Boosting coloration efficiency in an electrochromic device using an ITO/Ag/ITO multilayered electrode and porous WO3 chromic layer

ABSTRACT High-performance electrochromic devices (ECDs) were fabricated by introducing indium tin oxide/silver/indium tin oxide (IAI) multilayers as a transparent conducting oxide (TCO) electrode in conjunction with microstructural tailoring of tungsten trioxide (WO3) as an electrochromic (EC) layer. A 15-nm-thick Ag film with a preferential orientation of (111) and smooth morphology resulted in a high optical transmittance of 90.9% and sheet resistance 16.7 Ω/sq, making it a suitable TCO. In addition, the open network WO3 film used as an EC layer was deposited by carefully controlling the chamber pressure during the sputtering process. A high chamber pressure (2.0 Pa) resulted in a WO3 film with the lowest mass density (5.72 g/cm3) and a rough morphology. Half-cell ECDs consisting of a TCO bottom electrode (BE)/WO3 stack immersed in LiClO4 electrolyte were fabricated to verify the impact of using the IAI multi-layer stack and tailored WO3 film as a TCO and EC layer, respectively. Significant enhancement in terms of coloration efficiency was achieved for the ECDs by adopting the IAI TCO and open network WO3 EC layer compared to counterpart devices with an ITO TCO or dense WO3 EC film. This can be explained by the facile carrier extraction (injection) and transport of Li ions.


Introduction
Electrochromic devices (ECDs) are key technologies for energy-saving smart windows, next-generation information displays, switchable mirrors, and wearable electronics because they efficiently change their optical properties including transmittance, reflectance, and color under a low driving voltage range [1,2]. A typical ECD consists of an electrochromic (EC) layer, an electrolyte layer, and an ion storage (IS) layer sandwiched between two transparent conducting electrodes (TCOs) [3]. The injection/extraction of charge into an electrochromic layer under an applied electric field between the anode and cathode modulates its band gap, which changes the optical colligative properties. The performance of the ECD is strongly dependent on the structural and electrical properties of the transparent electrode and EC layer itself. First, the TCO film significantly affects the transmittance of the entire ECD, which provided a charge carrier to the electrochromic layer from an external circuit. Indiumtin oxide (ITO) is a representative TCO material used for commercial optoelectronic devices. It has a relatively low resistance ( < 10 −3 cm) and high transmittance to CONTACT  visible light ( > 80%) due to its wide bandgap ( ∼ 4.8 eV) [4]. However, a rather thick ITO is required to realize an acceptable low sheet resistance due to its larger electrical resistivity (by approximately two orders of magnitudes) compared to metals (2.5 × 10 −6 cm) such as Al and Cu. This increases the material cost and reduces the transmittance of visible light due to the interference effect [5,6]. This trade off problem between the electrical conductivity and optical transmittance of the ITO film can be mitigated using a multi-layer electrode scheme where a thin metal layer is inserted in the ITO electrode stacks [7][8][9]. The ITO/metal/ITO stack has a higher electrical conductivity compared to that of the single ITO film under an identical optical transmittance condition where the reflection from the metal layer in the visible range can be effectively suppressed for the ITO/metal/ITO stack. For this reason, various TCO/metal/TCO multi-layer structures have been investigated including ITO/metal (Au, Ag, Cu or Mo)/ITO [10,11], ZnO/Cu/ZnO [12], IZO/Au/IZO [13], ZnO/Ag/ZnO [14], AZO/Ag/AZO [15] and ZTO/Ag /ZTO [16]. An ITO/metal/ITO stack has been extensively examined in organic light emit-ting diodes [7,17,18] and solar cells [19][20][21][22]. However, case studies of ITO/metal/ITO stacks in ECD as a transparent conducting oxide (TCO) electrode are surprisingly limited. Second, the EC is a critical component governing the ECD performance through reversible insertion (or extraction) of external protons or ions into (or from) the host lattice. Among various EC materials, tungsten trioxide (WO 3 ) has