Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

REVIEW

Pressure Sensors www.afm-journal.de

Emerging Technologies of Flexible Pressure Sensors:


Materials, Modeling, Devices, and Manufacturing
Yan Huang, Xiangyu Fan, Shih-Chi Chen,* and Ni Zhao*
Pressure sensors that convert pressure
As an important branch of wearable electronics, flexible pressure sensors input into electrical signals are an essential
have attracted extensive research owing to their wide range of applications, member in the flexible electronics family
such as human–machine interfaces and health monitoring. To fulfill the and have been developed into a multi­
requirements for different applications, new material design and device fabri- discipline area, combining material sciences,
device and system engineering, and signal
cation strategies have been developed in order to manipulate the mechanical
processing, including the newly emerged
and electrical properties and enhance device performance. In this paper, the artificial intelligence technologies.[9–11] There
important progresses in flexible pressure sensor development over recent have been a number of reviews lately to
years are selectively reviewed from a material and application perspective. summarize the progress of flexible pressure
First, an overview of the fundamental working mechanism and the system- sensors. Some focus on flexible electronics
as a whole[2,3,12–15] while others are dedicated
atic design approach is presented. Particularly, how the theoretical modeling
to the emerging materials (e.g., graphene-
has been used as an auxiliary tool to achieve better sensing performance is based) for pressure sensors.[9,11,16,17] To bring
discussed. A number of applications, including human–machine interfaces, the flexible pressure sensor technology from
electronic skin and health monitoring, and certain application-driven func- laboratories to the real world, we note that
tions, e.g., pressure distribution visualization and direction-sensitive force some critical elements, such as system
detection, are highlighted. Lastly, various advanced manufacturing methods design, theoretical modeling, and manufac-
turing, are missing in the previous reviews.
used for realizing large-scale fabrication are introduced.
Also, how the emerging additive manufac-
turing technologies can impact the field of
flexible electronics has not been discussed.
1. Introduction This review is devised to provide a holistic view of the devel-
opment of flexible pressure sensors combining both scientific
With the rapid development of modern society, new and disrup- and engineering progresses in this area. We first introduce
tive technologies such as the Internet of Things (IoT) and wear- fundamental assessment principles of pressure sensors. Fol-
able medical devices have emerged in recent years, improving lowing that, latest material developments of high-performance
the well-being and quality of life. Flexible electronics plays an pressure sensors are categorized according to their sensing
important role in the aforementioned applications in the form mechanisms, including piezoresistive, piezocapacitive, superca-
of electronic skin, human–machine interaction interface, physi- pacitive, piezoelectric, and triboelectric. As a powerful approach
ological signal monitoring, etc.[1–4] Compared with the conven- to supplement the understanding of the sensor performance,
tional silicon-based electronics, flexible electronics have great theoretical modeling is discussed and the common methodolo-
adaptability to different kinds of substrates, be it soft or rigid, gies are introduced. Prevailing applications of flexible pressure
flat or curved. This unique advantage has attracted extensive sensors, including health care, electronic skin, and human–
research efforts on flexible electronics, and various applications machine interface, are presented, along with two new auxil-
have been developed, such as light-emitting diodes (LEDs), bat- iary functions, namely pressure distribution visualization and
teries, antennas, and sensors.[5–8] detection of stimulus directions. Lastly, we review the emerging
advanced manufacturing technologies for large-scale and on-
demand fabrication of flexible pressure sensors. The structure
Y. Huang, Prof. N. Zhao of this review and the relations between each part are shown in
Department of Electronic Engineering Figure 1.
The Chinese University of Hong Kong
Hong Kong
E-mail: nzhao@ee.cuhk.edu.hk
X. Y. Fan, Prof. S.-C. Chen 2. Figures of Merit of Pressure Sensors
Department of Mechanical and Automation Engineering
The Chinese University of Hong Kong Pressure sensors can be evaluated through several criteria,
Hong Kong including sensitivity, working range, hysteresis, response and
E-mail: scchen@mae.cuhk.edu.hk
recovery time, and stability. The knowledge of basic parameters
The ORCID identification number(s) for the author(s) of this article is critical for the selection of pressure sensors as every
can be found under https://doi.org/10.1002/adfm.201808509.
application scenario has its unique requirement on sensor
DOI: 10.1002/adfm.201808509 characteristics.

Adv. Funct. Mater. 2019, 1808509 1808509  (1 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

2.1. Sensitivity and Working Range


Shih-Chi Chen received his
Sensitivity (S) is a value to measure the ability of sensors to con- B.S. degree in mechanical
vert pressure into electrical signals. The definition of sensitivity engineering from the National
is expressed as S = (ΔE)/(ΔP × E0), where ΔE is the variation of Tsing Hua University, Taiwan,
electrical signals (e.g., current, voltage, or capacitance), ΔP is the in 1999. He received his
pressure variation (i.e., working range), and E0 denotes the orig- S.M. and Ph.D. degrees in
inal electrical signal when no pressure is loaded. The transfer mechanical engineering from
curve is usually plotted with pressure input on the x-axis the Massachusetts Institute
and electrical signal output on the y-axis to describe the of Technology, in 2003
sensor performance. A linear relationship between the output and 2007, respectively. In
and the input pressure is desired for the convenience of data 2007–2009, he was a research
processing. Devices based on different sensing mechanisms fellow at the Wellman Center
require different types of materials and device structures to for Photomedicine, Harvard Medical School. Dr. Chen
achieve high sensitivity. This will be illustrated later. is presently an associate professor in the Department of
Mechanical and Automation Engineering at the Chinese
University of Hong Kong. His current research interests
2.2. Hysteresis and Response Time include biomedical optics, precision engineering, and
nanomanufacturing.
Hysteresis of a pressure sensor is the phenomenon in which
the sensor cannot restore its electrical characteristics when the Ni Zhao is currently an
load is removed and it is related to the viscoelasticity of the associate professor in the
active materials and the property of the electrode/active material Department of Electronic
contact. Hysteresis is often defined by the maximum difference Engineering at the Chinese
between the output signals obtained at the same pressure in University of Hong Kong
one loading-unloading cycle divided by the full-scale output.[18] (CUHK). She received her
The response time is defined as the time needed to achieve bachelor’s and master’s
90% of the stable output and reflects the speed in which a degrees in materials sci-
sensor can respond to step input. It is also related to the viscoe- ence and engineering from
lastic property of the active material and the contact conditions Tsinghua University and
between the material and the electrode. Certain surface struc- McMaster University, respec-
tures and porous bulk have been shown to reduce the negative tively, and her Ph.D. degree
effect of the viscoelastic properties, both the response time[19–21] in physics from the University of Cambridge. Her research
and hysteresis.[22–26] A systematic investigation on this phenom- interests include optoelectronic and bioelectronic devices
enon is still missing. based on organic and nanostructured materials, and char-
acterizations of the electronic processes in these devices.

2.3. Stability and Durability


sensing materials are three indispensable parts of a pressure
Maintaining a stable and reliable input-output relation under sensor. Polyimide (PI),[27] polydimethylsiloxane (PDMS),[28] and
long-time or multicycle use is another important requirement polyethylene terephthalate (PET)[18] are common materials for
for pressure sensors. Continuous loading-unloading testing flexible substrates. Patterned films made from gold,[27] silver,[29]
cycles are needed in order to examine the stability of a sensor. copper,[30] liquid metal,[31] and indium tin oxide (ITO)[32] are
There is no specific number of cycles set as a standard for used as electrode materials for their good electrical conductivity.
the stability testing, but commonly more than 5000 cycles are In the following subsections, we mainly review the sensing
adopted in most of the reports of flexible pressure sensors. The materials according to their sensing mechanisms.
current/capacitance/voltage output is recorded before and after
the loading cycles to track the variations of sensor response.
In terms of device lifetime, since the output degradation often 3.1. Piezoresistive Sensors
originates from the irreversible collapse of the active material,
some reports included supplementary characterizations of the Piezoresistive sensors transduce pressure input into a resist-
morphology of the active material to better investigate the deg- ance change of the device (Figure  2a). In this type of devices,
radation mechanisms.[19] the active material is either laid on a pair of parallel interdigital
electrodes or sandwiched between two vertically aligned elec-
trodes. The active material should provide both sufficient charge
3. Materials Development transport paths for electrical current flows and good elasticity
to be compatible with various mechanical deformation during
In this section, we report the recent material developments of operation. Composite materials consisting of an elastic matrix
pressure sensors. In general, flexible substrates, electrodes, and and conductive fillers are the most popular material choice to

Adv. Funct. Mater. 2019, 1808509 1808509  (2 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 1.  Illustration of the structure of this review. Applications of pressure sensors such as wearable health monitoring platform, prosthetic e-skins,
human–machine interfaces, etc., require new functionalities and better performances. To this end, new materials and various structural designs
have been proposed and realized with the aid of theoretical modeling and advanced manufacturing methods. Reproduced with permission.[146]
Copyright 2018, WILEY-VCH Verlag GmbH & Co. Reproduced with permission.[201] Copyright 2018, WILEY-VCH Verlag GmbH & Co. Reproduced with
permission.[144] Copyright 2018, WILEY-VCH Verlag GmbH & Co. Reproduced with permission.[125] Copyright 2014, Springer Nature. Reproduced under
the terms and conditions of the Creative Commons Attribution 4.0 International License.[19] Copyright 2018, The Authors, published by Springer Nature.
Reproduced with permission.[123] Copyright 2015, WILEY-VCH Verlag GmbH & Co. Reproduced with permission.[219] Copyright 2018, WILEY-VCH Verlag
GmbH & Co. Reproduced with permission.[187] Copyright 2018, WILEY-VCH Verlag GmbH & Co.

meet the requirements. For the elastic component, polymers the sensor. For instance, a piezoresistive response can originate
(e.g., PDMS,[33] polyacrylic ester (PEA),[34] and polyurethane from the contact resistance between the active material and the
(PU)[35]), hydrogels (e.g., polyacrylamide (PAAm)[36]), and electrodes[47,48] or the bulk resistance inside the active mate-
bioderived materials (e.g., fibroin[37] and cotton[38]) are com- rial.[49,50] Compared with bulk resistance change, the contact
monly selected. For the conductive component, carbon-based resistance change can often lead to a larger variation under the
fillers (e.g., carbon black,[39] carbon nanotube,[40] MXene,[41] same pressure as a small increase of the contact area between
graphene,[42] and carbonized biomass[43]), metal nanoproducts a poorly contacted electrode and active material can drastically
(e.g., Au/Ag nanowires and nanoparticles[44,45]), and conductive reduce the contact resistance.[51] Accordingly, high-sensitivity
polymers (e.g., polypyrrole (PPy)[18] and polyaniline (PANI)[46]) piezoresistive sensors usually contain an active layer with a
have been investigated and showed good device performance. rough surface, which is achieved either spontaneously during
In many applications, simply modifying the composi- material growth or through the use of a mold (e.g., etched
tion of the elastic matrix and conductive filler cannot achieve silicon,[52,53] abrasive paper,[54,55] and leaves[56–58]) to shape the
the desired sensor characteristic (e.g., high sensitivity and wide layer surface morphology. It has also been found recently that
linear range), and therefore, microstructuring of the active wrinkle-like surface morphology can be formed by exploiting
material is applied according to the transduction mechanism of the mechanical instability caused by polymer swelling,[59] heat

Adv. Funct. Mater. 2019, 1808509 1808509  (3 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 2.  Piezoresistive pressure sensors. a) Schematic illustration of the working mechanism of piezoresistive pressure sensors. b) Active materials
with four different morphologies placed in a vertical sensor structure. c) Corresponding transfer curves of the sensors in (b). Reproduced under the
terms and conditions of the Creative Commons Attribution 4.0 International License.[19] Copyright 2018, The Authors, published by Springer Nature.
d) Transfer curves of piezoresistive films with the hierarchical structure (HIE) and without the hierarchical structure (DOME). Insets: scanning electron
microscopy (SEM) images of HIE and DOME structure. Scale bar: 50 µm. e) The illustration showing the effect of hierarchical structure on increasing
the number of contact between the electrode and the material. The corresponding hierarchical surface is shown in the dashed box. Reproduced with
permission.[61] Copyright 2016, WILEY-VCH Verlag GmbH & Co. f) Fabrication process of hollow graphene/silicone piezoresistive pressure sensor.
g) Increase of sensitivity by increasing the graphene growth time from 3 to 30 min (from G3P7 to G30P7). h) Increase of linear range by increasing
the ratio of A:B component of PDMS from 1:20 to 1:5 (i.e., from G5P20 to G5P5). Reproduced with permission.[51] Copyright 2017, WILEY-VCH Verlag
GmbH & Co.

treatment,[47] self-assembly,[60] and mismatch between the mod- and a larger area variation.[19,61,62] Similar to the active mate-
ulus of the elastic matrix and conductive fillers.[21,34] Park et al. rials, the roughness of the electrode can be tailored to increase
compared the sensitivity of CNT/PDMS composite films with the sensitivity, and gold micropillar arrays[63] and rough copper
microdome, micropyramid, and micropillar surface structures, electrodes have been made to achieve this end.[18] Another
among which the microdome structure shows the highest sen- approach to increasing the contact variation is to include a
sitivity in all pressure ranges as a result of the largest contact porous layer (e.g., an electrospun film[64]) or an anodic alu-
area variation, as shown in Figure 2b,c.[19] It was found that minum oxide (AAO) membrane as an insulating spacer
creating wrinkles on the surface of the microdome structure to between the electrode and active material.[60] In this case, the
form hierarchical surface morphology can further increase the porosity and the thickness of the insulating spacer should be
sensitivity, as shown in Figure 2d,e.[61] A number of material carefully controlled to tune the contact area.[65]
systems have been used to demonstrate that higher surface pro- It is worthwhile to note that a solid active layer with only a
file increases the sensitivity as it leads to a smaller initial area rough surface usually exhibits sensitivity decay in the high

