Journal of Materials Processing Technology 216 (2015) 348–356

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

journal homepage: www.elsevier.com/locate/jmatprotec

The effect of a non-circular drawing sequence on spheroidization of

medium carbon steel wires
Ho Seon Joo a , Sun Kwang Hwang b , Hyun Moo Baek a , Yong-Taek Im a,c,∗ , Il-Heon Son d ,
Chul Min Bae d
a
National Research Laboratory for Computer Aided Materials Processing, Department of Mechanical Engineering, KAIST, 291 Daehak-ro, Yuseong-gu,
Daejeon 305-701, Republic of Korea
b
Ultimate Manufacturing Technology R&BD Group, Daegyeong Regional Division, KITECH, 320 Techno sunhwan-ro, Yuga-myeon, Dalseong-gun,
Daegu 711-883, Republic of Korea
c
Office of the President, KIMM, 156 Gajeongbuk-Ro, Yuseong-Gu, Daejeon 305-343, Republic of Korea
d
Wire Rod Research Group, Technical Research Laboratories, POSCO, 1 Goedong-dong, Nam-gu, Pohang, Gyeongbuk 790-785, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: A non-circular drawing (NCD) sequence was applied as a prior cold working to improve spheroidization in
Received 12 June 2014 a medium carbon steel wire. To investigate the effect of the NCD sequence on spheroidization behavior
Received in revised form of the medium carbon steel wire, numerical simulations and experiments were carried out up to the
28 September 2014
fourth pass and were compared with the results of the conventional wire drawing (WD). From the three-
Accepted 4 October 2014
dimensional numerical results, it was found that the NCD sequence could impose higher and relatively
Available online 12 October 2014
more homogeneous plastic deformation on the wire than the WD process. Spheroidization annealing
at subcritical temperature of 705 ◦ C for 2, 4, and 8 h was applied for the specimens that were produced
Keywords:
Non-circular drawing sequence
by the WD and NCD processes, respectively. Tension and Vickers micro-hardness tests were conducted
Spheroidization to measure mechanical properties of the annealed specimens. The annealed specimen processed by the
Subcritical annealing NCD sequence showed the reduced ultimate tensile strength and micro-hardness, and the enhanced
Medium carbon steel wire reduction of area compared to those values for the WD process. According to upsetting test results, it was
Cold workability found that cold workability was also improved for the annealed specimen processed by the NCD process
Scanning electron microscopy compared to the one produced by the WD process. Scanning electron microscopy (SEM) was employed
to investigate the microstructure change and to quantitatively measure the spheroidization ratio of the
annealed specimens processed by both processes. Owing to the SEM results, the spheroidization of the
annealed specimens processed by the NCD sequence was improved compared to the one produced by the
WD process. It can be construed from the current investigation that the NCD sequence might be helpful
in improving the spheroidization of cementite and cold workability of medium carbon steel wires during
the subcritical annealing.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction
Medium carbon steel wires are widely used for manufacturing
mechanical parts such as bolts, nuts, rivets, and screws. Since
these parts are produced into various shapes by applying cold
forging processes, sufficient workability should be guaranteed
to impose large plastic deformation on the material during the
∗ Corresponding author at: National Research Laboratory for Computer Aided
Materials Processing, Department of Mechanical Engineering, KAIST, 291 Daehak-
ro, Yuseong-gu, Daejeon 305-701, Republic of Korea. Tel.: +82 42 350 3227;
fax: +82 42 350 3210.
E-mail addresses: ytim@kimm.re.kr, ytim@kaist.ac.kr (Y.-T. Im).
forging processes. In addition, the service life of cold heading dies
is an important issue from the aspect of manufacturing cost (Lee
et al., 2009). In this regard, spheroidization annealing is usually
applied to the medium carbon steel wires prior to the cold head-
ing. During the spheroidization annealing, the cementite phase
in the steel wire is changed into a spherical morphology. Such
spheroidized microstructure is known to improve cold workability
of the material, resulting in reduction of the forming load during
the manufacturing process. However, commercial spheroidiza-
tion annealing usually consumes a large amount of time and
energy. Therefore, various attempts were made to accelerate
spheroidization behavior of medium carbon steel wires owing to
the microstructure, prior plastic deformation, and heat treatment.
The spheroidization rate of the material is influenced by the ini-
tial microstructure prior to spheroidization annealing. According

http://dx.doi.org/10.1016/j.jmatprotec.2014.10.001
0924-0136/© 2014 Elsevier B.V. All rights reserved.
