A Highly Aligned Nanowire‐Based Strain Sensor for Ultrasensitive Monitoring of Subtle Human Motion

Abstract

Achieving highly accurate responses to external stimuli during human motion is a considerable challenge for wearable devices. The present study leverages the intrinsically high surface‐to‐volume ratio as well as the mechanical robustness of nanostructures for obtaining highly‐sensitive detection of motion. To do so, highly‐aligned nanowires covering a large area were prepared by capillarity‐based mechanism. The nanowires exhibit a strain sensor with excellent gauge factor (≈35.8), capable of high responses to various subtle external stimuli (≤200 µm deformation). The wearable strain sensor exhibits also a rapid response rate (≈230 ms), mechanical stability (1000 cycles) and reproducibility, low hysteresis (<8.1%), and low power consumption (<35 µW). Moreover, it achieves a gauge factor almost five times that of microwire‐based sensors. The nanowire‐based strain sensor can be used to monitor and discriminate subtle movements of fingers, wrist, and throat swallowing accurately, enabling such movements to be integrated further into a miniaturized analyzer to create a wearable motion monitoring system for mobile healthcare.

Full text

2001363 (1 of 9) © 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
Full PaPer
A Highly Aligned Nanowire-Based Strain Sensor for
Ultrasensitive Monitoring of Subtle Human Motion
Ning Tang, Cheng Zhou, Danyao Qu, Ye Fang, Youbin Zheng, Wenwen Hu, Ke Jin,
Weiwei Wu,* Xuexin Duan, and Hossam Haick*
Dr. N. Tang, Dr. W. Hu, Prof. K. Jin
School of Aerospace Science and Technology
Xidian University
Xi’an, Shaanxi 710126, China
Dr. N. Tang, Dr. Y. Zheng, Prof. H. Haick
Department of Chemical Engineering and Russell Berrie
Nanotechnology Institute
Technion-Israel Institute of Technology
Haifa 3200003, Israel
E-mail: hhossam@technion.ac.il
C. Zhou, Y. Fang, Prof. X. Duan
State Key Laboratory of Precision Measuring Technology & Instruments
Tianjin University
Tianjin 300072, China
D. Qu, Prof. W. Wu, Prof. H. Haick
School of Advanced Materials and Nanotechnology
Xidian University
Xi’an, Shaanxi 710126, China
E-mail: wwwu@xidian.edu.cn
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/smll.202001363.
DOI: 10.1002/smll.202001363
sensors are highly ecient and sensitive
for monitoring human motion.[5,6] How-
ever, their high cost and complex opera-
tion hinder their practical application. In
recent years, an approach that could over-
come these diculties has emerged, based
on flexible strain sensors[7–9] that operate
with various transduction mechanisms
such as capacitance,[10] triboelectricity,[11]
piezoelectricity,[12] and piezoresistivity.[13]
The development of flexible strain sensors
that interface with human skin has led to
advances in electronic monitoring of skin
health status and human–machine inter-
faces.[14–17] In order to improve the sensing
performance, many materials with nano-
structures have been combined with elas-
tomers to form sensing elements, such as
carbon nanotubes/polydimethylsiloxane
(PDMS),[18] graphene/rubber,[19] and metal
nanoparticles.[20] Despite the great poten-
tial of these materials and devices, most of
them involve costly multistep fabrication processes and show
very limited responses to subtle external stimuli.
Spray coating and vapor phase polymerization are simple
manufacturing methods for fabricating strain sensors.[21,22]
However, the low gauge factor of strain sensors fabricated by
spray coating and the high power consumption during fabri-
cation by vapor phase poly merization have hindered their fur-
ther development and application as wearable strain sensors.
In addition, mass production of integrated flexible strain sen-
sors assembled from nanostructure materials (e.g., nanowires,
nanotubes, or nanofibers) by traditional printing methods (e.g.,
spray, inkjet, or transfer printing)[23–26] would be scarcely repro-
ducible in many cases. To overcome these problems, eorts
have been made to improve the performance of strain sensors
by optimizing the functional materials or the alignment of
nanostructures that can transduce subtle external strains e-
ciently into electrical signals.[27–29]
Nanomaterials with well-aligned structures have intrinsi-
cally high aspect ratios and mechanical robustness that enable
them to amplify received bending signals and make the sensing
more reliable. They can also ensure uniformity in the shape
of sensing elements and decrease device-to-device variation in
performance.[30–32] Capillarity, a natural phenomenon caused
by surface tension, is useful for patterning polymeric materials
from solutions into nanostructures.[33,34] In this work, we intro-
duce a flexible strain sensor based on highly aligned nanowires
Achieving highly accurate responses to external stimuli during human motion
is a considerable challenge for wearable devices. The present study leverages
the intrinsically high surface-to-volume ratio as well as the mechanical
robustness of nanostructures for obtaining highly-sensitive detection
of motion. To do so, highly-aligned nanowires covering a large area were
prepared by capillarity-based mechanism. The nanowires exhibit a strain
sensor with excellent gauge factor (≈35.8), capable of high responses to
various subtle external stimuli (≤200µm deformation). The wearable strain
sensor exhibits also a rapid response rate (≈230ms), mechanical stability
(1000 cycles) and reproducibility, low hysteresis (<8.1%), and low power
consumption (<35 µW). Moreover, it achieves a gauge factor almost five
times that of microwire-based sensors. The nanowire-based strain sensor
can be used to monitor and discriminate subtle movements of fingers, wrist,
and throat swallowing accurately, enabling such movements to be integrated
further into a miniaturized analyzer to create a wearable motion monitoring
system for mobile healthcare.
1. Introduction
Detection of human posture and activity is vital for assessing
physical rehabilitation and kinematics.[1–4] Traditional
optic-based or microelectromechanical systems (MEMS)-based
Small 2020, 2001363

