ISIJ International, Vol. 45 (2005), No. 4, pp.

513–521

Experimental Evaluation of Chemical Composition and Viscosity

of Melts during Iron Ore Sintering

Satoshi MACHIDA, Koichi NUSHIRO, Koichi ICHIKAWA, Hidetoshi NODA1) and Hidenori SAKAI2)

Steel Research Laboratory, JFE Steel Co., 1, Kokan-cho, Fukuyama, Hiroshima 721-8510 Japan. 1) JFE Mineral Co.,1,
Kawasaki-cho, Chuo-ku, Chiba 260-0835 Japan. 2) JFE Mineral Co., 1, Kokan-cho, Fukuyama, Hiroshima 721-8510 Japan.
(Received on August 27, 2004; accepted on November 20, 2004 )

The viscosity of the melts formed in the sintering reaction was measured for evaluating the fluidity of the
melts, which play a crucial role in the sintering process. Using the sphere draw-up method carried out the
measurements of high temperature viscosity. With low-melting point compositions of the Fe2O3–CaO–SiO2
system, the viscosity of high SiO2 content melts (Fe2O3/CaO/SiO2
38.5/34.5/27.0 weight basis) was 5 times
higher than that of low SiO2 content melts (Fe2O3/CaO
80/20). The viscosity of low SiO2 content increased
with addition of Al2O3 and SiO2. Similarly, with melt compositions generated from actual ores, viscosity de-
pended on the contents of Al2O3 and SiO2.
Based on the results, the viscosity of the melted liquid was estimated from the chemical composition,
and the viscosity of the solid-liquid coexistence phase was also evaluated using a viscosity calculating equa-
tion for suspensions. The viscosity distribution of the solid-liquid coexistence phase at 1 300°C in the ternary
composition system Fe2O3–CaO–SiO2 was obtained.
KEY WORDS: iron ore; sinter; melt; slag; viscosity; viscosity estimation; suspension.

sintering reaction was measured for evaluating the fluidity

1. Introduction
With the heightened requirement to secure upward flexi-
bility in blast furnace productivity and decrease the reduc-
ing agents rate in blast furnace operation, higher quality of
sintered ore has been required in recent years. On the other
hand, where iron ore resources are concerned, production of
high quality Australian iron ore has shown a decreasing
trend, and improvements aimed at maintaining sintered ore
quality have also become necessary from this viewpoint.
While the sintered ore quality is influenced by many fac-
tors, it has been reported that the mineral textures and phys-
ical structure which determine sinter quality are largely de-
pendent on the behavior of the melts generated in the sin-
tering process.1–5) Therefore, various methods of controlling
the fluidity of the melts generated during sintering have
been proposed.6–10)
The importance of melt behavior in sintering, and partic-
ularly melt fluidity, is generally recognized. However, due
to the complexity of the reaction and unsteady temperature
conditions in sintering, fluidity is grasped not the viscosity
of liquid but the one of suspension. Then, it is difficult to
quantify the solid phase ratio in actual sintering melts, con-
sequently, the physical explanations of fluidity in sintering
is still imprecise. Although various reports on the viscosity
of melted slag in ironmaking and steelmaking have been
published,11–13) there have been few measurements of highly
oxidized iron compositions such as the melts generated in
sintering.14,15)
Therefore, in the present research, first, the viscosity of
the melted liquid with same compositions formed in the
of the melts, which play a crucial role in the sintering
process. Based on the results, the viscosity of the melted
liquid was estimated from the chemical composition, and
the viscosity of the solid-liquid coexistence phase was also
evaluated using a viscosity calculating equation for suspen-
sions.
2. Experimental Procedure
2.1. Viscosity Measurement
2.1.1. Apparatus and Procedure
The cylinder rotation method and oscillating cylinder
method have generally been used in measuring the viscosity
of high temperature melts.11,12) However, in the present vis-
cosity measurements, the sphere drawing up method was
used, because this method featured a comparatively simple
apparatus structure and operation.
Figure 1 shows a schematic diagram of the experimental
apparatus for viscosity measurement. This apparatus con-
sisted of a balance scale section and a heating section. The
balance was set on the upper part of the electric furnace
used for heating, which contained the sample. The viscosity
of sample was calculated by drawing force and speed of the
sphere which was immersed in the melt bath. The viscosity
value was calibrated in advance using a cold reference solu-
tion for viscosimeter calibration (Nippon Grease Co., Ltd.
JS5, 10, 50, 100, 200).
The sphere was made of pure platinum and was 20 mm in
diameter. A hook was welded to the sphere for hanging.
The platinum-rhodium crucible (Pt/Rh
90/10) had an

