ISIJ International, Vol. 53 (2013), No. 1, pp.

96–101

Effects of Normalizing Processes on Microstructure and Impact

Toughness in Ti-bearing Weld Metal of Multilayer MAG Welded
HSLA Steel

Shanping LU,* Xin WANG, Wenchao DONG and Yiyi LI

Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Shenyang,
110016 China.
(Received on June 13, 2012; accepted on September 3, 2012)

The influences of different normalizing heat treatments on microstructure, non-metallic inclusions and
impact toughness in MAG Ti-bearing weld metal of HSLA steel has been studied. It has been shown that
for the Ti-bearing weld metal the impact toughness after normalizing treatment decreases significantly
against prolonging holding time and increasing normalizing temperature. The Mn-depleted zone forms
around the Ti-bearing phase (MnTiO3) precipitated on Mn–Si oxide. Proeutectoid ferrite preferentially
nucleates at the Mn-depleted zone and the interface between austenite and proeutectoid ferrite becomes
the nucleation sites for pearlite thereafter. Mn-depleted zone formation increases the pearlite nucleation
sites, and makes the pearlite fine. The dissolve of Ti-bearing precipitate causes disappear of Mn-depleted
zone at strong normalizing processes (longer normalizing time and higher normalizing temperature), and
the number of ferrite nucleation sites decreases, then the pearlite become coarser, which causes the
deterioration of impact toughness.

KEY WORDS: normalizing process; weld metal; impact toughness; Ti-bearing precipitate.

1. Introduction 2. Experimental Procedures

Ti has a great affinity for carbon atom, nitrogen atom and
oxygen atom, therefore, Ti carbides, Ti nitrides, Ti carbon-
nitrides and Ti oxides can precipitate easily in steel. These
precipitates have great influence on the mechanical proper-
ties and microstructure of steels. The calculation of solubil-
ity products has shown that the consequence of these phases
is Ti oxides, Ti nitrides and Ti carbides.1,2) However, the
final precipitates depend on the processing the material
undergone and the other microalloy elements, such as Nb
and V. There will be Ti-bearing oxides in a weld metal by
manual metal arc welding (MMAW),3,4) submerged arc weld-
ing (SAW),3) metal active gas welding (MAGW)5) and in Mn–
Si–Ti deoxidized steel.6) Simulation results show that few Ti
nitrides precipitate in the Heat-Affected Zone (HAZ).7,8) The
Ti-bearing complex carbonnitrides can pin and refine the
grain,9) and the Ti-bearing oxides have an effect on the
microstructure, especially for acicular ferrite formation, dur-
ing the transformation.5,10–12) However, the behavior and
effect of Ti-bearing oxides in weld metal on the microstruc-
ture and property after heat treatment is seldom reported.
In the present study, the effect of normalizing heat treat-
ment on the behavior of Ti-bearing oxides and weld metal
microstructure and impact toughness has been investigated
and discussed.
* Corresponding author: E-mail: shplu@imr.ac.cn
DOI: http://dx.doi.org/10.2355/isijinternational.53.96
Two kinds of low carbon steel welding wires with a diam-
eter of 1.0 mm were used in the experiment to weld a 12 mm
thick S355J2W (base metal) steel plate by the metal active
gas (MAG) welding. The geometry of the joint was Y
groove with the 60° groove angle. The joint was welded by
4 welding layers and the shielding gas was 82%Ar+
18%CO2. The schematic diagrams of the test plate and
groove type were shown in Fig. 1. Table 1 gave the chem-
ical compositions of the welding wires and the base metal.
The chemical compositions of the weld metals were ana-
lyzed and shown in Table 2. Since there was CO2 in the
shielding gas, nearly 60% Ti was burnt in the welding pro-
cess, and the Ti content in the weld metal was lower than
that in the welding wire. Therefore, the correspondent weld
metals welded by welding wire 1 and 2 were defined as the
weld metal without Ti and Ti-bearing weld metal, respec-
tively. The mechanical properties of the steel plate were
shown in Table 3.
The heat treatment parameters for the welded joint were
shown in Table 4. Specimens for the impact property test of
the weld metal were extracted from these joints transversely
to the welding direction using an electrical spark wire cut-
ting machine and were machined to a dimension of 55
mm×10 mm×10 mm. The V notch was machined in the cen-
ter of the weld metal on the cross section of the joints. The
impact tests were examined on the swing impacting test

© 2013 ISIJ 96
 ISIJ International, Vol. 53 (2013), No. 1

Table 3. Mechanical properties of the base metal.

