Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 147
WELDING RESEARCH
SUPPLEMENT TO THE WELDING JOURNAL, MAY 2016
Sponsored by the American Welding Society and the Welding Research Council
Fiber Laser Welding of Al­Si­Coated
Press­Hardened Steel
The mixing and formation of the Al­Si coating’s delta­ferrite phase was
the key factor for premature failure of the laser­welded joints
BY D. C. SAHA, E. BIRO, A. P. GERLICH, AND Y. N. ZHOU
ABSTRACT
The effect of an Al­Si coating on the microstructure and mechanical properties of
fiber laser­welded, press­hardened steel is examined in the present work. It is shown
that mixing of the Al­Si coating significantly modified the fusion zone microstructure,
which was found to be a combination of martensite and delta ferrite in both the welded
as­received and hot­stamped conditions. In contrast, a uniform distribution of alpha fer­
rite and martensite was observed when the as­received sheet was welded then hot
stamped. The ferrite phase was found to be rich in Al and Si, which retarded the kinetics
of martensite formation, increasing the ferrite fraction after hot stamping of the welded
coupon. Micro­ and nano­scale hardness measurements on ferrite suggest it is a softer
phase compared to martensite; therefore, inducing strain localization and premature
fracture during tensile tests occurred. The strength of all­welded joints was found to be
dependent on ferrite phase fractions, since this promoted premature fracture, and dete­
riorated joint strength and ductility.
KEYWORDS
• Fiber Laser Welds • Press­Hardened Steel • Coating Mixing • Weld Solidification
• Delta Ferrite
Introduction
In recent years, automotive industries have reduced vehicle weight by replacing conventional low-strength
steels with advanced high-strength
steels (AHSSs), and ultrahigh-strength
steels (UHSSs). The higher specific
strength, thinner gauge of AHSS and
UHSS sheets provide superior mechanical properties compared to the other
high-strength, low-alloy steels (Ref. 1).
By decreasing the thickness of the
sheets, the weight of the vehicles can be
reduced, improving fuel efficiency and
decreasing greenhouse gas emissions to
benefit the environment (Refs. 2, 3).
Among the other AHSS, presshardened steel (PHS), such as the USIBOR® glade, offers high ultimate tensile
strength (1500 MPa) in the hotstamped condition. In the as-quenched
state, PHS exhibits high strength and
low formability, making it difficult to
form complex parts (Ref. 1). Consequently, these steels are intended for a
hot stamping process, i.e., forming at
the austenite temperature and subsequently quenching to form martensite.
In order to prevent surface oxidation
and decarburization, PHS sheets are
usually coated with various coatings,
such as Al-Si, Zn, or Zn-Ni alloys (Ref.
4). Among these coating alloys, Al-Si
offers better corrosion and high-temperature oxidation resistance (Ref. 4).
Automotive components commonly
made with PHS are A-pillars, B-pillars,
door rings, bumpers, roof rails, and
tunnels (Ref. 5), all of which are usually made from laser-welded blanks
(LWBs), also known as tailor-welded
blanks (TWBs). Other property tailoring methods such as partial tempering, partial quenching, and post tempering of the monolithic sheets can be
employed (Ref. 6). Utilization of PHS
materials combined with LWB technology provides weight saving, material cost reduction, improved part performance, and simplifies the manufacturing process (Refs. 7–9). Fiber laser
welding is considered to be preferred
in this application compared to other
forms of laser sources, considering it
offers a small beam diameter, flexible
beam delivery, higher precision and accuracy, less thermal distortion, and
high-plug efficiency (Refs. 10, 11).
Apart from the chemical composition of steels, the presence of a coating
also influences the properties of
LWBs. It has been reported that the
coated steels are prone to more welding defects such as weld concavity
(Ref. 12) and coating mixing into the
weld pool (Refs. 13, 14), compared to
uncoated steels. One of the major issues associated with using coated PHS
is the presence of the Al-Si coating on
the surface, which affects the welding
process. During laser welding, the
coating is mixed into the fusion zone
D. C. SAHA (dcsaha@uwaterloo.ca), A. P. GERLICH, and Y. N. ZHOU are with the Department of Mechanical and Mechatronics Engineering,
University of Waterloo, Waterloo, Ontario, Canada. E. BIRO is with ArcelorMittal Global Research, Hamilton, Ontario, Canada.
MAY 2016 / WELDING JOURNAL 147-s
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 148
WELDING RESEARCH
A
B
C
D
A
B
Fig. 2 — Typical base metal microstructure of the investigated steels.
A — As­received condition; B — hot­stamped condition.
A
B
Fig. 1 — SEM micrographs of the Al­Si coating and line scan­
ning of Al, Si, and Fe. A and C — As­received condition; B and
D — hot­stamped condition.
(FZ), and thereby changes the local
chemistry and equilibrium phases
(Ref. 15). Hence, heterogeneous phases can form and adversely affect the
mechanical properties (Refs. 14, 16).
To date, limited work has been published regarding fiber laser welding of
PHS. Kim et al. (Ref. 13) investigated
the laser welding properties of Al-Sicoated PHS and found an intermetallic
phase at fusion boundary. Other studies (Refs. 17, 18) also reported intermetallic phase formation in the FZ
while welding Al-Si-coated PHS. Studies done on laser welding of Al-Si-coated PHS have claimed that a second
phase formed in either the FZ or the
fusion boundary was an intermetallic
compound, without identifying it;
however, it was shown to be related to
the presence of Al content.
Recently, Kang et al. (Ref. 19) investigated the influence of beam size and
hot stamping parameters (temperature
and time) on the FZ microstructure formation. They found that the fraction of
martensite increases in the FZ when a
larger beam size, higher austenization
temperature, and time are considered.
Industries are now dealing with this issue by employing a dedicated welding
process known as laser ablation (Refs.
17, 18), which uses a short multi-impulse laser to remove the Al-Si coating,
leaving behind only an intermediate
Fe2Al5 and
Fig. 3 — Fiber laser­welded PHS steel without coating on the surface
FeAl3 layer.
showing. A — The full weld profile; B — the fully martensitic FZ.
