Practical applications of next-generation stretchable electronics hinge on the development of sustained power supplies to drive highly sensitive on-skin sensors and wireless transmission modules. Although the manufacture of stretchable self-charging power units has been demonstrated by integrating stretchable energy harvesters and power management circuits with energy storage units, they often suffer from low and unstable output power especially under mechanical deformation and human movements, as well as complex and expensive fabrication processes. This work presents a low-cost, scalable, and facile manufacturing approach based on laser-induced graphene foams to yield a self-powered wireless sensing platform. 3D porous foams with high specific surface area and excellent charge transport provide an efficient flow of triboelectric electrons in triboelectric nanogenerators. The surface coating or doping with second laser irradiation on these foams can also form a 3D composite to provide high energy density in micro-supercapacitor arrays. The integration of a triboelectric nanogenerator and power management circuits with micro-supercapacitor arrays can efficiently harvest intermittent mechanical energy from body movements into stable power output. 3D foams and their composites patterned into various geometries conveniently create various deformable sensors on large scale at low cost. The generated stable, yet high, power with adjustable voltage and current outputs drives various stretchable sensors and wireless transmission modules to wirelessly measure pulse, strain, temperature, electrocardiogram, blood pressure, and blood oxygen. The self-powered, wireless, wearable sensing platform paves the way to wirelessly detect clinically relevant biophysical and biochemical signals for early disease diagnostics and healthy aging.

Epidermal electronic devices that can conform to hierarchically textured skin surfaces allow wireless and sensitive detection of human health conditions for personalized healthcare and telemedicine. Their long-term operation to continuously detect high-fidelity physiological signals in real time hinges on the development of sustainable, stretchable, and environmentally friendly power supplies. One promising solution is to integrate high-performance epidermal electronic sensors and signal processing circuits with stretchable energy harvesters, energy storage devices, and power management circuits, as well as wireless data transmission components for fully self-powered wearable electronics. Currently, electrochemical energy storage devices such as batteries or supercapacitors have been integrated with energy harvesters (based on electromagnetic induction,1,2 piezoelectric,3,4 or triboelectric5,6 effects) to power wearable biophysical and biochemical sensors.7,8 However, the fabrication often involves patterning of nanomaterials (e.g., gold or silver nanowires, copper nanowires, graphene, and carbon nanotubes) on flexible substrates through complex fabrication processes (e.g., photolithography, with deposition or etching). The high-temperature processes in device fabrication can also damage the active functional material and deform the elastomeric/polymeric substrates. Consequently, the resulting self-powered wearable electronic devices are associated with compromised performance, expensive precursor materials, costly fabrication facilities, complex fabrication processes, and serious environmental safety concerns. Furthermore, most of the systems with varying materials and working principles suffer from low output power and unstable output signals upon mechanical deformation and human/skin movements/motions.

In this work, we report a laser-induced graphene (LIG) foam-based self-powered wireless, wearable sensing platform by using a low-cost, scalable laser scribing fabrication approach for continuous, real-time monitoring of human healthcare conditions. The photo-thermally converted LIG foams9 by a CO2 laser showcase high specific surface area and charge (ionic/electrical) conductivity, as well as mechanical, thermal, and chemical stabilities. The 3D porous network structure and excellent charge transport of the LIG allow its use as an electrode in the triboelectric nanogenerator (TENG) to efficiently drive the flow of triboelectric electrons for mechanical energy harvesting. With coating or a second laser irradiation process, the LIG can be further modified to form a composite with layered structure or embedded nanocrystals (e.g., metals or metal oxides). The LIG electrodes and their composites with high ion-accessible surface area and low ion transport resistance can also provide high energy and power densities in micro-supercapacitor arrays (MSCAs) for energy storage. The direct patterning of the LIG into various geometries also allows the creation of various deformable sensors in large scale at low cost. In the self-powered wearable sensing platform, the sensors can be directly driven by the sustained LIG-based power supply that integrates TENG with MSCAs to continuously scavenge mechanical energy from human motions for charging. With a Bluetooth-based wireless transmission component for real-time monitoring, the system with representative biophysical sensors that continuously capture, process, and wirelessly communicate the clinically relevant data from human skin to a mobile user interface demonstrates its potential in healthcare, fitness, and security monitoring.10,11

The LIG-based self-powered, wireless, wearable sensing platform exploits MSCAs charged by a TENG with a rectification circuit to power on-skin sensors, signal processing units, and wireless data transmission components [Figs. 1(a) and 1(b)]. As shown in the schematic circuit diagram [Fig. 1(c)], the kinetic energy harvested by the TENG from human motions is regulated by a bridge rectifier for direct current output to charge MSCAs. As a result, the inherent intermittence and fluctuation of the harvested energy are stored and converted into sustained power supply for continuous and long-term operation of the integrated system. Without loss of generality, the representative biophysical sensors selected in the current demonstration include those to detect pulse, strain, temperature, electrophysiological data such as electrocardiogram (ECG), blood pressure, and blood oxygen. Elastomeric devices substrate such as silicone elastomers (e.g., Ecoflex) or bio-adhesive hydrogels12 also provide an intimate interface to the skin for robust signal detection against motions [Fig. 1(b)]. The detected signals can then be processed and wirelessly transmitted to a mobile user interface via a Bluetooth-based wireless module.

FIG. 1.

Design concept of the LIG-based self-powered stretchable wireless sensing platform. (a)–(c) Schematic illustration and optical images of the LIG-based self-powered, wireless, wearable sensing platform, with the energy harvested by TENGs from kinetic human motion to charge MSCAs as the sustained power supply for powering on-skin sensors, signal processing units, and wireless data transmission components. Optical images of (d) the stretchable LIG-based top electrode with island-bridge layout in the TENG and (e) the stretchable LIG-based MSCAs attached on a human wrist. Scanning electron microscope (SEM) images of the highly porous LIG foam (f), Au@LIG foam (h), and MnO2@LIG foam (i). Scale bar, 5 μm.

FIG. 1.

Design concept of the LIG-based self-powered stretchable wireless sensing platform. (a)–(c) Schematic illustration and optical images of the LIG-based self-powered, wireless, wearable sensing platform, with the energy harvested by TENGs from kinetic human motion to charge MSCAs as the sustained power supply for powering on-skin sensors, signal processing units, and wireless data transmission components. Optical images of (d) the stretchable LIG-based top electrode with island-bridge layout in the TENG and (e) the stretchable LIG-based MSCAs attached on a human wrist. Scanning electron microscope (SEM) images of the highly porous LIG foam (f), Au@LIG foam (h), and MnO2@LIG foam (i). Scale bar, 5 μm.

