Advances and Prospects of Chemiresistive Breath Humidity Sensors

Abstract

Chemiresistive breath humidity sensors (CRBHSs) have emerged as a promising technology for non-invasive health monitoring, offering high sensitivity, a simple device architecture, strong miniaturization potential, and low power consumption. This review summarizes recent progress in CRBHSs from three core perspectives: sensing mechanisms, material systems, and device applications. First, we outline the fundamental sensing principles, emphasizing the Grotthuss proton-hopping mechanism and the resistance modulation associated with water adsorption/desorption. Next, we discuss structural engineering strategies for zero-dimensional (0D), one-dimensional (1D), two-dimensional (2D), and three-dimensional (3D) sensing materials, highlighting how dimensional design can balance water uptake, charge transport, mechanical compliance, and wearability. Finally, we review representative applications ranging from healthcare diagnostics and respiratory monitoring to emotion- and behavior-related assessment. Overall, this review integrates the mechanism–material–application relationship to provide a cohesive understanding of CRBHSs; identifies key challenges such as environmental stability and anti-interference performance; and outlines future directions, including performance optimization, flexible/wearable integration, and intelligent sensor systems.

1. Introduction

Breathing is one of the most fundamental physiological activities in the human body [1,2]. It not only enables gas exchange, i.e., taking in oxygen (O₂) [3] and expelling carbon dioxide (CO₂) [4], but also carries a wealth of accompanying information. Exhaled breath is characterized by physical parameters such as humidity, temperature, and airflow rate, as well as chemical constituents, including volatile organic compounds (VOCs), ammonia (NH₃), and CO₂ [5,6,7]. Together, these features provide non-invasive “signals” that can reflect an individual’s health status and support convenient, real-time assessment [8,9]. Among these multidimensional indicators, breath humidity is particularly attractive for clinical monitoring because it can be measured readily and shows strong associations with physiological conditions. Variations in exhaled humidity can directly indicate respiratory system function and may also correlate with metabolic state and disease progression [10]. Accordingly, accurate and real-time monitoring of breath humidity is important for early warning of potential abnormalities, tracking clinical status, supporting the auxiliary assessment of breath-related disorders, and enabling daily health management.

For decades, breath humidity has been measured primarily using laboratory-grade instruments such as dew-point meters [11] and precision hygrometers [12]. Although these systems can deliver high accuracy, they are typically bulky, expensive, and require specialized operation. In addition, their use is often constrained by testing conditions and instrumentation requirements, making real-time and continuous monitoring during natural breathing difficult [13,14]. As a result, they are not well suited for bedside, dynamic tracking in clinical settings, nor for portable, everyday self-monitoring, factors that have limited the broader adoption of breath humidity monitoring technologies [15,16]. In recent years, rapid advances in flexible electronics, micro-/nanofabrication, and low-power sensing have accelerated the development of flexible, miniaturized wearable devices, opening new opportunities for portable breath humidity monitoring [17,18,19,20]. Within this landscape, chemiresistive sensors have attracted considerable attention because they can respond sensitively to humidity variations while offering a simple device architecture, facile miniaturization and integration, and low power consumption [21,22,23,24]. Recent studies have further demonstrated that humidity, as an important physiological signal, can play a key role in supporting breath analysis and diagnosis [25,26]. Such inherent strengths align well with the flexibility, compactness, and long-term wearability required for wearable platforms, making chemiresistive breath humidity sensors (CRBHSs) strong candidates for next-generation non-invasive breath monitoring.

In recent years, human-centered humidity sensing has become an active topic in the review literature, and a number of articles have summarized progress from different angles. For example, some reviews emphasize application scenarios, highlighting advances in breath behavior monitoring, speech recognition, and skin moisture detection [27]. With the rapid development of flexible electronics, several comprehensive reviews have also examined flexible humidity sensors based on nanomaterials such as carbon materials, metal sulfides, and metal oxides [28,29,30,31]. Other breath analysis studies have discussed coupling humidity sensors with gas sensors for disease screening and noted the broader move toward flexible sensing platforms [10,32]. In addition, reviews focused on key sensing materials for breath-related detection have discussed material design and optimization strategies for chemiresistive sensors, particularly those based on semiconductors and polymers [33,34].

Although these reviews provide valuable perspectives on breath-related sensing, including application contexts, material systems, and sensing mechanisms, most place greater emphasis on multifunctional humidity sensing or on detecting disease biomarkers using gas sensors. To the best of our knowledge, a dedicated and systematic review of chemiresistive sensors specifically designed for breath humidity monitoring is still lacking. Accordingly, this review focuses on CRBHSs and summarizes the latest progress from three core perspectives: sensing mechanisms, sensing materials, and device-level/integrated applications, while also discussing current challenges and future development opportunities.

2. Sensing Mechanisms of CRBHSs

2.1. Fundamental Principles of Chemiresistive Humidity Detection

Humidity sensing in chemiresistive devices is commonly interpreted using the Grotthuss mechanism [35,36], which describes proton transport through a hydrogen-bonded water network. In a simplified form, this process can be represented as proton transfer between neighboring water molecules, as expressed in Equation (1).

H₂O + H₃O⁺ = H₃O⁺ + H₂O

(1)

The sensing response originates from dynamic charge-transport processes that occur when the active material interacts with water molecules under a given relative humidity (RH). In general, the signal magnitude increases with the amount of water adsorbed on the surface; therefore, the surface hydrophilicity of the sensing layer is a key determinant of performance. Water adsorption typically proceeds in two stages. At low RH, water molecules chemically adsorb onto hydroxyl groups or defect sites on the material surface, forming strong interactions that establish a primary adsorption layer [37]. As RH increases, additional molecules physically adsorb onto this primary layer, building up multilayer water films. Under an applied electric field, the adsorbed water can partially ionize to produce mobile protons (H⁺). These protons then migrate through the hydrogen-bond network via the Grotthuss proton-hopping process, markedly improving charge transport and carrier mobility [38]. As a result, the device exhibits an enhanced electrical response, enabling efficient transduction of humidity variations into measurable resistance (or conductance) changes.

For chemiresistive humidity sensors, the operating principle relies on monitoring resistance changes induced by interactions between the sensing layer and water vapor. Specifically, H₂O molecules adsorb onto the active surface and undergo physical and/or chemical interactions with the material, leading to changes in key electronic parameters, such as carrier concentration, carrier mobility, and/or interfacial barrier height, which in turn produce a measurable resistance (or conductance) variation [39]. Accordingly, the response of a chemiresistive humidity sensor can be defined by [28]:

Response = |(R − R₀)/R₀| × 100%

(2)

Its sensitivity can be calculated using Equation (3):

Sensitivity = |(R − R₀)/R₀|/(Δ%RH) × 100%

(3)

where R is the resistance measured at a given relative humidity under the test conditions and R₀ is the resistance measured in dry air. In general, the response describes the relative change in resistance with respect to the initial resistance before and after exposure, whereas sensitivity quantifies how strongly the resistance changes with humidity (i.e., the slope of the resistance change versus % RH), thereby reflecting the material’s humidity responsiveness. As a representative example, Mahlknecht et al. [40] synthesized sodium lauryl sulfate (SLS)-assisted polyaniline nanorods using a solid-phase method and blade-coated them onto porous cellulose paper to fabricate a flexible sensor for humidity detection, breath monitoring, and skin moisture sensing. The device demonstrated strong overall performance, including high sensitivity (6.92% (% RH)⁻¹), rapid response and recovery, low hysteresis, good long-term stability, and skin compatibility. These merits were attributed to the role of SLS in tailoring the polyaniline morphology to increase the effective surface area and improving electrical conductivity. In addition, hydrogen-bond interactions between polyaniline and SLS help stabilize the nanostructure and facilitate charge transfer over a broad humidity range (5–95% RH). Mechanistically, water molecules first establish a chemisorbed layer on the sensing surface, followed by the formation of physically adsorbed multilayers at higher RH levels. Subsequent hydrogen-bond-assisted ionization and Grotthuss-type proton hopping enhance proton transport, increasing ionic conductivity and thereby reducing the sensor resistance (Figure 1a,b).

