ISIJ International, Vol. 50 (2010), No. 10, pp.

1362–1372

A Sulphide Capacity Prediction Model of CaO–SiO2–MgO–Al2O3

Ironmaking Slags Based on the Ion and Molecule Coexistence
Theory

Cheng-bin SHI,1) Xue-min YANG,2) Jin-sha JIAO,2) Chuang LI3) and Han-jie GUO3)
1) Ph.D Candidate at School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing.
Meanwhile Joint-Training Student at Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, P. R.
China. 2) State Key Lab of Multiphase Complex Systems, Institute of Process Engineering, Chinese Academy of
Sciences, Beijing 100190, P. R. China. E-mail: yangxm71@home.ipe.ac.cn. 3) School of Metallurgical and Ecological
Engineering, University of Science and Technology Beijing, Beijing 100083, P. R. China.
(Received on April 28, 2010; accepted on June 28, 2010 )

A sulphide capacity prediction model of CaO–SiO2–MgO–Al2O3 ironmaking slags has been developed
based on the ion and molecule coexistence theory (IMCT) and verified by two groups of sulphide capacity
data of CaO–SiO2–MgO–Al2O3 ironmaking slags by different researchers. A hot metal pretreatment slags of
CaO–SiO2–MgO–Al2O3 with high binary basicity is also applied to verify the feasibility of the developed
IMCT model. The predicted sulphide capacity of CaO–SiO2–MgO–Al2O3 ironmaking slags at 1 773 K as well
as high alumina CaO–SiO2–MgO–Al2O3 ironmaking slags in a temperature range of 1 773–1 873 K by the de-
veloped IMCT model has higher accuracy than the measured as well as the predicted by other sulphide ca-
pacity prediction models. The calculated equilibrium mole numbers, mass action concentrations of structural
units or ion couples and optical basicity are recommended to represent slag composition for correlating with
sulphide capacity of the slags compared with mass percentage of components or binary slag basicity. The
developed IMCT model can calculate not only the total sulphide capacity of the slags but also the respective
sulphide capacity of free CaO and MgO in the slags. Largely increasing Al2O3 content from 15 to 25 % and
decreasing CaO content from 40 to 34 %, MgO content from 9 to 4 % can improve contribution of free CaO
from 97 to 99 % while decreasing contribution of free MgO from 3 to about 1 % to the total sulphide capac-
ity of CaO–SiO2–MgO–Al2O3 ironmaking slags.
KEY WORDS: sulphide capacity; CaO–SiO2–MgO–Al2O3 ironmaking slags; sulphur distribution ratio; sul-
phide capacity model; the ion and molecule coexistence theory; blast furnace ironmaking; desulphurization
potential; mass action concentration; structural units; ion couples.

been developed since then. The sulphide capacity predic-

1. Introduction
After optical basicity concept was proposed and intro-
duced by Duffy and Ingram1) in the 1970s to describe func-
tion of all components to metallurgical slag basicity, sul-
phide capacity of slags proposed by Richardson and Fin-
cham2,3) in the 1950s to represent desulphurization ability
of metallurgical slags has attracted a lot of attentions in
worldwide metallurgical circles. Although sulphide capac-
ity of slags was defined and proposed based on slag–gas
phase equilibrium,2,3) tremendous metallurgical researchers
have widely applied sulphide capacity4–19) or sulphide ca-
pacity index20) to describe desulphurization ability of slags
based on slag–metal reactions because sulphide capacity of
slags has very close relation with slag composition and re-
action temperature.4–21) Stimulated by an empirical relation
among sulphide capacity, optical basicity and temperature
developed by Sosinsky and Sommerville4) in the 1980s,
many prediction models of sulphide capacity4–19) or sul-
phide capacity index20) for various slags have been pro-
posed or developed to predict or fit the measured sulphide
capacity of slags equilibrated with metal. Besides S–S’s
model by Sosinsky and Sommerville,4) other sulphide ca-
pacity prediction models with wide reorganization, such as
Young’s model5) and Reddy’s model8) related to optical ba-
sicity, and KTH model9–12) in terms of the specially defined
interaction coefficient of component i to j x i–j
interaction have
tion models related to optical basicity, such as Young’s
model5) and Reddy’s model,8) have been widely applied as
easy calculation, however, the underestimation of sulphide
capacity by those models has been observed by some
researchers.6,7,16) Meanwhile, KTH model9–12) proposed by
researchers at the Department of Metallurgy, Royal of
Institute of Technology (KTH), must need x i–j interaction, while
x i–j
interaction for some components in slags is absent as meas-
urement difficulty.10) Under those circumstances, Shankar et
al.6) and Taniguchi et al.7) have recently proposed two mod-
els to accurately predict sulphide capacity of high alumina
CaO–SiO2–MgO–Al2O3 slags equilibrated with gas by
modifying the above-mentioned sulphide capacity mod-
els5,8) related to optical basicity. However, whether the new
model by Shankar et al.6) or by Taniguchi et al.7) can be ap-
plied to accurately predict sulphide capacity of other slags
should be further verified and proved.
Like sulphide capacity, sulphur distribution ratio between
slag and metal has been successfully applied and widely
recognized as a common parameter to depict desulphuriza-
tion ability of slags. The sulphur distribution ratio predic-
tion models13–15) have been developed by some empirical
formulas or indirectly proposed from sulphide capacity pre-
diction models to describe desulphurization ability of slags
from metal. However, the developed sulphur distribution
ratio prediction models13–15) can only be successfully ap-

