TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 384

WELDING RESEARCH

Influence of Al Interlayer Thickness on

Laser Welding of Mg/Steel
The impact of different Al interlayer thicknesses on interfacial reactions
and mechanical properties during laser welding of Mg to steel was investigated

BY C. W. TAN, B. CHEN, X. G. SONG, L. ZHOU, S. H. MENG, L. Q. LI, AND J. C. FENG

to steel was proven to be difficult by

ABSTRACT conventional fusion welding tech-
niques due to a huge difference in
The bonding of immiscible system Mg/steel was facilitated by adding an Al interlayer us­
their melting points and Mg-Fe immis-
ing laser welding. The influence of the added interlayer thickness on the microstructure and
mechanical properties of the dissimilar joints were investigated. The Al interlayer was dis­ cibility characteristics. The melting
solved into the Mg fusion zone and mutual diffusion with steel, causing metallurgical bond­ point of steel (1450ºC) is higher than
ing of Mg and steel at the interface. Fe­Al reaction layers with different thicknesses were the boiling point of magnesium
produced along the fusion zone­steel interface. Two different reaction layers divided by the (1091ºC), which could cause cata-
thickness of the reaction layer (2 m) were thereafter identified by TEM analysis. The phas­ strophic vaporization of the molten
es formed adjacent to the steel substrate were Fe(Al) solid solution and Al8(Mn, Fe)5 phase magnesium if they melt simultaneous-
on the Fe(Al) surface, while the Al6Fe phase formed when the reaction layer was more than ly. In addition, the maximum solid sol-
2 m. Three different fracture modes were distinguished from all the joints, which was ubility of Fe in Mg is only 0.00041
closely associated with the interfacial reaction. Insufficient atomic diffusion caused interfa­ at-%. Their immiscibility and low solid
cial failure. Fracture occurred in the fusion zone when the thickness of the Al interlayer was
solubility suggests they do not react
0.2–0.5 mm. The maximum value of the tensile­shear fracture load reached 133 N/mm,
which represented 40.3% joint efficiency with respect to the steel base metal. However, the with each other or mix in the liquid
excessive interfacial reaction resulted in cracking at the interface and a greater amount of state at ambient pressure. Thus, met-
brittle Mg­Al compounds, giving rise to decreased mechanical properties. allurgical bonding between these two
materials would only be possible if an-
other intermediate element that can
KEYWORDS interact with or possess substantial
solid solubility with both of them can
• Laser Welding • Dissimilar Metals • Magnesium Alloy • Microstructure
be added during the welding process.
• Mechanical Properties
The feasibility of joining Mg to
steel using various processes, such as
Introduction which has drawn considerable attention friction stir welding (FSW) (Refs. 3–8),
in the past few years (Ref. 2). Currently, resistance spot welding (RSW) (Refs.
Vehicle design and fabrication has steel sheets remain the most commonly 9, 10), diffusion bonding (DB) (Refs.
been increasingly innovated by automo- used material in the automotive indus- 11–13), cold metal transfer (CMT)
tive manufacturers and their suppliers try. They are considered the most cost- welding (Ref. 14), hybrid laser-arc
in the effort to maximize fuel efficiency effective material for vehicle applica- welding (Refs. 15–20), and laser weld-
to meet the new, more stringent fuel tions. Therefore, the fabrication of hy- ing-brazing or brazing (Refs. 21–27)
and emission standards, while main- brid lightweight structural components have previously been investigated.
taining critical safety requirements (Ref. of magnesium alloys and steel, becomes The FSW process has been reported
1). To realize this goal, automakers are a feasible method to further lower the to successfully realize the bonding of
working to develop new lightweight ma- vehicle weight. Reliable joining of Mg to Mg and steel by accelerating the diffu-
terials and integrate them into vehicle steel would facilitate the increased use sion of Al atoms from the Mg base
designs. Magnesium alloys have been of magnesium alloys and expand the metal to steel with the combined ef-
considered as promising materials for scope of their application in the auto- fect of external force and strong stir-
automotive applications due to their motive industry. ring. The Fe-Al phase formed, acting
low density and high specific strength, Direct joining of magnesium alloys as the transitional compound to bond

C. W. TAN (tancaiwang@163.com), B. CHEN, X. G. SONG (songxg@hitwh.edu.cn), L. ZHOU, and S. H. MENG are with the Shandong Provin­
cial Key Laboratory of Special Welding Technology, Harbin Institute of Technology at Weihai, Weihai, China. TAN, L. Q. LI, and J. C. FENG are
with the State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin, China.

