Thermochimica Acta 605 (2015) 4351

Contents lists available at ScienceDirect

Thermochimica Acta
journal homepage: www.elsevier.com/locate/tca

Hydration heat evolution and kinetics of blended cement containing

steel slag at different temperatures
Fanghui Han a,b , Zengqi Zhang a , Dongmin Wang b , Peiyu Yan a, *
a
b

Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry, Department of Civil Engineering, Tsinghua University, Beijing, China
China University of Mining and Technology (Beijing), Beijing, China

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 12 November 2014
Received in revised form 19 February 2015
Accepted 21 February 2015
Available online 23 February 2015

Hydration heat characteristics of blended cement containing up to 50% steel slag were studied at 25
C,
45
C and 60
C by isothermal calorimeter. Kinetics equations were used to explore kinetics of blended
cement. Kinetics parameters, n,K, and Ea, were calculated and analyzed. Results indicated that the activity
of steel slag was very low. Steel slag accelerated the hydration of aluminates but decelerated the
hydration of silicates in cement clinker. Small amount of steel slag affected little the hydration process of
cement. Elevated temperature obviously promoted the hydration of blended cement. The hydration of
blended cement containing no more than 35% steel slag was controlled by nucleation of hydrates in
acceleration period and by diffusion of ions in decay period. The hydration of blended cement containing
50% steel slag was mostly dominated by phase boundary reaction, then directly by diffusion. Ea increased
with increasing steel slag content.
2015 Elsevier B.V. All rights reserved.

Keywords:
Steel slag
Blended cement
Hydration heat
Kinetics

1. Introduction
Mineral admixtures are widely used in modern concrete either
in blended cements or added separately in the concrete mixer [1].
The use of mineral admixtures such as ground granulated blast
furnace slag, y ash and silica fume which have both pozzolanic
and latent hydraulic properties can improve the properties of
concrete and limit environmental impact as well as bring economic
benets. With the resource shortage and rising price of slag, y ash
and other quality mineral admixtures, developing more kinds of
mineral admixtures that leads to sustainable concrete design and
greener environment is a valuable work. Currently, steel slag shows
a good potential as mineral admixture in concrete.
Steel slag is the industrial waste discharged from the
steel-rening process in a conversion furnace. It accounts for
approximately 1520% by mass of the steel output [2,3]. More than
80 million tons of steel slag is discharged in China every year [4].
However, the current utilization rate of steel slag in China is only
22%, far behind the developed countries [5]. Thus, intensive
research work is needed for high utilization rate of steel slag. A
more attractive approach is the replacement of cement by steel
slag. The chemical compositions of steel slag consist of CaO
4560%, SiO2 1015%, Al2O3 15%, Fe2O3 39%, FeO 720%, MgO

* Corresponding author. Tel.: +86 13501215836; fax: +86 01062785836.

E-mail address: yanpy@tsinghua.edu.cn (P. Yan).
http://dx.doi.org/10.1016/j.tca.2015.02.018
0040-6031/ 2015 Elsevier B.V. All rights reserved.

313%, and P2O5 14% [6]. And its main minerals consist of C3S, C2S,
C4AF, C2F, C12A7, RO phase (CaO-FeO-MnO-MgO solid solution) and
free-CaO [68]. Due to the minerals of C3S, C2S, C4AF, C2F, and C12A7,
steel slag has hydraulic properties. But the activity of the
cementitious minerals in steel slag is much lower than that in
Portland cement [9]. It is related to the crystalline state, the
cementitious phases of steel slag crystallize much better than that
of Portland cement clinker owe to the low cooling rate of steel slag.
Meanwhile, steel slag has large amount of non-active components,
such as RO phase and Fe3O4.
The effect of steel slag on the early hydration of Portland
cement has been studied by several researchers. Kourounis et al.
[6] investigated the properties and hydration of blended cements
with steel slag and found that blended cement developed lower
early-age strength compared to Portland cement and its strength
decrease was high when the content of steel slag was high. The
addition of steel slag slowed down the hydration of blended
cements, and the blended cements showed longer setting time
than Portland cement. Tsakiridis et al. [8] studied the utilization
of steel slag for Portland cement clinker production and pointed
out that the addition of steel slag by 10.5% in the raw meal did not
affect the hydration process during Portland cement production.
Wang et al. [10] found that the dormant period of blended cement
containing steel slag is longer than that of Portland cement.
Monshi and Asgarani [11] found that blending 10% extra iron slag
to a cement composed of 49% iron slag, 43% calcined lime and 8%
steel slag kept the compressive strength of concrete for type I

44

F. Han et al. / Thermochimica Acta 605 (2015) 4351

Table 1
Chemical compositions of cement and steel slag (w/%).
Composition

SiO2

Al2O3

Fe2O3

CaO

MgO

SO3

TiO2

P2 O 5

Na2Oeq

f-CaO

LOI

Cement
Steel slag

20.55
12.77

4.59
2.12

3.27
23.49

62.50
49.17

2.61
3.54

2.93
0.23

1.02

0.91

0.53
0.45

0.83

2.08
1.86

Na2Oeq = Na2O + 0.658 K2O; w: mass fraction.

