 2012; Chaabouni & Boufi, 2017; Nechyporchuk, Pignon, twice higher than that of pure PMMA were realized even at a very high
Maria, & Belgacem, 2016; Silva et al., 2014; Abraham et al., 2013). CNFs content (i.e., 47 wt%).
Stable aqueous dispersions and homogeneous composites were pre-
pared when using latex with anionic surfactant and CNFs, while non- 2. Experimental
homogeneous distribution of CNFs within the matrix was resulted from
mixing negatively charged CNFs and positively charged matrix 2.1. Materials
(Nechyporchuk et al., 2016). Moreover, the aggregation of CNFs was
more pronounce with CNFs percentage increased in NR matrix Sisal fibers were gifted by Guangxi Sisal Fiber Group. 1, 2-di-
(Abraham et al., 2013). Therefore, realizing the improvements in the chloroethane (DCE), hydrochloric acid, sodium hydroxide, Na2S·9H2O,
dispersion of CNFs in a hydrophobic transparent polymer matrix re- sodium chlorite, chloroacetic acid, acetic acid, sodium carbonate, so-
mains a challenge. dium bicarbonate, epichlorohydrin, ammonium hydroxide (29.4%),
Alternatively, the Pickering emulsion strategy, which refers to an fluorescein isothiocyanate (FITC), and ethanol were purchased from
emulsion stabilized by solid particles rather than organic surfactants Aladdin Chemistry Co. Ltd. (Shanghai, China) and used directly.
(Pickering, 1907), offers potential to address this issue. Depended on Polymethyl methacrylate (PMMA) with n of 70 000 was purchased from
diversity of the solid particles, the Pickering emulsion has great ad- Chi Mei Industrial Factory (Taiwan).
vantages in building different functional and structural materials, and
even acting as microreactors (Liu, Geng, Xu, Wei, & Zhou, 2016, 2017; 2.2. Fabrication of CNFs from sisal fiber
Liu et al., 2019; Tang et al., 2019; Wang et al., 2019). Especially,
Pickering emulsions stabilized by CNFs have been used to build CNFs/ Fabrication of sisal fiber pulp: As-received sisal fibers were pre-
polymer microcapsules or microspheres, and even CNFs/polymer conditioned prior to cellulose extraction. The fibers were repeatedly
composites. For example, a Pickering emulsion co-stabilized by cellu- washed with distilled water, dried in an oven at 80 °C for 12 h, and then
lose nanocrystals (CNCs) and cellulose nanofibers (CNFs) was used to chopped to an approximate length of 5 cm. Approximately 10 g of the
fabricate CNCs/CNFs reinforced polyurethane PU microcapsules fibers were treated under moderate conditions (1 m M HCl, 100 mL,
(Svagan et al., 2014). Our group also reported a straightforward 165 °C) with 42 min of hydrolysis time. After hydrolysis, the fibers were
method of fabricating multifunctional model drug methyl red-loaded washed with distilled water until a colorless filtrate was obtained. The
polymethylmethacrylate (PMMA)/CNFs composite microspheres from a drained sisal fibers, 4 g of NaOH, 4 g of Na2S·9H2O, and 100 mL of
Pickering emulsion system co-stabilized by CNFs and upconversion distilled water were then mixed in a hydrothermal synthesis reaction
nanoparticles (Liu et al., 2016). More interestedly, Fujisawa et al. re- kettle, which was kept in an oven at 170 °C for 2 h. The following wash
ported a facile aqueous preparation process that yields nanostructured was performed with distilled water in some cases. Subsequently, the
polystyrene (PS)/CNFs composites via the formation of a CNFs-stabi- filtrated preproduct was treated with an acetic acid aqueous solution of
lized styrene/Water Pickering emulsion and subsequent polymerization sodium chlorite (Msodium chlorite : Macetic acid : Mwater = 1.34 : 1.30 : 1) at
(Fujisawa, Togawa, & Kuroda, 2017). Also, the PS/CNFs composite was 75 °C for 2 h. The final sisal fiber pulp (SFP) was obtained after repeated
easily collected via filtration, and subsequent melt pressing allowed the washing and then drying in an oven at 75 °C.
preparation of a transparent PS/CNFs composite film. The film ex- Chemical pretreatment: First, the dried sisal fiber pulp was well-
hibited high optical transparency, strength, and thermal dimensional dispersed in a water/ethanol solution of NaOH and stirred for 30 min
stability, owing to the reinforcement provided by the homogeneously under 30 °C. The chloroacetic acid was then added and the reaction was
distributed CNFs. Pickering emulsification approach combined with a performed at 70 °C. After the reaction, the excess carboxylmethyl cel-
curing reaction was also reported to fabricate hierarchical composites lulose was removed via multiple immersion and washing with distilled
of immiscible acrylic polymer and cellulose or chitin nanofibers water. The sisal fiber pulp modified with carboxyl groups was obtained
(Biswas, Sano, Shams, & Yano, 2017; Shams & Yano, 2015). Ad- after filtration.
ditionally, an oil-in-water Pickering emulsion of PLA/methylene Fibrillation of sisal fiber pulp: Approximately 1 g of the sisal fiber
chloride was firstly stabilized by regenerated cellulose (RC), and uni- pulp modified with the obtained carboxyl group was immersed in
formly dispersed RC/PLA microspheres was obtained through eva- 250 mL of water with a pH of 10–11 adjusted by NaOH for 4 h. The
poration of methylene chloride and vacuum filtration. The subsequent system was then stirred for 6 min under about 32000 × g using a high-
hot pressing of the microspheres allowed the preparation of a RC/PLA speed shear blender such as a household bean juice maker (AOKE WMS-
composite film (Zhang et al., 2018). However, there are some dis- PB108, China). Afterward, the viscous fluid was diluted to 400 mL and
advantages for the preparation of the CNFs-reinforced polymer in these centrifuged for 10 min at 14900 × g using a refrigerated centrifuge
studies. The evaporation of organic solvents could give rise to de- (Dynamica V18R, Australia), with the upper layer corresponding to the
mulsification which possibly led cellulose to desorb from the surface of aqueous dispersion of CNFs. A desired concentration of CNFs aqueous
emulsion droplets and affect the dispersion of cellulose in polymer dispersion could then be prepared from the obtained dispersion.
matrix. Additional polymerization or resin-curing process and pur- Meanwhile, the carboxymethylation degree of the final CNFs was de-
ification processes were involved in many cases. termination by conductimetric titration. The detailed process has been
Herein, we developed a CNFs-stabilized Pickering emulsion gel presented in Supporting Information part, and the titration curve was
strategy for building a large panel of CNFs/polymer nanocomposites. A exemplified in Fig.S1. The result showed that the carboxyl content of
Facile route for obtaining a good dispersion of CNFs in a hydrophobic the CNFs was 1.48 mmol g−1, equivalent to about 0.24 C arboxyl group
polymer matrix was designed. In contrast to the aforementioned stu- per anhydroglucose units. The definition of the degree of substitution
dies, our work is aimed at obtaining a Pickering emulsion gel resulting (DS) is the ratio of reacted carboxyl groups per anhydroglucose units, so
from the entanglement of CNFs stabilized on the surface of emulsion the DS of the final CNFs was 0.24. According to what Eyholzer; et al
droplets containing organic solvents of certain polymers. After the va- have stated in the reference (Eyholzer et al., 2010), the DS of 0.24 for
cuum filtration and solvent washing of the emulsion gel, hot pressing CNFs is reasonable.
was performed to generate the CNFs/polymer nanocomposites. We can Furthermore, in preparation for confocal laser scanning microscopy
easily adjust the CNFs content of the resulting composites by control- (CLSM), CNFs were labeled by Fluorescein isothiocyanate (FITC) in
ling the concentration of CNFs in the water phase and the volume ratio accordance with a previously described method (Dong & Roman, 2007).
of water and oil phase. Poly (methyl methacrylate) (PMMA) was served The detailed process has been presented in Supporting Information

