CO2 laser beam welding of dissimilar magnesium-based alloys

CO2 laser beam welding of dissimilar magnesium-based alloys

Materials Science and Engineering A 496 (2008) 45–51 Contents lists available at ScienceDirect Materials Science and Engineering A journal homepage:...

3MB Sizes 2 Downloads 10 Views

Materials Science and Engineering A 496 (2008) 45–51

Contents lists available at ScienceDirect

Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea

CO2 laser beam welding of dissimilar magnesium-based alloys Yajie Quan ∗ , Zhenhua Chen, Xiaosan Gong, Zhaohui Yu School of Materials Science and Engineering, Hunan University, Changsha 410082, China

a r t i c l e

i n f o

Article history: Received 14 December 2007 Received in revised form 21 April 2008 Accepted 23 April 2008 Keywords: Laser welding Magnesium alloys Microstructure Mechanical property Dissimilar metals welding

a b s t r a c t In this paper, dissimilar metals welding for butting magnesium alloys AZ31, AM60 and ZK60 was conducted by a 3 kW CO2 laser beam. The microstructure and mechanical properties of joints were analyzed by optical microscope (OM), energy dispersive spectrometer (EDS), scanning electron microscope (SEM), tensile machine and hardness machine. The experimental results show that the welding heat input and the chemical composition of base metal have great influences on the formation of weld bead. The results of tensile test show that the ultimate tensile strength (UTS) of the optimum joints for dissimilar metals welding is above 90% of the base metals and even higher than one of the base metals. Many precipitates distributed in the fusion zone, whose number is related to the alloy elements. The hardness test indicates that a sudden decrease of microhardness occurs in HAZ for sides of all the metals, and the hardness in the fusion zone of AZ31–AM60 joint is slightly higher than the others. The elements analysis reveals that there is an obvious compositional gradient only in the fusion zone of AM60–ZK60 joints. © 2008 Elsevier B.V. All rights reserved.

1. Introduction As an extremely light metal, magnesium alloys have recently becoming one of the most important structure materials because of its high specific strength, good castability, and superior properties for absorbing vibration and insulating electromagnetic interference [1]. However, applications of magnesium alloys are restricted by their poor formability. If the shape-complicated work piece can be divided into several easily made parts and then combined into one under appropriate welding conditions, this method will effectively improve these problems. Magnesium alloys can be jointed by usual welding methods, but their low melting point, high chemical activity, and high thermal conduction require to be welded at high power with adequate protection. Compared with conventional methods, laser beam welding has much higher power density and depth-to-width ratio. Although the fusion penetration depth in thicker workpieces is restricted by the energy limitation of laser beam, it is still the first choice for welding magnesium alloys due to its cost-effective applications, especially for thin plates [2–5]. Among many Mg-based alloys, the AZ-(Mg–Al–Zn), AM(Mg–Al–Mn), and ZK-(Mg–Zn–Zr)-based alloys seem to be most popular, with the AZ31 (AZ61, AZ91), AM60 and ZK60 alloys occupying the highest market. As different magnesium alloys are used within a given structure, there will definitely be a need to somehow

∗ Corresponding author. Tel.: +86 731 8884206; fax: +86 731 8821648. E-mail address: quanyj [email protected] (Y. Quan). 0921-5093/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2008.04.065

join them. But research in the field of dissimilar metals welding for butting magnesium alloys with different compositions is tiny and under development. Chi et al. [6] concentrated their discussion on electron beam welding. Only Weisheit et al. [7] have reported on a part work of laser beam welding of dissimilar magnesium alloys (such as QE22, ZC63 and WE54), but their works still focused on the similar metals. So the present paper will give an overview of the results of investigations on the weldability of dissimilar metals welding for butting magnesium alloys (AZ31, AM60 and ZK60) with a CO2 laser beam. The investigations focused on macro- and microstructure analysis and mechanical property test. 2. Experimental procedure Three kinds of homemade wrought magnesium alloys (AZ31, AM60 and ZK60) were chosen for the welding experiments. Butt welds were carried out on rectangular plates with the size of 50 mm × 30 mm × 2 mm. The nominal compositions of the samples are listed in Table 1. The top surface of each specimen was cleaned with acetone to remove grease and residue, and brushed with stainless steel wire to remove oxides before welding. Welding was conducted without filler metal using a 3.0 kW continuous wave CO2 laser with the following properties: mode, TEM01 ; divergence, <2 mrad; beam diameter on focusing optic, 38 mm; focused diameter, 0.25 mm. The focal length of the parabolic mirror used for focusing the laser beam was 127 mm. During the welding process, high pure argon was used to protect the top and back of the weld. The laser power used was 1.0 kW, and the welding speeds were 2.0, 2.5, 3.0 and 3.5 m/min, respectively. The

46

Y. Quan et al. / Materials Science and Engineering A 496 (2008) 45–51 Table 1 Nominal composition of base metal (wt.%)

Fig. 1. Geometry and position of tensile specimens.

beam defocusing was 0 mm, namely the focal point was adjusted on the surface of workpiece, because the plates were only 2 mm in thickness. So, a total of 12 sets of specimens were welded under various conditions.

