SCC evaluation of a 2297 Al-Li alloy rolled plate using the slow-strain rate technique

Kuo ZHAO, Jianhua LIU, Mei YU, Songmei LI

School of Materials Science and Engineering, Beihang University, Beijing 100083, China

KEYWORDS

Abstract The stress corrosion cracking (SCC) susceptibility of 2297 Al-Li alloy in 1 M NaCl+0.01 M H2O2 solution (CP solution) and 1 M NaCl+0.01 M H2O2+0.6 M Na2SO4 solution(CPS solution) was investigated by slow-strain rate tests at various strain rates ranging from 10-5 s-1 to 10-7 s-1.The roles of H2O2 and in the corrosion process were estimated by potentiodynamic polarization and electrochemical impedance spectroscopy. 2297 Al-Li alloy does not fracture ascribed to SCC in CP solution,while it undergoes SCC in CPS solution.In CPS solution,with a decreasing strain rate from 10-5 s-1 to 10-7 s-1,the SCC susceptibility firstly rises and then declines exhibiting a peak value at a strain rate of 10-6 s-1. H2O2 promotes the active dissolution while lowers the corrosion rate. The SCC fracture is associated with a decline in the dissolution rate of the crack tip by ,which leads to stress concentration.In CPS solution,a reduction in the local dissolution rate of the crack tip leads to stress concentration,resulting in SCC fracture.As the preferred initiation site for a crack,pits also show a noteworthy effect on SCC of 2297 Al-Li alloy.

1. Introduction

It has been decades that the modern aerospace industry focuses on weight reduction, which is an effective way to improve aircraft performance. High-performance aluminumlithium(Al-Li)alloys are effective lightweight materials.Compared with conventional aluminum alloys,Al-Li alloys exhibit a favorable combination of good properties such as relatively low density, high specific strength, and high stiffness.1-4By an addition of every weight percentage of Li, the density of the resultant alloy decreases up to 3%, and the stiffness increases up to 6%.2-4

Aluminum alloys are generally stable in aqueous solutions of 4 ≤pH ≤8.5, ascribed to the substantial corrosion resistance provided by an oxide film.5-7Similarly, Al-Li alloys are protected from corrosion by an air-formed oxide film.However, due to the electrochemical inhomogeneity resulted from intermetallic strengthening precipitates, Al-Li alloys are not immune to localized corrosion such as pitting corrosion,intergranular corrosion, exfoliation corrosion, and stress corrosion.8-10Stress corrosion cracking (SCC) is a cracking phenomenon resulted from a conjoint action of a tensile load(the stress level must be higher than a particular threshold value) and a specific corrosive environment.11-13SCC remains one of the most serious failures that can occur with aluminum alloys, as an SCC failure often leads to a sudden and catastrophic failure without any warning.12-15

In recent years, SCC of aluminum alloys in chloride environment has been the subject of extensive studies.15-20Ricker et al. studied the SCC susceptibility of AA7075-T6 alloy, and deduced that chloride ions interacted chemically with the passivated surface at the crack tip leading to SCC.16Hu et al.studied the SCC resistance of 2195 Al-Li alloy and AA2219 alloys using a constant-load test and slow-strain rate tensile(SSRT) tests.17Their results revealed that crack initiation was mostly from pitting.In many engineering applications,pitting is the precursor to SCC as it provides stress concentration and aggressive local solution chemistry. For instance, Connolly et al. observed pit-to-crack transition using threedimensional(3D)X-ray microtomographic images,and results convincingly indicated that the preferred site for crack initiation was on the pit wall close to the pit mouth.18

