A wearable patch for continuous analysis of thermoregulatory sweat at rest

Resting (thermoregulatory) sweating occurs continuously at low secretion rates and can reflect sympathetic nervous system activity and conditions such as autonomic dysfunction, diabetes, cerebrovascular disease, Parkinson’s disease, and psychological stress. Unlike exercise- or chemically-induced sweat, at-rest sweat is generated at far lower rates (few nL min−1 cm−2 to ~100 nL min−1 cm−2 at high-secretion sites) and may better preserve diffusive equilibria of biomarkers with blood, offering potentially more representative chemistry. The research question is how to continuously collect and analyze resting sweat rate and composition noninvasively, despite low volumes and evaporation, to enable real-time monitoring of physiology during routine sedentary activities. The purpose is to engineer a wearable microfluidic-electrochemical patch that rapidly captures resting sweat, prevents evaporation, and measures both secretion rate and analytes (pH, chloride, levodopa), thereby enabling continuous, autonomous monitoring relevant to stress, metabolic disturbances (e.g., hypoglycemia), and neurodegenerative disease management. This is important because current devices primarily target higher-rate exercise/stimulated sweat and cannot reliably monitor resting sweat dynamics or rate.

Prior wearable sweat sensors focus on exercise-, thermal-, or chemically stimulated sweat with rates of tens to hundreds of nL min−1 cm−2 or higher, enabling easier collection but not representative of resting physiology. Low-rate platforms have been attempted but lacked continuous analysis and accurate resting sweat rate quantification. Clinical resting sweat assessments historically relied on bulky ventilated chamber systems or long-duration collections (up to 24 h) that yield single-point analyte reads and are prone to evaporation errors. The literature establishes regional variability of sweat rates (palms/fingers highest) and highlights clinical relevance of resting sweating to CNS defects in infants, stroke, autonomic dysfunction, diabetes, and Parkinson’s disease, as well as psychological stress. Existing hydrogel interfaces in wearables have been used for soft contact and fluid capture, but have not been engineered to minimize lag time and dead volume for rapid uptake of very low-rate thermoregulatory sweat while preserving analyte concentrations and enabling selective flow-rate sensing. Thus, a gap exists for a compact, wearable system enabling continuous, quantitative resting sweat rate and composition monitoring.

