Materials Transactions, Vol. 52, No. 2 (2011) pp. 254 to 257 #2011 The Japan Institute of Metals
RAPID PUBLICATION
Liquation Cracking of Dissimilar Aluminum Alloys during Friction Stir Welding Sang-Woo Song1 , Sang-Hoon Lee1 , Byung-Chul Kim2 , Tae-Jin Yoon2; *1 , Nam-Kyu Kim1 , In-Bae Kim3 and Chung-Yun Kang2;3; *2 1 Industrial Technology Support Division, Korea Institute of Materials Science (KIMS), 66 Sangnamdong, Changwon, Kyungnam 641-010, Korea 2 National Core Research Center (NCRC) for Hybrid Material Solution, Pusan National University, Geumjeoung-gu, Busan 609-735, Korea 3 Department of Material Science & Engineering, Pusan National University, Geumjeoung-gu, Busan 609-735, Korea
A liquation cracking mechanism of dissimilar Al alloys during the friction stir welding (FSW) is suggested in this study. To identify the mechanism, the precipitates were analyzed and Al-Mg-Cu phase diagrams were calculated. Electron backscattering diffraction (EBSD) analysis and electron probe microanalysis (EPMA) were also conducted. In the same manner as constitutional liquation, at high heating rate, the main liquation-inducing precipitates were not dissolved in the matrix and reacted with Al to form the partially melted zone (PMZ), after which liquation cracking occurred where strain was applied to the PMZ. [doi:10.2320/matertrans.M2010343] (Received September 30, 2010; Accepted November 22, 2010; Published January 25, 2011) Keywords: friction stir welding, liquation, partial melting, dissimilar aluminum alloys
1.
Introduction
Since The Welding Institute (TWI) of the United Kingdom developed the friction stir welding (FSW) in 1991, it has primarily been applied to aluminum alloys and identified as an adequate technology for joining aluminum alloys, including similar and dissimilar alloys, while offering versatility, environmental friendliness and energy efficiency. In general, most aluminum alloys are susceptible to liquation cracking during conventional fusion welding because of their wide partially melted zone (PMZ), large solidification shrinkage, large thermal contraction and residual intermetallic compounds.1) Therefore, many studies have examined the liquation cracking or partial melting in the heat-affected zone (HAZ) during conventional fusion welding of aluminum alloys.2–4) For liquation cracking during FSW, most studies have focused on friction stir spot welding of magnesium alloys other than friction stir seam welding of aluminum alloys. They have reported that the liquation cracking occurred in the stir zone itself, in the TMAZ region at the extremity of the stir zone and beneath the tool shoulder.5–8) For friction stir seam welding of aluminum alloys, only a few studies have been carried out and they reported that liquation or melted film formation other than liquation cracking have been observed and suggested in 2024 and 7010 alloys.9,10) If liquation cracking occurs during FSW of specific Al alloys, it must be considered to control and choose welding parameters, such as tool dimension, backing material, cooling device, tool rotation speed, and peak temperature, to reduce its incidence. It was reported that the maximum temperature within the weld zone was determined to be over 450550 C and that the maximum temperature rise occurred at the top surface of the weld zone.11) These temperatures are below the melting *1Graduate
Student, Pusan National University author, E-mail:
[email protected]
*2Corresponding
point of aluminum, but sufficiently high to melt various precipitates, such as Al8 Mg5 , Al6 CuMg4 , Al2 CuMg, that remain in the Al-Mg-Cu alloy matrix.12,13) Partial melting of the matrix alone, however, is insufficient to cause liquation cracking, which therefore occurs when both partial melting and strain are applied simultaneously. Arora et al. reported that the computed strains and strain rates during FSW were in the ranges of 10 to 5 and 9 to 9 s1 , respectively,14) and hence these strains would contribute to liquation cracking in the PMZ. In our previous study, we found evidence for liquation cracking during FSW of dissimilar aluminum alloys (5052KS5J32 joint) and believe this to be the first case of liquation cracking in the interior of welds during friction stir seam welding of aluminum alloys.15) Therefore, with the goal of identifying the mechanism of liquation cracking during FSW, additional experiments were performed, such as analysis of precipitates in the base materials, measurement of peak temperatures, calculation of the Al-Mg-Cu ternary phase diagram, and analysis of scanning electron microscopy (SEM), electron backscattering diffraction (EBSD) and electron probe microanalysis (EPMA) results. 2.
