Wearable Sweatband Sensor Platform Based on ... - ACS Publications

Article pubs.acs.org/ac

Wearable Sweatband Sensor Platform Based on Gold Nanodendrite Array as Efficient Solid Contact of Ion-Selective Electrode Shuqi Wang, Yongjin Wu, Yang Gu, Tie Li, Hui Luo, Lian-Hui Li, Yuanyuan Bai, Lili Li, Lin Liu, Yudong Cao, Haiyan Ding, and Ting Zhang* i-Lab, Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences, 398 Ruoshui Road, Suzhou, 215123, People’s Republic of China S Supporting Information *

ABSTRACT: As chemical sensors are in great demand for portable and wearable analytical applications, it is highly desirable to develop an all-solid-state ionselective electrode (ISE) and reference electrode (RE) platform with simplicity and stability. Here we propose a wearable sensor platform with a new type of allsolid-state ISE based on a gold nanodendrite (AuND) array electrode as the solid contact and a poly(vinyl acetate)/inorganic salt (PVA/KCl) membrane-coated allsolid-state RE. A simple and controllable method was developed to fabricate the AuNDs on a microwell array patterned chip by one-step electrodeposition without additional processing. For the first time, the AuND electrodes with different real surface area and double layer capacitance were developed as solid contact of the Na+-ISE to investigate the relationship between performance of the ISE and surface area. As-prepared AuND-ISE with larger surface area (∼7.23 cm2) exhibited enhanced potential stability compared to those with smaller surface area (∼1.85 cm2) and to bare Au ISE. Important as the ISE, the PVA/KCl membrane-coated Ag/AgCl RE exhibited highly stable potential even after 3 months’ storage. Finally, a wearable sweatband sensor platform was developed for efficient sweat collection and real-time analysis of sweat sodium during indoor exercise. This all-solid-state ISE and RE integrated sensor platform provided a very simple and reliable way to construct diverse portable and wearable devices for healthcare, sports, clinical diagnosis, and environmental analysis applications.

A

Despite abundant attractive merits (compact size, small sample volume, low cost, easy maintenance, simple operation, etc.) of the all-solid-state ISE presented, the ubiquitous challenges of potential instability and poor reproducibility still remain and limit its practical applications. As the first proposed coated-wire-based all-solid-state ISE eliminates cumbersome liquid contacts (inner filling solutions) between the sensing membrane and its solid contact, the reliable potential response is strongly affected due to the interrupted ion-to-electron transduction at the “blocked” solid contact/membrane interface.10,11 Moreover, an unintentionally formed thin water layer beneath the ion-selective membrane could cause potential drifts because of its extremely small volume that is very sensitive toward pH and ion flux changes in solutions.12 To achieve efficient transduction of the charge carrier from ions to electrons and solve the undesired water layer problems, various materials have been employed as the solid contact of all-solidstate ISE. Among all the solid contacts, capacitive mechanismbased nanostructured materials exhibit some inherent advantages over conducting polymers, such as insensitivity to light, pH, etc., which is quite important for in situ environmental and

ll-solid-state ion-selective electrode (ISE) and reference electrode (RE) have emerged as indispensable sensing components in the growing field of wearable and portable sensors for healthcare, clinical diagnosis, and environmental analysis applications.1,2 The most important reason is their compatibility with advanced microfabrication technologies to manufacture miniaturized chips comprising multifunctional sensors and wearable electronics. In recent years, several groups have made state-of-the-art improvements in combining all-solid-state ISEs with wearable electronics for real-time onbody monitoring of (bio)chemical biomarkers in sweat. For example, Javey and co-workers1 introduced an ISE and RE integrated wearable sensor for electrolyte detection in sweat; Heikenfeld and co-workers3 designed an adhesive radio frequency identification (RFID) bandage-based ISE and RE for sweat Na+ detection; Wang and co-workers4 developed a temporary tattoo-based ISE and RE for epidermal Na + monitoring; and Diamond and co-workers5−9 published successive pioneering works on integrating microfluidic devices with various chemical sensors for wireless sensor networks (WSNs). However, most of these works focused on integration of the wearable sensors, while less effort has been made to design new types of all-solid-state ISE/RE and to improve their performance, especially electrode stability. © 2017 American Chemical Society

Received: April 27, 2017 Accepted: September 1, 2017 Published: September 1, 2017 10224

