ArticlePDF Available

Multiscale Study of Interfacial Intermetallic Compounds in a Dissimilar Al 6082-T6/Cu Friction-Stir Weld

Authors:

Abstract

The objective of this work was to characterize the Alx Cuy intermetallic compounds (IMCs) formed at the abutting interface during solid-state friction-stir welding (FSW) of 6082 aluminum alloy and pure copper. As IMCs are potential sources of flaws in case of mechanical loading of welds, their study is essential at various scale lengths. In the present case, they have been identified by neutron diffraction, electron backscattered diffraction, and transmission electron microscopy. Neutron diffraction analyses have shown that a shift of the tool from the interface, in particular towards the Cu part, generates an increase of the IMCs’ volume fraction. In accordance with an exacerbation of its kinetics of formation by FSW, a 4-μm-thick layer has precipitated at the interface despite the shortness of the thermal cycle. This layer is composed of two sublayers with the Al4Cu9 and Al2Cu stoichiometry, respectively. Convergent beam electron diffraction analyses have, however, disclosed that the crystallography of the current Al2Cu compound does not comply with the usual tetragonal symmetry of this phase. The Al2Cu phase formation results from both the local chemical composition and thermodynamics, whereas the development of Al4Cu9 is rather due to both the local chemical composition and the shortness of the local FSW thermal cycle.
1 23
Metallurgical and Materials
Transactions A
ISSN 1073-5623
Volume 43
Number 12
Metall and Mat Trans A (2012)
43:4655-4666
DOI 10.1007/s11661-012-1277-3
Multiscale Study of Interfacial Intermetallic
Compounds in a Dissimilar Al 6082-T6/Cu
Friction-Stir Weld
M.N.Avettand-Fenoël, R.Taillard, G.Ji
& D.Goran
1 23
Your article is protected by copyright
and all rights are held exclusively by The
Minerals, Metals & Materials Society and ASM
International. This e-offprint is for personal
use only and shall not be self-archived in
electronic repositories. If you wish to self-
archive your work, please use the accepted
author’s version for posting to your own
website or your institution’s repository. You
may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.
Multiscale Study of Interfacial Intermetallic Compounds in a
Dissimilar Al 6082-T6/Cu Friction-Stir Weld
M.N. AVETTAND-FENOE
¨
L, R. TAILLARD, G. JI, and D. GORAN
The objective of this work was to characterize the Al
x
Cu
y
intermetallic compounds (IMCs)
formed at the abutting interface during solid-state friction-stir welding (FSW) of 6082 alumi-
num alloy and pure copper. As IMCs are potential sources of flaws in case of mechanical
loading of welds, their study is essential at various scale lengths. In the present case, they have
been identified by neutron diffraction, electron backscattered diffraction, and transmission
electron microscopy. Neutron diffraction analyses have shown that a shift of the tool from the
interface, in particular towards the Cu part, generates an increase of the IMCs’ volume fraction.
In accordance with an exacerbation of its kinetics of formation by FSW, a 4-lm-thick layer has
precipitated at the interface despite the shortness of the thermal cycle. This layer is composed of
two sublayers with the Al
4
Cu
9
and Al
2
Cu stoichiometry, respectively. Convergent beam electron
diffraction analys es have, however, disclosed that the crystallography of the current Al
2
Cu
compound does not comply with the usual tetragonal symmetry of this phase. The Al
2
Cu phase
formation results from both the local chemical composition and thermodynamics, whereas the
development of Al
4
Cu
9
is rather due to both the local chemical composition and the shortness of
the local FSW thermal cycle.
DOI: 10.1007/s11661-012-1277-3
Ó The Minerals, Metals & Materials Society and ASM International 2012
I. INTRODUCTION
FRICTION-STIR welding (FSW) is a solid-state
bonding technique first developed for welding of Al
alloys (such as 2xxx and 7xxx), which are difficult to
fusion weld without hot cracking, porosity, or distor-
tion.
[13]
As FSW operates in the solid state, it was
naturally extended to the very strategic challenge of the
joining of dissimilar reactive materials
[3,4]
with the aim
of reducing the amount of intermetallic compounds
(IMCs) by comparison with traditional fusion welding
processes. IMCs are readily formed at the faying
interface during the welding of metals with high
chemical affinity. Due to their inherent brittleness, it is
necessary to prevent the extensive formation of contin-
uous IMCs layers along the interface in order to ensure
suitable mechanical properties to the joint.
[512]
This
goal requires a careful study of the IMCs microstructure
in order to investigate in detail their mechanism of
formation. Such knowledge should be pertinent so as to
decrease their kinetics of formation by means of
judicious modifications of the processing parameters
and/or of the interface chemical composition. More
precisely, it is important to underline that a continuous
reaction layer with an optimal thickness is most often
required in order to obtain sound bonds by diffusion.
[13]
Whatever the welding process and the chemical nature
of the parent alloys, too thick layers of intermetallics are
indeed deleterious for the mechanical behavior of the
joints.
[11,1418]
Favorable effects of small changes of
the chemical composition of the base materials on the
kinetics of formation and nature of the IMCs with
profitable consequences for the mechanical behavior of
the joints have also been reported.
[19,20]
For instance, in
the case of diffu sion welding of a Si-bearing Al alloy to a
Ti alloy, the presence of a silicon content of 1.2 pct was
shown to slow down the Al diffusion towards Ti; Si
enabled to master the growth of the diffusion layer at the
interface, improving the mechanical properties of the
joint.
[20]
Likewise, 1.5 pct Si and 1 pct Cu additions to
6000 Al alloys have been shown to change the nature of
IMCs formed at the 6000 Al alloy/steel interface
obtained by diffusion welding. Instead of Al
3
Fe, a Cu-
rich Al
12
FeSi IMC has formed. These modifications of
both the nature and the crystal structure of the IMCs in
the interfacial layer have lead to an increase of the weld
ductility, very likely because of a multiplication of slip
planes.
[19]
In addition, a sustained attention must be
paid to the effects of a change of the mechanism of the
chemical reaction by the presence of a coating or of a
transition layer with a judicious chemical composition at
the faying interface.
[16,17]
For instance, the insertion of a
thin foil of zinc at an Al alloy/pure Cu interface would
reduce the amount of harmful IMCs formed during
FSW, which gives rise to a significant increa se of the
average fracture load of the joints.
[16]
This result must
nevertheless be questioned because of both the rather
low melting point of Zn and the high chemical affinity
between Cu and Zn that overcomes the Al/Cu one.
[2124]
In the same way, and concerning this time coated
M.N. AVETTAND-FENOE
¨
L, Assistant Professor, R. TAILLARD,
Professor, and G. JI, Research Scientist, are with the Unite
´
Mate
´
riaux Et
Transformations (UMR CNRS 8207), Lille 1 University, 59655 Villeneuve
d’Ascq, France. Contact e-mail: marie-noelle.avettand-fenoel@univ-
lille1.fr D. GORAN, Scientist, is with the Bruker Nano GmbH, 12489
Berlin, Germany.
Manuscript submitted January 19, 2012.
Article published online June 27, 2012
METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4655
Author's personal copy
steel–Al alloy FSW joints, the presence of a Zn-based
coating causes the formation of an Al-Zn eutectic film,
which removes the oxide film at the Al surface and
therefore promotes the dissimilar bonding by means of
the formation of Al-rich Al
x
Fe
y
compounds.
[17]
The
nature of the interfacial IMC is further strongly depen-
dent on the thermomechanical aspect s of the welding
process, that is to say, on the homo geneity and amplitude
of the plastic deformation at and close to the interface, as
well as on the heating and cooling rates. Such a feature is
exacerbated for the FSW case where the distance of
atomic diffusion must be changed from place to place
because of both the variation between local thermal cycles
and the very high heterogeneity of the plastic flow
pattern.
[25,26]
Thermal diffusion occurs because a close
to 0.8 peak value of homologous temperature is required
in order to achieve a suitable material flow.
[3,4]
Mechan-
ical mixing and short-circuit diffusion may also happen as
very drastic plastic strains ranging from 2 to approxi-
mately 130
[2732]
are predicted to occur in local areas
within the nugget. The effectiveness of such phenomena is
the most likely as local and transient high supersatura-
tions of vacancies are generated at high strain rates
[33,34]
and therefore at the 1 to 650 s
–1[3,4,27,30,32,3539]
values
commonly assessed during FSW. Furthermore, the
favorable effect of a discontinuity of IMCs layers on the
joint strength has also been reported.
[40]
All of these
observations prove the chief interest of a thorough study
of the microstructure of the IMCs layers formed in
dissimilar welds in order to improve the mechanical
resistance of the bond by means of a change of the
processing parameters. Heretofore, only few studies have
dealt with the fine microstructure of the interfacial area in
dissimilar FSW welds, however, made of materials other
than Al and Cu alloys.
