Fully Transparent, Ultrathin Flexible Organic Electrochemical Transistors with Additive Integration for Bioelectronic Applications

Abstract Optical transparency is highly desirable in bioelectronic sensors because it enables multimodal optical assessment during electronic sensing. Ultrathin (<5 µm) organic electrochemical transistors (OECTs) can be potentially used as a highly efficient bioelectronic transducer because they demonstrate high transconductance during low‐voltage operation and close conformability to biological tissues. However, the fabrication of fully transparent ultrathin OECTs remains a challenge owing to the harsh etching processes of nanomaterials. In this study, fully transparent, ultrathin, and flexible OECTs are developed using additive integration processes of selective‐wetting deposition and thermally bonded lamination. These processes are compatible with Ag nanowire electrodes and conducting polymer channels and realize unprecedented flexible OECTs with high visible transmittance (>90%) and high transconductance (≈1 mS) in low‐voltage operations (<0.6 V). Further, electroencephalogram acquisition and nitrate ion sensing are demonstrated in addition to the compatibility of simultaneous assessments of optical blood flowmetry when the transparent OECTs are worn, owing to the transparency. These feasibility demonstrations show promise in contributing to human stress monitoring in bioelectronics.

deposited with hole patterns corresponding to channel openings. The lamination mechanism was based on the entanglement of polymer chains during heat pressing. [3,4] This process permits layer formation without deteriorative effects (decomposition or oxidization) on PEDOT:PSS channels and AgNW electrodes, while the previously used etching process for the openings can expose the active channels or source/drain electrodes of OECTs to reactive ions or harsh chemicals.
Moreover, it is important to retain the form factors of constituent materials for the development of OECTs. Remarkably, in this method, the structures of the encapsulation layer and base film were retained. Figure S3 shows a step profile of the laminated and base films, which approximately have the same thickness (1 µm) as before lamination. Therefore, the source/drain electrodes can be easily placed in a neutral strain position between parylene films of the same thickness, which is favorable for mechanical stability. [5] Additionally, various opening designs with a width of up to 20 µm can be retained. For example, parylene films with holes of size 50 µm (source/drain design), cross-patterned holes (alignment marks), and boundary lines were successfully laminated as encapsulation layers on the base parylene films (Figures S3 and S4).
The designability of hole patterns in the encapsulation layer enabled us to pattern the openings of the channels, gate electrodes, and source/drain contacts.
The adhesion of the encapsulation layer in flexible devices is critical for mechanical stability. To evaluate the adhesion between the parylene substrate and encapsulation layer, the shear bond strength was measured through tensile tests (Figure S5a, b). We used a polyethylene terephthalate (PET) film to support a 1 µm parylene layer. The PET films effectively prevented the delamination of parylene layers from the supporting films during the tensile tests. Furthermore, we coated the parylene substrate with fluorocarbon agents, which are required for selective wetting deposition. Figure S5b shows that the shear strength of 287±64 kPa for the parylene-bonded interfaces with intermediate fluorocarbons was comparable to that without fluorocarbons (307±36 kPa). The minimal effect of intermediate fluorocarbons on the shear strength may be ascribed to the penetration of parylene polymer chains by fluorocarbons with a thickness of a few nanometers.
Next, we investigated whether the adhesion is high enough for ultrathin devices by comparing two fracture forces (i.e., tensile fracture in parylene films and interface shear fracture) estimated from the tensile strength of parylene film ( Figure S5c) and shear bond strength. The shear fracture force was more than 100 times higher than the tensile fracture force, as shown in Figure S5d. The results indicate that the adhesion was sufficiently high. Furthermore, we confirmed that the unbonded parts of parylene films were fractured, while the bonded parts remained intact when the freestanding parylene films were bonded and used for measurements ( Figure S6). Such strong bonding ensures that the encapsulation layer can be implemented without deteriorating the mechanical stability of ultrathin OECTs.