© 2010 ISIJ 1362

 ISIJ International, Vol. 50 (2010), No. 10
plied to some limited slags without widespread acceptance
from viewpoint of metallurgical physicochemistry. It is nec-
essary and interesting to develop a sulphide capacity pre-
diction model according to intrinsic relation of sulphide ca-
pacity and sulphur distribution ratio from new viewpoint.
A sulphide capacity prediction model of CaO–SiO2–
MgO–Al2O3 ironmaking slags has been developed accord-
ing to the ion and molecule coexistence theory22–26) (IMCT)
developed by J. Zhang, i.e., IMCT model. The developed
IMCT model for predicting sulphide capacity has been ver-
ified by two groups of measured sulphide capacity data of
CaO–SiO2–MgO–Al2O3 ironmaking slags by different re-
searchers6,22): one is of CaO–SiO2–MgO–Al2O3 slags equil-
ibrated with hot metal at 1 773 K,22) the other is of high alu-
mina CaO–SiO2–MgO–Al2O3 slags equilibrated with gas
containing sulphur in a temperature range of 1 773–
1 873 K.6) Meanwhile, the IMCT model has also been veri-
fied by comparing with calculated sulphide capacities of
above-mentioned two slags by five other models as S–S’s
model,4) Young’s model,5) Shankar’s model,6) Taniguchi’s
model7) and KTH model.9–12) The relationships among pre-
dicted sulphide capacity by IMCT model, optical basicity,
binary basicity, mass percentage of components, calculated
equilibrium mole numbers and mass action concentrations
of structural units or ion couples based on IMCT22–26) of
two slags are also applied as evidences to verify the devel-
oped IMCT model. The ultimate objective of this study is
not only to develop a sulphide capacity prediction model of
CaO–SiO2–MgO–Al2O3 blast furnace (BF) ironmaking
slags in terms of IMCT,22–26) but also to provide a universal
method to propose a sulphide capacity prediction model for
various metallurgical slags.
2. Sulphide Capacity Data of CaO–SiO2–MgO–Al2O3
Ironmaking Slags
Two groups of sulphide capacity CS2⫺ data of CaO–SiO2–
MgO–Al2O3 ironmaking slags have been applied in this
study to develop and verify the IMCT model. One group
CS2⫺ data were measured from CaO–SiO2–MgO–Al2O3
ironmaking slags equilibrated with carbon saturated hot
metal at 1 773 K by the present authors,22) the other were
measured from high alumina CaO–SiO2–MgO–Al2O3 iron-
making slags equilibrated with gas in a temperature range
Table 1. Chemical compositions of CaO–SiO2–MgO–Al2O3 ironmaking slags equilibrated with carbon saturated hot metal at 1 773 K,
measured sulphide capacity from tested sulphur distribution ratio by Yang et al.22) and predicted sulphide capacity of the slags
by the developed IMCT model.
1363
of 1 773–1 873 K by Shankar et al.6) Two groups of CS2⫺
data of CaO–SiO2–MgO–Al2O3 ironmaking slags applied
in this study were repeatedly given out in Table 1 and
Table 2 for easy understanding and comparison. It should
be emphasized that CS2⫺,measured listed in Table 1 was calcu-
lated from LS,measured22) of CaO–SiO2–MgO–Al2O3 ironmak-
ing slags at 1 773 K according to the relation between CS2⫺
and LS described in Sec. 3.1 as well as elsewhere.14,19)
3. Sulphide Capacity Model of CaO–SiO2–MgO–
Al2O3 Ironmaking Slags Based on IMCT
3.1. Relationship between Sulphur Distribution Ratio
and Sulphide Capacity
To develop the CS2⫺ prediction model based on
IMCT22–26) from the developed LS prediction model22) be-
tween CaO–SiO2–MgO–Al2O3 quaternary slags and hot
metal at 1 773 K, it is necessary to briefly introduce the re-
lationship between LS and CS2⫺ from original definition of
CS2⫺ of a slag. It is well known that CS2⫺ of a slag was firstly
proposed under equilibrium condition of a slag without any
sulphur and gas phase containing sulphur by Richardson
and Fincham2,3) as follows
1 1
S2 (g) ⫹ (O2⫺ ) ⫽ O2 (g) ⫹ (S2⫺ )
2 2 ......(1)
Θ
Δ r Gm,S 2⫺ ⫽ 118 535 ⫺ 58.8157T
10 ,19 )
(J/mol)
The CS2⫺ defined by Richardson and Fincham2,3) can be de-
scribed as
pO2
CS2⫺⫽ (%S) (⫺) .......................(2)
pS2
However, the desulphurization reaction between slag and
metal can be traditionally written by ion exchange reaction
as
[S]⫹(O2⫺)⫽(S2⫺)⫹[O] ........................(3)
The following reaction can be obtained by combining Eq.
(1) and Eq. (3)
© 2010 ISIJ
 ISIJ International, Vol. 50 (2010), No. 10

Table 2. Chemical compositions of high alumina CaO–SiO2–MgO–Al2O3 ironmaking slags equilibrated with gas containing sulphur,
measured sulphide capacity by Shankar et al.,6) and calculated sulphide capacity by both of the developed IMCT model and
Shankar’s model.

1 1 Obviously, the uneasily measured CS2⫺ of slags equilibrated

[S] ⫹ O2 (g) ⫽ S2 (g) ⫹ [O] with gas can be calculated from the easily measured LS of
2 2 ....(4) slags equilibrated with metal by Eq. (7).
Θ
Δ r Gm,S ⫽ 17 907.96 ⫺ 26.33524T 14,19 ) (J/mol)
2
3.2. Sulphide Capacity Model Based on IMCT
The reaction equilibrium constant K QS2 of Eq. (4) can be According to the reported LS prediction model22) between
written as14,19) CaO–SiO2–MgO–Al2O3 slags and hot metal based on
IMCT,22–26) the desulphurization reactions between CaO–
⎛a pS2 ⎞ ⎛ a pS2 ⎞ SiO2–MgO–Al2O3 slags and metal can be presented as fol-
lg KSΘ ⫽ lg ⎜ O ⎟ ⫽ lg ⎜
O
⎟ lows
2 ⎜ aS pO2 ⎟⎠ ⎜ [%S] f S pO2 ⎟⎠
⎝ ⎝
(Ca 2⫹ ⫹ O2⫺ ) ⫹ [S] ⫽ (Ca 2⫹ ⫹ S2⫺ ) ⫹ [O]
935 ...(8a)
Θ
⫽⫺ ⫹ 1.375 (⫺) ................................(5) Δ r Gm,CaS ⫽ 105 784.6 ⫺ 28.723T 27,28 ) (J/mol)
T
(Mg 2⫹ ⫹ O2⫺ ) ⫹ [S] ⫽ (Mg 2⫹ ⫹ S2⫺ ) ⫹ [O]
The following equation can be deduced by inserting CS2⫺ ...(8b)
Θ
definition presented in Eq. (2) into Eq. (5) as Δ r Gm,MgS ⫽ 203 604.6 ⫺ 35.023T 27,29 ) (J/mol)
⎛ (%S) aO ⎞ The respective sulphur distribution ratio of ion couple
lg KSΘ ⫽ lg ⎜ ⫻
2
⎝ [%S] f SCS2⫺ ⎟⎠ (Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺), i.e., LS,CaO and LS,MgO, of
CaO–SiO2–MgO–Al2O3 slags equilibrated with metal in
⫽ lg LS ⫹ lg aO ⫺ lg f S ⫺ lg CS2⫺ terms of IMCT22–26) can be expressed by22)

∑n
935 Θ
⫽⫺ ⫹ 1.375 (⫺) .............(6) (%S)CaS 16 K CaS N CaO i fS
T LS,CaO ⫽ ⫽ ⫻ (⫺) ....(9a)
[%S] [%O] fO
Therefore, the relationship between LS and CS2⫺ of a slag
can be obtained from Eq. (6) as14,19) LS,MgO ⫽
(%S)MgS
⫽
Θ
16 K MgS N MgO ∑n i
⫻
fS
(⫺)
[%S] [%O] fO
935
lg CS2⫺⫽ lg LS ⫹ ⫺ 1.375 ⫺ lg f S ⫹ lg aO (⫺) ....(7) ..............................................(9b)
T

© 2010 ISIJ 1364

 ISIJ International, Vol. 50 (2010), No. 10

Therefore, LS of CaO–SiO2–MgO–Al2O3 slags equili- Generally speaking, it is necessary to know some param-
brated with metal can be descried as follows22) eters to calculate CS2⫺,CaO, CS2⫺,MgO and CS2⫺ of CaO–
SiO2–MgO–Al2O3 ironmaking slags by applying Eq. (12),
(%S)CaS⫹MgS (%S)CaS ⫹ (%S)MgS such as reaction equilibrium constant KQCaS and KQMgS of Eq.
LS ⫽ LS,CaO ⫹ LS,MgO ⫽ ⫽ (8), mass action concentration NCaO and NMgO of ion couple
[%S] [%S] (Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺), and total equilibrium mole

