Microstructures in friction-stir welded dissimilar magnesium alloys and magnesium alloys to 6061-T6 aluminum alloy

Microstructures in friction-stir welded dissimilar magnesium alloys and magnesium alloys to 6061-T6 aluminum alloy

Materials Characterization 52 (2004) 49 – 64 Microstructures in friction-stir welded dissimilar magnesium alloys and magnesium alloys to 6061-T6 alum...

3MB Sizes 0 Downloads 19 Views

Materials Characterization 52 (2004) 49 – 64

Microstructures in friction-stir welded dissimilar magnesium alloys and magnesium alloys to 6061-T6 aluminum alloy A.C. Somasekharan, L.E. Murr * Department of Metallurgical and Materials Engineering, The University of Texas at El Paso, El Paso, TX 79968, USA Received 9 September 2003; accepted 10 March 2004

Abstract Wrought Mg alloy AZ31B-H24 and semisolid-cast Mg alloy AZ91D (approximately 3% primary solid fraction) were friction-stir welded to Al alloy 6061-T6. Semisolid-cast (approximately 3% and approximately 20% primary solid fractions) Mg alloys AZ91D and AM60B were also joined using the same technique, with AZ91D on the advancing side. Numerous welds were made with the Mg alloys and the 6061-T6 Al alloy in alternating advancing and retreating sides. Light optical metallography was used to observe and confirm the weld zone characteristics unique to dissimilar welds. Dynamic recrystallization (DRX) was observed in the weld region as well as in the transition region, with a clear decrease in the grain size from the base material through the transition zone and into the weld zone. The welds were free of porosities. The welds between the dissimilar Mg alloys revealed a homogeneous, equiaxed, fine-grained structure in the weld zone. The weld zone in the welds of the Mg alloys to Al alloy 6061-T6 showed unique dissimilar weld, flow characteristics, such as complex intercalated microstructures with recrystallized lamellar-like shear bands rich in either Mg or Al. Elemental analysis performed on the weld region showed bands with equal parts of Mg and Al, as well as unique recrystallized bands with a predominance of either material. Vickers microhardness testing on all the welds revealed a lack of degradation of residual microhardness of the materials in the weld zone or the transition zone, with some unusual, erratic hardness spikes exhibiting hardness values as much as three times that of the base material hardness. D 2004 Elsevier Inc. All rights reserved. Keywords: Magnesium; Aluminum; Friction-stir welding; Dynamic recrystallization; Light microscopy; Elemental analysis; TEM analysis; AZ31B; AZ91D; AM60B; 6061-T6

1. Introduction Friction-stir welding (FSW), a solid-state welding technique used primarily for joining nonferrous metals and alloys, is shown schematically in Fig. 1. In FSW, a rotating tool is fed through the joint between * Corresponding author. Tel.: +1-915-7470520; fax: +1-9157478036. E-mail address: [email protected] (L.E. Murr). 1044-5803/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.matchar.2004.03.005

the two plates to be welded. The tool rotation promotes mechanical (solid-state) mixing of the materials on the advancing and retreating sides of the weld. The welding of the material is facilitated by severe plastic deformation in the solid state, involving dynamic recrystallization (DRX) of the base material. The process itself is described in greater detail in several earlier publications ([1– 8]). In this paper, the focus is on the microstructural characteristics of friction-stir welds of wrought Mg

50

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

Fig. 1. Simple schematic diagrams of the FSW process: (a) two dissimilar metal workpieces butted together, along with the tool (with a probe); (b) the progress of the tool through the joint, also showing the weld zone and the region affected by the tool shoulder.