Device architecture: A multilayer wearable patch comprises (1) a PDMS microfluidic with a skin-facing collection well and a long, intertwined spiral channel (two designs: 70×70 µm and 200×70 µm cross-sections) tailored for low hydraulic resistance and ~750 nL holding capacity; (2) patterned Au electrodes on flexible PET for impedimetric sweat rate sensing (interdigitated, wheel-shaped radial spokes) and for electrochemical sensing (pH, Cl−, levodopa) on four semicircular tips aligned with the channel entrance; (3) a hydrophilic filler embedded in the collection well, consisting of a rigid patterned SU8 spacer (with ~100 µm grooves) fully coated by a thin PVA film and topped with an agarose–glycerol (AG-GLY) hydrogel film that directly contacts skin to rapidly wick sweat. Design rationale: Channel dimensions were modeled against estimated sweat gland secretory pressure for resting rates (3 nL to 1 µL min−1 cm−2) to ensure rapid advance of the sweat front while avoiding excessive hydraulic resistance that would impede gland flow. The spiral path guides the advancing sweat front over evenly spaced radial electrode spokes so each contact produces a discrete admittance pulse, enabling quantized volumetric increments independent of ionic strength. The hydrophilic filler minimizes dead volume and lag time while maintaining mechanical integrity, preventing well collapse, and reducing mixing/dispersion into the sensing channel by using only a thin hydrogel layer over a rigid spacer. Sweat rate sensing: The interdigitated wheel electrodes are driven at 100 kHz to emphasize solution resistance over capacitive effects. As sweat advances, each spoke contact yields a step increase in admittance; counting pulses and knowing the fixed volume between spokes yields volumetric increment versus time and thus flow rate. Temporal resolution depends on spoke count and channel dimensions (e.g., for 24 spokes, 200×70 µm channel, ~4–20 s resolution at 50 nL min−1 and 2–9 min at 2 nL min−1; finer in 70×70 µm design). Electrochemical sensing: pH is measured potentiometrically using a polyaniline-based ISE versus a printed Ag/AgCl/PVB reference; Cl− is measured potentiometrically with Ag/AgCl ISE versus the same reference. Levodopa is measured amperometrically using a tyrosinase enzyme electrode on Au nanodendrites with thionine mediator and a protective Nafion–TBAB micellar membrane; a small bias (0.35 V vs shared Ag reference/counter) yields a Faradaic current proportional to levodopa. Fabrication: Au electrodes patterned by photolithography/evaporation; sensor functionalization via electrochemical deposition (polyaniline), Ag/AgCl formation for chloride and reference; Au nanodendrite growth and mediator/enzyme immobilization for levodopa. Microfluidics cast by PDMS on SU8 molds, O2 plasma bonding to PET, APTES treatment; PDMS presoaked to minimize permeation-driven evaporation. Hydrophilic filler fabricated by SU8 patterning (200 µm thick), PVA coating (~10 µm), and laminated AG-GLY film (~90–130 µm) saturated with DI water. Bench characterization: Syringe pump delivered NaCl solutions (10–200 mM) at known rates (e.g., 150, 400 nL min−1) to validate that admittance pulse timing—and thus calculated flow—matches input irrespective of ionic strength. Calibration and performance for pH (pH 4–8 buffers), Cl− (25–200 mM NaCl), and levodopa (0–50 µM) were measured; flow-rate dependence assessed (levodopa shows modest flow dependence at low flow due to mass transfer limits; pH not flow-dependent). On-body studies: Devices worn on multiple body sites (shoulder, chest, bicep, wrist, abdomen, finger, thigh, calf) in healthy subjects to map resting sweat rates (24 h optical tracking of dyed sweat front). Longitudinal fingertip monitoring (two 24 h trials) tracked stress events (public speaking/teaching) alongside heart rate and ambient temperature. A diabetic subject was monitored at rest for sweat rate, heart rate, and ISF glucose (Dexcom G6) across insulin-induced glucose declines. Levodopa sensing on a healthy subject followed ingestion of boiled broad beans (100 g and 200 g; ~0.6 wt% levodopa) with fingertip patches to track sweat levodopa kinetics. Environmental conditions (temperature, humidity) and adhesives are reported; IRB approval and informed consent obtained.

• Hydrophilic filler minimizes collection lag: without filler, a 5 mm diameter, 400 µm thick well requires >2 h to fill at 300 nL min−1 cm−2, >30 min at 1 µL min−1 cm−2, and ~200 h at 3 nL min−1 cm−2; with filler, initiation occurs within minutes (≈3 min at 300 nL min−1 cm−2; <1 min at 1 µL min−1 cm−2; ~30 min at 3 nL min−1 cm−2). Flow detection down to ~2 nL min−1 demonstrated.
• Sweat rate sensing is selective and composition-independent: Admittance pulse timing is invariant to NaCl concentration (10 vs 200 mM) at fixed flow; measured flow matches syringe pump inputs (e.g., 150 and 400 nL min−1) despite ionic strength variations.
• Electrochemical sensor performance: pH sensitivity ~60 mV/pH (pH 4–8); Cl− sensitivity ~55 mV/decade (25–200 mM); levodopa sensitivity ~0.2 nA per µM with response time <20 s and effective detection down to ~3 µM; levodopa signal shows modest dependence on flow at very low rates, negligible above ~100 nL min−1.
• Regional resting sweat rates: Fingertip/palm region exhibited highest rates (~0.1–1 µL min−1 cm−2); other sites (wrist, chest, thigh, etc.) ~1–20 nL min−1 cm−2; optical and electrical measurements consistent, with reproducibility across adjacent patches and orientations.
• Activity tracking: During routine daily tasks, wrist and finger sweat rates generally tracked heart rate, rising during walking/lab work and falling during sedentary periods; pH remained stable (~6.8–7.1), Cl− stabilized around ~22 mM (finger) and ~40 mM (wrist).
• Stress detection over 24 h: During public speaking/teaching, heart rate increased (by ~28 bpm in Trial 1; ~21 bpm in Trial 2) and sweat rate sharply rose from baseline (~2.8 nL min−1 cm−2) to ~57 nL min−1 cm−2 (Trial 1) and from <2.5 to >7.5 nL min−1 cm−2 (Trial 2), indicating clear detection of stress-induced perspiration.
• Hypoglycemia-induced sweating: In a diabetic subject, insulin-induced glucose decreases were accompanied by increased sweat rate at rest (e.g., up to ~5 µL min−1 cm−2 when glucose dropped below ~90 mg/dL), with heart rate relatively unchanged.
• Levodopa in sweat: After broad bean ingestion, levodopa appeared in sweat ~20 min post-intake; peak at ~35 min. Peak concentrations scaled with dose: ~13 µM (100 g) and ~35 µM (200 g), decreasing thereafter; no significant signal with foods low in levodopa.