Portland cement. Kriskova et al. [12] found that mechanical

activation signicantly increased the reactivity of steel slag due to
the increase in surface area and amorphous phase, and the
amount of heat released per surface area increased with milling
time. Wang et al. [13] also found that super-ne steel slag
exhibited a much higher activity at early ages, but its activity was
still obviously lower than that of Portland cement. Liu and Li [14]
pointed that the early hydrations of cement and steel slag tended
to hinder each other, especially in the case of large replacement
ratio of steel slag and high neness of steel slag. Li et al. [15]
proposed a new method for modifying the chemical and
mineralogical compositions of steel slag by adding industrial
wastes to molten basic oxygen furnace steel slag, the modication
reduced the free-CaO content and remarkably improved the
cementitious property of steel slag. Wang et al. [16] found that the
initial alkalinity of hydration condition could promote the early
hydration of active components in steel slag, but the hydration
degree of non-active components is very low even under strong
alkaline condition with pH value of 13.8.
Based on a review of current literatures, it is obvious that
blended cement has a low early-age strength and long initial
setting time by replacing part of cement with steel slag. In order to
improve the activity of steel slag, methods of mechanical
activation, mineral modication and alkalinity activation are used.
However, the hydration heat evolution of blended cement
containing steel slag which belongs to the main characteristics
of the early-stage hydration process was investigated very rarely in
the past. Meanwhile, there is little information regarding the
hydration kinetics of blended cement containing steel slag, which
is important to understand the hydration process and mechanism
of blended cement. Moreover, most investigations are carried out
with small amount or single content of steel slag or at room
temperature. Therefore, study of detailed heat evolution behavior
and kinetics of blended cement containing steel slag are needed. In
this paper, the hydration heat evolution rate and cumulative
hydration heat of blended cement containing 0, 10%, 20%, 35% and
50% steel slag were measured within 168 h at 25
C, 45
C and 60
C
by an isothermal calorimeter. By applying the kinetics equations
[1719], the reaction of blended cement was evaluated as a
function of hydration time, taking into account the steel slag
content and curing temperature. Then the hydration reaction
mechanism of blended cement containing steel slag was
determined.

Fig. 1. XRD pattern of steel slag.

2. Experimental
2.1. Materials
P.I 42.5 Portland cement conforming to Chinese National
Standard GB175-2007 and basic oxygen furnace steel slag were
used. The chemical compositions of cement and steel slag are
shown in Table 1. The mineralogical phases of steel slag, which
were determined by XRD analysis, using a TTR IIIX-ray diffractometer (Cu Ka, 45 kV, 200 mA), are given in Fig. 1. According to the
calculation method of alkalinity proposed by Mason [20], the
alkalinity of steel slag is 3.59, which is indicated that steel slag used
in this paper has high alkalinity. As shown in Table 1, CaO, SiO2,
Fe2O3, MgO, Al2O3 are the main oxides of steel slag, its chemical
compositions are similar to that of Portland cement. But there are
signicant differences in mineralogical phases between Portland
cement and steel slag. The main difference in steel slag is the high
contents of iron oxide and RO phase, which have no cementitious
properties and do not combine to form hydraulic phases.
The specic surface areas of Portland cement and steel slag
are 350 m2/Kg and 458 m2/Kg, respectively. The particle size
distribution which was measured by a laser particle size analyzer
(MASTER SIZER 2000) is presented in Fig. 2. It is evident that steel
slag was ner than cement in the range of ne particles, up to
6 mm, and steel slag was coarser than cement when the particle
size was greater than 60 mm, while the particle size of steel slag
between 6  60 mm are much less than cement.
2.2. Test method
The hydration heat evolution rate and cumulative hydration
heat of blended cement containing steel slag were measured with
an isothermal calorimeter (TAM Air from TA instruments). The
tests were performed at three constant temperatures (25
C, 45
C
and 60
C) within 168 h. TAM Air has eight parallel twin-chamber
measuring channels: one chamber containing the sample, another
containing the reference. In order to avoid considerable temperature differences between the paste and the isothermal
environment, the binder and water were kept at a temperature
close to the measurement temperature before mixing. After
manually mixing homogeneously, the pastes were immediately
placed into the chamber maintained at a constant temperature.
The hydration heat evolution rate and cumulative hydration heat of

Fig. 2. Particle size distribution of cement and steel slag.

F. Han et al. / Thermochimica Acta 605 (2015) 4351

Table 2
Mix proportions of pastes.
Sample

W/B

Mass fraction (%)