alloy AZ31B-H24 and semisolid-cast Mg alloy AZ91D (approximately 3% primary solid fraction) to the Al alloy 6061-T6, and of dissimilar welds of different semisolid-cast Mg (approximately 3% and 20% primary solid fractions) alloys, AZ91D and AM60B. AZ31B-H24 is a hard-rolled wrought Mg alloy, characterized by elongated grains, the grains being oriented in the rolling direction. The semisolid-cast Mg alloy AZ91D is a high-purity Mg alloy with excellent corrosion resistance and is the most widely used Mg die-casting alloy. The other semisolid-cast Mg alloy investigated, AM60B, has greater ductility and toughness than the AZ91 series alloys,

but slightly lower strength [9]. Semisolid casting is a process in which a thixotropic slurry of primary solid-fraction globules in a continuous liquid matrix is injected into a preheated metal mold to fabricate the desired component. This semisolid casting of Mg alloys enables the formation of the thixotropic microstructure that has a primary solid fraction of aMg sitting in a eutectic matrix of a-Mg grains, with Mg17Al12 intermetallic phase at the a-Mg grain boundaries [10]. One particular proprietary process, Thixomolding, has the advantage of being able to produce castings with solid fractions ranging from 0.05 to 0.60.

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

51

2. Experimental details

2.2. FSW of dissimilar Mg alloys

The chemical compositional limits of the tested materials are listed in Table 1 [9]. All the friction-stir welds were performed with a Gorton Mastermil Model 1-22 vertical milling machine that was set up for performing such welds. All welds were butt welds and were carried out using tools made of O1 tool steel that had been annealed at 825 jC then tempered at 400 jC to a hardness of Rockwell C 62. The tools were 19 mm in diameter, with a smooth level shoulder and with an unthreaded probe machined at one end. They were mounted on a vertical spindle (1j tilt) rotating at a constant rate. The plates were welded with no preweld preparation. The tool was fed at a constant traverse rate into the joint of the two plates to be welded.

The tool and the tool shoulder were maintained at 19 mm, and the tool probe was machined such that its height was approximately 95% of the plate thickness. The unthreaded probe had a diameter of approximately 6 mm. Unthreaded probes were used in all the welds. A tool rotational speed of 2000 rpm and a traverse speed of 1.5 mmps were used to perform all the dissimilar Mg welds. Low (approximately 3%) and high (approximately 20%) solid fractions AZ91D and AM60B were welded to each other. AZ91D was friction-stir welded to AM60B, with AZ91D on the advancing side, and AM60B on the retreating side. These single-sided welds were performed with no tool offset, and the tool was plunged into the joint so as to cover more than 95% of the thickness of the plates. The welds were approximately 12 mm long, and samples were procured approximately 5 mm from the starting edge. Crosssectional samples were then cut and mounted to analyze their microstructures and microhardness values.

2.1. FSW of Mg alloys to Al alloy 6061-T6 The tool and the tool shoulder were maintained at 19 mm, and the tool probe was machined so that its height was approximately 50% of the plate thickness. The probe had a diameter of approximately 5 mm for plates thinner than 2 mm, with a slightly larger diameter of approximately 6 mm for the welding of plates thicker than 2 mm. Double-sided welds were performed, with the tool (which was approximately 50% of the plate thickness) being run through either side of the joint. Seventy-five percent of the probe width was in the advancing side on all welds. All welds of Mg alloys to 6061-T6 were made using a tool rotational speed of 800 rpm and a traverse speed of 1.5 mmps. The welds were approximately 120 mm long, and samples were procured approximately 50 mm from the starting point of the weld. Cross-sectional samples were then cut and mounted to analyze their microstructures and microhardness values.

2.3. Light optical metallography All the weld samples were polished using a grit sequence of 220, 320, 500, 800 and 1200. The effect of water on the Mg side of the welds was minimized by constant cleaning with ethyl alcohol. These samples were further polished using a polishing cloth and polishing solutions of 1, 0.3 and 0.05 Am alumina in a solution of 75% ethanol and 25% glycerol. Again, the ethanol base was used to minimize contamination of the Mg samples. A chemical polishing was then performed on the samples to remove any remaining fine scratches using a solution of 2 ml HCl in 100 ml methyl alcohol. The Mg – Al samples were preetched with a solution of 2 g NaOH in 100 ml distilled water, and

Table 1 Chemical composition limits (%)

AZ31B-H24 AZ91D AM60B 6061-T6

Al

Mn

Zn

Si

Fe

Cu

Ni

Others

Mg

2.5 – 3.5 8.3 – 9.7 5.5 – 6.5 Bal.