⫽
Θ
16( K CaS Θ
N CaO ⫹ K MgS N MgO ) ∑n i
⫻
fS
(⫺) .....(10)
number Í ni of all structural units in 100 g of the slags. For-
tunately, NCaO, NMgO and Í ni can be calculated from the de-
[%O] fO veloped thermodynamic model22) for calculating mass ac-
tion concentration of all structural units in the slags based
Although inserting the representations of LS,CaO and on IMCT.22–26) Meanwhile, KQCaS and KQMgS can also be cal-
LS,MgO in Eq. (9) or LS in Eq. (10) into Eq. (7), both terms of culated by choosing accurate and authoritative formulas of
fS and aO in Eq. (7) can be deleted. However, the activity standard molar Gibbs free energy D rGQm,i of reactions for
coefficient of sulphur and oxygen in hot metal, fS and fO, in forming component i shown in Eq. (8).
Eq. (6), Eq. (8) and Eq. (9) are needed to know and can be
calculated by the Wagner’s equation as follows
4. Evaluation of Other Sulphide Capacity Models
lg f i ⫽ ∑ eij [% j] (⫺) .....................(11) To verify the developed IMCT model for prediction CS2⫺
of CaO–SiO2–MgO–Al2O3 ironmaking slags, it is neces-
The values of activity interaction coefficients of element j sary to compare and evaluate other CS2⫺ prediction models.
to i, eij, at the related temperatures can be found in previous The five CS2⫺ prediction models with high acceptance have
publication22) by authors. been summarized in Table 3, such as Young’s model,5)
The respective sulphur capacity of ion couples Shankar’s model,6) Taniguchi’s model7) based on optical
(Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺), and CaO–SiO2–MgO– basicity L and KTH model9–12) in terms of defined interac-
Al2O3 slags, i.e., CS2⫺,CaO, CS2⫺,MgO and CS2⫺, can be pre- tion coefficient of component i to j x interaction
i–j
, while the early
sented according to IMCT22–26) as proposed S–S’s model by Sosinsky and Sommerville4) is
also listed in Table 3 for comparison.
935
Θ
lg CS2⫺,CaO ⫽ lg(16 K CaS N CaO ∑ n )⫹ i
T
⫺ 1.375 (⫺) 4.1. Evaluation of Sulphide Capacity Models Based on
.......................................(12a) Optical Basicity
Although an CS2⫺ empirical formula containing L and
935
Θ
lg CS2⫺,MgO ⫽ lg(16 K MgS N MgO ∑ ni ) ⫹
T
⫺ 1.375 (⫺) temperature T had been firstly developed by Sosinsky and
Sommerville4) in the 1980s for some binary, ternary and
.......................................(12b) quaternary slags in a temperature range of 1 673–1 973 K,
Young et al.5) had found from experimental study that S–S’s
Θ
lg CS2⫺ ⫽ lg[16( K CaS Θ
N CaO ⫹ K MgS N MgO ) ∑n] i model4) can only be applied to slags in low range of CS2⫺,
while there is a large deviation of predicted CS2⫺ of slags in
935
⫹ ⫺ 1.375 (⫺) ..........................(12c) high CS2⫺ range. Therefore, Young et al.5) proposed the
T other CS2⫺ prediction model with better applicability in a
Table 3. Comparison of related sulphide capacity prediction models with the developed sulphide capacity prediction model based on
IMCT.

1365 © 2010 ISIJ

 ISIJ International, Vol. 50 (2010), No. 10
large L range from their experimental data as shown in
Table 3.
However, Shankar et al.6) had found that the predicted
CS2⫺ by Young’s model5) as well as by KTH model9–12) was
much smaller than measured CS2⫺,measured of some slags.
A new CS2⫺ prediction model was proposed by Shankar
et al.6) as listed in Table 3 for slags CaO(30–46%)–
SiO2(30–40%)–MgO(2–10%)–Al2O3(12–30%) in a temper-
ature range of 1 773–1 873 K.
Taniguchi et al.7) had also found the same results
reported by Shankar et al.,6) therefore, Taniguchi et al.7)
proposed another CS2⫺ prediction model as listed in Table
3 for slags CaO(10–63%)–SiO2(0–68%)–MgO(0–15%)–
Al2O3(0–65%)–MnO(0–30%) in a temperature range of
1 673–1 928 K.
In addition, the effects of SiO2 and Al2O3 as components
on CS2⫺ are also contrary from different researchers, such as
both components of SiO2 and Al2O3 have negative effect on
CS2⫺ in Young’s model5) and KTH model9–12); while both
components of SiO2 and Al2O3 show an active influence on
CS2⫺ in Taniguchi’s model.7) The effect of structural unit
SiO2 or Al2O3 on CS2⫺ can not be directly presented in the
developed IMCT model because no any desulphurization
reactions at slag–metal interface can be written using
(SiO2) or (Al2O3) with [S] according to the classical metal-
lurgical physicochemistry; while, their effects on CS2⫺ can
be potentially embodied in the calculated NCaO, NMgO and
Í ni of the slags in Eq. (12).
4.2. Evaluation of KTH Sulphide Capacity Model
The KTH model9–12) was independently proposed accord-
ing to the specially defined interaction coefficient of com-
ponent i to j x i–j
interaction. The KTH model
9–12)
for the investi-
gated CaO–SiO2–MgO–Al2O3 ironmaking slags is also
listed in Table 3. Certainly, applying KTH model9–12) re-
quires to know or measure x i–j interaction. However, considering
experimental difficulties and other factors, some x i–j interaction
can not be measured or estimated accurately. It is the major
reason to limit application of KTH model to some slags.
5. Verification and Discussion for Sulphide Capacity
Prediction Model Based on IMCT
5.1. Comparison of Calculated Sulphide Capacity of
CaO–SiO2–MgO–Al2O3 Ironmaking Slags at
1 773 K by Different Models
The comparison of CS2⫺,calculated by six models summa-
rized in Table 3 and CS2⫺,measured transferred via LS,measured22)
of CaO–SiO2–MgO–Al2O3 ironmaking slags at 1 773 K is
shown in Fig. 1(a). Only the predicted C IMCT
S2⫺,calculated by IMCT
model has good linear relation with CS2⫺,measured except one
points of C IMCT
S2⫺,calculated is smaller than CS2⫺,measured in high
range of CS2⫺,measured. The similar result is also found to
Fig. 1. Comparison of calculated sulphide capacity by six models and measured sulphide capacity of
CaO–SiO2–MgO–Al2O3 slags equilibrated with hot metal at 1 773 K by Yang et al.22) (a), high alumina
CaO–SiO2–MgO–Al2O3 slags equilibrated with gas containing sulphur in a temperature range of 1 773–1 873 K
by Shankar et al.6) (b), and CaO–SiO2–MgO–Al2O3 hot metal pretreatment slags equilibrated with gas containing
sulphur in a temperature range of 1 673–1 773 K by Taniguchi et al.7) (c), respectively.
© 2010 ISIJ
Lcalculated in previous publication22) based on IMCT22–26) of
the same slags.
Obviously, C S–S 4) Young
S2⫺,calculated by S–S’s model, C S2⫺,calculated by
Taniguchi
Young’s model, CS2⫺,calculated by Taniguchi’s model7) and
5)
C KTH
S2⫺,calculated by KTH model
9–12) are much smaller than
CS2⫺,measured of CaO–SiO2–MgO–Al2O3 ironmaking slags at
1 773 K. Meanwhile, CSShankar 2⫺, calculated by Shankar’s model
6)
show
a good relation with CS2⫺,measured in a low CS2⫺ range as
CS2⫺,measuredⱕ8⫻10⫺4, but show an underestimation in a
high CS2⫺ range as CS2⫺,measuredⱖ8⫻10⫺4 of the slags. The
underestimated CS2⫺,calculated of CaO–SiO2–MgO–Al2O3 iron-
making slags at 1 773 K by S–S’s model,4) Young’s model5)
as well as KTH model9–12) is similar to that found by
Shankar et al.6) and Taniguchi et al.7) for their slags, respec-
tively. Therefore, it can be deduced that S–S’s model,4)
Young’s model5) and KTH model9–12) can not be applied to
accurately predict CS2⫺ of CaO–SiO2–MgO–Al2O3 ironmak-
ing slags. In addition, the interaction coefficient x i–j interaction
applied in KTH model9–12) is determined or calculated from
CS2⫺,measured with known composition of a slag at a constant
temperature, experimental deviations, regression deviations
and non-uniform of measured slag composition can affect
accuracy of the regressed x i–j interaction in KTH model,
9–12)
fur-
thermore, negatively influence the accuracy of results from
KTH model.9–12) However, it is strange with some unknown
reasons that Taniguchi’s model7) can not be applied to per-
fectly predict CS2⫺ of the investigated CaO–SiO2–MgO–
Al2O3 ironmaking slags.
Therefore, the developed IMCT model based on
IMCT22–26) can be reliably applied to predict CS2⫺ of
CaO–SiO2–MgO–Al2O3 ironmaking slags at 1 773 K com-
pared with other five models listed in Table 3.
5.2. Comparison of Calculated Sulphide Capacity of
High Alumina CaO–SiO2–MgO–Al2O3 Ironmak-
ing Slags in a Temperature Range from 1 773 to
1 873 K by Different Models
To calculate CS2⫺ of high alumina CaO–SiO2–MgO–
Al2O3 ironmaking slags in a temperature range of
1 773–1 873 K by IMCT model, the mass action concentra-
tions of all structural units in the high alumina ironmaking
slags as shown in Table 26) should be firstly calculated by
using the developed thermodynamic model22) for calculat-
ing Ni and Í ni of the slags. Other related parameters, such
as KQCaS and KQMgS, can also be calculated as descried in de-
tail in Sec. 3.2.
The comparison of predicted CS2⫺,calculated by six models
listed in Table 3 and CS2⫺,measured6) of high alumina CaO–
SiO2–MgO–Al2O3 slags in a temperature range of 1 773–
1 873 K is shown in Fig. 1(b). The calculated CS2⫺,calculated by
S–S’s model4) or Young’s model5) or Taniguchi’s model7) or
by KTH model9–12) is much smaller than CS2⫺,measured6) of the
high alumina ironmaking slags. Obviously, not only
1366
 ISIJ International, Vol. 50 (2010), No. 10