[4144]
According to the systems, the
interfacial zones were shown to consist either in interdif-
fusion or in intermetallic areas. By way of examples, i) the
presence of an interdiffusion zone extending over a 1.5 to
1.8 lm distance was proved by energy dispersive spec-
troscopy (EDX)/transmission electron microscop y
(TEM) at the interface of Fe-Ni samples
[41]
; ii) electron
diffraction an alyses put into evidence a 250-nm-thick
interfacial layer of Al
4
Fe in an Al alloy–austenitic
stainless steel joint
[42]
; iii) a 1-lm-thick interface made
of intercalated lamellae enriched in Al or Ti-V was
observed by EDX/TEM in Al 6061-T6–Ti-6 pct Al-4
pct V friction stir joint
[43]
; iv) the presence of the Al
12
Mg
17
IMC in the 1-lm-thick interfacial layer of an AA6040 Al
alloy–AZ31 Mg alloy friction-stir weld was further
established by electron diffraction.
[44]
For the case of
Al-Cu FSW, a rather global charact erization of the
interface has only been carried out.
[5,45]
X-ray diffrac-
tometry combined with selective etching has revealed the
presence of two sublayers of Al
2
Cu and Al
4
Cu
9
within a
1-lm-thick continuous and uniform layer at a Al-1060–
pure Cu interface.
[5]
For the same system, SEM analyses
have also shown that according to the rotational speed,
comprised between 400 and 1000 rpm, the thickness of
the interface changes from 0.3 to 3 lm, respectively.
[45]
The too macroscopic scale of these investigations justifies
the current study of a 6082-T6 Al alloy/Cu-a1 combina-
tion more especially as it is further concerned by
numerous potential uses. Potential applications of such
joints will be in the fields of power generation and energy
transmission because of both their attractive thermal and
electrical conductivities and of significant savings. This
article focuses on a multiscale study of interfacial IMCs.
The macroscopic investigation was essentially achieved
by neutron diffraction, whereas at a finer scale, the study
was performed by both electron backscattered diffraction
and transmission electron microscopy.
II. EXPERIMENTAL PROCEDURE
Two 5-mm-thick 6082-T6 alumi num alloy and Cu-a1
plates were abutted and friction-stir welded perpendicu-
larly to their rolling direction. Joining was conducted
under a 1500-kg controlled axial force with a clockwise
rotation rate of 800 rpm and at a transverse spee d of
750 mm/min. The tool was made of a heat-treated
Z38CD5 steel with a hardness of 50 HRC. It comprised
a 15-mm-diameter concave shoulder prolonged by a pin.
This 4.8-mm long probe was a M 4 threaded tapered
cylinder with a diameter of 9.7 mm at the root and 3 mm
at the edge, respectively. The tool was 2.5 deg backward
tilted and located at the faying interface (WI), or at a
distance of 1 mm from this interface either in the Cu
advancing side (WCu) or in the Al retreating side (WAl).
Neutron diffraction experiments were performed in the
bulk of the three kinds of seams by using the high-flux
multi-detector (800 cells) G4.1 diffractometer at Labora-
toire Le
´
on Brillouin in France. A wavelength of 2.423 A
˚
was used for the determination of both the nature and
volume fraction of the intermetallic compounds in the
welds. The investigations were made at room temperature
within a volume of 0.9, 0.5, or 0.4 cm
3
for the WCu, WAl,
or WI bead, respectively. All the data refinements dealing
with the cell parame ters and atomic positions were made
by means of the Rietveld method and making use of the
FullProf Suite program (LLB, CEA-Saclay, France).
[46]
The Thompson–Cox–Hastings pseudo-Voigt model fits
well to the shape of the diffraction peaks of the phases.
The intermetallic compounds located at or close to the
faying interface in the WCu sample were characterized at
a fine scale by TEM by means either of a Philips CM30
microscope (Phil ips, Amsterdam, the Netherlands) oper-
ating at 300 kV or of a 200 kV FEI Tecnai G2 20
apparatus (FEI Company, Hillsboro, OR). Both systems
were equipped with precession electron diffraction and
EDX devices. Thin foils were prepared by the focused
ionic dual-beam (FIB) technique and using an FEI
STRATA DB 325 microscope. They were located at the
joint interface and at a depth of 670 (zone 1) or 1500 lm
(zone 2) from the top surface of the WCu weld (Figure 1).
During cutting, the thin foil edge facing the ion beam was
protected by a 2-lm-thick deposit of platinum-tungsten.
Electron backscattered diffraction (EBSD) analyses were
also carried out in the vicinity of the WCu joint. A Bruker
Nano high-resolution e
-
Flash HRä EBSD detector
(Bruker AXS Inc., Madison, WI) mounted on a Zeiss
Supra 55VP field-emission gun scanning electron micro-
scopy (FEG SEM; Carl Zeiss, Oberkochen, Germany)
was used.
4656—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
Author's personal copy
III. RESULTS AND DISCUSSION
As detailed in previous papers,
[47,48]
the malleability
of Cu is insufficient for a suitable mechanical stirring,
which requires an homologous temperature close to
0.8.
[49]
Indeed, according to the high thermal conduc-
tivity of both materials on the one hand, and to a well
established expression of the peak temperature for Al
alloys as a function of the rotation rate and of the travel
speed of the tool on the other hand, welding is actually
expected to proceed at a close to 723 K (450 °C) peak
temperature,
[50]
and thus at a close to 0.75 and to 0.5
homologous temperature for the 6082 alloy and for the
Cu base material, respectively.
[47]
In agreement with this
insufficient malleability, determination of fracture initi-
ation sites in the welds by means of both SEM and
digital image correlation during tensile tests has proved
that the most harmful flaw is a lack of consolidation.
[48]
Due to this deficient adhesion, the resulting mechani cal
properties were rather poor. For instance, the yiel d
strength of the WCu joint amounts to 71 pct and 75 pct
of that of the Al alloy and Cu base materials in the
transverse direction, respectively. In accordance with
this incomplete consolidation this joint efficiency mark-
edly decreases with plastic deformation. The ultimate
tensile strength indeed corresponds to only 49 pct and
58 pct of those of the Al alloy and Cu base materials,
respectively. Besides this lack of consolidation, the
deleterious effect of the softened zone on the Al side
as well as the fracture of the intermetallic compounds
(Figure 2) are at most of secondary importance for the
mechanical behavior.
In spite of this result, the study of the nature and
causes of formation of the present interfacial interme-
tallic compounds remains of the uppermost importance
because of the promises of FSW with respect to the
challenge imposed by dissimilar welding.
A. Macroscopic Analysis of the Joint
This section deals with the effect of the tool offset on
the material flow with its consequences on the IMCs
volume fraction.
Figure 3 shows the refinements of neutron diffraction
patterns of the three kinds of welds. Irrespective of the
tool location, only two intermetallic compounds,
namely c
1
-Al
4
Cu
9
and h-Al
2
Cu, were detected. How-
ever, it is worthy to note that contrarily to the Al
2
Cu
case, it remains difficult to resolve the diffraction peaks
of the Al
4
Cu
9
phase. This problem is probably caused by
the submicrometric grain size of Al
4
Cu
9
, which has been
clearly proven (Figures 4 through 6). Furthermore,
Fig. 1—Location of the thin foils cut by FIB at the FSW interface between Cu-a1 and 6082-T6 aluminum alloy.
METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4657
Author's personal copy
concerning the estimation of the volume fraction of each
IMC constituent, the processing of the neutron diffrac-
tion data was ticklish because of the small num ber of
peaks attributed to each compound (Figure 3). Never-
theless, Table I displays the IMCs volume fractions
assessed by this procedure.
Table I shows that the IMC volume fraction differs
according to the position of the tool. When the tool is
offset from the interface, the amount of intermetallics is
almost twofold the WI one. These differences are gene-
rated by the complex interaction of various parameters,
Fig. 2—Fractured intermetallic compounds observed on the surface
of a polished tensile test sample (back scattered electrons [BSE]/
SEM).
Fig. 3—Refined neutron diffraction pattern for each joint (These
colored spectra are clearly differentiated in the electronic version of
the paper).
Fig. 4—Orientation images (EBSD) and Cu pole figures of zone A in Fig. 9 (WD, TD, and Z are the welding, transverse, and plate thickness
directions, respectively). (Color figure online).
4658—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
Author's personal copy
that is to say, the heat input, the temperature field, and
the ease of plastic flow, whic h governs the distribution
of both the amplitude and rate of plastic strains, and
therefore, the local mixing between the components. As
explained in the introduction, all of these variables are
indeed expected to govern the IMC nature as well as
their kinetics of form ation. The heat generated by
metalworking will increase with the resistance of the
base material to deform. In the present case, the less
malleable component of the bead is Cu because of its
low homologous temperature and its higher flow stress
at welding temperature.
[21]
In addition, the volume
fraction of Cu in the stir zone directly depends on the
tool location with respect to the faying interface.
Therefore, this amount of Cu as well as the ensuing
heat input increases in the order W Al, WI, and WCu. In
case of dissimilar Al-Cu FSW, an augmentation of peak
temperature by some decades has indeed been observed
on the side of the Cu stronger alloy.
[51]
In accordance
with this argumentation and, therefore, with the kinetics
of thermal diffusion of Al and Cu, the WCu sample
contains the highest volume fraction of IMCs. This
temperature effect is, however, in contradiction with the
higher amount of IMCs in WAl than in WI. The Al
4
Cu
9
,
and therefore the Cu-rich chemical composition of the
predominant IMC found in WAl, is further inconsistent
with both the primitive availability in solute and the far
higher thermal mobility of Cu in Al than of Al in
Cu.