Cement

Steel slag

Cem
SS10
SS20
SS35
SS50

0.4

100
90
80
65
50

0
10
20
35
50

blended cement containing steel slag can be continuously

monitored as a function of time.
In order to investigate the hydration properties of steel slag, the
steel slag pastes were prepared by mixing steel slag with water or
sodium hydroxide solutions with pH value of 13.0 and 13.6. The
water to binder ratio and sodium hydroxide solution to steel slag
ratio (W/B) were 0.4 for all samples. The mix proportions of pastes
are shown in Table 2.
3. Results and discussion
3.1. Hydration heat evolution of steel slag
The hydration heat evolution rate and cumulative hydration
heat of steel slag with water or NaOH solution to steel slag ratio of
0.4 at 25
C are shown in Fig. 3. It can be seen from Fig. 3(a) that the
hydration process of steel slag can be divided into ve stages,
which is similar to that of cement. The rst stage was associated to
the rst peak of the curve, it corresponded to the rst minutes of
reaction and it was related to the partial dissolution of steel slag.
For pastes mixing steel slag with water, NaOH solutions of pH
13.0 and pH 13.6, the values of the rst peak were 56.2 J/gh, 66.5 J/
gh and 89.5 J/gh, respectively. Then the hydration heat evolution
rate decreased rapidly, and reaction entered the second stage
called dormant period in which the concentration of Ca2+ needed
to reach saturation state before further hydration of steel slag [21].
The third and fourth stages are associated to the second peak. The
hydration of active components in steel slag such as C2S, C3S made
reaction products precipitate, and then with the reduction of active
phases, the hydration heat evolution rate decreased. Finally,
the hydration process entered a low reactivity period and the
exothermic rate decreased to a very low level.
As shown in Fig. 3(a), the dormant periods of pastes mixing
steel slag with water and NaOH solution of pH 13.0 were 43.3 h and
36.4 h, respectively. For paste mixing steel slag with NaOH solution
of pH 13.6, the hydration process almost did not experience
dormant period. The occurring time of the second exothermic peak
was about 80 h, 40 h and 20 h for above mentioned three pastes,
respectively. The peak value of paste mixing steel slag with NaOH

45

solution of pH 13.0 is a little larger than that of paste mixing steel

slag with water. But increasing alkalinity of solution to pH
13.6 increased the value of second exothermic peak nearly by three
times. It is indicated that steel slag was stimulated by NaOH
solution. It is due to the fact that OH is helpful to disintegrate
vitreous and thus accelerate the hydration of C2S and C3S [22]. The
promoting effect was not obvious when pH 13.0 (Fig. 3(a) and (b)),
but the cumulative hydration heat still increased after 168 h by
increasing the alkalinity of solution to pH 13.6. The hydration of
cement could make the pH value a little higher than 13.0 in a short
time [23], even for concrete with large amount of mineral
admixture, the pH value of its pore solution was higher than
12.0 [24]. As shown in Fig. 3, the activity of steel slag was still low
when pH 13.0. Therefore, the hydration heat evolution rate and
cumulative hydration heat were much lower than that of cement.
3.2. Hydration heat evolution of blended cement containing steel slag
The hydration heat evolution rate and cumulative hydration
heat of blended cement containing steel slag at 25
C are shown in
Figs. 4 and 5, respectively. As shown in Fig. 4(a), in the initial
hydration time up to 1 h, the hydration heat evolution rate
increased with an increase in steel slag content in blended cement.
For blended cement, steel slag almost acted as inert material in the
initial period due to its low activity (Fig. 3), an increase in the
replacement ratio of steel slag increased the effective water to
cement ratio, and there was more space for the hydration products
of cement. The rst maximum peak was formed in a short time due
to the release of surface energy and the fast reaction of aluminates
when blended cement contacted with water [25]. At the same
time, the cumulative hydration heat increased with the increasing
amount of steel slag (Fig. 5(a)).
It can be seen from Fig. 4(b) that the ending time of dormant
period increased with increasing dosage of steel slag. This was in a
good agreement with the investigations that the setting time of
cement and concrete became longer by replacing part of cement
with steel slag, and an increase in steel slag content increased the
setting time [6,10,26,27]. The replacement of Portland cement by
steel slag resulted in a decreased hydration heat evolution rate and
a decreased peak value of the second maximum as well as
increased occurring time of the second exothermic peak compared
to Portland cement (Fig. 4(b)). Moreover, a reduction in the total
released heat with the increasing dosage of steel slag was observed
(Fig. 5(b)). The promoting effect of steel slag on the hydration of
blended cement is shown only in the initial hydration time, the
cumulative hydration heat of Portland cement exceed that of
blended cement after about 5.5 h (Fig. 5(a)). The reaction of steel
slag was much slower than that of clinker phases, and the amount
of cement decreased with increasing dosage of steel slag,
meanwhile, the specic surface area of steel slag was only a little

Fig. 3. Hydration heat evolution of steel slag at 25
C (a) hydration heat evolution rate and (b) cumulative hydration heat.

