1. health and fitness
2. disease
3. cold and flu

Nano Research
Nano Res
DOI
10.1007/s12274-014-0652-3
Ultra-sensitive graphene strain sensor for sound signal
acquisition and recognition
Yan Wang1,2#, Tingting Yang2,3#, Junchao Lao1,2, Rujing Zhang3, Yangyang Zhang3, Miao Zhu2,3, Xiao
Li2,3, Xiaobei Zang3, Kunlin Wang3, Hu Jin4, Li Wang1, Hongwei Zhu2,3()
Nano Res., Just Accepted Manuscript • DOI: 10.1007/s12274-014-0652-3
http://www.thenanoresearch.com on November 24 2014
© Tsinghua University Press 2014
Just Accepted
This is a “Just Accepted” manuscript, which has been examined by the peer-review process and has been
accepted for publication. A “Just Accepted” manuscript is published online shortly after its acceptance,
which is prior to technical editing and formatting and author proofing. Tsinghua University Press (TUP)
provides “Just Accepted” as an optional and free service which allows authors to make their results available
to the research community as soon as possible after acceptance. After a manuscript has been technically
edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP
article. Please note that technical editing may introduce minor changes to the manuscript text and/or
graphics which may affect the content, and all legal disclaimers that apply to the journal pertain. In no event
shall TUP be held responsible for errors or consequences arising from the use of any information contained
in these “Just Accepted” manuscripts. To cite this manuscript please use its Digital Object Identifier (DOI®),
which is identical for all formats of publication.
1
Ultra-sensitive graphene strain sensor for sound signal
acquisition and recognition
Yan Wang1, Tingting Yang2, Junchao Lao1, Rujing
Zhang2, Yangyang Zhang2, Miao Zhu2, Xiao Li2, Xiaobei
Zang2, Kunlin Wang2, Hu Jin3, Li Wang1, Hongwei Zhu2*
1Nanchang
2Tsinghua
32D
University, China
University, China
Carbon Graphene Material Co., Ltd., China
Page Numbers. The font is
ArialMT 16 (automatically
A wearable and highly sensitive strain sensor was designed from thin
films of graphene woven structures for sound signal acquisition and
recognition.
inserted by the publisher)
1
Nano Res
DOI (automatically inserted by the publisher)
Research Article
Ultra-sensitive graphene strain sensor for sound signal
acquisition and recognition
Yan Wang1,2#, Tingting Yang2,3#, Junchao Lao1,2, Rujing Zhang3, Yangyang Zhang3, Miao Zhu2,3, Xiao Li2,3,
Xiaobei Zang3, Kunlin Wang3, Hu Jin4, Li Wang1, Hongwei Zhu2,3()
1Department
of Physics and Institute for Advanced Study, Nanchang University, Nanchang 330031, China
2Center
for Nano and Micro Mechanics, Tsinghua University, Beijing 100084, China
3School
of Materials Science and Engineering, State Key Laboratory of New Ceramics and Fine Processing, Key Laboratory of
Materials Processing Technology of MOE, Tsinghua University, Beijing 100084, China
42D
Carbon Graphene Material Co., Ltd., Changzhou, Jiangsu 213149, China
#These
authors contributed equally to this work
Received: day month year / Revised: day month year / Accepted: day month year (automatically inserted by the publisher)
© Tsinghua University Press and Springer-Verlag Berlin Heidelberg 2014
ABSTRACT
A wearable and highly sensitive sensor for sound signal acquisition and recognition was fabricated from thin
films of special crisscross graphene woven structures. When it was stretched, high-density of random cracks
in the network appeared, which decreased the current pathway and increased the resistance. Therefore, the
film could act as strain sensors on human throat to measure one’s words through the muscle movement no
matter a sound was made or not. The ultra-high sensitivity of the sensor could realize fast and low frequency
sampling of speech by extracting the signature characteristics of sound waves. In this study, representative
signals of 26 English letters, typical Chinese characters and tones, even phrases and sentences, were tested
with obvious resistance changes. Furthermore, resistance changes of graphene sensor responded perfectly
with sounds. By future combining artificial intelligence with digital signal processing, we expect that this
graphene sensor will be able to deal with complex acoustic systems and large quantities of audio data.