[52,53]
In fact, at the 773 K (500 °C) expected peak
temperature the coefficient of interdiffusion actually
ranges from 5.8 to 1.5 9 10
–10
cm
2
.s
–1
in presence of
3.05 at. pct of Cu against only 1.7 9 10
–12
cm
2
.s
–1
in
Cu-15 at. pct Al to Cu-20.7 at. pct Al alloys.
[52]
In
addition, and again close to the actual peak tempera-
ture, the impurity diffusion coefficient of Cu in Al is four
orders of magnitude higher than that of Al in Cu.
[53]
Moreover, the intrinsic diffusivity of Cu is again
somewhat higher than that of Al in a thin film of
Al
4
Cu
9
.
[54]
The latter proof is, however, more or less
suitable according to several parameters such as the
effects of the IMC layer thickne ss and of the elaboration
process on the density of lattice defects within the IMC
layer, and/or a possible change of the diffusion mech-
anism over the range of temperature extrapolation and/
or the influence of the nature of the adjoining phases on
the overall kinetics of atom displacement.
The complementary role of mechanical mixing must
be taken into account in order to explain the WAl
results. It is indeed relevant to note that the formation of
an intermetallic compound relies on the apparition of
local mixtures of atoms with a suit able chemical
composition.
[55,56]
In addition, a sharp concentration
Fig. 5—Orientation images (EBSD) and Cu and Al
4
Cu
9
pole figures of zone B in Fig. 9 (WD, TD, and Z are defined in Fig. 4 caption). (Color
figure online).
METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4659
Author's personal copy
gradient inhibits the intermetallics nucleation.
[57]
Con-
cerning the W Al sample and compared with its two
other locations, the tool has to displace the lowest
volume fraction of the Cu more resistant base material.
The material flow is accordingly easier which by
enhancing the mechanical stir of both materials is likely
to generate more IMCs than in the WI joint in spite of a
somewhat lower temperature. It should be noticed that
this latt er explanation agrees well with the presence of
an incipient onion ring structure in the W Al joint whi ch
contrasts with the sharpness of the WI interface.
[47]
In
accordance with the occurrence of a ‘‘forced’’ plastic
deformation, the WCu sample further exhibi ts the most
well-defined onion ring structure encountered in this
study. It can be inferred from all of these results that by
comparison with the cases of a predominant volume
fraction of one component, mechanical mixing as well as
material flow are significantly hampered during the
metalworking of bulk bi-materials made of equivalent
volume fractions of ductile components with distinct
malleability. It is relevant to note that this remark agrees
with the salient decrease of the fineness of mixing of the
components with a diminution of the tool offset in the
case of steel/Al fricti on-stirred welds.
[58]
Besides, the Al
4
Cu
9
most frequent chemical compo-
sition of the reaction product for any tool offset is going
to be discussed in the following.
B. Local Analysis of the WCu Weld
The study of the microstructure of the welds at the
finer scales of SEM/EBSD and TEM is only reported for
the case of the WCu sample as it contains the highest
content in intermetallic compounds (Table I). This
section essentially deals with the IMCs nature according
to their location and kinetics of formation. It finally
establishes their accelerated formation by FSW.
As mentioned in Section II, the metallurgical and
chemical features of the bonding between the 6082 alloy
and Cu-a1 were investigated in zones 1 and 2 (Figure 1),
and therefore, at two depths from the top surface of the
weld. The comparison of both areas put into evidence
less well-defined subgrains boundaries within Cu in zone
2 than in zone 1. This observation arises from the
difference of thermo-mechanical history of each zone
according to its location within the weld. According to
their magnitude, the misorientations measured by EBSD
have for instance proved that either dynamic recrystal-
lization or dynamic recovery has occurred from place to
place (Figures 4 and 5). This result has further been
checked by the comparison between the pole figures of
the base metals and of the weld.
[59]
In addition, both the
substructure of the dislocations boundaries and the
invariable contrast of the dislocations cells indicate that
the two materials ha ve experienced dynamic recovery in
zone 2 (Figure 7).
Due to the difference in malleability and, therefore, of
plastic flow pattern of both parent materials, large
debris of copper have penetrated inside the aluminum
part of the weld.
[47]
Such a feature may give rise to the
formation of IMCs particles in their place (shown by the
white arrows in Figure 8). These large plate-shaped
precipitates, with a length of approximately 500 nm and
a width of approximately 200 nm, act as pinning centers
for the Al alloy grain boundaries. In accordance with
the global availability of atomic species, the electron
diffraction patterns obtained with the precession mode
are consistent with their Al
2
Cu nature. In other areas,
EBSD further shows that small c
1
-Al
4
Cu
9
crystals were
formed in close contact with some other massive debris
of Cu (as shown by bold arrow in Figure 5).
1. Interpenetration bands
In the Cu side, stripes of primitive Al have penetrated
into pure Cu over distances up to a few micrometers.
These bands are marked by arrows on the left-hand side
in Figure 6. They are made of nanograins with a size
between 50 and 100 nm. Electron diffraction an alyses
prove that their P
43m group space and their crystal
lattice agree with the c
1
-Al
4
Cu
9
ones. These results are
consistent with the issues of the EBSD/SEM analyses of
similar zones (Figures 4 and 5), which are typical of the
joint in the Cu side and close to the crown (Figure 9). As
depicted by the pole figures in Figure 5, this c
1
-Al
4
Cu
9
phase present s a quasi-random texture that contrasts
with the dominant [111] fiber texture due to the shear
Fig. 6—Overall microstructure of the joint interface in zone 1
(TEM). Red lines (dark in a gray micrograph) delineate the bound-
aries of the IMC layer.
Table I. Volume Fractions of IMCs within the Welds*
Phases (vol pct) WAl WI WCu
Al
4
Cu
9
2.6 2.2 3.6
Al
2
Cu 1.4 0.2 1.7
*The precision of these measurements is at least of 0.1 pct.
4660—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
Author's personal copy
Fig. 7—TEM micrographs on either side and in close proximity of the interfacial IMC layer in zone 2: (a) Cu side and (b) 6082 alloy side.
Fig. 8—Precipitation state in zone 1 (TEM): particles of Al
2
Cu (white arrows) and of Al(Fe,Mn,Cu)Si (black arrows).
METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4661
Author's personal copy
deformation of Cu. Despite the primitive Al-rich chem-
ical composition in the shear bands as well as the faster
kinetics of diffusion of Cu in Al than the reverse,
[52]
and
the comparison of the thermodynamical driving forces
for precipitation of the various Al
x
Cu
y
equilibrium
IMCs (see the predictions of the effective heat of
formation model in Section III–B–3), Al
4
Cu
9
forms
preferentially to the Al
2
Cu compound in these bands. In
this regard, it seems worthy to note that this preferential
reaction is assisted by both the relaxation of local
residual stresses and the IMCs solubility range. The
formation of c
1
-Al
4
Cu
9
indeed generates a volume
contraction of approximately 23 pct against 10 pct for
the h-Al
2
Cu case. It is, therefore, more suitable in order
to reduce the magnitude of the residual compressive
stresses in the Al bands, which have deformed more
easily than their surrounding Cu slugs. Besides, it has
also been stated that the kineti cs of growth of an
intermetallic compound increase with its composition
range and that this effect is as efficient as that of the
coefficient of diffusion within this IMC.
[60]
According to
the Al/Cu binary equilibrium phase diagram,
[61]
such an
effect can, therefore, explain the higher kinetics of
growth of c
1
-Al
4
Cu
9
compared with the h-Al
2
Cu one.
EDX analyses have furt her established that the
Al
4
Cu
9
intermetallic compounds are rich in oxygen
(Figure 10 ) and sometimes in silicon and magnesium.
The two latter elements very likely originate from the
dissolution during friction-stir welding of the Mg
2
Si
precipitates originally contained in the 6082-T6 alumi-
num alloy .
[6266]
2. Interface
Concerning the surface of contact between both
metals, Figure 6 shows that it can be 3- to 4 lm thick
and composed of dislocation-free grains. A gradient of
chemical compo sition is in particular noted across this
interface. Accor ding to EDX profiles (not supplied in
this article), which are displaying two plateaus, the
interface comprises two kinds of Al
x
Cu
y
compounds: an
Al-rich and a Cu-rich intermetallic layer located on the
Al alloy and on the Cu side, respectively. Both inter-
metallic layers display a thin sublayer of equiaxed
nanograins with an average diame ter of 300 nm at the
faying surface. This central sublayer is prolonged by
columnar grains that are perpendicular to the interface.
In apparent contradiction with the predominant Al
4
Cu
9
volume fraction assessed by the neutron diffraction
analyses of the bulk samples (Table I), Figure 6 depicts
IMC basaltic grains that are more developed on the Al
side. They are approximately 1 lm long by 0.5 lm wide
against 0.5 lm in length and 0.25 lm in width on the Cu
side. This discr epancy between the results of neutron
diffraction and TEM analyses is expected to arise from
both the miscellaneous nature of the thermomechanical
history in FSW welds and the large difference of volume
Fig. 9—Optical micrograph of the cross section of the WCu joint showing the limits of the nugget (yellow or white lines in the electronic version
or in a gray picture, respectively).
