Ultra-conformal drawn-on-skin electronics for multifunctional motion artifact-free sensing and point-of-care treatment

The study addresses the challenge of reliably capturing physiological and physical signals from the skin using wearable electronics, which are often compromised by motion artifacts due to weak adhesion and imperfect conformal contact at the skin–electrode interface. The authors propose a new platform, Drawn-on-Skin (DoS) electronics, in which functional electronic inks (conductive, semiconductive, and dielectric) are directly drawn onto the skin using simple stencils and ballpoint pens. This approach aims to provide ultra-conformal, robust, and stretchable interfaces that maintain integrity during motion, thereby enabling motion artifact-free sensing. The work explores the materials, device design, fabrication, characterization, and application of DoS electronics for multifunctional sensing and point-of-care treatment, including demonstration of immunity to motion artifacts and accelerated wound healing via electrical stimulation.

The paper situates DoS electronics within the context of wearable and printed bioelectronics that typically use soft, flexible, or stretchable patch-based formats. Prior technologies, while enabling skin contact, are susceptible to motion artifacts arising from mechanical disturbance at the electronics–skin interface, often due to inadequate adhesion or conformability. The authors reference advances in ultrathin, serpentine mesh electrodes, printed electronics, and flexible systems, including efforts to minimize motion and sweat artifacts. Despite these advances, existing approaches often require dedicated fabrication equipment, auxiliary hardware, computational artifact mitigation, or still suffer under sweat and motion. Electrolytically gated transistors and various on-skin sensing modalities are also discussed as related technologies. The literature underscores the need for a simple, customizable, and robust interface capable of active electronics and multifunctional sensing without motion artifacts.

Materials and inks: Conductive ink was formulated by mixing silver (Ag) flakes with PEDOT:PSS solution (optimized 1:2 Ag flakes:PEDOT:PSS). The semiconducting ink comprised poly(3-hexylthiophene-2,5-diyl) nanofibrils (P3HT-NF). The dielectric was an ion gel ink. Preparation details are provided in Methods and Supplementary Information. Drawing process: A stencil (Kapton and clear Magic Tape) cut by a Silhouette Cameo was adhered to the target skin. Inks were drawn within stencil outlines using a modified 1 mm-tip ballpoint pen. A liquid meniscus formed between pen tip and surface allows deposition without the tip contacting skin; shear from pen motion spreads ink. Draw speed ~10 mm/s. Solvent evaporated at room temperature for ~5 minutes, then stencil removed, leaving a dry film. Multiple passes enable thickness control (hundreds of nm to ~10 µm). Line width controlled by pen tip (down to ~300 µm) and stencil spacing (down to ~200 µm). Imperfections were corrected by redraw. Substrates and replicas: For mechanical/electrical characterization, inks were drawn on PDMS substrates and PDMS skin replicas (surface treated to be more hydrophilic for adhesion in tests; actual skin is partially hydrophilic and does not require treatment). SEM and AFM were used to characterize morphology. Biaxial stretch, twist, and poke tests assessed mechanical robustness. Device fabrication on skin replica: Using stencils, the team fabricated: electrolyte-gated thin-film transistors (Ag-PEDOT:PSS as S/D/gate, P3HT-NF channel, ion gel dielectric), two-terminal strain and temperature sensors (Ag-PEDOT:PSS and P3HT-NF), and resistive heaters (Ag-PEDOT:PSS). Channel lengths were a few hundred microns. Electrical characterization included I–V and transfer curves under 0% and 30% strain, and cyclic tests. Temperature–resistance characteristics established thermistor behavior and β constant. Heaters were calibrated by applying DC biases (1–7 V) and monitoring temperature with an IR camera. On-skin devices: On porcine and human skin, Ag-PEDOT:PSS ink formed heaters (demonstrated via IR imaging), interconnections in a functional RC circuit driving an LED, interdigitated skin hydration sensors (impedance/capacitance-based), and electrophysiological (EP) electrodes for EMG and ECG. Hydration sensor calibration used a commercial hydration meter, relating capacitance to % hydration. Devices could be encapsulated for wet environments while retaining function. EP data acquisition (wired): Skin was optionally prepped with alcohol swab. DoS EP sensors (15 x 15 mm) were drawn; conductive acrylic tape (ARclad 8001-77) provided connection pads for leads to an Intan RHD2216 amplifier and RHD2000 interface. ECG: two electrodes on wrists; EMG: one on forearm and one on bicep. Sampling at 2000 Hz; 60 Hz notch enabled; no post-processing beyond acquisition filters. Wireless ECG monitoring: A portable circuit (amplifier AD8232, Arduino Nano microcontroller, Bluetooth module HC-05, 3.7 V LiPo battery) interfaced to DoS EP sensors via conductive adhesive. Data streamed over Bluetooth and displayed in Processing; post-processing included R-peak detection, adaptive median filtering for outliers, 1 Hz high-pass Butterworth to zero baseline, and gain normalization. A simple stress test had the subject alternate 16 s standing and 16 s walking, repeated 10 times. Comparative studies: DoS EP sensors were compared against hospital-grade gel electrodes (Meditrace 450) and ultrathin (350 nm) serpentine mesh electrodes fabricated by microfabrication. Tests assessed: (i) performance under sweat (pre- and post-15 min outdoor walk at ~38 C), (ii) durability over 7 hours of daily activity, (iii) adhesion (peel with tape and rubbing tests). ECG SNR before/during sweat was computed; observational changes logged. Motion artifact experiments: - Skin deformation during ECG: With wired DAQ, local stretching, compressing, and releasing were applied at the site of one electrode (approx. 2 mm/s, 1 s per phase), with fingers positioned 1 cm apart to deform skin. A custom stretching apparatus also validated reproducible deformations. - Vibrating motor (VM) induced motion during resting EMG: A 10 mm VM driven at 1 kHz (5 Vpp, 1.5 V DC offset; variations used for amplitude and placement) was taped midway between two electrodes (20 cm apart). Time-domain and time–frequency (TF) analyses (Hamming window, 75% overlap) evaluated artifacts. Biocompatibility: Inks were drawn on shaved backs of mice; H&E histology after 48 h assessed inflammation or malignancy. Wound healing study: In three CD-1 mice, a 1 cm-wide epidermal wound (dermis preserved) was created. DoS electrodes drawn as opposing plates around the top half of each wound delivered pulsed DC stimulation (30 µA, 100 µs pulse duration, every 15 ms) for 1 h on days 1 and 3. On day 5, tissues were harvested for H&E histology. Scab/wound width was measured on days 1, 3, and 5; treated vs untreated halves compared. All procedures were IACUC-approved.