Experimental Procedure
The base materials used were 5052 aluminum alloy (Al-2.5%Mg) and KS5J32 aluminum alloy (Al-5.5%Mg0.3%Cu) rolled sheets, which are widely used as the inner panels of automobile doors. KS5J32 is the alloy code number of Kobe steel, which corresponds to the 5023 designation by the aluminum association.16) It is produced by using high purity primary aluminum and contains the optimum Cu content for high formability and medium strength after paint baking.17) The chemical compositions are listed in Table 1. The dissimilar 5052-KS5J32 joints were produced perpendicularly to the rolling direction with a tool rotation speed
Liquation Cracking of Dissimilar Aluminum Alloys during Friction Stir Welding
Cu
Si
Fe
Mn
Zn
Cr
Ti
(a)
5052-H34
2.51
—
0.28
0.34
0.16
0.08
0.15
0.02
Bal.
KS5J32-T4
5.68
0.20
0.11
0.09
0.01
0.03
—
0.02
Bal.
of 1500 rpm. The welding speeds were varied within the range from 100 to 700 mm/min and the KS5J32 aluminum alloy was placed on the retreating side of the tool. A lefthanded threaded tool with a probe diameter of 3.8 mm, shoulder diameter of 12 mm and probe length of 1.45 mm was used. This tool had a slightly larger shoulder and probe than those of conventional tools, of which the ratio of probe diameter to depth is approximately 1 : 1. The tool-to-workpiece angle was 3 from the vertical axis of the welds. The tool was rotated in the clockwise direction and the workpieces were tightly fixed to the backing plate. During the FSW process, the peak temperature was measured beneath the edge of the tool shoulder on the retreating side (KS5J32) using a 1.0-mm-diameter, stainless steel, sheathed thermocouple. Following the FSW, the joints were cross sectioned perpendicular to the welding direction. The cross sections were polished with a standard technique, etched with Keller’s etchant and observed by optical microscopy. Fractographic observations were conducted by SEM. Moreover, precipitates in the base material were analyzed by a Hitachi S-4800 SEM equipped with an energy dispersive X-ray spectroscopy (EDXS) analysis system. The ternary Al-Mg-Cu phase diagram was calculated by using Thermo-Calc software, EBSD analysis was performed using JSM7001F SEM equipped with HKL channel 5 EBSD system, and EPMA was also performed using Cameca SX 100. 3.
(b)
Al
Result and Discussion
5µm
10µm
(c)
(d)
5µm
5µm
Fig. 1 Fractographs of crack surfaces at various welding speeds: (a) 100 mm/min, (b) 300 mm/min, (c) 500 mm/min, and (d) 700 mm/min.
600
400 300 200 100 0
0
20
40
60 80 Time (sec)
100
120
140
Fig. 2 Peak temperature measured beneath the edge of the tool shoulder on the retreating side (KS5J32) during FSW (1500 rpm, 400 mm/min).
α: FCC, β: Al8Mg5, θ: Al2Cu, T: Al6CuMg4 ,S: Al2MgCu 700 650
+ : Al-5.5Mg-0.3Cu 300°C
α+β+T
L
α+S+T
550
α+T
L+α
500
L+α+T
α
450 400
α+S+T
350 α+θ 300
400°C
α+T
600 Temperature(°C)
Figure 1 shows fractographs of cracks occurring at various welding speeds with a rotation speed of 1500 rpm. At all welding speeds, the cracks were found beneath the tool shoulder on the retreating side, i.e., in the KS5J32 alloy, as shown in Fig. 5(a). Their fractographs show many protuberances that are somewhat similar to those of dendrites and they indicate the existence of liquid among the grain boundaries and are evidence of resolidified material.18) The typical surface-resolidified material evident in these fractographs is indicative of liquation cracking during FSW. Therefore, the peak temperatures were expected to reach a reaction temperature of liquation-inducing precipitates at all welding speeds with a rotation speed of 1500 rpm. The peak temperature measured close to the edge of the tool shoulder, where liquation cracking occurred, was about 480 C at a rotation speed of 1500 rpm and a welding speed of 400 mm/min (Fig. 2). This was much lower than the melting points of the 5052 and KS5J32 alloys, but high enough to cause reactions between precipitates and the Al-matrix, e.g., eutectic or peritectic reaction. The 5052 alloy contains various precipitates, such as Mg2 Si and Al6 (FeMn), but their reaction temperatures with Al-matrix are much higher than 480 C. For KS5J32, that is
Peak: 481°C
500
α+S+T
wt. % Mg
Mg
wt. % Mg
Materials
Chemical compositions of the base materials (mass%).
