Corrosion protection of Mg–11Li–3Al–0·5RE alloy using hybrid epoxy/silica conversion coatings

Materials

The substrate material was a novel Mg–11Li–3Al–0·5RE alloy, which was prepared by the smelting method. The mass composition of the alloy consists of 11%Li, 3%Al, 0·5%RE and the remaining balance of Mg. The alloy was cut into specimens with dimensions of 20×15×3 mm, which were ground successively to a final 1500 grit SiC paper, washed with distilled water and dried with warm air. Epoxy resin used in this study was the diglycidyl ether of bisphenol A (E-51), which was purchased from Wuxi Resin Corporation, China, and was dried at 40°C in a vacuum oven for 6 h before use. The modifier of epoxy resin, γ-isocyanatopropyltriethoxysilane (KBE-9007) was purchased from Shin-Etsu Corporation, Japan. Tetraethoxysilane and other chemicals were obtained commercially and used as received.

Corrosion protection treatments

The first step is the chemical treatment for the ceria or phytic acid conversion coatings on the alloy. According to our previous work,15,16 the processing parameters for the ceria conversion coating were 0·05M cerium nitrate solution at 35°C for 20 min while for the phytic acid conversion coating, 20 g L⁻¹ phytic acid solution was used at pH 6 and 35°C for 10 min.

Second, an epoxy/SiO₂ hybrid dope was prepared via epoxy modified SiO₂ precursor with diethylenetriamine added as curing agent in N-methyl-2-pyrrolidone. Epoxy modified SiO₂ precursor was prepared through mixing modified epoxy resin with nanoSiO₂ precursor solution at 50°C for 4 h according to our previous work.17 The nanoSiO₂ precursor solution was synthesised as follows: tetraethoxysilane, H₂O and ethanol were mixed in the approximate molar ratio of 1∶4∶15. This amount of water is the requirement for forming anhydrous silica according to the net reaction. A small quantity of NH₃.H₂O was added to adjust the value of pH to 9. The mixture was stirred at 50°C for 4 h, and then aged at room temperature to obtain the nanoSiO₂ precursor. The modification of epoxy resin was made as follows: 2 g γ-isocyanatopropyl-triethoxysilane was added to 50 g epoxy resin at 60°C, and then stirred for 4 h until the NCO group of γ-isocyanatopropyltriethoxysilane reacted with the OH group of epoxy resin completely.

The Mg–Li alloy specimens, with one or other conversion coating, were immersed into the epoxy/SiO₂ hybrid dope for 1 min. The specimens were then placed in a furnace at 70°C for 24 h in order to cure the epoxy/SiO₂ hybrid coating.

Corrosion resistance test

The corrosion resistance of the coated and uncoated Mg–11Li–3Al–0·5RE alloy was estimated using electrochemical measurements, hydrogen evolution and salt spray testing. Potentiodynamic polarisation curves performed in 3·5%NaCl solution from −2600 to −600 mV using an Autolab, VWP31Z with the experimental data were analysed using the EC-lab software at room temperature at a scan rate of 10 mV s⁻¹. A three electrode cell with sample as working electrode, saturated calomel electrode (SCE) as reference electrode and platinum sheet as counter electrode were employed in this test. Electrochemical impedance spectroscopy (EIS) was also measured at the open circuit potential in 3·5%NaCl from 10 to 22 days immersion over a frequency range from 105 to 10⁻² Hz, with an amplitude perturbation of 5 mV. The schematic diagram of the set-up for the hydrogen evolution method is shown in Fig. 1. The salt spray test was conducted in a CCX2000 salt spray cabinet (ATLAS, USA) and the coated Mg–11Li–3Al–0·5RE alloy substrates were evaluated by exposure of the samples to a salt fog atmosphere generated from 5 wt-% aqueous NaCl solution at 35±2°C.

OPEN IN VIEWER

Figure 1

Schematic diagram of set-up for hydrogen evolution method

Potentiodynamic polarisation

Ceria conversion coating plus epoxy/silica hybrid coating

Potentiodynamic polarisation was conducted in a 3·5 wt-%NaCl solution in order to evaluate the corrosion resistance of the coated alloy. Figure 2 shows the polarisation curves for Mg–11Li–3Al–0·5RE alloy in 3·5 wt-%NaCl solution with the Ce based hybrid coating. The silica content of the epoxy/silica hybrid coating was varied and the fitted results of the polarisation curves are shown in Table 1.