384-s WELDING JOURNAL / OCTOBER 2016, VOL. 95

TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 385
A A
Fig. 1 — Schematic of laser welding and the testing specimen:
A — Laser welding process of Mg to steel; B — tensile­shear test
specimen.
Mg and steel. The results achieved us-
ing RSW showed low interplanar mis-
matching of Fe2Al5/Fe (0%) and
Fe2Al5/Mg (4.8%) heterophase inter-
faces, indicating the Fe-Al phase could
act as an interlayer between Mg and
Fe.
However, without external force,
utilizing various interlayers has been
the main way to solve the metallurgi-
cal bonding problems of Mg/steel. The
benefits of using different interlayers
and elements, such as Cu, Ni, Sn, Zn,
and Al-12Si, have been explored.
Elthalabawy and Khan (Refs. 11, 12)
tried to join Mg to steel using diffu-
sion bonding (transient liquid bond-
ing), in which the Cu and Ni interlay-
ers were used. Solid-state diffusion,
eutectic formation, and the formation
of ternary intermetallic compounds
were observed at the interface. The in-
terfacial reaction was intense in the
liquid state, inducing the excessive for-
mation of brittle and thick layers,
which was detrimental to the joint
strength.
Cao et al. (Ref. 14) investigated the
feasibility of CMT welding of Mg to
galvanized mild steel. The Zn coating
was found to promote the wetting of
the AZ61 filler metal on the steel sur-
face. A sound CMT welded Mg/coated
steel joint was obtained with the joint
strength comparable to a Mg-Mg fu-
sion joint.
WELDING RESEARCH
C
Laser welding
(brazing), as an alter-
native advanced
welding technique, D
has received consid-
erable attention
due to its potential
for flexible joining
of dissimilar met-
als. Liu et al. (Refs.
15–17, 20) used hy-
brid laser-arc weld-
ing to bond Mg and Fig. 2 — Weld appearances of Mg/steel joints produced with dif­
steel with Cu, Ni, ferent thicknesses of Al interlayers: A — 0.1 mm; B — 0.3 mm;
and Sn interlayers, C — 0.5 mm; and D — 0.7 mm.
respectively. These
thin interlayers joint could be achieved with the pres-
were placed in between upper Mg and ence of Zn coating on the steel surface
lower steel sheets. Typically, laser with using laser conduction welding.
low power penetrated from the Mg To minimize the vaporization of
sheet and the interlayer into the lower Mg sheet and improve welding flexibil-
steel sheet. These added interlayers ity, laser brazing with a preset Mg-
were heated and melted to react with based filler was employed. Various in-
Mg and steel, resulting in the forma- terlayers, including Al-12Si (Ref. 24),
tion of a transitional layer in the fu- Ni (Ref. 26), and Sn (Ref. 27), were
sion zone and solid solution close to used. When using an Al-12Si coating,
the steel side. The tensile-shear formation of -Fe(Al,Si)3 was observed
strength was improved because the along the fusion zone-steel interface.
bonding status was changed from me- In the case of the Ni interlayer, forma-
chanical bonding to semimetallurgical tion of Fe(Ni) solid solution on the
bonding. A severe evaporation of the steel surface was found to be the key
upper Mg base metal occurred during for metallurgical bonding of Mg/steel
the process, which limited the applica- immiscible systems. The joint strength
tion of this welding technique. was reported to be higher than that
In addition, Wahba and Katayama welded with Al coated steel. After-
(Ref. 21) also reported the desirable ward, they selected another viable in-
OCTOBER 2016 / WELDING JOURNAL 385-s
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 386
WELDING RESEARCH
A B
D E
Fig. 3 — Cross sections of Mg/steel joints produced with different thicknesses of Al interlayers: A — 0.1 mm; B — 0.2 mm; C — 0.3 mm;
D — 0.4 mm; E — 0.5 mm; and F — 0.7 mm.
terlayer, Sn, following a review of bina-
ry and ternary phase diagrams. The re-
sults suggested that all tensile-shear
specimens fractured in the steel base
metal. The fracture load far exceeded
that achieved with the former two
kinds of interlayers. The Sn coating
was reported to promote wetting be-
tween the molten filler metal and steel
sheet, creating an oxide-free steel sur-
face for metallurgical bonding. Forma-
tion of the Fe(Al) solid solution and
Al8(Mn, Fe)5 on the Fe(Al) surface,
with their low interplanar mismatch,
was mainly responsible for high joint
strength.
Therefore, selection of an appropri-
ate intermediate alloying element for
joining Mg to steel was vital for enhanc-
ing good wetting-spreading ability and
interfacial bonding between the fusion
zone and steel without producing too
thick brittle reaction layers. In our pre-
vious studies, the influence of different
coating surfaces (Refs. 28, 29) and alloy-
ing elements from filler metal and steel
(Refs. 30–32) have been investigated.
The important finding was that the Al
element, as a key element diffusing
from the filler metal into the interface,
was accelerated and induced by the
chemical potential, resulting in metal-
lurgical bonding of the Mg/steel inter-
face. The Fe-Al phase formed at the
Mg/steel interface was expected to be
386-s WELDING JOURNAL / OCTOBER 2016, VOL. 95
Table 1 — STEM­EDS Results of Points Indicated in Fig. 7 (at­%)
Positions Mg
P1 95.8
P2 52.6
P3 — 5.5
favorable for the mechanical properties
after the aforementioned studies.
In the present work, the Al interlay-
er was employed as the transitional in-
terlayer, which was expected to induce
the Fe-Al reaction after melting and
produce the Fe-Al phase to realize
metallurgical bonding of Mg/steel.
The aim of this study was to investi-
gate the influence of different Al inter-
layer thicknesses on interfacial reac-
tions and mechanical properties dur-
ing laser welding of Mg to steel. The
microstructure characteristics were
observed and identified. The mechani-
cal properties were evaluated. Finally,
the bonding mechanism was expected
to be elucidated.
Experimental Procedures
Selected as the base metals were
1.5-mm-thick AZ31B Mg alloy (Mg-
3Al-1Zn-0.2Mn, wt%) and 1-mm-thick
Q235 mild steel (0.17C-0.7Mn-0.35Si-
C
F
Al Mn Fe
3.8 0.4 —
27.1 13.0 7.3
— 94.5
0.035S-0.035P). The sheets were cut
into rectangular strips 30 mm wide
and 100 mm long. The size of the Al
interlayer used in the present work
was 3 mm wide and 100 mm long, so
the Al interlayer could totally melt and
react with the Mg or steel sheets. Be-
fore welding, the surfaces of the Mg al-
loy and steel sheets were cleaned with
abrasive paper to remove surface ox-
ides, and then cleaned in acetone and
other contaminants from the surfaces.
A fiber laser system with a maxi-
mum power of 10 kW (IPG YLR-
10000) and a KUKA six-axis robot
were used in this study. The laser
beam had a wavelength of 1070 nm
and beam parameter product of 7.2
mm mrad. It was transmitted by a
200-μm core-diameter fiber and fo-
cused by a 200-mm lens to obtain a
spot size of 0.2 mm.
Figure 1 shows the schematic of
laser welding. The experiments were
carried out in a lap joint configuration
with the Mg sheet clamped on the
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 387