attracted much attention due to its reversible color conversion capability from optical transparency to a deep blue color as well as its superior electrochemical stability compared to the counterparts of organic EC materials [23,24]. However, the low coloration efficiency (CE) ( < 30.0 cm 2 C −1 ) and slow switching speed ( > 10.0 s) of WO 3 films remain major obstacles their use as an EC layer in practical applications [25,26]. Thus, various approaches to enhance the EC performance have been proposed including tailoring the film structure, electronic structure, and crystalline structure [27][28][29]. In particular, the rather porous nano/microstructure of WO 3 films with a high specific surface area improved the modulation efficiencies of transmittance and CE values by increasing electrostatic contact with the electrolyte compared to dense WO 3 films [27,28]. For example, Yang et al. re-ported that a porous WO 3 film prepared by a template-assisted sol-gel method exhibited a high CE (50.1 cm 2 C −1 ) and fast switching speed (5.1 s for coloration rate and 8.7 s for bleaching rate) [30]. In contrast, a magnetron sputtering method is known to easily produce the amorphous WO 3 film with few defects and high coloration efficiency. Although many EC studies have been reported, research on how the ITO/Ag/ITO multi-layer structure impacts electrochromic performance with the porous WO 3 film as an EC layer remains lacking. In this study, Ag thickness-dependent structural, electrical, and optical properties of the ITO/Ag/ITO (IAI) multi-layer structures were examined. The effect of the WO 3 film's microstructure on the electrical and optical properties of the EC devices with the optimized IAI multi-layer structure was also investigated, for which the porous structure of WO 3 film was tailored by modifying the chamber pressure during sputtering. The electrochemical and optical properties were characterized by cyclic voltammetry, chronoamperometry and spectroscopic measurements. The light modulation of the WO 3 electrochromic layer was improved substantially by using the IAI multi-layer electrode and porous structure of WO 3 EC simultaneously.

Fabrication of ITO/Ag/ITO electrode stack
Transparent conductive IAI electrodes with varying Ag thickness were prepared on glass substrates by DC

Preparation of inorganic nano-porous tungsten oxide electrochromic layer
WO 3 thin films were deposited on IAI multilayers by reactive DC magnetron sputtering using a W metal target (99.99%) and oxygen plasma under an Ar atmosphere at room temperature. The base pressure of the sputtering chamber was ∼ 4 × 10 −4 Pa. The W target was presputtered for 10 min prior to WO 3 deposition to eliminate surface contamination such as WO x . The input gas mixture of Ar and O 2 with a 9:1 ratio was provided in the sputtering chamber where the chamber pressure varied from 1.0 to 2.0 Pa to control the WO 3 film's microstructure. The DC power of the W target was fixed to 80 W, and the sample stage was continuously rotated to ensure film uniformity during reactive sputtering. More detailed deposition parameters for each layer of ECD are listed in Table 1.

Characterization and analysis methods
Spectroscopic ellipsometry (SE, Elli-SE, Ellipso Technology, South Korea) was used to measure the film thickness. The structural properties of different films were characterized by X-ray diffraction (GIXRD, SmartLab, Rigaku, Japan) equipped with a Cu Kα source (40 KV, 30 mA). The morphological properties of various thin films were characterized by atomic force microscopy (AFM, Multimode 8, Brucker). The cross-sectional images of the ECD cell were obtained by field-emission scanning electron microscopy (FE-SEM, Verios G4 UC, FEI Co.). The densities of the films were analyzed by highresolution X-ray reflectivity (XRR, SmartLab, Rigaku), where the XRR raw data were fitted using GlobalFit analysis software. Resistivity and carrier concentration of the transparent ITO and ITO/Ag/ITO films were characterized by measuring Hall effect using the van der Pauw technique (HL5500, BioRad). Optical transmission spectra were acquired with a UV/visible spectrometer (Thermo scientific, Evolution 220) in the wavelength range of 300-800 nm at room temperature.
