www.afm-journal.de www.advancedsciencenews.

com

Carbonized Cotton Fabric for High-Performance Wearable

full paper

Strain Sensors
Mingchao Zhang, Chunya Wang, Huimin Wang, Muqiang Jian, Xiangyang Hao,*
and Yingying Zhang*

and thin films,[16] electrically conduc-

Recent years have witnessed the booming development of flexible strain sen- tive polymers,[17] and carbon nanoma-
sors. To date, it is still a great challenge to fabricate strain sensors with both terials,[18–30] such as carbon nanotubes
large workable strain range and high sensitivity. Cotton is an abundant supplied and graphene. To date, a large variety of
natural material composed of cellulose fibers and has been widely used for tex- flexible strain sensors with either high
sensitivity or large workable strain range
tiles and clothing. In this work, the fabrication of highly sensitive wearable strain
have been achieved. Despite these great
sensors based on commercial plain weave cotton fabric, which is the most achievements, most of the reported strain
popular fabric for clothes, is demonstrated through a low-cost and scalable pro- sensors failed to simultaneously possess
cess. The strain sensors based on carbonized cotton fabric exhibit fascinating both large workable strain range and high
performance, including large workable strain range (>140%), superior sensitivity sensitivity, limiting their applications. For
example, a graphene-based strain sensor
(gauge factor of 25 in strain of 0%–80% and that of 64 in strain of 80%–140%),
with a surprisingly high gauge factor
inconspicuous drift, and long-term stability, simultaneously offering advantages (GF) of 1000 within 2%–6% strain was
of low cost and simplicity in device fabrication and versatility in applications. reported.[30] However, the graphene net-
Notably, the strain sensor can detect a subtle strain of as low as 0.02%. Based work would be irreversibly broken when
on its superior performance, its applications in monitoring both vigorous and the applied strain was larger than 7%,
subtle human motions are demonstrated, showing its tremendous potential for leading to the destruction of the sensors
under large deformation. On the other
applications in wearable electronics and intelligent robots.
side, strain sensors with large workable
strain range (280%) have been prepared
using carbon nanotube film.[29] However,
1. Introduction the GF is only 0.06 within a strain range of 60%–200%. More-
over, the fabrication process of the nanomaterial-based strain
Wearable and skin-mountable electronics have attracted broad sensors is generally complicated and may generate environ-
attention in recent years.[1–6] Flexible strain sensors, as a sig- mentally hazardous by-products. There is still a need for a cost-
nificant subpart of wearable electronics, present substantive effective, large-area-compatible, and biocompatible approach
potential applications in human motion detection, personalized for fabricating strain sensors with both high sensitivity and
health monitoring, and smart human–machine interaction.[7] large workable strain range.
Recently, flexible strain sensors based on piezoresistive effect, Carbon materials derived from natural biomaterials, such
which generally combine a conductive sensing element with as carbonized silk that we reported recently,[31] are poten-
elastic polymer matrix, have been widely reported.[8] The typical tial elements for the fabrication of low-cost, scalable, and
reported active sensing elements include metal nanowires[9–15] environmental-friendly electronics on account of their abun-
dant renewable resources and electrically conductive features.
Cotton is undoubtedly the dominating natural fiber utilized
M. C. Zhang, C. Y. Wang, H. M. Wang, for clothes and textiles in our daily life.[32] Plain weave cotton
M. Q. Jian, Prof. Y. Y. Zhang
Department of Chemistry and Center for Nano
fabric is one of the most popular cotton fabrics. It is composed
and Micro Mechanics (CNMM) of warp and weft yarns (each yarn includes dozens of twisting
Tsinghua University cotton fibers), which are aligned to form simple crisscross pat-
Beijing 100084, P. R. China terns. Although natural cotton fibers are not electrically con-
E-mail: [email protected] ductive, a simple thermal treatment can convert them into
M. C. Zhang, Prof. X. Y. Hao highly conductive carbon fibers while maintaining the original
Beijing Key Laboratory of Materials Utilization
of Nonmetallic Minerals and Solid Wastes integrity and flexibility of the fibers.[33,34] Despite their promise,
National Laboratory of Mineral Materials carbonized cotton fabrics have not yet been used in flexible
School of Materials Science and Technology electronics except for a few recent reports[35–37] on using car-
China University of Geosciences bonized cotton for supercapacitors.
Beijing 100083, P. R. China
E-mail: [email protected]
Herein, we report a carbonized plain weave cotton fabric
(CPCF) based strain sensor and demonstrate its superior
DOI: 10.1002/adfm.201604795 performance in detection of both large and subtle human body

