A High‐Resolution, Transparent, and Stretchable Polymer Conductor for Wearable Sensor Arrays

Arrays of stretchable and transparent electronic sensors realize next‐generation skin‐conformable wearables and soft robotic skins, which require a high‐resolution patternable stretchable conductor. However, the difficulty of simultaneously engineering desirable material properties (i.e., conductivity, stretchability, and patternability) has limited the development of such stretchable electronic materials. Herein, a high‐resolution patternable, stretchable, and transparent conducting polymer by decoupled engineering of the material properties is shown. The high conductivity of the conducting polymer is achieved by rationally designing an ionic additive. The high stretchability is realized by matching the mechanical properties of the conducting polymer to the substrate. The developed conducting polymer is then patterned in a resolution less than 10 µm by nanosecond UV laser ablation, which enables the feasible demonstration of stretchable and transparent sensor arrays for touch and strain. The findings in this work will accelerate the development of high‐density stretchable sensor arrays and stretchable semiconductor devices.


Introduction
Spatially distributed sensing on soft and deformable surfaces, including human skin, is essential for next-generation humancomputer interfaces (HCIs) and wearable healthcare sensors. Increasing the number of inputs gives more freedom in HCIs. Multipoint biological signal sensing allows us to detect the propagation of signals and to easily find the place of interest. Sensor arrays based on a thin elastomeric sheets are capable of spatial sensing, which can exhibit mechanical properties similar to that DOI: 10.1002/admt.202201992 of human skin: stretchability of >30% [1] and Young's Modulus of ≈1 MPa. [2] Ideally, the sensors should be transparent. The aesthetic impression to the user can determine the adherence of wearable devices, with transparent devices minimally influencing the original appearance. Furthermore, transparency allows us to integrate optical healthcare sensors to add multimodality [3] or displays for the indication of obtained signals from the sensors. [4] For the realization of stretchable and transparent sensor arrays, it is important to develop high-resolution patternable, transparent, and stretchable conductors. Stretchable and transparent conductors have been realized using liquid metals, [5] carbon nanotubes, [6] silver nanowires, [7] and conducting polymers. [8] Among them, conducting polymers, including poly(3,4-ethylenedioxythiophene)poly(styrenesulfonate) (PEDOT:PSS), are attractive because of the low intrinsic modulus, transparency, and biocompatibility. Stretchable conducting polymers have been mainly realized by mixing nonstretchable PEDOT:PSS with surfactants, [9,10] ionic additives, [11] sorbitol, [12] or ductile polymers, [13][14][15][16] to simultaneously realize high stretchability and conductivity. The patterning of stretchable conducting polymers was performed by utilizing inkjet printing [11,17,18] and photolithography [15,16,19] to yield a resolution of 40  However, further improvement of the materials properties of PEDOT:PSS has been difficult because there are few additive materials for PEDOT:PSS to simultaneously achieve high conductivity, stretchability, and high-resolution patternability. For example, photocrosslinkable polyethylene glycol was necessary to realize the photopatterning of stretchable conducting polymers. [15,16] Inkjet printing requires us to finely tune the wettability and drying of inks by surfactant and solvents. [17] Therefore, it is ideal to develop engineering methods to independently approach conductivity, stretchability, and patternability of conducting polymers.
Here we show a high-resolution, high-conductivity stretchable conducting polymer, enabled by decoupled engineering of the material properties. Our stretchable conducting polymer showed a conductivity of 330 S cm −1 by a carefully designed additive. A considerable crack-on-set strain of 160% was achieved by controlling the mechanical properties of the substrate. The patterning was realized in a resolution of 7 μm by a high-throughput nanosecond UV laser ablation. Two applications demonstrate the high feasibility of our high-resolution conducting polymer. A stretchable and transparent touch sensor array can be placed on arbitrary surfaces, including human skin and robots, to form a seamless human-computer interface. A strain sensor array can map pulse wave signals on our skin.

