1D Arrays of Highly Adhesive Nanosheets Inspired by Mytilus Edulis Foot Proteins

Assembling functional materials into one‐dimensional arrays is an efficient method for the integration of devices. However, the complicated assembly methods and poor stability of patterns still restrict their further applications. In this work, the well‐dispersible nanosheets are prepared by introducing a bionic adhesive strategy. The oriented arrays with uniform morphology and precise position are realized by the liquid‐film‐induced capillary bridge assembly method. Owing to the abundant interactions between catechol structure and targeted substrates, the micro‐wires present extraordinary stability in adhesion tests. Field‐effect transistors (FETs) based on these micro‐wires are also fabricated, and enhanced carrier mobility is achieved, exhibiting to be 41 times higher than pristine film‐structured FETs. Micro‐wires have a certain stability and improved electrical properties at the same time, which endows this strategy with practicality and application prospect.


Introduction
Field-effect transistors (FETs), based on the working principle of using electric fields to control the electrical behaviors of devices, [1][2][3] are regarded as one of the essential elements in various electronic applications such as active-matrix imagers, sensor arrays, and health monitoring systems. [4][5][6][7] Unfortunately, the performance of traditional planar FETs is restricted by relatively large conductive channels and isotropic alignments. [7] For further optimization, FETs with onedimensional (1D) structured channels are developed. [8,9] The ordered 1D arrays possess regular molecular packaging, which is conducive to charge transfer. In addition, the separated stripes can reduce the current leakage and crosstalk between adjacent FETs. [10] Semiconductive conjugated polymers, owing to their merits of flexibility, solution processability, and tunability for electronic properties, have been extensively used in FETs. [11,12] Whereas, the irregular molecular arrangements and relatively low carrier mobility hinder their further development. Constructing -conjugated composites by introducing nanomaterials such as graphene has been demonstrated as a feasible solution to improve electrical performances. [13,14] However, the homogeneous dispersion of graphene in organic polymers is still challenging, which greatly limits the performance of devices.
1D micro-wire arrays assembled by the liquid dewetting process have been used to assemble organic and inorganic materials and generated satisfying directional assembly effects. [15,16] Among them, the liquid-film-induced capillary bridge (LICB) method is an attractive strategy to realize the anisotropic alignment of materials through microstructure manipulation and interface wettability guidance. [17] Due to the controllable anchoring and movement of the gas-liquid-solid three-phase line with the evaporation of solvent, the functional materials can be assembled into arrays with regular direction and large scale. [16,18] However, there is less research on the assembly of 2D nanosheets into 1D microwires and it is difficult to realize large-area assembly. The possible reason is the relatively strong interactions within nanosheets than between nanosheets and solvent, which could induce phase separation during the directional movement of the meniscus. In addition, nanosheets with large specific surface areas tend to be absorbed onto flat substrates and are more difficult to be transported along a three-phase contact line than nanoparticles. In addition, the radial dimensions of nanosheets are also required to be at the micron level as the restriction of the height of the capillary liquid bridge. More importantly, poor adhesion of the nanosheets to the substrate can also significantly reduce the practicality of the device. Therefore, how to obtain well-dispersed nanosheets, large-scale micro-wires, and robust substrate adhesive 1D arrays is a series of crucial problems to overcome.
The bio-inspired adhesive strategy has been developed in functional coatings, [19] surface modification, [20] and adhesive hydrogels. [21,22] Living along the humid and rocky shore, the marine mussel can cling to all kinds of surfaces including inorganic and organic materials. [23] The predominant reason for this extraordinary adhesivity is the existence of sufficient catecholic amino acids, such as 3,4-dihydroxy-L-phenylalanine (L-DOPA), which are distributed in Mytilus edulis foot proteins ( Figure   1a). [20] Polydopamine (PDA), which has a catechol structure similar to DOPA, shows high affinity to various substrates by various interactions (Figure 1b); and thus, is widely applied as a functional modifier to endow materials with certain adhesion. [24][25][26] In this work, the 1D nanosheet arrays were assembled using a LICB assembly method, and a bioinspired adhesion strategy was used to endow the microwires with a certain degree of substrate adhesivity. The conductivity of graphene oxide is poor, while the carrier mobility of semiconducting conjugated polymers is relatively low. Constructing -conjugated composites has been demonstrated as a feasible solution to improve electrical performances. Due to the good dispersive ability of nanosheets, the obtained micro-wires were precisely positioned and orderly oriented, and large scale was realized by manipulating the wettability-induced assembly process. Meanwhile, owing to the abundant interactions between catechol structure and targeted substrates, the micro-wires presented extraordinary stability in adhesion tests. In addition, FETs possessing separated 1D arrays showed improved electrical performance compared with pristine film. The constructed FET devices based on 1D nanosheet arrays showed an average carrier mobility of 5.84 × 10 −3 cm 2 V −1 s −1 , which was 41 times higher than filmstructured FETs. The results provide a new perception of the fabrication of functional FETs with micro-wire-based channels.

