Composites Science and Technology 253 (2024) 110628

Contents lists available at ScienceDirect

Composites Science and Technology

journal homepage: www.elsevier.com/locate/compscitech

A breathable and sweat-absorbing pressure sensor based on PDMS gradient

foam to achieve high sensitivity and wide detection range
Long Chen , Han Ye , Ping Yu *
College of Mechanical and Electrical Engineering, Wenzhou University, Wenzhou, 325035, China

A R T I C L E I N F O A B S T R A C T

Keywords: Various designs of porous structures have been introduced to enhance the sensitivity of resistive pressure sensors.
PDMS gradient stiffness foam (PGSF) However, this method often comes at the cost of narrowing down the measurement range, limiting its use to low-
Breathability pressure detection. This study proposes a soft pressure sensor with two-layer gradient foam structure. By finely
Sweat absorption
adjusting the porosity and cross-linking degree of the PDMS, we aim to enhance the sensitivity without
Motion detection
Pressure sensor
compromising the range of the sensor. Testing results demonstrate that the sensitivity of the gradient foam in­
creases with a higher porosity, and the measurement range expands with a higher cross-linking degree. The soft
sensor with the PDMS gradient stiffness foam (PGSF) achieves a sensitivity of 0.202 kPa− 1 in the range of 0–40
kPa, 0.067 kPa− 1 in the range of 40–210 kPa and 0.020 kPa− 1 in the range of 210–850 kPa. Furthermore, the
sensor exhibits a minimum detection limit of 50 Pa and demonstrates excellent stability (withstanding over 8000
testing cycles), breathability and sweat absorption. Additionally, we conducted motion detection, health
monitoring and pressure mapping tests to showcase the potential applications of our proposed soft sensor in these
fields.

1. Introduction PDMS casting. The fabrication of microstructures such as pyramids and

columns are complex and costly. In contrast, porous foam has attracted
The rapid growth in the artifical intelligence has spurred the swift scholarly attention due to its simple fabrication techniques and lower
development of flexible sensing technologies, encompassing applica­ costs [18–25]. Wang et al. [26] prepared PDMS porous foam through a
tions such as personalized human health monitoring, human-machine salt-leaching template method, yielding materials with low stiffness and
interfaces, electronic skin and soft robotics. Against this backdrop, high compressibility, which provide a 3D cellular scaffold for conductive
flexible pressure sensors have become crucial [1–4]. Flexible pressure fillers to form a permeating network, significantly improving sensitivity
sensors are developed based on various sensing mechanisms such as and response time. However, the narrow detection range typically
resistive [5], capacitive [6,7], piezoelectric [8,9] and triboelectric [10, characterizes foam sensors produced in the past, despite their good
11]effects. Among them, flexible resistive pressure sensors are response to minor pressures. In conclusion, developing flexible resistive
renowned for their simplicity in manufacturing, uncomplicated struc­ pressure sensors with high sensitivity and broad detection range remains
ture and stretchability. Nonetheless, their design is often hindered by a significant challenge.
sensitivity constraints. Furthermore, breathability and sweat-absorbing ability are of great
To maximize the detection performance of the sensitive functional significance in flexible wearable sensors, crucial for enhancing user
materials in tactile sensors, special structures are often employed to comfort and ensuring sensor performance [27–30]. Such sensors typi­
intensify the stress or strain caused by external forces, thereby boosting cally need to be worn on the body for long durations to monitor phys­
sensitivity [12–15]. For instance, Park et al. [16] have enhanced sensi­ iological metrics in real-time. Sensors that afford good breathability and
tivity by fabricating orderly micropillar arrays on a SiO2 substrate using moisture absorption can alleviate potential discomfort over extended
photolithography and reactive ion etching. Guo et al. [17] improved wear periods and increase user acceptance of wearable devices. More­
sensitivity by replicating biomimetic micro tower arrays with high over, maintaining stable sensing performance across different humidity
aspect ratios and low distribution density from lotus leaf surfaces using levels is an important consideration for flexible wearable devices. In

* Corresponding author.
E-mail address: [email protected] (P. Yu).

https://doi.org/10.1016/j.compscitech.2024.110628
Received 12 January 2024; Received in revised form 2 April 2024; Accepted 24 April 2024
Available online 25 April 2024
0266-3538/© 2024 Elsevier Ltd. All rights reserved.
L. Chen et al. Composites Science and Technology 253 (2024) 110628