Close modal

With a CO2 laser that is commonly available in machine shops, LIG foams with 3D networks and high charge conductivity can be readily patterned into programmed geometries from a wide range of carbon-containing materials [Fig. 1(f)]. The same material system is explored for all of the device components, including the TENG, MSCAs, and on-skin sensors (see Experimental Section, supplementary material). Scanning electron microscope (SEM) images indicate that the highly porous LIG foam exhibits interconnected 3D honeycomb-like structures [Figs. 1(g) and S1]. The size of the honeycomb pore is estimated to be in the range of 1–5 μm. The formation of the porous structure is likely attributed to gas release during the rapid laser irradiation process. The liquid nitrogen sorption measurement based on the Brunauer–Emmett–Teller (BET) method13 also confirms the porous feature of the LIG foam, revealing a specific surface area of 380 m2 g−1 and a broad pore size distribution (Fig. S2). After confirming few-layered graphene sheets by transmission electron microscopy (TEM), the porous honeycomb cell walls of the LIG foam are further investigated by high-resolution TEM (HRTEM) to show a lattice space of 0.34 nm, corresponding to the (002) planes in graphitic materials (Fig. S3). Furthermore, the associated diffraction patterns indicate a polycrystalline nature in these graphene sheets, and the ring diffraction patterns are attributed to the scrolled or folded graphene sheets [Fig. S3(d)].14 The x-ray diffraction (XRD) pattern of the LIG shows a broad diffraction peak centered at 2θ = 25.5° of the graphite (002) plane, indicating a high degree of graphitization [Fig. S4(a)]. The Raman spectrum of the LIG also shows three prominent peaks centered at 1353 (D band), 1582 (G band), and 1686 cm−1 (2D band) [Fig. S4(b)], which are associated with the defects or bent sp2-carbon bonds, the bond stretching of sp2-carbon pairs, and second-order zone-boundary phonons, respectively.15 The G peak intensity of the LIG is roughly six times that of the D peak. Notably, different from two components of the bulk graphite, a sharp 2D peak with a full width at half maximum of 57.8 cm−1 in the LIG foam further indicates a high degree of graphene formation. The dominated sp2 bonded carbon in the LIG foam observed in the x-ray photoelectron spectroscopy (XPS) (Fig. S5) agrees well with XRD and Raman results. With a sheet resistance of only 80 Ω per square on bare glass, LIG foams with a 3D porous interconnected structure and high charge conductivity provide unique application opportunities in energy harvesting/storage devices and biophysical sensors.

Most notably, 3D porous LIG foams with good hydrophilicity offer a high degree of structural and chemical tunability by surface coating or functionalization.16,17 After loading a metal-complex/metal precursor on a formed LIG surface, additional scribing with the same laser setup allows facile formation of the LIG composite with embedded nanoparticles (see Experimental Section, supplementary material). As a proof-of-concept demonstration, an in situ formation of Au and MnO2 nanocrystals in LIG foams yields hybrid nanostructures M@LIG (M = Au, MnO2) for (1) improved electrical conductivity and (2) enhanced specific surface area and electrochemical active sites, respectively. The former is essential for high-performance energy harvesting/storage devices and electrical signals detection/transmission in a self-powered wearable sensing platform, whereas the latter contributes to high electrochemical performance. In addition to the preserved overall morphology and 3D porous interconnected structure (Figs. S6 and S7), SEM images of the M@LIG hybrid nanocomposite also show uniform distribution of nanocrystals with no nanoparticle aggregation. The characterization of M@LIG hybrid nanocomposites also highlights the greatly improved electrical conductivity in Au@LIG (reduced sheet resistance from 80 to 50 Ω per square by 38%) and higher specific capacitance in MnO2@LIG (from 4.1 to 6.15 mF cm−2 × 50% at a current density of 0.1 mA cm−2). The programmable properties and functionalities of LIG and its composites offer a unique opportunity to meet the varying requirements of different applications (e.g., power outputs of energy harvesting/storage devices and sensitivity/selectivity of sensors). LIG-based top electrodes in TENGs and LIG-based MSCAs configured in the islands-bridge layout can provide (i) enhanced stretchability compared to wavy structures or kirigami layouts and (ii) adjustable voltage/current outputs through serial/parallel connections.

Due to the low cost and biocompatibility of a variety of functional materials,18 TENGs have been widely explored to harvest ambient mechanical energy into electricity by utilizing contact electrification coupled with electrostatic induction (Fig. S8). TENGs based on polymers [such as Ecoflex,14 polydimethylsiloxane (PDMS),19,20 polytetrafluoroethylene (PTFE),21,22] are low cost and exhibit high efficiency. The LIG-based TENG consisting of a top LIG foam electrode with an island-bridge architecture [Fig. 2(a)] and a bottom crumpled Ag film electrode23 used in this study is found to be highly flexible and stretchable, due to out-of-plane buckling and twisting in the top (Fig. S9) and wavy structures in the bottom (Fig. S10). The output performance of the TENG device with a pre-strain (εpre) of 100% and a contact area of ∼25 cm2 is studied by using a linear mechanical motor (external forces of 100 N) at different deformation frequencies under an external load of 1 MΩ. As the deformation frequency increases from 0.25 to 2 Hz, the output voltage and the corresponding output current increase with the frequency due to increased flow rate of the charges. However, the output voltage and current decrease as the frequency increases from 2 to 2.5 Hz, which is possibly related to the not fully released transferred charge density and compressive force during the faster deformation process. As a result, the maximum output voltage of 280 V and current of 16.8 μA are obtained at 2 Hz [Fig. 2(b)]. Embedding Au nanoparticles in the LIG further yields Au@LIG-based TENG with the output voltage increased from 280 to 320 V by 14% (Fig. S11), which is higher than that of previously reported LIG-based TENGs.24,25 As the deformation frequency increases from 0.5 to 2.5 Hz (vertical force of 100 N), the maximum transferred charges from the LIG-based TENGs first increase to ∼135 nC (5.4 nC/cm2) at 2 Hz and then decrease (Fig. S12). The high transfer of charges with relatively small variation across the frequency range showcases their excellent electrification performance. As the output power (P = VI) generated from LIG-based TENGs depends on the resistance value of the external load, a maximum output power of ∼0.47 mW (or 0.02 mW cm−2) is found at the resistance of 2 MΩ in the range from 10 kΩ to 10 GΩ [Fig. 2(c)]. The maximum output power corresponds to an average voltage and current of 396 V and 12.05 μA, respectively. This output performance is remarkable compared with previously reported stretchable TENGs, including Au nanofilm/PDMS@crumpled Ag,23 crumpled graphene/VHB/PDMS@crumpled graphene/VHB/PMMA,26 saline/silicone rubber,27 PDMS or VHB/PAAm-LiCl hydrogel/PDMS or very high bond (VHB)28 (Table S1, supplementary material).