 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
to the work by Chattopadhyay and Sellars (1977), spheroidiza-
tion occurred more quickly in the fine pearlite compared to
the coarse one. Kamyabi-Gol and Sheikh-Amiri (2010) investi-
gated the influence of various initial microstructures on the rate
of spheroidization. The rate of spheroidization in a martensitic
microstructure was the highest among bainite, fine pearlite, and
coarse pearlite for the same spheroidization annealing time. The
unstable microstructure of martensite had a high tendency to disso-
lute in high temperature and formed a more stable microstructure
like ferrite and cementite spheres.
To improve spheroidization of the material by applying prior
plastic deformation in the material, a number of studies also have
been conducted. These studies were tried to obtain the improved
spheroidization microstructure by increasing the amount of prior
plastic deformation at the elevated and room temperatures. In the
case of warm and hot forming, Zhang et al. (2006) reported that the
improved spheroidization could be obtained for the large reduction
in height of the hot compression test. Nam and Lee (1987) found
that the spheroidization of cementite lamellae in high-carbon steel
wires could be achieved in short time by applying wire draw-
ing (WD) at elevated temperatures below the A1 temperature of
727 ◦ C. According to the work by Shin et al. (2003), equal chan-
nel angular pressing could improve the kinetics of spheroidization
significantly. The enhanced kinetics was discussed in terms of the
carbon dissolution from the cementite and the defect induced in
the cementite by the severe plastic deformation (SPD). For cold
forming, Guo et al. (2012) reported that the large deformation
accelerated break-ups of lamellar cementite and the formation of
spheroidized cementite particles. Hwang et al. (2011) numerically
and experimentally investigated that a shear drawing (SD) process
could impose large plastic deformation on the material at room
temperature. Ko et al. (2010) also reported the SD process could
accelerate spheroidization of the wire triggered by a large amount
of plastic deformation compared to the WD process.
To improve spheroidization behavior, various heat treatment
methods were also investigated. Lv et al. (2013) reported that cyclic
heat treatment method could effectively accelerate spheroidization
of cementite without introducing prior large plastic deforma-
tion. However, heat treatment might be complex and requires
large energy consumption for industrial applications. O’Brien and
Hosford (2002) reported that the subcritical annealing (SA) could
require shorter heat treatment time than the intercritical anneal-
ing for achieving higher workability of the medium carbon steel.
According to the work by Baranova and Sukhomlin (1981), the
spheroidized microstructure was easily observed in the material
processed by cold working prior to the SA. Therefore, it can be con-
strued that the SA might be more effective than the intercritical
annealing with the help of prior plastic deformation during the
process to accelerate spheroidization of the medium carbon steel
wires.
However, it is not simple enough to practically apply the
above-mentioned processes in industry because of heat treatment
or additional processing facilities required. Thus, an alterna-
tive simple process needs to be developed for improvement
of spheroidization behavior for industrial applications. In this
aspect, Hwang et al. (2014) recently proposed a non-circular
drawing (NCD) sequence to improve mechanical property of low
carbon steel wire in a simple manner. They reported that the NCD
sequence could impose large plastic deformation on the material
more effectively. In addtion, Lee et al. (2014) found that the multi-
pass NCD sequence could manufacture high-strength wires with
relatively improved ductility compared to the conventional WD.
However, their investigations were limited only to the low carbon
steel wires. Therefore, it is necessary to investigate spheroidization
behavior of the medium carbon steel wire processed by the NCD
sequence, especially to save the processing time and energy during
349
the spheroidization annealing required to achieve the sufficient
workability of the material for various industrial applications.
In the present study, the NCD sequence was applied to inves-
tigate spheroidization behavior of the medium carbon steel wire
in comparison with the conventional WD. For this purpose, three-
dimensional finite element (FE) analysis was carried out using an
in-house FE analysis program, CAMPform3D, developed by Kim and
Im (2002) to check the level of plastic deformation incurred by
the NCD sequence in the steel wire. To investigate the effects of
prior cold working, spheroidization annealing at subcritical tem-
perature of 705 ◦ C for 2 (SA-2), 4 (SA-4), and 8 (SA-8) h was applied
to the specimens obtained after the fourth pass WD and NCD pro-
cesses, respectively. Mechanical properties were measured by the
tension and Vickers micro-hardness tests. Upsetting test was also
conducted to evaluate cold workability of the processed wires. For
the microstructural investigation, scanning electron microscopy
(SEM) was employed and spheroidization ratio (SR) was calculated
from the SEM micrographs. As a result, the empirical relationship
between the SR and ultimate tensile strength (UTS) was determined
in the present study.