traditional flexible pressure sensors, materials such as PDMS (poly­ 2.5. Preparation of gradient foam with MWCNT
dimethylsiloxane), PET (polyethylene terephthalate), PI (polyimide)
and silicone rubber are used to fabricate thin films for the sensing layer The fabrication process was illustrated in Fig. 1. First, the base and
or encapsulation layer [31–34]. However, these materials do not take curing agent of PDMS were mixed according to the required ratio
into account the need for comfort in wearable flexible pressure sensors. (5:1,10:1,15:1). Then, NaCl and the PDMS mixture were thoroughly
To address these two issues, inspired by the multi-layered structure mixed at mass ratios of 6:1 and 3:1 respectively. The mixture was then
of human skin [35] and the gradient structured flexible sensors [36], we transferred to a mold（10*10*10 mm^3） and placed at 80 ◦ C for 4 h.
propose a breathable and sweat-absorbing pressure sensor based on After demolding, the sample was placed in water and ultrasonically
PDMS gradient foam. The sensor utilizes two layers of PDMS foam with treated for 4 h to remove the NaCl particles within the sample. The
different porosities and cross-linking degree, filled with MWCNT, to sample was then rinsed with deionized water and placed at 70 ◦ C for 8 h
reconcile the contradiction between the sensitivity and detection range to dry and obtain the PDMS foam.
of the tactile sensing. The high-stiffness foam deforms under high Dopamine hydrochloride (2 mg/mL) was dissolved in tris(hydrox­
pressure, thus expanding the detection range, while the low-stiffness ymethyl) aminomethane buffer solution (Tris-HCl, pH 8.5). The PDMS
foam can measure smaller pressures. To address breathability and foam was immersed in the buffer solution for 24 h at room temperature,
sweat absorption, the gradient PDMS foam is modified by adding PDA under stirring to prepare PDA/PDMS foam. It was then rinsed with ul­
containing hydrophilic groups, and breathable electrodes are fabricated trapure water and dried at 70 ◦ C for 6 h in an oven.
to complete the preparation of the entire sensor. This skin-inspired MWCNT was dispersed in anhydrous ethanol at different mass ratios
design concept of gradient stiffness-adjustable and comfortable pres­ for 2 h to prepare a homogeneous MWCNT dispersion. The prepared
sure sensing materials offers a promising strategy for human health PDA/PDMS foam was added to the MWCNT dispersion, and under
monitoring, human-machine interfaces and electronic skin applications. vacuum, allowed to stand for 10 min. After retrieval, ultrasonic
dispersion was carried out for 2 h. Afterward, the PDMS/PDA/MWCNT
2. Materials and methods foam was removed and dried at 80 ◦ C for 1 h in an oven. Breathable
conductive adhesive tape was attached to both ends of the foam to serve
2.1. Materials as electrodes, thus completing the preparation of the foam sensor.
Samples were named based on variations in porosity and cross-
Sylgard 184 silicone elastomer base and curing agent for preparing linking degree. The PDMS foam prepared with a NaCl/PDMS mass
PDMS were purchased from Dow Corning. Multi-walled carbon nano­ ratio of 3:1 and a base/curing agent mass ratio of 5:1 was named PHSF
tube (MWCNT) were obtained from Suzhou Carbon Nano-Tech Co., Ltd. (Polydimethylsiloxane High Stiffness Foam). The PDMS foam prepared
NaCl powder with a purity of 99.5 % was purchased from Shanghai with a NaCl: PDMS mass ratio of 6:1 and a base/curing agent mass ratio
Macklin Biochemical Co., Ltd. Tris(hydroxymethyl) aminomethane of 15:1 was named PLSF (Polydimethylsiloxane Low Stiffness Foam).
buffer solution (Tris-HCl, pH 8.5) was obtained from PHYGENE. Dopa­ The gradient foam sensor with a PHSF:PLSF height ratio of 1;1 was
mine hydrochloride (DA, 98 %) was obtained from Aladdin. Molds used named as PGSF (Polydimethylsiloxane Gradient Stiffness Foam).The
in the preparation process came from WeNext Technology Co., Ltd. gradient foam sensor with a PHSF: PLSF ratio of 7:3 was named WDRF
(Wide Detection Range Foam), and the gradient foam sensor with a
PHSF: PLSF ratio of 3:7 was named HSF (High Sensitivity Foam). To
2.2. Characterization obtain the PGSF, two custom conductive fabric tapes with micropore
array(Fig. S1) were attached to the top and bottom surfaces of the PGSF
The morphology and microstructure of the PDMS foam, PDA/PDMS with the aid of conductive silver glue. For the 5 × 5 Eskin array (Fig. S2)
foam and PGSF were characterized using a Sigma 300 scanning electron 25 PGSF were attached to the rubber substrate using conductive silver
microscope (SEM) from Zeiss, Germany. The FTIR spectra of PDMS foam glue.
were obtained using a Thermo Scientific Nicolet iS20 Fourier transform
infrared spectrometer. The water contact angles of the PDMS and PDA/ 3. Results and discussion
PDMS foams were measured with a Data Physics Instruments GmbH
OCA25 contact angle measuring instrument, assessing their hydrophil­ 3.1. Characterization of gradient foam
ic/hydrophobic properties. Stress-strain curves were obtained using a
WDW-100E universal testing machine at a compression rate of 1 mm/s. In this study, the sacrificial template method is used to adjust the
foam’s stiffness by controlling the PDMS foam’s porosity and cross-
linking degree. The cross-linking degree increases with content of the
2.3. Sensing performance measurement curing agent within limits[37]. The porosity refers to the ratio of the
void volume to the total volume of the material, including open porosity,
To measure the pressure sensing capability of the PGSF, a WDW- closed porosity, and microporosity. The opposite concept is density,
100E universal testing machine was used to apply forces of different which represents the degree to which the material’s interior is filled with
frequencies and magnitudes. An MS-3010D DC power supply was used solids and affects the rigidity of the foam. A high porosity indicates low
to convert resistance signals into current signals, and a Keithley density.
DMM6500 graphical sampling multimeter was used to record the cur­ The porosity of a material refers to the percentage of the volume of
rent signals. voids or pores in a bulk material to the total volume of the material in its
natural state, denoted as P. The calculation formula for porosity P is[38]:
V0 − V ( ρ)
2.4. Mechanical simulation analysis P= × 100% = 1 − 0 × 100% (1)
V0 ρ
Mechanical simulation analysis was conducted using the commercial P is the porosity of the material; V0 is the total volume of the material
software Workbench 2020 R2. Stress-strain data for low stiffness and in its natural state; ρ0 is the density of the material; V is the volume of the
high stiffness foams were imported based on standard experimental solid; ρ is the density of the solid. The resistance change of gradient foam
measurements into material properties. The Poisson’s ratio was set to is mainly caused by the deformation of the foam. Multi-walled carbon
0.28 according to experimental measurement. Frictionless and non- nanotubes (MWCNT) are uniformly dispersed in the pores of the PDMS
penetrating contact interactions were adopted. foam (Fig. 1c). When pressure is applied, the pores of the foam shrink,