WELDING RESEARCH

A B

C D

Fig. 4 — Microstructure morphologies of interfacial reaction layers with different thicknesses of Al interlayers: A — 0.1 mm; B — 0.3 mm;
C — 0.5 mm; and D — 0.7 mm.

steel sheet. The laser beam was irradi-
ated on the surface of the AZ31B Mg
alloy vertically. Argon shielding gas
was provided at a flow rate of 20 L/
min to prevent oxidation. The laser
beam was defocused to irradiate a
large area. The process parameters em-
ployed in the study were as follows:
laser power 800 W, welding speed 0.3
m/min, and defocused distance +20
mm.
After laser welding, typical cross
sections of the welded specimens were
cut and mounted in epoxy resin. Stan-
dard grinding and polishing prepara-
tion procedures were then utilized.
The appearances and cross sections
were observed using an optical micro-
scope (OM). The reaction layer at the
interface between the fusion zone and
steel were observed using scanning
electron microscopy (SEM) in back-
scattered electron (BSE) mode.
A transmission electron microscopy
(TEM) foil of the bonded region was
prepared using the focused ion beam
(FIB) technique. The preparation for
the FIB-TEM specimen was made us-
ing an in-situ lift out method. The
TEM with a Tecnai-G2 F30 operating
at a nominal voltage of 300 kV was
used to characterize the microstruc-
ture in detail.
Phase identification was investigat-
ed by selected-area electron diffraction
(SAED) combined with energy-disper-
sive spectroscopy (EDS) in scanning
transmission electron microscopy
(STEM) mode. A Vickers hardness
measurement was performed across
the fusion zone-steel interface and fu-
sion zone adjacent to the interface, re-
spectively. A test load of 0.1 Kgf and a
dwell time of 10 s were utilized. The
tensile-shear tests were performed at
room temperature using an Instron
5569 at a crosshead speed of 1 mm/
min.
Shims were clamped to each end of
the specimens to ensure shear loads in
the lap joint while minimizing bending
or torque of the specimens. Joint
strength was calculated via the tensile
testing of at least three specimens, and
the average was reported with the stan-
dard deviation provided via error bars.