Results and Discussion
The schematic diagram of Figure 1c illustrates the preparation process of adhesive nanosheets. First, graphene oxide (GO) was sulfonated into sulfonated graphene oxide (SGO) for acting as counterions for poly (3,4-ethylene dioxythiophene) (PEDOT). Sulfonic acid group was negatively charged, while PEDOT was positively charged, and electrostatic interaction between them could make PEDOT water dispersible. [27] In order to endow the nanosheets with certain adhesiveness, PDA was introduced into the system. The dopamine was polymerized into PDA and grafted on the surface of SGO by stirring under a mildly alkaline environment. The monomer of 3,4-ethylene dioxythiophene (EDOT) was polymerized into PEDOT after adding oxidizing agent under the ice bath. SGO was used as skeleton during the polymerization and assembly process of PEDOT. The morphology was characterized by scanning electron microscope (SEM), atomic force microscope (AFM), and transmission electron microscope (TEM). As shown in Figure 1d,e, the obtained nanosheets present a typical 2D lamellar structure. The morphologies of GO, SGO and PDA-SGO are shown in Figure S1, Supporting Information. The lateral dimension of PDA-SGO/PEDOT nanosheets was measured by AFM as 2.68 μm (Figure 1f,g), which was less than the width of silicon pillars and was facilitated to be assembled under micropillars.
Spectroscopic characterization techniques such as Fourier transform infrared (FTIR) spectrum, X-ray photoelectron spectroscopy (XPS), and Raman spectroscopy were carried out to further demonstrate the structures of fabricated nanosheets. As shown in Figure 1h, two stretching bands of 1719 and 1052 cm −1 were observed in the spectrum of GO, which could be assigned to oxygen-containing groups (C=O and C-O-C). The successful incorporation of the sulfonic acid groups in SGO could be demonstrated by symmetric and asymmetric stretching bands at 1069 and 1269 cm −1 , respectively. In PDA-SGO, the appearance of N-H and O-H stretching bands located at 3350 and 3191 cm −1 , C=N stretching band at 1553 cm -1 , and CNC stretching band at 1297 cm −1 could be attributed to the grafting of PDA on the surface of SGO. For PDA-SGO/PEDOT, a set of bands at 1385, 1508, 771, and 1055 cm -1 was exhibited for PEDOT, which could be ascribed to the C=C and C-C stretching bands on the thiophene ring, C-S band, and alkylididenedioxy. XPS spectra of GO, SGO, and PDA-SGO are shown in Figure S2a-c, Supporting Information. As shown in Figure S2d, Supporting Information, the emergence of N 1s peak in the PDA-SGO survey spectrum compared with SGO also proves the successful introduction of the adhesive unit of PDA. The representative bonds which are relevant to C-C, C-O, C=O, and C-S bonds were tested in C 1s spectra of PDA-SGO/PEDOT (Figure 1i). In the Raman spectra, the intensity ratio of the D peak and G peak (I D /I G ) usually represents the degree of defects and disorder. [28] As shown in Figure 1j, the I D /I G ratio of SGO mildly increased from 1.989 of GO to 1.992 because the sulfonic acid groups were connected to the terminal groups of GO flakes and slightly reduced oxygen-containing groups. [29,30] A small red shift emerged from 1328 to 1315 cm −1 when PDA was grafted onto SGO flakes. This red shift originated from strong interactions such as -interaction and hydrogen bonds between PDA and SGO. Meanwhile, the introduction of PDA and PEDOT both caused an increase in the degree of disorder (I D /I G = 2.246 and 2.288, respectively). The UV-vis spectrum showed that the absorption band of SGO (237 nm) red shifted to 249 nm, which indicates the noncovalent bond interactions between SGO and PEDOT [31] (Figure S3, Supporting Information). The characterization results indicated that PDA-SGO/PEDOT nanosheets were successfully prepared.
