Electrochemical Response of Clay/Polyelectrolyte Composite Barrier Coatings

Abstract

Composite materials made of polymer and clay are effective at blocking mass transport. In this study, the blocking efficacy of layer-by-layer (LbL) coatings of exfoliated montmorillonite (MMT) and polyethylenimine (PEI) was studied using cyclic voltammetry and a redox couple, indigo carmine (IC). The pH of the MMT solution was varied from 4 to 10 to prepare LbL coatings of different surface roughness on metal substrates. It was found that the coated electrode had a lower redox peak current value than without the coating, demonstrating the reduction of the mass transport of IC to the metal surface. The peak values decreased with decreasing the coating’s roughness and increasing the number of layers, indicating that the blocking capability can be controlled by changing the deposition conditions. Smooth LbL coatings deposited with MMT at pH 4 showed the highest blocking efficacy up to 97.5%. The IC adsorbed at the interface between the coating and the metal substrate was found to cause the peak current measured for the coated electrode. It was also confirmed that the same coating on the copper substrate reduced the corrosion of the copper during the electrochemical potential cycling.

1. Introduction

Layer-by-layer (LbL) deposition is a versatile manufacturing technology that allows for the fabrication of thin films composed of polyelectrolytes, nanoparticles, and clay by utilizing electrostatic attraction (or hydrogen bonding) between the components [1]. It also allows for control of the thickness and structure of the coating in nanometer scale by simply adjusting the number of layers, concentration, temperature, and pH of the dispersion solution. Moreover, LbL deposition is an environmentally friendly method because it is mostly based on aqueous solutions. Many thin coatings based on LbL deposition are now widely investigated for energy, biological, and environmental applications [2,3,4].

Clay is often used in various organic-inorganic composite materials because of its high mechanical strength and thermal stability, low cost, and availability [5]. Compared to conventional bulk composite materials composed of polymer and clay, a multilayer coating of clay and polymer deposited by the LbL assembly method is more effective at blocking gas diffusion, flame retardation, and metal corrosion protection [5,6]. For example, LbL films of polyvinylamine (PVAm), polyacrylic acid (PAA), polyethylenimine (PEI), and montmorillonite (MMT) were shown to achieve an oxygen permeability as low as 0.009 cm³/(m² day atm) [7]. The permeability decreased when the number of the coated layers increased and depended on the deposited polymers’ types and order. Another study showed that LbL coating of PEI, PAA, and MMT on copper suppressed the corrosion of the metal in H₂S atmosphere about 1000 times via the measurement of the weight reduction [8]. Additionally, LbL coatings of polymethylmethacrylate (PMMA) and MMT were shown to be effective at inhibiting the corrosion of iron in a salty environment by using electrochemical impedance microscopy. The degree of corrosion protection depended on the shape of the clay used [9].

In this paper, an LbL coating composed of MMT and PEI (LbL MMT-PEI) was deposited on metal substrates, and mass transport through the coating was investigated via cyclic voltammograms of the coated electrode in a solution of the redox couple, indigo carmine (IC). The gold electrode coated with LbL MMT-PEI showed a lower redox peak current than the bare electrode, and the current also decreased as the number of coated layers increased. Furthermore, the blocking efficiency of the LbL coating was varied by the structure, and the lower redox current of the coated metal electrode was due to less number of the ICs adsorbed on the surface. When the same LbL coating was applied to a copper substrate, the corrosion was suppressed even under an electrochemical cycling environment.

2. Materials and Method

2.1. Materials

Both PEI (Mw = 2000 g/mol and Mn = 1800 g/mol) and IC (Mw = 466.35 g/mol) were purchased from Sigma Aldrich (St. Louis, MO, USA). Southern Clay Products (Gonzales, TX, USA) provided the sodium montmorillonite (NA+montmorillonite). The MMT solution was prepared as follows. First, MMT was stirred in deionized (DI) water for 24 h using a magnetic bar, and then it was dispersed for about 1 h using a tip sonicator. Subsequently, it was rotated at 3000 rpm in a centrifuge; after removing the precipitate from the solution, the solution was used for the LbL deposition. A dispersion of 0.1 mol/L of PEI was made by stirring it for 24 h in DI water using a magnetic bar.

2.2. Layer-by-Layer Coating

Gold/titanium-coated silicone was used as a metal substrate for all the experiments unless otherwise stated. The thickness of the gold, titanium and silicone layer was 100 nm, 5 nm and 250 μm, respectively. Prior to the LbL coating, the metal substrate was sonicated in acetone and then in water for 5 min. The cleaned substrate was treated with O₂ plasma and then soaked in the PEI solution (pH 10) for 20 min to obtain a positively charged surface. The substrate was washed with DI water to remove any excess PEI remaining on its surface and then used for LbL deposition of MMT and PEI as follows. First, the PEI-coated substrate was immersed in the MMT solution for 10 min and then washed in three consecutive DI water bathes for 1 min each, and then it was dipped into the PEI solution for 10 min and washed in three DI water baths for 1 min each. As a result, one bilayer (BL) of MMT and PEI was formed on the substrate. This sequence was repeated to obtain a multilayer coating of the desired bilayer number.

