Development of a Reliable Accelerated Corrosion Test for Painted Aluminum Alloys Used in the Aerospace Industry

Abstract

New environmental regulations have led to major changes in aluminum corrosion protection by prohibiting, for example, Cr(VI). Thus, the assessment of the corrosion behavior of Cr-free systems under atmospheric conditions is a major topic of interest for the aerospace industry. One major difficulty in this task is the lack of robust and reliable accelerated corrosion test(s) in this field. The aim of the present study is to compare the results of various accelerated corrosion standards (ASTM B117, ISO 4623-2, VCS 1027,149) to results obtained after 5 years of exposure at a marine atmospheric site in Brest, France. Additional accelerated corrosion tests were designed by varying several parameters in the VCS 1027, 149, such as the salt concentration, the time of wetness, and the relative humidity. The different modes of failure obtained in accelerated corrosion tests on the painted samples were then compared to field exposures in a marine atmospheric site. The first results obtained showed that the developed tests are more representative of service conditions than standard tests.

1. Introduction

Aluminum alloys of the 2xxx and 7xxx series are commonly used in the aerospace sector. Despite their favorable mechanical properties, these alloys are prone to corrosion [1].

Long-term corrosion data under atmospheric conditions of these aluminum alloys, including new developments, are, for the moment, rather scarce [2,3,4,5,6,7].

To protect aluminum alloys from corrosion, organic coatings are usually applied. The usual corrosion protection system in the aerospace industry involves a conversion or anodized coating, a primer coating with corrosion inhibitors, and a top coat. In many cases, both the conversion coating and the primer contained Cr(VI) as pigments [8].

As stated by Snihirova et al. [9], Cr(VI) demonstrates high efficiency in corrosion protection. They can inhibit corrosion in active areas by forming a stable Cr(III)-oxide/hydroxide film, creating layers with enhanced electrical resistance, or developing mixed Cr(III)/Cr(VI) surface films.

However, due to environmental regulations (REACH—Registration, Evaluation, Authorization, and Restriction of Chemicals), the use of Cr(VI) is limited or even banished for many applications, including aerospace [10]. Consequently, different surface treatments such as tartaric sulfuric acid (TSA) or sol–gel have been developed to replace chromate conversion coating (CCC) or chromic acid anodizing (CAA).

Several hybrid or organic coating formulations modified with pigments or additives conferring corrosion inhibition are in or under development and have shown promising results [9,11,12,13,14,15,16,17,18]. So, there is a need to qualify new materials and surface finishes for different applications in the aerospace sector [19].

Several tests are used to assess the corrosion resistance of aluminum alloys, but they are not developed specifically for the aerospace industry. For example, the neutral salt spray test (NSST) is widely used, but it is well known that it is not representative of service conditions and thus not adapted to qualify new systems [20]. Similarly, the ISO 4623-2 test [21] developed to evaluate the sensitivity to filiform corrosion could be useful to rank materials with respect to specific types of corrosion. However, the correlation with field exposure is rarely documented. Automotive corrosion tests such as VDA 233-102 [22] or VCS 1027,149 [23] have been used for testing aluminum alloys, in particular for the 5xxx and 6xxx series. However, the experience with aerospace materials is rather limited.

Various factors can influence underpaint corrosion, including filiform and blistering types. Bautista identified three key parameters required to trigger underpaint corrosion [24]. High relative humidity significantly contributes to the corrosion of painted surfaces. Several studies have indicated that a relative humidity of 80–85% is ideal for the formation of underpaint corrosion and, in particular, filiform corrosion [24,25].

The presence of defects also has an impact. Filament development occurs in the presence of mechanical defects, air bubbles, inadequate film thickness at edges or specific points of the structure, and dirt on the metal surface.

