Synthetic Metals 206 (2015) 154–161

Contents lists available at ScienceDirect

Synthetic Metals
journal homepage: www.elsevier.com/locate/synmet

Reduced graphene oxide filled cellulose films for flexible temperature

sensor application
Kishor Kumar Sadasivuni, Abdullahil Kafy, Hyun-Chan Kim, Hyun-U Ko,
Seongcheol Mun, Jaehwan Kim *
Center for EAPap Actuator, Dept. of Mechanical Engineering, Inha University, 253 Yonghyun-Dong, Nam-Ku, Incheon 402-751, South Korea

A R T I C L E I N F O A B S T R A C T

Article history: Graphene based polymer nanocomposites are widely used in flexible electronics and in sensing
Received 28 January 2015 technology. However, ecofriendly materials with utmost performances are still under investigation. Here
Received in revised form 1 May 2015 we report a simple, flexible, partially transparent and efficient temperature sensor fabricated from
Accepted 16 May 2015
composite film of reduced graphene oxide (rGO) and cellulose. The rGO nanosheets are obtained by
Available online 2 June 2015
partial reduction of graphene oxide through low-temperature multi-step annealing. The relative
capacitance of the sensor changes as a function of temperature for all thermally reduced rGO filled
Keywords:
composites. Upon temperature stimulation, charges form at the rGO–cellulose interface. This mechanism
Reduced graphene oxide
Cellulose
allows the composite films to act as high performance capacitive sensors capable of detecting
Polymer-filler interaction temperature variations. The functional properties of the reusable, ecofriendly and inexpensive rGO–
Temperature sensing cellulose device, allow for potential temperature sensor application, which is one among the best ever
reported.
ã2015 Elsevier B.V. All rights reserved.

1. Introduction these parameters, control systems that keep temperature within

Cellulose, the most abundant natural polymer on earth, is an
inexhaustible source of raw material for scores of ecofriendly and
biocompatible products [1]. Cellulose and its derivatives have been
proposed for a variety of applications including coatings,
laminates, optical films, pharmaceuticals, food and textiles [2–
8]. It is also considered as a promising base material for flexible
sensors due to the advantage of its low cost, light-weight, relatively
high thermal stability, high sorption capacity, alterable optical
appearance and biocompatibility [9,10].
Among sensors that detect several environmental parameters
such as strain, pressure, temperature as well as the presence of
foreign particles, the temperature detection has much significance
in most of the physical, electronic, chemical, mechanical and
biological systems [11–14]. The sensing property of temperature
sensors is marked by a sudden increase/decrease in resistivity or
capacitance with temperature indicated by a positive/negative
temperature coefficient [15]. The foremost characteristics a
temperature sensor system should possess are high sensitivity,
wide-operating temperature range, long-term stability against
other ambient influences such as humidity. In order to optimize
* Corresponding author. Tel.: +82 32 831 7325; fax: +82 32 832 7325.
E-mail address: jaehwan@inha.ac.kr (J. Kim).
specified limits are often used [1]. The main criteria of selecting a
material for a perfect temperature sensor are its accuracy, faster
response, medium independent nature and easily conditioned
output. Polymeric sensor devices have an additional advantage
that the heating/sensing element can be principally in any shape
based on the nature of polymer [1,11,13]. Designing temperature
sensors with small hysteresis, good linearity, low temperature
coefficient and short response time is a great challenge, and
nowadays advanced nanoparticles especially conducting carbon
are employed for it [1,15–17]. For example, two dimensional
graphene nanosheets and their derivatives have been studied for
various sensor devices due to its ultra-high specific surface area
(
2630 m2/g) and excellent electronic properties [18,19].
Owing to the difficulties in developing sensing materials, the
current main strategy is to fabricate innovative structural
behaviors from naturally available materials like cellulose. The
presence of large number of hydroxyl groups along its skeleton
causes the formation of comprehensive networks of intra- and
inter-molecular hydrogen bonds in cellulose and makes two
structural regions – crystalline and amorphous – in the same
polymer [8]. This is helpful in obtaining flexible biomaterials in
which polar fillers are uniformly bonded with the neat matrix. The
polarity is induced by the presence of functional groups in filler
particles. For instance, graphene oxide (GO) [20–22] contains a
large number of functional groups such as hydroxyl, epoxides,