513 © 2005 ISIJ

 ISIJ International, Vol. 45 (2005), No. 4

Fig. 1. Experimental apparatus for viscosity measurement.

inner diameter of 50 mm and height of 90 mm. The plat-

inum sphere was suspended from one arm of the balance,
as shown in Fig. 1, using a platinum-rhodium wire (Pt/Rh

70/30). A diameter of 0.5 mm was adopted for the platinum
wire in consideration of deterioration by the slag, and cut-
ting due to heat. The sphere drawing distance in the viscosi-
ty measurement was roughly 20 mm, and the position of the
sphere was adjusted so as to avoid contact with the inner
wall of the crucible.
The scale was balanced in advance by adjusting the
weight set on the second arm of the balance with the sphere
immersed in the bath. In this condition, to draw up the
sphere, additional weights as 20 to 90 mg were added on the
second arm. The drawing speed of the sphere was measured
5 times at each load. Finally, the viscosity was obtained by Fig. 2. Phase diagram of Fe2O3–CaO–SiO2 system.18)
averaging these measured values at each temperature.
This experiment was performed sequentially at tempera- Table 1. Chemical compositions of measured samples.
tures of 1 400, 1 350, 1 300, and 1 250°C in the atmosphere.
In the actual sintering process, it is known that the partial
pressure of oxygen at the melting zone is lower than atmos-
pheric pressure by combustion of coke, and then, the iron
oxides in the melt are partially reduced.16) However, be-
cause this reduction state changes rapidly and control of the
experimental atmosphere for this change is exceptionally
difficult, in these measurements, first priority was given to
clarifying the basic physical properties of the melted liquid.

2.1.2. Preparation of Samples

In viscosity measurements, two types of samples were
used, one being prepared with reagents, and the other, pre-
pared with actual iron ore and a CaCO3 reagent. The
reagents were Fe2O3, CaCO3, SiO2, Al2O3, and MgO pow-
der. These materials were mixed at a specified blending
ratio, including the actual ore.
With all samples, the sample weight was 700 g and the
bath height was 70–75 mm.
In the equilibrium diagram for Fe2O3–CaO–SiO2 system,
which are the main components of sintered ore, the compo-
sition regions of melting points of 1 300°C and under are
shown by the dark colors in Fig. 2. The initial melts in sin-
tering are formed with these eutectic point composition.17) 2.2. Assimilation Test
Table 1 shows compositions L, M, and N in each region, To investigate the composition of the initial melts gener-
and the melting points of these compositions are around ated in sinter, an assimilation test was performed using iron
1 200°C. In this research, viscosity of melted liquid was ore and quicklime (CaO). The CaO columns (diameter:
measured for several compositions close to M and N. 8 mm, height: 10 mm) were made by compacting of CaCO3

