ISIJ International, Vol.

55 (2015),
ISIJ International,
No. 9
Vol. 55 (2015), No. 9, pp. 19011905

Generation Mechanism of TiN Inclusion for GCr15SiMn during

Electroslag Remelting Process
Liang YANG,1)* Guo-guang CHENG,1) Shi-jian LI,1) Min ZHAO2) and Gui-ping FENG2)
1) State Key Lab. of Advanced Metallurgy, University of Science and Technology Beijing, Beijing, 100083 China.
2) Dongbei Special Steel Group, Fushun, 113001 China.
(Received on May 3, 2015; accepted on June 1, 2015)

TiN inclusion with large size found in ESR ingot is of great harm to steel GCr15SiMn, it is significant to
elucidate the possibility of this inclusion in consumable electrode retaining in the subsequent ingot and
the effect of slag composition on the content of TiN inclusion in ingot after ESR refining. Based on three
remeltings and with the help of ultrahigh-temperature confocal scanning violet laser microscope, it demonstrates that TiN inclusion in electrode will completely decompose in the solid-liquid coexistence region
at the electrode tip, reflecting that TiN inclusion found in ESR ingot is regenerated. For different contents
of Ti in slag, there is a corresponding equilibrium value of Ti content in steel, when the content of Ti in
electrode is higher than this critical equilibrium value, it will decline in steel, otherwise Ti pick-up will occur.
The effect of increasing SiO2 content in slag on decreasing Ti content in steel is obvious due to the Si/Ti
exchange reaction taking place, further leading to the low content of TiN in ingot.
KEY WORDS: generation mechanism; TiN inclusion; GCr15SiMn; ESR process.

1.

in the subsequent ingot, second the relationship between

the slag composition and TiN content in ingot after ESR
refining. The discussed results will provide some guidance
for chemical composition design of electrode and ESR slag,
further promote the cleanliness of ESR ingot.

Introduction

GCr15SiMn is one of the most important special steel,

with higher demand for non-metallic inclusion. Electroslag
remelting (ESR) has been widely used as an important technology to produce this steel in order to improve its cleanliness. However, in some special steel plants, TiN inclusion
with large size is usually found in this steels consumable
electrode and ESR ingot, which will limit its application,
such as high speed train and large-scale equipment.
Many studies have focused on research of oxide behavior
during ESR process, which have provided some understanding on minimizing oxide,13) the evolution of oxide4,5) and
controlling the oxygen content.68) As known, the effect of
non-deformable TiN inclusion on bearing steel is worse
than that of oxide in the same size.9) However, now only a
few studies have been reported TiN behavior during ESR
process. Hua et al.10) investigated the inclusion of Ni-22Cr12Co before and after ESR refining, proposed that ESR
process has almost no effect on removal of TiN inclusion,
but this paper did not take the influence of ESR slag into
consideration. S. F. Medina et al.11) studied the thermodynamic aspects in the manufacturing of microalloyed steels
by ESR process, the result shows that the uniformity of
chemical composition in the steels with Ti requires the presence of SiO2 in slag to inhibit an Si/Ti exchange reaction.
In this paper, in order to clarify the generation mechanism
of TiN in ingot for GCr15SiMn, it intends to make clear two
aspects, first whether TiN inclusion in electrode will retain

2.

Experimental Procedure

Three remeltings (Run E1, E2 and E3) were carried out

in an electroslag furnace with a capacity to produce 6 000
kg ingot. The electrode of GCr15SiMn was produced by
the process of steelmaking, secondary refining and ingot
casting, with radius 0.16 m. The size of ingot is 0.31 m in
radius and 1.62 m in height, the weight is about 3 500 kg.
The conditions under which the three remeltings were
carried out are described in Tables 1 and 2. The melting rate
Table 1.

Conditions for three remeltings.

Average chemical composition of electrode ( wt )

Run Slag

C,
%

Si,
%

Mn,
%

Cr,
%

Al,
%

Ti,
ppm

N,
ppm

O,
ppm

1.01

0.58

1.07

1.51

0.02

30

65

E1

E2

1.01

0.5

1.05

1.54

0.025

48

56

E3

1.02

0.59

1.05

1.49

0.05

41

56

10

Table 2.

* Corresponding author: E-mail: yangliangbj@163.com

DOI: http://dx.doi.org/10.2355/isijinternational.ISIJINT-2015-253

1901

Chemical composition of slag A and B (%, wt).

