Accepted Manuscript

Effect of adsorption of polyethylene glycol (PEG), in aqueous

media, to improve cellulose nanostructures stability

Giovanni F. de Lima, Alana G. de Souza, Derval S. Rosa

PII: S0167-7322(18)31424-7
DOI: doi:10.1016/j.molliq.2018.07.080
Reference: MOLLIQ 9401
To appear in: Journal of Molecular Liquids
Received date: 18 March 2018
Revised date: 27 June 2018
Accepted date: 18 July 2018

Please cite this article as: Giovanni F. de Lima, Alana G. de Souza, Derval S. Rosa ,
Effect of adsorption of polyethylene glycol (PEG), in aqueous media, to improve cellulose
nanostructures stability. Molliq (2018), doi:10.1016/j.molliq.2018.07.080

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

Effect of adsorption of polyethylene glycol (PEG), in aqueous media, to

improve cellulose nanostructures stability

Giovanni F. de Lima1, Alana G. de Souza1, and Derval S. Rosa 1*

1
Engineering, Modeling and Applied Social Sciences Center (CECS), Federal
University of ABC, Santo André, SP
*dervalrosa@yahoo.com.br

PT
Abstract.
In recent year, the interest in environmentally friendly materials, like cellulose

RI
nanostructures (CNSs) has increased. Nanoscale dimensions, high surface area, low
density, and high mechanical strength are examples of CNSs properties that attract

SC
researchers’ attention. CNSs stability and dispersity are the main limitations that
prevent its application as nanofiller in polymeric matrices. In this work, the surface
NU
modification of CNSs was investigated. To overcome CNSs limitation, they were
modified with polyethylene glycol (PEG), in aqueous solution, in different proportions:
MA

2, 5 and 10 wt%. FTIR, elemental analysis, NMR, and XPS were carried out to verify
the adsorption efficiency. The results showed an excellent adsorption of PEG chains on
the CNSs surface, with particle size increase of 50-90 nm and a considerable increase in
ED

thermal stability. XRD and AFM analysis observed no significant variation in

crystallinity and morphology in modified CNSs. Thus, modification of CNSs with PEG
PT

is an environmentally friendly process and showed positive results regarding improving

CNSs stability, in aqueous media.
CE

1. Introduction
AC

The growing concerns of the necessity to preserve the environment and reduce the
exploitation of fossil resources, the interest on natural fibers and environmentally
friendly materials have increased [1]. Recent scientific and technological studies have
emphasized the importance of using renewable resources from industrial, forest, and
agricultural wastes as a source of raw materials to obtain biobased products [2,3].
Most of these natural residues mainly consist of cellulose, which is the most
abundant renewable and biodegradable polymer [4]. The worldwide annual production
is estimated at a hundred billion of tons [5]. Cellulose fibers are composed of two types

1
 ACCEPTED MANUSCRIPT

of structures: a crystalline and a disordered one [6]. The extraction of the crystalline
region usually results in cellulose nanostructures (CNSs).
Cellulose nanostructures is a new class of eco-material with many advantages,
such as nanoscale, renewability, high strength and modulus, low density, and high
surface area, and have attracted researchers attention [7]. Many previous studies have
been studying several cellulose sources, such as sugarcane bagasse [2], coconut palm oil
[8], sugar palm [9], eucalyptus kraft pulp [5], among others [6,10]. Eucalyptus residues
have attracted attention as a cellulose source due to its fast-growing, good fiber

PT
properties, renewability, and high availability, with over of 48 millions of tons of
residues produced in Brazil, in 2016 [11,12].

RI
However, CNSs are highly hydrophilic and have limited ability to disperse in

SC
hydrophobic solvents and polymer matrices [13]. Cellulose is a polar molecule, with
hydroxyl groups which tend to form strong hydrogen bonds, impairing its dispersity in
NU
water and its ability to be used as reinforcement for polymers [14–16]. Steric stabilizers,
such as polyethylene glycol (PEG), have been used to minimize the effects of particle
agglomeration [14,17,18]. There are some advantages of steric stabilization over
MA

electrostatic stabilization: it is still used in high ionic strength systems, and it is

adequate for multiple phase systems [14].
ED

The objective of this work was to evaluate the efficiency of PEG adsorption onto
CNSs surface, in aqueous media, to improve its dispersity and steric stability. The
characteristics of the modified and non-modified particles, like nanosizes, thermal
PT

stability, chemical composition, crystallinity index, the interaction between the CNSs
and PEG, and morphology, were determined using various techniques.
CE

2. Experimental
AC

2.1. Materials
Eucalyptus residues (Eucalyptus citriodora) was obtained after harvesting and
cutting in Mato Grosso (Brazil). Polyethylene glycol (PEG) (Mw = 1500 g.mol-1) and
Sodium Chlorite (99%) were purchased from Sigma-Aldrich. Sodium Hydroxide (99%)
and Potassium Hydroxide (99%) were obtained from Labsynth. All samples were
prepared using deionized water. The CNSs were extracted according to the procedure
outlined below.
2.2. Isolation of cellulose nanostructures

2
 ACCEPTED MANUSCRIPT

Eucalyptus sawdust was dried in an oven for 24 h at 60 ºC. After drying, 50 g of the
sample was treated with 1 L of sodium chloride solution (3,9 wt.%) for 2 h at 70ºC
under mechanical stirring and then washed several times, and finally dried at 60 ºC for
12 h. After this, 30g of the dried sample was treated with 600 ml of an aqueous solution
of sodium hydroxide (10 wt.%) and potassium hydroxide (10 wt.%) for 2 h at room
temperature under mechanical stirring and then washed several times, and finally dried
at 60 ºC for 12 h. Then, 10 g of the treated sample was mixed with 10 ml of ethanol
solution (80 wt.%). The ball milling of the mixture was performed for 6 h in a

PT
MA500/ball mill (Marconi) using alumina balls with weight ratio (balls/sample) of
70:1. The milled sample was named CNS.

