Solvent exchange process using the ultrafiltration method
The important factor in solvent exchange of PEDOT:PSS is that PEDOT:PSS should remain wetted with the solvent to prevent coagulation of PEDOT:PSS. For this, a solvent exchange process using the ultrafiltration method was designed. Figure 1a shows the novel solvent exchange process (SEP) by ultrafiltration. The principle of this solvent exchange process is similar to changing cold water to hot water. Similar to adding hot water while removing cold water, an organic solvent well mixed with water is first added to aqueous PEDOT:PSS solution of the same volume. If half of the mixed solvent is removed through the membrane filter, then a solvent exchange ratio of 50% is obtained. In this mixture, half the volume is water, and the other half is the organic solvent. When this method is repeated 5–6 times, the solvent exchange ratio reaches 95% or higher. During this solvent exchange process, excess PSS or impurities can also be removed through the membrane filter at the same time. Figure 1b shows the dispersion stability of PEDOT:PSS solutions solvent-exchanged from water to various solvents after the SEP. EtOH and ethylene glycol (EG) were particularly selected as solvents for the solvent exchange of water-based PEDOT:PSS solutions because EtOH is widely used in alcohol organic solvents and EG is one of the most widely used conductivity enhancers of PEDOT:PSS. The dispersion stability was excellent in EtOH and EG, which can be explained by the Hansen solubility parameter. PEDOT and PSS have coulombic interactions with each other. The sulfonic acid of PSS is highly hydrophilic and dissociates into SO3− and H+ in water. Thus, PEDOT:PSS has excellent dispersibility in water due to the strong hydrogen bonding between the water molecules and SO3–. Supplementary Fig. 2a shows a graph of the Hansen solubility parameters of the solvents used in the SEP, indicating that the index determining the solubility is composed of three forces: the dispersion force (∂D), dipole–dipole force (∂P), and hydrogen bonding force (∂H)21,22,23. Supplementary Table S1 shows the values of ∂D, ∂P, and ∂H for each solvent. Among these three parameters, ∂H is mostly related to the dispersion stability of PEDOT:PSS. The higher the ∂H value of the solvent is, the better PEDOT:PSS is dispersed in the solvent, as shown in Fig. 1b. This relates to the mechanism by which PEDOT:PSS is dispersed in the solvent. PEDOT and PSS are connected by coulombic interactions, and PEDOT:PSS can be dispersed in solvent by the hydrogen bonding force between PSS and the solvent. Therefore, the higher the ∂H value of the solvent is, the stronger the hydrogen bonding force between the solvent and PSS, so PEDOT:PSS can be more stably dispersed in a solvent with a high ∂H value. Through our solvent exchange process, it has been confirmed that PEDOT:PSS was very well dispersed in a solvent with a ∂H value of 19.4 or higher. The solvent exchange ratio and the conductivity without the addition of DMSO are presented in Fig. 2b. After five or more SEP steps, the solvent exchange ratio reached 95% or higher. Note that the conductivity increased from 0.94 to 250 S/cm (in EtOH) without the addition of DMSO as a conductivity enhancer. This conductivity enhancement is due to the better connection between the conductive PEDOT chains, as the PEDOT and PSS are segregated by polar solvents such as EtOH24,25,26. Since EG is a typical conductivity enhancer, it exhibits a high conductivity of 1108 S/cm, even without DMSO. After the addition of DMSO, the conductivity in the water, EtOH, and EG solvents reached similar maximum values (Table S2). This means that the main chain of PEDOT, which is primarily responsible for the conductivity, remains unchanged, even after the SEP. XPS analysis was performed to confirm the change in the ratio between PEDOT and PSS after the SEP. Supplementary Fig. 2c shows XPS spectra of S 2p of PEDOT:PSS films from the water, EtOH, and EG solvents. The PEDOT:PSS ratio decreased from 1:2.37 in water to 1:1.90 in EtOH and 1:1.82 in EG, demonstrating the effect of removing the residual PSS in water. As the highly acidic PSS was removed, the pH of the PEDOT:PSS solution increased from 2.4 in water to 4.5 in EtOH and 4.4 in EG (Supplementary Table S2). Since the degradation of the atmospheric stability of PEDOT:PSS is mainly due to the hygroscopic property of PSS27, the reduction of the excess PSS can improve the atmospheric stability of PEDOT:PSS. Thus, the atmospheric stability of PEDOT:PSS depending on the solvent of PEDOT:PSS was analyzed by measuring the sheet resistance over