The study demonstrates a compact wearable microfluidic-electrochemical patch that overcomes the principal barriers to resting sweat monitoring—evaporation, low secretion volumes, and slow collection—by trapping sweat in a low-resistance channel and accelerating uptake with a hydrophilic filler. The impedimetric spoke design decouples flow measurement from ionic concentration, enabling accurate, continuous secretion-rate tracking. Together with multiplexed sensing (pH, Cl−, levodopa), the platform provides realtime insights into physiology during sedentary activities. On-body trials show that resting sweat rate correlates with activity and acutely with psychological stress, enabling quantitative identification of stress events across a day. In metabolic monitoring, increased sweating accompanies insulin-induced hypoglycemia, suggesting utility for hypoglycemia detection. Levodopa kinetics in sweat following dietary intake indicate dose-dependent appearance and clearance, pointing toward noninvasive drug monitoring for Parkinson’s disease management. These findings directly address the initial hypothesis that resting sweat can provide unique, continuous physiological information and that a properly engineered wearable can capture both secretion rate and chemistry without external sweat stimulation.

A wearable patch integrating microfluidics, a hydrophilic filler, and multiplexed electrochemical sensors enables near-real-time, continuous analysis of thermoregulatory sweat at rest. The device rapidly captures low-volume sweat, quantifies secretion rate independent of ionic composition, and measures pH, chloride, and levodopa within physiologically relevant ranges. On-body studies validate detection of activity-related changes, stress events, hypoglycemia-associated hyperhidrosis, and dose-dependent levodopa kinetics. This work expands sweat sensing to sedentary contexts and opens avenues for noninvasive monitoring of autonomic function, metabolic disturbances, and precision pharmacotherapy. Future research should include larger population studies to establish quantitative correlations between resting sweat metrics and clinical endpoints (e.g., blood glucose, cortisol, plasma levodopa), refinement of device ergonomics for joint regions, expansion to additional biomarkers, and algorithm development to integrate sweat rate with analyte signals for robust interpretation.

• Potential overestimation of local sweat rate due to compensatory sweating under the covered area; requires further study.
• Analyte time lag and dispersion between secretion at skin and arrival at sensors (e.g., ~3 min at fingertip; longer at lower rates), limiting instantaneousness of composition measurements.
• Levodopa sensor shows flow-rate dependence at very low flow due to mass transfer limitations; conversion of current to concentration must account for instantaneous sweat rate.
• Patch spanning finger joints can inhibit bending; more compliant substrates may be needed for certain placements.
• Environmental conditions (temperature, humidity) and inter-individual variability affect resting sweat rates; generalizability requires larger cohorts.
• Hydrogel and filler design, while minimizing mixing, may still introduce small delays in analyte replacement within the sensing zone.
• Gravimetric comparisons are confounded by evaporation; while the patch mitigates evaporation, cross-validation with alternative gold-standard methods is limited.
• Current analyte panel is limited; broader biochemical coverage and long-term sensor stability in varied conditions remain to be validated.