Recent Trends of Functional Composites and Structures for Electromechanical Sensors: A Review

With the rapid advancement of modern technology, flexible and wearable electronics, particularly electromechanical sensors, have gained considerable attention for motion‐based applications in medical health monitoring and artificial intelligence. Correspondingly, extensive efforts have been dedicated to enhancing their performance and practicality. Electromechanical sensors based on functional composites and structures accurately detect and monitor the body, generating the corresponding output signals. However, there are several limitations in manufacturing composite sensors, such as the selection and combination of functional materials, geometrical structures, constructed conductive pathways, stability of output signals, and operational lifetime. Thus, this review summarizes notable trends in the electromechanical sensors using functional composites and structures. This also provides an overview of different types of electromechanical pressure and strain sensors, exploring their operational mechanisms regarding triboelectricity, piezoelectricity, piezocapacitance, and piezoresistivity. The unique characteristics of functional materials, including conductive polymers, nanostructured metals, and carbon nanomaterials and composites, are analyzed alongside various design concepts for highly flexible and stretchable sensors. Furthermore, potential applications concerning human motion and human–machine interfaces are also recommended. Additionally, several future outlooks are reviewed for insights into future prospects and strategies. Thus, this review can assist readers in understanding current electromechanical devices more accurately.

This study provides a comprehensive review of electromechanical devices with triboelectric, piezoelectric, piezoresistive, and piezocapacitive sensors and summarizes recent developments in composite sensors in detail, including their functional structures, active materials, operable sensing mechanisms, and designs.This review also suggests promising applications and highlights their contributions, such as detecting various human motions and temperatures and human-machine interfaces.Additionally, several future prospects and outlooks on developing electromechanical-sensing devices are discussed.Overall, this comprehensive review serves as a valuable resource for readers to acquire a more accurate and effective understanding of the current development in electromechanical devices as a whole.

Main Parameters of Electromechanical Sensors
Several key parameters must be accurately evaluated to effectively use the electromechanical sensors for practical applications, including the sensitivity (gauge factor [GF]), response time, linearity, stretchability, limit of detection (LOD), repeatability, and stability.

Sensitivity or GF
When quantitatively evaluating kinetic parameters, sensitivity plays a crucial role in determining efficiency and accurate sensing ability of the sensors.Sensitivity is typically determined using Equation (1). [6,7]In the equation, S is the sensitivity (kPa À1 ), P denotes the external mechanical deformations (e.g., pressure), and X denotes the output electrical signals.The equation provides the change rate (S) of the output signal (X ) with respect to the imposed stimuli (P).
In addition, another type of P caused by an external mechanical deformation is strain.Sensitivity is a dimensionless unit known as GF, which is important for the direct and effective analysis of sensor signal feedback caused by a target deformation.The slope of the relative change in electrical signals (i.e., capacitance and resistance) with respect to the applied stress or strain reflects the sensitivity (S, kPa À1 ) or GF of the sensors.In addition, reports have been published regarding the construction of conductive networks to achieve segmented linear sensing curves in the entire stress or strain range, which is a significant challenge in the designs of conductive network structures and elastomernanoparticle (NP) interfacial interactions for the formation of the desired entire-scope linear sensing features. [8]These can mainly relate to the rapid propagation of microcracks in electrically conductive pathways of the piezoresistive sensors.For example, piezoresistivity is one of the most extensively utilized transduction mechanisms; it converts mechanical deformations to electrical resistance variations upon an applied pressure or strain.In this case, the GF is defined by Equation (2) or can be derived using a linear fitting (Figure 2A,B), [9,10] where ΔR (Ω) is the resistance change with respect to the difference in the resistance values attained under an applied strain deformation (R, Ω) and initial value (R o , Ω), and ε is the applied strain.Once the classifications and structures of the electromechanical sensors are identified, the desirable ranges of the GFs for the as-designed sensors can be determined accordingly.For the thin film-based strain sensors, the GF can be enhanced by adjusting the modulus distribution of the surface and sublayer materials from a mechanicrelated structure. [11,12]In addition, the sensitivity and the  Response curves of ΔR/R 0 against compressive strain and pressure for composite foams and the schematic of the sensing mechanism (adapted with permission [9] ).(B) Response curves of ΔR/R 0 against different strains for the electromechanical sensors (the curves of linearity and GF values were plotted versus strain) and the schematic of the sensing mechanism.Adapted with permission. [10]tretching range of porous film-based strain sensors can be tuned simply and reliably; however, it is difficult to adjust the stability of the GF according to the signal feedback in a relatively broad stretching range.Generally, it depends on the rapid propagation of microcracks or the penetration of conductive layers corresponding to the direction of electron transfer.However, this approach can create nonlinear GF curves, resulting in the inability to provide effective, accurate, and rapid feedback, which are recommended for postprocessing of signal feedback using specific algorithms to alter it appropriately and accurately.To overcome this challenge, sensor surface morphology control is used.This approach is the most effective for adjusting the sensitivity and linearity of relative resistance change (ΔR/R o ) response to strain (ε), particularly for micro/nanoscale structures on elastomer fiber surfaces that have large specific surface areas.In addition, other potential approaches are also used for in-depth study and investigation to achieve signal responses appropriately and accurately.

Stretchability or LOD
In addition to the sensitivity and GF, stretchability (E) is a crucial parameter of electromechanical wearable sensors.It is based on the elastic modulus (Young's modulus) of the sensors represented by the linear region of the curve described by Equation (3) [6] and determined through tensile or compression tests; σ and ε represent the applied stress and strain, respectively.
Normally, stretchability is based on the maximum strain that a sensor can tolerate without failure in function and is related to its stability in the linear region.A sensor can exhibit high stretchability with low sensitivity (low GF), which indicates an extremely nonlinear behavior. [13]Thus, a sensor with both high sensitivity and stretchability can well detect and monitor both small and high strains (an important consideration for researchers).
Response time is another factor affecting sensors in generating a stable and identifiable output signal under external forces; [13] this is particularly true as it is as short as that probably facilitates better and more acceptable sensor abilities for dynamic real-time monitoring, particularly for human motions.Another critical parameter is the LOD, which represents the measurable region and includes the lowest and highest values of the external stimuli (e.g., applied pressure).For example, in tactile sensors, the typical pressures created from daily activities of the human body are classified into low-pressure (slight touch, <10 kPa) and higher-pressure levels (10À100 kPa). [7]Normally, tactile-type sensors with high LODs are critically important for providing sensitivity and accurate and safe operation of objects.The LODs of current piezoresistive pressure (or tactile) sensors depend significantly on their key active materials and types of designed structures (Table 1).An elastic microstructure film was successfully fabricated from polypyrrole hydrogel through a multiphase reaction for designed conductive material-based piezoresistive sensors.This enabled a sensor to detect low pressures (≈1 Pa) with high sensitivity (133 kPa À1 ). [14]The sensor was based on an Au micropillar array and exhibited a sensitivity range of 0.03À17 kPa À1 at a low pressure of ≈1 kPa and an LOD of 2 Pa. [15]In a study on pressure sensors with porous conductive materials, ultralight graphene-based cellular elastomers could detect ultralow pressures of ≈0.1 Pa with a sensitivity of 10 kPa À1 . [16]In another study, a flexible elastic carbonized melamine-based carbon foam was prepared at 800 °C without using a metal catalyst.It exhibited an extremely high sensitivity (100.3 kPa À1 ) at a large pressure of 3 Pa. [17]Notably, a sensor based on an MXene-sponge showed a high sensitivity of 147 kPa À1 , rapid response time of 138 ms, low LOD (9 Pa), and excellent durability. [18]In addition, SWCNT/PDMS, a sensor based on conductive polymeric composites, achieved an ultrahigh sensitivity with a rapid response time, good repeatability and stability and a low LOD (0.6 Pa). [19]Generally, the materials of electromechanical sensors influenced their stretchability, sensitivity, response time, pressure LOD, and potential applications.

Linearity
In addition to sensitivity and stretchability, linearity is another key parameter that enables accurate signal detection.More specifically, linearity describes the possible relationships between relative changes in electrical signals and applied strains (i.e., a straight line).Thus, it involves the transformation of mechanical stimulus to electrical signal, rapid and linear response, and signal stability.These are typical challenges for piezoresistive pressure sensors, which may be caused by the creep behavior of elastomers used in most pressure sensor systems.Linear response in resistance change without hysteresis can effectively increase the sensing performance of the sensors, resulting in Application Au-deposited polydimethylsiloxane (PDMS) [233] 50.7 20 -Tactile sensing Au micropillar array [15] 17 -2 Subtle loading Poly(pyrrole) [14] 133 50 1 Subtle loading Graphene [16] 10 0.2 0.1 Subtle loading Graphene/PDMS [234] 44.5 --

Human pulse detection
Carbonized melamine [17] 100.3 -3 Wrist pulse detection MXene/Poly(vinyl alcohol) nanowires (NWs) [18] 147 138  9 Human physiology detection Vanadium nitride-graphene [235] 40 130  -Health monitoring Single-walled carbon nanotubes (SWCNTs)/ PDMS [19] 1.8 <10 0.6 Small muscle movement Metal particle/Polyurethane (PU) [236] 2.46 30 -Finger motion high output signal to noise ratio, easy readout, and low power consumption.This challenge can be overcome through the use of porous piezoresistive pressure sensors with strong polymer-filler interaction, broad-linear sensing range, and highly reversible piezoresistivity. [9,10]In addition, piezoresistivity-based strain sensors generally exhibit linearity at low strains and nonlinearity at high strains, thereby facilitating several blocks in information processing.The nonlinearity of these strain sensors is attributed to their nonhomogeneous morphologies over stretching.However, the sensitivity and linearity were significantly enhanced by the constructed hybrid structures.For example, CNT-polymer composites, which are poly(ethylene terephthalate) (PET)-coated CNT-graphite flake hybrid thin films, and nickel (Ni)-coated CNT-epoxy nanocomposites have GFs of 7.8 and 155, respectively, with high linearity and sensitivity at a low stretchability (≈1.5%). [13,20]The electrical signals of strain sensors often originate from changes in the contact resistance caused by the nano/microscale cracks created at the local sites in the active layers [9,10] ; however, the response characteristics depend on the continuous propagation of the crack sites in the active layers.These approaches are complex and difficult to understand; thus, the strain-sensing behavior is difficult to control accurately.Thus, the structural construction of the high-performance sensors is probably suggested to explore designable substrate structures toward the realization of the control of the rational number and propagation of cracks in the active layer under large-stretching deformations.
Another type of electromechanical sensor that uses piezocapacitance has excellent linearity and low sensitivity, which are attributed to the limited linearity with a certain amount of strains, owing to Poisson's ratio revolutions at large strains.It has been reported that piezocapacitive pressure sensors possess linear signal-to-pressure responses (Figure 3A); [21,22] however, their sensitivities are extremely low, which is mainly caused by the incompressibility of flat elastomeric dielectric layers in the sensors.Generally, the use of flexible pressure sensors has contradicting effects on sensitivity and linearity owing to their constructed structures, being a big challenge.Thus, piezocapacitive pressure sensors enhance sensitivity using dielectric layers with mechanically stable microstructures, such as microcones, micropyramids, and microdomes.These structures are generally more compressible than flat dielectric layers; however, these mechanically stable microstructures undergo structural stiffening, reducing sensitivity under large pressure forces.
Generally, strain sensors exhibit "high linearity and stretchability" and "high sensitivity."Sensors with high stretchability frequently mention the integrity of the morphology upon a large strain.Those with high sensitivity require considerable changes in structures under applied strains, whereas those with high linearity require changes in homogeneous morphology under stretching.This inconsistency can substantially support the development and improvement of strain sensors with concomitant high linearity, sensitivity, and stretchability, which is an important challenge for researchers.

Hysteresis, Response Time, and Dynamic Durability
Hysteresis, response time, and dynamic durability (repeatability and cyclic stability) are important parameters for estimating the performance of electromechanical sensors under dynamic loadings.Hysteresis behavior demonstrates the inconsistent sensing performance of the sensors during loading and unloading.In addition, repeatability or cyclic stability is crucial for ensuring effective long-term application of the sensors under long-term stimuli.This parameter indicates the dynamic durability of the sensors against repeated stretchingÀreleasing cycles without inducing failures of both electrical and mechanical functionalities.Another key parameter is response time, which determines how rapidly an electromechanical sensor responds steadily to a stimulus, which is often constructed to evaluate a stable and identifiable output signal upon the external forces acting on the sensors.It refers to the time between the application of a stimuli to the generation of a stable output signal.This implies that a dynamic response is often defined according to respon-seÀrecovery times and is generally presented as a time-related increase to a maximum level or a recovery to the initial condition.Thus, electrical signals are outputted stably during repeated response-recovery cycles, enabling researchers to study the repeatability or stability of the sensors.
Hysteresis behavior, response delay, and cyclic instability are often considered in polymer-based tactile sensors, which may affect the viscoelastic nature of the polymer.Generally, the performance of polymer-based tactile sensors in terms of hysteresis, response time, and dynamic durability is often affected by the characteristics of the polymers, working mechanism, sensor structure, morphologies of the active materials, and potential interactions between the sensing component and polymer matrix. [23]In addition, the use of low-viscoelastic polymer is preferred to achieve stable and reliable performance of the tactile sensors.Moreover, piezocapacitive sensors exhibit lower hysteresis performance and faster response compared with piezoresistive sensors because of the stably overlapped region between the opposite electrodes in piezocapacitive sensors. [24]An Au-NWs/ latex strain sensor exhibited efficient durability with good repeatability for >5000 stretch-release cycles of 30% (0.5 Hz) based on the evaluations of stable signal outputs. [25]An iontronic flexible pressure sensor was fabricated and exhibited high performance in terms of high sensitivity (49.1 kPa À1 ), linear response (R 2 > 0.995), and a wide pressure range (485 kPa).In addition, the sensor exhibited a fast response-recovery speed (<5 ms), which was suitable for detecting vibration with frequencies up to 200 Hz, and exhibited cyclic stability (5000 cycles) under fluctuating and dynamic loads (≈60 kPa) (Figure 3A). [22]The morphology of the sensing material is also important for sensor performance.For example, conductive composites of carbon black and polypropylene exhibit higher stretchability and better distinct cyclic behavior than composites of CNT and polypropylene. [26]Specifically, using carbon black/polypropylene conductive composites could gradually increase the maximal and minimal values of the relative resistance change according to increasing cycle number, whereas using CNTs/polypropylene composites could result in an opposite trend.B) Self-healing test and digital captures of a gelatin-tannic acid-Ag-NWs hydrogel sensor.Adapted with permission. [35]