RI
2.3. Adsorption of PEG onto the CNSs surface

SC
An aqueous suspension of CNS (3 wt.%) was prepared. Then, PEG was added in
this suspension in proportions of 0, 2, 5 and 10 wt.% (PEG:CNS), and the terminologies
NU
adopted to this proportions were CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG. This
suspension was kept under mechanical stirring at room temperature for 24h.
MA

2.4 Characterizations
2.4.1. Dynamic Light Scattering (DLS)
ED

The particle size of CNS dispersion in water was recorded by dynamic light
scattering (DLS) system using ALV-CGS3. The sample was illuminated by a polarized
HeNe laser beam (22mW,  = 633 nm) and followed by detection of the resultant
PT

fluctuation of scattered light at 90º by an APD pair of detectors operating in pseudo-

correlation. The samples were diluted to a final concentration of 1,0 wt.%, and an
CE

ultrasonic bath was used to facilitate dispersion of the CNS in water.

2.4.2. Adsorption tests
AC

The adsorption of PEG was studied using a fixed CNS suspension (3g.l-1) volume
(20 ml) with different initial concentrations of PEG (6g.l-1, 15g.l-1, and 30g.l-1). The
suspension was kept under mechanical stirring at room temperature for 24h. The
samples were filtrated through a 0,45 μm pore-diameter to remove the CNSs modified
with PEG. After that, the remaining PEG concentration was determined using a UV-vis
spectrophotometer at 192,4 nm wavelength (the spectral properties of PEG show a high
absorbance within the 190-230 nm range) [19]. The calibration curves of PEG have
been prepared in the range 1 g/l to 30 g/l.

3
 ACCEPTED MANUSCRIPT

The Langmuir adsorption isotherm describes quantitatively the formation of a

monolayer adsorbate on the outer surface of the adsorbent, and after that, no further
adsorption takes place [20]. The Langmuir isotherm is valid for monolayer adsorption
onto a surface containing a finite number of identical sites and is represented by the
following equation.
𝑄𝑜 ∗𝐾𝐿 ∗𝐶𝑒 1 1 1
𝑞𝑒 = or, linearized = ( 𝑜 ) + (𝑄𝑜 ) (1)
(1+ 𝐾𝐿 ∗𝐶𝑒 ) 𝑞𝑒 𝑄 ∗𝐾 𝐿 ∗ 𝐶𝑒

where, qe (mg/g) is the amount of element adsorbed per unit weight of adsorbent, Q o

PT
(mg/g) is the solid phase concentration corresponding to complete coverage of available
sites (limiting absorption capacity), Ce (mg/l) is the residual liquid phase concentration

RI
at equilibrium and KL is the adsorption coefficient.

SC
The Freundlich adsorption isotherm is commonly used to describe the adsorption
characteristics for the heterogeneous surface. The linearized Freundlich equation in
NU
logarithmic form is:
1
log 𝑞𝑒 = log 𝐾𝑓 + ∗ log 𝐶𝑒 (2)
𝑛
MA

where Kf (mg/g) and 1/n are characteristic constants determined, respectively, by the
intercept and slope of the Freundlich equation on logarithmic plot and which represent,
respectively, the sorption capacity and adsorption intensity.
ED

2.4.3. Zeta Potential

Zeta potential measurements were conducted using a Zetasizer Nano-ZS (Malvern
PT

Instruments) to characterize the surface charge property of nanoparticles. The

measurement was based on the electrophoretic mobility of the particles which was
CE

converted to zeta potential using Helmholtz-Smoluchowski equation [21].

2.4.4. Fourier Transformed Infrared Spectroscopy (FTIR)
Fourier transform infrared spectroscopy (FTIR) was conducted using a Frontier
AC

94942 (PerkinElmer, USA) in the range from 400 to 4000 cm-1 with 64 scans and a
resolution of 4 cm-1. It was used an attenuated total reflectance accessory.
2.4.5. Elemental Analysis (EA)
The elemental analysis of the samples was conducted to determine the Carbon and
Hydrogen contents using a FlashEA 1112 elemental analyzer (Thermo Scientific).
2.4.6. Thermogravimetric Analysis (TGA)
Thermogravimetric analysis (TGA) was carried out using STA 6000 instrument
(PerkinElmer, USA). The samples were heated from 20 to 600 °C with a heating rate of

4
 ACCEPTED MANUSCRIPT

20 °C/min under a nitrogen flow. The initial decomposition temperatures, measured at

10% weight loss, were used to compare the thermal stabilities.
2.4.7. X-Ray Diffraction (XDR)
The X-ray diffraction (XRD) experiments were performed on a STADI-P (Stoe),
equipped with CuK (=1,54056 Å), in the 2 range 0-40º, and a counting time of 100
s at each 1.05º. The crystallinity index was calculated (Eq. (3)) according to the method
proposed by Segal [22].
𝐼200 −𝐼𝑎𝑚

PT
𝐼(%) = ∗ 100 (3)
𝐼200

where I200 = Peak intensity corresponding to (200) plane and Iam = Intensity of the valley

RI
between (200) and (110) planes.
2.4.8. Nuclear Magnetic Resonance (NMR)

SC
The chain structure of the samples was characterized by NMR spectroscopy,
conducted by a Bruker Advance II+ (9.4 Tesla). NMR spectra were obtained at room
NU
temperature, with 4 mm probe operating at 400,13 MHz for 1H and 100,62 MHz for
C, MAS of 13 MHz. The pulse for 1H was 2,5 s, the contact time was 4 ms, with 512
13
MA

scans and time between pulses of 2 s.

2.4.9. X-ray Photoelectron Spectroscopy (XPS)
The surface modification of the CNS was further analyzed by X-ray photoelectron
ED

spectroscopy (XPS). The analysis was carried out using a ThermoFisher Scientific,
model K-alpha+, equipped with monochromatic radiation A1 K at room temperature
PT

operated at 10 or 20 eV for high resolution scans. Data were treated with CASAXPS
software.
CE

2.4.10. Atomic Force Microscopy (AFM)

AFM analyzed the morphologies of the samples. For images examination, a droplet
AC

of the aqueous suspension was sprayed onto a freshly cleaned silicon wafer and then
airdried during 24 h, at room temperature. The AFM experiments were performed in
tapping mode, using an Agilent equipment (model 5500), at ambient relative humidity
and temperature. The scan was done at a rate of 0.5 Hz, with an image resolution of
256×256 pixel, using a silicon cantilever of constant force between 25 and 75 N m-1 and
resonance frequency of 120 kHz.

