Humid Gas Stress Corrosion Cracking in MIG-Welded 5083 Aluminum Alloy Plate

Abstract

In 5000 series (Al–Mg) alloys, hydrogen embrittlement (HE) becomes a concern when Mg content exceeds 5%. The local Mg content will become higher than 5% due to solidification segregation in the weld joint of 5083 alloy. In this study, resistance to HE of MIG-welded 5083 aluminum alloy at three areas (weld center, weld corner and HAZ) was investigated by newly developed humid gas stress corrosion cracking (HG-SCC) test. Slow strain rate tensile (SSRT) test on base metal and welded joint was also conducted for comparison.

It was newly revealed that HG-SCC crack extends only in the sample of weld corner, confirming HE in this area. For the other areas, HG-SCC crack did not propagate, confirming HE resistance of these areas. From the results of SSRT test, neither the base nor welding material showed reduction in strength or ductility. Therefore, it was concluded that HE in 5083 MIG welds is only detectable by HG-SCC test. The reason for this should be in the fact that HG-SCC test can specifically test different local areas, while the softest area (weld center) always deforms preferentially and fractures in SSRT test.

Fig. 6 SEM images of the fracture surface after HG-SCC test in HA environment of the #1 sample of each position in Table 4. (c1) and (c2): close-up images of fatigue and HG-SCC cracks of sample (c), respectively. The top and bottom sides of the photo (c) and bottom side of the photo (d) corresponds to the surface of the test piece.

1. Introduction

In recently years, 5083 aluminum alloy is widely used as a structural material because of the excellent combination of weldability, corrosion resistance and mechanical properties. One of its most important applications is in construction of liquid natural gas (LNG) tanks in LNG carriers.^1–4) Despite of these unique properties, stress corrosion cracking (SCC) becomes a concern when Mg content exceeds 5%.¹⁾ The SCC tends to become more marked as the distribution of precipitates of β phase (Al₃Mg₂ or Al₈Mg₅) on grain boundaries becomes denser.^5–7) The β phase acts as a strong anode against surrounding matrix.⁸⁾ The precipitation of β phase is based on the low (around 1.4%) solubility of Mg in aluminum at ambient temperature.^9–11) Since the increase in the Mg content in the alloy accelerates the precipitation (increases the density) of β phase on grain boundaries, the Mg content in commercially available 5083 alloys is limited within about 5%. However, in the welding process, segregation takes place during solidification and it is believed that the Mg concentration becomes the highest at the weld boundary and decreases as it gets closer to the center.^12–15)

In the 5000 series alloy, hydrogen embrittlement (HE) associated with intergranular cracking was confirmed in an Al–8%Mg alloy in which hydrogen was introduced by the cathodic charging method with no apparent corrosion.¹⁶⁾ Humid environment as a source of hydrogen may also be playing an important role in the initiation and propagation of SCC in 5000 series alloys.^17,18) Water vapor in air can be a source of hydrogen for hydrogen embrittlement (HE) through reactions such as indicated in eq. (1).   

\begin{equation} \text{2Al(s)} + \text{($3+X$)H$_{2}$O} \to \text{Al$_{2}$O$_{3}{\cdot}{}X$(H$_{2}$O)(s)} + \text{3H$_{2}$(g)}. \end{equation} (1)

The fugacity of the hydrogen gas is over 10⁵³ Pa at 25°C in air with relative humidity of 87%¹⁷⁾ that is sufficient to dissolve into aluminum as atomic state. Similar reaction with Mg can also occur in 5000 series alloys.

Serious accidents have not been reported yet based on HE in 5083 alloys in O-temper with Mg content less than 5% but there is a possibility that the Mg concentration of welded part locally exceeds 5% due to solidification segregation as mentioned above, which will increase the risk for HE. However, only a paper has been reported according to the search by the authors on the impact of the humid environment on the HE characteristics of metal inert gas (MIG) welded parts in 5083 aluminum alloys.¹⁹⁾

As the test method, humid gas stress corrosion cracking (HG-SCC) test was employed for the purpose of this study. This test method has been newly developed for screening the aluminum alloys to be used for compressed hydrogen gas containers for fuel cell vehicles,²⁰⁾ and has never been applied to 5000 series aluminum alloys, not to mention to welded joints. In this testing method, the sensitivity to HE can be assessed by introducing pre-cracks in different locations with respect to the welding centerline, which enables assessing the extent of HE with respect to different local areas. Therefore, the HG-SCC test was carried out in this study by considering the location of the pre-cracks. Slow strain rate tensile (SSRT) test, which have been frequently used to assess the extent of HE,^19,21–23) on base metal and welded joint was also conducted for comparison.