http://dx.doi.org/10.1016/j.synthmet.2015.05.018
0379-6779/ ã 2015 Elsevier B.V. All rights reserved.
 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161
carboxyls and carbonyls on its surface. Such polar graphitic
derivatives filled cellulose composites with very high mechanical,
thermal and gas permeability properties are reported [2–7,22–24].
However the low conductivity of GOs does not allow the
fabrication of electronic devices and reduced GOs are largely used
for those purposes.
Here we introduce a new type of temperature sensor device
where cellulose forms the base matrix and the thermally reduced
GO (rGO) as the filler. Though rGOs are synthesized by various
means, [23–27] thermal reduction [13,25–26] provides enhanced
electrical conductivity, higher specific surface area, less agglomer-
ation and higher dispersion rates [28–35]. In this way we realized a
new temperature sensitive material where the polymer chains
exfoliate into the interlayers [31] of thermally reduced rGOs. The
reduction temperature of GO is tuned to institute enough
conjugated and non-conjugated filler networks within the matrix
so as to derive the best composite material. The morphology and
structural property relationship of the cellulose/rGO composite is
determined with the help of Fourier transformation infrared
spectroscopy (FTIR), X-ray diffraction (XRD), scanning electron
microscopy (SEM) and Raman spectral analyses. A mechanism
involving tunneling of charge carriers is proposed between the
cellulose and thermally reduced rGOs (reduced at 50, 100, 200 and
400  C) for temperature stimulation. The sensing is correlated with
the dielectric behavior as well as the molecular level interactions
based on the reduction temperature of GO. Hence a flexible and
optically translucent temperature sensor capable to store energy
and to regulate the device temperature by detecting high-
temperature conditions is realized out of the cellulose/rGO
composite film.
2. Experimental details
2.1. Materials
Cotton pulp of 98% purity and degree of polymerization
4500 was obtained from Junsei Chemicals, South Korea. Natural
flake graphite, and other reagents such as dimethyl acetamide
(DMAc), LiCl, H2SO4, HNO3, HCl, KMnO4 and 30% H2O2 solution
were procured from Sigma-Aldrich. The 99.5% isopropyl alcohol
was purchased from Daejung, South Korea.
2.2. Sample preparation
First, a cellulose solution was prepared by dissolving cotton
pulp with LiCl/DMAc solvent according to the previously reported
method [1]. 2/8/90 mixture of cotton pulp (heated with LiCl to
remove adsorbed water molecules), LiCl and anhydrous DMAc
(99%; Aldrich) were heated at 155  C with mechanical stirring for
4 h in order to obtain a solution. This cellulose solution was
centrifuged at 11,000 rpm to eliminate the undissolved cellulose
fibers
Second, thermal reduction of GO (obtained by improved GO
synthesis [36]) to rGO was carried out prior to composite
preparation. For thermal reduction, nitrogen gas was passed
through GO for 30 min in a sealed container to create an inert
Table 1
Samples prepared for the analysis.
Sl. no. Notation Cellulose (wt%) rGO 50 (wt%)
1 Cellulose 100 – –
2 CrGO50 99 1 –
3 CrGO100 99 – 1
4 CrGO200 99 – –
5 CrGO400 99 – –
155
atmosphere on its surface and the resulting product was placed in a
pre-heated muffle furnace at different temperatures 50, 100,
200 and 400  C for 60 min.
Third, rGOs (reduced at different temperatures) were dispersed
in DMAc by 3 h of mechanical stirring with stirring speed 500 rpm,
followed by mixing with the cellulose solution. This composite
solution was casted on a glass plate using doctor blade and cured
using a deionized (DI) water and isopropyl alcohol mixture with
the ratio of 4:6 for 3 h. This slow curing process prevents
aggregation of rGOs and eliminate remnant LiCl and DMAc [37].
Finally the cured composite was rinsed in DI water bath twice and
dried in a vacuum oven at 60  C. Table 1 shows the various
composite samples used in this study.
2.3. Composite characterization
Optical transmittance of the samples was studied using UV–
visible spectrophotometer. For this the spectra of the films in the
range of 200–800 nm wavelength were recorded with a Hewlett
Packard (8452A) diode array. Cold FE-SEM images of the composite
films were taken with a cold field emission scanning electron
microscope (JEOL JSM-6400F) to study the sample morphology.
The AFM and TEM images of the synthesized GO are given in
Supplementary information (Fig. S1), from which the thickness
and flake size of the platelets were determined.
FTIR spectrometer (Bruker Optics, Billerica, MA) was used to
characterize the chemical composition of the samples under
vacuum. The FTIR spectra obtained were in the range of 500–
4000 cm1 by averaging 16 scans at 1 min intervals to minimize the
effects of dynamic scanning. The resolution was 4 cm1. Raman
spectra were checked using a Raman spectrometer (Bruker, RFS
100/S) where A YAG laser with a wavelength of 1064 nm was used
as excitation source, with a laser power of 130 mW. Here also the