The formation mechaferred (within 10 s) to flat dies, where
nisms of producing the second phase
quenching was done to ensure fully
in the FZ and their effect on mechanimartensite microstructure (Refs. 1, 21).
cal properties are not well understood.
The cooling rate was measured with a
In this study, the effect of the coating
thermocouple welded to the dies, and
will be examined in terms of the forconfirmed to be about 31°C/s, which is
mation of various phases in the FZ,
identical to the critical cooling rate of
and their impact on the microhardness
martensite formation of PHS as reportand tensile properties. In the present
ed in the literature (Refs. 1, 22).
study, all of the possible processing
Welds were made using an IPG Phocombinations of PHS joints will be
tonics ytterbium fiber laser system
considered, such as as-received welded
(model: YLS-6000-S2); the details of the
(ARW), as-received welded then hot
system can be found in previous studies
stamped (ARWHS), and hot stamped
(Ref. 23). The welds were inspected for
welded (HSW). The properties of indiconcavity, porosity, and other defects in
vidual phases within the FZ will be
accordance with General Motors engievaluated using Vickers microhardness
neering standards (GM4485M) (Ref.
and nanoindentation.
24). The minimum concavity of the
joint was estimated to be within the inExperimental Methodology dustrial acceptable range (below 20%)
(Ref. 25). Welds were made in the beadon-plate configuration, and the materiFiber laser welds were made on 1.00als, hot stamping conditions, and weldmm-thick Al-Si-coated PHS, where the
ing parameters can be found in Table 3.
yield strength and ultimate tensile
The mechanically ground and polstrength of the as-received sheets are
ished specimens were etched with 5%
about 424 and 570 MPa, respectively.
Nital solution for 2 s to reveal the miThe chemical composition and mechancrostructure. Microstructures were
ical properties are listed in Tables 1 and
characterized using an optical micro2, respectively. For hot stamping, asscope and a field-emission scanning
received sheets were austenized in a furelectron microscope (FE-SEM, model:
nace at 930°C for 5 min, and then trans-
Table 1 — Nominal Compositions of the Press­Hardened Steel (PHS) Used in This Investigation
C
Mn
P
S
Si
B
Mo
Al
Cr
Ti
CEN (Ref. 20)
0.23
1.22
0.013
0.001
0.27
0.0032
0.02
0.039
0.20
0.037
0.48
148-s WELDING JOURNAL / MAY 2016, VOL. 95
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 149
WELDING RESEARCH
A
B
A
B
C
D
C
D
Fig. 4 — SEM micrographs showing microstructures at different
locations of the as­received welded sample (ARW). A — Full weld
profile; B — HAZ; C — intercritical HAZ; D — FZ.
Zeiss Leo 1550). The equilibrium binary
phase diagram of the studied steel was
calculated using commercial thermodynamic software ThermoCalc.
Vickers microhardness was measured under 200-g load and 15-s dwell
time. In order to avoid interference
from the strain fields developed by the
adjacent indents, a spacing of at least
three times the diagonal was maintained between indents, as recommended by ASTM Standard E384 (Ref. 26).
Nanoindentation was performed on the
etched specimens using 5000-mN load
for 20 s using a Hysitron Triboindenter
TI-900 equipped with a scanning probe
microscope. Transverse tensile specimens were machined as per ASTM Standard E8 (Ref. 27) with the sample dimensions shown in Refs. 28, 29. The
tensile tests were carried out using a
computerized tensile testing machine at
a crosshead speed of 1 mm/min.
Results
Coating and Base Metal
Characteristics
Figure 1 shows typical morphology
of the coating in the as-received (Fig.
1A) and hot-stamped (Fig. 1B) condition. As-received material has a coating
thickness of about 15–20 mm, which
was transformed and diffused into the
steel to a depth of about 30–40 mm during hot stamping. It was reported that
the growth of the coating thickness is
highly dependent on the austenization
temperature and time. Both parameters
increase the coating thickness (Ref. 4).
Fig. 5 — SEM micrographs showing microstructures at different lo­
cations of the as­received welded, then hot­stamped sample
(ARWHS). A — Full weld profile; B — HAZ and FZ; C — supercritical
HAZ; D — FZ.
Table 2 — Tensile Properties and Coating Compositions of the Studied Steels
Conditions
0.2% Yield
(MPa)
UTS
(MPa)
Elongation
(%)
As­received BM
420
570
15
As­hot stamped
1122
1601
4
Coating of the as-received steel consists
of an Al matrix with elongated pure Si
particles (Fig. 1A), which formed by a
eutectic reaction during the aluminizing
process (Ref. 30). An intermediate layer
exists between the coating, and the
steel substrate has a thickness of about
4–6 mm (Ref. 31) with a composition of
Fe2SiAl7.
In addition, a thin layer (<1 mm
thickness) of FeAl3 was also observed
between the steel substrate and the
Fe2SiAl7 layer, as shown in the inset of
Fig. 1A (Ref. 30). Energy-dispersive
spectroscopy (EDS) line scanning of
the coating (Fig. 1C) shows a thin Sienriched layer at the outer surface of
the coating. A multilayered coating is
formed during heat treatment (Refs. 4,
30–34), as shown in Fig. 1B and D. A
total of five distinct microstructural
regions can be found in the coating in
the hot-stamped condition, with different morphologies and compositions
(Refs. 30, 31, 34).
The microstructure of the asreceived steel is comprised predominantly of ferrite, pearlite, and a small
amount of martensite at the grain
boundaries — Fig. 2A. A typical mor-
Coating
(76 g/m2)
Al
(wt­%)
Si
(wt­%)
90
10
phology of lath martensite can be found
in the hot-stamped condition — Fig.
2B. Laths are parallel to each other and
some of the laths contain distributed
carbides, which resulted from the autotempering (Ref. 35).
Microstructure of the Welds
To investigate the influences of the
Al-Si coating on the FZ microstructure,
the coating was removed chemically by
dipping into a solution of 10% sodium
hydroxide (NaOH) in deionized water.
Figure 3 shows the full weld profile and
the FZ microstructure; a fully martensitic microstructure was obtained before
and after hot stamping of the welded
coupon.