2.2. Humidity-Specific Response of CRBHSs

The response of CRBHSs is fundamentally determined by their ability to selectively transduce changes in breath humidity into resistance variations. In general, the sensing layer exhibits a pronounced resistive response primarily to water molecules, while showing comparatively low sensitivity to other common breath constituents such as CO₂, O₂, and volatile organic compounds (VOCs). This selectivity typically arises from the synergistic contribution of three key factors.

First, water molecules exhibit distinctive physicochemical properties, including strong polarity, a high propensity for hydrogen bonding, and a high dielectric constant. These features enable much stronger interactions with many sensing materials than those produced by nonpolar or weakly polar breath components (e.g., O₂, N₂, and CO₂). In particular, hydrophilic functional groups on the sensing surface, such as −OH, −NH₂, and −COOH, can form robust hydrogen-bond networks with water molecules [41]. By contrast, interactions with CO₂ and many VOCs (e.g., acetone and ethanol) are often dominated by weaker interactions, including van der Waals forces [42,43]. Moreover, the high dielectric constant of water can significantly modify the local dielectric environment of the sensing layer and influence charge transport. For example, in conducting polymers, increased hydration can reduce interchain dielectric loss and promote ionic conduction [44]. In comparison, gases such as CO₂ typically have a much smaller influence on the dielectric environment due to their relatively low dielectric constants.

Second, selectivity can be further strengthened through hydrophilic surface engineering of the sensing material. Common strategies include introducing or enriching hydrophilic groups to increase water adsorption while minimally affecting uptake of other gases. For instance, NaOH treatment of MXene can remove surface −F species and generate abundant −OH groups, resulting in a several-fold increase in water adsorption relative to untreated samples [45], whereas CO₂ adsorption changes only marginally. Similarly, sensors coated with polyethyleneimine (PEI) benefit from the formation of multiple hydrogen bonds between PEI amino groups and water molecules, yielding a very high humidity response sensitivity of 1.16% (% RH)⁻¹ while maintaining minimal responses to typical VOC interferents such as acetone (<5% response) and ethanol (<3% response) [46].

Third, the inherently high water-vapor content in exhaled breath provides a concentration advantage that supports humidity-dominant responses. At ~37 °C and ~100% RH, the absolute humidity of saturated breath is on the order of tens of grams per cubic meter, far exceeding the concentrations of most interfering species. By comparison, exhaled CO₂ is present at only a few percent by volume, and typical VOC biomarkers such as acetone and isoprene are often at parts-per-billion (ppb) to parts-per-million (ppm) levels, corresponding to much lower mass concentrations [47,48]. Consequently, even if a sensing material exhibits a minor cross-response to certain interferents, the much higher abundance of water molecules generally ensures that humidity remains the primary driver of resistance variation, effectively preserving a humidity-specific output.

3. Structural Tuning of Sensing Material Systems for CRBHSs

The dimensional characteristics of sensing materials in CRBHSs critically govern their interaction with water molecules, charge-carrier transport pathways, and mechanical adaptability [49,50,51]. Thus, dimensionality is a fundamental determinant of key device metrics, including sensitivity, response/recovery speed, stability, and wearability. To satisfy the stringent requirements of breath humidity monitoring, such as high sensitivity, fast dynamics, mechanical flexibility, and resistance to interference, researchers have developed systematic structural engineering strategies tailored to materials of different dimensionalities, spanning zero-dimensional (0D), one-dimensional (1D), two-dimensional (2D), and three-dimensional (3D) architectures (Figure 2). Optimization across these dimensions, from atomic-level design and microstructural control to macroscopic morphology, is enabling CRBHSs to progress from isolated performance improvements toward integrated platforms that combine high sensitivity, rapid response, long-term stability, and practical wearability.

3.1. 0D Materials

Zero-dimensional (0D) materials offer exceptionally high specific surface areas, abundant surface defect sites, and, in some cases, pronounced quantum-confinement effects, all of which can provide plentiful adsorption sites for water molecules [52,53]. In addition, 0D materials can be readily integrated with other material platforms to form functionalized composites or hybrid structures, enabling tunable interfaces and enhanced charge-transport characteristics. As a result, 0D-based sensing layers often exhibit high sensitivity and are widely explored for breath humidity detection.

3.1.1. Carbon Dots (CDs)

CDs are zero-dimensional carbon nanomaterials typically smaller than 10 nm. They offer several advantages for humidity sensing, including good water solubility, widely available precursor sources, high chemical stability, low toxicity, and excellent biocompatibility. In addition, their surfaces are naturally rich in hydrophilic functional groups, such as hydroxyl (−OH), amino (−NH₂), and carbonyl (C=O), which facilitate strong interactions with water molecules [16,54]. Importantly, the surface chemistry and sensing performance of CDs can be further tailored by rational selection of carbon/nitrogen precursors and optimization of synthetic conditions. CDs can also be integrated with other nanomaterials to form heterostructures, enabling interfacial engineering and enhanced functional performance.

Jlassi et al. [55] constructed a PVP–CD composite CRBHS by incorporating CDs (prepared from graphite waste via hydrothermal treatment) into a quaternized poly (vinylpyridine) (PVP) matrix, followed by spin-coating onto ITO substrates (Figure 3a). This design combines the high density of hydrophilic adsorption sites provided by CDs, bearing functional groups such as COO⁻, −OH, and −NH, with the structural stability and charge-transport capability of the PVP matrix. The resulting sensor demonstrated high sensitivity, good linearity, low hysteresis, and notable long-term stability. Moreover, the fabrication route is simple and cost-effective, supporting its potential for practical humidity sensing and breath-related monitoring applications.

3.1.2. Quantum Dots (QDs)

QDs exhibit strong electrical responsiveness to surface-adsorbed species owing to quantum-confinement effects. Combined with their favorable stability and the availability of relatively low-toxicity compositions, QDs have attracted growing interest for wearable breath humidity monitoring [56,57]. Their sensing performance can be tuned through particle-size control, surface/ligand engineering, and the construction of composite heterostructures that optimize adsorption and charge-transport processes.

Pi et al. [58] developed a Cs₃Bi₂Br₉ QD-based CRBHS that demonstrated rapid response/recovery and excellent long-term stability. When integrated into a face mask, the device was able to track breath humidity variations associated with different physiological states, highlighting its potential utility in breath monitoring and disease screening. Similarly, Chaloeipote et al. [59] fabricated a high-performance resistive humidity sensor based on GQDs/AgNP nanocomposites, in which hydrothermally synthesized graphene quantum dots (GQDs) and green-synthesized silver nanoparticles (AgNPs) formed a core–shell structure. The resulting Schottky junctions, together with synergistic effects between the two components, enhanced charge transfer and water adsorption. Consequently, the sensor exhibited fast response/recovery kinetics and good repeatability, making it well suited for wearable breath monitoring (Figure 3b).

3.1.3. Metal Oxide Nanoparticles (MOx NPs)

MOx NPs [60,61] have become an important class of humidity-sensitive materials. Unlike their bulk counterparts, MOx NPs exhibit pronounced nanoscale effects, including a large specific surface area and high surface energy. These features provide substantially more exposed active sites than bulk metal oxides. Their surfaces commonly present abundant oxygen-related species and functional groups (e.g., M–O⁻ and M–OH, where M denotes a metal ion). These sites act as preferential adsorption and reaction sites for water molecules. Owing to their large specific surface area, MOx NPs can accelerate water adsorption/desorption kinetics [62] and strengthen interfacial interactions between water molecules and the sensing surface. In addition, their nanoscale dimensions shorten electron-transport pathways within the sensing layer, facilitating rapid charge transfer under humidity variations and thereby improving sensor response and recovery. This intrinsic surface chemistry, together with nanoscale structural advantages, makes MOx NPs highly attractive for high-performance humidity sensing, particularly in chemiresistive platforms.

Lou et al. [63] designed two Schottky structure sensors (SiNW/ZnO/rGO and SiNW/TiO₂/rGO), in which ZnO and TiO₂ nanoparticles play key roles. ZnO contributes high electron mobility, facilitating charge transfer, whereas TiO₂ can introduce abundant surface oxygen vacancies that promote water adsorption and dissociation. Together with the large specific surface area of silicon nanowires (SiNWs) and the excellent conductivity of reduced graphene oxide (rGO), these nanoparticles form efficient Schottky heterostructures that enhance surface activity and sensing efficiency (Figure 3c). The resulting devices exhibited high sensitivity, stable response/recovery behavior, and good long-term reusability, enabling accurate capture of breathing-related humidity fluctuations and offering potential for sleep apnea monitoring.