2. Numerical method
Three-dimensional FE analysis was conducted to estimate the
amount of plastic deformation in the wire imposed by the WD
and NCD processes before experiments. The in-house FE program,
CAMPform3D, based on a rigid-viscoplastic constitutive model
introduced by Lee and Kobayashi (1973) and three-dimensional
auto remeshing scheme introduced by Kwak and Im (2002), was
applied to simulate the multi-pass of the WD and NCD, respectively.
Fig. 1 shows the schematic diagram of the multi-pass WD and
NCD processes, respectively, and die geometries used in the present
investigation. Three-dimensional numerical simulations were car-
ried out up to the fourth pass with the same semi-die angle of 6◦ and
20% area reduction per each pass for both processes. For the NCD
sequence, the oval die with the length ratio of the major to minor
axis of 1.33 was used for every odd pass similar to the previous
work by Hwang et al. (2014). The dies and specimen for the three-
dimensional numerical simulations were modeled by one-quarter
due to geometrical symmetry of the NCD sequence.
In numerical simulations, the dies were considered to be rigid
and the specimen was pulled along the drawing direction (z-axis in
Fig. 1) using sticking condition between the grip and the specimen.
Fig. 1. (a) Schematic diagram of the multi-pass WD and NCD processes and (b)
sketch of the dies for the WD and NCD, respectively.
350 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
Fig. 2. Fitted flow stress–strain curve obtained from the compression test.
To apply a frictional force at the interface between the drawing
die and the specimen, the constant shear friction model with a
shear friction factor of mf = 0.1 was used in simulations based on
the work by Baek et al. (2012). A constant drawing velocity used in
the simulation was 10 mm/s.
Flow stress of the material was obtained from the compression
test according to the ASTM standard (E09-09, 2009). The tests were
conducted twice to guarantee reproducibility of the test results by
Fig. 3. (a) Effective strain distribution of the drawn wires at the cross-section and (b) shear strain of the drawn wires by WD and NCD processes for the fourth pass, respectively.
employing Instron 5583 (UTM) with a constant crosshead speed
of 0.017 mm/s. The diameter and height of the compression spec-
imens were 4 mm and 6 mm, respectively. Commercially available
grease was used as lubricant. The obtained stress–strain curve was
used in numerical simulations in a table format that was interpo-
lated by applying linear fitting between each point, as shown in
Fig. 2.
Fig. 3(a) shows the effective strain distributions of the speci-
men produced by the WD and NCD processes, respectively. The
numerical results show that the NCD sequence could impose higher
and relatively more homogeneous effective strain on the specimen
compared to the one produced by the WD process. Average effec-
tive strain values of the specimens obtained by the WD and NCD
processes for the fourth pass were 1.074 and 1.337, respectively.
The standard deviations of effective strain values of the specimens
obtained by the WD and NCD processes were 0.080 and 0.052,
respectively.
Fig. 3(b) shows the shear strain distributions of the produced
wire by the WD and NCD processes, respectively. An average shear
strain value of the specimen obtained by the WD process was  xz
(or  yz ) = 0.196. Average strain values for the specimen produced
by the NCD sequence  xz = 0.275 and  yz = 0.305 for the major and
minor axes, respectively. The strain levels were almost the same for
the major and minor axes according to the present data. The shear
strain level at the surface region of the produced wire by the WD
process was slightly higher than the one of the NCD sequence while
the tendency was reversed at the center region of the wire. These
results indicate that the NCD sequence could impose higher plastic
 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
Table 1
Chemical composition of the medium carbon steel wire used in experiments.
Material Components (wt.%)
C Si
SWRCH45F 0.42–0.48 0.10–0.35
deformation with relatively homogeneous strain distribution in the
wire than the WD process for the same area reduction. Therefore, it
is more likely to improve spheroidization of the wire produced by
the NCD sequence because of the higher accumulated deformation
energy in the drawn wire.
3. Experimental method
In the current study, a commercially available medium carbon
steel wire used for the experiments was provided by POSCO. The
initial diameter of coiled wire was 14 mm and its chemical compo-
sition is given in Table 1. In order to straighten out the coiled wire
prior to conducting the WD and NCD processes, the initial diame-
ter of 14 mm wire was drawn to the wire into the final diameter
of 13 mm. The drawn wires were annealed at 870 ◦ C for 1 h and air
cooled to room temperature to remove previous deformation and
thermal history. Acid pickling was applied for removing the scale
from the annealed wire before the experiments. The drawn wire of
13 mm in diameter was used to conduct the experiments for the
WD and NCD processes for the first pass.