DOI: 10.1021/acs.analchem.7b01560 Anal. Chem. 2017, 89, 10224−10231

Article

Analytical Chemistry

Figure 1. Schematic of chip fabrication process and AuND solid-contact-based all-solid-state Na+ sensor chip. (A) Photolithography (Pl)-based chip fabrication process. (B) Schematic profile of AuND-based all-solid-state ISE. (C) Optical image of Na+ sensor chip.

physiological measurements.13−15 Most nanomaterial-based solid contacts form an asymmetrical electrical double layer at the membrane/solid contact interface, which can convert one side of ions charge to the other side of electron charge via the electrical capacitor.12 The interfacial potential of ISE strongly relies on the quantity of charge in the electrical double layer. Therefore, various nanomaterials with large surface area and high double layer capacitance have been widely investigated as effective solid contacts in recent years, including threedimensional microporous carbon,16 graphene,17 carbon nanotubes,18 and gold nanoparticles.19 These attractive nanomaterials could efficiently improve ion-to-electron conductivity and stabilize potential drifts. However, not only was the synthetic process complicated but also the adhesion strength to electronic conductors was weak. Moreover, the relationship between quantity of interfacial contact area and response characteristics was not sufficiently addressed.12 Therefore, it is highly desirable to develop a high-quality solid contact of ISE with a simple and controllable synthetic method that can meet the need for simplicity and stability. Reference electrode (RE) plays a crucial role in all electrochemical sensor systems, as it can establish a reliable, stable, and reproducible electrical potential in an electrochemical measurement.10,20 In recent years, different strategies have been reported for fabrication of all-solid-state Ag/AgCl RE for the miniaturized electrochemical sensor. Inorganic salt combined with polymers [poly(vinyl chloride), PVC; polypyrrole, PPy; poly(vinyl butyral), PVB; and poly(vinyl acetate), PVA] were widely utilized as the RE membrane coated onto Ag/AgCl to create a high and constant chloride concentration instead of liquid electrolytes.21−25 All of these polymer/ inorganic salt composite membrane-coated Ag/AgCl REs exhibit simplicity and require less maintenance, so this become the most promising approach for mass fabrication of electrochemical sensors. However, their performance still lacks high potential stability in terms of continuous, prolonged, and intensive usage. Thus, there is an imperative need to develop an all-solid-state RE compacted on a miniaturized chip and to investigate its analytical qualities. In this paper, we present a simple, reliable, and controllable method to fabricate a gold nanodendrite (AuND) array on a microwell patterned chip in situ by one-step electrodeposition

without additional processing. The AuND electrode exhibited three-dimensional branched structure and displayed obvious improvements in surface area and hydrophobicity. AuND electrodes with different real surface areas and double layer capacitance were employed as the solid contact of the ISE to evaluate performance characteristics for the first time. Additionally, a PVA/KCl membrane-coated all-solid-state RE was fabricated on the miniaturized chip. Finally, a wearable sweatband platform was designed to harvest and analyze [Na+] in sweat continuously.

■

EXPERIMENTAL SECTION Reagents. ISE membrane components Selectophore-grade sodium ionophore X, tetrakis[3,5-bis(trifluoromethyl)phenyl]boron sodium (NaTFPB), bis(2-ethylhexyl) sebacate (DOS), poly(vinyl chloride) (PVC of high molecular weight), and gold chloride trihydrate (HAuCl4·3H2O) were obtained from Sigma−Aldrich (Shanghai) Trading Co., Ltd. RE membrane components vinyl acetate, photoinitiator 2,2-dimethoxy-2phenylacetophenone (DMPP); tetrahydrofuran (THF, 99.9%), and lithium acetate (LiAcO) were obtained from Aladdin (Shanghai) Co., Ltd. Ag/AgCl ink (product no. 011464) for reference electrode was purchased from ALS (Japan) Co., Ltd. All other chemicals were commercially available and of analytical reagent grade. All reagents were used as received. The aqueous solutions were prepared freshly with deionized water (18.2 MΩ·cm). Chip Design and Fabrication. Figure 1A shows the geometry and fabrication process of a typical chip design containing a reference electrode (RE) zone (3.0 mm in diameter) in the center and two ISE zone (3.0 mm in diameter). The microwell array pattern (10 μm in diameter and 100 μm in interval, distributed equally as hexagonal) in the ISE zone for electrodeposition of AuND structures was fabricated by a simple photolithography process. The detailed fabrication process is discussed in section S1 in Supporting Information. Electrodeposition of AuND Solid Contact on Chip. The chip was first cleaned by plasma and ethanol, rinsed with deionized water, and dried in a nitrogen stream to remove the residues. Then electrochemical cleaning was applied in 0.5 M H2SO4 by cyclic voltammetry method in the range 0.2−1.6 V to produce a homogeneous electrode surface. Electrodeposition 10225


