See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/305717708

Crystallization Kinetics of 2CaO·Fe2O3 and CaO·Fe2O3 in the CaO–Fe2O3

System

Article  in  ISIJ International · July 2016

DOI: 10.2355/isijinternational.ISIJINT-2015-710

CITATIONS READS

14 402

4 authors, including:

Chengyi Ding Xuewei Lv

Chongqing University Chongqing University
23 PUBLICATIONS   133 CITATIONS    348 PUBLICATIONS   3,185 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Graduate Scientific Research and Innovation Foundation of Chongqing, China View project

Natural Science Foundation of China View project

All content following this page was uploaded by Chengyi Ding on 18 April 2017.

The user has requested enhancement of the downloaded file.

 ISIJ International, Vol. 56 (2016),
ISIJ International,
No. 7 Vol. 56 (2016), No. 7, pp. 1157–1163

Crystallization Kinetics of 2CaO·Fe2O3 and CaO·Fe2O3 in the

CaO–Fe2O3 System

Chengyi DING, Xuewei LV,* Yun CHEN and Chenguang BAI

School of Materials Science and Engineering, Chongqing University, No. 174 Shazheng Street, Shapingba District, Chongqing,
400044 China.
(Received on December 22, 2015; accepted on March 30, 2016)

In this study, the non-isothermal crystallization kinetics of 2CaO·Fe2O3 and CaO·Fe2O3 were investigated
by DSC measurement. Crystallization of CaO–Fe2O3 system includes three reactions and can be explained
using Fe2O3–CaO phase diagram. The mechanisms of 2CaO·Fe2O3 and CaO·Fe2O3 crystallization were
analyzed using Avrami and Mo models. Results of the Avrami model analysis indicated that growth of
2CaO·Fe2O3 and CaO·Fe2O3 include two stages, which are controlled first by a fibril-like mechanism fol-
lowed by a spherulitic-type mechanism. The Mo model analysis yielded results similar to those of the
Avrami model and further concluded that Avrami exponent/Ozawa exponent is constant despite the
changes in cooling rate. Calculations using Kissinger method revealed that the activation energy of
2CaO·Fe2O3 and CaO·Fe2O3 crystallization were − 464.16 kJ·mol − 1 and − 172.61 kJ·mol − 1, respectively.
Moreover, 2CaO·Fe2O3 crystallization occurs in a difficult manner but proceeds at a faster rate than
CaO·Fe2O3 crystallization. Increasing the cooling rate promotes CaO·Fe2O3 crystallization but inhibits
2CaO·Fe2O3 crystallization, which is beneficial during sintering.

KEY WORDS: sinter; 2CaO·Fe2O3; CaO·Fe2O3; crystallization kinetics; Avrami model; growth type.

