Introduction

Advanced bioelectronic technologies have driven long-term innovation in the field of soft materials, playing crucial roles in human–machine interactions1,2,3, health monitoring4,5, and neural implants6,7,8. Seamless attachment between bioelectronics and tissues/organs via mechanical adaption ensures high-quality recording of biological signals9,10,11. Despite of noble metal electrodes’ superior electrical conductivity, the high cost and mechanical brittleness greatly restricts the precision of electrophysiological signal acquisition and transmission, posing a significant barrier to further improve scalability and long-term stability of bioelectronic devices12,13,14.

Over recent years, a wide range of soft electronics composed of conductive polymers have been developed due to their superior electrical/ionic conductivity, softness, and biocompatibility15,16,17,18,19. Particularly, poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT: PSS) has garnered extensive attentions due to its tunable conductivity (1 ~ 104 S cm−1)20. This is typically achieved by enhancing PEDOT crystallinity through addition of dopant or post-film solvent treatment, which induces phase separation or removes excessive insulating PSS21,22,23. However, the modulus mismatch between soft substrates and the conductive polymers limits the further applications of their electrical properties24,25. Additionally, the conductive polymers face significant challenges in forming close contact with tissue interfaces26,27. To address above issues, non-conductive polymers were introduced and conductive hydrogels with enhanced adhesion were developed to yield the mechanical-compliant and more conductive bioelectronic interfaces28,29. Nevertheless, this enhancement often comes at the cost of electrical performance, reducing their effectiveness in capturing weak and sensitive biological signals. Therefore, balancing the enhancement of conductivity with the improved biointerface interactions remains a challenge.

In principle, the PSS within PEDOT: PSS film can strengthen the electrode-tissue adhesion through hydrogen bonding of functional groups and electrostatic interactions with positively charged ions (e.g., -NH2, Na+)30,31. Additionally, the distribution or content of PSS within PEDOT:PSS film also affects the crystallization of PEDOT as well as the conductivity32,33,34. Various strategies have been developed to form vertical phase separation (VPS) structures of two-component polymer mixtures, such as inducing one polymer aggregation on the surface through interfacial solvent evaporation and using additives like dimethyl sulfoxide and 3-glycidoxypropyltrimethoxysilane to modulate surface energies35,36. Precisely designed component distribution aligns with different functional interface properties, thereby improving the overall device performance. The VPS structure has yielded noteworthy advancements in several fields like organic thin film transistors37,38, organic solar cells39,40 and perovskite solar cells41,42. However, how to leverage this strategy to balance interface modification and conductivity in the bioelectronics remains unexplored.

Here, we report an improved approach to regulate the VPS structure of a pre-oriented PEDOT: PSS films through solid-liquid interface (SLI) doping43. By adjusting thickness and doping solvents, a thicker PEDOT: PSS film with a strong VPS structure was achieved, featuring a higher PSS/PEDOT ratio on the surface and a lower ratio at the bottom. Such unique microstructures with significant component gradient endow the film with superior conductivity ( > 8500 S cm−1), excellent electrochemistry and satisfactory biocompatibility. Additionally, the PSS-enriched surface enhances the interaction between the film and biological tissues. The combination of high electrical properties and enhanced interface interactions strengthens the sensitivity and stability of the bioelectrode in converting electrical signals and capturing biological signals. The versatile bioelectrode is well-suited for both large-scale manufacturing in wearable electronics and precise micro/nano-fabrication in bioelectronics. We further demonstrate the potential applications of the bioelectrodes with VPS structures for biomechanical sensors, and implantable electrophysiological recording. This approach lays the foundation for next-generation biointerface technologies.

Results

Preparation of PEDOT: PSS electrodes with VPS structures

Building on our previous work, we further employ a metastable liquid-liquid contact (MLLC) doping dispersion as the dopant solution44. Briefly, the MLLC dispersion is an ethylene glycol (EG)-diluted PEDOT: PSS formulation in which the excess insulating PSS is partially removed, yielding a reduced PSS/PEDOT molar ratio of ~1.73 (Supplementary Fig. 1). To realize the highly conductive electrodes, commercial PEDOT: PSS ink was blade coated to form a pre-oriented pristine film. Subsequently, MLLC-doping dispersion was sheared on its surface to induce the larger and more ordered crystalline PEDOT domains, forming a high-performance film with VPS structures (defined as the MS-P film, Fig. 1a). The pristine PEDOT: PSS film was sheared with EG solvent, and the resulting films were defined as the SS-P film. During the SLI doping, the EG solvent penetrated into the crystal domain of PEDOT: PSS effectively to weaken the electrostatic interactions between PSS and PEDOT, leading to phase separation between the two components (Fig. 1b). After annealing, the solvent evaporation prompted the accumulation of hydrophilic PSS chains at the surface, while PEDOT-rich domains formed high crystallization and aggregated at the bottom. This VPS structure enables better surface adhesion of PEDOT: PSS film to skin/tissue by the stronger physical interactions as well as enhances the electrical conductance, accordingly ensuring reliable and stable electrical signal acquisition (Fig. 1c). By patterning the MS-P films using a laser system, customized sensor arrays in various sizes are able to record the physical signals from the human skin and physiological signals from the tissue interface (Fig. 1d, e).

Fig. 1: Schematic of preparing the MS-P film with VPS structure and applications in bioelectronic interfaces.
figure 1

a The fabrication of the MS-P film through depositing MLLC-doping dispersion on the pristine film surface and their respective 3D microstructure schematic. b MD simulation models of phase separation between PEDOT and PSS due to EG penetration. c Schematic illustration of MS-P film with vertically separated phase for strong physical interaction with the skin. d Soft bioelectronics with MS-P electrodes, interact with the human skin to record physical and electrophysiological signals. e Flexible MS-P film-based array for in vivo detection of electrophysiological signals.