Article


measurements, the obtained real-time sweat [Na+] was calculated from the calibration curve. Several healthy and energetic volunteers participated in an indoor cycling exercise, which is an easy way to measure sweating.

was performed by using a three-electrode system comprising the chip as working electrode, a platinum wire as counter electrode, and Ag/AgCl (3 M KCl) as reference electrode. Electrodeposition was carried out in 5 mM HAuCl4 and 0.5 M H2SO4 at −0.6 V potential by a Gamry Reference 600+ potentiostat. The obtained AuND electrode chip was rinsed with deionized water, dried with a nitrogen stream, and stored in a glass desiccator before use. Preparation of Na+ All-Solid-State Selective Sensor on Chip. The Na+ ion-selective membrane (ISM) (as shown in Figure 1C) consisted of 1.0 wt % Na+ ionophore X, 0.6 wt % NaTFPB, 33 wt % PVC, and 65.4 wt % DOS. The membrane cocktail was prepared by dissolving 100 mg of ISM components in 1 mL of THF (99.9%). ISE was then prepared by dropcasting 4 μL of ISM cocktail onto AuND or Au (for comparison) electrode three times at 10 min intervals, and finally the electrodes were left to dry overnight in a glass desiccator. The prepared ISEs are denoted as AuND-ISE and Au-ISE, respectively. Preparation of All-Solid-State Ag/AgCl Reference Electrode on Chip. The all-solid-state Ag/AgCl reference electrode was made by using a polymer/inorganic salt composite membrane on top of Ag/AgCl for direct contact with solution. The bottom layer of Ag/AgCl was first prepared by drop-casting 4 μL of Ag/AgCl ink onto the reference electrode zone on chip, which was and left to dry overnight in a glass desiccator. Meanwhile, milled and dried potassium chloride (KCl) powder was prepared for use by annealing for 30 min at 450 °C and carefully grinding in a mortar. To prepare the polymer/inorganic salt composite membrane, poly(vinyl acetate) (PVA) and KCl composite was selected. Vinyl acetate monomer and photoinitiator DMPP were mixed in a tube at a ratio of 1:0.01 by weight, and prephotopolymerization in a UV case (UV lamp: 185 nm +254 nm, 250 W) took place for 5 min. Then 50 wt % KCl powder was added and well mixed into a white viscous liquid, following by drop-casting of the asprepared mixture onto Ag/AgCl electrode and irradiation with UV light until the membrane was sufficiently hard (as shown in Figure 1C). The prepared reference electrode is denoted as PVA/KCl-Ag/AgCl reference electrode. Electrochemical Characterization of Chip. Electrochemical characterization was carried out with a Gamry multichannel electrochemistry testing system (Interface 1000B and Reference 600+, Gamry Instruments, Inc.). Potentiometric measurements were recorded against a double-junction Ag/AgCl/KCl (3 M) reference electrode containing 1 M LiAcO as ionic salt bridge. Ionic activity coefficients were calculated according to the Debye−Hückel equation. Cyclic voltammetry (CV) was carried out in a threeelectrode system with a platinum wire as the counter electrode and Ag/AgCl (3 M KCl) as the reference electrode. On-Body Sweat Analysis. A simple on-body sweatband fluidic prototype platform was proposed for efficient sweat collection and wearable analysis during indoor exercise. The main component was a commercially available sweatband (made from silicone) containing a channel (tank) for chip loading and sweat flowing through the chip. The sweatband platform could be located on the forehead for real-time sweat measurements where the sweat rates were high and convenient for trials during exercise. Sensors were calibrated before and after on-body measurements to avoid possible potential changes that may affect the signal reading after exposure to sweat sample. After on-body

■

RESULTS AND DISCUSSION Characterization of AuND Solid Contact. Gold nanodendrites were fabricated in situ on the chip by the spaceconfined electrochemical deposition method. As shown in Figure 1A, by simply designing the microwell array as the substrate, AuND fabrication could be well-controlled by onestep electrodeposition. The main mechanism is that the microwell promotes the edge effect to produce nonequilibrium conditions for the electrocrystallization process. Gold dendritic structures are formed by the interplay of nonequilibrium conditions and the driving force; thus the gold atoms selforganized into a hierarchical architecture with branches and particles. The detailed mechanism was reported in our previous works.26 The morphology of AuNDs on the chip was characterized by use of a field-emission scanning electron microscope (FE-SEM; Hitachi-s4800). As shown in Figure 2A,