3. Results and Discussion

3.1. Dynamic Light Scattering

5
 ACCEPTED MANUSCRIPT

DLS measurements were performed to identify the differences in nanoparticle

dimensions before and after PEG modification. The apparent hydrodynamic radius (RH)
of an aqueous solution of CNSs was determined at various concentrations of PEG (i.e.,
0, 2, 5 and 10 wt.%) [23]. Fig. 1 shows the results obtained from DLS analysis for the
CNS sample and the modified CNS samples and Table 1 presents the values for the
hydrodynamic radius of these particles.

PT
RI
SC
NU
Fig. 1. Size distribution profiles of CNSs before and after PEG adsorption.
MA

The data reveals that all modified samples had the largest radius comparing to CNS
sample [24]. It occurs because when PEG is added to the CNS suspension, it adsorbs
ED

onto the nanoparticle. In other words, it forms a layer with the PEG chains that prevent
the approximation of the CNS, promoting its steric stabilization [21]. The adsorption of
the polymer on the surface of nanocellulose results in a larger size of the particle
PT

[14,17].
Table 1. Hydrodynamic radius of the samples.
CE

Sample RH avg (nm) RH min (nm) RH max (nm)

CNS 159 68 388
CNS/2%PEG 211 15 414
AC

CNS/5%PEG 249 31 2106

CNS/10%PEG 248 39 816

Based on the results in Table 1, the optimum conditions for the physical adsorption
of PEG onto CNSs surface was 2 wt.% (sample CNS/2%PEG). Under this condition,
the particles sizes were very close to the CNS (RH max is only 30 nm larger). With
increasing PEG concentration, increasing amounts of polymer chains in the solution,
which probably results from more than the PEG in the suspension, represented by the
higher nanosizes [25].

6
 ACCEPTED MANUSCRIPT

3.2. Adsorption tests

To determine the maximum adsorption capacity of PEG onto CNSs, a study of
adsorption models was achieved, comparing the Langmuir and Freundlich models.
Table 2 shows Langmuir and Freundlich parameters. According to the Freundlich
model, slope 1/n, ranging between 0 and 1, is a measure for the adsorption intensity or
surface heterogeneity, becoming more heterogeneous as its value gets closer to zero
[26]. So, according to this theory, it is considered that the surface of the CNSs consists
of heterogeneous points with different adsorption potentials, and the parameters values

PT
show the distribution characteristics of the adsorption potentials [27].
From the data in Table 2, that value of l/n = 0.59 and R2 = 0.88 it is possible to

RI
conclude that the sorption of PEG unto CNSs surface is favorable [20]. It was observed

SC
that the amount sorbed increases with increasing the PEG contents, up to the maximum
level observed. This maximum content was 5 wt%, which exhibited the greatest amount
NU
adsorbed [28]. The best adsorption occurred when all the hydroxyls of the cellulose
were associated with PEG molecules, resulting in a total surface coating and the
formation of a monolayer, which, by steric repulsion, prevents the agglomeration of the
MA

cellulosic nanoparticles.
The results obtained from the Langmuir model give a satisfactory correlation
ED

between the model predictions and experimental data. This describes, as mentioned
before, the formation of a monolayer adsorbate (PEG) on the surface of the adsorbent
(CNSs) [20]. KL value indicates the adsorption nature to be either unfavorable if K L>1,
PT

linear, linear if KL=1, favorable if 0<KL<1 and irreversible if KL=0. From the data
calculated in Table 2, the K0 is greater than 0 but less than 1 indicating that Langmuir
CE

isotherm is favorable [20]. The R2 value is 0.92 and proving that the sorption data fitted
well to Langmuir Isotherm model. Comparing the values of R2 for the two models, it is
AC

possible to conclude that the Langmuir model is the most adequate, having the value
closest to one. Thus, it is concluded that the adsorption of the PEG on the surface of the
CNSs occurs with the formation of a monolayer.

Table 2. Langmuir and Freundlich parameter for adsorption of PEG onto CNSs.

Freundlich Langmuir
2 o
1/n Kf (g/g) R Q (g/g) KL R2
0.59 5.88 0.88 7.41 0.43 0.92

3.3. Zeta Potential

7
 ACCEPTED MANUSCRIPT

Zeta potential (δ) is a measurement that evaluates the electrostatic stabilization of

particles in suspension, regarding electrophoretic mobility [24]. Low absolute values for
δ indicate an agglomeration and flocculation tendency of the particles in suspension.
The δ results for the samples CNS, CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG
were -33 ± 1, -26 ± 2, -25 ± 4, and -16 ± 2, respectively.
These results showed that, after modification with PEG, absolute zeta potential
values have decreased (from 33 to 16). It is reasonable to assume that the modified
CNSs samples are less susceptible to present high electrostatic repulsion compared to

PT
the CNSs without modification. It occurs because the PEG chains anchorage themselves
in CNSs surfaces, and it tends to form hydrogen bonds with hydroxyls groups of

RI
cellulose, resulting in a reduction of the superficial charges in the solution. Since δ is a

SC
measure of the superficial charges in a suspension, a decrease on its values indicates
that the free hydroxyls of the nanocellulose are being consumed, meaning that PEG is
NU
being modified on the CNS [29]. Also, PEG chain absorbed on CNSs surfaces can
affect its configurational entropy of the system. When a PEG chain emanates from a
surface (i.e., anchored) encounters a similar chain, there is a steric repulsion that arises
MA

from the loss of configurational entropy of the liquid system [14].