FIG. 2.

Characterization of the stretchable LIG-based TENG. (a) Photograph of an LIG-based TENG consisting of a top LIG foam electrode with an island-bridge layout and a bottom crumpled Ag film electrode. (b) Output performance of the TENG with a pre-strain of 100% and a contact area of ∼25 cm2 at different deformation frequencies for an external force of 100 N and a load resistance of 1 MΩ. (c) Output power generated from LIG-based TENGs with different external loads from 10 kΩ to 10 GΩ. (d) Output voltage and (e) corresponding current of the LIG-based TENG with different pre-strain levels from 100 to 250% (for a force of 100 N at a frequency of 2 Hz). (f) Output characteristics of the LIG-based TENG with different external load resistances from 10 kΩ to 10 GΩ as a function of the pre-strain level from 100% to 250%.

FIG. 2.

Characterization of the stretchable LIG-based TENG. (a) Photograph of an LIG-based TENG consisting of a top LIG foam electrode with an island-bridge layout and a bottom crumpled Ag film electrode. (b) Output performance of the TENG with a pre-strain of 100% and a contact area of ∼25 cm2 at different deformation frequencies for an external force of 100 N and a load resistance of 1 MΩ. (c) Output power generated from LIG-based TENGs with different external loads from 10 kΩ to 10 GΩ. (d) Output voltage and (e) corresponding current of the LIG-based TENG with different pre-strain levels from 100 to 250% (for a force of 100 N at a frequency of 2 Hz). (f) Output characteristics of the LIG-based TENG with different external load resistances from 10 kΩ to 10 GΩ as a function of the pre-strain level from 100% to 250%.

Close modal

In addition to their stretchability, the level of pre-strain εpre also affects the output performance of the resulting LIG-based TENG. Therefore, the electrical output characteristics of the LIG-based TENGs is evaluated with varying εpre from 100% to 250% (for the same driving force of 100 N at the same frequency of 2 Hz). Since the larger pre-strain leads to increased surface roughness and charge density for an enhanced triboelectric effect, the output voltage and current increase from 280 to 566 V [Fig. 2(d)] and from 16.8 to 40.8 μA [Fig. 2(e)], respectively. Similar to the result in Fig. 2(c), the output power also first increases and then decreases as the load resistance increases from 10 kΩ to 10 GΩ for different εpre values [Fig. 2(f)]. The output power increases dramatically with increase in εpre to achieve a maximum instantaneous power density of 0.024 W at 2 MΩ and εpre = 250%. As a result, the stretchable LIG-based TENG with εpre of 250% is chosen in the subsequent studies unless specified otherwise. As the applied tensile strain increases, the maximum output power first slightly decreases in the small strain range (up to 40%) and then rapidly decreases in the large strain range (from 40 to 100%) (Fig. S13). Though the maximum output power is reduced to appropriately 50% from the initial 0.024 to 0.012 W at the applied strain of 100%, it is still as high as 0.022 W at the applied strain of 40%, which is well beyond the stretching range on the skin surface. The decrease in maximum output power with increase in the applied strain is mainly due to the increased resistance in TENG electrodes and the decreased contact potential difference between the two electrodes from the reduced roughness. In addition, the as-prepared LIG-based TENG can be operated continuously for about 1000 cycles with an almost unchanged output voltage of 280 V and even after exposing the device in the ambient environment for 5 months (Fig. S14), suggesting excellent mechanical stability. Nevertheless, LIG-based TENGs that can be easily deployed on human skin or clothing are promising to harvest mechanical energy from human movements for various wearable electronics.

By storing the intermitted energy harvested from TENGs, LIG-based micro-supercapacitor (MSC) arrays (LIG-MSCs) can provide a sustained energy to power on-skin wearable sensors in the self-powered sensing platform. In a single LIG-MSC with polyvinyl alcohol (PVA)/H2SO4 ionic gel electrolyte, the use of LIG foam as active material, current collector, and conductive substrate is explored. The cyclic voltammogram (CV) plots at various scan rates within a voltage range from 0 to 1 V exhibit quasi-rectangle shapes without distinct redox peaks even at a high scan rate of 300 mV s−1 [Fig. 3(a)], indicating the ideal electric double layer capacitive behavior. The galvanostatic charge–discharge (GCD) tests at different current densities show almost symmetric triangle shapes and negligible voltage drop at the beginning of discharge (infrared, IR, drop) due to low internal resistance during the charge–discharge process [Fig. 3(b)]. After introducing faradic nanocrystals such as MnO2 into the LIG foam, enhanced electrochemical performance can be achieved in the resulting MSC based on the LIG composite. Compared to the CV curves of LIG-MSC, the MnO2@LIG-MSC with PVA/Na2SO4 ionic gel electrolyte exhibits a much larger enclosed area to clearly reveal the pseudocapacitive characters of reversible redox reactions [Fig. 3(c)]. The cyclic voltammetry curves of MnO2@LIG-MSC under various scan rates from 30 to 200 mV have been indicated in Figs. S15 and S16. The MnO2@LIG-MSC samples show a distinct pair of redox peaks in the range 0.2–0.5 V and 0.6–0.8 V, indicating their pseudocapacitive characteristics.29 The redox peaks during the anodic and cathodic sweeps are induced by MnO2 nanoparticles electron transfer via MnO2 + A+ + e MnOOA (A = H3O+, Na+).30 In addition, the contribution ratio of capacitive- to diffusion-controlled charge in the MnO2@LIG-MSC is increased from 66.0% to 93.0% as the scan rate increases from 20 to 200 mV s−1 [Figs. S15(c) and S15(d)]. This result is attributed to the faradaic redox process (diffusion-controlled charge) with intrinsically slower kinetics than the electrostatic charge storage process (capacitive-controlled charge) in the hybrid MnO2@LIG-MSC devices. Moreover, the MnO2@LIG-MSC shows enhanced specific areal capacitances over LIG-MSC as calculated from the GCD curves: 6.15, 5.41, 4.65, 3.95, 3.42, and 2.64 mF cm−2 vs 4.1, 3.6, 2.8, 2.3, and 2.2 mF cm−2 at a constant current density of 0.1, 1.0, 5.0, 10.0, 15.0, and 30 mA cm−2 [Fig. 3(d)]. However, it is worth noting that the areal capacitance of the LIG-MSC only decreases from 4.2 to 1.3 mF cm−2 even as the scan rate drastically increases from 50 to 2000 mV s−1 [inset, Fig. 3(d)]. The total surface area of active electrodes in MSCA cells is about 1.8 cm, and the capacitance of each capacitor in MSCAs is ca. 7.6 mF. This performance is remarkable compared with that of previously reported carbon-based MSCs, including LIG-MSCs in the PVA/H3PO4 gel (0.62 mF cm−2),31 photoreduced graphene (0.53 mF cm−2),32 and spray-coated graphene-PH1000 (0.87 mF cm−2).33 Moreover, the as-assembled LIG-MSC delivers excellent cycling stability with 93% capacitance retention after 10 000 charge–discharge cycles at a current density of 5 mA cm−2 (Fig. S16). In addition, the Ragone plot of the LIG-MSC indicates that our device can deliver a maximum energy density of 0.569 (or 0.306) μWh cm−2 at a power density of 0.014 mW cm−2 (or even as high as 4.167 mW cm−2), exhibiting higher energy and power densities than previously reported all-solid-state MSCs (Fig. S17).