Temperature (°C)
Table 1
255
α+S
L+α+β α+T
α+β+T
α+S
250
α+S α
α+θ
α+θ+S
α
α+θ+S α+θ
wt. % Mg
wt. % Cu
wt. % Cu
(a)
(b)
(c)
Fig. 3 Calculated Al-Mg-Cu ternary phase diagrams: (a) isopleth of AlMg-0.3%Cu, (b) isothermal section at 300 C, and (c) isothermal section at 400 C.
Al-5.5%Mg-0.3%Cu alloy, it had been reported that the binary phases Al2 Cu () and Al8 Mg5 () and the ternary phases Al6 CuMg4 (T) and Al2 CuMg (S) could be in equilibrium with the Al matrix.12,13) To obtain more information for the given composition in the Al-rich corner, the Al-Mg-Cu ternary phase diagram was calculated and is shown in Fig. 3. Figure 3(a) illustrates a vertical section (isopleth) at 0.3 mass% Cu from the computed Al-Mg-Cu ternary phase diagram. Figures 3(b) and 3(c) show isothermal sections at 300 C and 400 C, respectively, and provide information on
256
S.-W. Song et al. (a)
(b)
(c)
(d)
1µm
200nm
1µm
1µm
(e)
(f)
(g)
(h)
1µm
1µm
1µm
1µm
Fig. 4 SEM investigation of the precipitates in the KS5J32 alloy.
(a)
Advancing side 5052
Retreating side KS5J32
1mm
(d)
(c)
(b)
Al 200µm
(e)
Mg 200µm
Cu 200µm
200µm
Fig. 5
Analysis of cross-sectional area: (a) optical macrostructure, (b) EBSD analysis, and (c), (d) and (e) EPMA elemental maps.
the stable phases at the given compositions and temperatures. The dashed line and ‘+’ marks indicate the composition of KS5J32 (Al-5.5%Mg-0.3%Cu), which lies in the or þ T region depending on the temperature. As shown in Fig. 3, Al6 CuMg4 (T) is a stable phase up to the solvus temperature ( 360 C), above which it is dissolved in the Al matrix under equilibrium condition. According to the previous studies,1–4) however, under high heating rate, as is often the case in the HAZ of the welding process, the precipitates do not have enough time to dissolve completely in the Al matrix and remain in the Al matrix above the solvus temperature. Consequently, upon heating to the reaction temperature, the precipitates react with the surrounding Al matrix and form the liquid phase. For the case of Al6 CuMg4 , it reacts with the Al matrix around 470 C, as follows:12,13,19,20) (Al) + Al6 CuMg4 (T) ! L + Al2 CuMg (S). Although the KS5J32 alloy sheet was solid-solution treated to dissolve precipitates after the rolling process, many precipitates remained in the matrix, such as Al8 Mg5 (), Mg2 Si, Al2 CuMg (S), Al6 CuMg4 (T), Al6 (CuFe) (D), and Al7 Cu2 Fe (N). Figure 4 shows various precipitates in the KS5J32 alloy and the EDXS results for the precipitates are listed in Table 2. Figure 4 and Table 2 reveal the KS5J32
alloy to be comprised of Mg/Si-rich, Cu-rich or Fe-rich precipitates. The Mg/Si-rich precipitates appeared to be Mg2 Si phase (63.2% Mg, 36.8% Si in mass%) due to the absence of any ternary phase in the Al-Mg-Si system. The Fe-rich precipitates appeared to be Al6 (CuFe) phase (7% Cu, 24.6% Fe in mass%). Most of the Cu-rich precipitates appeared to be Al6 CuMg4 phase (22–27% Cu, 27.5–30% Mg in mass%) but a few appeared to be Al2 CuMg phase (46% Cu, 17% Mg in mass%). Al2 CuMg, Mg2 Si and Al6 (CuFe) phases react with the Al matrix at 518 C, 555 C and 620 C, respectively,12,13) but the peak temperature during FSW did not reach such temperatures in this study. In contrast, Al6 CuMg4 phase reacts with the Al matrix at relatively low temperature ( 470 C), and can therefore form the PMZ during FSW as follows: (Al) + Al6 CuMg4 (T) ! L + Al2 CuMg (S). Therefore, Al6 CuMg4 (T) was the main liquation-inducing precipitate reacting with the Al matrix, which is consistent with the ternary phase diagram in this study (Fig. 3) and that of the previous studies.21,22) Figure 5(a) shows the location of the liquation cracking, which occurred during FSW, near the top surface of the stir zone on the retreating side (KS5J32). Because the exact location could not be identified by optical microscope with
Liquation Cracking of Dissimilar Aluminum Alloys during Friction Stir Welding Table 2
EDXS analysis results for the precipitates in Fig. 4.