KEYWORDS
Graphene, strain sensors, sound detectors, high sensitivity
1. Introduction
mechanical wave of pressure and displacement,
Sensing strain and vibration with soft materials in
through a medium such as air, water and solids,
small scale has attracted increasing attention. Sound
and its perception and processing is an integral part
is a vibration that propagates as a typically audible
of human intelligence. Traditional sound collection
————————————
Address correspondence to Hongwei Zhu, [email protected]
2
and recognition involve the process of taking
carbon nanotube (CNT)/polymer composite is
samples of an analog sound and storing the results
0.06~0.82 [17], that of traditional metal is 1~5 [18],
in storage mediums or as binary data. Human can
that of semiconductor is ~100 [19], that of
use auditory sensor to receive external sounds
CNT-based sensor is ~1000 [20], and that of single
through vibration of the spreading air. However,
nanowire-based sensor is ~1250 [21]. While the GF
when one cannot make voice physically due to
of graphene woven fabrics (GWFs) from our
some diseases, how could others know what one
research group can be as high as 103 with 2-6%
wants to say? Therefore, it is necessary to develop a
strains, 106 with higher strains (>7%), and ~35 with
new device and method to collect and recognize the
tiny strain of 0.2%, which are the best up to present
communications
and enough to measure with a normal digital
through
some
other
signals
instead of sounding voice.
Multimeter. Additionally, due to the high sensitivity,
Signals, as well as voice, are characterized as
excellent flexibility and no irritation to human
real-world’s
Voice
bodies, the GWFs had been used as stain gauge, the
phonological or prosodic features include loudness,
key component in many monitoring sensors (finger
jitter, fundamental frequency, zero-crossing rate and
bending, hand clenching, expression change, breath
energy frequency ratios, etc [2]. Normally, audio
and pulse monitoring) [8, 22, 23].
the
observable
output [1].
recorder or computer recording software (like
voiceprint)
can
nearly
perfectly
collect
In this study, we demonstrate one interesting
and
example of potential applications of graphene
recognize the human speech sound through the
woven structure in highly sensitive sensing as a
greatly reduced spectral information. However,
wearable
these severely restrict the listener to obtain the full
graphene) from chemical vapor deposition (CVD)
information according to the distribution of spectral
on copper meshes were transferred onto a thin,
energy [3]. When someone is difficult to make or
biocompatiblly
hear a sound, commercial artificial electronic throat
polydimethylsiloxane (PDMS) [24, 25] and double
[4] and electronic ear [5, 6] can be used to produce
sided tape, assembling a highly sensitive GWFs
or collect and recognize speech sound.
based strain sensor. When some stress was applied
acoustic
detector.
and
GWFs
elastic
(crisscross
polymer,
e.g.,
Recently, another type of strain sensor stuck on
on the GWFs, high-density and random cracks
throat skin could record the muscular movement
appeared in the network, leading to the decrease of
(see Fig. 1) to collect and recognize speech sound
current pathway and the increase of resistance.
because
different-degreed
Because vocalization of human is based on the
stretching or shrinking strains with different words
muscle movement/vibration and different contents
spoken [7, 8]. As a voice recognition sensor in the
correspond
throat, sensitivity, flexibility, stretchability and
compression, a new GWFs sensor has been
softness of the sensor are very important [9-12]. In
established to monitor throat muscle movements in
strain sensors [13-15], the gauge factor (GF) is
voice collection and recognition based on the
defined as resistance changing with mechanical
relationship between GWFs stress and its resistance,
deformation to express the sensitivity, and the
compared with the traditional method based on
relationship between the change of electrical
acoustic spectrum. When the assembled sensor is
resistance (dR/R) and the variation of mechanical
affixed onto throat, GWFs will be stretched or
strain (dL/L) is GF=(dR/R)/(dL/L), which means the
compressed along with the skin to change the
strain sensor is more sensitive with a larger GF
resistance
value [16]. It has been reported that the GF of
Furthermore, because of the excellent electrical
throat
muscle
has
to
then
different
to
collect
muscles
the
tension
voice
or
signals.