Yield strength/ Tensile strength/ Elongation/ –20°C impact

MPa MPa % energy/J
370 505 30.0 175

Table 4. Normalizing processes of the joints.

Fig. 1. Schematic diagram of test plate and groove type.
Number Heat treatment processes
Table 1. Chemical compositions of base metal and welding wires
1 Normalizing at 900°C for 1 h
(wt %).
2 Normalizing at 1 000°C for 1 h
Material C Si Mn S P Cu Nb Cr Ni Ti
3 Normalizing at 1 100°C for 1 h
Base
0.067 0.39 1.32 0.002 0.010 0.24 0.019 0.41 0.21 <0.005 4 Normalizing at 1 200°C for 1 h
metal
Welding 5 Normalizing at 900°C for 0.5 h
0.040 0.41 1.09 0.0034 0.010 0.48 – 0.01 0.87 0.021
wire 1
6 Normalizing at 900°C for 2 h
Welding
0.048 0.40 1.09 0.0043 0.011 0.45 – 0.01 0.85 0.074 7 Normalizing at 900°C for 5 h
wire 2

Table 2. Chemical compositions of weld metal (wt %).

Material C Si Mn P S Ni Cr Ti Cu N O
WithoutTi 0.045 0.39 1.11 0.010 0.007 0.84 <0.05 <0.01 0.41 0.0031 0.057
Ti-bearing 0.051 0.36 1.07 0.009 0.008 0.83 <0.05 0.028 0.41 0.0034 0.038

machine at –20°C. The test was carried out three times and
the average values were calculated.
The microstructures of weld metal were observed by opti-
cal microscope, scanning electron microscope Hitachi S-
3400N and JSM-6301F. The specimens were etched by 4%
nital. The size distribution of the pearlite was determined on
polished surfaces by SEM at 20 kV with 800× magnifica-
tion. Five fields in each specimen were examined. The dis-
tribution of the pearlite was then analyzed by standard ana-
lyzing software (IASB Metallographic Analysis Software).
Transmission electron microscope (TEM) was used to
directly observe and analysis the inclusions. The thin foils
were prepared from small 3 mm diameter cylinders cut from
the weld metal. The cylinders were then sliced to a thickness
of 0.5 mm, using an electrical spark wire cutting machine
and the thickness was reduced to 90 μ m by grinding, using
“wet-dry” paper to P1500 grade. The thin foil specimens
were prepared by using twin-jet machine Struers Tenupol-5
and the electrolytic etch was 10%HClO4+alcohol. Then
argon ion thinned for 20 min by using Gatan 691 precision
ion polishing system. Selected area diffraction patterns
(SADPs) of the inclusions undergone the different heat
treatment were conducted by using a JEM 2100 electron
microscope equipped with an energy dispersive X-ray spec-
troscopy (EDS) operated at 200 kV.
3. Results and Discussion
3.1. Impact Property and Microstructure of the Weld
Metal in as Weld Condition
The weld metal impact energy at –20°C in as weld con-
dition are 119J and 143J for weld metals without Ti and with
Ti bearing, seperately.5) Since multi-pass welding (4 layers)
is applied in the experiment as shown in Fig. 1, the weld
metal microstructure composed of columnar grain zone
(CGZ) of the last weld layer and fine grain zone (FGZ) for
the first three layers, as shown in Fig. 2.
The microstructures of the weld metal are shown in Fig.
3. The microstructure in CGZ is subdivided into three con-
stituents, i.e., primary ferrite (PF) that includes both grain
boundary and intragranular polygonal ferrite, ferrite with
second phase (FS) that contains all grain boundaries nucle-
ated lath-like morphology of ferrite and intragranular acic-
ular ferrite (AF). The microstructure in FGZ is polygonal
ferrite. The AF can increase the weld metal impact property.
It is also shown that with small addition of Ti the size of the
PF in the CGZ is much finer and the acicular ferrite is well-
developed, as shown in Fig. 3(b). The size of polygonal fer-
rite in the FGZ for Ti-bearing weld metal is a little smaller,
as shown in Fig. 3(d). Therefore, the impact property for the
Ti-bearing weld metal in as weld condition is better than that
of weld metal without Ti.
3.2. The Effect of Normalizing Processes on the Tough-
ness of the Weld Metal
The impact energy of the Ti-bearing weld metal decreases
against increasing temperature, as shown in Fig. 4. After
normalizing, the columnar grain zone disappeared and the
impact property improved both for the Ti-bearing weld met-
al and weld metal without Ti addition. When the normaliz-
ing temperature is 1 100°C, the impact energy of the Ti-
bearing weld metal and the weld metal without Ti is nearly
the same. It also can be shown that the normalizing temper-
ature has significant effect on the Ti-bearing weld metal.
The impact property of the weld metal against prolonging
holding time at 900°C is shown in Fig. 5. Results showed
that the impact property of the Ti-bearing weld metal
decreases significantly against prolonging holding time

