Highly Transparent Conducting Electrodes Based on a Grid Structure of Silver Nanowires

: Transparent conducting electrodes (TCEs) formed with silver nanowires (AgNWs) have attracted attention as substitutes for indium tin oxide (ITO). However, the randomly deposited AgNW ﬁlm performs poorly in terms of the transmittance and sheet resistance to serve as a substitute of ITO. To improve the performance of the AgNW ﬁlm, we fabricated a grid-patterned AgNW by modifying the surface energy of the substrate. The hydrophobized surface was selectively etched by UV light through a quartz chrome mask, and a suspension of AgNWs in isopropyl alcohol/ethylene glycol mixture was coated on the substrate by a meniscus dragging deposition process. The grid-patterned AgNW ﬁlm has a lower percolation threshold and a 13% higher ﬁgure-of-merit value compared to the randomly deposited AgNW ﬁlm. The transparent thin ﬁlms with a grid structure of AgNWs exhibit the high electrical conductivity with a sheet resistance of 33 Ohm/sq at a transmittance of 92.7% ( λ = 550 nm).


Introduction
Transparent conducting electrodes (TCEs) have been actively researched for use in electronic devices, such as transparent heaters, displays, touch screens, solar cells, and wearable devices [1,2]. Indium tin oxide (ITO) is the most commonly used material for transparent electrodes. However, the scarcity of indium, the need for high vacuum in the processing of ITO, and the rigidity of ITO are the drawbacks of using the ITO electrode. Although efforts have been made to overcome these flaws, the improvements made are inadequate to satisfy the demand of future electronics [3]. Thus, numerous studies are devoted to replace ITO with other conductive materials. Potential candidates include metal oxides [4,5], carbon nanotubes [6,7], graphene [2,8], metal nanoparticle [9], and metal nanowires [10,11]. Among these materials, silver nanowire (AgNW) is being spotlighted, owing to its high conductivity and flexibility [12]. In addition, a suspension of AgNWs can be easily mass-produced by the polyol method [13][14][15].
Many researchers have fabricated transparent electrodes with AgNWs by different coating methods. Numerous methods, such as spin-coating [16], spray-coating [17], vacuum filtration [18], and Langmuir-Blodgett assembly [19], are used for depositing nanomaterials on a substrate. Most of the structures of AgNWs deposited by these methods have a random network structure. However, the transmittance and sheet resistance of the randomly deposited AgNW films are not adequate to replace ITO. To overcome this issue, researchers have tried to control the radius and length of metal nanowires [20], their orientation [11], or have mixed the metal nanowires with other conductive materials [21][22][23][24]. In addition, there have been efforts to fabricate patterned AgNW films by polymer-assisted metal nanowire printing [25], spin-coating on selectively oxidized surface of a polydimethylsiloxane (PDMS) substrate [26], dry-film photoresist process [27], and selective control of adhesion between AgNWs and substrates [28]. 2 of 11 In this study, a new approach for fabricating AgNW grid-patterned electrodes is realized based on dewetting-driven micropatterning, driven by surface energy gradient [29]. Precise control of dewetting phenomena was achieved by combination of optimizing rheological properties of AgNW dispersion by adding ethylene glycol (EG) and forming uniform liquid thin film on the octadecyltrichlorosilane (OTS) grid-patterned substrate. Meniscus dragging deposition (MDD), previously reported by our group [8] is used to produce uniform liquid thin film on OTS pre-patterned substrate and UV-ozone etching is used to pattern OTS layer, generating a large surface gradient between hydrophobic OTS area and hydrophilic etched area. Based on this strategy, wide-ranged control of transmittance and sheet resistance was easily achieved by adjusting coating parameters, including EG concentration, pattern design, and deposition velocity. Especially, reliable control of sheet resistance and transmittance in high transmittance region was easily achieved by lowering percolation threshold of AgNW networked film. In addition to this, our grid-structure AgNW electrodes also exhibit outstanding optical transmittance and conductivity compared to recently reported AgNW based electrodes, confirming validity of our dewetting-driven micropatterning process.

