Available via license: CC BY 4.0
Content may be subject to copyright.
Carbonized Cellulose Nanofibril with Individualized Fibrous
Morphology: toward Multifunctional Applications in
Polycaprolactone Conductive Composites
Ju Dong, Xingyan Huang, Guang-Lin Zhao, Jaegyoung Gwon, Won-Jae Youe, and Qinglin Wu*
Cite This: https://doi.org/10.1021/acsabm.1c00360
Read Online
ACCESS Metrics & More Article Recommendations *
sıSupporting Information
ABSTRACT: Drying cellulose nanofibril (CNF) from aqueous
suspensions often leads to aggregated fibril morphology, negatively
affecting its performance in ensuing applications. In this work, we
introduced a new solvent drying approach to acquire dry CNF from
aqueous suspensions and subsequently pyrolyzed the CNF precursor
to obtain carbonized CNF (CCNF) without loss of its fibrous
morphology. The fibrous CCNF was dispersed homogeneously in
polycaprolactone (PCL) thermoplastic resin, greatly enhancing PCL
composite tensile performance. After being further mixed with
carbon black (CB), the CCNF helped to minimize CB aggregation
due to formation of interconnected three-dimensional (3D)
structures. The CCNF/CB/PCL composite exhibited superior
electrical conductivity ascribed to electrons transporting more
efficiently among CB aggregates. The composite is also suitable for
applications such as 3D printed electromagnetic interference (EMI) shielding and deformation sensing. Specifically, the 3D printed
EMI shielding composite efficiently absorbed EM radiation in the frequency range of 4−26 GHz, and the 3D printed deformation
sensor exhibited excellent sensitivity, durability, and flexibility in monitoring mechanical distortions. Herein, this study sheds light on
the development of multifunctional conductive composites embedded with fibrous CCNF from sustainable resources.
KEYWORDS: cellulose nanofibril, composites, 3D printing, EMI shielding, deformation sensing, conductivity
1. INTRODUCTION
Electrically conductive polymer nanocomposites are being
widely used as key components in batteries,
1
sensors,
2,3
electromagnetic interference (EMI) shielding,
4
and electro-
static discharge protection.
5
Since most polymer matrix
materials are nonconductive, embedded nanofillers need to
be superconductive to offer sufficient conductivity to the
nanocomposites. Hence, carbon nanomaterials such as carbon
nanofibers,
6
carbon nanotubes,
7
and graphene
8
are commonly
used as superior conductive nanofillers for nanocomposites.
In particular, carbon nanofibers are compatible with both
isotropic and anisotropic polymer matrix materials owing to
the presence of limited functional groups on the fiber surface.
Besides, good chemical resistance, high aspect ratio,
9
and cost
effectiveness also make carbon nanofibers popular for the
conductive nanocomposite.
Carbon nanofibers grown from hydrocarbon feedstock or
carbon monoxide on the metal catalyst
10
generally have
controlled fiber structure and nanoscale dimensions. How-
ever, large-scale production of carbon nanofibers in such a
way is not practical due to the requirement of complicated
equipment and demanding processing conditions. An
alternative method would be similar to the manufacturing
of carbon fibers that includes making nanofiber precursors
through electrospinning, stabilization, and graphitization.
Polyacrylonitrile (PAN),
11
pitch,
12
and cellulosic materials
13
are three most commonly used carbon fiber precursors.
Unlike PAN and pitch that need to undergo electrospinning
to form nanofiber, cellulose nanofibril (CNF) is already in a
fiber form with nanoscale diameter and microscale length.
14
Thus, carbon nanofibers might be directly produced from
CNF through stabilization and carbonization. Carbonization
of cellulose macromolecules is actually a thermochemical
decomposition process that includes desorption of free
moisture and bounded moisture, cleavage of cellulose
polymer side groups, session of cellulose polymer backbones,
and graphitization.
15
During the transformation from organic
cellulose molecules to inorganic carbon, the CNF precursor
Received: March 23, 2021
Accepted: May 7, 2021
Articlewww.acsabm.org
© XXXX American Chemical Society A
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
Downloaded via 38.145.92.50 on May 20, 2021 at 13:56:29 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
needs to remain in a fibrous morphology at all stages to avoid
material heterogeneity and loss of nanomaterial properties.
16
Widely used CNF drying methods include freeze-drying,
17
oven drying,
18
spray-drying,
19
and supercritical fluid drying.
20
However, all of these methods are reported to have problems
to obtain dry CNF precursors with individualized fibrous
structures. In this work, we developed a solvent drying
approach to acquire water-free CNF from a concentrated
CNF aqueous suspension. With this novel drying approach,
we were able to prepare dry CNF precursors and carbonized
CNF (CCNF) without loss of individualized fibrous
morphology. Furthermore, we thoroughly studied fundamen-
tal properties of CCNF and used CCNF as a multifunctional
filler for thermoplastic polycaprolactone (PCL)-based con-
ductive nanocomposites. The CCNF-integrated conductive
nanocomposites were demonstrated to be suitable for EMI
shielding and deformation sensing applications.
2. EXPERIMENTAL SECTION
2.1. Materials. Thermoplastic PCL pellets (Capa 6800) were
purchased from Perstorp Holding AB (Malmö, Sweden). CNF
suspension (containing ∼25 wt % CNF) was purchased from Daicel
FineChem (Osaka, Japan). Carbon black (CB-TIMCAL SUPER
C65) was purchased from MTI Corp. (Richmond, CA, USA).
Polyvinylpyrrolidone (PVP, Mw= 40,000) was purchased from
Sigma-Aldrich (St. Louis, MO, USA). Acetone, toluene, and
chloroform were purchased from Hach (Loveland, CO, USA). All
materials were used directly without further purification.
2.2. Preparation of CCNF. Three drying methods including
vacuum oven drying, freeze drying, and solvent drying, were used to
prepare the CNF precursors. Vacuum oven drying was carried out
using a DZF-6020-ETL-110 oven (MTI Corp. Richmond, CA,
USA) at 105 °C for 12 h. Freeze-drying was carried out using
Freezone 4.5(Labconco Corp. Kansas City, MO, USA) for 60 h.