OPEN IN VIEWER

Figure 2

Potentiodynamic polarisation curves of a Mg–11Li–3Al–0·5RE alloy substrate, b ceria conversion coating and c–g Ce based hybrid coating: contents of SiO² were 0·5, 1, 3, 6 and 9 wt-% respectively

OPEN IN VIEWER

Table 1

Parameters of polarisation curves of Mg–11Li–3Al–0·5RE alloy, ceria conversion coating and Ce based hybrid coatings

	E corr/V(SCE)	I corr/A cm−2	R p/Ω cm−2
Mg–11Li–3Al–0·5RE alloy substrate	−1·65	1·05×10−3	34·1
Ceria conversion coating	−0·99	2·96×10−7	1·52×105
Ce based hybrid coating (0·5 wt-%SiO2)	−0·74	5·75×10−8	7·74×105
Ce based hybrid coating (1 wt-%SiO2)	−0·70	3·80×10−8	1·33×106
Ce based hybrid coating (3 wt-%SiO2)	−0·53	1·20×10−9	4·35×107
Ce based hybrid coating (6 wt-%SiO2)	−0·64	4·48×10−9	9·70×106
Ce based hybrid coating (9 wt-%SiO2)	−0·69	1·88×10−8	2·26×106

The ceria conversion coating alone significantly decreased the corrosion current density I _corr and increased the corrosion potential E _corr. This means that it provided a valuable degree of corrosion protection to the alloy. However, after the epoxy/SiO₂ hybrid coating was formed on top of the ceria conversion coating, the I _corr further decreased and E _corr further increased. Thus, the overall corrosion resistance of the alloy was distinctly improved by the combination of the ceria conversion coating and the epoxy/SiO₂ hybrid coating. Furthermore, the amount of silica in the hybrid top coating clearly affected the anticorrosion ability. Thus, the optimum content of SiO₂ was found to be 3 wt-%, resulting in a decrease in I _corr to 1·199×10⁻⁹ A cm⁻², or six orders of magnitude lower than that of the uncoated alloy substrate (1·046×10⁻³ A cm⁻²). At the same time, E _corr increased to −0·530 V(SCE), which was ∼1 V higher than that of the alloy substrate. These data clearly show that the corrosion resistance of the alloy with the Ce based hybrid coatings was efficiently enhanced at an optimum SiO₂ content of 3 wt-%.

Phytic acid conversion coating plus epoxy/silica hybrid coating

Figure 3 shows the polarisation curves for Mg–11Li–3Al–0·5RE alloy in 3·5 wt-%NaCl solution with a phytic acid conversion coating, and the PA based hybrid coating containing various SiO₂ contents. The fitted results of E _corr and I _corr are listed in Table 2. The PA based hybrid coating also reduced I _corr and ennobled E _corr in a similar manner to the ceria coating. However, PA based hybrid coating was somewhat more effective than the Ce based hybrid coating. At the optimum silica content (also 3 wt-%), the I _corr was reduced by around seven orders of magnitude and the corrosion potential ennobled by 1·12 V. Compared with the Ce based hybrid coating, PA based hybrid coatings are better, probably because the phytic acid conversion coatings, which contain hydroxyl groups and six phosphate carboxyl groups, might generate additional chemical bonding with the subsequent epoxy/SiO₂ hybrid coatings thus improving adhesion and reducing diffusion or species.

OPEN IN VIEWER

Figure 3

Potentiodynamic polarisation curves of a Mg–11Li–3Al–0·5RE alloy substrate, b phytic acid conversion coating and c–g PA based hybrid coating: contents of SiO² were 0·5, 1, 3, 6 and 9 wt-% respectively

OPEN IN VIEWER

Table 2

Parameters of polarisation curves of Mg–11Li–3Al–0·5RE alloy, phytic acid conversion coating and PA based hybrid coatings

	E corr/V(SCE)	I corr/A cm−2	R p/Ω cm−2
Mg–11Li–3Al–0·5RE alloy substrate	−1·65	1·05×10−3	34·1
Phytic acid conversion coating	−0·91	2·63×10−6	1·57×104
PA based hybrid coating (0·5 wt-%SiO2)	−0·68	4·50×10−8	8·25×105
PA based hybrid coating (1 wt-%SiO2)	−0·62	7·24×10−9	6·07×106
PA based hybrid coating (3 wt-%SiO2)	−0·52	9·93×10−10	4·49×107
PA based hybrid coating (6 wt-%SiO2)	−0·55	4·14×10−9	1·12×107
PA based hybrid coating (9 wt-%SiO2)	−0·65	2·37×10−8	6·52×105