2
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

Fig. 1. The schematic fabrication process of CNFs/PMMA composites based on Pickering emulsions stabilized by CNFs.

part, and a FITC content of 0.016 mmol/g of cellulose was calculated, and second step was performed at 120 °C and 75 MPa for 20 min. The
equivalent to 1 FITC moieties per 400 anhydroglucose units. resulting composites with a multiple scale three-dimensional (3D)
network of CNFs are referred to as C-CNFs/PMMA-x, where x denotes
the mass percent content of CNFs in the composites. Note that there is
2.3. Fabrication of CNFs/PMMA composites some mass loss of both nanosized CNFs and dissolved PMMA when the
Pickering emulsion was filtrated into the organic gel mat. So, the x was
Our design strategy for the fabrication of CNFs/PMMA composites is determined by TGA analysis according to the following formula.
shown in Fig. 1. A certain concentration of DCE solution with polymer
(PMMA) as the oil phase was prepared by dissolving polymer in DCE. A R composite − Rpmma
x= × 100%
certain concentration of CNFs aqueous dispersion as the water phase R cnfs − Rpmma
was mixed with the oil phase in a cylindrical glass container. Afterward,
the O/W Pickering emulsion stabilized by CNFs was obtained by Here, x is the finial content of CNFs in C-CNFs/PMMA composites, and
emulsifying the above mixture on a sonicator (KQ-400KDE) under Rcomposite, Rpmma and Rcnfs are respectively the residue mass of the C-
400 W (6 cycles of intermittent sonication with 5 min at each cycle were CNFs/PMMA composites, pure PMMA and pure CNFs. As a result, this
employed). The CNFs-stabilized emulsion formed a gel after being left loss was ˜22–32% and the denoted value (x) ranged from 68 to 78% of
undisturbed for 24 h. The specific experimental schemes are shown in the initial CNFs content.
Table 1. These Pickering emulsion gels are referred to as G-CNFs/
PMMA-x, where x represents the mass percent content of CNFs in the 2.4. Characterization
dispersed solid phase of the emulsion calculated based on the feed ratio.
Subsequently, the Pickering emulsion gel was vacuum filtered for ˜3 h The crystal structure of CNFs was characterized by X-Ray dif-
using a micropore filter (diameter: 50 mm), which was padded with a fractometor (PANalytical B.V., X’Pert PRO, Netherlands) which used
layer of polyterafluoroethylene organic filter membrane (pore size: 0.22 Cu-K target at 40 kV 300 mA, λ = 1.542 Å. Samples were scanned in 2θ
μm). The filtration process was extended to 24 h by dropping the mix- ranges varying from 5° to 40° (1° min−1).
ture solution of DCE and ethanol (VDCE : Vethanol = 9 : 1) into the gel to Fourier transform infrared (FT-IR) spectra of samples were mea-
eliminate the residual water, and finally an organic gel mat yielded. An sured as KBr pallets by a NICOLETNEXUS 470 infra-red spectrometer
incompletely dried mat, which was formed by drying the above organic (America) operating in transmission mode for wavenumbers ranging
gel mat at 50 °C in a mold to volatilize most of the solvent, was used to from 400 to 4000 cm-1 (resolution: 2 cm-1).
generate CNFs/PMMA composites using a two-step hot-pressing The morphology and structure of the samples were characterized
method. The first step was performed at 70 °C and 1.5 MPa for 15 min via field emission scanning electron microscopy (FE-SEM, Hitachi S-

Table 1
Experimental recipe of preparing Pickering emulsion gels stabilized by CNFs and the resulted composites.
Sample name Concentration of CNFs in water Concentration of PMMA in DCE Volume of water phase Volume of oil phase
(%) (w W−1) (mL) (mL)
Pickering emulsion gel Composites

G-CNFs/PMMA-17 C-CNFs/PMMA-13 0.3 4% 12 m L 4 mL


G-CNFs/PMMA-30 C-CNFs/PMMA-20 0.4 4% 12 m L 3 mL
G-CNFs/PMMA-47 C-CNFs/PMMA-37 0.4 2% 12 m L 4 mL
G-CNFs/PMMA-62 C-CNFs/PMMA-42 0.4 2% 12 m L 3 mL