For the most part, both anodic dissolution and hydrogen embrittlement are the major SCC mechanisms. Electrochemical measurements are an attractive option to evaluate the SCC mechanism.21-25For instance, potentiodynamic polarization and electrochemical impedance spectroscopy(EIS) measurements were used to investigate the SCC mechanism of ultra-high-strength steel 30CrMnSiNi2A and 23Co14Ni12Cr3Mo.21,23For aluminum alloys, Frankel et al.investigated the SCC behavior of AA5083 using SSRT tests at different polarized potentials, and pointed out that the mechanism of inhibition of SCC was mainly through the control of the cathodic reaction.25Moreover, electrochemical measurements can also be employed to evaluate the effects of anions and passive film on SCC.21-23,26

In the present work, the SCC behavior of 2297 Al-Li alloy was investigated by SSRT tests. Two kinds of solutions were used in SSRT tests. The first was NaCl and H2O2aqueous solution, and the second was NaCl, H2O2, and Na2SO4aqueous solution.The electrochemical behavior of 2297 Al-Li alloy was investigated by potentiodynamic polarization and EIS measurements.

2. Experimental

2.1. Sample preparation

The target metal in this study was a 2297 Al-Li alloy rolled plate with a thickness (t) of 95 mm. Its chemical composition is listed in Table 1,which was measured by an inductively coupled plasma atomic emission spectrometer (Optima-7000DV,USA). The various directions of the plate are conventionally labeled as longitudinal(L),long transverse(T),and short transverse (S). Samples were cut from the plate at t/2 in the T-S plane. The dimensions of the tensile samples are shown in Fig.1, with a gauge length and width of 20 mm and 5 mm,respectively. Cylindrical samples with a basal area of 1 cm2were used in electrochemical tests. All samples were manually ground with abrasive papers of 1500 grits along T in ethyl alcohol, and then ultrasonically cleaned in ethyl alcohol.

2.2. Microstructure characterization

The metallographic morphologies of 2297 Al-Li alloy were revealed by a 3D video microscope(3D-VM;RH 2000,Japan)after being etched by Keller solution. A scanning electron microscope (SEM; Apollo 300, UK) and a transmission electron microscope (TEM; JEOL-2100, Japan) were employed to observe the microstructure of 2297 AL-Li alloy. The element compositions of intermetallic particles were determined by energy dispersive X-ray spectroscopy (EDS; INCA Penta-FETx3, UK).

After SSRT tests, the side surface morphologies and crosssectional morphologies of fractured samples were observed by SEM and 3D-VM which could construct the 3D structure of micro cracks and corrosion paths.

2.3. Solutions

Two solutions were used in SSRT tests. The first solution was 1 M NaCl+0.01 M H2O2solution(termed CP solution),and the second solution was 1 M NaCl+0.01 M H2O2+0.6 M Na2SO4solution (termed CPS solution). Three solutions were employed in electrochemical tests: CP solution, CPS solution,and 1 M NaCl solution (termed C solution). Corrosion products on corroded samples were ultrasonically cleaned in 5 vol.% phosphoric acid +20 g·L-1chromium trioxide (CrO3)solution.

Table 1 Chemical composition (wt.%) of 2297 Al-Li alloy.

2.4. SSRT tests

In order to define the fracture energy (FE, area under the entire stress-strain curve) loss and differentiate environmentinduced fractures,three kinds of SSRT were conducted in this study.In an air test,an uncorroded sample was stressed in air.The SSRT test in which a sample was subjected to a tensile load in an aggressive solution was named as solution test. In a pre-exposure test, unstressed samples were firstly immersed in an aggressive medium for an equivalent time to that of a solution test, and then stressed in air. Similarly, samples used in air tests, solution tests, and pre-exposure tests are respectively named as AT samples, ST samples, and PT samples.

2.5. Electrochemical measurements

Electrochemical measurements were conducted on a Princeton 2273 electrochemical workstation using a traditional threeelectrode cell. A cylindrical sample was used as the working electrode.A saturated calomel electrode(SCE)and a platinum sheet served as the reference electrode and the counter electrode, respectively.