OCTOBER 2016 / WELDING JOURNAL 387-s

TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 388
WELDING RESEARCH
A ty of Al on the steel surface led to
B lurgical bonding of the immiscible Mg
C ately near the interface. The reaction
Fig. 5 — EDS line scan results of Mg, Al, Fe, and Mn across the fusion zone­steel interface
at different thicknesses of the Al interlayer: A — 0.1 mm; B — 0.2 mm; and C — 0.3 mm.
Results and Discussion
Joint Appearances
Figure 2 shows the joint appearances
with different Al interlayer thicknesses.
No spatter or obvious defect was evi-
denced in the figure because of suitable
heat input employed in the present
work. However, a nonuniform and
rough surface was first observed when
the interlayer was thin, as shown in Fig.
2A and 2B. It was mainly attributed to
the instability of the Mg molten pool at
the action of laser energy. With the in-
crease of Al thickness, the joint surface
was improved, obtaining a smooth and
uniform appearance indicated in Fig. 2C
and 2D.
388-s WELDING JOURNAL / OCTOBER 2016, VOL. 95
From the cross-sectional views
shown in Fig. 3, the Al foil was not ob-
served as a separate layer along the in-
terface after the process, indicating it
entirely melted and was dissolved into
the Mg liquid adjacent to the interface.
The opening was observed at the edge
of the fusion zone close to the steel in-
terface due to the bad wetting-spread-
ing ability of Mg and steel. The open-
ing defect was improved with the addi-
tion of Al foil. With an increasing
thickness of the interlayer, the joint
width became larger due to good wet-
ting of Al and steel. It was believed
that the affinity of Al and Fe was much
stronger than that of Mg and Fe. Thus,
the thin Al liquid adhering to the steel
interface reduced the difficulty of the
Mg molten pool spreading on the steel
surface. Good wetting-spreading abili-
more flowability, as demonstrated in
Fig. 3E and F.
Microstructural Analysis
Figure 4 shows microstructure mor-
phologies along the fusion zone-steel
interface with different thicknesses of
the Al interlayer. Formation of the in-
terfacial reaction layer was observed
from all the joints, confirming metal-
and Fe was realized by adding the Al
interlayer.
In addition, the thickness of the Fe-
Al reaction layer formed at the inter-
face was found to increase with the in-
creasing Al interlayer. When the thick-
ness of the Al interlayer was 0.1 mm,
no obvious reaction layer was noticed,
even at higher magnification as shown
in Fig. 4A, suggesting most of the Al
interlayer melted and was mixed with
the molten Mg fusion zone immedi-
layer was evidently observed when in-
creasing the thickness of the Al inter-
layer to 0.3 mm. These reaction prod-
ucts exhibited nonuniform and faceted
morphology. With a further increase
of the Al interlayer to 0.7 mm, the re-
action layer grew dramatically, reach-
ing its thickness of more than 10 m.
The crack was evidently observed at
a higher magnification indicated in the
inset of Fig. 4D, which deteriorated
joint strength. The results were in ac-
cordance with the previous report that
cracking may easily occur if the thick-
ness of the Fe-Al intermetallic layers
exceeded 10 m, which was detrimen-
tal to mechanical properties (Ref. 33).
EDS line scanning analyses were
performed to obtain concentration
profiles of the main alloying elements
across the interface between the Mg
fusion zone and steel. Figure 5 shows
the corresponding line scan results.
The Mg element decreased gradually
from the fusion zone to the steel side,
while the Fe content was varied in an
opposite way. Note that an apparent
Al concentration peaked at the inter-
face in all joints, indicating an occur-
rence of atomic diffusion or dissolu-
tion of the Al element, which there-
after induced interfacial reaction. Ad-
ditionally, the diffusion distance be-
came longer with the increasing Al
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 389
A B
C D
Fig. 6 — Effect of Al interlayer with different thicknesses on the microstructure in the
fusion side near the interface of Mg/steel: A — 0.1 mm; B — 0.3 mm; C — 0.5 mm; and
D — 0.7 mm.
A B
Fig. 7 — Overview of TEM images and
STEM­EDS analysis result: A — TEM spec­
imen; B — TEM image of the reaction
layer with its thickness of less than 2 m;
and C — STEM­EDS line scan result.
thickness, which caused formation of
more Fe-Al reaction layers.
Figure 6 shows the microstructure
variation adjacent to the fusion zone-
steel interface with the different thick-
nesses of the Al interlayer. The Al in-
terlayer was heated and melted by
laser irradiation during the process.
Part of the molten Al interlayer was
dissolved into the upper Mg molten
C 5.5 at-% Al and 94.5 at-% Fe. As
pool close to the interface. Mutual dif-
fusion of Mg-Al atoms then took place
in a fast rate due to their liquid state.
Mg-Al intermetallic compounds would
OCTOBER 2016 / WELDING JOURNAL 389-s
WELDING RESEARCH
thus form after solidification. Their
morphology was varied with the par-
ticipation of different quantities of Al
atoms.
The precipitation of scattered Mg-
Al intermetallic compounds (Mg17Al12
phase) was noticed in Fig. 6A when the
thickness of the Al interlayer was 0.1
mm. These dispersed particles played
an important role in the improvement
of mechanical properties by dispersion
strengthening (Ref. 34). Similar phe-
nomenon and effects were reported
with the formation of finer and denser
Al-Cu particles when FSW Al to Cu
(Ref. 35). With the increase of Al inter-
layer, the quantity of precipitated Mg-
Al reaction products increased. When
the thickness of the Al interlayer ex-
ceeded 0.5 mm, network structure
formed at the fusion zone, which
made the weld brittle and deteriorated
the mechanical properties of the joint.
Phase Identification
Transmission electron microscopy
analysis was performed to further
identify the composition and structure
of the reaction layers formed between
the fusion zone and steel.
Figure 7 shows a typical bright field
TEM image of the fusion zone-steel in-
terface when the thickness of the reac-
tion layer was not more than 2 m. A
TEM foil was prepared at the interface,
as shown in Fig. 7A. A noncontinuous,
ultra-thin reaction layer was found to
exist between the Mg fusion zone and
steel substrate indicated in Fig. 7B. The
elemental composition at different posi-
tions indicated in Fig. 7B are listed in
Table 1 based on STEM-EDS results.
Location P2 contained 13.0 at-%
Mn, which was much higher than the
base metals. Location P3 contained
shown in Fig. 7C, a relatively long dif-
fusion distance (a depth of about 550
nm) of Al element into the steel side
was observed from the STEM-EDS line
scan result. Meanwhile, the Mn ele-
ment was enriched at the interface
close to the Mg fusion zone side. Con-
sequently, the Al-Mn phase and Fe-Al
reaction layer was expected to form at
the interface.
Figure 8 shows a TEM micrograph
with SAED patterns corresponding to
Fig. 7. The ultra-thin reaction prod-
ucts consisted of two different phases,
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 390
WELDING RESEARCH
A B
Fig. 8 — TEM bright field image and SAED patterns corresponding to Fig. 7: A — TEM mi­
crograph; B–D — SAED patterns of interfacial phases.
as shown in Fig. 8A.
The phases close to the steel sub-
strate were found to be continuous
and uniform followed by noncontinu-
ous phases toward the fusion zone.
The nonuniform phases close to the