Similarly, Soni et al. [64] fabricated rGO/ZnO heterostructures by compositing ZnO with rGO to develop wearable humidity sensors for breath monitoring. Compared with pristine ZnO sensors, the heterostructured devices showed higher sensitivity and a faster response, reduced hysteresis, and excellent stability within the breath-relevant humidity range. Importantly, it could track rapid humidity transitions from exhalation to inhalation, meeting the requirements for real-time, accurate, and continuous monitoring in wearable applications.

Figure 3. Schematic illustration of the device fabrication processes for (a) PVP–CDs [55], (b) GQDs/AgNPs [59], and (c) SiNW/ZnO/rGO and SiNW/TiO₂/rGO [63].

Figure 3. Schematic illustration of the device fabrication processes for (a) PVP–CDs [55], (b) GQDs/AgNPs [59], and (c) SiNW/ZnO/rGO and SiNW/TiO₂/rGO [63].

3.2. 1D Materials

One-dimensional (1D) materials, such as nanotubes, nanowires, and nanofibers, offer high aspect ratios and anisotropic (directional) charge-transport pathways, which can facilitate efficient carrier migration and rapid signal transduction [65,66]. Their elongated architectures also promote the formation of interconnected, porous networks that improve water-vapor diffusion and adsorption dynamics. In addition, 1D building blocks can readily assemble into flexible and mechanically robust sensing films, making them a key material platform for wearable CRBHSs.

3.2.1. Nanotubes

Carbon nanotubes (CNTs) [67] are among the most widely used building blocks for flexible humidity sensors, owing to their hollow geometry, high specific surface area, and excellent electrical conductivity. A common strategy to further improve CNT-based humidity sensing is to incorporate hydrophilic polymer composites, which can (i) enhance water adsorption, (ii) suppress CNT aggregation through hydrogen-bond interactions, and (iii) improve mechanical flexibility and environmental adaptability. Together, these advantages make CNT–polymer systems particularly attractive for wearable applications.

Following this design concept, Liang et al. [68] developed a low-cost porous fiber paper-based sensor (PLFC) using A4 paper as the substrate (Figure 4a,b). By employing MWCNT–carboxymethyl cellulose composite electrodes and LiCl modification, the device achieved a fast response and excellent linearity. Notably, it enabled stable monitoring of breathing patterns and speech-related humidity variations in a non-invasive manner (Figure 4c), providing multidimensional information relevant to health assessment. In addition, the straightforward fabrication process supports cost-effective, scalable production, aligning well with key requirements for wearable health monitoring, namely, real-time operation, accuracy, multifunctionality, and low cost.

3.2.2. Nanowires

Nanowires, including metal oxide nanowires such as TiO₂ and ZnO, as well as metallic nanowires (e.g., gold nanowires, AuNWs), are attractive humidity-sensing materials due to their pronounced structural advantages [69,70]. Their high aspect ratios provide large specific surface areas and facilitate the formation of continuous conductive pathways. In addition, single-crystalline structures (common in many oxide nanowires) and favorable surface chemistries can support efficient carrier transport, while surface oxygen vacancies and/or functional groups enhance water adsorption and activation. Accordingly, performance optimization typically emphasizes aspect ratio control, array/network architecture design, and composite/heterostructure construction. High-aspect ratio nanowires can create interconnected conductive networks that improve charge-transport efficiency, and vertically aligned arrays or porous interconnected structures can accelerate water-vapor diffusion, together enabling improved sensing speed and sensitivity.

Sehrawat et al. [70] developed humidity sensors based on ZnO/KIT-5 composite nanomaterials synthesized via a hydrothermal approach. By integrating ZnO nanowires into a mesoporous KIT-5 matrix, the sensor achieved strong overall performance at an optimized ZnO loading. The improved sensitivity, low hysteresis, and long-term stability were attributed to synergistic effects between the high surface activity of ZnO nanowires and the porous transport pathways provided by the KIT-5 framework. Similarly Adhyapak et al. [71] reported AuNW-based sensors in which the high aspect ratio and interconnected nanowire network enhanced water adsorption and charge conduction, resulting in excellent humidity-sensing performance (Figure 4d).

3.2.3. Nanofibers

Polymer nanofibers and composite nanofibers can be fabricated by electrospinning and subsequently processed or functionalized through approaches such as inkjet printing and solution drop-casting [72]. Owing to their high specific surface area, porous microstructure, and excellent mechanical flexibility, nanofiber mats serve as attractive substrates and/or active sensing layers for wearable humidity sensors [73,74]. The porous architecture provides abundant adsorption sites and rapid mass-transport channels, the large surface area enhances interactions with water molecules, and the intrinsic flexibility accommodates the deformations encountered in wearable scenarios.

Building on these advantages, Yu Liu et al. [75] developed an Ag/SA/TPU composite nanofiber sensor by forming an Ag/SA sensing layer on electrospun TPU nanofibers via inkjet printing followed by reduction. The device can be directly integrated into face masks without additional processing and demonstrates high humidity sensitivity, strong mechanical robustness under repeated bending and friction, low temperature dependence, and good reversibility. These characteristics enable reliable identification of breathing patterns under different emotional and physiological states, as well as accurate monitoring of exercise-related perspiration.

Similarly, Meng et al. [76] fabricated porous carbon nanofibers (PCNFs) by treating electrospun PAN/PVP precursor fibers with a DMF/H₂O mixed solvent, followed by stabilization and carbonization, and then compositing the PCNFs with cellulose to form flexible sensing films. Benefiting from the highly porous structure and abundant oxygen-containing functional groups of PCNFs, the resulting sensor exhibited high sensitivity, good repeatability, and stable performance under both dynamic and static humidity conditions (Figure 4e). Collectively, these advances support the development of high-performance, low-cost wearable humidity sensors and provide a solid technical foundation for practical health-monitoring applications.

Figure 4. Schematic illustration of (a) the overall structure, (b) internal structure and (c) breathing pattern differentiation of the PLFC sensor [68]; (d) the AuNW-based CRBHS device [71]; and (e) the sensing mechanism of the PCNF-based sensor [76].

Figure 4. Schematic illustration of (a) the overall structure, (b) internal structure and (c) breathing pattern differentiation of the PLFC sensor [68]; (d) the AuNW-based CRBHS device [71]; and (e) the sensing mechanism of the PCNF-based sensor [76].

3.3. 2D Materials

Two-dimensional (2D) materials encompass a broad range of structures and exhibit distinctive physicochemical characteristics, attractive scalability, and excellent electronic properties. Because their thickness can be as small as a single atomic layer (or only a few layers), surface atoms are largely exposed, resulting in a very high surface area-to-volume ratio [77,78]. This feature enhances water adsorption efficiency and can induce pronounced variations in resistance, capacitance, and/or dielectric properties upon humidity changes. In addition, the interlayer van der Waals spacing and transport pathways can be tuned by controlling the layer number or intercalation state, enabling performance optimization. These materials can also adhere well to flexible substrates, supporting bending, stretching, and folding, which are key requirements for wearable sensing platforms.

3.3.1. MXenes

MXenes [79,80,81] are a family of two-dimensional transition metal carbides, nitrides, or carbonitrides that combine high electrical conductivity with abundant hydrophilic surface terminations (e.g., −OH and −O). Owing to these attributes, MXenes have become a major focus in chemiresistive breath humidity sensing. Their performance can be tuned through surface-termination engineering, interlayer-structure modulation (e.g., spacing, restacking control, and ion intercalation), and heterostructure construction, which together can synergistically improve water adsorption and charge transport.

Leveraging these advantages, Chen et al. [45] prepared a curled-sheet CF–Ti₃C₂Tx sensor via NaOH electrochemical intercalation, where both hydrophilic group regulation and interlayer-structure optimization contributed to markedly enhanced sensitivity compared with the pristine material. The device also exhibited a rapid response (≈2 s), low hysteresis, and good long-term stability (Figure 5a,b). In another approach, Han et al. [82] constructed a Ti₃C₂Tx/SnO₂ composite sensor using a “2D/0D” heterostructure design. Incorporating SnO₂ nanoparticles into the MXene framework improved the adsorption capability and facilitated charge transfer, enabling a response time as low as 2 s together with high sensitivity, good linearity, and strong selectivity at high humidity. Notably, the composite device also maintained good flexibility and stable cyclic performance, indicating its suitability for high-performance wearable breath humidity monitoring.