2. Experimental Method

2.1 Sample

A pair of O-tempered 5083 plates with a thickness of 8 mm were subjected to MIG welding with a welding wire of 5356 alloy with a diameter of 1.6 mm. The actual and standard chemical compositions of these alloys are provided in Table 1. The butt joint with full penetration MIG welding was carried out, parallel to the rolling direction of the plate. The welding parameters are indicated in Table 2. The interval time between completion of welding process and starting date of tests was about 60 d. To reveal the macro- and micro-structures, the weld cross section was ground up to #2000, mirror-finished with colloidal silica liquid, and then etched in a 10%NaOH aqueous solution at 70∼80°C for 60 s, followed by de-smut treatment in a 10%HNO₃ aqueous solution at room temperature.

Table 1 Composition in mass% of the 5083 aluminum alloy used in this study as base metal and the standard composition of 5356 aluminum alloy welding wire.

Table 2 Welding parameters.

2.2 Vickers hardness test

Vickers hardness test was carried out at a load of 9.8 N and a load holding time of 10 s on the mirror-finished long transverse – short transverse (LT-ST) section (perpendicular to the weld line) of the base and welded materials. The indenter was positioned in a row at the center in the plate thickness direction with the other two rows with an interval of ±1.5 mm from plate thickness centerline toward the top and back surfaces, as shown in Fig. 1. To obtain hardness distribution in the LT direction, indentations were made with an interval of 1 mm from the weld centerline toward both sides up to 20 mm in the LT direction. The data from the three rows in ST direction were averaged. The averaged value (HV) was regarded as the representative of the location and plotted as a function of the distance from the weld centerline in the LT direction.

Fig. 1

Schematic diagram of location of indents in hardness test in weld cross section.

2.3 HG-SCC test

The HG-SCC test was carried out based on Japan High Pressure Institute Standard, HPIS E103: 2018.²¹⁾ in four different areas: (a) base metal, (b) weld metal center, (c) weld metal adjacent to the interface with base metal on the surface (weld corner), and (d) heat-affected zone (HAZ). In the HG-SCC test, a predetermined stress intensity factor (K_IApp in MPa$\sqrt{\text{m}} $), which is equivalent with 0.2% proof stress (σ_0.2 in MPa) at a crack tip of 1 mm length and is equal to $0.056\sigma _{0.2} = \sqrt{\pi \times 0.001} \sigma _{0.2}$, is kept applied to a compact tension (CT) test piece for the period of 90 d.

For conducting HG-SCC test, local σ_0.2 values are needed. In this study, local σ_0.2 values in the locations (b) to (d) were assessed with HV values under an assumption that σ_0.2 is proportional to HV²⁴⁾ (σ_0.2 of welded joint, σ_0.2W, can be estimated from eq. (2)).   

\begin{equation} \sigma_{0.2W} = \sigma_{0.2B} \times \mathit{HV}_{w}/\mathit{HV}_{B} \end{equation} (2)

where σ_0.2B, HV_B and HV_w are σ_0.2 and HV of base metal, and HV of welded joint, respectively. The HG-SCC test was conducted in humid air (HA) environment with relative humidity (RH) of 90% or more at 25°C for a duration of 90 d. The welded plates were machined to 7 mm thickness from both sides, removing the weld crown. Four CT test pieces: (one from base metal far from welding area and three from welded material at different welding areas) were cut and machined to the size and shape as indicated in Fig. 2. After the macroscopic features on the surface that was mentioned earlier, the test pieces will be referred to as: base metal (a), weld center (b), weld corner (c) and HAZ (d), hereafter. All the test pieces were ground up to #2000 with abrasive paper then mirror-finished by buffing with diamond past on both surfaces.

Fig. 2

Configuration and morphology of compact tension (CT) test piece, cut from base metal (a) and welded materials (b) to (d). (I) overall view of test pieces, gray line represents weld line, (II) location of the notch with respect to the weld centerline, (III) details of the CT test piece. D: distance of the notch from the weld centerline. The values of D are 0 for (b), 7 mm for (c) and 12 mm for (d).