0.2 – 1.0 0.15 – 0.50 0.25 – 0.60 0.15 Max.

0.6 – 1.4 0.35 – 1.0 0.22 Max. 0.25 Max.

0.10 Max. 0.10 Max. 0.10 Max. 0.4 – 0.8

0.005 Max. 0.005 Max. 0.005 Max. 0.7 Max.

0.04 Max. 0.03 Max. 0.01 Max. 0.15 – 0.4

0.005 Max. 0.002 Max. 0.002 Max. –

0.3 Max. (total) 0.02 Max. (each) 0.2 Max. (total) Cr 0.04 – 0.35 Ti 0.15 Max.

Bal. Bal. Bal. 0.8 – 1.2

52

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

subsequently rinsed in a solution of 5 ml HNO3 in 95 ml distilled water. This served to bring out the lamellar-like shear bands and other fine microstructures from the intercalated weld zone prior to the application of the dedicated etchants. A picral etchant was employed to etch the Mg side of the welds in all samples. The picral etch is a solution of 14 ml picric acid, 2 ml glacial acetic acid and 2 ml distilled water. The picric acid solution contained 2 g picric acid in 20 ml ethanol (this ratio was maintained in all cases). Cotton swabs and/or cotton buds were used for applying this etchant on to the Mg alloy. A variation of the Keller’s reagent, with equal parts methanol, HF and HNO3, was used to etch the Al in the Mg – Al samples. The Al side of the weld was swabbed multiple times (cleaned with ethanol after a few seconds of swabbing) until visible etching was obtained. The etched samples were then analyzed using a Reichert MEF4 A/M (Leica) light optical microscope. The overall weld cross-section was analyzed at low magnifications ( <  80) to view the entire span of the weld zone, showing the base materials, the transition zones and the FSW zone. The weld zone was further analyzed at higher magnifications (>  80) to view the intercalated FSW microstructures. 2.4. Transmission electron microscopy The two dissimilar Mg welds of AZ91D with AM60B were subjected to transmission electron microscopy. The two weld cross-sections were polished with the same procedure used for the light optical metallography. The cross-sectional plate was thinned down to < 0.2 mm, then 3-mm discs were punched out of the thinned sample for electrojet polishing. Before the sample was electrojet polished, the midsection of the disc was thinned down further (mechanically dimpled) using a dimple grinder (a 1-Am alumina paste was used to facilitate the grinding). The electrojet polishing was performed using a Tenupol-3 dual electrojet polisher. The electrojet polishing solution for both the high and low solid fraction weld samples contained 50 ml HNO3 in 1.2 l methanol; the polishing performed at 22 jC and 0.1 V. AZ91D base material samples (both high and low solid fractions) were electrojet polished with a solution of 400 ml HNO3 in 800 ml ethanol, at 10 jC and 30 V.