2
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Fig. 1. The preparation process. (a) PDMS foam; (b) PDA/PDMS foam; (c) PDMS/PDA/MWCNT Foam; (d) The overall structure of PGSF.

resulting in more conductive paths being connected, leading to a results demonstrate that the PDA layer has been coated on the PDMS
decrease in resistance, SEM images in Fig. 2d depict the variation in pore foam pore walls.
size under different levels of deformation (100 %, 80 %, 60 %). The microstructure of the PGSF with MWCNT embedded is shown in
Foam with higher density and lower porosity has a higher elastic Fig. 2g. The surface hydrophilicity and wettability are evaluated by
modulus, resulting in less deformation under low pressure, while foam measuring their water contact angles. The contact angle of the original
with higher porosity has a lower elastic modulus and exhibits more PDMS foam is 140.7◦ , while it decreased to 63.7◦ for the PDA/PDMS
significant deformation under low pressure. However, under high foam, indicating significantly enhanced hydrophilicity. On the other
pressure, the deformation tends to saturate due to its higher porosity, hand, the water contact angle of the PGSF increased to 80.7◦ due to the
resulting in a narrower response range. To address this issue, we adopt a aggregation of MWCNT (Fig. S5). The significantly enhanced hydro­
gradient distribution approach by combining PDMS foams with low philicity can be attributed to the presence of many polar functional
stiffness (Fig. 2b) and high stiffness (Fig. 2c) to fabricate pressure sen­ groups (catechol and amino) in the PDA layer. The resistance change in
sitive resistive sensors with gradient compressibility (Fig. 2a). To illus­ PGSF is due to the compressive deformation of PDMS, which results in
trate more clearly, the elemental mapping of silicon (Si) in different the formation of more conductive pathways, and the change in sensing
layers of the PGSF is presented (Fig. S3). When MWCNT are doped into performance is closely related to the amount of deformation. To
pure PDMS foam, due to the hydrophobic nature of PDMS, it is chal­ demonstrate the advantage of the gradient foam and show the influence
lenging to add MWCNT to the core of the foam, resulting in poor sensing of porosity on its mechanical properties, we simplified the foam model
performance. Additionally, the contact between PDMS and MWCNT is and conducted simulation simulations, as shown in Fig. 3a, b, and c.
not stable and prone to separation. These are the issues that need to be
Z0 − Z
addressed and resolved during the sensor fabrication process. Adding ΔZ = × 100% (2)
Z0
PDA to improve the hydrophilicity of PDMS foam can effectively alle­
viate these issues [39]. Fig. S4 compares the Fourier-transform infrared ΔZ signifies the relative change in thickness, where Z is the model’s
spectrometer of pure PDMS foam and PDA/PDMS foam. The original current thickness and Z0 is its initial thickness. Under low pressure, the
PDMS foam exhibits three characteristic peaks at 1256.95 cm− 1, deformation of the low-porosity PDMS foam is small, while the defor­
1009.52 cm− 1 and 790.47 cm− 1, corresponding to the deformation vi­ mation of both the gradient foam and the high-porosity foam is similar.
bration of Si–CH3, stretching vibration of Si–O–Si and stretching vi­ When the pressure reaches 150 kPa, the deformation of the high-
bration of Si–CH3, respectively. However, the PDA/PDMS foam shows porosity foam approaches the threshold, and its compressive modulus
additional characteristic peaks due to the presence of the PDA coating. is comparable to that of pure PDMS. When the pressure exceeds 150 kPa,
The broad absorption in the range of 3600–3100 cm− 1 is caused by the due to the limited deformation of the high-porosity foam, no more
stretching vibration of –OH and –NH, while the absorption peaks at conductive paths can be formed, resulting in the inability to detect
1608.33 cm− 1 and 1509.52 cm− 1 are attributed to the stretching vi­ higher pressures. In contrast, the gradient foam and the low-porosity
bration of the benzene ring C– –C and the bending vibration of -N-H, foam can undergo larger deformations, which allows the detection of
respectively [40–43]. The hydrophilic PDA/PDMS foam is obtained by greater pressures. From the perspective of mechanical properties, foams
modifying with PDA to achieve hydrophilicity (Fig. 1b). The color with different porosities exhibit different compressive moduli, and
change from white to dark brown on the foam, as shown in Fig. 1a and b, Fig. 3d shows the stress-strain curves of PDMS foams with different
visually reflects the PDA coating. The microstructure of PDMS and porosities, which are obtained by changing the mass ratio of salt to
PDA/PDMS foam is evaluated using optical and scanning electron mi­ PDMS (NaCl to PDMS mass ratios of 3:1, 4:1, 5:1 and 6:1). From this
croscopy (SEM) images. The original PDMS foam exhibits a typical figure, we can found that when the mass ratio exceeds 6:1, the foam
porous structure with smooth pore walls (Fig. 2e), resulting from the becomes difficult to mold. And the mass ratio is less than 3:1, PDMS
dissolution of the NaCl template. In contrast, the PDA/PDMS foam had a foam cannot fully encapsulate NaCl particles during the preparation
rough surface with many nanoparticles deposited on the skeleton process, leading to unsuccessful fabrication.
(Fig. 2f) due to the self-polymerization of dopamine monomers. These

3
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Fig. 2. Microscopic characterization of foam. (a) Schematic diagram of the overall PGSF; (b) SEM image of PLSF; (c) SEM image of PHSF; (d) SEM images of the foam
during compression; (e–g) SEM images of PDMS foam, PDA/PDMS foam, and PGSF sensor, respectively.

