The Effect of Variable Humidity on Corrosion Fatigue of AA7085-T7451 with Surface Salt Deposits

doi:10.21203/rs.3.rs-4681592/v1

Download PDF

Article

The Effect of Variable Humidity on Corrosion Fatigue of AA7085-T7451 with Surface Salt Deposits

https://doi.org/10.21203/rs.3.rs-4681592/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Corrosion fatigue (CF) crack growth is quantified as a function of relative humidity (RH) using AA7085-T7451 samples with NaCl deposited to understand the effect of deliquesced surface electrolyte droplets on CF performance when humidity varies. Fracture mechanics testing holding mechanical driving forces for cracking constant and incrementally increasing humidity show that crack growth rate (da/dN) more than doubles once RH moves above 78–80% RH. When decreasing RH, some amount of drying below the efflorescence RH (ERH) and/or a sufficient time is needed to pass in order for da/dN to return to that expected for a dry crack tip. All in all, this study establishes time of crack tip wetness as an important parameter for predicting fatigue lifetimes in atmospheric conditions, a parameter that cannot be solely predicted by RH, as accelerated cracking can continue for a limited amount of time even when RH is brought below the ERH.

Physical sciences/Materials science/Structural materials/Metals and alloys

Physical sciences/Materials science

Corrosion fatigue (CF) is a failure mode in age-hardenable Al alloy aerospace components [1, 2] caused by the simultaneous interaction between cyclic mechanical stress and an aggressive environment. An understanding of the factors influencing CF in atmospheric environments is of particular relevance to aerospace design and life prediction, as CF can cause accelerated fatigue crack growth rates (da/dN) and reduce the threshold above which cracking can occur. The literature on CF aluminum alloys has extensively examined in air of various humidity levels without deposited surface salt, [3–28] pure water vapor, [3, 4, 8, 24, 26, 29–39] and aqueous chloride environments [10–13, 16–19, 22, 24, 27, 40–51]. While these test environments are useful in establishing a susceptibility to CF in humid air or chloride containing environments, they may not provide an adequate analog to aerospace service environments as aircraft are exposed simultaneously to temperature changes, humidity changes, ultraviolet light, and ozone all while in contact with contaminant aerosols—a combination which may affect CF crack growth over conventional laboratory test environments.

When exposed to moist air, contaminant aerosols containing salts can deliquesce to yield surface electrolyte droplets [52, 53]. When these electrolyte droplets are wicked into a crack, they have been shown to lead to increased da/dN over that expected from air alone [21, 54–57]. Specifically, previous work by the authors showed that da/dN in AA7085-T7451 with 300 µg/cm² of NaCl deposited and exposed to 80% RH can be up to 5× higher than in humid air without surface salt [57]. This atmospheric environment was also shown to match or exceed da/dN measured while fully immersed in an equivalent salinity bulk aqueous environment[57]. Additionally, spontaneous fluctuations in da/dN have been measured when surface electrolyte droplets are present [56–59]. Because all mechanical and environmental parameters were held constant, it was suggested that these fluctuations may be linked to changes in crack electrolyte volume as the crack grows and wicks electrolyte into the crack. RH may be a key factor for in determining CF behavior in atmospheric environments, as it has direct control over whether the surface is dry or wet and many key electrolyte properties including the chloride concentration and droplet volume based on the salt load present on the surface [56–59]. This in turn, determines other key electrolyte properties including density, conductivity, oxygen solubility, and oxygen diffusivity, all that may affect cathodic kinetics and subsequently da/dN[60, 61].

Additionally essential may be understanding time-based RH changes and the subsequent transient effect on da/dN as humidity through a flight cycle will change dramatically between take off and flight at altitude. One study examined wet/dry cycling for AA7075-T651 with contaminant salt on the surface and found during a drying cycle, a temporary increase in da/dN was observed just prior to reaching the efflorescence RH (ERH – the RH at and below which salt is no longer hydrated and the surface is considered dry) [21]. Similar phenomena have been observed during stress corrosion cracking testing on a 5xxx series alloy under cyclic RH, ASTM B117, and other standardized accelerated test environments [62]. These tests found crack velocity was fastest during the drying phase of these tests when RH is decreased from above to below the deliquescence RH (DRH – the RH at and above which salt becomes hydrated and the surface is considered wet). Crack growth that occurs between the DRH and ERH as RH is decreased may be of particular concern as resulting higher chloride concentration and electrolyte layer thinning allows for increased oxygen availability.

The objective of this research is to expand upon the static RH considerations of previous work [56–59] and to understand how transient RH may alter CF crack growth kinetics when NaCl is deposited on the surface. To achieve this objective, fracture mechanics based CF studies were performed on AA7085-T7451 with 300 µg/cm² of NaCl deposited on the surface, which allows direct comparison back to previous work [57, 58]. In one approach, the RH was manually changed in a stair step fashion to determine average da/dN at a given RH. In a second approach, humidity was continuously changed at different rates to determine the effects of transient RH during a wet-dry cycle.

3.1 Establishing Baseline Performance in Conventional Environments

As the results presented in Fig. 1 show, baseline fatigue performance was assessed in conventional environment of air, without salt, at various humidity levels and while immersed in 0.06 M or 23.1 wt.% (4.62 M) NaCl solutions. These results are also discussed in detail in previous publications by the authors [57, 58]. Below 1 Hz, data collected in 0.06 M NaCl (▲) showed a maximum, f-independent da/dN. Above 1 Hz, da/dN decreased with increasing f. This behavior is consistent with trends for other 7xxx series alloys [41, 43]. Between 0.01–0.03 Hz(data points denoted with a “*” in Fig. 1) specimens likely experienced corrosion product induced crack closure leading to a falsely low da/dN as has been reported for other 7xxx series alloys [41, 43]. The higher salinity environment of 23.1 wt.% NaCl was chosen to have a common chloride concentration with equilibrium NaCl electrolyte droplets at 80% RH. Data collected in this environment (◆ symbols in Fig. 1) exhibited lower da/dN than 0.06 M NaCl when f was above 0.3 Hz. Both full immersion environments tested showed significantly accelerated da/dN over air at any RH.