AM60B base material samples (both high and low solid fractions) were electropolished with a solution of 125 ml HNO3 in 1.2 l methanol, at 27 jC and 0.1 V. The electrojet-polished thin film samples were then cleaned in ethanol and stored individually for analysis. They were analyzed with a Hitachi H-8000 analytical transmission electron microscope (TEM) operating at 200 kV, equipped with a goniometer-tilt stage and a Noran energy-dispersive X-ray (EDX) spectrometer. 2.5. Vickers microhardness testing Microhardness samples were prepared in exactly the same way as the light optical metallography samples. The measurements (primarily through the midthickness) were performed with a Shimadzu digital microhardness tester. Multiple measurements were taken through 40 mm of the sample cross-section (stretching from one base material, through the transition and weld regions, and into the other material) using a load of 100 gf (1 N) applied for 15 s. At least 46 microhardness readings were taken through the midthickness for each of the weld samples. For obtaining a better profiling of microhardness values in the weld zone from the FSW of Mg alloys to 6061T6, two more sections of testing (approximately 0.5 mm above and below the midthickness) were performed from one side of the FSW zone to another. 2.6. Elemental analysis Elemental analysis of the Mg – Al welds was performed using a Phillips scanning electron microscope (SEM) equipped with an EDX system. The analysis was conducted on select points and regions representing features, such as lamellar-like shear bands, vortex flow patterns, etc. This analysis was conducted to gauge the distribution of either element in the FSW zone.

3. Results and discussion 3.1. FSW of AZ31B-H24 (A) to 6061-T6 (R) This sample was analyzed using light optical metallography. The weld zone cross-section, and base

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

material AZ31B-H24 can be seen in Fig. 2a and b, respectively. The weld zone map shows a weld with no porosities. The base material AZ31B-H24 is seen to have grains of unequal size and distribution, with larger grains distributed among smaller grains. Fig. 2c shows recrystallized Mg from the transition zone, along with the intercalated flow patterns. Grain growth is also noticeable in the dynamically recrystallized Mg alloy in the transition region. The recrystallized Mg grains in the transition zone can be seen to be fine, homogeneous and equiaxed. The frictional heat created by the rubbing of the tool shoulder and the mechanical stirring of the materials by the tool probe provided the driving force for the DRX of the materials in the weld zone. This enabled the severe plastic deformation in the weld zone to be accommodated by the flow of the material in the solid state. Hence, DRX facilitates the solid-state flow that enables FSW. Also to be considered is the adiabatic heat arising from the deformation-induced DRX [6] that contributes to the weld zone temperatures reaching approximately 0.8 TM (TM is the melting point of the material) [12]. Fig. 2 also shows the weld zone intercalated microstructures, with lamellar-like shear bands rich either in Al or Mg. These lamellar-like shear bands, seen in Fig. 2d, have been seen previously in other work on dissimilar FSW welds (e.g., Flores et al. [6]). The microstructure in Fig. 2e shows complex intercalated flow patterns, in which recrystallized Mg and Al alloys are swirled together to form a complex mesh, as seen in other dissimilar FSW work (e.g., Murr et al. [7] and Flores et al. [6]). Fig. 2 also comprises microstructures from the 6061-T6 side of the weld. Fig. 2f shows the microstructure in the transition zone between the Al alloy base material and the FSW zone. The grain size can be clearly seen to decrease as it transitions from the base material. There is then a sharp demarcation where it passes into the FSW zone, where the lamellar-like shear bands can be seen. The grain size in the 6061T6 material away from the transition zone is shown in Fig. 2g. No appreciable grain growth can be detected in this case. More data on this weld are contained in Fig. 3. Fig. 3b shows the three Vickers microhardness profiles that were generated, with that through the midthickness (line ‘‘b’’) being the most extensive. Some fairly

