Kovove Mater.

44 2006 25–33 25

Microstructural and mechanical properties

of double-sided MIG, TIG and friction stir welded 5083-H321
aluminium alloy

E. Taban1 , E. Kaluc2 *

1
Department of Mechanical Engineering, Engineering Faculty, Kocaeli University, Kocaeli 41040, Turkey
2
Welding Technology Research and Training Center, Veziroglu Campus, Kocaeli University, Kocaeli, 41040, Turkey

Received 26 May 2005, accepted 14 December 2005

Abstract
In this study, strain-hardened EN AW-5083-H321 aluminium alloy plates were welded using
MIG, TIG and friction stir welding (FSW) processes in order to investigate microstructural
and mechanical properties, respectively. As a microstructural investigation, light microscopy
and image analysis system were used for double-sided MIG, TIG and FSW specimens weld
zones. Measurements of hardness were carried out using the microhardness tester with 200 g
test load on metallografic specimens taken from each welded plate. Tensile and bend tests
were applied to specimens taken from welded plate with reference to EN 895 and EN 910
standards. Fracture surfaces were also examined using light microscopy and scanning electron
microscopy. As a result, it has been observed that FS welded joints have relatively superior
properties as/than MIG and TIG welded joints.

K e y w o r d s : 5083-H321, MIG, TIG, friction stir welding (FSW), microstructural and me-
chanical properties

1. Introduction other gases. The joining of aluminium alloys, espe-

cially those which are often difficult to weld, has been
Aluminium and aluminium alloys are common the initial target for developing the new solid state
structural materials due to their attractive combi- welding – friction stir welding (FSW) [1, 6]. The FSW
nation of properties including light weight, moder- technique (invented, patented and developed by TWI)
ate strength, good corrosion resistance, workability, enables the advantages of solid phase welding to be
proven weldability and good electrical and thermal applied to the fabrication of long butt and lap joints
conductivity. Hence these alloys are commonly used with very little post weld distortion. FSW operates
in automotive, railway, shipbuilding and defence in- by passing a rotating tool between two plates that are
dustries, aviation and aerospace industries [1–5]. restricted from movement. The tool is plunged into
For the joining of aluminium alloys, besides fusion the material to a preset depth and moved along the
welding processes, solid state welding processes have weld joint. Heat is generated through frictional con-
been widely being used in various industrial appli- tact between the rotating tool shoulder and abutting
cations. There are, however, a number of mechanical material surface. Despite being such relatively new
and metallurgical problems associated with the fusion process, friction stir welds have already been launched
welding of aluminium and its alloys such as decreasing into aircraft, shipbuilding, automotive, railway, de-
strength and metallurgical precipitations in the weld fence industries with great success and the process
metal and heat affected zone (HAZ), distortions, resid- being actively investigated for further applications [1,
ual stresses due to high heat input, gas porosity, lack 4–9].
of fusion, high coefficient of thermal expansion, so- While examining the literature on this subject, it
lidification shrinkage, high solubility of hydrogen and has been found that a number of research studies have

*Corresponding author: tel.: +90 262 3351148 (ext. 1147); fax: +90 262 3352812; e-mail address: ekaluc@kou.edu.tr
26 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33

T a b l e 1. Chemical composition and mechanical properties of 5083-H321 aluminium alloy

Chemical composition [wt.%]

Si Fe Cu Mn Mg Cr Ni Zn V Ti Al (min)
0.117 0.245 0.071 0.488 4.7 0.108 0.001 0.083 0.016 0.041 Rem.

Mechanical properties

Proof strength Rp 0.2 [MPa] Tensile strength Rm [MPa] Elongation (min) [%]
259 355 13.2