3
performance and high sensitivity of GWFs, every
did the same action with a voice or without a voice,
English letters, Chinese characters, tone or phrases
which could help the patient who has trouble in
can readily be distinguished based on their different
voicing out.
resistance changes and the characteristic peak of
In order to achieve better performance in voice
each vocalization. Additionally, the low sampling
collection and recognition, the strain sensor should
point in this study, also simplified the traditional
be able to distinguish every letter and the minimum
speech recognition. Interestingly, it was found that
structural unit of a word. Figure 2 shows test signal
the signal waveforms between vocalization and
waveforms of all 26 English letters, and the inset is a
un-vocalization were nearly overlapped with same
real test photograph. As expected, the wave-forms
speaking action.
are unique and repeatable for all letters. Every
2. Results and Discussion
specific letter has its own different amplitude,
Fig. 1a shows the key processes to fabricate and
duration time and resistivity variation. Since each
operate the GWFs strain sensor. During the CVD
individual speech organ is different, people can
growth of graphene [26], GWFs was formed when a
easily distinguish whether the voice comes from the
crisscross copper mesh was used [22]. During the
same person. Similarly, we tested three females and
transferring of GWFs, PDMS was used as an ideal
a male speaking the same English letter “A”. As
adhesive on double-sided tape due to the flexibility,
shown in Fig. S2(a), the key features of the test
biocompatibility,
chemical
signal waveforms were similar. Placing sensors in
inertness and stability of PDMS over a wide range
the same spot each time on different people with
of temperatures [27]. Additionally, the double sided
different anatomy got very similar shapes of
tape was only 100 μm thick and much thinner than
responses for the letter, while same person who
the previous medical tape [8], which led more
repeats his/her own words generates relatively
precise signals to distinguish different words from
different patterns and magnitudes. Louder sound
the tester. Though the strain sensor device was
requires
multi-layered, the total thickness is about 200 μm,
amplitude of the signal is expected to change.
which was thin enough to get ideal signals and
Compared
decrease attenuation. From the image of an
recognition,
integrated strain sensor in Fig. 1(c), it could be
particularly lower (less data points) than those from
found that the wearable device was very thin and
sound spectrum, but one could still get the
could be bent with excellent flexibility. As shown in
characteristic signal waveform of each letter, which
Fig. S1, this flexible strain sensor could fit human
simplified the sound collection compared with
skins very well only with a sizable strain from any
other traditional ways.
high
transparency,
direction. In Fig. 1(d) of a scanning electron
more
muscle
with
our
movement,
other
methods
sampling
thus
of
frequency
the
sound
were
After the clearly identification of every English
microscopy (SEM) image, the wafers on the stripe
letter,
surface
a
different English phrases and sentences was done.
polycrystalline structure. If the assembled strain
As shown in Fig. 3, the resistance changed
sensor was stuck onto a human throat, signals of
significantly when a tester said different words or
resistance changing could be obtained with the
phrases, such as “Graphene”, “Lab of Science”,
movement of throat muscle to monitor vocalization,
“Strain
which is shown in Fig. 1(e). Furthermore and
Sensors”. It was also found that each signal curve
interestingly, it was found that the nearly same
was apparently different and the repeatability was
signal waveforms were obtained when the tester
very good by contrasting the same word of
indicated
that
the
GWF
had
the
voice
Sensors”
recognition
or
through
“Graphene-based
testing
Strain
4
“Graphene” in Fig. 3(a) and 3(d) or the same phrase
response to loudspeakers that are relying on current
of “Strain Sensors” in Fig. 3(c) and 3(d). Some other
signal-caused magnetic energy changing. As shown
tests on monitoring English sentences are shown in
in Fig. 5, some audio experiments with English and
Fig. S2(e). The signal proportion obtained from our
Chinese music with the GWFs-based sensor on the
multiple strain sensors was at least 4% in resistance
vibrating membrane of a loudspeaker were done. It
change, which was enough to measure with normal
was found that GWFs’ collecting signals had a
instrument Keithley and even digital Multimeter.
synchronous response to audio frequency, and
Compared with ZnO (the maximal signal of current
could retain almost every characteristic peak.
change was about 5 nA or voltage change was about
However, the volume of the loudspeaker also
50 mV) [28, 29] and CNTs (the maximal resistance
played an important role on the signal strength. As
change was about 2%) [17, 30] or molded graphene
shown in Fig. S2(d), with a louder loudspeaker
film-based sensors (the maximal current change
volume, the loudspeaker vibration amplitude was
was less 0.8 μA) [7, 31],] our GWFs-based strain
larger
sensor
Comprehensively, same sentence signals could be
has
absolute
predominance
in
voice
recognition in the near future.
and
then
the
signal
was
stronger.
obtained with GWFs-based sensor directly on
Different from English alphabets or phonetic
throat (red) and the loudspeaker (black), compared
symbols with differences in monosyllabic and
with the original recorded data (inset), and every
polysyllabic,
basic characteristic peak had been well recorded.
monosyllabic
Chinese
with
characters
four
are
pronunciation
all
tones.