Slag

CaF2

Al2O3

CaO

SiO2

FeO

Ti

Others

50

30

13

5.3

0.45

0.005

1.2

58

30

10

0.1

0.12

0.005

1.8

2015 ISIJ

ISIJ International, Vol. 55 (2015), No. 9

1 733 K, greater than 1 691 K, so it means that for 10 m

TiN inclusion, or even some larger, it will totally decompose
in the solid-liquid coexistence region at the electrode tip.
Actually, when the contents of Ti and N in steel are under
the equilibrium value for TiN inclusion, TiN inclusion will
begin to decompose in steel, the decomposition chemical
reaction of TiN is given in Eq. (1).9)

of electrode is approximately 8 kg/min. After ESR process,

chemical compositions of edge, 1/2 radius and center area
in E1, E2 and E3 ingots were analysed.
At present, there are no measures taken to directly
observe the evolution of inclusion in electrode during ESR
refining, so in order to investigate the characteristics of TiN
inclusion in electrode during the course of heating up, the
ultrahigh-temperature confocal scanning violet laser microscope has been used to simulate this process. The heating
rate is set as 2 K/s, the wave length of violet laser is 408
nm, and the protective atmosphere in experiment is argon.
Meanwhile, the content of TiN inclusion in electrode and
ingots has been checked by Scanning Electron Microscope
and Energy Dispersive Spectrometer (SEM-EDS). As to
reduce the observed error, greater than 500 fields have been
checked for every specimen of electrode and ingots.
3.

15 218
+ 5.64 .. (1)
T
The contents of Ti and N on the surface of TiN inclusion (C[Ti],Surf and C[N],Surf) can be calculated through Eqs.
(2), (3).14)
TiN ( s ) = [Ti ] + [ N ] lgKTiN ,decomp. =

A[ N ] D[LTi ] ( C[ N ],Surf C[ N ], L )
=
A[Ti ] D[LN ] ( C[Ti ],Surf C[Ti ], L ) .................. (2)

Results and Discussion

C[ N ],Surf C[Ti ],Surf =

3.1.

Genetic Research of TiN Inclusion from Electrode

to Ingot
The morphology and area scanning of TiN inclusion
found in specimen cut from Run E2 electrode is shown in
Fig. 1 by using SEM-EDS, which size is about 10 m.
With increasing temperature of specimen, the evolution of
TiN inclusion in it is shown in Fig. 2. In Fig. 2(a), the grain
boundary of steel becomes clear at about 1 433 K, and TiN
inclusion is just on the boundary, which further illustrates
the conclusion in the authors previous work that TiN inclusion usually precipitates at later period of solidification of
this steel.12) In Fig. 2(b), small obvious pits begin to appear
in this inclusion, and in Fig. 2(c), the interface between
inclusion and steel becomes obscure at the solidus temperature of this steel about 1 583 K, as calculated according to
Ts = 1 811 tixi, where ti and xi are respectively the
temperature coefficient and mass percentage of solute i.13)
From 1 628 K, higher than the steels solidus temperature
45 K, the decomposition rate of TiN speeds up, and at about
1 691 K, this inclusion almost disappears, as shown in Figs.
2(d)2(j). The liquidus temperature of this steel is about

Fig. 1.

Where D[LTi ] and D[LN ] represent diffusion coefficients of

Ti and N in liquid steel. A[Ti] and A[N] are respectively the
atomic mass of Ti and N. C[Ti],L and C[N],L are the contents of
Ti and N in steel. f[Ti] and f[N] represent activity coefficients
of Ti and N, respectively.
Based on Eqs. (1)(3) and experimental data in Run E2,
the contents of Ti and N on the surface of TiN inclusion
are respectively 59 ppm and 200 ppm, greater than 48 ppm
and 56 ppm in steel at the solidus temperature about 1 583
K. However, due to difficult diffusion of Ti and N in steel,
only the interface between inclusion and steel will become
obscure at this time. With increasing temperature, the diffusion of Ti and N becomes easier, and the impetus of
decomposition gets enhanced, further leading to the quick
decomposition of TiN inclusion.
The size distribution of TiN inclusion in Run E2 electrode
is shown in Fig. 3, which shows that most TiN inclusions in
electrode are within 10 m. According to above discussion,
it concludes that for researched steel GCr15SiMn, all TiN

The morphology and area scanning of TiN in electrode

specimen.

2015 ISIJ

KTiN, decomp.
................... (3)
f[ N ] f[Ti ]

Fig. 2.