system. Muan10) finally demonstrated the phase diagram of

1. Introduction
In the 1960s and 1970s, the CaO–Fe2O3 system replaced
the silicate system as the focus in fluxed sinter research.1)
Fluxed sinter containing CaO–Fe2O3 system as bonding phase
demonstrates improved physical and metallurgical perfor-
mance. Research on calcium ferrite, especially its reduction,
is extensive.2–5) However, crystallization of calcium ferrite has
not been intensively investigated. The crystallization behavior
and crystal morphology of CaO–Fe2O3 system affect various
indices, such as strength and reducibility, which are used to
evaluate the quality of fluxed sinter. Research on the growth
type of CaO–Fe2O3 system is useful in indirectly investigating
the microstructure of the crystallization phase. Crystalliza-
tion of the CaO–Fe2O3 system has not been reported. With
the rapid development in thermal analysis methods, such as
thermogravimetry (TG) and differential scanning calorimetry
(DSC), thermal analysis kinetics has become an effective
method used to describe crystallization behavior.
Sosman and Merwin6) initially assessed the phase rela-
tionship in the CaO–Fe2O3 system in air atmosphere in
1916, however, CaO·2Fe2O3 was not considered in their
theoretical system, but proven by Tavasci7) in 1936. By
using phase stability analysis, Malquori and Cirilli8) dem-
onstrated that CaO·Fe2O3 is more stable than CaO·2Fe2O3
in a system equilibrated with Fe2O3. Edstrom9) reported the
temperature interval of phases generated in the CaO–Fe2O3
* Corresponding author: E-mail: lvxuewei@163.com
DOI: http://dx.doi.org/10.2355/isijinternational.ISIJINT-2015-710
CaO–Fe2O3 in air and at 1 atm O2 pressure, in which phase
stability, Fe2O3 decomposition, and 2CaO·Fe2O3 melting
were fully discussed. The morphology of calcium ferrite is
highly relevant to phase analysis but has been rarely studied,
especially through thermal analysis. Thus, the present study
investigated crystallization by using kinetic methods in ther-
mal analysis. This paper discusses only the binary system.
The influence of chemical components on crystallization
will be tackled in future work.
The Avrami,11–13) Ozawa,14) and Mo15–17) models are used
to analyze the crystallization of copolymers, minerals, and
metals. The present study used these models to analyze
the crystallization behavior of the CaO–Fe2O3 system. The
relationship of conversion ratio to crystallization time and
temperature was determined by the kinetics study through
thermal analysis. DSC scans with regard to the associa-
tion of heat flow with enthalpy changes provide a scien-
tific means to assess the relevance of conversion ratio to
crystallization time and temperature. The non-isothermal
crystallization kinetics of 2CaO·Fe2O3 and CaO·Fe2O3 are
further investigated by DSC measurement with Avrami,
Ozawa, and Mo models. Moreover, the activation energy
for generating 2CaO·Fe2O3 and CaO·Fe2O3 are calculated
using Kissinger18) method.
2. Theoretical Basis
2.1. Avrami Model
Studies on thermal analysis kinetics have gradually shifted

1157 © 2016 ISIJ

 ISIJ International, Vol. 56 (2016), No. 7

from the use of isothermal methods to non-isothermal meth-
ods.19) In particular, the Avrami model was derived from
the isothermal system and extended to the non-isothermal
system by substituting crystallization time to temperature.
Other non-isothermal models, such as the Ozawa and Mo
models, have also been widely employed.
The conversion ratio can be described as follows:
STT0
X (T ) = ................................ (1)
STT01
where X(T) is the conversion ratio that corresponds to tem-
perature T. STT0 is the area surrounded by the DSC curve
and the baseline from the onset temperature T0 to a fixed
temperature T. STT01 is the total area from the onset tempera-
ture to the final temperature T1. X(T) can be also expressed
as Fig. 1.
During crystallization under a fixed cooling rate β, the
relationship between time and temperature can be described
as follows:
T0 − T
t= .................................. (2)
β is similar to that of n, that is, m reflects a type of nucleation
the Avrami model was formulated using the conversion ratio
as a function of time and can be described as follows:
( )
1− X ( t ) = exp − Kt n ........................ (3)
where X(t) is the conversion ratio at a given temperature at
time t, K is the rate constant of crystallization, and n is the
Avrami exponent, which is determined by using the conver-
sion ratio X(t) and time t by plotting ln[ − ln(1 − X(t))] against
lnt, as follows:
ln  − ln (1 − X ( t ) )  = ln K + n ln t ................ (4)
where the Avrami exponent n represents a type of nucleation
and growth mechanism.20) Table 1 shows the crystallization
growth type based on different n values.
The Avrami model derived from isothermal crystalliza-
tion should be modified to include the parameter K of the
formula when extending to the non-isothermal process.
Fig. 1. Conversion ratio obtained in DSC curve.
Table 1. Correspondence between growth types and n.
Jeziorny21) modified K as follows:
ln K
ln K c = ................................ (5)
β
where Kc is the modified rate constant based on the non-
isothermal process.
2.2. Mo Model
The Ozawa model is another kinetics equation derived
from the Avrami model based on Evans22) equation to
describe non-isothermal crystallization. Compared with
the Avrami model, in which the conversion ratio is related
to crystallization time, the Ozawa model was established
based on the evolution of conversion ratio as a function of
temperature.
 P (T ) 
1− X ( T ) = exp  −  ..................... (6)
 βm 
 