Table 4. Chemical compositions of CaO–SiO2–MgO–Al2O3 hot metal pretreatment slags equilibrated with gas containing sulphur,
measured sulphide capacity by Taniguchi et al.,7) and calculated sulphide capacity by both of the developed IMCT model and
Taniguchi’s model.

Shankar’s model6) but also the developed IMCT model in

this study can be successfully applied to predict CS2⫺ of the
high alumina ironmaking slags. It should be emphasized
that C SIMCT
2⫺,calculated by IMCT model and CSShankar
2⫺, calculated by
6)
Shankar’s model show a little underestimation for the high
alumina ironmaking slags under condition of CS2⫺,measuredⱖ
8⫻10⫺4.
To further verify the application range of the developed
IMCT model, a hot metal pretreatment slags of CaO–
SiO2–MgO–Al2O3 with high binary basicity as 8.0 has also
been chosen to predict CS2⫺ by the developed IMCT model
and other models. The predicted CS2⫺,calculated by six models Fig. 2. Relationship between optical basicity or binary basicity
and CS2⫺,measured7) of CaO–SiO2–MgO–Al2O3 hot metal pre- and calculated sulphide capacity by IMCT model or
treatment slags equilibrated with gas containing sulphur in measured sulphide capacity of CaO–SiO2–MgO–Al2O3
slags equilibrated with hot metal at 1 773 K by Yang et
a temperature range of 1 673–1 773 K by Taniguchi et al.7) al.,22) respectively.
in Table 4 is also illustrated in Fig. 1(c). Obviously, all the
six models listed in Table 3 can be successfully applied to
predict CS2⫺ of the high basicity hot metal pretreatment respectively. It can be obviously observed from Fig. 2 that
slags. 1) the relation between C IMCT S2⫺,calculated by IMCT model or
Therefore, the developed IMCT model can be success- CS2⫺,measured and L has a higher correlativity than that with B
fully applied to predict CS2⫺ of the investigated CaO– of the slags; 2) C IMCT S2⫺,calculated by IMCT model has good rela-
SiO2–MgO–Al2O3 ironmaking slags by the present tion with L , while CS2⫺,measured has a much scattered relation
authors,22) the reported high alumina CaO–SiO2–MgO– with L of the slags as shown in Fig. 2(b). These results
Al2O3 ironmaking slags by Shankar et al.6) and CaO–SiO2– show that the developed IMCT model can accurately pre-
MgO–Al2O3 hot metal pretreatment slags with high binary dict CS2⫺ because CS2⫺,measured could be affected by superpo-
basicity by Taniguchi et al.7) The application range of the sition of various experimental errors, such as chemical
developed IMCT model for ironmaking or hot metal pre- analysis deviations of slags and hot metal, sampling error
treatment slags can be summarized as CaO(28–55%)– as non-uniform of slag composition and so on.
SiO2(7–40%)–MgO(1–12%)–Al2O3(10–40%) with binary The relation of C IMCT S2⫺,calculated by IMCT model or
basicity as 0.73–8.0 as well as sulphide capacity in a range C S2⫺,calculated by Shankar’s model6) or CS2⫺,measured6) against L
Shankar

of 1⫻10⫺4–19⫻10⫺4. To focus more attention of the devel- and B of high alumina CaO–SiO2–MgO–Al2O3 ironmaking
oped IMCT model on BF ironmaking slags in this paper, slags in a temperature range of 1 773–1 873 K is illustrated
the related results of CaO–SiO2–MgO–Al2O3 hot metal pre- in Fig. 3, respectively. It can be observed from Fig. 3 that 1)
treatment slags by the developed IMCT model in Table 4 the corresponding relation of C IMCT S2⫺,calculated by IMCT model
will not be discussed in detail in the following text. or C SShankar
2⫺, calculated by Shankar’s model
6) or C 2⫺ 6)
S ,measured against
B is to some degree scattered than that against L for the
5.3. Relation between Optical Basicity and Sulphide high alumina ironmaking slags; 2) a good corresponding
Capacity of CaO–SiO2–MgO–Al2O3 Ironmaking relation between CS2⫺,calculated by both IMCT model and
Slags Shankar’s model6) or CS2⫺,measured6) and L can be found in
It has been repeatedly proved that optical basicity L has Fig. 3(b) for the high alumina ironmaking slags at 1 773 K,
very good corresponding relation with CS2⫺ of various met- however, improving temperature from 1 773 to 1 873 K will
allurgical slags from models listed in Table 3. It is a power- make an obvious difference of the plotted relations as
ful proof to further verify correctness and feasibility of the shown in Figs. 3(c)–3(d); 3) C IMCT S2⫺,calculated by IMCT model is
developed IMCT model according to whether the predicted almost the same with C SShankar 2⫺, calculated by Shankar’s model
6)
at
C IMCT
S2⫺,calculated by IMCT model of both CaO–SiO2–MgO– various L for the high alumina slags at 1 823 K, but they
Al2O3 ironmaking slags at 1 773 K22) and high alumina are a little smaller than CS2⫺,measured6); 4) there is a visible
CaO–SiO2–MgO–Al2O3 ironmaking slags in a temperature difference among C IMCT Shankar
S2⫺,calculated by IMCT model, C S2⫺,calculated
range of 1 773–1 873 K6) has a good relation with L . 6)
by Shankar’s model and CS2⫺,measured of the high alumina 6)

The relationship between C IMCTS2⫺,calculated by IMCT model or ironmaking slags with a fixed L at 1 873 K, CS2⫺,measured6) is
CS2⫺,measured transferred via LS,measured22) and L or binary slag greater than C IMCT S2⫺,calculated by IMCT model, but smaller than
basicity B, i.e., B⫽(%CaO)/(%SiO2), of CaO–SiO2–MgO– C SShankar
2⫺, calculated by Shankar’s model
6)
in a wide range of L for
Al2O3 ironmaking slags at 1 773 K is illustrated in Fig. 2, the high alumina ironmaking slags at 1 873 K; 5) fitting de-

1367 © 2010 ISIJ

 ISIJ International, Vol. 50 (2010), No. 10

Fig. 3. Relationship between optical basicity or binary basicity
and calculated sulphide capacity by IMCT model or by
Shankar’s model or measured sulphide capacity of high
alumina CaO–SiO2–MgO–Al2O3 slags equilibrated with
gas containing sulphur in a temperature range of
1 773–1 873 K by Shankar et al.,6) respectively.
Fig. 4. Relationship between mass percentage of components for
CaO, MgO, SiO2 and Al2O3, and calculated sulphide ca-
pacity by IMCT model or measured sulphide capacity of
CaO–SiO2–MgO–Al2O3 slags equilibrated with hot metal
at 1 773 K by Yang et al.,22) respectively.

