Activation of Ground Granulated Blast Furnace Slag Cement by Calcined Alunite

Materials Transactions, Vol. 52, No. 2 (2011) pp. 210 to 218
#2011 The Japan Institute of Metals EXPRESS REGULAR ARTICLE
Activation of Ground Granulated Blast Furnace Slag Cement by Calcined Alunite
Hyung-Seok Kim 1 , Joo-Won Park 2 , Yong-Jun An 3 , Jong-Soo Bae 2 and Choon Han 2; *
1 Minerals & Materials Processing Division, Korea Institute of Geoscience & Mineral Resources, Daejeon, Korea
2 Department of Chemical Engineering, Kwangwoon University, Seoul, Korea
3 Department of Resource Recycling Engineering, University of Science and Technology, Daejeon, Korea
To enhance the early strength of grounded granulated blast furnace slag (GGBFS) blended cement, the activation characteristics of
GGBFS were examined by a potassium aluminum sulfate (PSA) clinker, consisting of KAl(SO 4 ) 2 and amorphous Al 2 O 3 by calcining
alunite [K 2 SO 4 Al 2 (SO 4 ) 3 4Al(OH) 3 ] at 650 C for 30 min. The PSA clinker reacted with calcium hydroxide and gypsum to form
ettringite (3CaOAl 2 O 3 3CaSO 4 32H 2 O, AFt) by following reaction: 2KAl(SO 4 ) 2 þ 2Al 2 O 3 þ 13Ca(OH) 2 þ 5(CaSO 4 2H 2 O) þ 74H 2 O !
3(3CaOAl 2 O 3 3CaSO 4 32H 2 O) þ 2KOH. Mortar was prepared by mixing a blended cement of GGBFS and ordinary Portland cement (OPC)
with PSA clinker as activator. The compressive strength of the GGBFS blended cement mortar was compared with that of OPC mortar. When
the PSA clinker and gypsum activator was added to the blended cement of GGBFS and OPC, the hydration products investigated by DTA and X-
ray diffraction were mainly ettringite and calcium silicate hydrate(C-S-H) gel. The early and long-term strengths of the GGBFS blended cement
were higher than those of OPC. Therefore, PSA clinker as activator was shown to improve the early and long-term strengths of GGBFS blended
cement. [doi:10.2320/matertrans.M2010350]
(Received October 14, 2010; Accepted November 11, 2010; Published December 22, 2010)
Keywords:
ground granulated blast furnace slag (GGBFS), activator, alunite, potassium aluminum sulfate, ettringite
1. Introduction
Grounded granulated blast furnace slag (GGBFS) is a
glassy granular material formed when molten GGBFS is
rapidly cooled, usually by immersion in water, and then
ground to improve its reactivity. The major components of
GGBFS are SiO 2 , CaO, MgO, and Al 2 O 3 , which are common
components in commercial silicate glasses. It has been used
as a pozzolanic admixture in Portland cement paste. 1–6)
GGBFS is effective in reducing the hydration heat of
cement and has a high resistance to freezing, thawing,
chemicals and seawater. 7,8) It is therefore, recommended for
concrete structures that require high durability.
GGBFS is also environmentally-friendly as it reduces the
use of ordinary Portland cement (OPC) clinkers in proportion
to the amount of GGBFS that is substituted for OPC. It also
decreases the amount of CO 2 (g) generated from the thermal
decomposition of limestone, a material used to produce OPC
clinkers. 9) However, the use of GGBFS as a cement additive
has its drawbacks, including delaying the setting time of
concrete, 10) which prolongs construction.
Although slag without an activator does react with water,
the rate of hydration is very slow. Its hydraulic reactivity
depends on chemical composition, glass phase content,
particle size distribution, and surface morphology. 1–3)
The reduction of initial strength of GGBFS cement could
be overcome if the fineness of GGBFS were increased to
promote hydration speed. However, increasing the fineness
of slag by pulverization could increase the manufacturing
cost of GGBFS.
A coating of aluminosilicate forms on the surfaces of slag
grains within a few minutes of exposure to water, and these
coatings were impermeable to water. Unless a chemical
activator is present, further hydration is inhibited. In general,
Portland cement, gypsum, and many alkalies have been
*Corresponding author, E-mail: chan@kw.ac.kr
used as activators and the rate of hydration is faster at high
alkali concentrations. The surface of slag is amorphous,
and its dissolving behavior is very similar to that of silicate
glasses. 3)
GGBFS is a low performance cementitious material,
which can achieve high compressive strength when an
alkaline activator is used. However, crucial investigations
about the activation of GGBFS had been already made. 11)
Alkaline activators are classified in six groups; 12) Misan
alkaline:
(a) Caustic alkalis, MOH
(b) Non-silicate weak acid salts: M 2 CO 3 ,M 2 SO 3 ,M 3 PO 4 ,
MF, etc.
(c) Silicates, M 2 OnSiO 2 .
(d) Aluminates, M 2 OnAl 2 O 3 .
(e) Aluminosilicates, M 2 OnAl 2 O 3 (2-6)SiO 2 .