Other Parameters
In addition to the abovementioned key parameters, wearable sensors exhibit several relevant characteristics (e.g., self-healing and transparency).When electromechanical sensors are damaged, they fail to function normally.However, because of their selfhealing ability, they can be fixed electrically and mechanically, making this an appealing approach to functional restoration.This indicates that self-healing sensors can automatically heal damages by reconstructing reversible chemical bonds or using healable agents present in polymers or composites.[29][30][31] For the extrinsic healing system, healable agents and catalysts are often used in the vascular networks or capsules, that is, vascular-or capsule-based self-healing materials.For instance, a microencapsulated healing agent was synthesized to induce polymerization at the crack faces, [29] resulting in a high recovery efficiency of 75%.However, this trend probably shows the multiple-healing incapability of the materials that may be because of the depleted usage of healing agents in the healed region of the first self-healing process.The intrinsic healing system normally relates to dynamic reversible covalent and noncovalent bonds, dynamic covalent bonds, and reversible noncovalent bonds, which were developed to achieve multiplehealing capability for the intrinsic self-healing materials.Furthermore, the intrinsic self-healing materials are more durable and reliable than the extrinsic self-healing materials.For example, a synthesized biomimetic tactile sensor exhibited self-healing efficiency with repeated mechanical and electrical behaviors at room temperature. [30]This resulted in a conductive composite recovery efficiency of up to 90% within 15 s after a fracture, which was achieved through the reassociation of hydrogen bonds on the cut surfaces and the piezoresistive effect caused by the inversely applied flexion and tactile forces.The self-healing capability of the tactile sensors is achieved through the contact of the fractured interfaces with each other, which enhances the sensors' lifespans and broadens their application scopes.Several investigations have been conducted to determine or identify self-healing abilities for producing advanced sensors.For instance, a self-healing polymer was fabricated using thermally sensitive ionic liquids, producing a sensor with an identical and repeatable thermal sensitivity after fracture and selfhealing. [32]In a previous study, various conductive materials based on graphene, SWCNTs, and silver NWs (Ag-NWs) were added into hybrid composites with self-healing characteristics, enhancing the speed of self-healing (3.2 s) and improving the self-healing efficiency (98%).A GF of 1.51 was obtained at a strain of 1000%. [33]an et al. prepared an electrically conductive hydrogel based on combining a CNT component and polymeric hydrogel.It exhibited a high sensitivity (a GF of 343 at a 110% strain).In addition, the electromechanical performances of the conductive hydrogelbased sensor could be self-repaired through a simple thermal treatment. [34]Wang et al. presented a self-powered strain sensor based on a gelatin-tannic acidÀAg-NWs hydrogel.The sensor exhibited a high electrical healing speed (0.65 s) and self-healing efficiency (95%) (Figure 3B). [35]In a previous study, a 3D porous foam incorporating an Ag-NWs@thiolated graphene conductive network into functionalized PU elastomer could self-heal to 66.7% and 87.0% after 12 and 24 h at room temperature, respectively.As a self-healable strain sensor, it exhibited a high sensitivity (GF = 11.8),good stretchability (60%), fast response, and good stability. [36]Thus, the development of electromechanical sensors with self-healing abilities has been encouraged.However, the design and manufacture of self-healable sensors with efficient features (i.e., sensitivity, stretchability, responsiveness, and stability) remain challenging.
Another desired feature is transparency, which is a major concern in terms of aesthetics, which is an essential requirement for potential applications of wearable electronics.Generally, the addition of inorganic nanostructured conductive fillers can lead to the nontransparency of the sensors when these fillers are highly loaded into the hybrid composites.Thus, sensors mounted on parts of a human body frequently require optical transparency to render them invisible during daily activities for aesthetic purposes.A transparent and flexible humanÀmachine interface was successfully developed using ultrathin and transparent nanomaterials (i.e., Ag-NWs, doped graphene, and piezoelectric polymers).These materials were manufactured on stretchable substrates and a polymeric matrix and exhibited high transparency and deformation. [37]The stretchable sensor integrated a mesh-structured Ag-NW-PDMS composite based on concomitant laser-cutting and drop-coating processes, exhibiting good transparency of 88.3%, high sensitivity (GF of 846 at 150% strain), and long-term stability. [38]In a previous study, a sandwich-like structured strain sensor based on a 99% metallic CNT-PDMS composite thin film achieved optical transparency of >92%.In addition, the linearity and sensitivity of the strain sensor could be well controlled based on the number of CNT sprays. [39]Walia et al. successfully produced a hydrogel-based sensor containing a palladium NW network and achieved a good transmittance of approximately 80% using the crackle lithography technique. [40]Lee et al. manufactured a novel strain sensor based on PU-poly(3,4-ethylene dioxythiophene) (PEDOT):polystyrene sulfonate (PSS)/SWCNT/ PU-PEDOT:PSS as a new stacked piezoresistive nanohybrid film.The film exhibited relatively high transparency of 63%, a GF of 62.3 at a strain of >100%, and good stability. [41][44] For tactile sensors, the applied voltage significantly affects the power consumption of the sensors; thus, the voltage must be considered. [45]Table 1 lists recent piezoresistive pressure (or tactile) sensors in terms of the functional materials and desired key parameters.Overall, the electromechanical sensors must have the desired features and key parameters to achieve accurate and promising performances.

Typical Classifications and Mechanisms of Electromechanical Sensors
Electromechanical sensors show outstanding advantages, such as good detection of external stimuli, clear output electrical signals, lightweight, easy to mount on garments or the human body, wearability, and comfortability.These sensors are generally classified depending on the mechanical stimulations or sensing mechanisms.Specifically, the mechanical stimulations produced by the mechanical movements of a human body or robot can be classified into two main types: tensile and pressure forces.Strain and pressure sensors are designed to detect and convert mechanical forces into electrical signals through four transduction or sensing mechanisms: triboelectricity, piezoelectricity, piezocapacitance, and piezoresistivity.Their operation mechanisms, advantages, and disadvantages are described in Figure 4 and Table 2.The details of the four types of electromechanical sensors are discussed herein.

Triboelectric Sensors
The triboelectric effect is generally created under mechanical deformations (e.g., touches, torsions, and sliding motions) to produce electricity.It is considered an electrification process when functional materials contact each other through electrostatic induction.Although it is a common phenomenon, the triboelectric effect-based transduction mechanisms of the corresponding sensors (i.e., triboelectric sensors) are still being investigated.Typically, electrical charges are generated on the surface of dissimilar active materials under the influence of friction; the number of electrical charges is highly dependent on the variation in the triboelectric polarities among the contacting materials. [46]In triboelectric sensors, the triboelectric charges are created when various materials come into contact with each other.Among these functional materials, a charge transfer occurs from a surface with lower electron affinity to one with higher affinity.Under an open-circuit-potential condition, these dissimilar active materials with triboelectric polarities come in contact with applied external forces, generating the triboelectric effect on both sides of the surfaces with opposite charges. [47]ypically, the surfaces with opposite charges automatically disconnect from each other when the forces are released.Subsequently, offsetting charges are created at each side of the top surface.This indicates that the charge on the two surfaces cannot neutralize completely, causing a potential difference (possibly owing to the air layers in the materials).The mechanism in triboelectric sensors is shown in Figure 4A.The transduction mechanism allows triboelectricity-type sensors to output electrical signals based on diverse external stimuli; thus, it is often utilized in self-powered electromechanical sensors.
Triboelectric nanogenerators (TENGs) are examples of an emerging approach to developing self-powered sensors using TENG-based triboelectric sensors for monitoring in human health, environment, industrial production, and daily life (Figure 5).Their working mechanism is based on the combined effects of triboelectrification and electrostatic induction.TENGbased electromechanical sensors can detect and monitor certain conditions of the human body and physiology in real time.In a previous study, a humidity-resistant TENG was successfully fabricated by incorporating a nanostructured chitosan-glycerol film, which was used as an insole for collecting energy from human activities. [48]Choi et al. proposed a stretchable multifunctional TENG, [49] which had a low power consumption owing to its sleeping microcontroller unit, thereby maintaining the standby power of the sensor at a lower level.Sun et al. [50] fabricated a highly stretchable and transparent ion gel-based TENG with a wide temperature tolerance range of -20-100 °C, which could be directly fixed on human skin for detecting and monitoring wrist motions.Regarding environmental monitoring applications, TENG-based intelligent sensors can detect humidity, rainfall intensity, and wind speed in real time.This approach can effectively replace human observations and provide convenience to people.In addition, TENG-based sensors and smart systems used in industries can contribute to the development of smart factories and the detection of toxic gases, thereby protecting human life and property.TENG-based sensors and systems can contribute significantly to building intelligent homes and places.Currently, self-powered wireless sensors provide extensive convenience to people.Despite the advantages of TENG-based triboelectric sensors in human health, environment, industrial production, and daily life, they still have several important challenges (Figure 6) in output performance, durability, power management, and sensitivity.For the output performance of the TENG-based power sources, advanced friction materials and corresponding structural design are considered promising approaches to reach high output performance.Simultaneously, optimization in the combination of TENG and the external environment is required for better energy collection efficiency.Long-term repeated friction in TENG-based power sources can decrease the performance and life of TENGs, owing to their working mechanisms.Thus, the preparation and design processes should consider extending the service life of electromechanical sensors without affecting their output performance.Moreover, generating more compatible power management circuits for TENGs should be regarded because a power management circuit is critical for providing a continuous and stable power supply.TENGs are often combined with other nanogenerators and used as active hybrids to harvest higher mechanical energy to increase efficiency.Excellent sensitivity is always a challenge for TENGs that are used as smart electromechanical sensors.Thus, the abovementioned requirements can be applied to output performance.In addition, new structures can be utilized to enhance sensitivity.For quantitative analysis of the sensors, stability is one of the most important performance characteristics of electromechanical sensors.Several environmental conditions, such as humidity, can significantly impact TENGs.In addition, ideal designs can maintain environmental stability.

Piezoelectric Sensors
The use of piezoelectricity is another transduction approach used in electromechanical sensors.Piezoelectricity is an intrinsic characteristic of ferroelectric materials under compression, tensile, or bending deformations in terms of voltage.These deformations induce a reorientation of the electric dipole moments inside a medium and then create an electrical charge on the crystal surfaces of the material corresponding to the applied mechanical deformation, [51,52] as illustrated in Figure 4B.The long-term use of mechanical deformations can cause changes in the thickness and the dipole population inside the medium.Subsequently, a current of electrical charges is produced to offset this constructed imbalance.Thus, the outcomes correspond to the multimodal sensitivity and reversibility of the transduction mechanism of the piezoelectric sensors.For example, piezoelectric polymers, inorganics, and composites [53,54] are highly regarded for their ability to produce flexible piezoelectric sensors with high piezoelectric coefficients.Specifically, the piezoelectric coefficient is used as a physical evaluation measure to demonstrate the affiant capacity of the energy conversion of the piezoelectric materials.Piezoelectric sensors have been applied successfully for detecting different dynamic pressures of slip-and sound-based stimuli, owing to their high sensitivity and fast response. [55]In terms of energy harvesting features, piezoelectric materials are promising candidates for developing low-power-consumption or self-powered electromechanical sensors.Generally, piezoelectric materials can be deformed and rapidly change polarization density according to applied mechanical stimulations, which generate voltages in the polar direction and vice versa.Thus, the mechanical properties of piezoelectric materials play an important role in increasing their piezoelectric responses to ensure adequate stress transfer during mechanical stimuli.

Piezocapacitive Sensors
Piezocapacitive sensors use the concepts and general principles of a capacitor and capacitance.A capacitor consists of a dielectric material sandwiched between two electrodes (Figure 4C), whereas capacitance (C) refers to the ability of a capacitor to store electrical charges, as defined in Equation (4).ε o and ε r represent the free space and relative static permittivity of the dielectric layer in the plates, respectively, and A and d are the overlapping area and distance of the two plates, [55] respectively.
In this equation, ε o is always constant, whereas ε r , A, and d are sensitive to changes in the applied mechanical deformation.Changes in ε r can be used to detect forces using special functional materials; however, this approach has not yet been intensively studied.Hence, the changes in capacitance and corresponding changes in d [56,57] caused by an applied deformation generally exhibit good linearity but low sensitivity, particularly for dielectrics with high Young's modulus values.Tactile-type piezocapacitive sensors have demonstrated low power consumption, high sensitivity, and good compatibility. [55,58]owever, their sensitivity and response speed are normally restricted by the viscoelasticity and incompressibility of the rubbers used in elastomeric dielectric-integrated sensors. [56]To overcome this limitation, an air gap is normally used to enhance the performance of piezocapacitive sensors in highly compressible dielectrics. [59]In addition, these sensors can attain high sensitivity when a small dielectric thickness is used, resulting in high capacitance. [56]espite the advantages of piezocapacitive pressure sensors, such as simple configuration, low power consumption, and good working stability, they have low sensitivity values (<10 kPa À1 ) and sensing ranges (<1 MPa).The sensitivity and sensing capacity of piezocapacitive pressure sensors must be increased to improve their performance and broaden their applications.Thus, some suitable strategies are suggested: 1) microstructures should be constructed on electrodes or dielectrics; 2) composite dielectrics should be used; and 3) physical holes should be inside the dielectric layer.Recently, iontronic sensors containing two flexible electrodes sandwiching an ionic gel layer [21,22] have been developed and used as piezocapacitive pressure sensors.In this approach, an electric double layer acts as a capacitor formed at the electrode and ionic gel interface under a small voltage.Based on Equation (4), the electric double layer-based capacitance can be determined using the interfacial contact area (i.e., C-A), which indicates a noticeable difference compared with the sensing mechanism of the abovementioned piezocapacitive pressure sensors.The iontronic sensors exhibit a nanointerfacial behavior.The ionic and electron charges generated from the ionic gel and electrode, respectively, are separated at an atomic length scale of ≈1 nm, which generally results in ultrahigh signal intensity and high sensing performance compared with common capacitors.In addition, iontronic sensors exhibit nonlinear responses determined by the constructed microstructures; hence, ideal designs of high-performance iontronic capacitive sensors should be used to achieve high sensitivity and linear response under a wide pressure range.

Piezoresistive Sensors
Another transduction method is the use of piezoresistivity.It is the most extensively utilized transduction mechanism for electromechanical sensors owing to its simple designed structures, low energy consumption, simple read-out mechanisms, and wide detection ranges, [60] which effectively convert external mechanical stimuli into electrical resistance signals under applied pressure or strain (Figure 4D).Its typical principle is based on Equation (5), [6] a well-known resistance equation (R, Ω).In Equation (5), ρ (Ω•m), L (m), and A (m 2 ) are the resistive conductivity, length, and crosssectional area, respectively.The changes in resistance are highly dependent on the geometrical structures and resistivity of the sensors, [61] whereas the relative resistance varies according to the applied stress or strain, that is, as a slope indicating the sensitivity (S, kPa À1 ) or GF [Equation (2)] of the piezoresistive strain sensors.In strain-type tactile sensors, the change in resistance can be determined using Equation (6), where ε and υ are the strain and Poisson's ratio, respectively.Additionally, Equation (7) describes the relationship between the change in contact resistance (R c ) [62] and applied force (F); it results in high sensitivity and operating range for the piezoresistive sensors.This is attributed to the changes in the geometrical structures and contact areas between the functional materials.
Once a piezoresistive sensor is impacted by a corresponding force, any or all of the parameters can change under the deformation, inducing relative changes in the resistance.This indicates that the response significantly depends on the designed geometrical structure of the sensor.Piezoresistive devices can be divided into strain-and pressure-type sensors (Figure 4E), which are based on the type of externally applied forces, such as tensile and compression deformations.The sensors are also classified into three types based on the compositional and structural characteristics of these sensors: 1) conductive polymeric composite-based sensors, in which conductive fillers and a viscoelastic polymer matrix are incorporated; 2) porous conductive material-based sensors, which contain 3D interconnected porous-structured conductive materials; and 3) architected conductive material-based sensors, which have geometry-designed conductive material systems (Figure 4E).The sensitivity or GF values depend significantly on advanced materials or structures constructed for the as-designed piezoresistive sensors.In addition, the GFs of piezoresistive electromechanical sensors are mainly based on several mechanisms related to the structural and compositional characteristics, including the disconnection between conductive pathways and crack propagation in conductive networks.Generally, diverse transduction/sensing mechanisms, active materials, and geometrical structures are selectively combined to enhance sensitivity consistently.Thus, developing a proper electromechanical sensor system for potential applications remains a problem.The selection of functional materials, design ideas for electromechanical sensors, and recent applications of these sensors are discussed in detail in the succeeding sections.