97 © 2013 ISIJ
 ISIJ International, Vol. 53 (2013), No. 1

while the impact property of the weld metal without Ti does without Ti, as shown in Fig. 9(b). Impact toughness of weld
not change obviously, as shown in Fig. 5. When the holding metal with fine pearlite will be better than that of weld metal
time is 5 h, the impact energy of the two different weld met- with coarse pearlite.
als becomes closer. The Ti-bearing weld metal is much more
sensitive to the normalizing treatment. 3.4. Effect of Normalizing Heat Treatment on the Ti-
Bearing Precipitates
3.3. The Effect of Normalizing Processes on Micro- The inclusions in weld metal have an important effect on
structure the mechanical properties and it is necessary to investigate
When the normalizing temperature is 900°C and 1 000°C,
the microstructure is polygonal ferrite and pearlite and when
the normalizing temperature is 1100°C, the pearlite (pearlite
and degenerated-pearlite (DP)) become coarser, as shown in
Fig. 6. The size of pearlite increases against increasing tem-
perature, as shown in Figs. 6(b) and 6(h). The SEM micro-
graphs of weld metal after different normalizing treatments are
shown in Fig. 7. The pearlite area at higher normalizing tem-
perature has become coarser, as shown in Figs.7(b) and 7(d).
After different normalizing holding time, microstructure
of weld metal without Ti and Ti-bearing weld metal has no
obvious difference, as shown in Fig. 8. With prolonging
holding time, the pearlite becomes coarser, as shown in
Figs. 8(b) and 8(d). The statistical result of pearlite size after
normalizing is shown in Figs. 9(a) and 9(b). The number of
pearlite in Ti-bearing weld metal less than 15 μ m signifi- Fig. 4. Effect of normalizing temperature on impact property of
weld metal.
cantly decreases and that of the pearlite in the size 15–35
μ m increases as the normalizing temperature increasing
from 900°C to 1 200°C. When normalized at 900°C, the
number of fine pearlite in the Ti-bearing weld metal is more
than that in the weld metal without Ti, as shown in Fig. 9(a).
When normalized at 1 200°C, the fine pearlite in the Ti-
bearing weld metal is much less than that in the weld metal

Fig. 2. The macrograph of the weld metal as in weld condition. Fig. 5. Effect of holding time on impact property of weld metal.

Fig. 3. Microstructures of the weld metal in as weld condition: Without Ti: a) Columnar grain zone c) Fine grain zone; Ti-
bearing: b) Columnar grain zone d) Fine grain zone.