Preparation of Substrates
Glass substrates (25 × 75 mm 2 , Fisher Scientific, Hampton, NH, USA) were cleaned by immersion in a Piranha solution at 180 • C for 6 h. They were then rinsed several times with deionized (DI) water (Human Power II+, Human Corporation, Seoul, Korea) and dried with nitrogen gas blowing. For further hydrophilization, the substrates were treated with air plasma (PDC-32G, Harrick Plasma, manufacturer, Ithaca, NY, USA) for 15 min. Meanwhile, OTS (Sigma-Aldrich, St. Louis, MO, USA) was mixed with anhydrous toluene (99.8%, Alfa Aesar, Haverhill, MA, USA) in a glove box to prepare a 0.3 vol.% OTS solution. OTS was stored in a glove box filled with nitrogen to prevent the polymerization of OTS by reaction with moisture in air. Thereafter, the substrates were immersed in the OTS solution for 1 h. Subsequently, the glass substrates were rinsed with toluene (Daejung, Siheung, Korea) and dried with nitrogen gas blowing. The substrates were then cut into three pieces (25 × 25 mm 2 ) and sonicated (CPX5800H-E, Emerson, St. Louis, MO, USA) with toluene for 2 h. OTS SAM-treated substrates were sequentially rinsed with toluene, ethanol, and DI water. They were then blown dried with nitrogen gas. The contact angles of the substrates were determined by measuring and averaging the contact angles at three points on the surface, using a contact angle measuring machine (Smart Drop, Femtofab, Seongnam, Korea).

Patterning of the Substrates
An OTS SAM-treated substrate was fixed between a grid-patterned quartz chrome mask and fabric-wrapped mask using a custom-made device. It was then exposed to UV light through the quartz chrome mask for 1 h to etch the OTS layer and form a grid pattern. Air was circulated with a pump to prevent the unnecessary oxidation of the substrate during its exposure.

Fabrication of AgNW Films
A 0.5 wt.% suspension of AgNWs (average diameter: 40 nm, average length: 20 µm, and dispersed in IPA) was purchased from DS Hi-metal, South Korea. The coating mixture was prepared by adding EG (Sigma-Aldrich, St. Louis, MO, USA) to 2 vol.% of the AgNW suspension. The AgNW thin films were coated on the OTS-patterned substrate by the MDD method. A piranha-treated deposition plate (76 × 52 mm 2 , MATSUNAMI, Osaka, Japan) was set on the coating substrate at an angle of 30 • . Subsequently, 20 µL of the coating mixture was injected between the deposition plate and substrate. The deposition plate was moved forward and backward three consecutive times using a motion controller (AL1-1515-3S, Micro Motion Technology, Incheon, Korea) at a steady speed of 4 cm/s. One cycle of forward and backward movement is defined as one deposition (deposition number (DN) = 1). A DN of 3 was commonly used for the deposition of AgNWs. After the liquid film was fully dried, the substrate was annealed at 200 • C for 1 h in ambient air.

Characterization of AgNW Films
Optical images of AgNW films were obtained by digital photography (Lumix DMC-LX5, Panasonic, Osaka, Japan) and field-emission scanning electron microscopy (Carl Zeiss, Oberkochen, Germany) images were obtained by a Carl Zeiss Sigma system. The absorbance of the AgNW dispersion and the transmittance spectra of the AgNW films were recorded by an UV-vis spectrophotometer (V-670, JASCO, Easton, MD, USA). The measurements were performed at three locations on the films. The sheet resistance (R Sh ) of the films was measured using a four-point probe station (M3P302-system, MSTECH, Anyang, Korea) with a Keithley 2612A source meter.