53

large microhardness values are noticeable in the FSW zone, indicating compensation and improvement of the usual degradations seen with 6061-T6 welds [11]. The intercalated microstructure with its complex vortexes and lamellae is responsible for the wide variations and spikes in the values of microhardness in the weld zone. The hardness value actually depends on the band of lamellae the measurement from which is taken. It was attempted to keep as uniform a run as possible, without looking for, or avoiding, any specific band. The erratic hardness spikes are an indication that the intercalated DRX bands or shear bands contain some very heavily deformed DRX regimes intermixed with the softer and relatively dislocationfree DRX zones. The results of the elemental analysis performed at several locations in the sample are shown in Fig. 3c. Al seems to be the prevalent material, in this case, in the mixing and recrystallization, although the tool was offset into the advancing Mg-containing AZ31B-H24 alloy side. A smaller percentage of Mg is present throughout the weld zone, indicating the mixing of the two alloys. 3.2. FSW of 6061-T6 (A) to AZ31B-H24 (R) Fig. 4 shows the weld zone map (Fig. 4a) and the associated transition and weld zone microstructures that are very similar in nature to the previous weld (Figs. 2 and 3). The microstructure in Fig. 4b shows recrystallized 6061-T6 abutting the FSW zone, while that in Fig. 4c shows recrystallized AZ31B-H24 adjoining the FSW zone, along with grain growth of Mg subsequent to DRX. As before, the transitioning of 6061-T6 into the FSW zone was characterized by a sharp and clear demarcation. In the case of the Mg alloy, the transitioning seems not to be so well defined and unique as it nears the FSW zone. Fig. 4d shows the intercalated microstructure seen previously in Fig. 2e, testifying to the mixing of dissimilar welded materials in vortex flow patterns. Fig. 4e shows the Vickers microhardness profiles generated for this particular welded sample. As before, there is no reduction in microhardness in the weld zone, and the joining of a Mg alloy to 6061-T6 seems to be compensating for the usual degradation observed in all 6061-T6 welds. The region that was stirred during the second run of the tool seems to have

54

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

Fig. 2. (a) Advancing and retreating sides of the FSW of AZ31B-H24 (advancing side) with 6061-T6 (retreating side); (b) AZ31B-H24 base material; (c) transition zone with recrystallized AZ31B-H24; (d) lamellar-like shear bands in the FSW zone; (e) intercalated microstructure with vortexes, in the FSW zone; (f) transition zone showing the transitioning of 6061-T6; (g) 6061-T6 base material.

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

55

Fig. 3. (a) Advancing and retreating sides of the FSW of AZ31B-H24 (advancing side) with 6061-T6 (retreating side); (b) Vickers microhardness profile of the FSW region of the weld of AZ31B-H24 (advancing side) with 6061-T6 (retreating side); (c) elemental analysis of the FSW region of the weld of AZ31B-H24 (advancing side) with 6061-T6 (retreating side).

extremely large microhardness (spike) values, similar to what have been seen in all the double-sided Mg – Al welds. However, there is essentially no reduction in microhardness in the immediate vicinity of the FSW and transition zones.

3.3. FSW of 6061-T6 (A) to AZ91D (approximately 3% solid fraction; R) Fig. 5a shows the weld map from this weld. No porosities were observed. The microstructure in Fig.

56

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

Fig. 4. (a) Advancing and retreating sides of the FSW of 6061-T6 (advancing side) with AZ31B-H24 (retreating side); (b) 6061-T6 DRX zone; (c) AZ31B-H24 DRX zone; (d) intercalated microstructures in the FSW zone; (e) Vickers microhardness profile of the FSW region from the weld of 6061-T6 (advancing side) with AZ31B-H24 (retreating side).

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

57

Fig. 5. (a) Advancing and retreating sides of the FSW of 6061-T6 (advancing side) with AZ91D (approximately 3% solid fraction; retreating side); (b) 6061-T6 transitioning in the DRX zone; (c) intercalated microstructures in the FSW zone.

5b shows the transitioning of 6061-T6 into the weld zone as the base material recrystallizes through the transition zone. Unlike the weld between AZ31B-H24 and 6061-T6, this weld does not show a sharp demarcation from the 6061-T6 into the weld zone. Instead, there is a more uniform and gradual transitioning flow into the weld zone dictated by the stirring of the tool from either side. The intercalated microstructure with vortexes and swirls, seen in Fig. 5c, is indicative of typical dissimilar weld characteristics. Corresponding microhardness and elemental analysis data are given in Fig. 6. The Vickers microhardness profile (Fig. 6b) is characterized by elevated microhardness readings in the weld and transition zones, much like the dissimilar welds of AZ31B-H24 with 6061-T6. Meanwhile, the elemental analysis of this weld cross-section (Fig. 6c) indicates that, although there is generally a fairly uniform distribution of Al and Mg in the weld zone, there are regions having higher percentages of Al than Mg, and vice versa. Hence, a preferred recrystallizing material is not visible here, indicating a much more uniform mixing of the material. 3.4. FSW of AZ91D (approximately 3% solid fraction; A) to 6061-T6 (R) The map of this weld is shown in Fig. 7a. No porosities were observed. Fig. 7b shows the transitioning flow of dynamically recrystallized 6061-T6

