Advances in 2D materials for wearable biomonitoring

Abstract

Over the past two decades, the discovery of graphene has sparked a significant increase in research on two-dimensional (2D) materials These materials exhibit exceptional properties, including a large surface area, flexibility, and tunable electrical conductivity, making them ideal for building up wearable biosensors. Such biosensors offer rapid response times, high sensitivity, biocompatibility, and outstanding mechanical strength. This review provides a comprehensive overview of wearable biosensors based on 2D materials, highlighting their unique properties, synthesis methods, and integration into flexible electronic systems. Significant advancements, existing challenges, and commercialization prospects are explored. The development of these biosensors promises to revolutionize health monitoring and advance personalized medicine by enabling continuous, real-time monitoring of physiological parameters.

1. Introduction

Over the past two decades, research on two-dimensional (2D) materials has expanded rapidly, driven by the pioneering work of Geim and Novoselov in 2004 with the discovery of graphene [1]. This remarkable material, renowned for its exceptional physical properties, has become a cornerstone in numerous research domains, significantly advancing nanoscience, condensed matter physics, and chemistry [2]. The in-depth exploration of graphene’s properties has facilitated its integration into various interdisciplinary fields, including electronics, optoelectronics, and sensing technologies [3,4]. However, despite the initial enthusiasm for graphene, its limitations—including susceptibility to oxidative conditions [5,6], potential toxicity concerns [7], and an unswitchable conductive nature [8]—have driven the search for more stable and versatile 2D materials [1].

In recent years, research has increasingly focused on alternative 2D materials, including transition metal dichalcogenides (TMDs) [9–11], boron nitride (BN) [12–14], transition metal carbides and nitrides (MXenes) [15–17], and phosphorene. These materials exhibit a wide range of physical and chemical properties, making them well-suited for diverse applications, from electronics to biomedical sensing. The large-scale production of these materials has gained significant attention, driving pioneering research and technological advancements toward commercial healthcare products, such as graphene-based skin patches for glucose monitoring, wearable sweat sensors for electrolyte and metabolite analysis, and flexible electrocardiography (ECG) electrodes for cardiac health monitoring [18–21]. The scalable production of high-quality, large-area 2D materials has been made possible through techniques such as chemical vapor deposition (CVD) [22–24] and solution-based exfoliation [25–28], facilitating the development of advanced devices.

Wearable biosensors play a crucial role in the transition toward personalized medicine, enabling continuous, real-time monitoring of various physiological signals [29,30]. These devices offer valuable insights into an individual’s health, facilitating early disease detection and effective management [31]. However, current wearable biosensors suffer from a lack of sensitivity and specificity due to the trace levels of biomarkers in biofluids (as minor as pmol/L levels), as well as the chemical similarity of biomarkers, which further complicates accurate detection. To address these challenges and enhance the performance of wearable and noninvasive biosensors, there is a growing trend toward utilizing advanced 2D materials. These materials demonstrate superior sensitivity, specificity, and energy efficiency owing to their high surface-to-volume ratio, exceptional electrical conductivity, tunable bandgap, and the ability to be functionalized for selective molecular recognition. Additionally, their mechanical flexibility and compatibility with miniaturized electronic systems make 2D materials particularly well-suited for integration into next-generation wearable devices, paving the way for more precise and reliable real-time health monitoring [32].

For example, MXenes, known for their hydrophilic surfaces and high electrical conductivity, have garnered significant attention for their potential in sweat sensing applications. The incorporation of such advanced materials is particularly valuable, as addressing the simultaneous detection of multiple biomarkers while ensuring high specificity has long been a major challenge in the field of biosensing. The significance of 2D materials in wearable biosensors is further underscored by their potential in fabricating flexible, lightweight, and portable devices [33,34]. These characteristics are essential for creating comfortable and unobtrusive wearable sensors that can be seamlessly integrated into daily life.

Nevertheless, despite their promising potential, several challenges remain in the development and commercialization of 2D material-based wearable biosensors. Issues such as performance degradation, mechanical instability, and environmental susceptibility must be addressed to enhance the reliability and longevity of these devices [33,35]. Furthermore, ensuring low-cost, scalable production while maintaining high quality is essential for the widespread adoption of 2D materials in commercial applications [36].

This review summarizes recent developments of 2D materials-based wearable biosensors for health monitoring. It explores the unique properties of various 2D materials, their synthesis, and their integration into flexible devices, along with significant progress in the field. Additionally, the review examines the challenges and commercialization prospects of these technologies, offering insights into future research directions and potential applications. By addressing existing challenges and exploring innovative solutions, the development of 2D material-based wearable biosensors holds great promise for transforming health monitoring and advancing personalized medicine.

2. 2D materials properties for biosensor applications

2D materials exhibit a variety of properties that make them exceptionally well-suited for developing advanced biosensors for wearable health monitoring systems [37,38] (Fig. 1). Their inherent flexibility and stretchability allow them to conform to various surfaces, including human skins and intracorporeal organs, fitting well to surfaces even under mechanical deformation.

Fig. 1.

Key Attributes of 2D Materials for Biosensor Applications. (Reproduced with permission from ref. [37]. Copyright 2018, Springer Nature; from ref. [38]. Copyright 2020, Wiley).

These superior characteristics can be fine-tuned by the controllable reactivity of 2D materials, achieved through band engineering and surface functionalization, enables precise tailoring for specific biosensing applications. High carrier mobility ensures rapid and efficient charge transport, essential for sensitive and fast biosensor responses, while diverse electronic conductivity properties allow for varied applications in sensing technologies. Additionally, the high specific surface area (SSA) of 2D materials provides an extensive active surface for molecular interactions, resulting in high sensitivity and low detection limits. These properties collectively position 2D materials at the forefront of biosensor technology, promising significant advancements in real-time health monitoring and personalized medicine.

2.1. Flexibility and stretchability

The inherent flexibility and stretchability of 2D materials make them well-suited for wearable biosensors, which should conform to various surfaces, including human skin. These materials maintain high performance even under mechanical deformation, such as bending and stretching [39,40]. For example, a graphene-based flexible biosensor modified with olfactory receptors has demonstrated the ability to detect a range of analytes while maintaining functionality through prolonged bending cycles and up to a bend radius of 3 mm [40]. This flexibility ensures that wearable sensors remain comfortable and effective during prolonged use, crucial for continuous health monitoring. The use of ultra-thin 2D materials can significantly reduce the overall thickness of devices while increasing their relative surface contact area. This larger contact area enhances Van der Waals forces, resulting in greater adhesion energy between bioelectronics and the skin.

Using 2D materials with ultra-low thickness can reduce overall device thickness, enabling high adhesion energy with the skin. In general, sophisticated flexible applications like wearable sensors and healthcare monitoring devices demand ultrathin electronic materials with high mechanical flexibility and sometimes stretchability for specific applications. Young’s modulus and breaking strain limit of a semiconducting material are two key mechanical parameters that determine the successful operation of a fabricated flexible device. Young’s modulus (E) and stiffness (K) are correlated with a general relation of K = E×A/L, where A and L are the cross-sectional area and length of the used semiconducting material, respectively [41,42]. This relationship implies that a semiconducting material with high E typically exhibits strong stiffness and consequently low flexibility. Therefore, it is important to balance the stiffness and flexibility of devices to ensure adaptability and durability.

However, an imbalance arises as conventional semiconductors in their bulk form exhibit brittleness and high stiffness, which does not satisfy the critical demands of flexible electronics subjected to high strain during operation [43]. Despite numerous efforts to reduce their thickness for flexibility, the presence of dangling bonds and surface defects imposes limitations. That is because the strong intra-layer covalent bonding in 2D materials results in a high strain limit, in-plane mechanical strength, and Young’s modulus [44,45]. In contrast, the ultralow cross-sectional area of atomically thin-layered 2D materials provides them low stiffness, making them promising candidates for flexible device applications. However, the presence of structural defects in 2D materials can lead to mechanical failure [46–48]. Consequently, large-area 2D materials are susceptible to fractures, necessitating the use of an appropriate supporting substrate when integrating them into practical devices [49–51].

However, optimizing the mechanical properties of these materials remains a critical area of research. For instance, graphene, one of the most prominent 2D materials, demonstrates exceptional flexibility, low stiffness, and a high Young’s modulus (~1 TPa) combined with an impressive strain limit (~25%). These properties make it highly adaptable for enhancing interactions between graphitic layers during carbon nanotube (CNT) fabrication, which is vital for its ability to withstand high strains in wearable biosensors.

Similarly, transition metal dichalcogenides (TMDs)—which consist of hexagonally close-packed transition metals sandwiched between chalcogen atoms—exhibit strong intramolecular covalent bonds, resulting in a substantial Young’s modulus (120–400 GPa) and breaking strain limits of 6–11% [52]. The monolayers of black phosphorus (phosphorene), with their puckered honeycomb structure, offer Young’s moduli ranging from 44 to 166 GPa and fracture strains of 30% for monolayers and 32% for multilayers [53]. Borophene, an emerging 2D material, exhibits anisotropic mechanical properties with Young’s moduli of 389 GPa and 166 GPa along the α- and β-directions, respectively [54]. Overall, structural modifications and advancements in material design continue to enhance the mechanical properties of these 2D materials, making them more suitable for integration into wearable biosensors.

The additional benefit of 2D materials is their ability to conformably contact rough surfaces due to their extremely thin thickness, low bending stiffness, and high fracture strain limit. To continuously and reliably monitor various physiological changes, such as electromyogram (EMG) signals, electroencephalogram (EEG) signals, body temperature, and skin strain, it is crucial to obtain a high signal-to-noise ratio and minimized motion artifacts [55]. High adhesion forces between the device and the surface is essential for achieving conformal contact, maximizing van der Waals interactions [56,57]. The recent advancements in device architectures incorporating stretchable interconnects have been significant in the field of semiconductors. However, these devices are considered semi-flexible as they are manufactured by integrating flexible substrates with rigid silicon chips [58,59]. Consequently, the utilization of 2D materials with ultralow thicknesses demonstrates their exceptional potential for enhancing device performance in wearable applications.

2.2. Controllable reactivity

Controllable reactivity, through band engineering and surface functionalization, allows 2D materials to be tailored for specific biosensing applications [60,61]. The tunable band structure of 2D materials, achieved by applying electric fields, strain, or functionalization, enables the customization of these materials for developing advanced biosensors for monitoring glucose levels, detecting biomarkers for diseases such as cancer [62], and tracking physiological parameters like heart rate [63,64] and body temperature [65]. The adaptability of 2D materials enhances the sensitivity and selectivity of these biosensors, making them highly effective for continuous and real-time health monitoring. Additionally, the large surface-to-volume ratio, along with an adjustable density of reactive sites enhances the interaction with target analytes [66,67]. Functionalizing the surface with biorecognition elements can further increase the specificity and sensitivity of the sensors.

A significant factor in the reactivity of 2D materials, particularly TMDs, is the presence of defects such as chalcogen vacancies. These defects, often originating during synthesis and fabrication processes due to weak bonds between the transition metal and chalcogen atoms, can substantially impact the electronic and optoelectronic properties of TMDs. For instance, in monolayer MoS₂, sulfur vacancies are commonly observed and often lead to the presence of donor states, resulting in an n-type doping effect [68–70]. These defects can cause undesired threshold voltage shifts, poor gate controllability, and high contact resistance in transistor applications.

To mitigate the impact of these defects, careful synthesis processes are required. Approaches such as low-temperature synthesis and oxygen-incorporated CVD have been developed to reduce defect formation. For example, the synthesis of 2D materials at lower temperatures helps to minimize thermally induced defects [71,72]. Additionally, the introduction of oxygen passivation in sulfur vacancy sites can effectively suppress the formation of donor states in monolayer MoS₂. This leads to enhanced photoluminescence, a higher work function, and improved contact resistance at the metal/MoS₂ interface. Consequently, there is a reduction in the Schottky barrier and an associated decrease in energy barriers for electron transport, ultimately enhancing detection sensitivity [73,74].

Further, defect sites in grown films can be healed by treating the 2D materials surface with specific chemical molecules that have a high chemical affinity to the exposed atoms at the defect sites. Organic or inorganic species containing chalcogens (e.g., sulfur, selenium, tellurium) or halogens (e.g., fluorine, chlorine, bromine, iodine) can potentially be used for passivating defects in TMDs [75–79]. For instance, bis (trifluoromethane)sulfonimide (TFSI) passivation has been shown to improve the photoluminescence quantum yield by suppressing the effects of defects in metal-organic CVD (MOCVD)-grown MoS₂ films [80–82]. Therefore, the ability to control reactivity is crucial for enhancing the multiple reactivity and sensitivity of 2D materials in wearable bioelectronics.

2.3. Carrier mobility and electronic conductivity

Carrier mobility and electronic conductivity are critical attributes of 2D semiconductors, significantly enhancing the performance of biosensors [83] (Fig. 2a). High carrier mobility ensures rapid and efficient charge transport, crucial for achieving sensitive and fast biosensor responses [84–88]. For instance, monolayer MoS₂ transistors exhibit ultra-high on/off ratios, superior mobility, and ultra-low off-state currents [89]. This high carrier mobility enables the detection of analytes at extremely low concentrations, which is essential in both electrical and electrochemical biosensing applications. For example, a biosensor based on a 2D MoS₂ field-effect transistor (FET) has demonstrated the capability to detect prostate-specific antigen (PSA) at concentrations as low as 1 pg mL⁻¹, which is several orders of magnitude below the clinical threshold [90–92].

Fig. 2.

Properties and Sensing Mechanisms of Various 2D Materials. (a) Comparison of conductivities of various atomically thin 2D materials. (b) Comparison of specific surface area (SSA) values for various low-dimensional materials. Monolayer graphene has the highest SSA at 2630 m² g⁻¹. High SSA and low electrical noise improve analyte detection at low concentrations. h-BN: Hexagonal boron nitride; 2H-MoS₂: 2H-phase molybdenum disulfide; MOF: metal-organic framework; CB: conduction band; VB: valence band; MWCNT: multi-walled carbon nanotube; SWCNT: single-walled carbon nanotube; FL: few-layer. Environmental and physiological conditions used in 2D material-based sensors: (c) Thermoresistive effect; (d) biopotential signal sensing (e) piezoresistive effect; (f) piezoelectricity effect; (g) chemiresistivity. (b) (Reproduced with permission from ref. [146]. Copyright 2019, American Chemistry Society). (c-g) (reproduced with permission from ref. [33]. Copyright 2023, American Chemistry Society).

The performance of biosensors is heavily influenced by the quality of the active materials used, with charge carrier mobility being a key determinant of device performance, including switching speed, response current, and energy consumption. In the early stages of flexible technology development, materials such as organic semiconductors, amorphous materials, and metal oxide thin films were explored [93,94]. However, their low charge carrier mobility, environmental instability, and limited flexibility hindered their widespread use. In contrast, 2D materials have shown promise due to their superior charge carrier mobility and mechanical flexibility. These materials enhance several performance indices, such as on/off ratios, power consumption, and device speed, which is important for devices requiring high-speed data acquisition, transmission, and computing.

Graphene, for example, exhibits the highest field-effect mobilities with intrinsic ambipolar characteristics, making it promising for the development of high-speed electronics such as flexible radio frequency (RF) circuits [95,96]. The exceptional mobility of graphene, reported to be up to 180,000 cm² V⁻¹ s⁻¹ [97], which is the highest experimentally reported value to date, ensures that it can maintain high performance even under mechanical deformation, such as bending and stretching, which is crucial for wearable biosensors. Other 2D materials also exhibit high carrier mobility, though not as high as graphene. For instance, phosphorene has a mobility of ~1,000 cm² V⁻¹ s⁻¹ [98], silicene about 2, 100 cm² V⁻¹ s⁻¹ [97], germanene about 2,800 cm² V⁻¹ s⁻¹ [99], and stanene about 3,000 cm² V⁻¹ s⁻¹ [100]. Borophene, with its anisotropic electronic features, exhibits a mobility comparable to that of graphene in some directions [101].

The charge carrier mobility of 2D materials is influenced by various factors, including atomic number, atomic size, crystal structure, chemical bonding, and lattice strain [102,103]. Strategies to modulate the carrier mobility and transport properties of 2D materials include chemical doping, strain engineering, surface functionalization, layer number reduction, and high-k dielectric encapsulation. Introducing dopants into 2D materials can significantly enhance carrier mobility by reducing impurity scattering and altering the electronic structure [102, 104]. Applying mechanical strain can modulate the band structure and carrier mobility [105], for example, tensile strain in MoS₂ can increase its carrier mobility by reducing the effective mass of the charge carriers. Functionalizing the surface of 2D materials with various chemical groups can modify the charge carrier density and mobility [106–108]. For instance, functionalizing MXenes with different surface groups can tune their electronic properties. Reducing the number of layers in 2D materials can enhance carrier mobility by minimizing interlayer scattering. Monolayer TMDs, for example, often exhibit higher mobility than their multilayer counterparts [109]. Encapsulating 2D materials with high-k dielectrics can reduce Coulomb scattering from charged impurities, thereby enhancing carrier mobility [110–112]. For example, MoS₂ exhibits enhanced carrier mobility when encapsulated with high-k dielectrics like HfO₂, reducing Coulomb scattering and improving the overall transport properties of the material [113].

The electronic conductivity of 2D materials is another crucial factor for biosensing applications. High conductivity ensures efficient signal transduction and minimal energy loss. Graphene, with its exceptional conductivity and transparency, has been extensively explored in several applications such as flexible electrodes, transparent conductors, mechanical actuators, and heat dissipaters [114,115]. Various other high-mobility 2D materials, such as borophene, phosphorene, stanene, and germanene, 2D covalent organic frameworks (COFs), MOFs, layered metal oxides [116], have been developed in parallel for use in flexible applications. The semiconducting nature of TMDs, with appropriate band gaps, enables their use in high-performance flexible thin-film transistors (TFTs) and p-n junctions for various applications where direct bandgap TMDs are suitable for flexible optoelectronics.

The band structure of layered TMDs is strongly influenced by the combined effect of quantum confinement and interlayer interactions [117–119]. As the number of layers decreases, the quantum confinement effect becomes more pronounced while the interlayer interaction becomes less effective. For instance, a reduction in the number of layers in TMDs such as MoS₂ leads to a gradual increase in the energy of indirect excitonic transition, whereas the direct band gap at the K point of the Brillouin zone barely changes [120]. Interestingly, as the thickness of MoS₂ decreases to a monolayer, the indirect band gap becomes higher than the direct transition energy, and the material converts into a direct bandgap 2D semiconductor (Eg ~1.9 eV) [121,122]. Similarly, the monolayer form of several other TMDs such as MoSe₂ (~1.5 eV) [123], WS₂ (~2.1 eV) [124], WSe₂ (~1.6 eV) [125], PtSe₂ (~1.8 eV) [126], and MoTe₂ (~1.1 eV) [127] exhibits a direct band gap nature, whereas their bulk phases exhibit an indirect bandgap nature with smaller values. This ability to tune the bandgap makes TMDs versatile for various biosensing applications, where specific bandgap values are required for optimal sensor performance.

Strategies to modulate the electronic conductivity of 2D materials include doping, strain engineering, and layer thickness control. For instance, doping graphene with nitrogen or boron can tailor its electrical properties, enhancing its sensitivity and selectivity in sensors [128]. Despite the effectiveness of covalent functionalization in enhancing selectivity, it may alter the electrical nature of 2D materials, making them less suitable for sensing. Non-covalent functionalization offers flexibility, modifying physical and electrical properties without altering the crystalline structure [129]. Non-covalent interactions, such as electrostatic, van der Waals (vdW), and π–π interactions, allow controllable changes in doping concentrations, maintaining conductivity crucial for electrically transduced sensors [130–133]. Miniaturized devices benefit from the exceptional charge transport properties of 2D materials, preserved in ultrathin sub-nanometer layers [134,135], ensuring superior electrical contacts in bioelectronics. This approach supports the development of highly sensitive and specific biosensors, leveraging the unique properties of 2D materials to advance healthcare monitoring.

MXenes, a newly emerged family of layered 2D transition metal carbides/nitrides, exhibit tunable electronic phases. Most MXenes show semiconducting behavior with a bandgap range of 0.24–1.8 eV [136], and their electronic band structures can be adjusted through surface functionalization. For example, pure and F-functionalized Ti₂C are metallic [137], while Ti₂C becomes semiconducting when functionalized with oxygen atoms [138]. This tunability of MXenes’ electronic properties is highly advantageous for designing biosensors with specific electronic requirements.

BP is another promising layered semiconductor with a thickness-dependent direct bandgap ranging from 2.0 eV (monolayer) to 0.3 eV (bulk), and an experimental electronic conductivity of approximately 1,000 cm² V⁻¹·s⁻¹ [139,140]. The strain and layer number-dependent tuning of its bandgap make BP an interesting candidate for optoelectronic applications in the visible to mid-infrared wavelength range [141–143]. Although hexagonal BN (h-BN) is more regarded as an insulator with a wide bandgap of about 6 eV, it possesses excellent dielectric characteristics, making it useful as a gate dielectric and encapsulating material in 2D material-based devices [144].

The combination of high carrier mobility, tunable electronic properties, and adjustable bandgap makes 2D materials ideal candidates for next-generation biosensing applications. By leveraging strategies to modulate carrier mobility, electronic properties, and bandgap, researchers can optimize the performance of biosensors and other flexible electronic devices, ensuring rapid, sensitive, and reliable detection of analytes at low concentrations.

2.4. Specific surface area

Specific Surface Area (SSA) is a critical property of 2D materials that significantly enhances their biosensing capabilities. Large surface-to-volume ratio ensures a substantial active surface area for interactions with target analytes, resulting in superior sensitivity and low detection limits [145]. This extensive surface area allows the entire material to be available for adsorption, maximizing interaction with analytes and enhancing detection capabilities. For instance, 2D TMDs based plasmonic biosensors exhibit significantly stronger responses compared to traditional sensors using metallic substrates [146]. This improvement is attributed to the prolonged plasmon lifetimes and the high SSA of TMDs, which enhance their interaction with analytes. Monolayer graphene, with the highest SSA of 2630 m² g⁻¹, exemplifies this benefit [147] (Fig. 2b). Its high SSA contributes to its remarkable sensitivity and low detection limits for analytes, allowing for the detection of ultralow concentrations.

The high SSA of 2D materials enables various functionalities through defect engineering, creating heterostructures, alloys, composites, and chemical or molecular doping. This extensive surface area is advantageous for chemical and electrochemical sensing, as it provides numerous active sites for reactions, enhancing interactions with target molecules and resulting in improved sensitivity and lower detection limits. In gas sensing applications, a larger surface area allows more gas molecules to interact with the active 2D sensing material, enhancing the sensor’s performance [148,149]. Similarly, in biosensors, a larger surface area can accommodate more biological molecules, enabling precise and sensitive detection of biomarkers [150]. The sp² hybridized carbon atoms in graphene’s structure make it easy to functionalize its surface by covalently attaching various chemical groups or molecules, thereby improving selectivity for ions or biomolecules [151].

The atomically thin nature of 2D materials ensures that charge carrier flow is restricted to the material’s surface, which is directly exposed to external stimuli. This leads to efficient signal acquisition and conversion with a fast response time, crucial in applications where rapid detection is critical, such as medical diagnostics, glucose monitoring, detection of disease biomarkers, and real-time physiological monitoring [152,153]. Moreover, the larger surface area of 2D materials is vital in energy harvesting devices such as piezoelectric nanogenerators (PENGs) and photovoltaic devices [154,155]. The magnitude of stress applied to piezoelectric materials is affected by their surface area, with a larger surface area maximizing the electrical output for a given amount of mechanical stress. In optoelectronic devices like photosensors and solar cells, a larger surface area increases light-matter interaction, resulting in higher photocurrent and power output [156–158].

2D materials provide a large surface area that enables interactions with small molecules, resulting in high sensitivity even for low analyte concentrations. The reactivities on their surfaces enhance interactions with target components and facilitate the immobilization of additional recognition elements, such as receptors and metallic nanoparticles, through covalent bonding, electrostatic interactions, and π-π stacking, which are crucial in electrochemical sensing, where the immobilization of enzymes or antibodies on 2D material surfaces can significantly improve detection specificity and sensitivity. In electrochemical sensing, the active sites on 2D materials confer electrocatalytic properties, improving current response and reducing redox potential when an analyte is present. The large surface area and high reactivity of 2D materials enable efficient electron transfer, enhancing the overall performance of the sensor. For example, functionalizing graphene with gold nanoparticles can create a biosensor with heightened sensitivity for glucose detection [159,160], while immobilizing antibodies on MoS₂ can lead to selective detection of specific proteins [161,162]. Excellent selectivity can be achieved with 2D material-based biosensors through diverse surface engineering strategies. Defect engineering, for instance, involves creating vacancies or introducing heteroatoms to modulate the electronic properties and reactivity of the material, thereby enhancing its interaction with specific analytes. Chemical modifications by attaching functional groups to the material’s surface, can tailor the sensor’s specificity and binding affinity. Additionally, functionalization with polymers or other organic molecules can improve the stability and biocompatibility of the sensors, making them suitable for a wide range of biosensing applications.

3. Sensing mechanisms of various 2D materials

Sensing mechanisms of various 2D materials leverage their unique properties to detect changes in environmental and physiological conditions (Fig. 2c–g). The thermoresistive effect capitalizes on the thinness and conformal attachment of 2D materials to detect temperature variations accurately. Changes in electrical resistance due to temperature fluctuations arise from thermal vibrations in metals and alterations in charge carrier mobility, concentration, and doping levels in semiconductors. Biopotential signal sensing involves measuring electrical signals generated by living organisms, such as nerve impulses or muscle contractions. The mechanical flexibility of 2D materials ensures conformal contact with human skin, providing accurate and reliable measurements with minimal motion artifacts. The piezoresistive effect, observed in materials like graphene and MoS₂, involves changes in electrical resistivity due to mechanical deformation. This effect, significantly impacting semiconductors, results in a high gauge factor (GF), indicating the material’s sensitivity to deformation and strain. Piezoelectricity in crystals like MoS₂ and h-BN generates an electric potential when deformed, making 2D materials well-suited for flexible and wearable sensors. Chemiresistivity utilizes alterations in electrical resistance resulting from interactions with target analytes. These diverse sensing mechanisms show the potential of 2D materials for biosensors in various applications (Table 1).

Table 1.

Sensing mechanisms of 2D bioelectronic materials.

Sensing Mechanism	Materials	Sensitivity	Applications
Thermoresistive Effect	Graphene, MoS2	Up to 0.25% °C−1 [545]	Body temperature monitoring, thermal imaging for disease diagnosis [522]
Biopotential Signal Sensing	Graphene, h-BN, MoS2	0.1 mV [35]	EMG, ECG, EEG monitoring [373]
Piezoresistive Effect	Graphene, MXenes	0.05 kPa [181]	Blood pressure monitoring, respiration monitoring [536]
Piezoelectricity	Graphene, SnS2, MoS2	0.1 μW/cm2 [419]	Self-powered wearable sensors, vibration sensors for cardiac monitoring [413]
Chemiresistivity	Graphene, MXenes, metal oxides	1 ppm [620]	Glucose monitoring, cortisol detection [703]

3.1. Thermoresistive effect

The thermoresistive effect describes a phenomenon where a material’s electrical resistance changes due to temperature variations (Fig. 2c). This change can be expressed: R=R₀ [1+TCR(T–T₀)], where R₀ is the resistance at a reference temperature T₀, and TCR refers to the temperature coefficient of resistance of the material. In metals, the thermoresistive effect primarily originates from increased thermal vibrations of atoms as temperature rises, resulting in a positive TCR value due to more electron scattering and increased resistance.

In semiconductors, the thermoresistive effect is more complex. It is influenced by factors such as charge carrier mobility, carrier concentration, and doping levels [163,164]. In intrinsic semiconductors, increased temperature typically leads to higher carrier concentration, which can decrease resistance [165]. However, in extrinsic semiconductors, doped with impurities, the effect can be dominated by changes in carrier mobility, similar to metals [166]. Thus, semiconductors can exhibit either positive or negative TCR values depending on these factors.

The advantage of utilizing 2D materials in sensors lies in their thinness, allowing them to conformally attach to various surfaces, including the human body. This characteristic ensures high sensitivity and accuracy in detecting temperature changes. For example, sensors made from 2D materials can respond accurately to body temperature fluctuations, making them well-suited for wearable health monitoring devices [167,168].

3.2. Biopotential signal sensing

Biopotential sensing refers to the measurement of electrical signals generated in living organisms, such as nerve impulses or muscle contractions (Fig. 2d). These signals result from neuron stimulation through the movement of ions across cell membranes, leading to changes in membrane potential. The resulting biopotential signals, like EEGs and EMGs, typically exhibit low amplitudes (as low as 10 μV) and require highly sensitive devices for precise detection [169–171]. The mechanical flexibility of 2D materials is advantageous, as it the close contact enhances the accuracy and reliability by reducing motion artifacts, which are common issues in biopotential sensing. Furthermore, the atomic thickness of 2D materials ensures minimal interference with the natural movements of the body, providing a seamless integration with human bodies for continuous health monitoring [172–174]. This makes 2D materials well-suited candidates for biopotential sensing applications (Table 2).

Table 2.

Biopotential sensing applications of 2D material-based sensors.

Applications	Materials	Advantages	Disadvantages
EMG Monitoring	Graphene, MoS2	Great signal fidelity due to high conductivity	Susceptible to motion artifacts and noise
ECG Monitoring	Graphene, CNTs	Comfortable for long-term wear, lightweight	Requires skin preparation for optimal contact
EEG Monitoring	Graphene	High spatial resolution for brain activity mapping	Limited durability, may detach during movement

3.3. Piezoresistive effect

The piezoresistive effect describes the phenomenon where a material’s electrical resistivity changes due to mechanical deformations such as pressure or strain (Fig. 2e). In conductors, electrical resistance (R) is influenced by geometric changes (R=ρ×L/A, where A is the cross-sectional area, L is length, and ρ is resistivity). However, in 2D materials, the piezoresistive effect has a more significant impact on the electrical resistance of semiconductors compared to geometric changes. This effect is quantified by the gauge factor, which measures the sensitivity of a material’s electrical properties to deformation. It is defined as GF=(ΔR/R)/ϵ, where ΔR/R represents the resistance change, GF is gauge factor and ϵ is the applied strain. For example, semiconductors exhibit a GF that is typically two orders of magnitude higher than that of metals, which have a GF between 1 and 5 [175,176]. The piezoresistive effect in semiconductors is due to changes in lattice constants and crystal symmetry under strain, leading to shifts in the bandgap and changes in the effective mass of carriers. This, in turn, affects the material’s electrical conductivity and carrier mobility [177].

In various 2D materials such as graphene, MoS₂, In₂Se₃, and WS₂, the piezoresistive effect is prominent and is attributed to their atomic-scale thickness, high Young’s modulus, and high elastic strain limit, which demonstrate significant contribution to bandgap [178–180]. These properties make 2D materials attractive for use in deformable sensors capable of detecting pressure and phonation. For example, under strain, the lattice deformation in these materials alters their electrical pathways, leading to significant changes in resistivity. This makes 2D materials well-suited candidates for high-sensitivity applications, where precise measurement of biomechanical deformation is crucial [181].

3.4. Piezoelectricity

The piezoelectric effect is a phenomenon in which an electric potential is generated in a crystal when it undergoes mechanical deformation due to an external force (Fig. 2f). This electromechanical interaction occurs in materials whose crystal structure lacks a center of symmetry. When an external force is applied to a piezoelectric material, it causes a deformation or strain in the crystal structure, leading to a rearrangement of the dipoles within the lattice [182–193]. Unlike piezoresistive materials, which exhibit resistance variation under mechanical deformation, the displacement of electric dipoles in piezoelectric materials generates an electric potential difference across the material, which is utilized widely in self-powered bioelectronics.

Piezoelectric effect-based devices offer several advantages over other sensing technologies. These sensors require low power to operate, making them suitable for wearable and battery-power free devices. Various 2D materials, such as MoS₂ [194,195] and h-BN [196], demonstrate piezoelectric properties. Although there are many rigid materials such as lead zirconate titanate, which are well-known for exhibiting the piezoelectric effect, 2D materials possess excellent mechanical properties due to their atomic thinness [197]. For instance, wearable devices can benefit significantly from the piezoelectric properties of 2D materials, as they can detect subtle mechanical changes and convert them into electrical signals with high accuracy, providing real-time data with minimal power consumption [198].

3.5. Chemiresistivity

The chemiresistive effect refers to the phenomenon where the electrical resistance of materials changes due to alterations in the nearby chemical environment (Fig. 2g) [199]. When the surface of the 2D material-based biosensors interacts with target analytes (including gases, liquids, and biochemicals), it results in a change in the electrical properties [200,201]. The high surface area allows for larger interaction between the sensing material and target analytes, leading to increased sensitivity. Moreover, the ease of surface modification in 2D materials enables tailored functionalization to enhance selectivity toward specific analytes. Functional groups or molecules can be covalently attached to the surface of 2D materials, enhancing their capacity to selectively interact with target analytes. For instance, Few-layer holey graphene oxide (FHGO) is utilized in heterostructures with black phosphorus (BP) and MXene. FHGO functions as a multifunctional passivation layer, providing protection against oxidative degradation for the underlying two-dimensional materials while selectively permitting the diffusion of target gas molecules, such as NO₂, through its micropores [201,202]. Similarly, MoS₂ and other TMDs offer tunable electronic properties and high surface reactivity, making them suitable for detecting various chemical species [203].