To investigate the joint regulation cross-linking degree and porosity 3.2. Sensing characteristics
on the stiffness of PDMS foam, we select PDMS foams with two different
porosities (NaCl to PDMS mass ratios of 3:1 and 6:1) and three different To evaluate the performance of the PGSF, we conduct a series of
base/curing agent ratios (15:1, 10:1, 5:1). We test their stress-strain measurements, including pressure sensing, response time, and stability
curves (Fig. 3e). By adjusting the porosity and cross-linking dregee, tests (Fig. 4a). Among them, S is pressure sensitivity, which is defined as
PDMS foams with varying mechanical and sensing performances can be [44]
obtained. This further suggests that we can prepare PGSF based on
S = δ(ΔI / I0 ) / δP (3)
specific pressure detection ranges required. Fig. 3f illustrates the stress-
strain curves of PGSF（before the addition of MWCNT）, PHSF, and where ΔI is the change in current under a certain pressure, I0 is the
PLSF samples. Under the same low-pressure conditions, PGSF exhibits a original current of the samples without pressure, and P is the applied
larger strain compared to PHSF, indicating higher sensitivity when pressure. The amount of MWCNT added also affects the PGSF. On the
incorporating conductive fillers. Furthermore, within a larger pressure one hand, the added MWCNT improves the mechanical properties of
range, PGSF maintains a higher strain while PLSF starts reaching its PDMS foam, and on the other hand, the sensing material has different
strain threshold. This enables PGSF to more effectively detect higher resistance change curves due to the different MWCNT contents. The
pressures. MWCNT content is crucial for the sensing of PGSF, and we selected
ethanol dispersions of MWCNT at different mass ratios (0.5 %, 1 %, 1.5
%, 2 %) (Fig. 4b). We found that the PGSF had the best linearity when

4
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Fig. 3. Mechanical property assessment. (a–c) Simulations of PHSF, PGSF, and PLSF deformation; (d) Porosity impact on foam stress-strain; (e) Porosity and cross-
linking degree impact on foam stress-strain curves; (f) Stress-strain curves comparisons of PHSF, PGSF and PLSF.

the mass ratio of MWCNT is 1 %, so we chose this mass ratio. sensor’s Ohmic properties show good linearity, highlighting the resis­
As shown in Fig. 4c, the PGSF exhibits different responses at different tance stability of the gradient foam sensor. Fig. 4e shows real-time
pressure ranges, with sensitivity values of 0.202 kPa− 1 in the low- monitoring of the sensor’s current under loads ranging from 1 kPa to
pressure range (0–40 kPa), 0.067 kPa− 1 in the medium-pressure range 600 kPa, displaying stable and repeatable signals, with the sensor’s
(40–210 kPa), and 0.020 kPa− 1 in the high-pressure range (210–850 current changing dramatically as the pressure increases. We also
kPa). The different responses are achieved by adjusting the porosity and investigated the sensing behavior related to frequency and found that
changing the proportion of the PDMS curing agent on the two layers. In the ΔI/I0 of PGSF is consistent at different frequencies (0.02 Hz, 0.05 Hz,
the low-pressure region, the PGSF exhibits higher pressure sensitivity, 0.1 Hz, 0.5 Hz, and 1 Hz) under a fixed pressure (20 kPa) as Fig. 4f
primarily because the foam with the higher porosity and lower elastic illustrates.
modulus layer deforms first, forming more conductive pathways and The minimum pressure detection limit and response-recovery time
causing a change in resistance. In the medium pressure region, the are important parameters of the sensor. Fig. 4g demonstrates the output
sensor still maintains high sensitivity because, within this range, the signal (ΔI/I0) of PGSF under repeated loading with an amplitude of 50
foam in the lower porosity layer begins to deform, while the high Pa. From this figure, we can find that the sensor can detect pressure as
porosity layer remains relatively stable. In the high-pressure region, the small as 50 Pa. In Fig. 4h, fast response and recovery times of 60 ms and
deformation of the 3D framework of the low-porosity foam creates new 80 ms, respectively, can be seen under a pressure of 400 kPa. This in­
conductive pathways, leading to decreased contact resistance and dicates that the sensor can rapidly respond to pressure changes and
increased current. revert to its original state. This rapid response characteristic gives the
Furthermore, we have conducted stability assessments of the PGSF, sensor significant advantages in applications where high sensitivity and
including Ohmic characteristic tests under different pressures (0 kPa, 10 quick response times are required.
kPa, 50 kPa, 100 kPa, 200 kPa) and voltages. Fig. 4d indicates that the Long-term stability evaluation is conducted on the PGSF. The test