2001363 (2 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
fabricated by a capillarity-based imprinting method to signifi-
cantly enhance the sensing of mechanical stimuli. Having inves-
tigated the implications of nanomorphology for mechanical
properties, we demon strate that the sensitivity of a nanowire-
based strain sensor can be made five times greater than that of
a strain sensor with microstructures. The developed nanowire-
based strain sensor also shows a high response to tiny external
stimuli (≤200µm), a fast response (230 ms), high cycling sta-
bility (1000 times), and low hysteresis (8.1%). As a proof of con-
cept, the nanowire-based strain sensor was examined by various
human-based characterization techniques, showing that it can
be used to monitor subtle joint movements. This sensor creates
an ecient and portable approach to monitoring various subtle
human movements.
2. Results and Discussion
2.1. Fabrication and Characterization of a Nanowire-Based
Strain Sensor
Capillarity-based imprinting was used to prepare highly
ordered nanowires.[35] The integration of the fabricated nano-
wires into a strain sensor and further into a wearable device
is illustrated schematically in Figure1a. Briefly, a reusable
PDMS mold containing well-defined patterns (Figure S1, Sup-
porting Information) was used to contact a flexible substrate
(e.g., polyethylene terephthalate (PET)) to form several parallel
microchannels. A few drops of commercially available poly(3,4-
ethylenedioxythiophene):poly(sty renesulfonate) (PEDOT:PSS)
solution were added at one edge of the mold and flowed easily
into the microchannels, gradually forming several separate
liquid strips owing to the Laplace pressure (detailed calculations
of Laplace pressure are summarized in the Supporting Infor-
mation). During evaporation, the solution was attracted to the
microchannel side walls, which provided sites for PEDOT:PSS
deposition. Once the solvent had completely evaporated, the
PDMS mold was gently removed, leaving ordered nanowires
on the flexible substrate. Interdigitated electrodes were then
deposited on the nanowires by vacuum evaporation and encap-
sulated with a thin PDMS film. The result is a nanowire-based
flexible strain sensor for detecting subtle motion was success-
fully fabricated.
From the color stripes illustrating the nanowire array in
Figure1b, fully aligned PEDOT:PSS nanowires were extended
over a relatively large area (>10mm × 10 mm). The fabricated
nanowires had no obvious defects, indicating the eectiveness
of the imprinting technique (Figure S2, Supporting Infor-
mation). Because of the mechanism of the capillarity-based
imprinting method, the concentration of PEDOT:PSS is impor-
tant for the dimensions of the prepared structures. Therefore,
PEDOT:PSS solutions at dierent concentrations (1.3, 0.65, and
0.32 wt%) were used to optimize the dimension and structure
of the nanowires. As the conductive polymer concentration
was reduced, the width of the nanowires decreased accordingly
(Figure S3, Supporting Information). Lower concentrations also
meant fewer defects on the micro/nanowires. However, when
the PEDOT:PSS concentration was reduced to 0.32 wt%, no
obvious nanowire was formed on the substrate. Therefore, in
subsequent experiments, the PEDOT:PSS concentration was
fixed at 0.65 wt%. Besides the concentration, the depth of the
microchannels on the PDMS mold also significantly aected
the width and height of the nanowires. Using the 5 µm deep
PDMS mold, ordered nanowires with uniform width and
smooth surfaces were fabricated (Figure1c). As measured from
the magnified scanning electron microscopy (SEM) and atomic
force microscopy (AFM) height images (Figure1d,e), the fabri-
cated nanowires had an average width (w) of 110± 5nm and a
height (h) of 130±5nm, much smaller than the structures fab-
ricated by a 100µm deep PDMS mold (Figure S4, Supporting
Information). These results demonstrate that the width and
height of the assembled nanowires fabricated by the imprinting
method can be tuned by adjusting the PEDOT:PSS concentra-
tion and the dimensions of the PDMS mold to meet the require-
ments of dierent applications. After the interdigitated elec-
trodes were deposited on the nanowires, a flexible strain sensor
consisting of ordered nanowires was successfully prepared
(Figure 1f; Figure S5, Supporting Information). According to
the pad length of the interdigitated electrodes and the interval
between adjacent nanowires, the number of the nanowires was
≈660 (n≈ 660). Therefore, assuming that each nanowire has
the same conductivity as the bulk material (σ≈ 1 S cm−1), the
resistance of these parallel nanowires is (1/Ne)(1/n)(1/σ)(l/wh) ≈
0.0235 MΩ, where Ne represents the number of intervals
between fingers. The measured resistance of the nanowire-
based strain sensor (Figure 1g) was ≈0.0299 MΩ, which is
close to the calculated value, implying that capillarity-based
imprinting can create nanostructures of high quality in a
simple low-cost manner that lends itself to mass production.
2.2. Mechanical Properties of the Nanowire-Based Strain Sensor
To measure its mechanical properties, the nanowire-based
strain sensor was fixed in a tensile machine and monitored by
the relative change in resistance (Figure S6, Supporting Infor-
mation). As shown in Figure2a,b, variation in resistance gradu-
ally increased with the increase of forward distance (dL) when
the bending direction was parallel to the direction of the nano-
wires, implying that the nanowire-based strain sensor can be
used to distinguish bending states at dierent levels. The strain
sensor showed a high response even with a 200µm deforma-
tion, indicating excellent ability to monitor subtle movements.
In addition, it responded positively when bent outward. Con-
versely, when the sensor was bent inward, the response in resist-
ance variation was negative. The literature[36,37] reports that the
structural model of PEDOT:PSS fabricated from a water solu-
tion is a nanocrystal of PEDOT surrounded by PSS. Therefore,
variations in resistance caused by bending are mainly due to the
changes in the reversible contact between adjacent PEDOT:PSS
nanocrystals in the assembled nanowires. Specifically, inter-
vals between neighboring PEDOT:PSS nanocrystals increase
when bent outward, increasing the charge transfer distance and
therefore the sensor resistance. In contrast, inward bending
compresses the nanocrystals, so they overlap each other, short-
ening the charge transfer distance and decreasing the sensor
resistance (Figure S7, Supporting Information). In order to test
the mechanical anisotropy of the nanowire sensor, we examined
Small 2020, 2001363