Fig. 10—Profiles of EDX analysis along the direction of the red arrow marked in the micrograph. This line, which appears in black in a grey
micrograph, crosses two interpenetration bands (indicated by white arrows) in the Cu part.
4662—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
Author's personal copy
investigated with each technique. EDX spectroscopy
combined with precessed electron diffraction analyses
proves the presence of Al
4
Cu
9
and of Al
2
Cu on the Cu
and the 6082 side, respectively (Figure 11). Nevertheless,
more accurate convergent beam electron diffraction
(CBED) analyses showed that Al
2
Cu possesses an
unusual symmetry (Figure 12). The thermodynamically
stable h-Al
2
Cu phase has the tetragonal structure
(I4/mcm, No. 140). Theoretically, and as shown in
Figure 12(a), the highest ‘‘ideal’’ symmetry (i.e., the
symmetry taking into account the position and the
diffracted intensity of each reflection on a zone axis
pattern [ZAP]) of this phase should be (4 mm) on the
[001] zero-order Laue zone (ZOLZ) ZAP obtained by
CBED. By way of contrast, the experimental CBED
pattern, displayed in Figure 12(b), only indicates the (m)
symmetry. This decrease of symmetry discloses that the
formed phase does not correspond to usual h-Al
2
Cu.
Further detai ls on this new phase will be addressed later
in another article. This compound of unusual crystal-
lography will be called ‘‘modified Al
2
Cu.’’ Moreover, it
is also worthy to note that rather large Al(Fe,Mn,Cu)Si
precipitates with both an equiaxed morphology and a
mean diameter of approx imately 20 nm are observed
inside the columnar grains of modified Al
2
Cu (Fig-
ure 13). This feature proves that these columnar grains
have formed within the Al alloy side as Al(Fe,Mn,Cu)Si
precipitates are originating from the 6082-T6 alloy.
[63]
This observation is the most veracious as it has often
been found that precipitates of this kind are not
modified during the friction-stir welding of 6082-T6
alloys.
[6266]
Not any microstructural evidence of an
effect of these particles on the nucleation and growth of
modified Al
2
Cu has further been noticed.
Besides, the detection of very scarce and micrometer-
sized alumina particles at the Al
2
Cu/6082 Al interface
(Figure 6) very likely originates from the fracture during
welding of an oxide layer present at the abutting
interface of the Al plate before joining. This fragment
was stuck into the surface of a slug of Al, which was
machined during friction-stir welding. Such a fracture of
an Al
2
O
3
layer entailed by a marked plastic deformation
of its aluminum substrate has already been claimed to
occur in other FSW studies.
[67]
In addition, and as
illustrated by Figure 6, these large particles of alumina
clearly hamper the growth of the Al
2
Cu layer.
Figure 6 shows that, except the presence of the deep
bands of penetration of Al on the Cu side, the interfaces
are rather smooth between the IMCs layers and base
materials. Such a feature suggests that the formation of
these layers mainly occurs in the wake of the traveling
tool.
To conclude on the current techniques of investiga-
tion of the microstructure, it seems worthy to emphasize
that the TEM experiments have confirmed the neutron
diffraction results while supplying further information
on the IMCs nature and location in the weld. However,
concerning the measurement of the IMCs volume
fraction, it is worthy to emphasize that due to its
too-fine scale of analysis, TEM generally leads to
nonstatistical results, which may be misleading in
Fig. 11—Experimental precessed electron diffraction patterns of the
Al
2
Cu and Al
4
Cu
9
phases.
Fig. 12—(a) Theoretical [001] ZOLZ CBED pattern of the Al
2
Cu
phase showing the (4 mm) symmetry. (b) (m) symmetry of the exper-
imental [001] ZOLZ CBED pattern recorded in the Al-rich side. The
contrasts visible inside the diffracted disks result from the thickness
variations, but they cannot affect the symmetry.
Fig. 13—Al(Fe,Mn,Cu)Si precipitates (marked by white arrows) into
Al
2
Cu IMCs (details of zone a in Fig. 6) (TEM).
METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4663
Author's personal copy
samples with a heterogeneous microstructure, as it is the
case for FSW joints.
3. Nature of the intermetallic compounds
The formation of both Al
4
Cu
9
and modified Al
2
Cu
during friction-stir welding implies that welding was
reactive. The observed stoichiometries are consistent
with previous observations where the intermetallic
phases at the interface were identified either by X-ray
diffraction in an Al-Cu friction stirred weld
[5,45,51]
or by
energy dispersive X-ray analysis (EDX/TEM) in an
Al-Cu joint made by friction-stir diffusion.
[68]
Such identifications of only two compounds may seem
rather curious as the formation of six Al
x
Cu
y
equilib-
rium IMCs, i.e., h-Al
2
Cu, g
2
-AlCu, f
2
-Al
3
Cu
4
, d, c
1
-
Al
4
Cu
9
, and a
2,
[61]
and of a metastable one: Al
3
Cu
2,
[69]
is
a priori possible with a close to 773 K (500 °C) pea k
temperature.
More strictly speaking , the modified Al
2
Cu, identified
for the first time in this study, has not been said to occur
in the Al-Cu equilibrium phase diagram. So, due to this
lack of thermodynamic data and at a first approxim a-
tion, we are going to merge it with the usual h-Al
2
Cu
phase. The formation of the Al
2
Cu phase on the Al side
of the weld is favored by both local chemical composi-
tion and thermodynamics. The very limited solubility of
Cu in Al
[61]
explains the formation of the Al
2
Cu phase,
which is also the Al-richest phase among the Al
x
Cu
y
ones. This formation also agrees wi th the prediction of
the effective heat of formation model according to which
the first intermetallic phase to nucleate is the one with the
most negative effective heat of formation at the compo-
sition of the lowest temperature of the liquidus.
[70]
In
addition, Al
2
Cu seems to be the first phase to nucleate at
an Al-Cu interface by chemical interdiffusion.
[71]
The
preferential growth of Al
2
Cu displayed in Figure 6 is
again consistent wi th the far higher diffusion rate of Cu
in Al than of Al in Cu.
[72]
As a result, the growth rate of
Al
2
Cu is predominantly explained by the diffusion
kinetics all the more as it presents a very narrow
composition range. However, the exactness of this piece
of reasoning dealing with diffusion remains questionable
because of both the very likely difference of interface
quality in various works and the rather unknown
diffusion kinetics within intermetallics.
Due to the strain and stress conditions during the
process, the existence of transitory IMCs, among which
perhaps the modified Al
2
Cu seems possible. The forma-
tion of out of equilibrium solid solutions pleads in favor
of this suggestion.
[73]
Concerning the nature of the IMC formed on the Cu
side, it is relevant to note that it is not a
2
, i.e., the Cu-
richest compound amon g the Al
x
Cu
y
ones as should have
been expected from the chemical composition of the base
metal. In fact, the IMC formed on the Cu side is the c
1
-
Al
4
Cu
9
phase, which occupies the second rank among
the Cu-rich compounds that can precipitate. Its forma-
tion occurs in spite of the predictions of the effective heat
of formation model according to which g
2
-AlCu phase
should be the second compound to precipitate.
[70]
However, g
2
-AlCu as well as the equilibrium f
2
-Al
3
Cu
4
phases, which actually form during isothermal
treatments over the same temperature range,
[11,74]
were
not observed at the faying interface. Compared with the
Al
4
Cu
9
case, the formation of both g
2
-AlCu and f
2
-
Al
3
Cu
4
is expected to be hampered by their higher Al
content. Always in comparison with Al
4
Cu
9
, the kinetics
of growth of a
2
, g
2
, and f
2
are further assumed to be
impaired by a limited composition range.
[60]
It must,
however, be noted that in another Al-Cu friction-stirred
weld, the g
2
-AlCu phase was found as isolated particles
near the bottom of the stirred zone.
[5]
This clue suggests
that the nature of the IMCs formed by reactive solid-
state FSW is governed by both the local short thermal
cycle (a few seconds) and the local chemical composition.
These two main causes may also be assisted by the severe
deformation amplitude and local internal stresses.
[75]
4. Kinetics of formation of the intermetallic compounds
The kinetics of intermetallic formation seems to be
exacerbated by friction-stir welding. Actually, with
annealing conditions equivalent to the current FSW
thermal history, i.e., approximately 673 K (400 °C) for
10 seconds, a negligible thickness of intermetallic forms
in cold-rolled bonded Al/Cu bimetallic plate.
[11]
Such an
enhancement of kinetics due to FSW can also be inferred
from the comparison with diffusion data obtained with
defect-free interfaces according to which the diffusion
length of Cu in Al is an order of magnitude smaller than
the measured intermet allic thickness. The current asser-
tion is again reinforced by the slower kinetics of diffusion
in Al
x
Cu
y
compounds.
[54]
Such an acceleration of the
kinetics of diffusion is the most significant as the
interface must not be de fect free in FSW-processed
samples. The increase of kinetics of IMCs formation by
FSW arises from the material flow that brings about
mechanical mixing, short circuit diffusion, and a high
supersaturation of vacancies.
[3,4,2732,35,39]
In support of
this settlement, the efficiency of defect-induced atomic
diffusion on the kinetics of IMC formation during Al/
steel friction-stir spot welding (FSSW) has been recently
established from the comparison with results obtained
with solid-state diffusion couples.