effect of dynamic scanning was minimized by averaging 20 scans at
1 min intervals and the spectra in the range of 100–3500 cm1
were recorded. XRD patterns were checked with a thin film X-ray
diffractometer using CuKa target radiation at 40 kV and 50 mA, at a
scanning rate of 0.015 /min. The diffraction angle was varied from
5 to 40 .
2.4. Dielectric and sensing characterization
The temperature responses of the composite films were also
recorded as the capacitance responses using the LCR meter at
different temperature levels. The measurement frequency was
maintained to be 500 Hz and the oscillation level 1 V. Each sample
was analyzed and 5–10 measurements were taken at each
temperature. The sample dimension was maintained as 3cm  3
cm  0.25 mm. Sensor data is plotted as normalized capacitance
change versus time curves. This relative capacitance (Ac) measure-
ment of the sensor is expressed by Eq. (1).
C  Co
AC ¼ ; (1)
Co
where C is the film capacitance in the presence of temperature and
C0 the initial capacitance. Finally the electrical properties of the
rGO 100 (wt%) rGO 200 (wt%) rGO 400 (wt%)
– –
– –
– –
1 –
– 1
156 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161
aluminum coated samples were also investigated in terms of
current–voltage (I–V) measurements. The changes in semiconduc-
tor behavior of cellulose/rGO upon reduction of GOs were checked
at different temperatures.
3. Results and discussion
3.1. Morphology-structure relationship
Fig. S2 illustrates the effect of thermal reduction of rGOs on the
transmittance properties of cellulose. The average transmittance
values for all the samples, cellulose, CrGO50, CrGO100,
CrGO200 and CrGO400 films in the UV–visible range are 88, 41,
22, 16 and 12% respectively at 500 nm. The influence of reduction
temperature on the morphology and composite structure is further
analyzed by SEM and spectroscopic techniques.
The cross sectional SEM pictures of the samples in Fig. 1 show
considerable filler effect on the composite morphology, whereas
the temperature reduction has little effect on dispersion. Cellulose
film possesses a rough morphology (Fig. 1a) and the roughness
reduced upon the addition of rGOs. The very smooth and well bind
surface in the composites indicates good reinforcement between
the rGO filler and cellulose matrix (Fig. 1b–e). As no aggregation of
rGO platelets is shown in the figures, it is proposed that the filler
flakes are well dispersed in cellulose at 1% loading.
Spectral analysis has very significance in exploring the
composite characteristics, and the structural composition was
studied here with the help of FTIR, EDS, Raman and XRD
techniques. The FTIR spectra for the rGOs reduced at different
temperatures as well as the composites are given in Fig. 2. rGOs
show C O stretching vibration at 1209 cm1, O H stretching at
3459 cm1, CO at 1060 cm1 and C¼O stretching at 1711 cm1
(Fig. 2a). The effect of thermal reduction is very clear from the
spectra as the intensity of O H (3459 cm1) group reduces by de-
oxydation with the increase in reduction temperature. At higher
temperatures, the C O and C¼O stretching vibrations in rGO
become almost indistinguishable [38,39]. This is due to the
variations in the number of ionic groups on the surfaces of rGOs
depending on the extent of GO thermal reduction. Fig. 2b shows
the characteristic FTIR spectra of cellulose and its composites. The
bands at around 3418 and 1607 cm1 are respectively due to the
hydroxyl stretching and carboxyl bending modes of vibration and
Fig. 1. SEM images of (a) cellulose, (b) CrGO50, (c) CrGO100, (d) CrGO200 and (e) CrGO400.
the peak at 1375 cm1 corresponds to the COH bond of
cellulose. In addition, bands at 1137 cm1 are due to the C H
bending modes and the peaks at 2932 cm1 denote the character-
istic CH stretching of  CH2 group (from cellulose). The similar
peaks in all composites and cellulose indicate the absence of any
kind of chemical interactions between rGO and cellulose when
blended. The peak at 3418 cm1 coming from the stretching
vibration of OH groups (from both cellulose and rGO) diminishes
from CrGO50 to CrGO400 since increase in reduction temperature
removes the hydroxyl functional groups in rGOs. However the
formation of an intercalated/exfoliated structure is proposed
where weak intermolecular interactions connect the filler-
polymer molecules.
The energy-dispersive X-ray spectroscopy (EDS) analysis of the
samples (Table 2) shows increased C/O ratio for cellulose/rGO
composites compared to neat cellulose. CrGO400 gives the highest
C/O value confirming the less number of oxide groups on the
surface of rGO400. Also a steady decrease in C/O is observed
starting from CrGO50 according to the reduction temperatures of
corresponding rGOs in the composites and these results were
further confirmed by thermo gravimetric analysis (Supplementary
information Fig. S3).
Fig. S4 shows the Raman spectra of rGOs and cellulose/rGO
samples with D and G band respectively located around
1307 and