Solvent drying was carried out, as shown in Figure 1. Specifically,
the CNF was first transferred from water suspension to acetone and
subsequently to toluene. The obtained CNF−toluene suspension
was treated with high-intensity ultrasonication using an MSK-USP-
3N-LD ultrasonic processor (MTI Corp. Richmond, CA, USA) for
30 min. Toluene was stripped offvia evaporation at room
temperature to acquire dry CNF. Acetone used in the solvent
drying process was recollected using a Yamato RE 300-AO rotary
evaporator (Yamato Scientific Co., Ltd., Tokyo, Japan).
The CNF precursor was thermally stabilized in a GSL-1100x tube
furnace (MTI Corp., Richmond, CA, USA) at 240 °C in air for 8 h
and then pyrolyzed at 1000 °C in nitrogen for 2 h. The CCNF
properties, including thermal stability, morphology, crystal structure,
elemental composition, and particle size, were characterized using a
thermogravimetric analyzer (TGA Q50, TA Instrument, New Castle,
DE, USA), a field-emission scanning electron microscope (FEI
QuantaTM 3D FEG dual beam SEM/FIB system, FEI Co.,
Hillsboro, OR, USA), a high-resolution transmission electron
microscope (JEOL 2011, JEOL USA Inc., Peabody, MA, USA),
an X-ray diffraction (XRD) device (PANalytical Empyrean
diffractometer, Malvern Panalytical Inc. Westborough, MA, USA),
an X-ray photoelectron spectroscope (ESCA 2SR, Scienta Omicron,
Uppsala, Sweden), and a particle size analyzer (Microtrac S3500,
Microtrac Inc., Largo, FL, USA).
2.3. Fabrication and Tensile Testing of CCNF/PCL
Composite Films. The CCNF/PCL composite films were prepared
by the solution casting method. Tensile testing specimens were
prepared in a dumbbell shape in accordance with the ASTM D638
standard (type V). The tensile test was carried out with a model
5582 Instron machine equipped witha1 kN load cell and Bluehill
software (Instron Inc., Norwood, MA, USA). The tensile speed was
at 10 mm/min with a span distance of 7.62 mm. Five specimens
were tested for each material to obtain an average value. Specimen
fracture surfaces were observed with a Zeiss Axiovert 200 optical
microscope (Carl Zeiss X-ray Microscopy, LLC., Pleasanton, CA)
and scanning electron microscope.
2.4. Preparation of the PVP-Coated CCNF/Carbon Black
Conductive Hybrid. The concentrated CNF suspension and PVP
were mixed in water and stirred for 12 h. Similar to the solvent
drying process for preparing dry CNF, PVP-coated CNF was first
transferred from water to acetone (Figure 2). CB was then added,
and acetone in the mixture was removed by centrifuging. The PVP/
CNF/CB compound was redispersed in toluene with intensive
ultrasonication. After toluene was completely removed through
evaporation, dry PVP/CNF/CB precursors were obtained, and they
were subsequently stabilized and carbonized to prepare the PVP-
coated CCNF/CB (PVP@CCNF/CB) conductive hybrid (Figure
2).
2.5. Fabrication of Three-Dimensional Printing Conductive
Filaments. Three-dimensional (3D) printing conductive filaments
(1.75 mm in diameter) containing various conductive nanofillers
(i.e., CB, CCNF/CB, and PVP@CCNF/CB) were extruded at 90
°C using a Filabot EX2 extruder (Filabot, Barre, VT, USA). The as-
extruded filaments were chopped to small segments and re-extruded
3 times to ensure that conductive nanofillers were dispersed
homogeneously in the PCL matrix.
The four-probe method was used to measure filament electrical
resistivity (ρ)
RA
l
ρ=(1)
where lis the filament length (3 cm), Ais the filament cross-
sectional area, and Ris the resistance measured by a Keithley
2100digital multimeter (Tektronix, Inc., Beaverton, OR, USA).
2.6. 3D Printed EMI-Shielding Composites. The EMI-
shielding composites were 3D printed using a FlashForge Creator
Pro printer (Flashforge USA, City of Industry, CA, USA). The
composite has a cylindrical geometry with 1.5 mm inner diameter,
3.5 mm outer diameter, and 2 mm height (Figure S1a). The 3D
printing parameters are listed in Table S2. EMI shielding properties
were measured by a Keysight N5230c vector network analyzer
(Keysight Technologies, Santa Rosa, CA, USA). The coaxial line
method was used for the scattering parameter measurements within
the frequency range of 4−26 GHz. Composites’outer surface
morphologies were observed with SEM.
2.7. Fabrication of the Deformation-Sensing Composite
Film. Composite electromechanical behaviors were investigated by
measuring specimen real-time electrical resistance as a function of
applied strain. Composite sensing stability was tested on the casted
film and 3D printed film. The casted film was in a rectangular shape.
The 3D printed film was in a rectangular geometry with a hexagon
array built in the center region (Figure S1b). The 3D printing
parameters are listed in Table S2. Finite element analysis (FEA) was
performed using ANSYS Workbench 2.0 (ANSYS Inc, Canonsburg,
PA, USA) to compare stress distribution on the surfaces. The FEA
parameters (Table S3) were based on the tensile data of neat PCL
films. The casted and 3D printed films were meshed with 990 and
198 elements for the numerical analysis, respectively. The simulation
Figure 1. Schematic of the CNF solvent drying process.
Centrifugation was carried out at 7000 rpm for 7 min. CNF
concentration was kept at 3 wt % in water, acetone, and toluene.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
B
Figure 2. Schematic representation of the preparation of the PVP@CCNF/CB conductive hybrid. CNF and PVP had an equal weight in CNF/
PVP/water suspension. CNF/PVP/CB weight ratio was 1:1:3 in a CNF/PVP/CB−acetone suspension. Centrifugation was carried out at 7000
rpm for 10 min. For stabilization and carbonization processes, dry CNF/PVP/CB was first heated up to 240 °C with a heating rate of 1 °C/
min and then stabilized at 240 °C for 8 h in air, followed by heating up to 1000 °C with a heating rate of 5 °C/min, and then carbonized at
1000 °C for 2 h under the protection of N2.