[76]
The latter deduction
is the most relevant as, due to a significant gap between
the tip of the probe and the faying interface, the FSSW
plastic flow has occurred only within the upper plate of
aluminum and without any deformation of the joint
interface.
IV. CONCLUSIONS
The main conclusions of the present paper are
summarized as follows:
1. Al
2
Cu and c
1
-Al
4
Cu
9
phases are formed in dissimi-
lar Al 6082-T6/Cu friction-stir weld under the used
conditions. Their formation is essentially governed
by both the thermomechanical history and the local
mixing of the chemical species.
2. The Al
2
Cu compound is termed ‘‘mod ified’’ as it
possesses an unusual crystal symmetry. This,
heretofore, unknown symmetry suggests its out of
equilibrium conditions of formation.
4664—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
Author's personal copy
3. The volume fraction of both c
1
-Al
4
Cu
9
and modi-
fied Al
2
Cu depends on the tool offset because of
the predominant effect of the material flow pattern.
4. The rather thick interface of the joint is composed
of equiaxed and columnar grains of both kinds of
intermetallics.
ACKNOWLEDGMENTS
Many thanks are due to the IS French Welding
Institute for providing the samples, to G. Andre
´
(LLB,
Saclay) for neutron diffraction experiments, and to
Dr. D. Troadec (IEMN, Villeneuve d’Ascq) for pre-
paring FIB specimens. The TEM facility in Lille
(France) is supported by the Conseil Regional du
Nord-Pas de Calais and the European Regional Devel-
opment Fund (ERDF).
REFERENCES
1. P.L. Threadgill, A.J. Leonard, H.R. Shercliff, and P.J. Withers:
Int. Mater. Rev., 2009, vol. 54, pp. 49–93.
2. M. Ericsson and R. Sandstro
¨
m: Int. J. Fatigue, 2003, vol. 25,
pp. 1379–87.
3. R.S. Mishra and Z.Y. Ma: Mater. Sci. Eng. R, 2005, vol. 50, pp. 1–
78.
4. R. Nandan, T. DebRoy, and H.K.D.H. Bhadeshia: Progr. Mater.
Sci., 2008, vol. 53, pp. 980–1023.
5. P. Xue, B.L. Xiao, D.R. Ni, and Z.Y. Ma: Mater. Sci. Eng. A,
2010, vol. 527, pp. 5723–27.
6. T. Laurida, V. Vuorinen, and J.K. Kivilahti: Mater. Sci. Eng. R,
2005, vol. 49, pp. 1–60.
7. M. Braunovic and M. Alexandrov: IEEE Trans. Compon. Packag.
Manufact. Technol., Part A, 1997, vol. 17, pp. 78–85.
8. H.J. Park, S. Rhee, M.J. Kang, and D.C. Kim: Mater. T. JIM,
2009, vol. 50, pp. 2314–17.
9. M. Abbasi, A.K. Taheri, and M.T. Salehi: J. Alloy. Compd., 2001,
vol. 319, pp. 233–41.
10. Y.J. Su, X.-H. Liu, H.-Y. Huang, X.-F. Liu, and H.-X. Xie:
Metall. Mater. Trans. A, 2011, vol. 42A, pp. 4088–99.
11. C.-Y. Chen, H.-L. Chen, and W.-S. Hwang: Mater. Trans. JIM,
2006, vol. 47, pp. 1232–39.
12. M. Aonuma and K. Nakata: Mater. Sci. Eng. B, 2010, vol. 173,
pp. 135–38.
13. A. Hirose, H. Imaeda, M. Kondo, and K.F. Kobayashi: Mater.
Sci. Forum, 2007, vols. 539–543, pp. 3888–93.
14. T. Tanaka, T. Morishige, and T. Hirata: Scripta Mater., 2009,
vol. 61, pp. 756–59.
15. M. Aonuma and K. Nakata: Mater. Sci. Eng. B, 2010, vol. 173,
pp. 135–38.
16. K. Savolainen, J. Mononen, T. Saukkonen, and H. Ha
¨
nninen:
Proc. Int. Symp. FSW, 2006, Paper 79.
17. K. Ueda, T. Ogura, S. Nishiuchi, K. Miyamoto, T. Nanbu, and A.
Hirose: Mater. T. JIM, 2011, vol. 52, pp. 967–73.
18. D.-H. Choi, B.-W. Ahn, C.-Y. Lee, Y.-M. Yeon, K. Song, and S.-
B. Jung: Intermetallics, 2011, vol. 19, pp. 125–30.
19. T. Ogura, K. Ueda, Y. Saito, and A. Hirose: Mater. Trans. JIM,
2011, vol. 52, pp. 979–84.
20. J. Wilden, J.P. Bergmann, and S. Jahn: Adv. Eng. Mater., 2006,
vol. 8, pp. 212–18.
21. Smithells Metal Reference Book, 7th ed., E.A. Brandes and G.B.
Brook, eds., Oxford, U.K., 1992.
22. W. Zhou, L. Liu, B. Li, Q. Song, and P. Wu: J. Electron. Mater.,
2009, vol. 38 (2), pp. 356–64.
23. A.I. Zaitsev, N.E. Zaitseva, R.Y. Shinko, N.A. Arutyunyan, S.F.
Dunaev, V.S. Kraposhin, and H.T. Lam: J. Phys. Cond. Matter.,
2008, vol. 20, p. 114121.
24. M. Kowalski and P.J. Spencer: J. Phase Equil., 1993, vol. 14 (4),
pp. 432–38.
25. M. Watanabe, K. Feng, Y. Nakamura, and S. Kumai: Mater.
Trans. JIM
, 2011, vol. 52, pp. 953–59.
26. S. Takeshi, O. Masafumi, E. Seiichi, and M. Kazuya: Q. J. Jpn.
Weld. Soc., 2000, vol. 18, pp. 365–72.
27. A. Arora, Z. Zhang, A. De, and T. DebRoy: Scripta Mater., 2009,
vol. 61, pp. 863–66.
28. P. Heurtier, C. Desrayaud, and F. Montheillet: Mater. Sci. Forum,
2002, vols. 396–402, pp. 1537–42.
29. H. Schmidt and J. Hattel: Modell. Simul. Mater. Sci. Eng., 2005,
vol. 13, pp. 77–93.
30. G. Buffa, J. Hua, R. Shivpuri, and L. Fratini: Mater. Sci. Eng.A.,
2006, vol. 419, pp. 381–88.
31. G. Buffa, J. Hua, R. Shivpuri, and L. Fratini: Mater. Sci. Eng. A. ,
2006, vol. 419, pp. 389–96.
32. K. Masaki, Y.S. Sato, M. Maeda, and H. Kokawa: Scripta
Mater., 2008, vol. 58, pp. 355–60.
33. M. Militzer, W.P. Sun, and J.J. Jonas: Acta Metall. Mater., 1994,
vol. 42, pp. 133–41.
34. I.E. Gunduz, T. Ando, E. Shattuck, P.Y. Wong, and C.C.
Doumanitis: Scripta Mater., 2005, vol. 52, pp. 939–43.
35. K.V. Jata and S.L. Semiatin: Scripta Mater., 2000, vol. 43,
pp. 743–49.
36. Ø. Frigaard, Ø. Grong, and O.T. Midling: Metall. Mater. Trans.
A, 2001, vol. 32A, pp. 1189–1200.
37. A. Gerlich, G. Avramovic-Cingara, and T.H. North: Metall.
Mater. Trans. A, 2006, vol. 37A, pp. 2773–86.
38. C.I. Chang, C.J. Lee, and J.C. Huang: Scripta Mater., 2004,
vol. 51, pp. 509–14.
39. A. Askari, S. Silling, B. London, and M. Mahoney: Friction Stir
Welding and Processing, TMS, Warrendale, PA, 2001, pp. 43–54.
40. R. Qiu, S. Satonaka, and C. Iwamoto: Mater. Des., 2009, vol. 30,
pp. 3686–89.
41. R. Ayer, H.W. Jin, R.R. Mueller, S. Ling, and S. Ford: Scripta
Mater., 2005, vol. 53, pp. 1383–87.
42. W.B. Lee, M. Schmuecker, U.A. Mercardo, G. Biallas, and S.B.
Jung: Scripta Mater., 2006, vol. 55, pp. 355–58.
43. K.S. Bang, K.J. Lee, H.S. Bang, and H.S. Bang: Mater. Trans.,
2011, vol. 52 (5), pp. 974–78.
44. A. Kostka, R.S. Coelho, J. dos Santos, and A.R. Pyzalla: Scripta
Mater., 2009, vol. 60, pp. 953–56.
45. P. Xue, D.R. Ni, D. Wang, B.L. Xiao, and Z.Y. Ma: Mater. Sci.
Eng. A, 2011, vol. 528, pp. 4683–89.
46. FullProf: WinPLOTR Software, A Graphic Tool For Powder
Diffraction, Version 2006.
47. M.N. Avettand-Fe
`
noe
¨
l, R. Taillard, C. Herbelot, and A. Imad:
Mater. Sci. Forum, 2010, vols. 638–642, pp. 1209–14.
48. M.-N. Avettand-Fe
`
noe
¨
l, R. Taillard, and G. Ji: Mater. Sci. Forum,
2012, vols. 706–709, pp. 959–64.
