Multilayer Porous Polymer Films for High‐Performance Stretchable Organic Electrochemical Transistors

Porous films offer a general and simple strategy for balancing the electron/hole transport, and ion doping/dedoping process in organic electrochemical transistor (OECT) channel. Here a universal 3D integrated approach that simultaneously achieves both enhanced transconductance (gm) and mechanical stretchability via constructing a multilayer breath‐figured porous polymer channel by poly(3‐hexylthiophene) (P3HT)/ polystyrene‐block‐poly(ethylene‐ran‐butylene)‐block‐polystyrene (SEBS) and poly(2,5‐bis(3‐triethyleneglycoloxythiophen‐2‐yl)‐co‐thiophene) (Pg2T‐T)/SEBS mixture is demonstrated. The formed multilayer elastic porous structure provides efficient and tunable ionic‐electronic coupling and transport pathways, while also introducing immunity toward mechanical tensile deformation. Remarkably, an obvious increase in gm [from 10.05 mS (2.13 mS) to 29.23 mS (7.38 mS) for Pg2T‐T (P3HT)] is acquired by assembling the OECT porous channel from a single layer to a 3D trilayer. Moreover, mechanical stretchability as high as 40% for Pg2T‐T and 60% for P3HT, is obtained with >21% gm retained. Furthermore, high gms (9.34 mS and 0.92 mS for Pg2T‐T and P3HT, respectively) are maintained after 600 stretching cycles (20% and 30% tensile strains for Pg2T‐T and P3HT, respectively). Overall, the 3D porous structure provides an effective strategy to enhance stretchability and electrical performance for OECTs, as well as opens possibilities for other electronics where both stretchability and a large surface‐to‐volume ratio are needed.


Introduction
Organic electrochemical transistors (OECTs), which rely on mixed ionic-electronic transport and coupling, have aroused DOI: 10.1002/aelm.202300119 tremendous research attention in various applications, including chemical/physical sensors, low-power neuromorphic computing, and humanmachine interfacing, due to their superiorities in low driving voltage (<1 V), electrolyte compatibility, and high amplification capability. [1][2][3][4] Since first demonstrated in 1987, [5] significant progress has been made in stretchable OECTs by achieving efficient ionic-electronic coupling and transport through material and device engineering. [6][7][8][9][10][11] Adopting plasticizers to soften the transistor channel is a common route to achieve reliable mechanical deformation. For example, Sujit Kumar et al. reported fully stretchable OECTs with high transconductance (g m ) of 27.43 mS that can be stretched to 30% strain via adding polyethylene glycol plasticizer in the poly (3,4ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) based channel. [12] In this context, similar work was conducted by adding an additive (xylitol) to PEDOT:PSS channel, where the stretchable OECT retained 91% of the initial drain current (I D ) when stretched up to 30% strain. [13] While considerable attention has been devoted to the design of stretchable semiconductors, advances in terms of device engineering for stretchable OECTs have been established. However, simultaneously maintaining demanding mechanical www.advancedsciencenews.com www.advelectronicmat.de robustness and decent amplification capability remains a significant challenging. [14][15][16][17] To further enhance the stretchability along with the electrical performance, Cicoira et al. combined pre-stretched polydimethylsiloxane (PDMS) substrates with transfer patterning, and fabricated stretchable OECTs which g m only reduced from ≈1.08 mS under 0% strain to ≈0.95 mS under 30% strain. [18] Besides, Joohyuk et al. developed a photo-patternable elastomer called TC-PDMS which acted as the passivation layer of stretchable OECTs and showed a g m of 2.31 and 0.43 mS under 10% and 60% strain, respectively. [19] While most stretchable OECTs depended solely on depletion mode PEDOT:PSS as the transistor channel, which severely limited further development of OECTs in applications where low power dissipation is needed. [20,21] Emerging, Bao et al. demonstrated stretchable accumulation-mode OECTs based on highly stretchable and strain-resistant Au conductors, with g m of 0.54 mS under 0% strain and 0.14 mS under 140% strain. [22] Similarly, Dai et al. fabricated stretchable OECTs with g m of 4.2 mS (0% strain) and biaxial stretchability up to 100% strain with ≈40% reduction of g m . [23] Beyond these, we recently communicated a method for achieving high-performance stretchable OECTs with stable ion and charge transport ability under 30%-140% strain by combining a honeycomb porous structure with the biaxially pre-stretched platform. [24] Furthermore, we have constructed a 3D multilayer porous channel structure for transistor-based gas sensors, which showed enhanced gas sensing properties and bendability down to a 1 mm radius. [25] However, it remains unclear if the 3D porous structure and polymer composite with rubber matrix can also maintain ionic-electronic coupling and transport in OECTs, along with enhanced stretchability.