As for bare materials, the presence and concentration of corrosion active ions, as well as the pH, also impact corrosion by quickly breaking down the passive oxide film of aluminum [24,26,27]. In addition to these factors, the presence of oxygen, the nature of the substrate, and the surface treatment can influence the initiation and propagation of underpaint corrosion [24,26,27]. There is, thus, a need to design reliable and representable testing conditions for aerospace alloys and surface finishes used for exterior and interior applications.

The objective of this study was to compare the results of different accelerated corrosion test conditions, including standardized tests (ASTM B117 [20], ISO 4623-2, VCS 1027,149 [23]), with developed test conditions. As a result, various accelerated corrosion tests were created by adjusting multiple parameters in the VCS 1027,149, such as salt concentration during the spraying phase, the time of wetness, and the relative humidity during the wet phase of the test. The results were then compared to field exposures at a marine site.

In this study, eight painted aluminum alloys were selected based on their known performance in service conditions or in marine exposure conditions, specifically considering the type of corrosion and the level of resistance of the painted systems (good, medium, and poor).

2. Materials and Methods

2.1. Materials

Eight painted aluminum alloys, including different surface treatments, primers, and top coats, were selected according to their expected corrosion behavior. Two systems are Cr-free systems following the REACH legislation, while the other systems are Cr(VI) systems with different surface treatments and primers. Systems 3, 4, and 7 consist of cladded AA 2024. The cladding (AA1050) tends to dissolve preferentially, thus protecting electrochemically the substrate (AA2024). Their compositions are presented in

Table 1. Description of painted samples.

Nr	Alloy	Surface Treatment	Primer	Top-Coat	Target Value of Thickness (µm)
1	AA 2024 T3 sheet	TSA	Cr-free primer 1 (epoxy primer)	Top-coat 1 (polyurethane)	55
2		CrIII conversion coating
3	AA 2024 clad	TSA	CrVI primer 1 (2 or 3 components amine cured epoxy primer)		55
4	AA 2024 T351 clad	Sol–gel	CrVI primer 2 (epoxy primer)	Top-coat 2 (polyurethane)	69
5	AA 2324 T39	TSA	CrVI primer 3 (3 components amine cured epoxy primer)		55
6	AA 7055 T7751	TSA	CrVI primer 4 (epoxy primer)		55
7	AA 2024 T351 clad	Sol–gel	CrVI primer 5 (epoxy primer)		56
8	AA 2324 T39	TSA	CrVI primer 4 (epoxy primer)		55

.

The painted systems have been exposed as 100 mm × 150 mm panels. Three replicates were tested each time, whether in accelerated tests or in natural exposure conditions.

All painted systems have been scribed to simulate a defect in the paint. One scribe of 100 mm in length with a nominal width of 1.3 ± 0.3 mm and a nominal depth of 300 µm has been applied to the systems.

2.2. Field Exposures

The samples were exposed for 5 years at 45° facing south at the marine field station of Saint-Anne in Brest (France). Environmental data and corrosivity data of the field station are presented in

Table 2. Environmental and corrosivity data on the period of February 2018 to June 2023 of Sainte-Anne station, Brest, France. * Time of Wetness (TOW): time of period during which the relative humidity is more than 80% and the temperature is above 0 °C.

Average Temperature (°C)	Average Relative Humidity (%)	TOW * (%) Over the Period	Precipitation (mm) Over the Period (Sum)	Chloride Deposition (mg/m², d) (Mean)	Corrosivity Class (ISO 9223 [29] for Al AA1050)	Corrosivity Class (ISO 9223 for Carbon Steel)
13	82	59	5485	960	C2/C3	C4/C5

. The measurements of the chloride deposition were conducted using the wet candle method according to ISO 9225 [28].

2.3. Accelerated Corrosion Tests

2.3.1. Standardized Accelerated Tests

NSST, as per ASTM B117, was performed on the painted system for a duration of 3000 h.

A filiform corrosion test as per ISO 4623-2, including an initiation with HCl vapor, was performed. The samples were exposed horizontally to HCl vapor for 1 h. Once initiated, the samples were exposed in a climatic chamber at 40 ± 1 °C and 82 ± 3% relative humidity for a duration of 6 weeks (1000 h).