3
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

4800, Japan) where each sample was gold-sputtered, polarizing mi- produced. A TEM image of the obtained CNFs (see Fig. S3) revealed that
croscopy (POM, Leica DM PxP, Germany), confocal laser scanning mi- the crossed CNFs are dispersed in the copper mesh and have a diameter
croscopy (CLSM, Leica Tcs Spe, Germany) where the samples of the gels of 4.8 ± 0.6 nm calculated by a Nano Measurer 1.2 software. The
or films were put on a glass slide with a coverslip and observed under diameter of 4.8 ± 0.6 nm was much smaller than that of 20–50 nm and
100 × or 200 × magnification., and transmission electron microscopy 5–20 nm respectively produced by grinding (Iwamoto, Nakagaito, &
(TEM, JEM-2100 F, JEOL, Japan) performed at 200 kV. For the TEM Yano, 2007) and TEMPO oxidization (Besbes, Alila, & Boufi, 2011;
imaging of CNFs, a drop of 0.001 wt% CNFs aqueous suspension was Besbes, Rei Vilar, & Boufi, 2011). Their length ranges from hundred
deposited on Cu grids and then stained with 2% uranyl acetate prior to namometers even to several micrometers. So the CNFs have very high
complete drying. aspect ratio. The long and entangled CNFs can form a strong network,
The optical transmittance of samples was measured at wavelengths owing to their fibrillar morphology and strong mutual hydrogen bonds.
from 300 to 800 nm using an ultraviolet − visible light spectro- The obtained CNFs were well dispersed in water, owing to their aqu-
photometer (UV-3600, Shimadzu). Transmission spectra were mea- eous dispersion (see Fig. S4a). The appearance of light and dark stripes
sured using air as a reference. after adding polarizing film indicated the liquid crystalline nature of
The practical amount of CNFs in the resulting composite was esti- CNFs (see Fig. S4b). In addition, the crystal structure was characterized
mated and the thermal degradation profile was assessed by means of via X-ray diffraction of CNFs (see Fig. S5). The lack of a doublet in the
thermogravimetric analysis (TGA). A 3 mg portion of sample was intensity of the main peak (θ = 22.6°) indicates that the CNFs obtained
measured (heating rate: 10 °C min−1) in a flowing nitrogen (N2) at- was present in the form of cellulose I, rather than cellulose II (Morán,
mosphere (flow rate: 60 mL min−1) on a thermal analyzer instrument Alvarez, Cyras, & Vázquez, 2008).
(Netzsch STA-449, Germany). Each measurement was performed on 3 CNFs as solid particles for stabilizing a Pickering emulsion have
individual samples for obtaining a mean data. been reported in our previous work (Liu et al., 2016). In the present
Differential scanning calorimetry (DSC) was measured with a work, we focus on the fact that a CNFs-stabilized Pickering emulsion
Netzsch DSC-204 instrument using indium standards for calibration can form a gel through the entanglement of the long CNFs. By inversion
with a heating rate of 10 °C min−1from 20 °C to 200 °C. N2 experiments we investigated the conditions promoting gelation of the
(20 mL min−1) was employed as the purging gas for the sample and CNFs aqueous dispersion, and found that gelation occurs only at CNFs-
reference cells. Heating, cooling, and reheating cycle were performed suspension concentrations of > 0.3 wt% (see Fig. S6a). In order to
and second heating cycle was studied. Samples were heated from 20 °C further prove that the gels formed, we perform rheology of the 0.4 wt%
to 200 °C, held at 200 °C for 5 min to eliminate thermal history in the CNFs aqueous dispersion and G-CNFs/PMMA-47 after ultrasonication.
first heating cycle. The measurements were performed on 3 individual The results shown in Fig. S6b presented that the storage modulus (G’)
samples for obtaining a mean data. and loss modulus (G’’) of the both samples had platform zones. And also
The mechanical properties of neat PMMA and C-CNFs/PMMA the higher G’ than G’’ proved the formation of crosslinked network
composite films were studied with a dynamic mechanical analyzer structure.
(DMA, TA Instruments Model Q800) in tensile mode. The Therefore, in the subsequent experiment, we employed concentra-
stress − strain measurements of these materials were performed in a tions of 0.3 wt% and 0.4 wt% for the suspension. A digital photo, POM,
strain rate of 1% min−1at 25 °C. The storage modulus and tan δ of the and CLSM image of the obtained Pickering emulsion gel of G-CNFs/
pure PMMA and C-CNFs/PMMA composites were performed at a con- PMMA-17 are shown in Fig. 2a, c, and e, respectively. This CNFs-sta-
stant frequency (1 Hz), amplitude of 15 μm, a temperature range from bilized emulsion was fixed in the bottom of an inverted vial, where the
30 to 110 °C, and with a heating rate of 3 °C min−1. For these experi- Pickering emulsion gel was formed. In the gel, the emulsion droplets
ments, rectangular composite films with a length of 10 mm a width of consisted of DCE solution of PMMA stably dispersed in a water con-
3 mm and a thickness of 0.10 mm were used, and the measurements tinuous phase (see Fig. 2c). In the CLSM image of the sample, the green
were performed on 3 individual samples. water phase (shown in Fig. 2e) shows that CNFs in the emulsion gel
The thermal dimensional stability of the samples was evaluated were still present in the continuous water phase and on the surface of
using a thermomechanical analyzer (TMA Q400, TA Instruments, New droplets (black dispersion phase). The entangled CNFs led to gelation of
castle, PA, USA). The linear coefficient of thermal expansion (CTE) of the water phase and subsequent formation of the emulsion gel. The
the film was determined using a 0.05 N load cell where samples were gelation of the CNFs-stabilized Pickering emulsion is important for the
heated from 0 to 170 °C (heating rate: 5 °C min-1) under a N2 atmo- subsequent filtration and solvent washing process, which yields an or-
sphere. ganic gel with PMMA incorporated into a 3D interconnected CNFs
Rheological measurements were conducted on TA ARES-RFS strain- network. Note that if the emulsion was filtrated directly, demulsifica-
controlled rheometer at 25 ℃. A flat-plate geometry with diameter of tion occurred and most of the PMMA and CNFs were mixed in the filter
25 mm was used. Samples were placed in the plate. Silicone oil was liquid. However, upon gelation of this emulsion, the emulsion droplets
applied on the edge of the sample to prevent water evaporation. The could still be retained after filtration.
linear viscoelastic of samples were preliminarily obtained from an A digital photo, POM, and CLSM image of the obtained organic gel
amplitude sweep using a constant angular frequency of 1.0 s-1 with mat with PMMA incorporated into the 3D interconnected CNFs network
strain amplitude varying from 0.01% to 100%. Complex viscosity, are shown in Fig. 2b, d and f, respectively. The digital photo shows that,
storage modulus (G’), and loss storages (G’’) were measured as a after filtration and solvent washing, the obtained mat became a thin
function of the angular frequency using oscillatory measurements in the membrane and seemed harder than the previous emulsion gel. From
range 0. 1 s-1 −100 s-1 with fixed strain amplitude of 0.8%. Fig. 2d, we can see that there were many closely packed emulsion
droplets in the mat. The droplets diameter decreased significantly after
3. Results and discussion filtration and solvent washing. The diameter distribution of these dro-
plets before and after filtration and solvent washing was investigated.
The method employed in this work differs from the TEMPO method The results were shown in Fig. S7. The diameter of the droplets in G-
described by Saito et al. (Tanaka, Saito, & Isogai, 2012). In the present CNFs/PMMA-17 was felt in the several micrometers to three hundred
work, chloroacetic acid was used to convert the hydroxyl groups of micrometers range, and the diameter distribution was fairly wide and
cellulose molecule to carboxylates, thereby resulting in negatively the mean diameter was about 50 μm. While, after filtration and solvent
charged SFP. The repulsive effect of these charges contributed sig- washing the diameter distribution of the droplets in the organic gel mat
nificantly to the separation of SFP into nanofibrils. After a high-speed with a mean diameter of about 14 μm became obviously narrow be-
shear treatment, individualized CNFs with high aspect ratios were tween several micrometers and forty micrometers. A similar result is