Figure 3. (a) TGA curves for three types of dried CNF precursors. Insets show that oven-dried CNF and freeze-dried CNF were heavily
agglomerated, while solvent-dried CNF was in the fibrous form. (b) SEM image of solvent-dried CCNF. (c) CCNF size distribution from 0.1 to
100 μm. Insets show photographs of dispersion state of 0.5 wt % CCNF−water suspensions. (d) Schematic of the proposed CNF solvent drying
mechanism. (e) FTIR spectra of three types of dry CNF. Insets show magnified spectra at 3600−3100 and 1750−1550 cm−1and (f) XRD
patterns of solvent-dried CNF and CCNF. Insets show deconvolution of the CCNF diffraction peak (002) and an HRTEM image of a CCNF
with short-ranged carbon lattice highlighted in red boxes.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
C
results are presented in normal stress and von Mises stress (Figure
S2) under 3% applied strain (Figure S3) in the longitudinal
direction.
Filament specimens used for the cyclic stretching−bending test
had a length in the range of 10−15 mm. Filament electromechanical
response to mechanical deformations was evaluated by stretching
and bending the filaments with the Instron tensile testing setup 5
times. The maximum tensile and bending strains were set at 3 and
2%, respectively. Electromechanical data including short (i.e., 5
cycles) cyclic stretching−bending, electrical resistance, and long (i.e.,
300 cycles) cyclic stretching were collected within the composite
elastic regime (strain = 3%).
3. RESULTS AND DISCUSSION
3.1. Fundamental Properties of the Solvent-Dried
CNF Precursor and Carbonized CNF. Figures 3 and 4
show fundamental properties of the CNF precursor and
carbonized CNF. In particular, Figure 3a shows thermogravi-
metric curves for three types of dry CNF precursors. Rapid
sample weight loss between 240 and 390 °C was due to CNF
thermal decomposition.
21
For solvent-dried CNF, slightly
lower (∼15 °C) onset degradation temperature and less char
residue indicated its higher sensitivity to temperature as heat
diffused in and decomposed the inner CNF easily through
the largely exposed surface area.
22
This is verified by SEM
images that oven-dried CCNF agglomerated to particulates
(Figure 4a), freeze-dried CCNF agglomerated to flakes
(Figure 4b), and both show absence of fibrous morphologies.
Nevertheless, solvent-dried CCNF preserved its fibrous
morphology (Figure 3b), showing single fiber and entangled
fiber bundles in SEM (Figure 4c) and HRTEM images
(Figure 4d). After being dispersed in water (Figure 3c),
solvent-dried CCNF had an average size within the
nanometer scale (<1 μm.), while majority of the freeze-
dried CCNF was in the range of 1−5μm. The oven-dried
CCNF had the largest sizes and the most widespread size
distribution from 5 to 20 μm.
A schematic drawing (Figure 3d) illustrates the proposed
mechanism for the formation of the fibrous structure of
solvent-dried CCNF. During solvent drying, the liquid
medium for CNF was changed from water to acetone and
then to toluene.
21
The CNF−water suspension exhibits an
opaque appearance (Figure 4e) due to the existing CNF−
water molecule hydrogen bonds. In acetone, CNF self-
assembled through intrinsic intra- and interhydrogen bonds.
23
After the CNF was redispersed in toluene under ultra-
sonication, toluene infiltrated into CNF molecular chains and
disengaged CNF interhydrogen bonds, resulting in a
transparent appearance. Toluene-occupied spaces that used
to belong to water molecules remained, and owing to the
presence of these free spaces left from removal of toluene,
solvent-dried CCNF attained fibrous structures instead of
agglomerations. Moreover, high-temperature carbonization
induced the cleavage of CNF side groups, which destroyed
intrahydrogen bonds
24
and placed furthermore distances
between CCNF backbones. Our hypothesis is supported by
the FTIR results (Figure 3e) that the decreased peak
intensity at 3600−3150 cm−1of solvent-dried CNF
represented diminished CNF self-assembled inter hydrogen
bonds.
25
The absorbance bands at 1750−1680 and 1650−
1600 cm−1were reported to be attributed to the CO group
from CNF impurities (e.g., pectin) and the hydroxyl group
from CNF-absorbed water molecules, respectively.
23
Likewise,
Figure 4. (a) SEM images of oven-dried CCNF. (b) SEM images of freeze-dried CCNF. (c,d) SEM and HRTEM images of solvent dried
CCNF, showing individual fiber and entangled fibers. Yield rates (mass of CCNF/mass of CNF precursors) are 19.0, 17.8, and 17.5%,
respectively. (e) Photographs of CNF dispersion states in different mediums during the solvent drying process with the CNF concentration kept
at 3 wt % in water, acetone, and toluene. (f) XPS survey spectrum of solvent-dried CCNF. Inset figure shows deconvolution of the carbon peak.
Inset table compares the C/O ratio between CNF and CCNF.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
D
decreased intensity of these two peaks indicated that the free
spaces prevented solvent-dried CNF from forming interhy-
drogen bonds with impurities and absorbed moisture.
26
Three diffraction peaks (101), (101̅), and (200) in the
solvent-dried CNF (Figure 3f) represent the transverse
arrangement of the crystallites in cellulose I, while the
diffraction peak (040) represents the CNF longitudinal
structure.
27
The newly formed broad diffraction peak (002)
in CCNF was a mixture of a sharp crystalline peak and
amorphous halo associated with the graphitized carbon
28
and
amorphous carbon,
29
respectively. This intermediate carbon
phase of CCNF is also observed in the HRTEM image
(Figure 3f inset) in which both disordered and short-ranged
carbon structures are seen along the fibers.