where P(T) is a function of temperature related to nucleation
mode, nucleation rate, and growth rate. The meaning of m
and growth mechanism.14,23–25)
m is determined by the conversion ratio X(T) and time T
by plotting ln[ − ln(1 − X(T))] against ln| β | as follows:
ln  − ln (1 − X ( T ) )  = ln P ( T ) − m ln β .......... (7)
The Mo model is a combination of the Avrami and Ozawa
models.26–29) The Avrami model describes the relationship
of conversion ratio to time, whereas the Ozawa model
denotes the relationship of conversion ratio to temperature.
As shown in Eq. (2), crystallization time can be expressed as
a function of temperature for crystallization under a constant
cooling rate. Hence, a certain time should correspond to a
certain temperature at the same conversion ratio. That is,
when X(t) = X(T), the following equation can be obtained:
ln  − ln (1 − X ( t ) )  = ln  − ln (1 − X ( T ) )  ......... (8)
This equation can be transformed into
ln K + n ln t = ln P ( T ) − m ln β ................. (9)
then,
ln β = ln F ( T ) − a ln t
...................... (10)
1
 P (T )  m n
where F ( T ) =  and a = , where F(T) corre-
 K  m
 
sponds to the cooling rate at unit crystallization time when
reaching a certain value of conversion ratio; this parameter
represents the crystallization rate at a certain conversion
ratio. A low F(T) value indicates a fast cooling rate. a rep-
resents the type of nucleation and growth mechanism. F(T)
and a can be obtained using the intercept and slope of the
plot of ln| β | against lnt at a specific conversion ratio.

n Growth type Growth dimensions 2.3. Determination of Activation Energy Using Kissinger
Method
3–4 Spherulitic-type Three-dimensional
The Kissinger method was established based on the
2–3 Disc-like Two-dimensional assumption that the process can be described by nth-order
reaction. This method is widely used to obtain the kinetic
1–2 Fibril-like One-dimensional
parameter of crystallization and expressed as follows:

© 2016 ISIJ 1158

 ISIJ International, Vol. 56 (2016), No. 7

3.2. Non-isothermal DSC Investigation

βi AR E
ln = ln − ( i = 1, 2, 3…) ........... (11) Non-isothermal DSC investigation was conducted by
Tpi2 E RTpi using a Netzsch thermal analyzer (Model STA 449C).
where Tpi is the peak temperature at a fixing cooling rate, Samples (10 mg) were tiled in a Pt crucible, with an Al2O3
A is the pre-exponential coefficient, and E is the activation crucible containing air as reference. The samples were
energy of the crystallization. heated to 1 300°C at 20 K/min and then cooled to 1 100°C at
The Kissinger method can be further converted to a cooling rate of 10, 15, 20, and 25 K/min in air atmosphere.
 βi 
d  ln 2  4. Results and Discussion
 Tpi  = − E .......................... (12)
 1  R 4.1. DSC Curves
d  Figure 3 shows the DSC curves of the samples at various
T
 pi  cooling rates (10, 15, 20, and 25 K/min). As the temperature
activation energy E can be calculated using the slope of the decreases from 1 300°C to 1 100°C, three exothermic peaks
βi mainly appear in order. According to the Fe2O3–CaO phase
1
plot of ln 2
against . diagram,10) three crystallization reactions are generated at
T pi Tpi n(Fe2O3)/n(CaO) = 1:1 when the temperature decreases
from 1 300°C to 1 100°C; these reactions match the three
3. Experimental Procedure
peaks shown on the DSC curves. The experimental crystal-
3.1. Materials lization temperature is lower than the theoretical tempera-
Testing samples were prepared by mixing CaCO3 ture because of the crystallization undercooling effect. The
(99.99% purity) and Fe2O3 (α-Fe2O3, 99.99% purity) at a crystallization processes are described in Figs. 4 and 5 and
mole ratio of 1:1 for 30 min. The mixture was pressed into correspond to Eq. (13). The amount of product at various
a cylindrical shape under 10 MPa to obtain a homogenous cooling rates remains constant.
sample. The samples were heated from room temperature
up to 900°C, which was maintained for 1 h to ensure the
decomposition of CaCO3. The temperature was further
increased to 1 200°C and maintained for 10 h to generate
CaO·Fe2O3. The samples were pulverized in a vibration mill.
Qualitative and quantitative analyses of X-ray diffrac-
tion (XRD) patterns (Model D/max2500/PC; Cu Kα) were
performed before thermal analysis to verify the presence of
CaO·Fe2O3 in the samples after roasting and cooling. The
XRD patterns of the samples were compared with those of
CaO·Fe2O3 in the standard PDF card.
The results (Fig. 2) show that the XRD pattern of the
samples fitted well with that of CaO·Fe2O3 in the standard
PDF card, and other impurity peaks are not evident; this
finding reflects that the CaO·Fe2O3 phase accounts for the
majority of the sample phases. Furthermore, XRD quanti-
tative analysis shows that the CaO·Fe2O3 content reaches
93%, confirming that the thermal analysis test is based on
the crystallization phase of CaO·Fe2O3. Fig. 3. DSC investigations at various cooling rates.

Fig. 2. XRD patterns of samples and CaO·Fe2O3 in standard PDF

card. Fig. 4. Fe2O3–CaO phase diagram.10)

1159 © 2016 ISIJ

 ISIJ International, Vol. 56 (2016), No. 7

Fig. 5. Crystallization processes when temperature drops from

1 300°C to 1 100°C.