49. T.R. McNelley, S. Swaminathan, and J.Q. Su: Scripta Mater.
,
2008, vol. 58, pp. 349–54.
50. S.R. Ren, Z.Y. Ma, and L.Q. Chen: Scripta. Mater., 2007, vol. 56,
pp. 69–72.
51. H.J. Liu, J.J. Shen, L. Zhou, Y.Q. Zha, C. Liu, and L.Y. Kuang:
Sci. Technol. Weld. Joi., 2011, vol. 16 (1), pp. 92–98.
52. W. Jost: Diffusion in Solids, Liquids, Gases, E.M. Loebl, ed.,
Academic Press, New York, NY, 1960.
53. D. Liu, L. Zhang, Y. Du, H. Xu, S. Liu, and L. Liu: CALPHAD,
2009, vol. 33, pp. 761–68.
54. H.T.G. Hentzell and K.N. Tu: J. Appl. Phys., 1983, vol. 54,
pp. 6929–37.
55. R. Taillard, C.E. Bruzek, and E. Florianova: Proc. Int. Conf. Solid
to Solid Phase Transformations, Nemacolin Woodlands, PA, TMS,
Warrendale, PA, 1994, pp. 1183–88.
56. D.Y. Ying and D.L. Zhang: J. Alloy. Compd., 2000, vol. 311,
pp. 275–82.
57. F. Hodaj and P.J. Desre
´
: Acta Mater., 1996, vol. 11, pp. 4485–90.
58. T. Watanabe, H. Takayama, and A. Yanagisawa: J. Mater. Pro-
cess. Technol., 2006, vol. 178, pp. 342–49.
59. D. Goran, G. Ji, M.N. Avettand-Fe
`
noe
¨
l, and R. Taillard: Mater.
Sci. Forum, 2012, vols. 702-703, pp. 574–77.
60. M. Kajihara: Acta Mater., 2004, vol. 52, pp. 1193–1200.
61. N. Ponweiser, C.L. Lengauer, and K.W. Richter: Intermetallics,
2011, vol. 19, pp. 1737–46.
METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—4665
Author's personal copy
62. R. Taillard, M.N. Avettand-Fe
`
noe
¨
l, C. Herbelot, and A. Imad:
Proc. Journe
´
es Annuelles de la Socie
´
te
´
Franc¸ aise de Me
´
tallurgie et
de Mate
´
riaux, Nancy, France, 2011.
63. A. Simar, Y. Bre
´
chet, B. de Meester, A. Denquin, C. Gallais, and
T. Pardoen: Progr. Mater. Sci., 2012, vol. 57 (1), pp. 95–183.
64. L.E. Svensson, L. Karlsson, H. Larsson, B. Karlsson, M. Fazzini,
and J. Karlsson: Sci. Tech. Weld. Joi., 2000, vol. 5, pp. 285–96.
65. Y.S. Sato, M. Urata, and H. Kokawa: Metall. Mater. Trans. A,
2002, vol. 33A, pp. 625–35.
66. G. Mro
´
wka-Nowotnik and J. Sieniawski: J. Mater. Process.
Technol., 2005, vols. 162–163, pp. 367–72.
67. T. Watanabe, A. Yoneda, A. Yanagisawa, S. Konuma, and O.
Ohashi: Yosetsu Gakkai Ronbunsu, 1999, vol. 17, pp. 223–33.
68. M. Girard, B. Huneau, C. Genevois, X. Sauvage, and G. Raci-
neux: Sci. Tech. Weld. Joi., 2010, vol. 15, pp. 661–65.
69. P. Ramachandrarao and M. Laridjani: J. Mater. Sci., 1974, vol. 9,
pp. 434–37.
70. R. Pretorius, T.K. Marais, and C.C. Theron: Mater. Sci. Eng. R,
1993, vol. 10, pp. 1–83.
71. D. Moreno, J. Garrett, and J.D. Embury: Intermetallics, 1999,
vol. 7, pp. 1001–09.
72. X.K. Peng, R. Wuhrer, G. Heness, and W.Y. Yeung: J. Mater.
Sci., 1999, vol. 34, pp. 2029–38.
73. L.E. Murr: J. Mater. Eng. Perf., 2010.
74. Y. Funamizu and K. Watanabe: Trans. Jpn. Inst. Met., 1971,
vol. 12, pp. 147–52.
75. S. Guyot: Ph.D. Dissertation, University of Lille 1, France,
2005.
76. M. Watanabe, K. Feng, Y. Nakamura, and S. Kumai: Mater.
Trans., 2011, vol. 52, pp. 953–59.
4666—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A
Author's personal copy
... In a number of FSW butt tests involving Cu and Al dissimilar materials, the researchers placed the Cu on the AS and obtained a joint with a good appearance and performance, such as Jiahu Ouyang et al. [5], Akbri et al. [10], Rathesh et al. [14], Genevois et al. [15], Li Xiawei et al. [16], Ke Liming et al. [20], Liu et al. [22], Esmaeili et al. [42], Felix. Xavier et al. [54], Akinlabi et al. [58], Avettand et al. [59] and Galvão et al. [60,76]. ...
... In a number of FSW butt tests involving Cu and Al dissimilar materials, the researchers placed the Cu on the AS and obtained a joint with a good appearance and performance, such as Jiahu Ouyang et al. [5], Akbri et al. [10], Rathesh et al. [14], Genevois et al. [15], Li Xiawei et al. [16], Ke Liming et al. [20], Liu et al. [22], Esmaeili et al. [42], Felix. Xavier et al. [54], Akinlabi et al. [58], Avettand et al. [59] and Galva o et al. [60,76]. To conclude, the placement of Cu and Al is a problem that must be studied when FSW is carried out between the dissimilar materials Cu and Al. ...
... For example, in the study of al-Roubaiy et al. [10], only one kind of IMC, Al 2 Cu, was detected. Two IMCs, Al 2 Cu and Al 4 Cu 9 , were detected by Avettandfenoel et al. [59], Saeid et al. [6], Genevois et al. [15], Xue et al. [3,26], Akinlabi et al. [58], Ahmed Elrefaey et al. [64], Beygi et al. [65] and Fotoohi et al. [9]. Three IMCs, AlCu, Al 2 Cu and Al 4 Cu 9 , were found in the research of Jiahu Ouyang et al. [5], Liu. ...
Article
Full-text available
With the rapid development of various industries, the connection of copper and aluminum is in high demand. However, as a solid-phase connection technology, friction stir welding has a potential application prospect in the connection of copper and aluminum. This paper comprehensively summarizes the most recent 20 years of the literature related to the friction stir welding of copper and aluminum. The application significance of copper and aluminum connectors is introduced, and the research field of the friction stir welding of copper and aluminum is analyzed and explored from the aspects of welding technology, microstructure and mechanical properties, as well as innovations and improvements in the welding process. In view of the research status of this field, the authors put forward their views and prospects for its future, aiming to provide a basis for researchers in this field.
... More recently, Galvao et al. [8] claimed that the IMC formation in dissimilar FSW of aluminium to copper can only be explained by the thermo-mechanically activated solid state diffusion phenomenon. Unlike the approach of Ouyang et al. [7], Galvao et al. [8] proved the absence of solidification structures in both the aluminium and copper rich sides and reported that the resultant high strain rate during the FSW process facilitates the formation of Al2Cu, AlCu and Al4Cu9, an approach that has also been supported in other recent publications [9][10][11]. Further, Xue et al. [12] and Galvao et al. [13] investigated the influence of process parameters on the evolution of IMCs during FSW of aluminium to copper. ...
... Overall, researchers [3][4][5][6][7][8][9][10][11][12][13][14][15] have not definitively demonstrated the negative or positive role played by the IMC particles. Furthermore, there is a lack of understanding towards the evolution of these IMCs during the FSW process of aluminium to copper. ...
Article
A novel approach for predicting the intermetallic compound (IMC) formation during friction stir welding (FSW) of AA6061 to commercially pure copper has been developed, in addition to their effect on mechanical properties. The temperature distribution of the aluminium to copper weld nugget determined by a finite element model, the use of an Al–Cu phase diagram and the elemental concentration of copper and aluminium in the weld nugget have been combined to predict and validate several IMCs present in the different zones of the weldment. The results of performing butt-welding of these dissimilar metals using the FSW process demonstrated that the highest ultimate tensile strength of 194.5 MPa was achieved at 1500 rev min−1 tool rotational speed, 100 mm min−1 traverse speed and a zero-tool offset.
... It had IMC components similar to those of the joint produced under model 1: the IMC near the Cu substrate was θ-CuAl 2 , while it was α-Al + CuAl 2 in the gap of the valley. However, some references have stated that Cu 9 Al 4 would also be produced at the Al-Cu interface [24][25][26], which was not found in this report. This was possibly caused by the higher generated temperature of Cu 9 Al 4 compared to θ-CuAl 2 [24], or its limited generated volume, which was difficult to detect. ...
... However, some references have stated that Cu 9 Al 4 would also be produced at the Al-Cu interface [24][25][26], which was not found in this report. This was possibly caused by the higher generated temperature of Cu 9 Al 4 compared to θ-CuAl 2 [24], or its limited generated volume, which was difficult to detect. The interfacial phase components were also tested by XRD, and the corresponding results are shown in Fig. 9(c). ...
