Efficiency-Improved UWB Transparent Antennas Using ITO/Ag/ITO Multilayer Electrode Films

In this paper, a transparent antenna made with a new structure of ITO/Ag/ITO is proposed. Overcoming the physical limitations of transparency and conductivity is an important problem with transparent materials. By studying and comparing previously reported transparent materials, a transparent electrode with thin film Ag inserted between two layers of ITO instead of a single layer is selected for a highly efficient transparent antenna. This electrode has low sheet resistance ( $3.1~\Omega $ /sq) relative to its high transparency (88 % at 550 nm), which is a factor that can increase the efficiency of the antenna. In general, it is difficult to measure sheet resistance (SR) using a 4-point DC probe for very thin films (thickness of transparent material is less than the skin-depth). Therefore, a form of reverse engineering that can estimate DC sheet resistance using RF SR was presented and verified. As a result, it was possible to predict and design the performance of the transparent antenna with the new material structure. The selected transparent material is applied to design the wideband transparent antenna and the design process for wideband performance is covered in the paper. The proposed antenna with ITO/Ag/ITO was implemented for verification. The peak efficiency of the fabricated antenna was 66 %, and the measured bandwidth was 123 % (from 2.5 GHz to 10.6 GHz), which is the best performance than previously reported transparent antennas.


I. INTRODUCTION
Recently, transparent antennas for use in various applications such as electronics products, mobile phones, and vehicle communication are being studied [1]. In these applications, a wideband high efficiency antenna is essential for improved communication. Moreover, to use a transparent antenna, high transparency over 85 % is required [2]. The issue is the tradeoff between the conductivity and transparency in transparent electrodes. In other words, the better the optical transparency is, the lower the electrical conductivity becomes. Therefore, many studies have been conducted to overcome this.
Transparent antennas have been studied using a variety of transparent materials [3]- [31]. There are two approaches to design a transparent antenna. One is to use oxide-based transparent material. Of these, indium tin oxide (ITO) is the most common material for transparent electrodes [3]- [7]. This ITO The associate editor coordinating the review of this manuscript and approving it for publication was Davide Ramaccia . has transparency (81 -85 %) and high sheet resistance (SR, 7 -10 /sq), resulting in low antenna efficiency.
To overcome the high sheet resistance, indium tin oxide with Ag or Au thin films are applied to the transparent antenna [8]- [10]. In [8], the transparent coplanar waveguide (CPW)-fed monopole antenna has high transparency (85 %), and the peak antenna efficiency is 70 %. In addition, the fractional bandwidth is 155.6 %, which is suitable for a wideband transparent antenna. However, the shape of the antenna is visible because it uses gold nanolayer deposition. In [9], [10], AgITO is used for a transparent slot antenna. This material has low sheet resistance (0.9 /sq). Because of the low sheet resistance, the antenna has high peak efficiency (71 -80 %), but the transparency is too low (52.5 %). Moreoever, the slot antenna has limited fractional bandwidth, and the double layer for the feeding network makes the transparency lower.
Another approach is to employ metal for the transparent antenna. AgHT is a thin Ag films that has transparency (80 -85 %) and high sheet resistance (8 /sq) [13]- [19]. Moreover, the dual-band antennas in [14]- [17] are not suitable for a wideband application. In [18], a transparent CPW-fed slot antenna with high peak efficiency (80 %) is researched. However, the transparency is too low (80 %), and the fractional bandwidth is too narrow (40 %). Mesh metal and nanohole structures have also been studied for use in transparent antennas [20]- [24]. These materials have low sheet resistance and high transparency. However, the high optical haze resulting from scattering of the nanowires and mesh patterns is undesired in high-resolution displays [25]. Moreover, they are usually fabricated through precise patterning or complicated chemical synthesis processes which makes fabrication procedures expensive and complex [26], [27].
To overcome the problems with conventional transparent materials, multilayer electrode films are used for transparent antennas [28]- [32]. Structures like this are sealed with bottom and top oxide films, which act as suitable barriers against chemical corrosion of the metals [33]. In [28], IZTO (40 nm)/Ag (10 nm) /IZTO (40 nm) is used for a transparent antenna. It has high transparency (86 %), but the sheet resistance (7 /sq) is too high. The transparent antenna has a narrow fractional bandwidth (40 %) and is not suitable for a wideband application. In [29]- [31] IZTO (45 nm)/Ag (10 nm) /IZTO (45 nm) is employed. It has low sheet resistance of 2.52 /sq [30], [31]. In [32], a monopoletype antenna operates over a wide fractional bandwidth (109 %), but the antenna efficiency (42.1 %), and transparency (80 %) are both low. Moreover, ITO (85 nm)/Ag (13 nm)/ITO (85 nm) has low transparency (74 %) and high sheet resistance (5 %). To design a transparent antenna that has over 85 % transparency, high antenna efficiency, and a broad operating bandwidth, a new material, and proper antenna structure had to be researched.
In this paper, a new multilayered transparent antenna using the best transparent material is proposed to improve electrical performance while maintaining a high transparency state. We selected the best transparent material for a highly efficient transparent, wideband antenna. To implement it, we applied this material to create a CPW-fed diamond-shaped monopole antenna. Moreover, to design the transparent antenna using the new transparent material, reverse engineering was used to estimate the electrical property of the selected material, which result is presented. To design a transparent antenna with the new material, it is necessary to know the RF sheet resistance  of the transparent material. Traditionally, the DC sheet resistance is employed in place of the RF sheet resistance because, for thin films (thickness of transparent material less than skindepth), RF and DC SR are the same [34]. However, measuring the sheet resistance requires a 4-point DC probe that costs more and takes more time. Moreover, damage occurs to the thin film sample due to the pressure from physical contact [35], [36]. Therefore, in Section II, we introduce for the first time, a method for predicting DC SR using RF SR by reverse engineering. In Section III, the design of the proposed transparent antenna made using the selected transparent material is covered. In Section IV, the implementation of the proposed antenna structure is verified. Section V presents conclusions about the proposed antenna.