3.3.2. Transition Metal Dichalcogenides (TMDs)

TMDs (e.g., MoS₂, WS₂, and ReS₂) [83,84] feature a layered structure that is highly responsive to water adsorption within interlayer galleries and on exposed surface sites. Together with their favorable mechanical flexibility and chemical stability, these attributes make TMDs promising candidates for wearable breath humidity monitoring. Their sensing performance can be further improved through layer-number optimization, surface electronic (electropositive/electronegative) modulation, and substrate/interface engineering to enhance adsorption, charge transfer, and mechanical robustness.

A variety of TMD-based humidity sensors have been reported with strong performance using different fabrication and modification strategies. Adepu et al. [85] prepared a ReS₂/cellulose-paper flexible humidity sensor via vacuum filtration, leveraging the porous paper substrate to provide abundant adsorption sites while achieving good cyclic reversibility and long-term stability. Duan et al. [86] developed high-performance rGO/WS₂ sensors using drop-casting and spray-assisted thermal reduction, respectively; these devices showed good linearity and were suitable for breath monitoring and non-contact sensing. Similarly, Jin et al. [87] reported a fully integrated wearable humidity sensor based on MoS₂/PVP nanocomposite inks inkjet-printed onto flexible LCP–copper interdigital electrodes. The device exhibited rapid response/recovery, stable operation over 30 days, and multifunctional capability, including real-time respiration-pattern recognition (normal, fast, and slow breathing) and touchless skin moisture monitoring (Figure 5c).

Overall, humidity sensors based on different TMDs have achieved notable progress in sensitivity, response dynamics, and application versatility. Continued advances in structural modulation and composite/heterostructure design are expected to further expand the role of TMD materials in wearable health monitoring and non-contact human–machine interaction.

3.3.3. Graphene and Its Derivatives

Graphene and its derivatives (e.g., graphene oxide (GO) and reduced graphene oxide (rGO)) [88,89,90] have attracted extensive interest for humidity sensing because of their atomically thin structures, high electrical conductivity, and large specific surface areas. However, pristine graphene is relatively chemically inert and hydrophobic, which can limit water adsorption and thus humidity sensitivity. Accordingly, performance optimization typically relies on oxidation/functionalization, composite formation, and related interface-engineering strategies to introduce hydrophilic sites, tune surface chemistry, and improve compatibility with flexible substrates and polymer matrices.

Oxidation control is one of the most direct approaches to enhance humidity response. For example, An et al. [91] synthesized laser-induced graphene oxide (LIGO) using liquid-assisted laser patterning, achieving an approximately fivefold increase in the O/C ratio compared with laser-induced graphene (LIG). This modification led to an ~10-fold improvement in humidity response, with response times of 12 s. Peng et al. [92] developed a GO/lignosulfonate (LS) composite sensing layer integrated with LIG electrodes, achieving a sensitivity of 147.73% (% RH)⁻¹ and rapid response/recovery of 12 s/2 s. In addition, Hou et al. [93] constructed a graphene (GR)/polymer composite by incorporating graphene into a multifunctional polymer matrix containing hydroxyethyl cellulose (HEC), xanthan gum (XG), glycerol, and waterborne polyurethane (WPU) (Figure 5d). The resulting sensor exhibited an ultrahigh response, excellent mechanical robustness, and good long-term stability, enabling diverse applications, including respiration monitoring, skin humidity sensing, and contactless gesture recognition.

Figure 5. (a) Schematic illustration of CF–Ti3C2Tx MXene curly-flake formation and (b) its application in human breath detection [45], (c) schematic of the MoS2/PVP device architecture [87], and (d) fabrication process of the GR/polymer CRBHS device [93].

Figure 5. (a) Schematic illustration of CF–Ti3C2Tx MXene curly-flake formation and (b) its application in human breath detection [45], (c) schematic of the MoS2/PVP device architecture [87], and (d) fabrication process of the GR/polymer CRBHS device [93].

3.4. 3D Materials

Three-dimensional (3D) materials featuring interconnected porous networks, high porosity, and favorable mechanical compliance can support efficient water-vapor diffusion, multidimensional adsorption, and the flexibility required for wearable applications [94,95,96]. A key advantage of 3D architectures is the ability to couple macroscale structural design with microscale compositional regulation, enabling a balanced optimization of sensing performance, breath permeability, and mechanical robustness. Consequently, 3D material systems represent an important direction for advancing wearable chemiresistive breath humidity sensors.

3.4.1. Metal–Organic Frameworks (MOFs)

Metal–organic frameworks (MOFs) are porous crystalline materials assembled from metal ions/clusters and organic ligands through coordination bonds [97,98]. They offer ultrahigh specific surface areas; high porosity; and modular, tailorable structures, features that are advantageous for humidity sensing. For example, ionic-liquid-functionalized MOF-303 has been reported to deliver a strong response at low humidity; topologically tuned MOF-545 with dual apertures can provide reliable performance over indoor-relevant humidity ranges; and α-Fe₂O₃/MIL-88B (Fe) composites have shown a wide conductivity window and good stability [99,100,101].

Zhang et al. [102] developed a humidity sensor based on a Fabry–Perot (F–P) cavity constructed from MOF-801/GO heterostructure membranes. By assembling MOF-801 with GO via spin-coating, the device leverages synergistic effects to enhance humidity-sensing performance (Figure 6a,b). It exhibits strong resistance to electromagnetic interference, excellent selectivity toward water vapor, and reliable stability. Importantly, the sensor enables effective humidity detection and non-contact breath monitoring, and can distinguish breathing patterns such as rapid, normal, and deep breathing.

Overall, MOF-based breath humidity sensors are promising for wearable breath sensing due to their high sensitivity, broad operating range, and the potential to tailor adsorption/transport pathways through framework design. Future progress will likely focus on accelerating sensing dynamics, improving flexibility and biocompatibility, adapting devices to specialized humidity environments, and advancing scalable fabrication and system-level integration.

3.4.2. Covalent Organic Frameworks (COFs)

Covalent organic frameworks (COFs) are porous crystalline materials analogous to MOFs in architecture, but are built from organic building blocks linked by covalent bonds [103,104]. Their ordered pores, high surface areas, and tunable structures make them attractive for humidity sensing, and performance can be enhanced through rational functionalization and framework engineering. For instance, a PIL@COF device formed by UV-triggered in situ polymerization of ionic liquids exhibits impedance changes exceeding one order of magnitude over 11–98% RH, with fast response/recovery (2 s/2 s) and low hysteresis (~2% RH). The device can also distinguish different breathing modes (fast, normal, and slow) [105]. COF-5 prepared by a gas-phase-assisted approach has been reported to undergo humidity-driven changes associated with π–π stacking and layer reconstruction, producing an order-of-magnitude resistance variation across 11–98% RH. Notably, even after 180° bending, the linear fitting R² remains above 0.978, indicating excellent mechanical robustness [106].

Xin Liu et al. [107] further developed a self-standing ionic COF membrane (TGCl–TPA) by incorporating guanidinium-functionalized linkers, which introduce abundant hydrogen-bonding sites (Figure 6c). This design promotes rapid water adsorption/desorption and enables dynamic structural responses, while maintaining good mechanical flexibility and durability. The membrane can function both as a humidity sensor and as a moisture-driven actuator, showing fast response/recovery and stable cycling behavior.

Collectively, these examples demonstrate that COFs can deliver outstanding humidity-sensing performance through deliberate structural design and functional modification, broadening their prospects for wearable health monitoring and human–machine interaction.

3.4.3. Hydrogels

Hydrogels possess distinctive physicochemical characteristics, including a three-dimensional cross-linked network, high hydrophilicity, intrinsic stretchability, and readily tunable deformability and stiffness. Many hydrogel systems also offer attractive functionalities such as optical transparency, electrical/ionic conductivity, self-healing behavior, and biocompatibility [108,109,110]. Collectively, these features provide a favorable medium for water transport and ionic conduction, enabling broad opportunities for hydrogel-based flexible devices, from material synthesis and structural assembly to multifunctional sensing applications.