© 2005 ISIJ 514

 ISIJ International, Vol. 45 (2005), No. 4
reagent and baking at 900°C in an electric furnace. Also the
iron ore columns (diameter: 19 mm, height: 10 mm) were
made by compacting of fine ore which was crushed under
125 m m, using four ore brands B, E, J, K, as shown in Table
2. The CaO column was placed on the iron ore columns and
they were baked in the electric furnace at 1 300°C for 5 to
10 min, extracted from the furnace, and quenched in water.
3. Experimental Results
3.1. Comparison of Viscosity with Low-melting Point
Composition in Fe2O3–CaO–SiO2 System
Figure 3(a) shows the relationship between temperature
and viscosity of composition M. In these measured results,
a linear relationship was obtained between the logarithmic
value of viscosity and the reciprocal of the absolute temper-
ature. When this was extrapolated to higher temperatures, a
viscosity at 1 550°C was 36% lower than the measured re-
sult reported by Sumita et al.14) However, the difference of
these viscosities would be considered due to the difference
of the measuring temperature range. Figure 3(b) shows the
results for composition N. The measured viscosities in the
present research were 6% lower than the reference values14)
of substantially same composition and temperature. As in-
determinacy on the order of 10% is considered unavoid-
able, we concluded that it is possible to obtain measure-
ments which are comparable with other data by this method
in a temperature range up to 1 400°C.
Comparing the viscosity of composition M and N, the
former that includes higher SiO2 is 5 times larger.
Table 2. Chemical composition of iron ores for assimilation
test.
Fig. 3. Relationship between temperature and measured viscosity.
515
3.2. Effect of SiO2, Al2O3, and MgO Addition on Vis-
cosity with Low-melting Point Composition in
Fe2O3–CaO System
To investigate the composition of the initial melts gener-
ated in sinter, an assimilation test was performed using iron
ore and quicklime (CaO). The CaO columns (diameter:
8mm, height: 10 mm) were made by compacting of CaCO3
reagent and baking at 900°C in an electric furnace. Also the
iron ore columns (diameter: 19 mm, height: 10 mm) were
made by compacting of fine ore which was crushed under
125 m m, using four ore brands B, E, J, K, as shown in Table
2. The CaO column was formed on the iron ore columns
and they were baked in the electric furnace at 1 300°C for 5
to 10 min, extracted from the furnace, and quenched in
water. To study the effect of SiO2, Al2O3, and MgO addition
to low-melting point composition N, as the melt composi-
tion generated in the initial stage of sintering, these compo-
nents were added in the range of SiO2: 1 to 4 mass%,
Al2O3: 0.5 to 8 mass%, and MgO: 0.25 to 2 mass%, as
shown in Table 1.
Figure 4(a) shows the effect of SiO2 addition. Viscosity
increased as the amount of SiO2 addition increased.
Furthermore, viscosity decreased with increasing in the
temperature of the specimen.
Figure 4(b) shows the effect of Al2O3 addition. At tem-
peratures above 1 350°C, the viscosity increased gradually
with addition of up to 6 mass% Al2O3, but at temperatures
of 1 300°C, viscosity increased notably with 0.5 mass% ad-
dition. A remarkable increase was also observed at temper-
atures above 1 350°C with 8 mass% addition. Thus, in com-
parison with SiO2 addition, when Al2O3 is added to compo-
sition N, the sample temperature is considered to have a
larger effect on viscosity.
Figure 4(c) shows the effect of MgO addition on viscosi-
ty. Although viscosity increases notably with 0.5 mass% ad-
dition of MgO, the increase in viscosity basically tends to
be small with 0.5 to 2.0 mass% addition of MgO. Moreover,
as with Al2O3, the temperature of the sample is also consid-
ered to have a large effect on viscosity in the case of MgO
addition.
Figure 4(c) also showed the viscosities with 1% MgO
© 2005 ISIJ
 ISIJ International, Vol. 45 (2005), No. 4

Fig. 4. Effect of additives on viscosity with low-melting point composition in Fe2O3–CaO system.

Fig. 5. Cross section of sample after the assimilation test. (Assimilating condition: 1 300°C)

Table 3. Chemical composition of initial melts in assimilation

test.