II. TRANSPARENT MATERIAL DESIGN A. TRANSPARENCY PROPERTY
To design a transparent antenna, first, an appropriate transparent material is needed. To select the transparent material, we investigated a transparent material with transparency > 85 % [37]- [44]. As shown in Fig. 1, an ITO (48 nm)/Ag  (17 nm)/ITO (42 nm) material has the lowest sheet resistance (3 /sq) and its transparency is 88 %. This material satisfies the need for > 85 % transparency and low sheet resistance can make the antenna efficiency high. To compare with other materials, three transparent electrodes were designed, as shown in Fig. 3. The electrodes were implemented using RF magnetron sputtering [44]. There were three configurations: Single-layer ITO (150 nm), ''Multilayer A'' ITO/Ag/ITO (48/12.5/42 nm), and ''Multilayer B'' ITO/Ag/ITO (48/17.5/42 nm). Corning glass with a relative permittivity of 5.27, loss tangent of 0.001, and thickness of 0.7 mm was used for the substrate. From an optical point of view, the wavelengths 400-750 nm encompass the visible range. In particular, light at 550 nm is that to which the human eye is most sensitive [45]. As shown in Fig. 3(a), the transparency was measured using a PerkinElmer' Lambda 950. The transparency to visible light of the new transparent electrode is shown in Fig. 3(b). In Fig. 3(c), the ITO singlelayer (SL) and multilayer A (MA) were higher in terms of the 400-750 nm average transparency. At 550 nm, the single-layer, MA, and multilayer B (MB) showed similar transparency (1 % difference). The optical characteristics at 400-750 nm were analyzed. The SR characteristics were analyzed based on the RF property as an electrical analysis method.

