Various Coating Methodologies of WO3 According to the Purpose for Electrochromic Devices

Solution-processable electrochromic (EC) materials have been investigated widely for various applications, such as smart windows, reflective displays, and sensors. Among them, tungsten trioxide (WO3) is an attractive material because it can form a film via a solution process and relative low temperature treatment, which is suitable for a range of substrates. This paper introduces the slot-die and electrostatic force-assisted dispensing (EFAD) printing for solution-processable methods of WO3 film fabrication. The resulting films were compared with WO3 films prepared by spin coating. Both films exhibited a similar morphology and crystalline structure. Furthermore, three different processed WO3 film-based electrochromic devices (ECDs) were prepared and exhibited similar device behaviors. In addition, large area (100 cm2) and patterned ECDs were fabricated using slot-die and EFAD printing. Consequently, slot-die and EFAD printing can be used to commercialize WO3 based-ECDs applications, such as smart windows and reflective displays.

Tungsten trioxide (WO 3 ) is a widely used EC material owing to its facile fabrication method via solution processing, such as a sol-gel technique at relatively low temperatures (~60 • C), which enables the use of conventional plastic substrates including polyethylene terephthaltate (PET) [3,13,36]. To apply WO 3 in information displays or smart mirror/windows, it is necessary to develop manufacturing techniques to suit the characteristics of the application. Although spin-casting is one representative method to form WO 3 films and investigate various EC properties in detail [36], there are demands to make shapes or letters for information transfer (display application). In addition, the previous methods may not be efficient for large area manufacturing with respect to film uniformity and quantity of materials consumption during film formation (smart window/mirror application). Therefore, an appropriate methodology is needed to commercialize WO 3 -based EC device (ECD) applications.
This paper introduces slot-die coating and electrohydrodynamic (EHD) jet printing of a WO 3 ink for a uniform film coating on a large area and patterning, respectively. A slot-die coating method allows the pre-metered and continuous coating of ink flowing from the downstream meniscus, forming at a horizontal slot-die head edge while the ink is supplied consistently in the slot-die head [15,37]. Therefore, it is considered to be a cost-effective and easily scalable technique for the high throughput and large area film production field [38][39][40]. On the other hand, EHD jet printing enables uniform and continuous line patterning by forming a jet stream from the nozzle tip, where the electrostatic force applied between a nozzle tip and a substrate can deform the meniscus of an ink to eject droplets consistently [11,13,14]. Among EHD printing modes, electrostatic force-assisted dispensing (EFAD) mode, which applies very low external voltage between the nozzle tip and substrate for the formation of continuous ink flow between the two, have been frequently applied to make patterns for electronic materials with good pattern fidelity and morphological uniformity [36].
In this study, the coating and morphological properties of WO 3 films fabricated by slot-die coating and EFAD printing were characterized and compared with films produced by spin-coating. In addition, the EC performance of the device fabricated by each printing method was compared. The ECDs showed similar morphological properties of WO 3 and device behaviors, such as optical modulation, switching speed, and coloration efficiency. These results imply that the slot-die and EFAD printing can be commercialized for diverse WO 3 -based ECDs applications. To demonstrate the feasibility, the large scale (100 cm 2 ) and patterned WO 3 -based ECDs were produced by slot-die and EFAD printing, respectively. The obtained ECDs can be used as smart windows and reflective display applications.

Materials
Except for ITO glass (sheet resistance: 15 Ω/sq, Asahi Glass Co., Tokyo, Japan), all materials were purchased from Sigma-Aldrich (St. Louis, MO, USA). Tungsten trioxide (WO 3 ) nanoparticles were prepared using a method reported elsewhere [3,13,14,32]. To synthesize WO 3 nanoparticles, tungsten (W) powder (7.0 g) was added to hydrogen peroxide solution (31 wt.% in water, 90 g) and allowed to react at 100 • C. After 5 h, the solvent was evaporated by using rotary evaporator, giving WO 3 nanoparticles. The characterizations of resulting WO 3 nanoparticles using photograph, scanning electron microscopy (SEM) and X-ray diffraction (XRD) are shown in Figure S1a-c, respectively. A WO 3 suspension was fabricated by mixing WO 3 nanoparticles, DI-water, and isopropyl alcohol (IPA) at a weight ratio of 0.3:0.35:0.35, respectively, followed by sonication for 4 h to obtain a homogeneous dispersion. Before WO 3 suspension deposition, ITO glasses were cleaned sequentially with acetone (15 min) and IPA (15 min) with sonication, and a UV/ozone treatment was then conducted for 10 min prior use. In addition, propylene carbonate containing LiClO 4 (0.5 M) and ferrocene (0.05 M) was used as an electrolyte.