Before the assembly of micro-wires, asymmetric wettability modification was performed on silicon templates. As shown in Figure 2a, the original silicon template exhibited symmetrically lyophilic properties. The top surface of the micro-pillars was protected by a photoresist in advance. Subsequently, fluorination treatment was conducted under the fluoroalkyl silane atmosphere. After cleaning the protective layer, micro-pillars with lyophilic top surfaces and lyophobic side walls and bottom were achieved. Compared with unmodified silicon templates, no apparent morphology changes could be observed in modified templates (the right inset of Figure 2a). For both symmetrical and asymmetrical modifications, the contact angles (CA) increased ( Figure S4a-f, Supporting Information). In particular, the symmetrically modified templates exhibited a more lyophobic state. It is worth mentioning that the larger CA for vertical direction was due to the microstructures on the silicon template ( Figure  S4d-f, Supporting Information). A droplet (2 μL) was dropped onto the surface of the asymmetrically modified template and inverted 180°, and the droplet did not fall ( Figure S4g, Supporting Information). The results indicated that the modified template presented an overall lyophobic state and lyophilic top surface.
The LICB assembly process was performed in order to obtain 1D micro-wire arrays. In the beginning, a droplet of nanosheet suspension was sandwiched between a micro-pillar template and a targeted substrate. Due to the lyophilic top surfaces, the liquid film was pinned under these highly adhesive micro-pillars. With the evaporation of the solvent, the integrated liquid film was ruptured into separated liquid bridges. Under the Laplace pressure, discrete liquid films were further dewetting until the solvent evaporated completely. Last, the micro-wire arrays were patterned on the substrate after removing the silicon template. As shown in Figure 2c, SEM was performed to investigate the morphology of the oriented 1D arrays. The assembled micro-wires of PDA-SGO/PEDOT were size-uniform and position-precise and composed of tightly packed nanosheets ( Figure S5, Supporting Information). The energy dispersive X-ray spectroscopy (EDS) mapping of SEM image proves the structure of micro-wires ( Figure  S6, Supporting Information). The C element was mainly distributed in the position of micro-wire. The characteristic elements of N, O, and S appearing on the substrate were mainly due to the fact that a thin liquid film dried up before it was completely separated during the dewetting process. The dimension of wire structures was determined by confocal microscope, as shown in Figure 2d. The micro-wires exhibited an average height Methodology and mechanism for the dewetting and assembly process of regularly arranged 1D arrays. a) The asymmetric wettability modification process was performed to obtain micro-pillars with an overall lyophobic state and lyophilic top surface. The schematic diagram on the top shows the micro-pillar surface structure before modification (left) and after modification (right), and the corresponding SEM images of the micro-pillar template before modification and after modification are provided. b) Schematic illustration of the assembly process for 1D arrays: construction of sandwich-like structures, formation of continuous liquid film, rupture into separated liquid bridges, and dewetting into micro-wires. c) SEM image of PDA-SGO/PEDOT micro-wire arrays. d) Confocal microscope image of PDA-SGO/PEDOT micro-wire arrays and e) corresponding height profile. f) Optical microscope image and g) Raman scanning map of PDA-SGO/PEDOT micro-wire arrays, in which micro-wires and blank regions can be distinguished clearly. h) The fitted Raman spectra of micro-wires and blank region. i) The micro-wires can be assembled in a large range. The SEM magnification is 300. Figure 3. a) Schematic illustration of micro-wire assembly process with low concentration, suitable concentration, and high concentration. The corresponding assembled microwires with b) low concentration, c) suitable concentration, and d) high concentration. e-k)The adhesive robustness on the glass of 1D arrays is evaluated by finger rubbing and tape peeling (scotch tape). Schematic illustration of 1D arrays with original state (e), finger rubbing (f), and tape peeling (g). The morphology and position of 1D arrays based on PDA-SGO/PEDOT remain basically unchanged (h 1 ) after finger rubbing (h 2 ) and tape peeling (h 3 ). Other control samples of GO (i 1 ), SGO (j 1 ), and SGO/PEDOT (k 1 ) show different degrees of damage after finger rubbing (h 2 -k 2 ) and tape peeling (h 3 -k 3 ). of 2.14 μm and width of 3.53 μm (Figure 2e). Raman mapping was scanned to further demonstrate the assembled patterns. By scanning the intensity at 1379 cm −1 , the position of the micro-wires and blank region were clearly distinguished (Figure 2f,g). The fitted Raman curves of micro-wires and blank region are shown in Figure 2h (original Raman curves are shown in Figure S7, Supporting Information). More importantly, the capillary bridgemediated assembly method allowed for the assembly of 1D arrays on a large-scale (Figure 2i).