Finally, a VCS 1027, 149 test was performed. This test corresponds to a cyclic corrosion test used in the automotive industry, which consists of temperature variations between 35 and 45 °C and humidity between 50 and 95% RH. As shown in Figure 1, which corresponds to one week of the test, there are two pollution phases by a salt solution (NaCl 1 wt% at pH 4 with a deposition rate of 15 mm/h). One pollution phase consists of 3 salt rains of 15 min. The duration of the test was 12 weeks.

2.3.2. Accelerated Corrosion Test-Design of Experiment (DoE)

A design of experiments (DoE) was used to study the influence of the salt concentration (SC), the wet time (WT), and the wet relative humidity (WRH). The choice of salt concentration was indeed supported by the conclusions of Guérin et al. [30] and Vera et al. [31], as well as Ambat et al., who proved that the corrosion rate of aluminum increased when the chloride concentration increased [32]. According to Bautista and Funke, the mode of degradation (filiform or blister) would be dependent on the relative humidity. Indeed, a relative humidity value of 80–85% is optimal for the development of filiform corrosion [24,27]. It is also widely accepted that the wet time will have an important effect on the corrosion of aluminum alloys.

Based on the results from standardized accelerated corrosion tests and on the parameters selected in

Table 3. Variable parameters of the DoE.

	Level −	Level +
Salt (NaCl) Concentration, %wt (SC)	0.5	2
Wet Time, h (WT)	4	8
Wet Relative humidity, % (WRH)	80	95

, eight testing conditions were designed. All tests were based on the VCS 1027,149 presented previously, for a duration of 12 weeks. From the results obtained on the first eight tests (from the DoE), four new tests were performed to study the effect of the temperature (constant or variable), the type of salt, or to mimic the behavior of the different systems according to the experience acquired on them as presented in

Table 4. Tests matrix.

Test	Test Label	Salt Conc. Wt%	Type of Salt	Wet Time, h	Wet RH, %	Temperature, °C
		SC		WT	WRH	T
1	SC0.5/WT4/WRH80	0.5	NaCl	4	80	35–45 °C
2	SC2/WT4/WRH80	2		4	80
3	SC0.5/WT8/WRH80	0.5		8	80
4	SC2/WT8/WRH80	2		8	80
5	SC0.5/WT4/WRH95	0.5		4	95
6	SC2/WT4/WRH95	2		4	95
7	SC0.5/WT8/WRH95	0.5		8	95
8	SC2/WT8/WRH95	2		8	95
9	SC0.5/WT4/WRH80/T45	0.5		4	80	45 °C (Constant)
10	SC0.5mix/WT4/WRH80	0.5	NaCl + MgCl2 + CaCl2	4	80	35–45 °C
11	SC1/WT4/WRH80	1	NaCl	4	80
12	SC0.75/WT4/WRH95	0.75		4	95

. The schematic of the two tests included in Table 4 is presented in Figure 2. All tests were performed in controlled climatic chambers equipped with salt rain deposition.

2.4. Evaluation Procedure

The extent of underpaint corrosion was measured using a microscope and image analysis (Lucia). The number of initiations along the scribe, the maximum length of underpaint corrosion (blister or filament) perpendicular to the scribe, and the total area of underpaint corrosion have been determined. Only the area of underpaint corrosion is reported in this study, as the area of underpaint corrosion is probably the most interesting parameter. Indeed, it is indicative of the length, width, and number of filaments (or blistering) formed along the scribe.