5
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Fig. 4. Basic sensing performance of PGSF. (a) Schematic and partial magnified view of the sensor testing system; (b) Sensing performance with different contents of
MWCNT (Multi-Walled Carbon Nanotubes); (c) The changes in ΔI/I0 with pressure variation; (d) Ohmic properties of the sensor under different pressures; (e) The
changes in ΔI/I0 of the sensor at different pressure; (f) The changes in ΔI/I0 of the sensor at 25 kPa and under different frequencies; (g) The changes in ΔI/I0 under
repeated loading with amplitude of 50Pa; (h) Response and recovery time of the sensor.

results show that even after more than 8000 loading and unloading temperature conditions (20 ◦ C–60 ◦ C) (Fig. 5e).
cycles, the resistance of the sensor doesn’t change significantly. This To satisfy the diverse needs of different application scenarios, we
result can be seen in Fig. 5a, which shows the relationship between have developed the HSF and WDRF by adjusting the gradient distribu­
resistance and the number of loading and unloading cycles. This in­ tion. The sensitivities of these sensors are illustrated in Fig. S7. Addi­
dicates that the sensor has excellent long-term stability and can maintain tionally, we have compared our gradient piezoresistance foam sensors
its performance and accuracy under continuous use and stress with those from other studies, as shown in Fig. 5g and Fig. S8. The HSF
environments. offers higher sensitivity within a wider measurement range (0–32 kPa:
The comfort of smart sensors in direct contact with human skin is 37.937 kPa− 1 and 32–150 kPa: 7.047 kPa− 1). On the other hand, our
primarily determined by their breathability and sweat absorption. Our WDRF significantly improves the detection range (0–5 MPa) and out­
developed PGSF, featuring an all-breathable design, slightly lags behind performs other pressure sensors in detecting low pressures (0–300 kPa:
traditional cotton and linen in breathability but significantly out­ 0.068 kPa− 1). The PGSF combines high sensitivity at low pressures with
performs conventional resistive pressure sensors based on silicone rub­ a wide detection range. The gradient distribution method allows for the
ber and PDMS substrates (Fig. 5c). Breathability is assessed by customization of foam sensors with different ranges and sensitivities
measuring air permeability time in a vacuum environment, as illustrated according to specific needs. This enhances the adaptability of the sen­
in Fig. 5b. The PGSF’s design optimizes gas and moisture exchange be­ sors, leading to improved application outcomes.
tween the skin and the environment, aiding in maintaining thermal
balance of the skin.
In addition to its excellent breathability, the PGSF also demonstrates 3.3. Applications
good sweat absorption, evaluated by measuring the water contact angle.
The surface of the PGSF is modified with a PDA layer, introducing polar To explore the potential of the PGSF in monitoring physiological
groups like catechol and amine, enhancing its hydrophilicity. Thus, signals, we prepare a simple wearable sensor by attaching a conductive
although the hydrophilicity of PGSF is slightly lower than that of the tape with an array of micropore to the PGSF to extract the electric signal.
original cotton fabric, it still exhibits better water and sweat absorption By wearing the sensor on different parts of the body, various physio­
than conventional silicone rubber and PDMS substrates (Fig. 5d). logical signals can be obtained. As shown in Fig. 6a, the sensor is
These characteristics make in terms of stability, we measure the mounted above the carotid artery to detect the physiological parameter
resistance changes of PGSF at different humidity levels (RH = 30 % - RH of heart rate. As shown in Fig. 6b, the PGSF is attached to the forehead
= 99 %). The results show that the sensor exhibits excellent resistance and cheek to recognize facial expressions such as smiles and frowns,
stability within a certain humidity range（Fig. 5f and Fig. S6）. Simi­ which are crucial for the prevention of early Parkinson’s disease.
larly, the sensor demonstrates good stability under different Detection of joint movements is equally important in sports exer­
cises. We can detect finger bending and determine the amplitude of joint

6
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Fig. 5. Study of the stability and comfort of PGSF. (a) Schematic of the PGSF under 50 kPa for up to 8000 cycles; (b) Experimental schematic for measuring
breathability of the PGSF; (c) Comparison of breathability between several common electronic skins, commercial fabrics, and the PGSF; (d) Comparison of hy­
drophilicity between several common electronic skins, commercial fabrics, and the PGSF; (e) The ΔI/I0 of the PGSF at different temperatures; (f) The ΔI/I0 of the
PGSF at different humidity levels; (g) Comparison of the present sensor with various other sensors with porous structures.