time when exposed to air (temperature: 25 °C, relative humidity: 30%). The sheet resistance of the aqueous PEDOT:PSS increased by 98% on day 7, while the sheet resistance of EtOH- or EG-exchanged PEDOT:PSS increased by 38 or 34%, respectively. In other words, the atmospheric stability of solvent-exchanged PEDOT:PSS was improved by nearly three times compared to the aqueous PEDOT:PSS due to the effect of removing excess PSS (Supplementary Fig. S1). Supplementary Fig. S2a, b shows SEM and AFM images, implying that the gel particles increase in size after solvent exchange and that the crystallinity is improved by the 3D network structure and π–π stacking. This phenomenon is particularly evident in EG and is similar to that in previous reported studies of PEDOT:PSS treated by polar solvents, including EtOH and EG28,29,30,31,32,33. The contact angles (CA) of PEDOT:PSS solutions depending on the solvent and addition of a wetting agent (WA) on PDMS are presented in Fig. 2d (Supplementary Table S3). Among all the CA measurements, the PEDOT:PSS solution with water as a solvent showed a large variation from 96.94° to 49.18°, while the PEDOT:PSS solution with the EG solvent showed ~14° less variation, and the PEDOT:PSS solution with the EtOH solvent exhibited a negligible change (~0.16°) after the addition of the WA at 1%. Although the PEDOT:PSS solution with water exhibited a lower CA after the addition of the WA (49.18°), the PEDOT:PSS solution with the EtOH solvent without the WA displayed an even lower CA, thereby avoiding the necessity of a WA for wetting the surface. Therefore, it is clear from Fig. 2d that the PEDOT:PSS solution with EtOH as a solvent without a WA exhibited better coating properties (lower CA) than the solutions with EG and water. It is noteworthy that only the solutions with water and EtOH were mixed with 5% DMSO, not the solution with EG, because EG itself is a conductivity enhancer.
Fabrication and characterization of a pressure sensor based on EtOH solvent-exchanged PEDOT:PSS
A step-by-step pressure sensor fabrication process using solvent-exchanged PEDOT:PSS with EtOH as a sensing material is schematically illustrated in Fig. 3a. The corresponding multilayer design layout and the sensor attached to the wrist artery for blood pulse monitoring are presented in Fig. 3b. A schematic illustration of the circuit model of the sensor mechanism with and without pressure is shown in Fig. 3c, where R1 and R2 represent the resistance before and after pressure is applied, respectively. To ensure PDMS microstructure formation, SEM measurements were performed, and corresponding images were obtained (Supplementary Fig. S3a–d). It was concluded that all pyramidal microstructures form sharp edges at the top without any structural deformation. To ensure a uniform coating of PEDOT:PSS on the uneven microstructure, drop-casting was preferred instead of spin-coating, which led to a nonuniform coating. As a systematic study, initially, pristine PEDOT:PSS solvent-exchanged with EtOH without dilution was used as prepared, without any additional changes. In this case, the SEM images showed thick coatings with a nonuniform thickness over the pyramids, as shown in Supplementary Fig. S4a,b. To obtain a more uniform and thin coating, the pristine solution was diluted twice with the same solvent (EtOH). Uniform and conformal nanocoating of PEDOT:PSS on the uneven surface without any agglomeration was observed, which is essential for practical applications (Fig. 4c, d). To further investigate the nanolevel thickness, uniformity, and presence of the PEDOT:PSS coating over the pyramids, SEM and EDX measurements were performed. As expected, due to the inclined surface, the thickness profile from the base to the top of a pyramid decreased from 90 to 10 nm according to cross-sectional SEM images, as shown in Fig. 4e–i. Similarly, the presence of the PEDOT:PSS coating was confirmed by EDX at the respective positions shown in Fig. 4e, f, h and Supplementary Table S4. The conformation and particle size (≈30 nm) of PEDOT:PSS are consistent with previous studies34,35. Notably, sharp edges with a uniform and conformal coating were observed at the top of the pyramids (also confirmed by EDX), as shown in Fig. 4h, which are crucial for minimizing large resistance deviations in response to pressure in the low pressure regime. On the other hand, commercial water-based PEDOT:PSS partially coated near the edge of the PDMS structure, leading to poor sensing performance (Figs. 4a, b and 5b). This can also be seen in Fig. 2d; the water-based PEDOT:PSS solution exhibited a high CA compared to the EtOH-based solution.