Fabrication of TCEs Based on Grid-Patterned AgNWs
The fabrication procedure for the grid-structured AgNW films is illustrated in Figure 1. The glass substrates were cleaned and hydrophilized by piranha treatment and air plasma treatment to facilitate the formation of OTS SAMs on the glass substrates. Hydrophobic glass substrates with a water contact angle of approximately 100 • -110 • were obtained after immersion in the OTS solution for 1 h. The OTS SAM was selectively etched by exposing the substrate to UV light through a grid-patterned quartz chrome mask, generating a gridpatterned hydrophilic region [30,31]. A uniform liquid thin film, containing AgNWs was formed on the OTS pre-patterned substrate via MDD, resulting in spontaneous dewetting of liquid thin film. Then, a grid-structured AgNW electrode was obtained after solvent drying. Japan) was set on the coating substrate at an angle of 30°. Subsequently, 20 μL of the coating mixture was injected between the deposition plate and substrate. The deposition plate was moved forward and backward three consecutive times using a motion controller (AL1-1515-3S, Micro Motion Technology, Incheon, Korea) at a steady speed of 4 cm/s. One cycle of forward and backward movement is defined as one deposition (deposition number (DN) = 1). A DN of 3 was commonly used for the deposition of AgNWs. After the liquid film was fully dried, the substrate was annealed at 200 °C for 1 h in ambient air.

Characterization of AgNW Films
Optical images of AgNW films were obtained by digital photography (Lumix DMC-LX5, Panasonic, Osaka, Japan) and field-emission scanning electron microscopy (Carl Zeiss, Oberkochen, Germany) images were obtained by a Carl Zeiss Sigma system. The absorbance of the AgNW dispersion and the transmittance spectra of the AgNW films were recorded by an UV-vis spectrophotometer (V-670, JASCO, Easton, MD, USA). The measurements were performed at three locations on the films. The sheet resistance (RSh) of the films was measured using a four-point probe station (M3P302-system, MSTECH, Anyang, Korea) with a Keithley 2612A source meter.

Fabrication of TCEs Based on Grid-Patterned AgNWs
The fabrication procedure for the grid-structured AgNW films is illustrated in Figure 1. The glass substrates were cleaned and hydrophilized by piranha treatment and air plasma treatment to facilitate the formation of OTS SAMs on the glass substrates. Hydrophobic glass substrates with a water contact angle of approximately 100°-110° were obtained after immersion in the OTS solution for 1 h. The OTS SAM was selectively etched by exposing the substrate to UV light through a grid-patterned quartz chrome mask, generating a grid-patterned hydrophilic region [30,31]. A uniform liquid thin film, containing AgNWs was formed on the OTS pre-patterned substrate via MDD, resulting in spontaneous dewetting of liquid thin film. Then, a grid-structured AgNW electrode was obtained after solvent drying.

Principle of Pattern Formation
AgNWs were dragged to the hydrophilized area via the dewetting of the liquid thin film of the coating mixture on the substrate. The solvents used in the coating mixture are IPA and EG, which have a large difference in surface tension (IPA: 21.4 mN/m, EG: 48.4 mN/m) [32]. Therefore, IPA vaporized at a higher rate than EG, which resulted in a gradual increase in the concentration of EG after the deposition of the mixture on the substrate. The higher EG concentration leads to the increment of the surface tension of the liquid thin film on the substrate. Therefore, the liquid film is dewetted from the hydrophobic OTS SAM area and moves to the plasma-etched area with a high surface energy from hydroxyl groups. Simultaneously, the AgNWs are transferred to the plasmaetched area by the dewetting-driven liquid flux, resulting in the formation of the gridstructured AgNW film. Figure 2 shows the optical images of the patterned AgNW films with various grid widths (1, 2, 5, and 10 µm) at a fixed spacing of 40 µm. When the width of the grid was 1 µm, the AgNWs tended to agglomerate at the junction of the grid (Figure 2a). In contrast, as shown in Figure 2c,d, when the width of grid was 5 and 10 µm, the AgNWs are scattered in the grid because the liquid was pinned at the border of the grid and it spread in the hydrophilic area [33] and the amount of AgNW also is not enough to fill the grid area. However, when the width was 2 µm (Figure 2b), the well-defined grid-structure of AgNW film was obtained (Figure 2b). Therefore, the following experiments were conducted with a fixed width of 2 µm.