13,14)
(f) Non-silicate strong acid salts, M 2 SO 4
Of all these activators, NaOH, Na 2 CO 3 ,Na 2 OnSiO 2 and
Na 2 SO 4 are the most widely available and economical
chemicals. Some potassium compounds have been used in
laboratory studies. However, their potential applications
will be very limited due to their availability and costs.
Conversely, the properties of sodium and potassium compounds
are very similar.
This study compares the compressive strength and hydration
properties of GGBFS mixed cement with the calcined
alunite (KAl(SO 4 ) 2 and amorphous Al 2 O 3 ) as the activator,
which could be regarded as aluminates or non-silicate strong
acid salts, with those of OPC.
2. Experimental Methods
2.1 Raw materials
OPC was obtained from a Korean cement company,
alunite from Gasa Island in Korea, and GGBFS from a
Korean steel mill. GGBFS was in an amorphous state, its
basicity [b ¼ðCaO þ MgO þ Al 2 O 3 Þ=SiO 2 ] was 1.85 and its

Activation of GGBFS Cement by Calcined Alunite 211
Table 1
Chemical composition of raw materials.
Raw
materials
Ig. loss SiO 2 Fe 2 O 3 Al 2 O 3 CaO MgO Na 2 O K 2 O SO 3
OPC — 19.68 5.48 5.08 60.66 1.92 — — 1.45
GGBFS — 34.30 0.11 14.50 43.15 5.61 0.26 0.24 0.04
Gypsum — 2.96 0.38 0.41 29.38 0.40 — — 15.74
Alunite 37.99 6.67 1.16 33.86 0.10 0.01 0.72 9.33 35.46
Table 2
Chemical composition of calcined alunite.
Composition SiO 2 Al 2 O 3 Fe 2 O 3 K 2 O Na 2 O TiO 2 CaO MgO P 2 O 3 SO 3
mass% 7.60 38.60 1.32 10.64 0.82 0.10 0.12 0.10 0.27 40.42
100
Table 3
The tested activators.
Weight, mass %
Intensity, arb. unit
90
80
70
60
0
P
10
Fig. 1
Fig. 2
473
673
1073
1273
fineness [blaine] was 4,060 cm 2 /g. The chemical compositions
of OPC, GGBFS, gypsum and alunite are shown in
Table 1.
When alunite is heated between 500 and 600 C, as shown
in Fig. 1, it becomes PSA clinker, consisting of KAl(SO 4 ) 2
and Al 2 O 3 by the following dehydration reaction: 15,16)
K 2 SO 4 Al 2 (SO 4 ) 3 4Al(OH) 3
! 2KAl(SO 4 ) 2 þ 2Al 2 O 3 þ 6H 2 O ð1Þ
The PSA clinker used as activator was prepared by
sintering alunite at 550 C for 1 h. Figure 2 and Table 2 show
the results of XRD analysis and the chemical composition of
PSA clinker, respectively.
873
Temperature, T/K
TGA
DTA
TG and DTA analyses of alunite.
Q
20
Q
P
30
P
P
40
Diffraction angle, 2θ/degree
P: KAl(SO 4 ) 2
Q: Quartz
Q
50
XRD pattern of calcined alunite.
Q
60
∆T
Activators
Add amount (mass%)
Gypsum 0, 7.5, 15
PSA clinker 0, 7.5, 15
PSA + Gypsum 0, 7.5, 15
0.9
1.0
0.0
0.8
Fig. 3
0.7
0.1
Activator
0.6
0.5
0.2
0.4
0.3
0.3
0.2
0.1
0.4
0.0
0.5
0.6
12 8 4
0.5
11 7 3
0.4
10 6 2
0.3
9 5 1
0.2
0.0
1.0
2.2 Method
Three activators were used: gypsum, PSA clinker, and a
mixture of the two with a mass ratio of 100 : 21 as shown
in Table 3. In the OPC-GGBFS-activator ternary system,
the activators were mixed at 0, 7.5 and 15.0%, as shown in
Fig. 3.
Mortars were prepared, stored and tested according to the
method described by KSL ISO 679. Blended cement, sand
and water were mixed with a mass ratio of 1:3:0:5 as
shown in Table 4. Prepared mortar was stored in 5 cm cubic
molds, in a humidity chamber. After 24 h, it was removed and
cured in water at 20 C. The compressive strengths of the
mortars were measured using a universal testing machine
(Heung Jin, Korea).
Hydration of the slag was performed by cement paste being
prepared with a 0.5 water/cement ratio and curing in water at
23 2 C. The hydration reaction of the cement paste was
interrupted with acetone and hydrate analyzed with XRD
(Philips, Netherlands) and TG-DTA (Shimadzu, Japan).
1.0
OPC
0.9
0.6
0.8
0.7
0.7
Slag
0.8
experimental
0.9
0.1
Ternary diagram of the admixture of raw materials.

212 H.-S. Kim, J.-W. Park, Y.-J. An, J.-S. Bae and C. Han
Table 4 The tested mortar (KSL ISO 679).