Selection of Functional Materials
Recently, advanced technologies have been developed for smart devices; thus, researchers have explored and used novel advanced materials and structures to construct highly stretchable and sensitive sensors.Thus, the properties of the material-dependent systems affecting the sensing mechanisms of the sensors are important to appropriately improve the desired features (e.g., low cost, lightweight, flexibility, feasibility, and sensing efficiency of the final product).Traditional conductive materials, such as inorganic materials, may not be appropriate for significant mechanical behaviors owing to their rigidity.Thus, exploring strategies for effective electromechanical sensor designs is necessary.Recent advances in stretchable and flexible materials may allow for diverse manufacturing of electromechanical sensors, integrating various functional materials instead of using individual materials, such as in functional composites and structure-based sensors.This implies that functional composites combining both favorable electrical and excellent mechanical features should be employed to achieve large-scale production of high-performance and stretchable sensors.Examples of potential materials for manufacturing electromechanical sensors developed recently, including polymeric substrates and conductive/electrode materials, are presented in this section.Details of the chemical structures, advantages, and disadvantages of common functional materials for constructing electromechanical sensors are described in Figure 7 and Table 3.

Stretchable Matrix or Polymeric Substrates
The substrate-supporting materials are integral parts of flexible and stretchable sensors.They do not contribute directly to the sensing but significantly support the flexibility and wearability of the sensors.Among them, glass, ceramics, and silicon have been used as traditional substrates; however, these supporting materials are generally brittle.This limits the practical applications of the sensors, particularly for tactile-type electronics.Thus, highly stretchable or flexible novel and potential materials with corrosion resistance are required for polymeric substrates and stretchable matrices.A mechanical property of conductive composites is tensile strength.The values of their Young's modulus are based on their elastomer matrices.Elastomers are used as stretchable matrices or substrates for combination with various conductive nanofillers to achieve a desirable stretchability of the composite materials and good stability of the conductive networks during large deformations.Thus, elastomers are generally classified into thermosets (i.e., PDMS and Ecoflex), thermoplastics (i.e., PU and block copolymer elastomers), and gels.Their properties are attributed to the various inter-and intramolecular interactions, including chemical (as thermosetting elastomers), physical (as thermoplastic elastomers), and ionic (as gels) cross-linking.These interactions confer ultrahigh elasticity to the elastomers, making them ideal candidates for applications in stretchable electronics. [63]o date, common polymeric substrates, such as silicone elastomers, have been extensively employed.For example, Ecoflex elastomers (chemical structure shown in Figure 7) are extremely tough and can resist up to 900% strain. [64]They are also ecofriendly and nontoxic to the human body.Thus, they are appropriate candidates for fabricating array-type tactile sensors for detecting drowsiness by incorporating a molybdenum disulfide (MoS 2 )/graphene foam. [65]In addition, Ecoflex layers are also used as friction surfaces for contact with new single-frictionsurface TENGs.A TENG was incorporated into a CNT-poly(pyrrole) mixture (conductive and antibacterial materials) and cotton textile (supporting material) and used as a wearable triboelectric generator to collect energy from finger contacts on the sensor. [66] sensor based on magnetic reduced graphene oxide (rGO)  @nickel NWs (magnetic rGO@Ni-NWs) fillers in an Ecoflex matrix demonstrated not only high flexibility but also improved transmittance and conductivity, which are essential features in several applications for transparent e-skins. [67]n addition to Ecoflex elastomers, PDMS, a silicone substrate, has been extensively used recently.It is a promising candidate for flexible and stretchable sensors owing to its nontoxicity, chemical inertness, simple process, high transparency (>95%), and stretchability (>100%). [68,69]Figure 7 shows the chemical structure of PDMS.The abovementioned TENG can recover after torsion, tension, or compression, owing to its flexible cross-linked molecular chains.Its nontoxic feature enables the PDMS-based sensors to be implanted in the human body, and they have been investigated for bioapplications on human skin.In addition, the aforementioned TENG can act as a perfect substrate for photoelectronic devices owing to its high transparency.It can effectively combine with various conductive materials to produce PDMS-contained composites for highly flexible and stretchable sensors. [70,71]urrently, numerous synthesized polymeric substrates are available; these include hydrogels and UV cross-linkable materials, [72,73] self-healing materials, [74] and block copolymer elastomers. [75]Other commercially available polymers include PE, PET, PI, and PU [76,77] (Figure 7).Among these, hydrogels can connect metallic electrodes and living tissues as biocompatible materials, normally required as soft and conductive materials.In addition, their elastic moduli are close to that of living tissues and can be well adjusted in a pressure range of 1À100 kPa.Moreover, hydrogels with self-healing abilities and softness properties can mimic the chemical and mechanical characteristics of human skin; however, achieving the desired effects in terms of water retention, fatigue resistance, and strong and stretchable adhesion (like those in human skin) remains a challenge in developing hydrogel iontronic devices. [72,73]The self-healing abilities of these substrate materials, which normally comprise external and intrinsic self-healing abilities, are attributed to the dynamic balance of the cross-linked networks and osmotic paths in the polymer systems.Several mechanically robust elastomers have been successfully designed for self-healing soft devices.For example, a 3D network structure hydrogel exhibited stretchable, transparent, and ionic conductors that enabled the transmission of large-frequency electrical signals over a long distance for artificial muscles, skins, and axons. [78]A self-patterned hydrogelbased pressure sensor, which was highly biocompatible, stretchable, and transparent, was well fabricated. [79]Generally, flexible electromechanical sensors derived from self-healing hydrogels satisfy the requirements for both substrate and functional materials.Moreover, UV cross-linkable elastomer materials are compatible with advanced technologies, such as 3D printing and photolithography, even for low-cost mass production. [72,73]riblock copolymer elastomers are normally assembled with a long soft chain in the middle and two short rigid chains at the ends consisting of SBS, SIS, poly(styrene)-poly(ethylene butylene)-poly(styrene), and poly(styrene)-poly(isobutylene)poly(styrene) (Figure 7), in which poly(styrene) microdomains act as physical cross-linking agents in the poly(butadiene) matrix; thus, the properties of the triblock copolymer elastomers exhibit good organic solubility, adhesiveness, viscoelasticity, and fluidity. [75]Therefore, they can effectively achieve the desired physical properties of the sensors, such as binding capability to various substrate materials (e.g., glass, metals, and polymers) and compatibility with printing technologies.For elastomer-based strain sensors, the strain hysteresis can occur under cyclic stress loadings, which are mainly caused by the viscoelastic behavior of the elastomers used, poor interfacial interactions, and the modulus mismatch among the elastomers and active conductive nanomaterials.To overcome this challenge, researchers suggest microchannel designs, elastomer fiber surface wrapping of thermoplastics, and construction of wrinkled surface structures. [80]o date, it is difficult to explore a universal, facile, and scalable approach in manufacturing stretchable strain sensors with low hysteresis and to clearly understand the working mechanism to reduce hysteresis that effectively supports the ideal designs of strain sensors and stretchable conductors.
PI substrates have been widely used as common films in flexible sensors owing to their outstanding thermal stability, mechanical features, and insulating properties. [81]Specifically, PIs have high glass-transition temperatures of 360À410 °C; [82] thus, the corresponding sensors can be manufactured and operated at high temperatures.Furthermore, these polymeric Nonabundant in natural sources Good charge transfer and stability substrates are elastic; thus, they can be used as potential supporting materials for flexible electromechanical sensors. [83]However, these films cannot recover under large deformations and are not transparent; thus, they are only utilized as highly stretchable substrates.Their characteristics are similar to those of PE and PET substrates, except these polymeric substrates are highly transparent. [84]Several other materials have also been developed as stretchable substrates for sensors, including silk fibers, papers, and 3D porous frameworks, which are highly flexible, low cost, lightweight, and endure relatively large deformations. [85,86]

Conductive Materials
The ability of the electromechanical sensors to achieve high flexibility and sensitivity depends not only on the polymeric substrates but also on the selection of functional conductive materials.The conductive materials are the main components in electronic devices for converting mechanical stimuli into electrical signals.For high-performance sensing devices, the potential materials should provide good chemical stability, conductivity, mechanical behaviors, and compatibility with largescale processing techniques.To date, conductive polymers and nanostructured conductive materials (graphene, CNT, metal NWs, and metal NPs) have been widely used by placing them on polymeric substrates or embedding them into a stretchable matrix (Figure 7).The details are presented herein.

Conductive Polymers
Conductive polymers (e.g., conductive polymeric composites, intrinsically conductive polymers, and ionic conductors) have gained significant attention for manufacturing stretchable and flexible electronic devices because they demonstrate outstanding mechanical features and well retain electrical behaviors during stretching.In a previous study, a large-area solution process was used to synthesize an intrinsically conductive polymer such as a poly(pyrrole) (details of the chemical structure are shown in Figure 7) hydrogel with hollow microspheres for use as an active component in a tactile-type piezoresistive sensor. [60]Owing to the low elastic moduli of these poly(pyrrole) microspheres and an efficient contact in the working mechanism, the sensor achieved a high sensitivity of 133 kPa À1 , low LOD of <1 Pa, and fast response speed.A conductive polymeric composite is a typical functional material for manufacturing piezoresistive strain sensors owing to its high anisotropy and conductivity. [87,88]For ionic conductors, ionogel and hydrogels (i.e., polymers that are biocompatible and swell in brine) have demonstrated high transparency and large stretchability. [89]These polymers are softer than tissues; thus, they are considered "mechanically invisible" and serve as biometric sensors for monitoring soft tissues. [89]The polymers use the available ions as charge carriers instead of electrons (as in traditional conductive polymers) and subsequently convert mechanical stimuli into signals of capacitance or resistance.For example, an ionic skin with high stretchability, transparency, and biocompatibility was successfully synthesized in a previous study. [90]The ionic skin-based sensor successfully detected a huge strain deformation of 700% and exhibited an LOD of <1 Pa.
In addition to the abovementioned conductive polymers, π-conjugated, flexible, and carbon-based conductive polymers represent promising functional candidates for producing soft sensors.Among these, PANI, PEDOT:PSS (chemical structure shown in Figure 7), and their derivatives are extensively used as conductive polymers for electronic device manufacturing. [91,92]hese polymers can be obtained using printing and spraying techniques and are often used to produce large-scale and low-cost electronic arrays, [73,93] owing to their excellent compatibility with solution processing techniques.Their mechanical features are more comparable to soft human skin; hence, they can be ideally used for skin-like sensors.However, their performance and stability are not superior compared with those of graphene, CNT, metal, and other conducting inorganic materials.

Carbonaceous Nanomaterials
Carbonaceous nanomaterials, such as graphene, CNT, and carbon black (Figure 7), have been extensively used owing to their excellent chemical stability and biocompatibility and good mechanical and electronic stability.
[96] In addition, it is commonly used for mass and low-cost production and is an ideal active material for developing various sensors (particularly pressure sensors). [94,97]However, several defects appear in the hexagonal honeycomb structure near its edges during stretching, which indicate significant changes in the structures of the electronic bands and the resistance of graphene nanosheets.Several studies have investigated the potential use of graphene in sensors; thus, many technical approaches have been developed, including chemical/mechanical exfoliation and chemical vapor deposition (CVD) techniques.Chemical exfoliation is an oxidation-reduction process in liquid and can be successfully used for the mass and low-cost production of graphene nanosheets.Spin/spray/dip coating, vacuum filtration, and inkjet printing are other approaches for the mass production of graphene nanosheets, owing to their compatibility with solution-processing techniques. [98,99]However, several defects can appear during the chemical treatment process, which can reduce the electrical and optical performance.Mechanical exfoliation (e.g., ultrasonication and ultrashear mixing) has been used to mass-produce graphene nanosheets without defects. [99]The CVD method can produce large graphene films with high electronic performance on SiC or any metallic surface.In a previous study, graphene-woven fabrics were successfully fabricated using the CVD technique with a Cu mesh before placing them on a PDMS elastomer substrate to become a piezoresistive strain sensor. [100]However, defects and contaminants may be introduced when graphene films are transferred to soft substrates, impacting their applications. [101,102]In a previous study, a combination of crumped graphene and nanocellulose was used to develop a high-performance strain sensor (GF = 7.1 at ε = 100%). [103]A tactile-type piezoresistive sensor was developed in a previous study by integrating two laser-scribed foamlike graphene films.The resulting structure of the graphene foam-based sensor demonstrated a high sensitivity of 0.96 kPa À1 at a low pressure of <50 kPa. [104]NTs are 1D cylindrical carbonaceous nanostructured materials.They exhibit high chemical stability, noticeable charge carrier mobility, excellent conductivity, and excellent mechanical strength and elasticity. [105]CNTs with a suitable chiral angle exhibit high sensitivities owing to changes in the electronic bands under mechanical stimuli. [106]In addition, CNTs can be deposited directly onto polymeric substrates using solutionprocessing techniques.These solution-based techniques focus on pristine and functionalized CNTs, which are key issues for the fabrication and application of CNT-containing sensors.The dispersion of CNTs into various organic solvents is related to two main forces: the attractive van der Waals forces between the CNTs and the interactions between the CNT bundles and the solvent.The Hansen solubility parameters are used to precisely determine the dispersion states of CNTs into different organic solvents, as listed in Table 4. [105] In addition to traditional techniques such as spin/spray/dip coating, vacuum filtration, and roll/inkjet printing, several large-scale fabrication methods have been developed for CNTs recently.For example, the Langmuir-Blodgett (LB) method has been used to produce a high-performance and large-scale CNT film. [107]The LB method can be simultaneously combined with a spray technique to produce large-area and high-performance nanostructured carbonaceous films consisting of graphene and CNTs. [108]For instance, a pressure sensor based on transparent SWCNT films was successfully constructed by Lipomi et al. using a spray-depositing method. [57]Yamada et al. developed a novel CNT-based strain sensor for human motion detection. [109]The as-overlapped CNT films were placed onto a PDMS elastomer substrate and then assembled into a piezoresistive sensor.Consequently, a high stretchability of 280%, fast response speed of 14 ms, and good durability were achieved when stretched to 150%, enabling the sensor to detect human motion as a wearable electronic device.Another technical approach for developing selfassembling ability comprises arranging 1D carbonaceous nanomaterials into an interconnected network or designed micropatterns.This is the main key to appropriately constructing sensors for carbonaceous nanomaterials.Thus, CNTs have been widely used in electromechanical sensors, and their industrial demand has been increasing.To date, strategies for synthesizing CNTs include laser-ablation, arc-discharge, floating catalyst, CVD, and plasma-enhanced CVD methods, which are improved synthetic methods for CNTs, as shown in Table 5.In addition to the aforementioned carbon-based conductive materials, carbon nanofibers are the most basic nanostructure of 1D sp 2 carbon, offering a fast current transfer and a large surface area.Thus, they are extensively utilized as energy storage and conductive materials to fabricate electromechanical sensors. [110]In particular, the electrospinning technique has gained substantial attention for large-scale production of uniform carbon nanofibers owing to its low cost, simple operation, and high performance. [111]The uniform polymer nanofibers are fabricated using the electrospinning method and then changed into sp 2 carbon materials (i.e., carbon nanofibers) after carbonization.