Principle of Pattern Formation
AgNWs were dragged to the hydrophilized area via the dewetting of the liquid thin film of the coating mixture on the substrate. The solvents used in the coating mixture are IPA and EG, which have a large difference in surface tension (IPA: 21.4 mN/m, EG: 48.4 mN/m) [32]. Therefore, IPA vaporized at a higher rate than EG, which resulted in a gradual increase in the concentration of EG after the deposition of the mixture on the substrate. The higher EG concentration leads to the increment of the surface tension of the liquid thin film on the substrate. Therefore, the liquid film is dewetted from the hydrophobic OTS SAM area and moves to the plasma-etched area with a high surface energy from hydroxyl groups. Simultaneously, the AgNWs are transferred to the plasma-etched area by the dewetting-driven liquid flux, resulting in the formation of the grid-structured AgNW film. Figure 2 shows the optical images of the patterned AgNW films with various grid widths (1, 2, 5, and 10 μm) at a fixed spacing of 40 μm. When the width of the grid was 1 μm, the AgNWs tended to agglomerate at the junction of the grid (Figure 2a). In contrast, as shown in Figure 2c,d, when the width of grid was 5 and 10 μm, the AgNWs are scattered in the grid because the liquid was pinned at the border of the grid and it spread in the hydrophilic area [33] and the amount of AgNW also is not enough to fill the grid area. However, when the width was 2 μm (Figure 2b), the well-defined grid-structure of AgNW film was obtained (Figure 2b). Therefore, the following experiments were conducted with a fixed width of 2 μm.

Effect of the Deposition Conditions
The thickness of the liquid thin film obtained by the MDD process can be defined by the Landau-Levich-Derjaquin (LLD) model, which is expressed in the form of Equation (1) [34].