vibrations by the relative current change output of the sensor, which can The pressure generated by a water-filled beaker and an empty beaker on
also serve as an early prevention indicator of Parkinson’s disease. The the array is demonstrated through a current change 3D histogram, as
relative current variation of the PGSF attached to the finger joint as the shown in Fig. 6f. Similarly, when we touched the E-skin array with our
finger flexes is shown in Fig. 6c. To evaluate the performance of the fingers, the array accurately detected the pressure distribution,
sensor in human voice tests, we conduct voice tests for phrases such as demonstrating that the E-skin array can precisely differentiate the
"Hello" and "Wenzhou University." The results in Fig. 6d demonstrate magnitude of force applied by each finger (Fig. S9). These results verify
that the vibration excited by the pronunciation of these two phrase can the excellent pressure-sensing capabilities of the electronic skin array.
be sensed by the PGSF. The human signal detection test results in Fig. 6 We also demonstrate the excellent detection capability of the PGSF by
show the tremendous potential of the sensor in human behavior attaching it to the root of a music box to detect sound (Fig. S10 and
detection. movie 2). These experimental results indicate that the PGSF pressure
To illustrate the benefit of the PGSF’s extensive detection range, gait array has broad potential in practical applications, exhibiting excellent
detection is conducted. The importance of the sensors in gait detection performance in highly sensitive mechanical sensing and sound detec­
cannot be overlooked. They can provide accurate and objective gait tion, among other fields.
data, analyze gait features in multiple dimensions, provide real-time Supplementary video related to this article can be found at http
feedback and personalized rehabilitation guidance, as well as support s://doi.org/10.1016/j.compscitech.2024.110628
long-term monitoring and trend analysis. With sensor technology, we
can better understand and evaluate the characteristics and changes in 4. Conclusion
human gait, providing stronger support for motion analysis, rehabili­
tation treatment, and disease management. In Fig. 6e and Movie 1, the In summary, we successfully develop a PGSF to simultaneously
PGSF is placed on the front part of the sole for gait detection, allowing achieve high sensitivity and wide detection range. The PGSF adopts a
for the detection of current changes during walking, running, and graded gradient hybrid structure and is realized by the method of using
jumping. salt as a sacrificial template. Testing results show the minimum detect­
Supplementary video related to this article can be found at http able limit of the PGSF is 50 Pa, with sensitivities of 0.202 kPa⁻1 in the
s://doi.org/10.1016/j.compscitech.2024.110628 range of 0–40 kPa, 0.067 kPa⁻1 in the range of 40–210 kPa and 0.020
An important function required in the human-machine interface is kPa⁻1 in the range of 210–850 kPa. For applications, a human motion
accurately mapping mechanical stimuli. To achieve pressure distribu­ and health monitoring sensor based on the PGSF is developed and tested.
tion detection, we designed a 5 × 5 pressure sensing array base on PGSF. Results show that the sensor can realize plantar pressure detection in

7
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Fig. 6. Demonstration of PGSF applications. (a) Sensor used for pulse detection; (b) The sensor used for detecting smiling and frowning; (c) Sensor used for detecting
finger extension and bending; (d) Sensor for voice recognition; (e) Sensor used for gait detection; (f) Pressure mapping array constructed with the PGSF.

which the pressure is up to 850 kPa and human carotid pulse detection Declaration of competing interest
in which the pressure is down to 200 Pa. Furthermore, the breathability
and sweat absorption of the PGSF are improved by preparing breathable The authors declare that they have no known competing financial
electrodes and adding PDA-modified PDMS foam, respectively. This interests or personal relationships that could have appeared to influence
graded hybrid foam structure has a broad application in interactive the work reported in this paper.
wearable devices, physical activity monitoring platforms, health moni­
toring, and other fields. Data availability

CRediT authorship contribution statement Data will be made available on request.

Long Chen: Writing – review & editing, Writing – original draft, Acknowledgment
Software, Methodology, Data curation, Conceptualization. Han Ye:
Validation, Software, Methodology, Data curation. Ping Yu: Writing – This work is supported in part by the National Natural Science
review & editing, Writing – original draft, Validation, Resources, Foundation of China under Grant 51905383, in part by the Natural
Methodology, Data curation, Conceptualization. Science Foundation of Zhejiang Provincial under Grant LGF21E050001,
in part by the Master’s Innovation Foundation of Wenzhou University
3162024003074.