been done on the friction stir weldability of 5083 alu- 1.6 mm diameter ER 5356 (AlMg5Cr(A)) filler metal
minium alloy [8, 10–18]. However, there has not been were used.
enough information reported concerning the compar- 6.45 mm thick plates were friction stir welded using
ison of the mechanical and microstructural properties AISI 4340 (0.40 % C, 0.70 % Mn, 0.30 % Si, 0.80 % Cr,
of welded zones of double-sided MIG, TIG and FS 1.80 % Ni, 0.25 % Mo) tool. Shoulder diameter was 20
welded 5083-H321 alloy. mm and the pin was standard M5 threaded and 3.6
The aim of the present paper is to determine and mm height. The tool rotational speed was set equal
discuss microstructural observations and provide in- to 1600 rpm and translation speed was 125 mm/min.
formation about the mechanical properties of double- The tool was tilted at 2◦ from the plate normal and ro-
sided MIG, TIG and FS welded 5083-H321 aluminium tated in the anti-clock direction. The specimens were
joints. After mechanical tests, the fracture surfaces of tightly fixed at the machine tool table. Before FSW,
weld zones have been examined using LOM and SEM weld groove was cleaned with acetone and brushed by
and discussed. stainless steel brush carefully.
After welding, a visual inspection was made on all
welded plates in accordance with ANSI/AWS B1.11.
2. Material and experimental procedure All weld defects were examined from both the front
and back side of weld bead carefully. Weld reinforce-
The base material used in this study was 6.45 ment and spatter dimensions were also measured. All
mm thick 5083-H321 aluminium alloy, which is a welded joints were subjected to distortion measure-
non-heat treatable Al-Mg alloy with relatively good ments. Angular displacements in butt joint welded
mechanical properties and corrosion resistance com- plates were inspected.
monly used in defence and shipbuilding industries. MIG, TIG and friction stir welded joints were
This alloy has a relatively large Mg content to pro- cross-sectioned perpendicular to the welding direc-
mote solid solution strengthening and to increase the tion for metallographic analyses. The cross-sections
rate of work hardening, the most important strength- of the metallographic specimens were polished with
ening mechanisms in this alloy. Manganese increases diamond suspension and colloidal chemical, etched
the recrystallisation temperature and improves the with modified Keller’s reagent (2 ml HF, 3 ml HCl,
strengthening effect of magnesium [19]. Chemical com- 20 ml HNO3 , 175 ml H2 O) at room temperature for
position and mechanical properties of the alloy are lis- about 30 s. They were visualised as macrographs and
ted in Table 1. subsequently observed by light microscope. Metallo-
5083-H321 aluminium alloy plates were cut into graphy and hardness variations with 200 g test load
6.45 × 150 × 350 mm coupons with a 30◦ level of were carried out on all as-welded samples in a section
each plate to provide 60◦ groove angle for a double across the weld of the first and second pass, respecti-
V-groove (symmetric X-groove) butt joint configura- vely. Future-Tech FM 700 Microhardness Tester hav-
tion according to EN ISO 9692-3 for MIG and TIG ing a loading capacity of 10 g minimum was used for
welding. The root face was 1 mm with root opening the hardness measurements.
of 2 mm. Only I-groove butt welding preparation was The tranverse tensile and bend test specimens were
applied for FSW process without root opening. taken and prepared from MIG, TIG and FS welded
MIG and TIG joints were realised semi-automati- plates with reference to EN 895 and EN 910 standards,
cally in industrial conditions. Chemical oxide removal respectively. Tensile and bend tests were carried out
processes were applied prior to and before welding all using servo-hydraulic controlled DARTEC test ma-
grooves have been cleaned from all contaminations by chine at room temperature. Weld reinforcement of the
using acetone and stainless steel brush. Detailed para- specimens was ground off.
meters are: voltage 24–26 V, welding current 180–200 After tensile test, fracture surfaces of broken
A, 99.999 % Argon as shield gas (flow rate 12 l/min), samples were examined using light macro-stereomicro-
 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33 27

DISTORTION [mm]
14 MIG
12 TIG
10 FSW
8
6
4
2
0
MIG, TIG AND FS WELDED PLATES

Fig. 1. Distortion phenomena of double-sided welded plates: the photographs for the distortion of MIG (a), TIG (b) and
FS (c) welded plates, the graph showing the comparison of welded plates distortion rates (d).

scopy and SEM in order to determine the fracture hibited much better dimensional accuracy mostly due
mode of welded joints. Also, bend test samples were to the fact, that the materials had not been melted
investigated using macro-microscopy in order to exam- during the welding process. It is known that the max-
ine the surface cracks of face and root sides on weld imum temperature for aluminium FS welds is far lower
beads. than that of MIG and TIG welding. Increase in the
temperature during welding results in greater distor-
tion. The total distortion of 14.75 mm was measured
3. Results and discussion for MIG welded plate, 6.65 mm for TIG welded and
0.5 mm for FS welded joints. The largest distortion
As a result of visual inspection on welded plates, amount was determined for MIG welded joints. Be-
it was observed that welds of both MIG and TIG wel- cause the heat input of TIG welding is local, the dis-
ded joints were in acceptable limits; there was not tortion rate of the TIG plate was less than that of
too much spatter, the sizes of spatter were very small MIG welded plates. Figures 1a,b,c show the photos of
and no surface cracks and discontinuities were detec- distortion of as-welded plates and Fig. 1d shows the
ted. The surface appearance of FS welded joints was comparison of the amount of distortion for three weld-
a regular series of semi-circular ripples. No surface ing methods.
cracks and discontinuities were detected. Investiga- The microstructural investigation was carried out
tions have shown that the distortions in MIG and TIG on the metallographic specimens of the joints using
welded plates were much higher than in FS welded light microscope with 50× and 200× magnifications.
joints nevertheless still they were within the accept- The investigation of the weld metal was performed
able limits. Double-sided MIG, TIG and FS welded with the order of base metal (BM)-heat affected zone
plates were compared for dimensional accuracy using (HAZ) for fusion welding processes and thermome-
angular distortion measurements. FS welded plates ex- chanically affected zone (TMAZ) for FSW- and weld
28 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33