fusion zone were identified as Al8Mn5,
while some Fe element was also de-
tected. Therefore, the Al8(Mn,Fe)5 in-
termetallic compound was exactly in-
dexed, in which some Mn atoms were
replaced by Fe atoms because the
atomic radius of Mn (0.112 nm) and
Fe (0.124 nm) were close (Ref. 27).
The phase adjacent- to the steel sub-
strate was indexed by TEM and identi-
fied as Fe(Al) solid solution with a
body-centered cubic (BCC) crystal
structure, respectively. The SAED was
-
taken along the [111] zone axis of the
phase. These findings were consistent
with the observation of the interfacial
reaction layer when laser brazing Mg
to Sn coated steel sheet (Ref. 27).
To prove the degree of solid solu-
tion Al in Fe, the interatomic misfit
and interplanar mismatch between
two phases can be calculated by
a  a0
=  100%
a0
where |a – a0| is the difference be-
tween interatomic or interplanar spac-
390-s WELDING JOURNAL / OCTOBER 2016, VOL. 95
C The hardness distribution across
D tively. Note that an abrupt change in
ings of the two phases and a0 is the in-
teratomic or interplanar spacing of the
steel substrate. The a0 was 0.287 nm.
The misfit was 4.2%; it reduced to
0.5% with the distance from the steel
interface of 2000 nm after calculation.
The results corresponded to the long
diffusion of Al into the steel substrate
obtained in Fig. 7C.
Figure 9 shows TEM images taken
at the fusion zone-steel interface when
the thickness of reaction products was
more than 2 m. The reaction layer
was clearly evidenced at the interface.
The relatively thick Al interlayer melt-
ed, providing more Al atoms at the in-
terface, making the atomic number of
Al dominate in the whole Fe-Al sys-
tem. As a result, it was possible that
the Al-rich Al-Fe intermetallic com-
pound formed, rather than solid
solution.
High magnification of the TEM im-
age at the interface indicated the reac-
tion layer was distinctly different from
the fusion zone and steel as shown in
Fig. 10B. Combined with the EDS and
SAED calibration results, the Al6Fe
phase with orthorhombic structure was
identified. The metastable state Al6Fe
was precipitated at the temperature of
about 820ºC under the condition of the
rapid thermal cycle. In addition, the
Mg17Al12 intermetallic compounds also
existed adjacent to the interface, corre-
sponding to the observation of SEM
morphology in Figs. 4 and 6.
Mechanical Properties
Figure 11 shows Vickers hardness
distribution profiles across the inter-
face and near the fusion zone.
the fusion zone/steel interface with
the Al thickness of 0.2 mm was meas-
ured, and the result is shown in Fig.
11A. The average hardness of the fu-
sion zone and steel substrate was ap-
proximately 100 and 209 HV, respec-
the hardness value at the interface was
noticed, which was higher than the fu-
sion zone and neighboring steel sub-
strate, reaching the maximum value of
587 HV. The distinct rise in the hard-
ness at the interface was closely asso-
ciated with the presence of interfacial
reaction products corresponding to
the observation in Fig. 4.
Figure 11B presents the hardness
distribution profile at the fusion zone
near the interface. The hardness of the
AZ31 Mg alloy base metal was about
60 HV when without Al foil. The thick-
er the added Al interlayer, the higher
the hardness close to the fusion zone.
It was evident that the hardness in-
creased toward the interface due to
the precipitation of more Mg-Al
compounds.
Figure 12 shows the tensile-shear
fracture load of the laser welded mag-
nesium-steel joints and the thickness
of the reaction layer with the variation
of thicknesses of the Al interlayer and
the fracture load with a function of
thickness of reaction layer. The joint
strength was given here as fracture
load because it had both tensile and
shear stresses during tests. Three dif-
ferent fracture modes were distin-
guished in the present work. The frac-
ture behavior depended upon the in-
terfacial bonding force and brittleness
of the fusion zone. The fracture load
of the joint produced with an Al thick-
ness of no more than 0.1 mm was
quite low, resulting in interfacial fail-
ure. The interfacial bonding was weak
due to insufficient atomic diffusion or
interfacial reaction.
Fracture occurred at the faying
surface of the Mg and steel sheets.
With increasing thickness of the Al
interlayer (0.2–0.5 mm), the thick-
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 391
A B
Fig. 9 — Diffraction patterns of the Fe­Al phase at different locations of 0, 100, and 2000 nm with the point indicated in Fig. 8.
A B
C patterns obtained when the thickness of the
ness of the reaction layer formed at
the interface was 2.3–5.7 m. The
fracture load was enhanced to maxi-
mum value of 133 N/mm, which rep-
resented 40.3% joint efficiency with
respect to the mild steel base metal,
as shown in Fig. 11B.
Joint efficiency was calculated by
comparing the maximum load of the
joint when fractured in lap shear ten-
Fig. 10 — TEM bright field image and SAED
reaction layer was more than 2 m: A–B —
TEM micrograph; C — SAED patterns of the
interfacial phases indicated in B.
sile testing with the maximum load-
carrying capacity of the weaker base
sheet section. The calculated joint effi-
ciency was reported in a previous
study (Ref. 30). In this case, the frac-
ture location was at the Mg fusion
zone, indicating the Mg/steel interface
was not the weak point.
Compared to high interfacial bond-
ing, the brittle precipitated Mg-Al in-
termetallic compounds made the crack
tend to initiate and propagate at the
fusion zone. The fracture load de-
creased sharply with further increasing
the thickness of the Al interlayer. The
crack formed in between the interface
of the Fe-Al reaction layer and steel
substrate, which became the weakest
part of the joint despite the presence
of brittle Mg-Al compounds in the fu-
sion zone close to the interface.
OCTOBER 2016 / WELDING JOURNAL
WELDING RESEARCH
C
Figure 13 presents the fracture sur-
face morphologies of joints welded at
different thicknesses of the Al inter-
layer. For the joint with interfacial fail-
ure mode, some residual Mg fusion
zone was attached to the steel sub-
strate, as shown in Fig. 13A. Higher
magnification indicated the fracture
surface was characterized by the tear-
ing ridge — Fig. 13B.
When the joint fractured at the Mg
fusion zone, the fracture surface exhib-
ited dimples together with some cleav-
age-like flat facets demonstrated in Fig.
13D, which was similar to the observa-
tion of Mg-Mg fusion welding (Ref. 36).
The flat surface suggested brittleness of
the Mg-Al intermetallic compounds in
the fusion zone. It was believed that the
more brittle the microstructure, the
flatter the fracture surface. With the in-
creasing thickness of the reaction layer,
the fracture was found to occur along
the interface between the reaction layer
and steel substrate due to the presence
of cracks formed in between them after
the process.
Smooth fracture surface without
adequate plastic deformation was ob-
served on the steel side, as shown in
Fig. 13E. Cracking was noticed at the
higher magnification arrowed in the
inset of Fig. 13E. Remnants of the fu-
sion zone adhering to the fracture sur-
face shown in Fig. 13F were found to
be more brittle compared to the result
in Fig. 13D.
Bonding Mechanism
Based on the analyses above, the
bonding mechanism of Mg to steel
391-s
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 392