Adv. Funct. Mater. 2019, 1808509 1808509  (4 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

pressure region when the rough surface is flattened by the of three layers: paper substrate, transferred graphite electrode,
stress and deformation during the loading process.[47,52–56] and spin-coated PDMS dielectric that still maintains the rough
Accordingly, micro- or nanostructured matrix is introduced surface morphology of the graphite-coated paper (if the spin-
to address this problem. For example, polymer-based aero- coating speed is well controlled). It is argued that good pressure
gels containing freeze-dried conductive components was used response may result from 1) the increase of the effective dielec-
to construct the active layer, exhibiting highly reversible mor- tric constant following the pressure induced closure of air gaps
phology after large-deformation cycles.[33,43,66–68] In addition, in the sensor and 2) the interlock of the rough graphite surfaces
woven or nonwoven fiber networks, such as textile and paper, upon loading, which increases the overlapped electrode areas.
were utilized as the skeleton of the active materials through In some applications, a pressure sensor is integrated with other
carbonization or surface decoration with conductive compo- optical components and therefore the optical transparency of
nents.[48,69] Recently, Luo et al. reported a hollow graphene/sili- the device is needed. To achieve high transparency and sensi-
cone composite (Figure 2f) that allows for independent tuning tivity simultaneously, Kim et al. utilized homogeneously and
of sensitivity and linear range as shown in Figure 2g,h.[51] With heterogeneously dispersed SiO2 nanoparticles on a dielectric
the combination of a rough surface and compressible bulk surface to form a microstructure.[79] The haze, i.e., wide angle
structure, a wide linear range (>50 kPa) and high sensitivity scattering, is minimized by controlling the particle size based
(>10 kPa−1) can be achieved. on their aggregation and separation tendency, and the transpar-
ency (>95%) and sensitivity (up to 1.0 kPa−1) are also optimized,
as shown in Figure 3d–f, respectively. The nonmonotonic rela-
3.2. Piezocapacitive Sensors tion between the sensitivity (or transparency) and particle size
is due to the different dispersion states at different particle
Piezocapacitive sensors transduce pressure input to a change sizes.
in the capacitance of a parallel plate capacitor, which is given Introducing pressure-dependent-permittivity in the dielectric
by C = ε 0εdr A , where A is the overlapping area of the two plates or layer is another way to improve sensitivity. In 1980, Grannan
effective area of the capacitor, ε0 is the permittivity of vacuum, et al. reported that adding conductive fillers into a dielectric
εr is the relative permittivity of the dielectric material between material to near its percolation threshold can increase the rela-
the plates, and d is the separation between the plates. Similar to tive permittivity of the material dramatically.[80] Wang et al.
piezoresistive sensors, piezocapacitive sensors consist of three exploited this idea and fabricated a flexible pressure sensor with
main components, i.e., electrode, substrate and active mate- a AgNW-filled PU film as the dielectric interlayer,[35] achieving a
rial that should be a dielectric medium. Over the past ten sensitivity of 5.54 kPa−1 for pressure below 30 Pa and 0.88 kPa−1
years, a number of conductive materials and composites, such for pressure above 30 Pa.
as ITO,[20,32] carbon nanotubes,[31] Au,[70] AgNW,[35,71,72] Mg,[73] Besides capacitor structures, a field effect transistor (FET)
AgNPs,[24] and ionic liquid,[74] have been explored for their structure can also be used to provide piezocapacitive response
potential use as the electrode material of piezocapacitive sen- while offering an additional advantage of self-amplification. Air
sors. A variety of low modulus dielectric materials, such as dielectric layers and suspended gate electrodes were utilized by
PDMS,[20,75] Ecoflex,[71] AgNW-polyurethane composites,[35] Zang et al. to make an ultrasensitive pressure-sensitive FET,[81]
glycerol sebacate,[73] iontronic film,[32] poly-(methyl meth- which presents an unprecedented sensitivity of 192 kPa−1, a
acrylate),[72] graphene oxide,[76] and CNT/PDMS,[74] have been low limit-of-detection pressure of <0.5 Pa, and a short response
studied as the dielectric medium. time of 10 ms. It is however worth noting that the adoption of
Deformation of dielectrics caused by external pressure leads air as the dielectric medium may result in unstable sensing
to a change in capacitance (Figure  3a). From this perspective, performance, which is highly susceptible to various interfer-
improving the compressibility of the dielectric layer directly ences around the sensor. Schwartz et al. integrated micro-
increases sensitivity. To this end, various dielectric layers structured PDMS dielectric in thin film polymer transistors,
have been designed and structured for lower elastic modulus, achieving a sensitivity of 8.4 kPa−1 by operating the transistor in
such as micropyramid PDMS,[20,70,77] porous PDMS with air the subthreshold regime.[75] Integrating pressure-sensitive com-
gap,[24,78] porous foam,[23,76] rough surface,[25] and nanoparticles- ponent into a transistor structure may also be a good solution
dispersed polymer.[79] In air-embedded structures, the dielectric to achieve sensor array configurations by virtue of its compat-
constant change may contribute significantly to sensitivity as ibility with low-cost large-area fabrication processes.
shown in Figure 3b, i.e., when the structured films are com-
pressed, the air volume is reduced, resulting in an increase
in the effective dielectric constant since air has a lower dielec- 3.3. Supercapacitive Iontronic Pressure Sensors
tric constant than elastomers.[20] In addition, the viscoelastic
behavior of an elastomer will be reduced after an air-embedded Supercapacitive iontronic pressure sensors also transduce a
structure is created in its bulk, resulting in a shorter response pressure input into a capacitance change; however, this type of
and relaxation time.[20] pressure sensors utilizes the formation of an electron double
Various methods have been employed to achieve a structured layer (EDL) at the dielectric/electrode contact to enhance the
dielectric layer. Lee et al. used a paper as a mold to create sur- piezocapacitive effect. In brief, an ionic fluid or gel containing
face roughness in a “half-cell” and then fabricated a capacitive numerous cations and anions is sandwiched between the two
pressure sensor by placing two identical half-cell capacitors electrodes. When the voltage is on, cations and anions are
face-to-face.[25] As shown in Figure 3c, each half-cell consists respectively attracted to the cathode and anode to form two

Adv. Funct. Mater. 2019, 1808509 1808509  (5 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 3.  Capacitive pressure sensors. a) Schematic illustration of the working principle of a capacitive pressure sensor. b) Schematic illustration of
the deformation process of capacitive pressure sensors with and without the air embedded dielectric layer under loading process. Reproduced with
permission.[23] Copyright 2016, ACS Publications. c) Schematic illustration of the half-cell capacitor. Reproduced with permission.[25] Copyright 2017,
WILEY-VCH Verlag GmbH & Co. d) Aggregation and separation tendency of the colloidal particles according to the particle size (top). Tilted-view
SEM images of silica nanospheres dispersed onto a PEDOT: PSS layer (bottom). e) Measured transparency of the films in d). For the 500 nm silica
nanospheres, the transparency was 97% at the wavelength of 550 nm. f) Transfer curves of the transparent capacitive sensors made by films in d). The
capacitive pressure sensor with the SiO2 particles of 500 nm shows the highest sensitivity of 1.0 kPa−1 below 2 kPa. Reproduced with permission.[79]
Copyright 2018, WILEY-VCH Verlag GmbH & Co.

EDLs. The working mechanism of this type of sensors relies on conductivity, high boiling point, and low viscosity.[83] It can work
the contact area variation between the electrodes and the active independently in microfluidic pressure sensors,[83] or in coor-
materials, as shown in Figure  4a. When pressure is applied, dination with the polymer to form ionic gels.[32] Similar to the
the contact area is increased, and more cations or anions are requirements of the electrolytes of ion batteries, the polymer
attracted to the contact surface, thus increasing the capaci- should have low crystallinity and a high dielectric constant.
tance. Owing to the EDL, the supercapacitive iontronic pressure The amorphous domains allow the anions and cations to move
sensors show a higher specific capacitance than conventional quickly through the free volumes to respond to the pressure.
capacitive sensors and are argued to be more suitable for signal Poly (vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP)
acquisition and analysis.[82] has been shown to be a good candidate.[84]
Ionic liquids, such as 1-ethyl-3-methylimidazolium As a type of microfluidic sensors, the sensing performance
bis(trifluoromethylsulfonyl)imide ([EMIM][TFSI]), play a of the supercapacitive iontronic pressure sensors heavily relies
significant role in this type of sensors due to its high ionic on the geometric cavity parameters, and the miniaturization

Adv. Funct. Mater. 2019, 1808509 1808509  (6 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 4.  Supercapacitive iontronic pressure sensors. a) Schematic illustration of the working mechanism of supercapacitive iontronic pressure sen-
sors. b) Cross-sectional view of the epidermal–iontronic interface with human skin as one electrode. c) Performance of the devices in b). Reproduced
with permission.[86] Copyright 2018, WILEY-VCH Verlag GmbH & Co. d) Device structure and e) transfer curves of all-fabric supercapacitive iontronic
pressure sensors. Reproduced with permission.[89] Copyright 2017, WILEY-VCH Verlag GmbH & Co. f) Device structure containing biomimetic dielectric
layer (SEM image in the dashed box) templated from a calathea zebrine leaf. g) Transfer curve of the sensor in (f). Reproduced with permission.[88]
Copyright 2018, WILEY-VCH Verlag GmbH & Co.

of the lateral pixels leads to a high threshold pressure and a have been studied for pressure sensing. Among these materials,
decreased sensitivity.[83,85] Therefore, much effort has been PVDF and its copolymers, such as P(VDF-TrFE), are promising
made to develop solid ionic gels that generate a large varia- candidates for wearable pressure sensors due to their flexibility,
tion of contact area, as is the case with piezoresistive sensors. durability, light weight, ease of fabrication, chemical inert-
Air gaps were introduced to the sensor structures between the ness etc.[96] The problem of using PVDF for pressure sensing
electrode and the ionic gel,[32] and interestingly, human skin lies in the coupled piezoelectric and pyroelectric effects. For
could act as the electrode in the epidermal devices as shown in this reason, protection from thermal interference is required
Figure 4b,c.[86] Ionic gels with the fabric structure (Figure 4d,e), in these sensors.[98] PVDF thin films need to be electrically
micropyramid[87] and microdome structures (Figure 4f,g),[88] poled to create maximum polarization in a direction orthog-
were investigated for the purpose of sensitivity improvement.[89] onal to the film plane to achieve high sensitivity. Using elec-
More research needs to be conducted in order to fully under- trospinning eliminates this need due to its high-voltage nature
stand and enhance the performance of such sensors. (Figure 5b).[96] PZT is one of the most commonly used piezo-
electric materials owing to its large piezoelectric and electrome-
chanical coupling coefficients. Intrinsically rigid as it is, PZT
3.4. Piezoelectric and Triboelectric Pressure Sensors can be integrated into flexible devices in the form of ultrathin
sheet and other nanostructured forms. Compared with sys-
Piezoelectricity was first discovered by the Curie brothers tems incorporating organic active materials, PZT-based sensors
in 1880, and it refers to the phenomenon where positive and provide high sensitivity, fast response (sub-millisecond), low
negative charges are spatially separated in certain materials hysteresis, and superior stability.[99]
in response to an applied mechanical stress, or vice versa Hybridizing piezoelectric materials with other functional
(Figure  5a). Piezoelectric sensors are suitable for dynamic materials may improve the device performance and generate
measurements of pressure due to their variation-dependent novel properties. Han et al. developed a pressure visualiza-
sensing mechanism. They are not applicable to static sensing tion and recording (PVR) system as shown in Figure 5c, which
because the sensor generated output voltage is impulsive. can directly detect, visualize and record applied pressure by
Many materials have been found to exhibit piezoelectricity. combining ZnO-nanowire-matrix pressure sensors (ZPSs)
They can be natural or synthetic, organic or inorganic. Alu- with WO3-film electrochromic device (ECD) array.[100] In the
minum nitride (AlN),[90] ZnO,[91] PbTiO3,[92] BaTiO3,[93] lead zir- PVR system, each pixel of the ECD array connects to a group
conate titanate (PZT),[94] polyvinylidenefluoride (PVDF),[95] and of ZnO nanowires via a linking electrode. Pressure applied to
poly(vinylidenefluoride-co-trifluoroethylene) (P(VDF-TrFE))[96,97] the ZnO nanowires induces piezoelectric charges and these

Adv. Funct. Mater. 2019, 1808509 1808509  (7 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 5.  Piezoelectric pressure sensors. a) Schematic illustration of the working principle of piezoelectric pressure sensors. b) Schematic illustration of
electrospinning. Reproduced with permission.[97] Copyright 2016, ACS Publications. c) Schematic illustration of the structure of the PVR system. Reproduced
with permission.[100] Copyright 2017, WILEY-VCH Verlag GmbH & Co. d) Pressure responses of a pure PbTiO3-nanowire-based sensor, a graphene-based
sensor and a PbTiO3 nanowires/graphene sensor under a pressure pulse. Reproduced with permission.[92] Copyright 2017, ACS Publications.