S2⫺,calculated by IMCT model and L

gree of relation between C IMCT
is higher than relation between C SShankar 2⫺, calculated by Shankar’s

model6) or CS2⫺,measured6) and L for the high alumina iron-

making slags at 1 773 K, 1 823 K and 1 873 K.
The fitting degree of corresponding relation between
S2⫺,calculated by IMCT model and L for the two CaO–
C IMCT
SiO2–MgO–Al2O3 ironmaking slags is higher than that of
relation between CS2⫺,measured and L , respectively. Therefore,
it can be deduced that the developed IMCT model can pre-
dict more accurate sulphide capacity of BF ironmaking
slags with various compositions at different temperatures
than other models listed in Table 3.
5.4. Relation between Component Content and Sul-
phide Capacity of CaO–SiO2–MgO–Al2O3 Iron-
making Slags
5.4.1. Relation between Component Content and Sulphide
Capacity of CaO–SiO2–MgO–Al2O3 Ironmaking
Slags at 1 773 K
The relationship between mass percentage of CaO, SiO2,
MgO and Al2O3 as four components and both C IMCT S2⫺,calculated
by IMCT model and CS2⫺,measured22) of CaO–SiO2–MgO– Fig. 5. Relationship between calculated equilibrium mole num-
Al2O3 ironmaking slags at 1 773 K is shown in Fig. 4, re- bers of structural units or ion couples (Ca2⫹⫹O2⫺),
(Mg2⫹⫹O2⫺), (SiO2) and (Al2O3), and calculated sul-
spectively. Although all relations in Fig. 4 are to some de- phide capacity by IMCT model or measured sulphide ca-
gree scattered, the tendency of increasing content of CaO pacity of CaO–SiO2–MgO–Al2O3 slags equilibrated with
and MgO as well as deceasing content of SiO2 and Al2O3 hot metal at 1 773 K by Yang et al.,22) respectively.
can effectively promote both C IMCTS2⫺,calculated by IMCT model
and CS2⫺,measured22) of the slags; meanwhile, the correspon- (Al2O3) has better corresponding relation with CS2⫺ than
ding relations of mass percentage of four components and mass percentage of components in the slags as shown
CS2⫺,measured22) are more scattered than that with C IMCT S2⫺,calculated in Fig. 4. In addition, very ideally corresponding relation
by IMCT model. This means that experimental errors and can be observed between ni22) of structural units or ion cou-
other unknown reasons will deteriorate the possibly existed ples (Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and (Al2O3) and
relation of CS2⫺ and mass percentage of components in the C IMCT
S2⫺,calculated by IMCT model compared with a scattered re-
slags. lation between ni22) of structural units or ion couples
The relationship between the calculated equilibrium (Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and (Al2O3) and
mole number ni22) of structural units or ion couples CS2⫺,measured22) of the slags. This result further suggests that
(Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and (Al2O3) in the slags C IMCT
S2⫺,calculated by the developed IMCT model is more reliable
in terms of IMCT22–26) and both C IMCT S2⫺,calculated by IMCT than CS2⫺,measured22) of CaO–SiO2–MgO–Al2O3 ironmaking
model and CS2⫺,measured22) of the ironmaking slags at 1 773 K slags at 1 773 K.
is shown in Fig. 5, respectively. Surely, ni22) of structural The relationship between the calculated mass action
units or ion couples (Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and concentrations Ni22) of structural units or ion couples