ED

3.4. Fourier Transformed Infrared Spectroscopy

FTIR spectra of PEG and modified samples are shown in Fig. 2. All the samples
showed native cellulose peaks and a very similar structure with differences due to the
PT

interaction between the CNSs and PEG. In all cases, it was possible to identify the
broad transmittance in the range of 3400-3100 cm-1, that can be ascribed to the
CE

stretching of the -OH groups. There was a slightly shift to high frequency on modified
samples possibly caused by hydrogen-bonding interactions between PEG chains and
AC

CNS surface hydroxyls [30]. An intense peak at 2900 cm-1 appeared due to the C-H
stretching [24,31]. The deformation, wagging and twisting modes of
anhydroglycopyranose vibration of cellulose were shown from 1800 to 800 cm-1 [24].
The PEG spectra showed several peaks in the region between 1500 and 850 cm-1,
characteristics to the C-H and C-O bonds.

8
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
ED
PT
CE

Fig. 2. Infrared Spectra of the CNSs, PEG, and modificated samples: (a) full spectra and (b) zoom on the
region from 1600 cm-1 to 700 cm-1.
AC

After the PEG adsorption, the samples showed distinct peaks, which confirmed the
presence of both materials. Also, the main peaks observed were 1034 cm-1, 947 cm-1,
and 841 cm-1, which is assigned to C-O bounding of PEG and cellulose interaction, C-H
stretching of β-glycosidic bonds of cellulose, and 841 cm-1 of out-of-plane bending
vibrations of O-H of the crystalline region in the PEG [29,30,32]. According to Li, Liu,
and Huang, the presence of peaks at 841 cm-1 and 947 cm-1 in modified samples is an
indicator of the success of the adsorption process [33].

9
 ACCEPTED MANUSCRIPT

3.5. Elemental Analysis

PEG functionalization via hydrogen bonding-based adsorption onto the CNSs
surface is an efficient way of modifying the surface of nanoparticles. PEG can provide
steric hindrance in aqueous systems [14]. Elemental analysis is a technique that
evaluates carbon and hydrogen atoms mass fraction, allowing the verification of surface
modifiers adsorption onto the nanoparticles and the degree of substitution concerning
oxygen reduction [14,33].
The results show that carbon and hydrogen content was 35% and 5.4%,

PT
respectively, for the CNS sample. Considering that CNSs sample was constituted only
of carbon, hydrogen, and oxygen (cellulose structure), the oxygen content was 59%, and

RI
the ratio O:C was 1.6 [34]. After the adsorption process with PEG, an increase in carbon

SC
content was verified (45%, 49% and 51% for CNS/2%PEG, CNS/5%PEG, and
CNS/10%PEG, respectively). Since PEG had a higher content of carbon than cellulose,
NU
this behavior supported the conclusion of PEG adsorption onto CNS surface [14].
Considering that the hydrogen content was 7.4%, 8.4%, and 8.7% for
CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG, respectively, and that PEG is
MA

composed only by carbon, hydrogen, and oxygen, it was possible to calculate oxygen
content in the samples. The estimated values were 47%, 42% and 40% for
ED

CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG, respectively. This element content

had decreased considerably, as had the O:C ratio (from 1.03 of CNS/2%PEG to 0.85
and 0.77 of CNS/5%PEG and CNS/10%PEG, respectively). This result was an evidence
PT

of the substitution of the free hydroxyls of the CNS by the PEG chains [35].
CE

3.6. Thermogravimetric Analysis

Thermogravimetric analysis (TGA) is a conventional method for detecting the
AC

thermal stability of different materials [36]. Fig. 3a presents the TGA curves of CNS
and modified CNS samples and Fig. 3b, its respective DTG curves. It can be observed
that the samples had an initial weight loss in 10 wt.% (T10%) reported to the breakage of
inter- and intramolecular hydrogen bonds accompanied by gradual loss of water [37].
The T10% of the modified samples showed a considerable increase, from 83ºC of CNS to
290ºC, 340ºC and 340ºC of CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG,
respectively. It occurred, probably, due to the hydrogen bonds between PEG and CNS,
which increased thermic stabilization of the samples [38].

10
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
ED
PT

Fig. 3. Thermograms of CNSs before and after the adsorption (a) TG curves and (b) DTG curves.
CE

Increasing the temperature, the weight was constant until the temperatures
AC

corresponding to the beginning of thermal degradation of the samples. The peaks of

maximum degradation for CNS, PEG, CNS/2%PEG, CNS/5%PEG and CNS/10%PEG
were 325, 432, 437 and 434 ºC, respectively. This peak of degradation is related to the
depolymerization and pyrolytic decomposition of the biopolymers (PEG and CNSs).
The increase in this temperature indicated the formation of strong intermolecular
interactions between PEG and CNSs [36]. The results showed that the samples are more
stable after the adsorption, which is excellent for future applications [33].

3.7. X-ray Diffraction

11
 ACCEPTED MANUSCRIPT

The X-ray diffraction patterns of CNS and PEG-modified samples are represented
in Fig. 4. A strong characteristic reflection at 2θ about 22.5° (crystallographic plane
(002) of cellulose type I crystal) was observed [24]. This peak remained presents in the
modified samples. This indicated that the crystalline structure of CNS was essentially
preserved after adsorption [39]. However, it was observed a small shift at the peak of
22.5° that can be justified by the little change in the crystalline structure of the samples.
It probably occurs because the bound PEG molecules do not have enough mobility to
crystallize in the same way that CNSs without PEG [30,40]. At higher PEG content (10

PT
wt%) the peak variation is more pronounced, probably because the excess of PEG
results in a microphase separation, and furthermore disrupted the uniform alignment of

RI
mesophases of CNSs, thus leading to a slight decline of crystallinity of the samples

SC
[30]. Other characteristic reflections for cellulose I type structure can be found at 14.7°
(101), 16.6° (101̅) and 34.7 (040). The PEG-modified samples had the same diffracted
NU
pattern as the CNSs.
MA
ED
PT
CE
AC

Fig. 4. XRD diffraction patterns of CNS before and after the adsorption with PEG.