Figure 2. (A) Scanning electron microscopic (SEM) images of AuNDs prepared in situ on chip by electrodeposition with different deposition times of (a) 60 s, (b) 180 s, and (c) 300 s, respectively. (B, C) Cyclic voltammogram (CV) curves of AuND electrodes with different deposition times (corresponding to a, b and c in panel A), recorded in (B) 0.5 M H2SO4 or (C) 0.1 M KCl vs Ag/AgCl RE. (B, Inset) Real surface area calculated by integrating the adsorption charges of reduction peaks.

regular array patterns of AuNDs were successfully formed on a microwell patterned chip. As can be seen, only one single threedimensional treelike AuND grew in one microwell. Upon simply increasing the deposition time, the treelike AuNDs grew denser and larger with more branches. High-resolution SEM images of the AuNDs shown in Figure S1 revealed that the dendritic structure was composed of small particles with a size range of 30−150 nm. Notably, although the overall sizes of single AuNDs were different with increasing deposition time, their fine nanostructures were similar in both size and morphology. 10226


Article

Analytical Chemistry The proposed method provided a very simple and effective way to fabricate a solid contact with controlled large surface area. Electrochemical characterizations of AuND electrodes with different deposition times confirmed their differences in real surface area. As Figure 2B shows, cyclic voltammogram (CV) curves of AuND electrodes with different deposition times, and of a bare Au electrode (3 mm in diameter) for comparison, were scanned in H2SO4 solution with a potential range from 0.2 to 1.6 V. All the AuND electrodes (a, b and c) and bare Au electrode exhibited well-defined reduction peaks for gold oxide. The real surface areas were calculated by integration of reduction peaks, with a specific charge of 390 μC/cm2 assumed, according to previous reports.27 As shown in Figure 2B (inset), real surface areas increased dramatically upon going from bare Au electrode (flat surface) to AuND electrode with increasing deposition time. Electrical double layer capacitance was also investigated by CV in a potential window of 0.1−0.6 V. The pronounced nearrectangular-shaped CV loops appearing in Figure 2C were indicative of the capacitive charging process of the AuND electrode. Since the double layer capacity originated from the electrode surface area (according to the plate capacitor equation C = ε0εA/d, where ε0 and ε are dielectric constants of vacuum and medium, A is plate area, and d is the medium thickness). The results in Figure 2C proved that the AuND electrode with a larger surface area possessed enhanced double layer capacitance. Therefore, it could be concluded that AuND electrodes with different surface areas and high double layer capacitance were well fabricated on chip by simply controlling the parameters of microwell patterns on the substrate (discussed in section S1 in Supporting Information) and the electrodeposition time. Due to their large surface area and high double layer capacitance, AuNDs could be used as a solid contact interface for developing stable and reliable all-solidstate ISE. The simplicity and controllable fabrication of nanomaterials with different surface areas were favorable not only for ISE miniaturization but also for evaluating their differences in potentiometric response properties. 3.2. Potentiometric Performance of AuND-ISE. AuND electrodes with different surface areas were used to fabricate the all-solid-state polymeric membrane Na+-ISE, and a bare Au electrode (3 mm in diameter) -based Na+-ISE was also prepared for comparison. After conditioning in 1 mM NaCl solution overnight, the potentiometric responses of AuND-ISE and Au-ISE were measured by successively increasing the concentration of Na+. Figure 3A,B shows typical potential response curves and the calibration curves of AuND-ISE and Au-ISE. Both of them showed a stable and fast response within 10 s. The linear response range was from 10−6 to 10−1 M with a slope of 56.58 ± 1.02 mV/decade (n = 5, R2 > 0.998) for AuND-ISE and 56.43 ± 1.17 mV/decade (n = 4, R2 > 0.998) for Au-ISE. The detection limits (calculated by tangent lines shown in Figure 3B) were 0.8 × 10−6 M for AuND-ISE and 2.5 × 10−6 M for Au-ISE. More data for the sensitivity of AuNDISEs with different solid contact surface areas and of Au-ISEs are summarized in Figure S2A. Their sensitivities were close to each other and showed no specific relationship with surface area of the solid contact. Additionally, the selectivity coefficients of AuND-ISE and Au-ISE were evaluated by using the separated solutions method. As shown in Table S1, the selectivity coefficients were comparable with each other. The results indicated that the selectivity of ISE was not influenced by the type of solid contact used but was mainly determined by