The XPS survey spectrum (Figure 4f) shows that CCNF is
mainly composed of carbon and oxygen, while the other
small amounts of coexisting elements such as sodium,
nitrogen, and sulfur are from alkali purification and acid
hydrolysis in CNF preparation. The increased C/O ratio was
ascribed to the loss of oxygen-based functional groups after
carbonization (i.e., dehydration and depolymerization). In
CCNF, oxygen elements cross-linked with the carbon
microstructures (i.e., C−OandCO), forming the
amorphous carbon region, while the graphitized carbon
region contained concentrated sp2 carbon, which was
reflected by the C−C peak with a broad and asymmetric
tail toward higher binding energy.
30
3.2. Tensile Properties of CCNF/PCL Composite
Films. The addition of solvent-dried CCNF significantly
improved tensile performance for PCL composite films
(Figure 5a). The maximum tensile toughness (317.0 MJ/
m3), modulus (117.7 MPa), strength (30.5 MPa), and strain
at break (1959.4%) were achieved at a 7 wt % CCNF
loading, which were 2.9, 1.4, 1.9, and 1.9 times higher
compared to the corresponding properties of the neat PCL
films.
Figures 5b and S4a−cshow a comparison of composite
tensile properties in terms of the CCNF filler type. It is
clearly observed that the solvent-dried CCNF/PCL compo-
site films had superior tensile properties over both oven-dried
and freeze-dried CCNF/PCL composite films at the given
loading rates. It also should be noted that PCL composite
obviously benefits from the solvent-dried CCNF on tensile
strength enhancement even when compared to some
commonly used nanofillers (Table S1).
Improved tensile properties were strongly associated with
the uniform dispersion state of CCNF in the PCL matrix,
mainly due to the fact that solvent-dried CCNF had fibrous
morphology. It is evident that solvent-dried CCNF/PCL
composite films show a relatively smooth fracture surface
(Figure 5c,d) with pulled-out fibers (Figure 5e), indicating a
strong fiber−matrix interface. Rough fracture surfaces are
observed for CCNF/PCL composite films with oven-dried
(Figure 5f) and freeze-dried (Figure 5h) CCNF materials,
showing a poor CCNF dispersion state resulting from fiber
agglomerations in the fracture region. Moreover, not only the
oven- and freeze-dried CCNF failed to reinforce the PCL
matrix but also tended to initiate failures at fiber-rich regions
(Figure 5g−i).
3.3. Conductivity of CCNF/CB Hybrids. CCNF derived
from a low carbonization temperature (<1400 °C) lacks of
sufficient electrical conductivity. Hence, a combination with
Figure 5. (a) Tensile stress−strain curves for solvent-dried CCNF/PCL composite films, with a loading rate from 0 to 7 wt %. (b)
Comparisons on tensile toughness for PCL composite films containing different types of CCNF. (c) Photograph of composite films after the
tensile test. and (d−i) fracture surface morphologies of composite films containing 5 wt % CCNF. Optical images: (d) Solvent-dried CCNF/
PCL. (f) Oven-dried CCNF/PCL. (h) Freeze-dried CCNF/PCL. SEM images: (e) solvent-dried CCNF/PCL. The arrows highlight pulled-out
fibers. (g) Oven-dried CCNF/PCL. (i) Freeze-dried CCNF/PCL.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
E
other conductive materials
31
is necessary to boost CCNF
conductivity.
32
In this work, we physically blended CCNF
and CB in an attempt to have CCNFs serving as bridges to
connect CB aggregates (Figure 6a). Furthermore, in the
precursor preparation step, we coated PVP on top of CNFs
through strong hydrogen bonds. After carbonization, both
organic CNF and PVP converted into inorganic carbon. The
XRD patterns show that the PVP-coated CCNF (PVP@
CCNF) (Figure S5a) has a crystallite structure identical to
pure CCNF, indicating that the resulting PVP-derived carbon
was only in an amorphous state.
33
The HRTEM images
(Figure 6b,c) show that a 17.3 nm thickness of PVP derived
carbon laid on the CCNF continuously, which plays a role as
a binding agent that linked CB particles to CCNF. Therefore,
in addition to PVP@CCNF, a pathway through the linked
CB particles on the CCNF helped electrons to move freely
(Figure 6d), which led to a significantly enhanced electrical
conductivity for the PVP-coated CCNF/CB hybrid (PVP@
CCNF/CB).
Figure 6e shows a comparison of conductive filaments
powering up LEDs in a closed loop (Figure S5b). It is clearly
observed that the PVP@CCNF/CB/PCL filament lighted up
all LEDs, while yellow and green LEDs barely emitted lights
for CB/PCL and CCNF/CB/PCL filaments. A conductivity
comparison between the CCNF-based filaments and
commercial CB-based filaments is shown in Figure 6f. It
should be noted that the electrical resistivity for the PVP@
CCNF/CB/PCL filament significantly dropped from 1116.44
to 0.0159 Ωm as the conductive filler loading increased from
10 to 30 wt %, making 30 wt % PVP@CCNF/CB/PCL the
most conductive filament. The extraordinary conductivity
indicates a great potential for 3D printing, EMI shielding, and
deformation-sensing composites.
3.4. Composite EMI Shielding Properties. Conductive
composites are able to weaken or completely shield EM
radiation from dielectric loss.
34
The nature of materials (e.g.,
permittivity), EM radiation characteristics (e.g., frequency),
and geometries (e.g., thickness) all together control EMI
shielding effectiveness.
35
Figure 7a,b shows composite real (ε′) and imaginary (ε″)
dielectric permittivity within the frequency range of 4−26
GHz. As the PCL matrix is nonconductive, ε′is only
associated with composite interfacial polarization ability,
which is closely related to the filler surface area and free
charge mobility within composites.
36
Positive ε′intended the
interfacial polarization arising from the difference in electrical
conductivities between the filler and matrix.
37
The CCNF
tended to interact with CB and form a chain-like
interconnected structure,
8
which further delocalized polarized
charge at the filler−matrix interface instead of causing
electron accumulations.
38
The decreased ε′and ε″values
with frequency were due to the possible interfacial polar-
izationa phenomenon resulting from insulated PCL,
creating boundaries within the CCNF/CB network.