Engineering Conductivity and Stretchability of a Conducting Polymer
The stretchable conducting polymer is composed of PEDOT:PSS doped with lithium bis(pentafluoroethanesulfonyl)imide (LiBETI), formed by spin coating the solution on a thermoplastic polyurethane (TPU) substrate (Figure 1a). Li-BETI is an ionic additive to improve the stretchability and conductivity by altering the morphology of PEDOT:PSS. As a control, other previously reported ionic additives (lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) and 4-(3-ethyl-1-imidazolio)-1-butanesulfonate (ION E)) were also investigated. [11,20] LiTFSI is one of the best additives to improve conductivity and stretchability. [11] Although higher conductivity of PEDOT:PSS:LiBETI without strain compared to PEDOT:PSS:LiTFSI has been reported, [21] the stretchability and conductivity under strain have not been investigated. ION E is a zwitterion that significantly enhances stretchability. [20] The addition of ION E to PEDOT:PSS:LiBETI further improved the stretchability. Additionally, we found that the mechanical property of the TPU substrate greatly enhances the stretchability of PEDOT:PSS compared with a control substrate (styrene-ethylene-butylene-styrene (SEBS)) used previously. [11,17,18] Figure 1b shows the effect of ionic additives on the conductivity of PEDOT:PSS. PEDOT:PSS with LiBETI showed a high conductivity of 330 S cm −1 , while that with LiTFSI showed 270 S cm −1 . PEDOT:PSS with no additive, ION E, and both ION E and LiBETI showed a conductivity of 1.3, 0.037, and 220 S cm −1 , respectively. The mechanism of higher conductivity with LiBETI compared to LiTFSI was confirmed by grazing incidence wide angle X-ray scattering (GIWAXS) ( Figure S1, Supporting Information). The clear peak around 0.89 Å −1 indicates the higher crystallinity of PEDOT in PE-DOT:PSS:LiBETI than that in PEDOT:PSS:LiTFSI. [11] Importantly, LiBETI still reduced the hardness of PEDOT:PSS to a similar degree as the other plasticizers, confirmed by nanoindentation ( Figure S2, Supporting Information). The reduced hardness was realized by weakening the hydrogen bonding between PSS or poly(4-styrenesulfonic acid) (PSSH) in PE-DOT:PSS, verified by Fourier transform infrared spectroscopy ( Figure S3, Supporting Information). In addition, all the additives improved the transparency of the film ( Figure S4, Supporting Information).
The stretchability of our conducting polymers was dramatically improved by the TPU substrate. We investigated the crackon-set strain as a measure of stretchability, defined as the strain that generates cracks in thin films on the elastomer. [22] Although PEDOT:PSS:LiBETI showed a crack-on-set strain of 30% on an SEBS substrate, the same conducting polymer showed a large crack-on-set strain of 160% on a TPU substrate (Figure 1b). Pristine PEDOT:PSS and PEDOT:PSS:LiTFSI also showed a similar improvement of stretchability by using TPU substrates. We confirmed this improved stretchability of conducting polymers on various TPU substrates ( Figure S5, Supporting Information). Our approach to obtain highly stretchable PEDOT:PSS is free from specially synthesized materials, which secures the large-scale production and easy availability to the other researchers.
The mechanism of the large improvement of stretchability is that TPU's have a greater similarity in mechanical properties with PEDOT:PSS, compared to SEBS. The Young's modulus of PEDOT:PSS:LiTFSI was reported to be 55 MPa. [11] The Young's modulus of TPU and SEBS used in this study are 30.6 and 5.5 MPa, respectively. In addition, Poisson ratio of PEDOT:PSS, TPU, and SEBS are reported as 0.35, [23] ≈0.45, [24] and ≈0.49, [25] respectively. Strain can be more uniformly applied to thin films when the substrates possess similar mechanical properties to the thin films. [23] The molecular-level uniform distribution of strain energy to PEDOT:PSS by TPU was confirmed by dichroic ratio, measured using polarized UV-vis absorption (Figure 1c), which indicates the strain-induced polymer chain alignment. [22] PEDOT:PSS:LiBETI on TPU showed a linear increase of dichroic ratio by strain. On the other hand, the dichroic ratio on SEBS showed a decrease after 50% strain where microcracks were generated, which is the result of nonuniform distribution of strain in the PEDOT:PSS film. Instead of mechanical properties matching of TPU with PEDOT:PSS:LiBETI, we first hypothesized that stronger adhesion of PEDOT:PSS with TPU compared with SEBS was the mechanism of the stretchability improvement. However, the adhesive force was measured to be similar ( Figure S6, Supporting Information).  (Figure 1d). The conductivity at a strain of 0% and 100% was 269 and 196 S cm −1 , respectively. On the other hand, PEDOT:PSS:LiTFSI showed a much reduced conductivity by strain. The conductivity at a strain of 0% and 100% was 243 and 97.9 S cm −1 , respectively. This phenomenon can be explained by the higher strain-induced polymer chain alignment of PEDOT:PSS with LiBETI than with LiTFSI, which was confirmed by the dichroic ratio ( Figure 1c). Additionally, the conductivity of PEDOT:PSS:LiBETI was enhanced by posttreatment, by spin coating a methanol solution of LiBETI. The conductivity of post-treated PEDOT:PSS:LiBETI was to 525 S cm −1 at 0% strain and 280 S cm −1 at 100% strain (Figure 1d), which may be caused by the removal of excessive insulating PSS and altering the film morphology. [26] Although PEDOT:PSS:LiBETI + ION E showed the highest crack-on-set strain (Figure 1b), the conductivity dropped significantly by strain. Excessive ionic additives may have hindered the percolation of conductive PEDOT within the composites.