Figure 3. (A) Typical time-dependent potential responses of AuNDISE (area = 1.70 cm2) and Au-ISE (area = 0.1 cm2). (B) Calibration curves vs double-junction Ag/AgCl/KCl (3 M) containing a 1 M LiAcO salt bridge. (Inset) Calibration curves for AuND-ISE (black line) and Au-ISE (red lines). (C) Typical chronopotentiometry curves, measured in 0.1 M NaCl solution, of (a) Au-ISE, area = 0.10 cm2; (b) AuND-ISE, area = 0.52 cm2; (c) AuND-ISE, area = 1.70 cm2; (d) AuND-ISE, area = 7.23 cm2. (D) Water layer test for AuND-ISE (1.85 cm2) and Au-ISE. Potential responses in A−D have been vertically shifted for clarity of presentation.

the ion-selective membrane, which is in agreement with literature results.28 Potential stability of the all-solid-state ISE was one of the most important criteria for evaluating its quality. Constantcurrent chronopotentiometry technique has been widely used to evaluate the short-term potential stability of different types of ISE. A current of ±1 nA was applied on ISE to represent a relatively harsh condition compared to the typical measurement using a potentiometer of high-input impedance. The chronopotentiometric E−t curves of AuND-ISEs with different surface areas and of Au-ISE are shown in Figure 3C. Two main features could be observed: a potential jump when the applied current changed the direction and a potential drift at longer times. The potential jump could be used to calculate the total resistance of the ISE (which is dominated by bulk resistance of the polymeric PVC-based membrane) according to Ohm’s law (R = E/I, where E represents potential change upon applied current I). In Figure 3C, the calculated total resistances were 3.25 MΩ for Au-ISE (Figure 3C, curve a) and 2.65, 2.25, and 2.40 MΩ for AuND-ISEs (Figure 3C, curves b−d). The mean value for AuND-ISEs was 2.78 ± 0.68 MΩ (AuND-ISEs with different surface area) and the Au-ISE value was 3.25 ± 0.45 MΩ. The total resistances of AuND-ISE and Au-ISE were similar, which indicated that the ISE membrane dominated the total resistance. The potential drift of the ISE was calculated from the slope of the E−t curves (ΔE/Δt). In Figure 3C, the potential drifts were 1.678 mV/s (curve a, Au-ISE, area = 0.10 cm2), 0.479 mV/s (curve b, AuND-ISE, area = 0.52 cm2), 0.135 mV/s (curve c, AuND-ISE, area = 1.70 cm2), and 0.040 mV/s (curve d, AuND-ISE, area = 7.23 cm2). More data are summarized in Figure S3A. Obviously, it could be concluded from these results that the potential stability of the ISE was dramatically improved by using AuND as the solid contact. The high stability of the AuND-ISE is mainly attributed to the 10227


Article


Figure 4. Long-term stability of AuND-ISE over 2 month’s storage under ambient conditions: (A) potential response curves, (B) calibration curves, and (C) sensitivities of AuND-ISE.