- Ink/material performance and conformity: - Ag-PEDOT:PSS conductive films exhibited sheet resistance of 1.2 ohm/sq at 0% strain, increasing to 9.9 ohm/sq at 30% strain, and relaxing to 2.7 ohm/sq upon release; minimal crack formation up to 30% strain. After 1000 cycles at 10% strain, only slight resistance increase. Stable resistance over months at 22 C and 4 C. SEM of skin replica showed the ink filling grooves/valleys for ultra-conformal contact. - Histology at 48 h for Ag-PEDOT:PSS and P3HT-NF on mouse skin showed no inflammation or malignancy in epidermis or dermis. - Transistors (electrolyte-gated P3HT-NF): - p-type characteristics. Under no strain: ION/IOFF = 1.69 x 10^3; field-effect mobility µFE = 7.07 cm^2 V^-1 s^-1; VDS = -0.55 V; VG swept 0 to 3 V. - Under 30% strain: ION/IOFF = 2.01 x 10^2; µFE = 5.36 cm^2 V^-1 s^-1. Threshold voltage shifted from 2.38 V (0% strain) to -2.03 V (30% strain). Some hysteresis expected for electrolyte gating. - Strain and temperature sensors: - Strain sensor: Resistance increased ~4x at 30% strain; gauge factor ~15 across stretching. Under cyclic loading: ~2x change at 10% strain; ~4.5x at 25% strain. - Temperature sensor: Relative resistance decreased from 1.0 to 0.57 over 20–45 C (NTC behavior). Thermistor constant β = 2589 K, comparable to commercial NTC thermistors. - Heaters: - Resistive DoS heater produced temperatures ~25 to 140 C with 1–7 V DC bias; calibration established voltage–temperature relationship. - Skin hydration sensor: - Interdigitated electrodes recorded skin impedance/capacitance with skin-textured conformality; strong linear correlation with a commercial hydration meter: % hydration = b*C + a, R^2 = 0.997. Lotion application sharply reduced impedance, detectable even under stretch. - EP sensing on skin: - DoS EP sensors captured EMG and ECG signals on human skin with and without stretch. ECG SNR ~45 dB remained approximately unchanged after stretching. Devices retained function when encapsulated for wet environments. - Wireless ECG stress test: Average heart rate decreased during standing (91.5 bpm) and increased during walking (95.1 bpm), with clear P, QRS, T waves even during walking. - Comparisons vs gel and ultrathin mesh electrodes under sweat, durability, adhesion: - Under sweat: DoS and mesh ECG signals remained consistent; gel electrodes showed dampened ECG amplitude, attributed to sweat pooling or reduced gel ion concentration. SNR change before vs during sweat ~1 dB (DoS, mesh) vs ~3 dB decrease (gel). - Over 7 h: DoS ECG remained consistent; gel amplitude decreased (gel drying); mesh noise increased at 7 h (partial delamination). - Adhesion: DoS sensors could not be removed by tape peel and remained intact under vigorous rubbing; gel and mesh were removable; mesh damaged by rubbing. DoS imperfections are easily redrawn/fixed. - Motion artifact testing: - Local skin deformation during ECG: DoS showed no abnormal deviations; gel and mesh exhibited motion artifacts. SNR during deformation: DoS 50 dB, gel 20 dB, mesh 12 dB. - Skin–electrode impedance (SEI) change during deformation: DoS 138 ± 41.5 ohm; gel 54.1 ± 13.2 ohm; mesh 1.81 ± 1.44 kOhm. No significant difference between DoS and gel (P = 0.99), indicating SEI change was not the main driver of superior DoS performance. - Vibrating motor (1 kHz) during resting EMG: TF analysis showed mesh had strong artifact near 1 kHz and lower frequency components; gel exhibited vibration-correlated lower-frequency content; DoS showed no artifact in time–frequency domain. Increasing vibration amplitude and moving VM closer/on top of DoS electrodes did not induce artifacts. - Wound healing: - In a mouse epidermal wound model (N=3), pulsed DC stimulation via DoS electrodes accelerated healing. By day 5, scab width reduced from 1.0 cm to 0.49 cm (untreated) versus 0.20 cm (treated). Histology showed smaller wound width and reduced scab on the treated half.

