Development of novel surface treatments for corrosion protection of aluminum: self-repairing coatings

2.1 Structure of microcapsules containing liquid repairing agents

Here, the processes of the formation of microcapsules containing liquid repairing agents will be described, and the influence of the formation conditions on the structure of the microcapsules (Yang et al. 2008; Sondari et al. 2010; Yamada et al. 2017; Chiba et al. 2017) will be discussed. First, a “prepolymer”, a precursor for the shells of the capsules, was synthesized by the reactions shown in Figure 2A. Tolylene-2,4-diisocyanate (TDI), glycerol, and cyclohexanone were mixed at 75°C with 600 rpm agitation for 24 h. Then, to prepare a “prepolymer solution”, this prepolymer in the cyclohexanone was mixed with isophorone diisocyanate (IPDI) alone or with 50 wt% of xylene.

Figure 2:

Chemical reactions of (A) synthesis of prepolymer with TDI and glycerol and (B) formation of polyurethane capsules.

Next, microdroplets of the prepolymer solution described above were dripped at regular intervals into an aqueous solution containing 0.5% glycerol and 3% sodium dodecyl sulfate (SDS) with agitation. The prepolymer solution dripped into the agitated glycerol/SDS solution forms spherical micelles because the prepolymer solution does not dissolve in the glycerol/SDS solution and the SDS acts as a surface-active agent. During the dripping and agitation, the prepolymer at the micelle surface reacts with the glycerol according to the scheme in Figure 2B, forming microcapsules with solid shells of polyurethane and liquid contents of the repairing agents.

The structure of the capsules is affected by the presence of xylene in the prepolymer solution. To examine the structure of capsules by scanning electron microscopy (SEM), the solution with the dispersed capsules was filtered, and the capsules were collected and dried. Figure 3 shows SEM images of the surface of capsules formed from the prepolymer solution (A) without xylene and (B) with xylene. The capsules formed without xylene are spherical, 5–10 μm in diameter (Figure 3A). The capsules formed with xylene are spherical and appear as coalesced clusters, shown by the white and black circles, respectively. The average of the diameter of the capsules is 3 μm.

Figure 3:

SEM images of surface of capsules formed (A) without the addition of xylene and (B) with xylene.

The effect of the addition of xylene on the structure of the capsules will be discussed below. It is possible that this effect is correlated to the surface tension of the prepolymer solution. The diameter of micelles in the glycerol/SDS solution is affected by the interface tension between the prepolymer and glycerol/SDS solutions. Lower interface tension tends to form smaller micelles, according to the surface energy change on the micelle formation. The addition of xylene acts to decrease the interface tension of the prepolymer solution, leading to smaller capsules. More details of the mechanism are described elsewhere (Chiba et al. 2017).

To determine the inner structure of capsules formed with or without xylene, polyurethane coatings with dispersed capsules were sliced with a cutter blade to expose the inside of the capsules before SEM observation. Figure 4 shows SEM images of vertical cross-sections of coatings with dispersed capsules formed (A) without and (B) with xylene. There are dark circular 5–10 μm diameter areas and small dots with a 1 μm diameter in the center of the circular areas in Figure 4A. The dark circular areas shown by a black circle may correspond to the dispersed capsules in the coating and the small dots correspond to whose composition is different from that around the dot. This result suggests that the capsules formed without xylene are solidified wholes, leading to an absence of liquid phases. The central dot is low-polymerized compounds.

Figure 4:

SEM images of cross-section of capsules formed (A) without the addition of xylene and (B) with xylene.

In contrast with Figure 4A,B shows the presence of smaller capsules and pores at the center of capsules shown by a white circle. It is reasonable to assume that the pores contain a liquid phase, and that the liquid has evaporated while exposed to a vacuum for the SEM observations. The volume of the liquid phase contained in capsules depends on the thickness of the shells of the capsules.

The reason why a thinner shell of capsules is formed from the prepolymer solution with xylene can be explained as follows (Chiba et al. 2017). During the formation of capsules, glycerol penetrates through the shell into the inside of capsules to react with the prepolymer. The penetration rate of glycerol is determined by the solubility of the glycerol in the prepolymer solution. The solubility of glycerol is much lower in the prepolymer solution with xylene than in the prepolymer solution without xylene, leading to a low penetration rate of glycerol into the capsule during the capsule formation. As a result, the capsules produced from the prepolymer solution with xylene have thinner shells and contain a larger volume of liquid repairing agents.