1604795 (1 of 7) wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2017, 27, 1604795
www.advancedsciencenews.com www.afm-journal.de

motions. By simply annealing the cotton fabric in an inert atmos- CPCF-based strain sensor. Pristine plain weave cotton fabric con-

full paper
phere, flexible and highly conductive CPCF is obtained, which is tains twisted cotton yarns composed of dozens of cellulose fibers
then fabricated into strain sensors. Due to the unique hierarchical with diameters ranging from 10 to 20 µm. In addition, these cel-
and conductive network structure of the CPCF, the strain sensor lulose fibers are composed of millions of cellulose molecules,
simultaneously exhibits large workable strain range (>140%) and which release small molecules, such as H2O, CO2, and CO, and
high sensitivity (GF of 25 under 0%–80% and that of 64 under transform into carbon with distorted graphited structure under
80%–140%). In particular, the CPCF strain sensor can detect a high-temperature treatment. The pristine cotton fabric exhib-
strain of as low as 0.02% with a gauge factor of 10 and can sense ited a faint yellow color (Figure 1b), and it turned into black
light items down to 0.3 mg. We further demonstrate its excel- (Figure 1c) after being thermal treated at 900 °C under an inert
lent performance in monitoring both subtle motions (phonation, atmosphere (see details in the Experimental Section), which
pulse, respiration, and facial microexpression) and large motions could be further fabricated into flexible strain sensors by encap-
(joints movements of body) of human body. sulating it into an elastic matrix (Ecoflex) (Figure 1d).
After carbonization, the fibers became much thinner, leading
2. Results and Discussion to a shrinkage of 49.2% and 84.7% in the surface area and the
weight, respectively, while the plain weave structure of the
2.1. Fabrication Process of the Carbonized Cotton Fabric cotton fabric retained very well, as shown in the scanning elec-
Strain Sensor tron microscope (SEM) image (Figure 1e). As seen in Figure 1f,
the transmission electron microscope (TEM) image confirmed
Figure 1a illustrates the hierarchical structure of the pristine that small distorted lattice fringes with an interlayer thickness
plain weave cotton fabric and the fabrication process of the of 0.35 nm, which belonged to sp2 bonded graphited carbon

Figure 1. Fabrication process and characterization of carbonized plain weave cotton fabric (CPCF). a) Illustration showing the structure of the pristine
cotton fabric and the fabrication process of the CPCF strain sensor. b) Photograph of a pristine cotton fabric. c) Photograph of the CPCF made from
(b). d) A flexible CPCF-based strain sensor. e–g) SEM image, TEM image, and Raman spectrum of the CPCF.

Adv. Funct. Mater. 2017, 27, 1604795 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com (2 of 7) 1604795
 www.afm-journal.de www.advancedsciencenews.com