4
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

Fig. 2. Digital images (a, b), POM (c, d) and CLSM images (e, f) of the obtained Pickering emulsion gel of G-CNFs/PMMA-17 (a, c, e) and the corresponding organic
gel (b, d, f), and SEM images of the surface (g) and fracture (h) of the fully dried organic gel. The scale bars are 50 μm.

shown in Fig. 2f, where the smaller and closely packed emulsion dro- allowed flow of the melted PMMA and completed blending of the
plets (black spherical area) are dispersed in the interconnected CNFs PMMA with the interconnected CNFs network under high temperature
network (continuous green area). This may have resulted from the fact and pressure. However, with only one-step hot pressing, the rapid
that (i) on the one hand, the network structure in the emulsion gel can evaporation of DCE solvent ineffectively promoted the exudation of
fix and protect the emulsion droplets, but (ii) on the other hand, can PMMA from the droplets and complete penetration into the network,
reduce the efflux rate of the water and solvent during filtration and which led to form obvious boundary between PMMA droplets and
solvent washing. With the removal of the water phase and solvent CNFs. The CLSM image (Fig. 3e) of the composites obtained via one-
washing, the DCE solution of PMMA in the droplets could slowly exude step hot pressing revealed individual PMMA droplets (marked by white
from the inside to the outside. This resulted in the larger droplets be- dotted circle in the figure). These droplets were characterized by a clear
coming unstable and even demulsified or transforming to more stable boundary which represented phase separation occurred between these
smaller droplets, whereas the initially smaller droplets were perfectly droplets and interconnected CNFs network (green area). On the con-
preserved. The structure of the obtained organic gel mat was further trary, although these droplets still existed, the boundary of two phases
investigated via scanning electronic microscopy (SEM; see Fig. 2g and h was ambiguous in the composites obtained via two-step hot pressing
for SEM images of the surface and fracture, respectively, of the fully (Fig. 3f). It can be concluded that the interfacial compatibility of PMMA
dried mat). Many circular particles, i.e., PMMA droplets, were packed droplets and the interpenetrated CNFs network was improved during
on the surface of the mat. However, the image of the fracture (Fig. 2h) two-step hot pressing.
revealed flat circular (deformed during vacuum filtration) PMMA dro- To clarify the interface difference between the CNFs and PMMA
plets inlayed in the continuous CNFs network, where initial individual droplets in the composites produced by one-step and two-step hot
CNFs eventually evolved into 2D-sheet-like structures during the drying pressing, the cross-sectioned fractures of the composites were in-
process, owing to the intensive hydrogen bonding interaction between vestigated by SEM imaging. For the composite subjected to one-step hot
CNFs. pressing, Fig. 4a shows clear gaps between the oval PMMA droplets
The C-CNFs/PMMA composites were generated by hot pressing the (marked by white solid arrows) and the CNFs network where initial
gel mats. Hot pressing was performed in two steps: 1) low temperature individual CNFs eventually evolved into sheet-like structures (marked
and pressure (at 70 °C and 1.5 MPa for 20 min, deflation at intervals of by white dotted arrows), demonstrating relatively loose connections
5 min), 2) high temperature and pressure (at 120 °C and 75 MPa for between the two phases. However, for the composites obtained through
20 min). In our experiment, we found that the first step played an im- two-step hot pressing, oblate PMMA droplets (marked by white solid
portant role in the formation of high-transparency C-CNFs/PMMA arrows) were compactly embedded in a continuous micrometer-sized
composites. The appearance of the composites obtained through one- network (marked by white dotted), as shown in Fig. 4b and c, indicating
step hot pressing (under high temperature and pressure) and two-step the condense inner morphology of the composites. So, the composites
hot pressing is shown in Fig. 3a and b, respectively. A highly trans- contained a unique self-assembled two-tier hierarchy of CNFs networks,
parent composite was obtained through two-step hot pressing, while an one being the micrometer-sized network around the droplets and an-
opaque composite was obtained through one-step hot pressing. The other being the interconnecting bulk network throughout the compo-
POM images (Fig. 3c and d) of the two composites further disclosed that site.
many PMMA droplets were still obviously dispersed in the composites The FT-IR spectra of PMMA, the CNFs and C-CNFs/PMMA-37
obtained via one-step hot pressing, but the surface of the composite composite films were shown in Fig. S9. From the spectrum of CNFs, we
obtained via two-step hot pressing was smooth and clean. Obviously, can see that the stretching vibrations of the ‒OH and C = O groups are
the two-step hot pressing can facilitate to improve interface in- respectively observed at 3350 cm−1 and 1620 cm−1. While in spectrum
compatibility the PMMA droplets and CNFs. The possible reason is the of the C-CNFs/PMMA-37 composite film, along with the above-men-
slow exudation of a portion of PMMA from the droplets into CNFs tioned characteristic peaks of CNFs, there were distinct characteristic
network during the first step of the two-step hot pressing. It allowed sharp peaks at 1723 cm-1 and 2990 cm-1 respectively corresponding to
further slow volatilization of the trace residue DCE solvent in the in- the stretching vibrations of ester carbonyl and C-H groups, which were
completely dried gel mat, accompanied with which PMMA (surrounded characteristic peaks of PMMA. So, the results convincingly proved the
by close-packed CNFs) Facilely exudated from the droplets and pene- successful production of the C-CNFs/PMMA composite films.
trated into the interconnected CNFs network. The following second step Fig. 5a shows digital photos of the neat PMMA and C-CNFs/PMMA

5
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

Fig. 3. Digital images (a, b), POM (c, d) and CLSM (e, f) images of the obtained composite respectively through one-step hot pressing (a, c, e) and two-step hot
pressing (b, d, f).