1585 cm1 and reflect the significant structural changes occur-
ring during thermal exfoliation. The Raman spectra intensity ratio
of the D to G bands (ID/IG) of fillers follows as 1.104, 1.252, 1.287 and
1.351 for rGO50, rGO100, rGO200 and rGO400, respectively. The
removal of oxygen functional groups from GOs upon reduction
creates graphene networks with smaller average size compared to
original GO and this enhances ID/IG. The ID/IG values of cellulose and
cellulose/rGO composites obtained from Fig. S4 are given in
Table 2. The CrGO50 and CrGO100 composites show almost similar
Raman spectra and CrGO400 shows the highest ID/IG ratio
following a similar trend of rGO400.
The XRD patterns of rGOs and its composites in Fig. 3 furnish
interesting facts about composite microstructure. Peak I and Peak
II marked in Fig. 3a correspond to the (0 0 2) reflection in graphite
and its derivative (rGO). The diffraction peak of pristine graphite at
2u = 26.52 determines an interlayer distance, d, of 3.354 Å [10].
The rGO50 peak gradually shifts toward right with the temperature
reduction. A strong and sharp peak I near 2u = 11 in rGO50 similar
 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161 157

Fig. 2. FTIR analysis of (a) rGOs reduced at different temperatures, (b) cellulose/rGO composites and cellulose.

Table 2 The crystalline phase of cellulose usually contributes a few peaks at

Structural parameters and elemental composition of the cellulose/rGO composites. 15.1, 16.8 and 20.5 but here a merged broad peak is observed at
lower angle as reported [41] and is assigned to the (1 0 1)
Samples C/O Ratio ID/IG Crystallinity (%)
crystalline plane. The crystallinity index (CI) was calculated from
Cellulose 0.78 – 49.2
the XRD data using Eq. (2) [42].
CrGO50 1.06 1.121 51.3
CrGO100 1.15 1.157 54.6
X
ACrys
CrGO200 1.30 1.172 56.5
CIXRD ð%Þ ¼ X P  100; (2)
CrGO400 1.48 1.188 57.8 ACryst þ Aamorp

where Acryst is the area of the crystalline peak at 22 and Aamorp the
area of the amorphous peaks, as shown in Fig. 3b. The crystallinity
to GO gives a d value of 8.0854 Å [21,40]. In the 100–200  C zone for
values of cellulose and cellulose/rGO composites are compared in
rGO100 and rGO200, both peaks I and II appear (Fig. 3a). The peak I
Table 2. The crystallinity values are found to vary from CrGO50 to
has higher intensity than peak II in rGO100, indicating the presence
CrGO400 because of the variations in nucleating capacity of fillers
of more number of oxide groups. In the case of rGO200 the peak II is
present in the cellulose. These results were in good connection
more intense, pointing toward its more reduced plane structure.
with differential scanning calorimetric results as shown in Fig. S6.
The reduced peak II intensity above 200  C marks an intensive
thermal exfoliation of GOs and reduction in the number of
3.2. Dielectric behavior
graphene layers. The structure of rGO400 is proposed to be long-
tailed because of the least peak intensity (2u = 23–24 ) and thus
In order to check the dielectric response of the samples, the
maximum interplanar distance (approaching toward graphite).
dielectric constant, e0 , of the cellulose matrix and its composites
These observations were further confirmed by UV–visible spectra
were measured with 20–10,000 Hz frequency range at 25  C and
given in Fig. S5.
25% relative humidity (RH). Fig. 4 shows the results. At low
Fig. 3b gives information about the crystalline structure and
frequency, 20 Hz, dielectric constant value increases from 980 for
dispersion of rGOs in cellulose. The broad diffraction peak at 19
CrGO50 to 2200 for CrGO100 and decreases to 1230 for
indicates the partly crystalline nature of cellulose. Overall, the XRD
CrGO200 and 620CrGO400. The dielectric constant of
diagram of all composites is similar to cellulose since the rGO
CrGO100 sample shows 44 times enhancement from that of the
content is too small to produce great variation in the peak.
neat cellulose (e0 = 50 at 20 Hz). CrGO100 exhibits the maximum
However, the peak corresponding to rGO disappears in the XRD
dielectric constant. As increasing the frequency, dielectric constant
spectra of cellulose/rGO composites confirming the significant
values of all cases decrease. At 1 kHz, dielectric constant of
reduction of GO and the exfoliation of the layered GO nanosheets.

Fig. 3. XRD spectra of (a) rGOs reduced at different temperature levels: (1) = 50  C, (2) = 100  C, (3) = 200  C, (4) = 400  C and (5) = graphite, (b) cellulose/rGO composites and
cellulose.
158 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161

