Kovove Mater.

55 2017 51–56 51
DOI: 10.4149/km 2017 1 51

The development of a banded microstructure in S355J2 steel bar

A. Nagode1 *, A. Resnik2 , R. Vertnik2 , M. Bizjak1 , B. Kosec1 , M. Gojić3 , G. Kosec4 , B. Šarler5 ,

B. Zorc1,6

1
University of Ljubljana, Faculty of Natural Sciences and Engineering, Aškerčeva cesta 12, SI – 1000 Ljubljana, Slovenia
2
Štore Steel d.o.o., Štore, Slovenia
3
University of Zagreb, Faculty of Metallurgy, Sisak, Croatia
4
Acroni d.o.o., Jesenice, Slovenia
5
University of Nova Gorica, Laboratory for Multiphase Processes, Nova Gorica, Slovenia
6
Institut za varilstvo d.o.o., Ljubljana, Slovenia

Received 13 February 2015, received in revised form 12 March 2015, accepted 8 February 2016

Abstract
A homogeneous microstructure plays an important role in the mechanical properties of
every alloy. However, because of chemical inhomogeneities, a banded microstructure can de-
velop in steels. This banding has a negative effect, especially on the impact toughness and
cold forming. In this paper, the development of a banded microstructure in hot-rolled low-alloy
structural steel S355J2 is explained. The banded microstructure in the cross-section consisted
of ferrite and pearlite bands, while towards the centre, bainitic-martensitic bands were also ob-
served. The deviations in chemical compositions of the bands were measured with EDS, which
showed that the bainitic-martensitic bands contained more alloying elements (Mn, Cr, Mo, Si)
than the ferritic. However, by using Oberhoffer reagent, the segregations of phosphorus were
also revealed. Phosphorus segregations coincided with the positive segregations of the alloying
elements. Because manganese and chromium have the prevailing effect on the distribution
of carbon in austenite, an increased concentration of carbon in bainitic-martensitic bands is
also expected. The CCT diagrams of S355J2 steel with three different chemical compositions,
i.e., with the nominal composition and compositions with negative and positive segregations,
were calculated. However, increased concentrations of carbon and phosphorus were not taken
into account in the calculations of the CCT diagrams. Since these elements can significantly
increase the hardenability, they play an important role in the formation of the bainite and
martensite.

K e y w o r d s : S355J2 steel, banded microstructure, phosphorus segregations, Oberhoffer

reagent, CCT diagrams

1. Introduction not necessarily lead to the development of microstruc-

tural banding; its occurrence also strongly depends
The mechanical properties of steels are dependent on the austenite crystal grain size and, especially, on
on their microstructure and the associated defects. the cooling conditions from the austenite region. A
The chemical inhomogeneities in the microstructure banded microstructure can also be observed in steels
due to the inter-dendritic (micro) segregations of the with a high content of manganese sulphide inclusions
alloying elements, as well as carbon and phosphorus, [1–4]. Elongated inclusions and a banded microstruc-
can cause the development of a banded microstruc- ture have a negative effect on the impact properties
ture. This form of microstructure consists of alter- in both the longitudinal and transverse directions. If
nating bands of different microstructural constituents the banded microstructure also consists of martensite
aligned parallel to the rolling direction of steel prod- bands, it is detrimental to any machining and subse-
ucts. The presence of inter-dendritic segregations does quent cold-forming operations [2, 5].
*Corresponding author: tel: +386 1 2000 433; fax: +386 1 470 45 60; e-mail address: [email protected]
52 A. Nagode et al. / Kovove Mater. 55 2017 51–56

T a b l e 1. Nominal chemical composition of the S355J2 steel (wt.%)

C Si Mn P S Cr Mo Ni Al Cu Nb B Fe

0.18 0.28 1.38 0.012 0.013 0.13 0.03 0.11 0.026 0.17 0.001 0.0006 Bal.