2001363 (3 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
the case in which the bending direction was perpendicular to
the direction of the nanowires; the outward and inward bending
results calculated from the I–V conductivity measurements
(Figure S8, Supporting Information) under dierent for-
ward distances displayed almost no change (Figure S9a,
Supporting Information). Meanwhile, after 3000 cycles of
Figure 1. a) Schematic illustration of preparation of the ordered nanowires and integration into the fabricated strain sensor. b) The colored stripes formed
by the nanowires at 5–10µm intervals were fabricated by the 5µm depth PDMS mold on the PET substrate. c) Optical microscopic images of the nano-
wire array fabricated on the PET substrate by the 5µm deep PDMS mold. d) High-magnification SEM image of a single nanowire. e) Side view 3D AFM
image of the nanowire array. f) Picture of the integrated flexible device. g) I–V conductivity measurement of the PEDOT:PSS nanowire-based strain sensor.
Small 2020, 2001363

2001363 (4 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
outward bending/relaxing testing, the response showed a neg-
ligible rise (Figure S9b, Supporting Information), indicating
that the strain sensor is mechanically stable and is insensitive
to bending. In our previous studies, we demonstrated that the
insensitivity of the nanowire-based sensor to curvature under
such bending directions is mainly due to the alignment of the
nanowires, which reduces the resistance caused by the strain.[38]
In addition, the strain distributed on to it is much less than
the strain on the PET substrate owing to the extremely small
configuration of the nanowire, making the sensor mechani-
cally robust.[39] Therefore, the aforementioned results allow
us to conclude that, owing to the preferential alignment of the
nanowires, the strain sensor is mechanically highly anisotropic,
which is critical for developing multidimensional strain sensors
for subtle movements.
Figure2c,d shows the response time of the nanowire-based
strain sensor under the outward and inward bending states,
respectively, when the bending forward distance dL was about
0.2mm. The strain sensor showed an immediate response time
of ≈230ms, which is fully sucient for motion monitoring sys-
tems and robotic applications.[40] The relationship between the
forward distance (dL) and the bending radius (r) was calculated
using the following equation
2/ /12
2
s
22
r
L
dL LhL
ππ
()
()
=− (1)
where L and hs denote the initial length and the PET thickness,
respectively (Figure 2e).[41] Table S1 (Supporting Information)
shows the dL set in the experiments and the corresponding
bending radius. The assumed length of the conductive path (l)
can be defined as l= θ• (r+ d), where θ is the central angle,
d= h/2, and h is the thickness of the sensor. Since the height of
the nanowire is much less than the PET thickness, h is approx-
imately equal to that thickness (hs= 100 µm). Therefore, the
mechanical strain Δε can be calculated as
00
00
l
l
rd
rd
rd
tt
ε
θθ
θ
()()
()
∆=
∆=
+− +
+ (2)
where θ0 and θt represent the central angles before and after
bending, respectively. Since r0 >> d, and θtrt≈ θ0r0, Equation(2)
can be converted to
11
11
0
00 00 0
0
0
d
r
d
r
d
r
r
r
d
rr
tt
tt
εθθ
θ
θ
θ
()
∆=
−=−