3.2 Humidity Testing using Constant RH Segments

Two trials (with two separate samples) of humidity testing were conducted where RH was held constant for a period of crack extension and then incrementally changed while holding all other environmental and mechanical parameters constant. The average da/dN measured during each RH segment is plotted as a function of RH for both Trial 1 (■) and Trial 2 (○) in Fig. 2. Arrows are shown to indicate the order of testing. Note that several humidity levels exhibited multiple stable da/dN and are represented by two data points connected by a vertical arrow. Comparison data taken from Fig. 1 for similar conditions with no surface salt applied (× symbols) and fully immersed in 0.06 M NaCl (average denoted by a right arrow) are shown. For both trials, testing began at 45% RH where a da/dN equivalent to air with no salt was measured (≈ 2 × 10^− 4 mm/cyc). RH was then increased to 76% RH (near the deliquescence RH for NaCl) and increments of 2% RH thereafter. During this wetting phase both trials showed similar behavior, albeit at slightly different humidity levels. First, da/dN accelerated 1.5× up to ≈ 3.5 × 10^− 4 mm/cyc upon increasing humidity between 76–80% RH. Following this initial acceleration, da/dN roughly doubles to ≈ 7 × 10^− 4 mm/cyc. Further increases in humidity resulted in da/dN values ranging between ≈ 7 × 10^− 4 mm/cyc and ≈ 1 × 10^− 3 mm/cyc, which is equivalent to that measured in 0.06 M NaCl (right arrow in Fig. 2). After the wetting phase, humidity was decreased to 76% RH, where da/dN remained high in both trials. For Trial 1 (■) only, RH was then decreased to 60% where da/dN fell over a 9 minute period to a value near that measured at 45% RH as will be discussed further below.

An overview of a–N for all of Trial 1 is shown in Fig. 3a. Vertical dashed lines demarcate where the humidity level changed. While increasing humidity, da/dN did not increase dramatically until 80% RH where da/dN changed from 3.7 × 10^− 4 mm/cyc to 7.1 × 10^− 4 mm/cyc after 50 minutes at 80% RH without changing any environmental or mechanical parameters. A more detailed view of the 80% RH segment is shown in Fig. 3b. After this increase in da/dN, crack growth stayed relatively consistent for the remainder of the test until dropping to 2.4 × 10^− 4 mm/cyc 9 minutes after decreasing RH to 60% as detailed in Fig. 3c.

Trial 2 is plotted similarly in Fig. 4. For this trial, da/dN at 76% RH was only slightly higher than that measured at the previous RH of 45%; and unlike Trial 1, da/dN accelerated to 7.2 × 10^− 4 mm/cyc once RH was changed to 78%, as shown in Fig. 4b. Following this, da/dN was relatively stable until reaching 84% RH, where da/dN changed spontaneously without changing experimental parameters from 9.9 × 10^− 4 to 7.0 × 10^− 4 mm/cyc (an almost 30% decrease) as shown in Fig. 4c.

It is important to note there were some differences in the testing environment between these two trials. First, Trial 1 had longer holds at each individual RH than Trial 2 (during Trial 1, 142 minutes elapsed between 76 and 80% RH, as compared to 80 minutes for Trial 2). However, given that Trial 2 showed a more rapid increase in da/dN, it is unlikely the relatively slow increase in da/dN seen in Trial 1 was due to insufficient hold times at each RH. Additionally, there was a 2–3°C difference in ambient temperature between the two tests. Because Trial 1 was warmer than Trial 2, Trial 1 may have been warmed to a degree that it encouraged partial surface electrolyte drying, preventing an adequate volume of electrolyte wicking into the crack tip until a higher RH. Lastly, the method of salt deposition varied between the two tests, with Trial 1 being printed, and Trial 2 being pipetted; however previous work by the authors has shown no measurable difference in da/dN between salt printing and salt pipetting [57, 58].

3.3 Humidity Testing with Continuously Changing RH

Experiments that continuously varied humidity between 85% and 25% RH were conducted to determine the effects of transient RH exposures, which are more representative of in-service conditions, on CF performance. These experiments used one of two ramp rates with the slower having a 60-minute ramp while the faster used a 15-minute ramp. Both ramp rates used a 25-minute hold time at the minimum and maximum RH before ramping was resumed. Results for each trial with continuously changing RH are shown in Fig. 5 and Fig. 6 for the slow and fast ramp rate, respectively. The corresponding humidity vs. time, da/dN vs. time, and crack extension vs. cycle count behavior are shown. For results using the slower ramp rate (Fig. 5), each cycle was conducted on a separate sample, while the faster ramp rate (Fig. 6) allowed for two cycles to be performed using the same sample.

Two cycles with the slower ramp rate were completed, the first of which is show in Figs. 5a-c. RH vs. time is plotted in Fig. 5a with shaded regions showing when RH was changed. Humidity was decreased from 85% and 25% RH over the course of 60 minutes, resulting in a large decrease in the instantaneous da/dN plotted in Fig. 5b. The decrease in da/dN is also evident in the crack size vs. cycle count (a-N) data plotted in Fig. 5c where linear regression was taken to find the average da/dN over several regions that showed relatively constant crack growth. During the drying phase for the first trial, da/dN averaged 9.26 × 10^− 4 mm/cyc before dropping rapidly to 1.88 × 10^− 4 mm/cyc as shown in Fig. 5c. The inflection point in the a-N data occurred at ≈ 50% RH, suggesting that was the RH needed to cause a significant slowing in crack growth. After the drying stage, humidity was held constant at 25% RH for 25 minutes and no large changes in da/dN were observed. Subsequently, humidity was increased back to 85% RH over the course of 60 minutes. Starting at the end of the ramping/wetting phase (the 2nd shaded region in Fig. 5b), increases in da/dN were recorded until reaching a da/dN in the range expected for a high RH environment with surface salt deposits. The a-N data recorded in Fig. 5c shows the average da/dN increased to 6.81 × 10^− 4 mm/cyc and the inflection point where this increase occurred corresponded to ≈ 85% RH. Another cycle of slower ramp rate testing was completed using a different sample as shown Figs. 5d-f which yielded similar results to the first cycle. For all of the fitted lines shown in Fig. 5c and f, \(\:{R}^{2}\) values were 0.97 or better.