© 2013 ISIJ 98
 ISIJ International, Vol. 53 (2013), No. 1
the Ti-bearing precipitates after normalizing treatment. The
morphology of the Ti-bearing precipitates has changed both
by prolonging the holding time and increasing the normal-
izing temperature, as shown in Fig. 10. Figures 10(a) and
10(b) shows the effect of holding time on the MnTiO pre-
cipitate on the Mn–Si oxides at 900°C, and Figs. 10(c) and
10(d) show the effect of normalizing temperature on the
MnTiO precipitate on the Mn–Si oxides. The White parti-
cles on black substrate are MnTiO precipitates and the black
substrate is the Mn–Si oxides. Figures 10(e) and 10(f) are
the EDS analysis to the white particle and black substrate.
Fig. 6. Effect of normalizing temperature on optical microstructure
of weld metal; (a) 900°C, without Ti; (b) 900°C, Ti-bear-
ing; (c) 1 000°C, without Ti; (d) 1 000°C, Ti-bearing; (e)
1 100°C, without Ti; (f) 1 100°C, Ti-bearing; (g) 1 200°C,
without Ti; (h) 1 200°C, Ti-bearing.
Fig. 9. Comparison of pearlite size: (a) 900°C; (b) 1 200°C.
99
The white particle is the MnTiO precipitate as shown in Fig.
10(e), and the black substrate is the Mn–Si oxide as shown
in Fig. 10(f). The Ti-bearing precipitates on the Mn–Si
oxides disappears after normalizing heat treatment both by
prolonging holding time, as shown in Figs. 10(a) and 10(b),
and by increasing normalizing temperature, as shown in
Figs. 10(c) and 10(d). It means that some elements in the Ti-
bearing precipitates diffuse. The selected area diffraction
Fig. 7. SEM micrographs of weld metal at various normalizing
temperatures; 900°C, without Ti; (b) 900°C, Ti-bearing; (c)
1 200°C, without Ti; (d) 1 200°C, Ti-bearing.
Fig. 8. SEM micrographs of weld metal normalizing at 900 for dif-
ferent holding times; (a) 0.5 h, without Ti; (b) 0.5 h, Ti-
bearing; (c) 5 h, without Ti; (d) 5 h, Ti-bearing.
© 2013 ISIJ
 ISIJ International, Vol. 53 (2013), No. 1

patterns (SADPs) of the inclusions shows that the matrix of

the inclusions is amorphous, as shown in Figs. 11(a) and
11(b), and the Ti-bearing precipitates are still MnTiO3, as
shown in Fig. 12. Composition analysis by using line scan
analysis of TECNAI F30 was carried out and shows that
there exists Mn-depleted zone between the inclusions and
matrix at 900°C/1 h normalizing, as shown in Fig. 13. Pre-
cipitation picture in Fig. 13 is same as that in Fig. 10(c), nor-
malized after 900°C/1 h, and the EDS analysis for MnTiO
precipitation is shown in Fig. 10(e). The Mn-depleted zone
at higher normalizing temperature (1 200°C/1 h) disappears,
as shown in Fig. 14. Inclusing picture in Fig. 14 is same as
that in Fig. 10(d), normalized after 1 200°C/1 h, EDS anal-
ysis is shown in Fig. 10(f).

4. Discussion
As is known to all, the heat treatment is an important
method to adjust the mechanical property for material, and
the basic principal is to change the dislocation through heat

Fig. 10. Figure 10 SEM Morphology and EDS analysis of typical Fig. 13. Mn profile between the Ti-bearing inclusions and matrix
inclusions in the Ti-bearing weld metal; (a) 900/0.5 h; (b) for Ti-bearing weld metal at 900°C/1 h.
900/5 h; (c) 900/1 h; (d) 1 200/1 h; (e) EDS analysis for
point A in Fig. 10(c); (f) EDS analysis for point B in Fig.
10(d).