The schematic of the fabricated ECD system is shown in Figure S1 in Supporting Information (SI). The electrochemical properties of ECDs were investigated by cyclic voltammetry (CV) and chronoamperometry (CA) techniques using an electrochemical workstation (BioLogic, SP-50). The cell size of the ECD was 2 cm × 4 cm. The schematic of the three-electrode system used for the CV and CA measurements is presented in Figure S2 (SI), and the system was connected to a potentiostat and a computer. The EC testing of LixWO 3 (WO 3 :Li) samples in 1.0 M liquid-electrolyte of LiClO 4 /PC solution was performed using a three-electrode cell, comprising a working electrode (WO 3 electrode film on ITO/glass), a counter-electrode (Pt wire) and a reference electrode (Ag/AgCl). The potential voltage windows for CV and CA measurements were −1.0 V (tinted) and +1.0 V (bleached). In CA cycling measurements, the duration was 30 s for both the staining and bleaching phases. The scan rate for CV measurements was 50 mV/s. Light transmission spectra (300 nm-900 nm) of the ECDs in the bleached and pigmented states were acquired using a UV-Vis spectrophotometer. To ensure that the ECDs were sufficiently stained and bleached, the transmittance spectra were measured after applying a voltage of ±1.0 V to the ECDs for 30 s each for staining and bleaching. Figure 1(a-d) depicts the evolution of SEM surface morphologies of Ag films on the ITO/Si substrate as a function of Ag deposition time. When the Ag deposition time (t dep ) is short ( ∼ 13 s), the deposited film had discontinuous, isolated island morphologies (Figure 1(a)). This suggests poor wetting between metallic Ag and ITO substances [31]. The Ag film becomes continuous and completely covers the underlying ITO film with increasing t dep ( ≥ 15 s). Thus, the smooth morphologies of the Ag films with thicknesses of 15 and 17 nm were obtained, corresponding to t dep of 15 and 17 s, respectively. However, the prolonged Ag deposition of 19 s caused surface roughening in the resulting 19 nmthick Ag film as shown in Figure 1(d). This thicknessinduced morphology roughening phenomena can be attributed to the buildup and release of residual internal stress involving the formation of protrusion [32]. These features were reflected in the roughness variations of deposited Ag films with different thicknesses as shown in Figure 1(e-h). The Ag film at t dep = 13 s suffered from huge root-means-square (RMS) values of 3.7 nm due to the island-like discontinuous morphologies. Smaller RMS values of 0.47 and 0.48 nm were obtained for the Ag films at t dep = 15 and 17 s by virtue of the formation of a continuous smooth film. Conversely, the Ag film at the t dep = 19 s had an increased RMS value of 0.76 nm because of the roughening morphology (Figure 1(h)). Figure 2 shows the GI-XRD patterns of the control ITO and IAI multi-layer films with Ag films with varying thickness. The control ITO film had discernible diffraction peaks at 30.5°and 51°, which were assignable to bixbyite (222) and (400) reflections. In contrast, the IAI multi-layer films had additional diffraction peaks at 38.1°, which correspond to (111) of the face-centeredcubic Ag phase. It is noted that the (111) reflection of Ag phase in IAI multi-layer samples gradually increased with increasing Ag thickness, whereas the other diffraction peaks had rather thickness-independent intensities. The development of (111) preferential orientation can be explained by the following thermodynamic reasoning. The (111) face has the smallest interface energy in the face-centered-cubic Ag crystal, which means that (111) faces are dominant with increasing t dep to minimize the total Gibbs free energy of the Ag film. Figure 3(a) shows the variations in resistivity (ρ) and carrier concentration (n e ) of the IAI multi-layer structure as a function of t dep . The control ITO film (80 nm) had the largest ρ value of 1.76 × 10 −3 Ωcm. The IAI sample with an Ag interlayer at t dep = 13 s exhibited a slightly reduced ρ value of 1.58 × 10 −3 Ωcm. The modest improvement should be attributed to the discontinuous, island microstructure of the Ag layer. Significant improvement (1.43-1.58 × 10 −4 Ωcm) in terms of ρ value was observed for the IAI samples with Ag films at t dep = 15 and 17 s. This was attributed to a change from the discontinuous island-like structure to the continuous formation of Ag film [33]. The dynamic change in ρ value of IAI multi-layer films comes from the n e variation. It increased from 1.1 × 10 20 cm −3 (control ITO sample) to 5.2 × 10 21 cm −3 for the IAI sample with ITO film at t dep = 19 s. The optimal transmittance of the IAI multilayers depending on the t dep was examined as shown in Figure 3(b). The control ITO sample had an average transmittance near 90% for the wavelength range from 300 to 800 nm. A reduced transmittance of 64% was obtained for the IAI sample with an Ag interlayer at t dep = 13 s due to the light scattering by the isolated island-shaped Ag interlayer. In contrast, the IAI sample with an Ag interlayer at t dep = 15 s exhibited a higher transmittance ( ∼ 91%) for visible light, which is consistent with its structural properties as mentioned earlier. This high transmittance suggests that the Ag interlayer at t dep = 15 s intercalated between two ITO layers has an anti-reflection effect with the upper and lower ITO layers at the given specific thickness [34]. The IAI sample with the thickest Ag film at t dep = 19 s, however, showed a drop in the transmittance ( ∼ 67%) due to its strong absorption of visible light. The figure of merit (FOM) of the IAI multi-layer structures, which can be a good indicator of the optimal capability of high transmittance and low resistivity, was calculated using the Haacke equation [35] as follows:

Effect of ITO/Ag/ITO electrode stack on the performance of ECDs
Here, TC is the FOM, T is the optical transmittance of the IAI multi-layer structure at a wavelength of 550 nm, and R s is the sheet resistance. The maximum value of TC was 23.1 × 10 −3 Ω −1 for the IAI multi-layer with an Ag film at t dep = 15 s with an optical transmittance of 90.9% and sheet resistance of 16.7 Ω/cm 2 as shown in Figure 3(c) and Table 2. We anticipated that the IAI multiplayer with an Ag film at the t dep = 15 s (hereafter referred to as IAI-15s) would exhibit the best performance for ECD applications, which is why we chose the IAI-15 stack as a benchmarking electrode to compare with the control ITO electrode.