Evaluation by hydrogen evolution rate

As shown in the above research, it is concluded that the optimum SiO₂ content was 3 wt-% for both Ce based hybrid coating and PA based hybrid coating. Therefore, hybrid coatings with 3 wt-%SiO₂ content were used to continue the following evaluations of corrosion resistance, including the analysis by hydrogen evolution rate and salt spray test. Figure 4 shows the hydrogen evolution rate for the uncoated and coated alloys during immersion in 3·5 wt-%NaCl solution. At the beginning of the immersion, small bubbles were visible on the surface of the uncoated alloy which subsequently increased in size. The results show that the hydrogen evolution rate of the uncoated alloy increased with immersion time sharply, i.e. it was quickly corrosive with a high hydrogen evolution rate. On the other hand, there were basically no bubbles attached to the surface of the specimens with either the ceria conversion coating or phytic acid conversion coating at the beginning of the immersion. After 1 h of immersion, only a few bubbles were found to be attached to the corner of samples where the conversion coating might be thinner. Corrosion was observed to start at uneven spots of the surface, which presumably formed microcells in the corrosion medium. Subsequently, corrosion gradually expanded to the whole surface while the conversion coating broke off. Therefore, the efficient protection of the conversion coating on the alloy was just for short period.

OPEN IN VIEWER

Figure 4

Hydrogen evolution rate of a Mg–11Li–3Al–0·5RE alloy substrate, b phytic acid conversion coating, c ceria conversion coating, d Ce based hybrid coating and e PA based hybrid coating

No bubbles were attached to the surface of the specimens with Ce based hybrid coating or PA based coating within 30 h immersion. Moreover, only a few bubbles were found to slowly attach to the corner of hybrid coating systems at longer exposure times. The volume of evolved hydrogen was nearly zero in 30 h in the initial stages of immersion after which the evolved hydrogen increased a little with the immersion time reaching only 8·4-9·7 mL cm⁻² after 57 h of immersion. As shown in Fig. 4, the hydrogen evolution rates of the specimen with either the Ce based hybrid coating or the PA based coating were obviously lower than the others. Comparing the Ce based and PA based hybrid coatings, the hydrogen evolution rate with the PA based hybrid coating was slightly lower than that with the Ce based hybrid coating, which suggests that the corrosion resistance of the PA based hybrid coating was better.

Evaluation by salt spray test

Salt spray testing was employed to further investigate the corrosion resistance of the coatings; Fig. 5 shows the salt spray test results of the coatings. No visible corrosion was observed with the three coatings for the first 6 days. After 8 days, a few corrosion pits appeared on the epoxy/SiO₂ hybrid coating distributed at the specimen edges. During this time, both the Ce based and PA based hybrid coatings remained intact with no obvious changes. Finally, after 10 days corrosion, pits started to appear on these coating systems, however, with a reduced corrosion area compared with the epoxy/SiO₂ hybrid coating. This means the corrosion resistance of Ce based and PA based hybrid coatings was better than that of epoxy/SiO₂ hybrid coating because of the additional presence of the conversion coating. The corrosion area for the Ce based and PA based hybrid coatings was almost the same until around 18 days after which the corrosion area of PA based hybrid coating was less than that of Ce based hybrid coating from 18-22 days, and the difference expanded with the extension of corrosion time. It is concluded that the PA based hybrid coating had the better corrosion resistance probably because of the additional chemical bonding that may be formed between phytic acid conversion coating and the subsequent epoxy/SiO₂ hybrid coating. This result is consistent with those of electrochemical potentiodynamic polarisation curves and hydrogen evolution rate.

OPEN IN VIEWER

Figure 5

Salt spray test results of a epoxy/SiO² hybrid coating, b Ce based hybrid coating and c PA based hybrid coating

Analysis of corrosion process of PA based hybrid coating by EIS

Since the corrosion of the PA based hybrid coating appeared to start after 10 days salt spray test, as shown in Fig. 5, the corrosion of this coating was further studied from EIS after 10 days immersed in 3·5%NaCl. Figure 6a shows the EIS results after 10 days immersion in 3·5 wt-%NaCl solution. The diagram showed an obvious trailing circular arc, the circular arc at high frequency region was attributed to the coating and the linear part at low frequency region was corresponding to the substrate/coating interface, suggesting that a diffusion phenomenon was happening at this time. The electrochemical impedance response was analysed using the ZSimpWin code using the equivalent circuit shown in Fig. 7, where R _s is solution resistance, R _p is coating resistance, C _p is coating capacitance and W is the diffusion component. After 10 days of immersion, it is assumed that the electrolyte species are beginning to significantly penetrate the coating system and electrochemical corrosion (development of anodes and cathodes) is occurring at the interface between the substrate and the conversion coating. Diffusion of reacting species (e.g. dissolved oxygen or ions) would be expected to be the controlling process given the impedance response observed under these conditions.