Fig. 11. Selected Area Diffraction Patterns (SADPs) of Mn–Si–O

matrix; (a) Bright Field image; (b) Electron Diffraction Fig. 14. Mn profile between the Ti-bearing inclusions and matrix
Pattern. for Ti-bearing weld metal at 1 200°C/1 h.

Fig. 12. SADPs of Ti-bearing precipitate; (a) Bright Field Image; (b) Electron Diffraction Pattern; (c) Electron Diffraction
Pattern.

© 2013 ISIJ 100

 ISIJ International, Vol. 53 (2013), No. 1
treatment. The strength property of the steel decreases and
the toughness increases in common. Pearlite plays an impor-
tant role on the cleavage behavior during the mechanical
property test. It has been demonstrated that initial cracking
in both smooth tensile and Charpy impact specimens depends
largely on the result of shear cracking of pearlite in ferrite-
pearlite steels.13) During the tensile test or the impact process,
the localized slip bands in ferrite promotes cracking of the
cementite plates, which is then followed by tearing of the adja-
cent ferrite laths. Cracks in adjoining cementite lamellae ini-
tially link up to form a fibrous microcrack. When the fibrous
crack becomes large enough, it acts as a Griffith type crack in
initiating unstable cleavage fracture. Therefore, the size and
distribution of pearlite plays an important role on the impact
property for high strength low alloy steel.
The normalizing heat treatment can change the morphol-
ogy of the inclusions, as shown in Fig. 10, and the TEM
experiment has shown that the Mn–Ti–O precipitate can dis-
solve both by prolonging the holding time and increasing
the normalizing temperature, as shown in Fig. 11. H. S. Kim
et al. calculated the effect of Ti on the liquidus and solidus
temperatures of Mn–Ti–O for low carbon steels.6) As the Ti
content increased from 0 to 120 ppm in low carbon steel, the
oxide changed from Mn2SiO4 to MnTi2O4/Ti2O3 and the liq-
uidus and solidus also increased with the Ti content. The
solidus and liquidus of the MnTiO3+Mn2SiO4 is between
1 020°C to 1 200°C. In our experimental here, though the Ti
content is different, MnTiO3 and Mn–Si oxide also found in
the weld metal. As the Normalizing temperature and holding
time is high and longer, MnTiO3 precipitate is unstable and
possible melt. C. Zhang14) and H. Goto et al.15) have report-
ed that oxygen can diffuse into the matrix at higher temper-
ature and Ti has much higher diffusion coefficient compared
with oxygen in low carbon steel. It is probable that the Mn–
Ti–O precipitate will melt and dissolve by Ti and O diffu-
sion with increasing normalizing temperature and prolong-
ing holding time, and then the Mn-depleted zone disappear.
The detailed calculation and experiment research on the
MnTiO3 dissolving behavior need to be investigated further.
The Mn-depleted zone has an important effect on the trans-
formation of austenite to ferrite. The Mn-depleted zone forms
due to absorbing Mn by Ti2O310,16) or MnS precipitated on the
Mn–Si oxides.17,18) In the present study, the Mn–depleted zone
forms due to the MnTiO3 precipitated on the Mn–Si oxides, as
shown in Fig. 13, which agree well with the Ref. 19.
From the analysis and results above, it shows that the
change of Ti-bearing precipitate has an important effect on
the phase transformation of weld metal. During the air cool-
ing for the Ti-bearing weld metal, the austenite grain bound-
ary and the Mn-depleted zone around the Ti-bearing precip-
itate are the nucleation sites of proeutectoid ferrite and then
the interface between austenite and proeutectoid ferrite
become the nucleation site for pearlite thereafter. Mn-