Fabrication of Tungsten Oxide Film
In this study, three different processes (spin coating, slot-die, and EFAD printing) were used to fabricate WO 3 films, and the thickness of the WO 3 films was maintained for a fair comparison. To fabricate the same film thickness, in the case of spin coating, the WO 3 suspension was spin-coated on ITO at 5000 rpm for 20 s. For the slot-die process, a WO 3 suspension was ejected into the slot-die head at a flow rate of 1 mL/min, and the moving speed was set to 5 mm/s. In EFAD printing, the flow rate of the WO 3 suspension was fixed to 0.5 µL/min, and the applied voltage between the nozzle tip and substrate and printing velocity were 10 V and 3 mm/s, respectively. To prepare WO 3 films with a size of 1.5 cm × 1.5 cm using spin-coating, slot-die and EFAD printing, the WO 3 suspension required approximately 100 µL, 1 µL and 1 µL, respectively. Prior to use, the as-spun WO 3 films were annealed thermally at 60 • C under vacuum for 10 h.

Device Fabrication and Characterization
The WO 3 films obtained by the three different processes were characterized by optical microscopy (OM, ECLIPSE LV100ND, Nikon, Tokyo, Japan), scanning electron microscopy (SEM, S-4800, Hitachi, Tokyo, Japan), X-ray diffraction (XRD, D/MAX2500 VL-PC, Rigaku, Tokyo, Japan), and cyclic voltammetry (Weis 500, WonA Tech., Seoul, Korea). For WO 3 -based ECDs assembly, the WO 3 -coated ITO glass, electrolyte solution, and counter bare ITO glass were sandwiched to form the (ITO/WO 3 /electrolyte/ITO) configuration, in which~88 µm thick double-sided tape was used as a spacer and adhesive. To investigate the performance of ECDs, DC and square-shaped wave voltages were supplied by a potentiostat (Wave Driver 10, Pine Instrument, Durham, NC, USA). In addition, a UV-VIS spectrophotometer (V-730, Jasco, Easton, MD, USA) was used to record the change in transmittance according to the applied voltages. All ECDs in this work were fabricated to a size of 1.5 cm × 1.5 cm to measure characteristics. 3 deposition processes (spin coating, slot-die, and EFAD printing) in this study. EC devices based on WO 3 -coated ITO glass were fabricated using propylene carbonate (PC) as an electrolyte containing 0.5 M LiClO 4 and 0.05 M ferrocene (Fc) (Figure 1a). In this device, the ion storage layer was unnecessary because Fc acted as anodic species. Thus, the device configuration became simple. In the spin-coating process (Figure 1b), the ITO glass should be fixed on a vacuum chamber, followed by casting the WO 3 suspension on ITO glass and spinning the chamber to form a uniform WO 3 film. Although spin coating is a facile method to fabricate the films precisely, it is limited to large area and patterning productions because of the necessity of large amounts of solution and additional post-process. The slot-die and EFAD printing are attractive technologies that allow the fabrication of large area and patterned films by the simple direct printing of a solution. Figure 1c,d show schematic illustrations of the slot-die and EFAD printing processes, respectively. The slot-die head was placed vertically at ITO glass, and the WO 3 suspension was injected through a connected tube. The WO 3 suspension was ejected from slot-die head to ITO glass, while the slot-die head was moved horizontally to form a uniform large-area WO 3 film. In EFAD printing, the WO 3 suspension was filled into a syringe with a nozzle. An electric field between the nozzle tip and ITO glass was applied to enable well-defined and continuous line printing while moving the sample stage. The three different films (obtained by spin coating, slot-die, and EFAD printing) underwent thermal annealing at 60 °C under vacuum prior to use.