257
43.45 (44.48) 81.09 (72.68)
9.25 (19.25)
33.07 (30.51) 9.66 (7.80)
(c)
70.06 (62.88)
9.89 (20.90)
20.05 (16.22)
—
—
Al6 CuMg4 (T)
(d)
75.97 (73.41)
4.03 (9.18)
20.00 (17.42)
—
—
Al6 CuMg4 (T)
(e)
79.62 (65.05)
2.12 (4.08)
—
—
18.26 (30.87)
Al6 (CuFe) (D)
38.04 (41.22)
—
Mg2 Si
EMPA were also conducted to determine where the liquation cracking occurred. Although the KS5J32 alloy was solid-solution treated to dissolve precipitates after the rolling process, many precipitates remained in the matrix, including Mg2 Si, Al6 CuMg4 (T), and Al6 (CuFe) (D). According to the ternary phase diagram, Al6 CuMg4 (T) is a stable phase at room temperature and reacts with the Al matrix at around 470 C. The main liquation-inducing precipitate was Al6 CuMg4 (T), which formed the PMZ (constitutional liquation) at around 480 C during the FSW process. Finally, the strain produced by the rotation of the tool was applied to the PMZ, causing liquation cracking to arise in the interior of the material flows, rather than at the interface between unaffected base metal and flows.
—
—
Al6 CuMg4 (T)
Acknowledgments
—
—
Al6 CuMg4 (T)
The authors thank Dr. Chang-Seok Oh for the calculation of the phase diagram and Dr. Jae-Hyung Cho for the EBSD analysis. This work was supported by a grant-in-aid for the National Core Research Center Program from MOST/ KOSEF No. R15-2006-022-02004-0 and 70000521.
atomic% (mass%) No (a) (b)
(f) (g) (h)
Al
Cu —
Mg
Si
Fe
Possible Phase
23.47 (25.01)
—
Mg2 Si
—
—
Al2 CuMg (S)
6.54 (6.81) 67.57 (58.99)
12.21 (25.10)
55.42 (51.97) 20.22 (15.91)
69.83 (61.78)
11.01 (22.95)
19.16 (15.27)
—
conventional technique, EBSD analysis was conducted. As shown in Fig. 5(b), the crack was located within the finegrained zone broken by the rotation of the tool, which means that liquation cracking occurred in the interior of the material flow induced by tool rotations, rather than at the interface between the unaffected base metal and the material flows. This indicated that the crack was not a kind of void due to inadequate filling of materials. However, it remained uncertain as to whether liquation cracking occurred within KS5J32 or 5052 in Figs. 5(a) and 5(b). Figures 5(c), 5(d) and 5(e) present the EPMA results at the same location as the EBSD analysis. Due to the much higher Mg content in KS5J32 ( 5:5% Mg) than in 5052 ( 2:5% Mg), the two alloys can be easily differentiated in the EPMA elemental maps. It can be seen from these figures that the yellow, light blue and purple areas indicate the KS5J32 alloy, and, therefore, that the cracking occurred within KS5J32, and not at the interface or within 5052. From the results, we propose the following liquation cracking mechanism in KS5J32-5052 dissimilar aluminum alloys during FSW. (1) The base material contains many liquation-inducing precipitates, i.e., low-melting particles. (2) The main liquation-inducing precipitate is Al6 CuMg4 (T phase) in the KS5J32 alloy. (3) The temperature increases rapidly up to 480 C during FSW. (4) Al6 CuMg4 (T phase) is not dissolved and remains in the matrix, because of high heating rate, although it is within the solid solubility limit. (5) Residual precipitates react with the matrix (constitutional liquation) and form the PMZ: Al + Al6 CuMg4 (T) ! L + Al2 CuMg (S) around 470 C. (6) Finally, liquation cracking occurs when strain is applied to the PMZ. 4.