4. 2D materials preparation

The field of wearable biosensors is experiencing significant advancements due to the integration of 2D materials, which possess unique properties that enhance sensor performance. These materials, exemplified by graphene, are characterized by their atomic thickness and immense surface areas, making them well-suited candidates for various applications in wearable technology. Over the past decade, the advance and synthesis of emerging 2D materials beyond graphene, such as TMDs, h-BN, MXenes, have opened new avenues in the design of wearable biosensors [204–206]. These materials exhibit unique optical, transport, and electromagnetic properties, along with flexibility and durability, which are essential for creating highly efficient and reliable biosensors.

The incorporation of 2D materials into wearable biosensors presents several challenges. The fabrication of high-quality 2D materials with uniform characteristics, such as the lateral dimensions, layer numbers, and surface chemistries, remains challenging [207]. Achieving uniform dispersion of these nanofillers in a polymer matrix without agglomeration, particularly at high filler content, is another difficulty.

To overcome these challenges, advanced synthesis methods that go beyond direct mixing process have been developed. These techniques can be broadly categorized into bottom-up and top-down approaches. Top-down methods, such as liquid phase exfoliation and chemical/mechanical exfoliation, involve breaking down bulk materials into atomically thin layers. For example, liquid phase exfoliation uses solvents and ultrasonic energy to separate layers of materials like graphite into individual graphene sheets. This method is advantageous for its scalability and cost-effectiveness, but it can result in variations in layer thickness and lateral dimensions. Chemical exfoliation, which involves intercalating chemicals between layers to weaken interlayer forces, also allows to produce few-layer 2D materials, but controlling the final product’s uniformity remains challenging.

Bottom-up methods, including CVD and molecular beam epitaxy (MBE), involve the synthesis of 2D materials atom by atom or molecule by molecule. CVD, for instance, allows for the growth of high-quality, large-area graphene on substrates by decomposing gaseous precursors at high temperatures. This method offers excellent control over the thickness and uniformity of the 2D materials but can be more complex and expensive. MBE provides precise control over the atomic layer deposition, resulting in highly crystalline 2D materials, yet it is less scalable and more costly compared to other methods.

Despite substantial progress, there remains a gap in optimizing 2D material structures for applications, which often require a combination of tailored properties. The quality of 2D nanofillers varies significantly with synthesis methods, affecting their crystallinity, geometry, and surface properties. For example, in mechanical reinforcement applications, a high content of aligned 2D fillers enhances strength and durability. Conversely, for conductive composites, a 3D interconnected network with a low percolation threshold facilitates efficient electron transport. Understanding and controlling these structural variations are crucial for maximizing the performance of 2D materials in biosensors, and flexible electronics.

Addressing these issues requires an in-depth grasp of the relationships between process, structure, and properties to simultaneously customize multiple characteristics for specific applications. For wearable biosensors, it is crucial to balance properties such as sensitivity, flexibility, durability, and biocompatibility. The assembly and homogeneous dispersion of 2D materials into hierarchical nanocomposites can lead to biosensors that meet these stringent requirements. This section covers the synthesis and rational assemblies of 2D materials, their multiscale structural characteristics, and their multifunctional properties. The focus is primarily on materials frequently used in preparing nanocomposites, such as graphene derivatives, h-BN, MoS₂, and Ti₃C₂T_x. The discussion highlights the relationships between process, structure, and properties at various length scales, which inform the rational design of customized multifunctional materials in wearable biosensors [208–211].

4.1. Top-down exfoliation

Preparing 2D materials involves several top-down approaches, such as mechanical exfoliation, ultrasonication, chemical exfoliation, and liquid exfoliation, which is critical for advancing the production of high-quality 2D materials for biosensing applications.

4.1.1. Physical exfoliation

Mechanical exfoliation is a widely used technique for producing high-quality 2D materials. This approach entails physically separating layers from bulk materials, typically using adhesive tape, to obtain atomically thin flakes [212]. Owing to the weak interlayer vdW interactions and strong covalent bonds within layers, this method effectively produces individual flakes with unique properties not seen in their bulk form. Although it is a simple and low-cost method, mechanical exfoliation often results in limited size and thickness uniformity, making it less suitable for large-scale production. To address these limitations, recent mechanical milling has been utilized to enhance throughput and simplify manufacturing, potentially making mechanical exfoliation feasible for large-scale production [213] (Fig. 3a).

Fig. 3.

Preparation of 2D Materials.(a,b) Physical exfoliation for producing 2D-material-based inks. (c) Ion intercalation weakens layer interaction, and agitation yields monolayer sheets. (d) Chemical exfoliation for obtaining 2D material monolayers. (e) Direct growth of high-κ dielectric on 2D semiconductors: ALD (upper panel) and plasma-enhanced atomic layer deposition, PEALD (lower panel) processes. (f) Depositing high-κ dielectric on a seed layer. (g) Growth of high-κ Bi₂SeO₅ dielectric by in situ oxidation of Bi₂O₂Se. (h) Substrate surface modification for wafer-scale transfer. (i) h-BN as a dielectric via vdW transfer. (j) Direct Transfer with patterning. (a) (Reproduced with permission from ref. [212]. Copyright 2015, Springer Nature). (b) (Reproduced with permission from ref. [33]. Copyright 2023, American Chemistry Society). (c) (Reproduced with permission from ref. [219]. Copyright 2013, Science). (d) (Reproduced with permission from ref. [229]. Copyright 2019, Elsevier). (e-f) (reproduced with permission from ref. [237]. Copyright 2022, Springer Nature). (g) (reproduced with permission from ref. [237]. Copyright 2022, Springer Nature). (h) (Reproduced with permission from ref. [238]. Copyright 2018, Wiley). (i) (Reproduced with permission from ref. [237]. Copyright 2022, Springer Nature). (j) (Reproduced with permission from ref. [243]. Copyright 2013, American Chemistry Society).

In conventional mechanical exfoliation, a bulk material like graphite is repeatedly pressed and peeled using adhesive tape until thin flakes are obtained. This approach can yield ultrathin layers but poses limitations as repeated folding can cause the material flakes to break, leading to low yield and small lateral dimensions [214]. Enhancements such as oxygen plasma exposure and soft baking at 100 °C can improve the interaction between the 2D materials and the substrate, which enhances yield and size without mechanical deformations [215]. Mechanical exfoliation methods have also evolved to include Au-assisted [216] and Ni-assisted techniques [217–219], leveraging the high affinity between certain atoms and Au or Ni to exfoliate large monolayer sizes with high yield.

Ultrasonication-based exfoliation is another method used to produce 2D materials by dispersing bulk materials in a liquid medium and exposing them to ultrasonic waves emitted by a transducer (Fig. 3b). The ultrasonic waves create unstable cavitation bubbles in the solution that collapse inward, generating high temperature, pressure, and velocity jets of liquid. These forces are strong enough to surpass the weak vdW forces between layers, resulting in the exfoliation of relatively small 2D flakes. This method is viable for producing nanosheets on a large scale, generating significant interest for its potential to produce air-stable inks or thin films. However, several conditions should be considered, such as solvent selection, sonication power, time, and the use of additives.

Liquid-phase exfoliation benefits from the large surface area-to-volume ratio and ease of surface modification [220] (Fig. 3c). Solvent selection is crucial, as solvents with surface tensions similar to those of 2D materials produce homogeneous dispersions [221,222] Additives like surfactants and polymers can further enhance the process by modifying solvent surface tension, promoting cavitation events, and preventing re-aggregation of exfoliated sheets. The use of surfactants, for instance, facilitates electrostatic repulsion between 2D sheets, improving yield without increasing sonication power [223]. Polymers can mediate surface energy mismatches between solvents and layered materials, enabling efficient exfoliation using non-traditional solvents [224].

Shear-induced exfoliation is an alternative and highly scalable method for exfoliating 2D materials. This method uses a solution containing 2D flakes that flow through a narrow channel at high speed, combining viscous forces from the solvent and shear forces across the 2D flakes [225,226]. This process breaks the vdW forces between layers, producing thinner sheets with low defect concentrations. Shear exfoliation requires balancing viscous fluids and shear forces to ensure efficient exfoliation without breaking the flakes, making it well-suited for large-scale production.

4.1.2. Chemical exfoliation

Chemical exfoliation one of the earliest methods for creating thin 2D sheets, often involves intercalation, where specific molecules or ions are inserted between adjacent layers of host materials, causing expansion and exfoliation by reducing the strength of vdW forces [227]. This process causes expansion and exfoliation by decreasing the vdW forces. The physical characteristics of the host 2D materials can undergo significant and remarkable alterations depending on the type of intercalate species used. For example, lithium intercalation using a solvent-free technique can induce a reversible phase transition in TMD flakes, transforming them from semiconducting to metallic phases [228,229]. A facile and cost-effective technique to produce few-layer graphene sheets involves exfoliating graphite in a binary-component system of peroxyacetic acid (PAA) and sulfuric acid [230] (Fig. 3d). This process is simple and feasible, transforming graphite into few-layer graphene sheets with nearly 100% yield without the need for mechanical action such as stirring or sonication.

Electrochemical exfoliation, a specific type of chemical exfoliation, can be scaled up to generate substantial amounts of TMD flakes in a liquid solution using an electrical current [231,232]. Ions generated in the electrolyte are inserted between the layers of the bulk 2D materials, weakening the vdW bonds and causing layer separation. This method is relatively simple to develop, assemble, and utilize under normal conditions. The electrochemical exfoliation process involves three primary stages: electrochemical production of ions within the electrolyte, diffusion and insertion of these ions between the bulk material layers, and an electrochemical reaction that transforms the ions into gases within the material, resulting in layer separation. Efficient electrochemical exfoliation of graphite using ammonium sulfate, potassium sulfate, and sodium sulfate has been reported, producing high yields of thin graphene sheets [233].

4.2. Bottom-up direct growth

Bottom-up methods such as CVD and MBE expand the possibilities for the mass production of high-quality 2D materials and applications in various fields [234]. CVD is a highly scalable method capable of producing large area 2D materials with excellent uniformity and purity. It involves the reaction of gaseous precursors on a heated substrate, forming a thin film of the desired material. CVD has been successfully used to synthesize various 2D materials. The ability to precisely control growth parameters such as pressure, temperature, and precursor flow rates makes CVD a versatile and widely adopted technique in the field.

MBE allows for the atomic-level control of film growth. In this technique, molecular or atomic beams of the constituent elements are directed at a heated substrate in an ultra-high vacuum environment, enabling the layer-by-layer construction of 2D materials. MBE is known for producing high-quality epitaxial films with sharp interfaces and minimal defects, making it suitable for electronic and optoelectronic applications.

Advanced dielectrics can lower the threshold voltage, reduce device power consumption, and eliminate hysteresis, thereby enhancing device stability [235–237]. Initial applications of atomic layer deposition (ALD) for depositing high-κ oxide gate dielectrics in 2D semiconductors involved creating thicker gate dielectric layers in prototype devices [238] (Fig. 3e). However, the lack of nucleation sites on the 2D semiconductor surface often results in uneven oxide films when directly grown by ALD, leading to local defect states that trap charges, cause hysteresis, and induce leakage in thin gate dielectrics.

To address these challenges, vdW interfaces, such as h-BN, have been utilized as gate dielectric insulating layers. These interfaces maintain atomic-level flatness, encapsulate the 2D semiconductor from the external environment, and significantly improve intrinsic characteristics like mobility and stability. A buffer layer or seed layer technique has been developed to integrate high-quality, ultra-thin high-κ dielectrics on 2D semiconductors [238] (Fig. 3f). Typically, an ultrathin metal or oxide layer is first deposited on the 2D semiconductor by thermal evaporation before applying the ALD gate dielectric. This buffer layer provides the necessary nucleation sites for depositing ultrathin high-κ dielectrics. Despite improvements, evaporated metal films can suffer from roughness and damage caused by energetic metal ions.

To mitigate these issues, organic materials such as perylene-3,4,9,10-tetracarboxylic dianhydride (PTCDA) have been proposed as seed layers. High-quality gate dielectrics with a 1 nm equivalent oxide thickness (EOT) prepared from PTCDA seed layers meet the practical application requirements in the industry. In situ oxidized high-κ dielectric layers, akin to natural Si oxide for Si, hold promise for 2D semiconductors, with oxidation methods such as O₂ plasma, and calcination playing a crucial role in enhancing surface properties and performance for high-performance transistors and other electronic applications. For example, Bi₂O₂Se can generate its own high-κ oxide through precise layer-by-layer oxidation [238] (Fig. 3g). The oxidized Bi₂SeO₅ has a high dielectric constant of 21 and maintains good leakage properties at 0.9 nm EOT. However, in situ oxides are only effective for specific oxidable 2D semiconductors, limiting the choice of materials.

Moreover, liquid-phase chemical synthesis methods, although challenging, offer potential for scalable production of 2D materials. Solution-based synthesis techniques, such as solvothermal and hydrothermal methods, can produce nanosheets with controlled thickness and lateral dimensions. These methods involve the reaction of precursors in a liquid medium under high temperature and pressure conditions, leading to the formation of 2D materials.

4.3. 2D materials transfer

4.3.1. Physical transfer

The process for transferring 2D materials onto a target substrate often begins by coating polymethyl methacrylate (PMMA) onto the 2D material synthesized on a donor substrate. The stack is then dipped into an etchant to dissolve the donor substrate, leaving the PMMA/2D material floating on the etchant surface. The PMMA/2D material is then transferred onto a target substrate, followed by PMMA removal in a suitable solvent [239] (Fig. 3h). This method allows privileged substrate selection and reduces mechanical damage to the 2D material. However, complete removal of PMMA is challenging and can degrade the electrical and optical properties. To address these issues, alternative methods using other polymers or solvent treatments have been explored. Recently, residue-free wet transfer of 2D materials using rosin and paraffin as transfer films has shown promise for high-performance flexible electronic devices. The use of CVD and MOCVD presents a promising approach for the large-scale synthesis of TMDs. CVD allows for precise control over vapor-phase chemistry during synthesis by monitoring the introduction of metal and chalcogen precursors, enabling the production of high-quality films. The CVD process is compatible with established silicon technology but typically requires high-temperature conditions (700–1000 °C) [240]. These elevated temperatures, essential for CVD treatment and representing a step in physical transfer, pose challenges for direct growth on low-thermal-budget substrates like flexible polymers and glass. Consequently, layer transfer from specialized growth substrates becomes necessary, underscoring the importance of clean, damage-free, and scalable transfer techniques.

4.3.2. vdW transfer

vdW transfer is based on the weak van der Waals forces between the layers of 2D materials and the substrate. These forces enable the layers to be separated effortlessly during the transfer process, all while maintaining the atomic structure of the material intact. h-BN is effective as a gate dielectric when the dielectric layer does not need to be too thin [241]. Mechanical exfoliation of h-BN produces high-quality material, though the sizes obtained are small, limiting large-scale preparation [242]. The dielectric constant of h-BN is around 5, similar to Al₂O₃, but it is not a high-κ dielectric. For applications requiring smaller equivalent oxide thickness (EOTs), the thin physical thickness of h-BN would lead to relatively large leakage current [243]. To overcome this limitation, the development of automated and precise mechanical exfoliation techniques is essential. Advanced methods using robotic systems have been proposed to perform high-speed optical microscopy and computer vision algorithms to identify and transfer 2D flakes, creating complex vdW heterostructures with high precision and minimal contamination [238] (Fig. 3i).

4.3.3. Direct transfer with patterning

Direct transfer with patterning in 2D materials can be achieved by using nanostructured substrates. Dettlaff-Weglikowska and co-workers placed graphene on a periodic grid of hydrogen silsesquioxane (HSQ) resist defined by electron-beam lithography [244] (Fig. 3j). This method allowed precise control over the features and periodicity of the grid, resulting in a uniform pattern with small features (approximately 10 nm). This approach is advantageous for fabricating devices with highly controlled nano-architectures. However, it is crucial to address challenges related to mechanical damage and interfacial contamination during the transfer process. The development of roll-to-roll transfer [245,246] and the use of thermal release tapes (TRTs) [247] and self-release layers (SRLs) [248] have shown potential for scaling up the transfer of large-area 2D materials.

5. Facial fabrication of flexible electronics

Advancements in facial flexible electronics fabrication include techniques such as initiated chemical vapor deposition (iCVD), diverse printing methods (e.g., inkjet, screen, and extrusion), graphene-assisted metal transfer, PMMA-assisted dielectric transfer, and strain-engineered architectural designs, all aimed at achieving high flexibility, robustness, and seamless integration of 2D materials.

Despite these advancements, challenges remain, particularly in ensuring the uniformity and scalability of high-quality 2D materials while maintaining consistent device performance. Future efforts are likely to focus on developing low-temperature growth and transfer methods to minimize defects and contamination. Additionally, strain engineering, incorporating mechanical deformations like bending, stretching, and twisting, will play a vital role in enhancing device durability and functionality, providing a comprehensive framework for advancing flexible electronics fabrication.

5.1. Facial fabrication

5.1.1. Initiated CVD

To achieve mechanical flexibility, selecting suitable substrates with appropriate mechanical properties is essential. Commonly used polymeric substrates, such as PDMS, parylene-C, polyethylene terephthalate (PET), polyimide (PI), and polyethylene naphthalate (PEN), often cannot withstand high temperatures, presenting challenges due to the thermal expansion coefficient mismatch between the polymeric substrate and 2D materials [249]. This mismatch can result in undesired film stress and cracking during fabrication, particularly during high-temperature processes like electrode and dielectric deposition. Therefore, low-temperature processes such as ALD and physical vapor deposition (PVD) are preferred for constructing electrodes and dielectrics on flexible substrates.

Thin polymeric materials offer mechanical flexibility but exhibit thermal expansion coefficients 10 to 30 times higher than inorganic materials like Si and MoS₂, leading to significant issues during high-temperature fabrication steps. Alternative ultrathin materials such as ultra-thin glass (UTG) have been explored, offering better tolerance to moderate temperatures (~400 °C) due to a lower thermal expansion coefficient mismatch [250]. However, UTG’s brittle nature and limited ability to withstand large strains offer only modest mechanical advantages.

Direct synthesis of most 2D materials, including graphene, TMDs, and h-BN, typically requires temperatures above 600 °C, making it challenging to grow these materials directly on flexible substrates. This necessitates the development of low-temperature growth techniques. Several works have successfully demonstrated the low-temperature growth of diverse 2D materials using volatile intermediates or metal-organic precursors. Conventionally, 2D materials are grown on rigid substrates and then transferred onto flexible substrates, allowing the use of high-quality materials on various flexible platforms [251]. However, this approach often results in residues, contamination, and molecule trapping at interfaces.

Recent advancements in transfer techniques aim to address these issues. For instance, using rosin [252] and paraffin [253] as transfer films has shown promise for achieving residue-free wet transfer, enhancing the performance of flexible electronic devices. Additionally, dry transfer methods relying on the difference in adhesion strength between the transfer film and the target substrate have been developed to achieve clean and damage-free transfer of large-area 2D materials.

To further improve the fabrication of stretchable devices, initiated CVD (iCVD) is employed. iCVD is a vacuum deposition technique that allows for the polymerization of monomers directly on the substrate surface [254–257]. This method provides several specific advantages, including high areal uniformity, conformal coating capabilities, and the ability to deposit polymers with precise control over composition and thickness. iCVD is useful for creating robust and elastic polymer dielectrics, essential for high-performance stretchable electronics.

Using iCVD, a dielectric comprising isononyl acrylate (INA) and 1,3,5-trimethyl⁻¹,3,5-trivinyl cyclotrisiloxane (V3D3) monomers is deposited. INA acts as a soft segment, while V3D3 maintains the cross-linked polymer structure, providing robust insulation performance [258] (Fig. 4a). The iCVD process forms a robust and conformal interface with a carbon nanotube (CNT) network. The fabricated intrinsically stretchable transistors exhibit high performance and low operation voltage. The wafer-scale fabrication of stretchable CNT transistor arrays demonstrates their potential for creating stretchable p-type inverters and logical gates (NAND, NOR, and XOR) that function even under 40% applied strain.

Fig. 4.

Facial Fabrication of Flexible Electronics.(a) Stretchable logic gates were fabricated with these transistors including the initiated chemical vapor deposition (iCVD) process for depositing the stretchable polymer dielectric using INA and V3D3 monomers, and a wafer-scale fabrication image of the stretchable CNT transistor arrays. INA: Isonicotinic acid; V3D3: 1,3,5-Trivinyl-1,3,5-trimethylcyclotrisiloxane. (b) An asymmetric micro-supercapacitor (AMSC) fabricated via inkjet-printing of chemically exfoliated c-MoS₂ and rGO inks. (c) Graphene-assisted metal transfer printing process for weakly and strongly adhering metals, which includes six metals on a wafer-scale Gr/Ge substrate; physical peeling of metal patterns with PVA film and corresponding Au patterns; printing metal patterns on a target substrate; and removing PVA by dissolution in water with transferred Au patterns on a SiO₂ wafer. (d) PMMA-assisted transfer of a large area Al₂O₃ layer to create a wrinkled dielectric. (a) (Reproduced with permission from ref. [257]. Copyright 2023, Springer Nature). (b) (Reproduced with permission from ref. [267]. Copyright 2020, American Chemistry Society). (c) (Reproduced with permission from ref. [268]. Copyright 2022, Springer Nature). (d) (Reproduced with permission from ref. [269]. Copyright 2020, Springer Nature).

5.1.2. Ink printing

The all-solution-based processing approach for synthesizing and printing 2D materials on flexible substrates at low temperatures holds significant promise for the advancement of flexible electronic devices. While spin coating is the most used method for depositing thin and transparent films, it is limited by a small coating area and cannot be used for patterning. Screen printing can be used for pattern design and mass production, where ink is pushed through a pre-patterned stencil onto the substrate [259]. This method requires inks with high viscosity and shear-thinning properties to flow through the stencil and retain pattern integrity under pressure. Automated and roll-to-roll modified systems enable rapid printing at speeds of 70–100 m min⁻¹ [260]. Developing high-viscosity inks without binders is crucial to eliminate post-treatment and maintain the functional properties of 2D materials. Roll-to-roll screen printing processes are extremely fast but are associated with high prototyping costs and require large quantities of ink. Digitally controlled inkjet printing is well-suited for laboratory-scale device fabrication, as it enables precise ink deposition, resulting in high printing resolution. However, strict ink parameters are necessary for uniform printing. To obtain reproducible and precise ink deposition, it is crucial to adjust the rheology, drying behavior, wettability, and substrate adhesion of the ink to suit specific deposition techniques.

Getting the right rheological properties for deposition methods can be challenging when relying solely on pure solvents. It is common to modify the physicochemical characteristics of solvents by adding polymeric additives like surfactants or polymers. For instance, graphene is difficult to exfoliate in water through liquid phase exfoliation alone. Therefore, researchers were able to produce a graphene dispersion by adding sodium cholate to water and subjecting graphite to sonication [261], making it suitable for mass production. Conversely, water-based TMDs dispersions are often not suitable for inkjet printing due to low viscosity [262–264]. To solve this issue, a combination of propylene glycol and Triton X-100 was used to enhance the viscosity and reduce the surface tension of water-based TMDs dispersions for inkjet printing [265,266]. Thus, selecting the appropriate deposition method and developing detailed processing steps are crucial for the successful deposition of 2D inks and achieving the desired patterns and functionalities for specific applications.

Inkjet printing is a direct ink-writing technique that accurately deposits inks to form patterns through digitally controlled ejection. It offers the added benefit of generating 3D structures [267]. However, producing high-quality inkjet prints can be challenging due to the significant influence of ink parameters and printing conditions on the deposition process. The printability of inks for inkjet printing can be predicted using the dimensionless inverse Ohnesorge number (Z), which depends on the ink’s viscosity, density, and surface tension, as well as the nozzle diameter. While Z values between 1 and 14 generally produce stable ink jetting, successful examples of inks with Z values beyond this range have also been reported. To adapt to the jetting process, inks should have low viscosity, which can lead to the accumulation of deposited material on the periphery during drying (the coffee ring effect). Various methods, such as using high-boiling-point solvents, adding surfactants or binders, using two solvents, and treating the substrate surface, can be employed to alleviate this effect and improve printability. To achieve uniform patterns with smooth lines and even edges, optimization of the drop frequency, substrate temperature, and printing speed is necessary. Inkjet printing of chemically exfoliated metallic 1T-MoS₂ and reduced graphene oxide (rGO) inks was developed to fabricate asymmetric micro-supercapacitors (AMSCs) using [268] (Fig. 4b). This method offers a low-cost and scalable solution for producing flexible electronic devices.

Inkjet printing also enables the creation of non-continuous microarray patterns. Modified techniques like electrohydrodynamic jet and aerosol jet printing offer higher spatial resolution and homogeneity, thereby enhancing the printing quality. Extrusion printing is another technique that enables the fabrication of complex 3D structures with new functionalities. This technique involves preparing paste-like inks with distinctive shear-thinning characteristics, accompanied by a sufficiently high elastic modulus and shear yield stress. The properties of 2D inks can be altered by introducing additives, increasing the solid content, and enhancing the aspect ratio of 2D sheets to elevate viscosity. Furthermore, the surface chemistry of 2D materials offers rich tunability, enabling the modification of ink properties through adjustments to the colloidal stability of the nanoflakes.

5.1.3. Graphene-assisted metal transfer printing

The graphene-assisted metal transfer printing process enables the efficient transfer of both weakly and strongly adhering metals onto various substrates. This method leverages the unique vdW surface properties of 2D materials to create freestanding metal films, which can then be transferred with high efficiency and minimal damage. The fabrication process begins with the preparation of a wafer-scale single-crystal monolayer graphene on a Ge (110) substrate using CVD. Metals such as Cu, Ag, Au, Pt, Ti, and Ni are then deposited onto the graphene/Ge substrate [269] (Fig. 4c). A polyvinyl alcohol (PVA) film is coated onto the metal/graphene/Ge surface. Due to the weak vdW forces at the metal/graphene interface, the metal films can be transferred with nearly 100% yield when the PVA film is peeled off. These metal patterns are subsequently printed onto a target substrate, and the PVA film is dissolved in water, leaving the metal patterns transferred onto substrates like Al₂O₃ wafers with high yield and minimal damage to the 2D materials.

This process is different from 3D printing, which involves the layer-by-layer deposition of materials to build up a three-dimensional object. Metal transfer printing involves the preparation of a metal film on a 2D material substrate, which is then transferred to a target substrate in a single step, resulting in flat, high-quality metal patterns suitable for electronic and photonic applications. By leveraging the unique properties of graphene and vdW interactions, the graphene-assisted metal transfer printing process provides a robust, efficient, and scalable method for fabricating high-quality metal patterns on various substrates, enhancing the development of advanced electronic devices.

5.1.4. PMMA-assisted transfer of dielectric layers

The PMMA-assisted transfer technique involves transferring a large area of Al₂O₃ dielectric layer with insulating properties to create a wrinkled dielectric, suitable for flexible electronics [270] (Fig. 4d). The fabrication process begins by coating PMMA onto the dielectric layer. The PMMA serves as a support film during the transfer process. The original substrate holding the dielectric layer is etched away, releasing the PMMA/dielectric stack. This stack is then transferred onto the target substrate. Finally, the PMMA is removed by dissolving it in a suitable solvent, leaving behind the dielectric layer on the new substrate. For successful transfer, the target substrate should have a compatible surface that ensures good adhesion with the transferred dielectric layer. Additionally, the etching process should be carefully controlled to avoid damaging the dielectric layer. This method effectively addresses common issues such as residues and contamination, which are often encountered in other transfer techniques, thereby enabling the development of high-performance flexible devices on a large scale.

Moreover, the PMMA-assisted transfer technique focuses on transferring pre-formed layers of material onto a new substrate. This method is beneficial for creating uniform dielectric layers over large areas, which is crucial for the performance and reliability of flexible electronic devices [271,272]. The PMMA-assisted transfer technique provides several advantages, including the ability to transfer large-area dielectric layers with minimal damage and contamination. This makes it suitable for scalable processes and the incorporation of high-performance materials into stretchable and flexible electronics.

5.2. Strain engineering

Flexible devices should withstand various mechanical deformations such as shearing, stretching, and bending. These deformations can induce performance degradation and fatigue failure over repeated cycles. Despite their ultrathin nature and high flexibility, 2D materials are susceptible to strain-induced damage, including crack formation and propagation, buckling, and delamination. To enhance the mechanical reliability of 2D materials-based flexible devices, strain management strategies are essential. This involves adopting appropriate device architectures and selecting suitable materials for substrates and dielectrics.

5.2.1. Architecture design

An essential characteristic of flexible devices is their capacity to maintain performance and inherent characteristics under various forms of deformation, such as bending, twisting, or stretching. While 2D materials are intrinsically more strain-resistant than conventional inorganic semiconductors, using pristine 2D materials alone does not provide the mechanical flexibility required for stretchable or flexible devices. For example, human skin can stretch up to 60%, necessitating a high degree of elasticity in devices that adhere to the skin.

To address this challenge, several structural innovations have been developed to improve the mechanical tolerance of 2D materials. These include serpentine, buckling, origami, and kirigami designs, which are incorporated into interconnects and active components [273] (Fig. 5a). The serpentine structure, with interconnected arcs, deforms uniformly and predictably under strain [274–276]. This design allows the structure to expand and contract significantly without mechanical failure, making it suitable for applications including wearable electronics, sensors, energy harvesters, and flexible displays. Early applications of serpentine structures were mainly in metal interconnects [275,276] and later extended to 2D materials for advanced flexible devices [277,278]. For instance, serpentine-shaped ribbons consisting of graphene and Au electrodes have been used to develop e-tattoos for electrodermal activity sensing, showing significant stretchability and stability over many cycles of applied strain.

Fig. 5.

Strain Engineering of Flexible 2D Material Devices. (a) Architecture designs for flexible devices, including serpentine, wavy (buckling) shapes, origami and kirigami for flexible devices. (b) Strain-engineered MoS₂ photodetector array on flexible PI film. (c) Schematic of bulging methods for straining 2D material-based devices. (d,e) Comparison of internal quantum yield of black phosphorus (BP) with various semiconductors (GaSb, PbSe, InAs) in the near-infrared (NIR) wavelength range, and strain effect on BP photoluminescence. (f) Schematic of thermal expansion mismatch and substrate bending to apply tensile strain to BP crystals. Transfer characteristics of the fabricated thin-film transistor (TFT) with (g) normalized field-effect mobility and (h) increasing tensile strain. (i) Vacuum assembly and dry transfer fabricating vdW heterostructures. (j) Modification of the supporting layer for the wet transfer. (a) (Reproduced with permission from ref. [33]. Copyright 2023, American Chemistry Society). (b) (Reproduced with permission from ref. [399]. Copyright 2021, American Chemistry Society). (c) (Reproduced with permission from ref. [33]. Copyright 2023, American Chemistry Society). (d,e) (Reproduced with permission from ref. [313]. Copyright 2021, Springer Nature). (f) (Reproduced with permission from ref. [33]. Copyright 2023, American Chemistry Society). (g,h) (Reproduced with permission from ref. [317]. Copyright 2022, American Chemistry Society). (i) (Reproduced with permission from ref. [323]. Copyright 2022, Springer Nature). (j) (Reproduced with permission from ref. [324]. Copyright 2022, Springer Nature).

Besides, the wavy structure involves patterning device components into a periodic sinusoidal wave shape in an out-of-plane architecture [279–281]. This is achieved by transferring materials onto a pre-stretched elastomeric substrate. When the pre-strain is released, the material forms a wavy pattern. The degree of pre-strain controls the wave pattern’s amplitude and wavelength. This design maintains material integrity and prevents fractures during stretching. For example, wavy graphene micro-ribbons have been used to create highly stretchable micro-supercapacitors that can withstand significant tensile strains without serious changes in resistance.

Origami structures leverage folding and unfolding along predefined crease patterns to achieve significant deformation while maintaining structural integrity [282–285]. These designs provide compact, foldable configurations useful for wearable electronics and biomedical devices. The flexibility and robustness of crease patterns accommodate large deformations without compromising device stability. For example, origami-based graphene hygroelectric generators maintain consistent electricity generation under complex deformations, including rolling and stretching. Additionally, origami-inspired morphable 3D structures using graphene/Au electrodes have been used to produce stretchable displays that maintain image quality under stretching.

Kirigami designs involve cutting and folding thin materials to create three-dimensional structures capable of complex deformations such as twisting and bending [286,287]. This approach allows the use of rigid or non-stretchable materials to produce ultra-stretchable devices. Kirigami-based structures endure substantial deformation while maintaining functionality. For instance, kirigami-inspired graphene strain sensors retain reliable electrical properties under tensile strains up to 240% [288,289]. Additionally, integrating PtSe₂ layers on kirigami-patterned substrates has demonstrated exceptional stretchability and strain-tunable photoresponse [290,291].

5.2.2. Mechanical design

The bending limit of a device on a flexible substrate is defined by the minimum bending radius (r_min), expressed as r_min=t_s/2ε_max, where t_s is the substrate thickness and ε_max is the fracture strain limit of the active material [292,293]. Thinner substrates result in a reduced bending radius, increasing device flexibility. However, substrates cannot be reduced below a certain thickness without compromising device handling [294]. Flexible devices typically use substrates around 10–50 μm thick, with device layers including active material, dielectric, interconnects, and passivation being roughly 1 μm or less. Within a bent substrate, a neutral mechanical plane (NMP) exists where minimum strain occurs, transitioning from tensile to compressive strain [295,296]. Positioning active 2D materials near the NMP reduces strain-induced failure risks. For thick substrates (t_s≫t_d), the NMP can be considered at the substrate’s center, while for thin substrates (t_s~t_d), all device layers should be considered in determining the NMP location.