Full size image

Morphology and microstructure of the MS-P films

Atomic force microscopy (AFM) phase imaging was used to investigate the morphologies of various PEDOT: PSS films with different preparation parameters (Fig. 2a and Supplementary Fig. 2). The AFM phase pictures were further analyzed using a programed software (named the GTFiber) to provide false color orientation maps and pole figures of the polymer crystal fibril distribution45,46, as shown in Fig. 2b. The SLI doping through shearing the pure solvent EG on the pristine PEDOT: PSS film (named the SS-P film) induces the phase separation between PEDOT-rich domains (bright color) and PSS-rich domains (dark color). As an entropy-driven phase separation, it slightly affects the orderliness of PEDOT chain arrangement after annealing compared with that of the pristine PEDOT: PSS film. Compared to pristine PEDOT: PSS ink, the MLLC-doping liquid is a dilute dispersion with a lower viscosity44. When shearing the MLLC-doping dispersion on the surface of pristine film, an evident orientation transition for the PEDOT chains from a parallel to a perpendicular alignment was captured, which was determined by the shearing evolution and evaporation of MLLC-doping dispersion (Supplementary Fig. 3)47,48. To be specific, the meniscus surface remained constant around the three-phase contact line (i.e., the boundary where the liquid, solid substrate, and air meet) at a low shear rate49. As the MLLC-doping dispersion was progressively dragged out by the blade and evaporated synchronously, the containing polymers tended to align roughly along the direction of coating. However, polymers within the meniscus are influenced by capillary flow and Marangoni vortex over a long period, inevitably generating a disordered alignment of fibrils upon deposition. With a further increase in coating speeds, the MLLC-doping dispersion is rapidly coated onto the pristine film, and the contact line is practically infinitely far away, allowing for subsequent quiescent drying. This is comparable to that colloidal particles in a droplet evaporate to form a coffee-ring pattern. Simply put, capillary flow drives the majority of polymers in the dispersion toward the three-phase interface, where they self-assemble and organize along the receding contact line. Such an SLI doping method offers unique advantages in reducing surface defects and enhancing polymer orientation compared to traditional post-treatment and additive-mixed doping techniques.

Fig. 2: Characterization of PEDOT: PSS films with varying degrees of phase separation.
figure 2

a, b AFM phase images (a), corresponding orientation mapping analysis (b), and pole figures of the fibril orientation distribution (bottom). The color map of fiber orientation ranges from 0° (red), 90° (cyan), to 180° (magenta), with the white arrow indicating the coating direction. Scale bars: 200 nm. c–e The depth profiles of PEDOT: PSS films include a schematic diagram (left) illustrating the distribution of PEDOT and PSS polymers, the atomic composition (middle), and the PSS-to-PEDOT ratio (right). These profiles are obtained using XPS ionic cluster gun sputtering, with S (2p) core levels measured as a function of etch cycles. f, g Molecular dynamics analysis about the interfacial adsorption energy between PEDOT: PSS film with different PSS-to-PEDOT ratio and the tissue and the calculated adsorption energy. h Contact angles of 0.1 M NaCl solution on different films. i Two-dimensional grazing incidence wide-angle X-ray scattering (GIWAXS) patterns for PEDOT: PSS films. j, k The vertical and horizontal GIWAXS profiles of films, where qxy and qz are the perpendicular and parallel wave vector transfers with respect to sample surface, respectively.

Full size image

UV-Vis spectroscopy was employed to investigate the compositional changes in thin films following SLI doping. As is well known, hydrophobic PEDOT is insoluble in EG, whereas PSS is highly hydrophilic. Supplementary Fig. 4 presents the UV absorption spectra of various films. The two absorption peaks (~200 and 225 nm) are attributed to the aromatic rings in PSS50. Notably, the absorbance profiles, particularly within the characteristic spectral regions of PSS, appear highly consistent between the SS-P and MS-P films. This suggests that the overall component content remains largely unchanged between the two samples.

Although surface X-ray photoelectron spectroscopy (XPS) provides significant chemical insights, a comprehensive investigation across the film’s depth is more evident for elucidating the vertical compositional gradient. Therefore, high-resolution XPS analysis was conducted, during which the sulfonate groups in PSS with the higher binding energy as well as the thiophene groups with lower binding energy in PEDOT can be distinguished, as depicted in Supplementary Fig. 5. The depth profiles illustrated in Fig. 2c–e highlight the distribution of these components within various PEDOT: PSS films. In the pristine film, a constant PSS to PEDOT (PSS/PEDOT) ratio of around 1.5 is observed throughout the depth in etching process. While both SS-P and MS-P films exhibit the VPS structure. In the SS-P film, the PSS/PEDOT ratio decreases obviously from 30.85 to 1.24 within the initial 1500 s during etching and stabilizes at ~1.20 for the remained components of the film. Similarly, the MS-P film also displays a distinct VPS structure. The PSS/PEDOT ratio reduces from 11.5 to 0.74 during the etching from surface to the bottom of this film, indicating more significant VPS structure of MS-P film than that of SS-P film. The PSS/PEDOT ratios on the surfaces of these two films presents huge difference, likely due to the various compositions of the doping dispersions. The MLLC-doping dispersion contains a small amount of polymer particles with a PSS/PEDOT ratio of 1.7344. Polymer particles reduce the polarity of the dispersion and delay the evaporation of the solvent, allowing sufficient time for the solvent to penetrate into the interior of the film. Thus, the MLLC-doping dispersion could not rapidly accumulate PSS at the film surface as the pure solvents do within a brief period. Unlike the previous study on the PEODT:PSS film with a weak VPS (surface PSS/PEDOT ratio: ~2.7) for electrochemical transistor research, the MS-P and SS-P films exhibit more pronounced component gradient distributions and enhanced interface PSS modification. These characteristics make them more suitable for biocompatible electrodes in bioelectronics, where they effectively promote intimate contact between the electrode and biological tissues51.