As­Received Welded (ARW)
Bead-on-plate welds were made with
the welding parameters as specified in
Table 3. From Fig. 4, it is noted that the
thermal cycle employed during welding
has drastically changed the HAZ and FZ
microstructures. Although the fiber
laser welding process involves rapid
heating and cooling, the FZ was not
MAY 2016 / WELDING JOURNAL 149-s
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 150
WELDING RESEARCH
A
B
A
B
C
D
C
D
Fig. 6 — SEM micrographs showing microstructures at different
locations of the hot­stamped welded sample (HSW). A — Full
weld profile; B — HAZ; C — subcritical HAZ; D — FZ.
A
B
C
D
Fig. 8 — SEM micrographs showing Al­Si coating mixing and re­
sultant microstructures at various locations of the ARWHS sample.
transformed to complete martensite
due to mixing of the coating into the
FZ. The FZ microstructure in the coated
sheet is a combination of skeletal ferrite
and martensite (Fig. 4D), and the area
fraction of ferrite and martensite (using
ImageJ software) (Ref. 36) was estimated to be about 20 and 80%, respectively.
It has been reported that the skeletal
pattern ferrite is a typical structure of
high-temperature delta-ferrite (d-ferrite) (Refs. 37, 38); therefore, the ferrite
Fig. 7 — SEM micrographs showing Al­Si coating mixing and re­
sultant microstructures at various locations of the ARW sample.
A
B
C
D
Fig. 9 — Al­Si coating mixing in the fusion boundary of the HSW
sample shows continuous network­like δ­ferrite.
structure found on the ARW sample can
be referred to as d-ferrite. From the solidification pattern, it can be presumed
the d-ferrite phase was solidified as the
first solidified phase; afterward, the
martensite phase appeared. The influence of the coating on the FZ microstructure modification is discussed
later. The BM contains only ferrite and
pearlite, which are already softer than
the martensite in the FZ; therefore, no
heat-affected zone (HAZ) softening was
Table 3 — Materials, Hot Stamping Conditions, and Welding Parameters Used in
This Study
Conditions
Power
(kW)
Focal Length
(mm)
Spot Size Speed
(mm)
(m/min)
Stamping Conditions
ARW
4
200
0.6
12
As­received welded
ARWHS
4
200
0.6
12
HSW
4
200
0.6
12
As­received welded
then hot stampd
Hot­stamped welded
150-s WELDING JOURNAL / MAY 2016, VOL. 95
detected. However, the HAZ of the
ARW sample consists of retransformed
martensite and ferrite (Fig. 4C) where
the peak temperature during welding
was between the Ac1 and Ac3 temperatures of the steel.
As­Received Welded then Hot
Stamped (ARWHS)
The as-received material was welded using similar welding parameters as
used for ARW sample (Table 3). The
welded coupons were then austenized
and quenched between dies, which resulted in the microstructures shown in
Fig. 5. It can be noted that after hot
stamping, the HAZ microstructure is
completely transformed to martensite
similar to the hot-stamped BM; therefore, identical microhardness can be
expected, since the microhardness is
dependent on martensite fraction.
It is suggested that the hot stamp-
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 151
WELDING RESEARCH
A
B
Table 4 — Average Al and Si wt­% in Fer­
rite Phase for Three Different Sample
Conditions
Conditions
E
2.52
0.08
0.71
0.07
ARWHS
1.70
0.21
0.67
0.10
2.64
0.23
0.60
D
F
Fig. 10 — EDS line scanning results of Al and Si across the martensite and ferrite phase. A
and B — center of the FZ in the ARW sample; C and D — center of the FZ in the ARWHS
sample; E and F — the fusion boundary in the HSW sample.
ing process could not transform the
steel into a fully martensite structure
due to the mixing of the coating alloy
into the FZ. The FZ microstructure
was found to be a combination of ferrite and martensite. The peak temperature (930°C) in the hot-stamping
process is identical to the intercritical
HAZ temperature (between Ac1 and
Ac3); therefore, during hot stamping,
the d-ferrite and martensite structure
can transform to ferrite (a-ferrite) and
austenite (g). This would explain the
ferrite phase observed in the FZ of the
ARWHS sample, which can be referred
to as a-ferrite. This will be discussed
in a later section. The area fraction of
the a-ferrite and martensite were estimated to be about 40 and 60%, respectively, compared to 20% d-ferrite, and
80% martensite in the ARW condition
— Fig. 4D.
Si(wt%)
ARW
HSW
0.13
C
Al (wt­%)
Hot­Stamped Welded (HSW)
Unless coating mixing occurs at the
fusion boundary, the FZ is also composed of 100% martensite, which has a
similar BM lath microstructure — Fig.
6D. However, the martensite laths are
clearly thinner and finer in the FZ,
which suggests the cooling rate was
much faster during welding. The HAZ
microstructure can be divided into
three distinct zones termed as subcritical HAZ, intercritical HAZ, and supercritical HAZ, depending on the temperature ranges. Among these zones, the
subcritical HAZ (Fig. 6C) experienced
temperatures between room temperature and the Ac1 temperature, such that
the BM martensite structure underwent tempering. During tempering, the
martensite structure is decomposed to
ferrite and cementite (Fe3C), which has
a lower strength compared to a fully
martensite structure (Refs. 39–42),
which degrades the tensile, fatigue, and
formability behavior of the welded
joints (Refs. 23, 28, 29).
The next zone is the intercritical
HAZ, where partial austenization occurred between the temperatures of Ac1
and Ac3 line. The microstructure was
found to be a combination of ferrite and
retransformed martensite — Fig. 6B. As
the temperature increased in the HAZ,
the fraction of martensite increased as
well, thereby improving the mechanical
properties in contrast to the subcritical
HAZ. When the local peak temperature
during welding exceeds the Ac3 temperature, the microstructure becomes fully
martensitic (with the highest microhardness) and is therefore referred to as
the supercritical HAZ. The FZ structure
was found to be fully martensitic,
whereas the fusion boundary microstructure is composed of martensite
and a network of d-ferrite phase.