2.2 Self-repairing of flaws in the coatings with dispersed capsules

The self-repairing properties of coatings with dispersed capsules were examined in the following procedure in which the two kinds of differently formed capsules were investigated: capsules formed (1) without xylene and (2) with xylene. After filtering by filter paper with 8 μm of pore size (Grade 40, Whatman) and drying, 10 mg of the capsules formed either without xylene or with xylene was mixed with 1.5 g of prepolymer solution and 0.22 g of ethylene glycol and the mixture was painted onto Al substrates (20 mm × 30 mm × 0.4 mm) in a supersonic wave environment. A polyurethane coating with the dispersed capsules was formed by the polymerization of the prepolymer during aging of the painted specimen for 48 h at 333 K. The aged coatings were damaged by scratching with a cutter blade (SB10K, OLFA), whose width was 0.38 mm, at 1.5 N of compressive force over a-15 mm length before curing for 24 h before SEM observations.

Figure 5 shows surface SEM images of the polyurethane coatings with capsules after scratching and curing, the capsules being formed (A) without xylene and (B,C) with xylene. It can be seen from Figure 5A that a clear scar of approximately 200 μm width is present at the center of the image, and that the Al substrate is exposed at the bottom of the scar. This strongly suggests that there is no failure of the scar and no repair of the coating. The image of the surface of the coating with capsules formed with xylene shows a linkage structure across parts of the scar (Figure 5B), whereas Figure 5C shows a scar that is nearly completely repaired.

Figure 5:

Surface SEM images at damage by scratching on coatings formed (A) without xylene, and (B,C) with xylene.

From the results above, it may be concluded that the coatings with dispersed capsules provide self-repairing properties without any triggering reaction. The repairing mechanism in this study will be explained next. When the coating with dispersed capsules is physically damaged, the capsules are destroyed, leading to the IPDI with solvent flowing out from the capsules. The IPDI fills the damaged area, and reacts with water vapor in the air and water adsorbed on the specimen surface and cutter blade to form polyurea, with a chemical structure similar to that of polyurethane (Figure 6). During curing, the solvent evaporates, and the polyurea forms strong chemical bonds with the polyurethane.

Figure 6:

Chemical reactions of self-repairing of coatings with dispersed capsules.

3.1 Structure of porous anodic oxide films with inhibitors

Here, #1050 Al alloy plate specimens (0.4% Fe, 0.1% Si, bal. Al, 20 mm × 20 mm × 0.3 mm) were electropolished in 78 vol% CH₃COOH/22 vol% HClO₄ solution at a constant voltage of 30 V for 60 s. Porous anodic oxide films were formed on the pretreated specimens by anodizing in 2 wt% of (COOH)₂ solution at a constant current density of 20 mA/cm² for 1 h. A SEM image of the vertical section of an anodized specimen showed that the specimen is uniformly covered with a porous-type anodic oxide film with a 25 μm thickness, as shown in Figure 8. The anodized specimen was immersed in 5 wt% glycerol/5 wt% inhibitor solution for 24 h to fill the pores with the solution. The inhibitor used was Kurilex L-501 (Kurita Water Industries Ltd.), in which the main components were nitrite, silicate, and water-soluble polymer. After wiping the surface of the specimen with a soft tissue, the specimen was immersed for 24 h in the prepolymer solution (see Section 2.1). During the immersion, the prepolymer reacts with glycerol to form polyurethane film by the reaction described in Section 2.2, and the inhibitor solutions contained in the pores of anodic oxide films were covered with the polyurethane film. After removing from the prepolymer solution, the specimen was kept at room temperature for 24 h prior to the SEM observations.

Figure 8:

SEM image of vertical cross-section of a #1050 Al alloy specimen anodized in 2 wt% (COOH)₂ at 20 mA/cm² for 1 h.