The slower evaporation allows more controllable and gradual migration of PSS from the bottom of the film to the surface. Additionally, the polymer particles in the MLLC-doping dispersion further prevent the excessive PSS/PEDOT ratio at the surface, ensuring sufficient conductivity of the exposed surface. According to the molecular dynamics simulation results (Fig. 2f, g), the SS-P and MS-P films have a larger absolute value of the surface adsorption energy, indicating the enrichment of surface PSS groups. The enhanced interfacial wettability of films further verifies the high concentration of PSS on the surface of films (Fig. 2h). The contact angle of 0.1 M NaCl on the pristine film was 94.6°, indicating interfacial hydrophobicity, while the contact angles of SS-P and MS-P films were 45.9° and 52.1°, respectively, suggesting an improvement in hydrophilicity. These results suggested that the enrichment of PSS at the surface caused by VPS is beneficial for strengthening the interaction between electrode surface and tissue.

Furthermore, grazing incidence wide-angle X-ray scattering (GIWAXS) was utilized to study the crystallinity and stacking behavior of the conjugated structures in films (Fig. 2i–k). Compared to the pristine film, the SS-P film demonstrated an increased crystallinity and a strong π-π stacking (020) peak for the out-of-plane (along qz) directions were observed at qz = 1.82 Å−1 (corresponding to the interplanar distance, d, of 0.345 nm), implying preferential formation of the face-on orientation. The preferential orientation was also evident in the MS-P film, as indicated by the shift of the peak to qz = 1.84 Å−1 (d(020) = 0.341 nm), which suggested a reduced spacing within the conjugated plane of polymers. Furthermore, the horizontal GIWAXS profile exhibited a prominent peak at qxy = 1.82 Å−1 (d(020) = 0.345 nm) for the MS-P film, assigned to the π-π stacking of PEDOT along the in-plane direction, which is also smaller than that observed in the SS-P film (d(020) = 0.349 nm). These structural characteristics indicated that the improved lamellar arrangement within PEDOT chains of the MS-P film was likely due to the optimized gradient distribution of components stemming from the VPS. Additionally, based on the full width at half maximum (FWHM) of the PEDOT π-π scattering peak (Supplementary Table 1), the FWHM of MS-P was 0.32 Å−1, narrower than that of SS-P (0.35 Å−1), indicating that the diffraction peak of MS-P was sharper and reflected a larger crystalline domain or a higher degree of ordering. The VPS structure improves the stacking of PEDOT chains while promoting a more compact and ordered alignment of PEDOT. This edge-on orientation of PEDOT facilitates efficient interchain charge transport, thereby enhancing the electrical performance of films45.

Raman spectroscopy (Supplementary Fig. 6) displayed evident peak shifts corresponding to the transformation of PEDOT from a benzenoid to a quinonoid structure by SLI doping, which led to a planar molecular arrangement and in turn enhanced the conductivity of films. The phase separation structure and shearing speeds of MLLC-doping (ranging from 2 to 30 mm/s) exhibited minimal impacts on the polymer arrangement. Additionally, UV polarization absorption revealed that the pristine PEDOT films were isotropic, while the films with SLI doping were observed with increased absorption along the shear direction due to the strengthened crystallization and alignment (Supplementary Fig. 7). Notably, the absorption spectra of SS-P film differed from that of MS-P film, reflecting the variations in micro-phase separation and crystallinity.

Electrical characterization of different PEDOT: PSS films

To explore the relationship between the electrical properties and microstructure of different PEDOT: PSS films, several electrical characterizations were carried out (Fig. 3a). The pristine film exhibits uniformly low conductivity (~1 S cm−1) irrespective of the measurement direction relative to the coating orientation. The SS-P film exhibits a markedly enhanced conductivity of 3973 ± 206 S cm−1 along the coating direction compared to the pristine film, with a slightly lower value observed perpendicular to the coating direction. SLI doping of EG on the pristine film surface induced significant conductivity enhancement by promoting phase separation between PEDOT and PSS-rich domains and enhancing π-π stacking crystallinity during annealing. However, the excessively high PSS/PEDOT ratio (~30) at the surface may limit further improvements in conductivity. Shearing with the MLLC-doping dispersion induces a better-defined VPS in the MS-P film, leading to a remarkable conductivity increase to approximately 8797 ± 253 S cm−1 and 8422 ± 324 S cm−1 in the parallel and perpendicular directions, respectively. By inducing the larger PEDOT crystalline domains and significantly shortening the π-π stacking distance, the VPS structure effectively enhances the electrical performance of MS-P film. Furthermore, the optimized surface PSS/PEDOT ratio (~11.5) facilitates interfacial charge transport. Overall, the slightly higher conductivity is observed in the parallel direction than in the perpendicular direction for films with SLI doping.

Fig. 3: Electrical properties of various PEDOT: PSS films.
figure 3

a Conductivity of various films measured parallel and perpendicular to the coating direction. b, c Temperature-dependent changes in R and ln(R) plot versus 1/T with Arrhenius fitting (K). d Schematic illustration of the electrochemical system and its equivalent circuit model. e, f The fitted electrochemical impedance spectroscopy (EIS) of various PEDOT: PSS films and corresponding volumetric capacitance. g–i Representative CV curves of pristine, SS-P and MS-P film (from the left to the right) during the 2000 charge/discharge cycles in 0.1 M NaCl solution.

Full size image

The manner in which a material’s resistance (R) varies with temperature (T) determines the stability of its electrical performance across wide temperature ranges, which is particularly important for electronic devices, sensors, and other applications. Figure 3b shows the R of various PEDOT: PSS films as a function of T between 25 and 100 °C. R of the pristine film reduced greatly as T increased (from 270 to 44 kΩ), suggesting the improved conductivity at higher T and the characteristics of semiconductor. This behavior was likely due to the increased thermal excitation of charge carriers. On the contrary, resistances of both SS-P and MS-P films were significantly lower than that of the pristine film, and exhibited a slight increase with T, indicating a conductor-like behavior. This is because lattice vibrations (phonons) intensify with T, which increases the scattering probability of charge carriers (holes) and results in higher R. Such a structural reinforcement narrows the electrical performance gap between organic conductive polymers and conductors. Thus, the conductance of PEDOT: PSS electrode with optimized phase separation structure was less interfered by the variation of T.