In this study, we report stretchable OECTs based on multilayer porous polymer films where both the electrical and mechanical performances can be subsequently enhanced by the implementation of extra porous channel layers. Specifically, composite of hydrophobic P3HT and hydrophilic Pg2T-T blending with rubber SEBS imparts the stretchability and ensures films water transferring process. OECTs based on the composite, are investigated and reveal enhanced g m with increased porous layer quantity [g m increases from 3.55 ± 0.90 and 10.90 ± 2.46 mS (1 layer), to 5.30 ± 1.49 and 15.35 ± 1.42 mS (2 layers), and 7.38 ± 2.36 and 29.23 ± 3.88 mS (3 layers), for P3HT and Pg2T-T based OECTs, respectively]. Moreover, the stretchability of OECTs is effectively improved by introducing a 3D porous trilayer based channel, where the breakdown strain increases from 40% (1 layer) to 50% (2 layers), and 70% (3 layers), for P3HT-based OECTs. The 3D structure of the multilayer porous channel not only provides a large surfaceto-volume ratio, which facilitates both ion and electron transport/coupling, but also enhances mechanical robustness due to an interconnected structure. This research demonstrates an effective method for enhanced electron-ion transport and stretchability in OECTs and provides an extendable approach for structures and devices that require a good electron transporting path, large surface-to-volume ratio, and sufficient mechanical robustness.

Porous Polymer Film Morphology and OECT Performances
Two typical polymer semiconductors, hydrophobic P3HT and hydrophilic Pg2T-T are utilized as the OECT channel materials, respectively. [26] First, porous film is fabricated by mixing one of the above-mentioned polymers with SEBS ( Figure S1, Supporting Information) in chloroform/methanol (v/v = 92:8) and spincoated on an ultrathin glass substrate (0.15 mm) with a sacrificial layer [polyvinyl alcohol (PVA) or dextran] in air under 88% relative humidity (RH) (Figure 1a). [27][28][29] Next, the substrate is immersed in water, where a free-standing porous film (p-film), which consists of a P3HT/SEBS blend or Pg2T-T blend, floats on the water surface. Then, the p-film is transferred to the substrate with prepatterned Au source/drain electrodes (50 nm). Note, under small deformations, the pure porous active layer films all exhibit significant fractures. ( Figure S2, Supporting Information), the addition of SEBS enables p-film with better mechanical robustness and easier transferring process. [30,31] Increasing weight ratio of SEBS leads inferior electrical performance of OECTs as evidenced by gradual decreased I on , resulting from the decrease of its conductive property with insulating rubber. Moreover, hysteresis of transfer characteristics for the hydrophobic P3HT increases significantly with increasing SEBS content. Consequently, blend ratios of 2:1 for P3HT: SEBS and 3:1 for Pg2T-T: SEBS were adopted to ensure a reproducible film-transferring process and optimal transistor performances, respectively (See Figures S3-S5, Supporting Information for the optimization process of the semiconductor to SEBS mass ratio). [32][33][34][35] [35] Multiple layers of the p-film are fabricated by repeating the above-mentioned steps. Detailed p-film and device fabrication processes are shown in Figure 1a and Experimental Section. Conventional dense film (d-film) is also fabricated and transferred based on the same procedure, except for spin-coated under 10% RH.