Fig. 6. Viscosity with initial melt compositions of assimilation

and 6% Al2O3 addition to composition N. In this composi-
tion, measured results were not obtained for 1 350°C or
under. However, the viscosity was higher than the one of
1% MgO addition or 6% Al2O3 addition respectively at
1 400°C.
3.3. Evaluation of Initial Melt Composition of Sinter
As shown in Table 2, ore K had highest contents of both
SiO2 and Al2O3, whereas ore J had lowest contents of both
components than ore K. The Al2O3 content of ore B was
approximately one half that of ore K, and ore E had ex-
tremely low contents of both SiO2 and Al2O3 in comparison
with the other ores. Figure 5 shows the sample cross-sec-
tion of ore J after baking. It is considered that the initial
melt have been generated in the boundary of the CaO
tests at 1 300°C.
reagent and iron ore. According to EPMA composition
analysis at this region, as shown in Table 3, the mole ratio
of CaO/Fe2O3 of each sample was 0.84–1.15 as the
composition range of monocalcium-ferrite (CaO · EFe2O3).
Moreover, the concentration of SiO2 and Al2O3 depended
on the content in the respective ores.
The test materials of four initial melt compositions ob-
tained from the assimilation tests, as shown Table 3, were
prepared using ore and CaCO3 reagent, and the viscosities
of samples were measured. Figure 6 shows the results of
viscosity measurements of each sample at 1 300°C. The
viscosity of sample No. 1, which had high contents of SiO2
and Al2O3, was the highest. And, viscosity was lowest in

© 2005 ISIJ 516

 ISIJ International, Vol. 45 (2005), No. 4
Fig. 7. Comparison of the measured and calculated viscosities in composition N with SiO2 addition.
sample No. 4, which had low contents of both compounds.
The viscosity is largely dependent on the contents of SiO2
and Al2O3 in this tests, and these results have the same ten-
dency as the measured results of the viscosity using
reagents described in the previous section.
4. Discussion
4.1. Calculation and Evaluation of Viscosity of Melted
Liquid with Extended Compositions
Iida et al.13) have proposed a calculating method using
network parameters which is applicable to molten fluxes in
steelmaking, blast furnace slag, etc. We attempted to apply
a viscosity calculating equation to the results as mentioned
in Sec. 3.2. Figure 7(a) shows a comparison of the mea-
sured and calculated viscosities for a system comprising
composition N with added SiO2 in which the calculated vis-
cosities were obtained by original method.13) The compara-
tive difference defined by subtracting the measured value
from the calculated value and dividing by the measured
value was a maximum of 80%. As described in Sec. 3.3, in
the measured values, the rate of increase in viscosity tended
to decrease with increasing SiO2 addition. But, as a result,
the comparative difference increased more in higher viscos-
ity. Moreover, the difference became even greater when
Al2O3 or MgO was added.
The equation proposed by Iida et al. would be applied to
slags consisting mainly of SiO2, CaO, and Al2O3, which is
significantly different from the high Fe2O3 compositions in
this research. Therefore, we attempted to modify the coeffi-
cients for calculation. Five components (Fe2O3, CaO, SiO2,
Al2O3, and MgO) were considered as objects of the calcula-
tion. The following Eq. (1) is the viscosity equation.13)
 E 
µ L
Aµ 0 exp   ...........................................(1)
 Bi * 
A
1.7451.962103T7.000107T 2
E
11.113.65103T
µ0
∑µ 0i X i
...............................(2)
α Fe2O3WFe2O3 α CaOWCaOα MgOWMgO
Bi *
.......(3)
α SiO2WSiO2 α *Al2O2 WAl2O3
517
a *Al2O3
aBibWAl2O3c .....................................(4)
a
1.20105T 24.3552102T41.16
b
1.40107T 23.4944104T0.2062
c
8.00106T 22.5568102T22.16
In Eq. (1), m o is the viscosity when the melt is a
monomolecular system. And Bi* is a modified basicity
index obtained from the composition of the melt and the in-
trinsic coefficient a i [i
Fe2O3 (aci.), CaO, SiO2, Al2O3,
MgO]. These terms have the values calculated from Eqs.
(2) and (3), respectively. T is absolute temperature, and Xi is
mole concentration. Table 4 shows the values of m 0i of the
constituent components used in this research.
In Eq. (3), the numerator is a term showing basic oxides,
while the denominator is a term showing acidic oxides. Wi
is the weight ratio of each component. Regarding the am-
photeric oxides Fe2O3 and Al2O3, when both Fe2O3 and
Al2O3 were assumed to be acidic, the comparative differ-
ence between the measured and calculated values became
smaller. Therefore, in calculating Bi*, Fe2O3 was assumed
to be acidic, as shown in Eq. (5).
α CaOWCaOα MgOWMgO
Bi *