Conclusions
In summary, since liquation cracking during FSW can cause severe loss of mechanical properties, such as strength and ductility, a fundamental study was conducted to examine the mechanism involved. To identify the liquation mechanism, the compositions of the precipitates were examined and ternary phase diagrams were calculated. EBSD analysis and
REFERENCES 1) S. Kou: Welding Metallurgy, 2nd Edition, (Wiley-Interscience, NewJersey, 2002) pp. 304–328. 2) C. Huang and S. Kou: Weld. J. 81 (2002) 211S–222S. 3) C. Huang and S. Kou: Weld. J. 82 (2003) 184S–194S. 4) C. Huang and S. Kou: Weld. J. 83 (2004) 111S–122S. 5) A. Gerlich, M. Yamamoto and T. H. North: J. Mater. Sci. 43 (2008) 2–11. 6) A. Gerlich, G. Avramovic-Cingara and T. H. North: Metall. Mater. Trans. A 37 (2006) 2773–2786. 7) M. Yamamoto, A. Gerlich, T. H. North and K. Shinozaki: J. Mater. Sci. 42 (2007) 7657–7666. 8) Y. K. Yang, H. Dong, H. Cao, Y. A. Chang and S. Kou: Weld. J. 87 (2008) 167S–177S. 9) C. Dalle-Donne, B. Braun, G. Staniek, A. Jung and W. A. Kayser: Materialwiss. Werkstofftech. 29 (1998) 609–617. 10) Kh. A. A. Hassan, P. B. Prangnell, A. F. Norman, D. A. Price and S. W. Williams: Sci. Technol. Weld. Joining 8 (2003) 257–268. 11) R. S. Mishra and Z. Y. Ma: Mater. Sci. Eng. R 50 (2005) 1–78. 12) L. F. Mondolfo: Aluminum Alloys: Structure and Properties, (Butterworth & Co. Ltd., London, 1976) pp. 497–504. 13) N. A. Belov, D. G. Eskin and A. A. Aksenov: Multicomponent Phase Diagram: Applications for Commercial Aluminum Alloys, (Elsevier, Oxford, 2005) pp. 85–88. 14) A. Arora, Z. Zhang, A. Deb and T. DebRoy: Scr. Mater. 61 (2009) 863– 866. 15) S. W. Song, B. C. Kim, T. J. Yoon, N. K. Kim, I. B. Kim and C. Y. Kang: Mater. Trans. 51 (2010) 1319–1325. 16) T. Inaba, K. Tokuda, H. Yamashita, Y. Takebayashi and T. Minoura: Kobelco Technol. Rev. 26 (2005) 56–62. 17) T. Inaba: Automobile Aluminum Sheet, Automotive engineering: lightweight, functional, and novel materials, ed. by B. Cantor, P. Grant and C. Johnston (Taylor & Francis Group, New York, 2008) pp. 19–28. 18) F. Matsuda and H. Nakagawa: Trans. JWRI 6 (1977) 81–90. 19) S. L. Chen, Y. Zuo, H. Liang and Y. A. Chang: Metall. Mater. Trans. A 28 (1997) 435–446. 20) T. Buhler, S. G. Fries, P. J. Spencer and H. L. Lukas: J. Phase Equilib. 19 (1998) 317–329. 21) G. B. Brook: Precipitation in Metals, Special Report No. 3, (Fulmer Research Institute, UK, 1963). 22) P. Ratchev, B. Verlinden, P. D. Smet and P. V. Houtte: Mater. Trans. JIM 40 (1999) 34–41.