A key limitation, however, is that hydrogels are prone to freezing at low temperatures and dehydration at elevated temperatures, which can severely compromise their mechanical integrity and sensing reliability. To address this, a range of strategies has been developed to improve their antifreezing capability and environmental stability [111], including the incorporation of polyols, ionic liquids, electrolytes, and modified polymer networks. Additional performance gains can be achieved through auxiliary optimization methods such as thin-film treatments, which have shown promising results in practical device implementations [112,113].

For example, introducing glycerol or LiCl into hydrogel matrices can facilitate water transport and increase ionic conductivity, while simultaneously enhancing environmental tolerance. Such modifications have been reported to reduce mass loss to ~11% after 48 h in a dry environment at 80 °C and to depress the freezing point to below −120 °C, supporting robust long-term operation [114]. Moreover, thin-film treatment can further improve sensing performance by increasing the effective surface area; in a polyacrylamide (PAM)/cassava hydrogel, this approach yielded an ultrahigh relative humidity sensitivity of 1.15% (% RH)⁻¹ and shortened the breath-response time to 1.41 s [115] (Figure 6d).

Figure 6. Schematic illustration of (a) Fabry–Perot (F–P) cavity fabrication and (b) its application in human breath detection [102], (c) preparation of the ionic COF membrane (TGCl–TPA) [107], and (d) hydrogel-film-based sensing for wireless breath monitoring [115].

Figure 6. Schematic illustration of (a) Fabry–Perot (F–P) cavity fabrication and (b) its application in human breath detection [102], (c) preparation of the ionic COF membrane (TGCl–TPA) [107], and (d) hydrogel-film-based sensing for wireless breath monitoring [115].

Overall, sensing materials with different dimensionalities (0D–3D) exhibit distinct characteristics for humidity detection. As summarized in

Table 1. The humidity detection effect of CRBHSs based on different sensing materials.

Dimension	Sensing Materials	Detection Range (%)	Response	Hysteresis	Sensitivity/(% RH)−1	Ref.
0D	TENG–CDs	11–94	–	–	64.07%	[16]
	C–dots@g–C3N5	7–85	1.5 s	–	–	[54]
	Cs3Bi2Br9QDs	5–90	5.56 s	–	–	[58]
	SnO2	11–98	8.2 s	1%	1.15%	[60]
	rGO/ZnO	11–90	1.5 s	3.95%	1.27%	[64]
1D	Paper–CNTs	33–98	500 ms	–	9.98%	[68]
	ZnO/KIT-5	11–98	18 s	1.2%	1.15%	[70]
	AuNWs	11–92	0.2 s	–	49.5%	[71]
	Ag/SA/TPU	12–96.9	–	–	24.6%	[75]
2D	CF–Ti3C2Tx	11–97	2 s	1.38%	1.16%	[45]
	Ti3C2Tx/SnO2	11–97	2 s	–	1.10%	[82]
	ReS2/cellulose paper	43–95	142.94 s	–	0.77%	[85]
	GO/WS2	0–91.5	31 s	3%	0.18%	[86]
	MoS2/PVP	11–94	0.8 s	–	1.69%	[87]
	Laser-induced graphene oxide (LIGO)	10–70	180 s	–	0.75%	[91]
	GO/lignosulfonate (LS)	–	12 s	0.58%	147.73%	[92]
3D	MOF-303	7–43	14.8 s	0.2%	–	[99]
	MOF-545	40–70	–	–	–	[100]
	α-Fe2O3/MIL-88B (Fe)	11–97	56 s	4.2%	1.09%	[101]
	PIL@COF	11–98	2 s	2%	1.09%	[105]
	Glycerol/LiCl modified hydrogel	40–85	–	–	2.3%	[114]
	PAM/cassava hydrogel	–	1.41 s	3.1%	1.15%	[115]

, key performance metrics—including operating detection range, response time, hysteresis, and sensitivity—vary substantially across material systems and device architectures. In this review, all sensitivity values reported in Table 1 have been recalculated or verified using the unified definition given in Equation (3).

4. Device Architectures and Integrated Applications of CRBHSs

4.1. Medical Care

Breath humidity is a core physiological parameter that reflects breath function and, to some extent, metabolic status. It is closely associated with the onset and progression of breath diseases and other pathological conditions [116,117]. By enabling accurate, real-time tracking of breath humidity, CRBHSs can support early screening, dynamic monitoring, and risk warning, offering a non-invasive and continuous testing solution for healthcare. These capabilities can improve diagnostic efficiency and quality of care, contribute to better patient outcomes, and potentially reduce healthcare costs.

4.1.1. Disease Surveillance

Abnormal fluctuations in breath humidity often provide early indications of disease. By capturing subtle deviations in the baseline level, fluctuation amplitude, and temporal rhythm of exhaled humidity, CRBHSs can provide quantitative indicators that assist clinical diagnosis and condition assessment, particularly for breath disorders [118,119]. This approach is grounded in the premise that disease-induced airway structural changes and secretion abnormalities can shift the normal range and dynamics of expiratory humidity.

For example, asthma and chronic obstructive pulmonary disease (COPD) are associated with characteristic humidity variations linked to airway inflammation [120], imposing stringent requirements on sensor sensitivity and dynamic response. Ni et al. [114] developed a stretchable CRBHS suitable for conformal integration with skin. Kim et al. [121] reported a flexible CNT@CPM core–shell humidity sensor that can be incorporated into wearable platforms (e.g., masks) to identify multiple breathing patterns in real time (Figure 7a,b). These devices exhibit high sensitivity and fast response/recovery, allowing accurate capture of dynamic expiratory humidity changes, together with robust long-term stability to support continuous monitoring. Compared with conventional diagnostic approaches, CRBHSs offer key advantages, such as simple operation, non-invasive sampling, and real-time continuous measurement, enabling uninterrupted signal acquisition during daily activities and avoiding the time constraints and invasiveness of traditional tests. As a result, they provide convenient and reliable data support for the prevention, diagnosis, and management of breath diseases, while improving user compliance and streamlining clinical workflows.

4.1.2. Breathing Warning

Breathing abnormalities such as sudden apnea and unilateral nasal obstruction pose serious risks for high-risk populations. Newborns may face life-threatening events due to immature breath control, and postoperative patients or individuals with chronic breath diseases are susceptible to undetected breathing disorders that can lead to complications [122]. Traditional monitoring approaches often rely on single-sensor modalities or bulky equipment, with delayed responses and limited ability to map spatial humidity distributions, making it difficult to deliver timely and precise warnings. To address these unmet clinical needs, CRBHSs have been extensively explored in laboratory settings. Current prototype systems can rapidly capture abnormal breath humidity dynamics and provide early warning signals, thereby laying a technical foundation for shifting care from passive rescue to proactive prevention.

Zhang et al. [123] developed a flexible ionogel-based humidity sensor array using a scalable coating process. Integrated with a Bluetooth module and assisted by AI algorithms, the system can be conformally mounted on wearables to identify events such as apnea and unilateral nasal congestion, transmit data to a smartphone app, and trigger timely alerts. Jin et al. [87] reported a fully integrated MoS₂/PVP-based wearable humidity sensor on a flexible LCP substrate with a miniaturized form factor, enabling wireless data transmission to smartphones. The device can differentiate multiple breathing patterns and support touchless monitoring and early warning of breath abnormalities, providing daily protection for high-risk groups such as newborns and postoperative patients (Figure 7c,d).

Collectively, these laboratory-developed CRBHS systems enable sensitive and reliable breath humidity monitoring under controlled conditions, capturing even subtle abnormal fluctuations to support proactive early warning. Although these prototypes have not yet undergone large-scale clinical validation or obtained medical device certification, their core functions, such as continuous, non-invasive monitoring with real-time alerts, align well with key needs in both clinical and home-care settings. With further optimization for biocompatibility, long-term stability, and compliance with clinical regulatory requirements, these systems hold strong promise for translation into practical wearable devices. Such advances could reduce the risk of life-threatening respiratory events in high-risk populations and enhance breathing safety in both hospital and home environments, helping to shift care from passive rescue to proactive health protection.