The feasibility of using IAI multi-layer stack as a transparent conductive electrode was tested by fabricating an ECD cell with the WO 3 layer as an EC layer. Figure 4   is referred to as a coloring process [see the details related to the experimental setup ( Figure S2) in SI]. The slope between J and V during the coloring process significantly increased for the ECD with the IAI-15 BE compared to that for the device with the ITO BE, which can be explained by the ∼ 9-fold smaller resistance of the IAI-15 electrode ( ∼ 16 /sq). When the BE voltage increased from −1 V to 1 V as a function of time, the reversal flow of Li ion occurs through the given cell, leading to a positive incremental change in J with respect to the applied V ( J/ V). This de-intercalation corresponds to the bleaching process, which is indicated in Figure 4 The detailed electro-chemical model responsible for the coloring and bleaching process will be addressed later. The dolphin-shaped area, which is the path of the coloring and bleaching process, is known to be a good indicator of the EC performance. Obviously, the dolphinshaped area was greatly enhanced for the ECD with the IAI-15 BE (Figure 4 (a) and (b)). Figure 4( Figure 4(d) showed the kinetic behavior for both ECDs during the bleaching and coloring states by providing a time-dependent rectangular voltage pulse (amplitude ± 1 V, 1 cycle duration = 60 s, also see Figure S3, SI). Response time is defined as the time needed to reach 90% of the transmittance modulation T of the given ECD at a fixed wavelength of 550 nm. The coloration response times for the ECDs with the control ITO and IAI BE were 20 and 16 s, respectively (Figure 4(d)). The falling time values of J with increasing time for the ECDs with control ITO and IAI BE during the bleaching process were approximately 20 and 7 s, respectively. The rapid bleaching response of the device with the IAI BE is consistent with its efficient CV characteristics. The superior device performance including CV, dynamic transmittance modulation and CA for the device with the IAI BE stems from the higher value of TC (23.1 × 10 −3 Ω −1 ) for the IAI multilayer. Figure 5 shows the SEM surface morphology of the WO 3 films prepared at different sputtering pressures (P chamber ). The WO 3 film at P chamber = 1.0 Pa had the smoothest surface morphology as shown in Figure 5(a). Its cross-sectional SEM image showed a featureless and dense packing structure ( Figure 5(b)). As the P chamber value increased from 1.0 Pa to 2.0 Pa during WO 3 sputtering, the surface morphologies became rougher, and their cross-section SEM images contained some cracks or holes (Figure 5(c) and (f)). It should be noted that the WO 3 films were in an amorphous phase state irrespective of the P chamber values from the XRD analysis (data not shown). Thus, the P chamber -dependent morphological and structural evolution cannot be explained by the amorphous-to-polycrystalline transition. To obtain insight into the P chamber -dependent morphological variations, the mass density of WO 3 films was calculated using XRR analysis. Figure 6 shows the XRR data for the WO 3 films at different P chamber values on SiO 2 /Si substrates. Critical  angles for the total reflection for the given films are known to be proportional to their electronic density and, consequently, their mass density. The WO 3 film at P chamber = 1.0 Pa had the highest mass density of 8.38 g/cm 3 . The mass density of the prepared WO 3 films decreased as P chamber condition increased. The lowest mass density of 5.72 g/cm 3 was obtained for the WO 3 film at P chamber = 2.0 Pa, which is consistent with the SEM analysis ( Figure 5). This result can be explained by the concept of mean free path (λ) of ad-atoms during reactive sputtering. At the low P chamber condition, ad-atom species such as WO, WO 2 , and WO 3 are likely to be less scattered by the Ar atom or radical in the plasma chamber, leading to the enhanced λ and kinetic energy of ad-atoms near the growing surface. Thus, the energetic ad-atoms can migrate easily onto the growing surface and form a dense WO 3 film. This is why the lowest P chamber condition favors the formation of a smooth and dense micro-structure. However, the collision event between ad-atoms and Ar species will increase with increasing P chamber , resulting in reduced λ and kinetic energy of ad-atoms. Consequently, a WO 3 film with a rather porous and less dense random network is likely