Conformal and Transparent Al₂O₃ Passivation Coating via Atomic Layer Deposition for High Aspect Ratio Ag Network Electrodes

Abstract

We demonstrated conformal Al₂O₃ passivation via atomic layer deposition (ALD) of a flexible Ag network electrode possessing a high aspect ratio. The Ag network electrode passivated by the ALD-grown Al₂O₃ film demonstrated constant optical transmittance and mechanical flexibility relative to the bare Ag network electrode. Owing to the conformal deposition of the Al₂O₃ layer on the high aspect ratio Ag network electrode, the electrode exhibited more favorable stability than its bare Ag-network counterpart. To demonstrate the feasibility of Al₂O₃ passivation via ALD on a flexible Ag network, the performances of flexible and transparent thin-film heaters (TFHs) with both a bare Ag network and that passivated by ALD-grown Al₂O₃ were compared. The performance of Al₂O₃/Ag network-based TFHs was minimally altered even after harsh environmental tests at 85% relative humidity and a temperature of 85 °C, while the performance of bare electrode-based TFHs significantly deteriorated. The improved stability and reliability of the Al₂O₃/Ag network-based TFHs indicate that the ALD-grown Al₂O₃ film effectively prevents the introduction of moisture and impurities into the Ag network with a high aspect ratio. The improvement in the stability of the Ag network through Al₂O₃ passivation implies that the ALD-grown Al₂O₃ film represents a promising transparent and flexible thin film passivation material for high quality Ag network electrodes with high aspect ratios.

1. Introduction

Transparent and flexible electrodes (TFEs) are being actively researched for use in flexible displays [1,2], flexible photovoltaics [3,4], flexible electronics, smart windows [5], flexible sensors [6,7], and film heaters [8], and TFEs are a key component of high-performance flexible devices. Various TFE materials, such as indium tin oxide (ITO) [9,10], carbon-based electrodes [11,12,13], conductive polymers [14,15], metal nanowires [16], and metal networks [17], have been used as effective TFE layers for flexible devices owing to their unique advantages. Among these TFE materials, ITO films deposited on flexible substrates via magnetron sputtering have been predominantly employed as transparent electrodes in flexible optoelectronic devices, owing to their low resistivity, high transmittance, easy scalability, and mature thin-film process [18,19]. However, the scarcity of indium elements and use of vacuum-based systems can increase the production cost of flexible ITO films. In addition, the brittleness of the ITO films is a critical challenge in the use of ITO films as TFEs. As cost-effective alternatives, carbon-based, conducting polymer-based, and metal nanowire-based TFEs have been extensively studied, as they can be prepared through a simple and cost-effective printing process [18,19]. However, solution-processed TFEs obtained via spin coating, bar coating, slot-die coating, and spray coating continue to demonstrate critical issues, such as inferior performance, low scalability, and non-uniformity, relative to the sputtered ITO films [20,21]. Another alternative candidate for ITO is a self-assembled Ag network with high conductivity, high optical transmittance, and outstanding flexibility [17,22,23]. In our previous works, the Ag network was employed as an electrode in transparent thin film heaters due to its low resistivity, high transparency and mechanical flexibility. However, Ag network electrodes are easily degraded by oxygen and humidity owing to their high aspect ratio and large effective surface area. When the Ag network is exposed to a high-temperature and humid environment, H₂O or carbonyl sulfide diffuses into the network and degrades its electrical and optical properties [24,25]. Therefore, the effective passivation of Ag network electrodes is essential to ensure their stability and reliability. However, the high aspect ratio of the Ag network prevents the conformal deposition of a passivation layer, resulting in exposed regions of the Ag network. For these reasons, the effective passivation of the Ag network remains a critical issue. In our previous paper, we reported the fabrication of PTFE, InSnTiO, and ZnO thin films via sputtering and atomic layer deposition (ALD) [5,26,27]. In particular, ZnO deposited via ALD significantly improved the performance and lifetime of the heater based on the Ag network. Although sputtering and ALD techniques have been employed as thin-film passivation processes, ALD is more compatible with Ag network electrodes considering that the conformal deposition of a passivation layer on the surface of the Ag network is critical for achieving high-performance thin-film passivation. Therefore, ALD has been widely employed as a thin-film passivation method [28,29,30] as it affords high density and high transparency for the resulting films [31,32,33]. Despite the importance of ALD-based Al₂O₃ passivation, the conformal deposition of Al₂O₃ on self-assembled Ag network electrodes has not been reported thus far. High-quality thin-film passivation to protect the Ag network against ambient conditions should be developed to facilitate the widespread application of Ag network electrodes in flexible optoelectronic devices.