Figure 1a. As shown in Figure 2c, the relative resistance change

full paper

(i.e., the (002) plane of graphite), were irregularly embedded

throughout the whole CPCF. The Raman spectrum in Figure 1g of the CPCF strain sensor exhibited almost no frequency
further ascertained that graphite-like carbon formed after the dependence within the tested frequency range (0.05–0.25 Hz)
thermal treatment. The CPCF showed good electrical conduc- to apply the strain.
tivity with a sheet resistance of 80 Ω sq−1, which could be attrib- The CPCF strain sensors showed long-time stability, little
uted to the formation of graphite-like carbon throughout the hysteresis, and high durability. When the CPCF strain sensor
CPCF and explained with the theory of percolation threshold.[34] was held at various certain strains for a period of time (10 s), the
relative resistance changes were recorded to evaluate whether
the signal could remain stable. As shown in Figure 2d, during
2.2. Electromechanical Performance of the CPCF Strain Sensor each period, the signals remained stable without distinct drifts.
The CPCF strain sensor showed a little hysteresis because of
The CPCF strain sensor showed excellent electromechanical the viscoelastic nature of the Ecoflex matrix (Figure S2, Sup-
properties. As shown in Figure 2a, a monotonic increase in porting Information). Figure 2e shows the response of the
relative resistance change was observed with the loading of CPCF strain sensor to 2000 time cyclic loading of 50% strain at
strain in the range of 0%–140%. The sensitivity of a strain a frequency of 0.2 Hz, indicating its remarkable stability.
sensor can be evaluated by GF, i.e., the slope of curve of the
relative resistance change versus strain. The curve shown in
Figure 2a includes two relatively linear parts, 0%–80% and 2.3. Working Mechanism and Detection of Ultralow Strain
80%–140%, with a GF of 25 and 64, respectively, demonstrating
both its high sensitivity and large workable strain range, which One of the particularly remarkable features of the CPCF strain
is preponderant among these recently reported strain sensors sensor is its ability to sense ultralow strain down to 0.02%.
(Figure S1, Supporting Information).[38–48] Note that, before Figure 3a shows that when a cyclic 0.02% strain was applied
measurement, each CPCF strain senor was applied to a tensile onto the CPCF strain sensor, the pattern of the curve remained
strain at a rate of 10 mm min−1 until it finally failed (rupture- quite stable. The corresponding GF is around 10, indicating the
training), and then 10 times of training were conducted until high sensitivity of the CPCF strain sensors for the detection of
a stable response was achieved. We found that the difference ultralow strains or deformations.
among devices was negligible as long as the thread counts, To understand the working mechanism of the superior sen-
the length and the width of the carbonized plain weave cotton sitivity, a CPCF sensor was applied to a series of various strains
fabric, and the thickness of the Ecoflex were kept the same. and the evolution of its structure was recorded (Figure 3b).
The CPCF strain sensors showed quite stable and reliable As mentioned above, the CPCF strain senor was subjected to
response to the applied tensile strain. Figure 2b shows the rupture training before measurement, i.e., continuous strain
relative resistance change of a CPCF strain sensor under cyclic was applied onto the CPCF until it was completely broken.
stretching–releasing of different strains with the same fre- During the first-time loading of tensile strain, irreversible
quency. The relative resistance change increased as the strain cracks among warp yarns of the CPCF formed and increased
increased, which was consistent with the results shown in with the increase of strain until the CPCF was completely

Figure 2. Basic electromechanical performance of the CPCF strain sensors. a) Relative resistance change as a function of tensile strain. b) Relative
resistance change under various cyclic strains at a frequency of 0.2 Hz. c) Relative resistance variation under cyclic stretching–releasing with a strain
of 50% at frequency of 0.05, 0.1, 0.125, and 0.25 Hz. d) Plot showing the negligible drift when being held at different strains. e) Performance of the
sensor under cyclic tensile loading, showing its stability and durability.

1604795 (3 of 7) wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2017, 27, 1604795
www.advancedsciencenews.com www.afm-journal.de