Besides banding, the phosphorus in steel can cause with a particle size of 3 µm and then with a particle
quench cracking and quench embrittlement and thus, size of 1 µm. After the polishing, the samples were
it is very harmful, especially when is not homoge- etched with 2 % Nital. However, in order to reveal
neously distributed [6]. Since the overall concentration the phosphorus segregation, the samples needed to be
of phosphorus in steel is usually very low, the mea- etched with Oberhoffer reagent (500 ml of distilled wa-
surement of phosphorus inhomogeneities in the mi- ter, 500 ml C2 H5 OH, 50 ml HCl, 30 g FeCl3 , 1 g CuCl2 ,
crostructure is challenging because of low detection 0.5 g SnCl2 ) [7].
limits of conventional analytical techniques based on The microstructure analyses were performed with
X-ray spectroscopy (EDS, WDS). Thus, phosphorus a light (LM) and scanning electron microscope (SEM),
segregation can be easily overlooked. Hence, it is es- i.e., Olympus BX61 and Jeol JSM 5610, respectively.
sential to understand the primary reasons for the de- The microstructure was observed in secondary elec-
velopment of microstructural banding in order to take tron mode (SEI) with an accelerating voltage of 20
appropriate actions so as to minimise its occurrence kV and a working distance of 20 mm. The same con-
and deleterious effects, especially with respect to cold- ditions were used with the EDS system Gresham Sci-
forming processes and quenching. entific Instruments Ltd., Model No. Sirius 10/SUTW.

2.4. Microhardness measurements

2. Experimental procedure
The microhardness of the microstructure con-
2.1. Material stituents was measured with a Vickers microhardness
tester Shimadzu No. 4960 using loads of 100 g and a
The banding microstructure was investigated in a loading time of 10 s. The microhardness was measured
low-alloy structural steel S355J2. This steel, with a from the edge of the samples towards the centre.
nominal chemical composition listed in Table 1, has
a minimum yield strength of 355 N mm−2 and is used 2.5. Calculation of CCT diagrams
mainly for the production of cold-formed sections. The
investigated steel was supplied as a hot-rolled bar with For the calculation of the CCT diagrams, a com-
a diameter of 65 mm. The rolling process under indus- puter programme JMatPro 7.0 [8] was used. Using this
trial conditions began at 1200 ◦C and ended at 875 ◦C. programme, CCT diagrams for S355J2 steels with dif-
ferent chemical compositions were calculated. In the
2.2. Hot-rolling and temperature calculation, a crystal grain size of 9, in accordance
measurements with the ASTM standard, was taken into account.

The temperature of the steel bar after hot rolling

was measured with a digital, two-colour pyrometer 3. Results
M90R-1 and a thermal camera FLIR ThermoVision
SC6000HS with the software package ThermaCAM 3.1. Temperature measurements
Research Pro2.8 SR-3 in order to obtain the cooling
curve. The pyrometer was used for measuring the tem- The results of the temperature measurements of
perature after hot rolling down to 700 ◦C. The tempe- the hot-rolled steel bar are expressed with a cooling
ratures below 700 ◦C were measured with a thermal curve (Fig. 1). From that curve, a transformation tem-
camera. perature Ar1 (690 ◦C) is clearly seen. This indicates the
high precision of the measurements.
2.3. Microstructure characterisation
3.2. Microstructure analysis
The samples for the microstructure characterisa-
tion were cut from the hot-rolled steel bar in such a The microstructure analysis of the sample etched
way that the normal to the sample surface was par- with 2 % Nital performed near the surface with a
allel to the axis of the bar. The samples were met- light microscope revealed a banded microstructure. It
allographically prepared, i.e., ground with abrasive consisted of alternating bands of ferrite (white) and
paper to 1200 grit and then diamond polished, first pearlite (dark), aligned parallel to the rolling direc-
 A. Nagode et al. / Kovove Mater. 55 2017 51–56 53

Fig. 1. The cooling curve of a hot-rolled bar obtained with

a pyrometer and a thermal camera.