© 2010 ISIJ 1368

 ISIJ International, Vol. 50 (2010), No. 10
(Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and (Al2O3) in the
slags22) in terms of IMCT22–26) and both C IMCT S2⫺,calculated
by IMCT model and CS2⫺,measured22) of the ironmaking slags
at 1 773 K is shown in Fig. 6, respectively. Obviously,
C IMCT
S2⫺,calculated by IMCT model and CS2⫺,measured
22)
of the slags
at 1 773 K has good corresponding relation with Ni22) of
structural units or ion couples (Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺),
(SiO2) and (Al2O3) like ni22) as shown in Fig. 6. It can be
observed by comparing Fig. 5 and Fig. 6 that increasing
NCaO and NMgO or 2nCaO and 2nMgO shows a promotion
tendency of both C IMCT S2⫺,calculated and CS2⫺,measured
22)
of the slags
at 1 773 K, respectively. This result can be explained by
the developed IMCT model depicted by Eq. (12c) that in-
creasing NCaO and NMgO or 2nCaO and 2nMgO can result in
improving of C IMCT S2⫺,calculated and CS2⫺,measured
22)
of the slags; op-
positely, increasing NSiO2 and NSiO2 or nSiO2 and nAl2O3 show
an obvious decreasing trend for both C IMCT S2⫺,calculated and
CS2⫺,measured22) of the slags because the sum of all Ni is
equal to 1.0 or Í ni is constant for a given amount of the
slags with a fixed composition, increasing NSiO2 and NAl2O3
or nSiO2 and nAl2O3 corresponds to decreasing NCaO and NMgO
or 2nCaO and 2nMgO in the slags.
5.4.2. Relation between Component Content and Sulphide
Capacity of High Alumina CaO–SiO2–MgO–Al2O3
Ironmaking Slags in a Temperature Range from
1 773 to 1 873 K
The relationship between mass percentage of CaO, SiO2,
MgO and Al2O3 as four components and C IMCT S2⫺,calculated by
IMCT model or C SShankar 2⫺, calculated by Shankar’s model6) or
CS2⫺,measured6) of high alumina CaO–SiO2–MgO–Al2O3 iron-
making slags in a temperature range of 1 773–1 873 K is
shown in Fig. 7, respectively. It can be found from Fig. 7
that 1) a scattered relation between mass content of CaO or
SiO2 and C IMCT Shankar
S2⫺,calculated by IMCT model or C S2⫺,calculated by
Shankar’s model6) or CS2⫺,measured6) of the high alumina iron-
making slags in a temperature range of 1 773–1 873 K can
be observed, which is similar to results shown in Fig. 4(a)
and Fig. 4(b) of CaO–SiO2–MgO–Al2O3 ironmaking slags
at 1 773 K; 2) no obvious corresponding relation between
mass content of MgO and C IMCT S2⫺,calculated by IMCT model or
Fig. 6. Relationship between calculated mass action concentra-
tion of structural units or ion couples (Ca2⫹⫹O2⫺),
(Mg2⫹⫹O2⫺), (SiO2) and (Al2O3), and calculated sulphide
capacity by IMCT model or measured sulphide capacity
of CaO–SiO2–MgO–Al2O3 slags equilibrated with hot
metal at 1 773 K by Yang et al.,22) respectively.
1369
C SShankar
2⫺, calculated by Shankar’s model
6) or C 2⫺ 6)
S ,measured of the high
alumina ironmaking slags can be established in Fig 7(c),
which is not in agreement with that illustrated in Fig 4(c)
of CaO–SiO2–MgO–Al2O3 ironmaking slags at 1 773 K;
3) a fuzzy relation between mass content of Al2O3 and
C IMCT Shankar
S2⫺,calculated by IMCT model or C S2⫺,calculated by Shankar’s
6) 6)
model or CS2⫺,measured can be observed in Fig. 7(d) for the
high alumina ironmaking slags; 4) the fitting degree of rela-
tion between C IMCT S2⫺,calculated by IMCT model and mass per-
centage of CaO or SiO2 or Al2O3 is higher than that with
C SShankar
2⫺, calculated by Shankar’s model
6)
or CS2⫺,measured6) of the high
alumina ironmaking slags as shown in Fig. 7.
The difference of relation between mass percentage of
MgO or Al2O3 and CS2⫺ of two ironmaking slags shown in
Fig. 4 and Fig. 7 can be explained from change of chemical
compositions for two slags. Comparing chemical composi-
tions of two slags listed in Table 1 and Table 2, the average
content of MgO and Al2O3 in the high alumina ironmaking
slags is 4% and 25%, while it is about 9% and 15% in
CaO–SiO2–MgO–Al2O3 ironmaking slags, respectively.
The lower content of MgO and higher content of Al2O3 in
the high alumina ironmaking slags will weaken the effect of
MgO and Al2O3 content on CS2⫺ and decrease desulphuriza-
tion potential or CS2⫺ of the high alumina ironmaking slags.
It should be pointed out that higher Al2O3 content with
lower MgO content will decrease fluidity of the high alu-
mina ironmaking slags. Therefore, results illustrated in Fig.
(4) and Fig. (7) are important and interesting to optimize
slag composition for possessing better fluidity as well as
higher desulphurization potential or higher sulphide capac-
ity during BF ironmaking process.
The relationship between the calculated ni of structural
units or ion couples (Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and
(Al2O3) in terms of IMCT22–26) and C IMCT S2⫺,calculated by IMCT
model or C SShankar 2⫺, calculated by Shankar’s model
6) or C 2⫺
S ,measured
6)
of high alumina CaO–SiO2–MgO–Al2O3 ironmaking slags
in a temperature range of 1 773–1 873 K is shown in Fig. 8,
respectively. Surely, ni of structural units or ion couples
(Ca2⫹⫹O2⫺), (SiO2) and (Al2O3) has better fitting degree
than mass percentage of components with CS2⫺ in the high
alumina ironmaking slags as shown in Fig. 7. A fuzzy cor-
Fig. 7. Relationship between mass percentage of components for
CaO, MgO, SiO2, and Al2O3, and calculated by IMCT
model or by Shankar’s model or measured sulphide ca-
pacity of high alumina CaO–SiO2–MgO–Al2O3 slags
equilibrated with gas containing sulphur in a temperature
range of 1 773–1 873 K by Shankar et al.,6) respectively.
© 2010 ISIJ
 ISIJ International, Vol. 50 (2010), No. 10
Fig. 8. Relationship between calculated equilibrium mole num-
bers of structural units or ion couples (Ca2⫹⫹O2⫺),
(Mg2⫹⫹O2⫺), (SiO2) and (Al2O3), and calculated sul-
phide capacity by IMCT model or by Shankar’s model or
measured sulphide capacity of high alumina CaO–
SiO2–MgO–Al2O3 slags equilibrated with gas containing
sulphur in a temperature range of 1 773–1 873 K by
Shankar et al.,6) respectively.
responding relation between 2nMgO or n(Mg2⫹⫹O2⫺) and CS2⫺,
in which increasing 2nMgO or n(Mg2⫹⫹O2⫺) has positive effect
to promote CS2⫺, can be observed in Fig. 8(c) for the high
alumina ironmaking slags except one point with the largest
2nMgO or n(Mg2⫹⫹O2⫺) as 0.27. In addition, very ideally corre-
sponding relations can be observed between ni of structural
units or ion couples (Ca2⫹⫹O2⫺), (SiO2) and (Al2O3) and
C IMCT
S2⫺,calculated by IMCT model compared with a scattered re-
lation between ni of structural units or ion couples
(Ca2⫹⫹O2⫺), (SiO2) and (Al2O3) and C SShankar 2⫺, calculated
Shankar’s model6) or CS2⫺,measured6) of the high alumina iron-
making slags. This result further suggests that C IMCT S2⫺,calculated
by IMCT model is more reliable than C SShankar 2⫺, calculated by
Shankar’s model6) as well as CS2⫺,measured6) of the high alu-
mina ironmaking slags in a temperature range of 1 773–
1 873 K.
The relationship between Ni of structural units or ion
couples (Ca2⫹⫹O2⫺), (Mg2⫹⫹O2⫺), (SiO2) and (Al2O3) in
terms of IMCT22–26) and C IMCT S2⫺,calculated by IMCT model or
C SShankar
2⫺, calculated by Shankar’s model
6)
or CS2⫺,measured6) of high
alumina CaO–SiO2–MgO–Al2O3 ironmaking slags in a
temperature range of 1 773–1 873 K is shown in Fig. 9,
respectively. Obviously, C IMCT S2⫺,calculated by IMCT model or
C SShankar
2⫺, calculated by Shankar’s model
6) or C 2⫺ 6)
S ,measured has good
corresponding relation with Ni of structural units or ion
couples (Ca2⫹⫹O2⫺), (SiO2) and (Al2O3) in the high alu-
mina ironmaking slags like that with ni as shown in Fig. 8.
A fuzzy corresponding relation between NMgO and CS2⫺ can
be observed in Fig. 9(c) for the high alumina ironmaking
slags except one datum with the largest NMgO as 0.04. It can
be observed by comparing Fig. 8 and Fig. 9 that increasing
NCaO or 2nCaO shows a promotion tendency for C IMCT S2⫺,calculated
by IMCT model or C SShankar 2⫺, calculated by Shankar’s model6)
or CS2⫺,measured6) of the high alumina ironmaking slags.
However, increasing NSiO2 and NAl2O3 or nSiO2 and nAl2O3 show
an obvious decreasing trend for C IMCT S2⫺,calculated by IMCT
model or C SShankar 2⫺, calculated by Shankar’s model
6) or C 2⫺
S ,measured
of the high alumina ironmaking slags.
© 2010 ISIJ
Fig. 9. Relationship between calculated mass action concentra-
tion of structural units or ion couples (Ca2⫹⫹O2⫺),
(Mg2⫹⫹O2⫺), (SiO2) and (Al2O3), and calculated sul-
phide capacity by IMCT model or by Shankar’s model or
measured sulphide capacity of high alumina CaO–
SiO2–MgO–Al2O3 slags equilibrated with gas containing
sulphur in a temperature range of 1 773–1 873 K by
Shankar et al.,6) respectively.
5.5. Contribution of Basic Oxides to Sulphide Capac-
ity of CaO–SiO2–MgO–Al2O3 Ironmaking Slags
The developed IMCT model can not only predict the
total CS2⫺ of a slag by Eq. (12c) but also can determine the
respective sulphide capacity CS2⫺,i of basic oxides or ion
couples with desulphurization potential in a slag from Eq.
(12a) or Eq. (12b). The other five models listed in Table 3
can only predict CS2⫺ of a slag, but not the respective sul-
by phide capacity CS2⫺,i of basic oxides or ion couples with
desulphurization potential.
5.5.1. Contribution of Basic Oxides to Sulphide Capacity
of CaO–SiO2–MgO–Al2O3 Ironmaking Slags at
1 773 K
The relationship between C IMCT S2⫺,CaO,calculated of ion couple
(Ca2⫹⫹O2⫺) from Eq. (12a) or C IMCT S2⫺,MgO,calculated of ion cou-
ple (Mg2⫹⫹O2⫺) from Eq. (12b) and C IMCT S2⫺,calculated of CaO–
SiO2–MgO–Al2O3 ironmaking slags at 1 773 K22) from Eq.
(12c) by IMCT model is shown in Fig. 10(a), respectively.
Obviously, C IMCT IMCT
S2⫺,CaO,calculated and C S2⫺,MgO,calculated by IMCT
model has very good linear relation with C IMCT S2⫺,calculated of the
slags, respectively. The intercept of linear relation between
C IMCT IMCT IMCT
S2⫺,CaO,calculated or C S2⫺,MgO,calculated and C S2⫺,calculated is very
small, the slop of the linear relation can be considered as
the respective contribution of ion couple (Ca2⫹⫹O2⫺) or
(Mg2⫹⫹O2⫺) to the total sulphide capacity of the slags as
97% and 3% in terms of IMCT,22–26) which is the same
contribution of ion couple (Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺)
to LIMCT
calculated in previous publication
22)
of the same slags
based on IMCT.22–26) Therefore, ion couple (Ca2⫹⫹O2⫺) or
free CaO in CaO–SiO2–MgO–Al2O3 ironmaking slags
plays a decisive role to C IMCT IMCT
S2⫺,calculated as well as L calculated
22)
of
CaO–SiO2–MgO–Al2O3 ironmaking slags at 1 773 K.
Considering a little scattered relation between
C IMCT
S2⫺,calculated by IMCT model and CS2⫺,measured via measured
6)
LS,measured22) of CaO–SiO2–MgO–Al2O3 ironmaking slags at
1 773 K as illustrated in Fig. 1, it is necessary to clarify the
respective contribution of ion couple (Ca2⫹⫹O2⫺) as well
1370
 ISIJ International, Vol. 50 (2010), No. 10