Resutls for the fast ramp rate (15-minute ramp, and 25-minute hold) are shown in Fig. 6. Overall, there was good agreement seen between the two back-to-back humidity cycles using the fast ramp rate. While this test showed similar trends to the slow ramp rate experiments, there were a few key differences. First, the drop in da/dN associated with drying occurred at a lower RH. Examination of the a—N data shown in Fig. 6c shows the inflection point associated with drying occurs near the end of the drying stage, as compared to midway through the drying stage for the slower ram rate (Fig. 5c and f). The RH associated with the sharp decrease in crack growth rate was ≈ 35% RH for the fast ramp rate and ≈ 50% RH for the slow ramp rate. A similar delay was seen during the wetting phase of the humidity cycle, as return to rapid crack growth was not seen until after the end of the RH increasing stage and the RH had been held at 85% RH for ~ 10 minutes.

To better compare the two ramp rates, da/dN is shown plotted as a function of RH in Fig. 7 for all trials shown in Fig. 5 and Fig. 6. Arrows show the direction of testing starting at the top right of the plot. As shown, da/dN initially remains high (≈ 1 × 10^− 3 mm/cyc) as humidity decreases toward 50% RH. Following this, the RH at which da/dN begins to decrease is clearly dependent on the ramp rate. When RH is ramped quickly, a decrease in da/dN likely associated with drying of the crack electrolyte was not observed until ≈ 25% RH. By contrast, when RH is ramped slowly, da/dN began decreasing at ≈ 50% RH and dropped to the level expected in dry air before the end of the ramp down. When increasing RH to re-wet the sample, differences between the two ramp rate were more subtle—with both ramp rates showing some small increases in da/dN during the wetting stage, but larger increases to ≈ 8 × 10^− 4 mm/cyc did not occur until the start of the 85% RH hold. Note that because Fig. 7 is not plotted as a function of time, crack growth rate as a function of time during the low and high RH holds is not apparent. Plots in Figs. 5 and 6 show those parts of the experiment more clearly.

3.4 SEM Fractography

Scanning electron microscopy (SEM) showed that changes in humidity correlate with a change in fracture mode as well as changes in da/dN. Figure 8 shows a composite fractograph taken from the segmented humidity test in Fig. 4. Vertical lines are drawn showing the approximate locations where humidity was changed. As can be seen, 45% RH exposure yielded flat transgranular cracking, which is similar in appearance to that reported in previous work for low RH testing with a similarly low da/dN [57]. On the other extreme, 80% RH yielded a more faceted morphology similar to that found previously when testing in 80% RH with salt applied to the surface or fully immersed in 0.06 M NaCl. Due to the similarity in size of these facets to the typical grain size, this faceted morphology was attributed to either intergranular cracking or cracking along preferential planes. At 76% RH, the portions of the sample closest to the surface and bulk environment exhibited the faceted cracking morphology associated with higher RH, while the center exhibited the flat transgranular mode associated with a lower RH. The boundary between these two regions is indicated by the yellow line in Fig. 8. The center flat transgranular region narrows as the crack advances, and eventually converts completely to faceted cracking during the 78% RH segment. Changes in cracking mode were also observed when ramping humidity. An example of this is shown in Fig. 9, which shows the wet-dry-wet transition from Fig. 5c. In this case, the faceted morphology is again observed at higher RH, which was at the beginning and the end of the segment, while the lower RH from the dry phase resulted in flat transgranular cracking.

The results of this study demonstrate the effect of humidity on CF when contaminant salts forming surface electrolyte droplets are present. As will be discussed, large changes in da/dN are connected with increasing humidity above the deliquescence RH (DRH), which wets surface salts and wicks electrolyte into the crack, and decreasing humidity below the efflorescence RH (ERH), which dries the electrolyte within the crack. When given the time to equilibrate, da/dN was found to reach levels equivalent to static humidity exposures Additionally, results show that changes in da/dN associated with decreasing RH are affected by the rate of humidity change. The effects of transient humidity exposures during a wetting stage (RH increasing) and a drying stage (RH decreasing) will be discussed in the following subsections.

4.1 Understanding Crack Growth Rate Changes During the Wetting Process

4.1.1 While RH is Below 76%, the DRH

When increasing humidity from low levels and prior to reaching 76% RH, salt loaded samples continued to show da/dN similar to that measured for samples with no salt, which indicates there was no electrolyte within the crack to provide an environment to accelerate crack growth. This is supported by the observed flat transgranular cracking (45% RH region at left in Fig. 8 and center region in Fig. 9). In general, RH changes made while staying below the DRH only resulted in relatively minor changes in da/dN, such as the trend seen in Fig. 7 from ≈ 1.6 × 10^− 4 mm/cyc to ≈ 2.3 × 10^− 4 mm/cyc when increasing RH from 25–60%.

4.1.2 Once RH is at and Above the DRH

Once humidity is increased to the DRH, surface salts on the sample deliquesce and form electrolyte droplets which are intersected and wicked into the crack tip..The wicked electrolyte is then able to create a sufficiently aggressive environment to cause accelerated crack growth.SEM fractography suggests that this can be a gradual process. The boundary between flat transgranular cracking and faceted (possibly intergranular) cracking in Fig. 8 was curved with the fracture morphology changing first near the edge of the sample—suggesting electrolyte wicked in from the sample sides during the 76% RH segment and accelerated crack growth in only part of the crack front.As the humidity rose to 78% RH, the entire crack front eventually transitioned to the faceted morphology, likely indicative of increased electrolyte coverage within the crack. This is expected as increased humidity leads to a greater equilibrium electrolyte volume, and further crack advance would allow for more electrolyte droplets to be intersected and wicked into the crack.. This could explain the mild da/dN increase rom 1.97 to 3.38 × 10^− 4 mm/cyc when humidity increased to 76% RH, and the subsequent larger increase to 7.22 × 10^− 4 mm/cyc when humidity rose to 78% RH (Fig. 4). The first increase in da/dN may have been the result of wicking a small volume of electrolyte into the crack tip, but not enough to accelerate the entire crack front through the full thickness of the sample. As more electrolyte wicked in the crack tip, the entire crack front filled with electrolyte resulting in a change to the faceted cracking mode and higher da/dN.