Morphologies and microstructures of p-and d-films based on both P3HT/SEBS and Pg2T-T/SEBS are characterized by scanning electron microscopy (SEM) (Figure 2). All d-films reveal ultra-smooth surfaces, while distinct pores are observed in both p-films. Moreover, the shapes/depths of the pores differ in these two kinds of p-films, [28] where pore diameters of 400-1200 and 200-1000 nm, depths of ≈250 and ≈40 nm ( Figure S6, Supporting Information), are observed in P3HT/SEBS and Pg2T-T/SEBS based p-films, respectively. Note that, percolated pores are obtained in P3HT/SEBS p-films since from the 2 and 3 layers based p-films, Si/SiO 2 substrate and underlying p-films can be observed through the pores (Figure 2a). While for Pg2T-T/SEBS p-films, a thin layer exists in the pores which prevents the direct observation of underlying layers. Nevertheless, both morphologies form a 3D porous structure and should facilitate the diffusion of ions due to the enlarged surface-to-volume ratio. [28,36] Notably, for Pg2T-T/SEBS multilayer porous structure, ions in the electrolyte should be able to dope the film easier compared to its dense counterpart as the layer in the pore is thin (≈5 nm, Figure  S7, Supporting Information). Moreover, 3 layers of p-film show a slight collapse in some regions of the 3D structure, which is more Transistor performances of these OECTs based on p-and d-films for both polymers are next investigated on the rigid substrate (Si/SiO 2 ) by adopting 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide ([EMIM][TFSI]) (for P3HT) or phosphate-buffered saline (PBS, 0.01 M) (for Pg2T-T) as electrolytes. [37,38] Figure 3a,b show representative transfer and g m characteristics of these OECTs. Note that, the transfer characteristics are collected on the 3rd scan to ensure stable characteristics. Peak g m (g m,p ), on-state current (I on ) [I D at V G (gate voltage) = -1.0 V for P3HT and -0.6 V for Pg2T-T, while V D (drain voltage) = -0.5 V in all cases], and threshold voltage (V th ), are summarized in Table 1. In 1 layer d-P3HT based OECTs, V th s of ≈-0.63 ± 0.03 V (forward sweep) and -0.45 ± 0.04 V (backward sweep), are almost identical to these in 1 layer p-P3HT based OECTs [-0.61 ± 0.04 V (forward sweep) and -0.47 ± 0.02 V (backward sweep)]. Interestingly, the I on of 1 layer p-P3HT OECTs is 0.59 ± 0.15 mA, which is 73.5% higher than those in corresponding 1 layer d-P3HT ones (0.34 ± 0.61 mA). Similarly, g m,p of 1 layer p-P3HT OECTs (3.55 ± 0.90 mS) is 66.7% higher than that in 1 layer d-P3HT OECTs (2.13 ± 1.12 mS). The same trend is also observed in 1 layer d-and . While enhanced I on (4.58 ± 1.03 mA) and g m,p (10.90 ± 2.46 mS) in 1 layer p-Pg2T-T based OECTs are obtained when compared to these in d-Pg2T-T based OECTs (I on = 3.87 ± 1.21 mA, g m,p = 10.05 ± 1.21 mS). Therefore, it is clear that p-film based OECTs all exhibit higher I on , and g m,p than their counterparts based on d-films, not mention that the effective thickness of the d-films is less than these of the p-films (vide infra). Such results are consistent with previous reported results on OECTs with porous channels. [24,28] Consequently, normalized peak transconductance g m,n is calculated based on g m,n = g m,p /(WdL -1 ), where d is the effective semiconductor thickness. Note that, for p-film based OECTs, d is estimated by excluding the voids in porous structure (See Supporting Information for detailed calculating method, where d are   112.5, and 20.75 nm for 1 layer p-P3HT and 1 layer p-Pg2T-T, respectively). [2,39] The calculated g m,n s for p-film based OECTs are 10.53 ± 2.66 S cm -1 (1 layer p-P3HT) and 175.1 ± 39.5 S cm -1 (1 layer p-Pg2T-T), which are ≈3 times higher than these in OECTs based on d-film (3.43 ± 1.81 S cm -1 for d-P3HT, and 84.19 ± 10.14 S cm -1 for d-Pg2T-T). The tremendously enhanced g m,n is owing to a faster and more comprehensive ion doping/dedoping process benefiting from a higher surface-to-volume