2.3. Film Characterization

The surface thickness and roughness of the LbL coating were observed and measured using scanning electron microscopy (SEM, S-4800, Hitachi High-Technology, Tokyo, Japan). The working distance and the voltage were around 8 mm and 5 kV, respectively. The thickness profile of a sample was measured at six different locations using a surface profilometer (Dektak 150, Bruker, Billerica, MA, USA) to obtain the mean and standard deviation of thickness and the arithmetic average roughness. Cyclic voltammograms (CVs) and bode plots of the coating were obtained using a potentiostat/galvanostat and electrochemical impedance microscopy (Biologic, VSP 200, Seyssinet-Pariset, France). The coated electrode, an Ag/AgCl electrode, and a platinum mesh were used as a working electrode, a reference electrode, and a counter electrode, respectively. The electrodes were immersed in an acidic aqueous solution (i.e., 0.2 M HCl) with 0.25 mM IC. The CV of the coated electrode was measured in the voltage range of −0.3 to 0.3 V (vs. Ag/AgCl) at a scan rate of 10 to 600 mV/s. The percent blocking efficacy of the LbL coatings was defined as follows:

blocking efficacy (%) = (I_b − I_c)/I_b × 100 (%),

(1)

where I_b is the IC’s peak current from an uncoated gold electrode and I_c is that from the coated electrode.

3. Results and Discussion

Figure 1 shows a schematic diagram of the LbL coating deposited on a metal surface. The LbL assembly can be formed because MMT is electrically negative and can be adsorbed on the electrically positive surface of the PEI-treated metal substrate via electrostatic attraction. This process was repeated to obtain the desired bilayer number by repeatedly immersing the sample in MMT and PEI solutions. As shown in the schematic diagram, PEI connects sheet-shaped MMTs, and small molecules in the solution would penetrate the coating by diffusion. When these molecules reach the metal surface, they will undergo oxidation or reduction reactions, depending on the voltage applied to the metal electrode. In this study, a redox couple (IC) was used, and its reversible redox reactions and corresponding molecular structure change with the transfer of two hydrogen and two electrons were shown in the schematic diagram.

Figure 2a shows the thickness of the LbL (MMT-PEI)_n (n = bilayer numbers) coating assembled using MMT dispersions of pH 4 and 10. The thickness increased as n increased: at n = 5, the thickness per bilayer (BL) was small (20 nm/BL), but, after that, the thickness per bilayer increased to more than 50 nm/BL. This growth rate is much higher than that of the previously reported LbL MMT-PEI coating (3 nm/BL) [6], possibly because of the differences of the PEI molecular weight and the substrates used (i.e., poly(ethylene terephthalate) vs. gold) [10]. It should be noted that the thickness of LbL (MMT-PEI)_n was hardly affected by the pH of the MMT dispersion solution.

However, the surface roughness was significantly influenced by the pH, and it can be seen that, at pH 10, the surface roughness increased more rapidly with the increase in the number of layers than at pH 4 (Figure 2b). This result can be attributed to the pH-dependent structure of the cationic polyelectrolyte PEI (pKa = 6.5~8.5) [7]. At high pH, it is in its coiled state. With the decrease in pH, it changes into a linear form due to the chain’s mutual repulsion. Therefore, the PEI layer already deposited at pH 10 on the electrode surface during the previous LbL process would spread out in plane when the electrode surface was exposed to the pH 4 MMT dispersion. However, it would remain aggregated if it were exposed to pH 10 MMT dispersion. Therefore, the surface roughness would be much higher for the LbL (MMT-PEI)_n deposited with pH 10 MMT solution. Unlike PEI, MMT’s sheet structure does not change significantly with pH, and the negative charge density of MMT is not notably affected by the change of pH from 3 to 11 [11]. LbL (MMT-PEI)_n assembled with pH = 4 MMT solution resulted in a smoother surface, and it would be beneficial for blocking mass transport, serving as a barrier layer.

Figure 2c shows a typical surface of LbL (MMT-PEI)_n=10 coating assembled with pH 4 MMT solution. It resembles a cobbled surface with some thin wrinkles, possibly the thin layers of folded MMT. The cross-section (Figure 2d) of the coating shows a wavy layered structure in which sheets are stacked to the thickness direction. This layered structure is similar to other clay-polymer thin films [5]. As expected, a rough surface and cross-section were observed for LbL (MMT-PEI)_n=10 assembled with pH 10 MMT solution (see Figure S1).

