Detection of Sweat-Related Metabolites (Glucose, Lactic Acid, and Urea) Using a SWCNT-Modified Gold Screen Printed Electrode Based Biosensor

Abstract

The increasing demand for continuous physiological monitoring has accelerated the development of high-sensitivity wearable electrochemical platforms. This study reports the fabrication of a multi-analyte electrochemical sensor based on single-walled carbon nanotubes (SWCNTs) for the detection of sweat-associated metabolites. To facilitate efficient heterogeneous electron transfer, glucose oxidase (Gox), lactate oxidase (Lox), and urease (Ure) were immobilized onto the SWCNT network through π–π interaction using 1-pyrenebutanoic acid succinimidyl ester (PBSE), followed by additional stabilization via 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC)/N-hydroxysuccinimide (NHS) coupling. The developed platform exhibited concentration-dependent resistance responses within the ranges of 0.02–0.20 mM for glucose, 20–100 mM for lactate, and 50–400 mM for urea under controlled experimental conditions. The resistance-based configuration enabled stable and reproducible signal modulation across these concentration intervals. Although direct testing with human sweat was not performed, the electrochemical behavior of key sweat-related metabolites was systematically evaluated as a preparatory step toward future wearable integration.

1. Introduction

Advances in nanotechnology and biotechnology have accelerated the development of electrochemical biosensors for clinical diagnostics and continuous monitoring [1]. Enzyme-based biosensors are particularly valued for their specificity and compatibility with miniaturized electrode systems [2]. The widespread clinical use of glucose meters and pregnancy test kits illustrates the practical maturity of biosensor technologies [3]. Glucose, lactic acid, and urea are key metabolites associated with metabolic regulation, physical performance, and renal function, respectively, and their monitoring provides valuable insights into systemic physiological status. Diabetes mellitus is among the most prevalent chronic diseases worldwide and is a leading cause of mortality [4]. Poor glycemic control can result in severe complications affecting major organs, underscoring the importance of maintaining normal blood glucose levels to improve diabetic prognosis [5,6]. However, traditional blood glucose meters rely on invasive blood sampling, leading to discomfort and pain that may discourage consistent monitoring [7]. Although approaches like alternative sampling sites and refined needle designs have been explored, patient discomfort has not been fully resolved [8]. Over the past two decades, technological advances have facilitated non-invasive glucose monitoring by eliminating the need for blood sampling [9]. Electrochemical biosensors, which measure biochemical reactions through changes in electrical signals such as potential, current, or conductivity, have gained considerable attention due to their rapid response times and potential for miniaturization [10]. In these sensors, selectivity is primarily governed by the biological recognition element, while sensitivity depends on the performance of the transducer. As such, the strategy for assembling these components is critical in biosensor development. Carbon nanotubes (CNTs), particularly single-walled carbon nanotubes (SWCNTs), possess exceptional mechanical strength, a large surface area, excellent electrical conductivity, and high chemical stability, making them ideal for a range of applications, including biosensors [11]. Within biosensing systems, CNTs enhance electrode activity and support stable enzyme immobilization, thereby promoting efficient electron transfer in biochemical reactions [12,13]. Sweat is an appealing biofluid for non-invasive monitoring as it contains key metabolites like glucose, lactic acid, and urea [14]. Representative concentration ranges of these metabolites in human sweat have been reported in previous studies. Glucose levels in sweat are typically observed in the micromolar range, commonly between 0.01 and 0.1 mM, although variations may occur depending on physiological conditions and sweat rate. Lactate concentrations are generally higher, often ranging from 5 to 25 mM during moderate physical activity. Urea levels in sweat have been reported in the approximate range of 10 to 60 mM. These physiological ranges provide important reference intervals for selecting appropriate detection windows in electrochemical sensor development. Low glucose concentration relative to blood, alongside varying sweat rates, temperature, pH, and humidity, poses significant challenges to accurate measurement [15]. Incorporating additional metabolites, such as lactic acid and urea, can improve the sensor sensitivity while providing a more comprehensive picture of an individual metabolic state [16,17,18].