as (Mg2⫹⫹O2⫺) to CS2⫺,measured of the slags. Assuming the
calculated contribution ratio of ion couple (Ca2⫹⫹O2⫺) as
well as (Mg2⫹⫹O2⫺) to C IMCT S2⫺,calculated of the slags as listed in
Table 1 does not change for each test runs, the contribution
of ion couple (Ca2⫹⫹O2⫺) CS2⫺,CaO,measured as well as
(Mg2⫹⫹O2⫺) CS2⫺,MgO,measured to CS2⫺,measured of the slags can
be also determined as 97% and 3% from linear relation be-
tween CS2⫺,CaO,measured or CS2⫺,MgO,measured and CS2⫺,measured
shown in Fig. 10(b), which is almost the same with the con-
tribution ratio of ion couple (Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺)
to LS,measured22) of the slags at 1 773 K.
5.5.2. Contribution of Basic Oxides to Calculated Sul-
phide Capacity of High Alumina CaO–SiO2–
MgO–Al2O3 Ironmaking Slags in a Temperature
Range of 1 773–1 873 K
The relationship between C IMCT S2⫺,CaO,calculated of ion couple
(Ca2⫹⫹O2⫺) and C IMCT S2⫺,MgO,calculated of ion couple (Mg ⫹
O2⫺) and CS2⫺,measured by Shankar et al.6) or C SShankar
Shankar’s model6) or C IMCT S2⫺,calculated by the IMCT model of
high alumina CaO–SiO2–MgO–Al2O3 ironmaking slags in
a temperature range of 1 773–1 873 K is depicted in Fig. 11,
respectively. The contribution of ion couple (Mg2⫹⫹O2⫺)
to CS2⫺,measured6) or C SShankar2⫺, calculated by Shankar’s model6) or
IMCT
C S2⫺,calculated by the IMCT model of the high alumina slags
can be obtained as 0.4% or 0.3% or 0.6% from slop of cor-
responding relation shown in Fig. 11, respectively. Mean-
while, contribution of ion couple (Ca2⫹⫹O2⫺) to
CS2⫺,measured6) or C SShankar
2⫺, calculated by Shankar’s model
6)
alumina slags can be calculated as 99.6% or 99.7% by sub-
tracting 0.4% or 0.3% from 100% as slop in Fig. 11(a) and
Fig. 11(b) is not small enough, while contribution of ion
couple (Ca2⫹⫹O2⫺) to C IMCT S2⫺,calculated by the IMCT model of
the high alumina slags is 99.4% from slop shown in Fig.
11(c). Therefore, the respective contribution of ion couple
(Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺) to CS2⫺ of the high alumina
Fig. 10. Contribution of ion couple (Ca2⫹⫹O2⫺) and (Mg2⫹⫹
O2⫺) to calculated by IMCT model or measured sul-
phide capacity of CaO–SiO2–MgO–Al2O3 slags equili-
brated with hot metal at 1 773 K by Yang et al.,22) re-
spectively.
ironmaking slags accounts for 99% and about 1%, respec-
tively. Comparing results shown in Fig. 10 and Fig. 11, it
can be deduced that decreasing CaO content from 40 to
34% and MgO content from 9 to 4%, and increasing Al2O3
content from 15 to 25% can effectively result in decreasing
contribution ratio of ion couple (Mg2⫹⫹O2⫺) in BF iron-
making slags from 3 to about 1%, increasing contribution
ratio of ion couple (Ca2⫹⫹O2⫺) in BF ironmaking slags
from 97 to 99%. In addition, the linear relation with very
high fitting degree in Fig. 11(c) compared with that in Fig.
11(a) and Fig. 11(b) further suggests that C IMCT S2⫺,calculated by the
developed IMCT model is more reliable than CS2⫺,measured by
Shankar et al.6) as well as C SShankar
2⫺, calculated by Shankar’s model.
6)
6. Conclusions
A sulphide capacity prediction model of CaO–SiO2–
2⫹
MgO–Al2O3 ironmaking slags has been developed based on
2⫺, calculated by the ion and molecule coexistence theory (IMCT). The pre-
dicted sulphide capacity by the developed IMCT model has
better accuracy than two groups of measured as well as pre-
dicted of CaO–SiO2–MgO–Al2O3 ironmaking slags by
other five models. The main summary remarks can be ob-
tained as follows:
(1) The developed IMCT model for predicting sulphide
capacity can be successfully applied to predict sulphide ca-
pacity of CaO–SiO2–MgO–Al2O3 ironmaking slags at
1 773 K as well as high alumina CaO–SiO2–MgO–Al2O3
of the high ironmaking slags in a temperature range of 1 773–1 873 K.
The predicted sulphide capacity of CaO–SiO2–MgO–Al2O3
ironmaking slags by the developed IMCT model has better
accuracy than measured sulphide capacity of CaO–SiO2–
MgO–Al2O3 ironmaking slags at 1 773 K by the present au-
thors and high alumina CaO–SiO2–MgO–Al2O3 ironmaking
slags in a temperature range of 1 773–1 873 K by Shankar et
al. or predicted sulphide capacity of two slags by other five
models. To expand application field of the developed IMCT
model, a hot metal pretreatment slags of CaO–SiO2–
MgO–Al2O3 with high binary basicity is also chosen to ver-
ify the feasibility of the IMCT model. The application
range of the developed IMCT model for ironmaking slags
or hot metal pretreatment slags can be summarized as
CaO(28–55%)–SiO2(7–40%)–MgO(1–12%)–Al2O3(10–40
%) with binary basicity as 0.73–8.0 as well as sulphide ca-
pacity in a range of 1⫻10⫺4–19⫻10⫺4.
(2) The calculated equilibrium mole numbers, mass ac-
tion concentrations of structural units or ion couples and
optical basicity, but not mass percentage of components or
binary basicity, can be applied to represent slag composi-
tion to correlate corresponding relation with sulphide ca-
pacity of CaO–SiO2–MgO–Al2O3 ironmaking slags.
(3) Not only the total sulphide capacity of CaO–SiO2–
MgO–Al2O3 ironmaking slags, but also the respective con-
tribution of ion couple (Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺), or

Fig. 11. Calculated contribution of ion couple (Ca2⫹⫹O2⫺) and (Mg2⫹⫹O2⫺) by IMCT model to measured or calculated by Shankar’s model as
well as by IMCT model of sulphide capacity of CaO–SiO2–MgO–Al2O3 slags equilibrated with gas containing sulphur in a temperature
range of 1 773–1 873 K by Shankar et al.,6) respectively.