46

F. Han et al. / Thermochimica Acta 605 (2015) 4351

Fig. 4. Hydration heat evolution rate of blended cement containing steel slag at 25
C (a) the rst peak and (b) hydration within 168 h.

larger than that of Portland cement and large particles existed in

steel slag (Fig. 2), so the particles of steel slag cannot act as
nucleation site for the hydration products to deposit and growth
during the process of nucleation and crystal growth. Therefore, the
hydration heat evolution rate and cumulative hydration heat
decreased with an increase in steel slag content. It is noted that the
cumulative hydration heat of blended cement containing 10% steel
slag was almost as much as that of Portland cement within 168 h. It
was consistent with the Tsakiridiss result that addition of 10.5%
steel slag had no inuence on the hydration process of Portland
cement [8].
Figs. 6 and 7 present the hydration heat evolution and
cumulative hydration heat of blended cement containing steel
slag at 45
C, respectively. As shown in Fig. 6(a), at the temperature
of 45
C, the peak value of the rst heat evolution was lower than at
25
C except blended cement containing 35% steel slag. This
phenomenon was related to the increasing rate of the C3A
hydration in the presence of gypsum with increasing temperature.
This reaction leaded to the formation of primary ettringite on the
surface of the grains, which prevented the further hydration for a
certain period of time [25]. For blended cement containing 35%
steel slag, the hydration heat evolution rate and cumulative
hydration heat were higher than other samples in the initial
hydration time with increasing temperature from 25
C to 45
C
(Fig. 6(a) and Fig. 7(a)). It might be due to the combined effect of
elevated temperature, increased water to cement ratio and
sufcient quantity of cement, which promoted signicantly the
initial hydration of sample SS35. It is apparent that the cumulative

hydration heat of Portland cement was more than that of blended

cement at about 4.0 h, which was obviously shortened (Fig. 6(a)
and Fig. 7(a)).
It can be seen from Fig. 6(b) that the reaction temperature
affected the hydration process of blended cement containing steel
slag. An increase in temperature from 25
C to 45
C reduced the
ending time of dormant period, shortened the occurring time of
the second hydration heat evolution peak by about two times and
increased the value of the second exothermic peak by about three
times. The intense reaction continued for more than 40 h at 25
C,
but it shortened for about 20 h at 45
C. Thus, the second
exothermic peak with a high but narrow shape was formed. As
shown in Fig. 7(b), the early-stage cumulative hydration heat of
blended cement had a considerable increase compared to samples
hydrating at 25
C. It was found that the 12 h accumulative
hydration heat increased by 118.14% for Portland cement with
increasing temperature from 25
C to 45
C, while 124.01%, 156.12%,
155.92% and 119.60% for samples SS10, SS20, SS35 and SS50,
respectively. It is elucidated that elevated temperature promoted
the early-age hydration of blended cement more obviously than
that of Portland cement and the promoting effect became great
with an increase in steel slag content in blended cement.
Compared to hydrating at 25
C, after 168 h of hydration, the
cumulative hydration heat increased by 2.22%, 0.68%, 18.14%,
35.36% and 27.99% for samples Cem SS10, SS20, SS35 and SS50,
respectively. It is indicated that the hydration of blended cement
containing large amount of steel slag developed signicantly at
later stage of hydration. It also can be seen from Fig. 7(b) that the

Fig. 5. Cumulative hydration heat of blended cement containing steel slag at 25
C (a) the initial hydration time and (b) hydration within 168 h.

F. Han et al. / Thermochimica Acta 605 (2015) 4351

47

Fig. 6. Hydration heat evolution rate of blended cement containing steel slag at 45
C (a) the rst peak and (b) hydration within 168 h.

cumulative hydration heat of Portland cement and blended cement

containing no more than 35% steel slag had little difference in the
later period. But there was still certain gap between Portland
cement and blended cement containing 50% steel slag due to small
mass fraction of cement and low activity of steel slag.
Figs. 8 and 9 show the hydration heat evolution rate and
cumulative hydration heat of blended cement at 60
C, respectively. It can be seen from Fig. 8(a) that the value of the rst exothermic
peak was lower at 60
C compared to hydrating at 25
C. This was a
consequence of the accelerated formation of primary ettringite on
the surface of cement grains at 60
C as it was explained before. It
should be noted that increasing temperature to 60
C promoted
obviously the initial hydration of samples SS20 and SS35 (Fig. 8(a)
and Fig. 9(a)).
As shown in Fig. 8(b), as expected, an increase in temperature
from 45
C to 60
C reduced the ending time of dormant period and
increased the value of the second exothermic peak further,
meanwhile, it shifted the peak to earlier time. After about 10 h
of hydration, the hydration of blended cement containing steel slag
entered a diffusion controlling process. The cumulative hydration
heat also increased signicantly as increasing temperature
(Fig. 9(b)). When temperature raised from 25
C to 60
C, the
12 h accumulative hydration heat increased by 179.10%, 194.98%,
251.32%, 268.77% and 233.43% for samples Cem SS10, SS20,
SS35 and SS50, respectively. The activities of cement and steel slag
were greatly stimulated at the temperature of 60
C. Elevated
temperature could obviously promote the early hydration of active