During laser beam welding, a higher
vapor pressure pushed away the molten
metal from the keyhole center (Ref. 43)
and the weld metal mixed with the Al-Si
coating generated a vortex flow at the
upper sides of the molten pool. Therefore, due to limited solubility of the hotstamped coating to the weld pool, the
coating was not mixed with the molten
metal homogeneously. Consequently,
the local Al concentration was higher at
the fusion boundary, which stabilizes
the high-temperature ferrite phase (dferrite) during rapid solidification.
Coating Mixing into the
Fusion Zone
In the current studies, it was found
that the presence of an Al-Si coating influences the formation of d-ferrite
phase in the FZ (Figs. 4–6), since no evidence of d-ferrite was observed in the
welded sample where this coating was
MAY 2016 / WELDING JOURNAL 151-s
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 152
WELDING RESEARCH
Fig. 11 — ThermoCalc binary phase diagrams of Fe­Al and il­
lustrated influence of Al on the phase transformation and δ­
ferrite formation.
chemically removed — Fig. 3. It was
also observed that the presence of the
coating on either surface may be mixed
into the FZ; however, a coating on the
top surface of the sheet has the most
prominent effect on the d-ferrite formation. Coating mixing was most noticeable along the fusion boundary, where
d-ferrite was formed as a continuous
structure due to the effect of Marangoni
convection flow (Refs. 44, 45).
Figures 7 and 8 distinguish the
coating mixing and resultant microstructure among the ARW and ARWHS conditions. Although both of the
FZ consist of a ferrite (d-ferrite or aferrite) and martensite structure, the
martensite fraction is clearly higher in
the ARW condition (about 80%). It is
worth mentioning here that the higher
fraction of martensite is associated
with higher heating parameters, i.e.,
temperature and time, during hot
stamping, as reported by Kang et al.
(Ref. 19). Their calculation shows that
when the Al and Si content increases,
the ferrite to austenite phase transformation temperature also increases.
Martensite formed in the FZ of the
ARW sample resulted from the higher
austenization temperature and rapid
cooling (~10,000°C/s (Refs. 23, 46))
during welding, whereas the ARWHS
specimen showed a decreased fraction
of the martensite (about 40%) because
of the lower austenization temperature (930°C) and slower cooling rate
(~31°C/s) in the hot-stamping process.
Another distinguishable feature is that
152-s WELDING JOURNAL / MAY 2016, VOL. 95
A
B
C
D
Fig. 12 — Nanoindentation study of the different phases present
in the weldment. A — The ARW sample; B — the ARWHS sam­
ple; C — the HSW sample; D — the comparison of the
nanohardness among the samples. Error bars represent the
standard deviation.
A
B
Fig. 13 — Microhardness distributions across the weldment. A — ARW and ARWHS sam­
ples; B — HSW sample. Error bars represent the standard deviation.
samples in the ARW condition contain
a larger fraction of d-ferrite along the
fusion boundary (Fig. 7D); conversely,
a uniform distribution of the a-ferrite
and martensite can be identified
throughout the FZ of the ARWHS
sample — Fig. 8D.
Unlike the above two cases, the
HSW condition exhibits coating mixing only along the fusion boundary —
Fig. 9. There is no trace of d-ferrite at
the center of the FZ, where the microstructure was fully martensite with
a high microhardness (505 ± 23 HV). A
higher fraction of localized d-ferrite at
the fusion boundary indicates that
there was limited mixing of the coating in the weld pool. As a result, the local concentration of Al at the fusion
boundary was higher compared to the
center area of the FZ. The d-ferrite
phase at fusion boundary was found to
have a skeletal-like continuous struc-
ture (Fig. 9C and D) rather than a
grain-like structure — Fig. 8D.
The compositional analysis of the
martensite and ferrite phase was performed using the EDS line scanning
technique as presented in Fig. 10. In addition, the chemical composition of
each phase was measured (Table 4); the
result shows rich Al content (1.70 to
2.64 wt-%) in the ferrite phase. The influence of Al content on phase transformation and d-ferrite formation is predicted using a ThermoCalc Fe-Al binary
phase diagram (Fig. 11), and is discussed later. However, Si content was
found to be at the same level in all three
different sample conditions.
Mechanical Properties
Hardness
Nanohardness of the individual
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 153
WELDING RESEARCH
Fig. 14 — Representative engineering stress versus engineering
strain curves of the base metals and welded samples (ARBM: as­
received base metal, HSBM: hot­stamped base metal, ARW: as­
received welded, ARWHS: as­received welded then hot stamped,
and HSW: hot­stamped welded).
phases of martensite, and ferrite (dferrite and a-ferrite) were further estimated by using instrumented nanoindentation tests. The indent impressions from individual grains on each
phase in three different samples are
presented in Fig. 12. A total of 12 to
16 indents were made on each phase,
i.e., on martensite and ferrite in the
FZ (ARW and AWRHS sample) and the
fusion boundary (HSW sample). The
average nanohardness of the martensite and ferrite phase ranges from 7.1
to 7.9 GPa and 4.5 to 4.9 GPa, respectively, for all three sample conditions.
Hardness impression on the martensite islands of the ARWHS sample exhibit the lowest nanohardness (7.1 ± 0.6
GPa) among the martensite phases on
the two other samples, which is due to
the lower cooling rate in the hot-stamping process and smaller stress-free
martensite grain. Conversely, nanohardness measured on the ferrite phase is
fairly in the same range, which indicates
that the strength of the ferrite phase
was not influenced by the hot-stamping
process. It should be concluded that the
second phase found in the FZ of all material conditions definitely was not consistent with the hardness of an intermetallic compound as reported in earlier literature (Refs. 13, 17, 18). The
nanohardness of the intermetallic compound phase (Fe (Al, Si), Fig. 1B) is
much higher (10.1 ± 0.5 GPa) compared
to the hardness of ferrite (4.1–4.9 GPa)
phase.