39
For
the dielectric loss tangent (ε″/ε′) in the frequency range of
4−15 GHz, ε″/ε′values were in the order CCNF/CB/PCL
> PVP@CCNF/CB/PCL > CB/PCL > neat PCL (Figure
7c). At frequencies higher than 15 GHz, the ε″/ε′ranking
changed to PVP@CCNF/CB/PCL > CCNF/CB/PCL >
CB/PCL > neat PCL, leading to a higher EM attenuation
factor for PVP@CCNF/CB due to the significantly increasing
ε″/ε′value.
40
Radiation reflection and absorption are two dominant EMI
shielding mechanisms.
41
Reflection occurs when radiation
interacts with shielding-material surface charges.
42
Absorption
is a dissipation of EM radiation energy into other internal
energies.
43
EM radiation can be reflected multiple times, and
the multireflected EM radiation residual is eventually
reabsorbable by the material.
35
The total EMI shielding
effectiveness (SET) is a sum of effectiveness by absorption
(SEA) and reflection (SER), which is expressed as
S
ESESE
TR
A
=+
(2)
Figure 6. (a) SEM image showing CCNF connecting CB aggregates in the CCNF/CB hybrid. (b) TEM image of the PVP@CCNF/CB hybrid
structure. (c) TEM image of the PVP-derived carbon layer on the CCNF surface. (d) Schematic of surface charges moving between CB
aggregates in the PVP@CCNF/CB hybrid. (e) 30 wt % conductive filler−PCL composite filaments power up light-emitting diodes (LEDs). (f)
Resistivity comparison between CCNF-based filaments (this work) and commercial CB-based filaments.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
F
i
k
j
j
j
j
j
y
{
z
z
z
z
z
P
P
S
E10log
TI
T
=
(3)
i
k
j
j
jy
{
z
z
z
T
R
S
E10log
1
A=−
−(4)
R
S
E10log(1)
R=− − (5)
where PIand PTare the powers of incident and transmitted
EM radiations, respectively.
The reflection coefficient R, transmission coefficient T, and
absorption coefficient Aare calculated by
R
S112
=(6)
T
S212
=
(7)
A
R
T
1=− − (8)
Measured Sparameters, S11 [reflected (PR) power over
incident power (PI)] and S21 (transmitted power over
incident power) are expressed by
S
P
P
11 R
I
=
(9)
S
P
P
21 T
I
=
(10)
Figure 7d shows the composite total shielding effectiveness
in terms of conductive filler types. As expected, neat PCL
barely showed shielding performance within the entire testing
frequency, while the CCNF-added CB/PCL composites
showed much improved total effectiveness. For example,
the maximum SETvalue for the PVP@CCNF/CB/PCL
composite reached above 15 dB at 26 GHz. Composite
absorption effectiveness (Figure 7e) followed the trend of the
total effectiveness; while composite reflective effectiveness
was in a bell shape, as shown in Figure 7f, with the maximum
reflection reaching ∼6 GHz. It is clearly observed that
radiation absorption had the major contribution to the total
shielding effectiveness. Limited radiation reflection was
partially related to the poor surface morphology of 3D
printed composites. Composite defects such as voids between
adjacent layers (Figure 7g), excessive stringing, and layer
Figure 7. Comparison of EMI shielding properties for composites with different conductive fillers: (a) ε′. (b) ε″. (c) Loss tangent. (d) SET. (e)
SEA. (f) SER. (i) Skin depth. SEM surface morphologies of 3D printed EMI shielding cylindrical specimens: (g) Voids between two successive
layers and (h) seams at a single layer.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
G
seams (Figure 7h) caused multiple EM radiation scattering
and thereby negatively impacted composite reflection
effectiveness.
The skin depth is defined as the distance up to which EM
radiation attenuates by 1/e or 37%.
44
It is expressed as
f
1
r0
δ
πμμσ
=
(11)
where fis the EM radiation frequency, μ0is the absolute
permeability of free space, which equals 4π×10−7H/m−1,
relative permeability μrequals 1 for the nonmagnetic material,
and σis the composite electrical conductivity. It is known
that the δvalue is inversely proportional to SEA; as a result,
the PVP@CCNF/CB/PCL composite with the minimum δ
value exhibited the maximum absorption, and alternatively,
the CB/PCL composite exhibiting the minimum absorption
had the maximum δvalue.
3.5. Composite Deformation Sensing Properties.
Figure 8a−c shows CCNF/CB/PCL filament electrical
resistance measured by cyclic (five cycles) stretching−
bending test at 1.5, 3.0, and 4.5%/min strain rates,
respectively. The ratio of real-time resistance to the initial
value (R/R0) dropped after the first cycle but reached a
relatively stable state within the first five cycles. As the
applied strain recovered to 0, the R/R0value dropped below
1, indicating that the filament electrical resistance was lower
than the initial value, which could be ascribed to the
formation of distortions and cracks during the straining
step.
45,46
As expected, the strain rate had no direct impact on
the maximum R/R0value as the electromechanical property
was mainly determined by the geometrical effect.
47
Due to
the uniform dispersion of conductive fillers, structural
integrity of conductive networks was well preserved as long
as stretching maintained in the elastic regime. Filament
electrical resistance exhibited a significant drop before the
Figure 8. Electrical resistance (R/R0) and FEA results of normal stress distribution on the sensor surface in response to the stretching−bending
force for 20 wt % CCNF/CB/PCL filaments. (a)1.5, (b) 3.0, and (c) 4.5%/min−1, (d) film composite, (e) 3D printed composite, (f)
Comparative R/R0stability data for the film type and 3D printed composites under cyclic tensile test for 500 cycles at a strain rate of 1.5%/
min−1. Insets show R/R0values for the 50th and 250th cycles, and (g) sensitivity and reliability to deformations with insets show film flexibility
to bending.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
H
yield point as the CCNF/CB hybrid aligned along with PCL
molecular chains, after which it started to increase following
the geometrical effect and finally reached infinity at fracture.
It should be noted that a hysteresis phenomenon was
observed due to the viscoelastic nature of the PCL matrix
(Figure S6).