As
mentioned
previously,
although
the
Therefore, it is necessary to identify Chinese
sampling frequencies of our GWFs-based strain
characters and phrases with our GWFs-based voice
sensor were far lower than that from any existing
recognition sensors, and the test results are shown
recording software, it still could keep every sound’s
in Fig. 4. In Fig. 4(a) and 4(b), respectively, the most
characteristic peaks very well. Fig. 6 shows the test
frequently used Chinese characters of “东”, “西”,
data of four typical animal sounds (dog, cow, bird
“南”, “北” and Chinese tones of “ā”, “á”, “ǎ”, “à”
and horse) and the insets are the original recorded
were tested. From the results, a conclusion could be
audio data. Compared with the original recorded
made that every Chinese character had its own
data, each measured signal from our GWFs-based
characteristic peak and different Chinese tones had
strain sensor had a high degree of consistency.
different signals. Generally, signal waveforms of the
Additionally, due to the excellent flexibility and
third tone often had up and down, and signal
high sensitivity, our GWFs-based strain might be
amplitudes of the forth tone always were the
used in not only the recognition of human sounds,
maximum. And another experiment of “前”, “后”,
monitoring of acoustic activity of animals, but also
“左”, “右” and “ō”, “ó”, “ǒ”, “ò” was done (Fig.
various deformations or vibrations monitoring in
S-2(b) and S-2(c)) to confirm the reliability of the
robotics, fatigue detection, etc.
data
generated
from
our
GWFs-based
voice
This work provides an interesting beginning of
recognition sensors. As shown in Fig. 4(c) and 4(d),
a proof of concept to the ability of the graphene
Chinese phrases and sentences could also be
sensors to generate different signals for different
recognized with our GWFs-based sensitive sensors.
tones and words. However, the results still require
Because the collected signals for GWFs-based
further validation, repetition and cross-validation
sensors came from different degrees of stretching or
between different subjects as well as an explanation
shrinking strain, it is worth to checking whether the
regarding the origin of these signals (muscle
signals of GWFs-based sensors have the same
movement or response to voice vibrations). In
5
addition, signal processing should be done through
Copper mesh (100 mesh, woven with wires of 100
mathematical extraction of signal shapes and
μm in diameter) was used as the CVD growing
features
automated
substrate. H2 (~50 mL/min) and Ar (300 mL/min)
than
just
was flowed until 1000 oC was reached. Then, CH4
comparison of signal shape. For example, Fast
(30 mL/min) was introduced into the reactor while
Fourier Transformation (FFT) can be applied to both
the flow of H2 was turned down to 5 mL/min and
signals and get the Pearson Correlation to extract
Ar was stopped to increase the number of graphene
features of the given voice signal for comparing two
nucleation because the GWFs would more sensitive
audio signals to check the audio signal similarity.
to deformation with smaller graphene grain. After 2
Utilization of machine learning algorithms on the
min, 200 mL/min of Ar was introduced for another
extracted features could provide a good proof that
18 min. After GWFs were completely grown around
the different signals can indeed be differentiated
the copper mesh, GWFs could be obtained after the
and identified as letters and words.
copper was etched away with FeCl3 and HCl
3. Conclusion
aqueous solution.
In summary, a flexible and wearable acoustic
4.2 Fabrication of GWFs-based strain sensor
detector was assembled by adhering graphene
A small amount of PDMS (the weight ratio of base
woven
elastic
to cross linker was 10:1) was spin-coated on a
polymer/double-sided tape film. The detector based
microns double-sided tape (100 μm in thickness).
on strain sensing exhibited ultra-high sensitivity to
As-made materials were degassed under vacuum
tiny strains/vibrations and could be used as
for 20 min to remove bubbles and solidified at 80 oC
electronic skin covering human throat to detect
for 1 h. A PDMS-double sided tape composite
sound waves at low sampling frequencies. The
substrate (about 200 μm in thickness) for GWFs was
results
completed.
and
mathematical
their
analysis
algorithms,
structures
showed
that
using
rather
on
the
the
graphene
sensor
After
GWFs
were
transferred
to
successfully collected and recognized all 26 English
PDMS-tape substrate, silver wires were connected
letters, some typical Chinese characters, phrases,
with silver paste and the multilayer device was
and sentences with wearability, excellent flexibility
assembled.