Supporting Information) shows clearly that

full paper
the gap of the cracks increases with the
loading of strain and decreases with the
unloading of strain.
In addition, we found that the thread
count (the number of threads per square
inch along weft and warp direction) of cotton
fabric plays an important role in influencing
the performance of the sensors. CPCF strain
sensors made from cotton fabric with dif-
ferent thread counts were investigated. The
result shows that the CPCF strain sensors
with denser thread counts have higher sen-
sitivities and smaller workable strain range
(Figure S4, Supporting Information).
As a proof of concept, light-weighted items
were put onto a suspended CPCF strain
sensor to demonstrate its ultrasensitivity.
Small items with certain weights will result
in subtle deformation of the CPCF strain
sensor, which will lead to the separation of
the tentacle-like touched fibers beneath the
item, thus causing an increase in the resist-
ance. Different items with different weights
would induce different deformations and
resistance changes. As seen in Figure 3e,
when an item was put onto the sensor, the
current dropped sharply to a certain value
and then partially recovered. This could be
understood that when an item was put onto
the strain sensor, it had a certain speed,
which would cause a bigger deformation.
After the item stopped, an equilibrium state
could be achieved and the current became
stable. A lighter item caused a smaller defor-
mation and thus a smaller drop in current,
while a heavier one led to a large drop. Var-
ious items, including a 0.3 mg PU foam, a
Figure 3. High sensitivity of the CPCF strain sensors. a) Relative resistance change under 1.1 mg copper wire, a grain of rice (18.6 mg),
cyclic loading of a strain of 0.02%. b) Photographs of a CPCF strain sensor during loading of
a 69.3 mg mung bean, and a 152.6 mg red
increasing strain. c) SEM image showing the cross-section of broken fibers. d) SEM image of
a broken yarn. e) Current change during loads of small items with different weights on a sus- bean, were measured. The decrement in cur-
pended strain sensor, insets: photographs of the small items, 0.3 mg PU foam, 1.1 mg copper rent matched their corresponding weight and
wire, 18.6 mg rice, 69.3 mg mung bean, and 152.6 mg red bean (left to right on the top), respec- the signals were quite stable. Accordingly, the
tively. Photograph of a grain of rice put on a suspended CPCF strain sensor (bottom right). sensitivity for weight sensing is calculated
as 127.6 kPa−1, indicating its excellent sensi-
broken, contributing to the monotonical increase of the elec- tivity for subtle deformation detection. Besides, the CPCF-based
trical resistance. At the same time, the weft yarn remained strain sensor showed good repeatability in sensing weighted
intact due to the lack of a transverse force. With the assistance item (Figure S5, Supporting Information).
of elastic Ecoflex matrix, these ruptures were uniformly distrib-
uted throughout the CPCF without islands and gaps. Figure
3c,d shows SEM images of ruptured yarns and fibers. When 2.4. Monitoring of Full-Range Human Body Activities
the CPCF strain sensor was released, the fractured yarns con-
tacted and the broken fibers in each yarn contacted each other, The high sensitivity and the large workable strain range enable
forming a conductive, tentacle-like pathway. Upon subsequent the CPCF strain sensors to monitor full-range human body
loading of strain, these tentacle-like fibers were separated again, motion. Due to the wide workable strain range, large defor-
thus leading to increase in resistance. This could explain that mation of the human body, such as motions of joints can be
0.02% stain (≈5 µm deformation) could be sensed by the ten- recognized by the CPCF strain sensors. When the CPCF strain
tacle-like pathway. The structure evolution of the CPCF strain sensor was fixed onto a knee, it could detect and monitor dif-
sensor during the loading and unloading of strains (Figure S3, ferent motions of knee joints, such as extending/flexing,

Adv. Funct. Mater. 2017, 27, 1604795 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com (4 of 7) 1604795
 www.afm-journal.de www.advancedsciencenews.com
full paper

Figure 4. Monitoring of human motion using the CPCF strain sensors. Response to motions of a) knee joint, b) finger bending, c) facial expression,
d) breath, e) pulse, and f) reading of English words. Insets in each panel show the photographs of the strain sensors.

marching, and jumping (Figure 4a). We fixed a CPCF strain which verified its good stability and repeatability. Figure 4d
sensor on an index finger to see its response to the bending of shows the excellent performance of the CPCF strain sensor for
the finger. When the extended finger bent to a certain angle, the respiration measurement. When the volunteer stayed in relaxa-
relative resistance change of the strain sensor rose to a certain tion, steady and shallow breathing was recognized, while after
value and then remained stable. Further bending would lead to exercise, the obvious thrust of deep breathing could be detected
a further increase in the resistance, forming a stepwise signal with apparent denser signal peaks.
(Figure 4b). Similar measurements were conducted on a wrist Furthermore, a CPCF strain sensor was fixed on a wrist with
with continual bending and rotation, and the strain sensor also the aid of bandage (see the inset in Figure 4e). As shown in
exhibited good performance (Figure S6, Supporting Informa- Figure 4e, the CPCF strain sensor could record distinct human
tion). This demonstrated the ability of the CPCF strain sensor pulse waveforms (diastolic, tidal, and percussion peaks), which
to sustain and monitor large deformation of human body. were significant health-related indexes. The pulse line shapes
On the other hand, based on its ability to sense ultralow measured by the CPCF strain sensor are similar to the recently
strain (0.02%), subtle strain deformation of human body can reported results,[49–51] except for showing more distinctive
be recognized by the CPCF strain sensors. Facial microexpres- peaks that can be ascribed to the high sensitivity of the CPCF
sion could be detected and recognized using the CPCF strain strain sensor. The CPCF strain sensors can also be used to rec-
sensors (Figure 4c). When the face of the volunteer remained ognize the motion of throat induced by reading different words.
relaxing, the resistance kept stable, but when the facial Figure 4f shows the performance of the CPCF strain sensors
expression altered, the sensor responded promptly and the in phonetic recognition. A volunteer was asked to read several
relative resistance change rose. When the volunteer repeated English words repeatedly, as shown in Figure 4f; the strain
cheek-bulging, the signal showed quite good repetitive patterns, sensor showed a distinctive and repeatable signal pattern for