4.2. Emotion and Behavior Monitoring

Emotional states and behavioral activities can modulate breathing patterns through neuroregulation, resulting in characteristic changes in breath humidity [124]. By quantifying these changes, CRBHSs can support recognition and monitoring of physiological states and enable new approaches for psychological assessment, health management, and human–computer interaction. Their non-invasive, real-time, and unobtrusive nature also allows data collection under naturalistic conditions during daily life.

4.2.1. Emotion

Emotion can influence breathing frequency and depth through coordinated sympathetic and parasympathetic regulation, leading to distinctive humidity signatures in exhaled breath [125]. During tension or anxiety, sympathetic activation often accelerates respiration and can elevate expiratory humidity transiently. In relaxed states, breathing tends to be steady with smaller humidity fluctuations, whereas sadness or emotional dysregulation may produce more irregular breathing and larger humidity variability. These humidity dynamics provide objective physiological correlates for emotion-related state assessment.

However, practical emotion monitoring remains challenging because conventional devices may suffer from poor skin compatibility and limited breathability, causing discomfort or irritation during long-term use. Some systems also exhibit a slow response or inadequate anti-interference performance, limiting their ability to capture rapid humidity fluctuations driven by emotional changes [126,127], while poor portability restricts real-world deployment. To address these issues, Li et al. [128] developed a skin-conformal CRBHS by anchoring MXene–polydopamine (PDA) moisture-sensitive materials onto electrospun elastomer nanofiber substrates for emotion-pattern recognition. The porous, elastomeric substrate provides breathability, conformal contact, and biocompatibility, improving comfort for prolonged wear. Meanwhile, the composite sensing layer delivers high sensitivity and a fast response, enabling accurate capture of breath humidity fluctuations while tolerating minor daily interferences. Its thin, lightweight form factor supports integration into smart wearables for real-time emotion monitoring and feedback, offering practical support for mental health management (Figure 8a).

4.2.2. Sleep

Sleep is essential for tissue repair, immune regulation, cognition, and emotional stability, whereas poor sleep quality and sleep disorders are linked to cardiovascular risk, cognitive decline, and other health problems. Accordingly, real-time and accurate sleep monitoring is important for effective health management [129,130,131].

Breath humidity dynamics correlate with sleep stages: during light sleep, respiration tends to accelerate and humidity fluctuations increase; during deep sleep, breathing becomes more stable and humidity variability decreases; and during rapid eye movement (REM) sleep, respiration often becomes irregular with increased humidity fluctuations. Obstructive sleep apnea–hypopnea syndrome (OSAHS) is characterized by recurrent apnea/hypopnea events that can cause abrupt drops in breath humidity [132], providing key signals for sleep-stage assessment and breath health screening.

Although polysomnography is the clinical gold standard, it is complex, intrusive, and facility-dependent, limiting routine at-home assessment and early warning. CRBHSs offer a non-invasive alternative suitable for daily use. Ding et al. [131] developed a wearable CRBHS integrating electrospinning and vacuum magnetron sputtering, achieving strong breathability, biocompatibility, and anti-interference performance with stable operation under temperature/pressure variation and sweat exposure. Ma et al. [47] designed a low-cost, paper-based humidity sensor with a dual-channel 3D structure capable of distinguishing oral versus nasal breathing and identifying sleep-related breath patterns, including OSA-associated abnormalities—providing an accessible solution for home monitoring and auxiliary diagnosis (Figure 8b,c).

These sensors enable real-time tracking of sleep status and breath health in both home-based monitoring and routine health-management settings. As a result, abnormalities such as atypical sleep-stage transitions and OSAHS-related events can be detected earlier, helping to reduce health risks and avoid delays in diagnosis and intervention. Overall, CRBHS platforms provide an effective, non-invasive approach for sleep monitoring and breath screening, further advancing wearable technologies for practical health diagnostics.

4.2.3. Exercise

Exercise is essential for maintaining health, and accurate assessment of exercise intensity is important for improving fitness while avoiding risks such as overexertion and dehydration. During physical activity, the metabolic rate increases, accompanied by elevated breath rate and ventilation. Exhaled humidity typically rises accordingly and often shows a positive correlation with exercise intensity. Thus, humidity dynamics in exhaled breath can serve as informative physiological signals for evaluating both exercise intensity and overall physiological state [133,134].

Conventional intensity-monitoring approaches, such as heart-rate devices or subjective perception, can suffer from response lag and limited real-time feedback, which may hinder timely intensity adjustment during exercise. To address these limitations, Liu et al. [135] developed an MXene/TPU composite film-based CRBHS. In this design, electrospun TPU nanofibers were modified with chitosan, and MXene nanosheets were deposited onto the fiber surface via electrostatic interactions. The resulting sensor combines fast response, a wide humidity-sensing range, low hysteresis, and excellent repeatability with strong mechanical robustness. Notably, it maintains reliable performance under bending and other deformations and remains stable despite mask deformation during exercise. When integrated into the middle layer of a face mask, the sensor is protected from saliva contamination while enabling non-invasive breath monitoring. Moreover, it can distinguish breath humidity patterns associated with different activity states (e.g., standing, walking, and running), demonstrating strong potential for exercise-related breath monitoring and intensity assessment (Figure 8d).

4.2.4. Speech

During speech production, vocal fold vibration and periodic airflow modulation lead to characteristic fluctuations in expiratory humidity. Different speech sounds generate distinct humidity signatures. For example, vowel pronunciation (e.g., “a” and “o”) typically involves sustained vocal fold vibration, producing larger-amplitude humidity variations with relatively stable periodicity. In contrast, consonant articulation (e.g., “b” and “p”) often involves shorter, more transient vocal fold activity, resulting in smaller-amplitude humidity changes and less regular temporal patterns [136,137,138]. These humidity waveforms can therefore serve as physiological features for speech recognition. By accurately capturing such signatures, CRBHSs enable non-contact speech recognition, offering new opportunities for speech disorder rehabilitation and human–computer interaction.

To demonstrate this potential, Gong et al. [136] developed a CRBHS based on MXene/bamboo cellulose fiber (BCF)/TPU composite membranes. The device was fabricated by loading MXene and BCF onto polydopamine (PDA)-modified electrospun TPU nanofibers via vacuum filtration, delivering strong humidity-sensing performance and durability. When coupled with a convolutional neural network (CNN), the system learned word-specific waveform features, improving prediction accuracy and enabling speech-pattern differentiation. Complementing this approach, Wang et al. [139] proposed a humidity-sensing breath microphone (HSRM) by embedding AuNP/PAH nanocomposite sensors into commercial masks (Figure 8e). Benefiting from the moisture-capturing capability of citrate-stabilized AuNPs and the humidity-responsive ammonium groups in PAH, the HSRM converts breath humidity fluctuations into quantifiable electrical signals; combined with a CNN model, it achieved 85.61% accuracy for non-vocal speech recognition, enabling communication support for patients with vocal impairments without relying on residual vocal fold vibration or skin contact.

Beyond recognition, CRBHSs can also support voice disorder rehabilitation by providing real-time feedback on humidity waveform features during pronunciation. Comparing these patterns with reference databases can help users identify deviations in airflow coordination, phonation stability, and breathing rhythm, enabling targeted training and adjustment. Overall, these advances expand the toolkit for non-contact speech interfaces and provide practical pathways toward assistive communication technologies and next-generation human–machine interaction systems.

Figure 8. (a) Humidity response patterns corresponding to different emotional states (fear, pain, normal, and wonder) [128]; (b) oral and nasal airflow pathways for four breath patterns measured by the PHS patch and (c) its response curves [47]; (d) response curves of the MXene/TPU CRBHS under continuous motion states [135]; and (e) application schematic of the AuNPs/PAH CRBHS system [139].

Figure 8. (a) Humidity response patterns corresponding to different emotional states (fear, pain, normal, and wonder) [128]; (b) oral and nasal airflow pathways for four breath patterns measured by the PHS patch and (c) its response curves [47]; (d) response curves of the MXene/TPU CRBHS under continuous motion states [135]; and (e) application schematic of the AuNPs/PAH CRBHS system [139].

5. Conclusions

CRBHSs hold significant promise for non-invasive health monitoring because of their high sensitivity, simple device architecture, ease of miniaturization, and low power consumption, offering practical routes for real-time detection of breath-related physiological signals. This review provides a unified perspective on CRBHSs by systematically summarizing three core aspects: sensing mechanisms, dimensional material systems, and integrated applications.