charges change the colors of the corresponding pixels in the sensors, this sensor achieves a static sensing sensitivity of
ECD array due to the electrochromic property of WO3. Hence, 9.4 × 10−3 kPa−1 as a result of the heterostructures. Compared
the distribution of external pressure can be directly visualized with flexible piezoresistive sensors of which the response time
by color changes. Notably, an obvious color change in ECD is on the scale of tens of milliseconds, this sensor shows much
requires large external pressure applied to the matrix, which faster response (the rise and fall time are about 5–7 ms).[92]
may be attributed to twofold factors, i.e., the low piezoelectric Mechanical to electric energy conversion can also occur in
the form of triboelectric effect, where certain materials become
efficiency  charge density  of ZPS and low coloration efficiency
 pressure  electrically charged after they come into frictional contact with
 optical density  a different material.[101] Triboelectric nanogenerators (TENGs)
of ECD. Another issue is the so-called “color
 charge density  have been intensively studied for low-power or self-powered
memory effect” of WO3 film. Although claimed as a favorable pressure sensors[102–107] since the first prototype was reported
property, due to the fact that the color change can maintain for in 2012.[102] The output of these sensors is affected by both the
hours and record the pressure distribution, the color memory magnitude and frequency of external pressure and thus they
effect is not desirable for applications in real-time visualization are suitable for dynamic sensing, e.g., vibration detection.
of pressure distribution. Chen et al. reported a pressure sensor Based on TENG and mechanoluminescent materials, a full
for static measurements based on PbTiO3 nanowires/graphene dynamic-range pressure sensor matrix with high sensitivity and
heterostructures.[92] When pressure is applied, the induced resolution was fabricated.[105] The device was equipped with
polarons in piezoelectric PbTiO3 nanowires work as charged two integrated sensing components, i.e., a triboelectric sensor
impurities, affecting the carrier mobility of graphene. As shown matrix and mechanoluminescent sensor matrix. The sensor
in Figure 5d, compared with conventional piezoelectric pressure was able to visualize pressure distribution in low (<100 kPa)

Adv. Funct. Mater. 2019, 1808509 1808509  (8 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

and high-pressure regimes (>1 MPa). A soft skin-like triboelec- of dielectric layers is involved, nonlinearity of the materials
tric nanogenerator (STENG) that can be used for tactile sensing should be considered in the model to provide precise predic-
was fabricated by adopting elastomers as the electrification tion of performance; however, this comes at the expense of
layer and ionic hydrogels as the electrode, achieving ultrahigh computational efficiency.[114] As for piezoresistive materials,
stretchability (1160%) and transparency (average transmittance this similar effect of porous structure was simulated and exper-
of 96.2% for visible light) simultaneously.[106] The advantage of imentally validated by Jung et al.[115] Interlocked microdome
the STENG lies in its capability to maintain the performance design, which increases the sensitivity and shortens response
until mechanical failure when compared with percolated net- time, was studied with FEA and a theoretical model by Park
works of carbonaceous materials or metal nanowires, of which et al.,[52] showing that the superior sensor performance benefits
performances degrade gradually and inevitably in the stretching from the giant tunneling piezoresistance present in this design.
process. The drawback is the weak interfacial bonding between Interestingly, Pang et al. compared various surface shape
the electrification layer and electrodes as hydrogels are hydro- design, including pyramid, hemisphere (dome), nanowire, and
philic and elastomers are hydrophobic. random distribution spinosum (RDS), and claimed that the
RDS pressure sensor exhibits the highest sensitivity and the
broadest linear range by simulating their resistance variations
4. Theoretical Modeling under pressure with a model shown in Figure 6e,f.[54] The four
structures were defined by the corresponding height distribu-
A model is a simplification of the real system. An intuitive tion matrix and each element in the contact area was equivalent
understanding of the experimental observation as well as para- to a resistor with a constant value of resistance, which con-
metric relationship among the sensor design parameters can be tributed to the total contact resistance, assuming all resistors
established through a properly simplified model which requires were connected in parallel. By observing the contact points in
lower computational cost, while producing reliable predictions the simulated loading process, it was learnt that the sharp mor-
that are consistent with experimental data. Modeling together phology which enhances the compressibility contributes to the
with computational analysis have been applied intensively sensor’s high sensitivity, while the random distribution contrib-
in pressure sensor studies[40,57,103,108] and help researchers to utes to the sensor’s wide linear range.
explain, illustrate and predict the electrical and mechanical Analytical and computational analyses have also been
properties of sensors under external pressure. Modeling also applied in the research that aims to achieve certain sensor char-
assists the design of complex geometry and optimization of acteristics other than high sensitivity or good linearity, such as
design parameters at both material and device levels. For bending insensitivity, wide response bandwidth, etc.
example, Cantarella et al. reported a mesa-shaped elastomeric Lee et al. utilized simulations to illustrate the structural
substrate that supports electronic devices. With the aid of a advantage of a new material design, i.e., a nanoporous structure
numerical parametric study based on finite element analysis made from electrospun composite nanofibers for a bending-
(FEA), they optimized the substrate structure and made it pos- insensitive pressure sensor.[116] The FEA simulation was con-
sible to localize the strains around the pillars and shield the ducted in ABAQUS, where the randomly stacked electrospun
electronics, as shown in Figure  6a.[109] In another exemplary nanofibers were simplified as a square lattice of ligaments
case, through numerical simulations and analytical models, Jin and the fibers were assumed to be well bonded. When the
et al. showed that the electrical hysteresis behavior of CNT thin nanofibers were bent to a radius to thickness ratio of two, the
film is governed by the ratio of the mean projected CNT length fibers were deflected to accommodate the large deformation.
over the film length, which was later verified in experimental As a result, the maximum strain was ≈7%, while as a compar-
studies.[110] ison, a continuous sheet would have a maximum lateral strain
Computational analysis has been widely used to illustrate of 25% according to the classical beam theory. Although the
the impact of material microstructures on device sensitivity. model is simplified, it is reasonable to emulate the nanofibers
Kang et al. reported a capacitive pressure sensor based on a as a lattice of ligaments as it captures the key structural char-
sponge-like porous structured dielectric layer.[111] To analyze the acteristics of the electrospun nanofibers. On the other hand,
effect of pore size on the compressibility of porous layers, the the simulation results also suggest that the fibers were almost
unit support was simplified as an isotropic elastic half pillar, not stretched under bending. By just combining the simula-
from which the Sneddon compliance as an indicator of com- tion results one cannot draw a direct relation between the
pressibility and sensitivity can be theoretically solved. The adopted material structure and the bending insensitivity (i.e.,
microporous structure has been applied widely in dielectric the relation between the mechanical deformation of the active
layers to design highly sensitive capacitive pressure sensors as material and the electrical response of the sensor needs to
mentioned in the previous sections. Its critical effect on device be clarified); furthermore, the influence of the material-
compressibility and sensitivity was investigated through FEA electrode interfaces and fiber-to-fiber interfaces needs to be
by Kwon et al. (Figure 6b–d).[23] Similar strategies for achieving considered. For instance, accordingly to another study,[69]
high sensitivity, e.g., micromyramid and microcones, have the resistance variation of a fiber-based pressure sensor is
also been studied using FEA by Tee et al.,[112] and Qiu et al.,[88] dominated by the contact resistance of the material-electrode
respectively. Ho et al. reported a crack-enhanced microfluidic interfaces and fiber-to-fiber interfaces.
capacitive sensor and analyzed how the crack deformation and Yang et al. reported a self-powered bionic membrane sensor
liquid wettability influence the performance by simulating the (BMS) for health-monitoring.[103] Inspired by the human ear-
movement of air-liquid interface.[113] When large deformation drum, the oval shape of the multilayered-structure sensor

Adv. Funct. Mater. 2019, 1808509 1808509  (9 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 6.  a) Simulation of bending and twisting to analyze the strain distribution of mesa-shaped elastomeric substrate. Reproduced with permis-
sion.[109] Copyright 2018, WILEY-VCH Verlag GmbH & Co. b) Full 3D representative model of the porous Ecoflex dielectric cube and its stress distribu-
tion under compression. c) Compressive strain−pressure curves of the solid Ecoflex cube (left), the porous Ecoflex cube (right) by experiment and
simulation. d) Transfer curves by experiment and simulation for parallel-plate capacitor models with solid and microporous Ecoflex dielectric layers.
Reproduced with permission.[23] Copyright 2016, ACS Publications. e) Simulated pressure distributions for different surface microstructures: pyramid,
hemisphere, nanowire, and random distribution spinosum microstructure under 5 kPa. f) Transfer curves of sensors made by the active materials with
different surface microstructures in (e). Reproduced with permission.[54] Copyright 2018, ACS Publications.

Adv. Funct. Mater. 2019, 1808509 1808509  (10 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

enables the BMS to rapidly measure pressure variation over morphologies, which was validated by surface profilometer as
an exceptionally wide frequency, i.e., from 100 to 3200 Hz. shown in Figure 7e.
The BMS is therefore able to acquire and recover the human
signals of both low-frequency, e.g., arterial pulse wave, and
high-frequency, e.g., throat sound. The working principle was 5. Application Scenarios and Derived Functions
elucidated with the aid of FEA, where the vibration patterns of
the oval polytetrafluoroethylene (PTFE) membrane in response There are a great variety of applications of pressure sensors.
to external pressure from 100 to 5000 Hz was investigated. As We divide them into three categories: 1) human–machine inter-
demonstrated in Figure  7a, the deformation distribution and face, 2) electronic skin, and 3) health monitoring. Driven by the
magnitudes of the tympanic PTFE membrane highly depend applications, the need for a number of specific functions, e.g.,
on the excitation frequencies. The frequency response of the pressure distribution visualization and classification, arises for
deformation amplitudes was studied via a series of simulated better user experience or system reliability and accuracy. The
experiments, showing a bandwidth of up to 3200 Hz. Figure 7b sensor development for the three main applications together
presents the FEA modeling of the electricity generation process with other important application-driven functions are reviewed
induced by the membrane, which clearly shows the electric in this section.
potential distribution at different stages of vibration. When the
PTFE membrane contacts the nylon layer under external pres-
sure, triboelectric charges are generated on both sides of the 5.1. Human–Machine Interface
membrane due to their different affinity for electrons. In this
state, there is no potential difference between the two layers Human–machine interface refer to devices and software
since the generated positive and negative charges balance each installed on machines that assist the human–machine com-
other due to electrostatic induction. Once the two layers are munication. Flexible pressure sensors play an important role
separated, a potential drop occurs between the two layers and in translating mechanical input from humans to electrical sig-
between the ITO electrode and ground, driving the electrons nals for controlling the machine, or for providing feedback. For
to flow from the ground to the ITO electrode. The electricity example, a smart glove with integrated pressure sensors was
generation process is simulated based on a simple parallel-plate able to control the movement of the robot arm to play music,[117]
model, showing the potential distribution when the planar and epidermal acoustic sensors and a speech recognition algo-
PTFE and nylon layers are closed and separated; note that in rithm could translate the commands from the human to con-
the model the deformation of the layers is not considered. trol a virtual character in computer games.[118] Flexible touch
Although various vibration patterns, i.e., deformation areas and panels have been developed with transparent pressure sensor
magnitudes, of the PTFE membrane in response to different arrays to fulfill the requirements of flexible displays.[119–122] Kim
frequencies were simulated, their effect on the electrical out- et al. reported a surface-capacitive touch system as a stretchable
puts, e.g., short-circuit current and open-circuit voltage, was not ionic touch panel, using touching fingers as a ground electrode
studied. Ideally, the bandwidth should be directly obtained from to determine the contact position.[120] A wearable touch panel
the electric response to input frequencies, but the frequency was demonstrated to write words, play music and chess, as
response presented in the work is the amplitude of deforma- shown in Figure  8a–c. In addition, pressure sensors can also
tion instead of the electrical outputs. It is worth mentioning be used in security applications, such as signature collectors for
that since a PTFE membrane normally deforms in a nonuni- the detection of signing characteristics.[123]
form way, the electrical outputs and deformation amplitude
may have a nonlinear relation. Therefore, the simulated sensor
bandwidth may not be entirely accurate. 5.2. Electronic Skin
Huang et al. demonstrated the design of a sensor fabrica-
tion process with the guidance of FEA,[33] which predicts the Electronic skin refers to a network of flexible electronics with
influence of mechanical cutting processes on the structure and various sensing functionalities that emulate the functions
morphology of an active material. In this work the sensor was of human skin, but not limited to the functions of human
made from wood and the effect of the multichannel structures skin.[124,125] For practical applications, it can help the disabled
in wood to enhance its performance was analyzed via FEA. people to restore their sensing ability lost in accidents, e.g.,
The mater wood was processed both vertically and horizon- prosthetics,[126,127] or build the biomimetic cognition of sur-
tally in reference to the natural wood channels. Experiments rounding stimulus on the surface of artificial robotics.[128,129]
found the horizontally cut materials have 10 times wider linear Pressure sensors are an indispensable part of electronic skin;
range than the vertically cut materials; and this is attributed and typically require large-area coverage with small sensing
to the new morphologies introduced by the cutting processes. pixels. Various prototypes of pressure sensor arrays installed
The simulation revealed the formation mechanism of the two on hand, tongue, or robotics have been developed in recent
types of morphology produced by two different cuts. During a years.[130–132]
horizontally cut, wood channels are peeled off because cracks The combination of different functional electronic com-
caused by the stress concentration at the channel wall intercon- ponents is often required to realize electronic skin. Com-
nection propagate along the channel walls, while the vertically monly used functional components include sensors,[28] flexible
cut channels do not have this effect (Figure 7c,d). Accordingly, energy supply,[133–135] and transmitting and processing com-
the different cut directions introduce different types of surface ponents (Figure 8d).[136] The last one is essential for stimulus

Adv. Funct. Mater. 2019, 1808509 1808509  (11 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Adv. Funct. Mater. 2019, 1808509 1808509  (12 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 8.  Application demonstrations of electronic skin and human–machine interface. A highly stretchable ionic touch panel is capable of a) writing words,
b) playing music, and c) playing chess. Reproduced with permission.[120] Copyright 2016, the American Associations for the Advancement of Science.
d) Image (left) and circuit schematic illustration (right) of a model hand with tactile electronic skin. Reproduced with permission.[136] Copyright 2015, the
American Associations for the Advancement of Science. e) Discoid cockroach with an artificial afferent nerve on its back. f) Illustration of the working
mechanism of the system in (e). Reproduced with permission.[138] Copyright 2018, the American Associations for the Advancement of Science.