1902

The evolution of TiN during the process of heating up.

ISIJ International, Vol. 55 (2015), No. 9

It can be seen that this kind of inclusion usually has a regular shape, which will easily induce the microcrack between
steel and inclusion, and further deteriorate the properties
of steel.
Based on the model proposed by J. H. Wei and A.
Mitchell,15) the relationship between slag composition and
Ti content in steel has been theoretically calculated.
In ESR process, the thermodynamic equilibrium will be
reached between steel and slag.11) At the interface between
steel and slag, the oxidation reaction is given as Eq. (4). The
exchange reaction between oxides and other solutes will be
inferred through the oxidation reaction.

inclusion in electrode will completely decompose during

electrode remelting and solutes Ti and N will join in the
following reaction between steel and slag. In other words,
TiN inclusion found in ESR ingot is regenerated, not from
the electrode.
3.2.

Relationship between Slag Composition and TiN

Content in Ingot
Chemical compositions of edge, 1/2 radius and center
area in Run E1, E2 and E3 ingots are listed in Table 3.
Compared to the data in Table 1, it is obvious that for Run
E1 and E2 using slag A, the contents of Ti in steel both
decline, while Ti pick-up has occurred for Run E3 using
slag B. Meanwhile, the content of N in Run E1 increases
compared with that in electrode, while in Run E2 and E3,
N contents are almost no change. This is possible due to the
protective atmosphere used in Run E2 and E3 remeltings.
The content of TiN inclusion in center area for Run E1,
E2 and E3 ingots is shown in Fig. 4, which illustrates that
TiN content in Run E3, corresponding to the highest Ti
content among three ingots, is greater than that in E1 and
E2. In addition, TiN content in Run E1 is slightly less than
that in E2, based on the data in Table 3, it concludes that
the effect of Ti content on the content of TiN is greater than
that of N content for this steel.
Part of TiN inclusions found in ingot is shown in Fig. 5.

Fig. 3.

x[ M ] + y[O] = ( M x Oy ) ........................ (4)

In Eq. (4), M represents Fe, Si, Mn, Al and Ti, respectively. For the oxide of solute Ti, according to thermodynamic calculation, it knows that the oxide Ti3O5 is the most
stable compared with other oxides of Ti under the same
temperature.16) So in this paper, it takes Ti3O5 as the oxide
for solute Ti.
According to the mass conservation of M, Eq. (5) can be
obtained.
C[ M ]e + xC( M x Oy )e = C[ M ]0 + xC( M x Oy )0 ............. (5)
Where C( M x Oy )0 and C[M]0 represent the initial contents of
MxOy in slag and M in steel, respectively. C( M x Oy )e and C[M]e

Fig. 4.

Size distribution of TiN inclusion in electrode.

Table 3.
Run

E1

E2

E3

Position

TiN content in center area of E1, E2 and E3 ingots.

Chemical compositions of ingots at different positions.

Chemical compositions of ingots (wt).

C, %

Si, %

Mn, %

Cr, %

Al, %

Ti, ppm

N, ppm

O, ppm

Edge

0.48

1.05

1.50

0.014

21

76

1/2 radius

0.99

0.5

1.05

1.52

0.014

17/23

74

10

Center

0.96

0.49

1.05

1.51

0.018

22

74

11

Edge

1.1

0.49

1.05

1.55

0.02

28

53

1/2 radius

1.08

0.52

1.05

1.55

0.019

27

51

Center

1.07

0.51

1.05

1.54

0.02

27

52

Edge

0.98

0.56

1.02

1.47

0.038

50

60

1/2 radius

0.57

1.02

1.48

0.039

48

59

Center

0.57

1.03

1.46

0.039

48

62

1903

2015 ISIJ

ISIJ International, Vol. 55 (2015), No. 9

Fig. 5.

Morphology of part TiN inclusions in ingot.

are respectively the equilibrium contents of MxOy and M at

the interface between steel and slag.
After the calculation for C( M x Oy )e and C[M]e, the diffusion
of M in steel and MxOy in slag will be considered through
penetration theory.
Assume that the position x = 0 is the interface between
steel and slag, and the slag is on the x > 0 direction, the
steel is on the x < 0 direction. For Ti + Ti3O5 system, Eq. (6)
can be get.

Table 4.
lgK FeO

6 317/T2.73

lgK SiO2

30 110/T11.40

lgKTi3O5

92 075/T29.84

lgK MnO

15 050/T 6.75

lgK Al2O3

64 900/T20.63

Table 5.