In general, CRBHSs operate on humidity-dependent resistance modulation, where proton transport described by the Grotthuss mechanism plays a central role. Their humidity-selective response is enabled by the unique physicochemical properties of water molecules, further reinforced by hydrophilic surface engineering of sensing materials and the relatively high water-vapor concentration in exhaled breath. A broad range of sensing materials across dimensionalities has been developed, including 0D materials (e.g., CDs and QDs), 1D architectures (e.g., nanotubes and nanofibers), 2D systems (e.g., MXenes and graphene derivatives), and 3D porous networks (e.g., MOFs, COFs, and hydrogels), each offering distinct advantages in adsorption behavior, charge-transport characteristics, and mechanical compliance.

Importantly, CRBHSs have demonstrated growing practical value in integrated applications, ranging from medical care (e.g., disease surveillance and breathing-warning systems) to emotion and behavior monitoring (e.g., emotion recognition, sleep tracking, exercise assessment, and speech-related interfaces). By linking sensing principles with materials engineering and application-driven device integration, recent research has substantially advanced CRBHS technologies and broadened their potential impact in clinical medicine and daily health management.

6. Prospects

Although CRBHSs have made remarkable progress, several key challenges must still be addressed to enable their large-scale deployment. These include scalable manufacturing, multidimensional signal acquisition, higher levels of device/system integration, limited selectivity and anti-interference capability, and more intelligent data processing and interpretation.

6.1. Optimizing Device Processing Methods

At present, CRBHS fabrication often relies on laboratory-scale processes (e.g., vacuum filtration and small-batch electrospinning), which can suffer from limited throughputs, relatively high costs, and device-to-device variability, factors that hinder scale-up and broad adoption. In addition, some key sensing materials (e.g., hydrogels and pristine MXenes) are sensitive to environmental conditions and mechanical stress. Hydrogels may freeze at low temperatures or dehydrate at elevated temperatures, while certain 2D materials can be susceptible to oxidation or wear, compromising long-term operational stability under practical breathing conditions.

Future efforts should therefore prioritize low-cost, scalable, and reproducible manufacturing strategies. This includes refining existing fabrication parameters and advancing mass-production-compatible methods such as roll-to-roll (R2R) continuous processing, printing/coating technologies, and 3D/inkjet printing to enable batch fabrication with consistent performance. In parallel, improving material stability through composite engineering and structural optimization will be essential. For example, incorporating glycerol or LiCl into hydrogels has been reported to markedly enhance environmental tolerance (e.g., reducing mass loss after prolonged exposure to hot/dry conditions and substantially depressing freezing points), supporting more reliable long-term operation [114].

Finally, greater emphasis should be placed on integrated forming and device assembly, particularly for wearable platforms. In situ composite strategies that combine elastic polymer substrates with sensing layers can improve mechanical compliance and wearing comfort while strengthening interfacial robustness. For instance, integrated fabrication routes, such as optimizing electrospinning to directly form a composite nanofiber sensing layer on a mask substrate, can eliminate post-assembly steps and improve stability under complex real-world conditions, including daily activities and exercise.

6.2. Advancing Multi-Parameter Synergistic Sensing

Most existing CRBHSs are designed primarily for single-parameter humidity detection, which limits their ability to capture the multidimensional physiological information embedded in the breathing process. In practice, abnormal respiratory states are often accompanied by coordinated changes in multiple indicators, such as humidity, temperature, airflow rate/pressure, and VOC biomarkers. Reliance on a single signal can therefore increase the risk of bias or misinterpretation. In addition, key issues such as condensation and temperature–humidity coupling remain incompletely resolved. These physical interferences can directly compromise measurement accuracy and further impede the translation of CRBHSs toward practical clinical applications.

To address these challenges, future development should move toward multi-parameter synergistic sensing through two complementary pathways. First, composite sensing-layer designs can enable “one platform, multiple functions” by heterogeneously integrating humidity-sensitive materials with other functional components, such as temperature-responsive layers (e.g., polyimide-based elements) and VOC-responsive conducting polymers. Second, multimodal sensor arrays can be realized by integrating multiple independent modules (e.g., humidity, temperature, and airflow/pressure sensors) onto a single flexible substrate, allowing synchronous acquisition of multiple physiological signals.

A key strategy to synergize these two pathways and advance multi-parameter sensing is the development of hierarchically integrated sensing architectures. In such designs, functional layers or modules are vertically or laterally integrated on a single substrate, for example, a core humidity-sensitive layer combined with a temperature-compensation layer, an airflow/pressure-sensing layer, and a gas-recognition layer targeting key exhaled biomarkers. Each module is engineered for a selective response, while data-fusion algorithms are used to suppress cross-interference and correlate multidimensional outputs into a coherent physiological profile. For instance, coupling an MXene-based humidity-sensitive layer with polyaniline can yield a dual-response module capable of simultaneously monitoring breath humidity and ammonia, thereby improving robustness and specificity for breath-based screening.

Mitigating condensation will also be essential for reliable multi-parameter sensing. This can be achieved by tuning surface energy and porosity and by constructing hierarchical porous structures that promote rapid vapor diffusion while discouraging liquid-water accumulation. Breathable yet protective encapsulation membranes may further suppress condensation while maintaining gas permeability. To address temperature coupling, miniaturized temperature sensors can be integrated into the platform, together with real-time calibration models. Temperature signals can then be used to dynamically correct humidity and gas responses and compensate for thermal drift, analogously to dedicated temperature-compensation modules in sensor arrays designed for real-world operation.

Overall, these integrated multi-signal platforms can provide a more comprehensive depiction of breath physiology and deliver richer datasets to support disease screening and health assessment. Such advances will help accelerate the transition of CRBHSs from single-parameter monitors to integrated physiological profiling systems.

6.3. Enhancing Selectivity and Anti-Interference Capability

A major gap in current CRBHS research is the lack of systematic evaluation against potential interfering species present in exhaled breath (e.g., organic vapors/VOCs, CO₂, and NH₃). Many studies do not perform standardized interference tests, which limits confidence in sensor performance under complex real-world conditions and, consequently, slows progress toward clinical adoption. Addressing this limitation will require targeted solutions at both the material and system levels.

Future efforts should focus on three key directions. First, material functionalization: engineering sensing surfaces to introduce water-preferential interaction sites (e.g., hydrophilic functional groups) while suppressing adsorption of non-target species through hydrophobic modification, selective barrier layers, or molecularly imprinted polymer coatings. Second, sensor-array strategies: constructing multi-material arrays in which individual elements exhibit differentiated responses to humidity and interferents, enabling multivariate analysis and pattern-recognition algorithms to extract humidity-specific signatures from complex mixtures. Third, standardized evaluation protocols: establishing representative interferent panels and concentration ranges that reflect real exhaled-breath compositions, thereby enabling community-wide benchmarking and improving cross-study comparability.

6.4. Improving Long-Term Stability Under Real Breathing Conditions

Long-term stability is essential for CRBHSs deployed in continuous health monitoring, but current devices often suffer from performance degradation over time. Factors contributing to this issue include environmental exposure, repeated mechanical wear during breathing, contamination from exhaled droplets or skin secretions, material aging under cyclic humidity and temperature fluctuations, and suboptimal fabrication processes that may introduce structural defects or inconsistent material bonding.

Future improvements should focus on material, structural, and packaging design. At the material level, modify sensing layers with anti-fouling and anti-oxidation components such as hydrophobic coatings or protective polymer matrices to resist contamination and degradation. Composite engineering can also enhance material durability, such as integrating rigid nanofillers into flexible polymers to balance flexibility and mechanical strength. Structurally, adopt fatigue-resistant designs such as flexible 3D nanofiber networks that can withstand repeated mechanical deformation during breathing without compromising sensing performance. For packaging, use biocompatible, breathable materials that isolate the sensing layer from physical damage and contaminants while allowing the unobstructed passage of water vapor and gases. Additionally, establish standardized accelerated aging tests that simulate real-world use scenarios, including cyclic humidity fluctuations, temperature shocks, and mechanical bending. These tests will validate long-term stability and establish reliability benchmarks for CRBHSs, ensuring consistent performance over extended periods of use. By closing this methodological gap, CRBHSs can achieve more reliable humidity detection in the presence of environmental and breath-borne interferents, an essential prerequisite for advancing from laboratory prototypes toward clinically deployable tools.