α Fe2O3 ( aci.)WFe2O3  α SiO2WSiO2  α *Al2O3 WAl2O3
...........................................(5)
Because the number of measurement points was limited
in this research, it was difficult to optimize the intrinsic co-
efficients for each component. Then, the original intrinsic
coefficients of the basic oxides CaO and MgO and the
acidic oxide SiO2 were used on the assumption that the be-
havior of the basic oxides CaO and MgO and the acidic
oxide SiO2 in the melt does not differ greatly from that in
steel slag. On the other hand, we attempted optimizations of
the intrinsic coefficients for the amphoteric compounds
Fe2O3 and Al2O3. As the optimization procedure, the value
of a Fe2O3 was obtained for composition NSiO2 and com-
position NMgO systems, and the value of a Al2O3 was de-
termined for composition NAl2O3 system. Table 4 shows
the original and modified values of the intrinsic coefficient
a i for the constituent components used in this work. To
© 2005 ISIJ
 ISIJ International, Vol. 45 (2005), No. 4

Table 4. Values for viscosity of monomolecular system m 0 and intrinsic coefficient a i.

Fig. 8. Comparison of the measured and calculated viscosities in composition N with MgO or Al2O3 addition.

evaluate melts in liquid-state, the composition range of the

components treated in these calculations was SiO2: 0 to 4
mass% and Al2O3 and MgO: 0 to 2 mass%.
Figure 7(b) shows the results for composition N+SiO2
system, comparing the measured and calculated viscosities
obtained with the modified intrinsic coefficients. The com-
parative difference between the two was a maximum of
46%, showing improved accuracy. However, because the in-
crease rate of the viscosity with SiO2 addition differs in the
measured and calculated values, further improvement in ac-
curacy is considered difficult. The reason of this improve-
ment is modification of intrinsic coefficient of Fe2O3.
Fig. 9. Estimation of calculated value using the optimized para-
Figure 8(a) and 8(b) show the comparisons of the mea-
meters on sample M.
sured and calculated values for composition NMgO sys-
tem and composition NAl2O3 system, respectively. The
maximum comparative difference was 52% for MgO addi- fect of basicity in the liquid phase, same as in the case of
tion and 44% for Al2O3 addition. As shown in Fig. 8, the Fe2O3. And this consideration agrees with the finding19) that
conformity of MgO addition is worse than that of Al2O3 ad- the Al2O3 contributes largely to increase viscosity at higher
dition, because the value of a MgO used in this research was basicities.
optimized for low Fe2O3 melt. Because viscosity increase On the other hand, when the intrinsic coefficient modi-
behavior with both Al2O3 and MgO addition differs in the fied formerly was applied to composition M (Table 1), the
measured and calculated values depending on the tempera- comparative difference was more than 300 %. We therefore
ture and amount of addition, further improvement in accu- attempted to optimize the intrinsic coefficient of Fe2O3 and,
racy is also considered difficult in these two substances, as a result, obtained another parameters shown in Table 4.
as with composition NSiO2. With the composition When Fe2O3 was assumed to be basic in composition M
NAl2O3MgO system, accuracy was comparatively good and the contribution of Al2O3 was excluded due to no mea-
(difference: 14%), as shown in Fig. 8(a). surement of composition MAl2O3 system, calculation ac-
The intrinsic coefficient for Al2O3 given by using Eq. (4), curacy was satisfactory. Figure 9 shows a comparison of
is 340 to 1 725 under the measurement conditions, and thus the measured results and the values calculated using the op-
differs greatly from the value of 0.667 obtained in this re- timized parameters for sample M. In the temperature range
port. It is considered that the difference is caused by the ef- of 1 300–1 400°C, the comparative difference was within