Furthermore, the performances of the casted film and 3D
printed honeycomb composites were compared in order to
understand composite geometry effect on the sensing
response to various mechanical deformations. Both types
had the same composition (i.e., 20 wt % CCNF/CB/PCL)
and dimension (i.e., length, width, and thickness), except that
the 3D printed one was lighter in weight due to the presence
of hexagonal holes in the center region. The FEA results
show that under the uniaxial force, the film type (Figure 8d)
carried load mostly at four corners. On the other hand,
stresses distributed uniformly across the entire honeycomb
structure (Figure 8e), and there was noticeable deformation
upon applied force and quick shape recovery upon the release
of force. This also aligns with the reproducible R/R0values
for the 3D printed composite from cycle to cycle (Figure 8f).
For example, a stable R/R0value of 0.93 was observed for
both the 50th and 250th cycles. Figure 8g also demonstrates
the flexibility and durability of the 3D printed composite in
response to bending and its ability to recover to the original
position upon the release of bending force. The 3D printed
composite showed equal sensitivities, regardless of the
bending direction as the R/R0value shifted to the same
level when being bent to the same angle in two directions.
4. CONCLUSIONS
In this work, we introduced a novel solvent drying approach
for acquiring dry CNF precursors. Through solvent drying,
the majority of the CNF self-assembled interhydrogen bonds
were displaced by toluene molecules, which resulted in
sufficient free spaces between CNF molecular chains and
fibrous morphology for dried CNF. The obtained CCNF was
in an intermediate state between graphitized and amorphous
carbons, having both short-ranged and disordered carbon
lattices distributed along a single fiber or entangled fiber
bundles. The retained fibrous morphology was essential for
CCNF serving as reinforcing fillers to disperse homoge-
neously in the PCL matrix and for remarkably improving
PCL composite film tensile performance. CCNF also helped
in forming a 3D network when blending with CB, creating
bridges for free charges to move between CB aggregates.
Conductive composites containing CCNF/CB-based fillers
(i.e., CCNF/CB and PVP@CCNF/CB) were suitable for
EMI shielding. EM radiation was mainly absorbed owing to
the CCNF interconnected structure effectively delocalizing
free charges. The conductive composites also find applica-
tions for deformation sensing, which exhibited excellent
stability and durability in response to various types of forces.
■ASSOCIATED CONTENT
*
sıSupporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acsabm.1c00360.
Autodesk Fusion 360 designs of 3D printed compo-
sites, equivalent (von Mises) tensile stress distribution
on the surface and the FEA method with 3% tensile
strain applied in the uniaxial direction, comparison of
tensile constants for PCL composite films containing
different types of CCNF, XRD pattern for PVP@
CCNF and the photograph of a closed LED loop
(Figure S5), and stress−time curve under the cyclic
stretching−bending test for 20 wt % CCNF/CB/PCL
filaments, a summarized comparison of tensile strength
of CCNF/PCL composite films (this work) vs PCL
composites filled with other types of reinforcing agents,
3D printing parameters for EMI shielding and
deformation sensing components, and PCL tensile
parameters used for FEA (PDF)
■AUTHOR INFORMATION
Corresponding Author
Qinglin Wu −School of Renewable Natural Resources,
Louisiana State University, Baton Rouge, Louisiana 70803,
United States; orcid.org/0000-0001-5256-4199;
Email: wuqing@lsu.edu
Authors
Ju Dong −School of Renewable Natural Resources, Louisiana
State University, Baton Rouge, Louisiana 70803, United
States
Xingyan Huang −School of Renewable Natural Resources,
Louisiana State University, Baton Rouge, Louisiana 70803,
United States; College of Forestry, Sichuan Agricultural
University, Chengdu 611130, China
Guang-Lin Zhao −Physics Department and Nano Materials
Laboratory, Southern University and A&M College, Baton
Rouge, Louisiana 70813, United States; orcid.org/0000-
0002-5710-4922
Jaegyoung Gwon −Department of Forest Products, National
Institute of Forest Science, Seoul 130-712, Korea
Won-Jae Youe −Department of Forest Products, National
Institute of Forest Science, Seoul 130-712, Korea
Complete contact information is available at:
https://pubs.acs.org/10.1021/acsabm.1c00360
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version
of the manuscript.
Notes
The authors declare no competing financial interest.
■ACKNOWLEDGMENTS
This collaborative study was carried out with support from
the National Institute of Forest Science (Seoul, Korea)
through a cooperative project to the LSU AgCenter and
Louisiana Board of Regents [LEQSF(2020−23)-RD-B-02 and
LEQSF(2015−17)-RD-B-01].
■REFERENCES
(1) Wang, J.; Yang, J.; Xie, J.; Xu, N. A Novel Conductive
Polymer-Sulfur Composite Cathode Material for Rechargeable
Lithium Batteries. Adv. Mater. 2002,14, 963−965.
(2) Yamada, T.; Hayamizu, Y.; Yamamoto, Y.; Yomogida, Y.; Izadi-
Najafabadi, A.; Futaba, D. N.; Hata, K. A stretchable carbon
nanotube strain sensor for human-motion detection. Nat. Nano-
technol. 2011,6, 296−301.
(3) Liu, H.; Chen, X.; Zheng, Y.; Zhang, D.; Zhao, Y.; Wang, C.;
Pan, C.; Liu, C.; Shen, C. Lightweight, Superelastic, and Hydro-
phobic Polyimide Nanofiber /MXene Composite Aerogel for
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
I
Wearable Piezoresistive Sensor and Oil/Water Separation Applica-
tions. Adv. Funct. Mater. 2021,31, 2008006.
(4) Li, N.; Huang, Y.; Du, F.; He, X.; Lin, X.; Gao, H.; Ma, Y.; Li,
F.; Chen, Y.; Eklund, P. C. Electromagnetic Interference (EMI)
Shielding of Single-Walled Carbon Nanotube Epoxy Composites.
Nano Lett. 2006,6, 1141−1145.
(5) Rosner, R. B. Conductive Materials for ESD Applications: An
Overview. IEEE Trans. Device Mater. Reliab. 2001,1,9−16.