1371 © 2010 ISIJ

 ISIJ International, Vol. 50 (2010), No. 10
free CaO and MgO in the slags can be quantitatively pre-
dicted by the developed IMCT model. Increasing Al2O3
content from 15 to 25% and decreasing CaO content from
40 to 34%, MgO content from 9 to 4% can improve contri-
bution of free CaO from 97 to 99%, while decreasing con-
tribution of free MgO from 3 to about 1% to the total sul-
phide capacity of CaO–SiO2–MgO–Al2O3 ironmaking
slags.
Acknowledgments
One author, Dr. Xue-min Yang, would like to express his
sincere gratitude to Prof. J. Zhang, who is the founder of
the ion and molecule coexistence theory, for his continuous
encouragement, support, valuable suggestions and discus-
sion.
Nomenclature
ai : Activity of component i (⫺)
B : Binary slag basicity of slags, i.e., the ratio of
(%CaO) to (%SiO2) (⫺)
CS2⫺ : Sulphide capacity of slags (⫺)
CS2⫺,measured : Measured sulphide capacity of slags (⫺)
CS2⫺,i : Respective sulphide capacity of ion couple i
or free components i in slags (⫺)
CS2⫺,i,measured : Calculated respective sulphide capacity of
ion couple i or free components i in slags
from measured data (⫺)
C IMCT
S2⫺,i,calculated : Calculated respective sulphide capacity of
ion couple i or free components i in slags by
IMCT model (⫺)
C IMCT
S2⫺,calculated : Calculated total sulphide capacity of slags
by IMCT model (⫺)
C SShankar
2⫺, calculated : Calculated total sulphide capacity of slags
by Shankar’s model (⫺)
C iS2⫺,calculated : Calculated total sulphide capacity of slags
by i’s model (⫺)
eij : Activity interaction coefficient of element j
to i in metal based on mass percentage as
concentration unit of element j (⫺)
fi : Activity coefficient of component i in metal
phase (⫺)
KiQ : Chemical reaction equilibrium constant of
forming component i or structural unit i (⫺)
LS : Sulphur distribution ratio between slag and
metal (⫺)
LS,measured : Measured total sulphur distribution ratio be-
tween slag and metal (⫺)
LS,i : Calculated respective sulphur distribution
ratio of ion couple i or free components i in
slags (⫺)
LIMCT
S,calculated : Calculated total sulphur distribution ratio
between slag and metal by IMCT model
(⫺)
Ni : Mass action concentration of structural unit
i or ion couple i in slags (⫺)
ni : Equilibrium mole number of structural unit i
or ion couple i in 100 g slags (mol)
pi : Partial pressure of component i in gas phase
(Pa)
R : Gas constant (8.314 J/mol · K)
T : Absolute temperature (K)
© 2010 ISIJ
D rGQm,i : Standard molar Gibbs free energy of form-
ing component i (J/mol)
Í ni : Total equilibrium mole number of all struc-
tural units in 100 g slags (mol)
(%i) : Mass percentage of component i in slag
phase (mass%)
[%i] : Mass percentage of component i in metal
phase (mass%)
Greek
L : Optical basicity of slags (⫺)
x i–j
interaction : Interaction coefficient of component i to j in
slags defined in KTH model (⫺)
Subscripts
f : Final content after desulphurization reaction
of component i.
REFERENCES
1) J. A. Duffy and M. D. Ingram: J. Non-Cryst. Solids, 21 (1976), No.
3, 373.
2) C. J. B. Fincham and F. D. Richardson: Proc. R. Soc. (London) A,
223A (1954), 40.
3) F. D. Richardson and C. J. B. Fincham: J. Iron Steel Inst., 178 (1954),
No. 9, 4.
4) D. J. Sosinsky and I. D. Sommerville: Metall. Mater. Trans. B, 17B
(1986), No. 6, 331.
5) R. W. Young, J. A Duffy, G. J. Hassall and Z. Xu: Ironmaking Steel-
making, 19 (1992), No. 3, 201.
6) A. Shankar, M. Gornerup, A. K. Lahiri and S. Seetharaman: Metall.
Mater. Trans. B, 37B (2006), No. 6, 941.
7) Y. Taniguchi, N. Sano and S. Seetharaman: ISIJ Int., 49 (2009), No.
2, 156.
8) R. G. Reddy and M. Blander: Metall. Mater. Trans. B, 18B (1987),
No. 6, 591.
9) S. C. Du, R. Nilsson and S. Seetharaman: Steel Res., 66 (1995), No.
11, 458.
10) M. M. Nzotta, S. C. Du and S. Seetharaman: ISIJ Int., 38 (1998),
No. 11, 1170.
11) M. M. Nzotta, S. C. Du and S. Seetharaman: ISIJ Int., 39 (1999),
No. 7, 657.
12) M. M. Nzotta, S. C. Du and S. Seetharaman: Metall. Mater. Trans. B,
30B (1999), No. 6, 909.
13) A. Shankar: Ironmaking Steelmaking, 33 (2006), No. 5, 413.
14) M. A. T. Andersson, P. G. Jönsson and M. M. Nzotta: ISIJ Int., 39
(1999), No. 11, 1140.
15) M. A. T. Andersson, P. G. Jönsson and M. Hallberg: Ironmaking
Steelmaking, 27 (2000), No. 4, 286.
16) M. Hino, S. Kitagawa and S. Ban-ya: ISIJ Int., 33 (1993), No. 1, 36.
17) R. Nilsson, S. Seetharaman and K. T. Jacob: ISIJ Int., 34 (1994), No.
11, 876.
18) S. Ban-ya, M. Hobo T. Kaji, T. Itoh and M. Hino: ISIJ Int., 44
(2004), No. 11, 1810.
19) M. Andersson: Ph.D Thesis, Division of Metallurgy, KTH, Sweden,
(2000), ISBN: 91-7170-621-6.
20) X. M. Yang, T. Z. Liu, Z. C. Guo, X. P. Yu and D. G. Wang: J. Iron
Steel Res., 7 (1995), No. 6, 1.
21) E. T. Turkdogan: Physicochemical Properties of Molten Slags and
Glasses, The Metals Society, London, (1983), 218.
22) X. M. Yang, J. S. Jiao, R. C. Ding, C. B. Shi and H. J. Guo: ISIJ Int.,
49 (2009), No. 12, 1828.
23) J. Zhang: Acta Metall. Sin. (Engl. Ed.), 14 (2001), No. 3, 177.
24) J. Zhang: J. Univ. Sci. Technol. Beijing, 9 (2002), No. 2, 90.
25) J. Zhang: Rare Metals, 23 (2004), No. 3, 209.
26) J. Zhang: Computational Thermodynamics of Metallurgical Melts
and Solutions, Metallurgical Industry Press, Beijing, (2007), 241.
27) R. Tsujino, J. Nakashima, M. Hirai and Y. Yamada: ISIJ Int., 29
(1989), No. 1, 92.
28) M. Ohya: Ferrous Metallurgy and Thermodynamics, Nikkan Kogyo
Shimbun-sha, Tokyo, (1971), 80.
29) J. Yang, M. Kuwabara, K. Okumura and M. Sano: ISIJ Int., 45
(2005), No. 12, 1795.
1372