© 2005 ISIJ 518

 ISIJ International, Vol. 45 (2005), No. 4
10%.
Next, Fig. 10 shows a comparison of the measured val-
ues of initial melt composition at 1 300–1 400°C and the
calculated viscosities obtained with the modified intrinsic
coefficient of sample N. The comparative difference was a
maximum of 50%. All values of calculated viscosities are
lower than those of measured. It may be caused by the in-
teraction between components in the melt. To further im-
prove calculation accuracy, it is considered necessary to in-
crease the number of measurements and to determine more
suitable intrinsic coefficients for all components. However,
qualitative evaluation of viscosity obtained from chemical
composition is considered possible, and the results are ade-
quate for understanding the melt fluidity of sinter.
4.2. Calculation of Viscosity of Suspension Melts
Various calculation models for the viscosity of suspen-
sions have been reported.20–23) However, among these, the
equation proposed by Mori et al.22) was adopted as a funda-
mental equation in this discussion. It is considered that the
equation provides comparatively high accuracy in the wide
solid phase ratio range. Viscosity, m , can be described by
Fig. 10. Comparison of the measured and calculated viscosities
with initial melt composition.
Fig. 11. Iso-thermal section through the Fe2O3–CaO–SiO2 system at 1 300°C.
C3S2: 3CaO · 2SiO2, C2S: 2CaO · SiO2, C2F: 2CaO · Fe2O3, CF: CaO · Fe2O3.
519
Eq. (6), based on the assumption that solid particles in
melts are spheres of single diameter.22)
 3 
µ
µ L 1  ......................(6)
 1 / S1 / φ VC 
f VC
0.52
where, m L is the viscosity of the liquid phase, S is the volu-
metric ratio of the solid phase, and f VC is the maximum
volumetric concentration of solids.
Few measurements of viscosity with the composition
near sample N and M have been reported, and in particular,
data at sintering temperatures around 1 300°C are seldom.
Therefore, the melt viscosity was calculated for a wide
range of compositions at 1 300°C. As shown in Fig. 11, the
Fe2O3–CaO–SiO2 ternary system can be divided into the
liquid phase regions (1) and (7) and the solid–liquid coexis-
tence regions (2) to (6) and (8) to (14). The liquid phase
composition and solid phase ratio were obtained from the
phase diagram and used to calculate viscosity. The value of
m L was calculated using the Eq. (1) in Sec. 4.1. For the in-
trinsic coefficient, the values of composition M were used
under basicities from 1 to 2 and the values of composition
N were used under basicities higher than 2. In calculating
the volumetric ratio, S, of the solid phase, the specific grav-
ity of liquid and solid were necessary. Measured values for
the specific gravity of the liquid at 1 300°C were not
found. Therefore, in consideration of the specific gravity
at 1 400°C as 3.9–4.2 g/cm3 24) and the temperature
dependency of the specific gravity with the composition
Fe2O3/FeO/CaO
7.35/1.29/24.96 (mass%) as 1.635
103 g/cm3/K,24) we have concluded the appropriate specific
gravity for the liquid phase as 3.9 g/cm3. And we have used
the specific gravity of the hematite (Fe2O3) solid as
5.287 g/cm3, 3.38 g/cm3 as 2CaO · SiO2 (hereinafter called
C2S), and 3.015 g/cm3 as 3CaO · SiO2 (hereinafter called
C3S).25) Because the available specific gravity for
2CaO · Fe2O3 (hereinafter called C2F) was not found, we
have put the value as 4.49 g/cm3, estimated on the propor-
tional distribution of the weight ratios of Fe2O3 and CaO.
The results of viscosity calculations at 1 300°C are
© 2005 ISIJ
 ISIJ International, Vol. 45 (2005), No. 4
Fig. 12. Calculated iso-viscosity line in the Fe2O3–CaO–SiO2 system at 1 300°C.
shown in Fig. 12. Due to the restriction of the maximum
volumetric concentration f VC of the solid phase in the melt,
a fluid condition appears at the composition inside the line