composite sheets with different mass percent content of CNFs. The et al., 2013; Liu et al., 2010). Moreover, the light transmittance of the
patterns and letters in the background indicate that the C-CNFs/PMMA C-CNFs/PMMA composite fabricated via our design strategy is higher
composite sheets are all transparent despite the high CNFs mass percent than that of the previously reported transmittance (Kiziltas et al., 2015;
content (i.e., 42%), which is significantly higher than that reported by Littunen et al., 2013). This may have resulted from the successful
Littunen and Kiziltas (Kiziltas et al., 2015; Littunen, Hippi, Saarinen, & construction of the 3D interconnected network structure of CNFs in the
Seppälä, 2013). This result indicated that the light transmittance was PMMA matrix.
only moderately affected by the CNFs at the present loading levels, The thermal properties of the C-CNFs/PMMA composites were also
owing to the nanosize and homogenous dispersion of the CNFs (Liu, Liu, investigated. Fig. 6a and b show the TGA and DTG curves of the com-
Yao, & Wu, 2010). Here, the high transparency of the C-CNFs/PMMA posites, PMMA, and CNFs (the corresponding data are listed in Table 3).
composites, despite the high CNFs content, resulted from the unique The initial CNFs weight loss occurring at temperatures below 200 °C
self-assembled two-tier hierarchy of CNFs networks within the PMMA (see Fig. 6a) is attributed to the removal of moisture from the cellulose.
matrix. The thermal degradation occurring at temperatures of 200–380 °C is
Fig. 5b shows the quantitatively measured profiles of light trans- attributed to the depolymerization of hemicellulose and cleavage of
mittance versus the wavelength of visible light for the neat PMMA and glycosidic cellulose linkages for CNFs (Lee et al., 2012). Thermal de-
C-CNFs/PMMA composite sheets examined via UV–vis spectroscopy gradation of the PMMA began at 300 °C, i.e., the onset decomposition
performed over a visible wavelength range of 300–800 nm. The percent temperature, and the maximum weight loss occurred at 368 °C. How-
transmittances at 400, 600, and 800 nm are tabulated in Table 2. The ever, the thermal degradation of the C-CNFs/PMMA composites can be
neat PMMA and CNFs sheets transmitted ˜92% and 90%, respectively, divided into two processes, which are attributed to the decomposition
of the incident light. The C-CNFs/PMMA composites with 13 wt.%, of CNFs and PMMA, respectively. During the first process, the compo-
20 wt.%, 37 wt.%, and 42 wt.% of CNFs resulted in 85%, 84%, 87%, sites began to degrade at a temperature considerably lower and slightly
and 85% light transmittance, respectively, at a wavelength of 600 nm. higher than the temperatures associated with PMMA and CNFs, re-
The light transmittance of the C-CNFs/PMMA composites decreased spectively (see Table 3). The maximum weight loss of the C-CNFs/
only slightly compared with that of the neat PMMA, and was in- PMMA composites at temperatures of up to 313 °C was slightly higher
dependent of the CNFs loading content. This trend differed completely than that of CNFs. The possible reason is that the addition of PMMA
from the results of previous studies where, with increasing cellulose- inhibited the degradation of CNFs and increased the decomposition
nanomaterial loading level, the composites became increasingly temperature of the composites. During the second process, the onset
opaque, owing to CNFs agglomeration in the polymer matrix (Littunen and maximum decomposition temperatures of the C-CNFs/PMMA

Fig. 4. SEM images of the fracture of the composite C-CNFs/PMMA-42 obtained through one-step hot pressing (a) and two-step hot pressing (b, c).

6
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

compositions. Tg values of 110 °C and 98–105 °C were measured for


PMMA and the C-CNFs/PMMA composites, respectively. The lowered
Tg phenomenon has also been reported for PMMA/CNC composites (Liu
et al., 2010) and PMMA/NFC (nanofibrillated cellulose) composites
(Littunen et al., 2013). The decrease of Tg value of C-CNFs/PMMA
composites is probably resulted from the plasticization of the residual
solvent which was not completely volatilized during the hot-pressing
process. However, the gradually increased Tg values from 98 °C to
105 °C were also found with the increasing content of CNFs in these C-
CNFs/PMMA composites. The increase of Tg value is probably resulted
from the hydrogen bonding interactions between PMMA and CNFs. So,
though the Tg of the C-CNFs/PMMA composites was affected by the
probable residual solvent, the interaction between PMMA and CNFs
was still improved by the formation of hydrogen bonding.
Moreover, as revealed by the thermal expansion behavior (Fig. 6d),
the thermal dimensional stability of the C-CNFs/PMMA composite was
superior to that of the neat PMMA. The thermal expansion of the
composite was considerably lower than that of PMMA. The coefficient
of thermal expansion (CTE) of the C-CNFs/PMMA-42 Composite
(12.7 ppm K-1, comparable to that of glass (7–10 ppm K-1)) was sig-
nificantly lower than that (149.2 ppm K-1, below the glass transition
region) of PMMA (Ashby, 1992). Therefore, the inclusion of CNFs
yielded considerable reduction (˜12 times) in the thermal expansion of
the PMMA.
The tensile properties of neat PMMA and C-CNFs/PMMA composites
are shown in Fig. 7. As the figure shows, the tensile properties of all the
C-CNFs/PMMA composites were superior to those of the neat PMMA.
The tensile stress at break of the composites increased gradually with
increasing CNFs content. The optimum tensile stress (59 MPa) and the
maximum elongation at break (4.3%) achieved at a CNFs loading of
42%, which represented both ˜200% improvement, compared with
those of the neat PMMA. The C-CNFs/PMMA composites exhibited
Fig. 5. Digital photos (a) and profiles of light transmittance versus the wave-
length of visible light (b) of the neat PMMA and C-CNFs/PMMA composite
higher stiffness and toughness than the neat PMMA. The improvement
sheets with different mass contents of CNFs. of tensile strength indicates that CNFs acted as reinforcement in the
PMMA matrix by transferring load from the polymer matrix to the
CNFs. In our C-CNFs/PMMA composites, the multiple scale inter-
Table 2
connected network structure of CNFs carried the load, and stress dis-
UV-vis transmittances of the neat PMMA and C-CNFs/PMMA composite sheets
sipation could occur at the interfaces of the CNFs network/PMMA
at 400 nm, 600 nm, 800 nm.
droplets through mutual sliding. This dissipation could also occur in the
Sample name Transmittance(%) CNFs network itself, where the soft PMMA is penetrated in associated
with volatilization of DCE during solvent washing and two-step hot
400 nm 600 nm 800 nm
pressing. Moreover, the maximum elongation at break was improved
Neat PMMA 91 92 92 with the CNFs content in composites increased, which is different from
CNFs 88 90 90 that reported by Huang et al (Huang, Kuboyama, Fukuzumi, &
C-CNFs/PMMA-13 81 85 85
Ougizawa, 2018). The difference maybe resulted from the unique self-
C-CNFs/PMMA-20 80 84 85
C-CNFs/PMMA-37 83 87 87 assembled two-tier hierarchy of CNFs networks in the C-CNFs/PMMA
C-CNFs/PMMA-42 81 85 86 composites. The micrometer-sized network around PMMA droplets can
be deformed and allow larger deformation to take place along the di-
rection of the stress loaded.
(337 °C and 377 °C, respectively) were both significantly higher than Meanwhile, the temperature dependency of the storage modulus
those of PMMA and dependent on the CNFs content. Similar results and tan δ of the pure PMMA and C-CNFs/PMMA composites were in-
were observed for PMMA/CNC (cellulose nanocrystal) composites with vestigated via DMA to explain the mechanical behavior of the compo-
loading levels of 2 wt.% –10 wt.% (Liu et al., 2010). This could be at- sites with change in temperature, especially in the plastic region of
tributed to hydrogen bonding interactions between CNC hydroxyl materials. The C-CNFs/PMMA-20 composites showed much greater
groups and carbonyl groups on the PMMA matrix (Dong et al., 2012; storage modulus throughout the temperature range tested (Fig. S10a).
Kuo, 2008), where hydrogen bonding had a significant effect on the Though with changing from a glassy state to a rubbery state, a drop in
thermal properties of the polymer blends. Moreover, owing to the high the storage modulus was observed for both the pure PMMA and C-
char yield (27.8%) of CNFs at 700 °C, the char yield of the C-CNFs/ CNFs/PMMA composites, the decrease for the C-CNFs/PMMA compo-
PMMA composites also increased from 1.9% to 12.7% with increasing sites was slighter than that of the pure PMMA. Seen from Fig. S10b and
CNFs content of the composites. So, we can conclude that the thermal c, tan delta tan δ peak position the C-CNFs/PMMA-20 composites was
property of the C-CNFs/PMMA composites was poorer than the pure shifted to the right and towards the higher temperature regions com-
PMMA during the beginning decomposition process, but the subsequent pared with that of the pure PMMA, which suggested that CNFs hindered
maximum temperature of the C-CNFs/PMMA composites was inversely the movement of molecular chain segments in the PMMA matrix.
increased with the addition of CNFs. Additionally, the intensity of the tan δ peak obviously decreased com-
The glass transition temperatures in Fig. 6c show that the addition pared to pure PMMA. This indicates that The C-CNFs/PMMA-20 com-
of CNFs lowered the glass transition temperature (Tg) of PMMA at all posites had the better elasticity and probable crosslinking structure,