Nanostructured Metals
In addition to the abovementioned carbonaceous nanomaterials, nanostructured metals are also extensively used as conductive materials or electrodes for producing flexible electronics owing to their excellent conductivity.Nanostructured materials comprise several various nanostructures, including metallic NWs, [112,113] metallic NPs, [114] metallic nanothin films, [115] and liquid metals. [116,117]Generally, 1D nanostructured metals with high aspect ratios are effectively applied in stretchable and flexible devices rather than 2D and 3D metals because they can form and maintain their percolation networks at relatively low concentrations, even under large stretching deformations.This implies that 1D metallic nanomaterials can slip and collide with each other rather than breaking, thereby preserving the percolation pathways.Their high conductivity can be well maintained when the sensors are stretched.Ag-NWs are potential substances that can be deposited onto various substrates owing to their excellent transparency, conductivity, and mechanical flexibility. [118,119]The strong interactions between Ag-NWs and supporting substrates ensure the efficient performance of the conductive networks under repeated mechanical deformations.For example, in a previous study, Ag-NWs were directly deposited on the surface of poly(dopamine)-modified PDMS substrate to fabricate a highly transparent, stretchable, and conductive thin film through spraying. [119]n another study, a PDMS-embedded Ag-NW sensor with a multiscale structure was used to construct a high-performance pressure sensor. [113]Ag-NWs are relatively expensive, and their natural sources are not abundant.Thus, due to their low cost and high performance, Cu-NWs have become a potential substance for flexible electronics.However, they are extremely sensitive to moisture or oxygen, leading to undesirable chemical corrosion, antioxidation, and degraded conductivity. [120,121]Conductive networks based on metallic NPs have also been fabricated through different technical approaches (e.g., screen and inkjet printing). [122,123]In a previous study, Cu-NPs were dispersed well in an oil-based ink, significantly enhancing its silk-screen printing ability onto a flexible PI substrate before being transferred into conductive patterns under heat treatment. [122]In another study, commercial abrasive papers were integrated with microcracked Au-NP films to produce a self-waterproof crack-based piezoresistive sensor. [86]Additionally, flexible metal electrodes could potentially possess high conductivity and excellent stability under large strains.Normally, metal electrodes are formed into nano/microsized structures such as NPs or thin films (<100 nm).These thin film-like products are not damaged when folded or bent but may form cracks when stretched.For example, researchers have used laser and thermal techniques to successfully fabricate flexible and transparent electronics by integrating flexible substrates (e.g., PI and PET) and synthesized electrodes (e.g., NiO x NPs, NiO x thin films, and CuO x NP ink). [124,125]Using this approach, various conductive materials can be further developed for better applications of electromechanical sensors, such as in the sports and healthcare sectors.

Conductive Hybrids
Each conductive material has weaknesses and strengths.Using a single functional material alone to produce stretchable sensors will likely lead to imperfections in the active materials. [5]herefore, the amount of conductive fillers (sensing materials) embedded into a polymer matrix (insulation material) is crucial, mainly regarding the contact continuity of the conductive fillers for constructing the flexibly conductive pathways in the polymer matrix.When the carbonaceous fillers are added to the polymer matrix, conductive pathways are constructed in the insulating polymer matrix through the connections of the adjacent fillers.Thus, the percolation threshold (Ф C ), which is a key parameter based on the classic percolation theory, should be considered.The relationship between the electrical conductivity of the fabricated composite and the filler content can be determined using the scaling law (Equation ( 8)), where σ is the electrical conductivity of the composite containing both the conductive fillers and the insulating polymer matrix, σ 0 and Ф are the electrical conductivity and volume fraction of the conductive fillers, respectively, and t is the dimension of the conductive networks in the polymer matrix.Thermal CVD requires a temperature range of 500-1000 °C This technique is highly adaptable but has two drawbacks: 1) the tubes are not correctly oriented or in the ideal straight form, and 2) the substrate material is destroyed by high temperatures

Arcdischarge
An electric arc discharge between two electrodes at >3000 °C causes carbon to sublimate from graphite SWCNTs, MWCNTs [243]   In the presence of appropriate metal catalyst particles (Fe, Co, or Ni), the shape of nanotubes develop The use of filler-based conductive hybrids is an effective approach, indicating that the "bridge effect" constructed in the hybrid conductive fillers can favorably generate the backbones of the conductive pathways, effectively enhancing the electrical conductivity of the entire composite structure containing both the conductive fillers and insulating polymer matrix.The use of conductive fillers with various geometries can generate cosupporting conductive pathways flexibly at a specific content ratio when these hybrid conductive fillers are homogeneously dispersed into the polymer matrix. [5]For filler-based conductive hybrids, Equation ( 8) cannot accurately determine the relationship between Ф C and the loading fraction of each conductive filler.Thus, a new theoretical model has been proposed to appropriately establish the relationship of the loading amount of conductive fillers using the excluded volume theory (Equation ( 9)), [126] where m A and m B are the loading fractions of filler A and B; Ф A and Ф B are the percolation thresholds of filler A and B in the fabricated composites, respectively.If the blend phase morphology and the migration of the conductive fillers within the insulating polymer matrix are determined, Equation ( 10) can be used to modify Equation ( 9), where P' c,A and P' c,B are the percolation thresholds of filler A and B, respectively, when added individually in the insulating polymer mixture, and X is the volume fraction of the continuous phase.Nanocomposites can merge the strengths and offset the weaknesses of each material through appropriate combinations of two or more various functional materials, indicating the potential of constructing conductive pathways for sensors.Generally, strain sensors based on metallic NPs exhibit high sensitivity but a narrow sensing range, which may be related to the irreversible breakages between the nanostructured particles.In addition, using only nanostructured metal has several disadvantages, such as high oxidation and chemical corrosion, impacting its potential applications.To overcome these limitations, 1D nanomaterials are combined with 2D or 3D nanomaterials to enhance their resistance further and improve their relevant features under stretching deformations, dispersion, and attachment of the nanostructured metal hybrids into polymeric substrates containing nano/micropatterns.For example, Choi et al. encapsulated an Ag-Au nanocomposite into an elastomeric substrate, which exhibited high conductivity, good stability, and biocompatibility. [112]In a previous study, Ag-NWs were used as active bridges between Ag-NPs disconnected networks, resulting in significantly high conductivity under large strains. [127][130] In a previous study, the sensitivities and conductivities of binary systems were significantly enhanced when PANI microparticles were successfully doped with Au-NWs based on the benefits of sea-island structures. [131]Furthermore, the addition of poorly or insulating conductive fillers can reduce Ф C s in the fabricated composites, owing to the volume exclusion effect and aggregation/dispersion of the fillers into the polymer matrix.In addition, the addition of nonconductive fillers can reduce the conductive filler percentiles.The use of insulating 2D inorganic fillers can significantly enhance the filler-filler, and filler-polymer interactions and improve the dispersion/distribution of electrically conductive fillers in the polymer matrix.Additionally, the aspect ratio and intrinsic electrical conductivity of fillers are important in the selection of filler-based conductive hybrids that can produce composites containing both conductive fillers and insulating polymer matrix with low Ф C and improved properties.Therefore, depending on the purpose, diverse conductive materials can be explored for suitable applications in electromechanical sensors.

Approaches for Composite Structures Designed for Electromechanical Sensors
The essential components of a typical electromechanical sensor are flexible/stretchable substrates, signal transfer materials, and sensing materials.The selection of suitable components should be carefully considered to construct a flexible and wearable sensor with desired efficiencies, such as a wide working range, high sensitivity, good stretchability, and high resolution.The conductive pathways constructed inside the sensors can change in various directions under tensile or compressive stresses, thereby affecting the connections of the conductive networks and the compatibility between the components.Thus, the structures and sensing mechanisms of the sensors must be appropriately designed to enhance their sensitivity, stretchability, and durability.Developing functional composites and structures with suitable electrical and mechanical efficiencies is critically important to design and manufacture high-performance electromechanical sensors, which is a big challenge for researchers.
Recent composite structures designed for electromechanical sensors have demonstrated numerous coupling approaches between functional materials, such as a conductive network attached to the substrate surface, conductive network dispersed in a stretchable matrix, conductive network covered by a polymer matrix, and conductive network scaled up to arrays.The schematics of various composite structures designed for electromechanical sensors are shown in Figure 8.

Conductive Network on Substrate Surface
Exploring the flexible and wearable sensors originating from functional components and structures with the desirable capabilities for detecting and monitoring small-and large-scale deformations is necessary for facilitating practical applications.Composite materials and structures, electrical capabilities of fillers, and mechanical flexibility of stretchable matrices are common considerations in the manufacture of flexible electromechanical sensors.However, it remains a challenge for researchers to achieve the desired characteristics of fabricated sensors.
In addition to advanced materials, a common approach for achieving the desired structures of sensors is attaching conductive networks to substrate surfaces.Normally, this approach satisfies the requirements for strong adhesion and interactions between the conductive components and substrate surfaces, achieving high sensor performance and stability.For example, in a previous study, a conductive gradient network was fabricated using a surface penetration technology, in which carbon black penetrated from the surface to the interior of a natural rubber latex glove. [132]The carbon black-contained depth increased from 2 to 80 μm for an ultrasonic immersion period of 1À30 min.In addition, the conductive gradient network of the sensors exhibited wide detection ranges for strain (0.05-300%) and pressure (1.7-2900 kPa) and excellent reliability and reproducibility.Therefore, a hierarchical nanostructured network-contained pressure sensor with Ag-NW/graphene nanocomposite-attached polyamide nanofibers [133] was used to construct an ultrathin flexible piezoresistive sensor with high sensitivity and wide detection range.The Ag-NWs were placed on a framework of polyamide nanofibers to create conductive pathways, whereas graphene was used for the Ag-NW-crossed active bridges.Thus, the hierarchical nanostructured network contained Ag-NW/graphene conductive pathways, which enabled the sensors to achieve a high sensitivity of 134 kPa À1 (at a pressure range of 0-1.5 kPa) while exhibiting a wide detection range (>75 kPa) and excellent durability.Sharma et al. proposed a hybrid ionic nanofibrous membrane-based capacitive sensor comprising a sensing layer consisting of Ti 3 C 2 T x MXene and an ionic salt of lithium sulfonamides on a nanofibrous backbone of poly(vinyl alcohol) attached between two Au-coated microstructured PDMS films. [134]It exhibited a high sensitivity, rapid response time of 70.4 ms, low LOD of 2 Pa, and high durability.In addition, a superhydrophobic stretchable film was proposed for 3D conductive network-based strain sensors. [135]Superconductive carbon black NPs were attached to electrospinning-related thermoplastic PU fibers to create a carbon black/PU conductive fiber film.Subsequently, a mixture of CNT and fluorinated silica (F-SiO 2 ) was spray-deposited on the conductive fiber film to fabricate a conductive carbon black/PU@CNT/F-SiO 2 composite film.Thus, the as-obtained conductive network-contained composite was produced through two processes: attachment and adhesion of the conductive components (i.e., carbon black and CNT/F-SiO 2 ) to the electrospinning-related thermoplastic PU fiber film.To further enhance the stretchability and stability of a conductive composite film, Ding et al. used a blended solution of PDMS and perfluorodecyltrichlorosilane to cover composite film and used it as a strain sensor. [135]It exhibited high sensitivities (GF = 12.05-60.42),a broad working strain range (0-100%), rapid response times (75-100 ms), and long-term stability.Thus, diverse fabrication processes, which are combinations of numerous and various technical approaches, can produce sensor structures with the desired flexibilities.
Commercial templates, including thin porous elastomer sponges and woven cotton fabric, have been used for attaching CNT components and producing flexible electronics with multidirectional sensing capabilities. [136,137]For instance, Kim et al. produced a high-performance and flexible pressure sensor based on a 3D microporous CNT network-attached thin porous PDMS sponge using concomitant O 2 plasma treatment and immersion methods. [136]It exhibited a wide pressure sensing range (10 Pa-1.2 MPa) and maintained proper sensitivities (0.01-0.02 kPa À1 ) while demonstrating good electromechanical stability and sensitivity under bending deformations.Zhang et al. attached a wearable woven cotton fabric to a SWCNT component using a facile solution process. [137]Owing to the unique hierarchical backbone of the twisted-yarn cotton fabric incorporated with the SWCNT conductive network, the strain sensors exhibited outstanding performances, including a large workable strain range, low LOD, excellent responses, long-term stability and durability, and good air permeability.Additionally, a study effectively deposited and attached a combination of Ag-NWs and lamellarstructured MXene nanosheets to nonwoven fabrics. [138]The interconnected conductive network containing 1D Ag-NWs bridging the 2D MXene nanosheets enabled the sensor to achieve extremely high sensitivity (GF = 1085, ε = ≈100%) and facilitated excellent long-term durability.Wang et al. explored a novel bilayer conductive mesh strain sensor [139] using concomitant 3D printing and ink-spraying techniques.Four major components were considered: the reentrant auxetic frame, auxetic transition zone, bilayer conductive mesh containing the bottom carbon black/Ecoflex conductive mesh, and upper SWCNT conductive coating and electrodes.These components attached and adhered to each other, forming a strain sensor.The as-obtained sensor had a high GF of approximately 13.4 owing to the synergistic effect of the bilayer conductive mesh, strain concentration, and auxetic deformation.
Yao et al. successfully developed microstructural graphenebased sensors.A low-cost commercial PU sponge, which was used as a flexible supporting framework, was attached to a graphene conductive component using the chemical and thermal reductions in GO nanosheets. [140]A similar approach was applied to a 3D composite sponge comprising graphene nanoribbons and a PU network. [141]This structure exhibited high sensitivity (0.26 kPa À1 ), low LOD (9 Pa), and good stability, which is attributed to the instantaneous changes in the contact areas between the constructed conductive networks during deformations.This low-cost and easily scalable graphene-attached PU sponge pressure sensor can be potentially applied in array-like pressure sensors without complex nanostructural designs.Huang et al. [142] and Liu et al. [143] used freeze-drying methods to fabricate conductive CNT/PU and graphene/PU composite foams with interconnected porous structures, respectively.Among the composite foams containing CNT and graphene, the 2D graphene nanosheets increased the pore wall thickness more effectively compared to those with CNT components, forming a robust porous structure with notable mechanical recoverability and stable sensing performance at >90% compressive strain.
A 3D network-contained thermoplastic elastomer substrate was prepared using a fused deposition molding 3D printing technique.It was immersed into a CNT dispersion using a traditional dip-coating method to form CNT conductive layers attached to the thermoplastic elastomer substrate surface [144] to construct a high-performance and flexible sensor with a 3D conductive sensing unit.Owing to the novelty of the 3D conductive network, the sensor achieved an extremely high sensitivity of 136.8 kPa À1 at a pressure of <200 Pa and a GF of 6.85 at a strain of 800%.It achieved excellent stability and durability because the attachment of the CNT component to the substrate surface induced little effect on the flexibility of the elastomeric composite of the sensor.However, the desired flexibility of the sensors is often based on synthetic polymer materials, which are not biodegradable and biocompatible, thus potentially causing environmental pollution.Accordingly, Lv et al. fabricated an environmental-friendly wearable sensor using natural materials, such as milk protein fabric, natural rubber, tannic, and vitamin C. [145] Among these, the milk protein fabric is a new functional fiber fabric produced based on milk proteins and cellulose.They have numerous advantages, including high strength, stretchability, skin-touch-ability, and diverse functional groups.The natural rubber and vitamin C-reduced graphene networks were attached to the milk protein fabric framework through the Ca 2þ and tannic components.Owing to the interfacial interaction and organized 3D conductive networks, the environment-friendly sensor exhibited good mechanical features and high sensitivity.Thus, these functional materials and structures can be used to produce environment-friendly flexible sensors.
To date, numerous techniques have been explored to fabricate conductive network-attached substrate surfaces used as electromechanical sensors.In these sensors, the conductive components are directly attached to the surface of a substrate using advanced techniques such as nanoimprint lithography, dropcasting, laser scribing, and roll-to-roll printing techniques, in which each method exhibits various advantages and disadvantages (Table 6).Depending on research purposes, trends and advances in the ideal designs of electromechanical sensors are developed appropriately to produce high-performance sensors.Among these, the conductive inks used in printed electronics generally satisfy certain requirements in terms of high flexibility and conductivity, low cost, and good compatibility with stretchable substrates; therefore, they are normally attached directly to the substrate surfaces through casting, screen printing, and rolling.For example, a Cu-NP ink was attached to a flexible PI substrate using silk-screen printing and was subsequently transferred onto conductive patterns under heat treatment. [122]n another study, graphene ink (i.e., graphene nanoplatelets) was screen-printed on plastic and paper substrates; a printed graphene pattern was then produced after postprinting treatments (i.e., thermal annealing and compression rolling). [146]The pattern exhibited high conductivity, good flexibility, and durability.These approaches have the potentials for mass production of lowcost and flexible printed electronics.
In addition to conductive inks, Hyun et al. designed an ordered zigzag structure on a PDMS substrate by filling and attaching an interpenetrating network between a poly(ethylene oxide) gel and Ag-NPs to the zigzag-patterned PDMS surface using concomitant spin coating, UV exposure, and reduction methods. [147]The as-obtained wavy structure enabled the sensor to provide high conductivity and stretchability simultaneously.
Nam et al. proposed a simple laser digital patterning procedure for the fabrication of a transparent Ni-based flexible sensor using solution-processed NiO x thin films [148] (Figure 9A).Owing to the low temperature used for the reductive sintering of the NiO x NPs, the Ni component was successfully attached to the PET surface.The sensor achieved superior electromechanical performance, high transparency, high thermal stability, high corrosion resistance, and excellent durability.To fabricate a transparent and stretchable sensor, a mesh-structured Ag-NWs-PDMS composite was fabricated using a combination of spin coating, laser scribing, and drop-coating methods. [38]A PDMS liquid was spin-coated to form a uniform thin film (≈100 μm) and subsequently drawn into a mesh-shaped structure using a laser scribing procedure.Subsequently, using a drop-coating method, the O 2 plasma technique was used to treat the surface of the PDMS elastomer film to deposit the Ag-NW components.The sensor achieved an extremely high sensitivity over a wide sensing range (GF = 846 at ε = 150%), slight hysteresis, high transparency of 88.3%, and long-term durability owing to the connected conductive networks on the surface of the Ag-NW/PDMS thin film.Generally, several conductive microstructures have been successfully designed further to improve the mechanical sensitivity of fabricated electromechanical sensors.Thus, elastic microstructures have been fabricated using potential micropatterns directly attached to conductive components.These can be used to construct conductive microstructured networks, such as 2D micropillar-shaped, [60] micropyramid-shaped, [62] and microdome-shaped [2] microstructures, as well as metal-impregnated tissue paper. [149]Consequently, the requirements for costeffective and advanced technical approaches are increasing and continue to pose considerable challenges.