Article
In this research, Al/Cu lapped joints with a thickness of 0.2 mm were joined via nanosecond pulsed laser welding. The influence of laser scanning paths on the weld formation, microstructure and bonding strength was investigated. Sound weld formations were produced when the outer spiral laser scanning path was used. Porosities and cracks were easily formed in joints produced by concentric circle and straight laser scanning paths due to the interval duration between the neighboring welding processes. The interfacial morphology was a wave structure when the laser scanning paths were an outer spiral and concentric circle, and continuous interfacial intermetallic compound (IMC) was formed at the interface. θ–CuAl2 was generated in the vicinity of the Cu substrate, while α-Al + θ–CuAl2 was generated in the outer region and the weld seam. When a straight line laser scanning path was adopted, the wave structure disappeared. Discontinuous IMC and cracks were also produced at the interface. The tensile-shear load results indicated that the outer spiral laser scanning line was the most beneficial for the bonding strength of the joint and could reach up to 198 N. A satisfactory weld formation, the pinning effect due to the wave structure at the interface and continuous interfacial IMC all contributed to its highest bonding strength.
Article
Full-text available
The influence of tin foil and Ni coatings on microstructures, mechanical properties, and the interfacial reaction mechanism was investigated during laser welding/brazing of Al/Cu lap joints. In the presence of a Zn-based filler, tin foil as well as Ni coating strengthened the Al/Cu joints. The tin foil only slightly influenced the joint strength. It considerably improved the spreading/wetting ability of the weld filler; however, it weakened the bonding between the seam and the Al base metal. The Ni coating considerably strengthened the Al/Cu lap joints; the highest tensile strength was 171 MPa, which was higher by 15.5% than that of a joint without any interlayer. Microstructure analysis revealed that composite layers of Ni3Zn14–(τ2 Zn–Ni–Al ternary phase)–(α-Zn solid solution)–Al3Ni formed at the fusion zone (FZ)/Cu interface. Based on the inferences about the microstructures at the interfaces, thermodynamic results were calculated to analyze the interfacial reaction mechanism. The diffusion of Cu was limited by the Ni coating and the mutual attraction between the Al and Ni atoms. The microstructure comprised Zn, Ni, and Al, and they replaced the brittle Cu–Zn intermetallic compounds, successfully strengthening the bonding of the FZ/Cu interface.
Article
The effects of tool-workpiece relative positions on the microstructures and mechanical properties of 1-mm-thick 6061-T6 aluminum to T2 copper friction stir welds are systematically investigated in this research. Compared with the case of placing Cu sheet on advancing side, placing the Al sheet on advancing side can promote the flow of nugget material and expand the volume of the mixing zone, which is more conducive to obtain high-quality dissimilar Al-Cu welds. Based on this, the 1-mm tool offset toward Cu side is found to yield higher hardness and more defects due to the formation of Cu9Al4, leading to poor tensile properties and a brittle fracture feature of the joint. When the tool axis moves 1 mm toward the Al side, the decrease in Cu content in the stir zone not only limits the adhesion of material to the shoulder surface but also prevents the phase transformation from CuAl2 to Cu9Al4 and induces thin lamellar structures in the weld, which improves the fracture patterns and leads to good tensile properties of joint. The present study is expected to clarify the process characteristics of the dissimilar FSW of Al and Cu thin sheets and thus to promote the high-quality application of Al and Cu materials in electrical and electronic technology.
Article
The microstructure of the interfa ce zone is the most important factor to determine the properties of the Cu/Al clad metal sheet joint. The composition distribution and phase constitution at the interface of the double-sided electron beam welding (DSEBW) joint of 6.5 mm thick Cu (1.5 mm)/Al (5 mm) clad metal sheet were revealed by means of technologies of scanning electron microscopy (SEM), electron backscatter diffraction (EBSD), X-ray diffraction (XRD) and transmission electron microscopy (TEM). It was confirmed that there formed a distinct transition zone with a length of about 5.8 mm at the interface, and it consisted of a thin layer of IMC and a large area eutectic structures. Moreover, the thickness distribution of IMC layer in the center and edges of the interface zone was nonuniform, and gradually decreased from the center (7.5 μm) to the edges (1.5 μm and 2.6 μm) by means of SEM and EBSD. Furthermore, the eutectic structure performed the superior laminar, and the grey IMC dispersed in Al-rich dendrites inside the eutectic structure. In addition, the interface zone mainly consisted of two slices, including Cu/IMC interface (Interfaces I and III) and IMC/eutectic structure interface (Interfaces II and IV). For interfaces I and III, this IMC was Al2Cu, and there was no other IMCs between Cu and Al2Cu IMC by means of TEM; For interfaces II and IV, the eutectic structure was composed of Al2Cu and α-Al, while there was also no other IMCs between Al2Cu IMC and (Al2Cu + α-Al) eutectic structure. In short, the microstructure characterizations at the whole interface of the DSEBW joint were Al2Cu IMC and (Al2Cu + α-Al) eutectic structure.
Article
Full-text available
Aluminium and copper are employed in various industrial applications due to their high plasticity, thermal conductivity, electrical conductivity and characteristics. By effectively joining dissimilar aluminium and copper, the unique properties of composite formed by these metals can be adequately addressed. Friction stir welding (FSW), an energy-efficient solid-state welding process is capable of joining dissimilar metals, has enormous potential in the future of various industries. This present work comprehensively summarises all pertinent topics related to aluminium to copper FSW, such as FSW process parameters, microstructural characterisation, mechanical properties, and electrical characteristics of aluminium-copper joints produced by FSW. In addition, the current report also discusses several applications of additives used in dissimilar FSW of Al-Cu and new FSW techniques, which generally aim to enhance Al-Cu joint properties. Moreover, numerical modelling of Al-Cu FSW is discussed profoundly to understand the effects of alterations in different process parameters on temperature gradients and microstructure evolution, which would be time-consuming or prohibitively expensive in practice by physical testing. Additionally, several recommendations for future research are proposed to facilitate the advancement and success of Al-Cu FSW studies.
Article
Friction stir butt welding of C1020 and A1050 were carried out. To avoid the deterioration of strength and workability of the joints, dispersion of copper fragments into aluminum was suppressed. The effect of tool tilt angle on root flaw at the bottom part was investigated. To obtain unmixed joints, the probe was plunged into only aluminum. Under the condition of tool tilt angle of 0°, equiaxed grains were observed at the bottom of A1050 near the bonding interface. It indicates that material flow and heat input could promote the reaction of copper and aluminum at the bottom part. In contrast, under the condition of tool tilt angle of 3°, unbonded area remained at the bonding interface at the bottom. The lack of penetration near the bonding interface was larger than that of tool tilt angle of 0°.
Article
Gradient nanostructured metals attract extensive attention due to their excellent mechanical properties, while they are prone to microcracks during forming and servicing process and eventually develop into a sudden fracture. Actually, experimental studies show that microcracks occurred in both the grain boundary (GB) and the intragrain, but most of the current reports only focus on the GB cracking, very limited work has studied the intragranular cracking. In this paper, the crystal plasticity finite element method (CPFEM) and cohesive zone model (CZM) are combined to study the cracking mechanism of polycrystalline aluminum (Al) with different grain-gradient structures under tensile load, where molecular dynamics (MD) method is used to determine the cohesive parameters of the intragrain and GB. The results not only show the crack initiation and propagation process of polycrystalline Al with different grain-gradient structures, but also reveal the mechanism of intragrain fracture, GB fracture and crack transgranular. The grain-gradient distribution with optimal comprehensive performance is obtained. Moreover, it is also found that the initial microcracks with different positions, numbers and angles have a great influence on the cracking mechanism and effective properties of the whole material. This study provides a solid theoretical basis for improving the quality and operation life of metal parts.
Article
Full-text available
Dissimilar Al-Cu joints are desirable due to the combination of the good electrical and thermal properties of copper with the economic advantages of aluminium. However, different chemical, mechanical and thermal properties turn dissimilar welding into a challenge with traditional fusion welding techniques. Friction stir welding appears to be a good alternative to welding dissimilar materials. Most of the works in Al-Cu FSW have been performed with Al on top. This work investigates the influence of heat input on the shear strength of aluminium-copper lap joints produced by FSW with a copper over aluminium lap configuration. Welds produced with the ω/ν rate of 200 rev.mm-1 or 500 rev.mm-1 resulted in the melting of the base materials and material leaking. Joints with shear tensile strength varying from 76 to 85% of the original resistance of AA6060 T5 were produced when the ω/ν rate was varied from 80 to 110 rev.mm-1. The fracture was governed by tension overload on the aluminium sheet, resulting in little influence of welding defects, such as tunnelling, on the joint strength.