2.5. Statistical Analysis of the Results

By comparing the results from laboratory tests against the results obtained in the field, the degree of acceleration of the test can be determined. It is possible to calculate the acceleration factor (Acc) according to Equation (1):

$$\mathrm{Acc}=\mathrm{Const}{\displaystyle \frac{\mathrm{X} (\mathrm{a}\mathrm{c}\mathrm{c}.\mathrm{t}\mathrm{e}\mathrm{s}\mathrm{t})}{\mathrm{X} \left(\mathrm{f}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}\right)}}$$

(1)

where x denotes the corrosion extent (corrosion maximum attack depth…), and the constant (Const) is determined by dividing the exposure time in the field by the exposure time in the accelerated test.

An accelerated test should accelerate the corrosion reactions on all materials by the same amount relative to the results obtained under field exposures. As the standard deviation increases as the mean increases, it can make the comparison of tests at different severity levels very difficult. To avoid that, it is possible to calculate the relative variation, or the coefficient of variation, which corresponds to the standard deviation divided by the mean. In fact, this value is the standard deviation as a percentage of the mean (Equation (2)).

$$\mathrm{Coefficient} \mathrm{of} \mathrm{variation}={\displaystyle \frac{\mathrm{S}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d} \mathrm{d}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n} \left(\mathrm{A}\mathrm{c}\mathrm{c}\right)}}$$

(2)

where Acc corresponds to the acceleration factor. The best-accelerated corrosion test will be defined by a low coefficient of variation and a sufficient acceleration factor.

3. Results

3.1. Standardized Accelerated Corrosion Tests

Figure 3 presents the total area of underpaint corrosion measured on painted aluminum alloys as a function of the three standardized accelerated corrosion tests.

The data indicated that extremely low propagation of underpaint corrosion occurred after 3000 h of NSST. This obviously underlined the fact that such tests cannot be used to discriminate aerospace materials. Nevertheless, system 2, which shows the largest extent of underpaint corrosion, is very likely sensitive to high humidity as NSST provides permanent wet conditions. This system also behaved quite poorly after the other corrosion tests. Yet, in some cases, such as for systems 3, 4, and 5, the ISO 4623-2 test was the most aggressive test. As further inferred from Figure 3, systems involving a Cr-free primer (1 and 2) or sol–gel system number 7 seem to be the least resistant ones.

In conclusion, clad and sol–gel systems (4 and 7) and Cr-free primer (1 and 2) showed the lowest resistance to underpaint corrosion.

Based on the guide of Kandell et al. [22], underpaint corrosion can be classified into four types of attack morphologies; for example, very thin filaments, wide corrosion filaments, mostly blistering and scattered filiform, and blistering, as illustrated in Figure 4. As noticed by Kandell et al., filaments are observed for a relative humidity varying between 50 and 80%. Blistering generally occurs at higher humidities.

As shown in

Table 5. Type of attacks from the scribes observed on painted materials after standardized accelerated corrosion tests.

	Mostly Filiform		Mostly Blistering
	Very Thin Filaments	Wide Corrosion Filaments	Mostly Blisters + Scattered Filiform	Blistering
ISO 4623-2 (1000 h)	4, 7	1, 2, 6, 8		3, 5
NSST (3000 h)				1, 2, 3, 4, 7, 8
VCS 1027, 149 (2000 h)	4, 7		2	1, 3, 5, 8

, NSST generated, as expected, only blistering since this type of corrosion occurs at very high humidity [25], while filiform corrosion was mostly detected in ISO 4623-2. Sol–gel systems showed mostly filiform corrosion in VCS 1027,149, while the other systems generated blistering corrosion. It should be noted that the type of corrosion observed on all systems was test-dependent.

3.2. Accelerated Corrosion Tests—Design of Experiment (DoE)

Figure 5 presents the total area of underpaint corrosion measured on painted systems as a function of the different accelerated corrosion tests obtained in the DoE matrix. The results from the VCS 1027, 149 test are also reported. As observed in standardized tests, systems 7, 1, 2 and, to a lesser extent, system 4 presented a lower resistance to underpaint corrosion independently on the corrosion test. Thus, systems including sol–gel (e.g., 4 and 7) and Cr-free primer (e.g., 1 and 2) were obviously more prone to underpaint corrosion than the other systems.