Alloys

Al

Zn

Mn

Zr

Mg

AZ31 AM60 ZK60

3.0 6.0

1.0 – 5.5

0.2 0.13 –

– – 0.45

Balance Balance Balance

After welding, the weld quality was visually assessed, then, only the specimens free of surface defects were machined into tensile samples, in the form of a gauge section 15 mm long and 4 mm wide, as illustrated in Fig. 1. Before the joints being machined, the face and backside of the weld was minimally ground in order to achieve a smooth surface. Each data point of ultimate tensile strength (UTS) represented an average of three samples, which were tensioned to fracture in a hydraulic tensile machine at a cross-head velocity of 5 × 10−2 mm/s. Subsequently a pair of specimens with higher UTS under the same welding speed was cross-sectioned, ground and polished for metallurgical examination. The mounted specimens were etched in a solution comprised of 5 ml acetic acid + 5 g picric acid + 10 ml water and 100 ml ethyl alcohol for 20–60 s until the microstructure was revealed. The microstructure and microhardness distribution were characterized by optical microscope (OM) or scanning electron microscope (SEM). The precipitated phase was determined through X-ray diffraction (XRD) analysis.

Fig. 2. Cross-section of the joints under different welding speeds.

Y. Quan et al. / Materials Science and Engineering A 496 (2008) 45–51

3. Results and discussion 3.1. Weld appearance Under all the welding speeds, the three pairs of magnesium alloys (AZ31–AM60, AM60–ZK60 and AZ31–ZK60) were jointed successfully. The welds have complete fusion penetrations and the top surface appearances are smooth with regular ripples. Fig. 2 shows the metallographic photographs of weld cross-sections with dissimilar metals welding for alloys AZ31, AM60 and ZK60. When the welding speed is at 2 m/min, weld beads of three pairs of materials are relatively wide, which have a uniform width on the top, the middle, and at the bottom (see Fig. 2(a), (e) and (i)). And there are craters occurred in the fusion zone caused by vaporization due to the over high heat input, especially in Fig. 2(a) and (e). With increasing the welding speed, the average width of all weld beads decreases. But, it is harmful to cause seriously delaminating in fusion zone at higher welding speeds (Fig. 2(h) and (l)). On the base

47

of equation (∂T/∂t)x = −2kv(T − T0 )2 /P [8], it can be concluded that the cooling rate increases with increasing the welding speed, so the melt alloys have not enough time to mix homogeneously. However, this phenomenon is not obvious for AZ31–AM60 as for the other two, which can be attributed to the minimum component difference between AZ31 and AM60. Therefore, not only the cooling rate but also the chemical composition has influence on the weld formability. 3.2. Tensile test In this study, the tensile test is the most convenient way to overall evaluate the welding quality under different parameters. The UTS of the base material and the welded joint are shown in Table 2. These results indicate that (i) the UTS have the tendency to increase at first and then decrease when the welding speed increases. (ii) The UTS of the welded joints are slightly lower than that of the base metal ZK60, when ZK60 is jointed to the other two alloys. (iii)

Fig. 3. Micro-photographs of transition zone: (a and b) AZ31–AM60, (c and d) AM60–ZK60 and (e and f) AZ31–ZK60.

48

Y. Quan et al. / Materials Science and Engineering A 496 (2008) 45–51

Fig. 4. OM photographs of fusion zone: (a) AZ31–AM60, (b) AM60–ZK60 and (c) AZ31–ZK60.

Fig. 5. SEM photographs of precipitates in fusion zone: (a) AZ31–AM60, (b) AM60–ZK60 and (c) AZ31–ZK60.

Fig. 6. Shrinkage porosities in fusion zone: (a) AZ31–AM60, (b) AM60–ZK60 and (c) AZ31–ZK60.

Fig. 7. Binary phase diagram: (a) Mg–Al and (b) Mg–Zn.