ratio. [28] Next, OECTs based on varying layers of p-films are fabricated and characterized. As shown in Figures 3c-f and Table 1, with the increasing of p-film layers, enlarged I on s are observed simultaneously, where they increase from 0.59 ± 0.15 and 4.58 ± 1.03 mA (1 layer) to 0.79 ± 0.22 and 6.22 ± 0.58 mA (2 layers), and 1.16 ± 0.32 and 13.55 ± 2.58 mA (3 layers) for p-P3HT and p-Pg2T-T based OECTs, respectively. A similar tendency in g m,p is also observed, where they increase significantly from 3.55 ± 0.90 and 10.90 ± 2.46 mS (1 layer) to 5.30 ± 1.49 and 15.35 ± 1.42 mS (2 layers), and 7.38 ± 2.36 and 29.23 ± 3.88 mS (3 layers) for p-P3HT and p-Pg2T-T based OECTs, respectively (Figures 3e,f). The near linearly enhanced I on and g m,p benefit from the enlarged d and facile ion injection pass ways introduced by the 3D porous multilayer structure. In contrast, a different tendency is observed in g m,n , where they decrease from 10.53 ± 2.66 and 175.1 ± 39.5 S cm −1 (1 layer) to 7.86 ± 2.21 and 123.7 ± 11.5 S cm −1 (2 layers), and 7.28 ± 2.33 and 156.6 ± 20.8 S cm −1 (3 layers) for p-P3HT and p-Pg2T-T based OECTs, respectively. Meanwhile, the g m,n s for d-P3HT and d-Pg2T-T based OECTs decrease from 3.43 ± 1.81 and 84.19 ± 10.14 S cm −1 (1 layer) to 0.53 ± 0.32 and 71.14 ± 8.42 S cm −1 (3 layers), respectively. Therefore, it is clear that the maximum g m,n s for both devices based on these two materials (for P3HT, maximum g m,n = 10.53 ± 2.66 S cm −1 ; for Pg2T-T, maximum g m,n = 175.1 ± 39.5 S cm −1 ) are achieved in OECTs with 1 layer p-film, indicating that shorter ion diffusion distance is still the key for higher g m,n . [40,41] **Meanwhile, control OECTs based on 3 layers of d-films for both materials are fabricated and characterized (Table 1 and Figure S8, Supporting Information). Specifically, OECTs based on 3 layers of d-P3HT exhibit a low current on/off ratio due to high off-current (≈10 −4 A), which is due to insufficient diffusion of the electrolyte ions in the bulk of the semiconductor, thus resulting in incomplete doping/dedoping. In general, the OECT based on 3 layers of d-Pg2T-T still shows relatively good transistor performance which with ≈18% and ∼ 12% lower I on and g m,p respectively, when compared to its porous counterpart. Additionally, the improvement of g m,n in p-P3HT based OECTs is more obvious, where they increase significantly from 3.43 ± 1.81 and 0.53 ± 0.32 S cm −1 (1 layer and 3 layers d-films, respectively), to 10.53 ± 2.66 and 7.28 ± 2.33 S cm −1 (1 layer and 3 layers p-films, respectively) (Figure 3e). A similar tendency in p-Pg2T-T based OECTs is also observed, where the g m,n increase significantly from 84.19 ± 10.14 and 71.14 ± 8.42 S cm −1 (1 layer and 3 layers d-films, respectively), to 175.1 ± 39.5 and 156.6 ± 20.8 S cm −1 (1 layer and 3 layers p-films, respectively) ( Figure 3f). Consequently, these results further verify that the improved OECT performance is due to the increased film mass, as well as the multi-layer porous microstructure which facilitates the ion doping/dedoping process. Additionally, we have computed the μC* value for various film morphologies, which govern the transportation of carriers and ions. The μC* value denotes the coefficient of linearity be-tween gm and WdL −1 (V th -V g ) ( Figure S9 and Table S1, Supporting Information). Compared with the μC* based on dense films (20.6 and 222.8 F cm −1 V −1 s −1 for d-P3HT and d-Pg2T-T, respectively), porous structure leads to notably enhanced μC* (107.3 and 354.6 F cm −1 V −1 s −1 for monolayer p-P3HT and p-Pg2T-T, respectively). Furthermore, multilayer porous films also demonstrate obviously enhanced μC* compared to dense counterparts, but with a slight reduction when compared to monolayer porous layers. The slight reduction can be attributed to the longer diffusion length for multilayer porous films. It is worth noting that the absolute gm still increases as the porous layer number increases.