Effect of the Deposition Conditions
The thickness of the liquid thin film obtained by the MDD process can be defined by the Landau-Levich-Derjaquin (LLD) model, which is expressed in the form of Equation (1) [34].
where µ is the viscosity of the coating liquid, V is the deposition velocity (DV), γ is the surface tension of the liquid, and R is the radius of the meniscus. The radius of the meniscus, R, can be controlled by adjusting the injection volume of the coating mixture. V and R can be changed to control the thickness of the liquid film without a change in the liquid composition. Therefore, in this study, only the deposition velocity was changed because increasing the injection volume of the coating mixture also increases the waste of AgNWs. Equation (1) explains how the properties of the electrodes change with the DV. According to the equation, a higher DV increases the film thickness. Therefore, the solvent evaporation time was lengthened. Further, a longer evaporation time maximized the effect of the Marangoni flow. The Marangoni flow occurs in the film of a liquid mixture because of non-uniform solvent evaporation, causing the deposition of AgNWs on the substrate [35]. Therefore, numerous AgNWs were already deposited and anchored on the substrate during the dewetting of the liquid film. Hence, AgNWs tended to form a hybrid structure of web networks in the square area of the grid pattern ( Figure 3). The grid-structured AgNW electrodes on the glass substrate were annealed at 200 • C for 60 min in order to enhance the electrical conductivity by removing the PVP layer and EG residue. The transmittance and sheet resistance also changed with the variation in DV (Figure 4). The transmittance decreased as the DV increased, whereas the sheet resistance increased, owing to the web network of AgNWs in the square area of the grid. At the lowest DV (0.17 cm/s) that could be programmed by the motion controller, a very thin liquid film formed and rapidly reached the critical thickness. Therefore, the liquid film was not maintained and it contracted during the coating process. Therefore, few AgNWs were deposited on the substrate, resulting in different tendencies in transmittance and sheet resistance.
surface tension of the liquid, and R is the radius of the meniscus. The radius of the meniscus, R, can be controlled by adjusting the injection volume of the coating mixture. V and R can be changed to control the thickness of the liquid film without a change in the liquid composition. Therefore, in this study, only the deposition velocity was changed because increasing the injection volume of the coating mixture also increases the waste of AgNWs.
Equation (1) explains how the properties of the electrodes change with the DV. According to the equation, a higher DV increases the film thickness. Therefore, the solvent evaporation time was lengthened. Further, a longer evaporation time maximized the effect of the Marangoni flow. The Marangoni flow occurs in the film of a liquid mixture because of non-uniform solvent evaporation, causing the deposition of AgNWs on the substrate [35]. Therefore, numerous AgNWs were already deposited and anchored on the substrate during the dewetting of the liquid film. Hence, AgNWs tended to form a hybrid structure of web networks in the square area of the grid pattern ( Figure 3). The grid-structured AgNW electrodes on the glass substrate were annealed at 200 °C for 60 min in order to enhance the electrical conductivity by removing the PVP layer and EG residue. The transmittance and sheet resistance also changed with the variation in DV (Figure 4). The transmittance decreased as the DV increased, whereas the sheet resistance increased, owing to the web network of AgNWs in the square area of the grid. At the lowest DV (0.17 cm/s) that could be programmed by the motion controller, a very thin liquid film formed and rapidly reached the critical thickness. Therefore, the liquid film was not maintained and it contracted during the coating process. Therefore, few AgNWs were deposited on the substrate, resulting in different tendencies in transmittance and sheet resistance.  The structure of AgNW networks can be changed by controlling the concentration of EG (Figure 5). At a low EG concentration ( Figure 5), there were more web networks of AgNWs than grid networks because the dewetting force was very weak to pull all the AgNWs to the hydrophilic area. At the optimal concentration ( Figure 5), the grid structure was mainly formed because of the increased dewetting force. However, the AgNWs began to concentrate at the junction of the grid and lose contact with the side of the grid when the EG concentration was too high ( Figure 5). The optoelectrical properties also changed upon varying the EG concentration ( Figure 4). The transmittance increased as the concentration increased, owing to the convergence of the AgNWs toward the etched area, whereas the sheet resistance decreased because of the disconnections at the side. In addition, the web networks diminished, causing an increase in transmittance and a decrease in sheet resistance. The structure of AgNW networks can be changed by controlling the concentration of EG (Figure 5). At a low EG concentration ( Figure 5), there were more web networks of AgNWs than grid networks because the dewetting force was very weak to pull all the AgNWs to the hydrophilic area. At the optimal concentration ( Figure 5), the grid structure was mainly formed because of the increased dewetting force. However, the AgNWs began to concentrate at the junction of the grid and lose contact with the side of the grid when the EG concentration was too high ( Figure 5). The optoelectrical properties also changed upon varying the EG concentration ( Figure 4). The transmittance increased as the concentration increased, owing to the convergence of the AgNWs toward the etched area, whereas the sheet resistance decreased because of the disconnections at the side. In addition, the web networks diminished, causing an increase in transmittance and a decrease in sheet resistance.

Performance of Grid-Structured AgNW Electrodes
The expanded images of the randomly deposited and grid-patterned AgNW films are shown in Figure 6. Large junctions and dense bundles of AgNWs can be observed in the grid-patterned AgNW film, unlike in the random film. The difference in the structures between these films resulted in differences in their performance as transparent electrodes. The AgNW electrode with a grid structure has bundles of AgNWs on all sides of the grid and junctions at the intersection of the grid. The bundles and junctions offer many points of contact for the AgNWs; they facilitate many electron movement paths and lower the sheet resistance of the electrode. The AgNWs are well structured along the grid-patterned