No matter what kind of magnesium alloy is welded to AZ31, the UTS of the joints are all higher than that of the base metal AZ31. On the one hand, these phenomena can be explained by the results of macro- and microstructure observations. Firstly, when the welding speed is too low, craters occurred in the fusion zone (see Fig. 2(a) and (e)). On the contrary, the over high welding speed will cause non-uniform fusion, as shown in Fig. 2(d), (h), (k) and (l). As we all know, these defects will lead to severe stress concentra-

tion. Secondly, the grains in fusion zone get coarser with decreasing the welding speed. The literature [6] has reported that the harmful influence of stress concentration and grain coarsening on the UTS reached at least 13% and 9%, respectively. On the other hand, we can analyze the results according to the change of the chemical composition in the joints. When ZK60 is jointed with another magnesium alloy, the average content of Zn in the weld seam will excess 1%. It is reported that Zn content, in

Y. Quan et al. / Materials Science and Engineering A 496 (2008) 45–51

49

Table 2 UTS of the base material and the welded joint Materials

AZ31 AM60 ZK60 AZ31–AM60 AM60–ZK60 AZ31–ZK60

UTS (MPa) v = 2 m/min

v = 2.5 m/min

272.0 309.0 281.6

262.0 (base metal) 328.7 (base metal) 360.9 (base metal) 294.8 322.0 314.2 341.9 275.6 314.2

v = 3 m/min

v = 3.5 m/min

327.5 321.7 307.1

excess of 1%, increases hot shortness, which may cause weld cracking to weak the joints [9]. Moreover, in ZK60 alloy, the content of high vapor pressure element Zn is relatively high, which is more easily evaporated during the welding process. The Zn loss will also decrease the UTS of the joints. When one of the base metals is AZ31, after welding the total content of alloy elements in fusion zone increases, which will strengthen the welded joints due to solid solution or precipitated particles. With regard to the strengthening mechanism, much more researches still need to be followed. 3.3. Microstructure observation

Fig. 8. X-ray diffraction pattern of fusion zone: (a) AZ31–AM60, (b) AM60–ZK60 and (c) AZ31–ZK60.

According to the results of tensile test, the welded samples with welding speed at 3 mm/min were chosen to observe the microstructures. During welding, partially melting zone (PMZ) is the portion of the base metal immediately adjacent to the complete fusion boundary and is heated to the temperature range between the liquidus and solidus temperature. Fig. 3(a)–(f), respectively shows the microstructure of the transition zone for sides of AZ31, AM60 and ZK60. It is found that there is no apparent PMZ only in the base metal of AZ31 (see Fig. 3(a) and (b)). In that of AM60 and ZK60, the fusion line extends into the matrix and appears to be a partially liquated zone near the fusion boundary. According to the phase diagrams of Mg–Al and Mg–Zn [10], the liquid + solid dual-phase temperature range of AM60 and ZK60 is wider than that of AZ31. At the same time, these two magnesium alloys have more content of alloy elements, which is apt to form precipitates. Therefore, the low melting point nature of the precipitated particles, coupled with the high thermal conductivity and low heat capacity of Mg-based alloys, lead to the wide PMZ, as shown in Fig. 3(c) and (d) (for the side of AM60)and Fig. 3(e) and (f) (for the side of ZK60). Fig. 4 displays the optical metallographs of the fusion centre. Many precipitates concentrate in the fusion zone, in a distribution that tends to grow from a few scattered particles to densely packed coarser ones as the total content of Al increases, which is obvious in the magnifying photographs as presented in Fig. 5. Observed from Fig. 6, it is found that some shrinkage porosities are appeared at the interface between the precipitates and matrix, whose number decreases as the total content of Al decreases. Therefore, it can be concluded that Al has the main influence on the formation of particles. This can be explained by the difference in solubility limits of different alloy elements in magnesium. Based on the Mg–Al and Mg–Zn binary phase diagram in Fig. 7, the solubility limit of Al decreases more quickly than that of Zn with decreasing the temperature, which will lead to that Al is the main element contained in the precipitated particles. This is in agreement with the results of XRD, which indicate these precipitates mainly are Mg17 Al12 particles (see Fig. 8). Energy dispersive spectrometer (EDS) for the precipitates indicates that the content of Zn element in the particles as shown in Fig. 5(b) and (c) is higher than in the matrix. It can be concluded that when ZK60 is welded to the Mg–Al alloys, the content of Zn element in the fusion zone is higher than the

50

Y. Quan et al. / Materials Science and Engineering A 496 (2008) 45–51

Fig. 9. Back-scattered electron images and compositional maps in fusion zone: (a) AZ31–AM60, (b) AM60–ZK60 and (c) AZ31–ZK60.

latter and part of the Al may be replaced with Zn for Mg17 Al12 , indicated by Mg17 (Al,Zn)12 , which was also reported in literature [6,11]. Because of the low melting point of the intermetallics, they may lead to shrinkage at the grain boundary solidified at last when the liquid fail to accommodate high tensile stress induced by thermal cycles and high external constraints during the rapid cooling.

shows a concentration gradient in the chemical composition. There is an Al-rich band on the side of AM60 and a Zn-rich one on the other side, which indicates that the distributions of Al and Zn are obviously non-uniform in fusion zone of AM60–ZK60. This may be explained by the marked difference in physical property resulting from the large difference of chemical composition between AM60 and ZK60.