Furthermore, we have demonstrated the versatility of this approach in fabricating N-type material (Homo-gDPP). To enhance its mechanical stability, we constructed a 3D structure by adding a composite of Homo-gDPP/SEBS with weight ratio of 2:1. As shown in Figures S10 a-d (Supporting Information), compared with dense film, porous structure exhibits higher I on and g m . Furthermore, as the number of stacked layers increased, significant enhancement on the device performance is exhibited. The surface morphologies of corresponding films are shown in Figure  S10 e-h and Figure S11 (Supporting Information). This result proves the universality of the pore-making method for both Ptype and N-type materials in this work.

Stretchable Multilayer Porous Films and OECT Performances
Next, stretchable OECTs based on multilayer p-and d-films are fabricated and characterized by adopting stretchable polyurethane (PU) as substrates. [40,42] Transistor performances of these stretchable OECTs under various tensile strains and directions are first investigated (The stretching testing diagram is illustrated in Figure S12 (Supporting Information), stretching along the length/width of the channel is indicated as L-direction and W-direction, respectively). As shown in Figure 4, Figure  S8-S23 (Supporting Information), representative transfer characteristics of different OECTs under various elongation strains are tested, along with their corresponding I on and g m summarized. Obviously, porous structure also leads to improved transistor performance on stretchable architecture as the assembled porous layer quantity increases and shows much higher performance compared to their dense counterparts.
As shown in Figure 4, both Pg2T-T and P3HT based stretchable OECTs exhibit good stretchability, especially for the ones with multilayer p-films. Transistor properties (including I on , I off , and g m ) of all the devices gradually decrease with increased stretching strain. While it is obvious that as the layer quantity increases, the decreasing tendency reduces. Specifically, in 1 layer d-P3HT based OECTs, the performance of the transistor degrades tremendously with increasing stretching strains, where the g m,p and I on decrease from 0.33 mS and 0.084 mA (strain = 0%) to 0.17 mS and 0.037 mA (strain = 10%), and 0.05 mS and 0.008 mA (strain = 20%). As a results, only 15.2% and 9.5% of the initial g m,p , and I on remain. While by introducing 1 layer p-P3HT as the OECT channel, the declining trend becomes weaker along with higher transistor performance, where the g m,p , and I on decrease from 0.88 mS and 0.26 mA (strain = 0%) to 0.81 mS and 0.22 mA (strain = 10%), and 0.46 mS and 0.12 mA (strain = 20%). Furthermore, increasing the p-film layer quantity would largely enhance the OECT performance stability under stretching. Especially, no significant change in the OECT performance is observed under relatively low stretching strain (10%) in 3 layer p-P3HT based OECTs, where g m,p s vary in a narrow range from 1.18 mS to 1.50 mS when experiencing either L-or W-direction stretching's, respectively. In multilayer structures, the overall integrity of the active layer doped with elastomer remains intact under small tensile strength. The deformation caused by stretching enlarges porous structure, and enables larger electrolyte/active layer contact area. Correspondingly, mechanical tensile deformations would also lead to a reduction in film thickness, enhancing the bulk doping/de-doping extent of the active layer especially for P3HT which has low μC* and resulting in increased switching ratio and gm. In contrast, the impact is relatively subdued in Pg2T-T films. (Figure 4e,f). Consequently, a more efficient doping/dedoping process is obtained, along with higher I on and g m,p under 10% strain. Such effect also leads to an enhanced current on/off ratio when a 10% elongation strain is applied (Figure 4a). [13,43] Note, in Pg2T-T based stretchable OECTs, monotonously decreased I on and g m,p are observed, due to better ion diffusion efficiency.