components in steel slag such as C2S and C3S. Moreover, fast

reaction of cement in a short time resulted in a hydration
environment with high alkalinity, which was benecial to the
hydration of steel slag (Fig. 3). Thus, for blended cement containing
steel slag, much more hydration heat generated at the early stage
of hydration at 60
C. Compared to hydrating at 25
C, the
cumulative hydration heat within 168 h increased by 15.49%,
7.47%, 26.47%, 22.97% and 32.26% for samples Cem SS10, SS20,
SS35 and SS50, respectively. It can be found that the amplitude of
increase of hydration heat of blended cement in the later period
was relatively small compared to Portland cement at 60
C. It is due
to the fact that most active phases of steel slag had hydrated in the
early period and the inert components (RO phase, Fe2O3) cannot
react even in the condition of strong alkalinity and elevated
temperature [21]. The hydration of steel slag was different from
that of pozzolanic mineral admixture, such as granulated blast
furnace slag or y ash, which will react with Ca(OH)2 produced by
cement at the later stage of hydration, moreover, elevated
temperature provided energy to activate alkali-hydroxide attack
on the slag or y ash particles [28]. For blended cement containing
small amount of granulated blast furnace slag, the cumulative
hydration heat was more than Portland cement at 60
C due to the
released latent heat of crystallization of amorphous phase in slag
[29]. However, steel slag did not react with the hydration products
of cement. Steel slag and cement affected each others hydration by
changing the hydration environment [10]. Thus, for steel slag in
blended cement, increasing temperature just promoted the

Fig. 7. Cumulative hydration heat of blended cement containing steel slag at 45
C (a) the initial hydration time and (b) hydration within 168 h.

48

F. Han et al. / Thermochimica Acta 605 (2015) 4351

Fig. 8. Hydration heat evolution rate of blended cement containing steel slag at 60
C (a) the rst peak and (b) hydration within 168 h.

hydration of active components and made these parts of heat emit

ahead of time. Therefore, for blended cement containing 10% steel
slag, the cumulative hydration heat was close to Portland cement
at 25
C and 45
C in the later period (Fig. 5(b) and Fig. 7(b)), and it
had a certain gap with Portland cement at 60
C (Fig. 9(b).
Meanwhile, the cumulative hydration heat of blended cement
containing steel slag cannot exceed that of Portland cement within
168 h at elevated temperatures (Fig. 7(b) and Fig. 9(b)).

1 indicates that the process is governed by phase boundary

kinetics, and the value is equal or larger than 2 indicates that the
kinetics corresponds to a diffusion process.
The degree of reaction, a, which can be represented as the ratio
of the cumulative hydration heat at time, t, to the ultimate total
hydration heat, Qmax, and it is shown in Eq. (2).

Q t
Q max

(2)

3.3. Hydration kinetics of blended cement containing steel slag

Several studies have described the hydration kinetics of
Portland cement and many kinetics models have been built to
mimic the hydration process of Portland cement [30]. In order
to investigate the hydration kinetics of blended cement containing
steel slag, a hydration kinetics equation was used in this paper. This
equation, the general form of which is presented in Eq. (1), has also
been used to study the hydration reaction of Portland cement and
blended cement containing granulated blast furnace slag [31], the
alkali-activated reaction of slag [17,32] and the pozzolanic reaction
of metakaolin [33] or rice husk ash [34].
h
in
1  1  a1=3 Kt
(1)
where a is the degree of hydration; K is the rate constant; t is the
hydration time; n is a constant related with the hydration
mechanism. The value of n is smaller than 1 indicates that
nucleation kinetics controls the reaction rate, while it is closer to

A formula of hydration kinetics proposed by Knudson (Eq. (3))

[19] is used to determine the ultimate total hydration heat, Qmax.
1
1
t

50 t  t0
Q t Q max Q max

(3)

where t50 is the hydration reaction time when the cumulative

hydration heat is 50% of the total hydration heat (i.e., half-life
period); (tt0) is the hydration time starting from the acceleration
period.
The curves of ln[1-(1-a)1/3] vs. ln(tt0) relationship for blended
cement containing steel slag are shown in Fig. 10. It can be seen
from Fig. 10 that Eq. (1) has been simulated well during the
acceleration period (Fig. 10 line AB) and decay period (Fig. 10 line
CD), but Eq. (1) was not applicable for deceleration period due to
the fact that the corresponding line of deceleration period was a
curve (Fig. 10 line BC). For deceleration period, Equation [1-(1-a)1/
3
] = Kln(tt0) has been tted onto the [1-(1-a)1/3] vs. ln(tt0)

Fig. 9. Cumulative hydration heat of blended cement containing steel slag at 60
C (a) the initial hydration time and (b) hydration within 168 h.

F. Han et al. / Thermochimica Acta 605 (2015) 4351

49

Fig. 10. Curves of ln[1-(1-a)1/3] vs. ln(tt0) and segmentally simulate lines for: (a) Cem, (b) SS10, (c) SS20, (d) SS35 and (e) SS50. Symbols represent the experimental data and
the red solid lines show the stimulated lines. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

relationship for blended cement containing steel slag. Fig. 11

presents an example of SS 20 to elucidate this relationship. The
kinetics parameters of blended cement containing steel slag at
different temperatures are shown in Table 3.
As shown in Fig. 10, apparently, for all samples studied in this
paper, the reaction fast entered the acceleration period, and after
the intense reaction, the time of entering the decay period was also
shortened with increasing temperatures. These results were
agreement with the hydration heat evolution rate and cumulative
hydration heat (Figs. 49).
It can be seen from Table 3 that the values of n for blended
cements containing steel slag were smaller than 1 in acceleration
period, which is indicated that the hydration reaction rate was
controlled by nucleation kinetics. While the values of n became

larger than 2 in decay period, elucidating that the hydration

process became a diffusion controlling process at later stage of
hydration. Due to the different controlling mechanism, the
reaction rate in acceleration period was about 10  20 times of
that in decay period. The value of n was also related to the reaction
resistance of blended cement. As shown in Table 3, the value of n
increased with an increase in replacement ratio of steel slag at
three examined temperatures. It is indicated that increasing steel
slag content increased the reaction resistance of blended cement.
The activities of active phases in steel slag such as C2S and C3S were
low due to the slowly cooling resulting in a perfect crystallization.
C3S is the main cementitious phase of cement, but cluster-gathered
C2S with large size and nger structure is the main cementitious
phase of steel slag [6]. Thus, the reaction resistance of steel slag