Potentiodynamic polarization measurements were initiated at-200 mV relative to the open circuit potential(OCP)with a scan rate of 1 mV·s-1to 200 mV (vs. OCP). Polarization data was analyzed by Tafel extrapolation. EIS measurements were performed at the OCP using a 10 mV AC stimulus signal with a frequency range from 100 kHz to 10 mHz. Impedance data was analyzed by Zsimpwin software.All the tests and measurements were repeated at least three times for data reproducibility. Scatter bands given in this work are standard deviations calculated from 3 parallel valid data.

3. Results

3.1. Microstructure of the alloy

The metallographic morphology in the T-S plane of the plate at t/2(Fig.2)shows that the plate has an elongated,deformed grain structure.A typical backscattered electron SEM image in the T-S plane of the plate at t/2 and the corresponding EDS spectrum are presented in Fig.3. The white speckles in Fig.3(a) reveal the distribution of intermetallic particles with sizes (equivalent diameters) ranging from 9.95 μm to 0.41 μm.Chemical compositions of nearly 50 intermetallic particles were determined by EDS. The intermetallic particles are AlCuMnFe phases,and their average composition(by weight)comprises Al 63.5%,Cu 25.6%,Mn 7.6%,and Fe 3.3%.TEM was employed to characterize the detailed microstructure of the alloy. Typical TEM images of the plate at t/2 are shown in Fig.4. TEM observation reveals a uniform distribution of plate-shaped T1phases (Al2CuLi) with sizes (lengths) ranging from 27 nm to 161 nm.

Fig.2 Typical metallographic morphology in the T-S plane of the plate at t/2.

3.2. SSRT tests associated with CP solution

Stress-strain curves of the alloy tested at different strain rates in CP solution are presented in Fig.5.It is shown that the fracture is ductile at the five strain rates. Both the ultimate tensile strength(σm)and the FE decreased remarkably with a decreasing strain rate. For comparison, stress-strain curves of the alloy tested at a strain rate of 10-6s-1in air and in CP solution are presented in Fig.6.Noticeable decreases in σmand the FE are detected for the ST sample and the PT sample in comparison with those of the AT sample.In pre-exposure tests,the FE loss is ascribed to severe intergranular attack, which preferentially propagates in the lamination.In solution tests,if an SCC phenomenon exists, the ST sample ought to show additional FE loss.However,the FE of the ST sample is larger than that of the PT sample,indicating that the fracture in CP solution at a strain rate of 10-6s-1is a consequence of mechanical overload, rather than an SCC phenomenon.

The side surface morphologies of the ST sample and the PT sample after relevant tests at a strain rate of 10-6s-1are shown in Fig.7. The PT sample shows an increased corrosion width and depth in comparison with those of the ST sample.It is well known that corrosion of aluminum alloys leads to a local acidification as a result of hydrolysis of metal ions.27-29Consequently, the dissolution of the material continues in pre-exposure tests after immersion.Thus,the PT sample shows additional σmloss and FE loss in comparison with those of the ST sample.

The FEs of the alloy tested at various strain rates in air and in CP solution are presented in Fig.8. The FE loss of the ST sample increases significantly with a decreasing strain rate.The pre-exposure times for PT samples at strain rates of 10-5s-1, 5×10-6s-1, 10-6s-1, 5×10-7s-1, and 10-7s-1are 145±15 min, 245±9 min, 1052±43 min, 1973±65 min, and 7664±288 min, respectively. A decrease in the strain rate would force a longer exposure time in the aggressive medium, thus further accentuating the reduction in the sectional area of the sample. Furthermore, for all analyzed strain rates,there is additional FE loss for the PT sample in comparison with that of the ST sample tested at the same strain rate. It is thereby concluded that 2297 Al-Li alloy does not fracture ascribed to SCC in CP solution at strain rates of 10-5s-1, 5×10-6s-1, 10-6s-1, 5×10-7s-1, and 10-7s-1.

Fig.3 SEM image of the plate at t/2 and corresponding EDS spectrum.

Fig.4 Typical TEM images of the plate at t/2.