1604795 (5 of 7) wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2017, 27, 1604795
www.advancedsciencenews.com www.afm-journal.de

each word. The CPCF strain sensor was further demonstrated Acknowledgements

full paper
to recognize the English alphabet (see Figure S7, Supporting
M.C.Z. and C.Y.W. contributed equally to this work. This work was
Information). This further demonstrated the excellent perfor-
supported by the NSF of China (Grant Nos. 51672153, 51422204, and
mance of the CPCF strain sensor in monitoring subtle defor- 51372132) and the National Key Basic Research and Development
mation of the human body. Program (Grant Nos. 2016YFA0200103 and 2013CB228506).

Received: September 15, 2016

Published online: November 2, 2016
3. Conclusion
In conclusion, we reported a strain sensor based on carbonized
plain weave cotton fabric, which could be fabricated through [1] X. Pu, L. Li, H. Song, C. Du, Z. Zhao, C. Jiang, G. Cao, W. Hu,
a low-cost, green, and scalable process, and demonstrated its Z. L. Wang, Adv. Mater. 2015, 27, 2472.
[2] S. Lim, D. Son, J. Kim, Y. B. Lee, J. K. Song, S. Choi, D. J. Lee,
superior performance in monitoring both vigorous and subtle
J. H. Kim, M. Lee, T. Hyeon, Adv. Funct. Mater. 2015, 25, 374.
human motions. The CPCF strain sensor exhibits high sensi- [3] J. Lee, H. Kwon, J. Seo, S. Shin, J. H. Koo, C. Pang, S. Son,
tivity in a wide workable strain range (GF of 25 under 0%–80% J. H. Kim, Y. H. Jang, D. E. Kim, Adv. Mater. 2015, 27, 2433.
and that of 64 under 80%–140%). Notably, a subtle strain of as [4] S. Lee, S. Shin, S. Lee, J. Seo, J. Lee, S. Son, H. J. Cho, H. Algadi,
low as 0.02% can be detected. The superior performance can be S. Al-Sayari, D. E. Kim, Adv. Funct. Mater. 2015, 25, 3114.
ascribed to the hierarchal network structure of the CPCF. The [5] T. Q. Trung, S. Ramasundaram, B. U. Hwang, N. E. Lee, Adv. Mater.
strain-induced reversible contact and separation of tentacle- 2016, 28, 394.
like cotton fibers enable its excellent sensing performance in [6] K. Xia, M. Jian, Y. Zhang, Acta Phys.-Chim. Sin. 2016, 32, 2427.
monitoring both large and subtle deformation. Based on that, [7] M. Amjadi, K.-U. Kyung, I. Park, M. Sitti, Adv. Funct. Mater. 2016,
we demonstrated its ability to detect subtle motion (phonation, 26, 1678.
[8] S. Park, M. Vosguerichian, Z. Bao, Nanoscale 2013, 5, 1727.
pulse, facial microexpression, and respiration) and large joints’
[9] M. Amjadi, A. Pichitpajongkit, S. Lee, S. Ryu, I. Park, ACS Nano
movement of the human body. We foresee that the high per- 2014, 8, 5154.
formance CPCF strain sensors, which can be fabricated from [10] C. Yan, J. Wang, W. Kang, M. Cui, X. Wang, C. Y. Foo, K. J. Chee,
abundant natural cotton fiber sources through a green, cost- P. S. Lee, Adv. Mater. 2014, 26, 2022.
effective, and simple process, will find great potential for appli- [11] F. Xu, Y. Zhu, Adv. Mater. 2012, 24, 5117.
cations in wearable electronics. [12] S. Gong, D. T. H. Lai, B. Su, K. J. Si, Z. Ma, L. W. Yap, P. Guo,
W. Cheng, Adv. Electron. Mater. 2015, 1, 1400063.
[13] S. Yao, Y. Zhu, Nanoscale 2014, 6, 2345.
[14] S. Mun, L. Zhai, S. K. Min, Y. Yun, J. Kim, J. Intell. Mater. Syst. Struct.