Meanwhile, PEG-modified samples showed a new peak at 19,5º, corresponding to

(120) reflections from the PEG crystallites, indicating its presence in the samples [30].
Regarding crystallinity, there were no significant changes in the samples structures.
The crystallinity index (CI) varied from 70% for CNS to 74%, 70% and 73% for
CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG. It is an indicator that the alignment of

12
 ACCEPTED MANUSCRIPT

the cellulose chains was stable at low loading of PEG (2 wt.%), thus imparting positive
effects on the crystallization of the PEG-modified samples after drying [14,30].

3.8. Nuclear Magnetic Resonance

Solid-state nuclear magnetic resonance analysis was carried out to put in evidence
the surface modifications caused by adsorption with PEG, and the peak assignment was
carried out based on the literature [33]. Fig. 5 shows 13C NMR spectra.

PT
RI
SC
NU
MA
ED

Fig. 5. Solid-state 13C NMR for the studied samples.

CNS sample shows cellulose characteristic peaks. The region from 62 to 66 ppm
PT

was related to carbon C6. This region presented two peaks, one associated with the
amorphous region (63 ppm) and one, to the crystalline region (65 ppm). The region
CE

between 104 and 107 ppm was assigned to carbon C1. The region from 82 to 90 ppm
was associated with carbon C4 (84 ppm for amorphous cellulose chains and 89 ppm for
AC

crystalline chains). The carbons C2, C3, and C5 were associated to the region from 71
to 77 ppm. It was not possible to discern these carbons, which do not contribute to the
β(1-4) linkage. All of these signals were coherent to the reports on the literature [41–
44].
It was possible to observe that PEG-modified samples presented the same signals
than CNS samples, with small variations on its intensity. New peaks could be observed
in these samples, at 70 ppm and 63 ppm, characteristics of PEG chains, and these values
agreed with literature values [45,46]. The increase of these peaks intensity, among the
decrease in carbon C4 peak intensity, was an indicator of PEG adsorption onto the

13
 ACCEPTED MANUSCRIPT

samples. According to Goetz et al. the interaction between cellulose nanostructures and
modifying agents occurs, principally, in the C4 carbon region [46].

3.9. X-ray Photoelectron Spectroscopy

The XPS data of CNSs before and after adsorption are shown in Figs. 6 and 7. This
analysis data was obtained from the surface until 10 nm of depth. Thus, it is an
important technique to verify surface modifications in CNSs [47]. The full spectra of
CNSs show that the surface is mainly constituted of carbon (signal at 284 eV) and

PT
oxygens atoms (532 eV) (Fig. 6a) [34]. As expected, after treatments, the elements
observed are the same than before the modification. The O/C ratio varied for all the

RI
samples, according to the addition of PEG by mass, as presented in Table 3. The O/C

SC
ratio decreased from about 2.5 to 0.5, 0.5 and 0.4, for CNS, CNS/2% PEG, CNS/5%
PEG and CNS/10% PEG, respectively (Table 3). This is in agreement with the chemical
NU
structure of the PEG, which have a high concentration of carbons.
MA
ED
PT
CE
AC

Fig. 6. Full XPS spectra of (a) CNS; (b) CNS/2%PEG; (c) CNS/5%PEG and (d) CNS/10%PEG.

The C1s XPS spectrum for CNS sample are shown in Fig. 7a. It was possible to
observe two main peaks at 288 eV (peak C3), associated to C-O-C bonds and/or C=O
(acetal and carbonyl moieties of cellulose), and at 286 eV (peak C2), associated to C-O
and/or C-H bonds (alcohols and ethers groups) [35,48]. The XPS analysis of all the

14
 ACCEPTED MANUSCRIPT

samples displays the presence of two peaks, even after the treatment (C2 and C3).
Ferreira et al. observed the same peaks, which were an indicator of a pure material,
since the peak C1 is attributed to impurities associated with the presence of residual
lignin, extractive substances and fatty acids. Also, the C4 signal is assigned to
carboxylic functions present in hemicellulose molecules [34,49]. The C1s XPS spectra
for PEG-modified samples are shown in Fig. 7b (CNS/2% PEG), Fig. 7c (CNS/5%
PEG) and Fig. 7d (CNS/10% PEG). It was possible to observe an increase in the
intensity of C3 peak (~288 eV) and, also, the appearance of the C2 peak for all modified

PT
samples at different intensities [50].

RI
SC
NU
MA
ED
PT
CE
AC

Fig. 7. C1s XPS spectra for: (a) CNS; (b) CNS/2%PEG; (c) CNS/5%PEG and (d) CNS/10%PEG.

These variations in the carbons are consistent with PEG molecular structure, which
is composed of carbon, hydrogen, and oxygen. The increase in the intensity of C2 and

15
 ACCEPTED MANUSCRIPT

C3 are associated with bonds between PEG and the CNSs, confirming the efficiency of
the methodology presented in this work [34].
Table 3. O/C ratio of CNSs before and after adsorption with PEG.

Sample O/C

CNS 2.6
CNS/2%PEG 0.5
CNS/5%PEG 0.5
CNS/10%PEG 0.4

PT
3.10. Atomic Force Microscopy
The neat CNS and PEG-modified samples were characterized by AFM (Fig. 8). No

RI
significant difference was observed in the AFM images regarding morphology, mainly

SC
for the samples NEC/2%PEG and NEC/10%PEG. In the Fig. 8 is also presented a
possible scheme of a steric stabilization of cellulose nanostructures. It suggests that the
PEG has no substantial impact on the structural integrity or the nanosizes of the
NU
particles [14,51].
CNS samples, as expected, present agglomerates, which justify the larger sizes in
MA

the DLS analysis. After adsorption, it is possible to observe more dispersed particles,
with smaller sizes due to the better distribution. However, the particles presented a
higher sizes heterogeneity. The best result, in terms of dispersion, was verified for the
ED

samples with 2 and 10 wt.% of PEG. Despite not presenting overlapping particles,
CNS/5%PEG showed a large number of nearby particles. This is justified by the steric
PT

stabilization promoted by PEG, even at lower intensity [52].