Once the humidity is above the DRH, and enough electrolyte has wicked into the crack to accelerate da/dN in the range of 7–10 × 10^− 4 mm/cyc, there was no clear effect of further increasing humidity. As there are many electrolyte properties that are expected to change as RH changes, including chloride concentration, oxygen solubility, and conductivity all of which were identified by Chen et al. to play a role in cathodic current capacity for the case of atmospheric crevice corrosion [63], it may be expected that RH could alter da/dN even after full wetting and crack filling occurs. The effects of RH on these electrolyte properties were predicted using OLI software and are shown in Fig. 10. As can be seen, when RH is increased from 76–85% RH, chloride concentration decreases from 5.3 to 3.7 M (corresponding to an increase in electrolyte volume), oxygen solubility increases (0.048 to 0.080 mol/m³), and conductivity decreases (23 to 22 S/m). The net effect of these changes in electrolyte on corrosion kinetics, and by extension cracking kinetics is difficult to predict. Chen et al. calculated the maximum cathodic current for thin electrolyte films outside of a crevice and found that increased humidity would lead to an overall increase in cathodic current availability [60]. Chen attributed this to increased water layer thickness but did note that the effect of water layer thickness was balanced by a decrease in conductivity such that the overall effect of an increased RH was relatively small. Whether or not this change in cathodic kinetics would affect overall cracking kinetics is dependent on what the limiting factor for cracking is. According to the Hydrogen Environment Assisted Cracking model (HEAC, a common model used to explain corrosion fatigue related phenomena), increased da/dN can be attributed to hydrogen created via corrosion reactions at the crack tip, which is then absorbed in the material ahead of the crack tip and diffused into the fracture process zone[24, 42–45, 64–71]. If the rate of hydrogen diffusion is limiting, or if there is already sufficient hydrogen available to saturate the effect of hydrogen in the fracture process zone, then changes in cathodic kinetics would not be anticipated to cause a change to cracking kinetics. The data collected in the present study suggests that this may be the case as only minor fluctuations in da/dN were seen when RH was above the DRH and da/dN had stabilized. For example, in Fig. 4 da/dN fluctuated between 6.8 and 9.9 × 10^− 4 mm/cyc at high RH, but there was no clear increasing or decreasing trend when RH was increased to 84% or decreased to 76%. Research that examines effects of RH at even higher humidity levels may be able to shed further light on what effects, if any, RH has corrosion fatigue processes when surface salts are present. The experimental setup used in this work was able to examine the effects of RH up to ~ 85% RH, which results in a relatively modest change in solution conductivity from (XX to XX) as seen in Fig. 10c. If RH was further increased to 98% RH, the conductivity would drop further to XX.

In summary, the results shown here demonstrate that samples with surface salt deposits see increased crack growth rate following an increase in humidity after completing a multi-stage process. First, humidity must be elevated above the DRH to allow for electrolyte droplets to form on the surface of the sample. Next, the crack must intersect and wick in some of the electrolyte droplets. These first droplets may result in increased crack growth rate to some intermediate da/dN, potentially due to incomplete filling of the crack as evidenced by SEM fractography. Finally, sufficient electrolyte is wicked into the crack tip to yield a maximum environmental contribution to da/dN for the given environment.

4.2 Understanding crack growth rate changes during the drying process

Once a surface salt has deliquesced, humidity levels below the efflorescence RH (ERH) are needed for sufficient evaporation to dry the electrolyte. The ERH is typically lower than the DRH because of the energy associated with nucleation and growth of the solid phase [72]. For NaCl, ERH values where complete crystallization of electrolyte have been measured to occur in the range of 41–53% RH [73–77]. As such, it would be expected for da/dN to remain accelerated when decreasing RH from 76–50% RH until efflorescence occurs. After the RH is decreased below the ERH, a value which is difficult to measure and highly dependent on the substrate, drying due to evaporation is expected to take place allowing crack growth to return to slower da/dN consistent with dry air crack growth rates. As shown in Fig. 7, the RH at which drying occurs appears to vary with respect to the rate that RH was decreased. Using the fast ramp rate, the RH needed to cause drying was ≈ 25% RH. Experiments using the slower ramp rate began to show decreased da/dN at ≈ 50% RH. One test going into the drying stage was conducted using constant RH segments and showed a decreased da/dN associated with drying occur after 9 minutes at 60% (Fig. 3), which is above the expected ERH of 50% RH. All in all, the exact RH below which crack growth returns to the rate expected in air with no electrolyte is highly variable. Experiments showed that this critical RH level appeared to dependent on the rate which RH decreased, suggesting that either a large amount of under-drying (i.e. bringing RH well below the ERH), or allowing for sufficient time to pass at a more intermediate RH is needed in order to completely dry the crack tip. With the limited data available it is not possible to independently determine either the amount of under-drying needed or the time needed for drying to occur.

Special attention to crack growth rate during drying (when RH is below the DRH, but above the ~ ERH) should be given. In a study on AA7075-T651 with 400 µg/cm² of NaCl applied to the surface, researchers observed that a temporary increase in da/dN just prior to electrolyte drying [21]. This was attributed to there being more oxygen available for cathodic kinetics as the electrolyte layer thickness decreased, which would yield an overall increase rate of corrosion reactions that generate hydrogen in the crack. The authors also noted that potential for a more concentrated electrolyte as the drying process occurred which again could increase the rate of corrosion reactions. The limited testing completed in the present study did not find a clear trend in da/dN during the drying stage when RH was between 50 and 76% RH. Some testing did show a small increase in da/dN during drying (Fig. 5b), others showed no change (Fig. 5e), while others showed a small decrease in da/dN (Fig. 6b). These small changes in da/dN are not larger in magnitude than other fluctuations in da/dN seen when RH is being increased or being held at humidity levels above the DRH, so are not likely indicative of any phenomena unique to the drying stage.

In summary, the results shown here indicate that time of crack tip wetness is a key parameter for predicting fatigue life in atmospheric conditions as a wetted crack tip may show five-fold faster da/dN than a dry one. Unfortunately, the time for crack tip wetness cannot be readily determined using relative humidity data alone. In the present work, drying and subsequent decreases in da/dN were seen to occur anywhere between 25% and 60% RH, depending on the time allotted for drying to occur.. Further work is warranted to truly understand the drying process and understand connections between RH and time for drying. The critically important finding here is that accelerated da/dN can persist atRH below expected ERH values, which implies that accelerated da/dN can persist for a limited period of time when an aircraft is brought from low to high altitude where dry conditions are historically expected/used for prediction.