Performance of Grid-Structured AgNW Electrodes
The expanded images of the randomly deposited and grid-patterned AgNW films are shown in Figure 6. Large junctions and dense bundles of AgNWs can be observed in the grid-patterned AgNW film, unlike in the random film. The difference in the structures between these films resulted in differences in their performance as transparent electrodes. The AgNW electrode with a grid structure has bundles of AgNWs on all sides of the grid and junctions at the intersection of the grid. The bundles and junctions offer many points of contact for the AgNWs; they facilitate many electron movement paths and lower the sheet resistance of the electrode. The AgNWs are well structured along the grid-patterned area with an average thickness of 130.5 nm (Figure 6f). In addition, owing to the open space of the grid structure, grid-structured AgNW film has higher transmittance than the randomly networked AgNW film. Consequently, the grid-structured AgNW film can display superior performance to that of the randomly deposited AgNW films, as shown in Figure 7. The grid-structured films have higher transmittance than the randomly deposited films at the same sheet resistance value. The dots represent the experimental values of transmittance at a wavelength of 550 nm and sheet resistance. The smooth line was obtained through nonlinear curve fitting. The fitted equations of both curves express the transmittance (T) as a function of the sheet resistance ( ) (Equations (2) and (3)).
In general, the transmittance and sheet resistance can be related to a figure-of-merit (FoM), which is defined as the electrical conductivity-to-optical conductivity ratio (Equation (4)) [36].
where σ is the electrical conductivity and σ is the optical conductivity of the conductive film.
The value of is called the FoM, and it is used for the quantitative analysis of the Consequently, the grid-structured AgNW film can display superior performance to that of the randomly deposited AgNW films, as shown in Figure 7. The grid-structured films have higher transmittance than the randomly deposited films at the same sheet resistance value. The dots represent the experimental values of transmittance at a wavelength of 550 nm and sheet resistance. The smooth line was obtained through nonlinear curve fitting. The fitted equations of both curves express the transmittance (T) as a function of the sheet resistance (R sh ) (Equations (2) and (3)). film as a TCE is better than that of the randomly deposited grid, because the FoM value of the former is 13% higher than that of the latter. Figure 7 illustrates the relationship between the transmittance and FoM for the grid-structured and randomly deposited films. The FoM value of the grid-structured film is higher for the overall transmittance. The percolation threshold theory also supports this result. According to this theory, interconnections between neighboring wires of small dimensions can form larger groups over a surface if the 2D surface concentration reaches a specific threshold that is large enough so that any potential boundary effect can be negligible [37], and this point is called the percolation threshold. Equation (5) expresses the theory by relating the sheet resistance and surface concentration. (5) where X is the surface concentration of AgNWs, is the percolation threshold, and α is the percolation exponent, which is correlated to the dimensionality of the percolated film.
To obtain the relationship between the transmittance and surface concentration, a controlled amount of AgNW was deposited by continuous back and forth dragging of deposition plate until the solvent was completely vaporized during the MDD, depositing all injected amount of AgNWs. The transmittance was then measured, and the surface concentration was calculated; they were then related by a linear fit. These are shown in Figure 8. In general, the transmittance and sheet resistance can be related to a figure-of-merit (FoM), which is defined as the electrical conductivity-to-optical conductivity ratio (Equation (4)) [36].
where σ DC is the electrical conductivity and σ OP is the optical conductivity of the conductive film. The value of σ OP σ DC is called the FoM, and it is used for the quantitative analysis of the performance of thin transparent electrodes. The higher the FoM is, the better the transparent conductive electrode is. By fitting the experimental values to Equation (4), the FoM values of the grid-structured and randomly deposited AgNW films were determined to be 118 and 108, respectively. We can confirm that the performance of the grid-structured film as a TCE is better than that of the randomly deposited grid, because the FoM value of the former is 13% higher than that of the latter. Figure 7 illustrates the relationship between the transmittance and FoM for the grid-structured and randomly deposited films. The FoM value of the grid-structured film is higher for the overall transmittance.
The percolation threshold theory also supports this result. According to this theory, interconnections between neighboring wires of small dimensions can form larger groups over a surface if the 2D surface concentration reaches a specific threshold that is large enough so that any potential boundary effect can be negligible [37], and this point is called the percolation threshold. Equation (5) expresses the theory by relating the sheet resistance and surface concentration.
where X is the surface concentration of AgNWs, X c is the percolation threshold, and α is the percolation exponent, which is correlated to the dimensionality of the percolated film.
To obtain the relationship between the transmittance and surface concentration, a controlled amount of AgNW was deposited by continuous back and forth dragging of deposition plate until the solvent was completely vaporized during the MDD, depositing all injected amount of AgNWs. The transmittance was then measured, and the surface concentration was calculated; they were then related by a linear fit. These are shown in Figure 8. A nonlinear curve fitting was performed to estimate the percolation threshold of each electrode ( Figure 8). The estimated percolation exponent α values are 1.33 (grid) and 1.31 (random), which are very close to the theoretical value of the 2D model α 1.33 . The percolation threshold of the grid-patterned electrode 1.246 is much lower than that of the randomly deposited one 2.944 , implying that the grid-patterned electrode is more AgNW-efficient. Finally, Figure 9 compares the optoelectrical properties of our AgNW grid film with those of other AgNW conductive films [8,[38][39][40][41]. Our AgNW grid film exhibits good performance over a wide range of transmittance values.