AFM analysis also allows the determination of particle size distribution. The results
CE

indicated a high frequency for the range of 100 to 150 mm, consistent with the
previously present characterizations, like DLS. The sample that presented the smallest
AC

and the most uniform results was CNS/2%PEG, indicating that, besides the success of
PEG adsorption, there was no excess of PEG in the sample, which would result in
higher particle sizes [14]. These results are coherent with the nanosizes measured by
Alves et al. [24]. In all systems, the particles showed a needle-like crystals shape.
Another relevant information was the aspect ratio (defined as the length-to-
diameter, L/d) which is a factor usually indicated as a controlling parameter for the
mechanical properties of nanocomposites [24]. Typically, cellulose nanocrystals have
L/D values between 3 and 10. In this work, the values found were: 6.1, 3.0, 3.0, and 5.4
for the CNS, CNS/2%PEG, CNS/5%PEG, and CNS/10%PEG, respectively. This

16
 ACCEPTED MANUSCRIPT

parameter depends on the original cellulose characteristics, extractions conditions,

modification of cellulose, and others [24].

PT
RI
SC
NU
MA
ED
PT
CE
AC

Fig. 8. AFM images of CNS and PEG-modified samples (5x5 μm).

4. Conclusions
A successful adsorption of cellulose nanostructures (CNSs) with polyethylene
glycol (PEG) is reported. The nanoparticles showed a little increase in its size
(compared to pure CNSs sample) and a decrease in zeta potential values, supporting the
conclusion that there was stability due to steric hindrance. The efficiency of the
methodology concerning the interaction between both materials was confirmed by
FTIR, NMR, XPS and elementary analysis, which indicated the presence of PEG in

17
 ACCEPTED MANUSCRIPT

CNSs modified samples. Also, it was observed that PEG was absorbed by hydrogen
bonding successfully on the surface of the CNS. The PEG-modified samples showed
high thermal stability and XRD results showed no significant changes in crystallinity
and morphology in the samples, respectively. The addition of 2 wt.% PEG was the one
that stood out regarding system stability. Thus, adsorption with PEG is an efficient and
environmentally friendly way to improve cellulose nanostructures stability.

Acknowledgments

PT
The authors thank the UFABC, REVALORES Strategic Unit, CNPq (nº 306401/2013-
4) and Multiuser Central Facilities (UFABC).

RI
References

SC
[1] W.P.F. Neto, H.A. Silvério, N.O. Dantas, D. Pasquini, Extraction and
characterization of cellulose nanocrystals from agro-industrial residue – Soy
NU
hulls, Ind. Crop. Prod. 42 (2013) 480–488. doi:10.1016/j.indcrop.2012.06.041.
[2] F.B. de Oliveira, J. Bras, M.T.B. Pimenta, A.A. da S. Curvelo, M.N. Belgacem,
MA

Production of cellulose nanocrystals from sugarcane bagasse fibers and pith, Ind.
Crops Prod. 93 (2016) 48–57. doi:10.1016/j.indcrop.2016.04.064.
[3] L. Sisti, G. Totaro, M. Vannini, P. Fabbri, S. Kalia, A. Zatta, A. Celli, Evaluation
ED

of the retting process as a pre-treatment of vegetable fibers for the preparation of

high-performance polymer biocomposites, Ind. Crops Prod. 81 (2016) 56–65.
PT

doi:10.1016/j.indcrop.2015.11.045.
[4] S.H. Osong, S. Norgren, P. Engstrand, Processing of wood-based microfibrillated
CE

cellulose and nanofibrillated cellulose, and applications relating to papermaking:

a review, Cellulose. 23 (2016) 93–123. doi:10.1007/s10570-015-0798-5.
[5] L. Chen, J.Y. Zhu, C. Baez, P. Kitin, T. Elder, Highly thermal-stable and
AC

functional cellulose nanocrystals and nanofibrils produced using fully recyclable

organic acids, Green Chem. 18 (2016) 3835–3843. doi:10.1039/C6GC00687F.
[6] P. Phanthong, G. Guan, Y. Ma, X. Hao, A. Abudula, Effect of ball milling on the
production of nanocellulose using mild acid hydrolysis method, J. Taiwan Inst.
Chem. Eng. 60 (2016) 617–622. doi:10.1016/j.jtice.2015.11.001.
[7] C. Xu, J. Chen, D. Wu, Y. Chen, Q. Lv, M. Wang, Polylactide/acetylated
nanocrystalline cellulose composites prepared by a continuous route: A phase
interface-property relation study, Carbohydr. Polym. 146 (2016) 58–66.

18
 ACCEPTED MANUSCRIPT

doi:10.1016/j.carbpol.2016.03.058.
[8] C. Xu, S. Zhu, C. Xing, D. Li, N. Zhu, H. Zhou, Isolation and properties of
cellulose nanofibrils from coconut palm petioles by different mechanical process,
PLoS One. 10 (2015) 1–11. doi:10.1371/journal.pone.0122123.
[9] R.A. Ilyas, S.M. Sapuan, M.R. Ishak, Isolation and characterization of
nanocrystalline cellulose from sugar palm fibres (Arenga Pinnata), Carbohydr.
Polym. (2017). doi:10.1016/j.carbpol.2017.11.045.
[10] F. Jiang, Y. Lo Hsieh, Cellulose nanocrystal isolation from tomato peels and

PT
assembled nanofibers, Carbohydr. Polym. 122 (2015) 60–68.
doi:10.1016/j.carbpol.2014.12.064.