3.4. Composition analysis

3.5. Microhardness test

The compositional maps from EDS analyses are shown in Fig. 9. The regions measured included the whole weld cross-section, as shown in the images on the left. The other images show the elemental distributions for Mg, Al and Zn. In the compositional maps, bright colors indicate regions of high concentration. Only Fig. 9(b)

Fig. 10 presents the result of microindentation hardness testing on the middle of the weld cross-section. A sudden decrease of microhardness occurs in HAZ for sides of all the metals because it is affected by annealing softening. In fusion zone, the hardness of three kinds of joints all decrease as compared with the base metal,

Fig. 10. Microhardness distribution of welded joints.

Y. Quan et al. / Materials Science and Engineering A 496 (2008) 45–51

but the value of AZ31–AM60 is slightly higher than the others, which is in agreement with the distribution of brittle precipitates in fusion zone. 4. Conclusion A 3 kW CO2 laser beam was used to weld the homemade magnesium alloys AZ31, AM60 and ZK60. The welds were subjected to a number of metallographic, tensile and microhardness examinations in order to evaluate the weldability of dissimilar metals welding for butting magnesium alloys AZ31, AM60 and ZK60. This study yielded the following results: (1) When the welding speed is at 2 m/min, all the weld beads are relatively wide, but when it is at 3.5 m/min, seriously delaminating occurred in the fusion zone, especially one of the base metals is ZK60. (2) When ZK60 is jointed to the other two alloys, the UTS of the welded joints are slightly lower than that of ZK60. When AZ31 is one of the base metals, the UTS of the joints are all higher than that of AZ31. (3) There is no apparent partially melting zone in the base metal of AZ31. On the side of AM60 or ZK60, the fusion line extends into the matrix, resulting in a wide partially liquated zone near the fusion boundary. In the fusion zone, many precipitates concentrate in a distribution that tends to grow from a few scattered particles to densely packed coarser ones as the total content of Al increases.

51

(4) The composition analyses indicate that the distribution of main alloy element Al or Zn is uniform in the fusion zone except for AM60–ZK60, which has an obvious concentration gradient in the chemical composition Al and Zn. (5) The results of microhardness tests show a sudden decrease of hardness occurs in HAZ for sides of all the metals. In fusion zone, the values also have a reduction, but the hardness of AZ31–AM60 is slightly higher than the others. Acknowledgements The authors would like to thank Hunan University, China, for financially supporting this research. The authors also thank W.J. Xia, X.Q. Li and Y.W. Wu for technical assistance. References [1] E. Aghion, B. Bronfin, Mater. Sci. Forum 350–351 (2000) 19–28. [2] J.H. Zhu, L. Li, Z. Liu, Appl. Surf. Sci. 247 (2005) 300–306. [3] M.B. Kannan, W. Dietzel, C. Blawert, S. Riekehr, M. Kocak, Mater. Sci. Eng. A 444 (2007) 220–226. [4] M. Dhahri, J.E. Masse, J.F. Mathieu, G. Barreau, M. Autric, Adv. Eng. Mater. 3 (2001) 504–507. [5] R.S. Coelho, A. Kostka, H. Pinto, S. Riekehr, M. Kocak, A.R. Pyzalla, Mater. Sci. Eng. A 485 (2008) 20–30. [6] C.T. Chi, C.G. Chao, T.F. Liu, C.H. Lee, Scripta Mater. 56 (2007) 733–736. [7] A. Weisheit, R. Galun, B.L. Mordike, Weld. J. 77 (1998) 149–154. [8] S. Kou, Welding Metallurgy, second ed., John Wiley & Sons, New Jersey, 2003. [9] Metals Handbook, vol. 6, ninth ed., Nippes Coordinator, American Society for Metals, Metals Park, OH, 1983. [10] M. Chase, Binary Alloy Phase Diagrams, ASM International, Ohio, 1996. [11] Z.D. Zhang, L.M. Liu, Y. Shen, L. Wang, Mater. Charact. 59 (2008) 40–46.