=−






=−





 (3)
Since r0 tends to infinity, Δε= d/rt= hs/2rt. Similarly,
Δε=−hs/2rt in the inward bending state. Combining this with
Equation(1)
//
12
s
2
s
22
hdLL
hL
L
t
ε
ππ
()
()
∆=
±− (4)
Sensitivity is important measure for assessing the ability of
a strain sensor to monitor subtle movements. It has been char-
acterized by the gauge factor (GF), defined as GF = (ΔR/R0)/Δε.
As shown in Figure2f, for a relatively small range of strains
(between −0.64% and 0.24%), the GF value is 15.56, increasing
to 24.17 and 35.79 in the ranges −1.08% to −0.68% and 0.42% to
Figure 2. Where the bending direction was set parallel to the direction of the nanowires, the real-time responses in resistance of the nanowire strain
sensor are shown at dierent a) outward and b) inward bending forward distances. Time response of the nanowire-based strain sensor under c)
outward and d) inward bending with a 0.2mm forward distance. e) Schematic image of the bending sensor to show the definition of the physical
parameters. f) Relative resistance changes of the nanowire-based sensor under dierent strains. GF values were obtained by linear fitting.
Small 2020, 2001363

2001363 (5 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
1.08%, respectively, which are typical for strain sensors in these
stages. A theoretical model of the PEDOT:PSS nanowire-based
strain sensor was built to elucidate the eects (Figure S10, Sup-
porting Information). When applied to a relatively small strain,
the response can be expressed as ΔR/R0= N0/Nt− 1, indicating
an approximately linear relationship between the response and
the rate of change in transfer paths, which is close to the experi-
mental results (the derivation of ΔR/R0 is shown in Figure S10,
Supporting Information). However, upon further outward
bending strain, the primary factor aecting the response
changed to a particle tunneling eect; thus, the resistance in
this stage can be expressed as
2
32
exp42
p
2
p
R
M
N
dh
Ae m
dm
h
φ
πφ
=⋅