Fig.5 Stress-strain curves of the alloy tested at different strain rates in CP solution.

3.3. SSRT tests associated with CPS solution

Fig.6 Stress-strain curves of the alloy tested at=10-6 s-1 in air and in CP solution.

Fig.7 3D images of side surfaces after SSRT tests at =10-6 s-1 associated with CP solution.

Fig.8 FEs of the alloy tested at different strain rates in air and in CP solution.

Fig.9 Stress-strain curves of the alloy tested at different strain rates in CPS solution.

Stress-strain curves of the alloy tested at different strain rates in CPS solution are presented in Fig.9. For the alloy tested at strain rates of 10-6s-1, 5×10-7s-1, and 10-7s-1,a fracture occurs shortly after the elastic limit, indicating that it is a brittle fracture. For comparison, stress-strain curves of the alloy tested at a strain rate of 10-6s-1in air and in CPS solution are presented in Fig.10. As expected, noticeable decreases in σmand the FE are detected for the ST sample and the PT sample in comparison with those of the AT sample.Importantly, the FE of the ST sample is smaller than that of the PT sample. The side surface morphologies of the ST sample and the PT sample after relevant tests at a strain rate of 10-6s-1in CPS solution are shown in Fig.11.It is shown that the corrosion depth of the PT sample is larger than that of the ST sample. Though it undergoes additional corrosion after immersion, the PT sample exhibits a higher FE than that of the ST sample, indicating an SCC-caused fracture for a solution test at a strain rate of 10-6s-1.

Fig.10 Stress-strain curves of the alloy tested at=10-6 s-1 in air and in CPS solution.

The fracture surface morphology of the ST sample after a solution test at a strain rate of 10-6s-1is presented in Fig.12,showing an intergranular(IG)brittle fracture surface.Pits are visible in the fracture surface of the ST sample, and some secondary cracks are detected within the pit cavity.The side surface morphology and cross-sectional morphology of the ST sample after a solution test at a strain rate of 10-6s-1are shown in Fig.13.Clearly,it can be seen that a secondary crack originates from the pit cavity. Moreover, the height difference of the fracture surface is greater than 1.5 mm. It is difficult to get clear SEM images of the whole uneven fracture surface (Fig.12). Thus, a calculation of the area with a cleavage-like fracture has not been carried out.The presence of secondary cracks and an intergranular brittle fracture surface reveal an SCC-caused fracture of the alloy in CPS solution.

The FEs of the alloy tested at various strain rates in air and in CPS solution are presented in Fig.14. For all analyzed strain rates, the FE losses for pre-exposure tests in CPS solution are higher than those for solution tests at the same strain rate, indicating SCC-caused fracture. Moreover,the FE losses for both solution tests and pre-exposure tests increase significantly with a decreasing strain rate. The pre-exposure times for PT samples at strain rates of 10-5s-1, 5×10-6s-1,10-6s-1, 5×10-7s-1, and 10-7s-1are 105±8 min, 209±15 min, 734±59 min, 1349±83 min, and 4717±229 min, respectively. As demonstrated in Figs. 11, 12,and 13, the ST sample underwent severe localized corrosion during a solution test, resulting in FE loss. Note that a decrease in the strain rate could induce an increase in the FE loss as it forces a longer exposure time in the medium,resulting in an FE drop increase. To estimate the SCC susceptibility,one reasonable method is the characterization of the fraction of the intergranular or brittle cracking area on the fracture surface. However, as the fracture surface in this study is uneven(Figs. 12 and 13), it is difficult to estimate the fraction of the IG or brittle cracking area. Besides, a partial fracture surface(mainly crack wall, if it exists) is corroded by the aggressive medium used in this study. Thus, it is preferable to evaluate the SCC susceptibility by the FE loss.

Fig.11 3D images of side surfaces after SSRT tests at =10-6 s-1 associated with CPS solution.