FIG. 3.

Characterization of stretchable LIG-based MSCs. (a) CV plots of the stretchable LIG-based MSCs (LIG-MSC) at various scan rates for voltage from 0 to 1 V. (b) Galvanostatic charge–discharge (GCD) of an LIG-MSC at different current densities. (c) CV curves of an MnO2@LIG-MSC with different scan rates. (d) The specific areal capacitance of an LIG-MSC and MnO2@LIG-MSC at different current densities and scan rates. (e) CV curves of LIG-MSCAs consisting of multiple cells connected in series and parallel. (f) Output voltage and capacitance with different MSC numbers in an LIG-MSCA. Normalized capacitance retention of LIG-MSCAs (C/C0) with (g) the applied strain and (h) the number of repetitive folding or twisting cycles.

FIG. 3.

Characterization of stretchable LIG-based MSCs. (a) CV plots of the stretchable LIG-based MSCs (LIG-MSC) at various scan rates for voltage from 0 to 1 V. (b) Galvanostatic charge–discharge (GCD) of an LIG-MSC at different current densities. (c) CV curves of an MnO2@LIG-MSC with different scan rates. (d) The specific areal capacitance of an LIG-MSC and MnO2@LIG-MSC at different current densities and scan rates. (e) CV curves of LIG-MSCAs consisting of multiple cells connected in series and parallel. (f) Output voltage and capacitance with different MSC numbers in an LIG-MSCA. Normalized capacitance retention of LIG-MSCAs (C/C0) with (g) the applied strain and (h) the number of repetitive folding or twisting cycles.

Close modal

Multiple LIG-MSCs connected in series or parallel with an island-bridge layout further yield the all-in-one stretchable MSCAs to meet specific output voltage/current needs for target applications. The CV curves of LIG-MSCAs consisting of multiple cells connected in series and/or in parallel indicate that the output voltage and current density can be well adjusted [Fig. 3(e)]. The LIG-MSCs connected in serial configurations generate an increased output voltage (e.g., threefold for three in the series), whereas those connected in parallel generate an increased output in the current density (e.g., threefold for three in parallel). The linear scaling in the output voltage and current with the number of LIG-MSCs holds true for nine LIG-MSCs [Fig. 3(f)] or possibly beyond, which can be used to easily adjust the output performance for target applications.

In addition to serial/parallel connections, the island-bridge layout also allows the LIG-MSCAs to robustly deform and stretch both uniaxially and biaxially up to 100% without any noticeable damage [insets in Fig. 3(g)]. Though the capacitance retention of the LIG-MSCAs slightly decreases with increase in the applied tensile strain, it still maintains ∼90%, 85%, and 72% under 100% stretching along the x, y, and both directions (i.e., biaxial), respectively [Fig. 3(g)]. Excellent capacitive retention in highly deformable states is attributed to (1) the strain isolation effect from the rigid polyethylene terephthalate (PET) film beneath the LIG-MSCs units and (2) the “filamentary serpentine” LIG with high electronic conductivity. Different from the previously published ZnP@LIG-based MSCAs that can only accommodate uniaxial strain,34 the LIG-based MSCAs with island-bridge designs in this work can be biaxially stretched up to 100%. It is worth noting that the LIG composites demonstrated in this and previous works also provide viable routes to deliver enhanced gravimetric (areal) capacitance and energy density. The stretchable LIG-MSCAs also show excellent durability, as evidenced by 83% and 89% in the normalized capacitance (C/C0) after 1000 folding and twisting cycles, respectively [Fig. 3(h)]. The superior stretchability and electrochemical capacitive performance of our stretchable LIG-MSCAs compare favorably over those reported previously (Table S2, supplementary material). The stretchable LIG-MSCAs with performance minimally affected by the strain provide robust energy storage devices for stretchable electronic devices.

The facile patterning capability from the laser scribing process enables simple creation of various LIG-based biophysical sensors to noninvasively measure arterial pulse, body temperature, biopotentials, blood pressure, and arterial oxygen saturation (Figs. 4 and 5). These representative sensors are chosen to monitor human health conditions for early diagnosis of cardiovascular diseases, chronic wounds, diabetes, hyperuricemia, gout and renal syndrome,35 which help decrease office visits and associated expensive lab testing. The LIG-based biophysical sensors fabricated on an ultrathin Ecoflex substrate (thickness of 100 μm) achieve seamless and conformal contact with the human skin through a biocompatible liquid bandage (Nexcare, 3M).

FIG. 4.

Characterization of wearable LIG-based epidermal pressure and temperature sensors. (a) The sensing mechanism of LIG-based epidermal pressure sensor. Output characteristics of the LIG-based pulse sensor from the foot of a healthy human subject (b) without and (c) with jumping up and down. (d) Dynamic electrical responses of an LIG-based temperature sensor with stepwise increased temperatures and (e) its calibration curve between the resistance and the temperatures ranging from 20 to 100 °C. (f) Performance of an LIG-based temperature sensor as a function of the bending radius of curvature.

FIG. 4.

Characterization of wearable LIG-based epidermal pressure and temperature sensors. (a) The sensing mechanism of LIG-based epidermal pressure sensor. Output characteristics of the LIG-based pulse sensor from the foot of a healthy human subject (b) without and (c) with jumping up and down. (d) Dynamic electrical responses of an LIG-based temperature sensor with stepwise increased temperatures and (e) its calibration curve between the resistance and the temperatures ranging from 20 to 100 °C. (f) Performance of an LIG-based temperature sensor as a function of the bending radius of curvature.

Close modal
FIG. 5.

