Perhaps the most widely recognized accelerated corrosion test is the ASTM B117 salt fog test. However, numerous studies have demonstrated that this test method has poor correlation to outdoor exposures, particularly for non-chromate primers. As a result, more realistic cyclic environmental exposures have been developed to more closely resemble actual atmospheric corrosion damage. Several existing tests correlate well with the outdoor performance of some materials and assemblies of interest, but not all. The risk, then, is that promising new environmentally friendly technologies may be incorrectly rejected based on corrosion processes observed in flawed laboratory tests.

The development of accurate laboratory corrosion tests requires a basic understanding of the relationship between specific environmental variables and specific modes of corrosion failure. This requires fundamental studies to relate simple environmental parameters to actual conditions at the surface of a material and how those conditions specifically effect corrosion and protection. The overall goal of this project was to develop an improved dynamic accelerated corrosion test environment that more accurately predicts material system corrosion behavior in operational environments. Throughout this program, the research team assessed the relationship between environmental parameters and the resulting corrosion modes of various material systems. The corrosion of metals (boldly exposed, in occluded sites, and in galvanic couples), coating adhesion, and mechanical properties of metals as a function of environmental parameters (including humidity cycling, chemistry, and temperature) were assessed. These findings were then used to identify controlling factors for different damage modes. The relationships between exposure parameters and the resulting corrosion damage were then used to develop an accelerated test method that results in corrosion damage more similar to that observed in outdoor exposures than previous tests.

The single most relevant factor in governing atmospheric corrosion is the relative humidity (RH). Significant effort has been expended throughout the course of this work to define how RH and cyclic variations in RH affect corrosion rates and corrosion modes. On bare carbon steel surfaces, it was shown that corrosion can continue at RH values as low as 10% for some deposited salt chemistries. However, it should be noted that below 50% RH the corrosion rate is “very low” according to the ISO standard 9223. Multi-Electrode Array (MEA) studies of steel and aluminum under cyclic RH conditions were also performed. Real time MEA data revealed that under cyclic RH conditions, spikes in corrosion rate occurred during wetting and drying cycles. These spikes were also observed for aluminum systems indicating the importance of cyclic RH.

The effect of cyclic RH conditions on a creviced steel surface and galvanic couples were also examined. Under conditions of high RH, the surface outside of the crevice became totally cathodic such that all anodic currents were within the occluded crevice. Upon drying, the internal and external portions of the crevice decoupled and corrosion current was observed external to the crevice mouth. This coupling and decoupling effect was also observed for galvanic couples between steel and aluminum. Specifically, coupling was seen above the deliquescence relative humidity (DRH), the point at which deposited salt absorbs sufficient water from the atmosphere to dissolve. Metallographic analysis of galvanic assemblies showed that intergranular corrosion, pitting, and fissure formation within the bolt hole of a galvanic fastener occurred only under conditions where RH was cycled. These findings suggest that for values of RH above 50%, there are at least two regions of RH that are important. The regions are divided at the DRH. Above the DRH, thick electrolyte layers are present, allowing long-range electrochemical interactions such as coupling of galvanic materials and crevice regions. Below the DRH, it is likely that as drying begins, the thin electrolyte film begins to break into smaller islands. While corrosion can still occur under these isolated electrolyte islands, long range electrochemical coupling forces can no longer effect corrosion. 

As with bare metal and metal assemblies, cyclic variation in RH plays a key role in driving degradation of coated samples. In an attempt to better understand the role of the number of cycles and time of wetness on adhesion, further testing was performed using RH cycles of varying lengths and proportions of high RH to low RH periods. Sensor data as well as image analysis suggest that coating delamination is a strong function of the time of wetness within a given cycle. Therefore, adhesion loss is less a function of the number of cycles than of corrosion rate.

A drying step is required to solidify corrosion product and lift the coating from the surface. Additionally, extended periods in the intermediate RH range resulted in the appearance of exfoliation corrosion. The duty cycle for accelerating delamination would include relatively long exposure in the wet portion of a cycle. For accelerating corrosion after delamination, a short drying cycle would be appropriate.