The presented DoS electronics directly address the core problem of motion artifacts in epidermal sensing by creating an ultra-conformal, robust, and stretchable interface that adheres intimately to skin texture, thereby minimizing electrode–skin relative motion. This yields artifact-free ECG and EMG recordings even under local skin deformation and vibration conditions that disrupt conventional gel and ultrathin mesh electrodes. The stable SNR, minimal SEI change impact, and resistance to sweat accumulation demonstrate robust physiological signal acquisition in realistic, dynamic environments without relying on additional hardware or computational artifact suppression. Beyond passive sensing, the platform’s ability to incorporate active components (electrolyte-gated transistors) and functional elements (strain/temperature sensors, heaters, hydration sensors) highlights versatility for integrated on-skin systems. The wound-healing results validate the platform’s therapeutic potential for point-of-care, customizable treatments, leveraging on-demand, freeform electrode geometries. Collectively, the findings support DoS electronics as a simple, equipment-light, customizable approach suitable for daily use and clinical contexts, with advantages in adhesion, durability, and motion artifact immunity over established electrode technologies.

This work introduces a simple, on-demand, drawn-on-skin electronics platform that achieves ultra-conformal contact, robust adhesion, and mechanical deformability for multifunctional sensing and therapy. Using conductive (Ag-PEDOT:PSS), semiconducting (P3HT-NF), and dielectric (ion gel) inks, the authors demonstrate active transistors, strain and temperature sensors, heaters, skin hydration sensors, and EP electrodes directly on skin or skin replicas. DoS EP electrodes provide motion artifact-free ECG and EMG acquisition under skin deformation and vibration, maintain performance under sweat and over hours of wear, and adhere strongly while remaining repairable by redraw. The platform also accelerates wound healing with pulsed electrical stimulation in mice. Future directions include: engineering pen/ink delivery for tighter uniformity and repeatability; integrating more complex circuitry for on-skin signal conditioning and wireless modules; long-term biocompatibility and wear studies; scaling to high-density arrays and smaller feature sizes; and clinical trials assessing diagnostic accuracy and therapeutic outcomes across broader populations and conditions.

- Manual drawing can introduce variability in film thickness and uniformity, although multiple passes and potential pen engineering can mitigate this. - Electrolyte-gated transistors showed hysteresis and performance shifts under strain; active semiconductor devices drawn directly on human skin may be affected by skin potentials, necessitating an insulating underlayer. - Comparative and motion artifact studies were performed on a limited number of subjects/sessions; broader, multi-subject clinical validation is needed. - Wound healing experiments were conducted in a small mouse cohort (N=3) and over a short duration; larger studies are required to generalize efficacy and optimize stimulation parameters. - Long-term (weeks to months) durability, skin compatibility under continuous wear, and responses to varied environmental conditions (e.g., prolonged moisture, temperature extremes) remain to be fully characterized.