The above results indicate that the porous structure is more stretchable than its dense counterpart owing to 3D pore defor-mation, which can effectively release the mechanical strain in the film. [44,45] Moreover, when applying p-films with increased layers, the breakdown strain can be effectively enhanced, which effectively increases from 30% (1 layer d-film) to 40% (1 layer pfilm), 50% (2 layers p-films), and 70% (3 layers p-films) in P3HT based OECTs (Figures 4a,c,d, Figure S13-S20, Supporting Information). Note that, the breakdown strain is defined as the strain where exponentially dropping on the OECT performance is observed and the device performance would not recover after such strain is released. Meanwhile, for Pg2T-T based OECTs, the breakdown strain increases from 40% (1 layer d-film devices) to 50% (all p-film devices) (Figures 4b, e, f; Figure S21-S28, Supporting Information). The underlying reason for the different breakdown strains between P3HT and Pg2T-T based OECTs is that the adhesion capability of the materials on the electrode also has a large impact on the stretchability. P3HT, with superior adhesion capability, exhibits better stretchability. Due to the inferior adhesion capability of Pg2T-T on Au electrodes, the p-Pg2T-T delaminates obviously from the Au electrode under strain. On the contrary, the p-P3HT holds intact interfacing with the underlying Au under strain and would only break together with the Au electrode ( Figure S29, Supporting Information). In addition, the SEM images of Pg2T-T under different elongation strain also support the results. Under 50% tensile strain, 3-layers Pg2T-T obviously fragmented, which also leads to the degradation of its electrical properties ( Figure S30, Supporting Information).
To investigate the underlying mechanism for the better stretchability of OECTs with multilayer p-films, SEM was utilized to investigate the micro morphologies of multilayer p-P3HT under 60% elongation strain and p-Pg2T-T under 40% elongation strain (Figure 4g; Figure S31, Supporting Information). Obvious cracks are observed in all the p-film. However, the number of cracks is effectively suppressed along with the increased porous layer quantities. It is hypothesized that in a multi-layer porous structure, cracks may be formed only in one layer at a specific position, while other layers in the position remain intact. As shown in Figure S32 (Supporting Information), when the uppermost semiconducting layer cracks, the underlying layers at the same position show no cracks and maintain effective connections. Consequently, carrier transport will not be fully prohibited at this position due to the remained carrier transporting path. Note that, the stretchability of the gold electrode is also characterized to ensure the variation of the transistor performance is due to the change in electrode conductivity ( Figure S33, Supporting Information). The gold electrode allows current up to 3.64 ± 2.77 mA even when stretched up to 70% strain, which is much higher than the max-imum I on measured for the stretchable OECTs. Therefore, it is verified that variation in the device performance during stretching is mainly due to the change in the channel. [46] Last, the performance of 3 layers p-films based OECTs based on both P3HT/SEBS and Pg2T-T/SEBS are further monitored over multiple stretching cycles (for P3HT, strain = 0% ↔ 30%; for Pg2T-T, strain = 0% ↔ 20%). Representative OECT transfer characteristics are shown in Figure 5a,b, and Figure S34-S37 (Supporting Information). By summarizing the variations of I on and g m,p with increasing stretching cycles (Figure 5c-f), it is observed that these OECTs all show excellent mechanical robustness, with a maximum g m degradation of 28.1%/40.7% and I on degradation of 41.2%/50.7% after 600 stretching cycles (W-direction) for p-P3HT and p-Pg2T-T based OECTs, respectively. Note, the variation of V th after various stretching cycles is not obvious ( Figure S38, Supporting Information). Overall, the performance of the OECTs stretched in the W-direction remained better than that of the devices stretched in the Ldirection. Specifically, OECTs stretched in the W-direction remain more than 72% (0.92 mS)/60% (9.34 mS) of the original g m,p (1.28 mS/15.48 mS) after 600 stretching cycles while the devices stretched in the L-direction hold only 42% (0.43 mS)/55% (8.66 mS) of the original g m,p (1.03 mS/15.86 mS) after 600 stretching cycles, for OECTs based on 3 layers of p-P3HT and p-Pg2T-T, respectively. The mismatched Young's modulus between channel and rigid source/drain electrodes (Au) leads to material delamination and/or local fracturing when stretching the device, which may impair the channel/electrode interface and lead to a deterioration in electronic performance.