B. REVERSE ENGINEERING TO DETERMINE THE ELECTRICAL PROPERTY
In general, the conductivity of a transparent electrodes cannot be known given the new structure of the material. A 4-point probe measurement analyzes the DC SR of the electrical characteristics. However, for very thin films, it is difficult to measure sheet resistance using a 4-point DC probe [35], [36]. Because the layer is so thin, the measurement error due to the DC probe contact is very large. If an electrode with thickness less than the skin-depth, is used in the operating frequency range, RF SR verification is possible using the following equation [34].
(1) VOLUME 9, 2021 Equation (1) is the formula of SR (R DC ) with conductivity (σ Films ). The thickness (t Films ) here can be seen as the effective thickness in terms of DC. However, in RF, skin-depth (δ Films ) occurs depending on the frequency (ω). Thus, equation (1) should be represented, including skin-depth, as shown in (2). In equations (1) and (2), the conductivity (σ Films ) is lower because the carrier concentration and hall mobility are low due to the thin metal and not bulky [46]. In the case of the thin film (1xx nm), the thickness of the transparent material is less than the skin-depth; so the RF and DC SR are equal in the UWB band as seen in equation (3) [34]. Therefore, in a thin-film material (1xx nm), the DC SR can be obtained by reverse engineering.
First, for when the conductivity of a transparent material cannot be known, the transmission lines of CPW using SL, MA, and MB were fabricated to analyze the RF SR characteristics for reverse engineering. The insertion loss per unit length for each transparent material was measured. Fig. 4(a) shows the fabricated CPW of the transmission line with the transparent materials (i.e., 30, 40, and 50 mm). Fig. 4(b)-(d) shows the insertion loss for the SL, MA, and MB. The loss factor for each frequency improves in the order SL, MA, and MB. As shown in Fig. 4(b)-(d), based on the insertion loss, it can be expressed as a representative value of the average insertion loss per unit length (3-6 GHz), as summarized in Table 1.   The average delta insertion loss (ADIL) per unit length is 0.2 dB/mm for SL, 0.13 dB/mm for MA, and 0.09 dB/mm for MB. Based on the measured ADIL, a 3-D electromagnetic (EM) simulator was set up to estimate RF SR as being in the same environment as the CPW transmission line manufactured. Then, the same ADIL could be extracted by the sweep of the SR parameter of the transparent material. Fig. 5(a) shows that the RF SR estimated with the measured ADIL is 8 /sq as a result of variation of the SR to 5.5-10 /sq for the SL. Similarly, Fig. 5(b) shows that the estimated SRs of MA and MB are 4 /sq and 3 /sq, respectively. As such, a multilayer may have higher conductivity than a single layer, which is illustrated by equalizing materials that may be configured in parallel, as shown in Fig. 6, where R SL total = R ITO and R ML total = R ITO R Ag R ITO . In this case, since R ML total is constructed in a parallel relationship, conductivity can be improved. The measured DC SRs of SL, MA, and MB were 8.4, 4.3, and 3.1 /sq, respectively. Thus, the RF SR can be equal to DC SR and the factor for predicting the efficiency in antenna design, as described in Section III. By reverse engineering, DC SR was estimated from the RF SR. Moreover, the selected MB had the lowest sheet resistance among the three transparent materials.

III. TRANSPARENT ANTENNA DESIGN
To design a wideband, highly efficient transparent antenna using the selected transparent material, a diamond-shaped monopole antenna was designed, as shown in Fig. 7. The proposed transparent antenna was designed using SL, MA, and MB. For high transparency, the transparent antenna had to be designed as a single layer antenna. Moreover, for wideband application, the printed monopole antenna was suitable. Therefore, a CPW-fed monopole antenna was designed. Fig. 8 shows the reflection coefficient and efficiency according to the C L and G P for the essential parameters in terms of antenna design when the SR is 3 /sq (MB). Fig. 8(a) shows the reflection coefficient for the process of becoming antenna C by adding the cutting area (CA) in the rhombus shape. When the CA length (C L ) is increased from 0 to 8 mm, it can be seen that the 10 dB IBW expands to the 9 GHz band. The upper C L is 0.7 mm less than the C L on both sides, which is a factor that can increase IBW by asymmetric current induction. The IBW is largest when the C L is 6 mm. As shown in Fig. 8(b), the efficiency increases by 4 % in the 7.1 GHz band when the C L is 6 mm rather than 0 mm, so the C L is most optimal at 6 mm. Fig. 8(c) and (d) show the results of the reflection coefficient and efficiency according to the gap change between the feeder line and ground. As shown in Fig. 8(c), the antenna is improved in terms of the IBW and the impedance matching when G P is 0.5 mm, 1.0 mm, and 1.5 mm. However, in Fig. 8(d), the efficiency is optimal when G P is 1.5 mm. Fig. 9 illustrates variations in the reflection coefficient and efficiency according to the SR of the antennas. The thickness of the ITO and Ag layers could be changed in the process of implementation. In this case, the sheet resistance becomes different. In Fig. 9 (a), the bandwidth in the lower frequency band of the reflection coefficient becomes narrow as the sheet resistance decreases. However, the high frequency performance is maintained as the sheet resistance is varied. Moreover, as the sheet resistance decreases, the antenna efficiency increases. The difference in the peak efficiency between SL and MB is 19 %. Although the peak efficiency is different, the proposed antenna has similar tendencies regarding the reflection coefficient and efficiency according to change in the SR. This is an advantage in terms of reducing a process error when depositing a transparent electrode.