4. Experimental Section 2015, 27, 1011.
[15] S. Yun, X. Niu, Z. Yu, W. Hu, P. Brochu, Q. Pei, Adv. Mater. 2012,
Fabrication of the Strain Senor: Commercially available plain 24, 1321.
weave cotton fabrics (0.0092 g cm−2, thread count: 40 × 40) were [16] T. Yang, X. Li, X. Jiang, S. Lin, J. Lao, J. Shi, Z. Zhen, Z. Li, H. Zhu,
rinsed using deionized water to remove solid impurity. After being Mater. Horiz. 2016, 3, 248.
dried thoroughly, the cotton fabrics were stored for further usage
[17] H. Niu, H. Zhou, H. Wang, T. Lin, Macromol. Mater. Eng. 2016, 301,
under a dry environment. Under a mixed argon (gas flow, 200 sccm
707.
purity, 99.999%) and hydrogen (gas flow, 20 sccm purity, 99.999%)
[18] A. Bessonov, M. Kirikova, S. Haque, I. Gartseev, M. J. A. Bailey,
atmosphere in a tube furnace, the cotton fabrics were subjected to
the following thermal treatment: (a) heat from ambient temperature Sens. Actuators, A 2014, 206, 75.
to 900 °C at the rate of 3 °C min−1, (b) kept at 900 °C for 200 min, and [19] W. Yuan, Q. Zhou, Y. Li, G. Shi, Nanoscale 2015, 7, 16361.
(c) cooled to room temperature within 200 min. The obtained CPCFs [20] S. H. Bae, Y. Lee, B. K. Sharma, H. J. Lee, J. H. Kim, J. H. Ahn,
were cut into rectangular shapes for the fabrication of strain sensors. Carbon 2013, 51, 236.
Copper wires were attached onto the two ends of the CPCF with silver [21] M. Amjadi, Y. J. Yoon, I. Park, Nanotechnology 2015, 26, 375501.
paste. Then, encapsulation was done by embedding the CPCF into [22] N. Hu, Y. Karube, C. Yan, Z. Masuda, H. Fukunaga, Acta Mater.
Ecoflex with a thickness of about 1 mm. 2008, 56, 2929.
Characterization of the CPCF Strain Sensor: A SEM (Quanta FEG [23] I. Kang, M. J. Schulz, J. H. Kim, V. Shanov, D. Shi, Smart Mater.
450, FESEM) was employed to characterize the surface morphology Struct. 2006, 15, 737.
of the CPCF and the cross section of the broken CPCF yarns. A field [24] R. Rahimi, M. Ochoa, W. Yu, B. Ziaie, ACS Appl. Mater. Interfaces
emission TEM (JEOL JEM2010F) was utilized to characterize the 2015, 7, 4463.
structure of the CPCF. The Raman spectrum was measured with [25] E. Steven, W. R. Saleh, V. Lebedev, S. F. Acquah, V. Laukhin,
a Raman spectrometer (Horiba Jobin Yvon HR Evolution) with a R. G. Alamo, J. S. Brooks, Nat. Commun. 2013, 4, 2435.
532 nm laser. The electromechanical performance of the CPCF strain [26] B. Wang, B. K. Lee, M. J. Kwak, D. W. Lee, Rev. Sci. Instrum. 2013,
sensors was measured with a universal testing machine (Shimadzu 84, 105005.
AGS-X) and a digital source meter (Keithley 2400) at a constant bias [27] W. Wang, T. Yang, H. Zhu, Q. Zheng, Appl. Phys. Lett. 2015, 106,
voltage of 3 V.
171903.
[28] Y. Wang, T. Yang, J. Lao, R. Zhang, Y. Zhang, M. Zhu, X. Li, X. Zang,
K. Wang, W. Yu, H. Jin, L. Wang, H. Zhu, Nano Res. 2015, 8, 1627.
[29] T. Yamada, Y. Hayamizu, Y. Yamamoto, Y. Yomogida, A. Izadi-
Supporting Information Najafabadi, D. N. Futaba, K. Hata, Nat. Nanotechnol. 2011, 6, 296.
Supporting Information is available from the Wiley Online Library or [30] Y. Wang, L. Wang, T. Yang, X. Li, X. Zang, M. Zhu, K. Wang, D. Wu,
from the author. H. Zhu, Adv. Funct. Mater. 2014, 24, 4666.