WELDING RESEARCH

A B metric composition of Al6Fe, which

was precipitated when the tempera-
ture decreased to 820ºC. Some of the
Fe atoms were replaced by a small
amount of Mn atoms diffusing from
the steel. After that, the eutectic struc-
ture was produced in the molten pool,
which was much denser than that
formed with the thin Al interlayer.

Conclusion
Fig. 11 — Hardness distribution profile: A — Across the interface of Mg/steel; B — near the 1) With the addition of the Al inter-
interface with the different thicknesses of the Al interlayer. layer, successful joining of Mg to steel
was realized by laser welding. A visually
acceptable and uniform joint was
A B achieved with the assistance of a good
affinity of Al and steel. The best joint
quality was obtained using the following
process parameters: 800-W laser power,
0.3-m/min welding speed, +20-mm de-
focused distance, and 0.3-mm-thickness
of Al interlayer.
2) Al interlayer was dissolved into
the Mg fusion zone and mutual diffu-
sion with steel, causing metallurgical
bonding of Mg and steel at the inter-
face, and more precipitation of the Mg-
Fig. 12 — Tensile­shear fracture load of the laser welded magnesium­steel joints: A — Al compounds. A reaction layer formed
Fracture load and thickness of the reaction layer with the variation of thicknesses of the Al along the interface of the fusion zone-
interlayer; B — fracture load with a function of the thickness of the reaction layer. steel joint. The thickness increased with
an increasing thickness of the Al inter-
with the addition of the Al interlayer face was saturated inducing crystalliza- layer. Cracking was observed between
was elucidated. tion of Fe(Al) when the Al interlayer the reaction products and steel sub-
Figure 14 shows the schematic dia- was very thin, as indicated in Fig. 14C. A strate when the thickness was larger
gram of the joining mechanism. First, thin layer of Fe(Al) formed first from than 10 m. Precipitation of the Mg-Al
melting of the Mg sheet and Al inter- the liquid by solid-state diffusion of Al intermetallic compounds was noticed at
layer occurred simultaneously when atoms in the liquid filler metal into the the fusion zone close to the interface.
suffering the laser beam irradiation, as steel. Upon further cooling, some Mn The quantity of the compounds in-
shown in Fig. 14B. The liquid Mg atoms diffusing from the steel substrate creased with the increase of the Al inter-
atoms and Al atoms dissolved into were then bonded with Al atoms, giving layer. The excessive precipitation caused
each other. At the same time, some Al rise to the Al8Mn5 phase under the tem- the brittleness of the fusion zone and
atoms close to the liquid/solid inter- perature of 680ºC (Ref. 37). decreased the fracture load.
face diffused into the steel surface. The phase nucleated and grew on 3) Two different reaction layers were
Because of the laser irradiation, the the first precipitated Fe(Al) surface, as identified by TEM analysis. When the
atoms in the steel adjacent to the in- shown in Fig. 14D. Part of the Mn thickness of the reaction layer was less
terface were also activated at a high atoms were replaced by the Fe atoms than 2 m, the phases close to the steel
temperature and slightly dissolved due to their similar atomic radius and substrate were the Fe(Al) solid solution
into the liquid. The diffusion distance metallurgical characteristics. When and Al8(Mn, Fe)5 phase on the Fe(Al)
was relatively short due to fast ther- the temperature decreased to 650ºC or surface, while the Al6Fe phase formed
mal cycle during laser welding. As a re- below, -Mg first precipitated from when the reaction layer was more than
sult, different percents of the main al- the remaining liquid. 2 m.
loying elements Mg, Al, and steel were Finally, a eutectic reaction occurred 4) Three different fracture modes
mixed from fusion zone to the in the liquid, producing a eutectic were distinguished from all the joints
liquid/solid interface. The different structure ( – Mg + Mg17Al12). In the produced with different thicknesses of
precipitation modes were distin- case of the thick Al interlayer, more Al the Al interlayers, which was closely as-
guished with the variation of thethick- atoms segregated at the front of liq- sociated with the interfacial reaction.
ness of the Al interlayer. uid/solid interface. Therefore, differ- Insufficient atomic diffusion caused in-
Upon cooling, as the temperature de- ent Fe-Al phases formed depending on terfacial failure with the fracture at the
creased to about 1000ºC, the solid solu- the ratio of Al and Fe atoms. The con- faying surface of the Mg and steel sheet.
bility of Al in Fe at the solid/liquid inter- tent of Al and Fe reached the stoichio- The suitable interfacial reaction oc-