The electrochemical response of the LbL (MMT-PEI)_n coating in an IC solution was investigated using cyclic voltammetry. As shown in Figure 3a, the CV of the uncoated gold electrode shows large and distinct oxidation and reduction peaks at 0.05 and 0.08 V (vs. Ag/AgCl), respectively, due to the redox reaction of IC on the electrode’s surface. Although the coated electrodes also show the faradaic peaks, the peak values are significantly smaller, demonstrating that fewer IC reach the surface. In addition, as the number of coating layers increases to 10 and 20, the redox peak value and the overall current decrease (Figure 3b), indicating that the number of ICs reaching the surface is reduced with higher n. The corresponding blocking efficacy based on the reduction of the peak current reached up to 97.5%.

Compared to the bare electrode, it would take a longer time for IC in the solution to diffuse through the thickness direction of the LbL coating and to approach the metal surface. The reason is that MMT in the LbL (MMT-PEI)_n is sheet shaped, and each layer is deposited in plane with the surface as shown in the SEM images. Furthermore, the coating also has a thickness of ~400 to 1200 nm and is made of a material with low conductivity. The potential difference between the oxidation and reduction peaks of IC in the coated electrode is within ~30 mV at a scan rate of 50 mV/s, which indicates that IC redox reaction is facile. Considering the slow diffusion of IC into the coating, this result suggests that IC adsorbed near the metal surface is responsible for the coated electrodes’ faradaic peaks.

CVs at different scan rates were measured (Figure 3c) to determine whether IC is confined on the surface or dissolved in the solution. IC’s peak current values in the coated electrodes were linearly proportional to the scan rate with a slope of 1, indicating these peaks are from the adsorbed IC near the metal surface [12]. In an uncoated gold electrode, IC’s peak current is proportional to the scan rate multiplied by 0.5, indicating that the IC redox reaction is diffusion controlled. Because the concentration of the surface-adsorbed species is linearly proportional to the peak currents, the reduced peak current values (to 0.45 from 6.06 μA/cm²) indicate that the concentration of adsorbed species on n = 20 is significantly lower than that of n = 10.

Figure 3d shows that the peak current value gradually increases as the number of cycles of CV increases and reaches equilibrium value at about 50 cycles. This indicates that the amount of IC reacting on the surface gradually increases as the cycle is repeated. This result can be attributed to the slow diffusion of IC and water molecules into the coating and, possibly, the molecular rearrangement of the coatings under voltage application. LbL coatings made with MMT solution at pH 7 and 10 show the same trend. However, the peak currents are higher than that of the smoother coating made with pH 4 MMT solution. This means that the rougher surfaces of the LbL coatings made with higher pH facilitate the diffusion and adsorption of IC inside the coatings. In the coated sample, the shape of the CV remains unchanged after stabilization, and the LbL assembly is electrochemically stable like other LbL electrodes [13,14]. These results show that LbL coating is effective at blocking material transfer of IC, and its blocking degree can be controlled by the deposition conditions.

A copper electrode was coated with the LbL coatings made of a pH 4 MMT dispersion, and the corrosion caused by the reaction with chloride ions was investigated. The CVs of the bare copper and the coated copper electrodes in Figure 4 show that an oxidation current and a reduction current occur, respectively, in the voltage range of −0.4 to 0.3 V. The reduction peak is between −0.08 to −0.06 V. This is due to the reaction of copper with chloride ions (Cu + Cl⁻ <-> CuCl + e⁻) [15,16]. Bare copper has the largest reduction peak value and cathodic peak area. In the case of the coated electrode, it can be seen that the values decrease. When the number of layers is 15, the peak value is reduced to about a half, and when the number of layers is 30, it is reduced to about a third. This is close to the level that is reduced (1/2 ~ 1/10) when copper is coated with a common organic inhibitor for copper corrosion such as imidazole [16]. This demonstrates that corrosion due to small ions such as Cl⁻ can be suppressed considerably with this type of LbL coating.

4. Conclusions

Using the LbL deposition method, a metal surface was coated with a multilayer of MMT and PEI composite. The coating’s roughness increased when increasing the pH of MMT dispersion, and the thickness was proportional to the number of BLs. In the IC solution, the coated electrode showed a redox peak current value significantly lower than that of the uncoated electrode, which means that the number of the IC molecules contacting the electrode surface was significantly reduced. A lower current value was measured specially for thickly stacked coatings with small surface roughness and a high number of BLs. Although small values, IC peak current was measured in all coated electrodes, showing that some ICs with a molecular weight of MW ~466 g/mol penetrate the coating and reaches the interface between the electrode and the coating. Cyclic voltammetry revealed that the redox peak appeared due to the reaction of the IC adsorbed on the interface between the metal surface and the coating. These LbL coatings were also found to be effective for inhibiting the copper corrosion in sodium chloride.