The HG-SCC test was conducted in four steps:

1. I)    Pre-crack with length of about 4 mm was introduced by cyclic loading less than 60% of K_IApp. Then the pre-crack length was measured with optical microscope on both surfaces, and then the average of the values for both surfaces was used to determine the tensile load that provides K_IApp for the specimen.
2. II)    Static stress intensity factor equal to K_1App was applied in constant displacement condition and kept in HA chamber with RH of more than 90% but without dew, held for 90 d. Environment temperature was kept constant at 25°C ± 5°C for the entire duration of the test. After completion of the duration, crack length was measured with optical microscope on both surfaces. The extension of crack length means the progress of HG-SCC on the surface.
3. III)    Post crack was introduced by cyclic loading up to a length more than 1 mm with stress intensity factor range less than 60% of K_IApp, followed by monotonic rapid loading up to break.
4. IV)    Fracture surface was observed with a scanning electron microscope (SEM), and the region sandwiched between the two fatigue cracks was regarded as HG-SCC crack surface. The length of HG-SCC was assessed as the average of the crack propagation distance at the position of 1/4, 1/2 and 3/4 thickness. Testing in a dry air chamber with RH less than 5% was conducted with the same procedures as explained above in I) to IV) for evaluating the effects of humid gas on crack progress.

2.4 SSRT (Slow strain rate tensile) test

Slow strain rate tensile (SSRT) test for base metal and welded material was conducted in two environments: humid air (HA) and dry nitrogen gas (DNG) to compare the results obtained by HG-SCC test. The detail of the test is indicated in Table 3. The principle of SSRT test for assessing the resistance to HE has been described in detail in many literatures.²³⁾ The test pieces for SSRT tensile tests illustrated in Fig. 3 were cut from central portion in the thickness direction of welded material by a milling machine so that the weld line is perpendicular to the tensile direction. The milled thickness was originally 1.5 mm and then reduced by grinding by abrasive paper (finally by #800) to 1 mm thickness. Then, test piece was etched with alkali solution of 10% NaOH and then de-smut-treated with 10% HNO₃ aqueous solution.

Table 3 SSRT (Slow Strain Rate Tensile) test condition. RH: relative humidity.

Fig. 3

Morphology and dimension of the SSRT test piece in mm.

3. Experimental Results and Discussion

The results of the Vickers hardness test are presented in Fig. 4(a), together with the macroscopic and microscopic images of the weld cross section in Figs. 4(b), 4(c) and 4(d). The solid-plotted curve in the Fig. 4(a) represents the hardness distribution, which starts to increase from the center of the cross-section (minimum) towards the boundary between weld metal and base metal and continues to HAZ area. The thick horizontal line demonstrates the average value of the five base metal measurements. It is observed that within ±9 mm from the center, the hardness becomes smaller than the base metal. However, in the range beyond ±10 mm from weld centerline, the hardness tends to be slightly larger than that of the base metal. From Fig. 4(b), it can be seen that welded metal part turned into black in macroscopic observation, while the base metal section remained white. Figure 4(d) shows that a layered microstructure is formed in the weld metal portion of the weld’s boarder line (weld corner), which is presumed to be caused by columnar growth of preexisting solid (base metal) grain into molten weld metal pool.²⁴⁾

Fig. 4

Features of the weld cross section. (a): Vickers hardness distribution, (b): macroscopic image with the location of indentation, (c): microscopic image of the boxed area in (b), (d) close-up image of the boxed area in (c).

Figures 5(a) to 5(d) show optical microscopic (OM) images of the crack end before and after HG-SCC test for the samples indicated in Figs. 2(I)(a) to 2(I)(d), respectively. Thin arrows point pre-crack ends while the thick arrow indicates HG-SCC crack end. In base metal (a), weld center (b) and HAZ (d), the pre- and post-fatigue cracks are at the same position. Thus, no HG-SCC has propagated in these areas. In contrast, in weld corner (c), adjacent to base metal, crack end is found to extend.

Fig. 5

Surface appearance of test pieces of positions (a), (b), (c) and (d) before (1) and after (2) the HG-SCC test for 90 d. Tip of pre-cracks and HG-SCC crack are indicated by thin and thick arrows, respectively.

In Figs. 6(a) to 6(d), are shown SEM images of the fracture surfaces of the four samples corresponding to Figs. 5(a) to 5(d). In Figs. 6(a), (b) and (d), the pre- and post-crack boarder line is mostly connected with each other by the thin white curves, showing that no HG-SCC crack has progressed, in these three samples, in accord with the results of optical microscopy mentioned above, although a small patch of HG-SCC area is shown in Fig. 6(d) near the surface (bottom of the image indicated by an arrow). In contrast, in weld corner sample (c), HG-SCC crack extension is clearly seen in Fig. 6(c), where the averaged crack propagation length was evaluated to be 2.68 mm, following the HPIS standard,²²⁾ which shows the sensitivity to HG-SCC at this position does not satisfy the qualification (not less than 0.16 mm).