The microhardness distributions
A
B
C
D
Fig. 15 — The optical micrographs and failure locations of the
tensile tested samples. A — ARBM and ARW sample (both sam­
ples failed at identical location, i.e., the BM, therefore, only one
micrograph is showing here); B — HSBM sample; C — ARWHS
sample; D — HSW sample.
across the three different welded samples are presented in Fig. 13. The asreceived and hot-stamped BM hardness
values are about 244 ± 4, and 575 ± 8
HV, respectively. Hardness profiles obtained from the ARW and HSW conditions exhibited the maximum hardness
in the supercritical HAZ, which was due
to the solid-state transformation, rapid
cooling, and a larger fraction of the
martensite phase compared to other locations. Although there are significant
differences in hardness among the asreceived and hot-stamped steels, the average FZ hardness of the ARW and
HSW samples were similar, about 516 ±
25 and 505 ± 23 HV, respectively.
A larger deviation of the hardness
in the FZ was believed to be the effect
of the coating mixing in the fusion
boundary. For instance, the circular
marked point on the hardness profile
of the HSW sample (Fig. 13B) indicates the fusion boundary, where there
is a higher fraction of network-like dferrite — Fig. 9C. The average microand nanohardness of the fusion
boundary and ferrite were measured
to be about 404 ± 63 HV and 4.5 ± 0.1
GPa, respectively. Unlike the HAZ of
the ARW condition, the HAZ of the
HSW sample undergoes tempering because of the martensite microstructure present in its BM. It has been reported that tempering of martensite
has reduced the strength of the joint
in all types of welding processes (Refs.
23, 28, 29).
In this study, it was found that the
HAZ hardness was reduced by about
35% (from 575 ± 8 HV (BM) to 374 ±
25 HV (HAZ)) compared to the hotstamped BM hardness. Similar softening was reported by Jia et al. (Ref. 47)
in hot-stamped welded 22MnB5 steel.
In addition, they also found that with
increasing welding speed, the HAZ and
FZ hardness increases due to lower
heat input. While comparing hardness
distribution of the ARWHS sample
with other conditions, there is no softening or variation of the hardness between BM and HAZ. Conversely, a
sharp drop in hardness was measured
in the FZ (Fig. 13A); the BM hardness
was 575 ± 8 HV, which decreased to
394 ± 24 HV, representing about a
32% reduction of hardness, in contrast
to the BM and HAZ. The degree of
softening in the FZ of the ARWHS
sample was the same as that observed
due to the HAZ softening in the HSW
sample — Fig. 13B.
Tensile Properties and
Failure Locations
Typical engineering stress vs. strain
curves of the BMs and three different
welded samples are presented in Fig.
14. The corresponding mechanical
properties and the failure macrographs of the joints can be found in
Table 5 and Fig. 15, respectively.
The hot-stamped BM (HSBM) shows
the highest yield strength (1122 MPa)
and ultimate tensile strength (1601
MPa), which can be expected from the
MAY 2016 / WELDING JOURNAL 153-s
Saha 5-16.qxp_Layout 1 4/15/16 3:04 PM Page 154
WELDING RESEARCH
Table 5 — Mechanical Properties of the Joints of Three Different Conditions
Conditions
Yield Strength Ultimate Tensile
(MPa)
Strength (MPa)
Elongation Joint Efficiency Fracture
(%)
(%)
Locations
ARW
412 ± 10
579 ± 16
15.6
101
BM
ARWHS
1050 ± 60
1131 ± 73
0.77
71
FB­FZ
HSW
1140 ± 8
1307 ± 15
0.85
82
FB
BM: base metal, FB: fusion boundary, and FZ: fusion zone
fully martensitic microstructure (Fig.
2B). While comparing the as-received
BM (ARBM) with its welded condition
(ARW), there is a negligible difference in
the yield strength and the ultimate tensile strength. However, the as-welded
sample shows a slight increase in the ultimate tensile strength due to the formation of martensite in the FZ (Fig.
4D), which exhibited higher microhardness — Fig. 13. It can be expected that a
higher fraction of martensite in the FZ
will reduce the joint elongation as half
of the joints were yielded; therefore,
necking will likely occur in either one
half of the joint, and as a result, the
joint elongation is reduced.
In both of the cases, the welded
joints failed at the BM (Fig. 15A), which
suggests the welding condition does not
influence the mechanical properties of
the joints. When the samples were hot
stamped after welding (ARWHS samples), the ductility of the joint was reduced dramatically, which is due to the
fact that the ferrite and pearlite in the
BM (Fig. 2A) were transformed completely to martensite; the FZ microstructure remained a mixture of aferrite and martensite — Fig. 5D.
Tensile test results of the ARWHS
condition exhibited the lowest elongation (0.77%) and lowest yield strength
(1050 ± 60 MPa) and ultimate tensile
strength (1131 ± 73 MPa) compared to
other hot-stamped conditions; a 30%
reduction of joint efficiency was estimated in this case (Table 5). Tensile
fracture occurred consistently between
the fusion boundary and FZ — Fig.
15C. Conversely, the HSW condition
shows higher tensile properties (joint
efficiency 82%) compared to the other
welded conditions, which is due to the
formation of a larger fraction of
martensite in the FZ. All of the welded
hot-stamped steel reached a maximum
of 0.85% elongation. As explained in
154-s WELDING JOURNAL / MAY 2016, VOL. 95
the earlier section, in this case, coating
mixing occurs only at fusion boundary,
which acted as a site for strain localization, where the softer d-ferrite phase
promoted fracture to occur in this region and reduced ductility. This is
clearly noted by the negligible necking
observed in Fig. 15C and D. Although
the subcritical HAZ has the lowest microhardness value (374 ± 25 HV) (Fig.
13), the soft d-ferrite has the lowest
nanohardness (4.5 to 4.9 GPa) compared to the martensite (Fig. 12) and
tempered martensite phase (Ref. 48).