C[Ti ]
C[Ti ]
= D[LTi ]
x0
t
x 2
............... (6)
C(Ti3O5 )
2C(Ti3O5 )
x0
= D(Ti3O5 )
t
x 2
Where C[Ti] and C(Ti3O5 ) represent the contents of Ti in
steel and Ti3O5 in slag, respectively. D(Ti3O5 ) is the diffusion
coefficient of Ti3O5 in slag.
According to Eq. (6) and its initial and boundary conditions, the content distribution of solute Ti in steel at different
positions and time can be calculated, as shown in Eq. (7).

The equilibrium constants of Eq. (4).

Diffusion coefficients of Ti, Si, Mn and O in steel.

Solutes

D[Li ] (cm2 s 1)

Solutes

D[Li ] (cm2 s 1)

Ti

11 500
3.1 10 3 exp
RT

Si

71123
6.65 1012 exp
T

Mn

62 535
5.97 1010 exp
T

12 000
3.34 10 3 exp
RT

melt rate and radius of electrode, respectively. m is the

viscosity of metal, and l is the length of cone edge in contact
with the slag. m and s represent the densities of steel and
slag, respectively.
The equilibrium constants of Eq. (4) and the diffusion

coefficients of Ti, Si, Mn and O in steel are shown in Tables

3C(Ti3O5 )e 3C(Ti3O5 )0
x

1 + erf
13,16)
C[Ti ] = C[Ti ]0 1
and 5.13,15,18)
2 (tD[LTi ] ) 4
3C(Ti3O5 )e 3C(Ti3O5 )0 + C[Ti ]e

At present, there is no reliable data for the diffusion coefL

L
ficient of Al in steel, in this paper, it takes D[ Al ] = D[ Si ] for

3C(Ti3O5 )e 3C(Ti3O5 )0
x
15)

calculation.
1 + erf
............. (7)
C[Ti ]0 1
2 (tD[LTi ] )
When taking CaF2 as an inert component, the main dif 3C(Ti3O5 )e 3C(Ti3O5 )0 + C[Ti ]e

ference between slag A and B is the content of SiO2, based
on above Eqs. (4), (5), (7), the effect of SiO2 in slag on the
The average residence time te and the thickness of the
liquid metal film at the electrode tip are given in Eqs. (8)
content of Ti in steel has been investigated, as shown in Fig.
and (9).17)
6. The data of Ti content corresponding to Run E1, E2 and
E3 in Table 3 is also marked in Fig. 6, which shows that
1
2
5
2 cos 3 m 3 R 3
the theoretically expected is in agreement with the experite = 3.35
g sin cos
mentally proved.
W
3
m
m

.......................................... (8)
It is obvious that, from Fig. 6, the content of SiO2 in
slag has a significant influence on Ti content in steel. With
1
increasing SiO2 content in slag, the content of Ti in steel will

l 2 cos 2 3
3m Wm
decline.
When the content of SiO2 in slag is lower than a
=

R 2
2 ( m s ) gsin cos
certain value, Ti pick-up in steel will occur in ESR process,
.......................................... (9)
like Run E3. According to Refs. 19), 20), the Si/Ti exchange
reaction has occurred in Run E3 following the expression
Where is the cone angle of electrode tip, g represents
acceleration due to gravity. Wm and R are the volumetric
given in Eq. (10), as a consequence of the low activity of
2015 ISIJ

1904

ISIJ International, Vol. 55 (2015), No. 9

about 12 ppm and 25 ppm of Ti content in steel respectively

corresponding to that of 0.005% and 0.05% in slag. When
the content of Ti in electrode is greater than this critical
equilibrium value, Ti will be oxidized and the oxidation
degree will rise with higher Ti content in electrode, while
under the opposite case, Ti content in steel will increase.
4.

Conclusions

In this paper, the generation mechanism of TiN inclusion has been investigated for clean GCr15SiMn ESR ingot
production. The conclusions in the present study are summarized as follows.
(1) During the course of consumable electrode remelting,
TiN inclusion in it will completely decompose in the solidliquid coexistence region at the electrode tip. TiN inclusion
found in ESR ingot is regenerated during solidification of
steel, not the one from electrode.
(2) The content of SiO2 in slag has an obvious effect
on Ti content in steel due to the Si/Ti exchange reaction
taking place. With the content of SiO2 in slag increasing,
Ti content in steel will decline, further lead to the low TiN
content in ESR ingot.
(3) When the content of Ti in consumable electrode
is higher than the critical equilibrium value in steel corresponding to the Ti content in slag, it will decline in steel,
and the decrease degree will rise with higher Ti content in
electrode.