 (5)
where M is the number of nanocrystals within a conducting
path and N is the number of the charge transfer path, hp is
Planck’s constant, and d is the average tunneling distance
between adjacent nanocrystals, A is the eective cross-sectional
area, and m and e are the charge carrier mass and charge,
respectively.[42] The resistance increases exponentially as the
tunneling distance increases, but the response can still be
fitted by a linear relationship with a higher gradient when a
small strain range from 0.42% to 1.08% is applied. Further-
more, we noticed that the GF value was higher during outward
than inward bending, which could be attributed to the dif-
ferent structural changes and the exponential dependence of
variation in the relative resistance.[43] Specifically, according to
Equation(5), the resistance of the strain sensor increases expo-
nentially when it is subjected to an outward bending strain.
However, it increases linearly (ΔR/R0= N0/Nt− 1) under an
inward bending strain, which results in a relatively small resist-
ance change. In view of its outstanding sensitivity, the sensor
can provide eective help in monitoring subtle human motion
during daily life. We have compared the GF value of this sensor
with other reported strain sensors (Table S2, Supporting Infor-
mation), concluding that in addition to having a much easier
and cheaper fabrication process, our strain sensor gives a supe-
rior performance.
To fully evaluate the nanowire-based strain sensor, stability,
reproducibility, and hysteresis were all considered and exam-
ined. The real-time response of the strain sensor to 500 testing
cycles—each of outward and inward bending—is shown in
Figure3a. Irrespective of the bending state, changes in resist-
ance throughout the cycles showed excellent stability, demon-
strating the durability of the sensor to repeated bending and
relaxing cycles. Variations in the resistance showed a slight
drift in contrast to the initial value during both outward and
inward bending cycles. This drift could be because the strain
causes rearrangement between the nanocrystals during
bending/relaxing. This is consistent with the above-mentioned
response mechanism. Moreover, these results demonstrate
that the nanowire-based strain sensor—if the bending direc-
tion is parallel to the nanowire—has good cycling stability
and its performance is unaected by the multiple bending
operations. Besides cycling stability, antiaging stability was
also investigated (Figure S11, Supporting Information). It can
be seen from the comparison of results that after 4 h of aging
(90% relative humidity (RH) at 50 °C), the performance of
the nanowire-based strain sensor did not change significantly.
This good antiaging stability could be attributed to the PDMS
film and PET substrate protecting the hygroscopic material
(PEDOT:PSS) from atmospheric moisture. The stretchability of
the nanowire-based strain sensor was then tested by horizontal
stretching (Figure S12, Supporting Information). The strain
sensor showed that no cracks were generated until a PET sub-
strate strain of 2.7% was induced. To check the reproducibility
of the sensor, it was examined after repeated bending and
relaxing over four cycles. Its real-time response to the same out-
ward or inward bending states sustained an almost equal signal
value during each testing cycle, implying good reproducibility
(Figure3b). More importantly, the hysteresis of the nanowire-
based strain sensor is quite small (≈8.1%), and this could
be repeatedly tested during continuous cycles (Figure3c). All
these results imply that the strain sensor can perform stably
and reliably in practical applications.
To study nanomorphological influences on the response
of the strain sensor, two microwire arrays (Microwire_1 and
Microwire_2) were fabricated by capillarity-based imprinting
for comparison purposes. These two microwire arrays were
prepared using the same concentration of PEDOT:PSS as the
aforementioned nanowires, but with dierent PDMS mold
widths and dierent solution volumes. The average morpho-
logical parameters (width/height) of these two microwires were
4µm/180 nm and 8 µm/400 nm, respectively (Figure4a,b). To
compare the performances of these PEDOT:PSS-based strain