Fig. 6

SEM images of the fracture surface after HG-SCC test in HA environment of the #1 sample of each position in Table 4. (c1) and (c2): close-up images of fatigue and HG-SCC cracks of sample (c), respectively. The top and bottom sides of the photo (c) and bottom side of the photo (d) corresponds to the surface of the test piece.

Although the SEM images are not shown here, observation on the fracture surface also confirmed that the crack does not extend in dry environment. Table 4 shows the summary of the HG-SCC test results for all the tested samples. It can also be seen in Table 4 that the crack does not propagate in dry environment in all the four samples (a), (b), (c) and (d).

Table 4 Summary of the HG-SCC test results. N: Number of the samples tested, Δa: crack propagation length, Δa_m: average of Δa.

Figure 7 shows stress-strain curves both for the welded material and base metal in the two environments, dry nitrogen gas (DNG) and humid air (HA), in an SSRT condition (1.39 × 10⁻⁶ s⁻¹). The mechanical properties of the welded sample as well as the base metal obtained from the SSRT tensile test are indicated in Table 5 together with the value of hydrogen embrittlement sensitivity index,^21–23) I(δ), defined by eq. (3),   

\begin{equation} I(\delta) = \frac{\delta_{\text{D}} - \delta_{\text{H}}}{\delta_{\text{D}}} \end{equation} (3)

where δ_D and δ_H are elongations in DNG and HA, respectively. From these results, it is seen that no hydrogen embrittlement occurs in either material in accord to the previous reports tested by other methods.¹⁾ The welded material shows lower strength than the base metal corresponding to the lower hardness in the central portion of the weld (Fig. 4(a)).

Fig. 7

Stress-strain curves of the specimens SSRT-tested.

Table 5 Mechanical properties and I(δ) of base metal (BM) and weld metal (WM) samples SSRT-tested in HA and DNG environments. σ_u: ultimate tensile strength, σ_0.2: 0.2% proof stress, δ: elongation to failure and I(δ) = (δ_D − δ_H)/δ_D: embrittlement sensitivity index where δ_H and δ_D are elongation to failure in HA and DNG, respectively.

Figure 8 shows the macroscopic view of the fractured test pieces of welded materials in the two environments. All the test pieces were fractured at the center. This can be also attributed to the lowest hardness of this portion. From the results of HG-SCC tests, the weld center area did not show any crack propagation in HA. Thus, the results of SSRT test showing no embrittlement both in the base metal and welded material are in good consistence to the results of HG-SCC for the position (a) and (b), respectively.

Fig. 8

Macroscopic view of the pieces SSRT-tested up to fracture.

The SEM images of the fractured surface of the welded materials SSRT-tested in DNG and HA environments are demonstrated in Fig. 9 at two magnifications. Dimples, evidence of ductile fracture, were seen throughout fracture surfaces, in accordance with high elongation to failure of the two samples showing no HE in SSRT test. Observation was also made on samples tested on base metal samples, but all the resultant images were almost the same as shown in Fig. 9.

Fig. 9

SEM images of the fracture surfaces of the welded materials SSRT-tested in DNG (a1) and HA (b1). (a2) and (b2) are magnified images of (a1) and (b1), respectively.

In this study, it has been revealed that HE in 5083 MIG welds is detectable by means of HG-SCC test, not of SSRT test. The reason for this should be in the fact that HG-SCC test can specify the different locations (weld metal center, weld corner and HAZ), while softest region (weld metal center) always deforms and fractures preferentially in SSRT test. Yet, there remain a few questions: why the crack propagates only in the weld corner, and how welded microstructure affects the crack propagation, etc. The mechanism for the fact that HG-SCC crack propagation occurs only in weld corner is probably based on the macroscopic segregation (the highest Mg concentration reported in the weld corner^12–15)) mentioned in the introduction. This will be investigated experimentally further in detail elsewhere, and may be related with possible β phase precipitation during cooling after solidification.

4. Conclusion

Resistance to hydrogen embrittlement was evaluated on MIG-welded 5083 aluminum plates as well as base metal by means of HG-SCC test in HA using samples with a pre-crack at different welding areas. Test in dry air was also carried out as a reference. Furthermore, SSRT tests on the same materials in HA and DNG were undertaken for comparison purposes. The major conclusions drawn from this study are as follows:

1. (1)    Local σ_0.2 of each area that is needed to fix K_1APP in the HG-SCC test was estimated from Vickers hardness test results.
2. (2)    There was no HG-SCC crack propagation in the weld metal center, HAZ and base metal in HA environment. However, HG-SCC crack was extended in the weld corner sample (about ±7 mm from the weld center), confirming HE. No crack extension was observed in any area when tested in dry air.
3. (3)    By means of SSRT test, neither welded nor base material show any HE, even in HA environment. This was due to the fact that only the softest region is assessed in the SSRT test.
4. (4)    In conclusion, HG-SCC test was revealed to be suitable to assess the sensitivity to HE on the local area, compared to SSRT test.