Recent progress in non-invasive electrochemical sensing has expanded the potential for monitoring metabolites from alternative biofluids such as sweat. Although wearable systems for sweat glucose detection have been widely reported, many remain constrained by limited detection ranges, stability concerns, or single-analyte configurations. Lactate sensing has primarily focused on exercise-related monitoring, where variations in pH and ionic conditions can affect signal reliability. Urea detection has also been explored, but most studies emphasize individual metabolite analysis rather than integrated platforms. Despite these advances, resistance-based architectures capable of evaluating multiple sweat-associated metabolites within a unified system are still relatively limited. This highlights the need for a stable electrochemical platform that can systematically characterize key sweat-related metabolites under controlled conditions as a preparatory step toward wearable applications.

Glucose is essential in numerous metabolic pathways associated with ATP production in humans. Excessive glucose levels can lead to severe complications, including blindness, heart disease, hypertension, and kidney failure [19,20,21]. Consequently, regular glucose monitoring is vital for preventing and managing diabetes. Urea serves as an important biomarker for kidney and liver function, prompting ongoing development of rapid, accurate detection methods in both urine and blood samples [22,23]. Meanwhile, lactic acid is a critical metabolic byproduct that accumulates in sweat during muscle exertion, and fluctuations in its level can reflect physical activity and metabolic health. In this study, we propose a highly sensitive electrochemical biosensor designed for real-time, non-invasive detection of glucose, lactic acid, and urea in sweat. By immobilizing specific enzymes onto a gold-based electrode composed of SWCNTs composites, we aim to capture subtle electrochemical changes indicative of these metabolites. The findings highlight the potential for a wearable sensor that continuously monitors key biomarkers at the molecular level, offering a promising strategy for personalized healthcare and early disease prevention. In this study, the electrochemical behavior of individual sweat-associated metabolites was systematically evaluated as a preliminary step toward future validation using real sweat samples.

2. Materials and Methods

2.1. Materials

Single-walled carbon nanotubes (SWCNTs) Glucose oxidase (Gox) from Aspergillus niger, lactate oxidase (Lox) from Aerococcus viridans, and urease (Ure) from Canavalia ensiformis were purchased from Sigma-Aldrich (St. Louis, MO, USA). N,N-dimethylformamide (DMF) and acetone were obtained from Daejung (Gyeonggido, Republic of Korea). All enzymes were stored at −20 °C prior to use to maintain catalytic activity. DRP-220AT gold-based screen-printed electrodes (SPEs) (Metrohm DropSens, FL, USA) were used as the primary transducers. Phosphate-buffered saline (PBS, pH 7.4) was used as the supporting electrolyte for all electrochemical measurements.

2.2. Preparation of Single Walled Carbon Nanotube (SWCNT) Dispersion

To prepare the CNT suspensions, SWCNTs were dispersed at concentrations of 0.025, 0.05, 0.1, 0.2, and 0.4 g/L in N,N-dimethylformamide (DMF). The desired amount of SWCNT powder was first weighed and added to the solvent, followed by vigorous mixing and ultrasonication using a sonicator (Ultrasonic 7, Asia Ultrasonic) operating at 300 W and 25 °C for 2 h. This procedure was employed to break up nanotube bundles and promote the formation of homogeneous and stable SWCNT dispersions. After ultrasonication, the suspensions were allowed to stand at room temperature for 24 h to assess their dispersion stability and sedimentation behavior. Immediately prior to film fabrication and electrical measurements, each SWCNT suspension was subjected to an additional ultrasonication step for 30 min to redispose any partially settled aggregates and ensure reproducible SWCNT loading. For each condition, 20 μL of the final suspension was carefully drop cast onto the designated substrates and dried at room temperature to form SWCNT thin films [24,25]. To address concerns regarding residual DMF toxicity, the modified electrodes were thoroughly rinsed with deionized water and dried under vacuum, ensuring the complete removal of solvent traces before enzyme immobilization.