Fig. 6. DSC curves at various cooling rates for the crystallization

of 2CaO·Fe2O3 (peak 1).
The samples are initially in the full liquid phase from
the start of cooling from point A. The first peak (peak 1)
indicates the liquid–solid transition, in which 2CaO·Fe2O3
precipitates from the liquid phase when the temperature
decreases between points A and B. CaO·Fe2O3 is then gener-
ated by the peritectic reaction at the interface of 2CaO·Fe2O3
and the liquid phase at the beginning of temperature B,
which corresponds to peak 2. Meanwhile, 2CaO·Fe2O3 and
the liquid phase gradually disappear. CaO·Fe2O3 is gener-
ated on adhesion of the 2CaO·Fe2O3 surface and inhibits the
contact of 2CaO·Fe2O3 to the liquid phase. Thus, the peritec-
tic reaction cannot proceed completely, and the liquid phase
remains low when the peritectic reaction is completed.
The remaining liquid phase is cooled from points C to D.
A eutectic reaction is generated, in which CaO·Fe2O3 and
CaO·2Fe2O3 precipitate from the total liquid phase when the
temperature decreases at point D. The temperature in this
process is apparently similar to that during peritectic reac-
tion, and the heat release for this part corresponds to peak 2.
Fig. 7. DSC curves at various cooling rates for peritectic crystal-
The generated CaO·Fe2O3 adds to the CaO·Fe2O3 produced lization of CaO·Fe2O3 (peak 2).
from the peritectic reaction. CaO·2Fe2O3 is further cooled
from points E to F. Eutectoid reaction then occurs, in which
CaO·2Fe2O3 is transformed into CaO·Fe2O3 and Fe2O3 and the endothermic peaks.
corresponds to peak 3. The final phases for the entire pro-
− solid reaction , peak1
cess are CaO·Fe2O3 and Fe2O3. The entire crystallization is liquid Liquid
 → 2CaO ⋅ Fe 2 O3 (S ) + liquid
described by Eq. (13). Peritectic
 reaction , peak 2
→ CaO ⋅ Fe 2 O3 (S ,1)
The temperatures for peaks 1 and 2 are nearly similar to
each other; thus, the two crystallization peaks overlap. The + liquid Eutectic
 reaction
→ CaO ⋅ Fe 2 O3 (S ,1)
peak should be separated using the Gauss rule to obtain +CaO ⋅ Fe 2 O3 (S , 2) + CaO ⋅ 2Fe 2 O3 (S )
the two complete peak patterns of the crystallization of Eutectoid
 reaction , peak 3
→ CaO ⋅ Fe 2 O3 (S ,1)
2CaO·Fe2O3 and CaO·Fe2O3. This study mainly analyzed
the kinetics of the two crystallization processes, which cor- +CaO ⋅ Fe 2 O3 (S , 2) + CaO ⋅ Fe 2 O3 (S, 3) + Fe 2 O3 (S )
respond to peaks 1 and 2. Figures 6 and 7 show the DSC ........................................ (13)
curves of the two individual peaks at different cooling rates.
Figure 3 shows that the generation temperature of the 4.2. Avrami Model
reaction D (CaFe4O7 = Fe2O3 + CaFe2O4), which corre- The relationship between conversion ratio and time can
sponds to peak 3, reaches approximately 1 160°C. We used be determined using the DSC curves of the two crystalliza-
Factsage to prove our results, and the “reaction” module tion reactions (Fig. 8).
can reveal the thermodynamic condition of reaction D being According to the Avrami model, Kc and n are determined
generated; the results show that the theoretical generation by the intercept and slope of plotting ln[ − ln(1 − X(t))]
temperature of reaction D ranges from 1 156°C to 1 157°C, against lnt (Fig. 9). The Avrami kinetic parameters of the
where Gibbs free energy (ΔG) is < 0. The enthalpy change two reactions are shown in Table 2.
in reaction D shows a negative value ( − 21.8 kJ·mol − 1); The two reactions include two stages, which were ana-
thus, exothermic peaks appear on the DSC curve but not on lyzed in this study. Kc increases with increasing cooling

© 2016 ISIJ 1160

 ISIJ International, Vol. 56 (2016), No. 7

Fig. 8. Conversion ratio corresponding to time at various cooling rates for the crystallization of 2CaO·Fe2O3 (peak 1)
and peritectic crystallization of CaO·Fe2O3 (peak 2).

Fig. 9. ln[-ln(1-X(t))] against lnt based on Avrami model for the two peaks.