Table 6. Types of attacks from the scribes observed on painted samples in all accelerated corrosion tests (from the DoE).

	Mostly Filiform		Mostly Blistering
	Very Thin Filaments	Wide Corrosion Filaments	Mostly Blisters + Scattered Filiform	Blistering
SC0.5/WT4/WRH80	4, 7	1	2	3, 5, 6, 8
SC0.5/WT4/WRH80/T45	4, 7	1	2, 5	3, 6, 8
SC0.5mix/WT4/WRH80	4, 7		1, 2, 5	3, 8
SC1/WT4/WRH80	4, 7	1	2, 8	3, 5, 6
SC2/WT4/WRH80	4, 7	1	2	3, 5, 6, 8
SC0.5/WT8/WRH80	4, 7	1, 2		3, 6, 8
SC2/WT8/WRH80	4, 7		1, 2, 8	3, 5, 6
SC0.5/WT4/WRH95	4, 7		1, 2, 8	3, 5, 6
SC0.75/WT4/WRH95	4, 7		1, 2, 8	3, 5, 6
SC2/WT4/WRH95	4, 7	1	2, 8	3, 5, 6
SC0.5/WT8/WRH95	4, 7		1, 2, 5, 6	3, 8
SC2/WT8/WRH95	4, 7		1, 2, 8	3, 6

summarizes the type of attack observed on painted systems in all tests according to the guide of Kandell [25]. From Table 6, it appears that systems 4 and 7 (sol–gel systems) exhibited very thin filaments, whatever the tests, while blisters are mostly observed on systems 3, 5, 6, and 8. Other systems seem to be dependent on test conditions. Indeed, systems 1, 2, 5, and 6 showed an important proportion of blisters with the increase in humidity (comparison between SC0.5/WT8/WRH95 and SC0.5/WT8/WRH80). Overall, the type of corrosion is dependent on the system and test conditions. It is the interaction of the three parameters that influences corrosion morphologies rather than any single parameter.

3.3. Field Exposures

As illustrated in Figure 6, which presents the total area of underpaint corrosion from the defects after 5 years of exposure in the marine atmosphere, only minor degradations were observed. The values are low compared to results obtained in developed accelerated corrosion tests.

Nevertheless, filiform corrosion was observed on systems 4 and 7 and blistering on systems 1, 2, 3, 6, and 8, as summarized in

Table 7. Types of attacks from the scribes observed on painted samples in field exposures.

	Mostly Filiform		Mostly Blistering
	Very Thin Filaments	Wide Corrosion Filaments	Mostly Blisters + Scattered Filiform	Blistering
Field exposure (5 years)	4, 7		5	1, 2, 3, 6, 8

. Both types of corrosion were detected in system 5.

It is interesting to note that the trend observed in field exposures is in good agreement with accelerated corrosion tests for the few systems that exhibited some degradations.

4. Discussion

In this section, the accelerated corrosion tests from the matrix are compared with 5 years of outdoor exposure in the marine atmosphere. It shall be emphasized that the comparison shall be carefully considered as the extent of degradation was low after 5 years of exposure.

Morphology

Table 8. Summary of the type of attacks observed on painted samples in accelerated corrosion tests (Standard and DoE) and 5 years in the field.