In this paper, we report the fabrication of a conformal and transparent Al₂O₃ thin film passivation film via ALD for reliable self-assembled Ag network electrodes. The stability of high aspect ratio Ag network electrodes is improved by the conformal coverage of ALD-grown Al₂O₃ films, without significant changes in their electrical and optical properties. To demonstrate the feasibility of Al₂O₃ thin-film passivation through ALD, we compared the optical and electrical properties of bare and Al₂O₃-covered Ag networks before and after subjecting them to an 85°C/85% relative humidity (RH) test. In addition, we compared the mechanical flexibility of the bare and ALD-grown Al₂O₃-covered Ag networks using a specially designed bending test system to characterize electrode flexibility. Finally, to demonstrate the potential application of the passivated Ag network electrodes, we fabricated flexible and transparent thin-film heaters (TFHs) with both bare and Al₂O₃-covered Ag networks to compare their performance and reliability before and after the 85 °C/85% RH test.

2. Materials and Methods

Embossed, self-assembled Ag networks were formed on flexible polyethylene terephthalate (PET) substrates using a bar-coating system (KP-3000VH, Kipae E&T Co., Hwaseong, Republic of Korea) to generate a network structure. The Ag nanoparticle solution was synthesized by mixing 0.2 g of resin (BYK-410) and 4 g of Ag nanoparticles in a solution of 30 g 1,2-dichloroethane and 15 g deionized water [17,22,23]. To increase the wettability of the PET substrate, a surface treatment was applied to decrease its surface energy using 3-aminopropyltriethoxysilane with 1% acetone solution. The Ag nanoparticle solution was then drop-cast onto the pre-treated PET substrate; subsequently, the solution was coated on the surface via a bar-coating system at a speed of 30 mm/s. Finally, to produce self-assembled Ag network films, the film was dried at 50 °C for 30 min [23]. Prior to Al₂O₃ deposition, the substrates were ultrasonically cleaned with isopropanol and washed with deionized water several times. Subsequently, 60 nm thick Al₂O₃ films were grown on the Ag network electrode with size of 5 cm × 5 cm using a commercial ALD system (CN1 Co., Ltd., Hwaseong, Republic of Korea). Trimethylaluminum (TMAl) and H₂O were used as the precursors for aluminum and oxygen, respectively. TMAl and H₂O were fed into the ALD chamber through separate inlet lines, while N₂ gas was purged. The feed times of TMAl and H₂O were 2.0 and 1.0 s, respectively. The injections of TMAl and H₂O were separated by purging (1.0 s) with N₂ at a flow rate of 800 sccm. The pressure in the reaction chamber was maintained at approximately 13 Pa. The deposition temperature was 195 °C, and the growth rate was 1.2 Å per one cycle (DEZn-N₂ purge-H₂O-N₂ purge) for the ALD process. Field emission-scanning electron microscopy (FE-SEM, JSM-7600F, JEOL, Tokyo, Japan) was used to obtain images of the surfaces’ network electrodes. The electrode optical properties were examined using an ultraviolet (UV)/visible light spectrometer (V-670, JASCO, Oklahoma City, OK, USA) in the range of 400 to 800 nm based on measurements at every 20 h during the temperature/humidity tests. Dynamic mechanical tests, such as bending, rolling, and twisting, were performed by measuring the change in resistance during each test using laboratory-designed test instruments. The bending test was performed via two methods: critical bending and repeated bending tests. The critical test was conducted by changing the bending radius; the test was repeated 10,000 times at a fixed bending radius of 5 mm. Rolling and twisting tests were conducted 10,000 times at a rolling radius of 10 mm and twisting angle of 35°. The network electrodes were also exposed to harsh environmental tests in a temperature and humidity chamber (TH3-ME, Lab companion, Daejeon, Republic of Korea) at a temperature of 85 °C and humidity of 85% for 300 h. To fabricate flexible TFHs using Ag network electrodes, a 100 nm-thick Ag film was deposited on the edges of the Ag network electrode using a thermal evaporation system. Copper tape was then attached to the Ag contact electrodes. Direct current (DC) power was applied to the flexible TFHs using a suitable DC power supply (OPS 3010, ODA Technologies, Santa Clara, CA, USA). Subsequently, changes in the temperature of the heater over time was measured using a thermocouple temperature measurement system for each applied voltage. The measurement was performed for 600 s, with and without an applied voltage for first 400 s and the remaining 200 s, respectively. Infrared (IR) images were captured using an IR camera (A35sc, FLIR, Wilsonville, OR, USA) while supplying DC power to the device.