When fabricating 2D materials-based devices on flexible polymeric substrates, one of the challenges is the degradation of their electrical and optoelectronic properties, such as on-state current, charge carrier mobilities, and photoresponse. Strain engineering offers a powerful solution to this problem by enhancing the inherent properties of the active 2D materials, enabling 2D materials to be strongly modified with their inherent electronic, optical, magnetic, and transport properties, leading to novel functionalities and capabilities in strained devices.

Applying strain to 2D materials deforms their lattice structure, manipulating the electronic band structure by changing the location of the conduction band minima or valence band maxima in the Brillouin zone [297]. This variation can modulate resistivity, alter the bandgap size, and even change the bandgap nature from indirect to direct or vice versa [298]. Additionally, strain-induced shifts in the conduction or valence bands can reduce inter-valley phonon scattering, leading to increased charge carrier mobility. Consequently, strain engineering provides an opportunity to dynamically tune and enhance the performance of 2D materials based flexible electronics.

Several methods have been developed to induce strain in 2D materials, including using the mismatch in thermal expansion coefficients between substrates and 2D materials [299–302], hetero-epitaxial growth [303,304], applying hydrostatic pressures [305], bending, bulging [306], indenting [307], stretching, creating blisters [308] or bubbles [300], crumpling [309,310], and transferring 2D materials onto textured surfaces [311]. These techniques have been employed to create local strain and study its effects on the fundamental characteristics of 2D materials.

For practical applications, it is crucial to apply strain over a large area in a controllable and dynamically tunable manner. Methods such as bulging, bending, and stretching are commonly used and have shown promise in introducing strain into 2D materials-based flexible devices fabricated on large-area substrates. These approaches offer a high degree of dynamic control and reproducibility. Bulging is promising for inducing higher levels of biaxial strain in 2D materials-based devices in a highly controllable manner. This involves mounting and sealing flexible devices on a gas-filled cavity, which bulges upon increasing the gas pressure inside [312] (Fig. 5b,c). This pressure-induced bulging expands the substrate and incorporates biaxial strain into the device components. To ensure effective strain transfer in all these approaches, it is essential to avoid slippage by ensuring proper bonding within the device layers (active 2D material, metal electrode, dielectric layer, and encapsulation) and strong adhesion between the devices and their supporting substrates. For example, a strain-modulated photodetector array based on a graphene/MoS₂/graphene metal-semiconductor-metal (MSM) structure was fabricated on a flexible polymeric substrate [313] (Fig. 5b). The photodetector array, biaxially stretched to a tensile strain of up to 1.19% using a pneumatic bulging approach, demonstrated near-infrared (NIR) imaging capabilities even in foggy environments. Strain-induced piezopotential in non-centrosymmetric 2D semiconductors, such as TMDs, can improve photoresponse performance by coupling photonic properties with the piezotronic effect. This coupling, known as the piezo-phototronic effect, enhances photocarrier generation, separation, transport, and recombination under externally applied strain. Initial experiments demonstrated significant performance improvements in monolayer MoS₂-based strain-gated flexible optoelectronics, and subsequent studies have further explored this effect in 2D layers and heterojunctions. BP shows significant strain sensitivity, affecting its photoluminescence and internal quantum yield. Comparison with other semiconductors such as GaSb, PbSe, and InAs in the NIR wavelength range highlights BP’s superior performance under strain. The strain effect on BP photoluminescence can be harnessed for advanced optoelectronic applications, enhancing both detection and emission capabilities [314] (Fig. 5d,e).

Flexible devices are often designed to minimize strain on active sensing materials. However, strain can positively regulate the optoelectronic properties of semiconductors. Bending the substrates containing electronic devices can induce uniaxial strain in the device components, with the strain value controlled by adjusting the radius of curvature. It is important to note that this method accommodates both compressive and tensile strains, with compressive strain occurring on one side and tensile strain on the opposite side. The uniaxial strain applied to a device on a bent substrate can be estimated using the formula ε=t/2 R, where ε is the applied strain value, t is the substrate thickness, and R is the radius of curvature [315,316]. However, bending is limited to applying only uniaxial strains. Applied strain to a 2D material can enhance optical absorption, increase carrier mobility, and generate piezopotential to improve photocarrier generation and transport. For instance, a flexible photodetector (PD) fabricated with CVD-grown WS₂ film demonstrated tunable photoresponse and increased photocurrent under tensile strain due to bandgap narrowing [317]. Similarly, monolayer MoS₂-based PDs showed increased photocurrent with applied tensile strain. High-performance strain-tunable infrared optoelectronics, including both light detection and emission, have been demonstrated using BP [318] (Fig. 5f–h).

2D materials like TMDs and graphene are highly sensitive to external strain, which significantly affects their electrical and optical properties. For example, multilayer MoS₂ can transition from semiconducting to metallic under compressive pressure due to increased S–S interactions in the compressed interlayer space [319]. Strain engineering has been shown to decrease the band gap and enhance mobility in WSe₂ and MoS₂ FETs on flexible substrates [320]. In flexible MoS₂ transistors fabricated with CVD-grown monolayers on freestanding PEN substrate, uniaxial tensile strain of 0.7% nearly doubled the on-state current and mobility, with electrical characteristics returning to their initial state after releasing the strain [318] (Fig. 5g,h).

Stretching is an effective approach to transferring the desired strain to active 2D materials in devices. This can be achieved through mechanical stretching, thermal expansion, or mounting the device on a piezoelectric crystal, allowing for biaxial strain in both the x and y directions. For example, using a pre-stretched elastomeric substrate can effectively engineer local strain in 2D materials. When the tension is released, the 2D material layers form wrinkles due to buckling-induced delamination, introducing uniaxial tensile strain because of severe local bending, significantly affecting their electronic, optical, and mechanical properties [321,322].

Transfer techniques are crucial for the large-scale synthesis and integration of 2D materials, particularly when using high-temperature methods like CVD. Dry transfer methods could minimize contamination and physical damage [323]. Recently, semi-automated and autonomous systems have been developed to address the mechanical complexity and precision required for fabricating vdW heterostructures [324] (Fig. 5i). These systems represent advanced dry transfer methods. A semi-automated transfer process using a homemade mechanical apparatus has been developed, while an autonomous robotic system capable of performing vdW 2D flake pick-up and transfer has also been showcased. This system uses a high-speed optical microscope, a motorized XY stage, and computer vision algorithms to identify and transfer 2D flakes, creating complex vdW stacks for device fabrication. The system identifies over 400 monolayer graphene flakes per hour with an error rate of less than 7% and has prepared heterostructures with up to 29 layers. Further advancements include a four-dimensional pixel assembly method for manufacturing vdW solids with precise design, large area coverage, and angle control. This method enabled the creation of vdW solids with up to 80 individual layers and facilitated efficient optical spectroscopic assays, revealing new excitonic and absorbance layer dependencies in MoS₂.

Conversely, wet transfer methods, such as the PMMA-assisted process, are commonly used for transferring 2D materials [325] (Fig. 5j). In this process, PMMA is coated onto the 2D material on its growth substrate. The stack is then dipped into an etchant to dissolve the donor substrate, leaving the PMMA/2D material floating on the etchant surface. This stack is then transferred onto the target substrate, and the PMMA is removed in a suitable solvent. This method allows for residue-free transfer with minimal damage, which is crucial for scalable and industry-compatible fabrication of flexible electronic systems. However, complete removal of PMMA can be challenging and can sometimes degrade the electrical and optical properties of the 2D material [326].

6. Integrated circuit designs using 2D materials

After the successful fabrication of 2D materials and subsequent strain engineering, their mechanical properties have been significantly enhanced. In addition to these improvements, advancements in circuit design for 2D materials have further expanded their potential for biosensing applications. These developments enable greater miniaturization and flexibility of 2D bioelectronics while exploring enhanced performance metrics, such as improved carrier mobility and efficiency.

The integration of integrated circuits (ICs) into flexible devices is critical for their functionality in performing data processing, power management, and communication. Wearable devices, often constrained by limited space and battery life, require ICs that are miniaturized and high performing [327]. Traditional Si-based ICs, being rigid and brittle, are not suitable for such applications. To overcome these constraints, 2D materials like graphene, TMDs, and BP are being explored to develop flexible ICs. These materials offer mechanical flexibility and improve electrical performance, paving the way for high-performance multifunctional circuits in deformable electronics. A notable example is a MoS₂-enabled flexible rectenna designed for RF energy harvesting [328] (Fig. 6a). This device, fabricated on Kapton film, effectively collects RF power in the 5.9 GHz Wi-Fi channel and produces a rectified output voltage reaching up to 250 mV. The rectenna utilizes a lateral Schottky diode patterned into a metallic-semiconducting (1 T/1 T’−2H) phase heterostructure, demonstrating a cutoff frequency of around 10 GHz at zero bias, indicating potential for battery-free applications. Additionally, a flexible RF amplifier utilizing mechanically exfoliated MoS₂ transistors has been demonstrated [329] (Fig. 6b). These transistors, fabricated on a flexible polyimide substrate, exhibited an intrinsic cutoff frequency of 13.5 GHz and a maximum oscillation frequency of 10.5 GHz. The RF amplifier achieved a relative voltage gain larger than unity for a 300 MHz input signal.

Fig. 6.

Flexible and Integrated Circuit Designs Using 2D Materials. (a) Circuit diagram of a MoS₂-based radio frequency (RF) energy harvesting device, and photograph of the MoS₂ rectenna on Kapton film. (b) Circuit layout of an RF amplifier with two MoS₂ thin-film transistors (TFTs) and photograph of the MoS₂ RF amplifier on a flexible substrate. (c) Circuit layout of an amplitude modulation (AM) demodulator with a BP field-effect transistor (FET) and its optical image. (d) Circuit diagrams of various logic gates: inverters, NOR, NAND, SRAM, and AND gates; and photograph of flexible integrated circuits with MoS₂. (e) Device layout of a 3-stage ring oscillator and photograph of the device on a flexible platform. (a) (Reproduced with permission from ref. [327]. Copyright 2019, Springer Nature). (b) (Reproduced with permission from ref. [328]. Copyright 2014, Springer Nature). (c) (Reproduced with permission from ref. [329]. Copyright 2015, American Chemical Society). (d) (Reproduced with permission from ref. [330]. Copyright 2020, Springer Nature). (e) (Reproduced with permission from ref. [331]. Copyright 2023, American Chemical Society).

Flexible electronics also benefit from amplitude modulation (AM) demodulators, which extract signal information from RF carriers. An example is a BP-based flexible AM demodulator fabricated on polyimide substrates [330] (Fig. 6c). This device showcases the potential of BP FETs in flexible RF applications, maintaining performance under mechanical deformation.

CVD is an optimal technique for synthesizing 2D materials for large-scale digital circuit fabrication due to its ability to produce high-quality, uniform films with controlled thickness. For instance, large-scale fabrication of logic gates using monolayer MoS₂ has been demonstrated, with integrated flexible devices including NOR, NAND, SRAM, and AND gates [331] (Fig. 6d). These circuits operated stably under bending deformation, highlighting their potential for high-performance flexible electronics.

Ring oscillators are another essential component in digital circuits, used for timing and signal processing. A flexible photoresponsive ring oscillator consisting of MoS₂ transistors was fabricated on a polyimide substrate [332] (Fig. 6e). This three-stage ring oscillator demonstrated excellent flexibility and fit on curved surfaces. Its output oscillating frequency was tunable from 0 Hz to 2.35 kHz based on light intensity, and it exhibited significantly reduced power consumption compared to conventional CMOS pixels, consuming only 12.4 nW at a light intensity of 10 mW cm⁻².

7. Biopotential sensing

Biopotential sensing measures electrical signals generated by the body, which are crucial for monitoring physiological functions. There are several types of biopotential sensors, each designed to detect specific bioelectric activities. Strain sensors measure mechanical deformation or strain in tissues, converting these mechanical changes into electrical signals. They are typically used to monitor physical changes, such as muscle contractions or respiratory movements. While strain sensors focus on physical deformation, electromyography (EMG) sensors directly measure the electrical activity produced by muscle fibers during contraction. EMG sensors provide insights into the neuromuscular activation of muscles, which is crucial for diagnosing neuromuscular disorders and assessing muscle function in rehabilitation settings.

Moreover, ECG sensors measure the electrical activity of the heart. They are widely used to monitor heart rate, rhythm, and detect abnormalities in cardiac function. ECG signals are crucial for diagnosing conditions like arrhythmias and myocardial infarctions. Electroencephalography (EEG) sensors capture the electrical activity of the brain. They are essential for diagnosing and monitoring neurological disorders, including epilepsy, sleep disorders, and brain injuries. EEG provides insights into brain function by recording voltage fluctuations resulting from neuronal activity.

Flexible and stretchable electronics have transformed personal healthcare and activity tracking by enabling continuous monitoring of physiological indicators. These devices can conform to dynamic surfaces such as human skin and tissues [333–335]. To fully realize the potential of these technologies, it is essential to develop devices with suitable electrical and mechanical properties and to integrate them into comprehensive systems. 2D materials, with their superior properties, have been at the forefront of advancements in wearable strain sensors for health monitoring [336,337].

7.1. Strain sensing

Wearable strain sensors are crucial for accurately collecting information from the human body. These sensors should conform to the skin’s surface, ensuring high signal-to-noise ratios, minimal motion artifacts, and comfort. Given the dynamic nature of human skin, which can stretch up to 15% elastically and endure up to 30% strain [338], sensors need to accommodate significant deformation. The strain involved in daily body movements can reach up to 100%, necessitating highly flexible and stretchable sensors. 2D materials are well-suited for developing wearable strain sensors attributed to their exceptional mechanical flexibility, electrical conductivity, and optoelectronic properties. These materials can be processed into solutions compatible with scalable manufacturing techniques, such as printing onto low-temperature substrates like polymers, making them suitable for large-scale production.

There are several types of strain sensors that can be developed using 2D materials: Piezoelectric strain sensors generate an electric charge in response to mechanical stress [339]. They are suitable for applications that require precise detection of small deformations. While piezoresistive strain sensors change their electrical resistance when strained. They are simple to fabricate and integrate with electronic systems, offering a straightforward approach to strain sensing. Capacitive strain sensors measure changes in capacitance caused by deformation, providing high sensitivity and stability. They are useful for applications that require consistent performance over long periods. Triboelectric strain sensors generate voltage from mechanical motion. The development of 2D material-based wearable strain sensors has significantly advanced personal healthcare and activity tracking.

7.1.1. 2D materials-based strain sensing mechanism

7.1.1.1. Piezoresistive sensing.

2D materials are advantageous for piezoresistive sensors owing to their atomic-scale thickness, high Young’s modulus, and high elastic strain limit [340]. These properties enable stable operation under various deformations, strong adhesion to skin surfaces, and minimal discomfort due to their lightweight and transparent nature. A practical example of a piezoresistive strain sensor is the fabrication of a polystyrene (PS) nanoparticle-doped rGO sensor [341] (Fig. 7a). The process involves depositing a mixture of PS nanoparticles and GO fragments onto a PDMS film, followed by laser-scribe patterning. Laser scribing efficiently reduces GO to rGO, creating electrically conductive channels. The PS nanoparticles alter the stacking pattern of rGO fragments, enhancing resistance variation under strain. Raman spectroscopy confirms the reduction of GO to rGO, revealing structural defects and multilayer rGO fragments in the conducting channel. Graphene-based strain sensors are effective for real-time monitoring of biological signals like pulse rates. For instance, a strain sensor utilizing MoS₂-enhanced laser-induced graphene can monitor human motion [342] (Fig. 7b). When wrapped in an adhesive bandage and placed above the radial artery, it can measure arterial pulse rates accurately. Similarly, graphene-based sensors can track pulse rates before and after physical activity, showing increased rates after exercise [343] (Fig. 7c,d). Another example is a sensor based on MXene (Ti₃C₂T_x)/rGO aerogel, which can monitor the jugular venous pulse on the neck, demonstrating high strain sensitivity with a gauge factor of 237 and steady responses for wrist twisting [344] (Fig. 7e,f).

Fig. 7.

2D Materials-based Strain Sensing Mechanism. (a) Piezoresistive sensing: Polystyrene (PS)-rGO strain sensor: The laser scribing and PS modifying create a porous structure in rGO, which enhances its piezoresistive properties by increasing the surface area and the number of conductive pathways. (b)The piezoresistive strain sensor attached to the wrist of a volunteer for real-time pulse rate monitoring. (c,d) Graphene-based strain sensor for real-time monitoring pulse rate during pre- and post-physical activity. (e,f) MXene/rGO-based piezoresistive strain sensor for arterial pulse monitoring. (g) Capacitive sensing: Graphical representation of the MXene/polyvinyl alcohol (PVA) hydrogel-based capacitive strain sensor. (h) Capacitance response of the sensor during epidermal movement. Inset shows the capacitance response in four stages of drinking water. (i,j) Piezoelectric sensing: In₂Se₃-based self-powered piezoelectric sensor for real-time monitoring breath rate. The piezoelectric effect is enhanced by the 2D nature of In₂Se₃, which provides a large surface area for stress interaction and charge accumulation. (a) (Reproduced with permission from ref. [340]. Copyright 2018, Elsevier). (b) (Reproduced with permission from ref. [341]. Copyright 2019, American Chemical Society). (c,d) (Reproduced with permission from ref. [342]. Copyright 2017, American Chemical Society). (e,f) (Reproduced with permission from ref. [343]. Copyright 2018, American Chemical Society). (g,h) (Reproduced with permission from ref. [348]. Copyright 2019, Wiley). (i,j) (Reproduced with permission from ref. [204]. Copyright 2019, American Chemical Society).

7.1.1.2. Capacitive strain sensing.

Capacitive strain sensors offer low hysteresis, superb stability, and exceptional linearity, making them an attractive option for various applications, including health monitoring and wearable technology [345]. These sensors operate based on the principle of capacitance, which refers to a system’s ability to store an electrical charge. In capacitive strain sensors, changes in mechanical deformation alter the distance between conductive plates or the dielectric properties, leading to a measurable change in capacitance that can be related to the strain [346–348].

The structure of a typical capacitive strain sensor includes two conductive plates separated by a dielectric material. When strain is applied, the physical dimensions of the sensor change, altering the capacitance. For instance, a capacitive strain sensor using MXene (Ti₃C₂T_x)/PVA electrodes demonstrates this principle effectively. The sensor is attached to the throat to detect epidermal movement and records consistent electrical impulses during drinking. The capacitance response closely aligns with the swallowing process, showcasing the sensor’s effectiveness in real-time monitoring [349] (Fig. 7 g,h). Capacitive strain sensors have been developed for a broad range of applications due to their versatility. For instance, a pressure sensor with high sensitivity integrates a conductive micro-structured air-gap gate with 2D TMD transistors. This sensor provides significant advantages over traditional dielectric-based capacitive sensors, including tunable sensitivity and a broad pressure-sensing range. The MoS₂ and WSe₂--based devices exhibit average sensitivities of 44 kPa⁻¹ and peak sensitivities up to 770 kPa⁻¹. Additionally, by using the air-gap gate as a pressure-sensitive gate for transistor sensors, the pressure sensitivity can be amplified to ~10³-10⁷ kPa⁻¹ at a pressure regime of 1.5 kPa.

7.1.1.3. Piezoelectric strain sensing.

The piezoelectric effect occurs in specific materials whose crystal structure lacks a center of symmetry, allowing them to generate an electric potential when mechanically deformed by an external force [350–352]. In a typical piezoelectric strain sensor, applying an external force induces a deformation or strain in the crystal structure of the piezoelectric material [353]. This deformation causes a rearrangement of the dipoles within the lattice structure, leading to a displacement of electric dipoles and generating an electric potential difference across the material. This process allows for the direct conversion of mechanical energy into electrical signals, which can then be measured and analyzed. Piezoelectric sensors offer several advantages over other sensing technologies. They are highly sensitive and precise, capable of directly measuring small changes in pressure or force with high accuracy. This direct conversion mechanism eliminates the need for additional signal conditioning, enhancing the accuracy and reliability of the measurements. Moreover, piezoelectric sensors require low power to operate, making them suitable for wearable and battery-powered devices.

An example of a 2D material-based piezoelectric strain sensor is the In₂Se₃-based self-powered piezoelectric sensor. This sensor monitors breath rate by showing distinct frequency and current outputs during different breathing states. It accurately records breath rates, providing valuable data for physiological monitoring [286] (Fig. 7i,j). Such sensors demonstrate advanced capabilities in real-time monitoring of biological signals, providing robust, sensitive, and versatile solutions for healthcare applications.

7.1.1.4. Triboelectric strain sensing.

Triboelectric strain sensors (TPSs) operate based on the principles of triboelectric nanogenerators (TENGs). TENGs work by coupling triboelectrification with electrostatic induction, resulting in the generation of electrical signals from mechanical energy [354]. The working mechanisms of TENGs can be categorized into four types: single-electrode mode, vertical contact-separation mode, lateral sliding mode, and freestanding mode [355] (Fig. 8a).

Fig. 8.

Working Modes and Mechanisms of Triboelectric Pressure Sensors. (a) The four working modes of TENGs: vertical contact-separation mode, single-electrode mode, lateral-sliding mode, and freestanding mode. (b) The sensing mechanism in vertical contact-separation mode for triboelectric pressure sensors (TPSs) is depicted, highlighting the generation of electrical signals through contact and separation. (c) The working mechanism of single-electrode TPSs, demonstrating how a single electrode interacts with the environment to produce electrical signals. (d) Schematic illustration of the electric double-layer sensing mechanism, showcasing the formation of electric double layers at the interface and their role in sensing. (a) (Reproduced with permission from ref. [354]. Copyright 2023, Elsevier). (b-d) (Reproduced with permission from ref. [355]. Copyright 2021, Elsevier.).

Wearable TPSs primarily detect low-frequency physical and physiological signals from the human body. These sensors are effective when pressure is applied perpendicular to the contact surface, making vertical contact-separation and single-electrode modes the most suitable for wearable applications. These modes are simple in structure and manufacturing, and they align well with the typical pressures experienced by wearable devices on the human body. During operation, the device undergoes contact and separation processes under applied force, generating different electrical signals corresponding to the pressure, thereby achieving pressure sensing.

In TPSs, two friction materials, known as friction electrode pairs, generate triboelectric charges. Induction electrodes, connected to an external circuit, facilitate electron transfer. When friction materials with different electron affinity come into contact, they acquire opposite charges, creating a potential difference. During contact and separation, the varying potential difference induces electrostatic charges to rearrange within the induction electrodes, causing charge transfer in the external circuit. This process repeatedly generates electrical signals (open-circuit voltage, short-circuit charge, and short-circuit current).

In the vertical contact-separation mode, two induction electrodes are attached to the back of the friction electrodes. These electrodes form an external load path, and cyclic charge transfer occurs during the contact-separation process [356] (Fig. 8b). This mode is effective in sensing vertical pressures caused by joint flexion, footfalls, pulse, and heartbeat. The vertical contact-separation mode TPSs exhibit high sensitivity, a wide pressure range, and good stability due to the dual friction layers and induction electrodes. The performance of these sensors can be enhanced by combining multiple materials and microstructures.

The single-electrode mode operates similarly to the vertical contact-separation mode but involves a single induction electrode connected to a ground electrode [356] (Fig. 8b). This setup forms a path for reciprocating charge transfer between the ground and the induction electrode during the sensing process. Single-electrode mode TPSs are typically used with ionic conductive polymers like ion gels, which operate on the triboelectrification and double electric layer mechanism (Fig. 8c). The charges in the friction electrode induce opposite charged ions in the ion gel, facilitating electron conduction through separation and contact processes.

In single-electrode mode TPSs, a friction layer and an induction electrode are attached to skin surfaces or wearable devices. These sensors detect pressures from free-moving parts like fingertips or joint regions. Although single-electrode mode TPSs may have lower output performance compared to vertical contact-separation mode, they are simpler to fabricate and suitable for more flexible sensing applications. The performance of single-electrode mode TPSs can be improved by designing the materials and microstructures of the friction layer and induction electrode.

The material composition of TPSs includes friction electrode pairs, induction electrodes, and often substrates or encapsulation layers [355, 357]. For wearable applications, materials should be flexible to ensure comfort and durability. The well-suited materials, including flexible polymers, natural fibers, metallic nanomaterials, and carbon-based conductive materials generate significant electrical signals under small pressure changes and ensure long-term stability. Friction electrodes are critical, as almost all materials generate charge transfer when rubbed together. The triboelectric polarity of materials, determined by their electron gain or loss properties, affects their performance as friction electrodes [358,359]. Factors such as surface state, temperature, humidity, and environmental conditions also influence the triboelectric properties. A standardized method using liquid metal mercury has been reported to measure triboelectric charges accurately, providing a quantitative triboelectric material sequence that aids in selecting appropriate materials for friction electrodes [360–362]. It is important to note that, according to the electric double layer theory, the sensing capability of TPS is negatively impacted in environments with heavy perspiration. This occurs because the sweat electrolyte solution impedes the electron transfer between electrodes.

7.1.2. 2D Materials-based strain sensing applications

The incorporation of 2D materials into strain sensors has led to significant advancements in wearable technology for health monitoring. These sensors leverage the unique properties of 2D materials to achieve high sensitivity, flexibility, and multifunctionality, making them well-suited for various biosensing applications. The Ti₃C₂@PEDOT/graphene composite sensor is a multifunctional wearable strain sensor capable of monitoring strain, temperature, and electrocardiogram signals. This composite exhibits a superior temperature coefficient of resistance (0.86%), high strain sensitivity, and lower skin contact impedance, making it an excellent candidate for monitoring various biosignals. This sensor has been successfully used to wirelessly monitor subtle and large human deformations such as the vibration of the Adam’s apple, arterial pulse, and large body joint bending motions [363] (Fig. 9a–c). A remote health monitoring system integrates a FePS₃/rGO strain sensor and a flexible supercapacitor for real-time breathing rate and body temperature monitoring. This all-in-one textile-based monitoring band, wrapped around the abdomen, continuously tracks a person’s breathing patterns and transmits data wirelessly to a mobile device.

Fig. 9.

Composite-based Strain Sensors for Multifunctional Monitoring. (a-c) Ti₃C₂@PEDOT/graphene composite-based wireless sensors for strain, temperature, and heartbeat monitoring. (d) Remote health monitoring system with integrated FePS₃/rGO strain sensor and flexible supercapacitor for real-time breathing rate and body temperature monitoring. (e) Schematic of crumpled Ti₃C₂ film and a self-powered system transmitting bending and pressing signals to a smartphone. (f,g) Polyaniline@Ti₃C₂-based flexible strain sensor with stacked structure for wireless detection of artery pulse and phonation. (h) Sound wave detection by multilayer e-skin strain sensors, with sound waveform and spectrograms using short-time Fourier transform. (i) Breathing signals recorded during normal, rapid, deep, and shallow breathing cycles. (j) Strain sensor attached to a human throat for phonation detection. (k) Different tone hums detected by the artificial throat are converted into high-volume 10 kHz, low-volume 10 kHz, and low-volume 5 kHz sounds. (a-c) (Reprinted with permission from ref. [362]. Copyright 2022, Springer Nature). (d) (Reproduced with permission from ref. [363] Copyright 2022, Springer Nature). (e) (Reprinted with permission from ref. [364]. Copyright 2022, Elsevier; and ref. [664]. Copyright 2021, Wiley). (f-g) (Reproduced with permission from ref. [365] Copyright 2020, Elsevier). (h) (Reproduced with permission from ref. [367] Copyright 2018, American Chemical Society). (i) (Reproduced with permission from ref. [363] Copyright 2022, Springer Nature). (j) (Reproduced with permission from ref. [686] Copyright 2016, American Chemical Society). (k) (Reproduced with permission from ref. [687] Copyright 2017, Springer Nature).

Additionally, a smart portable temperature sensor patch, integrated with the Ti₃C₂/FePS₃-based portable power source, can track body temperature in real-time when attached to the armpit [364] (Fig. 9d). A self-powered wearable strain and pressure sensor system based on a triboelectric nanogenerator (TENG) utilizes a crumpled Ti₃C₂ film on a stretchy balloon. This configuration enhances mechanical strength and interfacial adhesion on human skin, demonstrating excellent tensile properties and sensitivity. The Ti₃C₂-based sensor serves as the sensing unit, identifying strain or pressure signals and converting them into electrical signals, which are then transmitted wirelessly to a smartphone via Bluetooth [365] (Fig. 9e). A flexible strain sensor based on a polyaniline@Ti₃C₂ composite with a stacked structure shows high sensitivity and an ultralow detection limit. This sensor can wirelessly monitor human activities, including bending fingers and elbow joints, detecting sound, and monitoring the human wrist pulse. The unique tile-like stacking structure of the composite provides additional conductive pathways, maintaining the continuity of the conductive path and allowing for high sensitivity over a broad strain range [366] (Fig. 9f,g). Multilayer e-skin strain sensors with an interlocked hemisphere micro-structure are used for detecting sound waves [367]. These sensors exhibit excellent linearity and high sensitivity, making them suitable for applications such as phonation detection. The multilayered e-skin can precisely detect time-dependent variations of acoustic sounds, accurately capturing sound waveforms and spectrograms using short-time Fourier transform [368] (Fig. 9h). A wearable strain sensor based on FePS₃ and a portable energy power storage device continuously tracks breathing patterns. The strain sensor records breathing signals during different breathing cycles, such as normal, rapid, deep, and shallow breathing, and transmits the data wirelessly to a mobile device [364] (Fig. 9i). The same kind of strain sensor can also be attached to the human throat can detect phonation, showing repeatable feature peaks when speaking specific words. (Fig. 9j, k).

A Ti₃C₂-based pressure sensor mounted on the human wrist can detect radial artery pulse signals with regular waveforms. This sensor has been integrated into an artificial e-skin for wireless human-machine interface sensing. Additionally, a breathable and highly sensitive Ti₃C₂@protein-based wearable pressure sensor on silk fabric offers high surface area, flexibility, and excellent sensing performance. It can monitor human physical activity, and psychological signals, and act as artificial skin for real-time wireless biomonitoring [369] (Fig. 10a). A flexible and biocompatible smart contact lens with a graphene-based FET transducer and a hybrid Ag nanofiber and nanowire antenna has been developed for accurate cortisol concentration detection in human tears. A 3D graphene-based touch sensor with a capacitive sensing mechanism is shown on the palm, detecting the approach of a finger. This device features an ultrathin PET substrate, transparent electrode arrays, and a dielectric layer, designed to maximize the sensitivity of the capacitive touch sensor. It can clearly detect multitouch points and recognize conductive objects, making it suitable for large-area applications, including robotics and 3D gesture sensing systems [370] (Fig. 10b). A novel wearable pressure sensor based on Ti₃C₂-coated tissue paper has been developed for monitoring human physiological signals. This sensor can measure blood pressure, radial artery pulse, phonation, and different motion states, such as jumping, standing, and walking. Its wireless array is capable of monitoring respiration and detecting opioid overdoses, demonstrating its potential in healthcare applications [371] (Fig. 10c,d). Graphene electronic tattoos (GET) offer conformal contact with the skin without adhesives, making them well-suited for long-term wearability and real-time monitoring. These ultrathin and flexible sensors can measure various physiological signals during hand exercises and skin temperature changes using a graphene resistance temperature detector (GRTD). The GET sensors maintain low impedance and high durability, even during athletic activities, and can be easily removed after use [372] (Fig. 10e–h).

Fig. 10.

Strain Sensing in Wearable Electronics and Smart Devices. (a) Ti₃C₂-based pressure sensor for wireless wrist pulse monitoring. (b) Optical image of a 3D sensor on the palm with an approaching finger, and relative capacitance change detected. (c,d) Smart face mask with integrated Ti₃C₂ pressure sensor for wireless detection of opiate overdoses and monitoring respiration. (e-h) Photos of graphene electronic tattoo (GET) sensors under skin deformation, sensing during hand exercise, and skin temperature recording using GET and a thermocouple with an ice bag near the graphene resistance temperature detector (GRTD). (a) (Reproduced with permission from ref. [688]. Copyright 2019, American Chemical Society). (b) (Reproduced with permission from ref. [369]. Copyright 2017, American Chemical Society). (c,d) (Reproduced with permission from ref. [370]. Copyright 2021, American Chemical Society). (e-h) (Reproduced with permission from ref. [371]. Copyright 2017, American Chemical Society).

7.2. Electromyogram (EMG) sensing

EMG measures the electrical activity generated by skeletal muscles, providing crucial insights into the musculoskeletal and neurological systems. EMG signals are widely used in healthcare for detecting neurodegenerative diseases affecting motor functions, such as Parkinson’s disease and stroke. They are also instrumental in tracking rehabilitation progress in patients recovering from injuries or diseases. Recent advancements in biomedical technology have expanded the application areas for EMG, including using EMG signals as control inputs for exoskeleton systems [373,374], assisting patients in completing tasks that their impaired musculoskeletal systems would otherwise hinder [375,376].

The primary attributes desired in EMG signals are high signal-to-noise ratio (SNR) and good repeatability, both of which are challenging to achieve. EMG signal amplitude and frequency characteristics are susceptible to various internal and external factors, such as the number of detected muscle fibers, their location relative to electrodes, and the impedance between electrodes and fibers. External factors include changes in electrode impedance, shifts in electrode position, and artifacts, all contributing to signal instability. Insufficient EMG signal quality directly influences interpretation accuracy, necessitating improvements in traditional EMG technologies to ensure precise signal detection in disease analysis.