Figure 1 presents the WO 3 -based ECD configuration and schematic descriptions of various WO
The WO3 film morphologies obtained by three different processes were investigated by OM, as shown in Figure 2a-c (see the SEM images in the inset). The thickness of three WO3 films were determined to be ~300 nm (see cross-section SEM images in Figure S2). OM showed that all the films except for the EFAD film exhibited a similar shape. The EFAD film showed an overlapped line pattern because it needed to be printed several times to produce the same film thickness as the spincoating and slot-die films. Although the OM image of the EFAD film showed an overlapping line pattern, SEM revealed the three different films to have a similar morphology. Specifically, electrochromic performance is governed by the crystalline structure of WO3. Therefore, the crystalline structure of the WO3 films (spin coating, slot-die, and EFAD) was examined by XRD. Figure 2d shows the XRD patterns of the WO3 films before and after thermal annealing. The as-spun WO3 film (before thermal annealing) showed XRD reflections corresponding to the (002), (200), and (202) planes. After thermal annealing, each WO3 film (spin coating, slot-die, and EFAD) exhibited an enhanced (002) peak intensity, indicating the development of the monoclinic crystalline structure of WO3. Therefore, the electrochromic (EC) performance of each WO3 film-based ECD was expected to be similar because of their comparable film morphology and crystalline structure. An electric field between the nozzle tip and ITO glass was applied to enable well-defined and continuous line printing while moving the sample stage. The three different films (obtained by spin coating, slot-die, and EFAD printing) underwent thermal annealing at 60 • C under vacuum prior to use.
The WO 3 film morphologies obtained by three different processes were investigated by OM, as shown in Figure 2a-c (see the SEM images in the inset). The thickness of three WO 3 films were determined to be~300 nm (see cross-section SEM images in Figure S2). OM showed that all the films except for the EFAD film exhibited a similar shape. The EFAD film showed an overlapped line pattern because it needed to be printed several times to produce the same film thickness as the spin-coating and slot-die films. Although the OM image of the EFAD film showed an overlapping line pattern, SEM revealed the three different films to have a similar morphology. Specifically, electrochromic performance is governed by the crystalline structure of WO 3 . Therefore, the crystalline structure of the WO 3 films (spin coating, slot-die, and EFAD) was examined by XRD. Figure 2d shows the XRD patterns of the WO 3 films before and after thermal annealing. The as-spun WO 3 film (before thermal annealing) showed XRD reflections corresponding to the (002), (200), and (202) planes. After thermal annealing, each WO 3 film (spin coating, slot-die, and EFAD) exhibited an enhanced (002) peak intensity, indicating the development of the monoclinic crystalline structure of WO 3 . Therefore, the electrochromic (EC) performance of each WO 3 film-based ECD was expected to be similar because of their comparable film morphology and crystalline structure. The cyclic voltammograms (CVs) of WO3-coated films fabricated by three different methods were recorded at a scan rate of 25 mV/s from +1.0 to −1.5 V to estimate the electrochemical properties and electrochromic performances of the films ( Figure S3). When applied to a negative potential, significant increasing current densities were measured at each film indicating an activation process for the intercalation of Li + ions into the films. Despite the different coating methods, the shapes of the CV curves were similar for all the coating methods. To examine the EC behaviors of the three different WO3 film-based ECDs, the transmittance variations, according to the applied voltage, were recorded at 350-900 nm (Figure 3a-c). A noticeable decrease in the transmittance spectra was observed at −0.3 V, and the transmittance spectra of ECDs according to applied voltage exhibited similar behavior. The color changes at increasing bias were also analyzed with CIELAB color coordinates. In the bleached state the films had a slight yellowish color with L * , a * and b * of each device are (72. 26 Figure S4a-c). When increasing the applied voltage until −1.5 V, the devices become blue with the similar values of L * , a * and b * . In addition, the transmittance variations at 700 nm of the three different WO3 film-based ECDs as a function of the applied voltage were derived (Figure 3d). As the applied voltage was increased, the transmittance decreased, and ΔT of the three different WO3 film-based ECDs were similar 71.3%, 72.8%, and 72.1% at −1.5 V, respectively. The optical transitions of each ECD were also observed clearly, as shown in Figure 3e. The cyclic voltammograms (CVs) of WO 3 -coated films fabricated by three different methods were recorded at a scan rate of 25 mV/s from +1.0 to −1.5 V to estimate the electrochemical properties and electrochromic performances of the films ( Figure S3). When applied to a negative potential, significant increasing current densities were measured at each film indicating an activation process for the intercalation of Li + ions into the films. Despite the different coating methods, the shapes of the CV curves were similar for all the coating methods. To examine the EC behaviors of the three different WO 3 film-based ECDs, the transmittance variations, according to the applied voltage, were recorded at 350-900 nm (Figure 3a-c). A noticeable decrease in the transmittance spectra was observed at −0.3 V, and the transmittance spectra of ECDs according to applied voltage exhibited similar behavior. The color changes at increasing bias were also analyzed with CIELAB color coordinates. In the bleached state the films had a slight yellowish color with L * , a * and b * of each device are (72.26, 7.67, 16.8), (76.72, 6.24, 8.95), and (71.33, 7.68, 15.21) for spin-coated, slot-die coated, and EFAD films ( Figure S4a-c). When increasing the applied voltage until −1.5 V, the devices become blue with the similar values of L * , a * and b * . In addition, the transmittance variations at 700 nm of the three different WO 3 film-based ECDs as a function of the applied voltage were derived (Figure 3d). As the applied voltage was increased, the transmittance decreased, and ∆T of the three different WO 3 film-based ECDs were similar 71.3%, 72.8%, and 72.1% at −1.5 V, respectively. The optical transitions of each ECD were also observed clearly, as shown in Figure 3e. The EFAD film-based ECD showed slightly different dynamic behavior compared to the spin coating and slot-die film based ECDs. To measure the dynamic device behavior, the transmittance profiles of the ECDs were recorded at 700 nm upon the application of −1.5 V (coloration) and 0 V (bleaching) (Figure 4). The response times of coloration (tc) and bleaching (tb) were defined as the times at which 90% of the maximum transmittance contrast (ΔT) was achieved. The similar coloration (tc) and bleaching (tb) times were obtained as tc = ~14 s and tb = ~10 s (spin-coating), tc = ~12 s and tb = 8.5 s (slot-die) and tc = ~12 s and tb = 9 s (EFAD printing). To examine the coloration efficiency (η) of the three different ECDs, the correlation between the optical density (OD) and charge density (Q) was plotted, as shown in Figure 5. The η value, which is defined as ΔOD/ΔQ, corresponds to the slope of the linear fit in the linear regime. The similar η values of each ECD were recorded as ~40.2 (spincoating), ~38.5 (slot-die), and ~41.0 cm 2 /C (EFAD printing). The EFAD film-based ECD showed slightly different dynamic behavior compared to the spin coating and slot-die film based ECDs. To measure the dynamic device behavior, the transmittance profiles of the ECDs were recorded at 700 nm upon the application of −1.5 V (coloration) and 0 V (bleaching) (Figure 4). The response times of coloration (t c ) and bleaching (t b ) were defined as the times at which 90% of the maximum transmittance contrast (∆T) was achieved. The similar coloration (t c ) and bleaching (t b ) times were obtained as t c =~14 s and t b =~10 s (spin-coating), t c =~12 s and t b = 8.5 s (slot-die) and t c =~12 s and t b = 9 s (EFAD printing). To examine the coloration efficiency (η) of the three different ECDs, the correlation between the optical density (OD) and charge density (Q) was plotted, as shown in Figure 5. The η value, which is defined as ∆OD/∆Q, corresponds to the slope of the linear fit in the linear regime. The similar η values of each ECD were recorded as~40.2 (spin-coating),~38.5 (slot-die), and~41.0 cm 2 /C (EFAD printing).  By applying the advantages of slot-die and EFAD printing, which can print a WO3 suspension on a large and selective area, large area (100 cm 2 ) and patterned WO3-based ECDs were obtained, as shown in Figure 6. The color of the large area WO3-based ECD changed reversibly over the entire area when the bleached state and colored state were observed at 0 V and −1.5 V, respectively (Figure By applying the advantages of slot-die and EFAD printing, which can print a WO 3 suspension on a large and selective area, large area (100 cm 2 ) and patterned WO 3 -based ECDs were obtained, as shown in Figure 6. The color of the large area WO 3 -based ECD changed reversibly over the entire area when the bleached state and colored state were observed at 0 V and −1.5 V, respectively (Figure 6a). Reversible EC behavior was observed in the line patterned WO 3 -based ECD (Figure 6b). In addition, electrochromic letters (YU) were produced using the EFAD printing method (Figure 6c). Accordingly, slot-die and EFAD printing can be used to commercialize WO 3 -based ECD applications, such as smart windows and reflective displays.