6.5. Promoting Integrated and Multifunctional Applications

Multifunctional integration refers to incorporating capabilities beyond sensing that improve the practicality, user experience, and real-world deployability of CRBHSs. This direction complements multi-parameter synergy by directly addressing key barriers to adoption. Ultimately, the value of CRBHSs lies in their seamless incorporation into daily life and clinical workflows; however, most current systems remain “single-function, standalone devices” with two major limitations: (i) heavy reliance on external power sources and separate hardware for data transmission and (ii) insufficient integration with common wearable carriers. These constraints reduce convenience, limit suitability for long-term monitoring, and hinder broad deployment. Future development should therefore emphasize integrated, multifunctional design at both the device and system levels.

Key functional enhancements can be advanced along five interconnected directions. First, self-powered operation and micro-power supply solutions: integrating energy-harvesting approaches (e.g., piezoelectric generators that convert breathing-induced motion into electricity or sweat-powered biofuel cells) together with flexible thin-film batteries to reduce dependence on external power and enable uninterrupted long-term monitoring. Second, low-power wireless communication and end-to-end data transmission: embedding modules such as Bluetooth Low Energy (BLE) or WiFi to support real-time data transfer to smartphones and cloud/edge analytics platforms, thereby enabling remote monitoring, clinical data sharing, and home-based health management, particularly valuable for older adults and patients with chronic respiratory conditions. Third, deep integration with wearable carriers and biocompatibility optimization: designing ultrathin, stretchable devices that can be seamlessly integrated into masks, wristbands, sleep patches/mattresses, and sports apparel for truly non-invasive continuous monitoring. For example, positioning sensing units within the middle layer of a mask can help mitigate saliva contamination while enabling real-time tracking of breathing status during daily activities. Soft, hypoallergenic materials should be adopted to ensure comfort and biocompatibility during prolonged contact. Fourth, application-driven customization: developing tailored solutions such as miniaturized, skin-friendly sensors for neonatal intensive care units and devices with improved tolerance to extreme temperature/humidity for high-altitude or endurance-sport populations, thereby broadening practical use cases. Fifth, integration with microfluidic technologies: combining CRBHSs with microfluidics to enable portable lab-on-a-chip platforms for rapid multi-index analysis of exhaled breath samples, supporting point-of-care and on-site testing needs in clinical and field settings

By integrating these functionalities, CRBHSs can evolve from isolated sensing elements into comprehensive, user-centric monitoring systems. Such designs not only reduce practical barriers associated with current prototypes but also better match the requirements of continuous health monitoring and clinical practice, thereby accelerating translation from laboratory demonstrations to broadly adopted tools for daily health management and clinical support.

6.6. Intelligent Upgrading

The large volumes of dynamic, multidimensional data generated by multi-parameter CRBHSs during long-term monitoring cannot be fully leveraged through traditional manual analysis. Manual interpretation is often inadequate for extracting subtle patterns, delivering real-time abnormality warnings, or enabling adaptive device control, thereby limiting broader intelligent applications. Integrating artificial intelligence (AI) and machine learning (ML) is therefore crucial for transforming raw sensor outputs into actionable insights, improving detection accuracy, and enabling real-time decision-making. Deeper AI integration is expected to accelerate the evolution of CRBHSs toward data-driven, smart health-monitoring systems.

Future intelligent upgrades should emphasize four core application directions, supported by robust data infrastructure. First, AI-enabled pattern recognition: ML models (e.g., convolutional neural networks, long short-term memory networks, and support vector machines) can be trained on multi-parameter datasets to identify disease-relevant signatures, such as correlations among humidity fluctuations, airflow patterns, and breath biomarkers associated with respiratory disorders. Potential targets include early signals preceding asthma exacerbations, signatures related to emotional states (e.g., tension vs. relaxation), and patterns linked to sleep disorders such as obstructive sleep apnea syndrome. For example, AI-driven analysis of overnight humidity profiles may help differentiate sleep stages (light, deep, and rapid eye movement sleep), supporting sleep-quality assessment without reliance on specialized clinical equipment. These models could enable real-time early warning without manual intervention.

Second, AI-based adaptive calibration and control: AI can be used to dynamically adjust operational parameters, such as detection thresholds, sampling rates, and filtering/compensation factors, based on real-time environmental inputs (ambient temperature and humidity) and historical device-performance data. Such adaptive strategies can further mitigate condensation effects, temperature coupling, and long-term drift, thereby sustaining accurate detection in complex real-world environments.

Third, personalized health management: Individual-specific baseline models can be built from longitudinal monitoring data to tailor alert criteria and health recommendations, accounting for inter-individual physiological variability and improving the relevance of monitoring outcomes. With sufficiently large and diverse datasets, these personalized strategies can be refined and validated, forming a foundation for population-specific health guidance.

Fourth, enhanced intelligent human–computer interaction: Coupling CRBHSs with AI-based interpretation of humidity waveforms may enable novel interaction modalities. For instance, learning relationships between breath humidity signatures and speech articulation could improve non-contact speech recognition, especially in noisy environments, and provide alternative communication pathways for individuals with speech impairments. Beyond speech, AI-enhanced decoding of breath rhythm and humidity dynamics may support breath-controlled interfaces, expanding CRBHS applications beyond health monitoring.

In addition, establishing multicenter data-sharing platforms will be essential to support these advances. Aggregating datasets across institutions, regions, and populations can enable more robust model training and validation; uncover deeper associations among breath humidity, disease states, and behavioral patterns; and ultimately facilitate AI-assisted personalization and individualized health-management strategies.

6.7. Advancing Interdisciplinary Integration to Unlock Novel Application Scenarios

The continued advancement of CRBHSs will require deep interdisciplinary integration to address complex cross-domain challenges and to broaden both their application scope and practical value. Progress will increasingly depend on coordinated efforts spanning materials science, electronic engineering, biomedicine, computer science, energy technology, and environmental science.

In materials–electronics collaboration, future work should develop novel functional materials, such as high-stability composites, biodegradable polymers, earth-abundant precursors, and biomass-derived substrates, and integrate them with flexible electronics to realize miniaturized, robust, and multi-parameter sensing modules with uniform performance. In parallel, eco-friendly materials and manufacturing routes should be explored to reduce environmental impact and better align CRBHS development with sustainability goals.

In biomedicine–engineering partnerships, close collaboration with clinicians is essential to define clinically relevant targets (e.g., disease-specific biomarkers, monitoring duration, and deployment scenarios) while improving biocompatibility, wearability, and diagnostic specificity. A particularly promising direction is combined detection of “breath humidity + biomarkers” for specific diseases, for example, simultaneous monitoring of breath humidity with acetone and nitric oxide to support earlier screening and improved diagnosis of diabetes and asthma, respectively. Clinical validation of multi-parameter CRBHS platforms will be critical to ensure compatibility with diagnostic workflows and to strengthen clinical relevance.

In computer science–engineering integration, algorithms and embedded software should be optimized for low-power, real-time inference on portable devices, enabling intelligent data processing at the edge and reducing reliance on cloud computing. In energy–materials collaboration, CRBHSs can be coupled with energy-harvesting technologies (e.g., piezoelectric harvesting or sweat-powered systems) and low-power sensing materials to enable self-sustaining, long-term monitoring and extended device lifetimes for continuous health tracking. Finally, within electronic engineering, further advances in flexible electronics, miniaturized system design, and packaging are needed to create ultrathin, stretchable devices, accelerating adoption in wearable platforms and potentially enabling future implantable applications.

Overall, CRBHSs have established a solid technical foundation through advances in sensing mechanisms, materials, and application-oriented design, with the overarching goal of providing reliable, non-invasive tools for assisted diagnosis and health monitoring. Achieving broader real-world impact will require targeted solutions to persistent challenges, including scalability, multi-parameter synergy, selectivity, long-term stability, condensation and temperature coupling, multifunctional integration, and AI-enabled intelligence, together with sustained interdisciplinary collaboration across materials science, electronics, biomedicine, computer science, and energy technology. These combined efforts will accelerate the transition of CRBHSs from laboratory prototypes to accessible intelligent monitoring terminals, enabling early warning and continuous breath tracking and positioning CRBHSs as a key component of next-generation personalized healthcare.