Both processes were designed and applied up to the fourth pass
with the dies under the same condition used in the FE analysis
at room temperature. The wires were drawn up to the total area
reduction of 59% for both processes. The drawing velocity applied to
the experiments was 10 mm/s as mentioned earlier. Commercially
available MoS2 drying film lubricant (NABAMOLY #3003) was used
for the specimen and die interfaces to reduce the friction effect for
the WD and NCD processes.
To investigate the effect of prior cold working on spheroidiza-
tion of the material, subcritical annealing (SA) was applied to the
produced wires by the WD and NCD processes, respectively. The
thermal history of SA used in this study is given in Fig. 4. The drawn
wires were heated to 705 ◦ C (below A1 transformation temperature,
727 ◦ C) for 1 h, and were held at that temperature for different time
periods as 2 (SA-2), 4 (SA-4), and 8 (SA-8) h. Then, these wires were
cooled down to 540 ◦ C in the furnace and finally exposed to air
cooling until reaching room temperature.
To investigate mechanical properties of the processed wires
such as ultimate tensile strength (UTS) and reduction of area
(RA) at fracture, tension tests were carried out twice for each
case by employing Instron 5583 (UTM) with a constant crosshead
speed of 0.017 mm/s, which corresponds to an initial strain rate of
6.7 × 10−4 s−1 at room temperature. This particular strain rate was
Fig. 4. Thermal history of the subcritical annealing used for experiments.
351
Mn P S
0.60–0.90 0.030 max. 0.035 max.
chosen since it is within the intermediate deformation rate zone.
Tension specimen with the gauge length of 25 mm and the diam-
eter of 5 mm was machined off from the processed wire according
to ASTM standard (E8/E8M-11, 2011).
After carrying out the tension test, two parts of the fractured
specimen were fitted together and the reduced diameter was mea-
sured for determining the RA value at fracture. The minimum
diameter at the location of fracture was used to calculate the RA
value, as described in Eq. (1) for experimental measurements of
the ductility of the annealed wires:
Ai − Af
RA (%) = × 100, (1)
Ai
where Ai and Af refer to the area of initial and final cross-sectional
areas, respectively.
Vickers micro-hardness test was conducted for each experi-
mental condition by using Mitutoyo HM-122 to investigate local
strength of the drawn wires. The specimens were cut in perpen-
dicular to the drawing direction. They were cold mounted in epoxy
resin with a catalyst. Mechanical grinding by the SiC papers increas-
ing the grit number such as 500, 800, 1200, and 2000 and polishing
up to 1 ␮m by diamond suspensions, respectively, were applied to
prepare the specimens. The Vickers micro-hardness values were
measured at 11 points with regular interval of 0.77 mm along the
line, excluding 0.3 mm at both ends of the specimen, as shown in
Fig. 5. In the test, the applied load was 2 N and the duration time
was 10 s.
Upsetting test was conducted to evaluate cold workability of the
wires. The diameter and height of the cylindrical specimen were
5 mm and 10 mm, respectively. Tungsten carbide platens equipped
with a guide post apparatus were applied to conduct upsetting
tests. The tests were carried out using the hydraulic press with the
forming load limit of 300 kN. Commercially available grease was
used as lubricant between the specimen and platens interface to
reduce the friction effect. The height of the compressed specimen
was measured to calculate the compressive strain, as described in
Eq. (2) for experimental measurement of the cold workability of
the annealed specimen:
 
hi
ε = ln , (2)
hf
where hi and hf are the initial and final height of the specimen,
respectively. The upsetting test was continued until the surface
crack was observed. To obtain the reproducible data, the upset-
ting tests were conducted three times for the same experimental
condition.
Fig. 5. The location of measuring points for Vickers micro-hardness test.
352 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356

The scanning electron microscopy (SEM) was applied to investi-

gate the microstructure change and to measure the spheroidization
ratio (SR) of the specimens produced by the WD and NCD pro-
cesses for each annealing time. The specimens were sliced into two
at the middle part in parallel to the drawing direction and were
cold mounted in epoxy resin with a catalyst. The specimens were
prepared by mechanical grinding and polishing up to 1 ␮m by SiC
papers and diamond suspensions, respectively. These specimens
were etched by a nital and a picric solution for measurements. The
SEM micrographs were randomly obtained in the parallel plane of
the drawing direction for the same processing condition.