into the FSW zone. Although there is a sharp demarcation similar to that seen with the AZ31B-H24-to6061-T6 welds, a ‘‘funnel’’-shaped flow is seen on the Al side of the weld, indicative of the material being stirred by the tool and thus causing the DRX of the Al alloy. Fig. 7c shows lamellar-like shear bands in the FSW zone rich in either Al or Mg. Fig. 7d shows a dynamically recrystallized weld region within the FSW zone, rich in Mg. Grain growth is noted. Fig. 8 contains the corresponding microhardness distribution across this weld. The Vickers microhardness profile (Fig. 8b) shows elevated values comparable to the other welds, indicating that there is no diminution of microhardness in this zone. Thus, in every case where a Mg alloy was welded to 6061-T6, unique elevated microhardness values appeared in the FSW and transition zones. 3.5. FSW of AZ91D (A) to AM60B (R) Two welds were performed involving these dissimilar Mg alloys. One weld was made with both materials containing approximately 20% solid fraction, the other with each material containing approximately 3% solid fraction. The weld zone microstructural observations in both cases, collected respectively in Figs. 9 and 10, show a sharp demarcation in the advancing side (AZ91D) and a rather scattered flow in the retreating side (AM60B). In this,

58

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

Fig. 6. (a) Advancing and retreating sides of the FSW of 6061-T6 (advancing side) with AZ91D (approximately 3% solid fraction; retreating side); (b) Vickers microhardness profile of the FSW region from the weld of 6061-T6 (advancing side) with AZ91D (approximately 3% solid fraction; retreating side); (c) elemental analysis from the FSW zone of 6061-T6 (advancing side) with AZ91D (approximately 3% solid fraction; retreating side).

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

59

Fig. 7. (a) Advancing and retreating sides of the FSW of AZ91D (approximately 3% solid fraction; advancing side) with 6061-T6 (retreating side); (b) 6061-T6 transitioning into the weld zone; (c) lamellar-like shear bands in the FSW zone; (d) AZ91D-rich DRX zone.

the observations (Figs. 9a and 10a) are very similar to those seen on welds of AM60B material [1], and on other dissimilar welds, such as 2024 Al alloy/6061 Al alloy [7]. Fig. 9b is of the microstructure of the complex, approximately 20% solid fraction, base material and shows a primary solid fraction (unmelted alloy fraction) of a-Mg in an a-Mg grain eutectic (with Mg17Al12 intermetallic phase at the grain boundaries). This microstructure becomes dynamically recrystallized and forms a fine, homogeneous equiaxed grain structure in the FSW zone (Fig. 9d). Also to be noted is the intercalated flow patterns in the FSW zone. The transition zone away from this weld is made up of

recrystallized grains of varying sizes (Fig. 9c). Fig. 9e and f are TEM images from, respectively, the FSW zone and the AM60B base material. Fig. 9e shows some dislocation structures and Moire´ fringe patterns within the grains, as viewed from a triple point of grains in the FSW zone. Fig. 9f is from a region in the AM60B base material, showing dislocations propagating from the grain boundaries. The dislocation density in the base material is significantly greater than in the weld zone. Fig. 10b and c shows the microstructures of the approximately 3% solid fraction base materials AZ91D and AM60B. This complex microstructure undergoes DRX to form fine, homogeneous and