5.1 Material

A 115 mm hick plate of AA7085-T7451 manufactured by Arconic Inc. was used for this study. Material properties, microstructure morphologies, and baseline fatigue crack growth rates were previously reported on the same production lot of material [54, 57, 61, 70]. The yield strength (σ_YS) is 490 MPa and fracture toughness (K_IC) is 46 MPa√m. The grain size was 375 µm, 139 µm, and 66 µm in the longitudinal, transverse, and the short transverse directions, respectively. Samples were machined into L-T oriented eccentrically-loaded single edge tensile (ESE(T)) fracture mechanics specimens [78].

5.2 Salt deposition

NaCl was deposited on a 38 mm x 38 mm area on either side of the sample. The target load for this study was 300 µg/cm². For pipetting, a solution of ethanol saturated with NaCl is pipetted onto a warmed sample and allowed to evaporate following the same methodology as described in previous work [55–58]. Samples with pipetted salt received a nominal salt load density between 280 and 300 µg/cm², but actual salt loads likely differed. Salt printing was performed using a large format direct to substrate inkjet printer to print a solution of water, NaCl and ethanol using the same methodology as described in previous work [57, 79, 80]. Each batch of samples was validated by rinsing a test coupon with a fixed volume of DI water and performing conductivity measurements. The printed samples in this study received a salt load of 300–330 µg/cm².

5.3 Fatigue crack growth rate testing

ESE(T) samples were fatigue loaded on an MTS servo-hydraulic load frame operated in load control with the load applied in the L direction and crack propagation in the T direction, i.e. the L–T orientation. The ESE(T) samples had the following nominal dimensions: width (W) of 38.1 mm, thickness (B) of 3.2 mm, and saw-cut notch length (a_n) of 5 mm. The direct current potential difference (DCPD) method was used with Johnson’s equation to measure crack extension during testing [80]. The potential difference across the notch was normalized by a reference potential difference that was measured 25 mm above the crack mouth in order to mitigate any effects of temperature variation on the DCPD signal. The applied current was set to 4 A and was reversed at regular intervals to mitigate any thermoelectric effects. Testing was completed to measure da/dN as a function of RH using two different methods, as outlined below. All testing held ΔK and R constant at 9 MPa√m and 0.5, respectively, which were chosen in order to compare to previous work [57]. An f of 1 Hz was chosen to balance environmental effects with the time required to complete testing as extremely long testing at low frequency was found to cause an increased incidence of corrosion product induced crack closure complications.

After testing, samples were ultrasonically cleaned and optical and/or scanning electron microscopy was used to measure the final crack length. By assuming the error between the actual final crack length and the DCPD measured crack length accumulated linearly throughout the test, the actual ΔK applied throughout the test was determined. Data collected when the actual applied ΔK deviated from the target ΔK by more than 0.7 MPa√m were excluded and are not reported here. According to previous work in similar conditions, this is likely to yield a typical repeatability in measured crack growth rates of ± 20%.

5.4 Environmental control

RH was controlled using a digital pressure control valve to regulate airflow into a saturator using a method similar to previous work [21, 57, 58]. The flow rate was set to 1 SLPM which allows for complete replacement of the air in the environmental cell approximately every 30 seconds. ESE(T) samples were placed inside of an environmental cell that clamped onto the bottom of the sample using an O-ring. The top of the cell was sealed using a flexible mastic tape aside from a small exhaust port that was left open. Humidity was varied throughout testing using one of two methods as outlined below.

5.4.1 Humidity Testing using Constant RH Segments

RH was manually changed in a stair-step fashion with the goal of obtaining approximately 1 mm of linear crack growth before changing RH. When crack growth was observed to be fluctuating significantly, more time was allowed to pass with the goal of recording a stable da/dN. Crack growth rate was determined using a linear regression of crack extension as a function of cycle count. Following each segment, RH was incrementally stepped up or down. It is important to note that each constant RH segment was held for different lengths of time, as the primary goal of this testing was to maintain constant environmental conditions for long enough to ascertain an accurate da/dN. When da/dN is low, it takes a longer amount of time to reach a given distance of crack extension. Additionally, some testing showed fluctuations in da/dN while environmental conditions were held constant and constant RH segments were held for longer to better understand these fluctuations and generate a piecewise linear fit. Experiments conducted using this methodology were performed prior to the installation of a thermal control chamber which resulted in larger temperature variations between tests. Humidity was measured using an inline humidity probe positioned just prior to the environmental cell.

5.4.2 Humidity Testing with Continuously Changing RH

One of two humidity ramp profiles were used and are termed fast and slow ramp rate. Both profiles had a maximum humidity of 85% RH and a minimum humidity of 25% RH. The fast ramp rate used a 15-minute ramp time and the slower utilized a 60-minute ramp time. Both profiles contained a 25-minute hold at the maximum and minimum humidity values. All testing started with a sufficiently long high humidity (> 80% RH) hold to reach a stable, high da/dN after which the drying stage was started. Following testing, the incremental polynomial method was used to determine instantaneous da/dN as a function of RH. An incremental polynomial coefficient, n, of 10 was used as this was found to yield a good balance between decreasing noise and over-smoothing the data. Additionally, piecewise linear fits were used to determine the average da/dN values when crack growth was approximately linear. All testing using this methodology was temperature controlled to 22°C using a thermal chamber. Humidity was monitored and controlled using a humidity probe placed directly inside of the environmental cell to measure the RH as closely to the sample as possible and to minimize the effects of temperature fluctuations.

5.5 SEM Fractography

Scanning Electron Microscopy fractographic images were collected using a Thermo Scientific Apreo SEM. Images shown here were taken using an Everhart-Thornley Detector. In order to create high resolution images of a large area, multiple images were stitched together in order to create a composite image.

6 Data Availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the data forms part of an ongoing study. Should anyone wish to obtain the raw or processed data, please contact Prof. J. Locke at [email protected].

7 Acknowledgements

The research reported here was sponsored by the USAFA and The Ohio State University under agreement number FA700-18-2-0001. The U.S. Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. The opinions, findings, views, conclusions or recommendations contained herein are those of the authors and should not be interpreted as necessarily representing the official polices or endorsements, either expressed or implied of the USAFA and the U.S. Government. The authors are grateful for salt printing performed at Sandia National Laboratories by Jason Taylor. The authors would like to acknowledge Dr. Gary Bray, Dr. Lynne Karabin, and Greg Venema from Arconic Inc. for providing information about alloy properties. All electron microscopy was performed at the Center for Electron Microscopy and Analysis (CEMAS) at The Ohio State University.