Table 2. Avrami parameters for the 2CaO·Fe2O3 generation and crystallization. In addition, the n values of 2CaO·Fe2O3
peritectic crystallization of CaO·Fe2O3.
crystallization are between 1 and 2 for stage 1 and between
(a) 3 and 4 for stage 2. This finding indicates that 2CaO·Fe2O3
Stage 1 Stage 2
crystallization is controlled first by a fibril-like and then by
β (K/min)
2
a spherulitic-type mechanism. The n values of peak 2 also
n Kc R n Kc R2
lie nearly between 1 and 2 for stage 1 and between 3 and 4
10 1.70 0.73 0.9887 3.71 0.97 0.9978 for stage 2, showing that CaO·Fe2O3 crystallization is also
15 1.47 0.82 0.9932 3.42 1.02 0.9956 controlled first by a fibril-like and then by a spherulitic-type
20 1.09 0.83 0.9905 3.75 1.12 0.9945
mechanism.
25 1.63 0.95 0.9879 4.17 1.18 0.9944
4.3. Mo Model
X(t) (20%, 40%, 60%, and 80%) was selected to deter-
(b)
mine the corresponding crystallization times of the four
Stage 1 Stage 2 cooling rates in the two crystallization processes. Fitting
β (K/min)
n Kc R 2
n Kc R2 the plots of lnβ against lnt, F(T), and a can be calculated
10 2.01 0.48 0.9745 4.45 0.44 0.9983 using the intercept and slope of the fitted line as shown in
Fig. 10 and Table 3.
15 1.92 0.70 0.9843 3.88 0.74 0.9989
The F(T) values, which indicate that the cooling rate
20 1.34 0.80 0.9900 3.58 0.87 0.9969 corresponds to a specific conversion ratio at a unit crystal-
25 1.05 0.88 0.9992 3.40 0.96 0.9874 lization time, increase with increasing conversion ratio; this
Note: (a): peak 1, (b): peak 2 phenomenon follows the crystallization rules. Moreover,
the F(T) value of peak 1 is lower than that of peak 2 at the
same conversion ratio, indicating that peak 1 requires a low
rates, indicating that raising the cooling rate increases the cooling rate to reach a fixed conversion ratio. That is, the
crystallization rate; this conclusion is consistent with the crystallization rate of 2CaO·Fe2O3 is faster than CaO·Fe2O3.
actual situation. The Kc values of peak 1 are higher than This conclusion corresponds to the results of the Avrami
those of peak 2 at the same cooling rate, demonstrating that model-based analysis. The a values remain nearly constant
2CaO·Fe2O3 crystallization occurs faster than CaO·Fe2O3 at peaks 1 and 2. The a values for the crystallization of

1161 © 2016 ISIJ

 ISIJ International, Vol. 56 (2016), No. 7

Fig. 10. lnβ against lnt at specific cooling rate based on Mo model for 2CaO·Fe2O3 (peak 1) generation and peritectic
crystallization of CaO·Fe2O3 (peak 2).

Table 3. Parameters of crystallization kinetics based on the Mo

model.

Peak 1 Peak 2
X(t)
2
lnF(T) a R lnF(T) a R2
10 2.24 0.63 0.9990 3.05 0.97 0.9998
15 2.40 0.63 0.9912 3.25 0.97 0.9986
20 2.49 0.64 0.9840 3.39 0.98 0.9972
25 2.58 0.63 0.9891 3.54 0.97 0.9898

2CaO·Fe2O3 and CaO·Fe2O3 are 0.63 and 0.97, respectively.

Thus, a certain linear relationship exists between n and m.

4.4. Calculation of Activation Energy Using Kissinger βi 1

Method Fig. 11. Fitting the plots of ln against based on Kissinger
Tpi2 Tpi
According to Kissinger method, the peak temperature in mothed for 2CaO·Fe2O3 (peak 1) generation and peritec-
the DSC curves at various cooling rates must be obtained. tic crystallization of CaO·Fe2O3 (peak 2).

βi 1
After fitting the plots of ln against (Fig. 11), the
Tpi2 Tpi Table 4. Apparent activation energy E calculated based on
Kissinger mothed for the two peaks.
activation energy E can be calculated using the slope of the
fitted line. E/ kJ·mol −1
2CaO·Fe2O3 and CaO·Fe2O3 crystallization includes a Peak 1 Peak 2
chemical reaction and a liquid–solid transformation. There-
−464.16 −172.61
fore, the calculated value of the activation energy includes
two energy types, which are derived from chemical and
physical processes, as follows:
4.5. Effect of Cooling Rates on Two Crystallization
ETol = EChe + EPhy .......................... (14) Processes
where ETol is the apparent energy value; EChe and EPhy refer The enthalpy change can be calculated from the peak
to the apparent activation energy from the chemical reaction area in the DSC curve. The standard enthalpy changes in
and energy change from phase transformation. the two processes were already obtained ( − 557.25 and
The calculated activation energy is the activation energy − 1 228.7 kJ·mol − 1), the molar quantity of 2CaO·Fe2O3 and
of the chemical reaction; thus, energy accompanied by CaO·Fe2O3 can be derived as follows:
the liquid–solid transformation must be deducted from the
∆Hi ⋅ m
calculated ETol. The changes in energy in the phase trans- n= .............................. (15)
formation of the two processes are − 557.25 and − 1 228.7 ∆Hio
kJ·mol − 1 as revealed by the phase diagram module in Fact- where
sage. Table 4 shows that the apparent activation energy of
S
2CaO·Fe2O3 and CaO·Fe2O3 crystallization calculated using ................................ (16)
∆Hi =
the Kissinger method are − 464.16 and − 172.61 kJ·mol − 1, β
respectively. In Eq. (15), n refers to the molar quantity of 2CaO·Fe2O3