to form, resulting in rough morphology and smaller mass density in the SEM and XRR analysis [36]. In general, the cell performance of ECDs with the WO 3 EC layer is affected by various microstructural factors, such as porosity [37], surface roughness [38,39], specific surface area [40], electrical conductivity [41], and crystallinity [42]. Obviously, the kinetics of the ion diffusion, insertion/extraction, and charge transfer should impact the coloring/bleaching process. In this regard, novel approaches including increasing the contact area between WO 3 film and electrolyte, enhancing the electrical conductivity of films, and designing specific nanostructures have been investigated [37][38][39][40][41][42]. Here, the WO 3 film with rough morphology and porous open network structure was obtained by carefully controlling P chamber = 2.0 Pa, which can provide more reactive sites and shorter ionic diffusional path leading to better ion diffusion or insertion/extraction. Thus, it is anticipated that the WO 3 film with the smallest mass density and roughest morphology can facilitate a fast and efficient coloring process because the possible existence of nanoor micro-scale pores or path assists Li + diffusion [43]. Figure 7(a)-(c) show the W 4f XP spectra for the WO 3 films at different P chamber conditions of 1.0, 1.5, and 2.0 Pa, respectively. Two strong sub-peaks at the binding energy of 35.8 and 38.0 eV can be assigned to W 4f 7/2 and W 4f 5/2 of W 6+ , corresponding to the stoichiometric WO 3 . In contrast, two smaller sub-peaks at 34.1 and 36.3 eV indicate the existence of W 5+ , which comes from the oxygen-deficient WO 3−x . Interestingly, the fraction of stoichiometric WO 3 went up with increasing P chamber as shown in Figure 7(d). The role of V O in the oxygen-deficient WO 3−x can be assessed by the optical transparency. The WO 3 film at P chamber = 1.0 Pa appeared as dark blue to the naked eye whereas the WO 3 film at P chamber = 2.0 Pa had a high transparency to visible light (see Figure S4). The P chamber dependence on the non-stoichiometry during sputtering can be explained by the atomic mass disparity between W and O atoms [44]. The occupied O 2p and the vacant W 5d 0 in WO 3 constitute the valence band maximum and conduction band minimum, respectively, which is responsible for the photochromic transition [45]. The stoichiometric WO 3 is a colorless transparent insulator. The dark blue color of the WO 3 film at P chamber = 1.0 Pa is attributed to the pre-existence of considerable W 5+ ionic states. The coloration efficiency can be degraded further because the V O in non-stoichiometric WO 3 hinder the chemical conversion from W +5 to W 6+ upon the Li removal. Figure 8(a)-(c) shows the CV characteristics of ECDs with the WO 3 EC layer prepared under different P chamber conditions. The dolphin-shaped area was smallest for the ECD with the WO 3 EC layer at P chamber = 1.0 Pa (Figure 8(a)). The largest dolphinshaped area was observed for the device with a WO 3 EC layer at P chamber = 2.0 Pa. The influence of P chamber on the ECD cell performance underscores the importance of microstructure in the WO 3 EC layer. The coloration and bleaching process in the WO 3 EC layer can be described using the following chemical equation:

Effect of microstructure of WO 3 film as EC layer on the performance of ECDs
Here, Li + is an ion in the lithium perchlorate electrolyte. The injection (extraction) of electrons and Li + ions from the electrolyte causes coloration (bleaching) processes in the WO 3 film [46]. When a negative voltage is applied to the BE, the J value increases negatively with increasing cathodic potential, corresponding to the co-injection of e − and Li + into WO 3 to form tungsten bronzes. Anodic peaks occurring during the positivebias sweep indicate the extraction of e − and Li + from Li x WO 3 to the electrolyte. Reversible switching between the colored and bleached states requires Li + charging and discharging through the EC layer. Obviously, the better coloration capability of the ECD with the WO 3 EC layer at P chamber = 2.0 Pa should be attributed to the porous and less dense random network nature of the WO 3 film (XRR data) because the porous and open structure  provides an efficient percolation diffusional pathway for the charge carriers such as e − and Li + . Figure 8(d)-(f) show the optical transmittance spectra of ECDs with different WO 3 EC layers in the colored and bleached