sensors, variation in micro/nanowire resistance was monitored
Figure 3. a) The cycling stability and b) the reproducibility tests of the strain sensor at a forward distance of 1.4mm. c) Relative resistance variation
of the strain sensor as a function of the applied forward distance.
Small 2020, 2001363

2001363 (6 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
(Figure4c). The nanowire-based strain sensor achieved a gauge
factor almost five times that of the microwire-based sensors,
consistent with previous studies.[44,45] A key result from this
comparison was the strong eect of PEDOT:PSS wire mor-
phology on the sensitivity of the strain sensor. Specifically,
strain sensors made of microwires have relative responses
with similar dependence on applied strains (GF ≈ 8.4 and ≈4.3,
respectively), while the strain sensor made of nanowires is
much more sensitive (GF ≈ 29.5) in low strain (<0.6%) applica-
tions. The underlying mechanism could be the great eect of
the applied strain on the distance between adjacent PEDOT:PSS
nanocrystals in the nanowire-based sensor, resulting in a
smaller charge carrier transfer path. For microwire-based strain
sensors, one can conceive that the multilayer nanocrystals in
the microwire form bridges with each other even when strain
is applied, so the distances and the number of charge carrier
transfer paths are less aected by the applied strain, resulting
in a small variation in resistance (Figure4d).
2.3. Detection of Subtle Human Motion by the Nanowire-Based
Strain Sensor
Since the nanowire-based strain sensor has great flexibility,
reproducibility, and high sensitivity, it can be used as a wear-
able device to monitor dierent subtle body motions. In order
to test its performance in practical applications, it was attached
to the top surface of the forefinger. The response increased with
the bending angle of the forefinger (Figure5a). Even when the
finger underwent only a slight change in bending the sensor
responded, indicating the excellent performance of this sensor
in monitoring subtle movements. Subsequently, the perfor-
mance of the sensor for dierent wrist motions was checked
(Figure 5b,c). This motion involved outward and inward
bending at the joint, which drives the sensor to bend in the
corresponding directions. The sensor responded positively
and negatively to inward and outward bending of the wrist,
respectively, consistent with our experimental results. Notably,
the response speed of the sensor fully meets the requirements
of capturing changes in joint motions, implying high sensor
reliability (Figure 5c). During the bending and straightening
motions, variation in the resistance changed according to the
degree of joint bending, further indicating the excellent relia-
bility of the nanowire-based strain sensor. Moreover, the sensor
could clearly recognize a swallowing action when it was placed
on the throat (Figure5d). The results also indicate that when
the joint returned to its initial position, the response returned
completely to the original level, demonstrating the stability of
the device in monitoring subtle joint motions.
A flexible sensor integrated with portable electric devices to
form a sensing system for monitoring human motion in real
time is important in modern healthcare and for diagnosis of
diseases. Therefore, we integrated the nanowire-based strain
sensor with other electronic components on a printed circuit
board (PCB) to build a wearable electronic device that can detect
movement portably (see Figure S13 in the Supporting Infor-
mation). The red dashed boxes in Figure 5e indicate that the
wearable device is very small and does not aect normal move-
ment. A Smartphone-compatible application program was also
developed using the Android Studio platform (details can be
found in the Supporting Information), which receives and dis-
plays the joint motion signals in real time (green dashed box in
Figure 4. CLSM images of microwires with widths of a) 4 µm and b) 8 µm. c) Variation in relative resistance as a function of applied strain for
PEDOT:PSS-based strain sensors with dierent morphologies. d) Schematic image of the PEDOT:PSS micro/nanowire-based strain sensor under
outward bending states.
Small 2020, 2001363