rGO100 sheet surface which forms a capacitor network and thus

causes optimized electrical transmission (also evident from the I–V
curve in Fig. 6). In general, the sp2 carbon atoms from the p bonds
as well as the functional groups on rGO surface acts as passage for
the charge to move to the internal surface during the charging/
discharging processes. These functional groups also cause fast
redox processes at or near the electrode surface, which can provide
the pseudo-capacitance [43]. In the CrGO50 case, though enough
functional groups are present, the lack of sufficient sp2 carbon
atoms results in lower capacitance compared to CrGO100.
Upon increase in reduction temperature, the electron carrier
density of GO films increases due to p-doping and large number of
p bonds. High reduction temperature causes the dissipation of
charges accumulated at the interface between the rGOs and
cellulose, and this result in the decreased dielectric constant and
Fig. 4. Variation of dielectric constant of cellulose/rGO composites and pure
the composites behave as conductors. Thus, CrGO400 behaves like
cellulose with frequency variation at 1 V (RH = 25% and T = 25  C)
a semi-metal as explained in the I–V curve later on.
CrGO100 is 100, which is still 4 times higher than that of the neat
3.3. Temperature sensing
cellulose, 25. e0 of microelectronic polymers depends on the
amount of mobile (polarizable) electrical charges and the degree of
Fig. S7a schematically illustrates the key processes in fabricat-
mobility of these charges [35,43].
ing and operating a cellulose/rGO composite temperature sensor.
According to Maxwell–Wagner–Sillars (MWS) process poly-
Both sides of the composite film were coated with aluminum
mer-filler interfacial (like donor–acceptor complexes) interaction
electrodes by a thermal evaporation prior to the experiment.
is necessary to get changes in the composite dielectric properties.
Temperature was applied to the samples by keeping the heat
The huge interfacial area of composite provides numerous sites for
element with a distance of 5 mm. Fig. S7a gives the temperature vs
the reinforced MWS effect [44]. In the MWS effect, when a current
capacitance plots and Fig. S7b highlights the drastic increase in
flows across the two-material interfaces, charges accumulate at
capacitance with increasing the reduction temperature from 50 to
the interface between two dielectrics with different relaxation
100  C and further decrease at higher temperature.
time (t = e/s , where s is the conductivity). The dielectric
Cyclic temperature sensing curves represented in Fig. 5 explain
measurement also shows that the incorporation of the rGOs in
the increase in capacitance with the exposing temperature starting
the cellulose matrix increases e0 but leads to a gradual decrease for
from 25 to 40, 60 and 80  C, whereas in the same figure cooling
CrGO400. When the values are compared, CrGO100 possesses the
cycles show capacitance decrease (Fig. 5a–c). It is found that the
best dielectric constant value which is attributed to the motion of
capacitance of CrGO100 sample increased 6 times as moving from
free charge carriers due to interfacial polarization effects. This is
25 to 80  C compared to the neat cellulose. The sensitivity limit of
possible by the presence of moderate number of oxide functional-
CrGO100 obtained by checking the connection between the
ities (charge centers) and conjugation (carrier centers) on the
amplitude of the sensor signal with exposed temperature is