shown in Figure S3B. The calculated potential drifts were 4.66 mV/h for Au-ISE (curve a, area 0.10 cm2), 2.04 mV/h for AuND-ISE (curve b, area 0.52 cm2), 1.20 mV/h for AuND-ISE (curve c, area 1.85 cm2), and 0.22 mV/h for AuND-ISE (curve d, area 7.23 cm2). The potential drift of AuND-ISE was about 22-fold lower than that of Au-ISE (calculation and discussion shown in section S5 in Supporting Information). The results indicated that the stability was greatly improved by increasing surface area and surface hydrophobicity of the solid contact, which was in good agreement with the results obtained by chronopotentiometry. Moreover, the potential drift obtained from AuND-ISE with a surface area of 7.23 cm2 was 0.22 mV/ h, which was much lower than that from nanoporous goldbased solid contact (0.43 mV/h)30 and carbon nanotube-based solid contact (0.49 mV/h).31 More comparison data are shown in Table S2. Additionally, the long-term stability of AuND-ISE was investigated. Potentiometric measurements of AuND-ISE toward increasing concentration of Na+ were carried out every month, the results are shown in Figure 4. The potential response curves and calibration curves of AuND-ISE nearly overlapped, and their sensitivities were not significantly changed. Characterization of PVA/KCl-Ag/AgCl Solid Reference Electrode. The PVA/KCl-Ag/AgCl RE was fabricated on a miniaturized chip as shown in Figure 1C. The PVA/KCl composite film was about 500 μm thick (restricted by the PI reservoir), and it could remain hard and white after long-term measurements and storage in solution. The as-prepared PVA/ KCl-Ag/AgCl RE was conditioned in 3 M KCl solution to stabilize the response. When the RE first made contact with the solution, a small potential drift of a few millivolts lasting for about 1 h was observed; a typical conditioning curve is shown in Figure 5A. Afterward, the signal gradually turned stable within few hours and the potential drift level was below 0.2 mV for the next 60 h (Figure S5A). Remarkably, the conditioning time was much shorter compared to previous studies.23 After conditioning in 3 M KCl, the medium-term potential stability of RE was tested in a low concentration of 1 mM KCl. As shown in Figure S5B, the potential drift was 0.056 ± 0.028 mV/h (n = 3) for 15 h of continuous potentiometric measurements. The as-prepared RE was stored under ambient conditions and tested after 3 months. As shown in Figure 5A, a very short conditioning duration of 20 min was enough to stabilize the potential. Although the stabilized potential of PVA/KCl-Ag/AgCl RE gave a positive shift of 4 mV after 3 months’ storage, the stability of the RE was still outstanding for long-term usage (as it is a common practice to calibrate RE and ISE). The short-term conditioning duration and excellent stability may be attributed to the thin film of PVA/KCl

electrical double layer capacitor of the AuND that provided efficient ion-to-electron transduction at the membrane/solid contact interface. Moreover, the stability performance was clearly improved by increasing the surface area of the solid contact. Therefore, more stable AuND-ISE could be obtained by simply increasing electrodeposition time or decreasing interval distance of the microwells to achieve an AuND solid contact with larger surface area and higher double layer capacitance. Formation of a water layer between the solid contact and the ion-selective membrane was proved to be a main source of long-term potential instability of the solid contact ISE. The presence of the water layer could cause potential drifts, mechanical failure, and chemical hysteresis due to the distribution of ions or gases (O2 and CO2) into this thin water layer.12 Therefore, the method to test the water layer by conditioning the ISE alternately in solutions of the primary and interfering ions for several hours has been widely used. After being conditioned in 1 mM NaCl solution overnight, AuNDISE and Au-ISE were alternately measured in 0.1 M NaCl (primary ions), 0.1 M KCl (interfering ions), and back in 0.1 M NaCl solution for hours. As shown in Figure 3D, a positive potential drift upon changing from Na+ to K+ solution and a negative potential drift when changing from K+ back to Na+ solution were clearly observed for Au-ISE while not observed for AuND-ISE. According to previous work,29 the results suggested the presence of a water layer in Au-ISE. Since all the ion-selective membranes could take up water to some extent, these results indicated that the AuND solid contact could effectively reduce formation of the water layer, which was attributed to the rough morphology of the AuNDs that could improve surface hydrophobicity of the solid contact to avoid water accumulation (experimental data and a description of surface hydrophobicity of AuNDs are given in section S6 and Figure S4 in Supporting Information). In light of the good potential stability and the reduced water layer provided by using AuNDs as the solid contact and to demonstrate the practical performance of AuND-ISE, several tests such as potential reproducibility and medium- and longterm stability were performed. Figure S2B shows the reproducibility of AuND-ISE and Au-ISE by alternatively measuring their dynamic responses in 10−3 M and 10−2 M NaCl solution. The standard deviations of the AuND-ISE potential response values were ±0.34 mV in 10−3 M NaCl and ±0.73 mV in 10−2 M NaCl (n = 5), while the standard deviations of Au-ISE were ±7.21 and ±6.74 mV (n = 5). The medium-term stability of AuND-ISEs with different surface areas and of Au-ISE was evaluated by continuously measuring their potential responses in 0.1 M NaCl for more than 14 h, as 10228