Fig. 6. Effect of SiO2 content in slag on Ti content in ESR ingot.

Acknowledgements
The authors wish to express their thanks to the workers
for this production and the financial support provided by 863
Project in China (No.2012AA03A503).
Fig. 7.

Relationship between Ti content in electrode, slag and

ingot.

REFERENCES
1) A. Mitchell: Ironmaking Steelmaking, 1 (1974), 172.
2) J. Fu: Acta Metall. Sin., 15 (1979), 526.
3) C. B. Shi, X. C. Chen and H. J. Guo: Int. J. Miner. Metall. Mater.,
19 (2012), 295.
4) Y. W. Dong, Z. H. Jiang, Y. L. Cao, A. Yu and D. Hou: Metall.
Mater. Trans. B, 45B (2014), 1315.
5) C. B. Shi, X. C. Chen, H. J. Guo, Z. J. Zhu and X. L. Sun: Metall.
Mater. Trans. B, 44B (2013), 378.
6) F. R. Carmona and A. Mitchell: ISIJ Int., 32 (1992), 529.
7) C. B. Shi, X. C. Chen, H. J. Guo, Z. J. Zhu and H. Ren: Steel Res.
Int., 83 (2012), 472.
8) C. B. Shi, H. J. Guo, X. C. Chen, X. L. Sun and J. Fu: Special Steel,
34 (2013), 11.
9) J. Fu, J. Zhu, L. Di, F. S. Tong, D. L. Liu and Y. L. Wang: Acta
Metall. Sin., 36 (2000), 801.
10) H. J. Miao, Y. Wang, L. Zeng and Y. Li: Iron Steel, 48 (2013), 67.
11) S. F. Medina and A. Cores: ISIJ Int., 33 (1993), 1244.
12) L. Yang, S. J. Li, G. G. Cheng, M. Zhao and G. P. Feng: Materials
Science and Technology 2014, MS&T Administration, Pittsburgh,
PA, (2014), 379.
13) J. X. Chen: Manual of Chart and Data in Common Use of Steel Making, 2nd ed, Metallurgical Industry Press, Beijing, (2010), 510.
14) Y. Qu: Physical Chemical Calculation in Steelmaking, Metallurgical
Industry Press, Beijing, (1993), 121.
15) J. H. Wei and A. Mitchell: Acta Metall. Sin., 20 (1984), B261.
16) S. S. Babu, S. A. David, J. M. Vitek, K. Mundra and T. Debroy:
Mater. Sci. Technol., 11 (1995), 186.
17) M. E. Fraser and A. Mitchell: Ironmaking Steelmaking, 3 (1976), 279.
18) H. Goto, K. I. Miyazawa, K. I. Yamaguchi, S. Ogibayashi and K.
Tanaka: ISIJ Int., 34 (1994), 414.
19) S. F. Medina, F. Ruiz-Valdepenas, M. Chapa and A. Quispe: Steel
Res. Int., 69 (1998), 312.
20) S. F. Medina, F. Lopez and A. G. Coedo: Ironmaking Steelmaking,
24 (1997), 329.

15)

SiO2 in slag (about 6.21 10 through Eq. (11) and data

in Table 2) and high activity of Si in steel (about 1.06 by
using the data in Table 1 and Ref. 13) through the model
of Wagner).
5[ Si ] + 2(Ti3O5 ) = 6[Ti ] + 5(SiO2 ) ............. (10)
11 800
( NCaO + N MgO )
T
........ (11)
4 916
3 562

N MnO
N Al2O3
T
T

lg SiO2 = lg FeO

Where Ni is the molar percentage of slag component,

and i represents the activity coefficient of Raoult for slag
component.
Once there is a certain amount of SiO2 in slag, the Si/
Ti exchange reaction towards the formation of SiO2 will be
prevented, and with increasing content of SiO2 in slag, the
reaction will be reversed, leading to a decline in Ti of steel,
like Run E1 and E2.
Under using slag A in Run E1 and E2, Fig. 7 demonstrates that for different contents of Ti in slag, there is a
corresponding equilibrium value of Ti content in steel, with
1905

2015 ISIJ