Stretchable Matrix-Dispersed Conductive Network
Another fabrication method for conductive polymeric composites is the dispersion of conductive fillers inside stretchable matrices through mixing, blending, or chemical/physical cross-linking synthesis.The dispersion, interaction, and viscosity of the blended functional components are important in this approach for the fabrication of stretchable matrix-dispersed conductive networks.However, this approach generally requires higher loading fractions of conductive fillers to achieve a sufficiently high electrical conductivity, resulting in high viscosity and inhomogeneous dispersion of the fillers.Hence, loading amount is the key factor in achieving effective sensor performance and, similarly, good compatibility when mixed together.
Typically, gels or hydrogels comprise polymer networks containing water and have outstanding advantages.Highly stretchable conductive hydrogels have been widely studied, particularly in pressure and strain sensing, electronic skins, and implantable bioelectronics.Using a new cross-linked complex coacervate approach, Nguyen et al. fabricated highly stretchable and compressive conductive hydrogels. [150]Graphene nanoplatelets were used as conductive components and dispersed into a mixture of poly(ethylene imine) and poly(ethylene glycol) diglycidyl ether using a vortex-mixing machine.Conductive covalently crosslinked coacervate gels, which were highly stretchable and compressive conductive hydrogels, were produced.These conductive hydrogels were self-healable, with a stretchability of 1500% and compressive strength of 25 MPa.Zhu et al. proposed novel polyion complex/PANI hybrid hydrogels with approximately 65 wt% water and high conductivity. [151]The conductive phase of PANI was dispersed into a polyion complex matrix using phytic acid, which maintained the viscoelasticity of a tough matrix and exhibited good self-recovery and rate-dependent behaviors.Therefore, using stencil masks, these hybrid hydrogels can be alternately used between nonconductive/conductive patterns.Through this facile patterning method, they can be used for large-scale productions of wavy gel circuits and multichannel sensor arrays.
Another advanced fabrication and technical approach is using PEDOT:PSS nanofibrils for uniform dispersion into a poly(acrylamide) hydrogel matrix, which serves as a soft electrode component.This conductive hydrogel was photopolymerized with a poly(2-hydroxyethyl acrylate) elastomer using digitalized UV irradiation. [152]It exhibited a high conductivity and superior electrical stability owing to the hybrid hydrogels of the poly(2hydroxyethyl acrylate)/PEDOT:PSS/poly(acrylamide).The strong adhesion interfaces between the hydrogel and elastomer revealed the significant stability of the hybrid structure.This approach can also be used to fabricate stretchable electronics, such as 3D-printed capacitive and electroluminescent devices, using a digital light processing 3D printing method.Du et al. successfully synthesized a conductive film based on the cross-linking of poly(azomethine) (PAMÀCHO) and ethylene-diaminefunctionalized GO (GOÀNH 2 ) to form a dynamic covalent bond (e.g., imine bond and -CH═N-). [153]Owing to good dispersion and available functional groups of the active components, the conductive film achieved a high sensitivity (GF of 641) and excellent stretchability of 212-275%, with skin-like mechanical behaviors.Its healing efficiency reached up to 99% within 24 h through a mechanical approach; it likely could reach >95% based solely on the electrical conductivity.Thus, this conductive film achieved high performance and exhibited excellent self-repair in the sensor.These studies can lead to new strategies for developing multifunctional soft materials and potential applications of the sensors in health diagnoses and security protection.
In addition to electronically conductive hydrogels, Guo et al. mixed CNT components into a PDMS liquid matrix using a three-roll-mill machine to enhance the uniform CNT dispersion and form a composite dielectric layer using spin coating and thermal curing methods. [154]In this approach, the dispersion between functional components is important to achieve a stretchable matrix-dispersed conductive network.The loading amounts are the primary requirements to achieve high sensor performance.Thus, the flexible capacitive pressure sensor obtained from the direct mixing and spin-coating methods exhibited an increase in sensitivity.Owing to the facile production and high sensor performance, the CNT/PDMS composite dielectric layer was scaled up to prepare electrode arrays (3 Â 3 and 10 Â 10 pixels) for the sensor using a screen printing method instead of the traditional etching approach.In another study, a sensing layer was fabricated using magnetic rGO@Ni-NWs hybrid fillers dispersed in an Ecoflex elastomer matrix. [67]Subsequently, a microdome-shaped piezoresistive sensor containing the sensing layer was produced using the hot embossing technology, which resulted in a superior sensitivity of 1302.1 kPa À1 with a low device-to-device variation of 3.74%. [67]This study produced transparent, highly sensitive, and large-scale electronic skins.Ko et al. presented a stretchable conductive adhesive, which included a filler mixture of Ag particles/CNT and a PDMS matrix using a traditional blending method [155] (Figure 9B).The adhesive had a high conductivity of 6450 S⊡cm À1 , high stability, and a stretchability of 50%.Owing to its strong adhesion, gel-free nature, and dry adhesiveness, it can be used for printing on an elastic bandage (e.g., for electrocardiography monitoring and washing machines).In addition, a self-segregating conductive composite was also developed as a strain sensor printed on stretchable substrates.Yoshida et al. mixed carbon black into a PDMS liquid  [148] (B) Ag particle/CNT/PDMS stretchable conductive adhesive: (i) schematic of the preparation of stretchable conductive and printed adhesives and (ii) schematic of a stretchable conductive adhesive (adapted with permission [155] ).(C) Manufacture and structure of the triode-mimicking positive graphene/Ecoflex sensor: (i) schematic of the preparation of the sensor, (ii) schematic layout of a single sensor, and (iii) digital capture of the assembled sensor (adapted with permission [163] ).(D) PDMS-covered 3D porous conductive foam: (i) schematic of the PDMS-covered conductive foam and SEM images of (ii) 3D porous and (iii) PDMS-covered 3D porous conductive foams.Adapted with permission. [170]sing a deep eutectic solvent to enhance dispersion. [156]The as-obtained conductive network structure exhibited a superior electromechanical performance relative to other composites containing random conductive networks.Specifically, it exhibited 1) low hysteresis and high sensitivity, 2) high conductivity and low elastic modulus, and 3) excellent reliability and stability.The flexible and conductive composite was applied directly in a simple stencil printing process without any complex ink synthesis or posttreatment of the prepared devices.The carbon black was filled and dispersed into binary and ternary composites based on a thermoplastic PU and an olefin block copolymer using a facile and low-cost extrusion manufacturing process. [157]However, this process requires an additional amount of carbon black and significant kinetic control.Moreover, the use of a toxic organic solvent to disperse the nanofillers inside a stretchable matrix for the production of flexible conductive composites may induce environmental problems.Thus, a dry-mixed method is typically used to avoid using toxic organic solvents.Wang et al. presented an innovative method by modifying the traditional dry-mixed method using minor CNT bridging and high-frequency electric field enhancement at the percolation threshold of a graphene-PDMS composite. [158]They observed significant improvements in the electrical conductivity, GF (>8768), and strain range.In another study, Ag-Au nanocomposites were thoroughly dispersed inside an SBS elastomer (triblock copolymer elastomer) to form a soft conductive composite.The composite was then patterned using PDMS molds and assembled successfully into a multifunctional wearable electronic patch. [112]These approaches have paved the way for manufacturing potential hybrids of conductive composites in future stretchable electronics.
Conductive networks based on synthesized 3D porous frameworks represent another approach to dispersion between the functional components occurring in the mixtures and have been extensively studied in recent years, particularly for the dispersion of conductive materials in solvents or water.For example, Tang et al. produced Cu-NWs aerogel monoliths, [159] and Gui et al. produced conductive CNT sponges using the CVD method. [160]heng et al. proposed a novel pressure sensor with a bubbledecorated honeycomb-like graphene network structure. [161]he designed structures of the honeycomb-like network and bubbles were based on the evaporation of gases and water (CO, CO 2 , H 2 O) sealed in adjacent graphene nanosheets during the pyrolysis of the oxygen groups.Thus, the pressure sensor exhibited a high sensitivity of 161.6 kPa À1 , a low operating voltage, and a good stability, which were attributed to the switching effect based on "point-to-point" and "point-to-face" contact modes.Additionally, Qiu et al. successfully presented a 3D superelastic graphene, which was fabricated using a technical combination of graphene chemistry and ice physics. [162]In addition, an Ag-NW/poly(pyrrole) composite foam sensor had a temperature-independent coefficient and exhibited a good sensitivity of 0.33 kPa À1 , a short response time (≈1 ms), and good stability.Normally, synthesized conductive sponges, foams, and aerogels without elastomers are covered by stretchable matrices to improve the stretchability and stability of the interconnection against external deformations.Generally, the abovementioned studies can provide insights for the development of multifunctional flexible sensors.Electromechanical sensors have improved diverse aspects of the results for potential functions and facilitated promising applications, all of which are mainly based on the advanced materials and structures in the fabrication approaches of these sensors.Thus, developing sensors with more effective performances is always an effective strategy.