To quantitatively measure the spheroidization ratio, the SEM
micrographs were analyzed by using commercially available image
analysis program (Image J). This program was used to obtain the
local information in the area and the aspect ratio of the cemen-
tite particles. For the analysis, the SEM micrographs were changed
into a binary image to obtain the accurate information of each par-
ticle. To guarantee the reproducibility of the experimental data,
the image analyses of the SEM micrographs were carried out three
times for the same processing condition. According to the literature,
the spheroidized cementite particles were considered to be an ellip-
tical shape with an aspect ratio represented by the ratio of the major
to minor axes. According to Chattopadhyay and Sellars (1982) and
Chou et al. (1986), the cementite particles were considered to be
spheroidized up to the aspect ratio of 5:1. The cementite area was
used to calculate the spheroidization ratio (SR), as described in Eq.
(3) for experimental measurements of the spheroidization ratio of
the annealed specimen:
Vs
Spheroidization ratio (%) = × 100, (3)
Vt
where Vt and Vs are the cementite area of the total and spheroidized
cementite area, respectively. To exclude plate-shaped cementite
in the pearlite structure for the calculation of the spheroidization
ratio, circularity was introduced, as described in Eq. (4):
As
Circularity = × 4, (4)
Ps2
where As and Ps are the spheroidized cementite area and perimeter,
respectively. In the present study, the cementite particle under the
circularity of 0.3 was neglected in the calculation.

4. Results and discussion

Fig. 6(a) shows the engineering stress–strain curves obtained

from the initial specimen, drawn wires by the WD and NCD pro-
cesses, and the WD SA-2 and NCD SA-2, respectively. The UTS value
of the initial specimen was 715.7 MPa. The UTS values of the speci-
mens produced by the WD and NCD sequence up to the fourth pass
were 1076.7 and 1113.5 MPa, respectively. This result indicated
that the NCD sequence could impose higher plastic deformation
on the wire than the WD process. This observation was generally
in good agreement with prior FE analysis results which showed
Fig. 6. (a) Engineering stress–strain curves obtained from tension tests before and
higher effective and shear strains in the wire processed by the
after spheroidization annealing for 2 h, (b) UTS and (c) RA values obtained from
NCD sequence than the conventional WD. The UTS values of the tension tests for the specimens processed by the WD and NCD after spheroidization
specimens processed by the WD SA-2 and NCD SA-2 were 566.6 annealing.
and 523.8 MPa, respectively. The specimen processed by the NCD
sequence and SA showed 42.8 MPa lower than the one produced by
the WD and SA process according to the present investigation. specimens processed by the WD and NCD increased as annealing
The UTS values measured for various annealing times (SA-2, SA- times increased. In particular, the UTS value of the specimen like
4, and SA-8) for both processes are shown in Fig. 6(b). During the NCD SA-2 was lower than the WD SA-8.
SA, the UTS values significantly decreased for the annealing time of The values of reduction of area (RA) at fracture for each anneal-
2 h. The difference of the UTS values between the initial specimen ing time are shown in Fig. 6(c). According to the literature (Dieter,
and WD SA-2 was 510.1 MPa. It was 589.7 MPa for NCD SA-2. 2003), the RA value during the tension test could be the best mea-
The UTS values for the specimens produced by both processes surement data for workability since the necking strain and crack
slowly decreased and differences of these values between the propagation resistance of the material could be expected from the
 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
Fig. 7. (a) Distribution of measured Vickers micro-hardness with a standard deviation for spheroidization annealing of 2 h and (b) the average Vickers micro-hardness values
during spheroidization annealing.
RA value. In the present investigation, the RA values drastically
increased for the annealing time of 2 h and its values for the speci-
mens processed by the WD and the NCD processes were 63.8% and
71.7%, respectively. It was seen in Fig. 6(c) that the RA values do not
change significantly as the annealing times increased.
According to the work by Guo et al. (2012), remarkable
decrease of strength and increase of elongation at failure after the
spheroidization annealing was attributed to the release of internal
stress, the recrystallization of the deformed ferrite matrix, and the
elimination of lamellar pearlite and the generation of spheroidized
cementite in the microstructure. From the above-mentioned ten-
sion test results, it was found that the drawn wire by the NCD
sequence might show better workability compared to the one pro-
duced by the WD process.
Vickers micro-hardness distributions of the specimens
processed by the WD and NCD processes, and the WD SA-2
and NCD SA-2, respectively, are shown in Fig. 7(a) with the
average micro-hardness value and standard deviation for each
measured point. The micro-hardness distributions of the specimen
processed by the NCD sequence were higher than the one of the
WD process. The average micro-hardness values of the specimen
processed by the NCD sequence for the major and minor axes were
304.0 and 302.3 HV, respectively, and the one of the WD process
was 295.4 HV. The micro-hardness of the WD SA-2 and NCD SA-2
specimens shows more homogeneous distribution than the spec-
imens processed by the WD and NCD processes. Fig. 7(b) shows
that the average micro-hardness values significantly decreased
during the annealing time for 2 h and a tendency was similar with
the UTS values as observed in the tension tests. It was found that
the average micro-hardness value of the specimen processed by
the NCD sequence was lower than the one of the WD process after
the spheroidization annealing.