Optical Characterization of Fully Transparent Ultrathin OECTs
The high optical transparency of bioelectronic sensors is an attractive feature because optical assessments of biological organisms can be conducted directly above the sensors during an electrical assessment. The capability of multimodal assessments may advance the diagnostic quality of wearable monitoring systems. To evaluate the optical characterization of OECTs, the transmittance spectra in the visible range were obtained for specific device areas of encapsulation and substrate, source/drain electrodes, and channel, as shown in

Current Gain in Frequency Characterization
The cutoff frequency of OECTs was determined as the frequency at which the current gain (ΔIDS/ΔIGS and IDS and IGS denote the drain and gate currents, respectively.) becomes 1. The evaluation of cutoff frequency based on a current gain of 1 is essential for the frequency characterization of transistors because it represents the maximum frequency of input signals that are effectively amplified by transistors. [6] To characterize the frequency performance, first, as shown in Figure S9a, the frequency plots of the drain and gate currents were obtained to calculate the current gain at specific input signal frequencies. Figure S9b Figure 3f, the relationship between the extracted cutoff frequencies and channel length (L) was obtained. This demonstrates a good fit for the 1/L 2 function, which is reasonable for the transistor theoretical equation of the cutoff frequency: [6] fcutoff, gm, and CG denote the cutoff frequency, transconductance, and area-normalized capacitance, respectively; µ, VGS, and Vth denote the mobility, gate voltage, and threshold gate voltage, respectively. The maximum cutoff frequency around 560 Hz was slower than other candidates for transparent transistors (i.e., metal oxides, carbon nanotubes, and graphene), as shown in Figure   2b; however, it was fast enough to record most biological signals because the signal frequency was less than 1 kHz. [7] Furthermore, the frequency performance may be improved by reducing the resistance of silver nanowire electrodes through conductive coating techniques, such as graphene layering, [8] gold plating, [9] or secondary doping of PEDOT:PSS [10] .

Vertical Placement of Optical Sensors on OECTs
Multifunctional biomedical probes allow us to obtain versatile vital signals, leading to the extraction of more accurate information for medical diagnosis and brain-machine interfaces. [11] In particular, multifunctional probes combined with electrophysiological and optical sensors are among the most compelling candidates for wearable applications owing to their enhanced portability. [12] Thus far, the combination of EEG and functional near-infrared spectroscopy has demonstrated the ability to capture richer information related to human cortical activity. [12,13] The combined probes were devised for placing electrophysiological sensors next to optical sensors over the objected areas (forehead), as shown in Figure S11a. However, such in-plane placement of sensors inevitably reduced the spatial resolution in comparison to monofunctional probes because spacings between electrophysiological (optical) sensors must be required for one optical (electrophysiological) sensor. To circumvent this, we proposed a vertical placement enabled by transparent electrophysiological sensors, i.e., fully transparent ultrathin OECTs ( Figure S11b). In such vertical placements, the optical sensors are placed directly above OECTs without spacing owing to high optical transmittance, which causes minimal interruptions in the optical path for sensing. Figures S11c and S11d show that the optical blood flow on the fingertip was measured to obtain the human pulse wave, even when the OECTs covered the fingertip. This demonstrates that the sensors can be stacked vertically without losing the functionality of optical sensing, considering their mutual alignment. Thus, the spatial resolution of electrophysiological and optical sensors can be enhanced to the same extent as that of monofunctional probes when the electrophysiological sensors are optically transparent and flexible.      The change in resistance was calculated by measured resistance, R divided by resistance before the test, RO. A mechanical testing machine (EZ test; Shimadzu Co., Kyoto, Japan) was used to bend the transparent OECT (bending radius: approximately 4 mm) and release it back to its initial position during the resistance measurement (34461A; Keysight Technologies, USA). The total time for one cycle is 10 s. (b) Change in transconductance (measured valued, gm divided by one before the storage, gmO) of OECT after 4 months of storage at room temperature and after further cyclic bending tests. Although room temperature curable Ag paste was used at the contact point between the contact pad of the OECT and the wiring for the electrical resistance measuring instrument, destruction occurred in the area below the bending radius tested. Thus, only one of the three samples left was subjected to the bending test. As shown in Figure 4, the durability of the OECT was confirmed down to a bending radius of 0.8 mm. In the future, a connection method to stabilize the contact pad of the OECT and the external connection wiring should be investigated.