60

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

Fig. 8. (a) Advancing and retreating sides of the FSW of AZ91D (approximately 3% solid fraction; advancing side) with 6061-T6 (retreating side); (b) Vickers microhardness profile of the FSW region from the weld of AZ91D (approximately 3% solid fraction; advancing side) with 6061-T6 (retreating side).

equiaxed grains in the FSW zone (Fig. 10d). The recrystallized grains are smaller in size than the a-Mg grains of the base material. Fig. 10e and f are TEM images from within the weld zone. Dislocation structures are readily evident in this zone and at varying densities. However, the density still appears to be less than in the base approximately 3% solid fraction material (see Ref. [1]). The Vickers microhardness profiles for both the approximately 20% and the approximately 3% solid fraction welds are shown in Fig. 11. Both profiles show that the microhardness distribution is basically uniform throughout the weld cross-section. There is no observable change in microhardness, either in the transition or the FSW zone. Thus, it can be inferred that any loss of hardness due to a reduction in

dislocation density has been offset by the increase that would be expected from a reduction in grain size.

4. Summary and conclusions DRX is the primary mechanism in the FSW of metals and alloys, thus facilitating severe plastic deformation and solid-state flow in creating the weld. All the dissimilar welds examined exhibit recrystallization of the base materials. This recrystallization was enabled by frictional heat from the tool shoulder and tool probe, heat generated by the mechanical stirring of the materials by the probe and mostly adiabatic heat contributing to DRX through the deformation. Complex intercalated microstructures were observed in the

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

61

Fig. 9. (a) Advancing and retreating sides of the FSW of AZ91D (advancing side) with AM60B (retreating side; approximately 20% solid fraction); (b) AZ91D base material; (c) transition zone; (d) FSW zone; (e) TEM image from a triple point in the FSW zone showing dislocation structures and Moire´ fringes within the grains; (f) dislocations propagating from the grain boundaries in the AM60B base material.

weld zone, with swirls and vortexes indicative of the flow pattern of the dissimilar metals, especially in the case of the FSW of the Mg alloys to the Al alloy 6061T6. Lamellar-like shear bands rich in either Mg or Al

were seen in the weld zone. In the case of dissimilar Mg alloy welds, fine, homogeneous, equiaxed grains were seen in the weld zone, dynamically recrystallized from the complex base material microstructure. Grain

62

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

Fig. 10. (a) Advancing and retreating sides of the FSW of AZ91D (advancing side) with AM60B (retreating side; approximately 3% solid fraction); (b) AZ91D base material; (c) FSW zone; (d) AM60B base material; (e) TEM image from the weld zone (the FSW zone) showing dislocation structures within the grains; (f) dense dislocation structures within the weld zone.

growth subsequent to DRX was noted in the case of the Mg alloys for all Mg – Al welds. A sharp demarcation zone was seen in all the welds, but with different characteristics. In the dissimilar Mg welds, the sharp demarcation was on the advancing

side, which had the AZ91D material, and the scattered flow is viewed on the retreating side, which has the AM60B material. In the welds of Mg alloys to 6061T6, the demarcation was always sharp on the Al alloy side, with the softer Mg alloy having a more scattered

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

63

Fig. 11. Residual microhardness profile through the midthickness of the FSW region from the welds of approximately 20% and approximately 3% solid fraction welds of semisolid-cast Mg alloys AZ91D (advancing side) and AM60B (retreating side).

transition zone. Fine transitioning of 6061-T6 into the weld zone was observed. The grain size decreased as the material transitioned from the base material to the intercalated microstructures of the weld zone. The complex intercalated microstructures in the FSW zone contributed to the haphazard elevated readings seen in the weld zone. But it also contributed to the lack of degradation of microhardness. In the case of the dissimilar Mg alloy welds, the grain size decreased in the FSW zone, as did the dislocation density, although some dense dislocation substructures were seen. This served to keep a fairly uniform microhardness profile in the weld zone. The elemental analysis of the Mg – Al welds demonstrated that there was mixing of materials in the weld zone. In the case of the AZ31B-H24-to-6061-T6 welds, the preferred material of recrystallization was seen to be 6061-T6. The elemental analysis of the AZ91D-to-6061-T6 weld indicated a more uniform distribution of both Al and Mg in the FSW zone.