Flexible electrodes made from 2D materials offer high flexibility and are free from skin irritation, yielding improved EMG signal quality. These electrodes can conformally contact the skin, with many thin-film electrodes achieving thicknesses below the critical value of 25 μm, thus attaining high SNR. Flexible electronics, including wearables and epidermal devices, have become increasingly attractive in health monitoring and human-machine interfaces. Flexible electrodes are fundamental components of these systems, designed to monitor EMG measurements among other physiological signals. EMG signals, characterized by a broader bandwidth and amplitude range, require electrodes with high sensitivity, precision, and resolution to accurately capture small changes in physiological electrical activity.

Different from intramuscular EMG (iEMG) electrodes, which require implantation inside the muscle, surface EMG (sEMG) electrodes are used on the skin’s surface and offer a broad view of overall muscle activity. sEMG electrodes are preferred for their noninvasive, low-cost properties, making them suitable for biofeedback, ergonomics, and space medicine applications [377–380]. The latest advancements in flexible noninvasive electrodes (FNEs) for sEMG acquisition focus on different materials, structures, and properties, emphasizing adhesiveness, breathability, flexibility, long-term durability, biocompatibility, biodegradability, and high SNR. Passive and active FNEs are explored, highlighting various thin-film transistor (TFT) materials and amplifiers used in active FNEs [377]. By leveraging the unique properties of 2D materials, flexible sEMG electrodes have become critical components in the realm of wearable health monitoring and human-machine interaction systems.

7.2.1. EMG mechanisms

The average human muscle comprises approximately 20–50 motor units, consisting of muscle cells and motor neurons [381,382]. During activation, the density of internal and external ions (K⁺, Na⁺, Cl⁻, etc.), in muscle cells changes, resulting in fluctuations in the motor unit action potentials (MUAPs) of cell membranes [383]. The activation cycle of a muscle cell involves three sequential steps: resting potential, depolarization, and repolarization [384] (Fig. 11a).

Fig. 11.

Electromyography (EMG) Mechanisms.(a) Mechanism of muscle fiber activation and schematics of intramuscular EMG (iEMG) and surface EMG (sEMG) electrodes. (b) Bipolar EMG signal acquisition configuration. (c) Mechanism of dual biological channels. (a,b) (Reproduced with permission from ref. [383]. Copyright 2022, Wiley). (c) (Reproduced with permission from ref. [387]. Copyright 2022, Cell Press).

In the resting potential phase, the distribution of ions on both sides of the cell membrane is maintained by ion pumps, with K⁺ ions more concentrated inside the membrane and Na⁺ and Cl⁻ ions more concentrated outside. The typical potential difference across the cell membrane ranges from 80 to 90 mV [385]. During depolarization, neurotransmitters released from the nerve ending disrupt the ion balance, causing Na⁺ ions to flow in and K⁺ ions to flow out, changing the potential difference by approximately 100 mV [386]. After depolarization, ion pumps restore the muscle cell to its original state (resting potential) by reversing the ion exchange, returning the potential difference to the resting level. This cycle generates overlapping MUAPs that create EMG signals (10 to 480 Hz) [387]. By analyzing different characteristics of the EMG signal, various aspects of muscle function can be understood.

Bipolar measurements utilize two electrodes aligned with the muscle fiber direction to detect and amplify the action potentials’ difference on the same motor unit, relative to a reference electrode [384] (Fig. 11b). This differential amplifier structure significantly eliminates common-mode noise, such as ambient and power-line noise, resulting in a high SNR. However, the amplitude and frequency content of the detected sEMG signal are sensitive to electrode alignment errors. Additionally, large electrodes (8–10 mm diameter) are required for adequate sEMG amplitude, resulting in low spatial resolution and reduced detection depth, which attenuates signals from deep muscles. To overcome the limitations of single-signal methods and achieve high accuracy with fewer sensors, dual biological channels are used. When an expression forms in the brain, an action potential in the motor cortex transmits through neurons to muscle fibers, stimulating contraction and ion channel activity [388] (Fig. 11c). These ion currents, converted to electron currents by surface electrodes, form EMG signals reflecting muscle contraction. Additionally, muscle contraction deforms the skin surface, which can be sensed by mechanical sensors.

7.2.2. 2D materials-based EMG devices fabrication

The dual-channel EMG patch integrates both EMG and mechanical sensors to enhance the accuracy of speech recognition in noisy environments [389]. This bioinspired system uses laser-induced graphene (LIG) to capture EMG signals and mechanical deformation simultaneously. The patch consists of six layers: a hydrogel layer to reduce skin-electrode contact impedance, a LIG-based EMG electrode, a PI separation layer to prevent signal cross-talk, a LIG-based mechanical sensor, another PI substrate layer, and a thin-film polyurethane (PU) tape for encapsulation [388] (Fig. 12a). Each patch contains a pair of EMG electrodes and two mechanical sensors to monitor one channel of EMG signals and two channels of strain signals. This design improves durability and maintains performance after 10 million fatigue tests. Patches applied to the chin and throat collect speech information without tracing the source of EMG signals, instead focusing on the relevant features. A small planar dimension combined with an island-bridge structure design achieves large stretchability, matching the anatomical structures of muscles and joints. The sEMG electrodes and strain sensors are connected by stretchable serpentine interconnects arranged in a symmetrical T-shaped configuration [390] (Fig. 12b). The sensor system is connected to a flexible anisotropic conductive film (ACF) cable and integrated with stretchable medical adhesive tape to create a soft, breathable sensor patch. This ES-Patch can be directly mounted onto the forearm, maintaining intimate contact with the skin even under mechanical compression, providing reliable multimodal epidermal data acquisitions.

Fig. 12.

2D Materials-based EMG devices Fabrication. (a) Schematic of the dual-channel EMG patch, which contains two EMG sensors placed on the chin and throat to collect bio-signals. (b) Optical image of the epidermal soft sensor connected to a flexible anisotropic conductive film (ACF) cable and adhered to stretchable, transparent medical tape. (c) Size effects on conducting polymer particles. (d) Fabrication of the conductive nanocomposite electrodes. (a) (Reproduced with permission from ref. [387]. Copyright 2022, Cell Press). (b) (Reproduced with permission from ref. [389]. Copyright 2021, Wiley). (c) (Reproduced with permission from ref. [390]. Copyright 2020, Science). (d) (Reproduced with permission from ref. [391]. Copyright 2022, Springer Nature).

Various shapes and sizes of sEMG electrodes significantly impact signal frequency characteristics. Round-shaped electrodes are commonly used due to their ease of fabrication and consistent contact with the skin. These electrodes can be designed as single units or arrays. Smaller electrode sizes contribute to higher EMG image resolution by providing more detailed spatial information about muscle activity, although they can complicate the readout circuitry due to the increased number of data points that need to be processed.

The size of the electrode affects several key factors. Larger electrodes cover more skin surface, aiding in the detection of signals from a broader area but potentially reducing specificity to individual muscle fibers. They tend to average signals over a larger area, smoothing out variations in muscle activity but also diminishing resolution. Additionally, the size influences the amount of signal power detected; larger electrodes capture more overall signal power but can be more prone to noise. Conversely, smaller electrodes enhance selectivity, providing better resolution of signals from individual muscle fibers but are more sensitive to exact placement and alignment. For instance, Ag/AgCl electrodes with varying diameters are used depending on the application site. Larger electrodes (8–10 mm diameter) are often used for limbs and trunk muscles to ensure adequate signal capture. In contrast, electrodes with a diameter of less than 5 mm are preferred for facial muscles or other smaller and more precise applications. Smaller electrodes improve signal quality by minimizing interference and capturing more specific muscle activity.

Fig. 12c illustrates the relationship between particle size (α) and mean free path (λ) in the context of designing electrode materials [391], which aids in understanding how different electrode configurations and material properties can affect the performance of sEMG electrodes. Small particles (α < λ), represented in the lower left quadrant, are typically used in applications requiring high spatial resolution and detailed signal capture. Intermediate particles (d₁ < α, λ < d₂), shown in the central quadrants, balance high resolution and adequate signal strength, making them suitable for general-purpose sEMG applications. Large particles (d₂ < α, λ), depicted in the upper right quadrant, are used when a broad detection area and high signal power are more critical than spatial resolution. Understanding these relationships helps in designing electrodes that optimize the balance between sensitivity, resolution, and robustness, which is essential for reliable and accurate sEMG signal acquisition.

Typical gel-based electrodes face challenges such as skin irritation, dehydration over time, and inconsistent conductivity due to gel drying. Dry electrodes contact directly with the skin without electrolytic gel, suitable for long-term detections. However, they may have high skin-electrode impedance and motion artifacts [55]. Various materials are used, including Ag/AgCl, AgCl, Ag, and Au. Advanced materials and fabrication methods have been developed to overcome typical wet electrode disadvantages. Metallic nanomaterials-based electrodes offer excellent conductivity and SNR. For instance, PEDOT coated on Ag nanomesh enhances SNR and reduces impedance. Carbon material electrodes provide reliable long-term wearability. Novel fabrication methods, such as scalable production of rGO electrodes, offer improved spatial resolution and SNR. Scalable production refers to techniques that allow for the mass production of these materials while maintaining their quality and performance. For example, rGO can be produced using chemical reduction of GO, which is a cost-effective and efficient process that ensures high conductivity and flexibility of the resulting electrodes. This scalability is crucial for practical applications where large quantities of high-performance electrodes are needed.

Conductive nanocomposite electrodes, which combine conductive materials like graphene, CNTs, or metallic nanowires with flexible polymers, offer superior performance. These nanocomposites provide higher spatial resolution and SNR due to their enhanced electrical properties and better conformability to the skin’s surface, ensuring consistent contact and reduced motion artifacts [55]. Fig. 12d illustrates a detailed fabrication method for conductive nanocomposite electrodes [392]. The process involves creating microchannels in a PDMS substrate, filling them with a solution of conductive materials such as PEDOT, and then drying. This method results in a patterned, highly conductive, and flexible electrode that can be transferred onto a desired substrate for application. The final electrode structure offers a high level of mechanical flexibility and electrical performance, making it suitable for high-resolution and high-SNR EMG signal acquisition.

7.2.3. 2D materials-based EMG applications

EMG applications using 2D materials have advanced significantly, enabling high-quality, flexible, and stretchable sensors that provide accurate monitoring of muscle activity for various applications, including healthcare, human-machine interfaces, and speech recognition. The integration of advanced materials and fabrication techniques has enhanced the performance of EMG sensors, making them more effective and versatile.

The dual-channel EMG patch is placed on the chin and throat to collect bio-signals. This patch captures both EMG signals and mechanical signals, providing comprehensive data for accurate speech recognition [388] (Fig. 13a). The setup involves careful skin preparation and precise data acquisition to ensure high-quality signal collection. The frequency spectrum of dual-channel EMG signals illustrates that EMG signals (yellow) are concentrated in the high-frequency component, while mechanical signals (red) are in the low-frequency component [388] (Fig. 13b). This distinction is crucial for accurate signal processing and analysis. Both EMG electrodes and mechanical sensors are fabricated using laser scribing technology on a PI substrate, ensuring precise patterning and consistent performance.

Fig. 13.

2D Materials-based Wearable Surface EMG applications. (a) Illustration of patches on chin and throat, outputting two channels of EMG and four channels of mechanical signals. (b) Frequency spectrum of dual-channel EMG: EMG signals (yellow) and mechanical signals (red). (c) Hand gestures prediction system including an analog-to-digital converter (ADC) and bracelets with sEMG and force myography (FMG) sensors. (d) EMG and mechanical signals for different digits. (e,f) Assessment of muscle activity and motion using a stretchable, multifunctional EMG sensor patch. (g) SEM images and finite element analysis (FEA) of skin-interfaced sEMG electrodes under strain. (h) Optical image of gel-based and stretchable sEMG electrodes on the forearm. (i) Recorded sEMG signals from gel-based (G-electrodes) and stretchable (S-electrodes) electrodes. (j-l) Comparison of G-electrodes and S-electrodes sEMG signals: frequency spectrum, mean power frequency (MPF), and mean absolute value (MAV). (m) Electrode-skin impedance of commercial and graphene-based EMG electrodes on chin and throat. (n) EMG signals from chin (black) and throat (red). (o) Relative resistance changes of mechanical sensors with increasing sound intensity from a speaker. (a,b) (Reproduced with permission from ref. [387]. Copyright 2022, Cell Press). (c) (Reproduced with permission from ref. [392]. Copyright 2020, IOP Publishing). (d) (Reproduced with permission from ref. [387]. Copyright 2022, Cell Press). (e-l) (Reproduced with permission from ref. [389]. Copyright 2021, Wiley). (m-o) (Reproduced with permission from ref. [387]. Copyright 2022, Cell Press).

A hand gesture prediction system, including an analog-to-digital converter (ADC) and bracelets equipped with sEMG and force myography (FMG) sensors integrates multiple sensors to enhance the accuracy of gesture recognition [393] (Fig. 13c), utilizing advanced data processing techniques to interpret the collected signals. The EMG and mechanical signals for different digits are shown in Fig. 13d [388]. Each signal channel exhibits unique characteristics, enabling the differentiation of various digits. Advanced signal processing techniques, such as t-distributed stochastic neighbor embedding (t-SNE), are applied to analyze these signals and improve classification accuracy.

A stretchable, multifunctional EMG sensor patch can be developed for assessing muscle activity and associated motion [390] (Fig. 13e,f). This sensor patch is highly flexible and can be adhered to the back of the hand, to capture muscle activities and joint motions. The combination of sEMG and strain sensors enhances recognition accuracy for applications like sign language interpretation. Scanning electron microscope (SEM) images and finite element analysis (FEA) of skin-interfaced sEMG electrodes under strain reveal that the microstructured serpentine metal traces are crucial for sensing accuracy, mechanical reliability, and comfort, demonstrating the structure’s durability and effectiveness under mechanical stress (Fig. 13 g). An optical image of gel-based (G-electrodes) and stretchable (S-electrodes) sEMG electrodes on the forearm highlights the flexibility and improved skin interface of the stretchable electrodes (Fig. 13 h), making them more suitable for long-term and dynamic applications. The recorded sEMG signals from both G-electrodes and S-electrodes during muscle activity demonstrate the high-quality signal acquisition capability of the stretchable electrodes (Fig. 13i), which perform comparably to traditional gel-based electrodes. The frequency spectrum, mean power frequency (MPF), and mean absolute value (MAV) of sEMG signals from G-electrodes and S-electrodes are compared (Fig. 13j–l). The stretchable electrodes exhibit excellent performance metrics, validating their effectiveness for reliable sEMG measurements in various applications. The electrode-skin impedance for both commercial and graphene-based EMG electrodes was applied to the chin and throat [388] (Fig. 13 m). The graphene-based electrodes demonstrate lower impedance, ensuring better signal quality and less noise interference. The EMG signals collected from the chin (black) and throat (red) when the tester silently speaks the word “graphene” (Fig. 13 n) reflect the muscle movements associated with speech, with the chin signal being stronger due to larger movement The relative resistance changes of mechanical sensors with varying sound intensities from a speaker reveal that sensors prepared with 45% laser power exhibit the best resistance change rate and sensitivity to mechanical changes (Fig. 13o), emphasizing the importance of optimizing fabrication parameters to enhance sensor performance.

7.2.4. Challenges and perspectives of 2D materials-based EMG applications

The challenges of 2D materials-based EMG applications focus on enhancing signal quality, wearability, and integration with other sensing technologies [394–398] (Fig. 14). Gelled electrodes, which use an electrolytic gel between the skin and electrode, provide low skin-electrode impedance and stability during muscle movements, ensuring high SNR [384]. However, they have limitations such as low water and air permeability, drying out over time, and the need for inconvenient skin preparation. In contrast, dry electrodes directly contact the skin, requiring fewer preparations and being more suitable for long-term detections. The main disadvantage of dry electrodes is high skin-electrode impedance and susceptibility to motion artifacts due to air pockets at the interaction point. To mitigate this, dry electrodes are often integrated with a pre-amplifier. Various materials, such as Ag–AgCl, Ag, and Au, are used in surface electrodes. Advanced materials and fabrication methods, including metal thin-film electrodes, metallic nanomaterials, and carbon material electrodes, show excellent electrical conductivity and SNR.

Fig. 14.

Challenges and perspectives of 2D Materials-based EMG Applications. (Reprinted with permission from ref. [383]. Copyright 2022, Wiley; reprinted with permission from ref. [394]. Copyright 2020, Elsevier; reprinted with permission from ref. [395]. Copyright 2019, Science; reprinted with permission from ref. [396]. Copyright 2007, Springer; reprinted with permission from ref. [397]. Copyright 2018, IEEE).

Multisensory approaches that integrate EMG with various physical and physiological sensors have been widely reported. These multisensory systems are used in diverse applications, including healthcare monitoring, human activity recognition (HAR) [395], and prosthetic hand control. For example, combining EMG with inertial sensors enhances HAR applications by providing intuitive information about motion in three dimensions, benefiting prosthetic hand control and allowing for smoother and more accurate control of assistive devices [396]. The integration of EMG with other physical sensors, such as force and moment sensors, provides significant improvements in fields like real-time control and gait parameters measurement. This combination stimulates advancements in rehabilitation devices, enabling smoother trajectory control and assisting with tasks such as reach-and-grasp activities.

Combining EMG with visual motion capture systems offers a noncontact and precise method to obtain spatial information about human movements, which is valuable in biomechanical analysis and the design of assistive devices for preventing work-related musculoskeletal disorders (WMSD). The fusion of EMG with physiological sensors, such as ECG, EEG, and electrodermal activity (EDA), provides comprehensive biological models for detecting physiological states like disease, stress, and emotion. For example, combining EMG with ECG can help detect muscle fatigue and stress, while integrating EMG with NIR monitors muscle activity in real-time during neuromotor control and rehabilitation exercises.

In rehabilitation, EMG combined with FMG and NIR signals can control prostheses and assistive devices more effectively. The integration of EMG with gaze detection provides additional control mechanisms for upper limb prostheses [394], enhancing the accuracy and functionality of the devices. Advanced materials and fabrication methods have led to the development of new electrode types with enhanced performance. Nanocomposite, metallic nanomaterial-based, and carbon material electrodes demonstrate higher SNR and lower skin-electrode interface impedance. Doping 2D materials enhances electron transfer, reduces impedance, and improves the bioelectronic-skin interface through increased surface roughness.

Studies have demonstrated the advantages of advanced electrodes over traditional gel-based electrodes regarding signal quality and wearability. For instance, carbon-elastomer based dry electrodes offer long-term wearability and reliable performance, while nanofiber carbon electrodes prepared by electrospinning exhibit low impedance and high durability. The impedance of electrodes is a critical factor in determining the quality of EMG signals. Advanced electrode 2D materials have been shown to provide lower skin-electrode impedance and higher SNR compared to traditional materials.

Recent advancements in artificial intelligence and data processing capabilities have enabled the combination of more than two sensing modes. This multisensory fusion strategy offers significant benefits in healthcare monitoring, traffic safety, and rehabilitation, providing comprehensive biological signals for accurate state detection and improved control of assistive devices. However, the instability of EMG signals and the need for real-time interpretation with high accuracy remain ongoing challenges.

7.3. ECG monitoring

Heart rate is a vital indicator of one’s physical status, including sleep state, infections, medications, heart conditions, and blood-related issues such as anemia [399]. Heart rate variability (HRV), which measures how well the heart responds to stress, is another important parameter, with a high HRV indicating a healthy heart. Cardiovascular diseases, such as ischemic heart disease and stroke, have been the leading causes of death worldwide [400–402]. This makes wearable heart rate sensors extremely important.

Most heart rate sensors use photoplethysmography (PPG) or ECG. PPG utilizes light to detect blood flow through arteries and is convenient but can be affected by ambient light, skin color, and distance from the artery [403]. ECG is more accurate as it directly measures the electrical activity of the heart using multiple electrodes, but it is less convenient due to the need for multiple electrodes. ECG is a critical tool for assessing cardiac activities, depicted through the P-QRS-T wave, providing a non-invasive method to detect various arrhythmias. Accurate and early detection of arrhythmias is crucial, as they can lead to sudden cardiac death. Traditional ECG monitoring requires bulky devices, typically used in hospital settings. However, the advent of 2D materials has the potential to revolutionize ECG monitoring, making it more accessible and convenient. 2D material with superior electrical conductivity and large contact area, is effective in ECG sensors. Ultrathin, single-layer graphene sheets can conform closely to the skin, reducing air gaps and improving sensitivity. This enhances the performance of TENGs, PENGs, and piezoresistive sensors in converting mechanical vibrations from pulse waves into electrical signals.

7.3.1. 2D materials based ECG monitoring mechanism

7.3.1.1. Triboelectric ECG sensors.

Heart rate can be measured at various locations on the human body where arteries are located, such as the wrist, neck, and chest. When the heart beats, it sends electrical signals allowing the heart to undergo two phases—systole and diastole [399]. During systole, the heart contracts, pumping blood throughout the body. During diastole, the heart relaxes, allowing blood to flow back into the heart. These blood flow dynamics are used to measure heart rate in ECG. Wearable bioelectronics offer a novel approach for continuous ECG monitoring, enabling individuals to track their health status in real-time. The integration of IoT and Bluetooth technology facilitates the wireless transmission of health data from bioelectronic devices to mobile phones, allowing extensive data processing by mobile workstations. Upon user permission, this health information can be shared with family doctors, reducing the need for hospital visits [404] (Fig. 15a).

Fig. 15.

Working Mechanism of 2D Materials in ECG Monitoring. (a) A cardiovascular monitoring system using a self-powered textile triboelectric sensor. (b) Schematic of electricity generation by a textile triboelectric sensor in response to the radial artery pulse. (c) Stages of atherosclerosis development: initial lesion (I) to rupture (VI). (d) Standard contour plot of the pulse wave in human arteries. UT, upstroke time. RWTT, left ventricular ejection time. PPT, systolic–diastolic time. Properties and fabrication of stretchable graphene nanopapers: (e) Large surface area; (f) increased hydrophobicity after heating; (g) excellent conformability. (h) 2D schematic of pulse wave to electricity conversion using the triboelectric nanogenerator (TENG) principle. (i) Micromechanism of 2D materials in piezoresistive sensors. (j) Graphene in a photoplethysmogram (PPG). (a-d) (Reproduced with permission from ref. [403]. Copyright 2021, Wiley). (e-g) (Reproduced with permission from ref. [409]. Copyright 2019, Elsevier). (h) (Reproduced with permission from ref. [669]. Copyright 2022, Cell Press). (i) (Reproduced with permission from ref. [418]. Copyright 2017, Springer Nature). (j) (Reproduced with permission from ref. [312]. Copyright 2019, Science).

The ECG reflects mechanical variations caused by blood movement in vessels beneath the skin. Healthy vasculature, with abundant blood perfusion, induces mechanical force variations detectable by ECG. Conversely, conditions like atherosclerosis narrow the vascular lumen, reducing blood flow and mechanical force. The ECG records the cardiac contraction and diastole processes, providing detailed heart health status information [404,405] (Fig. 15b,c). Various ECG metrics offer insights into heart function, such as the time required for the pulse wave upstroke to reach the systolic peak (UT), which reflects aortic stiffness and elasticity. Reflected wave transit time (RWTT) and pulse pressure transit time indicate arrhythmias, while left ventricular ejection time (LVET) measures cardiac systolic and diastolic capacities [406–408] (Fig. 15d).

TENGs can measure arterial pulse waves by utilizing the pressure changes associated with vasoconstriction and vasodilation [332,404]. The voltage plot generated from a wearable TENG heart rate sensor closely resembles the arterial pressure wave, containing characteristic P-wave, D-wave, and valley, demonstrating the feasibility and accuracy of using wearable TENGs as heart rate sensors. Graphene is widely applied in ECG sensors due to its superior electrical conductivity and large contact area. Ultrathin, single-layer graphene sheets can be fabricated by stripping bulk graphene, followed by heat treatment to remove water and increase hydrophobicity [409]. This enhances graphene’s conformability to human skin, reducing air gaps and improving the sensitivity of bioelectronics [410] (Fig. 15e–g). Graphene’s conductivity makes it suitable as electrodes in TENGs, converting mechanical vibrations from pulse waves into electrical signals. This involves the surface charging effect, where contact between two triboelectric materials induces electron transfer, generating an electron flow upon separation (Fig. 15 h).

7.3.1.2. Piezoelectric ECG sensors.

PENGs operate on the principle of the piezoelectric effect, which involves the production of electric charges in response to applied tiny mechanical stress [411,412]. When these materials are deformed, such as by the pulsating movements of the heart, they generate an electric field. This field can be harnessed to power ECG sensors or transmit cardiac signals. The high piezoelectric coefficients of materials like polyvinylidene fluoride (PVDF) make them highly efficient at converting mechanical energy into electricity [413–415].

The integration of PENGs into ECG monitoring devices to generate power from heartbeats eliminates the need for external batteries, reducing device size and weight and minimizing the risks associated with battery replacements. Meanwhile, the high sensitivity of PENGs allows for the detection of subtle mechanical changes in the cardiovascular system, providing high-fidelity capture of arterial pulse waveforms. This sensitivity is crucial for detecting early signs of arterial stiffness and other cardiovascular conditions that conventional ECG devices might miss.

The flexible nature of 2D materials ensures that PENGs conform to the dynamic movements of the heart without causing discomfort or damage. This property is crucial for long-term continuous monitoring. Furthermore, PENGs’ ability to operate without external power sources reduces artifacts and noise in the recorded data, ensuring clear and consistent readings for cardiovascular assessment. This capability is vital for applications that require precise and reliable monitoring of cardiac function, such as pulse wave velocity (PWV) and other parameters essential for diagnosing and managing cardiovascular diseases.

7.3.1.3. Piezoresistive sensors.

Blood pressure (BP) monitoring, an essential aspect of cardiovascular health, traditionally relies on methods like auscultative techniques with an inflatable cuff. However, these methods are unsuitable for continuous monitoring due to their intermittent nature, obtrusiveness, and high-power consumption. An alternative approach involves using piezoresistive sensors to monitor BP and ECG simultaneously, offering continuous and real-time data through methods like pulse transit time (PTT). PTT calculates BP from the time interval between the R-peak of the ECG and a characteristic point of the pulse wave signal, providing systolic and diastolic BP at each heartbeat. The piezoresistive effect occurs when a material’s electrical resistivity changes due to mechanical deformation, allowing the detection of tiny pressure changes associated with heartbeats for monitoring heart activity [416–418].

2D materials, with their wide interlayer distances and inherent flexibility, are well-suited for piezoresistive sensors. These sensors convert external pressure into resistance signals, demonstrating high sensitivity and adaptability. They can be directly attached to muscles or joints, showing strong potential in ECG detection [419] (Fig. 15i). The structure of a piezoresistive ECG sensor typically includes interdigitated electrodes made from a conductive material like Au, deposited on a flexible substrate such as PI. A piezoresistive layer, often composed of materials like carbon black particles embedded in a fabric matrix, is placed over these electrodes. The working of this sensors involves the variation in contact points among the carbon particles and at the carbon/Au interface during pressure changes. When pressure is applied, the number of contact points increases, reducing resistance; when pressure is released, the number of contact points decreases, increasing resistance. This variation in resistance due to pressure changes is precisely what makes piezoresistive sensors effective for ECG applications. Unlike traditional optical methods using PPG sensors, which require high operating power for real-time cardiovascular status assessment [313] (Fig. 15j), piezoresistive sensors operate with much lower power consumption. This makes them suitable for integration into wearable devices for long-term health monitoring.

7.3.2. 2D materials-based ECG devices

Graphene, with its excellent electrical conductivity, large surface area, and optical permeability, has been extensively utilized in triboelectric, piezoelectric sensors, and PPG devices [197] (Fig. 16a). Laser-induced graphene (LIG), obtained by exposing PI to laser radiation, offers a straightforward method for creating digital patterns. LIG is compatible with various pattern transfer techniques and has adjustable physical and chemical properties, enabling the development of diverse wearable sensors. Devices incorporating LIG can detect low SNR ECG signals, successfully diagnosing arrhythmias and identifying signal evolutions during the agonal state [419].

Fig. 16.

Graphene-based ECG Devices. (a) Schematic of thin, antibacterial, and biocompatible stretchable graphene–hydrogel for wearable bioelectronics. (b) Schematic of rGO in a polyvinylidene fluoride (PVDF) triboelectric layer with a leaf vein structure for a TENG. (c) A flexible and transparent graphene-integrated quantum dots ECG monitoring bracelet. (d) A graphene array applied to a human chest for standard ECG monitoring. (e) Current signal measured by the graphene array during ECG monitoring. (f) Schematic of the electrical output assessment of the TENG in rats. (g) Fluorescence images of α-actinin (green) and CX43 (red) proteins in porcine cardiac sections. (h) Schematic of electrical mapping from Langendorff-perfused hearts in negative control (Sham), positive control (myocardial infarction, MI), and treatment group (TRI-TENG). (a) Reproduced with permission from ref. [196]. Copyright 2024, Springer Nature). (b) (Reproduced with permission from ref. [419]. Copyright 2024, Springer Nature). (c) (Reproduced with permission from ref. [312]. Copyright 2019, Science). (d,e) (Reproduced with permission from ref. [420]. Copyright 2021, Springer Nature). (f-h) (Reproduced with permission from ref. [419]. Copyright 2024, Springer Nature).

Despite graphene’s advantages, mass production via CVD or direct stripping methods is challenging, and traditional graphene can be unstable and defect-prone. Functionalized graphene has been developed to address these limitations. rGO can replace metal electrodes, with polydopamine (PDA)-modified rGO providing long-term adhesion on biological surfaces and ensuring seamless contact. In TENGs, rGO incorporated into PVDF with a leaf vein structure enhances the device’s performance, detecting ECG signals and distinguishing abnormal heart functions, such as irregular rhythms and pulsation amplitudes [420] (Fig. 16b).

Wearable devices for PPG-based ECG monitoring face challenges due to the rigidity of traditional Si photodiodes, leading to inaccuracies and limited application areas on the body. Transparent and flexible graphene-based sensors provide key benefits including broadband wavelength sensitivity and high responsivity. These properties allow the design of flexible, transparent sensors that preserve the form factor of the active sensing area while placing bulky readout electronics away from the sensor. Graphene-integrated quantum dots ECG monitoring bracelets leverage these advantages, offering high performance and user comfort [313] (Fig. 16c).

Graphene’s properties make it a well-suited material for epidermal electronics and healthcare monitoring. GET have been used for monitoring various physiological activities, including ECG, EMG, EEG, skin temperature, and hydration. These GETs, made using a refined ‘cut-and-paste’ method with poly (methyl methacrylate) (PMMA) as a support polymer and temporary tattoo paper as a transient substrate, provide intimate skin contact and high optical transparency. When applied to the chest, graphene arrays can monitor ECG signals accurately, demonstrating strong potential for noninvasive health monitoring [421] (Fig. 16d,e). The current signal measured by graphene arrays during ECG monitoring shows high accuracy and reliability, benefiting from the material’s superior electrical properties. Multilayer graphene structures enhance electrical pathways and mitigate defects, resulting in stable and consistent ECG readings. This advancement in graphene technology paves the way for continuous and accurate monitoring of vital signs, such as heart rate and blood pressure, in a wearable format.

The TRI-TENG (trinity, 3 functions in 1 device) is primarily composed of an elastomer bottom package, an rGO electrode, a PVDF triboelectric layer with a leaf vein structure, Ecoflex 00–50 spacer, and a PDA-rGO electrode. Ecoflex 00–50, a biocompatible elastomer, serves as a spacer and package to enhance the triboelectric effect and prevent leakage. Its minimal degradation ensures the stability of the TENG functionality. The TRI-TENG, with a diameter of 8 mm, undergoes cyclic contact and separation with the heart’s contraction and relaxation, resulting in charges with opposite signs on the PDA-rGO electrode and the epicardium surface. The PDA-rGO patch functions as a conductive patch, triboelectric electrode for energy conversion, and therapeutic electrode for electrical stimuli application to the epicardium. Wireless sensing of cardiac conditions is facilitated by connecting the rGO electrode to a Bluetooth-enabled device, enabling communication with a smartphone application. The TRI-TENG can be assembled in series with an array design to fit large heart surface areas [420] (Fig. 16f–h).

An MXene-sponge was fabricated using a simple and efficient dipping-coating process, applied in piezoresistive sensors with insulating PVA nanowires (NWs) as spacers [422] (Fig. 17a). This sensor exhibits rapid response time, high sensitivity (a low detection limit of 9 Pa), and superior durability. These properties make it suitable for monitoring health activities and measuring pressure distribution. The MXene-sponge with a 3D network structure is produced via a facile dipping-coating process. A melamine sponge with over 97% porosity and pore size of 30–130 μm is used as a substrate, providing good compression performance and the ability to fully restore to its initial state after significant compression.

Fig. 17.

MXene-based ECG Devices. (a) Schematic of a MXene-sponge. (b-d) Schematic of pressure sensors with MXene-sponge. b, Initial state. c, Compressive state, fiber network becomes dense. d, Release state, fiber network returns to the original state. (e) Optical image of a MXene/cellulose nanofiber film folded into heart shape, demonstrating its deformability. (f-h) Measurement of water contact angle on MXene surfaces. (i-k) Schematic of sensing mechanism of MXene-embedded ZnO nanowire arrays. i, Initial state. j, Compressive state with tiny pressure, interlock structure forms between adjacent layers. k, Compressive state with high pressure, low wrinkles and area saturation forms. (l) Current signals of an adult female before and after exercises. (a-d) (Reproduced with permission from ref. [421]. Copyright 2018, Elsevier). (e-h) (Reproduced with permission from ref. [422]. Copyright 2021, Elsevier). (i-k) (Reproduced with permission from ref. [423]. Copyright 2024, Springer). (l) (Reproduced with permission from ref. [422]. Copyright 2021, Elsevier).