Conductive Network-Covered Stretchable Matrix
In addition to the abovementioned approaches for the construction of sensors, there are many facile and cost-effective fabrication strategies for stretchable and flexible electromechanical sensors with appropriate features, such as sensitivity, stretchability, and durability.One common recent design strategy uses a polymer matrix to cover the conductive networks.Lee et al. fabricated a 99% metallic CNT-PDMS composite and used it as a transparent strain sensor. [39]This CNT-PDMS composite strain sensor was manufactured as a sandwich-like structure through sequential stacking of a PDMS layer, metallic CNT film, and PDMS layer.The CNT component was attached to the surface of the O 2 plasma-treated bottom PDMS layer with an amineterminated group using concomitant spray coating, solutiondepositing, and thermal annealing (90 °C).Subsequently, the PDMS liquid was cast onto the CNT-attached PDMS film to cover the metallic CNT film completely.Thus, the sandwiched-like strain sensor was designed to decrease the wrinkling and plastic deformation of the networked CNT film when the CNT film was placed on top of the bottom PDMS substrate without an additional PDMS cover layer, thereby providing high stability for the sensor.Consequently, the strain sensor achieved a high transparency of >92% and good performances in terms of hysteresis, sensitivity, linearity, and drift.Another study proposed a laser technique to scribe graphene effectively and subsequently create a low-cost, flexible pressure sensor with an extremely high sensitivity and a wide detection range [163] (Figure 9C).This pressure sensor was based on an encapsulation of triode-mimicking positive graphene and an Ecoflex elastomer matrix.It demonstrated long durability and had a signal amplification feature (similar to that of a mechanical triode) upon an external pressure bias.In another study, a CNT dispersion was drop-casted on a glass petri dish to create a uniform CNT conductive network.Subsequently, a hot prehydrogel mixture (i.e., K-carrageenan/ poly(acrylamide), a double-network hydrogel) was poured into the petri dish to cover the CNT film and thermally cured.A flexible sensor was then peeled off from the petri dish after a cooling process, revealing a highly conductive network (CNT film) and highly stretchable network (double-network hydrogel). [164]It also exhibited a good self-healing ability and a high GF of 343 (ε = 110%).Thus, it is a potential strategy for producing highly conductive, antifreezing, self-healing, and nondrying strain sensors. [164]Furthermore, a self-powered piezo-organic-e-skin sensor was developed as an active piezoelectric sensor.The PANI component was coated on aligned poly(vinylidene fluoride) nanofibers and then encapsulated by the PDMS matrix. [165]It manifested an excellent conversion of mechanical stimuli into electrical energy for sensing human finger touches (10 V at a pressure of 10 kPa) with an energy conversion efficiency of 53%, as well as excellent mechanical stability.
Another approach is the use of an elastomer matrix to directly cover synthesized 3D porous frameworks, such as aerogels, sponges, and foams.For example, a microstructured graphene aerogel was controlled to optimize the piezoresistivity effect of a graphene aerogel/PDMS nanocomposite strain sensor. [166]n another study, a graphene aerogel was produced by chemical reduction-induced self-assembly of GO nanosheets (i.e., L-ascorbic acid) and thermal annealing processes.The aerogel was subsequently covered with a PDMS matrix to form flexible polymer nanocomposites with high conductivity and excellent mechanical flexibility, which were used to produce low-cost, high-performance, graphene-based electromechanical strain sensors.Another study incorporated a device in a hollow structure constructed with a large surface roughness and bulk-free volume derived from 3D graphene-PDMS nanocomposites. [167]The number of graphene layers could increase from 4 to 14 by altering the graphene growth time.The sensitivity increased from 0.051 to 11.1 kPa À1 without any change in the elastic modulus.In addition, the elastic modulus and corresponding linear region of the sensors were well controlled in terms of the mixture of the PDMS liquid and curing agent, as well as in investigations of other desirable features, such as the density, thickness, adhesion, stability, and repeatability of the sensor.
In another study, using the CVD method, a commercial Ni framework was used as a 3D porous foam to develop graphenebased conductive networks on the Ni foam and construct 3D hollow network structures. [168]The approach was used to coat multilayer graphene on the Ni foam template, which was retained in the PDMS substrate. [102]This structure induced larger effects on the graphene-based constructed conductive networks, exhibiting a sensitivity of 0.09 kPa À1 at a pressure of >1000 kPa and a maximum GF of 25.6 under all applied strains.The Ni supporting template was also successfully used to form a flexible Ag-NW-PDMS composite structure [169] with not only porosity (Ni template) and flexibility (PDMS) but also conductivity (Ag-NWs).Thus, this structure exhibited stable steady-state responses to various pressures and stretches (even to mouthblowing).In another study, PDMS elastomer and 3D porous conductive foams were combined to produce functional composites containing conductive networks (i.e., PDMS-covered 3D porous conductive foam) using a low-cost and facile, combined freezedry-covering method, which achieved both wide stretching and sensing ranges, using as flexible and wearable sensors for detecting and monitoring human motions [170] (Figure 9D).3D porous conductive foams with cellular structures were prepared by combining low-cost polymers (gelatin/chitosan) and carbon-based fillers (GO-CNTs), which were cross-linked with glutaraldehyde and freeze-dried to produce a promising 3D porous framework with a cellular structure containing conductive bridges.Subsequently, these constructed networks were covered by an elastomeric mixture of PDMS liquid and curing agent (10:1) to produce functional composites with enhanced performances.Additionally, Wu et al. used an eco-friendly supporting framework derived from carbon nanofibercoated sugar particles to fabricate a carbon nanofibers/PDMS composite. [171]The carbon nanofiber-coated sugar foam was etched using warm water to remove the sugar particles and was then partially embedded in the PDMS matrix.The 3D microstructure of the carbon nanofibers/PDMS composite exhibited good sensitivity (GF = ≈6.5), a linear response, and high durability (ε = >70%).Layer-structured MXene-based electromechanical sensors have gained attention.For example, Adepu et al. demonstrated an SnS/Ti 3 C 2 T x (MXene) nanohybrid-based sensor for sign-to-text translation and sitting posture analyses. [172]The sensor possessed not only high sensitivity (7.49kPa À1 ) and GF (7.41) but also ultrahigh durability.The MXene nanohybrid-based piezoresistive sensor can broaden research scopes for flexible and wearable devices.
Furthermore, other cost-effective, facile, and scalable 3D supporting templates for clothes, cotton, paper, and medical tape have been effectively used for assembling conductive components (which are then covered by flexible elastomer matrixes).A study combined a cotton textile (3D template) and CNT-poly(pyrrole) mixture to produce flexible conductive networks, which were then covered with Ecoflex elastomer matrix, thereby providing a new single-friction-surface TENG. [66]In another study, an SWCNT ultrathin film was covered with silk-molded micropatterned PDMS to fabricate a flexible and transparent electrical skin as a cost-effective wearable electronic device. [173]It exhibited an ultrahigh sensitivity, a short response time, and high stability.A microstructured surfacecontained template comprising high-quality silk was found effective for constructing a micropatterned flexible PDMS thin film and was then integrated with a uniform free-standing ultrathin SWCNT film using a layer-by-layer exfoliation method.In another study, an aligned SWCNT thin film was arranged inside a flexible PDMS matrix.Subsequently, a strain sensor was assembled using a water-assisted CVD method. [109]Under stretching, the strain sensor could measure and withstand strains up to 280%.It exhibited a high durability, fast response (14 ms), and low creep (3.0% at 100% strain).In addition, commercial abrasive papers were integrated with microcracked Au-NPs films to fabricate self-waterproof crack-based piezoresistive sensors. [86]Pang et al. proposed a morphological surface with a spinosum microstructure of a random distribution based on a combination of an rGO-on-abrasive paper template and a PDMS elastomer, [174] The resulting sensor exhibited a high sensitivity of 25.1 kPa À1 over a wide linearity range of 0-2.6 kPa.The rapid advances in functional sensing devices have suggested numerous diverse innovations for practical applications of advanced active materials and structures for electromechanical sensors.Zhang et al. fabricated a carbonized cotton fabric-based multilayer piezoresistive pressure sensor. [175]The as-obtained sensor exhibited a high sensitivity of 13.89 kPa À1 , a wide pressure detection range, a fast response, and excellent repeatability.These high performances were attributed to the excellent conductivity of the carbonized cotton fabric and the multilayer structure comprising the 3D conductive network covered by the PDMS elastomer matrix.Generally, stretchable matrices can be used to cover conductive networks by pouring directly into a neat conductive layer or into conductive path-containing frameworks.This is generally conducted using the abovementioned two approaches: the conductive network-attached substrate surface and stretchable matrixdispersed conductive network.This is because the conductive network should be produced first before being partially or fully covered, depending on research purposes.This improves the stretchability and durability of the sensors against applied forces; however, achieving a highly sensitive performance from fabricated electromechanical sensors remains a challenge.

Array-like Electromechanical Sensors
Large-scale integration of high-performance conductive components into flexible substrates can produce electronic, sensing, and energy devices with promising applications (e.g., array-like electromechanical sensors and large-area integrated sensors) based on the aforementioned technical and structural approaches.Recent progress in the printing and transferring of singlecrystalline, inorganic nano/microstructures onto flexible substrates has allowed them to achieve high performance through various technical approaches.In particular, the contact printing of parallel NWs arrays can produce high-performance and bendable sensors. [176]However, certain limitations on processing-and assembly-related impediments remain.To overcome these challenges, a macroscale integration of parallel NW arrays (18 Â 19 pixels) was developed as a flexible pressure sensor array with >2000 bending cycles, [45] which is the largest integration of ordered active NW arrays as a model platform for the potential integration of nanomaterials.A piezoelectric sensor based on volatile organic compounds was fabricated as a promising platform. [177]In addition, SiO 2 NWs were aligned into dense and patterned arrays employing vapor-liquid-solid CVD processes and using catalysts at high temperatures. [178]Plasma thermal reactive ion etching was used to produce aligned SiO 2 NWs with aspect ratios extending up to 20 μm and lengths exceeding 1 μm, showing the potential applications in surface modification, optoelectronic, and electromechanical-based devices.Another technical approach for a dendritic NWs array is using a dendritic crystal that grows self-organically to effectively assemble uniform ZnO NWs (10À30 nm) into well-ordered 1D microscale arrays that form comb-type structures using a UV laser method. [179]n addition, Wang et al. successfully assembled an omnidirectional strain-sensing array by integrating Ag-NW/graphene and the vertical circular cavity structure of PDMS substrate. [180]n addition, owing to the available advantages of porous conductive networks, 3D porous composite-based sensors have been developed into array-like electromechanical sensors.For example, in a previous study, a graphene/PU composite sponge was used to form a sensor array of 13 Â 11 pixels using a modified and scaled-up fabrication process. [140]In another study, MoS 2 planar sheets were bonded to 3D graphene porous networks and subsequently covered with Ecoflex elastomer [65] (Figure 10A).The obtained structure of the conductive network was conformably arranged into a cracked paddy shape.The MoS 2 /graphene/ Ecoflex sensor exhibited a sensitivity of 6.06 kPa À1 and an excellent repeatability owing to the conformal MoS 2 planar sheets.This MoS 2 /graphene/Ecoflex composite was applied in a tactile sensor array (3 Â 3 pixels) using a PI membrane and Kapton tape (Figure 10A), demonstrating an accurate sensing ability, for potential applications in touch electronics.Yang et al. fabricated a graphene-woven microfabric/PDMS composite with an ultrahigh strain sensitivity (GFs of 500 (ε = <2%) and 10 4 (ε = >8%)) and an improved working range of >40%.These results were obtained by adjusting the integrated graphene network on the macrowovenfabric geometrical framework, as well as on the formation and propagation of cracks. [98]The macrowoven-fabric framework containing the graphene network evidently induced a high interfacial resistance among the interlaced ribbons and the formation of a microscale-controllable, locally oriented as a zigzag crack near the crossover location, leading to a synergistic effect that enhanced its sensitivity.The graphene woven microfabrics were used to construct an electromechanical sensor array of 8 Â 8 pixels for artificial e-skin applications.In another study, a carbonized cellulose paper-based pressure sensor was developed using a low-cost and environment-friendly approach. [181]It exhibited a high sensitivity of 2.56-5.67kPa À1 in a pressure range of 0-2.53 kPa, a short response time of <30 ms, a low detection limit of ≈0.9 Pa, and a good durability.Owing to the porous and corrugated structure of the carbonized crepe paper, it could be directly integrated on an electrode-containing flat printing paper to manufacture a microsized pressure sensor array (4 Â 4 pixels).Thus, the carbonized crepe paper-based pressure sensor provided new insights and technical approaches for developing green devices.In addition, another study proposed a multilayer piezoresistive pressure sensor array with a PDMS-anchored carbonized cotton fabric. [175]he deposition of conductive materials on a substrate surface, an approach extensively used to create electromechanical sensor arrays, is an alternative to the aforementioned 3D templates used in constructing composite sensors.It is used more significantly for the nano-/micropattern-constructed polymeric substrates.In a previous study, a flexible PDMS substrate was constructed containing an array of microscale pyramidal features, and the stretchable composite electrode included a conductive polymer (PEDOT:PSS) and aqueous polymeric dispersion elastomer (PU).Consequently, geometrical changes occurred when pressures were applied, exhibiting a sensitivity of 10.3 kPa À1 (ε = 40%). [62]ith this arrangement, pressure-induced geometrical changes could facilitate an optimum evaluation of the conductive electrode components, thereby enhancing the pressure sensitivity of the sensors.A similar approach was used in a polymeric array based on triblock copolymer elastomers (with good solubility, adhesiveness, viscoelasticity, and fluidity). [75]In addition, with the strengths of both the graphene components and PDMS with pyramid microstructures, a flexible tactile sensor was fabricated based on microstructured graphene arrays (sensitive layers). [182]he tactile sensor array achieved a high sensitivity (-5.5 kPa À1 ), low detection limit of 1.5 Pa, and short response time (0.2 ms).Its sensitivity could be well-adjusted through various parameters of the microstructured pattern.Another pressure sensor array (5 Â 5 pixels) incorporated a more complex design to satisfy the requirements for both the structural and active components while achieving a high sensitivity.It included two layers: an Audeposited PDMS-Ecoflex micropillar layer (top layer) and PANI conductive nanofibers on a PET substrate layer (bottom layer). [115]It exhibited a high sensitivity of 2.0 kPa À1 at a pressure of 0.22 kPa, a low detection limit of 15 Pa, a short response time of 50 ms, and a good durability (owing to the air gap between the layers).These studies demonstrated potential applications of stretchable pressure sensor arrays for wearable and artificial devices in electronic skin and human-machine interfaces.
Optimizing the microstructures constructed in dielectrics and electrodes is a structural approach for enhancing the performance of capacitive pressure sensors.Yang et al. successfully developed a novel 3D microconformal graphene electrode (microcylinder-type array) for high-performance and flexible capacitive pressure sensors. [183]The various morphologies of the graphene electrode (nano/microsized patterns) were investigated in detail.The electrode exhibited a high sensitivity and flexibility (3.19À7.68kPa À1 ), short response time (30 ms), low LOD (1 mg), and high stability.However, in terms of the requirements for long-term on-body utility concerning conformability, air permeability, and durability, enhancing the sensitivity remains an important strategy for ultrathin capacitive sensors.Yu et al. produced a highly sensitive and all-fabric capacitive pressure sensor based on a micropatterned thermoplastic PU nanofibers (using an electrospinning method) sandwiched with an Ag-NW electrode, such as a micropatterned nanofiber dielectric layer-contained breathable all-fabric network. [184]In addition to the outstanding performance of this sensor array, such as a high sensitivity (8.31 kPa À1 at a pressure of <1 kPa), a low detection limit (0.5 Pa), a wide detection range (0.5 PaÀ80 kPa) and excellent robustness, the obtained structure also exhibited good skin conformability, significantly high thinness, and air permeability.
Lin et al. proposed an innovative and self-powered pressuretype array based on a triboelectric active sensor. [185]Based on the working mechanism of this triboelectric pressure array, such as pressure-related responses in an open-circuit voltage and shortcircuit current, the sensor provided both static and dynamic pressure sensing on a single electronic device.It achieved a high sensitivity (0.31 kPa À1 ), low LOD (2.1 Pa), short response time Adapted with permission. [65](B) Applications of the flexible pressure sensor using a graded porous material: (i) digital photos of ant nests, (ii) schematic of the flexible pressure sensor, (iii) schematic of a bent finger with the sensor, and (iv) applications of the sensor with high sensitivity and wide detection range.Adapted with permission. [198](C) Application of the vanadium nitride/CNT strain sensor for detecting finger motions: (i) control circuit of the robot hand system containing five signal acquisition circuits and (ii) digital captures of instant controls of the robot hand by investigating the gestures from "five" to "one".Adapted with permission. [216]<5 ms), and long-term stability.Lee et al. proposed a packaged hemisphere-array-structured TENG [186] and used it as a selfpowered sensor array, similar to the approach used in another self-powered TENG (i.e., indium-tin-oxide electrode-PDMS). [187]hereby, sensor arrays based on TENGs can provide potential applications for human-electronic interfacing, artificial skins, and self-powered systems.
Regarding sensor array structures with micropatterns, Zhu et al. introduced a highly sensitive and wide-range pressure sensor based on PU mesoscale dome arrays (≈300 domes) embedded with gradient-distributed Ag-NWs [188] based on a glass template method.This novel hybrid architecture enabled the PU/Ag-NW pressure sensor to be effectively used as a highperformance and flexible electronic array, that is, showing good flexibility and stability and a short response time (20 ms).In another study, an active piezoelectric sensor based on aligned poly(vinylidene fluoride) (PVDF) nanofiber arrays was constructed by coating PANI components on the PVDF nanofiber arrays [165] demonstrating effective self-powered human-health monitoring.In another study, owing to the good microstructure of an octopus sucker, an octopus-inspired flexible bionic sensor was fabricated with an improved sensing performance.It exhibited a high sensitivity of 0.636 kPa À1 , rapid response time (≈40 ms), broadsensing range (8 Pa to 500 kPa), and excellent durability. [189]his study could provide reliable strategies for tracking human activities and using bionic manipulators to grasp objects.Overall, the structure and sensing layer are two key requirements for array-like electromechanical sensors, as they demonstrate important performance parameters, including sensitivity and sensing range.Owing to the large-scale integration of an entire flexible substrate into a sensor array, any corresponding approach should be low-cost, lightweight, flexible, and sustainable.However, further strategies are required to successfully integrate sensors with satisfactory electrical and mechanical properties for various applications, such as in soft robotics, biomedical devices, and stretchable electronics.