Fig. 5. Temperature sensing of cellulose/rGO composites and pure cellulose at (a) 40  C, (b) 60  C, and (c) 80  C at RH = 25% and (d) sensitivity limit of CrGO100 composite.
 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161
Fig. 6. I–V curves of (a) cellulose, (b) CrGO50, (c) CrGO100, (d) CrGO200 and (e) CrGO400 at RH = 25%.
shown in Fig. 5d. A nearly linear evolution of sensor response with
the temperature change decreasing from 20 to 3  C is obtained,
whereas noisy signals indistinguishable from the base line are
observed below 3  0.5  C.
3.4. Working mechanism of the sensors
To understand the underlying mechanism of the sensor devices
the capacitance change and the temperature variations are
correlated. The polymer networks are exfoliated between the
rGO sheets as explained in the morphology part. Since the
capacitance is a measure of the distance between the charge
carriers (here rGOs) the given composite sample possesses an
initial capacitance at room temperature. Upon heating the distance
of separation of filler particles increases and as a result a significant
change in capacitance is observed. This is happening in all
cellulose/rGO samples upon temperature stimulation. Also cooling
below decomposition temperature of the sample can recover the
capacitance to its original value. Thus temperature sensing occurs
based on the capacitance change.
CrGO100 displayed excellent temperature sensing and revers-
ibility with large deviation between heating and cooling capaci-
tance values due to the sufficient interaction of rGO100 with
cellulose and interfacial polarization effects. The oxygen-based
moieties and structure defects on rGO100 are electron-
159
withdrawing in nature and promote holes in the valence band
of rGO100, leading to the p-type semiconducting behavior in the
resultant CrGO100 [45,46] and play an important role in its sensing
property. The sensing ability increases from CrGO50 to
CrGO100 and decreases from CrGO100 to CrGO400 due to a
balanced state of oxide and conjugated moieties in CrGO100. As the
temperature level is changed, cellulose also exhibits capacitance
change since it is hydrophilic and absorbs humidity from the
surroundings. When temperature raises the amount of the
adsorbed water molecules (as observed from the TGA in Fig. S3)
reduces, which attributes to the increased capacitance. In the case
of composites in addition to this moisture adsorption the nature of
fillers (dispersion) also affects the capacitance change. We
observed that these temperature sensor response time and change
in capacitance are dependent on the size and mass of the samples,
practically. The working temperature of the sensors was examined
up to 80  C, but is expected to be useful up to 200  C and beyond
this level the material starts thermal degradation. The thermal
stability of the composites examined is included in the supporting
information (Fig. S3).
3.5. Electrical properties
In order to determine the potential of cellulose/rGO composites
for fabricating temperature sensors, its electrical properties must
160 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161
also be investigated. The electrical parameters obtained using
different techniques are found in reasonable agreement with other
characterizations done. It is observed that I–V characteristics of
cellulose and its composites show partially symmetrical behavior
with respect to the polarity induced by the applied voltage. The I–V
characteristics of all samples shown in Fig. 6a–e exhibit a nonlinear
and asymmetric response except CrGO400. This nonlinear behav-
ior in current intensity with the applied voltage is explained by the
conduction mechanism. Here the charge conduction is not only
governed by free carriers (electrons & holes) such as in intrinsic
semiconductors but also by the formation of polarons (combina-
tion of electron and thermal defects) and bipolarons [47]. Upon
increase in the applied voltage, the formation of polarons and
bipolarons increases rapidly contributing to higher current values
in the sample. The intensity values of cellulose, CrGO50,
CrGO100 and CrGO200 are independent at low voltage, but it
begins to increase exponentially once an appropriate value is
reached. This is due to the decrease in the width of depletion layer
at the junction of the filler-polymer, filler-filler and polymer-
polymer. At higher voltages, the I–V dependence is almost linear
because of the minimized depletion layer at the interface and
cellulose behaves as series of resistances.
In the CrGO400 case, the linear I–V curve obtained (Fig. 6e)
indicates Ohmic contact between the sample and the electrode.
The carbon–oxygen saturated bonds in GO change to the extended
unsaturated conjugated network as a result of high temperature
reduction. This saturated to unsaturated conjugated network
conversion is the highest in the case of CrGO400 film resulting in
high Ohmic behavior and this composite showed greater linearity
with a small band gap but no improvement in ionic conductivity
due to lower level of ionic character and filler concentration. In the
case of other composites sufficient oxide groups tightly bound each
other and electron transfer occurs easily between adjacent rGOs.
Fig. 6c shows the behavior of CrGO100 and this confirms the
significant level of interaction of cellulose with rGO100.
The resistance values of all composite films decrease signifi-
cantly and steadily with temperature at 25, 40, 60 and 80  C due to
enhancement of hopping or tunneling activated by the thermal
energy as seen in Fig. 6. For better understanding of the
phenomenon, a theoretical model was applied on the conduction
behavior of CrGO400 in the Ohmic regime. For this, the emission
current density, I, is related to the bias voltage, V, through the
combination of a power term and an exponential term, as given in
Eq. (3) [48]
 