2.3. Optimization of SWCNT Loading and Electrode Fabrication

In the development of carbon-modified electrodes, the dispersion stability of SWCNT suspensions and their adhesion to the electrode surface were systematically assessed. DMF showed the best dispersion efficiency and was therefore selected as the solvent. The drop-casting method was employed to immobilize the SWCNTs onto the electrode. In the drop casting method, 20 µL of the SWCNT suspension was deposited on the electrode via pipette and dried in an oven at 85 °C for 5 min to fix the SWCNTs onto the electrode. SWCNTs were selected for electrode modification and dispersed in DMF to prepare stable suspensions. Five concentrations (0.025, 0.05, 0.1, 0.2, and 0.4 g/L) were evaluated to determine their influence on film formation and electrical properties. Among these, 0.1 g/L produced the most uniform and reproducible coating on the electrode surface, providing consistent baseline resistance and adequate surface coverage. This concentration was therefore used for subsequent experiments. The optimized SWCNT layer was subsequently utilized as the conductive platform for 1-pyrenebutanoic acid succinimidyl ester (PBSE) functionalization and enzyme immobilization in the sweat biosensor system, as schematically illustrated in Figure 1. After deposition of the SWCNT layer, the modified electrodes were dried at room temperature. The electron transfer mediator, PBSE, was then investigated at concentrations of 2.0, 4.0, 6.0, and 8.0 g/L to optimize interfacial coupling efficiency. Based on resistance response and surface stability, 6.0 g/L was selected as the operating concentration. Subsequently, 20 µL of PBSE solution (6.0 g/L) was drop-cast onto the electrode surface and dried at 80 °C to ensure stable surface functionalization [12,26].

2.4. Enzyme Immobilization Procedure

Prior to enzyme immobilization, the SWCNT-modified screen-printed electrodes were prepared under controlled conditions to ensure reproducibility of the electrode modification process. PBSE was introduced as a linker molecule to facilitate the attachment of enzyme molecules onto the SWCNT surface via π–π interactions and subsequent covalent bonding. After PBSE functionalization, the electrodes were rinsed to remove excess linker. The enzyme solution was then drop-cast onto the electrode surface and incubated under ambient conditions to allow stable immobilization.

To minimize enzyme leaching during electrochemical measurements, covalent immobilization was achieved through PBSE-mediated π–π stacking on SWCNT sidewalls followed by 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC)/N-hydroxysuccinimide (NHS) coupling. After enzyme attachment, the electrodes were thoroughly rinsed with phosphate-buffered saline and deionized water to remove loosely bound biomolecules. The stability of the immobilized enzymes was indirectly confirmed by the reproducibility of resistance and amperometric responses during repeated measurements.

2.5. Electrode Characterization

The surface characteristics of the SWCNT, SWCNT/PBSE, and SWCNT/PBSE/Gox thin film electrodes were analyzed using field-emission scanning electron microscopy (FE-SEM; Inspect F50, FEI). For SEM analysis, samples were prepared following the procedure described in Section 2.4, with slight modifications. The modified materials (SWCNT, SWCNT/PBSE, and SWCNT/PBSE/Gox) were collected from the electrode surface and deposited onto a conductive substrate, followed by drying at room temperature to form a thin, uniform coating. Because CNT-based films exhibit poor conductivity and are susceptible to surface charging under high vacuum electron beams, the samples were sputter-coated with a thin Pt/Pd layer (thickness approximately 5–8 nm) using a sputter coater (E-1045, HITACHI). The coating step provided a conductive surface and enhanced image resolution by preventing localized charging artifacts. After metal coating, the samples were loaded into the SEM chamber and evacuated to the required operating vacuum level. Imaging was performed at accelerating voltages ranging from 5 to 15 kV to balance surface sensitivity and image clarity. Lower voltages were applied to visualize fine nanotube structures, while higher voltages were used to observe enzyme coated regions with greater topographical contrast [27]. All micrographs were captured under identical working distances and detector settings to enable direct comparison among the SWCNT, SWCNT/PBSE, and SWCNT/PBSE/Enzymes.

Electrochemical and electrical measurements were carried out using a portable potentiostat (EmStat4S, PalmSens BV, Houten, the Netherlands). The instrument operates within a potential range of ±3 V or ±6 V and supports a maximum current of ±30 mA or ±200 mA, enabling stable control of applied potential and current during measurements. All measurements were performed using a three-electrode configuration integrated within the screen-printed electrode (SPE) platform, consisting of a gold working electrode, a gold counter electrode, and a printed silver (Ag) reference electrode. Electrical resistance was evaluated from current–voltage (I–V) characteristics recorded under low applied potential conditions to ensure operation within the linear (ohmic) regime. The recorded I–V curves exhibited linear behavior within the selected potential window, confirming ohmic conduction through the SWCNT network. Resistance values were determined from the slope of the linear region of the I–V curves. Cyclic voltammetry (CV) measurements were conducted within a potential window of −0.2 V to 1.0 V at a scan rate of 100 mV/s. All experiments were carried out at room temperature unless otherwise specified.