(6) Yang, Y.; Gupta, M. C.; Dudley, K. L.; Lawrence, R. W.
Conductive Carbon Nanofiber-Polymer Foam Structures. Adv.
Mater. 2005,17, 1999−2003.
(7) Choi, H. Y.; Lee, T.-W.; Lee, S.-E.; Lim, J.; Jeong, Y. G. Silver
nanowire/carbon nanotube/cellulose hybrid papers for electrically
conductive and electromagnetic interference shielding elements.
Compos. Sci. Technol. 2017,150,45−53.
(8) Pan, F.; Chen, S.-M.; Li, Y.; Tao, Z.; Ye, J.; Ni, K.; Yu, H.;
Xiang, B.; Ren, Y.; Qin, F.; Yu, S.-H.; Zhu, Y. 3D Graphene Films
Enable Simultaneously High Sensitivity and Large Stretchability for
Strain Sensors. Adv. Funct. Mater. 2018,28, 1803221.
(9) Al-Saleh, M. H.; Sundararaj, U. Review of the mechanical
properties of carbon nanofiber/polymer composites. Compos. Appl.
Sci. Manuf. 2011,42, 2126−2142.
(10) Che, G.; Lakshmi, B. B.; Martin, C. R.; Fisher, E. R. Chemical
Vapor Deposition Based Synthesis of Carbon Nanotubes and
Nanofibers Using a Template Method. Chem. Mater. 1998,10, 206−
267.
(11) Kim, C.; Yang, K. S.; Kojima, M.; Yoshida, K.; Kim, Y. J.;
Kim, Y. A.; Endo, M. Fabrication of Electrospinning-Derived Carbon
Nanofiber Webs for the Anode Material of Lithium-Ion Secondary
Batteries. Adv. Funct. Mater. 2006,16, 2393−2397.
(12) Yang, H.; Yan, Y.; Liu, Y.; Zhang, F.; Zhang, R.; Meng, Y.;
Mei, M.; Xie, S.; Tu, B.; Zhao, D. A Simple Melt Impregnation
Method to Synthesize Ordered Mesoporous Carbon and Carbon
Nanofiber Bundles with Graphitized Structure from Pitches. J. Phys.
Chem. B 2004,108, 17320−17328.
(13) Bober, P.; Kovářová, J.; Pfleger, J.; Stejskal, J.; Trchová, M.;
Novák, I.; Berek, D. Twin carbons: The carbonization of cellulose or
carbonized cellulose coated with a conducting polymer, polyaniline.
Carbon 2016,109, 836−842.
(14) Moon, R. J.; Martini, A.; Nairn, J.; Simonsen, J.; Youngblood,
J. Cellulose nanomaterials review: structure, properties and nano-
composites. Chem. Soc. Rev. 2011,40, 3941−3994.
(15) Tang, M. M.; Bacon, R. Carbonization of cellulose fibers - I
low temperature pyrolysis. Carbon 1964,2, 211−220.
(16) Jazaeri, E.; Zhang, L.; Wang, X.; Tsuzuki, T. Fabrication of
carbon nanofiber by pyrolysis of freeze-dried cellulose nanofiber.
Cellulose 2011,18, 1481−1485.
(17) Jazaeri, E.; Tsuzuki, T. Effect of pyrolysis conditions on the
properties of carbonaceous nanofibers obtained from freeze-dried
cellulose nanofibers. Cellulose 2013,20, 707−716.
(18) Dong, J.; Mei, C.; Han, J.; Lee, S.; Wu, Q. 3D printed
poly(lactic acid) composites with grafted cellulose nanofibers: Effect
of nanofiber and post-fabrication annealing treatment on composite
flexural properties. Addit. Manuf. 2019,28, 621−628.
(19) Dong, J.; Huang, X.; Muley, P.; Wu, T.; Barekati-Goudarzi,
M.; Tang, Z.; Li, M.; Lee, S.; Boldor, D.; Wu, Q. Carbonized
cellulose nanofibers as dielectric heat source for microwave
annealing 3D printed PLA composite. Composites, Part B 2020,
184, 107640.
(20) Pakowski, Z. Modern methods of drying nanomaterials.
Transp. Porous Media 2007,66,19−27.
(21) Dong, J.; Li, M.; Zhou, L.; Lee, S.; Mei, C.; Xu, X.; Wu, Q.
The influence of grafted cellulose nanofibers and postextrusion
annealing treatment on selected properties of poly(lactic acid)
filaments for 3D printing. J. Polym. Sci., Part B: Polym. Phys. 2017,
55, 847−855.
(22) Han, J.; Zhou, C.; Wu, Y.; Liu, F.; Wu, Q. Self-assembling
behavior of cellulose nanoparticles during freeze-drying: effect of
suspension concentration, particle size, crystal structure, and surface
charge. Biomacromolecules 2013,14, 1529−1540.
(23) Oh, S. Y.; Yoo, D. I.; Shin, Y.; Kim, H. C.; Kim, H. Y.;
Chung, Y. S.; Park, W. H.; Youk, J. H. Crystalline structure analysis
of cellulose treated with sodium hydroxide and carbon dioxide by
means of X-ray diffraction and FTIR spectroscopy. Carbohydr. Res.
2005,340, 2376−2391.
(24) Sevilla, M.; Fuertes, A. B. The production of carbon materials
by hydrothermal carbonization of cellulose. Carbon 2009,47, 2281−
2289.
(25) O’Neill, H.; Pingali, S. V.; Petridis, L.; He, J.; Mamontov, E.;
Hong, L.; Urban, V.; Evans, B.; Langan, P.; Smith, J. C.; Davison, B.
H. Dynamics of water bound to crystalline cellulose. Sci. Rep. 2017,
7, 11840.
(26) Yadav, C.; Saini, A.; Maji, P. K. Energy efficient facile
extraction process of cellulose nanofibres and their dimensional
characterization using light scattering techniques. Carbohydr. Polym.
2017,165, 276−284.