The micro-hardness of a material is known to be proportional
to its strength, and it usually implies a local strength and resis-
tance to deformation of the material. In general, high-strength
and high-hardness material is usually hard to deform, resulting in
low workability. Therefore, the measured micro-hardness value is
conveniently used to evaluate the workability of the material as
an indicator similar to the compressive strain obtained from the
upsetting test. For comparison of both processes in the present
experiments, it was found that the wires processed by the NCD
sequence showed the improved workability for all the spheroidiza-
tion annealing times.
353
Upsetting specimens were compressed for various compressive
strains until surface crack appeared, as shown in Fig. 8. According
to the upsetting results, the specimens produced by the WD pro-
cess and NCD sequences were cracked at the compressive strain
of 2.27 and 2.43, respectively. Therefore, the specimen processed
by the NCD sequence showed better cold workability than the one
produced by the WD process. According to the work by Ma et al.
(2004), cold workability was particularly sensitive depending on
the microstructure and was greatest for the specimen with the
completely spheroidized microstructure. In this regard, it can be
construed that the spheroidization of the specimen processed by
the NCD might be better than the one by the WD.
In Fig. 9, microstructure changes of the initial specimen and
specimens produced by the WD and NCD processes are shown for
each annealing time. The initial microstructure consisted of soft
ferrite matrix (␣-Fe) and a great deal of pearlite which is ferrite
containing hard lamellar cementite (Fe3 C) before annealing. The
cementite of the specimen produced by both processes was elon-
gated. In the specimen processed by the NCD sequence lamellae
cementite was broken up and interlamellar spacing became nar-
rower compared to the WD process. This microstructure evolution
indicated that the likelihood to accelerate spheroidization behavior
of the drawn wire by increasing defects of cementite and shortening
diffusion path for carbon atoms.
For the initial SA-2, some cementite was changed into the
spherical morphology. However, there was still a lot of lamellar
cementite in the microstructure. The cementite in the WD SA-2
Fig. 8. Upsetting test results according to various compressive strains obtained
from the specimen processed by (a) WD and (b) NCD processes, respectively, for
spheroidization annealing of 2 h.
354 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
Fig. 9. SEM micrographs obtained from the initial specimen and specimens processed by WD and NCD processes during various spheroidization annealing times of 2, 4, and
8 h.
was broken up into spherical particles compared to the initial
specimen, but the specimen still had many lamellar cementite.
The lamellar cementite had not completely disappeared for the
initial specimen until annealing time of 8 h. The lamellar cemen-
tite of the specimen produced by the WD process was broken up
into the small spherical cementite compared to the initial mate-
rial although a small amount of lamellar cementite still remained
after the annealing time of 4 h. Most of lamellar cementite was
changed into spherical particles after the SA for 8 h. On the other
hand, most of the cementite in the specimen processed by the NCD
sequence was broken up into spherical particles for the case with
annealing for 2 h. The microstructures looked similar to the mor-
phology of the specimen produced by the WD process after the SA
for 8 h. The growth of spherical cementite was also observed as the
annealing time increased in the specimen processed by the NCD
sequence.
These microstructure changes might be attributed to deforma-
tion behavior for both processes. During the NCD sequence, tensile
and compressive stresses were repeatedly applied in the wires
because of the change of die geometry. The difference of loading
path might significantly influence on the microstructure evolu-
tion and mechanical properties of the drawn wire even though
the total area reduction was the same for both processes. In the
numerical simulation section, it was demonstrated that the NCD
sequence could impose higher shear and effective strains on the
wire than the conventional WD for the same pass. Therefore, it is
more likely to improve spheroidization behavior of the drawn wire
by the NCD sequence because the highly accumulated plastic defor-
mation could easily induce morphological defects of cementite in
the drawn wire compared to the WD.
According to the study by Tian and Kraft (1987), the initiation
and development of spheroidization were associated with morpho-
logical defects such as kinks, holes, and fissures in the cementite
plates. These defects in the cementite led to the break-up of large
cementite plates into small particles and had high tendency to form
a more stable microstructure like spheroidal cementite compared
to the regular cementite during spheroidization annealing. There-
fore, it is construed that higher plastic deformation incurred by the
NCD sequence could initiate the defects in the cementite more than
the one by the WD process, resulting in acceleration of spheroidiza-
tion of the material.