Discussion
Coating Mixing and Weld
Solidification
Al-Si coating mixing was detected for
all three different weld conditions as
confirmed by the composition analysis
(Fig. 10 and Table 4). The results exhibited a rich Al percentage at the ferrite
phase compared to the martensite
phase. However, the Si concentration
did not vary significantly among the
two phases, with all three samples indicating similar uniform Si concentration
profiles. A higher Al concentration is observed in the ARW (Fig. 10A and B) and
HSW samples (Fig. 10E and F); conversely, the Al distribution did not fluctuate much in the ARWHS sample —
Fig. 10C and D. Table 4 also indicates a
lower Al content in the a-ferrite phase
of the ARWHS sample, implying that
diffusion of Al occurred during austenization in the furnace. A decrease in
the Al content in the welded and hot
stamped sample (similar to the ARWHS
sample as defined in this study) was
also detected by Kang et al. (Ref. 19),
compared to an as-welded specimen.
The further found that with increasing
austenization temperature and time, Al
content in the a-ferrite decreases due to
diffusion of Al to the bulk material.
In the current investigation, a similar
phenomenon was observed. Al content
was found to decrease from 2.52 to 1.70
wt-% after hot stamping of the ARW
sample (Table 4). Since Al is a strong
ferrite stabilizer, an increase in Al can
shrink the single-phase austenite region
(g) and promote the formation of a twophase region (d + g) (Ref. 49). According
to the calculated phase diagram presented in Fig. 11, during cooling, a dferrite phase should be formed (L → d +
L) first from the liquid phase (ThermoCalc Scheil solidification calculation also
confirmed that the d-ferrite would be
the first phase); afterward, the liquid (L)
phase is completely transformed into
austenite (g) through a peritectic reaction (d + L → d + g).
The compositional analysis (Fig. 10
and Table 4) shows a higher percentage
(1.7 to 2.52%) of Al on the ferrite phase.
It was reported that an Al content of
more than 1.2 wt-% is sufficient to prevent the austenitic transformation (Ref.
50). The Fe-Al binary phase diagram
(Fig. 11) indicates if the Al content exceeds 0.78%, it prevents single austenite (g) phase formation, and if higher
than 1.05 wt-%, then it stabilizes only
d-ferrite phase. Therefore, all three different samples show a combination of
martensite and ferrite structure in either the FZ or the fusion boundary.
While comparing the FZ microstructures of ARW and ARWHS samples, it
can be noted that martensite grains are
larger in the ARW sample (Fig. 4D),
which is also due to the effect of heating
parameters and cooling rate. The morphology of the martensite structure is
also different. In the ARW sample, the
nucleation and growth of the martensite was faster, increasing the grain size;
an elongated slender martensite was observed after hot stamping — Fig. 5D.
The d-ferrite phase in the ARW sample
shows an irregular shape, which seems
to be highly stressed by the surrounding
martensite. It can be noted that while a
martensite plate forms and grows, it introduces a large compressive stress on
the surrounding ferrite phase (Ref. 35).
When the highly stressed ferrite phase
was austenized at 930°C for 5 min, the
compressive stress was released and an
unstressed a-ferrite phase was formed
via recrystallization.
Saha 5-16.qxp_Layout 1 4/15/16 2:10 PM Page 155
WELDING RESEARCH
Structure­Property Correlation
This section will deal with the hotstamped version of the steels, i.e., ARWHS and HSW specimens. In both
cases, the minimum hardness was estimated to be at the FZ and subcritical
HAZ, respectively. It was also observed
that hardness degraded in a similar
amount (ARWHS: 32% and HSW:
35%) in contrast to the BM — Fig. 13.
In addition, hardness also dropped at
the fusion boundary due to coating
mixing and formation of d-ferrite,
which accounts for a 30% reduction of
the hardness compared to the BM. In
addition to concavity, which acts as a
physical crack initiation zone, the
large variation of hardness between
the HAZ and FZ will effectively provide a metallurgical notch during tensile tests. It was interesting to observe
that failure did not occur at the subcritical HAZ where the highest hardness drop (35%) was measured, but
rather the origin of failure occurred at
the fusion boundary. This is a result of
the continuous network structure of
the fusion boundary d-ferrite, which
introduces a larger metallurgical notch
and promotes failure.
Therefore, the HSW sample experiences failure along the fusion boundary
and leaves the FZ completely intact. As
in the HSW sample, the Al-Si coating
mixing predominantly occurred on the
top surface and extended to half of the
sheet thickness; therefore, it can be suggested that half of the weld thickness
was fractured along the soft d-ferrite as
suggested by the fracture path (point 1,
Fig. 15D) and the other half failed by
tearing of the martensite phase (point
2, Fig. 15D) due to stress localization.
In contrast, tensile tests of the ARWHS sample show failure occurred predominantly at the FZ. The fracture initiated at the top surface of the fusion
boundary and propagated through the
FZ —Fig. 15C. As expected, the larger
hardness drop at the fusion boundary
promoted stress localization. Therefore,
it can be noted that the failure of the
laser-welded Al-Si hot-stamped steel is
dependent on the presence of d-ferrite
in the FZ microstructures.
Conclusions
The microstructure and the mechanical properties of fiber laser-welded PHS
were characterized considering the effect of Al-Si coating mixing in the weld
pool. The coating mixing role on weld
solidification was examined and the major findings of this study can be summarized as follows:
1) The FZ microstructure was modified due to the effect of coating mixing.
Significant coating mixing occurred in
the FZ of the ARW sample, and
promoted d-ferrite formation.
Therefore, the FZ microstructure was
found to be a combination of d-ferrite
and martensite. Conversely, the HSW
sample indicates a continuous networklike d-ferrite at the fusion boundary,
and a fully martensitic microstructure
in the center of the FZ.
2) A higher percentage of Al prompted ferrite formation. In addition, ferrite
formation was also influenced by the
hot stamping process. Lower cooling
rate increases the ferrite formation;
therefore, the fraction of martensite
was found to be decreased after hot
stamping of the ARW sample (i.e.,
ARWHS sample). In this case, the FZ
microstructure is a combination of
×-ferrite and martensite.
3) Vickers microhardness and
nanoindentation on the fusion boundary and ferrite revealed significantly
lower hardness values, 404 ± 63 HV
and 4.5 to 4.9 GPa, respectively, as
compared to other regions of the weldment. Likewise, the ferrite and
martensite dual phase structure shows
the lowest microhardness (about 394
± 24 HV) at the FZ.