of limitation (solid phase ratio S
0.52).
In the regions (1) and (7) as shown in Fig. 11, the viscos-
ity is relatively influenced by the SiO2 composition. In the
regions (2) and (6), the intervals of the iso-viscosity line are
narrow because the compositions of regions (2) and (6) are
close to those of suspended solid. In the regions (4) and (5),
the iso-viscosity lines extend toward the composition of
C2S, because C2S solid phase ratio increase gradually with
decreasing of Fe2O3 contents.
In the regions (8) to (14), the neighbor of the region (7),
also the iso-viscosity lines extend toward the Fe2O3, CS and
C2S.
On the other hand, the two liquid regions (1) and (7) in
Fig. 11 become a single region above 1 315°C. When vis-
cosity at 1 350°C or higher was calculated by the method
described above, inconsistencies occurred in the viscosity
values with the compositions where these regions combine.
This is due to the difference in the viscosity equations for
the two liquid phase regions, and in particular, suggests the
possibility that the intrinsic coefficient a of the amphoteric
oxide Fe2O3 changes sensitively in response to basicity.
Further study of this point is necessary.
4.3. Control of Melt Fluidity Behavior by Composition
of Melts
Low SiO2 content melts have lower viscosity, as also
mentioned in concerned with the viscosity measurements of
compositions M and N. From the viewpoint of improving
the fluidity of melts, preventing the formation of a high
SiO2 content melt can be considered effective. Then, SiO2
should be dispersed uniformly in the raw material with the
composition range near the region (1) as shown in Fig.11.
However, on the other hand, processes in which SiO2 or
Al2O3 are localized in the material have also been proposed
from the viewpoint of keeping enough fluidity of melt. In
this case, although the sintering with which the contents of
SiO2 or Al2O3 is partially low in the materials can be per-
formed satisfactorily, the sintering in the other part have
low fluidity of melts and sintering temperature should be
© 2005 ISIJ
controlled appropriately. Moreover, because the coke used
as a heat source contains SiO2 and Al2O3 in the form of ash,
fine dispersion is useful from the viewpoint of controlling
the melt composition.
5. Conclusions
In order to understand the basic physical properties of
the melts formed during iron ore sintering, the viscosity of
melted liquid using reagents and iron ores were measured
and the viscosity of suspension was calculated. The follow-
ing conclusions were obtained.
(1) With low-melting point compositions of the
Fe2O3–CaO–SiO2 system, the viscosity of high SiO2 con-
tent melts was 5 times higher than that of low SiO2 content
melts. Moreover, the viscosity of low SiO2 content in-
creased with addition of Al2O3 and SiO2. Similarly, with
melt compositions generated from actual ores, viscosity de-
pended on the contents of Al2O3 and SiO2.
(2) An intrinsic coefficient for low-melting point com-
positions of the Fe2O3–CaO–SiO2 system melts was pro-
posed in the calculation of viscosity, enabling qualitative
evaluation of viscosity based on the melt composition.
(3) The viscosity distribution of the solid-liquid coexis-
tence phase at 1 300°C in the ternary composition system
Fe2O3–CaO–SiO2 was obtained using a viscosity equation
for suspensions.
Nomenclature
h : Measured viscosity of melt (Pa s)
m : Viscosity of melt (Pa s)
m L: Viscosity of liquid phase (Pa s)
m 0: Viscosity of non-network forming melts (Pa s)
m 0i: Viscosity of component i of non-network form-
ing melts (Pa s)
Bi : Basicity index ()
Bi*: Modified basicity index ()
Wi: Weight fraction of component i ()
Xi: Mole fraction of component i ()
S: Volumetric ratio of solid phase ()
f VC: Maximum volumetric concentration of solid
()
520
 ISIJ International, Vol. 45 (2005), No. 4