B
I ¼ AV n exp  ; (3)
V
where A,B and n (1  n  3) are constants, where A and B are
respectively proportional to the tunneling frequency and the
average inter-particle separation.
The modeling explains the conduction mechanism as an
electron emission process, probably as a tunneling of electrons
from one particle to the next particle inside the cellulose matrix.
The I–V curves of CrGO400 fit well with this model as shown in
Fig. 6e. Here, the values of the power terms, n and B, were
maintained as 1 and 0.001, respectively, and the value of tunneling
frequency, A, varies to 1 107, 2.8  107, 8.8  107 and 1.5  106
at 25, 40, 60 and 80  C temperature levels. This indicates the faster
delocalization of thermally activated electrons in the sample so as
to increase the tunneling frequency with temperature. Thus, these
I–V curves give strong evidence for capacitance changes in the
material at different temperature levels. All the assumptions and
observations point toward the fact that the optimized thermal
reduction of GO is necessary to make cellulose composites suitable
for excellent capacitors, and the high dielectric constant composite
CrGO100 is an ideal candidate for fabricating capacitive sensors.
A commercial temperature sensor usually contains copper/
nickel/platinum metal wires connected together and the mecha-
nism of it involves conversion of thermal energy into electrical
energy. Apart from these devices, the cellulose/rGO film developed
is easy to handle, stable over time and resistant to contamination
with fastest response. Corrosion effect and resistance wire
problems are totally absent here. However, low sensitivity to very
small (DT < 3  0.5  C) and limited temperature range (T > 200  C)
and less accuracy in the presence of environmental conditions like
humidity are still under investigation and will be examined in
future in order to industrialize the product. The anticipated areas
of applications include computer processors, electronic test
equipment, cellular phones, food processing equipment, medical
field, industrial heat treating, packaging, air conditioning servicing
and environmental monitoring [49–51].
4. Conclusions
In this work, cellulose/rGO composite films were successfully
fabricated and their morphology and structural properties were
determined by FTIR, XRD, SEM and Raman spectra. The level of
thermal reduction of GOs was optimized and derived to be 100  C.
CrGO100 has the highest dielectric constant of 2200 at 20 Hz
depending on the extent of thermal reduction. Capacitance of the
cellulose/rGO samples monotonically increased with increasing
the exposed temperature and the temperature sensing character-
istics were explained by the principle of tunneling conduction.
Making high performance temperature sensors from cellulose/rGO
composite film by tuning the rGO reduction temperature is among
the future opportunities in flexible electronics. The proposed
temperature sensor was characterized between 25 and 80  C, it
showed a linearity with respect to the temperature change. With
advantages of cellulose/rGO composites, ecofriendly devices can
be made. We believe that such devices could find a wide range of
applications in electronic equipment as well as in wearable devices
that can be embedded in clothing.
Acknowledgement
This work was supported by the National Research Foundation
of Korea(NRF-2013M3C1A3059586).
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.
synthmet.2015.05.018.
References
[1] S.K. Mahadeva, S. Yun, J. Kim, Sens. Actuators A 165 (2011) 194.
[2] Y. Zhang, Y. Liu, X. Wang, Z. Sun, M. Junkui, T. Wu, F. Xing, J. Gao, Carbohydr.
Polym. 101 (2014) 392.
[3] M.U.M. Patel, N.D. Luong, J. Seppala, E. Tchernychova, R. Dominko, J. Power
Sources 254 (2013) 55.
[4] X. Zhang, X.X. Liu, W. Zheng, J. Zhu, Carbohydr. Polym. 88 (2012) 26.
[5] Y. Cheng, B. Feng, X. Yang, P. Yang, Y. Ding, Y. Chen, J. Fei, Sens. Actuators B 182
(2013) 288.
[6] K.K. Sadasivuni, A. Kafy, L. Zhai, H.U. Ko, S. Mun, J. Kim, Small (2014) , doi:
http://dx.doi.org/10.1002/smll.201402109.
[7] D. Han, L. Yan, W. Chen, W. Li, P.R. Bangal, Carbohydr. Polym. 83 (2011) 966.
[8] A. Kafy, K.K. Sadasivuni, H.C. Kim, A. Akther, J. Kim, Phys. Chem. Chem. Phys.
(2015) , doi:http://dx.doi.org/10.1039/C4CP05921B.
[9] H. Grange, C. Bieth, H. Boucher, G. Delapierre, Sens. Actuators 12 (1987) 291.
[10] P. Lian, X. Zhu, S. Liang, Z. Li, W. Yang, H. Wang, Electrochim. Acta 55 (2010)
3909.