transportation, especially at the device/neuron interface.[137] pathological information. Luo et al. reported that epidermal
Kim et al. reported an artificial afferent nerve consisting of pie- pulses measured by a flexible piezoresistive sensor can be
zoresistive pressure sensors, ring oscillators, and synaptic tran- used for beat-to-beat blood pressure (BP) tracking through the
sistors to mimic the function of biological mechanoreceptors, pulse transit time (PTT) method with the aid of electrocardi-
nerve fibers and biological synapses. The sensor not only cor- ogram (ECG) electrodes, as shown in Figure 9c,d.[69] Chu et
rectly identified braille characters, but also actuated real muscle al. experimentally verified that the pulse sensing system can
tissues when connected to the nerves of a discoid cockroach, as diagnose arrhythmia and detect BP through the oscillometric
shown in Figure 8e,f.[138] method.[144]
Several issues exist when flexible pressure sensors are used
in cardiovascular monitoring. For example, when measuring
5.3. Health Monitoring epidermal pulses on a wrist, recording quality signals requires
the sensor to be precisely located above the radial artery. Other-
Flexible pressure sensors have been broadly applied in health- wise, the measured signals can be distorted, preventing useful
care services to enable continuous, ubiquitous and unobtrusive physiological information from extraction.[75,145] Although the
tactile and pressure sensing. They can be used in voice recogni- alignment problem may be addressed by using densely packed
tion,[103] heart rate and respiration rate monitoring,[33,139] head pressure sensor arrays, the required power will be multiplied
position adjustment (Figure  9a,b),[89] plantar pressure map- by the number of sensors. To address this problem, Fan et al.
ping,[140] wound monitoring,[141] etc. proposed a wearable liquid capsule sensor platform embedded
Various piezoresistive and capacitive pressure sensors with a piezoresistive pressure sensor.[146] In this sensor design
have demonstrated the capability of precise pulse measure- the pulsation signals can be precisely collected and transmitted
ment,[26,70,75,142,143] which contains abundant physiological and through flexible elastomeric membranes of the capsule to the

Figure 7.  a) Demonstration of the working principle of the bionic membrane sensor. The ANSYS software was employed to characterize the PTFE
membrane vibrations under various applied pressure at frequencies of 100, 500, 1000, and 5000 Hz, respectively. b) Simulated vibration induced
electricity generation process of a PTFE membrane via COMSOL. Reproduced with permission.[103] Copyright 2015, WILEY-VCH Verlag GmbH & Co.
c) Representative models to simulate the cutting trajectory to obtain the horizontal sample (right) and vertical sample (left). d) Simulated results of the
cutting process of the horizontal and vertical samples. e) Step profiler results demonstrating the relative surface roughness of horizontal and vertical
samples. Reproduced with permission.[33] Copyright 2018, WILEY-VCH Verlag GmbH & Co.

Adv. Funct. Mater. 2019, 1808509 1808509  (13 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 9.  a) Photograph of the supercapacitive all-fabric facemask with 20 pressure sensing units. b) The facemask in (a) worn on a face that was
pressing a foam facial pillow. With such a mask, patients would be prevented from being exposed to excess face pressure by adjusting the head position
on time in prone-positioned surgeries. Reproduced with permission.[89] Copyright 2017, WILEY-VCH Verlag GmbH & Co. c) Photograph of a flexible
piezoresistive sensor (FPS) attached to a wrist. The inset shows the active area of the sensor. d) Flexible sensor system containing FPS and epidermal
electrocardiogram (ECG) sensors for cuffless blood pressure (BP) measurement. Reproduced with permission.[69] Copyright 2016, WILEY-VCH Verlag
GmbH & Co. e) Materials and overall assembly of the fully biodegradable strain and pressure sensor. f) The sensor in (e) was subcutaneously implanted
on the back of a rat. g) Pressure signal recorded after 2 weeks (top left) and 3.5 weeks (top right) after sensor implantation. The corresponding baseline
shows recorded respiration of the rat after 2 weeks (bottom left) and 3.5 weeks (bottom right) in vivo. Reproduced with permission.[149] Copyright 2018,
Springer Nature.

liquid-embedded pressure sensor, which substantially relaxes 5.4. Application-Driven New Functions
the typical stringent alignment requirements between the
sensor and the artery to 8.5 mm. Another important problem 5.4.1. Pressure Distribution Visualization
is motion artifact, which compromises the accuracy of pres-
sure measurement. Although algorithms based on data pro- Signal acquisition equipment and supporting software are
cessing to eliminate motion artifacts and recover signals have typically needed to translate the electrical signals into mean-
been proposed,[147] little progress on motion-tolerant pressure ingful physical values for sensor applications. However, in cer-
sensors has been reported. For certain healthcare applications, tain applications, e.g., human–machine interface, immediate
e.g., intraocular pressure measurements, sensors need to be visualization of the pressure distribution is required or highly
implanted inside the body[148] and continuously operate over a desired, and additional sensor capabilities are designed to meet
long period of time; accordingly, the implanted devices need to the needs. Three types of working mechanisms for 2D pressure
have good biocompatibility and biodegradability to avoid irri- visualization have been proposed; they are described in the fol-
tability and to bypass the need for an extraction surgery when lowing subsections:
they are no longer needed. Boutry et al. reported a stretchable Integration with Electro-/Thermochromic Devices: Electro-
and biodegradable strain and pressure sensor for orthopedic chromic materials can modulate their optical properties through
application, as shown in Figure 9e.[149] Figure 9f demonstrates reversible redox reactions by direct current potential[150] while
that the sensor can be implanted subcutaneously on the back pressure sensors can alter the electrical signals in the circuits.
of a rat. Pressure stimuli was applied in vivo and the signals Therefore, the combination of both properties can be exploited
are shown in Figure 9g, where a zoom-in view on the baseline to control the colors of pressure sensors according to the input
signals shows the rat respiration recording has good signal pressure. Chou et al. fabricated a new device by integrating
quality. an electrochromic device and a piezoresistive sensor with a

Adv. Funct. Mater. 2019, 1808509 1808509  (14 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 10.  a) Laminated structure illustration of the pressure-chromic device and the related circuit. Reproduced under the terms and conditions of the
Creative Commons Attribution 4.0 International License.[151] Copyright 2015, The Authors, Springer Nature. b) Schematic illustration of the pressure
induced thermochromic device, and c) series of optical images (top) and temperature distribution (bottom) images during loading and unloading
on the device. Reproduced with permission.[154] Copyright 2017, WILEY-VCH Verlag GmbH & Co. d) Schematic illustration of the working principle of
LED-based pressure sensor, and the demonstration of trace visualization in (e). Reproduced with permission.[155] Copyright 2018, ACS Publications.
f) Schematic illustration of the device structure of the LED-based pressure sensor array for pressure distribution visualization, and (g) imaging of the
“BINN” pattern. Reproduced with permission.[161] Copyright 2015, WILEY-VCH Verlag GmbH & Co. h) Illustration of the device structure of the elec-
troluminescent sensor, and the images of the sensor conforming to the end of a pencil. Reproduced with permission.[162] Copyright 2016, the American
Associations for the Advancement of Science. i) Schematic structure of the mechanoluminescence device using ZnS:Mn particles. j) Record of the
signing habits of four persons using the sensors in (i). Reproduced with permission.[123] Copyright 2015, WILEY-VCH Verlag GmbH & Co.

pyramid structure, as shown in Figure  10a.[151] During an arrays.[100] As a piezoelectric material, ZnO can generate polar-
operation, the loading of pressure decreases the sensor resist- ized positive and negative charges under pressure loading at
ance and simultaneously increases the voltage on the electro- the opposite sides of the material,[152,153] tuning the local band
chromic device; accordingly the output color changes from dark structures. Therefore, the electric current through the elec-
red to pale blue. Based on the piezo-phototronic theory, Han trochromic layer (i.e., WO3) can be adjusted, leading to a color
et al. achieved fast mapping of pressure distribution through an change from transparent to blue. Similar to the case of electro-
interactive matrix (10 × 10 pixels) by combining piezoelectric chromic materials, thermochromic materials allow for modula-
zinc oxide (ZnO) and electrochromic tungsten oxide (WO3) tion of colors by temperature variation. Kim et al. integrated a

Adv. Funct. Mater. 2019, 1808509 1808509  (15 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

heater, a sensor, and a thermochromic device to realize pres- capacitors as shown in Figure  11a.[166,167] The four capacitors
sure distribution mapping, as shown in Figure 10b,c.[154] In share the same top electrode, with four separate bottom elec-
this system, the pressure sensors incorporating a carbon black/ trodes; a vertical pressure load decreases the thickness of the
PDMS composite generate resistance change and heating effect dielectric layer, and an in-plane shear force varies the overlap-
at the same time upon pressure loading, and the resulting tem- ping area between the top and bottom electrodes, and accord-
perature rise is sufficient to induce a color change of the ther- ingly the output capacitances. These together allow three-axis
mochromic device. pressure sensing.[167] The direction-sensitive force detection has
Integration with LED: Organic LED devices are suitable for also been realized in piezoresistive pressure sensors through a
integration with flexible pressure sensors to develop interac- unique four-sidewall electrode design.[168] The combination of
tive electronic skins because of their mechanical flexibility different sensing mechanisms can be used to distinguish force
and current/voltage-driven light emitting properties. Pres- sources. The integration of stretch-sensitive CNT films as the
sure sensitive layers can be laminated between the electrode electrodes of a capacitive pressure sensor can help to distin-
and functional materials to provide current/voltage control as guish pressure, bending, and stretching inputs by measuring
shown in Figure 10d, and real-time display of pressure traces the resistances of the two electrodes and system capacitance
can be realized, as shown in Figure 10e.[155,156] Wang et al. fabri- simultaneously, i.e., the dielectric layer would show capacitance
cated an active matrix array and integrated a pressure sensitive variations under all three different stimuli while the two elec-
rubber above the cathode of a bilayer OLED, where the anode is trodes would show similar resistance responses to stretching
connected to the drain of a transistor.[157] When the pressure but different responses to bending.[78] The combination of pie-
increases, the brightness of the LED increases accordingly. zoresistive and triboelectric devices provides another opportu-
Traditional piezoelectric semiconductor, such as ZnO or nity to distinguish pressure and bending because triboelectric
GaN, can also be utilized in LED structures to develop pres- devices only respond to dynamic pressures while piezoresistive
sure sensors by exploiting the piezo-phototronic effect.[158,159] devices respond to both as shown in Figure 11b.[40]
For example, n-type ZnO nanowires can be coupled with p-type Some direction-sensitive pressures sensors exhibit much
silicon, or PEPOT: PSS to form a p–n junction with a tunable stronger responses to a pressure input over other types of
recombination rate of electrons and holes that are sensitive to mechanical stimulus, such as bending. This has been explained
the input pressure.[160,161] The resulted devices show a spatial by Lee et al. from the material perspective. They reported that
resolution below 10 µm, which is substantially smaller than bending insensitivity can be achieved through electrospun thin
that of thin-film structures, as shown in Figure 10f,g. films (total device thickness ≈2  µm), where nanofibers can
Integration with Phosphor Materials: Phosphorescent mate- change their relative positions to relieve the stresses induced
rials, such as ZnS, emit light in electric fields due to their by the bending deformation while maintaining the pressure
intrinsic heterojunctions. Dopants, such as copper and man- sensor’s sensitivity even when the bending radius is as small
ganese, function as wavelength centers within the ZnS lattice as 80 µm (Figure 11d).[116] They also reported that a thinner
to generate luminescence with different wavelengths. Because substrate layer can help to achieve better bending-insensitive
of their powder-like morphology, dopants can be mixed responses as a reduced thickness indicates small internal strain
with polymers to act as dielectric layers (the green layer) in as shown in Figure 11c. The same results were demonstrated
the piezocapacitive sensors as shown in Figure 10h.[162] Because by other research groups.[122] Luo et al. systematically inves-
of the asymmetry of the wurtzite structure of the ZnS, the tigated the influence of the film thickness and the filler con-
piezo-phototronic effect can also be induced in ZnS. It pro- centration in composite materials on bending insensitivity,
vides energy to help excite the electrons in Cu2+/Mn2+. When and it was found that a reduced thickness and proper filler
the excited electrons return to the ground state, luminescence concentration (much higher than the percolation threshold
is generated.[105,163] The dynamic pressure mapping of person- value) are beneficial to minimizing the influence of bending.[51]
alized handwriting has been realized by the device shown in More research and investigations, including material design,
Figure 10i, where the collection of graphics and pressure inten- device integration, and data processing, are needed to fully
sity recording are demonstrated in Figure 10j.[123] address this challenge.

5.4.2. Direction-Sensitive Force Detection 5.4.3. Other Functions

Pressure sensors generate responses not only to normal pres- In addition to the aforementioned functions, some other
sure inputs, but also to other types of mechanical stimuli that important properties, such as self-healing, biodegradability,
induce deformation, such as bending and shearing.[164,165] texture recognition, self-power, and wireless transmission,
This increases the difficulty of post data processing when the are highly demanded for specific applications. Self-healing
pressure sensors are used as electronic skins or in health care is the ability of a material to recover its mechanical and elec-
monitoring. Distinguishing stimulus is a common method to trical properties when the material is mechanically processed,
reduce the influence of undesirable stimulus. Distinguishing e.g., cut or stamped. This is related to the incorporation of
here means that a sensor outputs different types of signals dynamic molecular bonds existing in specific materials.[1,169,170]
according to the nature of the different input signals. For Researchers often evaluate such ability through the test of the
example, 3D force sensing can be realized in a piezocapacitive recovery time, resistance variation before and after the damage,
pressure sensor through the integration of four parallel-plate and whether supplementary energy, such as heat, is needed.

Adv. Funct. Mater. 2019, 1808509 1808509  (16 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 11.  a) Electrode design of the capacitive pressure sensor capable of distinguishing forces from three directions. Reproduced with permis-
sion.[167] Copyright 2014, WILEY-VCH Verlag GmbH & Co. b) Illustration on how integration of a pressure sensor and a triboelectric generator can be
used to distinguish pressure and bending input. Top: signals of triboelectric voltage and resistance under pressure. Bottom: signals of triboelectric
voltage and resistance under bending. Reproduced with permission.[40] Copyright 2017, WILEY-VCH Verlag GmbH & Co. c) Resistance variation of the
pressure sensors made on substrates with different thickness and under different bending radius. d) Pressure responses of the bending insensitive
sensor under bending states with different bending radii, when the sensor was hit by metal bars with 0.4 g and 1.6 g mass. Reproduced with permis-
sion.[116] Copyright 2016, Springer Nature.