Characterization of the wearable LIG-based electrocardiogram (ECG) and photoplethysmography (PPG) sensors. (a) Integration of an LIG-based electrode with signal processing and wireless transmission circuits yields a wireless electrophysiological signal system (b) to detect electrocardiogram (ECG) from the human chest and (c) during coughing. (d) Simultaneous measurement of ECG and pulse signals with two LIG-based electrodes and one pulse sensor on the wrist to obtain the pulse transit time (PTT). (e) Calculated beat-to-beat systolic (red, SDP) and diastolic (black, DBP) blood pressures based on the PTT method. (f) Schematic and (g) photograph of the photoplethysmography (PPG) sensor with one red and one infrared (IR) LED; a photodetector connected by the LIG-based circuit measures (h) the blood oxygen saturation and pulse rate from the human finger.

FIG. 5.

Characterization of the wearable LIG-based electrocardiogram (ECG) and photoplethysmography (PPG) sensors. (a) Integration of an LIG-based electrode with signal processing and wireless transmission circuits yields a wireless electrophysiological signal system (b) to detect electrocardiogram (ECG) from the human chest and (c) during coughing. (d) Simultaneous measurement of ECG and pulse signals with two LIG-based electrodes and one pulse sensor on the wrist to obtain the pulse transit time (PTT). (e) Calculated beat-to-beat systolic (red, SDP) and diastolic (black, DBP) blood pressures based on the PTT method. (f) Schematic and (g) photograph of the photoplethysmography (PPG) sensor with one red and one infrared (IR) LED; a photodetector connected by the LIG-based circuit measures (h) the blood oxygen saturation and pulse rate from the human finger.

Close modal

Based on arterial pulse-induced resistance change through the piezoresistive effect, the LIG-based epidermal pressure sensor [Fig. 4(a)] can continuously detect the arterial pulse signal from the foot of a human subject [Fig. 4(b)]. The measured radial artery pulse with an average of 72 beats per minute matches the reading by a conventional sensor (BH1792GLC-E2, Rohm Semiconductor). The sensor can also continuously detect pulse signals in motion (e.g., when the subject jumps up and down), as evidenced by the well-defined peaks and dips in Fig. 4(c). Furthermore, the stability of the LIG-based pulse sensor is demonstrated in a repeated loading–unloading pressure measurement (1000 Pa, frequency of 1 Hz) for 15 000 cycles (supplementary material, media I). The bandwidth and the line shapes remain almost unchanged after 15 000 cycles (Fig. S18). Moreover, the high signal-to-noise ratio is well maintained after sensor storage in an ambient environment for one month (Fig. S19). In addition to the pressure sensor, the LIG-based biaxial strain sensor with an Archimedean spiral-cutting kirigami pattern on Ecoflex can also measure dynamic forces in a wide deformation range up to 400% for detecting various human motions (Figs. S20–S22).

To measure the body temperature for personal healthcare, the LIG-based on-skin temperature sensor is demonstrated to rapidly detect stepwise increased temperature in the physiological range [Fig. 4(d)]. With an almost linear relationship between the resistance of the sensor and the temperature in the range from 20 to 60 °C [Fig. 4(e)], the LIG-based temperature sensor attached to a human foot dynamically captures the temperature change upon contact and removal of a glass beaker filled with hot water (Fig. S23). The performance of the LIG-based temperature sensor is also minimally affected by bending deformations (e.g., temperature variation of 5.6% even for a bending radius of 1 cm) [Fig. 4(f)] or cycling (e.g., 3.0% after 3000 cycles of repetitive bending with a radius of curvature of 1 cm). The LIG-based temperature sensors exhibit mechanical robustness and high stability to reliable monitoring of the body temperature.

As biopotential signals are important for early diagnosis of various cardiovascular diseases, the LIG-based electrodes are integrated with signal processing circuits and a wireless transmission module to yield an ECG sensing system [Fig. 5(a)]. The ECG data captured by the electrodes from humans are first converted into digital signals (analog–digital conversion, 8M-bit data storage) and then wirelessly transmitted to a mobile user interface via a Bluetooth chip (CC2541). The ECG data exhibit clear characteristic peaks [Fig. 5(b)] even during coughing [Fig. 5(c)]. Furthermore, concurrent measurements of the ECG and pulse signals provide cuffless detection of the blood pressure, according to the Hughes and Moens–Korteweg equations through the pulse transit time (PTT).36 By applying two LIG-based electrodes and one pulse sensor on the wrist, the beat-to-beat diastolic pressure (DBP) and systolic blood pressure (SDP) from a healthy human subject are measured as 75 and 120 mm Hg [Figs. 5(d) and 5(e)], which agree with those measured by a commercial sphygmomanometer (Fig. S24).

Using the LIG-based circuit to combine the GaAs-based red and infrared (IR) LEDs (center wavelength of 620 and 850 nm) with a silicon-based photodetector (400–1100 nm) yields a photoplethysmography (PPG) sensor to measure arterial oxygen saturation and heart rate variability [Fig. 5(f)]. The principle of an optoelectronic PPG sensor relies on the optical analysis of photometry using the Lambert–Beer law.37 As red signals with a smaller penetration depth are less robust to motion than IR signals, the light path between an LED and photodetector usually shifts under deformations or body movements. Therefore, this work exploits both the strain-isolation/island-bridge designs and the thinning technique in the optoelectronic chips (down to 20 μm)34,38 to allow the resulting device system to be attached to different locations of human skin such as finger and wrist (Fig. S25). As a result, the LIG-based blood oxygen sensor on the finger [Fig. 5(g)] reliably measures the blood oxygen saturation and pulse rate [Fig. 5(h)] even upon deformation (e.g., stretching/bending, speaking and arm moving, Fig. S26). By alternatively driving the red and IR LEDs for 9 s, each time using a custom-built circuit, the blood oxygen saturation and heart rate are calculated as 96.4% and 72% beats per minute [Fig. 5(h)], respectively. The results are consistent with those measured by a commercial oximeter (Beurer, PO30). The LIG-based blood oxygen sensor holds the potential to accurately and continuously monitor blood oxygen levels even upon skin motion.