3. Results

3.1. SWCNT Film Formation and Resistance Behavior

The electrical resistance of SWCNT films deposited on the screen-printed electrode (SPE) was measured to assess the development of conductive pathways as a function of nanotube concentration. The selection of SWCNTs and the dispersion solvent was initially evaluated based on suspension stability and film-forming behavior on the gold screen-printed electrode surface. Among the tested conditions, SWCNTs dispersed in DMF exhibited stable suspension characteristics without visible aggregation during preparation, enabling reproducible thin-film deposition. The resistance values obtained at 0.025, 0.05, 0.1, 0.2, and 0.4 g/L are presented in Figure 2. At 0.025 g/L, the resistance was approximately 4.1 Ω, indicating partial surface coverage and limited nanotube connectivity. Increasing the concentration to 0.05 g/L and 0.1 g/L resulted in resistance values of about 5.3 Ω and 5.7 Ω, respectively, reflecting the progressive formation of a more continuous nanotube network. At higher concentrations (0.2 and 0.4 g/L), the resistance decreased to approximately 5.1 Ω and 3.9 Ω, consistent with increased inter-tube contact and densification of the conductive layer. While these conditions improved overall conductivity, excessive network compaction may reduce effective surface accessibility and limit resistance modulation during sensing. Considering the balance between film uniformity, baseline stability, and signal responsiveness, 0.1 g/L was selected as the operating concentration for subsequent biosensor fabrication [10,28].

3.2. Surface Functionalization and Enzyme Immobilization Analysis of SWCNT-Modified Electrodes

The surface modification process of the SWCNT thin film electrodes was systematically analyzed to verify the successful attachment of 1-pyrenebutanoic acid succinimidyl ester (PBSE) linker molecules and the subsequent immobilization of glucose oxidase (Gox). SEM images in Figure 3 illustrate the morphological evolution of the electrodes at each fabrication step. The pristine SWCNT network (Figure 3a) displayed a highly porous and entangled tubular structure, providing abundant surface area and interconnected conductive pathways essential for electrochemical sensing. Following PBSE treatment, notable changes in surface compactness were observed (Figure 3b). The presence of a thin organic layer partially covering the nanotube bundles suggests effective π–π interactions between the pyrene moieties of PBSE and the SWCNT sidewalls [29,30]. This conformal coating not only stabilizes the nanotube network but also introduces reactive ester groups required for enzyme linkage. After Gox immobilization, the surface morphology changed considerably (Figure 3c). The formation of irregular aggregated clusters and flake-like structures is characteristic of densely adsorbed or covalently attached enzyme molecules. The disappearance of clearly distinguishable nanotube bundles indicates a high surface coverage of Gox, confirming the efficient activity of PBSE as a coupling mediator. This substantial increase in surface roughness and the formation of a continuous enzyme film strongly suggest that Gox molecules were stably immobilized rather than loosely physiosorbed. The morphological evolution from SWCNT, SWCNT/PBSE, SWCNT/PBSE/Gox indicates that the introduced functional groups remained active throughout the surface modification process. The greater roughness and thicker immobilized layer following Gox attachment are expected to increase the number of catalytic sites in direct contact with the conductive nanotube network, thereby enhancing electron transfer during glucose oxidation. These structural characteristics support the improved sensing performance observed in subsequent electrochemical evaluations and confirm the suitability of the SWCNT/PBSE interface for enzyme-based biosensing applications [31].

3.3. Effect of PBSE Concentration on Interfacial Resistance and Enzyme Coupling

To identify an appropriate PBSE concentration for stable electron transfer within the SWCNT-based biosensor, resistance values were compared for electrodes modified with PBSE concentrations of 2.0, 4.0, 6.0, and 8.0 g/L, as shown in Figure 4. PBSE functions as a bifunctional linker, enabling π–π interaction between its pyrene group and the SWCNT surface while providing reactive succinimide ester groups for enzyme immobilization [32,33]. At 2.0 g/L, the measured resistance was approximately 6.3 kΩ. Increasing the concentration to 4.0 g/L led to a further rise in resistance to around 7.1 kΩ, suggesting that a thicker PBSE layer formed on the SWCNT surface, which may partially restrict charge transport. When the PBSE concentration was adjusted to 6.0 g/L, the resistance decreased to approximately 5.0 kΩ. A similar value of about 5.3 kΩ was observed at 8.0 g/L. The reduction in resistance at 6.0 g/L indicates improved interfacial balance between linker coverage and electrical continuity. Considering both effective enzyme anchoring and preservation of conductive pathways, 6.0 g/L was selected as the optimal PBSE concentration for subsequent biosensor fabrication.