Fig.12 Fracture morphology of the alloy after a solution test at=10-6 s-1 in CPS solution.

Fig.13 Side surface morphology and cross-sectional morphology of the ST sample after a solution test at =10-6 s-1 in CPS solution.

Fig.14 FEs of the alloy tested at different strain rates in air and in CPS solution.

For a solution test in CPS solution, both localized corrosion and SCC are responsible for the FE loss. The FE loss caused by localized corrosion is therefore preferred to be removed.Thus,a decrease in the FE of the alloy in an aggressive solution over those in an inert medium (mostly in air) is inapplicable to evaluate the SCC susceptibility (ISCC) of the alloy in this study.Similar to the residual strength,the residual FE of the PT sample is adopted to evaluate the effect of corrosion on mechanical properties. Therefore, the corrosion susceptibility (IC) and the ISCCwere respectively calculated by the FE loss according to Eqs. (1) and (2). It should be mentioned that pits on the PT sample have already occurred prior to the loading in air, while the ST sample undergoes pitting during the loading in solution. Furthermore, tensile stress can accelerate the pit propagation. Those two effects cause errors for calculations of ICand ISCC. Fortunately, the error caused by the combination of those two effects is smaller than that caused by each individually. Hence, we believe this combined error is acceptable for evaluation of the SCC susceptibility of one alloy at strain rates.

where WAT,WPT,and WSTare the FEs of the AT,PT,and ST samples, respectively.

ICand ISCCof the alloy in CPS solution at different strain rates are shown in Fig.15. With a decreasing strain rate from 10-5s-1to 10-7s-1, ICincreases rapidly, while ISCCfirstly rises and then declines,exhibiting a peak value at a strain rate of 10-6s-1.Thus, it is concluded that 2297 Al-Li alloy undergoes SCC in CPS solution at strain rates of 10-5s-1,5×10-6s-1,10-6s-1,5×10-7s-1,and 10-7s-1,with a peak ISCCat a strain rate of 10-6s-1.

3.4. Electrochemical measurements

In order to investigate the corrosion behavior of the alloy,potentiodynamic polarization tests and EIS measurements were performed on the alloy in CP solution, CPS solution,and C solution. Samples tested in CP solution, CPS solution,and C solution are respectively named as CP sample,CPS sample, and C sample for distinguishing.

The potentiodynamic polarization curves of the alloy tested in the three solutions are presented in Fig.16. There is one obvious breakdown potential which is much nobler than the corrosion potential (Ecorr) in the polarization curve of the C sample, revealing the good protective property of the oxide film in C solution. Both the cathodic current density and the anodic current density of the CP sample are significantly higher than those of the C sample, indicating that the oxide film ruptures shortly in CP solution and CPS solution. The corrosion of aluminum in a NaCl solution is generally associated with a breakdown of the oxide film caused by chloride ions.8,18,33Oxygen is the only oxidizing species in a NaCl solution. Thus, the corrosion rates of aluminum alloys in such solution are usually determined by the oxygen concentration.In a NaCl solution containing hydrogen peroxide, oxygen is not the exclusive oxidizing species, as the cathodic reaction may also take place in the following way:

Fig.15 ISCC and IC of the alloy at different strain rates in air and in CPS solution.

Fig.16 Potentiodynamic polarization curves of the alloy tested in three solutions.

Table 2 Tafel extrapolation results of potentiodynamic polarization curves of the alloy.

Fig.17 Typical EIS results of the alloy tested in three solutions.

Fig.18 EEC used for fitting EIS tests data.

Fig.19 Fitted values of Rf and Rct in the EEC of the alloy tested in three solutions.

4. Discussion

4.1. Effects of Cl-, H2O2, and

The present results reveal that 2297 Al-Li alloy did not fracture ascribed to SCC in CP solution, while SCC-caused fracture is revealed in CPS solution.It has been demonstrated that pitting attack can act as nucleating sites of stress concentration,which eventually leads to SCC.18,38-41The fundamental steps of crack development involve pit initiation, pit growth, pit-tocrack transition, and crack growth.18There are two criteria for a pit-to-crack transition: the threshold stress intensity factor has to be exceeded,and the growth rate of the crack should be greater than that of the pit.