3. Results and Discussion

A conformal Al₂O₃ passivation layer was directly deposited on a self-assembled Ag network electrode using ALD. Figure 1a illustrates the ALD procedure for depositing an Al₂O₃ passivation layer on the Ag network. Owing to the high film density and transparency, Al₂O₃ films have been widely employed as a thin-film passivation layer in flexible optoelectronic devices [34,35]. In addition, ALD can control thin film characteristics at the atomic level and deposit a conformal film on three-dimensional structures, such as self-assembled Ag networks [36,37,38].

Moreover, ALD-grown Al₂O₃ layers can be prepared at relatively low temperatures and are suitable for application as polymer substrates [39]. TMAl was used as a precursor to deposit aluminum atoms, while H₂O was used as a precursor to deposit oxygen atoms. In the TMAl feed phase, the TMAl precursor was injected into the deposition chamber and reacted with the surface hydroxyl groups (OH*) to form Al−OH* bonding. Al is then deposited in the form of Al−O−Al(CH₃)₂*, producing methane. This reaction can be described as follows [39] (in the reaction formula, the asterisk indicates the surface species):

Al−OH* + Al(CH₃)₃ → Al−O−Al(CH₃)₂* + CH₄

(1)

The generated CH₄ gas was removed during the N₂-purge phase. Afterward, the H₂O precursor was injected into the chamber during the H₂O feed phase and reacted with the surface-located methyl groups (CH₃*). H₂O and CH₃* then formed Al−OH*, with the evolution of CH₄. The chemical equation for this reaction is as follows:

Al−CH₃* + H₂O → Al−OH* + CH₄

(2)

The CH₄ gas produced in the H₂O feed phase was also removed during the N₂-purge phase. Reactions A and B were performed alternately and repeatedly to form an Al₂O₃ thin film controlled in atomic units. Schematics and SEM images of the Ag network electrode before and after Al₂O₃ deposition are shown in Figure 1b,c, respectively. The thickness of the Al₂O₃ passivation layer was 60 nm, which is very thin relative to the width of the Ag network electrode. No significant difference was observed in the width of the electrodes.

The effect of the ALD-grown Al₂O₃ film on the optical and electrical properties of the Ag network electrode was evaluated through selective ALD using a shadow metal mask. Figure 2a shows an optical microscopy (OM) image exhibiting a clear boundary between the Al₂O₃-covered and uncovered regions on the Ag network electrode. The boundary was generated by removing tape on the Ag network to distinguish between the area covered with a Al₂O₃ film and the area without. The right side of the OM image represents the surface of the bare Ag network, while the left side represents the surface of Al₂O₃ deposited on the Ag network. Although a very thin Al₂O₃ film with a thickness of 60 nm was deposited, the boundary of the Al₂O₃ layer could be clearly observed. The surface images of the two regions are very similar, except for a slight brightness contrast, while the Ag network structure under the Al₂O₃ film could clearly be observed. This implies that the ALD-grown Al₂O₃ film possesses a high conformal coverage and optical transparency compatible with the irregular Ag network. The boundary region was characterized using atomic force microscopy (AFM), as shown in Figure 2b. The surface of the ALD-grown Al₂O₃ film exhibited numerous small sub-grains. This indicates that the Al₂O₃ film was well-deposited on the PET film, and the shadow metal mask could be used to select the passivation area. As shown in Figure 2c, the optical transmittance of the Al₂O₃-coated Ag network is slightly lower than that of the bare Ag network because of the light absorption of the Al₂O₃ passivation layer. However, the high optical transmittance of the Al₂O₃-coated Ag network of over 90% is sufficient for fabricating flexible optoelectronic devices. The insulating properties of the ALD-grown Al₂O₃ films were evaluated through current–voltage (I-V) characterization using the Al metal contact, as shown in the inset of Figure 2d. The operating current of the bare Ag network was 83 mA at a bias of 1.0 V. This indicates that the bare Ag network exhibits highly conductive properties through the development of Ohmic contacts on the conductive Ag network. However, following the deposition of the Al₂O₃ film on the Ag network, the operating current of the Al₂O₃/Ag network was 513 nA. The electrically insulating properties of the ALD-grown Al₂O₃ passivation resulted in a significant increase in resistance for the Al₂O₃-coated Ag network electrode. From the I-V curve, we conclude that the ALD-grown Al₂O₃ film completely covered the Ag network.