IV. IMPLEMENTATION AND MEASUREMENT
Diamond-shaped antennas with SL, MA, and MB were implemented, as shown in Fig.8(a). The transparent antennas clearly show the logo on the back of the antenna, as shown in Fig. 8(b). The shape of the antenna is difficult to distinguish with the human eye. Fig. 10 shows a scanning electron microscope image made with backscattered electrons to capture the nanoscale tomography of the three transparent materials. Fig. 10(a) shows the layer composition for the singlelayer ITO. Measurements show that the ITO was deposited at a thickness of about 155 nm, showing a process error of about 3 %. As shown in Fig. 10(b), MA was deposited using RF sputtering as ITO/Ag/ITO films with 47.4/13.8/42 nm from the top. It can be seen that the layers are distinguishable. Moreover, Fig. 10(c) is a cross-sectional photograph of MB, in which an increase can be seen compared to the Ag layer of the MA (about 48.6/16.5/44). This could lead to improved electrical properties. As a result, although there is a process error, the multilayer has little optical deteriorations due to the nanoscale layer in the middle, while the sheet resistance is reduced compared to the single layer.
As shown in Fig. 11(a), the 3-D radiation patterns of the proposed transparent antennas were measured in a mobile chamber. A dual-polarized horn antenna was employed as a reference antenna (LB-780-SF). As a result of measurement, it was determined that the radiation patterns of the proposed transparent antenna simulated and measured at the XZ and YZ plane at each frequency (3.0, 4.5, and 6.0 GHz), each show good agreement. Fig. 12 shows the measured reflection coefficient and efficiency for newly fabricated antennas: diamond shaped monopole antennas with an SL, MA, and MB. The chamber able to measure the antenna efficiency is for mobile use, which means that the frequency band can be measured is only available up to 6 GHz. Therefore, only 3 -6 GHz was measured considering the sweep point. The measured impedance (whole band) and efficiency (3-6 GHz) of all antennas are in good agreement with simulated results. Therefore, in the band where the efficiency was not measured, the efficiency could be predicted to correspond to the simulated results. The implemented antennas have wider impedance bandwidth than the simulated results do. The maximum efficiencies of the SL, MA, and MB antenna are 47, 62, and 67 %, respectively. Table 2 summarizes the performance of previously reported transparent antennas. In [9], an AgITO based slot antenna has wide impedance bandwidth (106 %) and high peak efficiency (70 %) due to low sheet resistance (0.9 /sq). However, the transparency is too low (52.5 %). In [12], the UWB antenna with wide bandwidth performance (128 %) studied has peak efficiency that is high enough (60 %). Unfortunately, the transparency is too low (72 %). In [28], [29], and [31], IZTO/Ag/IZTO is employed for transparent antennas. The transparent antenna in [28] was designed using IZTO 40 nm/Ag 10 nm/IZTO 40nm. The branch-type antenna has a narrow bandwidth. Moreover, in [29] and [31], a monopole antenna with IZTO 45 nm/Ag 10 nm/IZTO 45 nm was studied. This material has a low sheet resistance of 2.5 /sq. However, the proposed transparent antenna efficiency is higher than conventional antennas [29], [31]. Among the previously reported transparent antennas and new transparent antennas in this paper, the proposed transparent antenna using ITO 48 nm /Ag 17.5 nm /ITO 42 nm has high antenna efficiency, broad impedance bandwidth, and proper transparency.

V. CONCLUSION
In this paper, a new diamond-shaped antenna with ITO/Ag/ITO structure was proposed to overcome existing optical and electrical limitations. In the middle, rather than a single layer, Ag was thinly deposited at nanoscale to maximize conductivity performance without deterioration of the transparency. Moreover, for the new material, a reverse engineering method was proposed to overcome the limits of a method that made it difficult to measure thin films using the sheet resistance analysis method. A transparent antenna designed with DC SR estimated using reverse engineering was verified. As a result, it was possible to predict the DC SR of the thin film of the new material and design the antenna. Therefore, it has been verified that the transparent antenna using ITO/Ag/ITO is superior to other materials in terms of efficiency. In addition, the improved efficiency and widened IBW of the proposed antenna was verified by adding a cutting area in the conventional rhombus shape. The proposed antenna has high transparency, high efficiency, and broad wideband performance. His research interests include invisible antennas, millimeter-wave antennas, and wireless communication systems. VOLUME 9, 2021