[11] J. Kunzelman, T. Chung, P. Matherx, C. Weder, J. Mater. Chem. 18 (2008) 1082.
[12] M. Mohiuddin, K.K. Sadasivuni, S. Mun, J. Kim, RSC Adv. (2015) , doi:http://dx.
doi.org/10.1039/C5RA03043A.
[13] S. Wang, D.D.L. Chung, Composites Part B 30 (1999) 591.
 K.K. Sadasivuni et al. / Synthetic Metals 206 (2015) 154–161
[14] J.T. Robinson, F.K. Perkins, E.S. Snow, Z. Wei, P.E. Sheehan, Nano Lett. 8 (2008)
3137.
[15] H. Al-Mumen, F. Rao, L. Dong, W. Li, Micro Nano Lett. 8 (2013) 681.
[16] F. Arai, C. Ng, P. Liu, L. Dong, Y. Imaizumi, K. Maeda, H. Maruyama, A. Ichikawa,
T. Kukuda, 4th IEEE Conference on Nanotechnology (2004) 146.
[17] X.J. He, J.H. Du, Z. Ying, H.M. Cheng, X.J. He, Appl. Phys. Lett. 86 (2005) 062112.
[18] Y. Kim, T.K. An, J. Kim, J. Hwang, S. Park, S. Nam, H. Cha, W. Park, J. Baik, C.E.
Park, J. Mater. Chem. C 2 (2014) 4539.
[19] D. Ponnamma, Q. Guo, I. Krupa, M.A.S. Al-Maadeed, K.T. Varughese, S. Thomas,
K.K. Sadasivuni, Phys. Chem. Chem. Phys. (2015) , doi:http://dx.doi.org/
10.1039/C4CP04418E.
[20] K.K. Sadasivuni, D. Ponnamma, S. Thomas, Y. Grohens, Prog. Polym. Sci. 39
(2014) 749.
[21] D. Ponnamma, K.K. Sadasivuni, M. Strankowski, Q. Guo, S. Thomas, Soft Matter
9 (2013) 10343.
[22] K.K. Sadasivuni, M. Castro, A. Saiter, L. Delbreilh, J.F. Feller, S. Thomas, Y.
Grohens, Mater. Lett. 96 (2013) 109.
[23] K.K. Sadasivuni, A. Saiter, N. Gautier, S. Thomas, Y. Grohens, Colloid Polym. Sci.
291 (2013) 1729.
[24] D. Ponnamma, K.K. Sadasivuni, M. Strankowski, P. Moldenaers, S. Thomas, Y.
Grohens, RSC Adv. 3 (2013) 16068.
[25] J. Xia, F. Chen, J. Li, N. Tao, Nat. Nanotechnol. 4 (2009) 505.
[26] M.J. McAllister, J.L. Li, D.H. Adamson, H.C. Schniepp, A.A. Abdala, J. Liu, M.H.
Alonso, D.L. Milius, R. Car, R.K. Prud’homme, I.A. Aksay, Chem. Mater. 19 (2007)
4396–4404.
[27] L. Chen, Z. Xu, J. Li, C. Min, L. Liu, X. Song, G. Chen, X. Meng, Mater. Lett. 65
(2011) 1229.
[28] H.C. Schniepp, J.L. Li, M.J. McAllister, H. Sai, M. Herrera-Alonso, D.H. Adamson,
R.K. Prud'homme, R. Car, D.A. Saville, I.A. Aksay, J. Phys. Chem. B 110 (2006)
8535.
[29] Z. Wang, Y. Xiao, X. Cui, P. Cheng, B. Wang, Y. Gao, X. Li, T. Yang, T. Zhang, G. Lu,
ACS Appl. Mater. Interfaces 6 (2014) 3888.
[30] P. Guo, H. Song, X. Chen, Electrochem. Commun. 11 (2009) 1320.
[31] B. Zhao, P. Liu, Y. Jiang, D. Pan, H. Tao, J. Song, T. Fang, W. Xu, J. Power Sources
198 (2012) 423.
161
[32] N.D. Luong, N. Pahimanolis, U. Hippi, J.T. Korhonen, J. Ruokolainen, L.S.
Johansson, J.D. Nam, J. Seppala, J. Mater. Chem. 21 (2011) 13991.
[33] K. Gao, Z. Shao, X. Wu, X. Wang, J. Li, Y. Zhang, W. Wang, F. Wang, Carbohydr.
Polym. 97 (2013) 243.
[34] W. Ouyang, J. Sun, M. Memon, C. Wang, J. Geng, Y. Huang, Carbon 62 (2013) 501.
[35] K.K. Sadasivuni, D. Ponnamma, B. Kumar, M. Strankowskie, R. Cardinaels, P.
Moldenaers, S. Thomas, Y. Grohens, Composites Sci. Technol. 104 (2014) 18.
[36] D.C. Marcano, D.V. Kosynkin, J.M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L.B.
Alemany, W. Lu, J.M. Tour, ACS Nano 4 (2010) 4806.
[37] S. Yun, Y. Chen, J.N. Nayak, J. Kim, Sens. Actuators B 129 (2008) 652.
[38] L. Shen, H.S. Shen, C.L. Zhang, Mater. Des. 31 (2010) 4445.
[39] P. Pokharel, Q.T. Truong, D.S. Lee, Composites Part B 64 (2014) 187.
[40] V. Alzari, V. Sanna, S. Biccai, T. Caruso, A. Politano, N. Scaramuzza, M. Sechi, D.
Nuvoli, R. Sanna, A. Mariani, Composites Part B 60 (2014) 29.
[41] C.M. Popescu, G. Singurel, M.C. Popescu, C. Vasile, D.S. Argyropoulos, S.
Willford, Carbohydr. Polym. 77 (2009) 851.
[42] L. Segal, J.J. Creely, A.E. Martin, C.M. Conrad, Text. Res. J. 29 (1962) 786.
[43] Q. Du, M. Zheng, L. Zhang, Y. Wang, J. Chen, L. Xue, W. Dai, G. Ji, J. Cao,
Electrochim. Acta 55 (2010) 3897.
[44] J.K. Yuan, S.H. Yao, Z.M. Dang, A. Sylvestre, M. Genestoux, J. Bai, J. Phys. Chem. C
115 (2011) 5515.
[45] L. AI-Mashat, K. Shin, K. Kalantar-Zadeh, J.D. Plessis, S.H. Han, R.W. Kojima, R.B.
Kaner, D. Li, X. Gou, S.J. Ippolito, W. Wlodarski, J. Phys. Chem.C 114 (2010)
16168.
[46] J.D. Fowler, M.J. Allen, V.C. Tung, Y. Yang, R.B. Kaner, B.H. Weiller, ACS Nano 3
(2009) 301.
[47] S.K. Saha, Y.K. Su, C.L. Lin, D.W. Jaw, Nanotechnology 15 (2004) 66.
[48] T. Blythe, D. Bloor, Electrical Properties of Polymers, 2nd ed., Cambridge
University Press, 2005, pp. 270–272.
[49] M.E. Roberts, A.N. Sokolov, Z. Bao, J. Mater. Chem. 19 (2009) 3351.
[50] T. Someya, T. Sekitani, S. Iba, Y. Kato, H. Kawaguchi, T. Sakurai, Proc. Natl. Acad.
Sci. U. S. A. 101 (2004) 9966.
[51] J. Jeon, H.B.R. Lee, Z. Bao, Adv. Mater. 25 (2013) 850.