Conclusions
WO3-based ECDs were fabricated by spin coating, slot-die, and EFAD printing techniques. The morphology of the WO3 films obtained by the above three processes showed a similar shape, except for the EFAD film, which showed an overlapping line. Although the EFAD film exhibited a different morphology, the crystalline structure was no different compared to spin coating and slot-die, which

Conclusions
WO 3 -based ECDs were fabricated by spin coating, slot-die, and EFAD printing techniques. The morphology of the WO 3 films obtained by the above three processes showed a similar shape, except for the EFAD film, which showed an overlapping line. Although the EFAD film exhibited a different morphology, the crystalline structure was no different compared to spin coating and slot-die, which has a monoclinic structure. In addition, the device performance of three different WO 3 -based ECDs was similar. For example, each WO 3 -based ECD produced from spin-coating, slot-die, and EFAD printing showed similar optical modulation (~71.3%,~72.8%, and~72.1%), response times (t c =~14,~12,~12 s), and coloration efficiencies (~40.2,~38.5,~41.0 cm 2 /C). Large area (100 cm 2 ) and patterned WO 3 electrochromic devices were demonstrated by taking advantage of the slot-die and EFAD printing processing methods. Slot-die and EFAD printing are attractive technologies for commercializing WO 3 -based ECDs into smart windows and reflective displays.