The relationship between the natural logarithm of resistance lnR and the inverse temperature (1/T) was analyzed, according to the following equation:

$$R={R}_{0}\exp \left(\frac{{E}_{a}}{{k}_{B}T}\right)$$

Where R0 is the initial resistance of the material, Ea is the activation energy representing the energy required for charge carriers to overcome potential barriers, kB is the boltzmann constant. As shown in the Fig. 3c, the pristine film shows the highest resistance variations with temperature and the line slope (Ea/kB), indicating that the conductivity is greatly affected by T and relies on thermal excitation for conduction. In contrast, both MS-P and SS-P films exhibit gentle slopes and their resistances hardly change with T, demonstrating that the conductive polymers exhibit metal-like behavior46,52.

Subsequently, the electrochemical analysis of these films was conducted using a three-electrode-based electrochemical impedance spectroscopy (EIS). The EIS data of different films were fitted using an equivalent circuit model (Fig. 3d) comprising a series resistor (Rs), a parallel capacitor (Cp), a parallel resistor (Rp), and a Warburg element (W), with detailed results shown in Supplementary Fig. 8 and Table 2. The pristine film shows a high Rs (108.6 Ω), while the much lower Rs in SS-P (32.14 Ω) and MS-P (28.93 Ω) confirms enhanced charge transport after SLI doping. The doped films of SS-P and MS-P exhibit significantly higher capacitance values (0.0020 and 0.0023 F, respectively). Compared to the pristine film, that exhibits a high polarization resistance (Rp = 2993 Ω) and diffusion resistance (WR = 129.6 Ω), the significantly reduced values of Rp (~120 Ω) and WR (31.23 Ω for SS-P, 37.65 Ω for MS-P) in the doped samples highlight greatly improved charge transfer and ion diffusion capabilities resulting from the applied modifications. The EIS fitting data and volumetric capacitance of different thin films were compared together in Fig. 3e and f. The pristine PEDOT: PSS film, characterized by the highest impedance at low frequencies (0.1–10 Hz) and the lowest volumetric capacitance, presented a significant interfacial charge transfer impedance. In the case of SS-P, the impedance reduced to 1000 Ω while the volume capacitance increased by nearly five times. Notably, the MS-P film exhibited the best electrochemical performance, both in terms of impedance and volumetric capacitance. It is likely that the MLLC-doping dispersion shear treatment has significantly refined the internal microstructure and surface morphology of the thin film, thereby further rendering the charge transfer pathways more conducive to efficient transport.

We employed an essential electrochemical technique of cyclic voltammetry (CV) to analyze the charge storage capacity and electrochemical stability of the films (Fig. 3g–i). Compared to the pristine film, the SS-P film presented a larger current density, and the CV curve became more rectangular, indicating an improved capacitive behavior and charge storage capacity. The MS-P film displayed the largest current density and a highly rectangular CV curve, implying an excellent capacitive behavior and minimal resistive contributions. Notably, compared to the pristine film, the current density of the SS-P and the MS-P films remained excellent stability over 2000 cycles. The enhanced electrochemical performance and stability can be attributed to the optimized conductivity and surface-enriched PSS, which reinforced the interface structural stability. The cross-linked PSS network mitigated structural damage caused by interfacial charge injection and extraction to a certain extent.

Biocompatibility of PEDOT: PSS films

In bioelectronic applications, biocompatibility is as critical as the outstanding electrical properties and interface interactions of PEDOT: PSS films. To ensure the safety of proposed bioelectronic materials, comprehensive evaluation tests are essential. We initially assessed the short-term cytotoxicity of various PEDOT: PSS films using in vitro experiments. Figure 4a showed a slight reduction in cell density in the PEDOT: PSS groups compared to the control group, while the cell morphology remained consistent. The live/dead cell assay revealed that the density of live cells exhibiting green fluorescence was marginally lower in the PEDOT: PSS samples compared to the control group. The 3-(4,5-dimethylthiazol2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay was performed to measure cell viability, which showed that viability was slightly decreased to 85.5% for the pristine PEDOT: PSS group,76.7% for the SS-P group and 84.1% for the MS-P group (Fig. 4b). The significantly lower cell viability observed in the SS-P group may be attributed to the excessive accumulation of PSS (PSS/PEDOT ratio: ~30) on the film’s surface. This enrichment likely increases the acidity of the solution at the interface, adversely affecting cell survival.

Fig. 4: Biocompatibility of PEDOT: PSS films.
figure 4

a Microscopic images of L929 cells at 24 h seeded on various PEDOT: PSS films. Bright-field microscopy images (first row) and fluorescent microscopy images of live/dead staining (second row). b Cell viability was assessed using MTT assay at 24 h. *P < 0.05, **P < 0.01, and ***P < 0.001 (n = 3). Statistical analysis was performed using Tukey’s test, with significance compared to the control group, and non-significant differences excluded. c H&E staining slices of hindlimb (first column) and heart musucles (second column) with different PEDOT: PSS. d Fluorescent images immunochemically labeled by 4’,6-diamidino-2-phenylindole (DAPI), the biomarker neurofilament (NF) and inflammatory biomarker (ED1) of different PEDOT: PSS films.