392-s WELDING JOURNAL / OCTOBER 2016, VOL. 95

TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 393
A B
C D
E F
Fig. 13 — Fracture surfaces morphologies of the Mg/steel joints at different fracture modes:
A, C, and E — Overviews of the fracture surface showing different fracture modes; B, D, and
F — higher magnification of the area indicated by B, D, and F, respectively.
curred when the thickness of the Al in-
terlayer was 0.2–0.5 mm, resulting in
fusion zone fracture. The maximum val-
ue of tensile-shear fracture load reached
133 N/mm, which represented 40.3%
joint efficiency with respect to the steel
base metal. However, the excessive in-
terfacial reaction resulted in the crack-
ing at the interface and greater amount
of brittle Mg-Al compounds, giving rise
to decreased mechanical properties.
Acknowledgments
The authors wish to acknowledge
support from the National Natural Sci-
ence Foundation of China (No.
51504074 and 51405099), Natural
Science Foundation of Shandong
Province, China (No. BS2015ZZ008),
and China Postdoctoral Science Foun-
dation (No. 2015M571406).
References
1. Schubert, E., Klassen, M., Zerner, I.,
Walz, C., and Sepold, G. 2001. Light-
weight structures produced by laser beam
joining for future applications in automo-
bile and aerospace industry. Journal of
Materials Processing Technology 115: 2–8.
2. Mordike, B., and Ebert, T. 2001. Mag-
OCTOBER 2016 / WELDING JOURNAL 393-s
WELDING RESEARCH
nesium: Properties — applications — po-
tential. Materials Science and Engineering: A
302: 37–45.
3. Liyanage, T., Kilbourne, J., Gerlich, A.
P., and North, T. H. 2009. Joint formation
in dissimilar Al alloy/steel and Mg alloy/
steel friction stir spot welds. Science and
Technology of Welding and Joining 14:
500–508.
4. Schneider, C., Weinberger, T., Inoue,
J., Koseki, T., and Enzinger, N. 2011.
Characterisation of interface of
steel/magnesium FSW. Science and Tech-
nology of Welding and Joining 16: 100–107.
5. Santella, M., Brown, E., Pozuelo, M.,
Pan, T. Y., and Yang, J. M. 2012. Details of
Mg-Zn reactions in AZ31 to galvanised
mild steel ultrasonic spot welds. Science
and Technology of Welding and Joining 17:
219–224.
6. Chen, Y. C., and Nakata, K. 2010. Ef-
fect of surface states of steel on mi-
crostructure and mechanical properties of
lap joints of magnesium alloy and steel by
friction stir welding. Science and Technolo-
gy of Welding and Joining 15: 293–298.
7. Jana, S., Hovanski, Y., and Grant, G.
J. 2010. Friction stir lap welding of mag-
nesium alloy to steel: A preliminary inves-
tigation. Metallurgical and Materials Trans-
actions A 41: 3173–3182.
8. Wei, Y., Li, J., Xiong, J., Huang, F.,
and Zhang, F. 2012. Microstructures and
mechanical properties of magnesium alloy
and stainless steel weld-joint made by
friction stir lap welding. Materials & De-
sign 33: 111–114.
9. Liu, L., Xiao, L., Feng, J., Li, L., Es-
maeili, S., and Zhou, Y. 2011. Bonding of
immiscible Mg and Fe via a nanoscale
Fe2Al5 transition layer. Scripta Materialia
65: 982–985.
10. Liu, L., Xiao, L., Chen, D. L., Feng,
J. C., Kim, S., and Zhou, Y. 2013. Mi-
crostructure and fatigue properties of
Mg-to-steel dissimilar resistance spot
welds. Materials & Design 45: 336–342.
11. Elthalabawy, W. M., and Khan, T. I.
2010. Eutectic bonding of austenitic
stainless steel 316L to magnesium alloy
AZ31 using copper interlayer. The Interna-
tional Journal of Advanced Manufacturing
Technology 55: 235–241.
12. Elthalabawy, W. M., and Khan, T. I.
2010. Microstructural development of
diffusion-brazed austenitic stainless steel
to magnesium alloy using a nickel inter-
layer. Materials Characterization 61:
703–712.
13. Elthalabawy, W., and Khan, T.
2011. Liquid phase bonding of 316L
stainless steel to AZ31 magnesium alloy.
Journal of Materials Science & Technology
27: 22–28.
14. Cao, R., Yu, J. Y., Chen, J. H., and
Wang, P.-C. 2013. Feasibility of cold-met-
al-transfer welding magnesium AZ31 to
galvanized mild steel. Welding Journal
TAN SUPP OCT 2016_Layout 1 9/14/16 3:17 PM Page 394
WELDING RESEARCH
A B
C D
F G
Fig. 14 — Schematic of the bonding mechanism: A —
Preparation for laser welding with the Al interlayer; B —
laser irradiation; C–E and F–G — solidification at different
temperature ranges with the thickness of the formed reac­
tion layer less than 2 m and more than 2 m, respectively.
92(9): 274-s to 282-s.
15. Qi, X. D., and Liu, L. M. 2011. Inves-
tigation on welding mechanism and inter-
layer selection of magnesium/steel lap
joints. Welding Journal 90(1): 1-s to 7-s.
16. Liu, L., and Qi, X. 2010. Strengthen-
ing effect of nickel and copper interlayers
on hybrid laser-TIG welded joints between
magnesium alloy and mild steel. Materials