Fig. 2. Light macrostructure of MIG welded 5083-H321 alloy (a) and first pass microstructures of the alloy: BM (b), WM
(c), HAZ + WM (d).

metal (WM). Figures 2 and 3 show the macrographs loys is limited to recovery, recrystallisation and grain
and micrographs of first passes of double-sided MIG growth. Thus, loss in strength in the HAZ is not nearly
and TIG welded zones, respectively; Figure 4 shows as severe as that experienced in heat treatable alloys.
the macrograph and micrographs of both first and HAZ in fusion welded samples contains typical fea-
second passes of double-sided friction stir welded zones tures [19]. Because of the same filler metal ER 5356
of 5083-H321 specimens. used for TIG welding, similar weld metal structure
It is known that, in case of fusion welding pro- was observed, however no porosity was determined.
cesses such as MIG and TIG welding, due to the fu- Similar base metal and HAZ morphologies appeared
sion of material and high temperatures experienced by as seen in Fig. 3.
adjacent material relatively wide HAZ appears. The The structure of FS welds contains features that
5xxx series alloys are work hardenable alloys and so are not found in fusion welds. The reason is that, un-
the base metal microstructure is observed as typical like fusion welding processes, no filler metal was used
of rolling/work hardening microstructure. It was seen and no groove preparation was used in FSW process.
that MIG and TIG weld metals consist of dendrites The structure of the weld zone, due to severe me-
with fine precipitates of Mg2 Al3 (dark) as typically chanical stresses experienced by the material, shows
ER 5356 filler metal microstructure [20]. A number a variety of zones. A fine-grained structure is charac-
of microvoids and porosities were also observed in the terised by the presence of so-called onion rings, which
weld metal of MIG welded samples (see Figs. 2a,c). appear across the section and allow a partial recov-
These are the potential sources for cracks during the ery of mechanical properties. This onion ring struc-
service conditions. Unlike the case of heat treatable ture of the nugget is typical of high quality stir weld,
alloys, whose strengthening precipitates may dissolve in which neither porosity nor microvoids are detec-
or coarsen, the HAZ damage in non-heat treatable al- ted. In the double-side welded friction stir welding of
 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33 29

Fig. 3. Light macrostructure of TIG welded 5083-H321 alloy (a) and first pass microstructures of the alloy: BM (b), WM
(c), WM + HAZ (d).

this alloy, two nugget zones were clearly observed in cimens for both first and second pass as a function of
macro-section of the joints as seen in Fig. 4a. Fig- lateral distance from the weld line for each weld at 1
ure 4b shows the base metal 5083-H321 alloy micro- mm depth of the cross-section with 200 g test load for
structure. The onion ring patterns were clearly seen 10 s using Future-Tech FM 700 microhardness tester.
in Figs. 4c,d. Immediately adjacent to the nugget, Figure 5a shows the hardness distributions of welded
there is the plastically deformed and heat affected so- joints.
-called “thermo-mechanically affected zone (TMAZ)”, Figure 5b shows the average tensile properties of
which has only been affected by the heat flow [12]. The double-sided MIG, TIG and FS welded joints. It can
transition between the two nugget zones and thermo- be seen from the figure that the tensile properties of
mechanically affected zone are clearly visible in FS each joint are lower than those of the base material.
welded 5083-H321 alloy as shown in Fig. 4d. The sight Examination of the tensile test results of double-sided
of the top surface of the weld bead shows semicircular MIG, TIG and FS welded joints have demonstrated
rings referred as banded microstructure. that the average joint efficiency values of the welded
The deformed microstructure of the base material joints are 79 %, 81 % and 80 % of the base mate-
has been replaced by equiaxed grains with little sub- rial, respectively. It has been observed that MIG and
structure. This is typical of a recrystallised microstruc- TIG welded specimens were broken from the weld
ture and similar to that found previously in 5xxx, 6xxx metal zones. The weld metal of fusion welded non-
and 7xxx series aluminium stir welds in accordance -heat treatable aluminium alloys is typically the weak-
with the former publication by Peel et al. [10]. est part of the joint [19]. This situation is confirmed in
The hardness assessment was made on double- the hardness distribution in Fig. 5a. Heat provided by
-sided MIG, TIG and FS welded metallography spe- welding process is responsible of the decay of mecha-
30 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33