Full size image

The mechanical modulus mismatch between bioelectronic materials and biological tissues often leads to fibrosis and inflammatory reactions in surrounding tissues, and may even cause nerve damage to growing animals. The in vivo responses of implanted electrodes of different PEDOT: PSS films fabricated on soft PET substrate were evaluated. Electrodes were implanted into the hindlimb and heart of mice, and tissue changes were analyzed after 4 weeks. Histological sections of the hindlimb and heart tissues revealed abundant myofibrillar cytoplasm, with no infiltration of inflammatory cells in the interstitial spaces (Fig. 4c). Additionally, no significant differences were observed in neurofilament staining (Fig. 4d). When analyzing inflammation levels using inflammatory biomarkers (ED1), the inflammatory signals around nerve bundles of different PEDOT: PSS electrodes are almost at the same level. These findings suggest that all PEDOT: PSS films had minimal impact on tissue and nerve damage. Hence, while taking into account the electrical properties and interface enhancement, we were convinced that the MS-P film exhibits superior biocompatibility as a bioelectrode.

Bioelectronic applications

To make a compromise between the hydrophilicity and conductivity of the film, the MS-P film was chosen for subsequent applications. Due to high conductivity, the MS-P film electrode is able to obtain low contact impedance with skin or tissue. Notably, this film also displayed better electrical stability against bending (Supplementary Fig. 9). We fabricated a flexible piezoelectric sensor, which was designed as a skin-mountable keypad that can be integrated with a calculator (Fig. 5a, b). The device consisted of two patterned PEDOT: PSS films as electrodes, which were separated by PET spacers. The flexible electrodes were electrically isolated to prevent current flowing until under certain pressure values, at which they come into contact and generated an electrical signal. Each sensor unit was assigned a specific symbol for basic mathematical operations, and the electrical signal was visualized in real-time, with the process and location of each press displayed on an organic light-emitting diode OLED screen (Fig. 5c, d). Here, we presented a complete multiplication process and output the result. Finally, the clear key was pressed to proceed with the next logical calculation (Supplementary Fig. 10). In this application, we have successfully integrated flexible organic materials with hardware modules to convert mechanical signals into electrical signals for complex logical calculations and visualization. However, the softness of substrate and spacer materials, as well as the principle of contact-induced current, seriously limits the feedback sensitivity of sensors.

Fig. 5: Wearable electronics with MS-P electrodes.
figure 5

a 3D schematic diagram of the flexible resistive array sensor for calculator keyboard. b Optical image of pressing a keypad button on the human arm. c, d Current response upon pressing a“=” keypad button and the pressed symbol was displayed in real-time on an OLED screen, followed by automatically logical operations performed by the system to generate the final outputs. e 3D schematic diagram of the flexible capacitive array sensor for piano keyboard. f Image of a flexible piano keyboard controlled by pressing the MS-P electrode-based capacitive sensor. g, h Capacitive response upon sequentially pressing each keyboard button (from Do to Si) (g) and corresponding time-frequency diagram. The pressing process was displayed on an OLED screen and recorded in the form of a staff at the bottom of the screen (h). i Optical image of the flexible MS-P electrode for ECG/EMG monitoring. j The illustration of flexible MS-P electrodes mounted on the human body for three-lead ECG measurement. k Representative ECG traces from the commercial electrodes (Ag/AgCl) and MS-P electrodes. The right-side enlarged graph displayed the data of a single heartbeat. l The illustration of the flexible electrodes mounted on a human arm for EMG monitoring by clenching fists and bending arm movements. m, n EMG signals measured by the commercial electrodes and MS-P electrodes under clenching fists (m) and bending arm movements (n).

Full size image

Therefore, we designed another iontronic sensor array as a wearable electronic piano keyboard, which was composed of a thin dielectric layer with the ionic liquid sandwiched between the two flexible electrodes (Fig. 5e, f). Specifically, a micro pillar structure was intentionally introduced onto the surface of the dielectric layer, in order to respond to deformation under compression and generate capacitance changes as quickly as possible. The finite element analysis results suggested that the stress was concentrated on the micro pillars, causing larger deformation than the bulk dielectric film (Supplementary Figs. 11, 12). Such a microstructure design is beneficial for improving the sensitivity of sensors. In addition, the electric double layers at the interface between electrodes and dielectric layers can further enlarge capacitance, thereby achieving high sensitivity and a wide range of pressure response. Similarly, we defined each sensor as a representative musical note (Do, Ri, Mi, Fa, So, La, and Si), and the OLED screen displayed the pressing process of the corresponding sensor, which was recorded in the staff below the display screen (Fig. 5h). At the same time, the electrical signal was converted into an acoustic signal through a buzzer, functioning as an electronic piano. The output sound of each music note was recorded as an audio file for further analysis. With sequentially pressed, these sensors exhibited relatively consistent changes in capacitance signals. In the waveform of audio signals, the sound of each note presented similar amplitude and obvious intervals, indicating that the sound signals converted from the electrical signals were clear and stable (Fig. 5g). As the sensors were pressed sequentially from Do to Si, the frequency of the sound gradually increased. The frequency components below 500 Hz mainly consist of the fundamental frequency and lower harmonics, corresponding to the ascending pitch of the musical notes. In order to demonstrate the high sensitivity and low latency of signal conversion in the integrated electronic system, we played a famous song called “Little Star” through these flexible buttons. Overall, the MS-P film-based electronic system for skin fitting can effectively detect the electrical signal of each sensor under pressing and accurately convert it into the corresponding audio signal, showing high sensitivity and stability.

It is considered that, owing to the VPS structure, the MS-P film exhibited ultra-high conductivity and low interface impedance. Especially, the enriched PSS on the surface enhances the formation of hydrogen bonding, providing more stable contact between the conductive polymer film and tissue. As shown in Fig. 5i, j, the PEDOT: PSS electrode on a thin PET substrate can be seamlessly attached on curved tissues due to their mechanical conformability. These features allow the flexible electrodes to maintain a stable electrical interface with living tissue and accurately capture weak electrophysiological signals, such as electrocardiogram (ECG), electromyogram (EMG). By attaching the electrodes onto various parts of the human body, the MS-P film electrodes exhibited sensing capabilities comparable to commercially wet Ag/AgCl gel electrodes. These electrodes reliably displayed typical ECG cycles, including the P-wave, QRS complex, and T-wave. Notably, the MS-P electrode signals demonstrated a slightly reduced noise level and smoother waveforms, suggesting a relatively high signal-to- noise ratio (SNR) (Fig. 5k).