Patterning of a Conducting Polymer
The high-resolution patterning of our stretchable conducting polymer was realized by nanosecond UV laser ablation. Figure 2a shows the optical microscope images of the patterned PE-DOT:PSS:LiBETI. The resolution is as high as 7 μm. The successful ablation was confirmed by X-ray photoelectron spectroscopy (XPS, Figure 2b). After laser ablation, the peak corresponding to sulfur (≈170 cm −1 ) in PEDOT:PSS diminished, while the peak corresponding to carbon (≈280 cm −1 ) in the SEBS substrate was more prominent. Additionally, the clear edge of ablated PE-DOT:PSS was confirmed by a scanning electron microscope (Figure S7, Supporting Information). Although laser ablation patterning of conducting polymers has been demonstrated, [27][28][29] patterning in high-resolution (<100 μm) has never been shown to the best of our knowledge. Figure 2c shows the sheet resistancestrain characteristics of the conductors patterned with different line widths. The conductivity of the patterned stretchable conducting polymer was not affected by the laser ablation. Lines with thin width (<10 μm) showed a significant resistance increase at a strain less than 50% due to the crack propagation from the patterned edges. Still, our 10 μm wide conductors maintained low sheet resistance in a strain smaller than skin stretchability, which is enough for realizing on-skin wearable devices. The achieved resolution (7 μm) is as small as the previously reported method using photolithography [15,16] and is smaller than inkjet printing. [11,17,18] The higher resolution could be obtained by reducing the laser spot size and minimum step size of the Galvano scanner. We compare the performance with that of the previously reported patternable stretchable PEDOT:PSS (Table S1, Supporting Information). [11,[15][16][17][18][30][31][32] Our approach utilizes only commonly available materials, yet can enable modest conductivity, high stretchability, and high patterning resolution (<10 μm). Furthermore, our approach has a wider materials window because the patterning is independent from the additives, which can promise higher electrical conductivity and stretchability.

Stretchable Transparent Sensor Arrays
We demonstrate two applications to show the feasibility of our high-resolution stretchable conducting polymer. One is a stretchable, transparent touch sensor array (7 × 7) shown in Figure 3a,b. The sensor can be placed on arbitrary surfaces, including the soft and deformable skins of robots. Transparency of a touch sensor is important because it allows us to integrate other components including optical sensors, [3] and displays. Additionally, transparency can maintain the original appearance of our skin, which can affect the aesthetic quality. [4] Figure 3c shows the schematic structure. The sensor comprises of layers of our high-resolution stretchable conducting polymer and insulating elastomers to make an array of parallel electrode pairs. Our touch sensor utilizes the mutual capacitance sensing mechanism (Figure 3d). When a finger is approaching the parallel electrodes, the capacitance decreases because the finger disrupts the electromagnetic field between the two electrodes. We confirmed the capacitance change was induced by an approached finger ( Figure S8, Supporting Information). The pixel size (6 mm) and pitch (6.5 mm) were determined by the typical size of our fingers.