and high sensitivity. The graphene sensor could
4.3 Custom-made data processing system
perfectly retain every signal characteristic of all
Surface of the GWFs was observed with an optical
peaks, which was highly consistent with that from
microscope (Axio Scope A1) and scanning electron
the loudspeaker. Furthermore and interestingly, it
microscopy (LEO 1530). Audio tests of the strain
was found that the nearly same signal waveforms
sensors were performed with a computer-controlled
were obtained when the tester did the same action
loudspeaker. Voice collection and recognition of the
with a voice or without a voice, which could help
GWFs-based sensor were performed with a digital
the patient who has trouble in voicing out. Because
meter (Keithley 4200-SCS) with a test step of 5 ms
of the excellent flexibility, ultra-high sensitivity,
and a Source-Drain voltage of 1V DC.
good reversibility, friendly biocompatibility, the
GWFs-based sensor might have great potentials in
Acknowledgements
earthquake monitoring, animal communication,
This work was supported by Beijing Science and
and robot voice development.
Technology Program (D141100000514001), National
4. Materials and Methods
4.1 Preparation of GWFs
Science Foundation of China (51372133), and
National Program on Key Basic Research Project
(2011CB013000, 2014CB932401).
6
S. D.; Wu J.; Won S. M.; Tao H.; Islam A.; Yu K. J.; Kim T.-i.;
Electronic Supplementary Material: Supplementary
material (Photographs of GWF strain sensors
adhered on human skin with deformation, and
supplementary recognition signals for different
people and sound volumes, Chinese characters,
tones, and sentences) is available in the online
version
of
this
article
at
http://dx.doi.org/10.1007/s12274-***-****-*
(automatically inserted by the publisher).
References
[1]
Rabiner, L. A tutorial on hidden Markov models and
selected applications in speech recognition. P. IEEE. 1989, 77,
257-286.
[2]
Sharma N.; Gedeon T. Objective measures, sensors and
computational techniques for stress recognition and classification:
A survey. Comput. Meth. Prog. Bio. 2012, 108, 1287-1301.
[3]
Shannon R. V.; Zeng F. G.; Kamath V.; Wygonski J.;
Ekelid M. Speech Recognition with primarily temporal cues.
Science 1995, 270, 303-304.
[4]
Shykhon M.E.; Morgan D.W.; Beng R. D.; Hines E.L.;
Gradner J.W. Clinical evaluation of the electronic nose in the
diagnosis of ear, nose and throat infection: a preliminary study. J.
Laryngol. Otol. 2004, 118, 706-709.
[5] Porwancher R.; Sheth A.; Remphrey S.; Taylor E.; Hinkle
C.; Zervos M. Epidemiological study of hospital- acquired
infection with vancomycin-resistant enterococcus faccium:
possible transmission by an electronic ear-probe thermometer.
Cont. Hosp. Ep. 1997, 18, 771-773
[6]
Hammock M. L.; Chortos A.; Tee B. C.-K.; Tok J. B.-H.;
Bao Z. N. 25th Anniversary article: the evolution of electronic
skin (E-skin): a brief history, design considerations, and recent
progress. Adv. Mater. 2013, 25(42), 5997-6038.
[7]
Wang X.W.; Gu Y.; Xiong Z. P.; Cui Z.; Zhang T.
Silk-molded flexible, ultrasensitive, and highly stable electronic
skin for monitoring human physiological signals. Adv. Mater.
2014, 26, 1336-1342.
[8]
Wang Y.; Wang L.; Yang T. T.; Li X.; Zang X. B.; Zhu M.;
Wang K. L.; Wu D. H.; Zhu H. W. Wearable and highly sensitive
graphene strain sensors for human motion monitoring. Adv.
Funct. Mater. 2014, 24, 4666-4670.
[9]
Kim D.-H.; Lu N. S.; Ma R.; Kim Y.-S.; Kim R.-H.; Wang
Vhowdhury R.; Ying M.; Xu L. Z.; Li M.; Chung H.-J.; Keum H.;
McCormick M.; Liu P.; Zhang Y. W.; Omenetto F. G.; Huang Y.
G.; Coleman T.; Rogers J. A. Epidermal electronics. Science 2011,
333, 838-843.
[10] Xu S.; Zhang Y. H.; Jia L.; Mathwson K. E.; Jang K.-I.; Kim
J.; Fu H. R.; Huang X.; Chava P.; Wang R.; Bhole S.; Wang L. Z.;
Na Y. J.; Guan Y.; Flavin M.; Han Z. S.; Huang Y. G.; Rogers J. A.