8
L. Chen et al. Composites Science and Technology 253 (2024) 110628

Appendix A. Supplementary data [21] L. Li, X. Bao, J. Meng, C. Zhang, T. Liu, Sponge-hosting polyaniline array
microstructures for piezoresistive sensors with a wide detection range and high
sensitivity, ACS Appl. Mater. Interfaces 14 (26) (2022) 30228–30235.
Supplementary data to this article can be found online at https://doi. [22] C. Liu, Q. Tan, L. Kong, Y. Deng, X. Ma, L. Xu, Y. Xu, W. Zhou, L. Zhong, Q. Qiang,
org/10.1016/j.compscitech.2024.110628. T. Han, W. Chen, B. Liu, A low-cost 3D flexible pressure sensor with a wide working
range prepared from articles for daily use for monitoring human motion, Mater.
Lett. 330 (2023) 133331.
References [23] H. Liu, T. Xu, C. Cai, K. Liu, W. Liu, M. Zhang, H. Du, C. Si, K. Zhang,
Multifunctional superelastic, superhydrophilic, and ultralight nanocellulose-based
[1] W.-T. Guo, X.-G. Tang, Z. Tang, Q.-J. Sun, Recent advances in polymer composites composite carbon aerogels for compressive supercapacitor and strain sensor, Adv.
for flexible pressure sensors, Polymers 15 (9) (2023) 2176. Funct. Mater. 32 (26) (2022) 211308.
[2] J. Hu, G. Dun, X. Geng, J. Chen, X. Wu, T.-L. Ren, Recent progress in flexible micro- [24] Y. Liu, B. Wu, Q. Zhang, Y. Li, P. Gong, J. Yang, C.B. Park, G. Li, Micro/nano-
pressure sensors for wearable health monitoring, Nanoscale Adv. 5 (12) (2023) structure skeleton assembled with graphene for highly sensitive and flexible
3131–3145. wearable sensor, Compos. Appl. Sci. Manuf. 165 (2023) 107357.
[3] X. Guo, C. Jin, X. Li, A. Zhang, W. Hong, Z. Yan, X. Zhang, S. Yu, S. Yu, H. Zhang, [25] B. Wang, X. Yin, R. Cheng, J. Li, G. Ying, K. Chen, Compressible, superelastic and
Electromechanical performance of flexible sensor based on 3D printed biomimetic fatigue resistant carbon nanofiber aerogels derived from bacterial cellulose for
all-PDMS-coated multistage structure: toward kinesiology monitoring and human- multifunctional piezoresistive sensors, Carbon 199 (2022) 318–328.
machine interaction, Sensor Actuator Phys. (2024) 115193. [26] Y.J. Fan, X.S. Meng, H.Y. Li, S.Y. Kuang, L. Zhang, Y. Wu, Z.L. Wang, G. Zhu,
[4] X. Guo, T. Zhang, Z. Wang, H. Zhang, Z. Yan, X. Li, W. Hong, A. Zhang, Z. Qian, Stretchable porous carbon nanotube-elastomer hybrid nanocomposite for
X. Zhang, Tactile corpuscle-inspired piezoresistive sensors based on (3- harvesting mechanical energy, Adv. Mater. 29 (2) (2017) 1603115.
aminopropyl) triethoxysilane-enhanced CNPs/carboxylated MWCNTs/cellulosic [27] X. Liu, Z. Huang, C. Ye, Z. Luo, L. Chen, X. Yao, F. Liang, T. Yang, H. Bi, C. Wang,
fiber composites for textile electronics, J. Colloid Interface Sci. 660 (2024) C. Cai, L. Lyu, X. Wu, Graphene-based hydrogel strain sensors with excellent
203–214. breathability for motion detection and communication, Macromol. Mater. Eng. 307
[5] J. Yan, Y. Ma, G. Jia, S. Zhao, Y. Yue, F. Cheng, C. Zhang, M. Cao, Y. Xiong, P. Shen, (8) (2022) 2200001.
Y. Gao, Bionic MXene based hybrid film design for an ultrasensitive piezoresistive [28] J. Xu, Z. Wang, X. Wang, Y. Wu, R. Xing, T. Yu, Y. Li, J. Ao, Y. Tao, B. Bai, M.D.
pressure sensor, Chem. Eng. J. 431 (2022) 133458. D. Dickey, D. Zhang, J. Yang, Breathable encapsulated liquid metal foam-based soft
[6] K. Meng, X. Xiao, W. Wei, G. Chen, A. Nashalian, S. Shen, J. Chen, Wearable stress sensor, Adv. Mater. Technol. 8 (6) (2023) 2201193.
pressure sensors for pulse wave monitoring, Adv. Mater. 34 (21) (2022) 2109357. [29] Z. Li, M. Zhong, L. Zang, H. Ye, Dual-Mode metamaterial absorber for independent
[7] M. Hou, W. Hong, H. Chen, T. Zhang, X. Zhu, J. Hao, Z. Yin, Z. Yan, X. Li, A. Zhang, sweat and temperature sensing, J. Electron. Mater. 52 (6) (2023) 4106–4116.
Z. Lin, Y. Zhao, D. Kong, J. Ding, X. Guo, Simple fabrication of high-sensitivity [30] F. Lopresti, B. Patella, V. Divita, C. Zanca, L. Botta, N. Radacsi, A. O’Riordan,
capacitive tactile sensor based on a polydimethylsiloxane dielectric layer using a G. Aiello, M. Kersaudy-Kerhoas, R. Inguanta, V. La Carrubba, Green and integrated
biomimetic gray kangaroo leg structure, J. Phys. Appl. Phys. 57 (21) (2024) wearable electrochemical sensor for chloride detection in sweat, Sensors 22 (21)
215401. (2022) 8223.
[8] W. Deng, Y. Zhou, A. Libanori, G. Chen, W. Yang, J. Chen, Piezoelectric [31] N. Kang, S. Zhang, F. Tang, J. Wang, L. Li, Silver-Hydrogel/PDMS film with high
nanogenerators for personalized healthcare, Chem. Soc. Rev. 51 (9) (2022) mechanical strength for anti-interference strain sensor, Colloids Surf. A
3380–3435. Physicochem. Eng. Asp. 654 (2022) 130071.
[9] Z. Yi, Z. Liu, W. Li, T. Ruan, X. Chen, J. Liu, B. Yang, W. Zhang, Piezoelectric [32] J. Kim, J. Lee, B. Choi, Fabrication and characterization of strain gauge integrated
dynamics of arterial pulse for wearable continuous blood pressure monitoring, polymeric diaphragm pressure sensors, Int. J. Precis. Eng. Manuf. 14 (11) (2013)
Adv. Mater. 34 (16) (2022) 2110291. 2003–2008.