More information regarding self-healing devices can be found becomes an important function when real-time extraction and
in some published review papers.[1,171,172] Biodegradability is a display of the sensing information is required.[108,180,181] Induc-
crucial property for pressure sensors developed for implants; if tive coupling using an LC resonator has been applied for pas-
a device is biodegradable, it means no additional surgical proce- sive wireless data transmission. In this design, an inductor
dures are required to remove the implant; and the device would and a capacitor are connected in series to form the LC reso-
be degraded within the body through reactions with body fluid. nator.[141,182] As the applied pressure changes, the capacitance
The nontoxic reaction products should be absorbed or excreted or inductance of the sensor changes; accordingly, the frequency
outside the body. This can be achieved by using biodegradable response of the wireless sensor is changed. The sensor design
polymers and metals.[73,173,174] incorporating an LC resonator is easy for miniaturization due
For the e-skin to mimic the tactile sensation of human to its simple structure. However, the transmitted data needs to
hands, the ability of texture recognition is needed. In order be further processed as the signal only reflects the raw phase or
to fulfill this requirement, microstructures such as parallel frequency response of the device. Other wireless technologies,
ridges and pyramids, are fabricated on the surface of pressure e.g. Bluetooth, ZigBee, Infrared Data Association (IrDA), Near
sensors to transmit and amplify the vibration induced by the Field Communication (NFC), etc., have also been used in wear-
texture.[175,176] When the sensors are moving across the sur- able sensing systems. Detailed comparison of these technolo-
face texture, the electrical signals generated by the sensors gies can be found in previous review papers.[183,184]
are recorded over time. The signals are then converted to the
frequency-domain output through fast Fourier transform.[177]
It is demonstrated that the frequency with the highest ampli- 6. Manufacturing
tude can be related to the periodicity distance of the surface
texture,[178] with the minimum detectable stripe width around From the application perspective, large-area fabrication in com-
15 µm.[176] bination with small pixels and batch production are highly
For the sensor system design, some practical functions are desired. Importantly, for batch processes, each pixel must
considered by researchers such as self-power and wireless have high repeatability to generate reliable results. Unlike fab-
transmission. Self-power here means that the sensor can work ricating devices in laboratories, manual fabrication of active
without external power supply. The piezoelectric and triboelec- materials with complex surface/bulk structures is not viable
tric sensors can realize this function because of the intrinsic in terms of both cost and throughput. To tackle this challenge,
properties of the active material.[179] Wireless transmission advanced manufacturing methods have been developed and

Adv. Funct. Mater. 2019, 1808509 1808509  (17 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

demonstrated great potentials in manipulating and fabricating mesh fabricated through electrospinning can function as a
various material structures in a fast and controllable way. They layer of dielectric or piezoelectric materials with tunable proper-
can work independently to fabricate functional materials, or ties.[210,211] The nonconductive polymer mesh can be converted
synergistically to produce integrated devices.[185,186] We sum- to piezoresistive materials through carbonization,[37,209] or sur-
marize the recent efforts that utilize these methods in the face coating.[207,212] Kweon et al. fabricated a sponge-like PVDF-
fabrication of flexible pressure sensors to provide a sensor HFP network as the active material, with PEDOT grown on the
development guideline in consideration of scaling up the fab- fiber surface to provide conductivity, and demonstrated a sensor
rication processes for practical and commercial applications. In array with 16 × 10 pixels.[180] Several studies have introduced
the following sections, we introduce three categories of manu- conductive fillers in the ink solution of the electrospinning
facturing methods according to the involved physical processes. system to realize composite fiber networks. Notably, a uniform
dispersion of the filler still remains a challenge; and a high
filler loading ratio is difficult to achieve.[116] The application
6.1. Mechanical Force-Assisted Manufacturing Methods of electric fields during the electrospinning process is benefi-
cial to the formation of the crystalline β phase in PVDF, which
In this section, we introduce manufacturing methods that increases the intensity of the output piezoelectric signals.[213,214]
employ pumping or apply contact pressure via nozzles or molds. In addition, using separated parallel collectors, or rotating the
Examples include direct inkjet printing, microcontact printing, collector can achieve aligned piezoelectric nanofiber arrays,
roll-to-roll printing, and screen printing. Among them, the roll- which have higher voltage output than the randomly dispersed
to-roll printing and screen printing processes are particularly fiber arrays, as shown in Figure 12b.[96,215] The output films are
suitable for fabricating 2D patterns, such as electrodes.[71,187–191] free standing with numerous aligned fibers inside as shown in
For example, all-printed high-density thin film transistor (TFT) Figure 12c. This can also be realized through near-field elec-
arrays, e.g., 9.3 points in.−1, were fabricated through roll-to-roll trospinning (i.e., electrohydrodynamic direct-writing), which
printing, where the pressure sensing function was achieved shows advantages in precise pattern control that conventional
by integrating with piezoresistive rubbers.[191] Recently, a few electrospinning processes cannot achieve. The set-up of near
ultrahigh-resolution roll-to-roll systems have been developed field electrospinning is the same as conventional ones, with
based on flexures and advanced control, which may be used the nozzle to collector distance decreasing from centimeters to
to produce next-generation high-resolution pressure sensor millimeters as shown in Figure 12d.[216] Because of the reduced
arrays.[192–194] For direct inkjet printing,[195] the method prints working distance, the fibers can either reach the collector in the
both 2D and 3D structures by the extrusion of polymer-based stable zone (vs the whipping instable zone),[217] or be dragged
inks through a nozzle following the programmed trajectory. In by a collector that moves at high speeds to overcome the whip-
addition to the electrode fabrication,[196–198] the direct writing ping forces caused by environmental influences.[218] Currently,
technology has a unique advantage in building 3D foam-like there are only limited material systems available for the near-
structures, such as conductive aerogel, which can sustain large field electrospinning process, and more efforts are needed in
deformation.[199–201] Chen et al. produced a CNT coated PU this direction.
foam with hierarchical porosities through the sequential direct
writing process, etching, and dip-coating. The high elasticity of
this structure results in a wide sensing range of 40 kPa.[202] Guo 6.3. Light-Assisted Manufacturing Methods
et al. reported an all printed tactile sensor containing a silicone
base, silver/silicone electrodes, a spring-like sensing layer, and Light can be used to provide energy for 1) material carboniza-
an isolating layer through the continuous direct writing pro- tion or sintering to achieve conductivity, 2) material subtrac-
cess, as shown in Figure  12a.[203] However, since the diameter tion, and 3) polymerization. In electrode fabrication, laser focus
of the printed filament heavily depends on the nozzle size and can generate localized heat to selectively convert nonconduc-
position, printing speed, and the precursor composition, the tive materials into conductive carbon materials.[219] Pristine
fabrication difficulty are greatly increased when a filament with textile or paper can be transformed into interdigital electrodes
small diameters, e.g., <50 µm, is needed.[204] through laser scribing to help the fabrication of all-textile pres-
sure sensors or paper electronics.[220] The resulted carbon mate-
rial usually possesses a nano/microporous structure, which can
6.2. Electric Field-Assisted Manufacturing Methods be used as scaffolds to support composite materials, as shown
in Figure 12e.[221] Similarly, the highly concentrated laser power
Electroplating, electrospraying, and electrospinning are three can be used to sinter the unconnected metal nanoparticles to
common manufacturing methods enabled by the application of achieve conductive paths.[222] Material subtraction via light can
electric fields. The first two methods are useful in surface deco- be achieved in the forms of melting, evaporation or sublimation
ration[205,206] and electrospinning is useful in creating porous depending on the applied laser power. Ultrafast lasers are often
architectures with micro- or nanofibers.[207–209] In an electro- used to fabricate high-resolution structures (<400 nm) at the
spinning process, high voltages are applied to the droplets at expense of cost. Recent development in parallel ultrafast laser
the nozzle; when the electric force becomes larger than the sur- processes may substantially lower the cost of high-precision
face tension force, a jet shoots out from the Taylor cone. The laser processing.[223] For example, lasers have been used to
diameter of the fiber is usually much smaller than ones made fabricate molds with conical or pyramid cavities to replace the
from 3D printing, e.g., <1  µm. Flexible and porous polymer conventional silicon molds for soft lithography.[224]

Adv. Funct. Mater. 2019, 1808509 1808509  (18 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Figure 12.  a) Fabrication process of an all 3D-printed tactile sensor. Reproduced with permission.[203] Copyright 2017, WILEY-VCH Verlag GmbH & Co.
b) An electrospinning setup for highly aligned nanofibers of PVDF. Here, E means the direction of the electric field and ΔV is the applied bias. c) Optical
image (scale bar, 1 cm) and SEM image (in the dashed box, scale bar, 10 µm) of aligned arrays fabricated through the equipment in b). Reproduced
with permission.[96] Copyright 2013, Springer Nature. d) The set-up of the near-field electrospinning. Reproduced with permission.[218] Copyright 2014,
Royal Society of Chemistry. e) Fabrication process of the carbon/silver nanocomposite using laser pyrolization and selective aqueous silver ionic solu-
tion trapping. Reproduced with permission.[221] Copyright 2016, ACS Publications. f) Schematic illustration of dynamic optical projection stereolithog-
raphy. Reproduced with permission.[227] Copyright 2012, WILEY-VCH Verlag GmbH & Co. g) Laser scanning confocal images of the PAAm composite
micropattern. Reproduced with permission.[228] Copyright 2018, WILEY-VCH Verlag GmbH & Co.

Adv. Funct. Mater. 2019, 1808509 1808509  (19 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

Light, e.g., ultraviolent light, can convert certain liquid Acknowledgements


precursors from low molecular weight materials into solid
Y.H. and X.Y.F. contributed equally to this work. This work was supported
polymers of high molecular weight through the crosslinking
by Research Grants Council of Hong Kong (Grant Nos. CUHK 14207615
effect.[225] The laser point can cure the liquid precursor point and CUHK 14209081) and Hong Kong Innovation and Technology
by point to fabricate the patterned structures.[226] In order to Commission (Grant Nos. MRP/056/17X and ITS/428/17FP).
increase the throughput, digital-mircormirror devices (DMDs)
have been used to realize stereolithography. In this method,
millions of micromirrors on a DMD function as program-
mable pixels of a designed image; each micromirror has two
Conflict of Interest
stable states at 12° and −12° corresponding to the “on” and “off” The authors declare no conflict of interest.
states in a sliced mold, as shown in Figure 12f. Comparing with
other methods, e.g., focused ion-beam lithography and two-
photon polymerization (TPP), this approach achieves a higher Keywords
fabrication speed facilitated by the “batch fabrication process”
enabled by the DMD.[227] After passing through the projection applications, manufacturing, materials, modeling, pressure sensors
optics, the guided light (e.g., UV light) polymerizes all mate-
Received: November 30, 2018
rials at the focal plane at once; and rapidly produce 3D struc- Revised: January 6, 2019
tures in a layer by layer fashion. Structured PAAm hydrogels Published online:
with ionic fillers or gold nanowires inside have been fabricated
as the active material for sensors using this method, as shown
in Figure 12g.[228,229] Notably, several parallelized TPP methods
have been developed recently which may be used to fabricate [1] S. Wang, J. Y. Oh, J. Xu, H. Tran, Z. Bao, Acc. Chem. Res. 2018, 51,
1033.
high-resolution pressure sensors with high throughput and
[2] W. Wu, H. Haick, Adv. Mater. 2018, 30, 1705024.
yield in the near future.[230,231]
[3] X. Wang, Z. Liu, T. Zhang, Small 2017, 13, 1602790.
[4] Y. J. Hong, H. Lee, J. Kim, M. Lee, H. J. Choi, T. Hyeon, D.-H. Kim,
Adv. Funct. Mater. 2018, 28, 1805754.
7. Conclusion [5] H. Xu, L. Yin, C. Liu, X. Sheng, N. Zhao, Adv. Mater. 2018, 30,
1800156.
We have reviewed the latest development of the material and [6] S. Zhao, J. Li, D. Cao, G. Zhang, J. Li, K. Li, Y. Yang, W. Wang, Y. Jin,
device design, modeling, and manufacturing methods of flex- R. Sun, C.-P. Wong, ACS Appl. Mater. Interfaces 2017, 9, 12147.
ible pressure sensors according to their working mechanisms. [7] J. Kim, R. Kumar, A. J. Bandodkar, J. Wang, Adv. Electron. Mater.
Important applications in the areas of human–machine inter- 2017, 3, 1600260.
[8] M. Amjadi, K.-U. Kyung, I. Park, M. Sitti, Adv. Funct. Mater. 2016,
face, electronic skin, and health monitoring have been dis-
26, 1678.
cussed. A number of application-driven functions are reviewed
[9] S. Chen, K. Jiang, Z. Lou, D. Chen, G. Shen, Adv. Mater. Technol.
with an emphasis on how direction-sensitive detection and 2018, 3, 1700248.
pressure distribution visualization may enhance pressure [10] Y. Zang, F. Zhang, C.-A. Di, D. Zhu, Mater. Horiz. 2015, 2, 140.
measurement accuracy and human–machine interaction. [11] Y. Lee, J. Kim, H. Joo, M. S. Raj, R. Ghaffari, D.-H. Kim, Adv. Mater.
Lastly, we summarized the advanced manufacturing methods Technol. 2017, 2, 1700053.
used for large-scale fabrication in this area. [12] T. Wang, H. Yang, D. Qi, Z. Liu, P. Cai, H. Zhang, X. Chen, Small
Future research work and efforts for practical applica- 2018, 14, 1702933.
tions and commercialization should focus on addressing [13] T. Yang, D. Xie, Z. Li, H. Zhu, Mater. Sci. Eng., R 2017, 115, 1.
the following challenges. The first one is the development of [14] S.-T. Han, H. Peng, Q. Sun, S. Venkatesh, K.-S. Chung, S. C. Lau,
Y. Zhou, V. A. L. Roy, Adv. Mater. 2017, 29, 1700375.
a multiscale, consistent theoretical description of sensing
[15] T. Q. Trung, N.-E. Lee, Adv. Mater. 2016, 28, 4338.
mechanisms that bridge the material properties and device
[16] S. Yao, P. Swetha, Y. Zhu, Adv. Healthcare Mater. 2018, 7, 1700889.
performances. The second one is the selection and develop- [17] W. Zeng, L. Shu, Q. Li, S. Chen, F. Wang, X.-M. Tao, Adv. Mater.
ment of proper active materials which can be mass-produced 2014, 26, 5310.
in a repeatable, ecofriendly, and inexpensive fashion with high [18] L. Pan, A. Chortos, G. Yu, Y. Wang, S. Isaacson, R. Allen, Y. Shi,
sensing performances. The third one is the design and devel- R. Dauskardt, Z. Bao, Nat. Commun. 2014, 5, 3002.
opment of large-area array-device to achieve smaller pixel sizes [19] J. Park, J. Kim, J. Hong, H. Lee, Y. Lee, S. Cho, S.-W. Kim, J. J. Kim,
and minimal cross-talks. The fourth one is the integration of S. Y. Kim, H. Ko, NPG Asia Mater. 2018, 10, 163.
pressure sensors with other flexible electronics, including [20] S. C. B. Mannsfeld, B. C. K. Tee, R. M. Stoltenberg, C. V. H. H. Chen,
interconnects and data collection and processing components. S. Barman, B. V. O. Muir, A. N. Sokolov, C. Reese, Z. Bao, Nat.
Mater. 2010, 9, 859.
The aforementioned challenges provide an opportunity for
[21] Y. Wei, S. Chen, X. Yuan, P. Wang, L. Liu, Adv. Funct. Mater. 2016,
researchers from different disciplines, e.g., chemistry, mate-
26, 5078.
rial science, mechanical engineering, electronic engineering [22] Y. Shu, H. Tian, Y. Yang, C. Li, Y. Cui, W. Mi, Y. Li, Z. Wang,
and computer engineering, to contribute efforts collectively and N. Deng, B. Peng, T.-L. Ren, Nanoscale 2015, 7, 8636.
cooperatively to realize the goal of low-cost, high-performance, [23] D. Kwon, T.-I. Lee, J. Shim, S. Ryu, M. S. Kim, S. Kim, T.-S. Kim,
and high reliability wearable electronics for improving the I. Park, ACS Appl. Mater. Interfaces 2016, 8, 16922.
overall human well-being. [24] A. Chhetry, H. Yoon, J. Y. Park, J. Mater. Chem. C 2017, 5, 10068.