Soft Microfluidic assemblies of sensors, circuits, and radios for
the skin. Science 2014, 344, 70-74.
[11] Son D.; Lee J.; Qiao S.; Ghaffari R.; Kim J.; Lee J. E.; Song
C.; Kim S. J.; Lee D. J.; Jun S. W.; Yang S.; Park M.; Shin J.; Do
K.; Lee M.; Kang K.; Hwang C. S.; Lu N.; Hyeon T.; Kim D.-H.
Multifunctional wearable devices for diagnosis and therapy of
movement disorders. Nat. Nanotecnol. 2014, 9, 397-404.
[12] Gong S.; Schwalb W.; Wang Y.; Chen Y.; Tang Y.; Si J.;
Shirinzadeh B.; Cheng W. A wearable and highly sensitive
pressure sensor with ultrathin gold nanowires. Nat. Commun., in
press, DOI: 10.1038/ncomms4132.
[13] Sokolov A. N.; Tee B. C-K.; Bettinger C. J.; Tok Jeffrey
B.-H.; Bao Z. N. Chemical and engineering approaches to enable
organic field-effect transistors for electronic skin applications.
Acc. Chem. Res. 2011, 45, 361-371.
[14] Meital S.-B.; Hossam H. Flexible sensors based on
nanoparticles. ACS Nano. 2013, 7, 8366-83378.
[15] Sun D.-M.; Liu C.; Ren W.-C.; Cheng H. M. A revies of
carbon nanotube- and Graphene-Based flexible thin-film
transistors. Small 2013, 9, 1188-1205.
[16] Rolnick H. Tension coefficient of resistance of matals. Phys.
Rev. 1930, 36, 506-512.
[17] Yamada T.; Hayamizu Y.; Yamamoto Y.; Yomogida Y.;
Najafabadi A.-I.; Futaba D. N.; Hata K. A Stretchable carbon
nanotube strain sensor for human-motion detection. Nat.
Nanotechnol. 2011, 6, 296-301.
[18] Frank; Russel A.; Pierce H. F.; Eric C. Electric Resistance
Strain Gauges. U. S. Patent 2, 2429087, Oct. 14, 1947.
[19] Rasmussen P. A.; Thaysen J.; Hansen O.; Eriksen S. C.;
Bisen A. Optimised cantilever biosensor with piezoresistive
read-out. Ultramicroscopy 2003, 97, 371-376.
[20] Cao J.; Wang Q.; Dai H. Electromechanical properties of
metallic, quasimetallic, and semiconducting carbon nanotubes
under stretching. Phys. Rev. Lett. 2003, 90, 157601.
7
[21] Zhou J.; Gu Y.; Fei P.; Mai W.; Gao Y.; Yang R.; Bao G.;
Tzur G.; Sabanay I.; Mahalu D.; Safran S.; Bershadsky A.;
Wang Z. L. Flexible piezotronic strain sensor. Nano. Lett. 2008,
Addadi L.; Geiger B. Force and focal adhesion assembly: A close
8, 3035-3040.
relationship
[22] Li X.; Sun P. Z.; Fan L.L.; Zhu M.; Wang K. L.; Zhong M.
substrates. Nature Cell Biol. 2001, 3, 466–472.
L.; Wei J. Q.; Wu D. H.; Cheng Y.; Zhu H. W. Multifunctional
[28] Lee M.; Chen C.-Y.; Wang S.; Cha S. N.; Park Y. J.; Kim J.
graphene woven fabrics. Sci. Rep. 2012, 2, 395-402.
M.; Chou L.-J.; Wang. Z. L. A hybrid piezoelectric structure for
[23] Li X.; Zhang R. J.; Yu W. J.; Wang K. L.; Wei J. Q.; Wu D.
wearable nanogenerators. Adv. Mater. 2012, 24, 1759-1764.
H.; Cao A. Y.; Li Z. H.; Cheng Y.; Zheng Q. S.; Ruoff R. S.; Zhu
[29] Xiao X.; Yuan L.; Zhong J.; Ding T.; Liu Y.; Cai Z.; Rong
H. W. Stretchable and highly sensitive graphene-on-polymer
Y.; Han H.; Zhou J.; Wang Z. L. High-strain sensors based on
strain sensors. Sci. Rep. 2013, 2, 870-875.