2001363 (7 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
Figure 5. Real-time monitoring of dierent human movements. Optical images and relative resistance changes of the nanowire-based strain sensor in
response to a) forefinger bending (5°, 45°, 90°), b,c) wrist bending—b) 60°, 30°, 0°, −30°, −60°; and c) ≈±30°, ≈±45°, ≈±60°, and d) throat swallowing.
e) Optical pictures and schematic diagram (blue dashed box) of the wearable system. The red dashed boxes indicate the integrated sensing device,
and the green dashed box indicates the Smartphone screen showing the real-time monitoring of movements. The integrated wearable device consists
of 1) a STM 32 microcontroller, 2) an AD5933 chip, 3) a crystal oscillator chip, 4) a Bluetooth module connector, and 5) a source module connector.
The (6) in the red dashed box indicates the location of the nanowire-based strain sensor.
Small 2020, 2001363

2001363 (8 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
Figure5e). Thus, the portable subtle motion monitoring system
bridges the technology gap between signal generation from the
nanowire-based strain sensor and data transmission, providing
a convenient approach for monitoring human motion. All the
above-mentioned experiments and operations were conducted
at low voltage; the power consumption of the nanowire-based
strain sensor was calculated at <35 µW, which is much lower
than other reported flexible sensors for motion detection.[46,47]
Collectively, all these results demonstrate that this nanowire-
based strain sensor has many excellent advantages, such as easy
manufacturing, low cost, high sensitivity, and low power con-
sumption; hence, it can provide helpful information regarding
personal healthcare based on its ability to monitor subtle joint
motions accurately, especially for patients in the absence of a
doctor.
3. Summary and Conclusions
In summary, we demonstrated a highly sensitive strain sensor
based on aligned nanowires that can be fabricated with a few
minutes by capillarity-based imprinting. The results showed
that strain sensor based on nanomorphology has greatly
improved sensitivity, so it can detect slight bending changes up
to 200µm with a strong and rapid response. Owing to the high
alignment of the nanowires, the strain sensor gave excellent
mechanical performance, including cycling stability and repro-
ducibility. In addition, it has a wide range of applications in
monitoring dierent subtle human motions. Considering the
ease of fabrication, low cost, and powerful performance, this
flexible nanowire-based strain sensor could open exciting ave-
nues in monitoring subtle changes in human (notably patient)
motion with enhanced sensitivity. This type of monitoring
should be particularly useful in “gait clinics.”
4. Experimental Section
Materials and Apparatus: 1H,1H,2H,2H-Perfluorodecyltrichlorosilane
(PFDTS, 96%) was obtained from Aladdin Industrial Corporation and
used without further purification. The conductive polymer PEDOT:PSS
was purchased from Sigma–Aldrich as an aqueous suspension
(≈1.3 wt%) and diluted directly with deionized water. PDMS was
prepared from Sylgard 184 silicone elastomer. SEM, AFM, and confocal
laser scanning microscopy (CLSM) involved FEI FP 2031/12 Inspect
F50, Dimension Icon (Bruker), and OLYMPUS OLS5000 equipment,
respectively. Mechanical properties were measured by a tensile testing
machine (ZQ-990B). Electrical measurements at constant voltage (+1V)
were obtained using an Agilent Technologies B1500A. The responses
were calculated by the equation
RR
R
R
R
t
Resp
onse 0
00
=−=∆ (6)
where R0 and Rt are the resistances of the nanowire-based strain sensor
measured in the initial steady state and bending state, respectively.
Preparation of Large-Area-Ordered Micro/Nanowires: The reusable
patterned PDMS mold was built by casting the prepolymer base and
curing agent at 10:1 (w/w) ratio on to a silicon template upon which
an antiadhesion layer (PFDTS) had been deposited. After degassing to
remove air bubbles and curing at 80°C for 2 h, the PDMS was cooled
to room temperature and peeled from the silicon wafer. Thus, a PDMS
mold containing well-defined parallel microgrooves was prepared.
Subsequently, the PDMS mold was treated with O2 plasma (20 mTorr;
20 sccm; 30 W; 40 s) and then placed in contact with a PET substrate.
The PEDOT:PSS solution was dropped on the open end of the PDMS
mold and flowed spontaneously into the microchannels by capillary
pressure. After the solvent had completely evaporated, the PDMS mold
was gently removed from the PET substrate.
Fabrication of Interdigitated Electrode: Interdigitated electrodes of
50nm Au thickness were deposited by vacuum evaporation through a
homemade copper shadow mask with a pad length (Lpad) 5 mm and
width (W) 200 µm. The number of fingers was 10 (Ne= 9, where Ne
represents the number of intervals between fingers) and the spacing
(lspace) between the adjacent fingers was kept at ≈200µm.
Motion Detection with a Smartphone: The Smartphone-enabled
wearable system based on the nanowire-based strain sensor was used as
a portable tool for recording human motion. Using the developed Android
application programs, real-time electrical signals can be received and
plotted on the phone screen. First, the “Bluetooth” keypad was designed
for the Smartphone to search and connected to the portable detection
device. Thereafter, the microcontroller in the PCB received commands to
activate the resistance analyzer, which sent out feedback signals from the
nanowire-based sensor to the Smartphone. A coordinate graph including
the signals can be displayed in real time on the phone screen.
The experiments involving human subjects have been performed with
their full, informed consent.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
This research received funding from the Phase-II Grand Challenges
Explorations award of the Bill and Melinda Gates Foundation
(OPP1109493) and Horizon 2020 ICT grant under the A-Patch project.
This work was supported by the China Postdoctoral Science Foundation
(2019TQ0242; 2019M660061XB).
Conflict of Interest
The authors declare no conflict of interest.
Keywords
healthcare, motion monitoring, nanowires, strain sensors, wearable
electronics
Received: March 2, 2020
Revised: April 8, 2020
Published online:
[1] W.Wu, H.Haick, Adv. Mater. 2018, 30, 1705024.
[2] H. Jin, Y. S. Abu-Raya, H. Haick, Adv. Healthcare Mater. 2017, 6,
1700024.
[3] X.Wang, Z.Liu, T.Zhang, Small 2017, 13, 1602790.
[4] M. Khatib, O. Zohar, W. Saliba, H. Haick, Adv. Mater. 2020, 32,
2000246.
[5] C. M. N.Brigante, N.Abbate, A.Basile, A. C.Faulisi, S.Sessa, IEEE
Trans. Ind. Electron. 2011, 58, 3234.
Small 2020, 2001363