T: Temperature (K) (1999), 711.

a i: Intrinsic coefficient of component i 11) K. Seki, F. Oeters: Trans. Iron Steel Inst. Jpn., 24 (1984), 445.
12) T. Yasukouchi, K. Nakashima and K. Mori: Tetsu-to-Hagané, 85
a *Al2O3: Modified intrinsic coefficient of Al2O3 (1999), 571.
Subscripts 13) T. Iida, H. Sakai, Y. Kita and K. Shigeno: ISIJ Int., 40 (2000), S110.
14) S. Sumita, T. Mimori, K. Morinaga and T. Yanagase: J. Jpn. Inst.
i: Component (Fe2O3(aci), CaO, SiO2, Al2O3,
Met., (1980), 94.
MgO) 15) N. Hori, Y. Saito, K. Nakashima and K. Mori: CAMP-ISIJ, 15
(2002), 797.
REFERENCES 16) H. Soma, M. Wajima, Y. Hosotani and K. Tashiro: Tetsu-to-Hagané,
68 (1982), 2200.
1) Y. Hida, J. Okazaki, K. Ito and M. Sasaki: Tetsu-to-Hagané, 73 17) T. Inazumi: Koubutsukougaku, Asakurashoten, Tokyo, (1976), 199
(1987), 1893. 18) R. Roth, M. Clevinger and D. Mckenna: Phase Diagram for
2) M. Matsumura and T. Kawaguchi: CAMP-ISIJ, 3 (1990), 1065. Ceramists, Vol. 1, American Ceramic Society, Columbus, (1964),
3) M. Hoshi and T. Kawaguchi: CAMP-ISIJ, 7 (1994), 137. 228.
4) T. Haga, S. Kasama, T. Kozono and N. Ogata: Tetsu-to-Hagané, 82 19) M. Nakamoto, J. Lee, T. Tanaka and T. Usui: CAMP-ISIJ, 16 (2003),
(1996), 981. 1021.
5) E. Kasai, Y. Sakano and T. Nakamura: Tetsu-to-Hagané, 86 (2000), 20) A. Einstein: Ann. Phisik, 19 (1906), 286.
136. 21) J. Robinson: J. Phys. Colloid Chem., 53 (1949), 1042.
6) T. Haga, Y. Oshio, K. Nakamura, T. Kozono and S. Kamikawa: 22) Y. Mori and T. Ototake: J. Chem. Eng. Jpn., 20 (1956), 488.
Tetsu-to-Hagané, 83 (1997), 103. 23) S. Yagi and T. Ototake: Kogyo-Kagaku Zasshi, 61 (1958), 267.
7) T. Haga, Y. Oshio, Y. Hida, H. Fukuda and N. Ogata: Tetsu-to- 24) K. C. Mills: Slag Atlas, 2nd ed., Verein Deutscher Eisenhüttenleute
Hagané, 83 (1997), 233. (VDEh), Verlag Stahleisen GmbH, Düsseldolf, (1995), 321.
8) T. Kawaguchi, K. Kuriyama, S. Sato and K. Takata: Tetsu-to- 25) R. I. Gulyaeva, S. N. Shin, P. A. Kuznetsov and A. I. Okunev: Melts.,
Hagané, 76 (1990), 1642. 4 (1990), No. 2, 80.
9) I. Shigaki, M. Sawada, K. Yoshioka and T. Takahashi: Tetsu-to- 26) Iron and Steel Handbook, 4th ed., Vol. 1, ISIJ, Tokyo, (2002), (CD-
Hagané, 71 (1985), 1880. ROM), 7.3.4.
10) M. Nakano, S. Kasama and Y. Hosotani: Tetsu-to-Hagané, 85

521 © 2005 ISIJ