The insulating elastomer is multilayered thermoplastic elastomers (SEBS and TPU). TPU is necessary to improve the stretchability of PEDOT:PSS:LiBETI. However, only TPU did not work as an insulator, confirmed by impedance spectroscopy (Figure S9, Supporting Information), possibly because of the ionic conductivity of TPU induced by an ionic additive, LiBETI. Therefore, a nonpolar SEBS dielectric is necessary to realize an insulating property. The detailed fabrication process of the sensing area is available in the Experimental Section and Figure S10 in the Supporting Information. The bottom and top conducting polymers were separately fabricated on different substrates and integrated by lamination, to avoid damaging the bottom conducting polymer during the laser ablation of top electrodes.
Reliable connection of the sensing area to the readout circuit was realized by wiring to a flexible flat cable (FFC) connector (Figure 3b). A printable, stretchable silver ink minimized the resistance between the sensing area and the circuit. A 50 μm thin FFC was introduced between the silver inks and a readout circuit. The FFC ensures the mechanical robustness of the connection which can easily fail due to the significant mechanical mismatch between a stretchable sensor array and a rigid readout circuit. The electrical connection between a stretchable silver ink and an FFC was prepared by Gallium-based liquid metal. The flexible-to-stretchable interface showed a high robustness ( Figure S11, Supporting Information). The electrical connection did not fail even when the TPU substrate attached to FFC was stretched by >150%, which is comparable with a recent report on flexible-stretchable interfaces. [33] The details about the connection are available in the Experimental Section. Our readout circuit can cancel the leakage current between electrodes caused by moisture in the air or from our body ( Figure S12, Supporting Information). [34] The fabricated sensor could clearly sense the touch by a single finger and two fingers (Figure 3e,f and Movie S1, Supporting Information). The touch sensing was successful under 30% strain ( Figure S13, Supporting Information). It is important to note that our transparent and stretchable touch sensor array is the thinnest (≈100 μm) among the previously reported transparent and stretchable touch sensors (Table S2, Supporting Information). [35][36][37][38][39][40] Such a reduced thickness can significantly improve the conformability of the devices to arbitrary surfaces, including human or robotic skins. [41] Our sensor does not cause discomfort of wearing because it is as soft as human skin. The stiffness of a thin film is determined by the Young's modulus and thickness. [41] Our sensor has a Young's modulus and thickness of ≈30 MPa and ≈100 μm, whereas human skin has that of ≈1 MPa and ≈1 mm, respectively. [42] The thinness is realized by two components developed in this study: high-conductivity and patternable stretchable PEDOT:PSS and ultrathin (≈5 μm) elastomeric dielectrics. Notably, most thickness comes from the TPU substrate, suggesting that further thickness reduction is possible.
Finally, we fabricated capacitive stretchable strain sensors. The sensor consists of two stretchable conducting polymer layers sandwiching an insulator layer to form a capacitor structure (Figure 4a). Our stretchable dielectric consists of TPU and SEBS because TPU supports the stretchability of PEDOT:PSS and SEBS forms a good dielectric ( Figure S9, Supporting Information). Figure 4b shows the relative capacitance change by strain, indicating the linear response to the strain. [43] A 4 μm thick SEBS dielectric prevented the leakage current between the top and bottom electrode under strain, while a 1 μm thick SEBS dielectric was unable to suppress the leakage. The gauge factor is 0.78, which is comparable with previously reported capacitive strain sensors, as the gauge factor can be theoretically calculated. [43] Although the gauge factor is lower than resistive strain sensors, the high linearity allows us for the reliable sensing. Besides, we achieve high transparency and resolution in the strain sensor. The sensor showed high repeatability against 1000 cycles of 30% strain (Figure 4c). The capacitance at 0% strain did not go back to the original value due to the viscoelastic deformation of TPU substrate, which can be improved by crosslinking. By UV laser ablation of the conducting polymer layer, we fabricated a 1 × 4 strain sensor array. The pixel pitch is 1 mm. The readout circuit is shown in Figure S14 in the Supporting Information. The sensor was laminated on skin using commercially available skin-adhesive, and showed strong enough adhesion ( Figure S15, Supporting Information). The multipoint sensing enabled us to achieve a mapping of pulse waves detected by placing the sensor on a human wrist. All the obtained signals in four channels clearly showed clear features of a pulse wave, including systolic peak, tidal wave, dicrotic notch, and dicrotic wave. [44] Mapping of the pulse wave would allow us to improve the signal integrity by increasing sample size and calculating local pulse wave velocity, which is an important indicator for cardiovascular health. [44,45]

Conclusion
In summary, we have decoupled the materials engineering and patterning of a stretchable conducting polymer to enable highly conductive, stretchable, and high-resolution patternable conducting polymers. Our conducting polymer showed a high conductivity by careful additive design and high stretchability by matching the mechanical properties with the substrate. The high-resolution patterning was enabled by a nanosecond UV laser ablation, and the feasibility is demonstrated by the stretchable and transparent sensor for touch and strain, which is useful as next-generation wearables for human-computer interface and healthcare. The extensive materials window in the UV laser ablation patterning of conducting polymers would enable further improvement and fine tuning of the materials properties, including electrical conductivity, stretchability, and work function. This is highly beneficial for other types of high-resolution stretchable sensor arrays and stretchable semiconductor devices.