[10] T. Zheng, G. Li, L. Zhang, Y. Lei, Breathable, transparent, waterproof, flexible and [33] V. Kumar, G. LeeMonika, J. Choi, D.-J. Lee, Studies on composites based on HTV
high-output triboelectric nanogenerators for sport monitoring and speech and RTV silicone rubber and carbon nanotubes for sensors and actuators, Polymer
recognition, J. Mater. Chem. A 12 (1) (2023) 495–508. 190 (2020) 122221.
[11] T. Zheng, G. Li, L. Zhang, W. Sun, X. Pan, T. Chen, Y. Wang, Y. Zhou, J. Tian, [34] Y. Zhao, Y. Yuan, W. Gan, M. Yang, Optical fiber Fabry-Perot humidity sensor
Y. Yang, A waterproof, breathable nitrocellulose-based triboelectric nanogenerator based on polyimide membrane: sensitivity and adsorption kinetics, Sensor Actuator
for human-machine interaction, Nano Energy 114 (2023) 108649. Phys. 281 (2018) 48–54.
[12] W. Jin, Z. Yu, G. Hu, H. Zhang, F. Huang, J. Gu, Effects of three-dimensional [35] E. Lamers, T.H.S. van Kempen, F.P.T. Baaijens, G.W.M. Peters, C.W.J. Oomens,
circular truncated cone microstructures on the performance of flexible pressure Large amplitude oscillatory shear properties of human skin, J. Mech. Behav.
sensors, Materials 15 (13) (2022) 4708. Biomed. Mater. 28 (2013) 462–470.
[13] X. Lei, D. Guo, L. Ye, L. Xu, S. Yu, L. Tong, Y. Zhao, Y. Li, L. Yi, F. Liu, Highly [36] L. X, X. Meng, Y. Shi, H. Tang, W. Bao, Multilayer microstructured high-sensitive
sensitive, ultrawide range, and multimodal flexible pressure sensor with ultrawide-range flexible pressure sensor with modulus gradient, IEEE Trans.
hierarchical microstructure prepared by a one-step laser printing process, Adv. Electron. Dev. 69 (4) (2022) 2030–2037.
Mater. Technol. 8 (13) (2023) 2202065. [37] Z. Wang, A.A Volinsky, N.D Gallant, Crosslinking effect on polydimethylsiloxane
[14] R. Wang, R. Jiao, Y. Li, Y. Li, Y. Wang, Q. Xu, Z. Li, H. Yu, Miura-ori elastic modulus measured by custom-built compression instrument, J. Appl. Polym
microstructure-enhanced flexible piezoresistive pressure sensor for human 131 (22) (2014).
physiological signal monitoring, IEEE Sensor. Lett. 7 (9) (2023). [38] V. Karageorgiou, D.J.B. Kaplan, Porosity of 3D biomaterial scaffolds and
[15] B. Wei, G. Chen, Q. Wang, A high-performance flexible piezoresistive sensor based osteogenesis, Biomaterials 26 (27) (2005) 5474–5491.
on a nanocellulose/carbon-nanotube/polyvinyl-alcohol composite with a wrinkled [39] H. Lee, S.M. Dellatore, W.M. Miller, P.B. Messersmith, Mussel-inspired surface
microstructure, IEEE Sensor. J. 22 (16) (2022) 15834–15843. chemistry for multifunctional coatings, Science 318 (5849) (2007) 426–430.
[16] S. Chun, Y. Choi, D.I. Suh, G.Y. Bae, S. Hyun, W. Park, A tactile sensor using single [40] M.N. Biutty, S.I. Yoo, Enhanced performance of triboelectric nanogenerator by
layer graphene for surface texture recognition, Nanoscale 9 (29) (2017) controlled pore size in polydimethylsiloxane composites with Au nanoparticles,
10248–10255. Macromol. Res. 29 (1) (2021) 98–104.
[17] Y. Wan, Z. Qiu, Y. Hong, Y. Wang, J. Zhang, Q. Liu, Z. Wu, C.F. Guo, A highly [41] F.-L. Gao, P. Min, X.-Z. Gao, C. Li, T. Zhang, Z.-Z. Yu, X. Li, Integrated temperature
sensitive flexible capacitive tactile sensor with sparse and high-aspect-ratio and pressure dual-mode sensors based on elastic PDMS foams decorated with
microstructures, Adv. Electr. Mater. 4 (4) (2018). thermoelectric PEDOT:PSS and carbon nanotubes for human energy harvesting and
[18] X. Lü, T. Yu, F. Meng, W. Bao, Wide-range and high-stability flexible conductive electronic-skin, J. Mater. Chem. A 10 (35) (2022) 18256–18266.
graphene/thermoplastic polyurethane foam for piezoresistive sensor applications, [42] J. Jiang, L. Zhu, L. Zhu, B. Zhu, Y. Xu, Surface characteristics of a self-polymerized
Adv. Mater. Technol. 6 (10) (2021) 2100248. dopamine coating deposited on hydrophobic polymer films, Langmuir 27 (23)
[19] D. Kannichankandy, P.M. Pataniya, S. Narayan, V. Patel, C.K. Sumesh, K.D. Patel, (2011) 14180–14187.
G.K. Solanki, V.M. Pathak, Flexible piezo-resistive pressure sensor based on [43] C. Yu, J. Lu, J. Dai, Z. Dong, X. Lin, W. Xing, Y. Wu, Z. Ma, Bio-inspired fabrication
conducting PANI on paper substrate, Synth. Met. 273 (2021) 116697. of Ester-functionalized imprinted composite membrane for rapid and high-efficient
[20] S. Kim, M. Amjadi, T.-I. Lee, Y. Jeong, D. Kwon, M.S. Kim, K. Kim, T.-S. Kim, Y. recovery of lithium ion from seawater, J. Colloid Interface Sci. 572 (2020)
S. Oh, I. Park, Wearable, ultrawide-range, and bending-insensitive pressure sensor 340–353.
based on carbon nanotube network-coated porous elastomer sponges for human [44] Y. Wang, H. Mao, Y. Wang, P. Zhu, C. Liu, Y. Deng, 3D geometrically structured
interface and healthcare devices, ACS Appl. Mater. Interfaces 11 (26) (2019) PANI/CNT-decorated polydimethylsiloxane active pressure and temperature dual-
23639–23648. parameter sensors for man–machine interaction applications, J. Mater. Chem. A 8
(30) (2020) 15167–15176.