8 Author Contributions

The study was conceived by J.S.L and S.G.D.. Experimental setup guidance was provided by E.S. and S.G.D.. Experimentation and data acqusition was done by B.F., G.C.M., and G.A.M. Data analysis and interpretation was done by B.F. and J.S.L. Writing and reviewing was done by B.F., J.S.L., and G.C.M.

9 Competing Interests

The authors declare no Competing Financial or Non Financial interests

S. Findlay, N. Harrison, Why aircraft fail, Materials today, 5 (2002) 18–25.
G.A. Shoales, S.A. Fawaz, M.R. Walters, Compilation of damage findings from multiple recent teardown analysis programs, in: ICAF 2009, Bridging the Gap between Theory and Operational Practice: Proceedings of the 25th Symposium of the International Committee on Aeronautical Fatigue, Rotterdam, The Netherlands, 27–29 May 2009, Springer, 2009, pp. 187–207.
F.J. Bradshaw, C. Wheeler, The effect of environment on fatigue crack growth in aluminium and some aluminium alloys, Appl Mater Res, 5 (1966) 112–120.
T. Broom, Atmospheric Corrosion-Fatigue of Age-Hardened Aluminum Alloys, Journal of the Institute of Metals, 89 (1960–1961) 183–190.
L.B. Brown, S.G. Dorman, J.T. Burns, Incorporation of high-altitude environmental effects in the linear elastic fracture mechanics‐based modeling of aluminum alloy 7075, Fatigue & Fracture of Engineering Materials & Structures, 45 (2022) 2516–2530.
J.T. Burns, J.M. Larsen, R.P. Gangloff, Effect of initiation feature on microstructure-scale fatigue crack propagation in Al–Zn–Mg–Cu, International journal of fatigue, 42 (2012) 104–121.
J.T. Burns, V.K. Gupta, S.R. Agnew, R.P. Gangloff, Effect of low temperature on fatigue crack formation and microstructure-scale propagation in legacy and modern Al–Zn–Mg–Cu alloys, International journal of fatigue, 55 (2013) 268–275.
J.T. Burns, J. Boselli, Effect of plate thickness on the environmental fatigue crack growth behavior of AA7085-T7451, International Journal of Fatigue, 83 (2016) 253–268.
R.D. Carter, E.W. Lee, E.A. Starke, C.J. Beevers, The effect of microstructure and environment on fatigue crack closure of 7475 aluminum alloy, Metallurgical Transactions A, 15 (1984) 555–563.
L.V. Corsetti, D.J. Duquette, The effect of mean stress and environment on corrosion fatigue behavior of 7075-T6 aluminum, Metallurgical and Materials Transactions B, 5 (1974) 1087–1093.
J.A. Feeney, J.C. McMillan, R.P. Wei, Environmental fatigue crack propagation of aluminum alloys at low stress intensity levels, Metallurgical Transactions, 1 (1970) 1741–1757.
F.P. Ford, Corrosion fatigue crack propagation in aluminum-7% magnesium alloy, Corrosion, 35 (1979) 281–287.
A.D.B. Gingell, J.E. King, The effect of frequency and microstructure on corrosion fatigue crack propagation in high strength aluminium alloys, Acta materialia, 45 (1997) 3855–3870.
A. Hartman, On the effect of oxygen and water vapor on the propagation of fatigue cracks in 2024-T3 alclad sheet, International Journal of Fracture Mechanics, 1 (1965) 167–188.
J.L. Jones, M.D. McMurtrey, L.B. Brown, J.T. Burns, The effect of crack wake roughness on the environmentally assisted fatigue behavior of an aerospace Al-alloy in low moisture environments, Engineering Fracture Mechanics, 199 (2018) 343–357.
M. Khobaib, F.C. Chang, E.E. Keppler, C.T. Lynch, Accelerated atmospheric-corrosion testing, in: Atmospheric Corrosion of Metals, ASTM International, 1982.
E.U. Lee, A.K. Vasudevan, G. Glinka, Environmental effects on low cycle fatigue of 2024-T351 and 7075-T651 aluminum alloys, International Journal of Fatigue, 31 (2009) 1938–1942.
F.-S. Lin, E.A. Starke Jr, The effect of copper content and degree of recrystallization on the fatigue resistance of 7XXX-type aluminum alloys II. Fatigue crack propagation, Materials Science and Engineering, 43 (1980) 65–76.
F. Lin, E.A. Starke Jr, Mechanisms of corrosion fatigue crack propagation of 7XXX aluminum alloys in aqueous environments, Hydrogen Effects in Metals, (1980) 485–492.
S.P. Lynch, Mechanisms of environmentally assisted cracking in Al Zn Mg single crystals, Corrosion Science, 22 (1982) 925–937.
J. Niebuhr, S.G. Dorman, S. Arunachalam, J. Rausch, S. Fawaz, Effect of simulated accelerated atmospheric environment conditions on fatigue crack growth rates, Corrosion, 75 (2019) 506–512.
P.S. Pao, M.A. Imam, L.A. Cooley, G.R. Yoder, Comparison of corrosion-fatigue cracking of Al-Li alloy AA 2090-T8E41 and alloy AA 7075-T651 in salt water, Corrosion, 45 (1989) 530–535.
J. Petit, C. Sarrazin-Baudoux, Some critical aspects of low rate fatigue crack propagation in metallic materials, International journal of fatigue, 32 (2010) 962–970.
R.S. Piascik, R.P. Gangloff, Environmental fatigue of an Al-Li-Cu alloy: part I. Intrinsic crack propagation kinetics in hydrogenous environments, Metallurgical Transactions A, 22 (1991) 2415–2428.
R.E. Ricker, D.J. Duquette, The role of hydrogen in corrosion fatigue of high purity Al-Zn-Mg exposed to water vapor, Metallurgical Transactions A, 19 (1988) 1775–1783.
J. Ruiz, M. Elices, Effect of water vapour pressure and frequency on fatigue behaviour in 7017-T651 aluminium alloy plate, Acta materialia, 45 (1997) 281–293.
R.E. Stoltz, R.M. Pelloux, Inhibition of corrosion fatigue in 7075 aluminum alloys, Corrosion, 29 (1973) 13–17.
R.J.H. Wanhill, Fractography of fatigue crack propagation in 2024-T3 and 7075-16 aluminum alloys in air and vacuum, Metallurgical transactions A, 6 (1975) 1587–1596.
F.J. Bradshaw, C. Wheeler, The influence of gaseous environment and fatigue frequency on the growth of fatigue cracks in some aluminum alloys, International Journal of Fracture Mechanics, 5 (1969) 255–268.
J.T. Burns, R.W. Bush, J.H. Ai, J.L. Jones, Y. Lee, R.P. Gangloff, Effect of water vapor pressure on fatigue crack growth in Al–Zn–Cu–Mg over wide-range stress intensity factor loading, Engineering Fracture Mechanics, 137 (2015) 34–55.
J.T. Burns, J.J. Jones, A.D. Thompson, J.S.W. Locke, Fatigue crack propagation of aerospace aluminum alloy 7075-T651 in high altitude environments, International Journal of Fatigue, 106 (2018) 196–207.
M. Gao, R.P. Wei, P.S. Pao, Chemical and metallurgical aspects of environmentally assisted fatigue crack growth in 7075-T651 aluminum alloy, Metallurgical Transactions A, 19 (1988) 1739–1750.
P.S. Pao, M. Gao, R.P. Wei, Environmentally assisted fatigue-crack growth in 7075 and 7050 aluminum alloys, Scripta metallurgica, 19 (1985) 265–270.
Y. Ro, S.R. Agnew, R.P. Gangloff, Effect of environment on fatigue crack wake dislocation structure in Al-Cu-Mg, Metallurgical and Materials Transactions A, 43 (2012) 2275–2292.
Y. Ro, S.R. Agnew, G.H. Bray, R.P. Gangloff, Environment-exposure-dependent fatigue crack growth kinetics for Al–Cu–Mg/Li, Materials Science and Engineering: A, 468 (2007) 88–97.
J. Ruiz, M. Elices, Environmental fatigue in a 7000 series aluminium alloy, Corrosion science, 38 (1996) 1815–1837.
J. Ruiz, M. Elices, The role of environmental exposure in the fatigue behaviour of an aluminium alloy, Corrosion science, 39 (1997) 2117–2141.
R.P. Wei, P.S. Pao, R.G. Hart, T.W. Weir, G.W. Simmons, Fracture mechanics and surface chemistry studies of fatigue crack growth in an aluminum alloy, Metallurgical and Materials Transactions A, 11 (1980) 151–158.
R.P. Wei, M. Gao, P.S. Pao, The role of magnesium in CF and SCC of 7000 series aluminum alloys, Scripta metallurgica, 18 (1984) 1195–1198.