A clean melamine sponge is soaked in an MXene solution to absorb the 2D nanosheets. MXene’s large contact area and strong vdW forces ensure it is well-anchored onto the sponge. The color of the sponge changes from white to black after MXene coating. The MXene-sponge, due to its ultra-low mass density, can be supported by fragile structures without deformation. It also exhibits good mechanical strength under folding and twisting. The compress-release process shows that the MXene-sponge can recover its original shape under more than 90% compression. The 3D porous structure, low density, high elasticity, and strong mechanical strength of the MXene-sponge make it a powerful candidate for pressure-sensing materials.

The fabrication process of the MXene-sponge-based piezoresistive sensor involves several steps. First, interdigitated electrodes are deposited onto a thin PI film using an inkjet printing template sacrifice method. PVA NWs are then evenly deposited onto the PI substrate as spacers via electrospinning. A piece of MXene-sponge is placed on the PVA NWs network, and a thin thermoplastic polymer (PE) film is fixed and packaged to ensure a conformal contact between the MXene-sponge and the interdigitated electrodes [422] (Fig. 17b). The pressure-sensing mechanism of the MXene-sponge is dependent on variations in the contact area of the fiber network. When pressure is applied, MXene-coated microfibers contact each other, increasing the contact area and conductivity (Fig. 17c). Upon releasing the pressure, the microfibers return to their original state, reducing the contact area (Fig. 17d).

The MXene/CNF-foam is prepared using a vacuum filtration method followed by a hydrazine-induced foaming process. LiF/HCl is used to delaminate the Ti3AlC2 precursor, obtaining MXene nanosheets. CNF and MXene solutions are mixed using ultrasonic shaking, and a vacuum-filtration process creates a MXene/CNF-film. This film can be folded into a heart shape due to its excellent toughness [423] (Fig. 17e). Hydrazine vapor released during thermal treatment induces the formation of a cellular structure in the MXene/CNF-film, resulting in significant volume expansion and a rough surface. The CNF improves the mechanical properties of the foam, achieving a tensile strength of 68.9 MPa. The MXene/CNF-foam is placed on a flexible PI substrate with copper wire electrodes for pressure sensing, tightly encapsulated with an elastic PP film to reduce noise and ensure stability. Water contact angle measurements show that MXene/CNF-foam converts from hydrophilic to hydrophobic, enhancing its suitability for pressure sensing [423] (Fig. 17f–h).

In the initial state, the MXene-embedded ZnO nanowire arrays and the wrinkled membrane maintain their structure without any applied pressure [424,425]. Under tiny pressure, an interlocked structure forms between adjacent layers, increasing the contact area and enhancing conductivity [424]. Under high pressure, the wrinkles are compressed to their lowest state, maximizing the contact area and conductive pathways [425] (Fig. 17k). The pressure sensor effectively detects tiny pulse waves when placed on the wrist of an adult female, distinguishing between calm and post-exercise states, showing the difference in pulse rates [423,426] (Fig. 17l).

7.4. Wearable EEG devices

Brain–machine interfaces aim to bridge the gap between the human brain and external electronic systems, integrating biological electric pulses (biopotentials) with electronic devices [427]. EEG analysis, a pivotal non-invasive technique first reported by Hans Berger in 1929 [428,429], detects cortical electrical activity (brain waves) to interpret human intentions and translate them into commands for electronic robots and other devices. EEG sensors, being non-invasive, can be used for extended periods without requiring delicate surgical procedures. However, their signal collection is significantly challenged by the skull, which acts as a physical barrier, and the presence of hair and variable scalp conditions.

EEG sensors are typically classified into two types: wet and dry sensors. While wet electrodes often provide lower contact impedance with the scalp, they are less desirable for commercial applications due to their cumbersome deployment outside clinical environments [430,431]. Dry on-skin electrodes are more versatile and user-friendly, making them suitable for applications outside clinical settings. Dry EEG sensors can be made wearable, reusable, and easily integrated into caps or helmets, facilitating their use in various environments [432]. The successful implementation of dry electrodes hinges on achieving lower sensor/scalp impedance, prolonged reliability, and minimal skin reaction or discomfort. Traditionally, EEG electrodes are placed on the scalp, in the ear, and around the ear. Scalp EEG, which involves placing multiple electrodes along the scalp, is the most common method. The electrode numbers can vary based on the measurement type, ranging from high-resolution systems with numerous electrodes to simpler setups with just a few electrodes.

Despite its clinical importance, common scalp EEG systems can be obtrusive and bulky, limiting their use in non-clinical settings. The need for numerous electrodes, gels or pastes for electrode-skin contact, and long cables can make the setup complex and uncomfortable. To address these challenges, research has focused on developing less intrusive methods like in-ear and around-ear EEG [433–435], which offer advantages such as easier application, independence from hair type, and improved portability for long-term monitoring.

Recent advancements in conductive soft nanocomposites have shown promise in improving signal collection at the skull [436,437]. Among these materials, graphene stands out as an exceptional 2D material with superior biosensing capabilities for EEG applications. Flexible graphene sensors are gaining significant interest for continuous health monitoring and neural interfaces [438,439]. These sensors are typically developed from rGO on flexible polymer substrates like nylon, polydimethylsiloxane, or polyimide. Thin-film tattoo sensors made of CVD graphene on PMMA substrates have demonstrated encouraging performances for hands-free control of electronic devices.

Flexible graphene sensors are suitable for non-hairy areas and controlled or clinical environments. For prolonged applications in various environments, including all weather conditions, electrodes with higher mechanical stability are more appropriate. Few-layer epitaxial graphene deposited on SiC is a promising material for EEG electrodes that can operate reliably in external environments [440]. The combination of graphene’s excellent conductivity and SiC’s physical robustness, chemical inertness, hemocompatibility, and neural cell compatibility makes this material a strong candidate for high-reliability EEG sensors. The integration of graphene with SiC substrates brings additional advantages, as Si can be made flexible by thinning it down below 50 μm thickness. This flexibility is crucial for developing wearable sensors that can conform to the scalp’s shape and maintain good contact.

7.4.1. Wearable EEG electrodes

Scalp EEG is the most used method for monitoring brain activity and involves electrodes attached to the individual’s scalp. These electrodes are usually positioned following the international 10–20 system standard, where the distances between adjacent electrodes represent 10% or 20% of the total length between specific anatomical landmarks, such as the nasion and inion [441] (Fig. 18a). Scalp EEG electrodes can be classified as dry or wet, based on the presence of an electrolyte at their interface with the skin. The quality of the electrical contact at this interface is crucial and is generally evaluated by measuring contact impedance. Achieving low skin-electrode impedance is vital for capturing high-quality signals, and research on electrode materials and designs focuses on overcoming the highly resistive stratum corneum to reduce impedance effectively.

Fig. 18.

Common Wearable EEG Electrodes. (a) Scalp and ear EEG electrodes. (b) Wet electrodes (paste-based, hydrogel-based, sponge, and semi-dry) and dry/stiff electrode on the stratum corneum layer. (c) Tripolar concentric ring electrode arrays with self-adhesive graphene gel. (d) Schematic of a multilayer semi-dry electrode. (e) Ear-EEG positioning behind the ear. (f) In-the-ear EEG. (g) Screen-printed EEG electrodes attached to the skin with double-sided tape. (h) Silicone rubber generic earpiece. (a,b) (Reproduced with permission from ref. [440]. Copyright 2023, American Chemistry Society). (c) (Reproduced with permission from ref. [453]. Copyright 2022, IEEE). (d) (Reproduced with permission from ref. [454]. Copyright 2019, MDPI). (e) (Reproduced with permission from ref. [462]. Copyright 2022, Elsevier). (f) (Reproduced with permission from ref. [590]. Copyright 2012, IEEE). (g) (Reproduced with permission from ref. [464]. Copyright 2019, Front.; and ref. [707]. Copyright 2020, IEEE). (h) (Reproduced with permission from ref. [465]. Copyright 2013, IEEE).

7.4.1.1. Wet electrodes.

Wet electrodes are commonly used in EEG due to their ability to reduce contact impedance and enhance signal quality. Typically, these electrodes utilize ionic conductors in the form of pastes, gels, or electrolytic solutions, to reduce the contact impedance between the electrode and the skin, thereby improving signal quality [568] (Fig. 18b). Among these, Ag/AgCl electrodes paired with conductive pastes are considered the gold standard for EEG measurements due to their high SNR and low impedance. Integrating 2D materials into wet electrode designs holds significant promise for further improving EEG device performance.

Paste-based EEG electrodes use a thick paste containing ingredients such as aqueous electrolyte solutions, thickeners, humectants, and preservatives [441–443]. The paste hydrates the skin surface, enhancing the effective contact area and reducing impedance. However, typical paste-based electrodes have drawbacks such as the requirement for skin preparation, lengthy setup time, and signal degradation over time as the paste dries. Graphene can be integrated into these pastes to enhance their conductive properties due to its high electrical conductivity and large surface area.

Gel-based electrodes employ organogels or ionic hydrogels as conductive media. These gels are more elastic and easier to engineer or functionalize compared to pastes [443–445]. The addition of 2D materials can significantly enhance the performance of these gels. For example, graphene and other conductive nanomaterials can be dispersed within the gel matrix, resulting in stable impedance over extended periods [446–448]. This approach has already shown promise, with some gels maintaining low impedance for up to eight hours. Sponge electrodes, typically made of a conductive sponge saturated with a saline solution, are generally unsuitable for long-term operation due to drying out and increased impedance [449]. CNTs dispersed in a hydrophilic PU matrix have been developed to maintain low impedance over time [450, 451].

Semi-dry electrodes feature an electrolyte reservoir and a porous electrolyte-adsorbed material that enables gradual release of electrolyte onto the skin [452,453]. They combine the high conductivity of wet electrodes with the stability of dry electrodes, offering rapid setup and self-application without requiring skin preparation. Graphene-based gels can play a crucial role in these designs by serving as the electrolyte-adsorbed material. The high surface area and excellent retention properties of graphene enable a controlled and continuous release of the electrolyte, ensuring consistent low impedance contact. A ring concentric tripolar electrode array incorporating an adhesive graphene-based gel [454] (Fig. 18c) has been developed. This array accurately localizes and distinguishes between different sources on the head, allowing for recordings with high spatial resolution. Flexible semi-dry scalp EEG electrodes enabled hairy measurements [455] (Fig. 18d) are made from a conductive elastomer porous foam, which continuously releases an electrolytic solution to maintain stable and low contact impedance over extended periods.

7.4.1.2. Dry electrodes.

Dry electrodes for EEG offer several advantages over wet electrodes, including the elimination of gels or pastes, reduced skin irritation, and ease of use [456,457]. These electrodes rely on traces of sweat and environment moisture at the electrode-skin interface, eliminating the need for gels or pastes. This feature makes them less likely to change over time and minimizes irritation to the skin, although compensating for the absence of electrolytes to improve skin contact is necessary. Dry electrodes are easier for the general population to use since they don’t need a trained technician for application and can be applied directly to the skin without special preparation like skin abrasion or gel application. However, their loose attachment to the head is more susceptible to motion artifacts and can cause mechanical stress on the electrode structure during long-term use.

Dry EEG electrodes are categorized into non-contact (capacitive) and contact (resistive) types. Contact electrodes come in various shapes, with the most common being the disk electrode, which is usually attached to the scalp using cap or adhesive [458]. These disks, often made of Ag/AgCl, do not consider the presence of hair on the head, which poses a challenge for EEG recordings. This challenge is addressed by comb, fingered, or spike electrodes designed to reach the scalp more effectively. These electrodes, such as 3D-printed combs containing embedded carbon black for conductivity, eliminating the need for conductive gel or coatings despite having a higher skin-electrode impedance. Bristle electrodes have small brush-like bristles to evenly distribute pressure on the scalp, fitting the irregular surface and increasing the contact area [441]. These flexible polymer bristles, achieve electrode-skin impedances comparable to those of wet electrodes and are suitable for long-term use. Another approach to reducing large interface impedance from the stratum corneum is using microspikes or needles [432,459], lowering contact impedance and improving signal quality. Methods like vacuum casting and 3D printing are employed to fabricate these microneedle electrodes. Other types of dry electrodes include conductive foams, which conform to the uneven surface of the scalp, and temporary tattoos, which ensure conformal and barely noticeable contact with the skin through van der Waals forces. Foam-based electrodes, maintain low skin-electrode impedance, while inkjet-printed temporary tattoo electrodes offer electrical stability for extended periods [460,461].

Non-contact electrodes, also known as dry capacitive electrodes, cannot establish direct electrical contact with the skin [456,462]. These electrodes are integrated into textiles, quickly installed, and reused, which removes the risk of skin irritation. However, they are more susceptible to motion artifacts and require larger sizes, smaller gaps between the electrode and the skin, or preamplifiers to obtain performance similar to those of contact electrodes. Innovations like insulated electrodes with adhesive bottom layer design ensure close contact with the skin, enhancing capacitance and reducing impedance.

7.4.1.3. Ear EEG.

Around-ear EEG involves positioning electrodes near the ear to capture brain signals, offering a noninvasive method for brain monitoring [463] (Fig. 18e). The ear-EEG devices offer advantages such as minimal preparation and cleanup, and the absence of hair in the ear canal ensures consistent electrode placement. Personalized earpieces with electrodes on the outer surface are used to obtain EEG recordings [464] (Fig. 18f). These earpieces are molded to the individual’s ear, scanned, and manufactured using additive manufacturing techniques. The electrodes are typically Ag/AgCl, and a gel layer is applied to improve contact. Studies have shown that ear-EEG can produce signals comparable to scalp EEG, with slightly lower amplitude but similar signal-to-noise ratios.

Various around-ear electrodes have been developed, typically involving 2 or 3 electrodes affixed to the skin behind the ear using glue or tape [465] (Fig. 18g). These systems often use screen-printed electrodes on flexible substrates and have shown promise for seizure monitoring in epileptic patients. The feasibility of using these systems for continuous EEG monitoring has been demonstrated, and a behind-the-ear prototype using a generic geometrical model has been developed, eliminating the need for adhesives or headbands. Generic earpieces offer a more practical and cost-effective alternative to personalized earpieces for ear-EEG [466] (Fig. 18h). Generic earpieces use a cone-shaped earplug with biocompatible rubber. While they offer lower signal quality compared to personalized earpieces due to less optimal skin and electrode contact, they represent progress toward portable in-ear EEG devices.

7.4.1.4. On-skin EEG electrodes configurations.

On-skin EEG electrodes are emerging as a pivotal technology for continuous and non-invasive monitoring of brain activity, providing valuable insights into neurological and physiological states [467,468]. Temporary tattoo electrodes (TTEs) offer a promising platform for skin-worn devices due to their body-compliant nature and continuous skin contact. TTEs are developed using a layered structure composed of a sub-micrometer to micrometer-thick polymer decal film, which is releasable from its paper support by wetting and dissolving a sacrificial layer. This film can be laminated onto the skin, ensuring complete conformability and imperceptibility to the user. The design of TTEs involves a PU/allyl resin-based decal transfer film with a thickness of 1.5 μm, enhancing mechanical stability while maintaining unperceivability on the skin. The electrodes and interconnections are made of PEDOT, a conducting polymer known for its compatibility with skin mechanics and potential MEG compatibility. PEDOT is processed through inkjet printing, creating ultrathin layers that preserve mechanical conformability while reducing electrical resistance. The integration of TTEs with thin and flexible PEN sheets and a plastic clip enhances the robustness of the connection, ensuring reliable brain biopotential recordings [469] (Fig. 19a).

Fig. 19.

On-Skin EEG Electrodes Configurations. (a) Layered structure of an all-polymer printed temporary tattoo electrode (TTE) released on the scalp at the Oz position. (b) Non-contact electrodes. (c) Carbon and PDMS EEG electrodes with varied sizes and thickness. (d) Carbon and PDMS EEG electrodes structural configuration. (e) Graphene electronic tattoo (GET) attached to the skin. (f) On-skin rGO electrodes. (g) Gas-permeable, multifunctional stretchable on-skin EEG. (h,i) Epitaxial graphene sensors on a metal pin button with carbon tape. (j) On-skin electrochemical impedance spectroscopy (EIS) setup of epitaxial graphene (EG) sensors on the forearm. (k) Nyquist plot comparing epitaxial graphene-based EEG sensors with commercial foam-based and spring-loaded sensors. (a) (Reproduced with permission from ref. [468]. Copyright 2020, Springer Nature). (b) (Reproduced with permission from ref. [469]. Copyright 2020, IEEE). (c) (Reproduced with permission from ref. [470]. Copyright 2012, IEEE). (d) (Reproduced with permission from ref. [471]. Copyright 2014, Springer Nature). (e) (Reproduced with permission from ref. [371]. Copyright 2017, American Chemistry Society). (f) (Reproduced with permission from ref. [477]. Copyright 2020, Elsevier). (g) (Reproduced with permission from ref. [478]. Copyright 2018, Wiley). (h-k) (Reproduced with permission from ref. [439]. Copyright 2021, IOP Publishing).

Non-contact or dry capacitive electrodes do not require direct contact with the skin for electrical connectivity, making them suitable for integration into clothing and rapid installation [470] (Fig. 19b,c,d). These electrodes minimize the risk of skin irritation but are more susceptible to motion artifacts compared to wet or gel-based adhesive electrodes. Capacitive electrodes measure EEG signals by capacitance, where air, hair, and an insulating layer create high impedance between the scalp and the conductive layer. To mitigate this, larger electrodes, smaller gaps between the skin and the electrode, or preamplifiers are used. An enhanced non-contact electrode utilizes a PDMS/CNT as conductive adhesive to ensure close skin contact, achieving low impedance of ~120 kΩ at 0.1 Hz [471,472].

and conductive option for monitoring physiological signals. These electrodes leverage the unique properties of rGO, such as large contact areas and high conductivity, to reduce skin-to-electronic contact impedance and enhance flexibility. Advancements in printing methods have facilitated the development of wearable electronics that are comfortable, flexible, and biocompatible, making functionalized conductive graphene suitable for applications in wireless flexible circuits and muscle activity recording [473–477]. Large-scale rGO electrodes using simultaneous room-temperature reduction and patterning processes achieve low impedance and good adhesion with the skin [478] (Fig. 19e,f). Multifunctional gas-permeable on-skin EEG systems based on a sugar-templated porous substrate are designed to be comfortable for long-term wear while maintaining high functionality [479] (Fig. 19g). The electrochemical properties of epitaxial graphene (EG) electrodes were evaluated for on-skin applications using a three-electrode configuration [440] (Fig. 19h,i). The impedance spectroscopy setup on the forearm is depicted in Fig. 19j, with the Nyquist plot comparing EG-based EEG sensors with commercial foam-based and spring-loaded sensors [440] (Fig. 19k). This comparison highlights the performance of EG sensors in terms of impedance and signal quality.

7.4.2. On-skin 2D materials-based EEG electrodes fabrication

Bioinspired ultrathin epidermal soft electronics, or e-skin, have been developed for prosthetics, soft robotics, and health monitoring [480–486]. A notable example is the graphene-integrated ultrathin open-mesh structured e-tattoo sensor. This advanced sensor design allows for simultaneous monitoring of hydration, skin temperature, and biopotentials. The fabrication of this e-tattoo sensor involves a wet transfer process commonly used for polymer/graphene based thin-film sensors, followed by a dry patterning step directly on tattoo paper. This method ensures the sensor is both flexible and durable, capable of conforming to the skin’s surface without causing discomfort. These e-skin sensors function as dry, skin-attached electrodes for EEG monitoring [487] (Fig. 20a), demonstrating superior performance with lower contact impedance at frequencies below 1000 Hz and a higher signal-to-noise ratio compared to traditional commercial gel electrodes.

Fig. 20.

On-Skin 2D Materials-based EEG Electrodes Fabrication. (a) EEG e-skin preparation and attachment to the forehead. (b,c) Kapton-based Pt-transition metal dichalcogenide (Pt-TMD) tattoo design and fabrication, including Pt evaporation, transition metal atom conversion (TAC) conversion, CVD process, and final PtSe₂ and PtTe₂ tattoos. (d) Photographs of Pt-TMD tattoos with poly(methyl methacrylate) (PMMA) and Kapton supports with electrical contacts. (e) Schematic of dynamic changes at the Pt-TMD–skin interface over time. (f) Schematics of semi-embedded graphitized electrospun fiber/monolayer graphene (GFG) skin electrodes, including fabrication and reinforcement. (g) Schematic of laser-patterned 3D pillar MXene electrode arrays. (h) Photograph of electrode array geometries for EEG monitoring, and high-resolution CT scans for Pt electrode and MXene electrodes. (i) CV for 2-mm planar MXene and Pt electrodes at 50 mV s⁻¹. (j) Schematic of the double-layer stacked structure fabrication of transparent PEDOT:PSS/graphene. (k) Schematic of enhanced electrical conductivity by π-π interaction and charge delocalization between graphene and PEDOT:PSS. (a) (Reproduced with permission from ref. [486]. Copyright 2022, Wiley). (b-e) (Reproduced with permission from ref. [487]. Copyright 2021, American Chemistry Society). (f) (Reproduced with permission from ref. [489]. Copyright 2020, American Chemistry Society). (g-i) (Reproduced with permission from ref. [490]. Copyright 2021, Science). (j,k) (Reproduced with permission from ref. [491]. Copyright 2021, Springer Nature).

PtSe₂ and PtTe₂ are grown using a thermal-assisted conversion CVD process at moderately low temperatures (about 400 °C), which allows growth directly on select polymers like Kapton. Pt-TMD tattoos are grown directly over Kapton or on Al₂O₃/Si wafers and then transferred onto an ultrathin temporary tattoo form using a 200 nm thick PMMA layer for support. The CVD process involves Pt predeposition on the sample’s surface. For Kapton structures, a 6 nm thick Pt layer is used, resulting in ~25 ± 3 nm thick multilayer Pt-TMD with multiple out-of-plane growth planes. For Al₂O₃/Si growth, a 3 nm thick Pt layer results in ~15 ± 3 nm thick layered dichalcogenide. The final devices are shaped using a cutter-plotter tool and can be supported by either a 25 μm thick Kapton or a 200 nm thick PMMA [488] (Fig. 20b,c).

The tattoos are not entirely transparent due to the number of TMD layers. The optical transparency can be improved with thinner TMD layers. The sheet resistance and the optical transmittance at 550 nm of PtSe₂ and PtTe₂ layers vary with thickness. PtSe₂ films become more semiconducting as thickness decreases, with reduced sheet resistance. PtTe₂ films retain metallic properties. These tattoos can be easily placed on the skin for characterization and electrophysiology, with PMMA providing intimate skin contact and Kapton structures held in place by Tegaderm or kind removal silicone tape [489] (Fig. 20d).

The structural properties of thick Pt-TMDs result in uneven height distribution and nanoscale rough surfaces, leading to high-impedance conditions due to trapped air. Over time, tattoos establish a closer connection with the skin, reducing air gaps. The simplified Randles circuit model indicates that after ~900 seconds, the interface stabilizes, providing reliable impedance values. Tattoos maintain performance over time, with improved contact from long-term intimate skin contact [488] (Fig. 20e). Semi-embedded GFG electrodes are fabricated using electrospinning. These electrodes demonstrate robust performance under mechanical stress, high conductivity (~150 Ω sq⁻¹), and transparency (~80% transmittance). The fabrication process includes electrostatically spinning the polymer on chemical gas deposited graphene [490] (Fig. 20f).

MXene is used for fabricating high-density microarray sensors. These arrays, fabricated through laser patterning and integration with MXene ink, demonstrate excellent on-skin impedance and high SNR for EEG monitoring. The MXtrode arrays successfully acquire high-resolution scalp EEG with low skin impedance [491] (Fig. 20g). MXene electrodes are compared with commercial Pt electrodes through high-resolution CT scans. The MXene electrodes show no imaging artifacts in MRI scans, confirming their compatibility with clinical imaging techniques. The versatility and low cost of MXene electrodes enable diverse bioelectronic applications (Fig. 20h). Cyclic voltammograms (CV) measurements demonstrate the electronic performance of 2-mm planar MXene and Pt electrodes, highlighting the suitability of MXene for biopotential monitoring with superior electronic performance (Fig. 20i). Graphene synthesized via CVD on copper foil is transferred using PEDOT as the carrier layer. The addition of surfactants enhances wettability and electrical conductivity. The fabrication process includes spin-coating PEDOT on graphene/Cu foil, etching the Cu foil, and transferring the PTG thin film onto various substrates [492] (Fig. 20j). Interfacing PEDOT with graphene enhances electrical conductivity through molecular packing and electron coupling. The synergistic interaction results in high electrical conductivity and improved biopotential monitoring performance (Fig. 20k).

7.4.3. 2D materials-based EEG electrodes applications

EEG remains a cornerstone in brain monitoring techniques due to its early adoption and significant contributions to neuroscience and neuromedicine [493,494]. Beyond medical diagnostics, EEG is instrumental in brain–computer interfaces (BCI), biometrics, and close-loop therapy for schizophrenia, depression, and insomnia. Recent advancements have further extended its applications to emotion recognition, cognitive enhancement, and identity authentication. EEG operates by recording electrical impulses from the cerebral cortex using electrodes. These impulses, originated from neurons’ action potentials, are monitored as voltage fluctuations. The EEG recordings contain both brain activity signals and various artifacts, which should be distinguished for accurate interpretation. The frequency ranges of EEG signals—delta, theta, alpha, beta, and gamma waves—correspond to different mental states and cognitive processes.

2D materials-based EEG electrodes have shown significant promise. MXene electrodes, for example, offer excellent performance for human epidermal sensing applications. Scalp EEG using MXene electrodes, compared to standard gelled Ag/AgCl EEG electrodes, has demonstrated comparable signal quality with distinct advantages. An eight-channel MXtrode array, with mini-pillar electrodes arranged around a central opening for standard EEG electrodes, provides a practical solution for high-quality EEG recordings [491] (Fig. 21a–c). The impedance at the electrode-skin interface for dry MXtrodes is slightly higher than that of gelled electrodes, but the EEG recordings reveal clear alpha rhythms in the eyes-closed state, indicating reliable performance (Fig. 21c).

Fig. 21.

2D Materials-based EEG Electrodes Applications. (a,b) Photograph of MXene electrode array, and EEG signals in eyes-closed and eyes-open conditions. (c) Power spectral density of eyes-closed and eyes-open EEG recordings, highlighted with the 8–12 Hz alpha band. (d) Self-adapting gel forms a seamless interface with the scalp, avoiding gaps from hair and scalp folds. (e) Polarization potential characteristics of poly(acrylic acid) sodium salt (PAAS)-MXene. (f) PAAS-MXene pre-solution with water-like viscosity. (g) Nyquist plot comparing interface impedance of on-site gelled PAAS-MXene, off-site gelled PAAS-MXene, and commercial gel. (a-c) (Reproduced with permission from ref. [490]. Copyright 2021, Science). (d-g) (Reproduced with permission from ref. [494]. Copyright 2022, American Chemistry Society).

Further advancements in MXene technology include formulating MXenes into hydrogels, such as poly(acrylic acid) sodium salt (PAAS)-MXene (Fig. 21d), which improves stability and water retention. These hydrogels form a seamless interface with the scalp (Fig. 21e), reducing interfacial impedance and enhancing EEG signal acquisition. The PAAS-MXene pre-solution, with its water-like viscosity (Fig. 21 f) [495], ensures easy application and rapid gelation on the scalp, providing reliable adhesion and eliminating the need for an EEG cap. The interface impedance of on-site gelled PAAS-MXene is significantly lower than that of off-site gelled PAAS-MXene and commercial gels, confirming the effectiveness of this strategy for seamless adhesion and superior electrical characteristics (Fig. 21 g).

7.4.4. Challenges and perspectives on 2D materials-based EEG electrodes

The advancement of 2D materials has opened new frontiers in the development of wearable EEG devices, promising enhanced performance and user comfort. However, several challenges remain in realizing their full potential [496–500] (Fig. 22).

Fig. 22.

Challenges and Perspectives on 2D Materials-based EEG Electrodes. (Reproduced with permission from ref. [495]. Copyright 2018, MDPI; reproduced with permission from ref. [496]. Copyright 2023, MDPI; reproduced with permission from ref. [497]. Copyright 2017, MDPI; reproduced with permission from ref. [498]. Copyright 2022, Springer Nature; reproduced with permission from ref. [499]. Copyright 2018, MDPI).

7.4.4.1. High density.

High-density EEG systems, incorporating numerous channels for detailed signal capture, are essential for improving the spatial resolution of EEG recordings [496]. While a 256-channel setup significantly enhances the granularity of brain signal capture, it introduces challenges related to data management, including storage and transmission of large data volumes. Efficient data compression and transmission protocols are critical to ensure that high-density EEG data can be effectively utilized and shared.

7.4.4.2. Virtual channels.

In addition to physical electrode channels, advanced algorithms can create virtual channels to enhance EEG system functionality without physically increasing the number of electrodes. Virtual channels leverage computational methods to interpolate and predict EEG signals at locations without physical electrodes, reducing the physical burden on subjects and mitigating issues related to electrode placement and maintenance [497]. Implementing virtual channels requires sophisticated algorithms and robust validation to ensure data accuracy and reliability.

7.4.4.3. Spatial resolution.

Achieving high spatial resolution through precise electrode placement is another critical goal in EEG technology [498]. Careful positioning of electrodes according to standardized International 10–20 system, can optimize the spatial resolution of EEG recordings. Advancements in electrode materials and designs, including the use of 2D materials, have the potential to enhance spatial resolution further by allowing closer electrode spacing and improved signal quality. Ongoing research into optimal configurations and materials for electrode placement is necessary to maximize EEG system spatial resolution.

7.4.4.4. Electrode-tissue adhesion.

Ensuring effective adhesion of EEG electrodes to the scalp is crucial for maintaining signal quality and stability [499]. Developing multifunctional hydrogel adhesives offers promising solutions by combining strong adhesion with properties like electrical conductivity and biocompatibility. These hydrogels can provide stable electrode attachment during prolonged recordings, reducing the risk of signal artifacts caused by electrode movement. Innovative designs like Janus hydrogels with asymmetric adhesion properties can prevent tissue adhesion and secondary trauma, enhancing both comfort and reliability of EEG setups.

7.4.4.5. Modular design.

Modular designs for customizable and scalable EEG setups offer flexibility and ease of use. Modular EEG systems enable the configuration of electrode arrays tailored to specific research or clinical needs [500,501]. This approach facilitates the expansion or reduction of electrode numbers without compromising system integrity. Modular designs also support integrating various types of electrodes, including those based on advanced 2D materials, optimizing performance for different applications. Continued innovation in modular EEG designs will support the adaptability and scalability of EEG technology, making it more accessible and versatile.

7.4.4.6. Signal accuracy.

Significant progress in data characterization and interpretation has marked the evolution of EEG technology, yet substantial challenges persist in optimizing EEG hardware setup and functionality. Electrodes, the critical interface between the brain’s electrical activity and recording equipment, are central to these challenges. Advances in EEG electrodes are crucial for minimizing artifacts, improving signal quality, enhancing user comfort, and broadening the scope of applications.

Achieving high SNR and low contact impedance is essential for accurately detecting the microvolt amplitude signals generated by cortical activity. Ag/AgCl wet sensors currently set the benchmark for noninvasive on-skin sensors, exhibiting the lowest on-skin impedance of approximately 30 ± 5 kΩ at 100 Hz [440]. Replicating such low impedance with dry sensors remains a significant challenge. However, emerging e-skin sensors made from conductive polymers (PEDOT/PSS), CVD graphene, epitaxial graphene, and laser-induced porous graphene show promising competitive values close to those of wet sensors.

7.4.4.7. High-resolution monitoring through hair.

Most reported 2D material-based sensors are thin-film designs, making them ideal for contact on hairless the forehead and forearm [446,450,502]. However, controlling EEG-based robotics through eyeball movements or visual flickering, which requires oscillating neuron signals, is most effective with sensors positioned on the parietal lobes near the back and top of the head. These areas are typically covered with hair, presenting a significant challenge for signal acquisition.

The presence of hair complicates the contact between the electrode and the scalp, increasing impedance and reducing signal quality. To date, MXene-based 3D mini-pillar array sensors have shown promise in addressing this challenge, successfully acquiring EEG signals through hair with thickness of ~5 mm [446]. This innovative approach demonstrates potential, yet further development and optimization of 2D material electrodes are necessary to overcome the impedance and signal quality issues posed by hair-covered scalp regions. Advancing these technologies will be crucial for enhancing the accuracy and reliability of EEG signal acquisition in hair-covered areas, thereby expanding the applicability of 2D materials-based EEG electrodes.

7.4.4.8. Durability and comfort.

Prosthetics and other external robotics, as well as health monitoring systems, often require frequent use and continuous operation of EEG sensors. Metal-free graphene integrated into soft polymers has enabled the development of e-skin and e-textrode sensors capable of operating continuously for several days [503–505].