3.4. Resistance-Based Sensing Response

Figure 5 illustrates resistance values obtained after immobilizing various concentrations of enzymes (Gox, Lox, and Ure) onto electrodes coated with PBSE linker SWCNT. A notable increase in resistance response (R, kΩ) was observed as enzyme concentrations approached 0.1 g/L for all three enzymes. This indicates that the optimal concentration for Gox, Lox, and Ure is 0.1 g/L. Figure 5 shows the comparative resistance measurements obtained after immobilizing various concentrations (0.01, 0.05, 0.1, 0.2, and 0.5 g/L) of enzymes (Gox, Lox, and Ure) onto electrodes modified with PBSE linker SWCNT. For Gox (Figure 5a), the resistance increased from approximately 4.2 kΩ at 0.01 g/L to a peak value of 7.0 kΩ at 0.1 g/L, then decreased to around 3.5 kΩ at 0.5 g/L. Similarly, for Lox (Figure 5b), the resistance increased from 3.5 kΩ at 0.025 g/L to the highest value of 6.2 kΩ at 0.1 g/L, subsequently decreasing to approximately 3.4 kΩ at 0.5 g/L. For Ure (Figure 5c), resistance showed an increasing trend from 6.0 kΩ at 0.01 g/L to a maximum of 9.5 kΩ at 0.1 g/L and then slightly decreased at higher concentrations. These results indicate that the enzyme electrode interaction is most efficient at an enzyme concentration of 0.1 g/L, providing optimal conditions for the biosensor application [34].

3.5. Cyclic Voltammetry Measurement Results

Figure 6 shows the cyclic voltammetry (CV) graphs of glucose, lactate, and urea detection performed to evaluate the electrochemical characteristics of the fabricated electrodes. Figure 6a illustrates the CV cycle for glucose detection, showing clear current flow at the oxidation and reduction potentials of glucose, which appear at approximately 0.45–0.50 V and −0.05 V, respectively. Similarly, Figure 6b presents the CV cycle for lactate detection, indicating current flow at the lactate oxidation and reduction potentials, observed at approximately 0.50–0.55 V and −0.10 V. Figure 6c demonstrates current flow at the oxidation and reduction potentials of urea, occurring around 0.52–0.58 V and −0.12 V, respectively. As observed in Figure 6, the oxidation and reduction reactions occurring at the electrodes correspond to peaks appearing in response to cyclic voltages. Furthermore, an increase in substrate concentrations resulted in elevated oxidation peak currents, indicating enhanced electron transfer reactions. These results confirm the capability of the developed biosensor to effectively detect the three target metabolites. The interaction between the immobilized enzymes and the metabolites facilitates accurate measurements and real-time molecular-level detection of metabolic compounds secreted in sweat. Additional CV measurement results are shown in Figure 6, demonstrating the sensor’s electrochemical response according to different metabolite concentrations. In Figure 6a, glucose detection revealed distinct increases in oxidation peak currents as glucose concentration rose from 0.02 mM to 0.20 mM, indicating an enhanced electron transfer reaction at higher glucose concentrations. Similarly, Figure 6b depicts lactate detection, where increasing lactate concentrations from 20 mM to 100 mM resulted in clearly elevated oxidation peak currents, further confirming efficient metabolite-electrode interactions. Lastly, Figure 6c presents urea detection with concentrations ranging from 50 mM to 400 mM. A proportional increase in peak current was observed with increasing urea concentrations, validating the biosensor’s sensitivity to variations in urea levels. Overall, these additional CV data support the sensor’s capability to quantitatively detect different metabolic analytes in varying concentrations, highlighting its suitability for practical biosensing applications. The voltammetric profiles reflect the effect of surface modification on interfacial charge transfer. PBSE functionalization resulted in subtle changes in peak shape and intensity, consistent with the formation of an interfacial linker layer on the electrode surface. Following enzyme immobilization, a moderate reduction in peak current was observed, which is attributed to partial surface coverage and modified charge transfer kinetics at the electrode-electrolyte interface. Despite this attenuation, the redox signals remained well defined, indicating that the SWCNT network maintained sufficient electrical continuity to support stable electrochemical activity.