Fig. 3. Phosphorus segregations (a) etched with Oberhoffer

reagent (bright bands), (b) etched with Oberhoffer (upper
part) and 2 % Nital (bottom part).

sions (marked with arrows). The microstructure ob-

servation showed that the MnS inclusions were mostly
present in the bainitic-martensitic bands.
The microstructure of the sample etched with
Oberhoffer reagent is shown in Fig. 3a. This reagent
revealed the phosphorus segregations (bright areas).
It is clear that the phosphorus is in the microstruc-
ture in alternating bands, parallel to the rolling direc-
tion as well. Figure 3b shows the microstructure of the
sample after etching with 2 % Nital (bottom part of
the sample) and Oberhoffer reagent (upper part of the
sample). It is clear that the phosphorus segregations
Fig. 2. Banded microstructure (a) near the surface, (b) in coincide with the bainitic-martensitic bands. Again,
the centre of the hot-rolled bar (inclusions are marked with it was confirmed that the manganese sulphide inclu-
arrows). sions were present in the bainitic-martensitic bands,
where phosphorus was also detected by the Oberhof-
fer reagent.
tion of the bar (Fig. 2a). However, towards the cen- The average results of the EDS analyses in ferritic
tre, bainitic-martensitic bands also appeared in the and bainitic-martensitic bands (Fig. 4) are presented
microstructure (Fig. 2b). The proportion of bainitic- in Table 2. They indicate that in the ferrite the con-
martensitic bands increases with the distance from tents of silicon, chromium, manganese, and molybde-
the edge of the sample towards the centre. Figure num were lower (negative segregations) in comparison
2b also shows elongated manganese sulphide inclu- to the results of the nominal chemical composition of
54 A. Nagode et al. / Kovove Mater. 55 2017 51–56

T a b l e 2. Results of EDS analysis (wt.%)

Site of interest Si Cr Mn Mo Fe

1 0.20 0.12 1.0 0.05 Bal.

2 0.38 0.18 1.6 0.19 Bal.

Fig. 4. SEM micrograph of banded microstructure marked

with analysed regions.

T a b l e 3. Results of Vickers microhardness measurements

Distance (mm) 0.5 5 10 15 20 25 30

HV0.1 169.8 160.4 167.7 176.7 162.6 219.2 145.9

Fig. 5. Calculated CCT diagrams for steel with (a) the

nominal chemical composition, (b) with the composition
the investigated steel listed in Table 1, while in the in the negative segregations, and (c) with the composition
bainitic-martensitic bands, the contents of these ele- in the positive segregations.
ments were higher (positive segregations). Note that
the concentration of phosphorus is below the detec-
tion limit of the EDS analysis and, therefore, it was sured in the negative segregations (Fig. 5b) and posi-
not taken into account. tive segregations (Fig. 5c), respectively, were also cal-
culated. The chemical compositions of the steel in
3.3. Microhardness the negative and positive segregations were measured
with EDS in the ferritic (negative segregations) and
The results of the microhardness measurements bainitic-martensitic bands (positive segregations), re-
from the edge of the bar towards the centre are pre- spectively.
sented in Table 3. The lowest hardness value, mea-
sured mostly in the ferritic bands, was 138.5 HV, while
the maximum hardness was 293.0 HV, in the bainitic- 4. Discussion
-martensitic band.
A banded microstructure is a common phenome-
3.4. CCT diagrams non in rolled-steel products. S355J2 steel is a low-
-alloy structural steel that is commonly supplied in
Figure 5a shows the CCT diagram of the steel hot-rolled bars. Since microstructure banding may
with the nominal chemical composition given in Ta- have a negative effect on the impact properties and
ble 1. However, for comparison, CCT diagrams for subsequent cold-forming operations, it is necessary
the same steel with the chemical composition mea- to minimise it.
 A. Nagode et al. / Kovove Mater. 55 2017 51–56 55