Fig. 10 shows the distribution of spheroidized particle areas for
the specimen produced by the WD and NCD processes depending
on the various annealing times. The distribution of spheroidized
particle sizes for the WD SA-2 and NCD SA-2 shows the largest
group of spheroidized particles under the size of 1.0 ␮m2 . This
result reflects the increase in newly spheroidized particles from
the break-up of lamellae cementite occurring in the beginning of
the SA in order to decrease their surface energy. As a result of
this, the cementite fragments were progressively changed into the
spherical shape. In particular, the area fraction of spheroidized
cementite particles is higher for the NCD SA than the WD SA. In
case of the SA-4, it was not significantly changed from the SA-2 for
both WD and NCD processes since the new spheroidized particles
were formed at the same rate as previous particles were coars-
ening by Ostwald ripening (Ratke and Voorhees, 2002). Owing to
the distribution observed in the NCD SA-8, spheroidized cemen-
tite which was larger than 1.0 ␮m2 considerably increased rather
than newly spheroidized particles. However, the spheroidized par-
ticles for the WD SA cases are mostly under the size of 1.0 ␮m2 .
 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
Fig. 10. Distribution of the spheroidized particle area by the area fraction method for the specimens processed by WD and NCD processes during spheroidization annealing.
Fig. 11. Empirical relationship between the UTS value and spheroidization ratio
(SR).
From these observations, it was construed that the NCD sequence
might be more effective than the WD process in producing the
medium carbon steel wires with improved spheroidization during
heat treatment.
The relationship between the spheroidization ratio (SR) and UTS
value is shown in Fig. 11. In this figure, the UTS values are linearly
decreased by increasing the SR. The UTS value of the initial speci-
men was not nearly changed as the SR increased. For the WD SA-2
and 8 cases, spheroidization ratios were 43.3% and 76.6%, respec-
tively. Furthermore, the SR of the NCD sequence showed 82.9%
for the NCD SA-2. From this experimental result, the SR could be
improved by the prior cold working and the ratio could be esti-
mated from the UTS value by using the empirically derived equation
in Fig. 11.
5. Conclusions
The current study was conducted to investigate the effect of
the NCD sequence on the spheroidization of the material. The
355
numerical and experimental results of the present study show that
the NCD sequence could impose higher and relatively more homo-
geneous plastic deformation on the medium carbon wire than the
WD process. Thus, it can be construed that spheroidization can
be improved during the heat treatment. The annealed specimen
processed by the NCD sequence showed lower UTS and micro-
hardness values with higher RA value at fracture than those of
the WD process. Moreover, mechanical properties of the annealed
specimen produced by the NCD sequence and the SA for 2 h was
similar to the one processed by the WD and SA for 8 h. Therefore,
the energy saving during the SA incurred by the NCD sequence is
remarkable. In addition, the upsetting test results represent that
the annealed specimen processed by the NCD sequence shows
better cold workability than the one processed by the WD. The
SEM results also indicated that the prior cold working could accel-
erate spheroidization of cementite during the SA and improved
spheroidization ratio could be obtained for the NCD sequence in
less time compared to the WD process. The empirical relationship
between the spheroidization ratio and ultimate tensile strength
was derived in the present study to estimate the change of mechan-
ical strengths due to spheroidization. In conclusion, it was demon-
strated that the non-circular drawing sequence might be efficient
in manufacturing medium carbon steel wires with improved
spheroidization and cold workability by subcritical annealing.
Acknowledgments
This work was supported by a research grant from POSCO
G01130131. The assistance provided by Messrs. Yong Sik Park
and Myung Su Choi during various experiments at the technology
research center at POSCO was very much appreciated.
References
ASTM Committee E28.04, 2009. E09-09: Standard Test Methods of Compression
Testing of Metallic Materials at Room Temperature. ASTM International.
ASTM Committee E28.04, 2011. E8/E8M-11: Standard Test Methods for Tension
Testing of Metallic Materials. ASTM International.
356 H.S. Joo et al. / Journal of Materials Processing Technology 216 (2015) 348–356
Baek, H.M., Jin, Y.G., Hwang, S.K., Im, Y.T., Son, I.H., Lee, D.L., 2012. Numerical study
on the evolution of surface defects in wire drawing. J. Mater. Process. Technol.
212, 776–785.
Baranova, V.A., Sukhomlin, G.D., 1981. Spheroidization of cementite in steel. Met.