To further evaluate the sensing performance of the proposed biosensor, the electrochemical responses obtained from cyclic voltammetry measurements were analyzed in terms of concentration dependent signal variation. The sensor exhibited measurable electrochemical responses within the concentration ranges of 0.02–0.20 mM for glucose, 20–100 mM for lactate, and 50–400 mM for urea. These ranges correspond to physiologically relevant metabolite concentrations typically found in human sweat. Repeated measurements performed under identical experimental conditions produced stable electrochemical signals, indicating good reproducibility of the sensing interface. The stable response behavior is attributed to the conductive SWCNT network and the PBSE assisted enzyme immobilization, which together facilitate efficient electron transfer during the enzymatic reactions. Although detailed parameters such as limit of detection, response time, and long-term operational stability require additional calibration experiments and device-level testing, the present results confirm the feasibility of the proposed sensing platform for multi-analyte metabolite detection in sweat environments.

3.6. Pulsed Amperometric Detection (PAD) Measurement

Quantitative sensing performance was further evaluated using pulsed amperometric detection (PAD). Figure 7 shows the real-time current responses obtained upon sequential addition of substrate solutions (20 µL each) to the enzyme-modified electrodes. In Figure 7a, glucose concentrations ranging from 0.02 mM to 0.2 mM demonstrate stepwise increases in current signals. Despite a minor decrease observed at 0.08 mM, overall, the current continues to increase significantly at higher concentrations (0.16 and 0.2 mM), confirming the biosensor’s ability to reliably detect glucose concentrations in the given range. Similarly, Figure 7b demonstrates responses for lactic acid with concentrations from 20 mM to 100 mM. The current signal exhibits a clear upward trend as lactic acid concentrations increase, validating the sensitivity of the sensor for lactic acid detection within physiological concentration ranges. In Figure 7c, responses for urea, ranging from 50 mM to 400 mM, indicate relatively moderate yet consistent increases in current values as the substrate concentration increases. The sensor demonstrates reliable detection capability even at higher urea concentrations, supporting its practical application for the monitoring of urea levels. Collectively, these results validate the suitability and effectiveness of the developed biosensor for detecting various metabolite concentrations in real-time, highlighting its potential for practical biomedical and clinical diagnostics applications. Although the absolute current magnitude appears moderate, the biosensor exhibits reproducible and concentration-dependent stepwise responses immediately after each substrate injection, confirming reliable enzymatic activity and signal transduction.

3.7. Schematic Diagram of the Biosensor

Figure 8 presents a schematic illustration of the proposed multi-analyte nano-biosensor system designed for molecular-level detection of sweat-associated metabolites. The platform consists of multiple nano-biosensor units, each functionalized with a specific enzyme to selectively respond to a target metabolite such as glucose, lactate, or urea. These individual sensing units are integrated into a single detection module, enabling simultaneous monitoring of multiple biomarkers. Each nano-biosensor shares a common structural configuration composed of a substrate layer, an electrode structure, and a biochemical reaction interface. The substrate provides mechanical support and electrical insulation, while the electrode structure includes source, drain, and gate electrodes configured to enable electrical signal transduction. A sensing channel formed between the source and drain electrodes serves as the active region where enzymatic reactions influence charge transport behavior. Sweat samples introduced through the sample inlet are guided through a microfluidic pathway that distributes the analyte solution to each sensing unit. As the target metabolite interacts with the immobilized enzyme layer, the resulting biochemical reaction induces changes in the local electrical properties of the sensing channel. These changes are detected as variations in current under a constant applied voltage, allowing quantitative evaluation of metabolite concentration.