As a proof-of-concept for system demonstration, an LIG-based TENG with power management circuits is integrated with LIG-MSCAs to provide a self-charging power unit to drive an LIG-based wireless sensing platform for continuous health monitoring. In this work, the LIG-based TENG shoe sole insert incorporated into a flip-flop converts mechanical energy from human motion into electrical energy [Fig. 6(a)]. Through a bride rectifier, the harvested electrical energy charges the LIG-MSCAs for energy storage and power supply, which can directly power LIG-based sensors, signal processing units, and wireless transmission components [Fig. 6(b)]. The output performance of the TENG shoe sole driven by body motion varies with the driving frequency (e.g., walking or running; 0.5, 1.0, and 1.5 Hz) and body weight (e.g., 60, 70, and 80 kg). For a body weight of 80 kg and an external load resistance of 1 MΩ, the output voltage increases to reach a maximum of 190 V as the frequency increases from 0.5 to 1.5 Hz [Fig. 6(c)]. The output voltage of the TENG shoe sole driven by body motion increases with higher driving frequencies, which is mainly due to the higher flow rate of charges introduced by the higher deformation frequency (or the average velocity).39,40 In addition, the weight of the body is one of the critical parameters of the wearable TENG output performance. As the body weight increases from 60 to 70 and then to 80 kg, the output voltage also increases from 155 to 174 and then to 190 V [Fig. 6(d)]. The output power density of the LIG-based TENG also first increases and then decreases as the load resistance increases from 100 kΩ to 1 GΩ, resulting in a peak power density of 9.21 μW cm−2 (or 0.23 mW) at 2 MΩ (Fig. S27).

FIG. 6.

Demonstration of the integrated LIG-based wireless sensing platform for continuous, long-term health monitoring. (a) Photograph and (b) circuit diagram of the LIG-based TENG shoe sole incorporated into a flip-flop in a self-powered sensing platform. (c) Output voltage of the flip-flop-embedded TENG with the driving frequency (from 0.5 to 1.5 Hz) and (d) body weight (50, 70, and 90 kg). (e) Output voltage of LIG-MSCAs with different numbers of MSCs connected in parallel/series charged by the flip-flop-embedded TENG driven by people with 90 kg at 1.5 Hz (discharge at 10 μA). Dynamic current change of the LIG-based pulse sensor powered by self-charging power units when a human subject is (f) running at a constant speed of 10 km per hour and at a frequency of 1.5 Hz, (g) taking rest, (h) first running at 10 km per hour and then taking rest, and (i) first taking rest and then running at 10 km per hour.

FIG. 6.

Demonstration of the integrated LIG-based wireless sensing platform for continuous, long-term health monitoring. (a) Photograph and (b) circuit diagram of the LIG-based TENG shoe sole incorporated into a flip-flop in a self-powered sensing platform. (c) Output voltage of the flip-flop-embedded TENG with the driving frequency (from 0.5 to 1.5 Hz) and (d) body weight (50, 70, and 90 kg). (e) Output voltage of LIG-MSCAs with different numbers of MSCs connected in parallel/series charged by the flip-flop-embedded TENG driven by people with 90 kg at 1.5 Hz (discharge at 10 μA). Dynamic current change of the LIG-based pulse sensor powered by self-charging power units when a human subject is (f) running at a constant speed of 10 km per hour and at a frequency of 1.5 Hz, (g) taking rest, (h) first running at 10 km per hour and then taking rest, and (i) first taking rest and then running at 10 km per hour.

Close modal

Charged by the TENG shoe sole, the LIG-MSCAs with multiple LIG-MSCs connected in parallel/series (Fig. S28) can generate specific power output for target applications (e.g., μA for sensor operation and mW for signal processing and Bluetooth communication). In addition to the boosted value, the charging curves also indicate linearly and stably increased output voltage with the charging time [Fig. 6(e)]. Specifically, it takes 872 s to charge a single LIG-MSC from 0 to 1 V and then 396 s to discharge at a current density of 10 μA. Moreover, the charge and discharge times of the LIG-MSCAs show a linear dependence on the number of LIG-MSCs, which confirms the excellent performance uniformity and scalability of the as-prepared LIG-MSCs. The LIG-MSCA with an array of 3 * 3 MSCs exhibits a stable high voltage of 3 V after charging for 6678 s by the LIG-based TENG shoe sole, which then takes 3186 s to discharge at a current density of 10 μA.

The sustained power supply from the integrated LIG-based TENG and MSCAs is then demonstrated to provide continuous and wireless monitoring of pulse signals [Figs. 6(f)–6(i)]. For a human subject with a body weight of 80 kg and running at a constant speed of 10 km h−1 and at a frequency of 1.5 Hz, the wirelessly measured pulse signal with a sensor powered by the sustained energy supply on a smartphone exhibits stable base current in the electrical signals [Fig. 6(f)]. In comparison, the base current from the one powered by the MSCAs alone gradually decreases with monitoring to last for only 5 min [Fig. 6(g)], due to power consumption and decreased output voltage in the MSCAs during real-time pulse monitoring. In addition, whenever the sustain power supply is turned on by initiating the TENG to harvest the body's mechanical energy, stable base currents can be observed to continuously monitor the pulse signals [Figs. 6(h) and 6(i)]. The slight difference in base current during real-time pulse monitoring can be attributed to the variation in human activities. These results indicate that the sustained power supply from LIG-based TENG and MSCAs is sufficient for real-time data acquisition, processing, and wireless transmission, eliminating the need for an external power supply or charging unit.

Moreover, the integrated TENG and MSCAs as a self-charging power unit can be exploited for real-time monitoring of vital human signs, including temperature, ECG, and blood oxygen levels. The LIG-based strain sensor powered by a self-charging power unit as a wrist-mounted device accurately monitors different wrist bending angles with high sensitivity [Fig. 7(a)]. The self-powered temperature sensing system also immediately responds to temperature change in the presence of ice water [Fig. 7(b)]. Furthermore, the self-powered electrophysiological sensing system accurately measures the surface ECG and drives wireless transmission modules to wirelessly transmit the date to the smartphone [Fig. 7(c), supplementary material, media II]. At the same time, the self-charging power units can power the two ultrathin LEDs and drive the photodetector in the PPG sensor for blood oxygen measurement [Fig. 7(d)].

FIG. 7.

Self-powered stretchable sensing platform to measure varying physiological signals. Resistance responses of the LIG-based (a) strain sensor at different bending angles of the wrist and (b) temperature sensor on the human foot in the presence of ice water. Wirelessly recorded (c) ECG and (d) blood oxygen signals in real time, powered by self-charging power units.

FIG. 7.

Self-powered stretchable sensing platform to measure varying physiological signals. Resistance responses of the LIG-based (a) strain sensor at different bending angles of the wrist and (b) temperature sensor on the human foot in the presence of ice water. Wirelessly recorded (c) ECG and (d) blood oxygen signals in real time, powered by self-charging power units.

Close modal

In summary, we present the design and demonstration of a self-powered, wireless, wearable sensing platform based on 3D porous interconnected LIG foam and its composites with high surface area and excellent charge transport properties fabricated by an efficient, low-cost laser scribing process. The integration of s stretchable LIG-based TENG with LIG-MSCAs results in a sustained power supply that converts the intermittent harvested energy into a stable power output. These self-charging power units can directly drive various LIG-based biophysical sensors to wirelessly measure pulse, strain, temperature, ECG, blood pressure, and blood oxygen in real time. Together with other LIG-based biophysical and biochemical sensors, the results from this study provide a general yet programmable platform for the next generation self-powered wearable electronics.