	Duration	Mostly Filiform		Mostly Blistering
		Very Thin Filaments	Wide Corrosion Filaments	Mostly Blisters + Scattered Filiform	Blistering
ISO 4623-2	1000 h	4, 7	1, 2, 6, 8		3, 5
NSST	3000 h				1, 2, 3, 4, 7, 8
VCS 1027,149	2000 h	4, 7		2	1, 3, 5, 8
SC0.5/WT4/WRH80	2000 h	4, 7	1	2	3, 5, 6, 8
SC0.5/WT4/WRH80/T45	2000 h	4, 7	1	2, 5	3, 6, 8
SC0.5mix/WT4/WRH80	2000 h	4, 7		1, 2, 5	3, 8
SC1/WT4/WRH80	2000 h	4, 7	1	2, 8	3, 5, 6
SC2/WT4/WRH80	2000 h	4, 7	1	2	3, 5, 6, 8
SC0.5/WT8/WRH80	2000 h	4, 7	1, 2		3, 6, 8
SC2/WT8/WRH80	2000 h	4, 7		1, 2, 8	3, 5, 6
SC0.5/WT4/WRH95	2000 h	4, 7		1, 2, 8	3, 5, 6
SC0.75/WT4/WRH95	2000 h	4, 7		1, 2, 8	3, 5, 6
SC2/WT4/WRH95	2000 h	4, 7	1	2, 8	3, 5, 6
SC0.5/WT8/WRH95	2000 h	4, 7		1, 2, 5, 6	3, 8
SC2/WT8/WRH95	2000 h	4, 7		1, 2, 8	3, 6
Field exposure	5 years	4, 7		5	1, 2, 3, 6, 8

compares the aspect of underpaint corrosion in the accelerated corrosion tests (Standard and DoE) and in outdoor exposures.

Table 8 emphasizes the differences in failure modes observed when comparing NSST and ISO 4623-2 with the accelerated corrosion tests.

As shown in Table 8, only blistering was generated by NSST, while filiform corrosion was observed on a few systems in field exposure. This confirms that NSST is not representative of the field. The same conclusion can be made with ISO 4623-2, where filiform degradation was more accentuated than in field exposure.

The modes of degradation observed on systems 4 and 7 are system-dependent. Only filiform corrosion was noted on these two systems, regardless of the exposure conditions (accelerated tests or field).

Cyclic tests generally give more agreement with the mode of degradation observed at St-Anne station, with some variations depending on the test conditions.

The acceleration factors of the different tests with reference to marine exposure (5 years) were calculated from the maximum of the total area of underpaint corrosion. As shown in

Table 9. Coefficient of variation based on maximum values.

	Duration	Coefficient of Variation	Acceleration Factor
ISO 4623-2	1000 h	0.9	1566
NSST	3000 h	1.7	33
VCS 1027,149	2000 h	1.1	605
SC0.5/WT4/WRH80	2000 h	0.7	572
SC0.5/WT4/WRH80/T45	2000 h	1.0	473
SC0.5mix/WT4/WRH80	2000 h	1.1	169
SC1/WT4/WRH80	2000 h	0.8	1565
SC2/WT4/WRH80	2000 h	0.8	595
SC0.5/WT8/WRH80	2000 h	0.9	316
SC2/WT8/WRH80	2000 h	1.0	1220
SC0.5/WT4/WRH95	2000 h	1.3	661
SC0.75/WT4/WRH95	2000 h	0.4	496
SC2/WT4/WRH95	2000 h	0.6	703
SC0.5/WT8/WRH95	2000 h	0.4	487
SC2/WT8/WRH95	2000 h	1.2	695

, a group of accelerated corrosion tests composed of SC0.75/WT4/WRH95 and SC0.5/WT8/WRH95 gave a reasonable correlation with a coefficient of variation around 0.4.

5. Conclusions

From the results, it appears that the modes of failure observed in NSST and ISO 4623-2 are not representative of field exposures with a very high coefficient of variation and a low acceleration factor for the NSST.

Two cyclic corrosion tests (SC0.75/WT4/WRH95, SC0.5/WT8/WRH95) demonstrated interesting results, with failure modes similar to those observed in field exposures. This could be explained by the high relative humidity encountered in these two cyclic tests and known to generate blistering corrosion as it was observed in field exposure in Brest.

The developed cyclic corrosion tests also show a low coefficient of variation with sufficient acceleration factor, making them possible candidates to replace NSST and ISO 4623-2 for the qualification of aerospace materials.