& Design 31: 3960–3963.
17. Liu, L., Qi, X., and Wu, Z. 2010. Mi-
crostructural characteristics of lap joint be-
tween magnesium alloy and mild steel with
and without the addition of Sn element.
Materials Letters 64: 89–92.
18. Qi, X., and Song, G. 2010. Interfa-
cial structure of the joints between magne-
sium alloy and mild steel with nickel as in-
terlayer by hybrid laser-TIG welding. Mate-
rials & Design 31: 605–609.
19. Qi, X. D., and Liu, L. M. 2010. Com-
parative study on characteristics of hybrid
laser-TIG welded AZ61/Q235 lap joints
with and without interlayers. Journal of
Materials Science 45: 3912–3920.
20. Liu, L., Qi, X., and Zhang, Z. 2012.
The effect of alloying elements on the
394-s WELDING JOURNAL / OCTOBER 2016, VOL. 95
shear strength of the lap
joint of AZ31B magnesium
alloy to Q235 steel by hy-
brid laser-TIG welding
technique. Metallurgical
and Materials Transactions
A 43: 1976–1988.
21. Wahba, M., and
Katayama, S. 2012. Laser
welding of AZ31B magne-
sium alloy to Zn-coated
E steel. Materials & Design
35: 701–706.
22. Miao, Y., Han, D.,
Yao, J., and Li, F. 2010. Ef-
fect of laser offsets on
joint performance of laser
penetration brazing for
magnesium alloy and steel.
Materials & Design 31:
3121–3126.
23. Miao, Y. G., Han, D.
F., Yao, J. Z., and Li, F.
2010. Microstructure and
interface characteristics of
laser penetration brazed
magnesium alloy and steel.
Science and Technology of
Welding and Joining 15:
97–103.
24. Nasiri, A. M., Li, L.,
Kim, S. H., Zhou, Y., Weck-
man, D. C., and Nguyen, T.
C. 2011. Microstructure
and properties of laser brazed magnesium
to coated steel. Welding Journal 90(11):
211-s to 219-s.
25. Nasiri, A. M., Chartrand, P., Weck-
man, D. C., and Zhou, N. Y. 2012. Thermo-
chemical analysis of phases formed at the
interface of a Mg alloy-Ni-plated steel joint
during laser brazing. Metallurgical and Ma-
terials Transactions A 44: 1937–1946.
26. Nasiri, A. M., Weckman, D. C., and
Zhou, Y. 2013. Interfacial microstructure
of diode laser brazed AZ31B magnesium to
steel sheet using a nickel interlayer. Weld-
ing Journal 92(1): 1-s to 10-s.
27. Nasiri, A. M., Weckman, D. C., and
Zhou, Y. 2015. Interfacial microstructure
of laser brazed AZ31B magnesium to Sn-
plated steel sheet. Welding Journal 94(3):
61-s to 72-s.
28. Li, L., Tan, C., Chen, Y., Guo, W.,
and Hu, X. 2012. Influence of Zn coating
on interfacial reactions and mechanical
properties during laser welding-brazing of
Mg to steel. Metallurgical and Materials
Transactions A 43: 4740–4754.
29. Tan, C. W., Chen, Y. B., Li, L. Q., and
Guo, W. 2013. Comparative study of mi-
crostructure and mechanical properties of
laser welded-brazed Mg/steel joints with
four different coating surfaces. Science and
Technology of Welding and Joining 18:
466–472.
30. Li, L., Tan, C., Chen, Y., Guo, W.,
and Song, F. 2013. Comparative study on
microstructure and mechanical properties
of laser welded–brazed Mg/mild steel and
Mg/stainless steel joints. Materials & De-
sign 43: 59–65.
31. Tan, C. W., Li, L. Q., Chen, Y. B.,
Nasiri, A. M., and Zhou, Y. 2014. Mi-
crostructural characteristics and mechani-
cal properties of fiber laser welded-brazed
Mg alloy-stainless steel joint. Welding Jour-
nal (10): 399-s to 409-s.
32. Tan, C., Song, X., Meng, S., Chen,
B., Li, L., and Feng, J. 2015. Laser welding-
brazing of Mg to stainless steel: joining
characteristics, interfacial microstructure,
and mechanical properties. The Interna-
tional Journal of Advanced Manufacturing
Technology. DOI 10.1007/s00170-015-
8165-0.
33. Laukant, H., Wallmann, C., Müller,
M., Korte, M., Stirn, B., Haldenwanger, H.
G., and Glatzel, U. 2005. Fluxless laser
beam joining of aluminium with zinc coat-
ed steel. Science and Technology of Welding
and Joining 10: 219–226.
34. Chen, X. L., Yan, H. G., Chen, J. H.,
Su, B., and Yu, Z. H. 2013. Effects of grain
size and precipitation on liquation cracking
of AZ61 magnesium alloy laser welding
joints. Science and Technology of Welding
and Joining 18: 458–465.
35. Tan, C. W., Jiang, Z. G., Li, L. Q.,
Chen, Y. B., and Chen, X. Y. 2013. Mi-
crostructural evolution and mechanical
properties of dissimilar Al–Cu joints pro-
duced by friction stir welding. Materials &
Design 51: 466–473.
36. Shen, Y. 2016. The influence of
cryogenic and heat treatment on the me-
chanical properties of laser-welded AZ91D.
The International Journal of Advanced Man-
ufacturing Technology 1–5.
37. Pawar, S., Zhou, X., Hashimoto, T.,
Thompson, G. E., Scamans, G., and Fan, Z.
2015. Investigation of the microstructure
and the influence of iron on the formation
of Al8Mn5 particles in twin roll cast AZ31
magnesium alloy. Journal of Alloys and
Compounds 628: 195–198.