Sci. Heat Treat. 23, 798–802.
Chattopadhyay, S., Sellars, C.M., 1977. Quantitative measurements of pearlite
spheroidization. Metallography 10, 89–105.
Chattopadhyay, S., Sellars, C.M., 1982. Kinetics of pearlite spheroidisation during
static annealing and during hot deformation. Acta Metal. 30, 157–170.
Chou, C.C., Kao, P.W., Cheng, G.H., 1986. Accelerated spheroidization of hypoeu-
tectoid steel by the decomposition of supercooled austenite. J. Mater. Sci. 21,
3339–3344.
Dieter, G.E., 2003. Handbook of Workability and Process Design. ASM international,
Materials Park.
Guo, W.Y., Li, J., Jiang, X.F., 2012. Subcritical spheroidization of medium-carbon
50CrV4 steel. J. Mater. Eng. Perform. 21, 1003–1007.
Hwang, S.K., Baek, H.M., Lee, J.W., Son, I.H., Im, Y.T., Bae, C.M., 2014. The effect
of microstructure and texture evolution on mechanical properties of low car-
bon steel in a non-circular drawing sequence. J. Mater. Process. Technol. 214,
318–325.
Hwang, S.K., Jin, Y.G., Son, I.H., Rhee, K.H., Lee, D.L., Im, Y.T., 2011. Flow characteristics
of continuous shear drawing of high carbon steel. Int. J. Mech. Sci. 53, 479–484.
Kamyabi-Gol, A., Sheikh-Amiri, M., 2010. Spheroidizing kinetics and optimization
of heat treatment parameters in CK60 steel using Taguchi robust design. J. Iron
Steel Res. Int. 17, 45–52.
Kim, S.Y., Im, Y.T., 2002. Three-dimensional finite element analysis of non-
isothermal shape rolling. J. Mater. Process. Technol. 127, 57–63.
Ko, Y.G., Namgung, S., Shin, D.H., Son, I.H., Rhee, K.H., Lee, D.L., 2010. Spheroidization
of medium carbon steel fabricated by continuous shear drawing. J. Mater. Sci.
45, 4866–4870.
Kwak, D.Y., Im, Y.T., 2002. Remeshing for metal forming simulations-Part II:
three-dimensional hexahedral mesh generation. Int. J. Numer. Meth. Eng. 53,
2501–2528.
Lee, C.H., Kobayashi, S., 1973. New solutions to rigid-plastic deformation problems
using a matrix method. Trans. ASME, J. Eng. Ind. 95, 865–873.
Lee, H.C., Saroosh, M.A., Song, J.H., Im, Y.T., 2009. The effect of shrink fitting ratios on
tool life in bolt forming processes. J. Mater. Process. Technol. 209, 3766–3775.
Lee, J.W., Baek, H.M., Hwang, S.K., Son, I.H., Bae, C.M., Im, Y.T., 2014. The effect of
the multi-pass non-circular drawing sequence on mechanical properties and
microstructure evolution of low-carbon steel. Mater. Des. 55, 898–904.
Lv, Z.Q., Wang, B., Wang, Z.H., Sun, S.H., Fu, W.T., 2013. Effect of cyclic heat treatments
on spheroidizing behavior of cementite in high carbon steel. Mater. Sci. Eng. A
574, 143–148.
Ma, X., Humphreys, A.O., Nemes, J., Hone, M., Nickoletopoulos, N., Jonas, J.J., 2004.
Effect of microstructure on the cold headability of a medium carbon steel. ISIJ
Int. 44, 905–913.
Nam, S., Lee, D., 1987. Accelerated spheroidization of cementite in high-carbon steel
wires by drawing at elevated temperatures. J. Mater. Sci. 22, 2319–2326.
O’Brien, J., Hosford, W., 2002. Spheroidization cycles for medium carbon steels.
Metal. Mater. Trans. A 33, 1255–1261.
Ratke, L., Voorhees, P.W., 2002. Growth and Coarsening: Ostwald Ripening in Mate-
rial Processing. Springer, New York.
Shin, D.H., Han, S.Y., Park, K.T., Kim, Y.S., Paik, Y.N., 2003. Spheroidization of low
carbon steel processed by equal channel angular pressing. Mater. Trans. 44,
1630–1635.
Tian, Y., Kraft, R.W., 1987. Mechanisms of pearlite spheroidization. Metal. Trans. A
18, 1403–1414.
Zhang, S.L., Sun, X.J., Dong, H., 2006. Effect of deformation on the evolution
of spheroidization for the ultra high carbon steel. Mater. Sci. Eng. A 432,
324–332.