Article


membrane. The simple fabrication process, short conditioning duration, and capability for miniaturization were very attractive for analytical measurements in portable and wearable electrochemical devices. On-Body Trials. Sweat sodium has been a very useful biomarker for routine screening of cystic fibrosis in newborns33 and for evaluating hydration or heat stress in athletes, soldiers, first responders, and others working in extreme conditions.2,34 A wearable monitoring platform could provide real-time and continuous information for physicians or individuals to evaluate risks to health over time, for example, the different states of dehydration (hypertonic, hypotonic, and isotonic dehydration). Thus, it is important to develop a wearable all-solid-state ISE and RE platform for reliable sweat [Na+] monitoring. As shown in Figure 6A−C, the proposed sweatband fluidic platform could be comfortably worn on the forehead of a human subject for efficient sweat collection and wearable analysis during indoor exercise. The sweat appearing on the forehead of a subject could efficiently flow into the channel and be transported to the chip surface, driven by gravity and the oscillation of the movement. Before and after commencing onbody trials, calibration of the integrated all-solid-state ISE and RE was performed with NaCl standard solutions in a concentration range of 10−4−10−1 M to avoid possible potential changes that may affect the signal reading after exposure to sweat sample. To establish a stable baseline before sweat arrival and wetting the ISE and RE surface, a 10 μL drop of 10−4 M NaCl was added at the surface of the chip (without this procedure, the chip would produce a very noisy potential signal before the sweat passed through, as shown in Figure S6A). Real-time sweat sodium monitoring was performed on a subject during indoor exercise of cycling. The overlapping calibration curves shown in Figure 6D indicated that the as-prepared integrated all-solid-state ISE and RE were very stable even after several hours’ exposure to the sweat sample. Figure 6E shows a real-time sweat [Na+] profile of a subject exercising for 1.5 h of indoor cycling. The perspiration began after about 10 min of exercise with a constant load (warm-up stage). At the beginning of the sweating, a rapid increase in Na+ concentration was observed, which indicated the arrival of the initial sweat. After several minutes, the signal gradually turned stable and exhibited a small decrease; this phenomenon was also observed in other on-body trials in Figure S6. After 5 min of rest and 100 mL of water intake, a high intensity of exercise with an increased load was carried out by the subject. Another increase in sweat Na+ concentration was observed when the subject went through excessive sweating. The phenomenon might indicate dehydration of the subject, which could be an important sign for the subject to take a rest and drink water. The trends were similar to the previous report by Gao et al. 1 The sweat Na+ concentration levels of several subjects during indoor cycling were in a range of 17−40 mM, which was in the normal physiological range (18.2−70.8 mM).34 In the near future, the proposed sweatband platform will be combined with wearable electronics for signal conditioning and processing, incorporate more ISEs for analysis of other electrolytes analysis (such as K+ and NH4+), and be validated through larger-scale on-body trials.

Figure 5. Potentiometric responses of PVA/KCl-Ag/AgCl solid reference electrode vs double-junction Ag/AgCl/KCl (3 M) containing a 1 M LiAcO salt bridge in (A) 3 M KCl solution and (B) different concentration solutions of ions range from 10−1 M to 10−6 M. (C) pH response and (D) light sensitivity of PVA/KCl-Ag/ AgCl RE. Potential responses in panels B and C have been vertically shifted for clarity of presentation.

composite that needs less time to build and maintain a constant ion flow at the membrane/solution interface and to the extremely low leakage rate of ions into sample solutions.32 The influence of electrolytes, pH, and light on RE potential stability was evaluated. Figure 5B showed the potentiometric responses of PVA/KCl-Ag/AgCl RE toward different salts over a wide concentration range from 10−1 to 10−6 M. For KCl, NaCl and KNO3, the potential changes were negligible in the range from 10−1 to 10−3 M, while in their low concentration range a small potential variation of 2 mV appeared. For Na2SO4, a large potential variation was observed initially but could gradually go back to a small potential variation of 2 mV. Indeed, when potential values were studied in the whole concentration range in Figure 5B, their calibration slopes were always lower than 1 mV/decade, which indicated that the PVA/ KCl-Ag/AgCl RE exhibited low potential variability toward ions of different charges and strength. The influence of pH on PVA/KCl-Ag/AgCl RE response was investigated by adding small aliquots of acidic (HCl) and basic (KOH) stock solution, respectively, during potentiometric measurements in 0.01 M KCl solution. The results can be seen in Figure 5C. Larger potential changes were observed as HCl and KOH were added to concentrations of 10−2 M (pH ≈ 2) and 10−3 M (pH ≈ 11), respectively, while very small potential changes (