However, these e-skin sensors are generally designed for single-use applications and are not suitable for repeated use. The challenge of reusability has been partially addressed by e-textrode sensors fabricated from graphitized electrospun fiber and monolayer graphene, which can withstand washing and repeated use up to approximately ten times. A significant limitation for the repeated use of graphene sensors is the delamination of graphene from the polymer substrate during prolonged use. Epitaxial graphene, grown on Si/SiC, has demonstrated excellent stability over long-term and repeated use, owing to the strong adhesion between graphene and the underlying substrate [506]. This robust adhesion helps maintain the integrity of the sensor, ensuring consistent performance.

7.4.4.9. Moisture and sweat-proof stability.

The stratum corneum, which is the outermost epidermis layer, presents significant challenges for EEG electrodes due to contamination with microbes, oils, dirt, and sweat. Sweat contains physiological salts that can induce electrode degradation, compromising signal quality and device longevity. 2D materials-based EEG electrodes have shown promise in addressing these challenges due to their inherent properties.

GO-based sensors exhibit good stability in the presence of sweat, attributed to their hydrophilic surfaces. The formation of a thin water boundary layer increases hydrophilicity and enhances ion intercalation, thereby improving contact impedance with the skin. Epitaxial graphene on SiC substrates also demonstrates improved performance with repeated testing, thanks to small grains of graphene that enhance stability and hydrophilicity in the presence of sweat [507,508]. Similarly, MXene-based sensors with ordered surface functional groups benefit from their aqueous processability and surface hydrophilicity, maintaining stability in moist conditions [509–511].

To further enhance stability, introducing gas permeability or breathability through patterning in an open-mesh structure or on a nanoporous membrane can be beneficial. These design improvements help in managing moisture and sweat, ensuring consistent and reliable signal acquisition.

8. Electrochemical sensing

Electrochemical detection methods, such as amperometry, potentiometry, and voltammetry, are widely used in wearable biosensors owing to their superior selectivity, high sensitivity, fast response, and suitability for simple designs. Potentiometry measures ion concentration changes by detecting the potential difference between a sensing electrode and a reference electrode. Amperometry detects analytes by measuring the current generated by redox reactions at the working electrode. Voltammetry, which can measure multiple analytes simultaneously, applies a voltage scan between electrodes to induce redox reactions with target analytes and measures the resulting peak currents [512].

Various materials, including different forms of carbon, metals, and metal oxides, have been employed to create these interfaces. The introduction of nanostructured materials marked a significant advancement, as these materials integrate the optical, electronic, and catalytic properties of nanomaterials with the biorecognition capabilities of biological entities, resulting in improved performance. Traditionally, nanostructured materials have been deposited in monolayer configurations, limiting the surface area available for biomolecule immobilization [513]. Recent developments in layered and hierarchical nanostructures offer new opportunities to create multiscale structures with controlled functions and enhanced electrical, optical, and mechanical properties. The multifunctionality of 2D nanomaterials, including tunable structures and numerous active sites, is advantageous for detection and transduction, providing real-time, accurate, and stable monitoring of various biomarkers such as temperature, ions, and sweat (Table 3).

Table 3.

Electrochemical sensing applications of 2D material-based sensors.

Applications	Materials	Advantages	Disadvantages
Glucose Monitoring	Graphene, MoS2	Enables continuous monitoring without finger pricks	Sensitivity to environmental factors (e.g., sweat)
Cortisol Detection	PVA Hydrogel	Quick response to physiological changes	Limited to sweat analysis, may not reflect blood levels
Uric Acid Detection	Graphene	High sensitivity allows for early detection of conditions	Complexity in sensor fabrication, potential for high costs

8.1. Temperature sensing

Temperature sensors convert thermal variations into electrical signals for a wide range of applications, due to their simplicity, wide measurement range, stability, and high precision [514,515]. The evolution of temperature sensors can be traced through three major stages: traditional discrete sensors, analog integrated sensors, and intelligent sensors [516]. Traditional sensors, including resistance temperature detectors (RTDs), thermistors, and thermocouples, provide direct contact with the measured object, offering high accuracy [517]. However, their large size, low sensitivity, and batch preparation limitations pose challenges for modern needs in miniaturization and multidimensional array configurations.

Analog integrated temperature sensors emerged in the 1980s [518], integrating sensor and analog signal output functions into a single chip. These sensors feature small size, low cost, fast response, and micro power consumption, suitable for remote temperature measurements without requiring complex calibration processes. The advent of intelligent temperature sensors, or digital sensors, marked a significant advancement by incorporating microelectronic, computer, and automatic testing technologies. These sensors can output temperature data and control signals compatible with various microcontrollers, enabling sophisticated temperature measurement and control functions through software.

Recent advancements in micro-electro-mechanical systems (MEMS) technology have revolutionized sensor fabrication, combining semiconductor manufacturing processes and materials to produce highly integrated, small-scale devices [519,520]. MEMS sensors leverage 2D material films to detect external physical changes, offering advantages such as short response time, high resolution, and low power consumption. These devices are increasingly prevalent in precision control, intelligent vehicles, wearable devices, electronic medicine, and mobile photography, benefitting from their capability for batch fabrication and integration with neural network algorithms for real-time signal optimization.

2D materials-based temperature sensing performance has been explored, focusing on the temperature-sensitive mechanisms of graphene and its derivatives. Various structures and materials, including single-layer and multilayer graphene, graphene derivatives, composites, and novel structural sensors, have been examined. The performance of different graphene-based temperature sensors is compared to provide a comprehensive overview and vision for their future industrialization. These advancements highlight the potential of graphene-based sensors to revolutionize temperature sensing technologies.

8.1.1. Temperature sensing mechanisms

Graphene, a 2D zero-bandgap nanomaterial, exhibits unique thermal and electrical properties, making it a well-suited candidate for temperature sensing applications. The thermoresistive effect involving changes in resistance with temperature variations is prominently observed in 2D materials, which exhibit significant sensitivity to temperature changes [521]. These materials are often incorporated into flexible substrates to create highly responsive temperature sensors. The temperature-sensitive behavior of graphene is primarily influenced by its internal carrier mobility and concentration, which can vary under thermal excitation. Additionally, graphene’s coefficient of thermal expansion differs from that of other materials, causing it to stretch, shrink, or fold in response to temperature changes. The temperature sensitivity of graphene-based sensors is influenced by several mechanisms, including electro-phonon coupling, thermal expansion effects, intrinsic carrier excitation, and electron-charged particle interaction.

8.1.1.1. Electro-phonon coupling.

In metals and semiconductors, the conduction process involves the directional transport of charge carriers, primarily electrons. During this process, carriers scatter, altering their direction of motion, which affects conductivity and mobility. The scattering probability of carriers is influenced by interactions such as electron-electron interactions, lattice vibrations, and ionized impurities. Temperature impacts carrier transport mainly through electron-phonon scattering, originating from lattice vibrations [522,523].

As temperature decreases, phonon numbers diminish, making electron-phonon scattering a secondary factor, while ionized impurity scattering becomes dominant [524]. At low temperatures, the primary contributors to resistivity and mobility changes are impurities. However, as temperature rises, lattice vibrations disrupt the periodic potential field, causing electron scattering and transitions between different states in the energy band. This results in decreased carrier mobility and increased conductivity, displaying metallic characteristics. At higher temperatures, carrier concentration changes become the primary factor, leading to increased conductivity and semiconductor-like behavior. The temperature sensitivity of graphene-based sensors can be attributed to the thermal excitation of carriers in rGO [525,526] (Fig. 23a). As temperature increases, carrier hopping and tunneling between rGO sheets become more probable, enhancing carrier mobility and decreasing resistance, thus behaving as a negative temperature coefficient (NTC) sensor. Polyethyleneimine (PEI) improves the adhesion and stability of the rGO film on the substrate, ensuring durability and flexibility. Simulation results comparing thermal conductivity and thermal resistance under different conditions, using equilibrium molecular dynamics (EMD) and experimental data reveal the importance of understanding thermal transport properties in designing efficient temperature sensors [526] (Fig. 23b,c).

Fig. 23.

2D Material-based Temperature Sensing Mechanisms. (a) Schematic of carrier transport between neighboring rGO sheets. Enhanced temperature results in closer fitting between rGO and the substrate. (b,c) Simulation results of thermal conductivity and thermal resistance, comparing equilibrium molecular dynamics (EMD) and experimental (Exp.) results, with and without certain conditions. (d) Schematic of graphene resonators used for bolometric detection. (e) Schematic of thermal rectifier temperature sensors. (f) Schematic of a pyroelectric bolometer. (g) Schematic of the thermoelectric detection mechanism, illustrating how tuning the thermoelectric and plasmonic behavior of graphene converts electronic heat into a voltage through the thermoelectric effect. (h) Schematic of graphene-disk plasmonic resonators (GDPRs, red) connected by graphene nanoribbons (GNRs). (a) (Reproduced with permission from ref. [525]. Copyright 2019, Wiley). (b,c) (Reproduced with permission from ref. [526]. Copyright 2016, Springer Nature). (d) (Reproduced with permission from ref. [530]. Copyright 2019, Springer Nature). (e) (Reproduced with permission from ref. [533]. Copyright 2020, Elsevier). (f) Reproduced with permission from ref. [534]. Copyright 2019, Springer Nature). (g) (Reproduced with permission from ref. [535]. Copyright 2017, Springer Nature). (h) (Reproduced with permission from ref. [536]. Copyright 2018, Springer Nature).

8.1.1.2. Thermal expansion effect.

Graphene and its substrate often have different thermal expansion coefficients [528,529]. As temperature changes, this mismatch generates thermal stress due to constrained thermal expansion, particularly at high temperatures. The increased kinetic energy of molecular motion at higher temperatures enlarges the distance between molecules, while lower temperatures reduce it. Consequently, graphene undergoes thermal displacement and strain, generating stress that affects electrical conductivity and resistance.

Molecular dynamics simulations reveal the role of functional APTES on the thermal conductance of graphene films [527]. The FGO substrate minimizes perturbation on graphene morphology, maintaining high thermal conductivity. The presence of APTES molecules increases both thermal conductivity and interfacial thermal resistance for multi-layer graphene films due to enhanced cross-plane phonon coupling (Fig. 23b). For single-layer graphene, the APTES molecules reduce thermal conductivity due to scattering caused by a saddle-like surface around chemical bond.

8.1.1.3. Intrinsic excitation.

At lower temperatures, graphene’s conductivity decreases with increasing temperature due to the combined effects of electrophonon coupling and Coulomb potential. As temperature continues to rise, intrinsic excitation becomes more prominent. Higher temperatures cause some valence electrons to break free from covalent bonds, creating holes and increasing carrier concentration [530]. This intrinsic excitation leads to a transition from metallic to semiconducting properties, with increased conductivity as temperature rises. Graphene nanomechanical bolometers (GNBs) measure absorbed power by monitoring changes in the resonance frequency of suspended graphene membranes [530] (Fig. 23d). The temperature change induces thermomechanical stress, shifting the resonance frequency. Patterning graphene into a trampoline geometry enhances thermal resistance, leveraging graphene’s low heat capacity for sensitive light absorption measurements.

8.1.1.4. Electron-charged particle interaction.

Temperature influences the interaction between electrons and thermally vibrating charged particles. These interactions alter electron velocity and direction, creating a Coulomb potential field around the interacting ions and disrupting the periodic potential field near electrons [531–533]. This interaction changes the mobility of graphene carriers, affecting its resistivity.

A thermal rectifier designed as a Si membrane with spatially asymmetric porosity exploits the difference in thermal conductivity between forward and reverse configurations. Phonon scattering near cold ends due to porosity affects mean free paths, resulting in higher thermal conductivity in the forward configuration [533] (Fig. 23e). The rectification coefficient quantifies this difference, showcasing the topologically non-homogeneous design.

Pyroelectric bolometers (Fig. 23 f) connected by graphene represent advanced designs leveraging graphene’s unique properties for enhanced temperature sensing. The pyroelectric effect, where materials generate an electrical voltage in response to temperature changes, is another mechanism utilized in wearable temperature sensors [535]. Thermoelectric detection in graphene leverages plasmonic behaviour to convert electronic heat into voltage [536] (Fig. 23 g). Plasmons are generated by scattering scanning near-field microscopy to create near-field photocurrent, isolated from background light-induced current. This mechanism leverages graphene’s properties to achieve high sensitivity and efficiency in temperature sensing. Plasmonic graphene photodetectors combine graphene-disk plasmonic resonators (GDPRs) with graphene nanoribbons (GNRs) [537] (Fig. 23 h). GDPRs generate thermalized carriers, while GNRs facilitate transport through disorder potentials. Plasmon excitation increases electron temperature, enhancing carrier transport efficiency. This mechanism highlights the potential for thermal activation of electrical conductivity in graphene devices.

8.1.2. 2D material-based temperature biosensor fabrication

2D materials offer advantages for temperature sensing applications due to their superior electrical, thermal, and mechanical characteristics. The fabrication processes for these sensors are diverse, enabling the development of flexible, highly efficient temperature sensors suitable for wearable devices and healthcare monitoring. The functional lifespan and performance of fiber sensors were evaluated through tests including halogen lamp cyclic, mechanical deformation, spill, humidity, and machine-washing tests [538]. The fiber sensors were fabricated using a thermal drawing process (TDP), which allows for mass production on a kilometer-long industrial scale. The process involves adjusting the radial scale by controlling feed and capstan speeds [538] (Fig. 24a). Graphene fibers (GFs) with a diameter of ~50 μm were prepared from wet-spun GO films using a twist-draw process, similar to yarn production in the textile industry. The twisting procedure imparts a spring-like structure to the loosely stacked graphene sheets, enabling large tensile elongation beyond 15% in the axial direction [539] (Fig. 24b). Stretchable TFTs were developed using a styrene-ethylene-butadiene-styrene (SEBS) hydrogenated elastomer substrate. The TFTs featured a bottom-gate/top-contact structure with SEBS thin film serving as the gate dielectric. Single-walled CNTs (SWCNTs) were used as source-drain and gate electrodes, and semiconducting SWCNTs were used as the semiconductor. The stretchable temperature sensor demonstrated conformability when placed on the epidermis and during wrist bending, highlighting its functionality with an operating hairdryer approaching the sensor [540] (Fig. 24c–e). A wearable thermal patch integrated with an array of temperature sensors and a heater was developed [541] (Fig. 24f). The device included readout circuits for processing and externally transmitting data. The system allows monitoring of temperature distribution and thermal gradients, diagnosis of disease progression, and performing heat treatment to accelerate wound healing. An SU-8/Al₂O₃ composite was used to enhance the dielectric properties of a temperature sensor. Two-layer graphene was integrated under the substrate to support temperature sensing capability and function as a heater with a static bias current. The device exhibited good mechanical flexibility, conformable skin contact, and high optical transparency (~80%), enabling direct observation of the skin area [541] (Fig. 24g). The fiber sensor’s mechanical stability was assessed by altering the radii of curvature at various folding angles. The sensor maintained consistent temperature response at room temperature when bent at 90° and 180° angles with varying radii of curvature, demonstrating performance consistency over the selected temperature range (25–45 °C) [538] (Fig. 24h).

Fig. 24.

Graphene-based Temperature Sensors. (a) Fabrication of a temperature sensor using fiber drawing process. (b) Integration of graphene fibers into a cotton fabric for temperature sensing. (c-e) Graphene-based transistors and interconnects. (f,g) Capacitive sensor for temperature monitoring. (h) Resistance variations of the temperature sensor bent at a 90° angle with radii of curvature of 1.5 cm and 0.5 cm. (i) Self-healing temperature sensor with PBA–PDMS/ folded graphene film. (a) (Reproduced with permission from ref. [537]. Copyright 2023, Springer). (b) (Reproduced with permission from ref. [538]. Copyright 2020, Springer Nature). (c-e) (Reproduced with permission from ref. [539]. Copyright 2018, Springer Nature). (f-g) (Reproduced with permission from ref. [540]. Copyright 2022, Science). (h-i) (Reproduced with permission from ref. [541]. Copyright 2022, Springer).

A self-healing composite was created using a supramolecular framework of PBA–PDMS combined with folded graphene film. This composite exhibited high tensile strength and thermal conductivity, with self-healing efficiencies of 100% and 98.65% for tensile strength and thermal conductivity, respectively. The self-healing mechanism is attributed to efficient supramolecular interaction between polymer molecules and graphene, providing stability and functionality in next-generation electronic devices [542] (Fig. 24i). A flexible temperature sensor utilizing rGO-based nanocomposites on a PI substrate demonstrated a temperature coefficient of resistance (TCR) of −1.64 × 10⁻³ Ω K⁻¹ [543] (Fig. 25a). A skin-attachable flexible temperature sensor, developed on a PI substrate using a spray-dipping method [526] (Fig. 25b), demonstrated high sensitivity (1.30% °C⁻¹) and could detect temperature changes as small as 0.1 °C. The sensor’s excellent durability makes it suitable for long-term skin temperature monitoring.

Fig. 25.

Temperature Sensing 2D Material Deposition and Substrate Integration. (a) Fabrication of a rGO-Ag based flexible temperature sensor. (b) Structure of the rGO layered temperature sensor, and optical images of the sensor in normal and bent states. (c) Variation of normalized resistance with temperature. (d) Fabrication of an interdigital capacitive temperature sensor. (e) Schematic of e-skin temperature sensor arrays. (f) Sensor’s resistance response. (g) Schematic of single-layer graphene (SLG) temperature sensors. (h) Temperature sensor’s resistance response in the biological temperature range. (i) A stretchable graphene thermistor (top) including highly conductive Ag nanowires as electrodes and resistive graphene as temperature-sensing channels, and a black phosphorus (BP)@laser-induced graphene (LEG) based temperature sensor (bottom). (j-l) A transparent and stretchable gated temperature sensor responding to human muscle movement temperature, and human physical activity. (a) (Reproduced with permission from ref. [542]. Copyright 2017, Elsevier). (b) (Reproduced with permission from ref. [528]. Copyright 2019, Wiley). (c) (Reproduced with permission from ref. [543]. Copyright 2014, American Chemistry Society). (d) (Reproduced with permission from ref. [544]. Copyright 2014, IEEE). (e,f) (e,f) (Reproduced with permission from ref. [543]. Copyright 2014, American Chemistry Society). (g) (Reproduced with permission from ref. [545]. Copyright 2017, Springer Nature). (h) (Reproduced with permission from ref. [546]. Copyright 2021, Wiley). (i) (Reproduced with permission from ref. [547]. Copyright 2015, American Chemistry Society, and ref. [546]. Copyright 2021, Wiley). (j-l) (Reproduced with permission from ref. [520]. Copyright 2016, Wiley).

For e-skin applications, a temperature sensor was characterized between 21 and 80 °C, yielding a temperature sensitivity of ~0.25% °C⁻¹ [544] (Fig. 25c). A passive wireless integrated temperature sensor was fabricated using GO synthesized by a modified Hummers method. This sensor demonstrated effective temperature sensitivity due to the properties of GO and the interdigital capacitive design [545] (Fig. 25d). A flexible sensor array developed to detect three-axis force directions and temperature distribution mimics human skin’s sensory capabilities [544] (Fig. 25e,f). Monolayer and multilayer graphene temperature sensors have shown promising results, with resistance influenced by the number of conduction electrons, making them suitable for monitoring human body temperature [546,547] (Fig. 25g–h). Stretchable graphene-based thermistors could deliver high sensitivity and quick response to human body temperature changes [548] (Fig. 25i). A transparent, flexible temperature sensor using a FET configuration on a PDMS substrate could record temperature changes during physical activities, providing real-time monitoring of skin temperature [521] (Fig. 25j–l).

Wearable thermoelectric generators were fabricated using chemically exfoliated NbSe₂ (p-type) and WS₂ (n-type) films transferred to a PDMS substrate, and non-oxidized graphene nanosheets doped with surfactants were utilized to achieve high Seebeck coefficients for both nand p-type materials [549,550] (Fig. 26a). A tri-layer composite made up of a poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)) pyroelectric material placed between two electrodes was developed to enhance heating and cooling rates, crucial for thermoelectric generators [412] (Fig. 26b). Direct ink writing (DIW) was used to apply conductive rGO/poly(3-hydroxybutyrate) (PHB) ink on a flexible Kapton substrate, achieving good adhesion and stable temperature sensor performance [551] (Fig. 26c,d). A graphene fiber/PEDOT hybrid fiber synthesized using a hydrothermal process was used to create a thermoelectric generator for wearable energy harvesting applications [552] (Fig. 26e). A multifunctional sensor matrix, fabricated using CVD-grown graphene on Cu foil, transferred onto a PDMS substrate, and patterned using photolithography and etching, included capacitive humidity sensors and resistive temperature sensors [553] (Fig. 26f,g). rGO temperature sensors, suitable for robot skin and IoT applications, exhibited a TCR of 0.6345% °C⁻¹ and maintained stability under various pressures and deformations [554] (Fig. 26h). Composite rGO/PHB ink, prepared using an in situ reduction method, resulted in high-performance and stable temperature sensors [551] (Fig. 26i). Fig. 26j–l show the resistance changes with temperature for various sensors, including optical images and their applications on robotic skin [554]. The rGO-based sensors exhibited linear resistance responses with high sensitivity, making them suitable for precise temperature monitoring in advanced applications. These 2D material-based temperature sensors, with their advanced fabrication processes, offer promising solutions for precise temperature monitoring in various cutting-edge applications.

Fig. 26.

Advanced Composite and Hybrid Temperature Sensors. (a) Schematic of p-n doping through surface treatments of non-oxidized graphene. (b) Tri-layer composite of graphene, poly(vinylidene fluoride-co-trifluoroethylene) (P(VDF-TrFE)), and CNT/PDMS as a hybrid generator. (c) Direct ink writing (DIW) conductive rGO/poly(3-hydroxybutyrate) (PHB) ink on a flexible Kapton substrate. (d) Real-time temperature measurement of forehead skin. Vertical lines show fast response when loading and removing the device from a heated object. (e) Graphene/PEDOT:PSS fiber-based thermoelectric devices preparation. (f) Fabrication of a stretchable, multimodal all-graphene sensor matrix by photolithography, spray coating, and lamination. (g) Fabrication of graphite flakes utilizing super critical fluid method. (h) Fabrication of rGO-based flexible temperature sensors. (i) Synthesis and fabrication of PHB and rGO composites, including printed Ag electrodes with drop-coated rGO-PHB composite sensing elements. (j-l) Resistance changes with temperature, optical image of the sensor and its application on robotic skin. (a) (Reproduced with permission from ref. [549]. Copyright 2020, Wiley). (b) (Reproduced with permission from ref. [411]. Copyright 2014, Wiley). (c,d) (Reproduced with permission from ref. [550]. Copyright 2020, Wiley). (e) (Reproduced with permission from ref. [551]. Copyright 2020, American Chemical Society). (f-g) (Reproduced with permission from ref. [552]. Copyright 2016, Wiley). (h) (Reproduced with permission from ref. [553]. Copyright 2018, MDPI). (i) (Reproduced with permission from ref. [550]. Copyright 2020, Wiley). (j-l) (Reproduced with permission from ref. [553]. Copyright 2018, MDPI).

8.1.3. 2D material-based temperature sensors applications

Skin-like multifunctional sensors, critical in human-machine interaction, robotic, and healthcare applications, still face challenges in scalable and cost-effective production using conventional materials. The emerging two-dimensional transition metal carbide, Ti₃C₂T_x MXene, integrated with favorable thermoelectric properties, metallic-like conductivity, and a hydrophilic surface, offers a promising solution. These sensors are designed to precisely detect and distinguish temperature and pressure stimuli without crosstalk by decorating elastic and porous substrates with MXene sheets. The combination of thermoelectric and conductive MXene with a thermally insulating, elastic, and porous substrate integrates efficient Seebeck and piezoresistive effects, resulting in sensors with an ultralow detection limit (0.05 K) [555], high signal-to-noise ratio, and excellent cycling stability for temperature detection, as well as fast response, high sensitivity, and outstanding durability for pressure detection. Based on these dual-mode sensing properties, a multimode input terminal and an electronic skin are created, showing potential in robotic and human-machine interaction applications. This work provides a scalable fabrication method for multifunctional tactile sensors.

The MCP foam sensor is fabricated by decorating active MXene sheets on an elastic PDMS foam substrate to enable efficient temperature-pressure sensing [555] (Fig. 27a). MXene sheets are synthesized by selective etching of the Ti₃AlC₂ MAX phase followed by ultrasonic exfoliation. The process for smart textiles includes polydopamine (PDA) modification of the pristine elastic textile by self-polymerization of dopamine, MXene nanosheets decoration on the fiber surface by dip-coating, and surface treatment using PDMS [556] (Fig. 27b). The PDA layer chemically activates the textile, allowing firm adsorption of MXene nanosheets through hydrogen bonding, and the PDMS layer enhances interface interaction, prevents oxidation and degradation, and maintains breathability.

Fig. 27.

2D Material-based Wearable Temperature Sensors. (a) Schematic of the fabrication process of chitosan activation of PDMS foam. (b) Schematic of the stepwise preparation process and structure of the PM/PDMS temperature sensing textile. (c) Voltage and resistance response of temperature-pressure sensors. (d) Schematic of the e-skin with interlocked microdome array. (e) Schematic of waterproof and breathable properties of the PDA (PDA)/MXene/PDMS textile. (f) Schematic of loading MXene-based smart fabrics (M-fabric), achieving favorable breathability, exceptional Joule-heating performance, and a highly responsive humidity sensing ability. (g-h) SEM images of wood-multifunctional fire protection nanocoating with a dip-coating MXene, showcasing flame-retardant coating incorporated into the temperature sensors. g, Before burning. h, After burning. (i) Voltage response of temperature sensors under skin being bended and pressed conditions. (j) Current response of polymerized acrylamide/liquid metal/MXene/glycerol hydrogel to bear human breath and ice. (a) (Reproduced with permission from ref. [554]. Copyright 2023, American Chemical Society). (b) (Reproduced with permission from ref. [682]. Copyright 2021, Elsevier). (c) (Reproduced with permission from ref. [554]. Copyright 2023, American Chemical Society). (d) Reproduced with permission from ref. [684]. Copyright 2015, Science). (e) (Reproduced with permission from ref. [555]. Copyright 2021, Elsevier). (f) Reproduced with permission from ref. [558]. Copyright 2020, American Chemical Society). (g,h) Reproduced with permission from ref. [479]. Copyright 2022, Elsevier). (i) Reproduced with permission from ref. [485]. Copyright 2023, Science). (j) (Reproduced with permission from ref. [559]. Copyright 2023, Elsevier).

The MCP sensor’s porous network of MXene combines electrically conductive and thermoelectric features with ultralow thermal conductivity (0.065 W m⁻¹ K⁻¹) [555,557] (Fig. 27c). Inspired by human skin’s complex sensory system, unique microstructures, and receptors, multifunctional tactile sensors are designed to detect various static and dynamic stimuli, temperature, and surface textures [558] (Fig. 27d). The piezoresistive sensors with interlocked microdome arrays can detect multiple stimuli, similar to human skin.

The PM/PDMS smart textile’s breathability and waterproofness involve water vapor transmission through gaps inside the textile while the PDMS layer renders excellent waterproofness [556] (Fig. 27e). MXene-coated cellulose fiber non-woven fabrics (M-fabric) are created by dip-coating commercial cellulose fabrics with MXene dispersion. The successful synthesis of delaminated MXene sheets allows the MXene to be deposited onto cellulose fibers, converting them into black M-fabric while retaining the original structure [559] (Fig. 27 f). The multifunctional fire protection nanocoating (MFNC) features a TE layer and FR layer on a flammable substrate, constructed by dip-coating MXene, UPC, and MMT via layer-by-layer assembly. The FR layer acts as the epidermis layer, protecting from damage and burning, while the TE layer acts as the dermis layer, sensing temperature variations and warning of fire danger. MFNC exhibits sensitive temperature-sensing capability and outstanding flame retardancy [480] (Fig. 27g–h). A hybrid MXene/Clay/poly(N-isopropylacrylamide) (PNIPAM) hydrogel temperature sensor, inspired by biological thermoreceptors, is fabricated with high stretchability and deformation-insensitive temperature sensing properties. The sensor can accurately distinguish temperature changes even when the skin is stretched or pressed [485] (Fig. 27i).

Conductive gels, such as cellulose nanofiber stabilized liquid metal droplets, are utilized to construct polyacrylamide/MXene/glycerol hydrogels with stretchability (1000%) and high environmental adaptability (−25 to 80 °C). These hydrogels serve as flexible electrodes for triboelectric nanogenerators and versatile sensors combining temperature and deformation sensing [560] (Fig. 27j). A CVD-grown graphene thin layer as the top electrode, allows rapid cooling and heating of the pyroelectric P(VDF-TrFE) layer in response to changes in ambient temperature. The CNT/PDMS composite bottom layer acts as a stretchable and flexible substrate [561] (Fig. 28a). This sandwich structure, with a piezoelectric effect, can generate electrical voltages from mechanical deformations, functioning as a hybrid generator suitable for wearable electronics and robotics.

Fig. 28.

2D Material-based Temperature Sensors Applications. (a) Alternating voltages from piezoelectric and pyroelectric effects. (b) Structure of 3D FET temperature sensors. (c) Normalized resistance variation with temperature at zero grid voltage. (d,e) High-temperature pressure sensor with Si₃N₄ protective layer and normalized resistance vs. temperature. (f) Schematic of nanocomposite temperature sensors. (g) Current-voltage curves at different temperatures. (h) Schematic of CNTs/poly(ionic liquid)(PIL)/rGO hybrid material. (i) Temperature effect on resistivity of AgNW/thermally rGO/ PVDF (AgNW/TRG/PVDF) composite. (j) Resistance of GO/Ni composites over three heating and cooling cycles. (k) Array of four temperature sensors attached to skin. (a) (Reproduced with permission from ref. [411]. Copyright 2014, Wiley). (b,c) (Reproduced with permission from ref. [561]. Copyright 2021, IOP Publishing). (d,e) (Reproduced with permission from ref. [562]. Copyright 2023, IEEE). (f) (Reproduced with permission from ref. [563]. Copyright 2016, IEEE). (g) (Reproduced with permission from ref. [564]. Copyright 2018, Elsevier). (h) (Reproduced with permission from ref. [565]. Copyright 2015, Wiley). (i) (Reproduced with permission from ref. [566]. Copyright 2014, Springer). (j) (Reproduced with permission from ref. [567]. Copyright 2020, IEEE). (k) (Reproduced with permission from ref. [568]. Copyright 2016, Springer Nature).

Using traditional micro-nano manufacturing technology, 3D GFETs are fabricated through processes involving Si wafer preparation, deposition of sacrificial and gate layers, and the transfer of CVD-grown graphene. This comprehensive process involves multiple steps of lithography, etching, and deposition to achieve the final 3D GFET structure [562] (Fig. 28b). Fig. 28c illustrates the normalized resistance of 3D GFETs under zero grid voltage, showing high sensitivity with a temperature coefficient of resistance (TCR) of about 0.41% °C⁻¹ at 150 °C [562]. High-temperature pressure sensors are protected with a Si₃N₄ layer to ensure stability and performance at elevated temperatures [563] (Fig. 28d,e).

Nanocomposite temperature sensors utilize advanced materials for enhanced performance. Fig. 28 f provides a schematic of these sensors, which integrate different 2D materials for superior sensitivity and stability [564]. Current-voltage curves at various temperatures demonstrate the thermal generation of excess carriers and the role of electron-phonon and phonon-phonon scattering in temperature sensing [565] (Fig. 28 g). Hybrid materials like CNTs/poly(ionic liquid)/rGO combine properties to create high-performance temperature sensors [566] (Fig. 28 h). The effect of temperature on the resistivity of Ag/Thermally rGO/PVDF composites shows a positive temperature coefficient (PTC) effect [567] (Fig. 28i). The resistance variation of GO/Ni composites over multiple heating and cooling cycles is crucial for wearable devices [568] (Fig. 28j). Finally, an array of printed temperature sensors demonstrates practical applications in monitoring skin temperature [569] (Fig. 28k).

8.1.4. Challenges and perspectives of temperature sensors

Although substantial research has been conducted on the use of graphene in wearable temperature sensors, significant challenges remain. One major issue is the annealing process for producing rGO from graphene platelets and nanowalls, which requires precise and low heating temperatures to maintain mechanical strength. Current nanocomposite-based sensors face drawbacks such as compromised electrical conductivity, lack of structural integrity, low repeatability, and poor long-term stability. To address these issues, the use of suspended graphene for flexible thermistors is recommended due to its high elasticity, optical transparency, electron mobility, and thermal conductivity.

Multimodal and self-powered wearable temperature sensors represent a significant advancement in the field of wearable electronics, integrating multiple sensing mechanisms and energy-harvesting capabilities to create high-performance and versatile devices [570,571] (Fig. 29). Multimodal sensors, which utilize electrothermal temperature sensing, rely on the Seebeck phenomenon, where a temperature differential across a material generates a voltage [521]. This effect is pivotal in converting temperature changes directly into electronic impulses, making it well-suited for temperature monitoring applications. Materials with low thermal conductivity are preferred as they enhance the sensitivity and performance of the sensor. In 2016, a groundbreaking approach was proposed involving the use of freestanding graphene foam as an electrothermal sensor that operates without an external power supply. In this setup, the temperature sensing layer comprises a PU matrix integrated with graphene nanosheets. This combination leverages the excellent thermal and electrical properties of graphene, enabling the sensor to detect temperature variations accurately and efficiently. This multimodal sensor not only excels in temperature monitoring but also holds potential for various other applications due to its high sensitivity and flexibility [571].