Fig. 4. Light macrostructure of FS welded 5083-H321 alloy (a) and microstructures of the alloy: BM (b), WM (c), TMAZ
+ WM of double passes (d).

nical properties, due to softening induced in the alloy. detected in all face and root bend samples of FS wel-
The strength increasing effect of strain hardened 5083- ded joints. Figure 5c shows the bend test failure caused
-H321 aluminium alloy was eliminated by the high by under-crack defect detected in MIG welded joint,
heat input of fusion welding and porosities in weld and Fig. 5d illustrates the weld metal crack of TIG
metal are the reasons of this situation. However, it welded joint. This provides one more advantage of
was observed that the failure locations were in thermo- FSW process compared to fusion welding processes.
-mechanically affected zones (TMAZ) from advancing Figures 6a–f illustrate the macrographs and SEM
side of FS welded specimens after tensile test. The photographs of fracture surfaces of the tensile test spe-
weakest parts of FS welded joints were observed as cimens. Figure 6b is the enlarged SEM photograph of
TMAZ as is also confirmed in Fig. 5a with the de- MIG welded sample showing a dimple pattern that in-
crease of hardness values in TMAZ. It is interesting dicates ductile fracture and the porosity in this mat-
that the hardness plots of MIG, TIG and FS welded rix occurred. Figure 6d also contains the SEM photo-
joints were in the same order of mechanical efficien- graph of dimple patterns and some microvoids, which
cies. The order observed was as TIG, FSW and MIG. caused fracture to initiate in TIG welded joint. Fig-
The 80% joint efficiency of FSW 5083-H321 speci- ures 6e,f exhibit the fracture phenomena of FS welded
mens is a better result than the strength value pro- specimens. Tensile testing of these joints showed that
posed and discussed in an earlier publication carried cracking tended to occur initially at the upper region
out and reported by Liu et al. [13] in which the joint of the joint and propagated toward the bottom region.
efficiency was determined as 73%. The interesting fracture occurred from onion ring frag-
The face and root bend test of the samples re- ments as layers with ductile-brittle mixed type was
vealed, that although some samples of MIG and TIG similar to cast alloy structure that caused to reduce
welded joints contained cracking defects, no crack was the joints strength.
 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33 31

100
98 a
96
94
92

Hardness [HV0.2]
90
88
86
84 MIG1.PASS
82 MIG2.PASS
80 TIG1.PASS
78 TIG2.PASS
76 FSW1.PASS
74 FSW2.PASS

-16-14-12-10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16
Distance from centerline [mm]

350 Rp0.2 b
Rm
300
Strength [N/mm ]
2

250

200

150

100

50

0
BM MIG TIG FSW

Base Metal and Welded Specimens

Fig. 5. First and second pass hardness distributions of MIG, TIG and FS welded joints, respectively (a), comparison of
average tensile properties of the base metal with welded joints (b), the failures observed in the root and face bend test
specimens of MIG (c) and TIG (d) welded joints.
32 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33

Fig. 6. Light macrographs and SEM images of the fracture surfaces of the welded specimens: light macrograph (a) and
SEM photo (b) of MIG welded joint, light macrograph (c) and SEM photo (d) of TIG welded joint, light macrograph (e)
and SEM photo (f) of FS welded joint.