The PEDOT:PSS solvent-exchanged with EtOH (diluted twice with EtOH) coating the PDMS pyramidal microstructure array was considered as a sensing material for the other electrical and mechanical characterizations. I–V measurements were carried out by applying a voltage across the two (top and bottom) electrodes such that a current flowed from the flat PEDOT:PSS substrate to the flat Au substrate through the pyramid tops. As pressure was applied to the top electrode, the pyramidal structures deformed, thereby increasing the contact area and, as a result, decreasing the resistance. Further increasing the pressure led to a negligible change in resistance due to the limited deformation of the pyramid microstructures. Conversely, as the pressure was decreased, due to the elastic nature, the PDMS structure tended to return to its original position, causing the resistance to increase. Figure 5a, b presents the sensitivity of the proposed pressure sensor, defined as S = (ΔR/R0)/ΔP, where ΔR is the change in resistance under applied pressure P, and R0 is the initial resistance under zero pressure. In the case of water-based PEDOT:PSS, the sensor showed poor sensitivity with large hysteresis, which is unfavorable for pressure sensing applications. In contrast, the sensor with PEDOT:PSS solvent-exchanged with EtOH (diluted twice with EtOH) showed high sensitivity with excellent reliability upon the increase and decrease (from 0 to 1 kPa and 1 kPa to 0 Pa) of pressure, almost retaining the original values when the pressure was removed, as shown in Fig. 5b. The plot exhibited two distinct pressure sensitivity regimes, one for low-pressure values <100 Pa, with a sensitivity S1 of −21 kPa−1, and the other for medium pressure values of 0.1–1 kPa, with a sensitivity S2 of −0.016 kPa−1. Note that the PEDOT:PSS solvent-exchanged with EtOH is not coated thinly and uniformly without dilution, and the pressure sensitivity is very poor, with a sensitivity S0 of −0.0001 kPa−1 (Supplementary Figs. S4, S5). The performance of the proposed sensor was compared with that of other sensors reported in the literature, as summarized in Supplementary Table S536,37,38,39,40,41,42,43,44,45,46,47. Although a few reports demonstrate very high sensitivities >20 kPa−1, the procedures to fabricate these sensors are very complex, and the structures obtained by nonconventional techniques are inadequate for producing reliable and repeatable data. Depending on the designed parameters of the microstructures, such as the shape, width, height, and spacing, the sensitivity and pressure regimes vary20,47,48,49. Our sensor showed a fast response time of only 90 ms, with a maximum load of 1 kPa, which is essential for practical pressure sensor applications (Fig. 5c). Figure 5d presents the mechanical stability of the pressure sensor using PEDOT:PSS solvent-exchanged with EtOH. Even after applying a cyclic pressure of 0.3 kPa for 10,000 cycles, the sensor exhibited high mechanical stability without peeling-off or resistance change.
Application of the proposed pressure sensor
Figure 5e–g presents the blood pulse and pressure sensor array applications of the proposed pressure sensor. To demonstrate a blood pulse rate sensor, the sensor was fabricated by laminating the PEDOT:PSS-coated PDMS microstructures with face-to-face alignment of the Au-coated PET sheet, followed by extracting external wires from the edges of the films with the aid of ACF bonding. Among various vital signs of the human body, the pulse is considered a primary vital sign of the human body to prevent cardiac diseases. The proposed pressure sensor attached to a wrist showed pulse waveforms in real-time monitoring, with a pulse frequency of 78 beats/min, which corresponds to that of a healthy adult. Moreover, the sensor exhibited distinguishable systolic and diastolic phases with P1, P2, and P3 peaks indicating incident, systolic, and diastolic waves4. Furthermore, Fig. 5h–j shows a 4 × 4-pixel array pressure mapping panel that enabled spatial resolution of the pressure. Figure 5h shows the multilayer structure of the array sensor. The PEDOT:PSS-coated pyramidal PDMS structure was patterned by a benchtop programmable nanosecond pulse laser to form four electrodes side-by-side in a row. Similarly, the Au-coated PET was also patterned to form four electrodes in a column (Supplementary Fig. S6). The four formed electrodes were connected to external wires with the aid of ACF bonding, while the patterned PEDOT:PSS layer was connected with silver paste. Then, the array sensor was sandwiched by attaching the top electrodes perpendicular to the bottom electrodes. The load of 0.6 g corresponded to 1.5 kPa applied to the second row and third column of the sensor array, and the difference in the applied pressure of the pixel was observed via the increased magnitude of the resistance, presented in Fig. 5h, i. Simultaneously, the spatial pressure distribution on pixels other than the applied pressure pixel exhibited a lower magnitude, thereby enabling precise localization of the applied load. This demonstrates the viability of the proposed pressure sensor for obtaining spatial pressure distributions.