Large body movements often cause slight displacement at the interface between soft electrodes and tissues, leading to fluctuations in electrical signals. We captured EMG signals during arm bending and fist clenching (representing contraction and relaxation of the biceps brachii and forearm muscles, respectively) to assess the stability of MS-P electrode detection (Fig. 5l). Evidently, the EMG signals from the biceps brachii were highly distinguished and strong, whereas those from the forearm muscle were broad and weak (Fig. 5m, n). This difference is due to the fact that the biceps brachii is the larger and more powerful muscle, while the forearm muscles are responsible for finer motor control and produce weaker and dispersed signals. Based on these results, the MS-P film electrodes can accurately detect these EMG signals, demonstrating high sensitivity and detection stability comparable to commercial Ag/AgCl electrodes.

To further verify whether MS-P is promising for implantable electrodes, here we fabricated patterned microelectrode arrays via an ultraviolet laser system to record in vivo electrophysiological signals. By means of laser material reduction, fine electrode patterns can be obtained for recording electrophysiological signals (Supplementary Fig. 13)51,53,54,55. First, three MS-P microelectrodes were patterned on the flexible PET substrate for better contact with the hindlimb muscle of a rat (Fig. 6a, b), the remaining part of electrodes except the circle contact electrodes were encapsulated by a clinically used TPU tape to avoid additional interference signals. When the sciatic nerve was physically stimulated, it triggered the propagation of action potentials along the nerve fibers, leading to muscle contraction. The EMG signals from the hindlimb were recorded by using a three-channel implanted electrode array (Fig. 6c, d). The high sensitivity of the electrode was captured during both strong and weak muscle stimulations across all three channels (Ch1, Ch2, and Ch3) (Fig. 6e). We calculated the SNR for each data stream by dividing the peak amplitude of spikes into two times the root mean square value of the noise level. These SNRs illustrated that the signals captured by the electrode array were of high quality with low noise interference (Supplementary Fig. 14). Furthermore, the signals exhibited a remarkable degree of temporal and amplitude consistency, verifying the array’s capability to reliably monitor muscle activity with synchronized detection across multiple channels. Such results were further validated in the time-frequency spectrograms as shown in Fig. 6f, through performing a Fourier transform on the data from the three channels. The coherence and resolution observed in these time-frequency spectrograms underscore the electrode array’s high sensitivity and reliability in capturing and analyzing dynamic muscle activity with precision.

Fig. 6: Implantable bioelectronic applications.
figure 6

a, b 3D schematic diagram (a) and optical image (b) of implantable patches consisted of 6 MS-P electrode arrays (3 channels) for EMG recording. c Schematic illustration of the experiment for EMG recording in vivo during the stimulation of the sciatic nerve. d Optical image of flexible microelectrodes on the muscle of the rat’s hind legs. e, f EMG signals recorded from different channels and corresponding time-frequency spectra. g, h 3D schematic diagram (g) and optical image (h) of implantable patches consisted of 12 MS-P electrode arrays (6 channels) for in vivo ECG recording. i Schematic illustration of the experiment for in vivo ECG recording. j Optical image of flexible microelectrodes on a rat’s heart. k ECG signals recorded from different channels by the MS-P electrodes.

Full size image

Another flexible six-channel implantable electrodes were fabricated to evaluate the stability and sensitivity of detecting ECG signals in vivo (Fig. 6g–j). The consistent cardiac rhythms detected across all six channels exhibited strong signal uniformity, suggesting that the electrode array provided synchronized and coherent readings from different locations of the cardiac surface (Fig. 6k). Furthermore, slight variations in shape and amplitude of the ECG signals recorded from six channels were clearly captured. Such differences likely stem from the distinct recording locations on the heart’s epicardium, where variations in muscle contractions influence the signal. This capacity to accurately discern even minor electrophysiological variations across different heart regions underscores the stability of the MS-P film electrodes on the curved heart surface.

Overall, this implantable micropatterned electrode exhibited high sensitivity and stability in detecting EMG and ECG signals, attributed to the high conductivity and low impedance of MS-P film. Significantly, the PSS layer enriched at the interface ensured stable contact between the electrode and tissue in the wet environments, facilitating effective signal transmission. This advancement highlights the potential for further development in implantable electronics, paving the way for more reliable and efficient biomedical devices.

Discussions

We report an efficient SLI doping method to regulate phase separation within pre-oriented PEDOT: PSS films. The resulting structure introduces a VPS structure along the MS-P film thickness, with PSS enriched at the surface and PEDOT crystallizing toward the bottom. Accordingly, an improvement in morphological order and tight π-π stacking between conjugated PEDOT were achieved. Such a VPS microstructure leads to a remarkable electrical enhancement, with MS-P films exhibiting ultrahigh conductivity of 8797 ± 253 S cm−1 and metal-like temperature response behavior. The highly conductive film can be easily patterned into electrode arrays, making them suitable for biomechanical sensors. Additionally, due to the enhanced contact interface interaction and good biocompatibility, MS-P films are ideal candidates for flexible electrodes in both ex vivo and in vivo electrophysiological signal detection, offering high sensitivity and stability. Our approach offers a conductive and biocompatible interface for large-area manufacturing and micro/nanofabrication. This strategy not only provides a practical solution for building long-term robust bioelectronic interfaces but also addresses the ongoing challenge of sensitivity in implantable bioelectronics. These findings hold significant potential for advancing bioelectronic devices and paving the way for improved biomedical applications.