depleted zone formation increases the pearlite nucleation
sites. Therefore, the pearlite size in the Ti-bearing weld met-
al under the same heat treatment is much smaller, comparing
to the weld metal without Ti at relatively low normalizing
temperature and short holding time. With the Ti-bearing pre-
cipitates dissolving with the too strong the normalizing pro-
101
cesses (higher temperature and longer holding time), the Mn-
depleted zone disappears, as shown in Fig. 14, and the nucle-
ation sites of proeutectoid ferrite and pearlite thereafter
decrease. Therefore, the pearlite will become much coarser. As
the pearlite is fine and distributed uniformly in the matrix, the
impact toughness of weld metal will increase. In the present
study, the pearlite size become coarser while prolonging hold-
ing time and increasing normalizing temperature, and the
impact toughness of Ti-bearing weld metal deteriorates, as
shown in Figs. 4 and 5. For the Ti-bearing weld metal, the
main factor affecting the impact toughness is the Ti-bearing
precipitate, while for the weld metal without Ti, there is no
change of the inclusions, and the impact toughness of weld
metal has no obvious change, as shown in Figs. 4 and 5.
5. Conclusions
(1) With the prolonging holding time and increasing nor-
malizing temperature, the impact toughness of Ti-bearing
weld metal decreases significantly, and the impact toughness
of weld metal without Ti does not change obviously.
(2) With prolonging holding time and increasing nor-
malizing temperature, Ti-bearing precipitate will dissolve,
and the Mn-depletion zone will disappear, which is the main
factor for the deterioration of the impact toughness of Ti-
bearing weld metal.
Acknowledgements
The authors are grateful for the financial support from the
National Key Technologies R&D Program of China under
grant No.2009BAG12A07-D06.
The authors are grateful for the help from C. F. Li (Division
of Microelectronic Interconnect Materials, IMR), W. Y. Wang
and X. M. Dong (Solid Atomic Imaging Division, IMR) in the
TEM experiment.
REFERENCES
1) K. Oikawa, K. Ishida and T. Nishizawa: ISIJ Int., 37 (1997), 332.
2) D. C. Houghton: Acta Metall., 41 (1993), 2993.
3) W. W. Bose-Filho, A. L. M. Carvalho and M. Strangwood: Mater.
Charact., 58 (2007), 29.
4) M. H. Avakonandeh-Gharavol, M. Haddas-Sabzevar and A. Haerian:
Mater. Design, 30 (2009), 1902.
5) X. Wang, W. C. Dong, S. P. Lu, G. Z. He, X. Zhao, D. M. Zhou and
Y. Y. Li: Acta Metall. Sin., 25 (2012), 1 (in English).
6) H. S. Kim, H. G. Lee and K. S. Oh: ISIJ Int., 42 (2002), 1404.
7) I. Rack, V. Gila and M. Kodak: Metall. Mater. Trans. A, 28 (1997),
199.
8) K. S. Bang, C. Park and S. Liu: J. Mater. Sci., 41 (2006), 5994.
9) H. R. Wang and W. Wang: J. Mater. Sci., 44 (2009), 591.
10) J. S. Byun, J. H. Shim, Y. W. Cho and D. N. Lee: Acta Mater., 51
(2003), 1593.
11) B. Beidokhti, A. H. Koukabi and A. Dolati: J. Mater. Process. Technol.,
209 (2009), 4027.
12) T. Koseki, S. Ohkita and N. Yurioka: Sci. Technol. Weld. Join., 2
(1997), 65.
13) Y. J. Park and I. M. Bernstein: Metall. Mater. Trans. A, 10 (1979),
1653.
14) C. Zhang and P. R. Norton: J. Nucl. Mater., 300 (2002), 7.
15) H. Goto, K. Miyazawa, K. Yamaguchi, S. Ogibayashi and K. Tanaka:
ISIJ Int., 34 (1994), 414.
16) J. H. Shim, Y. W. Cho, S. H. Chung, J. D. Shim and D. N. Lee: Acta
Mater., 47 (1999), 2751.
17) H. Mabuchi, R. Uemori and M. Fusioka: ISIJ Int., 36 (1996), 1406.
18) H. S. Kim, H. J. Lee and K. S. Oh: Met. Mater., 6 (2000), 305.
19) Z. Y. Hu, C. W. Yang, M. Jiang, G. W. Yang, W. J. Wang and X.
H. Wang: Acta Metall. Sin., 47 (2011), 971.
© 2013 ISIJ