Conclusions
In this study, AgNW films with a grid structure were formed by the MDD, which is the efficient method for fabricating random networks. An AgNW suspension containing EG was used to form a thin liquid film on the selectively hydrophilized substrate surface with OTS. The EG concentration increased as IPA vaporized, and the surface tension of A nonlinear curve fitting was performed to estimate the percolation threshold of each electrode ( Figure 8). The estimated percolation exponent (α) values are 1.33 (grid) and 1.31 (random), which are very close to the theoretical value of the 2D model (α = 1.33). The percolation threshold of the grid-patterned electrode (X c = 1.246) is much lower than that of the randomly deposited one (X c = 2.944) , implying that the grid-patterned electrode is more AgNW-efficient. Finally, Figure 9 compares the optoelectrical properties of our AgNW grid film with those of other AgNW conductive films [8,[38][39][40][41]. Our AgNW grid film exhibits good performance over a wide range of transmittance values. A nonlinear curve fitting was performed to estimate the percolation threshold of each electrode ( Figure 8). The estimated percolation exponent α values are 1.33 (grid) and 1.31 (random), which are very close to the theoretical value of the 2D model α 1.33 . The percolation threshold of the grid-patterned electrode 1.246 is much lower than that of the randomly deposited one 2.944 , implying that the grid-patterned electrode is more AgNW-efficient. Finally, Figure 9 compares the optoelectrical properties of our AgNW grid film with those of other AgNW conductive films [8,[38][39][40][41]. Our AgNW grid film exhibits good performance over a wide range of transmittance values.

Conclusions
In this study, AgNW films with a grid structure were formed by the MDD, which is the efficient method for fabricating random networks. An AgNW suspension containing EG was used to form a thin liquid film on the selectively hydrophilized substrate surface with OTS. The EG concentration increased as IPA vaporized, and the surface tension of

Conclusions
In this study, AgNW films with a grid structure were formed by the MDD, which is the efficient method for fabricating random networks. An AgNW suspension containing EG was used to form a thin liquid film on the selectively hydrophilized substrate surface with OTS. The EG concentration increased as IPA vaporized, and the surface tension of the film increased. The high surface tension of liquid film and surface energy gradation of OTS patterned substrate results in pre-programmed dewetting of liquid thin film, creating grid-patterned AgNW electrodes. Various structures of AgNWs were formed by controlling