ZeO nanowire/polystyrene hybridized flexible films. Adv. Mater.
[24] Carpi F.; Gallone G.; Galantini F.; Rossi D. D. Silicone–Poly
2011, 23, 5440-5444.
(hexylthiophene)
Enhanced
[30] Luo S.; Liu T. SWCNT/Graphite Nanoplatelet Hybrid Thin
Electromechanical Transduction Properties. Adv. Funct. Mater.
Films for Self-Temperature-Compensated, Highly Sensitive, and
2008, 18, 235-241.
Extensible Piezoresistive Sensors. Adv. Mater. 2013, 25,
[25] Chan Y.; Mi Y.; Trau D.; Huang P.; Chen E. Micromolding
5650-5657.
of PDMS scaffolds and microwells for tissue culture and cell
[31] Zhu B.; Niu Z.; Wang H.; Leow W. R.; Wang H.; Li Y.;
patterning: A new method of microfabrication by the
Zheng L.; Wei J.; Huo F.; Chen X. Microstructured graphene
self-assembled
arrays for highly sensitive flexible tactile sensors. Small 2014, 10,
Blends
as
micropatterns
Elastomers
of
with
diblock
copolymer
micelles. Polymer 2006, 47, 5124–5130.
studied
using
elastic
micropatterned
3625-3631.
[26] Li X.; Cai W.; An J.; Kim S.; Nah J.; Yang D.; Piner R.;
Velamakannni A.; Jung I.; Tutuc E.; Banerjee S. K.; Colombo L.;
Ruoff R. S. Large-area synthesis of high-quality and uniform
graphene film on copper foils. Science 2009, 324, 1312-1314.
[27] Balaban N. Q.; Schwarz U. S.; Riveline D.; Goichberg P.;
8
Figures
a
CVD
Copper mesh
Etching
Graphene growth
Transferring
Graphene woven fabric
b
200 µm
Ag
Air vibration
Muscular movement
c
d
Ag
Graphene woven fabric
PDMS
Double sided tape
e
Vocalized
Not vocalized
(R-R0)/R0
0.2
0.1
0.0
0.0
0.8
1.6
2.4
Time (s)
Figure 1
GWFs-on-PDMS strain sensor. (a) Key steps of preparing the graphene strain sensor; (b) Three ways of collecting and
recognizing human voices; (c) Photograph of a bent strain sensor; (d) SEM image of GWFs; (e) Signals of vocalization (black) and
un-vocalization (red) are nearly overlapped when a tester made the same speech action.
9
0.16
0.20
0.04
0.00
(R-R0)/R0
(R-R0)/R0
(R-R0)/R0
0.08
0.08
0.04
0.5
1.0
1.5
2.0
0.5
1.0
1.5
2.0
2.5
E
-0.04
2
3
1.0
1.5
2.0
0
1
2
3
4
Time (s)
G
0.16
0.08
0.00
1
0.04
0.04
(R-R0)/R0
(R-R0)/R0
0.00
0
0.5
F
0.24
0.04
0.08
Time (s)
0.32
0.08
0.12
0.00
0.0
Time (s)
Time (s)
D
0.16
0.00
0.0
2.5
C
0.04
0.00
0.0
(R-R0)/R0
0.08
B
0.12
(R-R0)/R0
A
0.12
0.00
-0.04
0.0
0.5
Time (s)
1.0
1.5
0.0
2.0
0.5
1.0
1.5
2.0
2.5
Time (s)
Time (s)
0.16
0.06
0.00
0.10
0.05
0.0
0.5
1.0
0
1
Time (s)
0.24
0.04
0.08
0.04
0.00
0
1
2
0.0
3
0.5
0.20
M
0.12
0.06
1.5
2.0
N
0.15
0.18
1.0
Time (s)
Time (s)
0.10
0.05
0.00
0.00
0.00
0.0
0.5
0.0
1.0
0.5
1.0
1.5
2.0
-0.05
2.5
0
Time (s)
Time (s)
1
2
Time (s)
0.20
0.24
(R-R0)/R0
0.18
0.12
0.06
Q
0.12
P
0.15
0.10
0.05
R
0.18
(R-R0)/R0
O
(R-R0)/R0
0.24
(R-R0)/R0
0.04
-0.04
3
(R-R0)/R0
(R-R0)/R0
(R-R0)/R0
0.08
Time (s)
L
0.12
2
K
0.08
0.00
0.00
-0.06
J
0.12
0.15
(R-R0)/R0
(R-R0)/R0
(R-R0)/R0
0.12
I
(R-R0)/R0
0.20
H
0.18
0.08
0.04
0.12
0.06
0.00
0.00
0.00
0
1
2
0.00
0
3
1
2
0
3
1
0
1
2
Time (s)
Time (s)
Time (s)
Time (s)
2
0.16
0.08
0.00
(R-R0)/R0
0.16
0.04
0.00
-0.08
0.5
0
Time (s)
0.08
0.04
0.24
0
(R-R0)/R0
0.08
1
2
0.16
X
0.16
0.04
0
Time (s)
(R-R0)/R0
W
0.08
0.00
2
Time (s)
0.12
0.00
1
V
0.12
0.12
0.00
1.0
0.06
U
1
2
Time (s)
Y
Z
0.08
0.12
(R-R0)/R0
0.0
(R-R0)/R0
0.16
T
(R-R0)/R0
0.08
S
(R-R0)/R0
(R-R0)/R0
0.24
0.08
0.04
0.04
0.00
0.00
-0.06
0
1
0
Time (s)
Figure 2
1
Time (s)
0.00
0
1
Time (s)
2
0
1
2
3
Time (s)
Recognition of 26 English letters.