During solution tests, samples are simultaneously exposed to an aggressive solution and stress.A pit has to exceed a minimum depth to obtain a stress intensity higher than the threshold. The electrochemical test results demonstrate that the presence of H2O2promotes the corrosion rate of 2297 Al-Li alloy, resulting in large pits. As shown in Figs. 7(a) and 11(a),ST samples associated with CP solution and CPS solution both underwent pitting corrosion during relevant solution tests. It is well known that pits can causes stress concentration,12,38,41and this might result in a high enough stress intensity for crack initiation.

On the other hand,after crack initiation,the growth rate of a pit,especially for the wall,must be lower than that of a crack for the crack to grow. In CP solution, the new bare surface in the corrosion path and crack walls (if a crack exists) tend to locally dissolve, as indicated by the long and wide corrosion attack in Fig.7.As a result,the crack tip will be corroded,lowering the crack growth rate.Ultimately,the crack is merged by the corrosion attack.Consequently,a fracture is resulted from a mechanical overload at all analyzed strain rates in CP solution, because of the reduction in the sample sectional area caused by the severe corrosion attack.

4.2. Strain rate effect

In CPS solution,with a decreasing strain rate from 10-5s-1to 10-7s-1, ISCCof 2297 Al-Li alloy firstly rises and then declines, exhibiting a peak value at a strain rate of 10-6s-1. It is well known that SCC results from the synergistic effects of a mechanical load and a corrosive environment.11,12,18During the crack initiation stage, there is an incubation period in which the stress intensity and the corrosion environment interact with each other until local conditions in the crack tip region reach a critical state.12

As demonstrated by Figs. 5 and 9, with a decreasing strain rate, σmof the ST sample associated with CPS solution decreases significantly, revealing that the pit number and volume increase.This result is consistent with the FEs at different strain rates. A decrease in the strain rate forces a longer immersion time in an aggressive medium. Thus, ICincreases with a decreasing strain rate.

In the high-strain rate region from 10-5s-1to 10-6s-1,the immersion time increases with a decreasing strain rate, resulting in enhanced pitting corrosion. As a result, there are more pits formed due to corrosion for crack initiation. Thus, ISCCincreases with a decreasing strain rate from 10-5s-1to 10-6s-1. In the low-strain rate region from 5×10-7s-1to 10-7s-1,the immersion time in CPS solution before the stress reaches σmis much longer than that in the strain rate region from 10-5s-1to 10-6s-1. The pitting corrosion of the ST sample in the low-strain rate region is thus more severe in comparison with that of the ST sample in the high-strain region.The fracture in the low-strain region is more related to mechanical overload rather than SCC. This is consistent with the stress-strain curves. With a decreasing strain rate, the strain region of the stress plateau in the stress-strain curves becomes more shortened. Particularly, the stress-strain curve at a strain rate of 10-7s-1barely shows a stress plateau.SSRT test results associated with CPS solution suggest that pits, as preferred crack initiation sites, show a noteworthy effect on the SCC behavior of 2297 Al-Li alloy. When there are certain pits,the interaction of the stress intensity and a corrosive environment promotes SCC of 2297 Al-Li alloy.

5. Conclusions

(1) 2297 Al-Li alloy did not fracture ascribed to SCC in CP solution, while it undergoes SCC in CPS solution. In CPS solution, with a decreasing strain rate from 10-5s-1to 10-7s-1, the SCC susceptibility of 2297 Al-Li alloy firstly rises and then declines, exhibiting a peak value at a strain rate of 10-6s-1.

Acknowledgements

This study was co-supported by the National Nature Science Foundations of China (No. 51671013) and Beijing Nova Program of China (No. Z161100004916061).