To verify the effect of Al₂O₃ passivation on the mechanical flexibility of the Ag network electrode, mechanical bending tests were performed with both the electrodes. The change in resistance, represented by the following equation, was used as an indicator of the mechanical properties [6]. In the following formula, ΔR is resistance change; R is the resistance at the time of measurement, and R₀ is the initial resistance value.

$$∆R=\left(R-R_0\right)/R_0$$

(3)

Figure 3a shows the stress applied to the thin film during the inner/outer bending test. When the sample was bent, a compressive force was applied to the thin film. In contrast, the outer bending state provides tensile stress to the thin film, which renders the thin film fragile [6]. The photographs in Figure 3b illustrate the bending test procedure. The resistance of the electrode was measured in situ as a function of the bending radius. The results of the critical bending test of the bare Ag and Al₂O₃/Ag network electrodes are shown in Figure 3c,d respectively. The Al₂O₃/Ag network electrodes exhibited no gradation in resistivity over the entire range of bending radii; the outer critical bending radius was 2 mm, which is identical to that of the bare Ag network electrode. This indicates that passivation with a very thin Al₂O₃ layer does not influence the mechanical flexibility of the Ag network electrode. Figure 3e and f show the results of repeated bending tests of the bare Ag and Al₂O₃/Ag network electrodes, respectively. The bending test was repeated 10,000 times at a fixed bending radius of 5 mm. During the inner and outer bending tests, both electrodes exhibited very small change in resistance (2~5%). Regardless of the bending mode, both samples exhibited very small change in resistance [40].

Figure 4a,b demonstrate the rolling and twisting tests, respectively. The resistance of the electrode was measured at every rolling and twisting cycle. Figure 4c shows the results of repeated rolling tests of the bare Ag and Al₂O₃/Ag network electrodes performed 10,000 times at a rolling radius of 1 cm. Both samples exhibited no significant changes in resistance, indicating the outstanding flexibility of the Al₂O₃/Ag network electrode. A repeated twisting test was conducted 10,000 times at a twisting angle of 35°. The results of the twisting test also demonstrate no change in resistance in the bare Ag and Al₂O₃/Ag network electrodes. These dynamic mechanical test results demonstrated that the ALD-grown Al₂O₃ passivation layer did not significantly degrade the flexibility of the Ag network electrode.

To verify the passivating effect of the Al₂O₃ film, the bare Ag and Al₂O₃/Ag network electrodes were subjected to harsh environmental conditions, i.e., a temperature of 85 °C and RH of 85%, over a period of 300 h [41]. The transmittance changes of the electrodes as a function of time are shown in Figure 5a. The graphs exhibit changes in the average magnitude of transmittance in the range of wavelengths from 400 to 800 nm, and the transmittance values at an intermediate wavelength of 550 nm. Owing to moisture degradation, the transmittance of the bare Ag network electrode continuously decreased to ~10% over 300 h. However, with the deposition of a thin Al₂O₃ film on the Ag network electrode, the sample demonstrated relatively constant transmittance values owing to the effective passivation effect of the ALD-grown Al₂O₃ layer. Figure 5b shows photographs of the bare Ag and Al₂O₃/Ag network electrodes before and after the environmental degradation test. The bare Ag network electrode showed a slight decrease in transparency. Although the bare Ag network exhibited a 10% decrease in optical transparency at 550 nm after 300 h, it still maintained high transparency in the visible region. In contrast, the Al₂O₃/Ag network electrodes exhibited constant colors and transmittance values. Because of the high stability of the ALD-grown Al₂O₃ layer, the stability of the Al₂O₃/Ag network electrode was improved compared to that of the bare Ag network electrode.