7
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

Fig. 6. TGA (a), DTG (b) and glass transition temperatures (c) curves of the PMMA, CNFs and C-CNFs/PMMA composites with different mass contents of CNFs, and
(d) thermal expansion behaviors of PMMA and C-CNFs/PMMA-42 Composite.

which maybe resulted from the self-assembled two-tier hierarchy of desirable combination of high tensile strength (60 MPa) and strain
CNFs networks dispersed in the PMMA matrix. Therefore, the self-as- (2%), which are four and two times higher than those of neat polymer,
sembled two-tier hierarchy of CNFs networks endows the C-CNFs/ respectively. The composites also exhibited a high optical transparency
PMMA composites with the unique mechanical property. (87%), high flexibility, excellent thermal properties, and low CTE si-
milar to that of glass (12.7 ppm K-1, 1/12th that of the neat polymer),
owing to the well-dispersed CNFs. Consequently, the results of this
4. Conclusions work can be applied to various polymers where reinforcement by CNFs
is possible and can provide a Facile route for realizing a good dispersion
We have developed a simple CNFs-stabilized Pickering emulsion gel of CNFs in a hydrophobic polymer matrix.
pathway for fabricating a CNFs/PMMA composite. The obtained com-
posites featured a unique multiple scale 3D network structure of CNFs.
The CNFs/PMMA composite layer resulted from the exudation of Acknowledgment
PMMA in emulsion drops into the outer original interconnected CNFs
network formed in the Pickering emulsion gel. This exudation occurred The authors gratefully acknowledge the financial support of the
when DCE effused and volatilized from the inner region of the drops National Natural Science Foundation of China (21664006, and
during vacuum filtration, solvent washing, and two-step hot pressing. 21604025) and Natural Science Foundation of Guangxi Province (No.
The multiple scale 3D network structure of CNFs synergistically led to a 2016GXNSFAA380004 and 2015GXNSFAA139256).

Table 3
TG analysis of the CNFs, PMMA and C-CNFs/PMMA composites.
Sample First cycle Second cycle TEnd (oC) Char yield (%)

TOnset (oC) TMax (oC) TOnset (oC) TMax (oC)

CNFs 209.5 302.1 – – 327.8 27.8


PMMA – – 300.9 368.6 413.8 1.9
C-CNFs/PMMA-13 245.2 300.5 332.5 368.6 410.5 5.2
C-CNFs/PMMA-20 248.5 301.3 333.9 376.4 411.8 6.2
C-CNFs/PMMA-37 251.9 310.4 335.2 376.5 412.5 11.4
C-CNFs/PMMA-42 254.6 313.0 337.9 377.2 413.2 12.7

TOnset means the onset temperature; TMax means the maximum degradation temperature; TEnd means the ending degradation temperature.