Acknowledgement

The authors are grateful to UACJ Corporation for their supply of the materials for all the tests. Part of the work was financially supported by Light Metal Education Foundation, Inc., which is highly appreciated.

REFERENCES

• 1)   H.  Yoshida and  T.  Fukui: J. JILM 38 (1988) 496–512.
• 2)  N. Oiwa, T. Iijima, A. Kida and S. Ohga: Proc. 12th Int. Conf. on Aluminum Alloys, (Jpn. Inst. Light Metals, Tokyo, 2010) pp. 82–87.
• 3)   M.  Czechowski: J. KONES 21 (2014) 41–45.
• 4)  Y. Oya and Y. Kojima: 12th Int. conf. on Aluminum Alloys, (Jpn. Inst. Light Metals, Tokyo, 2010) pp. 1508–1512.
• 5)  ASTM B. 928/B 928M-07: Standard Specification for High Magnesium Aluminum-Alloy Sheet and Plate for Marine Service and Similar Environments (2007).
• 6)   R.  Goswami,  G.  Sapons,  P.S.  Paos and  R.L.  Holtz: Mater. Sci. Eng. A 527 (2010) 1089–1095.
• 7)  Microstructure and Properties of Aluminum, (Jpn. Inst. Light Metals, 1991) p. 88.
• 8)   Y.K.  Yang and  T.R.  Allen: Metall. Mater. Trans. A 44 (2013) 5226–5233.
• 9)   Y.K.  Yang and  T.R.  Allen: Mater. Charact. 80 (2013) 76–85.
• 10)   M.  Seifi,  I.  Ghamarian,  P.  Samimi and  P.C.  Collins: Corros. Sci. 138 (2018) 219–241.
• 11)   N.  Winzer,  A.  Atrens,  W.  Dietzel,  V.S.  Raja,  G.  Songe and  K.U.  Kainer: Mater. Sci. Eng. A 488 (2008) 339–351.
• 12)  ASTM G67-18. Standard Test Method for Determining the Susceptibility to Intergranular Corrosion of 5XXX Series Aluminum Alloys, ASTM International (2004).
• 13)   J.A.  Lyndon,  R.K.  Gupta,  M.A.  Gibson and  N.  Birbilis: Corros. Sci. 70 (2013) 290–293.
• 14)  D.E.J. Talbot and J.D.R. Talbot: Corrosion Science and Technology, (CRC Press, 2018) pp. 248–252.
• 15)   S.  Mukae,  M.  Katoh,  K.  Nishio and  K.  Kusaba: J. Jpn. Weld. Soc. 46 (1977) 158–165.
• 16)   T.  Ohnishi: J. JILM 27 (1977) 473–479.
• 17)   R.E.  Ricker and  D.J.  Duquette: Metall. Trans. A 19 (1988) 1775–1783.
• 18)   J.R.  Pickens,  J.R.  Gordon and  J.A.S.  Green: Metall. Trans. A 14 (1983) 925–930.
• 19)   S.J.  Kim,  S.K.  Kim and  J.C.  Park: Surf. Coat. Tech. 205 (2010) S73–S78.
• 20)  High Pressure Institute of Japan: Standard Test Method for Humid Gas Stress Corrosion Cracking of Aluminium Alloys for Compressed Hydrogen Containers, HPISE 103:2018 (2018).
• 21)   S.  Osaki,  H.  Kondo and  K.  Kinoshita: Mater. Trans. 47 (2006) 1127–1134.
• 22)   W.  Gao,  D.  Wang,  M.  Seifi and  J.J.  Lewandowski: Mater. Sci. Eng. A 730 (2018) 367–379.
• 23)  G. Itoh, M. Watanabe, K. Kuroyanagi and P. Zhao: 12th Int. Conf. on Aluminum Alloys, (Jpn. Inst. Light Metals, Tokyo, 2010) pp. 377–380.
• 24)  M.F. Ashby and D.R.H. Jones: Engineering Materials (An Introduction to Their Properties and Applications), Translated by R. Horiuchi et al., (Uchida Rokakuho, Tokyo, 1985) pp. 97–98.