Fig. 29.

Challenges and Perspectives on Wearable Temperature Sensors. (Reproduced with permission from ref. [520]. Copyright 2016, Wiley; reproduced with permission from ref. [570]. Copyright 2020, American Chemistry Society; reproduced with permission from ref. [620]. Copyright 2023, Elsevier).

E-skin mimics human skin by converting external stimuli such as pressure, heat, and humidity into electrical signals [572,573]. The development of textile-based e-skin focuses on achieving multifunctional sensing capabilities, distinguishing signals from various stimuli, and designing large-area sensing arrays. A key component of integrated smart systems (ISSs) is the power supply system, essential for continuous operation. To eliminate the need for conventional batteries and capacitors, researchers have developed textile-based energy-harvesting and storage devices. Among these, thermoelectric textiles that harvest energy from body heat are promising, as they can provide a continuous power supply without location constraints. However, few reports exist on e-skin powered by thermoelectric materials, and integrating textile-based thermoelectric powering with multisensing systems remains a rarity [557].

A novel approach involves modifying a 3D spacer fabric substrate with PEDOT:PSS to create a large-area wearable self-powered pressure-temperature e-skin [571–574]. The unique 3D structure of the SF substrate, combined with its excellent compressibility, flexibility, high permeability, low cost, and productivity, makes it a well-suited choice. PEDOT, an organic thermoelectric material, works synergistically with SF for energy harvest from the heat difference between the environment and the body, thus enabling continuous self-powering and converting temperature stimuli into voltage signals. Simultaneously, the pressure deformation of PEDOT/SF alters its conductivity, allowing it to convert pressure stimuli into current signals. This integration of self-powering and pressure-temperature sensing into a single pixel simplifies the structure and fabrication of ISSs.

8.2. Sweat sensing

Wearable electrochemical sweat sensing systems constitute a remarkable advancement in real-time health monitoring and diagnostics. The integration of electronic modules and wireless communication strategies into these sensors enables continuous, real-time analysis while maintaining low power consumption and user comfort. Sweat is a rich source of biomarkers, providing crucial insights into an individual’s physiological state without the need for invasive procedures [575]. The ability to monitor electrolytes, metabolites, trace metals, drugs, and small molecules in sweat offers significant advantages for personalized health management and disease detection [576–580].

8.2.1. 2D materials-based sweat sensing mechanism

By leveraging the unique properties of 2D materials, these sensors enable real-time, continuous analysis of various biomarkers in sweat. Sweat samples can be obtained through passive or active approaches. In the passive approach, individuals engage in intense physical activity to induce sufficient sweat secretion [581]. The active approach involves electrical stimulation. Iontophoresis, a widely used method of active sweat induction, allows the acquisition of sweat samples while the body remains sedentary. A current is generated under the skin surface by applying a voltage between iontophoretic electrodes, delivering an agonist (e.g., pilocarpine) to the sweat gland and stimulating sweat secretion [582]. This method has been applied to monitor levels of chloride, ethanol, and glucose [583] (Fig. 30a). Simultaneous analysis of sweat and interstitial fluid (ISF) can expand the scope of biomarker detection and improve clinical accuracy. A dual-function biosensor utilizes iontophoresis for simultaneous biofluid sampling and analysis. This system operates by extracting ISF through reverse iontophoresis in the cathodic chamber and inducing sweat in the anodic chamber at different skin locations. It enables the simultaneous alcohol detection in glucose and sweat in ISF, combining the advantages of both biofluids to enhance detection accuracy [584] (Fig. 30b).

Fig. 30.

2D Materials-based Sweat Sensing Mechanisms. (a) The underlying mechanism of iontophoresis. (b) A dual-function biosensor utilizing iontophoresis for simultaneous biofluid sampling and analysis, which actively induces sweat and tissue fluid production. (c) Enhanced sweat alcohol detection using an iontophoresis method with a hex-wick design. (d) A self-sufficient, fully integrated sweat sensor based on iontophoresis, featuring a programmable current source. (e) A 2D chemical sensor to detect sweat components using different analytes and receptors. (f) The sensing mechanism for a noninvasive wearable device for levodopa monitoring. (g) A wearable sweat sensor for detecting trace amounts of metallic zinc in real-time. (h) A fully printed and wearable microfluidic nanosensor for copper detection in sweat. (a) (Reproduced with permission from ref. [582]. Copyright 2019, Elsevier). (b) (Reproduced with permission from ref. [583]. Copyright 2018, Wiley). (c) (Reproduced with permission from ref. [584]. Copyright 2018, RSC). (d) (Reproduced with permission from ref. [585]. Copyright 2017, National Academy of Sciences USA). (e) (Reproduced with permission from ref. [34]. Copyright 2023, American Chemistry Society). (f) (Reproduced with permission from ref. [586]. Copyright 2019, American Chemistry Society). (g) (Reproduced with permission from ref. [587]. Copyright 2015, Elsevier). (h) (Reproduced with permission from ref. [588]. Copyright 2022, Elsevier).

A wearable sweat alcohol biosensing device based on iontophoresis has innovatively maintained data integrity of sweat samples. The device isolates sweat irritants from the skin with a membrane, preventing dilution. Sudomotor axon reflex sweating minimizes mixing between fresh and old sweat, while carbachol as an agonist ensures a longer, stable rate of sweat production. The “hex wick” material prevents skin contamination and quickly transports the sample to the sensor, reducing latency [585] (Fig. 30c). Conventional iontophoresis can corrode electrodes and cause skin discomfort. A programmable current source periodically induces sweat with an upper limit on the ion electroosmotic current as a safety mechanism, effectively avoiding overheating and burning of the skin [586] (Fig. 30d,e).

Monitoring drug levels is crucial for adjusting doses and understanding pharmacokinetics. Wearable sweat sensing assays face challenges due to the low concentration of drug molecules in biofluids. A pioneering wearable sensing platform for drug detection in sweat was reported in 2018, utilizing DPV for sensitive determination of drug concentration. Roll-to-roll printing technology enables mass production of high-performance electrode arrays. Integrated signal transduction, modulation, processing, and Bluetooth transmission allow noninvasive, real-time monitoring of methylxanthine drugs in sweat. This platform significantly facilitates pharmacokinetic studies. Adding Au dendritic nanostructures to electrodes improves stability, achieving noninvasive levodopa detection in sweat [587] (Fig. 30 f). Human sweat contains trace elements related to health status. In 2015, a wearable electrochemical sensor for trace metal detection was developed using a bismuth/sodium fluoride-coated electrode. The sensor uses square wave anodic stripping voltammetry to trace zinc in sweat [588] (Fig. 30 g). A fully printed and wearable microfluidic nanosensor has been developed for copper detection in sweat, demonstrating the capability for real-time monitoring of this trace metal [589] (Fig. 30 h).

8.2.2. 2D materials-based sweat sensor fabrication

2D materials-based sweat sensors leverage the exceptional properties of 2D materials to create advanced, flexible, and sensitive devices for real-time monitoring of sweat composition. These sensors enable continuous health monitoring, providing insights into physiological conditions and responses.

Strenuous exercise or pharmacological stimulation is often required to produce sweat for analysis. The low secretion rate and rapid evaporation of sweat at rest limit the volume available to be collected in a sensor. For patients unable to exercise or sensitive to the burn of iontophoresis, a sweat collection method using a waterproof, skin-contact microfluidic system is a good solution. However, this method can only capture sweat after warm water showering or bathing and cannot continuously analyze thermoregulatory sweat at rest. A recently developed microfluidic-based sweat patch achieves continuous measurement of the at-rest thermoregulatory sweat composition and rate. The device consists of a microfluidic layer, an electrochemical sensing electrode, and a laminated hydrophilic housing. The sweat collection wells of the microfluidic layer are filled with SU8 padding coated with hydrogel, which quickly absorbs sweat at a low secretion rate, effectively solving the problem of dead volume. This wearable sweat sensor fills a gap in devices for the continuous detection of resting sweat, aiding in the study of correlations between sweating rates and sweat composition [590] (Fig. 31a).

Fig. 31.

2D Materials-based Sweat Sensor Fabrication. (a) Structure of microfluidic sweat sensor for continuous analysis of body sweat at rest. (b) Fast, stress-free molecularly imprinted polymer (MIP) cortisol sensor. (c) Wearable laser-engraved sensor for detecting uric acid and tyrosine in sweat. (d) 2D materials-based sweat sensor fabrication: exfoliation of bulk MoS₂ into 2D flakes, graphene growth by laser ablation, transducer electrode patterning, and formation of the active layer and contacts. (a) (Reproduced with permission from ref. [589]. Copyright 2021, Springer Nature). (b) (Reproduced with permission from ref. [590]. Copyright 2021, Wiley). (c) (Reproduced with permission from ref. [591]. Copyright 2020, Springer Nature). (d) (Reproduced with permission from ref. [592]. Copyright 2022, American Chemical Society).

Sweat contains a wealth of components, including cytokines, hormones, and proteins. By detecting these molecules, a deeper understanding of body homeostatic mechanisms and overall health can be obtained. Continuous monitoring of small molecule biomarkers in sweat over a long duration is challenging due to changes in composition and pH over time and at low concentrations. Detection can be performed using antibodies or aptamers for affinity-based sensing, as well as molecularly imprinted polymer (MIP)-based methods. A wearable touch-based cortisol sensor uses a permeable, perspiration-wicking porous PVA hydrogel for fast and effective sweat collection. MIP electrochemical sensor technology achieves simple, fast, and reliable sweat cortisol detection through a highly selective combination with a cortisol-imprinted electropolymerized polypyrrole coating. This MIP-based fingertip cortisol sensing provides a reliable and practical method for rapid and stress-free pressure monitoring [591] (Fig. 31b). Measuring uric acid and tyrosine in sweat is challenging due to their low concentrations. Utilizing the advantages of graphene in electrochemical sensing, rapid and accurate in situ detection of uric acid and tyrosine in human sweat is achieved by using DPV to assess ultralow levels based on the amplitude of the peak oxidation current. This wearable laser-engraved sensor enables sensitive detection of uric acid and tyrosine in sweat [592] (Fig. 31c).

Bulk MoS₂ with a purity of 99% was ground into a paste with ethanol, dried, and then dispersed in an ethanol and water-based solvent. The solution was sonicated to separate the particles and exfoliate the MoS₂ flakes into few and monolayered sheets. Centrifugation separated the exfoliated flakes from the unexfoliated bulk particles. The resulting solution was found to contain exfoliated 2D MoS₂ flakes (Fig. 31d). Graphene for the active layer was grown on the surface of PI tape through direct laser ablation. The laser ablation process results in the formation of five-, six-, and seven-membered rings of carbon, creating a porous structure of graphene. The graphene powder was obtained by scratching off the LIG sheet from the PI surface. The 2D nanocomposite solution for the active layer of the sensor was prepared by adding powdered graphene to the prepared MoS₂ solution, resulting in a suspension that was sonicated to homogenize the composite before deposition as the active layer thin film. The final ratio of MoS₂ to graphene in the composite was selected to cancel out the effect of temperature in the final composite [593] (Fig. 31d).

8.2.3. 2D materials-based sweat sensing applications

Conventional flexible substrates including PI, PET, PDMS, PU, and PMMA provide excellent properties for wearable devices but often lack breathability and stretchability. In contrast, textile-based platforms offer natural breathability, promoting natural sweating and evaporative cooling while maintaining high flexibility, softness, and comfort. These textiles can be integrated into existing garments, making sweat monitoring more versatile [594] (Fig. 32a).

Fig. 32.

2D Materials-based Sweat Sensing Applications. (a) Fully integrated wearable sensing platform using polyethylene terephthalate (PET) as the substrate. (b) Embroidered electrochemical sensor for biomolecular detection on a cotton T-shirt. (c) Photograph of a carbon-based wearable electrochemical sweat analysis patch. (d) Thread-based multiplexed sensor patch for sweat monitoring. (e) Schematic of GO/silk fibroin (SF)/LiBr ink and the current change rate of the humidity sensor. (f) Schematic of a battery-free, biofuel-powered e-skin that efficiently harvests energy from the human body. (g) Structure of a capillary force-driven microfluidic sweat sensor. (h) Flexible rGO sensor for real-time moisture sensing with different relative humidity (RH) ranges. (i-k) Flexible microfluidic graphene sensor array on the skin under mechanical deformation and the circadian rhythm of sweat stress hormone. (l-n) Schematic of graphene-polymer (GP) hybrid electrochemical devices for diabetes monitoring. (o) Humidity sensing performance of a flexible and stretchable WS₂ sensor. (p,q) Optical microscopy image of the MoS₂-based pH sensor and its sensing performance driven by a MoSe₂ nanogenerator. (a) (Reproduced with permission from ref. [593]. Copyright 2016, Springer Nature). (b) (Reproduced with permission from ref. [594]. Copyright 2016, RSC). (c) (Reproduced with permission from ref. [595]. Copyright 2019, Science). (d) (Reproduced with permission from ref. [596]. Copyright 2020, Springer Nature). (e) (Reproduced with permission from ref. [597]. Copyright 2022, Springer Nature). (f) (Reproduced with permission from ref. [598]. Copyright 2020, Science). (g) (Reproduced with permission from ref. [599]. Copyright 2018, American Chemical Society). (h) (Reproduced with permission from ref. [600]. Copyright 2017, American Chemical Society). (i-k) (Reproduced with permission from ref. [601]. Copyright 2020, Cell Press). (l-n) (Reproduced with permission from ref. [602]. Copyright 2020, Springer Nature). (o) (Reproduced with permission from ref. [603]. Copyright 2017, RSC). (p,q) Reproduced with permission from ref. [410]. Copyright 2020, American Chemical Society.

Textile-based platforms provide an alternative to conventional substrates by offering natural breathability and flexibility. For instance, a smart textile electrochemical sweat lactate sensing device was assembled by weaving functional stretchable Au fiber electrodes into the fabric, maintaining high performance even under tensile strain. This approach addresses inhomogeneity and variability issues caused by core wicking of inks on textured fabric surfaces by using functionalized threads woven together [595] (Fig. 32b).

MXene, with its hydrophilic surface and high electrical conductivity, excels in applications like sensors, catalysis, and energy storage. Intrinsically N-doped carbon, silk-derived textile offers a promising material for working electrodes in wearable electrochemical sensors owing to its hierarchical structure and porous mesh weave, which facilitate good contact with reactants and effective electron transfer [596] (Fig. 32c).

To improve uniformity and performance in fabric-based electrochemical sensors, a bottom-up approach using functionalized threads woven together as a substrate, encapsulated as a patch, forms a reliable and accurate wearable multiplexed sweat-sensing platform [597] (Fig. 32d). Graphene-based materials have significant potential for biosensors. A flexible humidity sensor based on a metal-air redox reaction was developed using a GO/silk fibroin/LiBr ink, presenting high sensitivity to humidity changes with a quick response time. This device can be easily integrated into clothing or wearable devices for continuous health monitoring [598] (Fig. 32e). Wearable sweat sensors require reliable power sources. A perspiration-powered integrated electronic skin (PPES) uses nanomaterial integration to improve power strength and long-term stability, enabling selective continuous monitoring of key metabolic analytes in sweat [599] (Fig. 32 f). Microfluidic sweat sensors enhance sweat collection by channeling generated sweat through a microfluidic tube into a reservoir for storage and analysis, minimizing sample leakage, evaporation, and contamination [600] (Fig. 32 g). A flexible rGO sensor was developed for real-time moisture sensing across different relative humidity ranges, offering high sensitivity and rapid response, suitable for various wearable applications [601] (Fig. 32 h). A graphene-based flexible microfluidic sensor array can monitor multiple biomarkers simultaneously, providing continuous real-time monitoring of lactate, glucose, uric acid, ascorbic acid, K⁺, and Na⁺ in sweat. Integration with mobile healthcare systems enables tracking cortisol dynamics, providing relevant health information [602] (Fig. 32i–k). A groundbreaking electrochemical sensor based on graphene was developed for monitoring sweat-based glucose and pH and for controlled transcutaneous drug delivery. The device uses Au-doped CVD-graphene on a serpentine mesh for improved electrochemical activity, mechanical reliability, and optical transparency, providing stable electrical signal transfer for daily glucose monitoring and potential diabetes treatment [603] (Fig. 32l–n). Atomically thin 2D TMDs are promising for wearable chemical sensors. A wearable sensor based on WS₂ film demonstrated high humidity sensing performance under high flexibility, enabling real-time breath monitoring and potential healthcare applications [604] (Fig. 32o). A self-powered pH sensor was developed using a monolayer of CVD-grown MoSe₂ and mechanically exfoliated MoS₂ flakes. The MoSe₂ PENG drives the MoS₂ pH sensor, exhibiting superb sensing performance with a rapid response rate. The ultralow power consumption and excellent electrical properties of 2D materials highlight their promise in wearable chemical sensors for healthcare applications [605] (Fig. 32p,q).

8.2.4. Challenges and perspectives of sweat sensors

Integrating multifunctional perspiration analysis into wearable sensors has shown promising progress. The integration of wireless data transmission devices and smartphone applications has improved user accessibility and the utility of these sensors. Additionally, merging electrochemical sensors with Si integrated circuit assemblies can enhance the accuracy of physiological assessments by reducing signal noise through advanced signal processing [606]. The incorporation of iontophoresis devices allows for sweat collection through chemical stimulation while the body is at rest, and microfluidics can expedite the collection and analysis of sweat, enabling detection at low concentrations. Furthermore, the development of sensors capable of simultaneously detecting multiple analytes is advancing, allowing for a more comprehensive analysis of physiological states and improved calibration and accuracy.

Improving the reliability of sweat samples is crucial. Wearable sensors should adapt to varying sweating rates and compensate for these variations to ensure accurate biomarker distribution measurements. Different methods of sweat sample collection, such as exercise-induced or iontophoretic stimulation, can result in varying compositions [607], which should be considered for accurate analysis. Preventing changes in sweat composition during testing is essential, as small amounts of exposed sweat can evaporate quickly, altering biomarker concentrations. Isolating sweat from skin contaminants is also necessary to avoid interference with sensor readings.

Developing methods to induce sweat secretion in sedentary environments is essential for disease detection and continuous monitoring [608–610]. Iontophoresis shows promise but can damage the skin with prolonged use. Alternative technologies for biomarker extraction and improved microfluidic devices using hydrophilic materials may address the challenge of low sweat secretion rates at rest [611].

8.3. Noninvasive glucose sensing

The development of practical glucose sensors has been a subject of extensive research since the first enzyme-based electrode for glucose detection was introduced by Clark and Lyons in 1962 [612]. This early device relied on glucose oxidase immobilized within a semipermeable membrane, leading to the first commercial glucose analyzer in 1975 [613,614]. The advancement of glucose sensing technologies has significantly improved the management of diabetes, affecting over 350 million people worldwide [615], with an estimated global prevalence of over 640 million by 2040 [616].

Electrochemical glucose sensors have evolved through three generations, each improving sensitivity, selectivity, miniaturization, and integration with modern technologies. First-generation sensors were based on glucose oxidase and hydrogen peroxide detection. Second-generation sensors introduced synthetic electron acceptors and screen-printing technologies, enabling disposable and cost-effective devices. Third-generation sensors focused on direct electron transfer and the integration of advanced materials to enhance performance and stability.

Common sensing devices in this field include flexible electrochemical sensors that utilize enzymatic and non-enzymatic detection methods. Enzymatic sensors typically employ glucose oxidase or glucose dehydrogenase to catalyze the oxidation of glucose, producing a measurable electrochemical signal. Non-enzymatic sensors, on the other hand, rely on the direct oxidation of glucose on the surface of nanomaterials, such as metal oxides or nanoparticles, integrated with 2D materials.

8.3.1. Noninvasive glucose sensing mechanisms

Noninvasive glucose sensing mechanisms have attained remarkable attention because of their potential to improve diabetes management by eliminating the need for invasive blood sampling [617,618]. These technologies leverage various principles such as reverse iontophoresis, photoelectrochemistry, and innovative materials to provide continuous and reliable glucose monitoring.

One prominent technique, reverse iontophoresis, utilizes a small electric current to extract both uncharged and charged polar biomolecules, such as glucose, across intact skin [619]. This method leverages electromigration to move ions directly under the influence of an electric field and electroosmosis to transport neutral molecules like glucose. By placing an anode and cathode on the skin surface, a low current is applied, inducing the flow of Cl⁻ and Na⁺ ions towards the cathode and anode, respectively. The skin’s negative charge at physiological pH attracts positive ions (Na⁺), resulting in an electroosmotic flow of interstitial fluid with dissolved glucose towards the cathode. The extracted glucose is collected in a reservoir near the cathode and quantified by an externally attached electrochemical glucose sensor [620] (Fig. 33a).

Fig. 33.

Noninvasive Glucose Sensing Mechanisms. (a) Schematic of glucose extraction using reverse iontophoresis. (b,c) Photograph of the GlucoWatch biographer and SugarBeat continuous glucose monitoring (CGM) system for continuous glucose monitoring. (d) Working principle of a glucose sensor based on a graphene/PtO/n-Si heterostructure. (e) A graphene-WO₃-Au triplet junction and its energy levels for glucose detection. (f-h) Printable Tattoo-Based iontophoretic-sensing system for glucose monitoring. (a) (Reproduced with permission from ref. [619]. Copyright 2022, MDPI). (b,c) (Reproduced with permission from ref. [621]. Copyright 2002, Wiley). (d) (Reproduced with permission from ref. [623]. Copyright 2019, MDPI). (e) (Reproduced with permission from ref. [624]. Copyright 2014, Elsevier). (f-h) (Reproduced with permission from ref. [625]. Copyright 2015, American Chemical Society).

The pioneering wrist-watch device, GlucoWatch, approved by the FDA in 2001 [621], exemplified this approach. It employed iontophoresis electrodes to extract glucose from interstitial fluid through the skin, storing it in a hydrogel pad where it reacted with glucose oxidase to form H₂O₂. This reaction was then detected by an electrochemical glucose sensor. Despite its initial success, issues such as long warm-up times and skin irritation led to its market withdrawal. However, the fundamental principles it introduced continue to inspire modern innovations in noninvasive glucose monitoring using advanced 2D materials. Despite its innovation, the device faced issues such as poor accuracy during sweating and skin irritation, leading to its withdrawal in 2008. The concept continued, leading to the development of the SugarBeat CGM system, which is a patch-type sensor worn on the upper arm. It provides glucose readings every five minutes after a 25-minute warm-up period, displaying the data on a mobile app [622,623] (Fig. 33b,c).

A unique glucose sensor design utilizing a graphene/PtO/n-Si heterostructure enhances electrocatalytic activity and sensor stability. The sensor works by oxidizing glucose molecules on the surface of the PtO thin film, forming gluconolactone, H₂, and electrons. The current is measured from the graphene surface after applying a forward bias to the Si terminal. The interaction between PtO and graphene shifts the graphene’s Fermi level and causes p-doping, improving sensitivity by increasing active sites for glucose oxidation and enhancing charge carrier mobility and concentration [624] (Fig. 33d).

Photoelectrochemistry combines light with electrochemical processes to enhance sensing capabilities. Combining semiconductors like WO₃ with graphene and depositing Au nanoparticles creates a Schottky junction, reducing recombination rates and improving photocatalytic capabilities. In a glucose detection system, photogenerated holes oxidize glucose to gluconic acid, while photogenerated electrons convert water to H₂ on the cathode. The generated photocurrent serves as the sensing signal, demonstrating the effectiveness of the photoactive materials in responding to biomolecules [625] (Fig. 33e).

Non-invasive glucose sensing platforms using sweat, saliva, and tears are well-suited for diabetes management as they avoid blood contact and exposure to the immune system. A notable approach is a tattoo-based iontophoretic-sensing system. This platform uses a low current density to extract skin interstitial fluid (ISF) and selectively senses glucose using a glucose oxidase-modified Prussian Blue transducer at a low applied potential. Additionally, an ultrathin flexible skin-like biosensor system based on a biocompatible paper battery with an Au electrode has been designed for accurate, non-invasive monitoring of intravascular blood glucose. These biosensors exhibited high sensitivity and were validated in human clinical trials, showing high correlation with clinically measured glucose concentrations [626] (Fig. 33f–h).

8.3.2. Noninvasive glucose sensors fabrications

Noninvasive glucose sensors have revolutionized diabetes management by enabling continuous monitoring without the need for invasive blood sampling [627–629]. These advanced sensors utilize various fabrication techniques and innovative materials to enhance sensitivity, stability, and functionality.

The microfluidic device design integrates an electrochemical biosensor into the reservoir of the microfluidic system, allowing for real-time monitoring of sweat metabolites such as glucose and lactate [630]. This combination of electrochemical epidermal sensing and microfluidics ensures a short sweat sampling time, fast flow rate, and efficient transmission over the detector surface. This method overcomes the limitations of traditional silicon elastomer-based platforms, which require complex manufacturing processes and expensive microfabrication equipment. The use of CO₂ laser engraving for patterning further simplifies and reduces the cost of production [631] (Fig. 34a,b). The contact lens sensor fabrication process involves several steps, including applying and exposing a photoresist to ultraviolet light through a mask, developing the photoresist, evaporating thin metal films, and performing a lift-off process to leave the desired metal pattern. The sensor, cut from a polymer substrate and molded to the shape of a contact lens, is then functionalized with glucose oxidase enzymes [632] (Fig. 34c–e). These devices can be miniaturized into disposable strips, providing flexible options for glucose monitoring. Using microneedles, the device can transdermally deliver drugs like metformin and rosiglitazone into skins, controlled thermally in a multistage manner [633,634]. The glucose sensor, based on a Prussian blue-deposited porous Au electrode, is calibrated for glucose concentrations typical in human sweat and maintains stability under mechanical deformation, selectively detecting glucose even in the presence of other biomolecules and drugs [635] (Fig. 34f). The sensor operates by creating localized and transient alkaline conditions (high pH) for glucose sensing, even in neutral fluids. This transient high pH is achieved by applying a voltage to the metal contact, which absorbs and releases H⁺ ions to modulate the local pH, allowing for effective glucose detection [636] (Fig. 34g, h).

Fig. 34.

Additive Manufacturing for Noninvasive Glucose Sensors. (a,b) Microfluidic device design and operation. (c) Contact lens sensor fabrication process: Photoresist applied and exposed to ultraviolet (UV) light through a mask; development of photoresist and evaporation of thin metal films. Metal pattern remains after lift-off; sensor cut from polymer substrate, molded to contact lens shape, and functionalized with enzymes. (d,e) Images of cut-out sensors. (f) Wearable/disposable sweat monitoring device with microneedle-based transdermal drug delivery. (g,h) Non-enzymatic glucose sensor with interdigitated contacts: When on, Pd contact increases pH, making Au/Co₃O₄ contact reactive, oxidizing glucose, and increasing current (I_g). When off, pH is physiological (pH 7), no sensing occurs, and I_g is zero. (a,b) (Reproduced with permission from ref. [690]. Copyright 2017, American Chemical Society). (c-e) (Reproduced with permission from ref. [691]. Copyright 2011, Elsevier). (f) (Reproduced with permission from ref. [692]. Copyright 2017, Science). (g,h) (Reproduced with permission from ref. [693]. Copyright 2019, Springer Nature).

Bottom-up fabrication techniques for nanostructured and printed glucose sensors with MXenes and Au nanoparticles were developed for non-invasive blood glucose detection in tears [637] (Fig. 35a). Nanostructured biosensors like Cu@Au-GO nanoflowers-based sensors offer high sensitivity and stability, utilizing the porous Au matrix for enzyme immobilization [638] (Fig. 35b). Enzymatic glucose biosensors combine PDDA-capped Au nanoparticles with functionalized graphene and MWCNTs, achieving efficient electron transfer and high sensitivity [639] (Fig. 35c). Wearable glucose biosensors employ MoS₂ and Au nanofilms on polymer substrates, demonstrating high detection limits and flexibility [640] (Fig. 35d). Screen-printing processes using graphene ink create highly conductive and stable designs for flexible electronics, enhancing the sensitivity of glucose sensors [641] (Fig. 35e). 3D printing allows for the creation of complex geometries and intricate medical devices using 2D Ti₃C₂T_x inks [642] (Fig. 35f). Inkjet printing technologies, both drop-on-demand (DoD) and continuous inkjet (CIJ), are used to produce predesigned images by depositing ink droplets onto substrates. This method is scalable and cost-effective for printed electronics [643] (Fig. 35g). Printing with graphene ink demonstrates the potential for producing homogenous and well-covered products [644] (Fig. 35h).

Fig. 35.

Bottom-Up Fabrication for Nanostructured and Printed Glucose Sensors. (a) Filtration and self-assembled synthesis of 2D material glucose Biosensors. (b) Fabrication of a Cu@Au-GO nanoflowers-based electrochemical glucose nano-biosensor. (c) Schematic of enzymatic glucose biosensor fabrication: Preparation of the enzyme-loading hybrid nanocomposite via one-step co-assembly; fabrication process of the glucose biosensor via direct electrophoretic deposition (EPD) of nanocomposite onto a glassy carbon electrode (GCE) and subsequent photo-cross-linking. (d) Fabrication of a wearable glucose biosensor composed of a glucose oxidase (GO_x)/Au/MoS₂/Au nanofilm on a flexible polymer substrate. (e) Screen-printing process using graphene ink for flexible printed electronics. (f) 3D printing glucose biosensors. (g) Continuous inkjet and drop-on-demand inkjet printing piezoelectric material and bubbles introduction with a thermal element. (h) Inkjet printed Medal using graphene ink. (a) (Reproduced with permission from ref. [694]. Copyright 2022, Elsevier). (b) (Reproduced with permission from ref. [695]. Copyright 2020, Elsevier). (c) (Reproduced with permission from ref. [696]. Copyright 2019, Elsevier). (d) (Reproduced with permission from ref. [697]. Copyright 2019, Elsevier). (e) (Reproduced with permission from ref. [698]. Copyright 2019, American Chemica Society). (f) (Reproduced with permission from ref. [699]. Copyright 2016, RSC). (g) (Reproduced with permission from ref. [700]. Copyright 2018, RSC). (h) (Reproduced with permission from ref. [701]. Copyright 2017, Spinger Nature).

Various wearable glucose sensors have been developed for non-invasive monitoring. A system for cortisol and glucose detection on polyamide substrates addresses the relationship between cortisol and glucose metabolism, providing ultra-sensitive detection [645] (Fig. 36a, b). An iontophoretic skin-like patch integrates a paper battery for enhanced ISF glucose monitoring, showing high sensitivity and correlation with blood glucose levels [646] (Fig. 36c). Tattoo-based glucose monitoring offers continuous and practical alternatives to traditional methods [626] (Fig. 36d). Stretchable electrochemical sensor arrays combine CNTs with elastomeric materials for glucose sensing under strain, showcasing linear power density dependence on glucose concentration [627] (Fig. 36e). Single nanowire BFCs provide biochemical energy for in vivo devices, demonstrating high power outputs suitable for glucose sensors [628] (Fig. 36f). Pacifier and mouthguard-based sensors offer convenient, non-invasive glucose detection in saliva, demonstrating a strong correlation to blood glucose levels [629] (Fig. 36g). A wearable sensor array for multiplexed analysis of perspiration can detect multiple biomarkers, such as Na⁺, K⁺, lactate, and glucose [594] (Fig. 36h). Sweat-powered electronic skins enable multiplexed sensing with wireless data transmission, showcasing continuous operation capabilities [586,599,630] (Fig. 36i,j). Roll-to-roll printed sweat sensing patches provide high-yield, real-time glucose monitoring with microfluidic channels [631] (Fig. 36k). Bionic contact lenses and NovioSense tear glucose sensors offer convenient and non-invasive glucose monitoring methods [632,633] (Fig. 36l,m).

Fig. 36.

Other Wearable Glucose Sensors. (a,b) Cortisol and glucose detection system deposited on a polyamide substrate. (c) Depiction of an iontophoretic skin-like and paper battery biosensor for noninvasive glucose monitoring. (d) Tattoo-based noninvasive glucose monitoring. (e) Electrochemical sensor and biofuel cell array consisting of potentiometric ammonium sensors, amperometric glucose sensors, and enzymatic glucose biofuel cells. (f) Configuration of a single Nafion/poly(vinyl pyrrolidone) nanowire (NPNW)-based biofuel cell (BFC) containing GO_x and laccase in the anode and cathode regions, and the BFC immersed in the fuel solution. (g) Schematic of the assembled wireless glucose pacifier biosensor and mouthguard-based sensor. (h) Scheme of the sensor array (including glucose, lactate, sodium, potassium, and temperature sensors) for multiplexed perspiration analysis. (i) Photograph of the sweat-powered e-skin for multiplexed wireless sensing (top) and epidermal battery-free microfluidic system for simultaneous colorimetric, electrochemical, and volumetric sweat analysis (bottom). (j) Wearable sweat sensor. (k) Roll-to-roll printed sweat sensing patches, including serpentine and spiral electrode patterns. (l) A bionic contact lens including a biocathode, an anode, an interface chip, a glucose biosensor, a simple display, and an antenna. (m) NovioSense electrochemical tear glucose sensor. (n) Photograph of the mouthguard biosensor integrated with a wireless amperometric circuit board and reagent layer of the chemically modified printed Prussian blue carbon working electrode containing uricase for uric acid detection. (o) Wearable oral sensor for saliva glucose detection. (p) Dual glucose/insulin biosensor chip, with sensor configuration and localized detection of glucose and insulin on a single sensor, along with measurement intervals. (a,b) (Reproduced with permission from ref. [702]. Copyright 2017, Elsevier). (c) (Reproduced with permission from ref. [703]. Copyright 2017, Science). (d) (Reproduced with permission from ref. [625]. Copyright 2015, American Chemical Society). (e) (Reproduced with permission from ref. [626]. Copyright 2016, American Chemical Society). (f) (Reproduced with permission from ref. [627]. Copyright 2010, Wiley). (g) (Reproduced with permission from ref. [628]. Copyright 2019, American Chemical Society). (h) (Reproduced with permission from ref. [593]. Copyright 2016, Springer Nature). (i) (Reproduced with permission from ref. [598]. Copyright 2020, Science; reproduced with permission from ref. [629]. Copyright 2019, Science). (j) (Reproduced with permission from ref. [585]. Copyright 2017, National Acad Sciences). (k) (Reproduced with permission from ref. [630]. Copyright 2019, Science). (l) (Reproduced with permission from ref. [631]. Copyright 2013, American Chemical Society). (m) (Reproduced with permission from ref. [632]. Copyright 2018, American Chemical Society). (n) (Reproduced with permission from ref. [633]. Copyright 2015, Elsevier). (o) (Reproduced with permission from ref. [634]. Copyright 2016, Elsevier). (p) (Reproduced with permission from ref. [635]. Copyright 2019, Wiley).