Conclusions
Here, elastomeric semiconductor materials are blending P3HT and Pg2T-T with rubber SEBS that have high mechanical stability and relatively high electrochemical performance. We have reported OECTs with both high-performance and decent stretchability are fabricated and characterized by introducing multilayer breath figure p-films as the transistor channel. High g m,p (7.38 ± 2.36 mS and 29.23 ± 3.88 mS for P3HT and Pg2T-T, respectively), high stretchability (up to 70%), and high mechanical robustness are realized by carefully optimizing the porous film component, layer quantities and device structure. The multilayer ensures that even if cracks form in one of the layers at a specific location, the carrier can transport through other layers which remain intact at that location. In addition, the 3D structure of the multilayer simultaneously facilitates both electron transport/coupling and ion doping/dedoping process. This study provides a simple but effective, reproducible, and stable method to prepare highperformance stretchable OECTs.
OECT Fabrication: Si/SiO 2 wafer (first sonicated in IPA for 15 min and then dried at 80°C in an oven) was used as the rigid substrate. While for elastic substrates in stretchable OECTs, a water solution containing PU was drop-casted on a cleaned glass substrate, and then placed in a desiccator and pumped under vacuum for 3 min to remove all the gas bubbles. After allowing the water solvent to evaporate at room temperature for 48 h, the stretchable substrate was readily peeled off from the glass and cut into a dimension of 2 × 2 cm 2 as a stretchable substrate. Next, 50 nm Au acting as source and drain electrodes was thermal evaporated to the rigid or stretchable substrate via a shadow mask with 3 nm Cr served as the adhesive layer (W = 1.5 mm, L = 50 μm). Note, on stretchable substrate, no Cr was evaporated to avoid weakened mechanical tensile strains. For the formation of porous polymer film, an ultrathin cover glass (0.15 mm thickness) was first cleaned by sonicating it in isopropanol for 15 min. PVA (20 mg mL −1 ) dissolved in deionized (DI) water was then spin-coated on the UV-Ozone treated cover glass at 3000 rpm for 60 s, followed by baking on a hot plate at 110°C for 1 min. Such PVA served as a water-soluble sacrificial layer for the transfer printing of hydrophobic P3HT film. Dextran (20 mg mL −1 in water) was also processed the same way as that of PVA, and acted as a sacrificial layer for the transfer printing of hydrophilic Pg2T-T film. P3HT/SEBS and Pg2T-T/SEBS composites with different weight ratios (P3HT, Pg2T-T: 5 mg mL −1 ) were dissolved in a mixed solvent of chlo-roform and methanol (volume ratio: 92:8) and stirred under room temperature overnight. Then, the mixture solution was spin-coated on top of the cover glass with sacrificial layer at 5000 rpm for 10 s under a relative humidity (RH) of 88%. The high RH during spin-coating contributes to the formation of the desired porous morphology due to breath figure effect. While d-film was formed by spin-coating the same solution under dry atmosphere (RH < 30%). Next, the obtained porous or dense polymer film on the cover glass was immersed in DI water, resulting in a freestanding film on the surface of water, and then was carefully transferred on to the prepared rigid or stretchable substrate with source/drain electrodes. The transfer process was repeated multiple times to yield a multilayer semiconducting channel for OECT.
Characterizations: The morphologies of p-and d-films were characterized by field emission scanning electron microscopy (ZEISS Gemini 300). The contact angles were measured by using a JY-82 contact angle test platform. All of the electrical characteristics of OECTs were performed in the ambient environment on a probe station connected to an Agilent B1500 parameter analyzer. Stretchable OECTs were characterized under both stretched and relaxed states. Note that the electrical properties of different OSC-based OECT were measured in different electrolytes (P3HT: EMIM TFSI, Pg2T-T: PBS) using Ag/AgCl gate electrodes.

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