50

F. Han et al. / Thermochimica Acta 605 (2015) 4351

Table 4
Apparent activation energy of the overall reaction of blended cement containing
steel slag.

Fig. 11. Curves of [1-(1-a)1/3] vs. ln(tt0) and simulate lines in deceleration period
for sample SS20. Symbols represent the experimental data and the red solid lines
show the stimulated lines. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)

was higher than that of Portland cement. As expected, the reaction

rate, K, decreased with increasing steel slag content in three
hydration periods (Table 3). It was found that increasing
temperature reduced the values of n in acceleration period. This
was related to the overall acceleration of hydration kinetics with
increasing temperature. Elevated temperature promoted the
hydration of Portland cement and active components of steel slag
(Figs. 69) and increased the reaction rate, K, of blended cement as
well (Table 3). However, increasing temperature increased the
values of n in the decay period because that diffusion of reactants
through the reaction product layer to occur reactions on the
surfaces of unhydrated particle was the main controlling factor.
Elevated temperature leaded to form a dense microstructure with
plenty of hydration products, so it resulted in a higher reaction
resistance compared to hydrating at 25
C. It is noted that
increasing temperature increased the reaction rate in the decay
period although higher temperature leaded to higher reaction
resistance. It might be because that elevated temperature provided
sufcient energy to overcome the reaction resistance.
Equation [1-(1-a)1/3] = Kln(tt0) was the derivation of empirical
formula dx/dt = K/t and it was applicable for the dual controlling
reaction of chemical reaction and diffusion [31]. It can be seen from
Fig. 10 that blended cement containing no more than 35% steel slag
Table 3
Kinetics parameters of blended cement containing steel slag at different
temperatures.
Sample

T(
C)

Acceleration period

Deceleration period

Decay period

Cem
SS10
SS20
SS35
SS50

25

0.49
0.53
0.50
0.58
0.91

0.02743
0.02249
0.02358
0.01750
0.00659

0.19996
0.20904
0.19202
0.14703
0.10671

2.70
2.44
2.45
2.35
2.34

0.00147
0.00134
0.00142
0.00104
0.00068

Cem
SS10
SS20
SS35
SS50

45

0.35
0.37
0.42
0.54
0.77

0.07470
0.06480
0.05565
0.03511
0.01846

0.26205
0.23540
0.20152
0.13577
0.13914

3.85
3.89
3.94
3.03
4.21

0.00277
0.00188
0.00131
0.00081
0.00074

Cem
SS10
SS20
SS35
SS50

60

0.29
0.34
0.44
0.48
0.70

0.11275
0.10081
0.07419
0.06640
0.03113

0.29604
0.28262
0.24880
0.24664
0.23775

2.89
4.30
5.11
1.38
1.29

0.00433
0.00280
0.00154
0.00170
0.00146

Sample

W/B

Temperature range (
C)

Ea(kJ/mol)

Cem
SS10
SS20
SS35
SS50

0.4

2560
2560
2560
2560
2560

40.01
41.43
42.14
43.58
46.39

experienced the process of dual controlling reaction. But for

blended cement containing 50% steel slag, the value of n was close
to 1 in acceleration period and it became larger than 2 in decay
period. Moreover, as shown in Fig. 10(e), curves BC were almost
straight lines at three examined temperatures. It is elucidated that
the early-stage kinetics behavior of sample SS50 is mostly
dominated by phase boundary reaction, after which it directly
entered diffusion controlling process. For sample SS50, the
hydration heat evolution rate was rapid in the initial hydration
time (Figs. 4, 6, 8(a)), then lots of hydrates deposit on the
unhydrated particles, the reactions occurred on the phase
boundary between crystal and pore solution. Moreover, further
improvement of hydration degree was hindered due to low mass
fraction of cement and low activity of steel slag (Fig. 3).
The apparent activation energy, Ea, of blended cement
containing steel slag was determined from the Eq. (4):