© 2016 ISIJ 1162

 ISIJ International, Vol. 56 (2016), No. 7

Table 5. Molar quantity of 2CaO·Fe2O3 or CaO·Fe2O3 at each by a fibril-like and then by a spherulitic-type mechanism.
cooling rate. Mo model analysis confirmed the accuracy of the Avrami
n/mol (× 10 − 6) results and showed that n/m is constant despite the changes
in cooling rate.
10 15 20 25
(K·min −1) (K·min −1) (K·min −1) (K·min −1) (3) The apparent activation energy of 2CaO·Fe2O3 and
CaO·Fe2O3 crystallization calculated based on the Kissinger
Peak 1 1.44 1.26 1.59 1.01
method are − 464.16 kJ·mol − 1 and − 172.61 kJ·mol − 1,
Peak 2 0.18 0.28 0.37 0.36 respectively.
(4) 2CaO·Fe2O3 crystallization occurs in a difficult
manner but proceeds at a faster crystallization rate than
or CaO·Fe2O3, m represents the mass quantity of samples CaO·Fe2O3 crystallization. Moreover, increasing the cool-
(10 mg), ΔHi and ΔHio are the enthalpy change and standard ing rate promotes CaO·Fe2O3 crystallization but inhibits
enthalpy change of 2CaO·Fe2O3 or CaO·Fe2O3 crystalliza- 2CaO·Fe2O3 crystallization.
tion process, respectively, and S is the peak area corre-
sponding to each process. Table 5 shows that the molar Acknowledgment
quantity of 2CaO·Fe2O3 decreases, but that of CaO·Fe2O3 The authors are grateful for the financial support provided
increases with increasing cooling rate. That is, increasing by the Natural Science Foundation of China (Nos. 51104192
the cooling rate promotes CaO·Fe2O3 crystallization but and 51544203).
inhibits 2CaO·Fe2O3 crystallization, which is beneficial dur-
ing sintering. REFERENCES
The reduction behavior of calcium ferrite during sintering
1) B. Phillips and A. Muan: J. Am. Ceram. Soc., 42 (1959), 413.
has been fully studied, whereas the crystallization behav- 2) T. Murakami, T. Kodaira and E. Kasai: ISIJ Int., 55 (2015), 1197.
ior of sinter ore that cools from the liquid phase is rarely 3) T. Murakami, T. Kodaira and E. Kasai: ISIJ Int., 55 (2015), 1181.
4) H. Park and V. Sahajwalla: ISIJ Int., 54 (2014), 49.
studied. This paper discussed the effect of cooling rates 5) T. Usui, Y. Nakamuro, M. Nishi, M. Naito, H. Ono and P. S. Assis:
on the crystallization phase 2CaO·Fe2O3 and CaO·Fe2O3. ISIJ Int., 55 (2015), 1617.
CaO·Fe2O3 has higher reducibility and strength indices than 6) R. B. Sosman and H. E. Merwin: Wash. Acad. Sci. J., 6 (1916), 532.
7) B. Tavasci and G. Mulas: Ann. Chim. Appl., 26 (1936), 291.
2CaO·Fe2O3, the results based on this study indicates that 8) G. Malquori and V. Cirilli: 3rd Int. Symp. on Chemistry of Cement,
increasing the cooling rate promotes CaO·Fe2O3 crystalliza- Cement and Concrete Association, London, (1952), 137.