Morphology was observed and dimensions were measured for micro-wires assembled under different concentrations ( Figure  S8, Supporting Information). At low concentrations (1-2 mg mL −1 ), the micro-wires are basically disconnected, which is attributed to inadequate nanosheets for forming a continuous liq-uid bridge. As the concentration increases, the micro-wires become wider and higher. At 4 mg mL −1 , the micro-wires exhibit relatively uniform and smooth morphology. Further improving the concentration, excessive assembly occurs at 5 mg mL −1 , in which the surface of micro-wires is rough and meandering. By adjusting the concentration of suspension, micro-wires are available in heights ranging from 0.91 to 2.62 μm and widths ranging from 0.83 to 3.58 μm ( Figure S9, Supporting Information). The dewetting and shrinking processes under different concentrations are illustrated in Figure 3a. The liquid bridges present two meniscus lateral surfaces because of the surface tension of liquid and the wettability of pillars' top and substrate. With the evaporation of solvent, the nanosheets migrate along with three-phase-contact line and finally stack together to construct www.advancedsciencenews.com www.advmatinterfaces.de micro-wires. Under low concentrations (1-2 mg mL −1 ), the nanosheets accumulate loosely and the height cannot reach the neck of liquid bridge. As shown in Figure 3b, the micro-wires are thin and short in this state. When the concentration reaches a suitable value (4 mg mL −1 ), the micro-wires become dense, straight, and continuous (Figure 3c). Further increasing concentration, the height of stacked nanosheet exceeds the neck of liquid bridge, and under the action of gravity, it collapses on the bottom structure randomly, resulting in the phenomenon of overassembly (Figure 3d). For electrical application, incontiguous, loose, and random packaging of nanosheets are not conducive to carrier transportation. These results indicate that the optimal conditions for assembly 1D structures can be evaluated by adjusting the concentration of nanosheet suspension.
The adhesion tests of finger rubbing and tape peeling tests were conducted to investigate the stability of assembled microwires on the target substrate (Figure 3e-g). SEM images of original 1D arrays composed of PDA-SGO/PEDOT, GO, SGO, and SGO/PEDOT are shown in Figure 3d 1 -g 1 . Experiencing finger rubbing and tape peeling, the position and morphology of the PDA-SGO/PEDOT micro-wires remain largely unchanged ( Figure 3h 1 -h 3 ). The fingers rub along the direction parallel to the micro-wires, and the surface of the nanosheets is smoothed; In the process of tape peeling, a part of nano-sheets is pulled up by tape; whereas, all the other contrast samples have varying degrees of damage (Figure 3i-k). Notably, the reason that GO presents different initial stripe spacing is the GO micro-wires assembled at the gap between the micro-pillars of the silicon template. Finger rubbing can cause warping of the nanosheets and even displacement of micro-wires for GO, SGO, and SGO/PEDOT (Figure 3i 2k 2 ); Tape peeling can take away large nanosheets from the substrate while leaving a small amount on the substrate (Figure 3i 3k 3 ). Therefore, the robust stability of PDA-SGO/PEDOT microwires on substrate verifies the effectiveness of the PDA bionic adhesion strategy.
The 1D arrays containing PDA-SGO/PEDOT nanosheets were patterned onto Si wafers with an oxide layer. Au electrodes were evaporated onto micro-wires through a mask, which acted as source (S) and drain (D) electrodes, respectively. The schematic illustration of the obtained FET device with bottom gate/top contact configuration [32] is shown in Figure 4a. The silicon in the wafer acted as the bottom gate electrode for the three-electrodes FET, the silicon dioxide layer acted as the dielectric layer, and the micro-wires formed a parallel arrangement of channels. The representative SEM image of the 1D array channels between S-D electrodes is illustrated in Figure 4b. To evaluate the semiconductor parameters of the fabricated FETs, electrical measurement was performed on 1D channels-based FETs. As shown in Figure 4c,d, typical output and transfer plots of PDA-SGO/PEDOT-4 were presented. The average carrier mobility (μ) of PDA-SGO/PEDOT-4 was calculated as 5.84 × 10 −3 cm 2 V −1 s −1 , (ranging from 5.03 × 10 −3 to 6.38 × 10 −3 cm 2 V −1 s −1 ). Compared with other micro-wires assembled with different concentrations ( Figure S10, Supporting Information, Figure 4e), regular arrangements and uniform stacking of nanosheets in PDA-SGO/PEDOT-4 facilitate carrier transportation. The representative output and transfer curves of GO and commercially available PEDOT:PSS micro-wires are shown in Figure S11a-d, Supporting Information, which exhibits an average mobility of 0.016 × 10 3 and 0.94 × 10 3 cm 2 V −1 s −1 , respectively (Figure 4f). The results indicate that the improvement in electrical performance is attributed to the -stacking between PEDOT and GO. [14] Furthermore, the electrical measurements were conducted on FETs based on film-structure channels, and the output and transfer curves are shown in Figure S11e,f, Supporting Information. The calculated carrier mobility was determined as 0.14 × 10 −3 cm 2 V −1 s −1 , as shown in Figure 4f. The decline of carrier mobility in the pristine film shows random alignment cannot form an efficient carrier transportation path.