states under sweep voltages of ±1.0 V after 10-time cycle loading. As expected, the optical modulation range of ECDs was improved using the WO 3 EC layer prepared at higher P chamber conditions. The optical modulation range for the given ECD cell increased from 35.6% (WO 3 EC layer at the P chamber = 1.0 Pa) to 86.2% for the WO 3 EC layer at P chamber = 2.0 Pa. Figure 9(a) shows the relationship between optical density and wavelength. Coloration efficiency (CE,η) is defined as the ratio of change in optical density ( OD λ ) to the corresponding injected/ejected charge density (Q in ) per unit area (A), which is described as follows: Here, T bλ and T cλ stand for the transmittance of bleached and colored samples, respectively. Using these formulas, the CE,η values for the ECDs with a WO 3 EC layer at P chamber = 1.0, 1.5 and 2.0 Pa were calculated at 35.7, 37.6, and 67.1 cm 2 C −1 , respectively ( Figure 9(b)). The ECD cell with the WO 3 EC layer at the P chamber = 2.0 Pa exhibits the best CE,η value (67.1 cm 2 C −1 ). Considering that CE,η values of 33.8 and 28.0 cm 2 C −1 were reported in the WO 3 EC layer prepared by RF magnetron sputtering and spray pyrolysis, respectively, our approach involving the IAI stack as a TCO and porous WO 3 as an EC layer using sputtering technique shows considerable promise. The switching time from the coloration to bleaching (or vice-versa) is also an important metric in the practical implementation of ECD. Figure 8(c) shows J versus time during the coloration and bleaching process using the CA method. The ECD with the WO 3 EC layer at P chamber = 1.0 Pa suffered from a slow response time ( ∼ 20 s) during the bleaching process due to the sluggish transport of Li + and e − through the dense WO 3 film. The improvement in terms of response time ( ∼ 5.7 s) during the bleaching process was achieved for the device with a WO 3 EC layer at the P chamber = 2.0 Pa, which is attributed to the lowest mass density WO 3 film (Figure 9(c)).
Finally, it is noted that the performances of ECDs in this work are quite competitive compared to other studies regarding ECDs with WO 3 EC materials (Table 3). Though nanoparticle or nanowires-based approaches exhibited the highest CE (78.6 ∼ 132 cm 2 /C) owing to the nanoscale tailoring structure, the rather slower response time and complicated fabrication routes need to be resolved [25,47,50,51]. The good T of 86.2% at the wavelength of 600 nm, relatively faster response time (T c /T b = 16/5.7 s) as well as the reasonable CE,η of 67.1 cm 2 /C for the given ECDs were achieved by tailoring the microstructure of the sputtered WO 3 EC layer in conjunction with the TCO stack of IAI, which has mass production compatibility.

Conclusion
The feasibility of an IAI multi-layer stack was examined as a TCO for high-performance ECDs. The thicknessdependence of the structural and electrical properties of Ag films on the inserted ITO/SiO 2 /Si substrate was examined. We found that a 15 nm-thick Ag film characterized by the cubic Ag phase with a preferential orientation of (111) and smooth morphology was optimal. The IAI stack with the 15 nm-thick Ag film exhibited the highest FOM of 22.4 (R S of 16.6 Ω/cm 2 and optical transmittance of 90.6%), which is suitable as TCO electrode for high-performance ECDs. Furthermore, the WO 3 films with different mass densities were sputtered by controlling the P chamber value during film preparation. As the P chamber value increased, the prepared WO 3 film showed decreasing mass density. Specifically, it decreased from 8.38 g/cm 3 (P chamber = 1.0 Pa) to 5.72 g/cm 3 for the WO 3 film at P chamber = 2.0 Pa. The responsible growth model was discussed using the λ concept and kinetic energy of ad-atoms near the growing surface. The half-cell ECD with the IAI BE and porous WO 3 EC layer exhibited the best coloration efficiency and improved response time, which can be attributed to (i) high FOM of the IAI stack and (ii) enhanced transport capability of Li ions (e − ) through the open structural network of WO 3 . We concluded that the implementation of IAI stack and porous WO 3 EC is an effective approach to producing highefficiency ECDs.

Disclosure statement
No potential conflict of interest was reported by the author(s). Electronic Engineering at Hanyang University, Seoul, Republic of Korea as a professor. He has more than 185 authored (or coauthored) international journal papers and 112 international patents under his name.