Adv. Funct. Mater. 2017, 27, 1604795 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com (6 of 7) 1604795
 www.afm-journal.de www.advancedsciencenews.com
full paper

[31] C. Wang, X. Li, E. Gao, M. Jian, K. Xia, Q. Wang, Z. Xu, T. Ren, [41] S. Ryu, P. Lee, J. B. Chou, R. Xu, R. Zhao, A. J. Hart, S. G. Kim, ACS
Y. Zhang, Adv. Mater. 2016, 28, 6640. Nano 2015, 9, 5929.
[32] S. Gargoubi, R. Tolouei, P. Chevallier, L. Levesque, N. Ladhari, [42] H. Tian, Y. Shu, Y.-L. Cui, W.-T. Mi, Y. Yang, D. Xie, T.-L. Ren,
C. Boudokhane, D. Mantovani, Carbohydr. Polym. 2016, Nanoscale 2014, 6, 699.
147, 28. [43] S. Wang, X. Zhang, X. Wu, C. Lu, Soft Matter 2016, 12, 845.
[33] A. G. Dumanlı, A. H. Windle, J. Mater. Sci. 2012, 47, 4236. [44] M. Li, H. Li, W. Zhong, Q. Zhao, D. Wang, ACS Appl. Mater. Inter-
[34] Y. R. Rhim, D. Zhang, D. H. Fairbrother, K. A. Wepasnick, K. J. Livi, faces. 2014, 6, 1313.
R. J. Bodnar, D. C. Nagle, Carbon 2010, 48, 1012. [45] N. Hu, T. Itoi, T. Akagi, T. Kojima, J. Xue, C. Yan, S. Atobe,
[35] L. Li, Q. Zhong, N. D. Kim, G. Ruan, Y. Yang, C. Gao, H. Fei, Y. Li, H. Fukunaga, W. Yuan, H. Ning, Carbon 2013, 51, 202.
Y. Ji, J. M. Tour, Carbon 2016, 105, 260. [46] B. C. Zhang, H. Wang, Y. Zhao, F. Li, X. M. Ou, B. Q. Sun,
[36] M. Tian, M. Du, L. Qu, K. Zhang, H. Li, S. Zhu, D. Liu, J. Power X. H. Zhang, Nanoscale 2016, 8, 2123.
Sources 2016, 326, 428. [47] Y. Tang, Z. Zhao, H. Hu, Y. Liu, X. Wang, S. Zhou, J. Qiu, ACS Appl.
[37] J. Xue, Y. Zhao, H. Cheng, C. Hu, Y. Hu, Y. Meng, H. Shao, Mater. Interfaces 2015, 7, 27432.
Z. Zhang, L. Qu, Phys. Chem. Chem. Phys. 2013, 15, 8042. [48] M. Zhang, C. Wang, Q. Wang, M. Jian, Y. Zhang, ACS Appl. Mater.
[38] K. K. Kim, S. Hong, H. M. Cho, J. Lee, Y. D. Suh, J. Ham, S. H. Ko, Interfaces 2016, 8, 20894.
Nano Lett. 2015, 15, 5240. [49] Y. Cheng, R. Wang, J. Sun, L. Gao, Adv. Mater. 2015, 27, 7365.
[39] E. Roh, B. U. Hwang, D. Kim, B. Y. Kim, N.-E. Lee, ACS Nano 2015, [50] S. Gong, W. Schwalb, Y. Wang, Y. Chen, Y. Tang, J. Si, B. Shirinzadeh,
9, 6252. W. Cheng, Nat. Commun. 2014, 5, 3132.
[40] Q. Sun, W. Seung, B. J. Kim, S. Seo, S. W. Kim, J. H. Cho, Adv. [51] J. Ge, L. Sun, F. R. Zhang, Y. Zhang, L. A. Shi, H. Y. Zhao,
Mater. 2015, 27, 3411. H. W. Zhu, H. L. Jiang, S. H. Yu, Adv. Mater. 2016, 28, 783.

1604795 (7 of 7) wileyonlinelibrary.com © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Funct. Mater. 2017, 27, 1604795