Y.-G. Chun, S.-I. Pyun, S.-M. Lee, The influence of loading frequency on the fatigue crack propagation behaviour of Al-Zn-Mg alloy at low cyclic stress intensity level in 3.5 wt% NaCl solution, Journal of materials science letters, 10 (1991) 1439–1442.
Z.M. Gasem, R.P. Gangloff, Effect of temper on environmental fatigue crack propagation in 7000-series aluminum alloys, in, Trans Tech Publ, pp. 1479–1488.
Z.M. Gasem, Frequency-dependent environmental fatigue crack propagation in the 7XXX alloy/aqueous chloride system, University of Virginia, Ph.D., 1999.
Z. Gasem, R.P. Gangloff, Rate-limiting processes in environmental fatigue crack propagation in 7000-series aluminum alloys, Chemistry and electrochemistry of corrosion and stress corrosion cracking, (2001) 501–521.
A.M. Green, J.F. Knott, Effects of environment and frequency on the long fatigue crack growth of aluminium alloy 7475, in, Elsevier, 1989, pp. 1747–1756.
N.J.H. Holroyd, D. Hardie, Factors controlling crack velocity in 7000 series aluminium alloys during fatigue in an aggressive environment, Corrosion Science, 23 (1983) 527–546.
M. Khobaib, C.T. Lynch, F.W. Vahldiek, Inhibition of corrosion fatigue in high strength aluminum alloys, Corrosion, 37 (1981) 285–292.
D. Kovalov, B. Fekete, G.R. Engelhardt, D.D. Macdonald, Prediction of corrosion fatigue crack growth rate in alloys. Part II: effect of electrochemical potential, NaCl concentration, and temperature on crack propagation in AA2024-T351, Corrosion Science, 152 (2019) 130–139.
Y. Nakai, A. Alavi, R.P. Wei, Effects of frequency and temperature on short fatigue crack growth in aqueous environments, Metallurgical Transactions A, 19 (1988) 543–548.
J.S. Warner, R.P. Gangloff, Alloy induced inhibition of fatigue crack growth in age-hardenable Al–Cu Alloys, International journal of Fatigue, 42 (2012) 35–44.
J.S. Warner, S.S. Kim, R.P. Gangloff, Ionic Inhibition of Environental Fatigue Crack Growth in 7075-T6, Key Engineering Materials, 345 (2007) 989–994.
J.S. Warner, S. Kim, R.P. Gangloff, Molybdate inhibition of environmental fatigue crack propagation in Al–Zn–Mg–Cu, International Journal of Fatigue, 31 (2009) 1952–1965.
C.R. Bryan, E.J. Schindelholz, Properties of brines formed by deliquescence of sea-salt aerosols, in, NACE, pp. NACE-2018.
C.R. Bryan, D.G. Enos, Understanding the Risk of Chloride Induced Stress Corrosion Cracking of Interim Storage Containers for the Dry Storage of Spent Nuclear Fuel: Evolution of Brine Chemistry on the Container Surface, in, NACE, pp. NACE-2016.
G. Dorman, E. Sarah, J.H. Niebuhr, J.S. Locke, B. Free, The development of a method for complex environment corrosion fatigue testing, in.
J. Rausch, S.G. Dorman, S. Fawaz, Test methodology for the investigation of environmental effects on corrosion fatigue, Corrosion, 75 (2019) 525–535.
J.S. Warner, Inhibition of environmental fatigue crack propagation in age-hardenable aluminum alloys, University of Virginia, Ph.D., 2010.
B. Free, G. Marino, E. Schindelholz, S.G. Dorman, J.S.W. Locke, Measurement of atmospheric corrosion fatigue crack growth rates on AA7085-T7451, International Journal of Fatigue, 167 (2023) 107368.
B. Free, Atmospheric Corrosion Fatigue of AA7085-T7451 in Aerospace Environments, The Ohio State University, Ph.D., 2023.
J.S. Locke, Comparison of age-hardenable Al Alloy corrosion fatigue crack growth susceptibility and the effect of testing environment, Corrosion, 72 (2016) 911–926.
Z. Chen, F. Cui, R.G. Kelly, An analytical modeling method for calculating the current delivery capacity of a thin-film cathode and the stability of localized corrosion under atmospheric environments, ECS Transactions, 3 (2007) 443.
C.L. Alexander, C. Liu, A. Alshanoon, R.M. Katona, R.G. Kelly, J. Carpenter, C. Bryan, E. Schindelholz, Oxygen reduction on stainless steel in concentrated chloride media, Journal of The Electrochemical Society, 165 (2018) C869.
N. Brown, P. Kramer, F. Friedersdorf, M. Schindelholz, J. Siegel, Environmentally assisted cracking measurements in structural aluminum alloys under accelerated test conditions, Corrosion, 72 (2016) 1351–1362.
Z.Y. Chen, F. Cui, R.G. Kelly, Calculations of the cathodic current delivery capacity and stability of crevice corrosion under atmospheric environments, Journal of the Electrochemical Society, 155 (2008) C360.
J.R. Scully, G.A. Young Jr, S.W. Smith, Hydrogen embrittlement of aluminum and aluminum-based alloys, in: Gaseous hydrogen embrittlement of materials in energy technologies, Elsevier, 2012, pp. 707–768.
R.P. Gangloff, Environment sensitive fatigue crack tip processes and propagation in aerospace aluminum alloys, in, EMAS Stockholm (Sweden), 2002, pp. 3401–3433.
R.P. Gangloff, Diffusion control of hydrogen environment embrittlement in high strength alloys, Hydrogen effects on material behavior and corrosion deformation interactions, (2003) 477–497.
J.R. Scully, G.A. Young Jr, S.W. Smith, Hydrogen solubility, diffusion and trapping in high purity aluminum and selected Al-base alloys, in, Trans Tech Publ, 2000, pp. 1583–1600.
G.A. Young, J.R. Scully, Hydrogen production, absorption and transport during environment assisted cracking of an Al-Zn-Mg-(Cu) alloy in humid air, in, 2002, pp. 22–26.
G.A. Young, J.R. Scully, The effects of test temperature, temper, and alloyed copper on the hydrogen-controlled crack growth rate of an Al-Zn-Mg-(Cu) alloy, Metallurgical and Materials Transactions A, 33 (2002) 1167–1181.
L.M. Young, Microstructural dependence of aqueous-environment-assisted crack growth and hydrogen uptake in AA 7050, University of Virginia, 1999.
L.M. Young, G. Young, The effects of electrode potential and temper on environment assisted cracking in AA 7050, Light weight alloys for aerospace applications IV, (1996) 3–18.
E. Schindelholz, R.G. Kelly, Wetting phenomena and time of wetness in atmospheric corrosion: a review, Corrosion reviews, 30 (2012) 135–170.
E. Schindelholz, L.-k. Tsui, R.G. Kelly, Hygroscopic particle behavior studied by interdigitated array microelectrode impedance sensors, The Journal of Physical Chemistry A, 118 (2014) 167–177.
M.D. Cohen, R.C. Flagan, J.H. Seinfeld, Studies of concentrated electrolyte solutions using the electrodynamic balance. 1. Water activities for single-electrolyte solutions, Journal of Physical Chemistry, 91 (1987) 4563–4574.
I.N. Tang, Deliquescence properties and particle size change of hygroscopic aerosols, in, Brookhaven National Lab., Upton, NY (USA), 1979.
D.J. Cziczo, J.B. Nowak, J.H. Hu, J.P.D. Abbatt, Infrared spectroscopy of model tropospheric aerosols as a function of relative humidity: Observation of deliquescence and crystallization, Journal of Geophysical Research: Atmospheres, 102 (1997) 18843–18850.
E. Schindelholz, B.E. Risteen, R.G. Kelly, Effect of relative humidity on corrosion of steel under sea salt aerosol proxies: II. MgCl2, artificial seawater, Journal of The Electrochemical Society, 161 (2014) C460.
ASTM E647: Standard Test Method for Measurement of Fatigue Crack Growth Rates, in, ASTM International, 2015.
E. Schindelholz, R.G. Kelly, Application of inkjet printing for depositing salt prior to atmospheric corrosion testing, Electrochemical and Solid-State Letters, 13 (2010) C29.
R.F. Schaller, C.F. Jove-Colon, J.M. Taylor, E.J. Schindelholz, The controlling role of sodium and carbonate on the atmospheric corrosion rate of aluminum, npj Materials Degradation, 1 (2017) 20.