9) J. O. Edstrom: Jernkontorets Ann., 140 (1956), 101.
tion but inhibits 2CaO·Fe2O3 crystallization. Namely, our 10) B. Phillips and A. Muan: J. Am. Ceram. Soc., 41 (1958), 445.
studies focus on the effect of cooling rates on the crystal- 11) M. Avrami: J. Chem. Phys., 7 (1939), 1103.
lization phase. Our future work aims to concentrate on the 12) A. Melvin: J. Chem. Phys., 8 (1940), 212.
13) M. Avrami: J. Chem. Phys., 9 (1941), 177.
influence of cooling rate on the morphology of calcium fer- 14) T. Ozawa: Polymer, 12 (1971), 150.
rite, specifically on the appropriate cooling rate required to 15) T. X. Liu, Z. S. Mo, S. G. Wang and H. F. Zhang: Polym. Eng. Sci.,
37 (1997), 568.
generate acicular calcium ferrite. When applied in sintering, 16) S. Y. Liu, Y. N. Yu, Y. Cui, H. F. Zhang and Z. S. Mo: J. Appl.
a larger amount of acicular calcium ferrite can be formed Polym. Sci., 70 (1998), 2371.
under a controlled cooling rate. 17) X. Y. Qiao, X. H. Wang, X. J. Zhao, Z. S. Mo and H. F. Zhang:
Synth. Metal., 113 (2000), 1.
18) H. E. Kissinger: Anal. Chem., 29 (1957), 1702.
19) J. H. Flynn: Thermochim. Acta, 203 (1992), 519.
5. Conclusions 20) G. P. Balamurugan and S. N. Maiti: J. Appl. Polym. Sci., 107 (2008),
2414.
The non-isothermal kinetics of the CaO–Fe2O3 system 21) A. Jeziorny: Polymer, 19 (1978), 1142.
was examined by using DSC investigations determined by 22) U. R. Evans: Trans. Faraday Soc., 41 (1945), 365.
23) P. Supaphol, P. Thanomkiat and R. A. Phillips: Polym. Test., 23
the Avrami and Mo models. Based on the results of this (2004), 881.
study, the following conclusions are presented: 24) Y. Long, R. A. Shanks and Z. H. Stachurski: Prog. Polym. Sci., 20
(1995), 651.
(1) Crystallization includes three reactions. Crystal- 25) M. Eder and A. Wlochowicz: Polymer, 24 (1983), 1593.
lization of 2CaO·Fe2O3 and CaO·Fe2O3, which correspond 26) Q. X. Zhang, Z. H. Zhang, H. F. Zhang and Z. S. Mo: J. Polym. Sci.
to the first two peaks in the DSC curves, was studied by Part B Polym. Phys., 40 (2002), 1784.
27) Z. B. Qiu, Z. S. Mo, Y. N. Yu, H. F. Zhang, S. R. Sheng and C. S.
thermal kinetics. Song: J. Appl. Polym. Sci., 77 (2000), 2865.
(2) Avrami model analysis showed that the two crystal- 28) J. L. Cai, T. Li, Y. Han, Y. Q. Zhuang and X. Q. Zhang: J. Appl.
Polym. Sci., 100 (2006), 1479.
lization processes include two stages, namely, 2CaO·Fe2O3 29) S. Y. Liu, Y. N. Yu, Y. Cui, H. F. Zhang and Z. S. Mo: J. Appl.
and CaO·Fe2O3 crystallization, which are controlled first Polym. Sci., 70 (1998), 2371.

1163 © 2016 ISIJ

View publication stats