OPEN IN VIEWER

Figure 6

Impedance diagrams of PA based hybrid coating in 3·5 wt-%NaCl solution at a 10 days immersion, b 13 days immersion, c 16 days immersion, d 18 days immersion, e 20 days immersion, f 21 days immersion and g 22 days immersion

OPEN IN VIEWER

Figure 7

Equivalent circuit of EIS of PA based hybrid coating at 10 days immersion

Figure 6b–d shows the impedance diagrams of the specimens at 13, 16 and 18 days immersion in 3·5 wt-%NaCl solution respectively. Comparing Fig. 6b, c and d with a, it is found that the straight line diffusion feature decreases with time. The trailing circular arc still existed at 13 and 16 days immersion, at this time the ions in corrosion medium gradually penetrated the epoxy/SiO₂ coating and reached phytic acid conversion coating, and then Cl⁻ corroded part of the conversion coating. As shown in Fig. 6d, the trailing circular arc dissolved completely at 18 days immersion. At this time, the passages of ions expanded because the electrolyte continually penetrated the coating, and the electrolyte reached the interface of Mg–11Li–3Al–0·5RE alloy and the conversion coating. Then the corrosion of Mg–11Li–3Al–0·5RE alloy started at the corroded part of the conversion coating, and more corrosion products were generated. Therefore, the corrosion process could not be controlled by diffusion process, and it was controlled by charge transfer process. The impedance expanded, which means that the charge transfer resistance increased and the corrosion rate reduced compared with the beginning of the salt spray test.

Figure 6e and f shows the impedance diagrams of the specimens at 20 and 21 days immersion in 3·5 wt-%NaCl solution respectively. As shown in Fig. 6e and f, the maximum impedance was obtained at 20 days immersion, and then the impedance shrank at 21 days immersion.

Figure 6g shows the impedance diagrams of the specimen at 22 days immersion in 3·5 wt-%NaCl solution. At this period, more and more electrolyte reached the interface, capacitive reactance model between electrolyte and corrosion product film was built and the diffusion character disappeared. Therefore the equivalent circuit in Fig. 7 was not suitable, and the new equivalent circuit fitted through ZSimpWin software is shown in Fig. 8, in which C ₂ was film capacitance of corrosion products, R ₂ was film resistance of corrosion products, C ₃ was the inner oxidation film capacitance, R ₃ was the inner oxidation film resistance, C ₁ was epoxy/SiO₂ hybrid coating capacitance, R ₁ was epoxy/SiO₂ hybrid coating resistance and R _s was solution resistance. At this time, the impedance obviously shrank, probably because lots of Cl⁻ destroyed the corrosion product film on the interface resulting in the aggravated corrosion process. This impedance diagram showed diffusion character, but it is different from the diffusion at 10 days immersion. The diffusion at 10 days immersion existed in the coating, while the acceleration of reaction speed of Mg–11Li–3Al–0·5RE alloy led to the new concentration grads and resulted in the diffusion at 22 days immersion.

OPEN IN VIEWER

Figure 8

Equivalent circuit of EIS of PA based hybrid coating at 22 days immersion

The anticorrosion of PA based hybrid coating was analysed as follows. Because the main coating is high cross-linking epoxy resin, some pore canals will be formed in the coating. These pore canals provide the passages for ions. The ions in corrosion medium reach the interface between epoxy/SiO₂ hybrid coating and phytic acid conversion coating through the passages, corrode the conversion coating and then reach the interface between the conversion coating and the alloy. In epoxy/SiO₂ hybrid coating, there are SiO₂ particles which can efficiently intermit the passages for ions, so that the hybrid coating has better corrosion resistance than epoxy resin. Furthermore, the corrosion should be retarded when there is conversion coating between Mg–11Li–3Al–0·5RE alloy and epoxy/SiO₂ hybrid coating. The corrosion product film will be formed on the interface of the conversion coating and the alloy, and can partially restrain the corrosion of Mg–11Li–3Al–0·5RE alloy. But with more and more corrosive ions reaching the interface, the corrosion product film is gradually destroyed, resulting in the aggravated corrosion process. These are the reasons why the impedance expanded and then shrank during the corrosion process.