8
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

10(9), 2571–2576.
Iwamoto, S., Nakagaito, A. N., & Yano, H. (2007). Nano-fibrillation of pulp fibers for the
processing of transparent nanocomposites. Applied Physics A, 89, 461–466.
Jahed, E., Khaledabad, M. A., Almasi, H., & Hasanzadeh, R. (2017). Physicochemical
properties of carum copticum essential oil loaded chitosan films containing organic
nanoreinforcements. Carbohydrate Polymers, 164, 325–338.
Kiziltas, E. E., Kiziltas, A., Bollin, S. C., & Gardner, D. J. (2015). Preparation and char-
acterization of transparent PMMA-cellulose-based nanocomposites. Carbohydrate
Polymers, 127(10), 381–389.
Kuo, S. W. (2008). Hydrogen-bonding in polymer blends. Journal of Polymer Research,
15(6), 459–486.
Larsson, K., Berglund, L. A., Ankerfors, M., & Lindström, T. (2012). Polylactide latex/
nanofibrillated cellulose bionanocomposites of high nanofibrillated cellulose content
and nanopaper network structure prepared by a papermaking route. Journal of
Applied Polymer Science, 125(3), 2460–2466.
Lee, K., Tammelin, T., Schulfter, K., Kiiskinen, H., Samela, J., & Bismarck, A. (2012). High
performance cellulose nanocomposites: comparing the reinforcing ability of bacterial
cellulose and nanofibrillated cellulose. ACS Applied Materials & Interfaces, 4(8),
4078–4086.
Littunen, K., Hippi, U., Saarinen, T., & Seppälä, J. (2013). Network formation of nano-
fibrillated cellulose in solution blended poly (methyl methacrylate) composites.
Carbohydrate Polymers, 91(1), 183–190.
Liu, B. F., Yang, D. C., Man, H., Liu, Y. Q., Chen, H., Xu, H., ... Bai, L. J. (2017). A green
Fig. 7. Tensile properties of neat PMMA and C-CNFs/PMMA composites with pickering emulsion stabilized by cellulose nanocrystals via raft polymerization.
different mass content of CNFs. Cellulose, 25(1), 1–9.
Liu, H. X., Geng, S. M., Xu, Y., Wei, C., & Zhou, L. (2016). Facile fabrication of versatile
PMMA/CNFs–NaYF4:Yb/Er composite microspheres by Pickering emulsion system.
References Materials Letters, 166, 55–58.
Liu, H. Y., Liu, D. G., Yao, F., & Wu, Q. L. (2010). Fabrication and properties of trans-
parent polymethylmethacrylate/cellulose nanocrystals composites. Bioresource
Abraham, E., Deepa, B., Pothan, L. A., John, M., Narine, S. S., Thomas, S., ... Anandjiwala, Technology, 101(14), 5685–5692.
R. (2013). Physicomechanical properties of nanocomposites based on cellulose na- Liu, J. F., Wang, P., He, Y. N., Liu, K. Q., Miao, R., & Fang, Y. (2019). Polymerizable
nofibre and natural rubber latex. Cellulose, 20(1), 417–427. nonconventional gel emulsions and their utilization in the template preparation of
Ahmadi, M., Behzad, T., Bagheri, R., Ghiaci, M., & Sain, M. (2017). Topochemistry of low-density, high-strength polymeric monoliths and 3D printing. Macromolecules, 52,
cellulose nanofibers resulting from molecular and polymer grafting. Cellulose, 24(5), 2456–2463.
2139–2152. Ljungberg, N., Bonini, C., Bortolussi, F., Boisson, C., Heux, L., & Cavaillé, J. Y. (2005).
Alidadi-Shamsabadi, M., Behzad, T., Bagheri, R., & Nari-Nasrabadi, B. (2014). New nanocomposite materials reinforced with cellulose whiskers in atactic poly-
Preparation and characterization of low-density polyethylene/thermoplastic starch propylene: effect of surface and dispersion characteristics. Biomacromolecules, 6(5),
composites reinforced by cellulose nanofibers. Polymer Composites, 36, 2309–2316. 2732–2739.
Ashby, M. F. (1992). Materials selection in mechanical design. Oxford: Pergamon Press. LópLez-Suevos, F., Eyholzer, C., Bordeanu, N., & Richter, K. (2010). DMA analysis and
Samir, Azizi, Alloin, M. A. S., & Dufresne, A. (2005). Review of recent research into wood bonding of PVAc latex reinforced with cellulose nanofibrils. Cellulose, 17(2),
cellulosic whiskers, their properties and their application in nanocomposite field. 387–398.
Biomacromolecules, 6(2), 612–626. Morán, J. I., Alvarez, V. A., Cyras, V. P., & Vázquez, A. (2008). Extraction of cellulose and
Ballner, D., Herzele, S., Keckes, J., Edler, M., Griesser, T., Saake, B., ... Gindl-Altmutter preparation of nanocellulose from sisal fibers. Cellulose, 15(1), 149–159.
(2016). Lignocellulose nanofiber-reinforced polystyrene produced from composite Nechyporchuk, O., Pignon, Frédéric, Maria, B. D. R. A., & Belgacem, M. N. (2016).
microspheres obtained in suspension polymerization shows superior mechanical Influence of ionic interactions between nanofibrillated cellulose and latex on the
performance. ACS Applied Materials & Interfaces, 8(21), 13520–13525. ensuing composite properties. Composites Part B: Engineering, 85, 188–195.
Besbes, I., Alila, S., & Boufi, S. (2011). Nanofibrillated cellulose from TEMPO-oxidized Nordqvist, D., Idermark, J., Hedenqvist, M. S., Gällstedt, M., Ankerfors, M., & Lindström,
eucalyptus fibres: Effect of the carboxyl content. Carbohydrate Polymer, 84(3), Tom (2007). Enhancement of the wet properties of transparent chitosan-acetic-acid-
975–983. salt films using microfibrillated cellulose. Biomacromolecules, 8(8), 2398–2403.
Besbes, I., Rei Vilar, M., & Boufi, S. (2011). Nanofibrillated cellulose from Alfa, Euca- Pickering, S. U. (1907). Emulsions. Journal of the Chemical Society Transactions, 1907(91),
lyptus and Pine fibres: Preparation, characteristics and reinforcing potential. 2001–2021.
Carbohydrate Polymer, 86, 1198–1206. Qing, Y., Sabo, R., Cai, Z., & Wu, Y. (2013). Resin impregnation of cellulose nanofibril
Biswas, S. K., Sano, H., Shams, M. I., & Yano, H. (2017). Three-Dimensional-Moldable films facilitated by water swelling. Cellulose, 20(1), 303–313.
nanofiber-reinforced transparent composites with a hierarchically self-assembled Sain, S., Ray, D., & Mukhopadhyay, A. (2014). Improved mechanical and moisture re-
“reverse” nacrelike architecture. ACS Applied Materials & Interfaces, 9, 30177–30184. sistance property of in situ polymerized transparent PMMA/cellulose composites.
Cai, J., Chen, J. Y., Zhang, Q., Lei, M., He, J. R., Xiao, A. H., ... Xiong, H. G. (2016). Well- Polymer Composites, 36(9), 1748–1758.
aligned cellulose nanofiber-reinforced polyvinyl alcohol composite film: mechanical Saito, T., Kuramae, R., Wohlert, J., Berglund, L. A., & Isogai, A. (2013). An ultrastrong
and optical properties. Carbohydrate Polymers, 140, 238–245. nanofibrillar biomaterial: the strength of single cellulose nanofibrils revealed via
Chaabouni, O., & Boufi, S. (2017). Cellulose nanofibrils/polyvinyl acetate nanocomposite sonication-induced fragmentation. Biomacromolecules, 14(1), 248–253.
adhesives with improved mechanical properties. Carbohydrate Polymers, 156, 64–70. Sakakibara, K., Moriki, Y., Yano, H., & Tsujii, Y. (2017). Strategy for the improvement of
Diaz, J. A., Wu, X. W., Martini, A., Youngblood, J. P., & Moon, R. J. (2013). Thermal the mechanical properties of cellulose nanofiber-reinforced high-density poly-
expansion of self-organized and shear-oriented cellulose nanocrystal films. ethylene nanocomposites using diblock copolymer dispersants. ACS Applied Materials
Biomacromolecules, 14(8), 2900–2908. & Interfaces, 9(50), 44079–44087.
Dong, H., Strawhecker, K. E., Snyder, J. F., Orlicki, J. A., Reiner, R. S., & Rudie, A. W. Sakurada, I., Nukushina, Y., & Ito, T. J. (1962). Experimental determination of the elastic
(2012). Cellulose nanocrystals as a reinforcing material for electrospun poly(methyl modulus of crystalline regions in oriented polymers. Journal of Polymer Science,
methacrylate) fibers: Formation, properties and nanomechanical characterization. 57(165), 651–660.
Carbohydrate Polymers, 87(4), 2488–2495. Sato, A., Kabusaki, D., Okumura, H., Nakatani, T., Nakatsubo, F., & Yano, H. (2016).
Dong, S., & Roman, M. (2007). Fluorescently labeled cellulose nanocrystals for bioima- Surface modification of cellulose nanofibers with alkenyl succinic anhydride for high-
ging applications. Journal of the American Chemical Society, 129(45), 13810–13811. density polyethylene reinforcement. Composites Part A: Applied Science &
Eichhorn, S. J., Dufresne, E., Aranguren, M., Marcovich, N. E., Capadona, J. R., Rowan, S. Manufacturing, 83, 72–79.
J., ... Peijs, T. (2010). Review: current international research into cellulose nanofibres Shams, M. I., & Yano, H. (2015). Doubly curved nanofiber reinforced optically trans-
and nanocomposites. Journal of Materials Science, 45(1), 1–33. parent composites. Scientific Report, 5(2), 16421.
Eyholzer, C., Bordeanu, N., LopLez-Suevos, F., Rentsch, D., Zimmermann, T., & Oksman, Silva, M. J., Sanches, A. O., Medeiros, E. S., Mattoso, L. H. C., McMahan, C. M., &
K. (2010). Preparation and characterization of water-redispersible nanofibrillated Malmonge, J. A. (2014). Nanocomposites of natural rubber and polyaniline-modified
cellulose in powder form. Cellulose, 17, 19–30. cellulose nanofibrils. Journal of Thermal Analysis & Calorimetry, 117(1), 387–392.
Fujisawa, S., Togawa, E., & Kuroda, K. (2017). Facile route to transparent, strong, and Soeta, H., Fujisawa, S., Saito, T., Berglund, L., & Isogai, A. (2015). Low-birefringent and
thermally stable nanocellulose/polymer nanocomposites from an aqueous Pickering highly tough nanocellulose-reinforced cellulose triacetate. ACS Applied Materials &
emulsion. Biomacromolecules, 18(1), 266–271. Interfaces, 7(20), 11041–11046.
Horii, R., & Wada, M. (2005). The thermal expansion of cellulose. Cellulose, 12(5), Soman, S., Chacko, A. S., & Prasad, V. S. (2017). Semi-interpenetrating network com-
479–484. posites of poly(lactic acid) with cis-9-octadecenylamine modified cellulose-nanofi-
Huang, T., Kuboyama, K., Fukuzumi, H., & Ougizawa, T. (2018). PMMA/TEMPO-oxidized bers from Areca catechu husk. Composites Science and Technology, 141, 65–73.
cellulose nanofiber nanocomposite with improved mechanical properties, high Stelte, W., & Sanadi, A. R. (2009). Preparation and characterization of cellulose nanofi-
transparency and tunable birefringence. Cellulose, 25(4), 2293–2403. bers from two commercial hardwood and softwood pulps. Industrial & Engineering
Iwamoto, S., Kai, W., Isogai, A., & Iwata, T. (2009). Elastic modulus of single cellulose Chemistry Research, 48(24), 11211–11219.
microfibrils from tunicate measured by atomic force microscopy. Biomacromolecules, Stone, D. A., & Korley, L. T. J. (2010). Bioinspired polymeric nanocomposites.