K 01 t502
Ea T 1  T 2

exp
(4)
RT 1 T 2
K 02 t501
where t501 and t502 are the hydration time when the cumulative
heat released is 50% of the total hydration heat at curing
temperature T1 and T2, respectively. R is gas constant (8.314 J/mol).
Table 4 presents the apparent activation energy of the overall
reaction of blended cement containing steel slag in the temperature range of 2560
C. It is evident that an increase in replacement
ratio of steel slag increased the apparent activation energy. It was
related to the low activity of steel slag and large amount of inert
phases as it was explained before. It is noted that the difference in
apparent activation energy between Portland cement and blended
cement containing 10% steel slag was small. It is indicated that
small amount of steel slag affected little the hydration process or
properties of Portland cement, which was consistent with the
hydration heat evolution rate and cumulative hydration heat
(Figs. 49) as well as previous investigations [8,11]. However, Wu
et al. [35] determined Ea = 49 kJ/mol for blended cement containing
50% granulated blast furnace slag (GBFS), which was higher
compared to Ea = 46.39 kJ/mol with respect to blended cement
containing 50% steel slag. The hydration properties of steel slag
were poor when it was mixed with water (Fig. 3), but for GBFS,
almost no reactions occurred as it contacted with water. The glass
phase of GBFS had certain activity and the crystallization activation
energy of vitreous body was about 421256 kJ/mol [36], which
means that much more energy should be provided for the
hydration of GBFS. However, elevated temperature and alkalinity
environment can signicantly activate the hydration of GBFS,
which will react with Ca(OH)2 produced by cement hydration in
blended cement [29]. Compared to GBFS, the hydration mechanism of steel slag in blended cement was different, steel slag did
not react with hydration products of cement. Incorporation of steel
slag in Portland cement could be seen as incorporation of a kind of
material with low hydration activity, in other words, blended
cement containing steel slag could be seen as a low quality of
cement. Therefore, the apparent activation energy of blended
cement containing steel slag was lower than that of blended
cement containing GBFS at the same level of replacement.

F. Han et al. / Thermochimica Acta 605 (2015) 4351

4. Conclusion
(1) The hydration process of steel slag can be divided into ve
stages and it is similar to that of Portland cement. But compared to
cement hydration, the ending time of dormant period and the
occurring time of second exothermic were much longer, and
the value of second exothermic peak was much smaller. The
activity of steel slag was still low when pH value of solution was
13.0, but increasing alkalinity of solution to pH 13.6 stimulated
signicantly the activity of steel slag.
(2) The hydration heat evolution rate and cumulative hydration
heat increased with increasing steel content in blended cement in
the initial hydration time. At the temperature of 45
C or 60
C, the
peak value of the rst heat evolution was lower than that at 25
C
due to the fast reaction of C3A at elevated temperatures.
(3) Increasing steel slag content in blended cement increased
the occurring time of second exothermic and decreased the value
of second exothermic peak as well as reduced the cumulative
hydration heat. An increase in temperature from 25
C to 45
C or
60
C increased the value of second exothermic peak, shifts the
peak to earlier time and obviously increased the cumulative
hydration heat of blended cement. The promoting effect of
elevated temperature was obvious for blended cement containing
high content of steel slag.
(4) Small amount of steel slag affected little the hydration
process of Portland cement, but the retarding effect of steel slag on
the hydration of Portland cement became obvious when the
replacement ratio of steel slag was high.
(5) Increasing temperature could promote the hydration of
active components in steel slag and made these parts of heat emit
ahead of time. The cumulative hydration heat of blended cement
containing steel slag cannot exceed that of Portland cement within
168 h at elevated temperatures.
(6) For blended cement containing no more than 35% steel slag,
the hydration reaction rate was controlled by nucleation kinetics in
the acceleration period and then by diffusion in the decay period,
but in the deceleration period, the hydration experienced dual
controlling reaction of chemical reaction and diffusion. The earlystage kinetics behavior of blended cement containing 50% of steel
slag was mostly dominated by phase boundary reaction, after
which it directly entered diffusion controlling process.
(7) An increase in steel slag content in blended cement
increased the value of n and decreased the reaction rate, K. The
value of n reduced and the reaction rate, K, increased with
increasing temperature in acceleration period, while in decay
period, an increase in temperature increased the value of n,
although the reaction rate, K, increased as well.
(8) The apparent activation energy of blended cement
containing steel slag increased with an increase in steel slag
content.
Acknowledgment
Authors would like to acknowledge the National Natural
Science Foundation of China(grant Nos. U1134008 and 51278277).
References
[1] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious
materials, Cem. Concr. Res. 41 (2011) 12441256.
[2] B. Das, S. Prakash, V.N. Misra, An overview of utilization of slag and sludge from
steel industries, Resour. Conserv. Recycl. 50 (2007) 4057.