(Not answered)

Download PDF

Editorial decision: revise
05 Aug, 2024
Review #1 received at journal
02 Aug, 2024
Review #2 received at journal
31 Jul, 2024
Reviewer #2 agreed at journal
21 Jul, 2024
Reviewer #1 agreed at journal
16 Jul, 2024
Reviewers invited by journal
09 Jul, 2024
Editor assigned by journal
09 Jul, 2024
Submission checks completed at journal
09 Jul, 2024
First submitted to journal
08 Jul, 2024
Unknown event
04 Jul, 2024

You are reading this latest preprint version

The Effect of Variable Humidity on Corrosion Fatigue of AA7085-T7451 with Surface Salt Deposits

Status:

Version 1

Abstract

Figures

Introduction

Results

3.1 Establishing Baseline Performance in Conventional Environments

3.2 Humidity Testing using Constant RH Segments

3.3 Humidity Testing with Continuously Changing RH

3.4 SEM Fractography

Discussion

4.1 Understanding Crack Growth Rate Changes During the Wetting Process

4.1.1 While RH is Below 76%, the DRH

4.1.2 Once RH is at and Above the DRH

4.2 Understanding crack growth rate changes during the drying process

Methods

5.1 Material

5.2 Salt deposition

5.3 Fatigue crack growth rate testing

5.4 Environmental control

5.4.1 Humidity Testing using Constant RH Segments

5.4.2 Humidity Testing with Continuously Changing RH

5.5 SEM Fractography

Declarations

6 Data Availability

7 Acknowledgements

8 Author Contributions

9 Competing Interests

References

Additional Declarations

Status:

Version 1