A mouthguard biosensor for glucose detection in saliva incorporates Pt and Ag/AgCl electrodes and a wireless transmitter. This configuration enables telemetric measurement of salivary glucose, offering a noninvasive alternative to blood glucose testing [634] (Fig. 36n). Wearable oral sensors measure salivary glucose levels, providing continuous monitoring alternatives to traditional blood tests [635] (Fig. 36o). Dual glucose/insulin biosensor chips allow localized detection, offering comprehensive diabetes management [636] (Fig. 36p). Multianalyte detection platforms enhance disease diagnosis and management by measuring multiple biomarkers, providing a more complete understanding of patient health.

8.3.3. 2D materials-based glucose sensor applications

Glucose detection often employs glucose oxidase (GO_x) to catalyze the oxidation of glucose to gluconic acid and H₂O₂. This process can be harnessed in several ways: monitoring the consumption of O₂ or the production of H₂O₂, detecting electron exchange between enzymes and redox mediators, and monitoring electron exchange between embedded enzyme and electrodes through precisely designed nanostructures [637, 638]. For instance, reactive Pt nanoparticles are used to enhance the sensitivity of glucose detection by facilitating H₂O₂ oxidation, although this can lead to side reactions with substances like ascorbic acid and uric acid, reducing selectivity. Introducing Prussian blue mediator lowers the reaction overpotential and improves detection efficiency [639] (Fig. 37a). The effective gate voltage changes $\left({\Delta \text{V}}_{\text{G}}^{\text{eff}}\right)$ in response to various H₂O₂ concentrations were measured using graphene gate electrodes with and without Pt. The presence of Pt significantly enhances the sensitivity of the graphene gate electrodes towards H₂O₂, indicating the potential for improved glucose detection in practical applications [639] (Fig. 37b). CV studies of TiO₂ and GO_x-modified glassy carbon electrodes in phosphate-buffered saline (PBS) solution demonstrate the electrochemical activity of these electrodes in glucose detection. The GO_x-modified electrodes show distinct redox peaks corresponding to the enzymatic oxidation of glucose, highlighting their effectiveness in glucose sensing applications [640] (Fig. 37c). A comparative analysis of the glucose sensing performance of pristine MoS₂, MoS₂/GO_x, and MoS₂/Au/GO_x electrodes reveals significant improvements with the addition of GO_x and Au nanoparticles. The MoS₂/Au/GO_x electrodes exhibit superior sensitivity and response time, demonstrating the synergistic effects of combining MoS₂ with Au nanoparticles and GO_x for enhanced glucose detection [641] (Fig. 37d). The amperometric response of NiO/MoSe₂ modified glassy carbon electrodes (GCE) shows a linear increase in current with successive additions of glucose, indicating efficient glucose oxidation. This demonstrates the high electro-catalytic activity of NiO/MoSe₂ for non-enzymatic glucose sensing, providing a promising approach for practical applications [642] (Fig. 37e).

Fig. 37.

2D Materials-based Glucose Sensing Performance. (a) Schematic of generations of GO_x-based sensors. (b) Effective gate voltage change $\left({\Delta \text{V}}_{\text{G}}^{\text{eff}}\right)$ versus H₂O₂ concentrations with or without Pt on graphene gate electrodes. (c) Cyclic voltammograms of TiO₂, GO_x-modified glassy carbon electrodes in phosphate-buffered saline (PBS) solution (pH 7.4). (d) Response comparison between pristine MoS₂, MoS₂/GO_x, and MoS₂/Au/GO_x electrodes. (e) Amperometric curve of NiO/MoSe₂/GCE with successive glucose additions and oxidation process. (f) Sensing mechanism of flavin adenine dinucleotide (FAD)/pyrroloquinoline quinone (PQQ)-glucose dehydrogenase (GDH) and nicotinamide adenine dinucleotide phosphate (NAD(P))-GDH glucose sensors. (g) Graphene field-effect transistor (FET) conductance versus electrolyte gate voltage in pH 6, 7, 8, and 9 solutions. (h) Real-time pH detection by conductance changes with electrolyte gate potential. (a,b) (Reproduced with permission from ref. [638]. Copyright 2015, Springer Nature). (c) (Reproduced with permission from ref. [639]. Copyright 2011, American Chemical Society). (d) Reproduced with permission from ref. [640]. Copyright 2017, Elsevier). (e) (Reproduced with permission from ref. [641]. Copyright 2021, Springer Nature). (f) (Reproduced with permission from ref. [642]. Copyright 2018, Wiley). (g,h) (Reproduced with permission from ref. [643]. Copyright 2010, American Chemical Society).

The sensing mechanisms of FAD/PQQ-GDH and NAD(P)-GDH glucose sensors involve the catalytic oxidation of glucose by these dehydrogenase enzymes. These sensors exploit the specific interactions between glucose and the enzymes, facilitating electron transfer and enabling accurate glucose detection in various conditions [643] (Fig. 37 f). The conductance of graphene FETs based on the electrolyte gate voltage is shown for solutions with different pH levels. The Dirac point shifts positively with increasing pH, demonstrating the sensitivity of graphene FETs to changes in the chemical potential of the solution, making them suitable for chemical sensing applications [644] (Fig. 37 g). Real-time pH detection using graphene FETs involves monitoring conductance changes as the electrolyte gate potential is varied. The ability to detect pH changes in real-time confirms the viability of graphene FETs as direct chemical sensors, capable of providing accurate and responsive measurements in various chemical environments (Fig. 37 h).

8.3.4. Challenges and perspectives of glucose sensor

One of the key challenges is ensuring the stability, biocompatibility, and accuracy of glucose sensors during real-world use. While nanomaterials have facilitated electron transfer between the electrode and the enzyme’s redox center, progress in developing mediator-free glucose biosensors has been limited. Future efforts should focus on enhancing the stability of these sensors and addressing the requirements of real sample analysis and biocompatibility for glucose sensing.

The incorporation of electrochemical glucose sensors with flexible biosensors has resulted in the development of painless and non-invasive methods for measuring glucose in ISF, sweat, saliva, or tears. These systems, however, require further optimization and large-scale validation to ensure reliability and accuracy. Challenges such as surface contamination, inconsistent liquid extraction, and the physiological impacts like pH and temperature on monitoring accuracy need to be addressed.

Multiplexed assays capable of detecting various biomarkers about diabetes, including glucagon, ketone bodies, and insulin, could offer a more comprehensive insight into a patient’s health. However, integrating different surface chemistries and detection principles in a single system remains challenging. Additionally, developing flexible glucose sensors that can measure various biomarkers while avoiding cross-talk between analytes is a critical area for future research. Providing reliable energy sources for these biosensors is another challenge. While enzymatic biofuel cells show promise by harvesting energy from body metabolites, their power efficiency remains limited. Advancements in flexible, printable, and ultra-thin battery technologies could offer a viable solution [645,646].

8.4. Advanced electrochemical biosensing technologies

8.4.1. Graphene-based CRISPR chips

Single-nucleotide polymorphisms (SNP) account for over 50% of disease-causing mutations in humans. SNP play a crucial role not only in human health genetics but also in infectious disease prevention, aging, pharmacology, and agriculture [647,648]. Specific SNP have been linked to reduced effectiveness of the rubella vaccine by affecting key cytokine pathways and have been implicated in outbreaks of severe acute respiratory syndrome (SARS). Additionally, SNP have been associated with the hallmarks of human aging, drug metabolism, and crop resistance in agriculture.

Traditional methods for SNP genotyping, such as SNP microarrays, TaqMan SNP genotyping, or next-generation sequencing, although high throughput, require DNA amplification, experienced technical staff, and sophisticated optical equipment [649,650]. These requirements limit their widespread use outside traditional laboratory settings, making mass testing at the point of care or in the field challenging. Combining graphene with CRISPR technology, which offers precise and efficient gene-editing capabilities [651], opens new avenues for SNP detection.

The CRISPR-chip is fabricated through several stages using conventional microelectromechanical systems processing. After fabrication and packaging, the chips are cleaned and non-covalently functionalized with the molecular linker, 1-pyrenebutanoic acid (PBA), via π–π aromatic stacking. The graphene surface channel is then functionalized with dCas9 using carbodiimide crosslinking chemistry. To prevent non-specific adsorption of charged molecules, the surface is blocked with amino-PEG5-alcohol and ethanolamine hydrochloride. Finally, the immobilized dCas9 is complexed with a single-guide RNA (sgRNA) specific to a DNA target, forming the anchored dRNP complex [652] (Fig. 38a).

Fig. 38.

Advanced Biosensing Technologies Using Graphene and CRISPR. (a) Schematic of CRISPR-chip functionalization. (b) Current response of bfp-targeting CRISPR-chip with bfp PCR product, higher than Scram-targeting CRISPR-chip. (c) Sensing mechanism of molecular electrostatic microsystem (MolEMS): Probes on cantilevers recognize targets, and electrostatic actuation detects recognition events in gFET channel. (d) Current response at different concentrations of ATP, Hg²⁺, ATP, and single-stranded DNA (ss-DNA-T) i under electrostatic actuation of MolEMS with probes. (e) Comparison of MolEMS with SARS-CoV-2 nucleic acid detection methods: quantitative reverse transcription PCR (qRT-PCR), US CDC or China-NMPA-approved qRT-PCR, reverse transcription loop-mediated isothermal amplification (RT-LAMP), clustered regularly interspaced short palindromic repeats (CRISPR), recombinase polymerase amplification (RPA), surface plasmon resonance (SPR), and electrochemical (EC) methods. (f) Different surface functionalization membranes for multiplexed measurements. (g) Schematic of an individual sensing unit with ion-sensitive membrane (ISM). Electrostatic potential as a function of distance from graphene surface. ISM: ion-sensitive membrane; VM: membrane potential; VGS: gate-to-source voltage; VDS: drain-to-source voltage; VS: source potential. (h) Diffraction signal generated by periodic variation of optical susceptibility at the interface, influenced by adsorbed molecules like cetyltrimethylammonium bromide (CTAB) and species in the electrolyte double layer induced by graphene grating. (i) Diffraction spectra near anti-symmetric CH2 vibration during voltage scanning. (a,b) (Reproduced with permission from ref. [651]. Copyright 2019 Springer Nature). (c-e) (Reproduced with permission from ref. [652]. Copyright 2022 Springer Nature). (f,g) (Reproduced with permission from ref. [653]. Copyright 2022 Springer Nature). (h,i) (Reproduced with permission from ref. [654]. Copyright 2015 Springer Nature).

The current response of the CRISPR-chip functionalized with dRNPs targeting the bfp sequence shows a significantly larger signal compared to the non-specific dRNP-Scram-functionalized CRISPR-chips, indicating the specificity of the CRISPR-chip’s signal output to the immobilized dRNP complex. Real-time monitoring demonstrates a fast response time within 2.5 minutes in the presence of the target dsDNA, confirming the rapid and specific detection capabilities of the CRISPR-chip [652] (Fig. 38b).

The Molecular Electrostatic Microsystem (MolEMS) gFETs employ probes on cantilevers that recognize specific targets. Upon recognition, electrostatic actuation transfers the recognition events to the gFET channel, leading to efficient biorecognition and signal transduction. MolEMS achieves high sensitivity while maintaining excellent anti-fouling properties, crucial for operation in challenging biological environments such as full serum [653] (Fig. 38c). MolEMS demonstrates the ability to detect various analytes, including ATP, Hg²⁺, and single-stranded DNA (ss-DNA-T), at extremely low concentrations (down to 5 × 10−²⁰ M). This high sensitivity is realized by the high density of active arrays and the efficient signal transduction mechanism of the gFET channel under electrostatic actuation [653] (Fig. 38d).

MolEMS gFETs offer advantages such as rapid detection, easy operation, high sensitivity, specificity, and portability, making them competitive with other COVID-19 detection methods, including qRT-PCR, RT-LAMP, CRISPR, RPA, SPR, and electrochemical methods [653] (Fig. 38e). Different surface functionalization membranes are employed for multiplexed measurements, enhancing the versatility and applicability of the sensing technology in detecting multiple analytes simultaneously [654] (Fig. 38f).

The operating principle of an individual sensing unit with an ion-sensitive membrane (ISM) is based on the channel modulation of the graphene electrolyte-gated FET as cations diffuse into the membrane. The Nernst equation governs the ion transport through the interface, with the potential increasing with higher ion concentrations, leading to a more n-doped graphene channel and a leftward shift of the I-V characteristic [654] (Fig. 38g). The novel diffraction spectroscopy technique utilizes graphene grating electrodes to achieve high detection sensitivity and interfacial specificity. The diffraction signal originates from the periodic variation of optical susceptibility at the interface, influenced by adsorbed molecules like cetyltrimethylammonium bromide (CTAB) and species in the electrolyte double layer induced by the graphene electrode [655] (Fig. 38h).

A wireless graphene sensor designed for detecting whole-cell bacterial is demonstrated by monitoring the tooth surface. The sensor, modified with antimicrobial peptides (AMPs) and combined with a coil antenna, detects bacteria such as Helicobacter pylori with high sensitivity. This portable, wire-free sensor offers continuous, non-destructive analysis, highlighting its potential for real-time monitoring applications [655] (Fig. 38h). The bias-dependent diffraction intensity at 3,000 cm⁻¹ is observed during voltage scanning, providing insights into the electrochemical deposition and dissolution processes at the electrolyte/electrode interface. This approach enhances the understanding of interfacial phenomena and supports the development of advanced sensing technologies [655] (Fig. 38i).

In measuring of dopamine, the voltage is applied to the sensor electrode, which involves the oxidation of dopamine to dopamine-o-quinone, and its reduction back to dopamine [656] (Fig. 39a). The SNP-chip is an advanced graphene FET designed to detect SNP in unamplified DNA samples. This system integrates a reader and a cartridge connected to a computer for analysis. The graphene channel, situated between the source and drain electrodes, is functionalized with Cas enzymes using a chemical linker, PBA. This linker π–π stacks with graphene and is chemically activated to attach the Cas enzyme covalently. After enzyme immobilization, the graphene surface is passivated with poly(ethylene glycol) (PEG) amine and ethanolamine. The Cas enzyme forms a complex with guide RNA (gRNA) designed to target specific DNA loci. Upon incubation of DNA samples atop the graphene surface, the Cas–gRNA complex binds to target sequences, anchoring the DNA to the graphene. Mismatched DNA sequences, however, dissociate due to lower affinity. The device is rinsed to remove non-specifically bound DNA, allowing for accurate signal read-out [657] (Fig. 39b, c).

Fig. 39.

2D Materials based Wireless Sensing (a) Dopamine detection principle on graphene electrodes. (b,c) Single-nucleotide polymorphisms (SNP)-chip, a CRISPR-powered graphene field-effect transistor (gFET), detects SNP in unamplified DNA samples. (d) Schematic of a graphene-based electrochemical device with thermoresponsive microneedles for diabetes monitoring and therapy, providing sweat control. (e) Generic miniature pixel for glucose detection: Electrodes 1 (Ag/AgCl) and 4 (Ag) extract glucose. Electrodes 1 (Ag/AgCl), 2 (Pt nanoparticle-decorated graphene), and 3 (Pt) detect glucose electrochemically. Pt on graphene enhance signal. Electrodes contact an enzyme-encasing gel reservoir, defining the pixel active area. (f) Passive wireless telemetry system with a planar meander line inductor and interdigitated capacitive electrodes on a graphene/silk film. (g) Schematic of phase transition process of electrochemistry sensors and their SEM images. When detecting designated chemical compounds, there exists a phase transition (h,i) Graphene printed on bioresorbable silk forms contacts with a wireless coil. The nanosensing architecture is transferred onto a tooth’s surface. The magnified sensing element shows wireless readout and pathogenic bacteria binding by peptides on the graphene nanotransducer. (a) (Reproduced with permission from ref. [655]. Copyright 2019, Wiley). (b,c) (Reproduced with permission from ref. [656]. Copyright 2021, Springer Nature). (d) (Reproduced with permission from ref. [602]. Copyright 2016, Springer Nature). (e) (Reproduced with permission from ref. [660]. Copyright 2018, Springer Nature). (f) (Reproduced with permission from ref. [661]. Copyright 2012, Springer Nature). (g) (Reproduced with permission from ref. [704]. Copyright 2015, Springer Nature, reproduced with permission from ref. [705]. Copyright 2011, American Chemical Society). (h,i) (Reproduced with permission from ref. [661]. Copyright 2012, Springer Nature).

These advancements in graphene and CRISPR-based biosensing technologies demonstrate significant progress in noninvasive, continuous, and real-time monitoring. By integrating graphene with CRISPR technology, the SNP-chip exemplifies the potential for amplification-free electronic detection of target genes with SNP specificity.

8.4.2. 2D Material-based wireless sensing

2D materials have become transformative candidates for advanced wireless biosensing technologies owing to their superior electrical, thermal, and mechanical characteristics [658–660]. The flexible Au-doped graphene electrodes on an Au mesh create an efficient electrochemical interface. This patch monitors glucose and pH levels in sweat, detecting hypoglycemia-induced vibrations and delivering drugs when necessary. Temperature sensors activate hyperglycemia drugs, ensuring precise management of blood glucose levels [603] (Fig. 39d).

A miniature pixel designed for glucose detection consists of a glucose oxidase-bearing hydrogel reservoir, an electrochemical glucose sensor (graphene-based film decorated with Pt nanoparticles), and miniaturized electrodes on a flexible substrate. Glucose extracted transdermally reacts with glucose oxidase, producing H₂O₂, which is detected electrochemically. The active area of the pixel ensures single follicular hits, maximizing glucose detection efficiency. The pixel array guarantees redundancy, enhancing reliability [661] (Fig. 39e).

This system integrates large-area graphene monolayers with water-soluble silk fibroin films using a transfer printing process. Electrode patterns are incorporated via shadow mask-assisted electron beam evaporation of Au. The architecture is composed of a resonant LRC parallel circuit with an inductive coil for wireless transmission and interdigitated capacitors contacting graphene-based resistive sensors. This passive wireless telemetry system operates without onboard power sources and external connections, utilizing a full-wave electromagnetic simulation tool for design [662] (Fig. 39 f).

The process of 3D printing GO inks to produce a 3D graphene aerogel structure presents several challenges. Aerogels, which are ultra-low-density porous solids, are created by replacing the liquid in the pores of a wet gel with air. For the GO structure to become an aerogel, the GO ink should remain wet through printing and gelation so that the liquid in the gel can be removed via supercritical- or freeze-drying, avoiding gel collapse due to capillary forces. This requires printing the GO ink into a bath of liquid that is less dense than water and immiscible with aqueous GO inks. The fabrication scheme involves combining a GO suspension with a silica filler to create a homogeneous, highly viscous, and thixotropic ink. This ink is then extruded through a micronozzle to form 3D structures, which are printed in an isooctane bath to prevent drying and pore collapse [663,664] (Fig. 39 g). Standard literature methods are then used to process the printed structures, followed by silica filler etching to obtain periodic 3D graphene aerogel microlattices.

The sensor design involves graphene printed on bioresorbable silk, forming contacts with a wireless coil. This nanosensing architecture is transferred onto the tooth’s surface, with the magnified sensing element showing wireless readout and pathogenic bacteria binding by peptides on the graphene nanotransducer. The ultra-thin sensors biotransfer from the silk platform to biomaterials, achieving high adhesive conformability and specificity in biological recognition. The system operates battery-free with remote wireless sensing capability, modulating the electrical conductivity of the graphene film upon target recognition [662] (Fig. 39 h,i).

9. Challenges and perspectives

9.1. Challenges

9.1.1. Performance degradation

Flexible devices utilizing 2D materials often exhibit degraded performance compared to their rigid counterparts due to defects, grain boundaries, and impurities. The high surface roughness of flexible substrates like PET and PI increases channel scattering, leading to reduced carrier mobility and lower on-current values. Additionally, the poor water vapor transmission ratio (WVTR) of these substrates causes instability and larger hysteresis in devices. To mitigate these issues, thin oxide buffer layers such as Al₂O₃ can be employed, but further advancements are required to improve the stability and performance of 2D materials-based devices.

9.1.2. Miniaturization

Scaling down biosensors while maintaining performance is challenging due to high contact resistance, difficulties in dielectric deposition, and surface defects. Advanced fabrication techniques and material engineering are necessary to address these issues and achieve effective miniaturization. This includes the development of novel nanofabrication methods and precise control over material properties at the nanoscale to ensure reliable performance in smaller devices.

9.1.3. Mechanical failure

Wearable electronics face mechanical failure due to repeated bending and stretching, which induces strain and friction that can create defects in 2D materials [663,664]. Effective strain management strategies, including the development of appropriate device architectures and the use of encapsulating agents, are crucial to enhance mechanical durability and ensure long-term functionality.

9.1.4. Motion artifact and signal interference

Wearable biosensors are prone to motion artifacts and signal interference, compromising accuracy and reliability. Robust sensor designs and advanced signal processing algorithms are essential to mitigate these issues. Techniques such as real-time signal processing, adaptive filtering, and machine learning algorithms can help reduce noise and improve signal accuracy. Algorithms like Kalman filters and wavelet transforms are effective in filtering out motion artifacts, while convolutional neural networks (CNNs) and support vector machines (SVM) can identify patterns and enhance signal classification [665–669].

9.1.5. Power management

Efficient power management is vital for wearable biosensors. Traditional batteries are bulky and not suitable for wearable applications[670,671]. The development of flexible and stretchable devices, integration of energy harvesting technologies, and energy-efficient data collection methods are necessary to ensure continuous and reliable operation. Innovations in low-power electronics, such as subthreshold operation and dynamic voltage and frequency scaling (DVFS), can significantly reduce power consumption. Additionally, the utilization of tunnel FETs (TFETs) and the incorporation of energy harvesting technologies like TENGs and piezoelectric materials can provide self-sustaining power sources for wearable sensors [672–677].

9.2. Perspectives

9.2.1. Durability

Enhancing the durability of 2D materials-based wearable biosensors is crucial for ensuring their long-term functionality. Advancements in material synthesis, such as defect engineering and doping, can significantly improve the structural integrity of these materials. Additionally, novel device fabrication techniques, including encapsulation with robust yet flexible coatings, can protect the sensors from environmental factors such as moisture, sweat, and mechanical abrasion. Implementing self-healing polymers in the sensor design can further enhance durability by allowing the material to recover from minor damages, ensuring prolonged use in various conditions.

9.2.2. Portability

Wearable biosensors should be lightweight and portable. Miniaturization through advanced fabrication methods like photolithography and nanoprinting can significantly reduce the size of these devices. Integration with ultra-thin, flexible substrates such as PI and PDMS can enhance the portability, making them convenient for everyday use. Embedding flexible printed circuit boards (FPCBs) within the sensor design can further contribute to the miniaturization and portability of these devices, allowing seamless integration into wearable forms such as patches, wristbands, and clothing.

9.2.3. Biocompatibility

Ensuring biocompatibility is crucial for the prolonged use of wearable biosensors. Using biocompatible materials such as biopolymer coatings and hydrogels can prevent adverse reactions when the sensors are in contact with the skin. Surface modification techniques, such as plasma treatment and chemical functionalization, can improve the biocompatibility of 2D materials.

9.2.4. Flexibility and stretchability

The flexibility and stretchability of 2D materials-based wearable biosensors represent a significant leap forward in personalized healthcare technology. By seamlessly integrating with the human body, these sensors offer real-time, non-invasive monitoring capabilities, enhancing both the user’s experience and the accuracy of health data collection.

9.2.5. Sensitivity and specificity

Achieving high sensitivity and specificity is paramount for the accurate performance of wearable biosensors. The unique electrical and electrochemical properties of 2D materials, such as TMDs, and MXenes, significantly enhance the sensitivity of these sensors [678]. Functionalizing 2D materials with specific binding sites or antibodies further improves the specificity of the sensors. By targeting specific biomolecules, these functionalized 2D materials can distinguish between different analytes with high precision, reducing false positives and negatives. This functionalization enables the development of biosensors that can reliably monitor various health indicators in real-time [689, 706].

9.2.6. Energy efficiency

Energy-efficient wearable biosensors are crucial for extending battery life and reducing the need for frequent recharging. Advancements in low-power electronics, such as the development of energy-efficient transistors and circuits, are crucial for enhancing the energy efficiency of these devices. One method to achieve energy efficiency in wearable biosensors is with energy-efficient transistors. Subthreshold operation, where transistors operate in the subthreshold region below the threshold voltage for strong inversion, significantly reduces power consumption. This approach is useful for ultra-low-power applications. Additionally, Tunnel FETs leverage quantum tunneling to achieve lower subthreshold swings compared to traditional MOSFETs, resulting in reduced power consumption and enhanced energy efficiency. The utilization of advanced complementary metal-oxide-semiconductor (CMOS) technologies, with low leakage currents and optimized gate oxides, further improves the energy efficiency of transistors.

Low-power circuit design strategies also contribute to energy efficiency. Dynamic voltage and frequency scaling (DVFS) optimizes power consumption by adjusting the voltage and frequency according to the computational load. DVFS techniques dynamically scale down power usage during periods of low activity. Asynchronous circuits, which do not rely on a global clock signal, reduce the power consumed by clock distribution and synchronization, leading to significant energy savings.

Energy harvesting and storage are vital for maintaining a consistent power supply for wearable biosensors [679,680]. TENGs convert mechanical energy from body movements into electricity, providing a self-sustaining power source. Piezoelectric materials, which generate electric charge in response to mechanical stress, offer another method for energy harvesting from motion or vibrations [681–686]. Advanced energy storage solutions, such as supercapacitors and thin-film flexible batteries, efficiently store harvested energy, ensuring a consistent power supply for the sensors.

Implementing sleep modes and power gating in sensor circuits further reduces power consumption. Sleep modes allow sensors to wake up periodically to take measurements and then return to sleep mode, conserving energy during inactive periods. Power gating involves shutting off power to certain parts of the circuit when they are not in use, reducing leakage currents and overall power consumption.

9.2.7. Multifunction

Developing multifunctional biosensors capable of detecting multiple analytes simultaneously offers the potential for comprehensive health monitoring. This multifunctionality can be realized through innovative material design, such as the creation of heterostructures that combine different 2D materials with complementary properties. Advanced device architectures, including microfluidic integration, can facilitate the simultaneous detection and analysis of various biomarkers.

9.2.8. AI/ML integration, and real-time monitoring

AI and machine learning (ML) algorithms play a pivotal role in improving the accuracy and efficiency of biosensing systems. A key application is noise filtering, which utilizes various advanced techniques. Kalman filters estimate the state of a dynamic system from incomplete and noisy measurements, making them ideal for real-time smoothing and signal prediction. The wavelet transform decomposes sensor signals into different frequency components, effectively isolating and removing noise while preserving important signal features. Principal Component Analysis (PCA) reduces the dimensionality of sensor data, filtering out noise by retaining only the most significant components that capture the underlying data patterns.

In pattern recognition, Support Vector Machines (SVM) classify sensor data by finding the optimal hyperplane that separates different classes of signals, proving effective in distinguishing various physiological signals. Convolutional Neural Networks (CNNs), well-suited for analyzing image-like sensor data, automatically identify relevant features and patterns, making them valuable for complex biosensing applications. Recurrent Neural Networks (RNNs), and Long Short-Term Memory (LSTM) networks, are used for time-series data analysis. They capture temporal dependencies and patterns in continuous sensor signals, thereby improving biosensing accuracy.

For real-time monitoring, signal processing frameworks such as Apache Flink and TensorFlow Lite handle large volumes of data with low latency, enabling immediate analysis and feedback. Edge computing, by deploying AI/ML models on edge devices, ensures that data processing and analysis occur locally on the sensor device. This approach reduces the need for data transmission to centralized servers, improving real-time responsiveness. Additionally, anomaly detection algorithms, like Isolation Forests and Autoencoders, identify deviations from normal physiological patterns, triggering real-time alerts for abnormal conditions (Fig. 40).

Fig. 40.

Challenges and Perspectives of 2D Materials enabled Wearable Biomonitoring. Challenges in current 2D materials-based wearable biomonitoring include measurement reliability, mechanical robustness, miniaturization, motion artifacts, signal interference, systematic integration, wearing comfort, and effective power management. Future advancements in the field should focus on enhancing device durability, leveraging AI and machine learning, improving sensitivity and selectivity, expanding functionalities, and enhancing biocompatibility and energy efficiency. These efforts will contribute to making wearable biomonitoring solutions more reliable and clinically viable for continuous health monitoring.

10. Conclusion

The realm of 2D materials has unveiled numerous possibilities for developing advanced composites with superior mechanical strength and multifunctional properties. Despite considerable progress, the practical application of these materials in wearable biosensors remains a challenge. Recent innovations in manufacturing techniques offer valuable insights into the structured assembly of 2D nanosheets and their composites, paving the way for more effective integration into wearable technologies. Techniques like liquid phase exfoliation and chemical vapor deposition (CVD) produce high-quality 2D materials with uniform thickness and large surface areas. This increases the active sites available for biomolecule interactions, leading to improved sensitivity in detecting low concentrations of analytes.

Key parameters influencing composite performance operate across multiple scales. At the nanoscale, the properties of 2D nanosheets—such as their anisotropic structures, defects, functionalization, and layer number—play a pivotal role and are strongly affected by synthesis methods like liquid-phase exfoliation and CVD. At the microscale, advanced assembly techniques ensure uniform dispersion and effective interconnection of 2D nanofillers, creating highly porous cellular structures that enhance interfacial properties and percolative characteristics. At the macroscale, alignment and filler content are crucial in determining the bulk properties of composites. Integrating these multiscale factors not only translates the remarkable properties of 2D nanosheets into composites but also imparts multifunctional characteristics without sacrificing mechanical integrity. This multifunctionality opens new possibilities for applications in flexible electronics, energy storage, and conversion.

In summary, despite challenges in wearable biosensors and flexible electronics, advancements in materials science and fabrication offer promising solutions. By improving durability, portability, biocompatibility, flexibility, and sensitivity, the next generation of wearable devices will meet health monitoring needs and enable new applications, leading to more reliable, efficient, and user-friendly technologies that enhance health outcomes and quality of life.

Biography

Dr. Jun Chen is currently an associate professor with tenure in the Department of Bioengineering at the University of California, Los Angeles (UCLA). His research focuses on soft matter innovation for healthcare and energy. He has published two books and 380 journal articles, with 280 of them being corresponding authors in Nature Reviews Bioengineering (1), Nature Materials (2), Nature Electronics (10), Nature Biotechnology (2), Nature Biomedical Engineering (1), Nature Communications (10), Science Advances (6), Chemical Reviews (2), Chemical Society Reviews (2), Joule (3), Matter (20), among others. He also filed 14 US patents, including one licensed. With a current h-index of 125, Dr. Chen was identified to be one of the world’s most influential researchers in the field of Materials Science on the Web of Science. Among his many accolades are the V. M. Watanabe Excellence in Research Award (1 faculty per year in UCLA Samueli School of Engineering), Shu Chien Early Career Award, MRS Outstanding Early Career Investigator Award, Stephanie L Kwolek Prize, Asian American Academy of Science and Engineering (AAASE) Rising Star Award, ASME Rising Star of Mechanical Engineering, BMES CMBE Rising Star Award, UCLA Faculty Mentor Award, UCLA Society of Hellman Fellows Award, Georgia Tech Alumni 40 Under 40, ONR Young Investigator Award, AHA Innovative Project Award, AHA Transformational Project Award, AHA’s Second Century Early Faculty Independence Award, NIH UCLA CTSI KL2 Translational Science Award, BBRF Young Investigator Award, Okawa Foundation Research Award, Advanced Materials Rising Star, Materials Today Rising Star Award, ACS Nano Rising Stars Lectureship Award, Chem. Soc. Rev. Emerging Investigator Award, Nano Research Young Innovator Award, ACS PMSE Young Investigator Award, Clarivate Highly Cited Researchers 2019/2020/2021/2022/2023/2024/2025, among others. Beyond his research activities, Dr. Chen serves as an associate editor for Biosensors and Bioelectronics, FlexMat, Soft Science, Med-X, cMat, and Textiles. Additionally, he is a member of the advisory and editorial boards of over 15 journals, including Matter, Materials Today, Materials Today Energy, Cell Reports Physical Science, The Innovation, Biomedical Technology, among others.

Data availability

No data was used for the research described in the article.