– The microstructure analysis of the metallo- tive segregations of carbon, phosphorus, and the al-
graphically prepared samples after etching in 2 % loying elements will transform into ferrite during the
Nital showed a banded microstructure, which near cooling from the austenite region, while the austen-
the surface consisted of ferritic and pearlitic bands ite regions with positive segregations will later trans-
(Fig. 2a). However, towards the centre, bainitic- form into pearlite. In this way, a secondary, banded
-martensitic bands also appeared (Fig. 2b). The por- microstructure of ferrite and pearlite bands is devel-
tion and the width of these bainitic-martensitic bands oped.
increased towards the centre of the bar. The EDS anal- – Additionally, the alloying elements (Mn, Cr, Mo,
yses showed higher concentrations of the alloying ele- Si) have a strong effect on the hardenability since
ments (Mn, Cr, Mo, and Si) in the bainitic-martensitic they move the CCT diagram to the right. Moreover,
bands (positive segregations) in comparison to the fer- it is also known that phosphorus also significantly en-
ritic and pearlitic bands (negative segregations). How- hances the hardenability [6, 11], and can cause both
ever, an EDS analysis is not a proper method for the quench cracking and quench embrittlement [6]. Since
measurement of light elements (e.g., carbon) and el- the concentration of alloying elements as well as phos-
ements with a concentration below 0.1 % (e.g., phos- phorus and carbon increases towards the centre of the
phorus); anyway, it is well known that manganese ingots, the proportion of bainitic-martensitic bands is
lowers the activity of carbon in austenite and thus, higher towards the centre of the hot-rolled bar. More-
Mn-rich regions would tend to attract carbon from over, the concentration of alloying elements and phos-
neighbouring areas. Hence, in the Mn-rich regions, phorus in some austenite regions was high enough
an increased concentration of carbon is also expected. for a critical cooling rate to be achieved and, con-
While chromium has a similar effect to that of man- sequently, martensite, as well as bainite, was formed
ganese, phosphorus increases the carbon activity, and despite lower cooling rate in the centre of steel bar.
as a result rejects the carbon from the phosphorus-rich – The formation of bainitic-martensitic bands can
regions [9]. be explained with the calculated CCT diagrams: a
– Phosphorus has a very low equilibrium portion CCT diagram with the nominal steel composition
ratio and so it has a strong tendency to segregate dur- given in Table 1; a CCT diagram with the compo-
ing the solidification [9]. Since the concentration of sition measured in the negative segregations with a
phosphorus is lower than the detection limit of an EDS lower concentration of Mn, Cr, Si, and Mo (Table 2,
analysis, the phosphorus segregations were revealed Site of Interest 1); and a CCT diagram with a compo-
metallographically by using Oberhoffer reagent. From sition measured in the positive segregations with an
Karl [10], the lower limit of phosphorus detection us- increased concentration of Mn, Cr, Si, and Mo (Ta-
ing Oberhoffer reagent is 0.003 %. The microstruc- ble 1, Site of Interest 2). The increased concentration
ture analysis after etching with Oberhoffer reagent of alloying elements retards the diffusion-controlled
showed the phosphorus-rich regions as being brighter transformation, and therefore the ferritic and pearlitic
(Fig. 3). It was found that the phosphorus segregations regions move to a higher timing in comparison to
coincided with positive segregations of manganese, the CCT diagram of steel with the nominal compo-
chromium and molybdenum. Moreover, the elongated sition and even more compared to the CCT diagram
manganese sulphide inclusions were also observed in of the steel with the composition of negative segrega-
bainitic-martensitic bands with higher concentrations tions. In addition, the martensite start temperature
of manganese and phosphorus. This indicates that al- (Ms ) is lowered with an increased concentration of al-
though the manganese concentrates in the MnS in- loying elements. Hence, at a normal cooling rate of
clusions the austenite around the inclusions still con- approximately 0.4 ◦C min−1 (bold line in Fig. 5) in
tains enough alloying elements to retard the diffusion- regions with increased concentration of alloying ele-
-controlled transformation and enable the formation ments (positive segregations), some bainite in addi-
of bainitic-martensitic bands. tion to the ferrite is also expected (Fig. 5c), while in
– Regarding the activity of carbon, manganese the regions with a lower concentration of alloying el-
and phosphorus have the opposite effect. Because the ements (negative segregations) only ferrite and some
concentrations of manganese and chromium in steels pearlite can form (Fig. 5b). However, in this calcula-
are much higher than phosphorus, manganese and tion, the changes in the concentration of carbon and
chromium have the prevailing effect on the distribu- phosphorus were not taken into account since they
tion of carbon in austenite. Therefore, it is assumed could not have been analysed with EDS. It is known
that regions that are rich in alloying elements (Mn, that both elements significantly increase the harden-
Cr, Mo) and phosphorus are also rich in carbon, de- ability of steel since they move the CCT diagram fur-
spite the fact that phosphorus increases the carbon ther to the right, to a higher timing, and as such en-
activity. On the other hand, the austenite regions de- able the formation of martensite and, therefore, the
pleted in alloying elements and phosphorus contain formation of bainitic-martensitic bands at the same,
less carbon. Hence, the austenite region with nega- and with an even slower, cooling rate in the centre of
56 A. Nagode et al. / Kovove Mater. 55 2017 51–56