2001363 (9 of 9)
www.advancedsciencenews.com
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
[6] L. A. Schwarz, A. Mkhitaryan, D. Mateus, N. Navab, Image Vision
Comput. 2012, 30, 217.
[7] J. C. Yeo, H. K. Yap, W. Xi, Z. Wang, C. H. Yeow, C. T. Lim, Adv.
Mater. Technol. 2016, 1, 1600018.
[8] H.Jeon, S. K.Hong, M. S.Kim, S. J.Cho, G.Lim, ACS Appl. Mater.
Interfaces 2017, 9, 41712.
[9] C. M.Boutry, Y. Kaizawa, B. C. Schroeder, A.Chortos, A. Legrand,
Z.Wang, J.Chang, P.Fox, Z.Bao, Nat. Electron. 2018, 1, 314.
[10] R. Nur, N. Matsuhisa, Z. Jiang, M. O. G. Nayeem, T. Yokota,
T.Someya, Nano Lett. 2018, 18, 5610.
[11] L.Lan, T.Yin, C.Jiang, X.Li, Y.Yao, Z.Wang, S.Qu, Z.Ye, J.Ping,
Y.Ying, Nano Energy 2019, 62, 319.
[12] M. Kozioł, B. Toroń, P. Szperlich, M. Jesionek, Composites, Part B
2019, 157, 58.
[13] M. Khatib, O. Zohar, W. Saliba, S. Srebnik, H. Haick, Adv. Funct.
Mater. 2020, 30, 1910196.
[14] E.Roh, B. U.Hwang, D.Kim, B. Y.Kim, N. E.Lee, ACS Nano 2015,
9, 6252.
[15] Q. Sun, W. Seung, B. J. Kim, S. Seo, S. W. Kim, J. H. Cho, Adv.
Mater. 2015, 27, 3411.
[16] M.Segev-Bar, H.Haick, ACS Nano 2013, 7, 8366.
[17] M. Khatib, T. P.Huynh, Y. Deng, Y. D. Horev, W. Saliba, W. Wu,
H.Haick, Small 2019, 15, 1803939.
[18] X. Wang, J. Li, H. Song, H. Huang, J. Gou, ACS Appl. Mater.
Interfaces 2018, 10, 7371.
[19] H.Liu, H.Gao, G.Hu, Composites, Part B 2019, 171, 138.
[20] M.Segev-Bar, N.Bachar, Y.Wolf, B.Ukrainsky, L.Sarraf, H.Haick,
Adv. Mater. Technol. 2017, 2, 1600206.
[21] M. Zahid, E. L. Papadopoulou, A. Athanassiou, I. S. Bayer, Mater.
Des. 2017, 135, 213.
[22] P. M.Losaria, J. H.Yim, J. Ind. Eng. Chem. 2019, 74, 108.
[23] G.Konvalina, A.Leshansky, H.Haick, Adv. Funct. Mater. 2015, 25, 2411.
[24] C.Casiraghi, M. Macucci, K. Parvez, R. Worsley, Y.Shin, F.Bronte,
C.Borri, M.Paggi, G.Fiori, Carbon 2018, 129, 462.
[25] P.Costa, M. F.Carvalho, V.Correia, J. C.Viana, S.Lanceros-Mendez,
ACS Appl. Nano Mater. 2018, 1, 3015.
[26] L. Ma, W.Yang, Y. Wang, H.Chen, Y. Xing, J. Wang, Compos. Sci.
Technol. 2018, 165, 190.
[27] T. P.Huynh, P.Sonar, H.Haick, Adv. Mater. 2017, 29, 1604973.
[28] J. H. Lee, J. Kim, D. Liu, F. Guo, X. Shen, Q. Zheng, S. Jeon,
J. K.Kim, Adv. Funct. Mater. 2019, 29, 190162.
[29] H.Hwang, Y.Kim, J. H.Park, U.Jeong, Adv. Funct. Mater. 2020, 30,
1908514.
[30] J. F. Fennell, Jr., S. F. Liu, J. M. Azzarelli, J. G. Weis, S. Rochat,
K. A.Mirica, J. B.Ravnsbaek, T. M.Swager, Angew. Chem., Int. Ed.
2016, 55, 1266.
[31] S.Gong, W. L.Cheng, Adv. Electron. Mater. 2017, 3, 1600314.
[32] S.Yao, P.Ren, R. Song, Y.Liu, Q.Huang, J.Dong, B. T.O’Connor,
Y.Zhu, Adv. Mater. 2020, 32, 1902343.
[33] K. Y.Suh, Y. S.Kim, H. H.Lee, Adv. Mater. 2001, 13, 1386.
[34] Y. Y. Broza, X. Zhou, M. Yuan, D. Qu, Y. Zheng, R. Vishinkin,
M.Khatib, W.Wu, H.Haick, Chem. Rev. 2019, 119,11761.
[35] D.Ho, J.Zou, B.Zdyrko, K. S.Iyer, I.Luzinov, Nanoscale 2015, 7, 401.
[36] T.Takano, H.Masunaga, A.Fujiwara, H.Okuzaki, T.Sasaki, Macro-
molecules 2012, 45, 3859.
[37] Q. Wei, M. Mukaida, Y. Naitoh, T. Ishida, Adv. Mater. 2013,
25, 2831.
[38] N.Tang, C.Zhou, L. Xu, Y.Jiang, H.Qu, X. Duan, ACS Sens. 2019,
4, 726.
[39] K. Takei, T. Takahashi, J. C. Ho, H. Ko, A. G. Gillies, P. W. Leu,
R. S.Fearing, A.Javey, Nat. Mater. 2010, 9, 821.
[40] Z. Han, L. Liu, J. Zhang, Q. Han, K. Wang, H. Song, Z. Wang,
Z.Jiao, S.Niu, L.Ren, Nanoscale 2018, 10, 15178.
[41] S. I.Park, J. H. Ahn, X.Feng, S.Wang, Y. Huang, J. A. Rogers, Adv.
Funct. Mater. 2008, 18, 2673.
[42] J.Zhao, G.Wang, R.Yang, X.Lu, M.Cheng, C.He, G.Xie, J.Meng,
D.Shi, G. J. A. n.Zhang, ACS Nano 2015, 9, 1622.
[43] S.Chen, Y. Song, D. Ding, Z. Ling, F. Xu, Adv. Funct. Mater. 2018,
28, 1802547.
[44] C.Farcau, H.Moreira, B.Viallet, J.Grisolia, D.Ciuculescu-Pradines,
C.Amiens, L.Ressier, J. Phys. Chem. C 2011, 115, 14494.
[45] M. Amjadi, A. Pichitpajongkit, S. Lee, S. Ryu, I. Park, ACS Nano
2014, 8, 5154.
[46] P.Sahatiya, S.Badhulika, RSC Adv. 2016, 6, 95836.
[47] S. X. Li, H. Xia, Y. S. Xu, C. Lv, G. Wang, Y. Z. Dai, H. B. Sun,
Nanoscale 2019, 11, 4925.
Small 2020, 2001363