4. Discussion

In this study, an electrochemical nano-biosensor based on SWCNT was developed to enable the simultaneous detection of glucose, lactate, and urea in sweat. The key fabrication parameters influencing sensor performance included CNT dispersibility, enzyme immobilization efficiency, electrical conductivity, and sensitivity to substrate concentration changes. Optimization of SWCNT loading revealed a sharp decrease in electrical resistance between 0.1 and 0.4 g/L, indicating the establishment of a stable conductive network within this percolation range. Among these concentrations, 0.1 g/L provided the most uniform film with minimal agglomeration, ensuring reliable performance during subsequent PBSE and enzyme immobilization steps. This assessment was supported by comparative surface observations, where the film prepared at 0.1 g/L exhibited a more continuous nanotube network with reduced bundle clustering. In contrast, higher loadings led to localized densification of nanotube domains, whereas lower concentrations showed incomplete surface coverage, which could compromise electrical homogeneity. PBSE, serving as the electron-transfer mediator between SWCNTs and enzymes, showed increasing resistance at higher concentrations due to partial insulating effects. A concentration of 6.0 g/L achieved an optimal balance, supporting efficient enzyme attachment while maintaining sufficient electron transport across the electrode surface. All enzymes (Gox, Lox, and Ure) exhibited maximum resistance at 0.1 g/L, followed by a decrease at higher concentrations. This behavior suggests that while low enzyme levels introduce insulating effects, increased enzymatic activity and improved charge transfer at higher concentrations contribute to conductivity recovery. Gox, in particular, demonstrated the most stable electrochemical behavior at 0.2 g/L. CV and PAD analyses confirmed typical Michaelis–Menten–type electrochemical responses, with oxidation currents increasing proportionally to substrate concentration [35,36]. The sensor displayed clear sensitivity within physiologically relevant ranges for glucose (0.02–0.2 mM), lactate (20–100 mM), and urea (50–400 mM), demonstrating potential applicability for future wearable, non-invasive monitoring systems.

Compared with previously reported sweat-based electrochemical sensors, the present platform provides comparable detection ranges while integrating multi-analyte capability within a resistance-based sensing configuration. Earlier studies have frequently relied on amperometric or potentiometric sensing strategies focused on a single metabolite, such as glucose or lactate, measured using enzyme-modified electrodes [32,33,34]. These systems generally demonstrate good sensitivity but are typically designed for the selective detection of only one target metabolite. In contrast, the SWCNT-based sensing platform developed in this study enables concentration-dependent modulation of electrical resistance through enzyme-mediated reactions of multiple sweat-associated metabolites. This approach allows the simultaneous evaluation of different metabolic indicators while maintaining stable electrochemical responses. Although further validation using artificial or real sweat samples is required, the present results demonstrate a stable and scalable sensing strategy that extends beyond conventional single-analyte electrochemical sensor designs.

The multi-sensor platform presented here further enables independent detection of three metabolites on a single substrate, addressing the limitations of single-analyte systems and enhancing physiological assessment in dynamic sweat environments. Consistent operation of all enzyme-based sensors highlights the feasibility of expanding this design into integrated multi-analyte biosensing platforms. Nevertheless, this work was conducted under controlled laboratory conditions.

The present study therefore represents a controlled electrochemical validation of sweat-related metabolites rather than direct analysis of real sweat. Real sweat contains interfering components such as proteins, salts, and variable pH [10]. Future studies should therefore evaluate sensor performance using artificial sweat, human sweat samples, and long-term wearable testing. Overall, the SWCNT electrode combined with PBSE mediation provides a robust framework for sensitive and stable detection of sweat metabolites. These results underscore the potential of this biosensing platform for real-time, personalized healthcare applications.

5. Conclusions

In this study, we developed a nano-biosensor comprising single walled carbon nanotubes (SWCNTs), immobilized with distinct enzymes capable of detecting specific metabolites in human sweat. Glucose oxidase (Gox), lactate oxidase (Lox), and urease (Ure) enzymes were selected as bioreceptors to detect their respective target metabolites glucose, lactic acid, and urea. The experimental results demonstrated successful detection of glucose, lactic acid, and urea at various concentrations ranging from low to high levels, using SWCNT-based electrodes immobilized with oxidase enzymes (Gox, Lox, and Ure). Unlike conventional multi-analyte sweat sensors that rely solely on amperometric or potentiometric readouts, the present work introduces a resistance-based transduction strategy using an SWCNT network. This approach enables continuous monitoring through intrinsic conductivity modulation of the nanocarbon network, reducing circuit complexity while maintaining sensitivity across multiple metabolites. This validated the effectiveness of the developed nano-biosensor in detecting targeted metabolic analyzes. The experiments conducted in this study evaluated each metabolite independently. For further improvements, future studies could sequentially involve artificial sweat and actual human sweat samples, which would enhance the practical effectiveness of the sensor. Moreover, building on the findings of this research, future work might explore the possibility of simultaneously detecting multiple metabolites (glucose, lactate, and urea) independently within a single sample, leading to a more versatile and effective sensor platform based on carbon nanotube electrodes.