RH duty cycle plays a strong role in both coating adhesion and corrosion rate at a coating defect. MEA data were used to demonstrate the effect of coating inhibitors on the corrosion current distribution. The addition of Class C or Class N primer resulted in the localization of galvanic attack to the vicinity of the galvanic interface at high RH values. Aluminum attack was more distributed at RH values below DRH. As with the uncoated samples, decoupling of a galvanic couple occurs below the DRH.

Similar to coating delamination and galvanic decoupling, cracking is enhanced under conditions of cyclic RH. However, an increase in crack growth is only observed under conditions where the RH is falling. Test solution chemistry was found to be an important factor in determining crack growth rates. Specifically, the presence of sulfate species in the test solution was found to drastically increase observed crack growth rates. The degree of coating delamination and cracking can be controlled by the amount of time in the RH range between 50% RH and the DRH of the deposited salt.

Atmospheric chemistry is another major factor in governing corrosion rates. Salt speciation will determine the value of DRH and salt composition will have an effect on specific corrosion and cracking rates. Likewise, salt loading density will directly affect electrolyte formation geometry and overall material coverage. Based on work in this effort it was observed that the composition of salts that are found on sample assemblies (both in laboratory environments and outdoors) mimic the exposure conditions. In other words, in coastal environments, deposit composition closely resembles that of seawater while in accelerated lab tests, the composition closely resembles that of the spray solution.

Nevertheless, slight differences in composition are observed. For example, samples exposed at Pt. Judith and Daytona Beach had slightly reduced amounts of chloride, sulfate and nitrate indicating their reactiveness to the exposed surfaces. Thus, these ionic constituents are important in the corrosion reactions observed. After three weeks of exposure to laboratory environments, the maximum amount of salt loading is fairly similar for ASTM B117, ASTM G85-A5 and ASTM G85-A4. In contrast, loading levels for the GM9540P test are nearly a factor of 10 or more lower. This dilution is consistent with the observed reduction in corrosion rate.

Measurements were also made on samples exposed to outdoor environments for a period of one year. It was observed that Pt. Judith had the highest salt content. Examination of in-situ corrosion rate data, short term mass loss data, and visual observations reveal that the highest damage rates were observed at Pt. Judith. Further examination of sensor data reveals that an increase in cumulative corrosion coincides with increases in time of wetness and conductance. Conductance data relates directly to the salt loading observed, revealing that the chemical environment at Pt. Judith is more aggressive than the other ambient locations. Of particular interest is that the salt load loads for LAX are high, but the damage observed after long term exposure is mild. Examination of the data reveals, however, that the cumulative time of wetness is relatively low at LAX. Thus despite the relatively high salt loading, the conductance of the electrolyte layer remains low, resulting in a lower corrosion rate and demonstrating the importance of RH.

Due to the importance of RH levels and cycling, sensors and software programs were used to define appropriate time intervals within each RH range. Sensors were used to continuously measure RH at the outdoor exposure sites for several months. The RH levels were binned in four categories: (RH < 50%, rising RH with 50% < RH < DRH, falling RH with 50% < RH < DRH, and RH > DRH. In this case, it was assumed that NaCl dominated chemistry and so 76% RH was used as the DRH value. From this data, the percentage of time spent in each bin was calculated and normalized to the total exposure time. Results of this binning indicate that the severity of attack is not simply a function of percentage of time in each RH range but also the absolute amount of time in each range in each cycle.

Using the findings described above, an accelerated test protocol was developed and written. The test protocol provides detailed information regarding the testing apparatus, sample preparation, solution preparation, testing procedure, as well as post-test inspections and reporting. It was written in the format of the standard ASTM tests to facilitate easy discussion and adoption. Initial testing revealed that the resulting exposure environments were more mild than anticipated and limited corrosion was observed. Based on RH values achieved during testing and the percentage of time in each RH range, it was expected that exposure conditions would result in damage similar to what is observed at LAX. Photographs of the panels after testing revealed this to be the case. The mild exposures were primarily a function of test chambers performing in an unexpected manner. Differences in chamber performance between team member laboratories have been noted throughout this effort. Differences in high and low RH values as well as ramp rates varied significantly between labs despite efforts to reconcile differences. In order to achieve desired exposure conditions, rigorous control of relative humidity is required and must be called out in test specifications. This will require the use of more sophisticated test equipment but will ensure lab to lab reproducibility and the creation of damage states observed in the field.