4) The strength of the fiber laserwelded PHS is dependent on the fraction of ferrite phase formation, the
microhardness decreased by about
32% as a result of coating mixing and
leading to failure across the FZ in the
ARWHS sample, whereas the HSW
specimen failed at the fusion boundary. Therefore, an insignificant effect
of martensite tempering on the joint
performance was observed.
Acknowledgments
The authors would like to acknowledge AUTO21, Canada’s Automotive
Research and Development Program,
NSERC, and Innovation in Automotive
Manufacturing Initiative (IAMI) in
Canada for financing this project. The
authors are also thankful to Arcelor-
Mittal Dofasco, Inc., in Hamilton,
Canada, for providing the materials to
carry out this work. The authors would
like to thank Prof. Michael Worswick
from the Mechanical and Mechatronics
Engineering Department, University of
Waterloo, for the hot stamping facilities, and Dr. Yuquan Ding as well for
his kind cooperation with the nanoindentation study.
References
1. Karbasian, H., and Tekkaya A. E.
2010. A review on hot stamping. Journal of
Materials Processing Technology 210(15):
2103–2118.
2. Drew, S., Greg, F., Michael, W., Susan,
C. A., Rita Van, D., Nicholas, Z. M., Jeff, A.,
Dorothy, K., and George, M. 2011. Climate,
health, agricultural and economic impacts
of tighter vehicle-emission standards. Nature Climate Change 1(1): 59–66.
3. Kim, H. J., McMillan, C., Keoleian, G.
A., and Skerlos, S. J. 2010. Greenhouse gas
emissions payback for lightweighted vehicles using aluminum and high-strength
steel. Journal of Industrial Ecology 14(6):
929–946.
4. Fan, D. W., and De Cooman, B. C.
2012. State-of-the-knowledge on coating
systems for hot stamped parts. Steel Research International 83(5): 412–433.
5. Yao, Y., Meng, J. P., Ma, L. Y., Zhao,
G. Q., and Wang, L. R. 2013. Study on hot
stamping and usibor 1500P. Applied Mechanics and Materials 320: 419–425.
6. Kolleck, R., and Veit, R. 2010. Tools
and technologies for hot forming with local adjustment of part properties. Materials Science Forum 638-642: 3919–3924.
7. Saunders, F. I., and Wagoner, R. H.
1996. Forming of tailor-welded blanks.
Metallurgical and Materials Transactions A
27(9): 2605–2616.
8. Chan, L. C., Cheng, C. H., Chan, S.
M., Lee, T. C., and Chow, C. L. 2005.
Formability analysis of tailor-welded
blanks of different thickness ratios. Journal of Manufacturing Science and Engineering 127(4): 743–751.
9. Min., K. B., and Kang, S. S. 2000. A
study on resistance welding in steel sheets
for tailor welded blank evaluation of flash
weldability and formability (2nd Report).
Journal of Materials Processing Technology
103(2): 218–224.
10. Quintino, L., Costa, A., Miranda, R.,
Yapp, D., Kumar, V., and Kong, C. J. 2007.
Welding with high power fiber lasers — A
preliminary study. Materials & Design 28
(4): 1231–1237.
11. Thomy, C., Seefeld, T., and Vollertsen, F. 2005. High-power fibre lasers —
MAY 2016 / WELDING JOURNAL 155-s
Saha 5-16.qxp_Layout 1 4/14/16 6:18 PM Page 156
WELDING RESEARCH
Application potentials for welding of steel
and aluminium sheet material. Advanced
Materials Research 6-8: 171 to 178.
12. Westerbaan, D. 2013. Fiber laser
welding of advanced high strength steels.
M. Sc. thesis. Waterloo, ON, University of
Waterloo.
13. Kim, C., Kang, M. J., and Park, Y. D.
2011. Laser welding of Al-Si coated hot
stamping steel. Procedia Engineering 10(0):
2226–2231.
14. Moon, J. H., Seo, P. K., and Kang, C.
G. 2013. A study on mechanical properties
of laser-welded blank of a boron sheet steel
by laser ablation variable of Al-Si coating
layer. International Journal of Precision Engineering and Manufacturing 14(2): 283–288.
15. Kwon, M. S., Kim, Y. G., Kim, Y. J.,
Kang, C. Y., Kong, J. P., Oh, M. H., Shin, H.
J., and Oh, S. T. 2014. Tailor welded blank,
manufacturing method thereof, and hot
stamped component using tailor welded
blank. U.S. Patent 20140154521 A1.
16. Pic, A., Duque Múnera, D., and
Pinard, F. 2008. Press hardened steel based
laser welded blanks: The ultimate tool for
crashworthiness. Metallurgical Research &
Technology 105(01): 50–59.
17. Vierstraete, R., Ehling, W., Pinard,
F., Cretteur, L., Pic, A., and Yin, Q. 2010.
Laser ablation for hardening laser welded
steel blanks. Industrial Laser Solutions
25(2): 6.
18. Ehling, W., Cretteur, L., Pic, A.,
Vierstraete, R., and Yin, Q. 2009. Proc. 5th
Int. WLT Conf. Lasers Manuf., Munich,
Germany.
19. Kang, M., Kim, Y. M., and Kim, C.
2015. Effect of heating parameters on laser
welded tailored blanks of hot press forming steel. Journal of Materials Processing
Technology.
20. Kasuya, T., and Yurioka, N. 1993.
Carbon equivalent and multiplying factor
for hardenability of steel. Welding Journal
72(6): 263-s to 268-s.
21. Merklein, M., and Lechler, J. 2006.
Investigation of the thermo-mechanical
properties of hot stamping steels. Journal
of Materials Processing Technology 177(1-3):
452–455.
22. Bardelcik, A., Salisbury, C. P., Winkler, S., Wells, M. A., and Worswick, M. J.
2010. Effect of cooling rate on the high
strain rate properties of boron steel. International Journal of Impact Engineering
37(6): 694–702.
23. Saha, D. C., Westerbaan, D., Nayak,
S. S., Biro, E., Gerlich, A. P., and Zhou, Y.
2014. Microstructure-properties correlation in fiber laser welding of dual-phase
and HSLA steels. Materials Science and Engineering: A 607: 445–453.