Human Motion Detection
Electromechanical sensors have improved daily lives by providing intelligent and advanced functions and health monitoring of body activities, such as human motions.However, realizing effective electronic devices for personal healthcare and thermal management of the human body remains a big challenge.In this context, the increase in devices connected to the Internet of Things has promoted significant developments in portable and wearable devices.
For example, a smart glove has been used to track the bending gestures of hand fingers with long-term stability and high accuracy. [190]Thus, conductive elastomers with extremely low hysteresis were produced using a 3D printing technique, in which highly stretchable dielectric elastomers were combined with a polar hydrophobic ionic liquid and then polymerized using UV light.The strain sensors were integrated into the finger surfaces or directly printed on the finger surfaces of the intelligent glove and demonstrated highly sensitive and accurate electric sensing signals upon minor bending of the fingers.Generally, using sophisticated stretchable conductors is a promising, simple, quick, and large-scale printing process, concomitantly providing high stretchability, precision, multiple layers, and recyclability for commercial wearable electronics.Thus, liquid metals have been developed to facilitate a simple, quick, and large-scale soft stamper-based printing process and to realize liquid metal-based capacitors and conductors.For instance, sensors created from liquid metal-elastomer composites were integrated into a wearable sensing glove to monitor the grasping motions of a human hand (all finger joints) effectively. [191]The advantage of liquid metal microstructure control was highlighted by creating all-soft-matter stretchable capacitive sensors with tunable sensitivities.Similarly, the liquid metal wire-based capacitors were used as strain sensors placed onto the finger joints and effectively detected their bending movements. [192]The liquid metal wirebased capacitive sensor was integrated into the finger joints to detect Chinese number gestures from 0 to 10 according to the motion of each finger and the various bending levels (i.e., no bending: À1, slight bending: 0, and fully bending: 1).
Aside from finger motion detection, electromechanical sensors have also been effectively used to monitor the motions of other human joints.Among the organic piezoelectric materials, poly(vinylidene fluoride) and its copolymers are utilized mainly owing to their good piezoelectric coefficients; however, their piezoelectric coefficients are still lower compared with those of piezoelectric ceramics, which limit the potential practical applications of self-powered sensors.Thus, two approaches can appropriately enhance piezoelectric properties of poly(vinylidene fluoride) and its copolymers: 1) increasing the piezoelectric phase content and 2) further improvement of the remnant polarization.For example, a piezoelectric film pressure sensor was fabricated by electrospinning poly(vinylidene-fluoridetrifluoroethylene) and then coupled with a 2D active nanomaterial (e.g., MXene) to produce a self-powered linear pressure sensor. [193]The active MXene has many surface functional groups and high electrical conductivity that enable interaction with the poly(vinylidene-fluoride-trifluoroethylene) dipoles and increase the poly(vinylidene-fluoride-trifluoroethylene) polarization during the electrospinning process, respectively.In addition, a new wearable film sensor was developed by coating MXene onto a low-entropy structured piezoelectric poly(vinylenefluoride-trifluoroethylene) mat containing aligned nanofibers, which was capable of sensing multidirectional mechanical stimuli, owing to the constructed piezoresistive-piezoelectric hybrid effect. [194]These functional fiber mats provided an anisotropic inplane conductive network for 2D in-plane strain sensing behavior and arranged ferroelectric crystals in nanofibers with piezoelectricity, which detected out-of-plane dynamic pressure effectively.In addition, a self-powered piezo-organic-electronic-skin sensor was fabricated and used as an active piezoelectric sensor by coating a PANI component on aligned poly(vinylidene fluoride) nanofibers and then encapsulating the combination with a PDMS matrix. [165]Different human activities were detected and monitored, including neck stretching, wrist bending, arm compressions, throat movements while drinking water, swallowing, and coughing.Furthermore, various specific phonation recognition, heart-pulse measurements, and short-time Fourier transform analyses demonstrated that this approach is convenient and efficient for tracking human health conditions.Thus, a self-healing hydrogel comprising a CNT film and conductive hydrogel could be placed on the human body to output resistance signals in real time in response to movements, such as finger and elbow bending, demonstrating its effectiveness in producing self-healing hydrogels with high performance for future flexible electronics. [164]A highly sensitive wearable sensors based on rGO-patterned paper rings [195] detected pulses and motions of human joints in the wrists, knees, and fingers.A 99% metallic CNT-PDMS composite thin film was used as a transparent strain sensor to detect human emotions and activities by placing it on different spots of the human skin to detect various motions, such as motions of a finger or wrist joint, swallowing, frowning, and smiling. [39]Another flexible pressure sensor based on a 3D microporous CNT network-attached thin porous PDMS sponge was effectively applied in a flexible foot insole, serving as a wearable healthcare gait tracking device. [136]In another study, a fibershaped graphene skeleton and CNT branches were combined to produce a carbonaceous hybrid-based highly sensitive fiber for tracking sitting postures and correspondingly deterring lumbar disc and cervical spondylosis herniation. [196]In another study, a morphological surface with a spinosum microstructure of a random distribution based on a combination of an rGO-on-abrasive paper template and PDMS elastomer displayed potential applications for detecting physiological signals of the human body (e.g., heartbeats, phonation, and respiration) and human movements (e.g., pushups, arm bending, and walking. [174]The sensor array was also applied to detect supination, neutral, and pronation gait states.The results showed the potential applications of the sensor in monitoring human activities, constructing replacement prosthetic devices, tracking daily and sports activities, and continuous health monitoring. The use of green resources is crucial for energy and environmental sustainability; hence, organic piezoelectric materials are important in the fabrication of flexible and wearable electronics owing to their good biocompatibility, high flexibility, low density, and simple processes.However, appropriately controlling the piezoelectric performance of biomaterials in the fabrication of green, low-energy-consumed flexible pressure sensors with high mechanoelectrical output performance is still a challenge.Liu et al. [197] reported a green and high-output self-powered force/ pressure sensor, which included an oriented silk fibroin mat and a microstructured electrode.The sensor exhibited a fast response of 3.4 ms, sensitivity of 30.6 mV N À1 , good durability, high signal linearity, and power density of 5.9 mW m À2 .In addition, this sensor was also applied in oral healthcare by monitoring the biting force of a dummy dental cast.In another study, a bioinspired strategy was also proposed for producing a piezoresistive-typed pressure sensor using a graded porous material. [198]The sensors were fixed on a bicycle wheel to track the tire-pavement pressure and on human skin to detect biosignals, including the venous and arterial blood pressure pulses (Figure 10B).In another study, Dinh Le et al. proposed a skin-attachable ultrasensitive and ultrarapid acoustic sensor with a self-cleaning ability based on a rGO/PDMS composite film with bioinspired microcracks and hierarchical surface textures. [199]Owing to the synergetic effect of the spider-slit-organ-like multiscale jagged microcracks and lotus-leaf-like hierarchical structures, acoustic vibrations were detected even under high signal-to-noise ratios in an audible frequency range of 20-20 000 Hz. Thus, the excellent performance demonstrated an anti-interference-based perception of human voices with high precision, even in noisy environments, while assisting voice-related control.A highly sensitive and stretchable rGO-organohydrogel composite strain sensor was produced in another study. [200]The sensor exhibited a high hydrophobicity (122°) and can be used for real-time and continuous tracking of human activities in extremely cold (À60 °C), dry, and underwater conditions.In another study, graphene-PANI-embedded PE oxide composites were used as flexible pressure sensors [201] after being fabricated using a modified electrohydrodynamic jetting method.The resultant real-time detection of underwater activities demonstrated the potential for healthcare, environmental, and bio-related monitoring.Further, this sensor was applied to detect minute signals during physical activities, such as finger and hand gestures, acoustic vibrations, cough, and facial expression, as well as to detect real-time pulse waves of near-body states.In addition, these sensors were employed to measure the bio-potentials or currents generated by various organ activities for the early detection, diagnosis, and recovery tracking of various diseases, [202,203] such as Parkinson's disease [204] and nerve/muscle-related dysfunctions. [205]Generally, these sensors were developed and effectively applied to monitor both healthy and diseased individuals.

Temperature Monitoring
Skin-related temperature monitoring determines human body conditions.It can also reveal information for health and disease diagnostics (e.g., inflammation and fever).Therefore, temperaturesensing devices should be nontoxic and mechanically soft to minimize adverse effects on the skin and negative impacts on signal detection.The temperature-sensitive layers contained in/on the sensors are an important factor in evaluating their thermalsensing capacities.A temperature sensor based on a combination of a micropatterned silk-protein substrate and PEDOT:PSS conductive ink was presented with high sensitivity and accuracy and demonstrated transparency, flexibility, durability, conformability, and biodegradability. [206]The sensor exhibited a high sensitivity of À0.99% Â °CÀ1 at a temperature range of 20-50 °C, owing to a temperature-sensitive layer containing photoactive silk sericin and rGO.This suggested that it can monitor skin temperatures with rapid and accurate measurements.Furthermore, this sensor degraded under proteolytic conditions (≈10 d); thus, it can be used for personal health tracking to ensure efficient disease detection and diagnosis.In another study, Li et al. developed high-performance sensor arrays containing a laser-carbonized carboxymethylcellulose/GO composite for detecting mechanical and thermal stimuli using a laser. [207]The sensor array exhibited a temperature coefficient of resistance of À0.289% Â °CÀ1 at a temperature range of 25-100 °C.Moreover, an integrated device based on laser-carbonized carboxymethylcellulose strain and laser-carbonized carboxymethylcellulose/GO temperature sensing arrays enabled real-time measurements of both thermal and mechanical deformations.
Ionic conductors are also promising for producing temperature sensors; however, ionic conductive hydrogels and organogels often suffer from liquid leakage and evaporation problems.Zhou et al. produced a liquid-free ion-conducting elastomer composed of dry lithium bis(trifluoromethane sulfonimide) and a PU elastomer. [208]It exhibited an ultrahigh stretchability (>900% and GF = 3.21), excellent ionic conductivity, a high sensitivity (2.22% Â °CÀ1 at a temperature range of 25-60 °C), and environmental sustainability.Thus, the as-obtained sensor can accurately and repeatedly detect human motion and temperature changes, demonstrating its potential for digital medical diagnosis and monitoring.Chen et al. proposed wearable strain and temperature sensors based on multifunctional conductive hydrogel/thermochromic elastomer hybrid fibers with core-shell segmental configurations [209] for monitoring human motions and body/surrounding temperatures.The sensors were fabricated from a conductive hydrogel containing an rGO-doped poly(2-acrylamido-2-methyl-1-propanesulfonic acid-co-acrylamide) hybrid and a thermochromic elastomer containing an SIS block copolymer and thermochromic microcapsules, which served as stretchableand thermal-sensitive materials, respectively, using a dual-core coaxial wet-spinning method.The hybrid fiber had multifunctions, such as human motion tracking and body/room temperature detection.This study could be extended for the mass production of wearable strain and temperature sensors.Liu et al. proposed ionic conductive hydrogels with dynamic crosslinks, [210] which were related to a chitosan-poly(acrylamide-coacrylic acid) double-network hydrogel with dual-dynamic cross-links composed of a chitosan physical network and ionic coordination.The sensor was freezing-tolerant, highly sensitive, and durable.More importantly, its sensitivity effectively detected strains and pressures at room and subzero temperatures.Thus, this study could provide a platform for constructing and applying high-sensitive pressure and strain hydrogel sensors with good durability over a broad temperature range.In another study, a 3D honeycomb-structured conductive organohydrogel was fabricated by integrating carbonaceous materials (i.e., carbon black and CNT) into a poly(vinyl alcohol)/glycerol organohydrogel based on physical cross-linking networks (e.g., hydrogen bonds). [211]The organohydrogel-based sensors exhibited high stretching sensitivity to tensile strains (≈600%) and temperatures (À0.935%Â °CÀ1 ) and could detect full-range human physiological signals and respond to changes in temperature, thereby serving as multifunctional wearable electronic devices.Hong et al. fixed the temperature-sensing field-effect transistors on PET islands arranged on a stretchable PDMS elastomer substrate to stretch the sensor to 50% without reducing its performance. [212]enerally, temperature-typed sensors can be developed as thermal flow sensors to determine arterial blood flow rates. [213]In addition, the tracking of the thermal diffusivity and conductivity of the human skin has also been investigated using laminatedon-skin microthermal heaters. [214]

Human-Machine Interfaces
Human-machine interfaces facilitate bidirectional communication between humans and robots, mobiles, household objects, computers, and other machines through sensing functions and corresponding feedback.As discussed above, in the context of health monitors, there are various modes for collecting data from the human body.Each approach provides corresponding details of the information.The electromechanical sensors that enable human-machine interactions use sensing signals to provide different user feedback.The studies discussed below are arranged into two specific categories of human-machine interface applications: robots and others.