See the supplementary material for SEM images of the LIG foam and its composites; TEM, XRD, Raman, XPS, and BET images of LIG foam; schematic, working principle of the crumpled LIG-based TENG under cyclic compressive force; optical and SEM images of TENG electrodes; the electrical output characteristics of stretchable LIG-based TENG; the electrochemical behavior of LIG-MSC and MnO2@LIG-MSC; characterization and calibration of LIG-based sensors to measure temperature, strain, blood pressure, and blood oxygen; and the output performance of LIG-based TENG shoe sole and its use to charge LIG-MSCAs with multiple LIG-MSCs connected in parallel/series.

This work is supported by the National Natural Science Foundation of China (Nos. 52002162 and 12174172), the Natural Science Foundation of Fujian (No. 2021J011040), and the Fuzhou science and technology project (No. 2020-S-29). H.C. acknowledges the support from the National Science Foundation (NSF) (Grant No. ECCS-1933072); the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award No. R61HL154215; the National Institute of Biomedical Imaging and Bioengineering of the National Institutes of Health under Award No. R21EB030140; and Penn State University.

The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.

C.Z. and H.C. participated in the conceptualization, investigation, and review writing and editing of this research topic. H.C., X.D., F.L., C.H., B.Z., B.Z., J.W., and Y.X. participated in the investigation and data curation of this study.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
J. M.
Donelan
,
Q.
Li
,
V.
Naing
,
J. A.
Hoffer
,
D. J.
Weber
, and
A. D.
Kuo
,
Science
319
(
5864
),
807
(
2008
).
2.
A. D.
Kuo
,
Science
309
(
5741
),
1686
(
2005
).
3.
Z.
Li
,
G.
Zhu
,
R.
Yang
,
A. C.
Wang
, and
Z. L.
Wang
,
Adv. Mater.
22
(
23
),
2534
(
2010
).
4.
Y.
Yu
,
H.
Sun
,
H.
Orbay
,
F.
Chen
,
C. G.
England
,
W.
Cai
, and
X.
Wang
,
Nano Energy
27
,
275
(
2016
).
5.
X.-S.
Zhang
,
M.-D.
Han
,
R.-X.
Wang
,
F.-Y.
Zhu
,
Z.-H.
Li
,
W.
Wang
, and
H.-X.
Zhang
,
Nano Lett.
13
(
3
),
1168
(
2013
).
6.
M.
Ha
,
J.
Park
,
Y.
Lee
, and
H.
Ko
,
ACS Nano
9
(
4
),
3421
(
2015
).
7.
S.
Park
,
S. W.
Heo
,
W.
Lee
,
D.
Inoue
,
Z.
Jiang
,
K.
Yu
,
H.
Jinno
,
D.
Hashizume
,
M.
Sekino
,
T.
Yokota
,
K.
Fukuda
,
K.
Tajima
, and
T.
Someya
,
Nature
561
(
7724
),
516
(
2018
).
8.
H.
Ouyang
,
J.
Tian
,
G.
Sun
,
Y.
Zou
,
Z.
Liu
,
H.
Li
,
L.
Zhao
,
B.
Shi
,
Y.
Fan
,
Y.
Fan
,
Z. L.
Wang
, and
Z.
Li
,
Adv. Mater.
29
(
40
),
1703456
(
2017
).
9.
J.
Lin
,
Z.
Peng
,
Y.
Liu
,
F.
Ruiz-Zepeda
,
R.
Ye
,
E. L. G.
Samuel
,
M. J.
Yacaman
,
B. I.
Yakobson
, and
J. M.
Tour
,
Nat. Commun.
5
(
1
),
5714
(
2014
).
10.
Y.
Yu
,
J.
Nassar
,
C.
Xu
,
J.
Min
,
Y.
Yang
,
A.
Dai
,
R.
Doshi
,
A.
Huang
,
Y.
Song
,
R.
Gehlhar
,
A. D.
Ames
, and
W.
Gao
,
Sci. Rob.
5
(
41
),
eaaz7946
(
2020
).
11.
R.
Li
,
H.
Qi
,
Y.
Ma
,
Y.
Deng
,
S.
Liu
,
Y.
Jie
,
J.
Jing
,
J.
He
,
X.
Zhang
,
L.
Wheatley
,
C.
Huang
,
X.
Sheng
,
M.
Zhang
, and
L.
Yin
,
Nat. Commun.
11
(
1
),
3207
(
2020
).
12.
X.
Zhao
,
X.
Chen
,
H.
Yuk
,
S.
Lin
,
X.
Liu
, and
G.
Parada
,
Chem. Rev.
121
(
8
),
4309
(
2021
).
13.
G.
Zhang
and
X. W.
Lou
,
Adv. Mater.
25
(
7
),
976
(
2013
).
14.
J. C.
Meyer
,
A. K.
Geim
,
M. I.
Katsnelson
,
K. S.
Novoselov
,
T. J.
Booth
, and
S.
Roth
,
Nature
446
(
7131
),
60
(
2007
).
15.
A. C.
Ferrari
,
J. C.
Meyer
,
V.
Scardaci
,
C.
Casiraghi
,
M.
Lazzeri
,
F.
Mauri
,
S.
Piscanec
,
D.
Jiang
,
K. S.
Novoselov
,
S.
Roth
, and
A. K.
Geim
,
Phys. Rev. Lett.
97
(
18
),
187401
(
2006
).
16.
R.
Ye
,
Z.
Peng
,
T.
Wang
,
Y.
Xu
,
J.
Zhang
,
Y.
Li
,
L. G.
Nilewski