Acknowledgements This work was supported in part by a Graduate Research Assistantship, a Graduate Research Scholarship, and a Mr. and Mrs. Macintosh Murchison Endowed Chair at UTEP (LEM). Technical support

from David Brown (Manufacturing) and Erika Esquivel (Microscopy) is very much appreciated. References [1] Esparza JA, Davis WC, Murr LE. Microstructure-property studies in friction-stir welded thixomolded magnesium alloy AM60. J Mater Sci 2003;38:941 – 52. [2] Murr LE, Sharma G, Contreras F, Guerra M, Kazi SH, Siddique M, et al. Joining dissimilar aluminum alloys and other metals and alloys by friction-stir welding. In: Das SK, Kaufman JG, Lienert TJ, editors. Aluminum 2001: Proceedings of the 21st TMS Annual Meeting Aluminum Automotive and Joining Aluminum Symposia.The Minerals, Metals and Materials Society. Warrendale (PA): TMS, 2001. pp. 197 – 211. [3] Prado RA, Murr LE, Shindo DJ, McClure JC. Friction-stir welding: a study of tool wear variation in aluminum alloy 6061 + 20% Al2O3. In: Jata K, Mahoney M, Mishra R, editors. Friction stir welding and processing. Warrendale (PA): TMS (The Minerals, Metals and Materials Society), 2001. pp. 105 – 16. [4] Trillo EA, Esquivel EV, Murr LE, Magness LS. Dynamic recrystallization-induced flow phenomena in tungsten – tantalum (4%) [001] single-crystal rod ballistic penetrators. Mater Charact 2002;48:407 – 21. [5] Esparza JA, Davis WC, Murr LE. Friction-stir welding of magnesium alloy AZ31B. J Mater Sci Lett 2002;21:917 – 20. [6] Flores RD, Murr LE, Shindo D, Trillo EA. Friction-stir welding of metals and alloys: fundamental studies of solid-state and intercalated flow. Proceedings of International Conference on Processing and Manufacturing of Advanced Materials, Las Vegas, NV, 2000 Dec 4 – 8.

64

A.C. Somasekharan, L.E. Murr / Materials Characterization 52 (2004) 49–64

[7] Murr LE, Li Y, Trillo EA, Flores RD, McClure JC. Microstructures in friction-stir welded metals. J Mater Process Manuf Sci 1998;7:145 – 61. [8] Prado RA, Murr LE, Soto KF, McClure JC. Self-optimization in tool wear for friction-stir welding of Al 6061 + 20% Al2O3 MMC. Mater Sci Eng, A Struct Mater: Prop Microstruct Process 2003;349:156 – 65. [9] Avedesian MM, Baker H, editors. ASM specialty handbook: magnesium and magnesium alloys. Materials Park (OH): ASM International (The Materials Information Society), 1998.

[10] Czerwinski F, Zielinska-Lipiec A, Pinet PJ, Overbeeke J. Correlating the microstructure and tensile properties of a thixomolded AZ91D magnesium alloy. Acta Mater 2001;49(7): 1225 – 35. [11] Valerio-Flores O, Kennedy C, Murr LE, Brown D, Pappu S, Nowak BM, et al. Microstructural issues in a friction-stirwelded aluminum alloy. Scr Mater 1998;38(5):703 – 8. [12] Tang W, Guo X, McClure JC, Murr LE, Nunes A. Heat input and temperature distribution of friction stir welds. J Mater Process Manuf Sci 1998;7(2):163 – 72.