Methods

Materials and film preparation

Firstly, the MLLC-doping dispersion was prepared following the method described in previous literature. Briefly, pristine PEDOT: PSS solution (Clevios PH1000, Heraeus) was slowly dropped onto the surface of EG and left undisturbed at room temperature for 1 h to ensure thorough interfacial doping and removal of PSS from the original ink. Then, the supernatant was collected to obtain the MLLC-doping dispersion with reduced PSS content. Importantly, before use, the aqueous PEDOT: PSS solution was filtered through cellulose acetate syringe filters with a 0.45 μm pore size. The shearing blade was positioned fully parallel to the substrate, with a gap height of 100 μm between them. Pristine PEDOT: PSS films were fabricated on glass substrate using blade coating, with the substrates pre-heated to 120 °C on a hotplate and annealed during the coating process. To prepare SS-P films, EG was blade coated onto the pristine PEDOT: PSS film at a coating speed of 20 mm/s. For MS-P films, the MLLC-doping dispersion was sheared on the pristine film with various coating speeds. To generate patterned MS-P electrode, a Kapton film shadow mask prepared by using a cutting machine (Silhouette CAMEO) is placed on a PET film. Subsequently, the MS-P film preparation and doping process will follow the above process. As for implantable micro electrodes, we obtained them by laser etching MS-P thin films using a nanosecond UV laser system (355 nm, FOTIA, INNO Laser, China). P(VDF-HFP) (Maya agent) and ionic liquid [EMIM][TFSI] (Sigma-Aldrich) were dissolved in acetone (Sinopharm Chemical Reagent Co., Ltd) at a 1:1:10 weight ratio and stirred for 5 h for homogeneity. The dielectric layer with a cylindrical structure was formed by dropping the polymer solution onto the microstructure PDMS template (DOW CORNING) and spin coating at 500 rpm for 20 s to remove the residual solvent.

Microstructural characterizations

Depth profiling XPS was performed using an Ar ion cluster sputtering gun, and PEDOT: PSS films on glass were analyzed with a SHIMADZU AXIS Supra instrument equipped with an Al Kα source. AFM images were captured in tapping mode using a Veeco Multimode AFM. Changes in the molecular conformation of PEDOT: PSS were examined with a Raman spectroscopy system (inVia-Reflex, Renishaw). UV-vis absorption was taken with a Cary 7000i spectrophotometer. GIWAXS measurements were conducted at the Shanghai Synchrotron Radiation Lightsource using a photon energy of 10.26 keV and a sample-to-detector distance of 281 mm. The incident angle was set to 0.12° to probe the entire film while minimizing substrate scattering. Samples were scanned by rotating the substrate to position the incident beam both parallel and perpendicular to the printing direction.

Electrical characterization

The sheet resistance of PEDOT: PSS films was determined using a four-point probe with collinear probes spaced 2 mm apart. Film thickness was measured using a DEKTAK-XT profilometer and averaged from at least three different areas. The conductivity (σ) was calculated for each electrode measurement according to the following calculation: σ = \(\frac{1}{{R}_{sh{eet}}* t}\), where t is the film thickness. The resistance of the films at different temperatures was measured by placing the samples on a heated stage. The temperature was adjusted and held for 3 min before measuring the I–V curves using a source meter (keithley 2612B). CV, cyclic electrochemical current pulse injection, and EIS were carried out using a CHI660E electrochemical workstation with a standard three-electrode electrolytic cell containing 0.1 M NaCl. Platinum wires (1 mm diameter) were used as working and counter electrodes, with an Ag/AgCl electrode as the reference. In CV measurements, a potential range of -0.5 to +0.5 V was applied and cycled at a constant scan rate of 100 mV/s. EIS was conducted across a frequency range of 0.1 to 105 Hz with a 10 mV single sinusoidal signal at Edc = 0 V. EIS data were analyzed using an equivalent circuit model consisting of a parallel capacitor (Cp), a parallel resistor (Rp), a series resistor (Rs) and a series warburg component (W). The volumetric capacitance (C*) of PEDOT :PSS films was determined from the slope of the fitting line.

Molecular dynamics analysis

For MD simulation of polymer phase separation, all-atomic modeling of PEDOT: PSS was performed using the commercial MD software, Materials studio (Dassault Systemes, France). A PSS chain with 30 repeating units and a PEDOT chain with six repeating units were separately prepared. All PSS sulfonate groups were assumed deprotonated, while the PEDOT oligomer was considered fully doped with a net charge of +2e. Subsequent EG absorption and interaction energy calculations were conducted using the open-source code, LAMMPS (Sandia Lab, USA). Interatomic Interactions of the polymeric constituents were described using a polymer consistent force field (PCFF)56, and the TIP3P potential was used for EG molecules57. The interaction between polymeric constituents and EG molecules was described with Lennard-Jones potential (9–6) implemented in the PCFF. For the charge distribution of the PEDOT and PSS, two sets of parameters developed by W. Michaels et al.58 and A.d. Izarra et al.59 were used.

For molecular dynamics analysis about the interfacial adsorption energy, the structure of PEDOT: PSS is as described above, with the Gly-Pro-Hyp tripeptide used to represent skin tissue. The Gly-Pro-Hyp model was built using Materials Studio, and its initial conformation was optimized to ensure structural reliability. The PCFF was employed to describe intermolecular interactions, including Lennard-Jones and electrostatic forces. Initial configurations were generated using Packmol, containing PEDOT: PSS molecules at ratio (PSS to PEDOT) of 2, 10, and 30, along with the Gly-Pro-Hyp tripeptide uniformly distributed in a solvent-free simulation box. Molecules were arranged to avoid overlaps or unreasonable contacts. Simulations were performed in LAMMPS following a multi-stage process. Energy minimization was first conducted using the conjugate gradient method. The system was then equilibrated in the NVT ensemble at 300 K and 1 atm for 20 ns to achieve thermal stability. A 50-ns production run followed in the NPT ensemble, maintaining 300 K and 1 atm. A time step of 1 fs was used, with long-range electrostatic interactions handled by the Particle-Particle Particle-Mesh method and a 10 Å cutoff for Lennard-Jones and electrostatic interactions. Trajectory files were saved every 100 steps during production runs for subsequent analyses, including calculations of interfacial interaction energies and molecular conformational changes.