,
J.
Lin
, and
J. M.
Tour
,
ACS Nano
9
(
9
),
9244
(
2015
).
17.
B.
Qiu
,
M.
Xing
, and
J.
Zhang
,
J. Am. Chem. Soc.
136
(
16
),
5852
(
2014
).
18.
Q.
Zheng
,
Q.
Tang
,
Z. L.
Wang
, and
Z.
Li
,
Nat. Rev. Cardiol.
18
(
1
),
7–21
(
2021
).
19.
X.
Chen
,
Y.
Wu
,
A.
Yu
,
L.
Xu
,
L.
Zheng
,
Y.
Liu
,
H.
Li
, and
Z. L.
Wang
,
Nano Energy
38
,
91
(
2017
).
20.
S.
Lin
,
X.
Chen
, and
Z. L.
Wang
,
Chem. Rev.
(published online) (
2021
).
21.
S.
Li
,
Y.
Fan
,
H.
Chen
,
J.
Nie
,
Y.
Liang
,
X.
Tao
,
J.
Zhang
,
X.
Chen
,
E.
Fu
, and
Z. L.
Wang
,
Energy Environ. Sci.
13
(
3
),
896
(
2020
).
22.
Y.
Shi
,
F.
Wang
,
J.
Tian
,
S.
Li
,
E.
Fu
,
J.
Nie
,
R.
Lei
,
Y.
Ding
,
X.
Chen
, and
Z. L.
Wang
,
Sci. Adv.
7
(
6
),
eabe2943
(
2021
).
23.
H.
Chen
,
L.
Bai
,
T.
Li
,
C.
Zhao
,
J.
Zhang
,
N.
Zhang
,
G.
Song
,
Q.
Gan
, and
Y.
Xu
,
Nano Energy
46
,
73
(
2018
).
24.
P.
Zhao
,
G.
Bhattacharya
,
S. J.
Fishlock
,
J. G. M.
Guy
,
A.
Kumar
,
C.
Tsonos
,
Z.
Yu
,
S.
Raj
,
J. A.
McLaughlin
,
J.
Luo
, and
N.
Soin
,
Nano Energy
75
,
104958
(
2020
).
25.
H.
Chen
,
W.
Yang
,
P.
Huang
,
C.
Li
,
Y.
Yang
,
B.
Zheng
,
C.
Zhang
,
R.
Liu
,
Y.
Li
,
Y.
Xu
,
J.
Wang
, and
Z.
Li
,
Sustainable Energy Fuels
5
(
14
),
3737
(
2021
).
26.
H.
Chen
,
Y.
Xu
,
J.
Zhang
,
W.
Wu
, and
G.
Song
,
Nano Energy
58
,
304
(
2019
).
27.
X.
Wang
,
Y.
Yin
,
F.
Yi
,
K.
Dai
,
S.
Niu
,
Y.
Han
,
Y.
Zhang
, and
Z.
You
,
Nano Energy
39
,
429
(
2017
).
28.
X.
Pu
,
M.
Liu
,
X.
Chen
,
J.
Sun
,
C.
Du
,
Y.
Zhang
,
J.
Zhai
,
W.
Hu
, and
Z. L.
Wang
,
Sci. Adv.
3
(
5
),
e1700015
(
2017
).
29.
X.
Zhao
,
Y.
Hou
,
Y.
Wang
,
L.
Yang
,
L.
Zhu
,
R.
Cao
, and
Z.
Sha
,
RSC Adv.
7
(
64
),
40286
(
2017
).
30.
Z.-S.
Wu
,
W.
Ren
,
D.-W.
Wang
,
F.
Li
,
B.
Liu
, and
H.-M.
Cheng
,
ACS Nano
4
(
10
),
5835
(
2010
).
31.
X.
Shi
,
F.
Zhou
,
J.
Peng
,
R.
Wu
,
Z.-S.
Wu
, and
X.
Bao
,
Adv. Funct. Mater.
29
(
50
),
1902860
(
2019
).
32.
S.
Wang
,
Z.-S.
Wu
,
S.
Zheng
,
F.
Zhou
,
C.
Sun
,
H.-M.
Cheng
, and
X.
Bao
,
ACS Nano
11
(
4
),
4283
(
2017
).
33.
Z.-S.
Wu
,
Z.
Liu
,
K.
Parvez
,
X.
Feng
, and
K.
Müllen
,
Adv. Mater.
27
(
24
),
3669
(
2015
).
34.
C.
Zhang
,
Z.
Peng
,
C.
Huang
,
B.
Zhang
,
C.
Xing
,
H.
Chen
,
H.
Cheng
,
J.
Wang
, and
S.
Tang
,
Nano Energy
81
,
105609
(
2021
).
35.
J.
Kim
,
A. S.
Campbell
,
B. E.-F.
de Ávila
, and
J.
Wang
,
Nat. Biotechnol.
37
(
4
),
389
(
2019
).
36.
Y.
Xu
,
B.
Sun
,
Y.
Ling
,
Q.
Fei
,
Z.
Chen
,
X.
Li
,
P.
Guo
,
N.
Jeon
,
S.
Goswami
,
Y.
Liao
,
S.
Ding
,
Q.
Yu
,
J.
Lin
,
G.
Huang
, and
Z.
Yan
,
Proc. Natl. Acad. Sci. U. S. A.
117
(
1
),
205
(
2020
).
37.
H. U.
Chung
,
B. H.
Kim
,
J. Y.
Lee
,
J.
Lee
,
Z.
Xie
,
E. M.
Ibler
,
K.
Lee
,
A.
Banks
,
J. Y.
Jeong
,
J.
Kim
,
C.
Ogle
,
D.
Grande
,
Y.
Yu
,
H.
Jang
,
P.
Assem
,
D.
Ryu
,
J. W.
Kwak
,
M.
Namkoong
,
J. B.
Park
,
Y.
Lee
,
D. H.
Kim
,
A.
Ryu
,
J.
Jeong
,
K.
You
,
B.
Ji
,
Z.
Liu
,
Q.
Huo
,
X.
Feng
,
Y.
Deng
,
Y.
Xu
,
K.-I.
Jang
,
J.
Kim
,
Y.
Zhang
,
R.
Ghaffari
,
C. M.
Rand
,
M.
Schau
,
A.
Hamvas
,
D. E.
Weese-Mayer
,
Y.
Huang
,
S. M.
Lee
,
C. H.
Lee
,
N. R.
Shanbhag
,
A. S.
Paller
,
S.
Xu
, and
J. A.
Rogers
,
Science
363
(
6430
),
eaau0780
(
2019
).
38.
H.
Li
,
Y.
Xu
,
X.
Li
,
Y.
Chen
,
Y.
Jiang
,
C.
Zhang
,
B.
Lu
,
J.
Wang
,
Y.
Ma
,
Y.
Chen
,
Y.
Huang
,
M.
Ding
,
H.
Su
,
G.
Song
,
Y.
Luo
, and
X.
Feng
,
Adv. Healthcare Mater.
6
(
9
),
1601013
(
2017
).
39.
C.
Ning
,
K.
Dong
,
R.
Cheng
,
J.
Yi
,
C.
Ye
,
X.
Peng
,
F.
Sheng
,
Y.
Jiang
, and
Z. L.
Wang
,
Adv. Funct. Mater.
31
(
4
),
2006679
(
2021
).
40.
S.
Niu
,
S.
Wang
,
L.
Lin
,
Y.
Liu
,
Y. S.
Zhou
,
Y.
Hu
, and
Z.
Lin Wang
,
Energy Environ. Sci.
6
(
12
),
3576
(
2013
).

Supplementary Material