Experimental Section
Fabrication of Stretchable Conducting Polymers: SEBS (Asahi-Kasei, H1062) substrates were prepared by drop casting the cyclohexane solution (80 g L −1 ) on glass substrates. Cyclohexane was used because it can dry fast due to the lower boiling point than toluene. The solution was dried in a glass petri-dish overnight. TPU (Nippon Miractran, P22MBRNAT) substrates were prepared by drop casting the tetrahydrofuran solution (80 g L −1 ) on glass substrates. The resultant thickness of elastomer substrates was ≈80 μm. The solution of stretchable conducting polymer was prepared by mixing aqueous solutions of PEDOT:PSS (Heraeus, PH1000), ionic additives (40 g L −1 ), and fluorine surfactant (Capstone, FS30) in a volume ratio of 100:25:1. LiBETI and LiTFSI were purchased from Tokyo Chemical Industry Co., Ltd. (TCI). ION E was synthesized following the previous report. [20] The solution was spin coated typically at 1000 rpm for 1 min on SEBS or TPU substrates after the surface oxygen plasma treatment (PIE Scientific, Tergeo-Plus Plasma Cleaner) in a condition of 50 W for 1 min. After all the following spin coating here, films were dried using a hot plate at 110°C for 10 min. The conductivity of PEDOT:PSS:LiBETI was enhanced by additional spin coating a methanol solution of LiBETI (10 g L −1 ) at 0 rpm for 1 min followed by 3000 rpm 1 min.

Characterization of Stretchable Conducting Polymers:
The conductivity was measured by cutting the conducting-polymer-coated elastomer substrates in a width of 5 mm. Lines of liquid metal (GaIn) were painted in 10 mm to have an electrical contact. The resistance was measured using a digital multimeter (Keysight, U1252B). The thickness of the conducting polymers was measured by a profilometer (Bruker, Dektak XT). Crack-onset strain was measured by observing with an optical microscope (Leica, DM 4M) at every 10% tensile strain. The strain was applied to the conducting polymers on elastomers using a small vise. Strain-dependent conductivity was measured using an X-axis stage (COMS, PM80B-200X-HQ) and an LCR meter (NF Corporation, ZM2376). The transparency and dichroic ratio were measured using a UV-vis spectrophotometer (Shimadzu, UV-2600i). For all the characterization, the thickness of PEDOT:PSS with and without additives was set to 100 nm.
GIWAXS Measurement: GIWAXS measurements were carried out at BL40B2 in SPring-8 (Hyogo, Japan). Samples were prepared by spincoating 100 nm thick PEDOT:PSS with and without additives on bare silicon wafers. The wavelength of the X-ray beam was 0.1 nm, and the camera lengths were 343.6 mm in BL40B2. 2D scattering images were acquired using a photon-counting detector (Pilatus3 2M, Dectris, Ltd.). The samples were mounted in a helium cell to reduce radiation damage. The data acquisition time was 5 s. GIWAXS was performed at an incident angle of 0.10°, which was lower than the critical angle of total external reflection at the silicon surface. Therefore, the incident X-rays passed through the sample and reflected on the silicon wafer surface. The components of the scattering vector, q, parallel and perpendicular to the sample surface were defined as q y = (2 / )sin(2 )cos( f ) and q z = (2 / )(sin( i ) + (sin( f )), respectively. Here, i is the incident angle of the X-ray beams, f is the exit angle with respect to the surface, is the X-ray wavelength, and 2 is the angle between the scattered beam and the plane of incidence. [46] Laser Ablation Patterning of the Stretchable Conducting Polymers: Laser ablation was carried out using a nanosecond UV (wavelength: 355 nm) laser marker (Keyence, MD-U1000C). The outline of the pattern was ablated in the following condition. Power: 30%; Pulse frequency: 120 kHz; Scan speed 100 mm s −1 . The other area was filled with the laser in the following condition. Power: 80%, Pulse frequency: 120 kHz; Scan speed; 1000 mm s −1 ; Scan pitch: 4 μm; Defocusing distance: −10 mm.