Figure 9A and B show surface SEM images of an anodized only specimen and anodized specimens with inhibitor filling and polyurethane film covering, respectively, and Figure 9C and D are schematic illustrations of vertical cross-sections of the specimens from Figure 9A and B. Many 100 nm diameter pores are observed on the surface of porous anodic oxide films and the pores are regularly arranged with an interval of 240 nm (Figure 9A). The SEM image of the surface of the polyurethane-covered anodized specimen (Figure 9B) is similar to that of Figure 9A, showing many circular patterns with a 230 nm interval. However, the covered specimen shows less contrast than the uncovered specimen. Figure 9C and D assist in a clearer understanding of the difference in the contrast between Figure 9A and B.

Figure 9:

(A) SEM images of the surfaces of an anodized specimen and (B) a polyurethane-covered anodized specimen. (C,D) Schematic illustration of the vertical sections of specimens from A and B, respectively.

3.2 Self-repairing by the inhibitor stored in the pores

Here, a study about the self-repairing of damaged Al alloys covered with porous-type anodic oxide films with inhibitors stored in pores in the film will be discussed. In this study, corrosion behavior after scratching with a cutter blade at the compressive force of 2.55 N in the shape of two lines crossing at right angles was examined on #1050 Al alloy specimens with three different treatments: (A) electropolishing, (B) anodizing and (C) anodizing with inhibitor solution storage. The conditions of electropolishing, anodizing, and inhibitor storage are described in Section 3.1.

After scratching the surface with a cutter blade, the three kinds of specimens were kept in a desiccator for 24 h, and subjected to immersion tests in 1.57 × 10⁻³m–CuSO₄/0.57 m–KCl solution for 24 h at 20°C. The surface of the specimens after the immersion test was observed under an optical microscope.

Figure 10 shows the appearance of the surface of three differently pretreated Al alloys immersed in 1.57 × 10⁻³m–CuSO₄/0.57 m–KCl solution for 24 h after the scratching damage: (A) electropolishing, (B) anodizing, and (C) anodizing with inhibitor solution storage in the pores. The crossing areas surrounded with dotted lines in all the figures correspond to the area damaged with the cutter blade. The specimen with electropolishing shows reddish brown colors across the entire surface and white discontinuous lines at the damaged area (Figure 10A). The reddish brown color is due to the Cu deposited on the surface and the white deposits are corrosion products, mainly composed of Al hydroxides.

Figure 10:

Appearance of the three differently pretreated Al alloys after immersion in 1.57 × 10⁻³m–CuSO₄/0.57 m–KCl solution for 24 h following pretreatment by (A) electropolishing, (B) anodizing and (C) anodizing and inhibitor storage.

There are many reddish brown dots over the entire area of the specimen covered with the porous-type anodic oxide film, and the dot distributes at a high density in the damaged area (Figure 10B).

Differently from the results in Figure 10A–C shows no change in the surface appearance at both the damaged and undamaged areas, strongly suggesting the absence of corrosion.

The corrosion mechanism of the three kinds of specimens in solutions containing Cu²⁺ and Cl⁻ ions is schematically illustrated in Figure 11. The electropolished specimen is covered by an approximately 3.5-nm-thick oxide film (Takahashi & Nagayama, 1985), and there are many imperfections in the oxide film on the second phases and Al matrix (Figure 11A). When immersed in the Cu²⁺/Cl⁻ solution after scratching with the cutter blade, Cu particles are deposited at both scratched and undamaged areas [Eq. (1)], and severe corrosion proceeds across the entire specimen surface by Eqs. (2) and (3). The anodized specimen is covered with a porous anodic oxide film, 25 μm thick, and there are imperfections in the oxide film on the second phases (Figure 11B). The number of imperfections in the anodic oxide film seems to be much smaller than that in the oxide film formed after electropolishing. As a result, the corrosion of the anodized specimen occurs mainly at the scratched areas and at the imperfections in the oxide films as a secondary process. The specimen with anodizing and inhibitor solution storage has no imperfections because imperfections originated from the second phase contact with inhibitors in the pores of the oxide film and are covered with polyurethane film (Figure 11C). When this specimen is scratched, the inhibitors flow out from the destroyed porous layer to cover the Al substrate at the scratched area. The inhibitor forms a thin silicate layer on the specimen surface and blocks the deposition of Cu particles from the solution (Hiraga et al. 2013).

Figure 11:

Schematic illustration of the corrosion mechanism of the three differently pretreated Al alloys during immersion in Cu²⁺/Cl⁻ solution, pretreatment by (A) electropolishing, (B) anodizing and (C) anodizing and inhibitor storage.