Adv. Funct. Mater. 2019, 1808509 1808509  (20 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

[25] K. Lee, J. Lee, G. Kim, Y. Kim, S. Kang, S. Cho, S. Kim, J.-K. Kim, [56] M. Jian, K. Xia, Q. Wang, Z. Yin, H. Wang, C. Wang, H. Xie,
W. Lee, D.-E. Kim, S. Kang, D. Kim, T. Lee, W. Shim, Small 2017, M. Zhang, Y. Zhang, Adv. Funct. Mater. 2017, 27, 1606066.
13, 1700368. [57] J. Shi, L. Wang, Z. Dai, L. Zhao, M. Du, H. Li, Y. Fang, Small 2018,
[26] W. Cheng, J. Wang, Z. Ma, K. Yan, Y. Wang, H. Wang, S. Li, Y. Li, 14, 1800819.
L. Pan, Y. Shi, IEEE Electron Device Lett. 2018, 39, 288. [58] B. Su, S. Gong, Z. Ma, L. W. Yap, W. Cheng, Small 2015, 11, 1886.
[27] N. Luo, J. Zhang, X. Ding, Z. Zhou, Q. Zhang, Y.-T. Zhang, [59] N. Gao, X. Zhang, S. Liao, H. Jia, Y. Wang, ACS Macro Lett. 2016,
S.-C. Chen, J.-L. Hu, N. Zhao, Adv. Mater. Technol. 2018, 3, 5, 823.
1700222. [60] W. Chen, X. Gui, B. Liang, R. Yang, Y. Zheng, C. Zhao, X. Li,
[28] Q. Hua, J. Sun, H. Liu, R. Bao, R. Yu, J. Zhai, C. Pan, Z. L. Wang, H. Zhu, Z. Tang, ACS Appl. Mater. Interfaces 2017, 9, 24111.
Nat. Commun. 2018, 9, 244. [61] G. Y. Bae, S. W. Pak, D. Kim, G. Lee, D. H. Kim, Y. Chung, K. Cho,
[29] L.-Q. Tao, K.-N. Zhang, H. Tian, Y. Liu, D.-Y. Wang, Y.-Q. Chen, Adv. Mater. 2016, 28, 5300.
Y. Yang, T.-L. Ren, ACS Nano 2017, 11, 8790. [62] Z. Wang, S. Wang, J. Zeng, X. Ren, A. J. Y. Chee, B. Y. S. Yiu,
[30] C. Chen, J. Song, S. Zhu, Y. Li, Y. Kuang, J. Wan, D. Kirsch, L. Xu, W. C. Chung, Y. Yang, A. C. H. Yu, R. C. Roberts, A. C. O. Tsang,
Y. Wang, T. Gao, Y. Wang, H. Huang, W. Gan, A. Gong, T. Li, J. Xie, K. W. Chow, P. K. L. Chan, Small 2016, 12, 3827.
L. Hu, Chem 2018, 4, 544. [63] H. Park, Y. R. Jeong, J. Yun, S. Y. Hong, S. Jin, S.-J. Lee, G. Zi,
[31] D. J. Lipomi, M. Vosgueritchian, B. C. K. Tee, S. L. Hellstrom, J. S. Ha, ACS Nano 2015, 9, 9974.
J. A. Lee, C. H. Fox, Z. Bao, Nat. Nanotechnol. 2011, 6, 788. [64] W. Liu, N. Liu, Y. Yue, J. Rao, F. Cheng, J. Su, Z. Liu, Y. Gao, Small
[32] B. Nie, R. Li, J. Cao, J. D. Brandt, T. Pan, Adv. Mater. 2015, 27, 2018, 14, 1704149.
6055. [65] C. Luo, N. Liu, H. Zhang, W. Liu, Y. Yue, S. Wang, J. Rao, C. Yang,
[33] Y. Huang, Y. Chen, X. Fan, N. Luo, S. Zhou, S.-C. Chen, N. Zhao, J. Su, X. Jiang, Y. Gao, Nano Energy 2017, 41, 527.
C. P. Wong, Small 2018, 14, 1801520. [66] J. Xue, J. Chen, J. Song, L. Xu, H. Zeng, J. Mater. Chem. C 2017, 5,
[34] J. Mu, C. Hou, G. Wang, X. Wang, Q. Zhang, Y. Li, H. Wang, 11018.
M. Zhu, Adv. Mater. 2016, 28, 9491. [67] X. Xu, R. Wang, P. Nie, Y. Cheng, X. Lu, L. Shi, J. Sun, ACS Appl.
[35] J. Wang, J. Jiu, M. Nogi, T. Sugahara, S. Nagao, H. Koga, P. He, Mater. Interfaces 2017, 9, 14273.
K. Suganuma, Nanoscale 2015, 7, 2926. [68] P. Song, H. Qin, H.-L. Gao, H.-P. Cong, S.-H. Yu, Nat. Commun.
[36] J. Duan, X. Liang, J. Guo, K. Zhu, L. Zhang, Adv. Mater. 2016, 28, 2018, 9, 2786.
8037. [69] N. Luo, W. Dai, C. Li, Z. Zhou, L. Lu, C. C. Y. Poon, S.-C. Chen,
[37] Q. Wang, M. Jian, C. Wang, Y. Zhang, Adv. Funct. Mater. 2017, 27, Y. Zhang, N. Zhao, Adv. Funct. Mater. 2016, 26, 1178.
1605657. [70] C. Pang, J. H. Koo, A. Nguyen, J. M. Caves, M.-G. Kim, A. Chortos,
[38] M. Cao, M. Wang, L. Li, H. Qiu, M. A. Padhiar, Z. Yang, Nano K. Kim, P. J. Wang, J. B.-H. Tok, Z. Bao, Adv. Mater. 2015, 27, 634.
Energy 2018, 50, 528. [71] S. Yao, Y. Zhu, Nanoscale 2014, 6, 2345.
[39] Y. Huang, X. He, L. Gao, Y. Wang, C. Liu, P. Liu, J. Mater. Sci.: [72] Y. Joo, J. Byun, N. Seong, J. Ha, H. Kim, S. Kim, T. Kim, H. Im,
Mater. Electron. 2017, 28, 9495. D. Kim, Y. Hong, Nanoscale 2015, 7, 6208.
[40] H. Chen, Z. Su, Y. Song, X. Cheng, X. Chen, B. Meng, Z. Song, [73] C. M. Boutry, A. Nguyen, Q. O. Lawal, A. Chortos,
D. Chen, H. Zhang, Adv. Funct. Mater. 2017, 27, 1604434. S. Rondeau-Gagné, Z. Bao, Adv. Mater. 2015, 27, 6954.
[41] Y. Ma, N. Liu, L. Li, X. Hu, Z. Zou, J. Wang, S. Luo, Y. Gao, Nat. [74] S. G. Yoon, S. T. Chang, J. Mater. Chem. C 2017, 5, 1910.
Commun. 2017, 8, 1207. [75] G. Schwartz, B. C. K. Tee, J. Mei, A. L. Appleton, D. H. Kim,
[42] Z. Chen, T. Ming, M. M. Goulamaly, H. Yao, D. Nezich, H. Wang, Z. Bao, Nat. Commun. 2013, 4, 1859.
M. Hempel, M. Hofmann, J. Kong, Adv. Funct. Mater. 2016, 26, [76] S. Wan, H. Bi, Y. Zhou, X. Xie, S. Su, K. Yin, L. Sun, Carbon 2017,
5061. 114, 209.
[43] Y. Si, X. Wang, C. Yan, L. Yang, J. Yu, B. Ding, Adv. Mater. 2016, 28, [77] S. Jang, E. Jee, D. Choi, W. Kim, J. S. Kim, V. Amoli, T. Sung,
9512. D. Choi, D. H. Kim, J.-Y. Kwon, ACS Appl. Mater. Interfaces 2018,
[44] S. Gong, W. Schwalb, Y. Wang, Y. Chen, Y. Tang, J. Si, 10, 31472.
B. Shirinzadeh, W. Cheng, Nat. Commun. 2014, 5, 3132. [78] S. Park, H. Kim, M. Vosgueritchian, S. Cheon, H. Kim, J. H. Koo,
[45] S. Lee, S. Shin, S. Lee, J. Seo, J. Lee, S. Son, H. J. Cho, H. Algadi, T. R. Kim, S. Lee, G. Schwartz, H. Chang, Z. Bao, Adv. Mater. 2014,
S. Al-Sayari, D. E. Kim, T. Lee, Adv. Funct. Mater. 2015, 25, 3114. 26, 7324.
[46] G. Ge, Y. Cai, Q. Dong, Y. Zhang, J. Shao, W. Huang, X. Dong, [79] H. Kim, G. Kim, T. Kim, S. Lee, D. Kang, M.-S. Hwang, Y. Chae,
Nanoscale 2018, 10, 10033. S. Kang, H. Lee, H.-G. Park, W. Shim, Small 2018, 14, 1703432.
[47] C. Yang, L. Li, J. Zhao, J. Wang, J. Xie, Y. Cao, M. Xue, C. Lu, ACS [80] D. M. Grannan, J. C. Garland, D. B. Tanner, Phys. Rev. Lett. 1981,
Appl. Mater. Interfaces 2018, 10, 25811. 46, 375.
[48] F. Yin, J. Yang, H. Peng, W. Yuan, J. Mater. Chem. C 2018, 6, 6840. [81] Y. Zang, F. Zhang, D. Huang, X. Gao, C.-a. Di, D. Zhu, Nat.
[49] Y. Pang, H. Tian, L. Tao, Y. Li, X. Wang, N. Deng, Y. Yang, T.-L. Ren, Commun. 2015, 6, 6269.
ACS Appl. Mater. Interfaces 2016, 8, 26458. [82] M.-F. Lin, J. Xiong, J. Wang, K. Parida, P. S. Lee, Nano Energy 2018,
[50] D. Lee, H. Lee, Y. Jeong, Y. Ahn, G. Nam, Y. Lee, Adv. Mater. 2016, 44, 248.
28, 9364. [83] B. Nie, R. Li, J. D. Brandt, T. Pan, Lab Chip 2014, 14, 1107.
[51] N. Luo, Y. Huang, J. Liu, S.-C. Chen, C. P. Wong, N. Zhao, Adv. [84] W. Liu, X. K. Zhang, F. Wu, Y. Xiang, IOP Conf. Ser.: Mater. Sci. Eng.
Mater. 2017, 29, 1702675. 2017, 213, 012036.
[52] J. Park, Y. Lee, J. Hong, M. Ha, Y.-D. Jung, H. Lim, S. Y. Kim, H. Ko, [85] B. Nie, S. Xing, J. D. Brandt, T. Pan, Lab Chip 2012, 12, 1110.
ACS Nano 2014, 8, 4689. [86] Z. Zhu, R. Li, T. Pan, Adv. Mater. 2018, 30, 1705122.
[53] B. Zhu, Z. Niu, H. Wang, W. R. Leow, H. Wang, Y. Li, L. Zheng, [87] S. H. Cho, S. W. Lee, S. Yu, H. Kim, S. Chang, D. Kang, I. Hwang,
J. Wei, F. Huo, X. Chen, Small 2014, 10, 3625. H. S. Kang, B. Jeong, E. H. Kim, S. M. Cho, K. L. Kim, H. Lee,
[54] Y. Pang, K. Zhang, Z. Yang, S. Jiang, Z. Ju, Y. Li, X. Wang, D. Wang, W. Shim, C. Park, ACS Appl. Mater. Interfaces 2017, 9, 10128.
M. Jian, Y. Zhang, R. Liang, H. Tian, Y. Yang, T.-L. Ren, ACS Nano [88] Z. Qiu, Y. Wan, W. Zhou, J. Yang, J. Yang, J. Huang, J. Zhang,
2018, 12, 2346. Q. Liu, S. Huang, N. Bai, Z. Wu, W. Hong, H. Wang, C. F. Guo,
[55] C.-W. Tsao, X.-C. Guo, W.-W. Hu, RSC Adv. 2016, 6, 113344. Adv. Funct. Mater. 2018, 28, 1802343.