To further demonstrate the potential of the ALD-grown Al₂O₃ passivation layer, a TFH was fabricated using both the bare Ag and Al₂O₃/Ag network electrodes. A schematic of the TFH structure is shown in Figure 6a. The contact electrode was thermally evaporated on the left and right edges of the samples to improve the contact between the TFE and copper. Copper tape was subsequently attached to the Ag contact electrode, and DC power was applied to the TFE through the copper. The relationship between the applied power and generated change in heat (ΔQ_g) is as follows [22]:

$$∆Q_g=\frac{V^2}{R}∆t=Q_{conv}=h_{conv}{A}_{conv}\left(T_s-T_i\right)$$

(4)

where V is the applied DC voltage; R is the resistance of the TFH; t is the heating time; h_conv is the convective heat-transfer coefficient, and A_conv is the surface area. T_s and T_i are the saturation and initial temperatures, respectively. We can then describe the T_s as follows:

$$T_s=\frac{V^2∆t}{R·h_{conv}{A}_{conv}}+T_i$$

(5)

The T_s is proportional to V² and inversely proportional to resistance. The temperature profile of the TFH devices is shown as a function of the power application time.

Figure 6b shows the temperature change profile of the bare Ag network before and after 85 °C–85% RH test. Prior to the test, the temperature of the bare Ag network-based heater was increased to over 90 °C with an applied power of 5 V. However, significant degradation could be observed in the temperature profile following the 85 °C–85% RH test. The maximum temperature of the bare Ag network-based TFH following the 85 °C–85% RH test was only 74 °C. It can be inferred that the corrosion of the Ag network caused by humidity and high temperature increased the resistance of the electrode and lowered the performance of the TFH. The temperature profile of TFHs with an Al₂O₃-coated Ag network electrode is displayed in Figure 6c. The conductive Ag network electrode has a significant impact on the initial temperature profile in TFHs, regardless of the presence of the Al₂O₃ passivation layer. As a result, the temperature profiles of both samples are quite similar before undergoing the 85 °C–85% RH test. The Al₂O₃-passivated TFH maintained almost the same performance as that before the test. Figure 6d shows the irradiation image and photograph of the TFH as it was applied to flexible and wearable heaters. The TFHs worked effectively in a twisted or worn state owing to the high flexibility of the Ag network.

4. Conclusions

A transparent Al₂O₃ film was deposited on an Ag network electrode via ALD to prevent its corrosion and degradation. ALD is widely used to deposit thin films with high aspect ratios and precise thickness control on sophisticated structured patterns. The use of ALD-grown Al₂O₃ film did not significantly alter the optical properties of the Ag network electrode. In addition, the deposition of the Al₂O₃ passivation layer did not affect the mechanical flexibility of the Ag network electrode, as confirmed via dynamic mechanical tests, including bending, rolling, and twisting tests. To demonstrate the feasibility of using the Al₂O₃ passivation layer, the electrodes were also exposed to harsh environmental conditions, i.e., a temperature of 85 °C and a relative humidity of 85%, following which the optical transmittance of the bare Ag network decreased by approximately 10%, while the Al₂O₃ passivated Ag network demonstrated a relatively constant transmittance. Moreover, when the electrodes were applied to TFH devices, the passivation effect of Al₂O₃ could be more explicitly demonstrated. Prior to the 85 °C–85% RH test, the temperature of both TFHs was increased to values over 90 °C by applying a potential of 5 V. However, following the 85 °C–85% RH test, the temperature of the bare Ag network TFH decreased to 70 °C, implying the significant deterioration of the electrode. Conversely, the temperature of the Al₂O₃-passivated TFH remained over 90 °C, demonstrating considerable durability and reliability under elevated humidity and temperature conditions. Therefore, the Al₂O₃ thin film prevented the corrosion of the Ag network and deterioration of the electrode as well as thermal stability of the Ag network. In addition, the flexible and transparent Al₂O₃-passivated Ag network electrodes can be adapted to various applications involving prolonged atmospheric exposure, such as smart windows, vehicles, or bio-devices. Therefore, the Al₂O₃ film is a favorable choice as a passivation layer for the Ag network, which can be used in both flexible and rigid devices.