9
X. Liu, et al. Carbohydrate Polymers 224 (2019) 115202

Macromolecules, 43(22), 9217–9226. compatibility in melt-extruded cellulose nano-fibers reinforced polyethylene via
Sturcova, A., Davies, G. R., & Eichhorn, S. J. (2005). Elastic modulus and stress-transfer surface adsorption of poly(ethylene glycol)-block-poly(ethylene) amphiphiles.
properties of tunicate cellulose whiskers. Biomacromolecules, 6(2), 1055–1061. European Polymer Journal, 72, 270–281.
Svagan, A. J., Musyanovych, A., Kappl, M., Bernhardt, M., Glasser, G., Wohnhaas, C., ... Wang, G. L., Xi, M. Z., Bai, L. J., Liang, Y., Yang, L. X., Wang, W. X., ... Yang, H. W. (2019).
Landfester, K. (2014). Cellulose nanofiber/nanocrystal reinforced capsules: a fast and Pickering emulsion of metal-free photoinduced electron transfer-ATRP stabilized by
facile approach toward assembly of liquid-core capsules with high mechanical sta- cellulose nanocrystals. Cellulose, 26, 5947–5957.
bility. Biomacromolecules, 15, 1852–1859. Wu, X. W., Moon, R. J., & Martini, A. (2014). Tensile strength of Iβ crystalline cellulose
Tanaka, R., Saito, T., & Isogai, A. (2012). Cellulose nanofibrils prepared from softwood predicted by molecular dynamics simulation. Cellulose, 21, 2233–2245.
cellulose by tempo/naclo/naclo2 systems in water at ph 4.8 or 6.8. International Xiao, S. L., Gao, R. N., Gao, L. K., & Li, J. (2016). Poly(vinyl alcohol) films reinforced with
Journal of Biological Macromolecules, 51(3), 228–234. nanofibrillated cellulose (NFC) isolated from corn husk by high intensity ultra-
Tang, J., Zhou, X., Cao, S. X., Zhu, L. Y., Xi, L. L., & Wang, J. L. (2019). Pickering in- sonication. Carbohydrate Polymers, 136, 1027–1034.
terfacial catalyst with CO2 and magnetic dual-response for fast recovering in biphasic Zhang, Y. C., Jiang, Y., Han, L., Wang, B. J., Xu, H., Zhong, Y., ... Sui, X. F. (2018).
reaction. ACS Applied Materials & Interfaces, 11, 16156–16163. Biodegradable regenerated cellulose-dispersed composites with improved properties
Volk, N., He, R., & Magniez, K. (2015). Enhanced homogeneity and interfacial via a pickering emulsion process. Carbohydrate Polymers, 179, 86–92.

10

You might also like