Finite element analysis

FEA was performed using the commercial package COMSOL 6.2. The dielectric layer with microstructure (P(VDF-HFP) & ionic liquid [EMIM][TFSI]) was modeled as an incompressible neo-Hookean material with Young’s modulus of ~2 GPa. A force of ~ 0.5 N is applied to the surface of the dielectric microstructure and compressed downward. All contact interactions were assumed to be frictionless without penetration.

In vitro biocompatibility

L929 mouse fibroblast cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) (DMEM; Thermo Fisher Scientific, MA, USA; 11885-084) enriched with 10% fetal bovine serum (Sigma-Aldrich, MO, USA; F2442) and an Antibiotic-Antimycotic mixture (Thermo Fisher Scientific; 15240-062) at 37 °C in an incubator with 5% CO2. Various PEDOT: PSS films were deposited onto a PET substrate via blade-coating and subsequently annealed on a hot plate at 130 °C. The PEDOT: PSS films were then cut to occupy 10% of the surface area of the cell culture plate and attached to the bottom. Cells were seeded at a density of 10,000 cells/mL into a 24-well plate (Costar, MA, USA; 3516).

Cell morphology was observed after 24 h of incubation using a Nikon Eclipse TS100 bright-field microscope (Nikon, Japan). Cell viability was assessed using the LIVE/DEAD Viability/Cytotoxicity Kit (Invitrogen, CA, USA; L3224). In brief, cells were stained with 4 µM ethidium homodimer-1 and 2 µM calcein for 1 h, and live cells (green fluorescence) and dead cells (red fluorescence) were visualized using a Nikon Eclipse Ti fluorescence microscope (Nikon Instruments Inc., NY, USA). For quantitative evaluation of viability, cells (in triplicate for each sample) were treated with 0.5 mg/mL MTT solution (Thermo Fisher Scientific; M6494) in PBS at 37 °C for 1 h. Absorbance at 540 nm was measured using an Epoch Microplate Spectrometer (BioTek Instruments Inc., VT, USA) and normalized to the control group for comparison.

Histological analysis immunostaining

All three mice were used to assess the effects of chronic implantation. Four weeks after implantation, they were euthanized using CO2 gas, and the left hindlimb and heart containing the electrodes were collected. Tissues were fixed in 4% paraformaldehyde for 24 h, cryoprotected by sequential immersion in 15 and 30% sucrose solutions overnight, embedded in Optimal Cutting Temperature (OCT) compound (Sakura Finetek, CA, USA), and frozen at −70 °C. The frozen samples were sectioned into 10 µm slices using a Microm HM525 NX Cryostat (Thermo Fisher Scientific, MA, USA), rinsed with distilled water, and stained with hematoxylin and eosin (H&E) for histological analysis.

For immunofluorescence staining, cryosectioned samples were fixed in absolute ethanol at −20 °C for 10 min, washed with PBS, and blocked with 1% BSA for 30 min. Primary antibodies, anti-neurofilament (BioLegend, CA, USA; 801601) and anti-ED1 (Abcam, MA, USA; ab125212), were incubated overnight at 4 °C, followed by 2 h incubation with secondary antibodies. After PBS washes, nuclei were counterstained with DAPI (Vectashield mounting medium; Vector Laboratories, CA, USA). Fluorescent images were captured and processed using a Nikon Eclipse Ti microscope with NIS-Elements software.

Fabrication and characterization of flexible sensor array

The MS-P sensing electrode array was fabricated on PET substrates in the desired configuration using the previously described patterning process. Copper wires were connected to each electrode with silver paste to interface with external equipment, and Kapton tape was applied to secure the wires to the substrates, enhancing mechanical durability. The exposed silver paste was covered with UV-curing resin to prevent electrical noise. The resistive sensor array for the calculator keyboard was constructed by assembling two MS-P flexible electrodes, separated by PET spacers cut using a programmable cutting machine. Similarly, the capacitive sensor array with electric double layers for the piano keyboard consisted of two MS-P electrode arrays, sandwiching a microstructured dielectric. Keypad tests for wearable electronics were conducted with Arduino and Arduino LCD 1602, while current and capacitance signals were recorded using the Keysight 34465 A and TH2830 LCR Meter, respectively.

Physiological signal detection

Skin-mountable MS-P electrodes were fabricated according to the process mentioned above. A flexible transparent bandage was applied on the backside of the electrodes to adhere the sensor directly to the skin. Electrophysiological signals (ECG, EMG, and EOG) were obtained using AD8232 and Arduino UNO by three electrodes, which adhere to different parts of the skin, respectively. Implanted electrode arrays is obtained by laser etching and locally encapsulated with TPU. Epidermal epicardial electrogram (ECG) and EMG data for in vivo detections were collected by an 8-channel printed circuit board (PCB) with a sampling frequency of 2000 Hz and a reference voltage of 2.4 V. This self-designed PCB for data acquisition was composed of two parts, i.e., sampler and receiver. The sampler was composed of an analog front-end (AFE, ADS1198, Texas Instruments, USA), a BLE-integrated MCU (STM32WB55CGUx, STMicroelectronics, Switzerland), an RF filter (MLPF-WB55-01E3, STMicroelectronics, Switzerland), and a low dropout regulator (SPX3819, MaxLinear, USA). An STM32WB55CGUx, an MLPF-WB55-01E3, an SPX3819, and a USB to TTL chip (CH340N, Nanjing Qinheng Microelectronics Co., Ltd., China) were used to fabricate the receiver. All experiments related to the human body were performed in compliance with the protocol approved by the ethical committee of College of Biomedical Engineering and Instrument Science, Zhejiang University ((2024)-7). Informed consent was obtained from all the participants. All animal procedures were approved by the Laboratory Animal Welfare and Ethics Review Committee of Zhejiang University (ZJU20240573).