24. General Motors Engineering Standards. 2002. Weld Specifications Laser
Welds-Butt Joints. GM4485M.
25. Westerbaan, D., Parkes, D., Nayak,
S. S., Chen, D. L., Biro, E., Goodwin, F., and
Zhou. Y. 2014. Effects of concavity on tensile and fatigue properties in fibre laser
welding of automotive steels. Science and
Technology of Welding and Joining 19(01):
60–68.
26. ASTM Standard, E384-11E1. 2007.
Standard test method for Knoop and Vickers hardness of materials. West Conshohocken, Pa.: ASTM International.
27. ASTM Standard E8. 2011. Standard
test methods for tension testing of metallic materials. West Conshohocken, Pa.:
ASTM International.
28. Xu, W., Westerbaan, D., Nayak, S.,
Chen, D., Goodwin, F., Biro. E., and Zhou,
Y. 2012. Microstructure and fatigue performance of single and multiple linear
fiber laser welded DP980 dual-phase steel.
Materials Science and Engineering: A 553:
51–58.
29. Xu, W., Westerbaan, D., Nayak, S. S.,
Chen, D. L., Goodwin, F., and Zhou, Y.
2013. Tensile and fatigue properties of
fiber laser welded high strength low alloy
and DP980 dual-phase steel joints. Materials & Design 43: 373–383.
30. Fan, D. W., Kim, H. S., Oh, J. K.,
Chin, K. G., and De Cooman, B. C. 2010.
Coating degradation in hot press forming.
ISIJ International 50(4): 561–568.
31. Windmann, M., Röttger, A., and
Theisen, W. 2014. Formation of intermetallic phases in Al-coated hot-stamped
22MnB5 sheets in terms of coating thickness and Si content. Surface and Coatings
Technology 246: 17–25.
32. Drillet, P., Drillet, D., and Kefferstein, R. 2009. Process for manufacturing
stamped products, and stamped products
prepared from the same U.S. Patent
8733142 B2.
33. Grauer, S. J., Caron, E. J. F. R.,
Chester, N. L., Wells, M. A., and Daun, K. J.
2015. Investigation of melting in the Al–Si
coating of a boron steel sheet by differential
scanning calorimetry. Journal of Materials
Processing Technology 216: 89–94.
34. Suehiro, M., Maki, J., Kusumi, K.,
Ohgami, M., and Miyakoshi, T. 2003. Properties of aluminum coated steels for hotforming. Nippon Steel Technical Report:
15–20.
35. Bhadeshia, H. K. D. H., and Honeycombe, R. W. K. 2011. Steels — Microstructure and Properties. Butterworth-Heinemann, Oxford, UK, Elsevier.
36. Rasband, W. S. 1997. ImageJ, National Institutes of Health, Bethesda, Md.
0 38. Amirthalingam, M., Hermans, M.,
and Richardson, I. 2009. Microstructural
development during welding of siliconand aluminum-based transformation-induced plasticity steels — Inclusion and elemental partitioning analysis. Metallurgical
and Materials Transactions A 40(4):
901–909.
156-s WELDING JOURNAL / MAY 2016, VOL. 95
39. Biro, E., Vignier, S., Kaczynski, C.,
McDermid, J. R., Lucas, E., Embury, J. D.,
and Zhou, Y. N. 2013. Predicting transient
softening in the sub-critical heat-affected
zone of dual-phase and martensitic steel
welds. ISIJ International 53(1): 110–118.
40. Biro, E., McDermid, J., Embury, J.,
and Zhou, Y. 2010. Softening kinetics in
the subcritical heat-affected zone of dualphase steel welds. Metallurgical and Materials Transactions A 41(9): 2348–2356.
41. Xia, M., Biro, E., Tian, Z., and Zhou,
Y. N. 2008. Effects of heat input and
martensite on HAZ softening in laser welding of dual phase steels. ISIJ International
48(6): 809–814.
42. Baltazar Hernandez, V. H., Nayak, S.
S., and Zhou, Y. 2011. Tempering of
martensite in dual-phase steels and its effects on softening behavior. Metallurgical
and Materials Transactions A 42(10):
3115–3129.
43. Mickael, C., Muriel, C., Philippe Le,
M., Sadok, G., and Mikhaël, B. 2013. A new
approach to compute multi-reflections of
laser beam in a keyhole for heat transfer
and fluid flow modelling in laser welding.
Journal of Physics D: Applied Physics 46(50):
505305.
44. Limmaneevichitr, C., and Kou, S.
2000. Visualization of Marangoni convection in simulated weld pools. Welding Journal
79(5): 126-s to 135-s.
45. Limmaneevichitr, C., and Kou, S.
2000. Experiments to simulate effect of
Marangoni convection on weld pool shape.
Welding Journal 79(8): 231-s to 237-s.
46. Gould, J., Khurana, S., and Li, T.
2006. Predictions of microstructures when
welding automotive advanced highstrength steels. Welding Journal 85(5): 111s to 116-s.
47. Jia, J., Yang, S. L., Ni, W. Y., Bai, J.
Y., and Lin, Y. S. 2014. Microstructure and
properties of fiber laser welded joints of ultrahigh-strength steel 22MnB5 and its dissimilar combination with Q235 Steel. ISIJ
International 54(12): 2881 to 2889.
48. Baltazar Hernandez, V. H., Panda, S.
K., Kuntz, M. I., and Zhou, Y. 2010.
Nanoindentation and microstructure
analysis of resistance spot welded dual
phase steel. Materials Letters 64(2):
207–210.
49. San Martín, D., Palizdar, Y., GarcíaMateo, C., Cochrane, R. C., Brydson, R.,
and Scott, A. J. 2011. Influence of aluminum alloying and heating rate on
austenite formation in low carbon-manganese steels. Metallurgical and Materials
Transactions A 42(9): 2591–2608.
50. Briand, F., Dubet, O., and Choet, C.
2011. Process for Laser-ARC Hybrid Welding Aluminized Metal Workpieces. U.S.
Patent 20110226746 A1.