51

[3] E. Furlani, G. Tonello, S. Maschio, Recycling of steel slag and glass cullet from
energy saving lamps by fast ring production of ceramics, Waste Manage. 30
(2010) 17141719.
[4] J. Li, Q. Yu, J. Wei, T. Zhang, Structural characteristics and hydration kinetics of
modied steel slag, Cem. Concr. Res. 41 (2011) 324329.
[5] H. Yi, G. Xu, H. Cheng, et al., An overview of utilization of steel slag, Proc.
Environ. Sci. 16 (2012) 791801.
[6] S. Kourounis, S. Tsivilis, P.E. Tsakiridis, et al., Properties and hydration of
blended cements with steel making slag, Cem. Concr. Res. 37 (2007)
815822.
[7] C. Shi, Characteristics and cementitious properties of ladle slag nes from steel
production, Cem. Concr. Res. 32 (2002) 459462.
[8] P.E. Tsakiridis, G.D. Papadimitrious, S. Tsivilis, C. Koroneos, Utilization of steel
slag for Portland cement clinker production, J. Hazard Mater. 152 (2008)
805811.
[9] Q. Wang, P. Yan, J. Yang, B. Zhang, Inuence of steel slag on mechanical
properties and durability of concrete, Constr. Build. Mater. 47 (2013)
14141420.
[10] Q. Wang, P. Yan, S. Han, The inuence of steel slag on the hydration of cement
during the hydration process of complex binder, Sci. China Technol. Sci. 54
(2011) 388394.
[11] A. Monshi, M.K. Asgarani, Producing Portland cement from iron and steel slags
and limestone, Cem. Concr. Res. 29 (1999) 13731377.
[12] L. Kriskova, Y. Pontikes, . Cizer, et al., Effect of mechanical activation on the
hydraulic properties of stainless steel slags, Cem. Concr. Res. 42 (2012)
778788.
[13] Q. Wang, J. Yang, P. Yan, Cementitious properties of super-ne steel slag,
Powder Technol. 245 (2013) 3539.
[14] S. Liu, L. Li, Inuence of neness on the cementitious properties of steel slag, J.
Therm. Anal. Calorim. 117 (2014) 629634.
[15] Z. Li, S. Zhao, X. Zhao, T. He, Cementitious propertt modication of basic oxygen
furnace steel slag, Constr. Build. Mater. 48 (2013) 575579.
[16] Q. Wang, J. Yang, P. Yan, Inuence of initial alkalinity on the hydration of steel
slag, Sci. China Technol. Sci. 55 (2012) 33783387.
[17] A. Fernndez-Jimnez, F. Puertas, Alkali-activated slag cements: kinetics
studies, Cem. Concr. Res. 27 (1997) 359368.
[18] C. Shi, P.V. Krivenko, D. Roy, Alkali-Activated Cements and Concretes, Spon
Press, 2006.
[19] T. Knudsen, On particle size distribution in cement hydration, Paris,
Proceeding of 7th International Congress on the Chemistry of Cement, vol
I1980, pp. 170.
[20] B. Mason, The constitution of some open-heart slag, J. Iron Steel Inst. 11 (1994)
6980.
[21] Q. Wang, P. Yan, Hydration properties of basic oxygen furnace steel slag, Constr.
Build. Mater. 24 (2010) 11341140.
[22] N. Yang, Physical and chemical foundation of alkaline cementitious material, J.
Chin. Ceram. Soc. 24 (1996) 209215 (in Chinese).
[23] H.F.M. Taylor, Cement Chemistry, 2nd ed., Thomas Telford, London, 1997.
[24] Rasheeduzzafar, S.E. Hussain, Effect of microsilica and blast furnace slag on
pore solution composition and alkali-silica reaction, Cem. Concr. Compos. 13
(1991) 219225.
[25] V. Tydlitt, T. Matas, R. Cern, Effect of w/c and temperature on the early-stage
hydration heat development in Portland-limestone cement, Constr. Build.
Mater. 50 (2014) 140147.
[26] I.A. Altun, I. Ylmaz, Study on steel furnace slags with high MgO as additive in
Portland cement, Cem. Concr. Res. 32 (2002) 12471249.
[27] A. Rai, J. Prabakar, C.B. Raju, R.K. Morchalle, Metallurgical slag as a component
in blended cement, Constr. Build. Mater. 16 (2002) 489494.
[28] D.M. Roy, G.M. Idorn, Hydration, structure, and properties of blast furnace slag
cements, mortars, and concrete, ACI J. (1982) 444457.
[29] F. Han, R. Liu, D. Wang, P. Yan, Characteristics of the hydration heat evolution of
composite binder at different hydrating temperature, Thermochim. Acta 586
(2014) 5257.
[30] J.J. Thomas, J.J. Biernacki, J.W. Bullard, S. Bishnoi, et al., Modeling and
simulation of cement hydration kinetics and microstructure development,
Cem. Concr. Res. 41 (2011) 12571278.
[31] X. Wu, Kinetic study on hydration of blast furnace slag cement, J. Chin. Ceram.
Soc. 16 (1988) 423428 (in Chinese).
[32] D. Ravikumar, N. Neithalath, Reaction kinetics in sodium silicate powder and
liquid activated slag binders evaluated using isothermal calorimetry,
Thermochim. Acta 546 (2012) 3243.
[33] J. Cabrera, M.F. Rojas, Mechanism of hydration of the metakaolinlimewater
system, Cem. Concr. Res. 31 (2001) 177182.
[34] Q. Feng, H. Yamamichi, M. Shoya, S. Sugita, Study on the pozzolanic properties
of rice husk ash by hydrochloric acid pretreatment, Cem. Concr. Res. 34 (2004)
521526.
[35] X. Wu, D.M. Roy, C.A. Langton, Early stage hydration of slag-cement, Cem.
Concr. Res. 13 (1983) 277286.
[36] R. Roy, Crystal chemistry of non-metallic materials, Lecture Notes for Materials
Science, The Pennsylvania State Univarsity, 1981, pp. 249.