RI
[11] IBÁ, Anuário Estatístico da Indústria Brasileira de Árvore - 2017., Indústria

SC
Bras. Árvores. (2017) 100. doi:10.1017/CBO9781107415324.004.
[12] J.O. Brito, F.G. Silva, M.M. Leão, G. Almeida, Chemical composition changes in
NU
eucalyptus and pinus woods submitted to heat treatment, Bioresour. Technol. 99
(2008) 8545–8548. doi:10.1016/j.biortech.2008.03.069.
[13] E. Kloser, D.G. Gray, Surface grafting of cellulose nanocrystals with
MA

poly(ethylene oxide) in aqueous media, Langmuir. 26 (2010) 13450–13456.

doi:10.1021/la101795s.
ED

[14] D. Cheng, Y. Wen, L. Wang, X. An, X. Zhu, Y. Ni, Adsorption of polyethylene

glycol (PEG) onto cellulose nano-crystals to improve its dispersity, Carbohydr.
Polym. 123 (2015) 157–163. doi:10.1016/j.carbpol.2015.01.035.
PT

[15] K.A. Iyer, G.T. Schueneman, J.M. Torkelson, Cellulose nanocrystal/polyolefin

biocomposites prepared by solid-state shear pulverization: Superior dispersion
CE

leading to synergistic property enhancements, Polym. (United Kingdom). 56

(2015) 464–475. doi:10.1016/j.polymer.2014.11.017.
AC

[16] B. Lindman, G. Karlström, L. Stigsson, On the mechanism of dissolution of

cellulose, J. Mol. Liq. 156 (2010) 76–81. doi:10.1016/j.molliq.2010.04.016.
[17] D. Li, X. Xu, J. Xu, W. Hou, Poly(ethylene glycol) haired layered double
hydroxides as biocompatible nanovehicles: Morphology and dispersity study,
Colloids Surfaces A Physicochem. Eng. Asp. 384 (2011) 585–591.
doi:10.1016/j.colsurfa.2011.05.012.
[18] H. Zhu, W. Luo, P.N. Ciesielski, Z. Fang, J.Y. Zhu, G. Henriksson, M.E.
Himmel, L. Hu, Wood-Derived Materials for Green Electronics, Biological
Devices, and Energy Applications, Chem. Rev. 116 (2016) 9305–9374.

19
 ACCEPTED MANUSCRIPT

doi:10.1021/acs.chemrev.6b00225.
[19] L. Pietrelli, Effect of MW and pH on poly(ethylene glycol) adsorption onto
carbon, Adsorption. 19 (2013) 897–902. doi:10.1007/s10450-013-9503-x.
[20] D. A.O, Langmuir, Freundlich, Temkin and Dubinin–Radushkevich Isotherms
Studies of Equilibrium Sorption of Zn 2+ Unto Phosphoric Acid Modified Rice
Husk, IOSR J. Appl. Chem. 3 (2012) 38–45. doi:10.9790/5736-0313845.
[21] D. Daltin, Tensoativos: química, propriedades e aplicações, São Paulo: Blucher,
2011. doi:541.34514.

PT
[22] L. Segal, J.J. Creely, A.E. Martin, C.M. Conrad, An Empirical Method for
Estimating the Degree of Crystallinity of Native Cellulose Using the X-Ray

RI
Diffractometer, Text. Res. J. 29 (1959) 786–794.

SC
doi:10.1177/004051755902901003.
[23] T. Sato, T. Fukasawa, K. Aramaki, O. Glatter, R. Buchner, Molecular to
NU
diffusion dynamics and static structures of aqueous micellar solutions: A
SAXS/DLS/DRS study, J. Mol. Liq. 159 (2011) 76–82.
doi:10.1016/j.molliq.2010.06.001.
MA

[24] L. Alves, B. Medronho, F.E. Antunes, M.P. Fernández-García, J. Ventura, J.P.

Araújo, A. Romano, B. Lindman, Unusual extraction and characterization of
ED

nanocrystalline cellulose from cellulose derivatives, J. Mol. Liq. 210 (2015) 106–
112.
[25] M. Khimani, H.W. Tseng, V.K. Aswal, L.J. Chen, P. Bahadur, Salt-assisted
PT

microstructure evaluation of hydrophilic block copolymer F98: A thermal and

scattering study, J. Mol. Liq. 246 (2017) 363–371.
CE

doi:10.1016/j.molliq.2017.09.085.
[26] L. Gajdos, L. Pietrelli, A. Ciccarello, J. Derco, L. Svobodu, Elimination of
AC

Polyethylene Glycol from Aqueous Solution Using Activated Carbon, 16 (2007)

385–388.
[27] X. Zhao, K. Urano, S. Ogasawara, Adsorption of polyethylene glycol from
aqueous solution on montrnorfllonite clays, Colloid Polym. Sci. 267 (1989) 899–
906.
[28] W.B. Arbuckle, M.Y. Osman, Polyethylene Glycol Adsorption Equilibrium on
Activated Carbon, Environ. Eng. Sci. 17 (2000) 147–159.
[29] J. Araki, M. Wada, S. Kuga, Steric stabilization of a cellulose microcrystal
suspension by poly(ethylene glycol) grafting, Langmuir. 17 (2001) 21–27.

20
 ACCEPTED MANUSCRIPT

doi:10.1021/la001070m.
[30] M. Gu, C. Jiang, D. Liu, N. Prempeh, I.I. Smalyukh, Cellulose
Nanocrystal/Poly(ethylene glycol) Composite as an Iridescent Coating on
Polymer Substrates: Structure-Color and Interface Adhesion, ACS Appl. Mater.
Interfaces. 8 (2016) 32565–32573. doi:10.1021/acsami.6b12044.
[31] M.S. Jahan, A. Saeed, Z. He, Y. Ni, Jute as raw material for the preparation of
microcrystalline cellulose, Cellulose. 18 (2011) 451–459. doi:10.1007/s10570-
010-9481-z.

PT
[32] S. Sundararajan, A.B. Samui, P.S. Kulkarni, Shape-stabilized poly(ethylene
glycol) (PEG)-cellulose acetate blend preparation with superior PEG loading via

RI
microwave-assisted blending, Sol. Energy. 144 (2017) 32–39.