Adv. Funct. Mater. 2019, 1808509 1808509  (21 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

[89] R. Li, Y. Si, Z. Zhu, Y. Guo, Y. Zhang, N. Pan, G. Sun, T. Pan, Adv. S. H. Jung, T. Bretl, J. Xiao, M. J. Slepian, Y. Huang, J.-W. Jeong,
Mater. 2017, 29, 1700253. J. A. Rogers, Sci. Adv. 2016, 2, e1601185.
[90] M. Akiyama, Y. Morofuji, T. Kamohara, K. Nishikubo, M. Tsubai, [119] R. Li, B. Nie, P. Digiglio, T. Pan, Adv. Funct. Mater. 2014, 24,
O. Fukuda, N. Ueno, J. Appl. Phys. 2006, 100, 114318. 6195.
[91] X. Wang, J. Zhou, J. Song, J. Liu, N. Xu, Z. L. Wang, Nano Lett. [120] C.-C. Kim, H.-H. Lee, K. H. Oh, J.-Y. Sun, Science 2016, 353, 682.
2006, 6, 2768. [121] X. Wang, H. Zhang, L. Dong, X. Han, W. Du, J. Zhai, C. Pan,
[92] Z. Chen, Z. Wang, X. Li, Y. Lin, N. Luo, M. Long, N. Zhao, J.-B. Xu, Z. L. Wang, Adv. Mater. 2016, 28, 2896.
ACS Nano 2017, 11, 4507. [122] M. Kang, J. Kim, B. Jang, Y. Chae, J.-H. Kim, J.-H. Ahn, ACS Nano
[93] M. Kang, J. H. Park, K. I. Lee, J. W. Cho, J. Bae, B. K. Ju, C. S. Lee, 2017, 11, 7950.
Phys. Status Solidi A 2015, 212, 2005. [123] X. Wang, H. Zhang, R. Yu, L. Dong, D. Peng, A. Zhang, Y. Zhang,
[94] H.-J. Tseng, W.-C. Tian, W.-J. Wu, Sensors 2013, 13, 5478. H. Liu, C. Pan, Z. L. Wang, Adv. Mater. 2015, 27, 2324.
[95] J. S. Lee, K.-Y. Shin, O. J. Cheong, J. H. Kim, J. Jang, Sci. Rep. 2015, [124] M. L. Hammock, A. Chortos, B. C.-K. Tee, J. B.-H. Tok, Z. Bao, Adv.
5, 7887. Mater. 2013, 25, 5997.
[96] L. Persano, C. Dagdeviren, Y. Su, Y. Zhang, S. Girardo, [125] J. Kim, M. Lee, H. J. Shim, R. Ghaffari, H. R. Cho, D. Son,
D. Pisignano, Y. Huang, J. A. Rogers, Nat. Commun. 2013, 4, Y. H. Jung, M. Soh, C. Choi, S. Jung, K. Chu, D. Jeon, S.-T. Lee,
1633. J. H. Kim, S. H. Choi, T. Hyeon, D.-H. Kim, Nat. Commun. 2014,
[97] S.-H. Park, H. B. Lee, S. M. Yeon, J. Park, N. K. Lee, ACS Appl. 5, 5747.
Mater. Interfaces 2016, 8, 24773. [126] R. Rui Pedro, L. Pedro Alhais, T. d. A. Anibal, T. Mahmoud,
[98] A. V. Shirinov, W. K. Schomburg, Sens. Actuators, A 2008, 142, 48. M. Carmel, J. Micromech. Microeng. 2018, 28, 034001.
[99] C. Dagdeviren, Y. Su, P. Joe, R. Yona, Y. Liu, Y.-S. Kim, Y. Huang, [127] A. P. Gerratt, H. O. Michaud, S. P. Lacour, Adv. Funct. Mater. 2015,
A. R. Damadoran, J. Xia, L. W. Martin, Y. Huang, J. A. Rogers, Nat. 25, 2287.
Commun. 2014, 5, 4496. [128] Z. Zhan, R. Lin, V.-T. Tran, J. An, Y. Wei, H. Du, T. Tran, W. Lu, ACS
[100] X. Han, W. Du, M. Chen, X. Wang, X. Zhang, X. Li, J. Li, Z. Peng, Appl. Mater. Interfaces 2017, 9, 37921.
C. Pan, Z. L. Wang, Adv. Mater. 2017, 29, 1701253. [129] C. Bartolozzi, L. Natale, F. Nori, G. Metta, Nat. Mater. 2016, 15,
[101] Z. L. Wang, J. Chen, L. Lin, Energy Environ. Sci. 2015, 8, 2250. 921.
[102] F.-R. Fan, L. Lin, G. Zhu, W. Wu, R. Zhang, Z. L. Wang, Nano Lett. [130] R. A. Nawrocki, N. Matsuhisa, T. Yokota, T. Someya, Adv. Electron.
2012, 12, 3109. Mater. 2016, 2, 1500452.
[103] J. Yang, J. Chen, Y. Su, Q. Jing, Z. Li, F. Yi, X. Wen, Z. Wang, [131] A. Chortos, J. Liu, Z. Bao, Nat. Mater. 2016, 15, 937.
Z. L. Wang, Adv. Mater. 2015, 27, 1316. [132] M. Kaltenbrunner, T. Sekitani, J. Reeder, T. Yokota, K. Kuribara,
[104] J. Luo, F. R. Fan, T. Zhou, W. Tang, F. Xue, Z. L. Wang, Extreme T. Tokuhara, M. Drack, R. Schwödiauer, I. Graz, S. Bauer-Gogonea,
Mech. Lett. 2015, 2, 28. S. Bauer, T. Someya, Nature 2013, 499, 458.
[105] X. Wang, M. Que, M. Chen, X. Han, X. Li, C. Pan, Z. L. Wang, Adv. [133] J. Ye, H. Tan, S. Wu, K. Ni, F. Pan, J. Liu, Z. Tao, Y. Qu, H. Ji,
Mater. 2017, 29, 1605817. P. Simon, Y. Zhu, Adv. Mater. 2018, 30, 1801384.
[106] X. Pu, M. Liu, X. Chen, J. Sun, C. Du, Y. Zhang, J. Zhai, W. Hu, [134] B. He, Q. Zhang, L. Li, J. Sun, P. Man, Z. Zhou, Q. Li, J. Guo,
Z. L. Wang, Sci. Adv. 2017, 3. L. Xie, C. Li, X. Wang, J. Zhao, T. Zhang, Y. Yao, J. Mater. Chem. A
[107] K. Dong, J. Deng, W. Ding, A. C. Wang, P. Wang, C. Cheng, 2018, 6, 14594.
Y.-C. Wang, L. Jin, B. Gu, B. Sun, Z. L. Wang, Adv. Energy Mater. [135] H. Park, J. W. Kim, S. Y. Hong, G. Lee, D. S. Kim, J. h. Oh,
2018, 8, 1801114. S. W. Jin, Y. R. Jeong, S. Y. Oh, J. Y. Yun, J. S. Ha, Adv. Funct. Mater.
[108] J. Lee, H. Kwon, J. Seo, S. Shin, J. H. Koo, C. Pang, S. Son, 2018, 28, 1707013.
J. H. Kim, Y. H. Jang, D. E. Kim, T. Lee, Adv. Mater. 2015, 27, 2433. [136] B. C. K. Tee, A. Chortos, A. Berndt, A. K. Nguyen, A. Tom,
[109] G. Cantarella, V. Costanza, A. Ferrero, R. Hopf, C. Vogt, M. Varga, A. McGuire, Z. C. Lin, K. Tien, W.-G. Bae, H. Wang, P. Mei,
L. Petti, N. Münzenrieder, L. Büthe, G. Salvatore, A. Claville, H.-H. Chou, B. Cui, K. Deisseroth, T. N. Ng, Z. Bao, Science 2015,
L. Bonanomi, A. Daus, S. Knobelspies, C. Daraio, G. Tröster, Adv. 350, 313.
Funct. Mater. 2018, 28, 1705132. [137] C. Wan, G. Chen, Y. Fu, M. Wang, N. Matsuhisa, S. Pan,
[110] L. Jin, A. Chortos, F. Lian, E. Pop, C. Linder, Z. Bao, W. Cai, L. Pan, H. Yang, Q. Wan, L. Zhu, X. Chen, Adv. Mater. 2018, 30,
Proc. Natl. Acad. Sci. USA 2018, 115, 1986. 1801291.
[111] S. Kang, J. Lee, S. Lee, S. Kim, J.-K. Kim, H. Algadi, S. Al-Sayari, [138] Y. Kim, A. Chortos, W. Xu, Y. Liu, J. Y. Oh, D. Son, J. Kang,
D.-E. Kim, D. Kim, T. Lee, Adv. Electron. Mater. 2016, 2, 1600356. A. M. Foudeh, C. Zhu, Y. Lee, S. Niu, J. Liu, R. Pfattner, Z. Bao,
[112] B. C.-K. Tee, A. Chortos, R. R. Dunn, G. Schwartz, E. Eason, T.-W. Lee, Science 2018, 360, 998.
Z. Bao, Adv. Funct. Mater. 2014, 24, 5427. [139] K.-Y. Shin, J. S. Lee, J. Jang, Nano Energy 2016, 22, 95.
[113] D. H. Ho, R. Song, Q. Sun, W.-H. Park, S. Y. Kim, C. Pang, [140] C. Deng, W. Tang, L. Liu, B. Chen, M. Li, Z. L. Wang, Adv. Funct.
D. H. Kim, S.-Y. Kim, J. Lee, J. H. Cho, ACS Appl. Mater. Interfaces Mater. 2018, 28, 1801606.
2017, 9, 44678. [141] W. Deng, L. Wang, L. Dong, Q. Huang, IEEE Sens. J. 2018, 18,
[114] H. Bin, Y. Zhe, Z. Yanmin, Z. Jie, W. Qigang, W. Zhipeng, J. Micro- 4886.
mech. Microeng. 2018, 28, 105001. [142] C.-L. Choong, M.-B. Shim, B.-S. Lee, S. Jeon, D.-S. Ko, T.-H. Kang,
[115] S. Jung, J. H. Kim, J. Kim, S. Choi, J. Lee, I. Park, T. Hyeon, J. Bae, S. H. Lee, K.-E. Byun, J. Im, Y. J. Jeong, C. E. Park, J.-J. Park,
D.-H. Kim, Adv. Mater. 2014, 26, 4825. U.-I. Chung, Adv. Mater. 2014, 26, 3451.
[116] S. Lee, A. Reuveny, J. Reeder, S. Lee, H. Jin, Q. Liu, T. Yokota, [143] Y. Gao, H. Ota, E. W. Schaler, K. Chen, A. Zhao, W. Gao,
T. Sekitani, T. Isoyama, Y. Abe, Z. Suo, T. Someya, Nat. Nano- H. M. Fahad, Y. Leng, A. Zheng, F. Xiong, C. Zhang, L.-C. Tai,
technol. 2016, 11, 472. P. Zhao, R. S. Fearing, A. Javey, Adv. Mater. 2017, 29, 1701985.
[117] S. Liu, X. Wu, D. Zhang, C. Guo, P. Wang, W. Hu, X. Li, X. Zhou, [144] Y. Chu, J. Zhong, H. Liu, Y. Ma, N. Liu, Y. Song, J. Liang, Z. Shao,
H. Xu, C. Luo, J. Zhang, J. Chu, ACS Appl. Mater. Interfaces 2017, Y. Sun, Y. Dong, X. Wang, L. Lin, Adv. Funct. Mater. 2018, 28,
9, 24148. 1803413.
[118] Y. Liu, J. J. S. Norton, R. Qazi, Z. Zou, K. R. Ammann, H. Liu, [145] Y.-H. Kao, T.-Y. Tu, P. C.-P. Chao, Y.-P. Lee, C.-L. Wey, Microsyst.
L. Yan, P. L. Tran, K.-I. Jang, J. W. Lee, D. Zhang, K. A. Kilian, Technol. 2016, 22, 1437.