(27) French, A. D. Idealized powder diffraction patterns for
cellulose polymorphs. Cellulose 2013,21, 885−896.
(28) Luo, W.; Schardt, J.; Bommier, C.; Wang, B.; Razink, J.;
Simonsen, J.; Ji, X. Carbon nanofibers derived from cellulose
nanofibers as a long-life anode material for rechargeable sodium-ion
batteries. J. Mater. Chem. A 2013,1, 10662−10666.
(29) Deng, L.; Young, R. J.; Kinloch, I. A.; Zhu, Y.; Eichhorn, S. J.
Carbon nanofibres produced from electrospun cellulose nanofibers.
Carbon 2013,58,66−75.
(30) Zhu, H.; Shen, F.; Luo, W.; Zhu, S.; Zhao, M.; Natarajan, B.;
Dai, J.; Zhou, L.; Ji, X.; Yassar, R. S.; Li, T.; Hu, L. Low
temperature carbonization of cellulose nanocrystals for high
performance carbon anode of sodium-ion batteries. Nano Energy
2017,33,37−44.
(31) Kim, J. M.; Guccini, V.; Seong, K.-D.; Oh, J.; Salazar-Alvarez,
G.; Piao, Y. Extensively interconnected silicon nanoparticles via
carbon network derived from ultrathin cellulose nanofibers as high
performance lithium ion battery anodes. Carbon 2017,118,8−17.
(32) Li, Z.; Liu, J.; Jiang, K.; Thundat, T. Carbonized
nanocellulose sustainably boosts the performance of activated
carbon in ionic liquid supercapacitors. Nano Energy 2016,25,
161−169.
(33) Wang, P.; Zhang, D.; Ma, F.; Ou, Y.; Chen, Q. N.; Xie, S.; Li,
J. Mesoporous carbon nanofibers with a high surface area
electrospun from thermoplastic polyvinylpyrrolidone. Nanoscale
2012,4, 7199−7204.
(34) Al-Saleh, M. H.; Saadeh, W. H.; Sundararaj, U. EMI shielding
effectiveness of carbon based nanostructured polymeric materials: A
comparative study. Carbon 2013,60, 146−156.
(35) Zhang, C.-S.; Ni, Q.-Q.; Fu, S.-Y.; Kurashiki, K. Electro-
magnetic interference shielding effect of nanocomposites with
carbon nanotube and shape memory polymer. Compos. Sci. Technol.
2017,67, 2973−2980.
(36) Zhang, Y.; Wang, Z.; Zhang, B.; Zhao, G.-L.; Guo, S. M. The
electromagnetic interference shielding effectiveness of high aspect-
ratio SiC nanofibers/epoxy composites. RSC Adv. 2015,5, 93499−
93506.
(37) Thomassin, J.-M.; Jérôme, C.; Pardoen, T.; Bailly, C.;
Huynen, I.; Detrembleur, C. Polymer/carbon based composites as
electromagnetic interference (EMI) shielding materials. Mater. Sci.
Eng., R 2013,74, 211−232.
(38) Zhang, Y.; Zhang, B.; Li, K.; Zhao, G.-L.; Guo, S. M.
Electromagnetic interference shielding effectiveness of high entropy
AlCoCrFeNi alloy powder laden composites. J. Alloys Compd. 2018,
734, 220−228.
(39) Wu, Y.; Wang, Z.; Liu, X.; Shen, X.; Zheng, Q.; Xue, Q.; Kim,
J. Ultralight Graphene Foam/Conductive Polymer Composites for
Exceptional Electromagnetic Interference Shielding. ACS Appl.
Mater. Interfaces 2017,9, 9059−9069.
(40) Nwigboji, I. H.; Ejembi, J. I.; Wang, Z.; Bagayoko, D.; Zhao,
G. L. Microwave absorption properties of multi-walled carbon
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
J
nanotube (outer diameter 20−30 nm)−epoxy composites from 1 to
26.5 GHz. Diamond Relat. Mater. 2015,52,66−71.
(41) Al-Saleh, M. H.; Sundararaj, U. Electromagnetic Interference
(EMI) Shielding Effectiveness of PP/PS Polymer Blends Containing
High Structure Carbon Black. Macromol. Mater. Eng. 2008,293,
621−630.
(42) Liu, Z.; Bai, G.; Huang, Y.; Ma, Y.; Du, F.; Li, F.; Guo, T.;
Chen, Y. Reflection and absorption contributions to the electro-
magnetic interference shielding of single-walled carbon nanotube/
polyurethane composites. Carbon 2007,45, 821−827.
(43) Sainia, P.; Choudhary, V.; Singhc, B. P.; Mathurc, R. B.;
Dhawan, S. K. Polyaniline-MWCNT nanocomposites for microwave
absorption and EMI shielding. Mater. Chem. Phys. 2009,113, 919−
926.
(44) Kumar, R.; Mondal, D. P.; Chaudhary, A.; Shafeeq, M.;
Kumari, S. Excellent EMI shielding performance and thermal
insulating properties in lightweight, multifunctional carbon-ceno-
sphere composite foams. Compos. Appl. Sci. Manuf. 2018,112, 475−
484.
(45) Leigh, S. J.; Bradley, R. J.; Purssell, C. P.; Billson, D. R.;
Hutchins, D. A. A simple, low-cost conductive composite material
for 3D printing of electronic sensors. PLoS One 2012,7,
No. e49365.
(46) Wang, C.; Li, X.; Gao, E.; Jian, M.; Xia, K.; Wang, Q.; Xu, Z.;
Ren, T.; Zhang, Y. Carbonized Silk Fabric for Ultrastretchable,
Highly Sensitive, and Wearable Strain Sensors. Adv. Mater. 2016,28,
6640−6648.
(47) Zhang, M.; Wang, C.; Wang, H.; Jian, M.; Hao, X.; Zhang, Y.
Carbonized Cotton Fabric for High-Performance Wearable Strain
Sensors. Adv. Funct. Mater. 2017,27, 1604795.
ACS Applied Bio Materials www.acsabm.org Article
https://doi.org/10.1021/acsabm.1c00360
ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
K