4. Conclusions properties, hardness distributions, mechanical prop-

erties and examining of fracture surfaces of double-
The conventional fusion welding processes – MIG -side MIG, TIG (with X-groove) and FS (with I-
and TIG and a new solid state welding process FSW -groove) welded 5083-H321 alloy have been studied in
were successfully applied to the joining of 5083-H321 the present work. Following conclusions are drawn:
aluminium alloy from both sides. Microstructural (1) The microstructure of double stir zones was
 E. Taban, E. Kaluc / Kovove Mater. 44 2006 25–33 33

mainly composed of onion ring structures in double References

nugget zones with fine and equiaxed grains and TMAZ
were observed. [1] MATHERS, G.: The Welding of Aluminium and Its
(2) The hardness of the weld zone confirmed the Alloys. Cambridge, Woodhead Publishing Limited
joint efficiencies of welded plates as observed in tensile 2002.
[2] ANDERSON, T.: Welding Journal, 81, 2002, p. 77.
test.
[3] ANDERSON, T.: Welding Journal, 83, 2004, p. 28.
(3) The present study has demonstrated that the [4] JOHNSEN, M. R.: Welding Journal, 78, 1999, p. 35.
tensile properties of FS welded joints were satisfac- [5] KALLEE, S. W.—DA VENPORT, J.—NICHOLAS,
tory as/than TIG and MIG welded joints. Though all E. D.: Welding Journal, 81, 2002, p. 47.
failure locations were detected as weld metal in fusion [6] DAWES, C. J.—THOMAS, W. M.: Welding Journal,
welding processes, all failure locations of FS welded 75, 1996, p. 41.
specimens were occurred at TMAZ advancing side. [7] COLLIGAN, K. J.—KONKOL, P. J.—FISHER, J.
(4) Bend tests of welded plates have shown that J.—PICKENS, J. R.: Welding Journal, 82, 2003, p.
34.
FS welded specimens do not include any defect like
[8] THREADGILL, P.: TWI Bulletin Reprint 513/2/97.
fusion welded specimens. Cambridge 1997.
(5) The LM and SEM examination of fracture sur- [9] DAWES, C. J.—THOMAS, W.: TWI Bulletin 6. Re-
faces revealed porosities in MIG and TIG welds that print 493/6/95. Cambridge 1995, p. 124.
have caused mechanical results to decrease; however [10] PEEL, M.—STEUWER, A.—FREUSS, M.—WI-
FS welds do not include any weld defects. Further- THERS, P. J.: Acta Materialia, 51, 2003, p. 4791.
more, ductile-brittle mixed type fracture has been oc- [11] JAMES, M. N.—HATTINGH, D. G.—BRADLEY, G.
curred in FS welded joints weld zones as layers from R.: International Journal of Fatigue, 25, 2003, p. 1389.
the onion rings. [12] LARSSON, H.—KARLSSON, L.: Svetsaren, 2, 2000,
p. 6.
(6) No filler metal and groove preparation were [13] LIU, H. J.—FUJII, H.—MAEDA, M.—NOGI, K.:
needed in FSW process, the surface appearance ap- Transactions of Nonferrous Metals. Society of China,
proaches to that of a rough-machined surface. This 13, 2003, p. 14.
situation reduces the production costs. [14] DICKERSON, T. L.—PRZYDATEK, J.: Interna-
(7) The FSW process is a solid state welding pro- tional Journal of Fatigue, 25, 2003, p. 1399.
cess with process temperatures lower than fusion tech- [15] SHIGEMATSU, I.—KWON, Y. J.—SUZUKI, K.—
niques, thus avoiding problems such as porosity, crack- IMAI, T.—SAITO, N.: Journal of Materials Science
ing and distortion. These are also the advantages of Letters, 22, 2003, p. 353.
[16] JAMES, M. N.—BRADLEY, G. R.—LOMBARD,
FSW process in comparison with fusion welding pro-
H.—HATTINGH, G.: Fatigue Fract Eng Mater
cesses such as MIG and TIG. Struct., 28, 2005, p. 245.
[17] REYNOLDS, A. P.—LOCKWOOD, W. D.—SEI-
DAL, T. U.: Materials Science Forum, 331–337, 2000,
p. 1719.
Acknowledgements [18] MCLEAN, A. A.—POWELL, G. L. F.—BROWN, I.
H.—LINTON, V. M.: Science and Technology of Weld-
The authors would like to thank FNSS Defence In- ing and Joining, 8, 2003, p. 462.
dustry Company for providing aluminium alloys investi- [19] CROSS, C. E.—KOHN, M. L.: ASM Handbook Weld-
gated in present paper and for financial support of Kocaeli ing, Brazing and Soldering, ASM Int., USA, Vol. 6,
University Research Fund in scope of research project 2000, p. 537.
(Project No: 2004/51). Acknowledgements to Prof. Dr. [20] STEVENS, R. H.: ASM Handbook Metallography and
Sadi Karagoz and his team for the SEM photographs. Microstructures. ASM Int., USA, Vol. 9, 2000, p. 362.