Robots
With the rapid development of advanced technologies, research on the use of prosthetic limbs (robotics) has focused extensively on assisting people with disabilities related to limbs, such as weak limb functions or limb amputations.Electromechanical sensors have been explored as electrical stimulators for controlling prosthetic limbs or robotics.Human-machine interface devices can provide new avenues for interacting with human users to control prosthetics, robotics, and healthcare.
Among them, sensor-like TENGs can be utilized as humanmachine interfaces to manufacture smart and interactive products, as electromechanical sensors exhibit advantages for robotic control applications.Chen et al. reported a triboelectric sensor with self-powered and flexible abilities.It was developed into two sensor patches comprising 2D (grid layer and four electrodes) and 1D triboelectric sensors to control in-plane and out-ofplane robotic movements, respectively. [215]These sensors were placed on a human forearm and finger to monitor their continuous sliding information to control the acceleration, velocity, and trajectory of a robotic arm (3D robotic manipulator).These flexible sensors comprised environment-friendly materials, including a starch-based hydrogel, PDMS elastomer, and silicone rubber.This study demonstrated the potential applications of the sensors in robotic control, electronic skins, and touching screens owing to their facile design and low-cost materials.Lim et al. successfully proposed a wearable human-machine interface system consisting of transparent and stretchable sensors and stimulators. [37]The piezoelectric motion sensor was based on poly(methylmethacrylate)@graphene@poly(L-lactic acid)/SWCNT layer@graphene@ poly(methylmethacrylate), whereas the electrotactile stimulator and main components included epoxy@graphene/Ag-NWs/graphene@epoxy@PDMS.This system exhibited excellent performance and low power consumption.In addition, it was applied to control various motions of a robotic arm and exhibited feedback stimulation under successful executions of commands.
Zhang et al. presented another facile and cost-effective approach, which used metallic sandwiched-aerogel hybrids, including sandwich-like 2D vanadium nitride nanosheets with vertically aligned N-doped CNT arrays. [216]The aerogel hybrids were successfully used as flexible strain sensors, exhibiting a high sensitivity (GF = 386 at ε = 10%), rapid response, and extraordinary durability, which could be attributed to the excellent structural compatibility of the 2D vanadium nitride nanosheet-based main bone structure during the repetitive deforming process.The fabricated sensors were effectively investigated for monitoring physical signals and an actual real-time control system for a robot hand (Figure 10C), indicating high potential for applications in human-machine interactions.Additionally, highly sensitive wearable sensors based on rGO-patterned paper rings were also applied to investigate a 3D-printed robotic hand, and an rGO-patterned paper keyboard was utilized to activate light-emitting diodes effectively. [195]Another smart composite based on a combination of carbon black particles and a wrinkle Ecoflex matrix was successfully fabricated to enable a stretchable sensor to achieve a wide strain range of 500%, a good sensitivity (GF = 67.7), a short response time (120 ms), and high recyclability. [217]The carbon black/Ecoflex composite strain sensor could effectively monitor various human activities and vehicle security crash anthropomorphic situations.Moreover, it exhibited excellent real-time detection and tracking of robotic finger motions.The abovementioned approaches manifested potential applications in rehabilitation training for disabled people and posture simulations.Hu et al. used a graphene fiber skeleton and CNT branches to produce a highly sensitive carbonaceous hybrid-based fiber [196] with a high GF of >1127, short response time of 70 ms, and excellent reliability and stability.More importantly, this sensor was successfully applied in a real-time human-machine control system to track a sitting posture to prevent cervical spondylosis and lumbar disc herniation.Therefore, this study provided a feasible and scalable approach for manufacturing ultrasensitive fiber-based sensors applicable for tracking human physiological signals and control systems of human-machine interfaces.
Zhu et al. [218] and Lopes et al. [219] explored and successfully demonstrated a smart and soft glove with multimodal sensing and feedback functions.The triboelectricity-type soft modular glove was used to sense static and dynamic contact, stretching, and vibrations [218] and was designed to serve as a sensing and feedback device for diverse communications among humans, machines, and the virtual world through intelligent perceptions.This glove not only provided real-time detection of clever hand motions and direct feedback but also facilitated smart object recognition and better feedback when coupled with a machinelearning algorithm.Thus, using the gloves could significantly enhance the communication and perception of more comprehensive information.In another study, a novel Ag-In-Ga (i.e., Ag ink and eutectic gallium-indium liquid metal alloy) circuit was produced for printing on a transfer-tattoo paper or hydrographic paper before being transferred to a human body or a 3D surface. [219]Both the interfacing and printing processes were conducted at room temperature.The use of the electronic tattoo on the skin surface for acquitting electromyography signals, an interactive circuit with touch buttons, and LEDs transferred over the 3D printed shell of a robotic prosthetic hand, and a proximity measurement skin transferred over a 3D surface.A biaxial strain sensor based on a CNT film and a microdome array-patterned PDMS substrate was effectively designed and developed for monitoring human motions and human-robot interfaces. [220]In more detail, a robot hand control system was demonstrated using a smart glove integrated with this strain sensor.All fingers of the robot were actuated using a simple tendon-driven system with five servo motors controlled by the voltage outputs from this stretchable sensor.In addition, the different robot fingers were manipulated according to the bidirectional responses of the intelligent glove.
Furthermore, human-machine interactive sensors based on hydrogels and ionogels are used as multifunctional stretchable ionic skins to mimic human skin sensations.Currently, physically cross-linked ionogels with good stretchability, self-adhesion, and self-healing ability are investigated; however, most of them have poor elastic recovery under large deformations and weak thermomechanical stability at high temperatures.Therefore, to achieve excellent performance in terms of mechanical features, adhesion, self-healing ability, and stability, Hao et al. successfully developed ionogels using hyperbranched polymer covalent-crosslinked poly(zwitterion-ionic liquid)-co-poly(acrylic acid) and multiple dynamic bonding cross-linked networks. [221]The resultant ionogels exhibited an extremely high stretchability (>10 000%), ultrastrong adhesion (>6.8 MPa), an ultrafast self-healing ability (10 s), good thermal stability (À60À250 °C), and good 3D-printability.Owing to their excellent performances, they were used in a highly stretchable strain sensor and self-powered TENG touch sensor.In particular, the obtained ionogel-based sensors were applied for the recognition of human motions and for real-time wireless control of robots.They could self-heal and recover immediately upon mechanical damages, demonstrating their potential for use in wearable sensors and human-machine interfaces.Chen et al. [222] proposed a multifunctional nanocomposite hydrogel based on mixing nanosized hybrids of Ag-NPs-tannic acid@GO into a polyacrylamide hydrogel matrix.In addition to its outstanding performances, that is, a stretchability of 1250%, the conductivity of 0.15 S⊡m À1 , GF of 3.1, high linearity, longterm durability, and high precision for human motion monitoring, the nanocomposite hydrogel was effectively used in artificial intelligence, such as human-robot interface sensors, electronic skin, and information encryption sensors.This hydrogel enabled a robot arm to operate the touch screen of a smart phone and to write information encryption.Thus, this study was used as a basis for nanocomposite hydrogel-based multifunctional wearable sensors.In another study, Sim et al. fabricated semiconductor nanomembrane electronics containing sol-gel-on-polymerprocessed indium zinc oxide based on a one-step formation approach. [223]This approach allowed the sensor to be used as a multifunctional stretchable human-machine interface device through simple manufacturing and robust interfacing.Thus, this device can be directly worn by humans and can be placed on robotics.It can offer intelligent feedback and be a temperature delivery device for microheaters.Generally, electromechanical sensors receive significant attention in the fabrication of human-robot interface equipment.Soft sensors have recently focused on wearable applications for monitoring human motions and human-robot interfaces for effectively measuring multidirectional deformations.Developing functional materials with suitable stretchability, rapid sensitivity, wide-range measurement capabilities, and good linearity has been an important challenge.

Others
In addition to human-robot interfaces, intelligent objects based on the interactions between human activities and other machines have also been developed to improve the convenience of living, including those based on electromechanical sensors.Typically, these sensors use interactions with the human body to control household objects such as fans, lights, and microwave ovens.In particular, human-mechanical interface devices based on TENGs have been extensively investigated to produce smart home systems.Zhang et al. created a sensor-like TENG for a breath-based human-machine interface to control household objects. [224]The TENG acted as a self-powered sensor to deliver control commands through the breathing action, and subsequently, the signals were processed and transmitted to a fan and light fixture without requiring physical movements or language (Figure 11A).Cao et al. [225] explored a multifunctional electronic textile based on a self-powered triboelectric gesture textile, including two fabric layers stacked into a printed layer of PU/CNT electrodes.Owing to the advantages of the interactions between the CNT component and fabrics, the electrodes exhibited excellent stability against mechanical deformations and even after being washed.Importantly, the sensor array acted as a trigger to appropriately control electrical appliances, such as light Figure 11.(A) Application of the sensor-like TENG in a smart wireless human-machine interface system: sensor-like TENG was mounted on a mask, and TENG-based human-machine interface system used to (i, ii) wirelessly control an electric fan, and (iii, iv) wirelessly turned on/off a lamp via deliberate breathing.Adapted with permission. [224](B) Application of the PU/CNT-based sensor in wireless control of household objects, such as light bulbs, electric fans, and microwave ovens.Adapted with permission. [225](C) Application of self-powered triboelectric sensors for triggering a wireless alarm system: (i) finger slightly touching the sensor, (ii) stepping on top of the sensor embedded underneath a carpet, (iii) grabbing the sensor as applied to a door handle, and (iv) switching a panel light by touching slightly the sensor.Adapted with permission. [226]ulbs, electric fans, and microwave ovens (Figure 11B).Furthermore, multidimensional sensors can be integrated with a Bluetooth module, and Arduino assists in creating a portable measurement system for wireless recording (based on changes in the capacitance or resistance in response to various applied deformations).
Zhu et al. produced flexible thin-film devices that relied on contact triboelectrification to create a voltage signal in response to physical contact without the use of an external power source. [226]Diverse applications were effectively demonstrated and indicated that the flexible sensor could sensitively trigger functional devices in response to external stimuli, such as finger touching, foot pressing, and hand grabbing (Figure 11C).When this sensor was impacted, the output voltage would trigger a siren to activate an alarm with a flashing light.It could also be used as a touch-enabled switch for a panel light (Figure 11C).These manifested various immediate uses for human-electronics interfaces, remote operations, surveillance, automatic control, and security systems.Pu et al. successfully developed mechanosensation human-machine interfaces based on sensor-like TENGs to control household objects through eye movements. [227]These sensors contained a multilayered structure with a PET-based thin layer as a tadpole-shaped supporting substrate.Fluorinated ethylene propylene was used as the electrification layer and was coated by indium tin oxide to serve as a black electrode and was then laminated onto the PET substrate.Natural latex was used as another electrification layer owing to its ultrahigh elasticity, air permeability, long-term durability, and skin-friendly nature for sensitive eyes.The sensor was mounted on glasses to monitor human-machine interfaces, such as interactions between eye movement and household objects (table lamp, electric fan, and doorbell), and a wireless hands-free typing system; thus, it provided a promising design concept for smart sensor technologies and suggested practical applications for mechnosensational human-machine interfaces.
In addition, the control of virtual objects has also been investigated in the context of using electronic devices as humanmachine interfaces for virtual control.Normally, virtual object control based on human-machine interfaces can improve the quality of life.Electromechanical sensors have indicated usefulness for virtual object control, particularly for human-computer interactions.Touch panels are often mentioned for their stretchability and biocompatibility for integration with the human body.Rather than using stiff and brittle electrodes, Kim et al. developed an ionic touch panel based on a combination of polyacrylamide hydrogel and lithium chloride salts, achieving a high transmittance of 98% for visible light and sustaining a large deformation (1000% strain) without losing its functionalities. [228]A surfacecapacitive touch system was used to sense a touched position based on the current flow from the ionic touch panel to the finger, as demonstrated by writing words, playing games, and playing piano.Takamatsu et al. explored a textile-based humanmachine interface device used as a wearable keyboard. [229]It was prepared by patterning a conductive polymer (PEDOT:PSS) onto a knitted textile and then coating the brush-painted electrodes with a PDMS substrate to study the changes in capacitance.The wearable keyboard acted as a pressed sensor with a high spatial resolution, even when placed on the human forearm.Similarly, He et al. fabricated a motion sensor-like TENG for virtual control.
The TENG was created from a glove-based human-machine interface with PEDOT:PSS-coated textiles. [230]Consequently, different voltage responses were achieved when the PEDOT: PSS-containing TENG was bent.The subjects drew letters in a computer program and scrolled through a website with various motions of the self-powered glove-based intuitive interface.In another study, a flexible pressure sensor based on a 3D microporous CNT network-attached thin porous PDMS sponge was employed in a flexible piano pad as a human interface entertainment device. [136]he sensors were also explored for brain-computer interfacing and object operations in video games.For instance, Norton et al. used long-term epidermal electrodes comprising Au and PI constructed into fractal layouts and mounted on multiple positions of the human head for conventional electroencephalogram recording systems. [231]Correspondingly, a speller-based paradigm was set up to provide an algorithm for guessing the letters from the obtained electroencephalogram signals when the electrodes were mounted onto human skin.The letters of "computer" were spelled, and the algorithm indicated similar classification results for both long-term epidermal and conventional electrodes, that is, the algorithms could classify the corresponding letters through the subject's focus and thoughts while flashing various characters via the screen and observing the electroencephalogram signals obtained via the long-term epidermal electrodes.Another approach developed a mechanoacoustic sensing capacity for virtual control. [232]Liu et al. developed a stretchable hybrid electronic device for detecting speech signals from the vocal cords.The classified signals effectively controlled a virtual character in a video game. [232]They proposed applications for skin-integrated digital techniques based on the acoustics of the human body.Generally, new electromechanical sensors are emerging to obtain information from the human body and its surroundings based on smart, cost-effective, friendly, and real-time approaches for effective integration with machines, making them essential to such approaches.

Conclusions and Outlook
In summary, we investigated electromechanical sensors and reviewed their developments in the context of the four major operating mechanisms of triboelectricity, piezoelectricity, piezoresistivity, and piezocapacitance.Their applications revealed excellent working performances and consequently endowed good results in detections of human motions and temperatures and in human-machine interfaces.Advances in novel functional materials and structures show potential for the improvement of electromechanical sensors in the future, as well as the need to meet requirements for further development.These developments can substantially accelerate the development of advanced technologies and human-machine interface systems.Owing to the excellent advantages of tactile-typed electromechanical sensors, highly integrated tactile sensing systems have emerged as an important theme for both scientific research and practical applications.This suggests that flexible integrated circuits and other electronic devices should match with the growth of tactile-typed sensors to satisfy the requirements for potential applications in medical health monitoring and human-machine interactions.However, several challenges remain, particularly in regards to demonstrating multidirectional deformations.Thus, future efforts for soft-skin-mountable sensors should be developed to resolve these challenges.Therefore, possible relationships between functional material structures and characterizations, as well as requirements for producing high-performance sensors, must be comprehensively acquired and understood.Furthermore, new manufacturing technologies and approaches are required to develop electromechanical sensors.Recently, several potential methods based on nature have been explored to prove the efficient utilization of high-performance sensors.Printing technology is another promising approach for developing highly functional, economic, fashionable, and practical sensors, even for mass production.However, these methods should address some of the existing challenges, such as by 1) exploration of new sensing devices based on functional nanomaterials, hybrid composites, and simplified thin-film processing approaches; 2) investigation of biological contacts between the human body and substrate; 3) identification of multimodal sensing conditions; and 4) development of novel methods and structures based on 3D printing technologies.Much remains to be developed for effective printing procedures based on integrating electromechanical sensors, such as those with large areas, low costs, and recyclability.Therefore, understanding the details of the mechanics of these electronic devices is crucial to use them effectively.

Figure 1 .
Figure 1.Schematic and applications of electromechanical sensors designed from functional composites and structures containing conductive materials into stretchable matrix.

Figure 2 .
Figure 2. (A)Response curves of ΔR/R 0 against compressive strain and pressure for composite foams and the schematic of the sensing mechanism (adapted with permission[9] ).(B) Response curves of ΔR/R 0 against different strains for the electromechanical sensors (the curves of linearity and GF values were plotted versus strain) and the schematic of the sensing mechanism.Adapted with permission.[10]

Figure 5 .
Figure 5. Schematic of TENG-based triboelectric sensors used in various applications.

Figure 6 .
Figure 6.Tasks and challenges of TENG-based triboelectric sensors in various applications.

Figure 7 .
Figure 7.Chemical structures of common functional materials.

Figure 8 .
Figure 8. Schematic of various composite structures for electromechanical sensors: (A) conductive network on the substrate surface, (B) conductive network in the stretchable matrix, and (C) array-like electromechanical sensors.

Figure 9 .
Figure 9. (A) Transparent Ni-based conducting panels: (i) SEM images of various mesh-type Ni electrodes on a glass substrate and (ii) digital captures of different pattern-based Ni electrodes on a PET substrate.Adapted with permission.[148](B) Ag particle/CNT/PDMS stretchable conductive adhesive: (i) schematic of the preparation of stretchable conductive and printed adhesives and (ii) schematic of a stretchable conductive adhesive (adapted with permission[155] ).(C) Manufacture and structure of the triode-mimicking positive graphene/Ecoflex sensor: (i) schematic of the preparation of the sensor, (ii) schematic layout of a single sensor, and (iii) digital capture of the assembled sensor (adapted with permission[163] ).(D) PDMS-covered 3D porous conductive foam: (i) schematic of the PDMS-covered conductive foam and SEM images of (ii) 3D porous and (iii) PDMS-covered 3D porous conductive foams.Adapted with permission.[170]

Figure 10 .
Figure 10.(A) MoS 2 /graphene foam/Ecoflex nanostructure-based pressure sensor array: (i) schematic of a flexible sensor array of 3 Â 3 pixels fabricated using MoS 2 /graphene foam/Ecoflex, (ii) digital captures of MoS 2 /graphene foam/Ecoflex with various movements, and (iii) digital capture of a sensor array (3 Â 3 pixels).Adapted with permission.[65](B) Applications of the flexible pressure sensor using a graded porous material: (i) digital photos of ant nests, (ii) schematic of the flexible pressure sensor, (iii) schematic of a bent finger with the sensor, and (iv) applications of the sensor with high sensitivity and wide detection range.Adapted with permission.[198](C) Application of the vanadium nitride/CNT strain sensor for detecting finger motions: (i) control circuit of the robot hand system containing five signal acquisition circuits and (ii) digital captures of instant controls of the robot hand by investigating the gestures from "five" to "one".Adapted with permission.[216]

Table 1 .
Comparison of currently used piezoresistive pressure (or tactile) sensors.

Table 2 .
Operation mechanism and advantages and disadvantages of various electromechanical sensors.

Table 3 .
Advantages and disadvantages of functional materials for the construction of electromechanical sensors.

Table 4 .
Hansen solubility parameters a) for CNTs dispersed in different solvents.

Table 5 .
Descriptions of synthetic methods for CNTs.The floating catalytic CVD is known as the aerosol-aided approach; it uses an ultrasonic bath to keep the solution homogeneous outside the reactor, which controls the size, quality, and purity of the CNTs produced

Table 6 .
Advantages and disadvantages of some fabrication techniques for electromechanical sensors.