SC
doi:10.1016/j.solener.2016.12.056.
[33] Y. Li, R. Liu, Y. Huang, Synthesis and Phase Transition of Cellulose-graft-
NU
Poly(ethylene glycol) Copolymers, J. Appl. Polym. Sci. 110 (2008) 1797–1803.
doi:10.1002/app.
[34] E. hadji B. Ly, J. Bras, P. Sadocco, M.N. Belgacem, A. Dufresne, W.
MA

Thielemans, Surface functionalization of cellulose by grafting oligoether chains,

Mater. Chem. Phys. 120 (2010) 438–445.
ED

doi:10.1016/j.matchemphys.2009.11.032.
[35] G. Siqueira, J. Bras, A. Dufresne, Cellulosic bionanocomposites: A review of
preparation, properties and applications, Polymers (Basel). 2 (2010) 728–765.
PT

doi:10.3390/polym2040728.
[36] H. Sun, X. Shao, Z. Ma, Effect of Incorporation Nanocrystalline Corn Straw
CE

Cellulose and Polyethylene Glycol on Properties of Biodegradable Films, J. Food

Sci. 81 (2016) E2529–E2537. doi:10.1111/1750-3841.13427.
AC

[37] G.R. Shetty, B.L. Rao, S. Asha, Y. Wang, Y. Sangappa, Preparation and
characterization of silk fibroin/hydroxypropyl methyl cellulose (HPMC) blend
films, Fibers Polym. 16 (2015) 1734–1741. doi:10.1007/s12221-015-5223-z.
[38] H. Moustafa, C. Guizani, A. Dufresne, Sustainable biodegradable coffee grounds
filler and its effect on the hydrophobicity, mechanical and thermal properties of
biodegradable PBAT composites, J. Appl. Polym. Sci. 134 (2017) 1–11.
doi:10.1002/app.44498.
[39] M.E. Achaby, Z. Kassab, A. Barakat, A. Aboulkas, Alfa fibers as viable
sustainable source for cellulose nanocrystals extraction : Application for

21
 ACCEPTED MANUSCRIPT

improving the tensile properties of biopolymer nanocomposite films, Ind. Crop.

Prod. 112 (2018) 499–510. doi:10.1016/j.indcrop.2017.12.049.
[40] J. Araki, M. Wada, S. Kuga, Steric Stabilization of a Cellulose Microcrystal
Suspension by Poly ( ethylene glycol ) Grafting, (2001) 21–27.
[41] H. Khanjanzadeh, R. Behrooz, N. Bahramifar, W. Gindl-Altmutter, M. Bacher,
M. Edler, T. Griesser, Surface chemical functionalization of cellulose
nanocrystals by 3-aminopropyltriethoxysilane, Int. J. Biol. Macromol. 106 (2017)
1288–1296. doi:10.1016/j.ijbiomac.2017.08.136.

PT
[42] C.F. Liu, J.L. Ren, F. Xu, J.J. Liu, J.X. Sun, R.C. Sun, Isolation and
characterization of cellulose obtained from ultrasonic irradiated sugarcane

RI
bagasse, J. Agric. Food Chem. 54 (2006) 5742–5748. doi:10.1021/jf060929o.

SC
[43] D.O. Castro, J. Bras, A. Gandini, N. Belgacem, Surface grafting of cellulose
nanocrystals with natural antimicrobial rosin mixture using a green process,
NU
Carbohydr. Polym. 137 (2016) 1–8. doi:10.1016/j.carbpol.2015.09.101.
[44] A. Kaboorani, B. Riedl, Surface modification of cellulose nanocrystals (CNC) by
a cationic surfactant, Ind. Crops Prod. 65 (2015) 45–55.
MA

doi:10.1016/j.indcrop.2014.11.027.
[45] R. da Silva, M.R. Sierakowski, H.P. Bassani, S.F. Zawadzki, C.L. Pirich, L. Ono,
ED

R.A. de Freitas, Hydrophilicity improvement of mercerized bacterial cellulose

films by polyethylene glycol graft, Int. J. Biol. Macromol. 86 (2016) 599–605.
doi:10.1016/j.ijbiomac.2016.01.115.
PT

[46] L. Goetz, M. Foston, A.P. Mathew, K. Oksman, A.J. Ragauskas, Poly(methyl

vinyl ether-co-maleic acid)-Polyethylene glycol nanocomposites cross-linked in
CE

situ with cellulose nanowhiskers, Biomacromolecules. 11 (2010) 2660–2666.

doi:10.1021/bm1006695.
AC

[47] P. Uschanov, L.S. Johansson, S.L. Maunu, J. Laine, Heterogeneous modification

of various celluloses with fatty acids, Cellulose. 18 (2011) 393–404.
doi:10.1007/s10570-010-9478-7.
[48] M.N. Belgacem, A. Gandini, Surface modification of cellulose fibres, Polímeros.
15 (2005) 114–121. doi:10.1590/S0104-14282005000200010.
[49] F. V Ferreira, I.F. Pinheiro, R.F. Gouveia, G.P. Thim, L.M.F. Lona,
Functionalized cellulose nanocrystals as reinforcement in biodegradable polymer
nanocomposites, Polym. Compos. (2017) 1–21. doi:10.1002/pc.24583.
[50] E.E. Pérez, S. Domenek, N. Belgacem, C. Sillard, J. Bras, Green process for

22
 ACCEPTED MANUSCRIPT

chemical functionalization of nanocellulose with carboxylic acids,

Biomacromolecules. 15 (2014) 4551–4560. doi:10.1021/bm5013458.
[51] T. Kaldéus, M. Nordenström, A. Carlmark, L. Wågberg, E. Malmström, Insights
into the EDC-mediated PEGylation of cellulose nanofibrils and their colloidal
stability, Carbohydr. Polym. 181 (2018) 871–878.
doi:10.1016/j.carbpol.2017.11.065.
[52] G. Nabiyouni, A. Barati, M. Saadat, Surface Adsorption of Polyethylene Glycol
and Polyvinyl Alcohol with Variable Molecular Weights on Zinc Oxide

PT
Nanoparticles, Iran. J. Chem. 8 (2011) 20–30.

RI
SC
NU
MA
ED
PT
CE
AC

23
 ACCEPTED MANUSCRIPT

Highlights

 PEG was successfully adsorbed onto cellulose nanostructures (CNS)

surface in solution
 PEG dosage of 2% showed the best results for improving CNS
dispersity
 Stabilization with PEG is an environmentally friendly methodology
to improve CNS stability
 CNS crystallinity and morphology were not affected by the

PT
adsorption of PEG

RI
SC
NU
MA
ED
PT
CE
AC

24