Conclusion
In this work, we implemented the large-scale assembly of 1D arrays based on adhesive nanosheets. The assembled 1D microwire arrays exhibited homogeneous morphology and precise position. Inspired by the marine mussel, PDA was incorporated into the composite system. Extraordinary stability of micro-wires could be enabled owing to the bionic adhesion strategy. In particular, the morphology and position of micro-wires remained unchanged with finger rubbing and tape peeling. The dimension of micro-wires was tunable by controlling the concentration of suspension. The FET devices based on PDA-SGO/PEDOT nanosheets presented enhanced carrier mobility compared with GO and PEDOT:PSS because of the presence of -interaction. In addition, FETs possessing separated 1D arrays showed improved electrical performance compared with pristine films. The bioinspired 1D arrays in this work provided an effective approach for achieving field-effect transistors and facilitated the stability and practicability of devices.
Preparation of PDA-SGO/PEDOT Nanosheets: First, SGO was prepared. NaNO 2 (0.36 g) was dissolved in ice water (200 mL) and 4-ABS (0.92 g) was added gradually with stirring and cooling. After 4-ABS was completely dissolved, HCl (1.38 mL) was dripped into the aforementioned solution and stirred in an ice bath for 30 min. GO (0.5 g) was slowly dispersed in the above solution and stirred in the ice bath for 4 h. The product was centrifuged several times and dried afterward in the freeze-dryer. Then, PDA was grafted on SGO to obtain PDA-SGO. SGO powder was dispersed with a concentration of 1 mg mL −1 and tip sonicated for 30 min. DA·HCl (20 mg) was dissolved in Tris-HCl buffer solution (0.1 m, 20 mL) and stirred for 30 min to obtain PDA. SGO suspension was mixed with PDA solution and reacted for 2 h. Next, EDOT/ethanol solution (76 μL of EDOT, 100 mL of ethanol) was prepared and added to the above mixture with stirring for 30 min until EDOT was completely dissolved. FeCl 3 solution (0.5 mg mL −1 , 4 mL) was dripped into the reaction system drop by drop. After reaction under an ice bath for 24 h, the production was cleaned with water and ethanol several times and freeze-dried, which was marked as PDA-SGO/PEDOT.
Assembly of Micro-Wires-Based Arrays: The silicon micropillars-based template was fabricated on nitrogen-doped, 〈100〉 oriented silicon wafers using a direct laser writing apparatus (Heidelberg DWL200). Asymmetric wettability was performed by the following steps. First, a layer of photoresist was spin-coated on a cover glass with a spinning speed of 4500 rpm for 30 s. Then, the top surface of the template was covered by the glass and irradiated under UV light (365 nm) for 20 s. After removing the cover glass, a layer of cured photoresist was coated on the top surface of the micro-pillars. The silicon template was modified under PFDTMS atmosphere at 80°C for 12 h. The protective film of the photoresist was easily cleaned by ethanol. A 10 μL droplet of PDA-SGO/PEDOT suspensive droplet was dropped onto the silicon template and covered by the target substrate, yielding a sandwich-like structure. The assembly system was placed at room temperature for 24 h to completely evaporate the solvent. The obtained PDA-SGO/PEDOT micro-wires prepared with 4 mg mL −1 suspension were denoted by PDA-SGO/PEDOT-4.
Fabrication of FET Devices: The silicon wafer with a layer of silicon dioxide (300 nm) was used as substrate. The micro-wire arrays were assembled by the aforementioned method. The Au electrodes were thermally evaporated onto the as-prepared 1D arrays as source (S) and drain (D) electrodes, respectively.
Measurement of FETs: The electrical characteristics of the FETs were measured by a Keithley 4200-SCS semiconductor characterization system. All measurements were performed at room temperature in the atmosphere. The carrier mobility was calculated from the following equation: where L and W denote the length and width of the channel. I SD is the current between S and D electrodes. C i and V G represent the gate capacitance per unit area and gate voltage, respectively.

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