a higher concentration of alloying elements (Mn, Cr,

Mo, and Si) in the bainitic-martensitic bands in com-
parison to the ferritic bands. However, the EDS anal-
yses did not show the presence of phosphorus be-
cause the quantity was below the detection limit of
our EDS detector. Phosphorus can be very detrimen-
tal since it can be responsible for banding and it can
also cause both quench cracking and embrittlement.
The etching with Oberhoffer reagent revealed bands
of phosphorus segregations that coincided with posi-
tive segregations of the alloying elements. The calcu-
lated CCT diagrams of steel with chemical composi-
tions of positive segregations did not provide evidence
for martensitic transformations at a normal cooling
rate. Therefore, the formation of martensite can be at-
Fig. 6. Calculated CCT diagram of S355J2 steel with mea- tributed to the increased concentration of carbon and
sured and extrapolated cooling curves. phosphorus, which were not taken into account in the
calculation of the CCT diagrams. Carbon and phos-
phorus retard the diffusion-controlled transformation
the steel bar. This shows again the importance of re- and thus significantly increase the hardenability of the
vealing the phosphorus segregations since phosphorus steel. A real cooling curve put in the CCT diagram
can cause alternating bands of different microstruc- of S355J2 steel with a homogeneous chemical com-
tural constituents with different mechanical properties position showed that the microstructure would con-
(Table 3). sist of ferrite, pearlite, and a small amount of bainite.
– Taking into account the real cooling rate deter- The transformation temperatures of the austenite into
mined by a temperature measurement of the steel bar ferrite, pearlite, and bainite would be 775, 650, and
with a pyrometer and a thermal camera after hot- 570 ◦C, respectively.
rolling processes, a microstructure of steel with ho-
mogeneous chemical compositions can be predicted.
Namely, the measured cooling rate, which was extrap-
References
olated to lower temperatures, was used as an input for
the JMatPro software, and a CCT diagram of the steel
[1] Zorc, B., Bernetič, J., Nagode, A.: Eng. Fail. Anal.,
with nominal chemical compositions was calculated 40, 2014. p. 8. doi:10.1016/j.engfailanal.2014.02.014
(Fig. 6). The bold cooling curve in the CCT diagram [2] Krauss, G.: STEELS: Processing, Structure and Per-
shows that mainly ferrite, pearlite, and very small formance. Ohio, ASM International 2005.
amount of bainite were expected in the microstruc- [3] Majka, T. F., Matlock, D. K., Krauss, G.: Metall.
ture if the chemical composition of the steel was to Mater. Trans. A, 33, 2002, p. 1627.
be homogeneous. The transformation temperatures of doi:10.1007/s11661-002-0172-8
the austenite into ferrite, pearlite, and bainite were [4] Krebs, B., Germain, L., Gouné, M., Hazotte, A.:
IJMR, 102, 2011, p. 200. doi:10.3139/146.110467
also determined and were found to be 775, 650, and
[5] Yadegari, S., Turteltaub, S., Suiker, A. S. J., Kok,
570 ◦C, respectively. P. J. J.: Comp. Mater. Sci., 84, 2014, p. 339.
doi:10.1016/j.commatsci.2013.12.002
[6] Roberts, G., Krauss, G., Kennedy, R.: Tool Steels. 5th
5. Conclusions Edition. Ohio, ASM International 1998.
[7] Schumann, H.: Metallographie. Leipzig, VEB DVG
In this paper, we explain the development of 1987.
the banded microstructure in hot-rolled S355J2 steel. [8] JMatPro 7.0. Occam Road, United Kingdom, Sente
Software Ltd.
Aligned bands of ferrite and pearlite were observed;
[9] Krauss, G.: Metall. Mater. Trans. B, 34, 2003, p. 781.
however, towards the centre of the steel bar bainitic- doi:10.1007/s11663-003-0084-z
-martensitic bands were also present. Their portion [10] Karl, A.: Prakt. Metallogr., 15, 1978, p. 469.
and width increased when moving from the edge of [11] Jacobs, F. A., Kraus, G.: J. Heat Treating, 2, 1981, p.
the bar towards the centre. The EDS analyses showed 139. doi:10.1007/BF02833230