Adv. Funct. Mater. 2019, 1808509 1808509  (22 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

[146] X. Fan, Y. Huang, X. Ding, N. Luo, C. Li, N. Zhao, S.-C. Chen, Adv. [178] S. Chun, A. Hong, Y. Choi, C. Ha, W. Park, Nanoscale 2016, 8,
Funct. Mater. 2018, 28, 1805045. 9185.
[147] R. Yousefi, M. Nourani, S. Ostadabbas, I. Panahi, IEEE J. Biomed. [179] F. R. Fan, W. Tang, Z. L. Wang, Adv. Mater. 2016, 28, 4283.
Health Inf. 2014, 18, 670. [180] O. Y. Kweon, S. J. Lee, J. H. Oh, NPG Asia Mater. 2018, 10, 540.
[148] I. E. Araci, B. Su, S. R. Quake, Y. Mandel, Nat. Med. 2014, 20, [181] K. Meng, J. Chen, X. Li, Y. Wu, W. Fan, Z. Zhou, Q. He, X. Wang,
1074. X. Fan, Y. Zhang, J. Yang, Z. L. Wang, Adv. Funct. Mater. 2018,
[149] C. M. Boutry, Y. Kaizawa, B. C. Schroeder, A. Chortos, A. Legrand, 1806388.
Z. Wang, J. Chang, P. Fox, Z. Bao, Nat. Electron. 2018, 1, 314. [182] L. Y. Chen, B. C. K. Tee, A. L. Chortos, G. Schwartz, V. Tse,
[150] V. K. Thakur, G. Ding, J. Ma, P. S. Lee, X. Lu, Adv. Mater. 2012, 24, D. J. Lipomi, H. S. P. Wong, M. V. McConnell, Z. Bao, Nat.
4071. Commun. 2014, 5, 5028.
[151] H.-H. Chou, A. Nguyen, A. Chortos, J. W. F. To, C. Lu, J. Mei, [183] S. Majumder, T. Mondal, M. Deen, Sensors 2017, 17, 130.
T. Kurosawa, W.-G. Bae, J. B. H. Tok, Z. Bao, Nat. Commun. 2015, [184] B. W. An, J. H. Shin, S.-Y. Kim, J. Kim, S. Ji, J. Park, Y. Lee, J. Jang,
6, 8011. Y.-G. Park, E. Cho, S. Jo, J.-U. Park, Polymers 2017, 9, 303.
[152] Z. L. Wang, Nano Today 2010, 5, 540. [185] M. Vatani, M. Vatani, J. W. Choi, Appl. Phys. Lett. 2016, 108,
[153] C. Pan, M. Chen, R. Yu, Q. Yang, Y. Hu, Y. Zhang, Z. L. Wang, Adv. 061908.
Mater. 2016, 28, 1535. [186] M. G. Mohammed, R. Kramer, Adv. Mater. 2017, 29, 1604965.
[154] G. Kim, S. Cho, K. Chang, W. S. Kim, H. Kang, S.-P. Ryu, [187] W. Yang, N.-W. Li, S. Zhao, Z. Yuan, J. Wang, X. Du, B. Wang,
J. Myoung, J. Park, C. Park, W. Shim, Adv. Mater. 2017, 29, R. Cao, X. Li, W. Xu, Z. L. Wang, C. Li, Adv. Mater. Technol. 2018,
1606120. 3, 1700241.
[155] S. W. Lee, S. H. Cho, H. S. Kang, G. Kim, J. S. Kim, B. Jeong, [188] K. Chen, W. Gao, S. Emaminejad, D. Kiriya, H. Ota, H. Y. Y. Nyein,
E. H. Kim, S. Yu, I. Hwang, H. Han, T. H. Park, S.-H. Jung, K. Takei, A. Javey, Adv. Mater. 2016, 28, 4397.
J. K. Lee, W. Shim, C. Park, ACS Appl. Mater. Interfaces 2018, 10, [189] S. H. Kim, S. Jung, I. S. Yoon, C. Lee, Y. Oh, J.-M. Hong, Adv.
13757. Mater. 2018, 30, 1800109.
[156] R. Shimotsu, T. Takumi, V. Vohra, Sci. Rep. 2017, 7, 6921. [190] C. Yeom, K. Chen, D. Kiriya, Z. Yu, G. Cho, A. Javey, Adv. Mater.
[157] C. Wang, D. Hwang, Z. Yu, K. Takei, J. Park, T. Chen, B. Ma, 2015, 27, 1561.
A. Javey, Nat. Mater. 2013, 12, 899. [191] W. Lee, H. Koo, J. Sun, J. Noh, K.-S. Kwon, C. Yeom, Y. Choi,
[158] M. Peng, Z. Li, C. Liu, Q. Zheng, X. Shi, M. Song, Y. Zhang, S. Du, K. Chen, A. Javey, G. Cho, Sci. Rep. 2016, 5, 17707.
J. Zhai, Z. L. Wang, ACS Nano 2015, 9, 3143. [192] X. Zhou, D. Wang, J. Wang, S.-C. Chen, Precis. Eng. 2016, 45, 332.
[159] C. Pan, L. Dong, G. Zhu, S. Niu, R. Yu, Q. Yang, Y. Liu, Z. L. Wang, [193] X. Zhou, H. Xu, J. Cheng, N. Zhao, S.-C. Chen, Sci. Rep. 2015, 5,
Nat. Photonics 2013, 7, 752. 10402.
[160] M. Chen, C. Pan, T. Zhang, X. Li, R. Liang, Z. L. Wang, ACS Nano [194] S. Kim, H. Sojoudi, H. Zhao, D. Mariappan, G. H. McKinley,
2016, 10, 6074. K. K. Gleason, A. J. Hart, Sci. Adv. 2016, 2.
[161] R. Bao, C. Wang, L. Dong, R. Yu, K. Zhao, Z. L. Wang, C. Pan, Adv. [195] J. A. Lewis, J. E. Smay, J. Stuecker, J. Cesarano, J. Am. Ceram. Soc.
Funct. Mater. 2015, 25, 2884. 2006, 89, 3599.
[162] C. Larson, B. Peele, S. Li, S. Robinson, M. Totaro, L. Beccai, [196] A. D. Valentine, T. A. Busbee, J. W. Boley, J. R. Raney, A. Chortos,
B. Mazzolai, R. Shepherd, Science 2016, 351, 1071. A. Kotikian, J. D. Berrigan, M. F. Durstock, J. A. Lewis, Adv. Mater.
[163] X. Y. Wei, X. Wang, S. Y. Kuang, L. Su, H. Y. Li, Y. Wang, C. Pan, 2017, 29, 1703817.
Z. L. Wang, G. Zhu, Adv. Mater. 2016, 28, 6656. [197] R.-Z. Li, A. Hu, T. Zhang, K. D. Oakes, ACS Appl. Mater. Interfaces
[164] J. Park, Y. Lee, J. Hong, Y. Lee, M. Ha, Y. Jung, H. Lim, S. Y. Kim, 2014, 6, 21721.
H. Ko, ACS Nano 2014, 8, 12020. [198] Z. Lei, Q. Wang, P. Wu, Mater. Horiz. 2017, 4, 694.
[165] Y. Wei, S. Chen, X. Dong, Y. Lin, L. Liu, Carbon 2017, 113, 395. [199] S. Yoon, J. A. Park, H.-R. Lee, W. H. Yoon, D. S. Hwang, S. Jung,
[166] K. Noda, K. Matsumoto, I. Shimoyama, Sens. Actuators, A 2014, Adv. Healthcare Mater. 2018, 7, 1800050.
215, 123. [200] P. Laszczak, L. Jiang, D. L. Bader, D. Moser, S. Zahedi, Med. Eng.
[167] L. Viry, A. Levi, M. Totaro, A. Mondini, V. Mattoli, B. Mazzolai, Phys. 2015, 37, 132.
L. Beccai, Adv. Mater. 2014, 26, 2659. [201] Y. Jiang, Z. Xu, T. Huang, Y. Liu, F. Guo, J. Xi, W. Gao, C. Gao, Adv.
[168] Y. Jung, D.-G. Lee, J. Park, H. Ko, H. Lim, Sensors 2015, 15, Funct. Mater. 2018, 28, 1707024.
25463. [202] Q. Chen, P.-F. Cao, R. C. Advincula, Adv. Funct. Mater. 2018, 28,
[169] Y.-L. Rao, A. Chortos, R. Pfattner, F. Lissel, Y.-C. Chiu, V. Feig, J. Xu, 1800631.
T. Kurosawa, X. Gu, C. Wang, M. He, J. W. Chung, Z. Bao, J. Am. [203] S.-Z. Guo, K. Qiu, F. Meng, S. H. Park, M. C. McAlpine, Adv.
Chem. Soc. 2016, 138, 6020. Mater. 2017, 29, 1701218.
[170] C.-H. Li, C. Wang, C. Keplinger, J.-L. Zuo, L. Jin, Y. Sun, P. Zheng, [204] H. Yuk, X. Zhao, Adv. Mater. 2018, 30, 1704028.
Y. Cao, F. Lissel, C. Linder, X.-Z. You, Z. Bao, Nat. Chem. 2016, 8, 618. [205] B. W. An, S. Heo, S. Ji, F. Bien, J.-U. Park, Nat. Commun. 2018, 9,
[171] V. K. Thakur, M. R. Kessler, Polymer 2015, 69, 369. 2458.
[172] D. G. Bekas, K. Tsirka, D. Baltzis, A. S. Paipetis, Composites, Part [206] S. M. Doshi, E. T. Thostenson, ACS Sens. 2018, 3, 1276.
B 2016, 87, 92. [207] C. Cheng, J. Zhang, S. Li, Y. Xia, C. Nie, Z. Shi, J. L. Cuellar-Camacho,
[173] T. Lei, M. Guan, J. Liu, H.-C. Lin, R. Pfattner, L. Shaw, N. Ma, R. Haag, Adv. Mater. 2018, 30, 1705452.
A. F. McGuire, T.-C. Huang, L. Shao, K.-T. Cheng, J. B. H. Tok, [208] Y. J. Yun, W. G. Hong, W.-J. Kim, Y. Jun, B. H. Kim, Adv. Mater.
Z. Bao, Proc. Natl. Acad. Sci. USA 2017. 2013, 25, 5701.
[174] V. R. Feig, H. Tran, Z. Bao, ACS Cent. Sci. 2018, 4, 337. [209] T. Xu, Y. Ding, Z. Wang, Y. Zhao, W. Wu, H. Fong, Z. Zhu, J. Mater.
[175] J. Park, M. Kim, Y. Lee, H. S. Lee, H. Ko, Sci. Adv. 2015, 1, Chem. C 2017, 5, 10288.
e1500661. [210] Z. Hong-Di, L. Yan-Jie, Z. Jun, Z. Jian-Wei, Q. Qing-Hao,
[176] Y. Cao, T. Li, Y. Gu, H. Luo, S. Wang, T. Zhang, Small 2018, 14, Z. Chang-Zhe, L. Xin, Z. Jun-Cheng, L. Yun-Ze, J. Phys. D: Appl.
1703902. Phys. 2018, 51, 085102.
[177] S. Chun, H. Jung, Y. Choi, G. Bae, J. P. Kil, W. Park, Carbon 2015, [211] T. Lei, L. Yu, G. Zheng, L. Wang, D. Wu, D. Sun, J. Mater. Sci. 2015,
94, 982. 50, 4342.

Adv. Funct. Mater. 2019, 1808509 1808509  (23 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.afm-journal.de

[212] Z. Lou, S. Chen, L. Wang, K. Jiang, G. Shen, Nano Energy 2016, 23, [224] A. dos Santos, N. Pinela, P. Alves, R. Santos, E. Fortunato,
7. R. Martins, H. Águas, R. Igreja, Adv. Electron. Mater. 2018, 4,
[213] L. ChaBum, A. T. Joshua, Smart Mater. Struct. 2014, 23, 095044. 1800182.
[214] B. Li, C. Xu, J. Zheng, C. Xu, Sensors 2014, 14, 9889. [225] S. Wang, J. Xu, W. Wang, G.-J. N. Wang, R. Rastak,
[215] S. Ma, T. Ye, T. Zhang, Z. Wang, K. Li, M. Chen, J. Zhang, Z. Wang, F. Molina-Lopez, J. W. Chung, S. Niu, V. R. Feig, J. Lopez,
S. Ramakrishna, L. Wei, Adv. Mater. Technol. 2018, 3, 1800033. T. Lei, S.-K. Kwon, Y. Kim, A. M. Foudeh, A. Ehrlich,
[216] D. Sun, C. Chang, S. Li, L. Lin, Nano Lett. 2006, 6, 839. A. Gasperini, Y. Yun, B. Murmann, J. B. H. Tok, Z. Bao, Nature
[217] Y. K. Fuh, B. S. Wang, C.-Y. Tsai, Sci. Rep. 2017, 7, 6759. 2018, 555, 83.
[218] Y. Duan, Y. Huang, Z. Yin, N. Bu, W. Dong, Nanoscale 2014, 6, 3289. [226] X.-H. Qin, X. Wang, M. Rottmar, B. J. Nelson, K. Maniura-Weber,
[219] X. Zang, C. Shen, Y. Chu, B. Li, M. Wei, J. Zhong, M. Sanghadasa, Adv. Mater. 2018, 30, 1705564.
L. Lin, Adv. Mater. 2018, 30, 1800062. [227] A. P. Zhang, X. Qu, P. Soman, K. C. Hribar, J. W. Lee, S. Chen,
[220] M. Liu, X. Pu, C. Jiang, T. Liu, X. Huang, L. Chen, C. Du, J. Sun, S. He, Adv. Mater. 2012, 24, 4266.
W. Hu, Z. L. Wang, Adv. Mater. 2017, 29, 1703700. [228] M.-j. Yin, Y. Zhang, Z. Yin, Q. Zheng, A. P. Zhang, Adv. Mater.
[221] R. Rahimi, M. Ochoa, B. Ziaie, ACS Appl. Mater. Interfaces 2016, 8, Technol. 2018, 3, 1800051.
16907. [229] M.-J. Yin, M. Yao, S. Gao, A. P. Zhang, H.-Y. Tam, P.-K. A. Wai, Adv.
[222] J. Kwon, H. Cho, H. Eom, H. Lee, Y. D. Suh, H. Moon, J. Shin, Mater. 2016, 28, 1394.
S. Hong, S. H. Ko, ACS Appl. Mater. Interfaces 2016, 8, 11575. [230] Q. Geng, C. Gu, J. Cheng, S.-C. Chen, Optica 2017, 4, 674.
[223] C. Gu, D. Zhang, D. Wang, Y. Yam, C. Li, S.-C. Chen, Precis. Eng. [231] S. D. Gittard, A. Nguyen, K. Obata, A. Koroleva, R. J. Narayan,
2017, 50, 198. B. N. Chichkov, Biomed. Opt. Express 2011, 2, 3167.

Adv. Funct. Mater. 2019, 1808509 1808509  (24 of 24) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

You might also like