Article
Full-text available
Lap joining of a pure aluminum plate and a low carbon steel plate was performed using friction stir spot welding. The aluminum plate was placed over the steel plate, a rotating welding tool was inserted into the aluminum plate, and the tip of the tool was dwelled above the aluminum/steel interface. Dwell time was controlled in the range of 0 to 120 seconds. The microstructure of the welding interface was examined by optical microscopy and scanning electron microscopy. Chemical composition analysis was carried out by energy dispersive X-ray spectroscopy. Welding was achieved for all dwell times. Refined grains were formed by plastic flow in the aluminum matrix close to the welding interface. Intermetallic compound layer was produced along the welding interface. Precise backscattered electron image observation and energy dispersive X-ray spectroscopy analysis revealed that the intermetallic compound layer consisted of an Al13Fe4 phase layer and an Al 5Fe2 phase layer. The thickness of the layers increased in proportion to the square root of the dwell time. The parabolic coefficient K was 1.30 × 10-14 and 6.06 × 10-13 m 2/s for the Al13Fe4 layer and the Al 5Fe2 layer, respectively.
Article
Full-text available
The phase equilibria and reaction temperatures in the system Al–Cu were re-investigated by a combination of optical microscopy, powder X-ray diffraction (XRD) at ambient and elevated temperature, differential thermal analysis (DTA) and scanning electron microscopy (SEM). A full description of the phase diagram is given. The phase equilibria and invariant reactions in the Cu-poor part of the phase diagram could be confirmed. The Cu-rich part shows some differences in phase equilibria and invariant reactions compared to the known phase diagram. A two phase field was found between the high temperature phase η1 and the low temperature phase η2 thus indicating a first order transition. In the ζ1/ζ2 region of the phase diagram recent findings on the thermal stability could be widely confirmed. Contrary to previous results, the two phase field between δ and γ1 is very narrow. The results of the current work indicate the absence of the high temperature β0 phase as well as the absence of a two phase field between γ1 and γ0 suggesting a higher order transition between γ1 and γ0. The structure of γ0 (I-43m, Cu5Zn8-type) was confirmed by means of high-temperature XRD. Powder XRD was also used to determine the structure of the high temperature phase η1-Al1−δCu. The phase is orthorhombic (space group Cmmm) and the lattice parameters are a = 4.1450(1) Å, b = 12.3004(4) Å and c = 8.720(1) Å; atomic coordinates are given.
Article
The effects of Zn-based alloys coating (Zn, Al-Zn and Al-Mg-Zn) on the bondability of steel/aluminum alloy dissimilar metals joints were evaluated, in order to achieve strength in lower welding current. In the joint with Zn-based alloys insert, the oxide film on the aluminum alloy was sufficiently removed through eutectic reaction of Zn-based alloys and aluminum. In the joint with Zn-coated steel (GI), higher welding current is necessary to discharge the zinc coating and the oxide film from the bonding interface sufficiently. The thinner aluminum plate after welding and the thick reaction layer cause the decrease of cross tensile strength in the joints with no coating steel (SPCC) and Al-Zn-coated steel. Using Al-Mg-Zn-coated steel, higher strength was achieved in a lower welding current. This is because Al-Mg-Zn-coating melted at lower temperature than Zn and Al-Zn-coating, and the removal of the coating material and the oxide film on the aluminum alloy were sufficiently performed in the lower welding current.
Article
Interdiffusion in the aluminum-copper system was investigated in the temperature range of 400° to 535°C with diffusion couples of pure aluminum and copper metals.In the aluminum-copper equilibrium phase diagram there are five intermediate phases in this temperature range, namely γ2, δ, ζ2, η2 and θ. Electron probe micro line analysis of specimens studied indicated that all of these phases were located in the diffusion layer.It appears that the layer growth of each phase is controlled by the process of volume diffusion since the rate of layer growth obeys the parabolic law. From the temperature dependency of the rate constants of layer growth, the activation energies of the individual phases were obtained.The interdiffusion coefficient for each intermediate phase was calculated by the method introduced by Heumann, and the values obtained were consistent with those derived from Kidson’s equation for the δ, ζ2 and η2 phases.Aluminum oxide powder was used for the measurement of the Kirkendall effect. It is clear from this measurement that diffusion in the multilayer system is controlled by the vacancy mechanism and that aluminum diffuses more rapidly than copper.
Article
Friction stir welding (FSW) was performed about AA6061-T6 and Ti-6Al-4V alloy sheets. A unique shaped tool with circular truncated cone of probe was used. Mechanical properties and interfacial microstructure were evaluated using tensile test, hardness test, optical microscopy (OM), scanning electron microscopy (SEM) and scanning transmission electron microscopy (STEM), respectively. Root area of probe in stir zone (SZ) reveals a mixture of finely recrystallized grains of Al and Ti particles pushed away from the base metal by strong stirring of probe. The joint interface of tip area of probe was relatively flat because stirring between aluminum and titanium alloy was not occurred due to the gap of the probe and titanium alloy front. It is considered that the insufficient stirring due to inclined side of the probe was contributed to the decrease of weld strength. After tensile test, fracture surface was analyzed by SEM. In the probe root area, dimples of Al were observed. In the probe tip area, the initial surface of titanium alloy plate was observed. However, in the middle area, similar amount of Ti and Al was detected. As result, it was confirmed that the fracture sequence was very complex and the fracture position was different according to the probe position.
Article
The thermal cycle during friction welding is an important factor that significantly affects on the mechanical properties and microstructures in welded joints. The weld heat input has been normally estimated using the torque during friction welding, however most of industrial friction welding machines are not able to measure the torque which cannot be available to use process control.In this paper, the calculation of heat input rate is proposed for friction welding under a new concept for the dissimilar welded joints. This heat input concept, which is solved governing equations combined with high temperature strength, and thermal conduction is successfully expanded into joints by dissimilar materials with large differences of thermal behaviors. It is revealed that mechanical properties are correlated with the proposed heat input rate to dissimilar joints by pure aluminum and plain low carbon steel. Mechanical properties are evaluated by notch tensile strength. This weld heat input rate is also related with intermetallic phase formation that is confirmed in the transmission electron microscope observations.As a conclusion, the optimum welding condition is obtained with increased the heat input rate.1) By using the simple assumptions made in this study, it is possible to obtain the heat input rate, to a certain extent, to predict the mechanical properties of the friction welded joints to dissimilar materials. It is also clarified this proposed heat input rate predicts formation of intermetallic compounds at the weld interface.2) The calculated heat input rate correlates with the tensile strength of the welds. High joint strength is obtained in the case of welds of higher heat input rate.3) Intermetallic phase does not form when friction time is less than 1 second. This is also confirmed from the results of calculation. Intermetallic phase at weld interface is determined as Fe4A13 using diffraction pattern of TEM observation.
Article
Nanoindentation measurements were successfully applied to the interfacial reaction layers in dissimilar metal joints of 6000 series aluminum alloys containing alloying elements to steel in order to characterize their mechanical properties. The nanoindentation hardness of the reaction layer formed at the aluminum side was lower than that formed at the low carbon steel (SPCE) side of the investigated joints. At the aluminum side, the nanoindentation hardness changed by the addition of alloying elements. The hardness of the resulting Al12Fe3Si intermetallic compound (IMC) (and the same IMC containing Cu) was lower than that of Al3Fe In comparison with the hardness values obtained from bulk Al-Fe binary series IMCs, it is considered that hardness changes of interfacial reaction layers are derived from the crystal structural changes produced by the alloying elements. The result of micro-testing of Al-Fe series IMCs indicates that the modification of the interfacial reaction layer by alloying elements contributes to higher ductility and the improvement of joint strength through crystal structural change.
Conference Paper
The current work focuses on the particular case of dissimilar 6082 Al alloy/pure copper butt-friction stirred joints. It takes advantage of voluntarily non optimized welding conditions in order to test the potential of an original approach of identification of the welding defects by means of a single tensile test. The sequence and mechanism of the fracture events arise from their localization on the fracture surfaces thanks to strain maps obtained by digital image correlation. This technique of flaws identification is proved to be particularly efficient at least with the present highly damaged welds.
Conference Paper
Texture and microstructure of FSW joined Al and Cu sheets were investigated by means of electron backscatter diffraction (EBSD) technique. The analysis has revealed a strong texture evolution on both sides of the weld interface as well as a very complex microstructure. Grains were found to be fully recrystallized on both sides of the weld and with different average diameters at different specific zones of the weld.
Article
Compared to most thermomechanical processing methods, friction stir welding (FSW) is a recent technique which has not yet reached full maturity. Nevertheless, owing to multiple intrinsic advantages, FSW has already replaced conventional welding methods in a variety of industrial applications especially for Al alloys. This provides the impetus for developing a methodology towards optimization, from process to performances, using the most advanced approach available in materials science and thermomechanics. The aim is to obtain a guidance both for process fine tuning and for alloy design. Integrated modeling constitutes a way to accelerate the insertion of the process, especially regarding difficult applications where for instance ductility, fracture toughness, fatigue and/or stress corrosion cracking are key issues. Hence, an integrated modeling framework devoted to the FSW of 6xxx series Al alloys has been established and applied to the 6005A and 6056 alloys. The suite of models involves an in-process temperature evolution model, a microstructure evolution model with an extension to heterogeneous precipitation, a microstructure based strength and strain hardening model, and a micro-mechanics based damage model. The presentation of each model is supplemented by the coverage of relevant recent literature. The “model chain” is assessed towards a wide range of experimental data. The final objective is to present routes for the optimization of the FSW process using both experiments and models. Now, this strategy goes well beyond the case of FSW, illustrating the potential of chain models to support a “material by design approach” from process to performances.