10
a
0.12
b
“graphene”
“lab of science”
0.08
(R-R0)/R0
0.08
(R-R0)/R0
0.12
0.04
0.00
0.04
0.00
-0.04
-0.04
0
1
0
2
1
c
0.08
d
“strain sensors”
0.24
“graphene based strain
sensors”
0.16
(R-R0)/R0
(R-R0)/R0
0.04
0.00
0.08
0.00
-0.04
0
1
Time (s)
Figure 3
2
Time (s)
Time (s)
2
0
1
2
3
4
Time (s)
Recognition of English words and phrases. Relative resistance changes with throat muscle motions when the tester said
(a) “graphene”; (b) “lab of science”; (c) “strain sensors”; and (d) “graphene based strain sensors”.
11
a
b
0.08
0.08
à
北
南
(R-R0)/R0
(R-R0)/R0
东
0.04
西
ā
á
0.04
0.00
0.00
0
5
10
15
20
25
0
5
10
Time (s)
c
ǎ
0.18
15
20
25
30
Time (s)
d
“石墨烯”
“石墨烯传感器”
0.08
(R-R0)/R0
(R-R0)/R0
0.12
0.06
0.00
-0.08
0.00
-0.16
0
1
Time (s)
Figure 4
2
3
0
1
2
3
4
Time (s)
Recognition of Chinese characters, tones, phrases. Relative resistance changes for (a) Chinese characters and (b) tones.
Relative resistance changes with throat muscle motions when the tester said (c) “石墨烯” and (d) “石墨烯传感器”.
12
c
a
2.0
Music
0.2
(R-R0)/R0
(R-R0)/R0
e
2.5
0.3
0.1
English audio
1.5
1.0
0.5
0.0
0.0
0
b
10
20
Time (s)
30
0
20
Time (s)
30
40
d
0.6
3
0.5
0.4
Chinese audio
(R-R0)/R0
(R-R0)/R0
10
0.3
0.2
2
1
0.1
0.0
0
0
10
20
Time (s)
Figure 5
30
0
2
4
6
8
10
Time (s)
Recognition signals compared with the sounds from the loudspeaker. Relative changes of resistance when the
loudspeaker played (a) music; (b) Chinese audio and (c) English audio, respectively; (d) A same sentence tested by the sensors stuck
on the throat (red), on the loudspeaker (black). The insets show the original recording data; (e) Photograph of a strain sensor stuck on
the vibrating membrane of a loudspeaker.
13
0.6
(R-R0)/R0
0.4
Dog
0.2
b
0.6
(R-R0)/R0
a
0.4
0.0
Cow
0.2
0.0
0
4
8
12
16
0
4
0.06
0.04
d
Bird
0.02
0.00
12
16
0.10
Horse
0.05
0.00
0
4
8
12
16
0
Time (s)
Figure 6
8
Time (s)
(R-R0)/R0
c
(R-R0)/R0
Time (s)
4
8
12
Time (s)
Recognition of animal sound records. Relative resistance changes when the loudspeaker played sound records of (a) dog,
(b) cow, (c) bird and (d) horse, respectively. The insets show the original recording data.
14