Fabrication of a Touch Sensor Matrix and a Strain Sensor Array: Figure  S7 in the Supporting Information shows the fabrication process of a touch sensor matrix and strain sensor array. First, the bottom layer of PEDOT:PSS:LiBETI was formed and patterned on a TPU substrate. A toluene solution (80 g L −1 ) of SEBS (Asahi-Kasei, H1051) was spin coated at 1000 rpm for 1 min onto the PEDOT:PSS:LiBETI. In advance of the spin coating of SEBS, a cured sheet of polydimethylsiloxane (PDMS) (Dow, SILPOT 184) was placed on the edge part of patterned PEDOT:PSS to leave the contact area open. After removing the PDMS sheet, the film was dried at 110°C for 10 min. This 1.6 μm thick SEBS layer promotes the adhesion of the following dielectric SEBS. A 1.2 μm thick dielectric SEBS (Asahi-Kasei, H1051) was formed by spin coating the toluene solution (80 g L −1 ) at 1500 rpm for 2 min on a nanoground glass (NGG) substrate which possesses very high hydrophilicity. [47] The dielectric was delaminated from an NGG substrate to a PDMS temporal substrate with an assist of deionized water. The SEBS thin film was then transferred to the bottom PEDOT:PSS electrodes whose surface is covered by another SEBS. Next, an SEBS solution (H1051 in toluene, 80 g L −1 ) was spin coated at 1000 rpm for 1 min onto a glass substrate which surface was treated with trichloro (3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyl)silane (Fluorinated Self-Assembled Monolayer, FSAM). An N,N-dimethylformamide solution (100 g L −1 ) of TPU was spin coated at 1000 rpm for 30 s onto the SEBS/FSAM glass substrate. PEDOT:PSS:LiBETI was formed and patterned on TPU/SEBS/FSAM glass substrate. The top electrodes were covered with adhesive 500 nm thick SEBS (H1221) formed by spin coating the cyclohexane solution (20 g L −1 ) at 1000 rpm for 1 min and 1.2 μm thick dielectric SEBS (H1051) formed by lamination from an NGG substrate. The top electrodes were then transferred to bottom electrodes covered with dielectric SEBS.
Interconnects from the Stretchable Touch Sensor to the Readout Circuit Board: A silver-based stretchable conductive ink (Chimet, Ag X520Mod. 15 EI) was stencil printed using a 50 μm thick polyimide (PI) shadow mask to extend wirings from the top and bottom PEDOT:PSS:LiBETI electrodes. The high conductance of silver-based stretchable inks can minimize the series resistance, which can influence the electrical readout. 3 nm thick chromium and 50 nm thick gold were thermally deposited under a high vacuum (<1 × 10 −4 Pa) with a PI shadow mask on a 50 μm thick polyethylene terephthalate substrate to fabricate a FFC. At one end of the FFC, a strip of 125 μm thick PI was laminated by double-sided tape to increase the total thickness and connect it to a commercially available FFC connector. The FFC was adhered to a touch sensor matrix by an instant adhesive (Cemedine, SuperX). The electrical connection between silver ink electrodes and FFC was ensured by painting a liquid metal. The readout circuit board is described in Figure S12 in the Supporting Information.
Pulse Wave Measurement: Pulse wave was measured by attaching the strain sensor to the skin by glue (COSCOS, Face glue H). The experiment was performed under the approval of the Keio University bioethics committee (Approval number: 2021-111).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.