Table 5
Effect of GGBFS on cement’s compressive strength.
Materials
Admixture ratio (mass ratio)
OPC + slag + activator 1
Sand 3
Water 0.5
Mortar
Dosage (mass%) Compressive strength (MPa)
OPC GGBFS 3 days 7 days 28 days 56 days
S1 100 0 18.3 23.7 32.4 36.8
S2 70 30 16.0 18.4 36.3 44.0
S3 60 40 15.8 20.7 40.5 48.4
S4 50 50 12.7 21.6 43.2 48.8
S5 40 60 12.4 24.3 44.0 46.9
Table 6
Effect of gypsum on cement’s compressive strength.
Mortar
Dosage (mass ratio)
Compressive strength (MPa)
OPC GGBFS gypsum 3 days 7 days 28 days 56 days
G1 0.625 0.3 0.075 12.3 21.1 35.1 37.1
G2 0.525 0.4 0.075 11.5 26.2 38.1 40.0
G3 0.425 0.5 0.075 11.2 28.6 38.5 43.5
G4 0.325 0.6 0.075 15.6 29.6 41.3 44.6
G5 0.55 0.3 0.15 9.3 12.0 17.5 19.3
G6 0.45 0.4 0.15 8.7 13.1 18.8 28.1
G7 0.35 0.5 0.15 8.5 14.1 21.1 37.3
G8 0.25 0.6 0.15 8.1 17.4 27.3 35.6
G: Gypsum (CaSO 4 2H 2 O)
Table 7
Effect of PSA on cement’s compressive strength.
Mortar
Dosage (mass ratio)
Compressive strength (MPa)
OPC GGBFS Activator 3 days 7 days 28 days 56 days
P1 0.625 0.3 0.075 19.5 35.4 44.6 46.5
P2 0.525 0.4 0.075 15.3 33.5 43.5 49.4
P3 0.425 0.5 0.075 15.6 32.7 40.1 45.3
P4 0.325 0.6 0.075 18.1 23.7 29.8 31.9
P5 0.55 0.3 0.15 3.5 4.6 5.1 7.9
P6 0.45 0.4 0.15 3.1 4.3 4.7 6.6
P7 0.35 0.5 0.15 — — — —
P8 0.25 0.6 0.15 — — — —
P: PSA (potassium sulfoaluminate + amorphous Al 2 O 3 )
3. Results and Discussion
3.1 Compressive strength property
Table 5 shows the compressive strengths of the mortars at
each curing time according to the mixture of OPC and
GGBFS. The compressive strengths of OPC after 3, 7, 28 and
56 days were 18.3, 23.7, 32.4 and 36.8 MPa, respectively.
After 3 days, the compressive strength of the blended cement
consisting of OPC and GGBFS had decreased from 88% to
68% of that of OPC with increasing GGBFS content from
30 to 60%. However, increasing GGBFS content led to the
compressive strength of GGBFS blended cement surpassing
that of OPC after 28 days.
Table 5 shows the compressive strengths of the blended
cements being lower than those of OPC for up to 7 days.
This is why gypsum is generally used as an activator
to enhance the compressive strength of GGBFS blended
cement.
Table 6 shows the compressive strengths of mortars at
various curing times based on the proportions of the mixture
of OPC, GGBFS and gypsum.
When 7.5% gypsum was used as activator, the compressive
strength of GGBFS blended cement after 3 days was
lower than that of OPC when the proportion of GGBFS was
below 60% but surpassed that of OPC after 7 days. With 15%
gypsum, the compressive strength of GGBFS blended cement
was always lower than that of OPC indicating that excessive
addition of gypsum reduces the early and long-term strengths
of GGBFS blended cement.
Although the long-term strength of blended cement
consisting of OPC, BFS and gypsum improved with
increasing BFS content, its early strength was lower than
that of OPC. Therefore, a PSA clinker of KAl(SO 4 ) 2 and
amorphous Al 2 O 3 was used as activator instead of gypsum.
The results of compressive strength tests of GGBFS cement
blended by using solely PSA clinker as activator are shown in
Table 7. Table 8 shows results from the use of a mixed
activator of gypsum and PSA.
Blended cement with PSA clinker, GGBFS and OPC
contents of 7.5, 30 and 62.5%, respectively, had a higher
compressive strength than OPC. With GGBFS contents of
>40%, the compressive strength of the blended cement was

Activation of GGBFS Cement by Calcined Alunite 213
Table 8
Effect of PSA and gypsum admixture on cement’s compressive strength.
Mortar
Dosage (mass ratio)
Compressive strength (MPa)
OPC GGBFS Activator 3 days 7 days 28 days 56 days
PG1 0.625 0.3 0.075 24.1 32.1 43.3 46.9
PG2 0.525 0.4 0.075 20.0 30.7 45.5 49.3
PG3 0.425 0.5 0.075 22.1 33.6 46.9 49.4
PG4 0.325 0.6 0.075 18.2 27.5 42.3 45.6
PG5 0.55 0.3 0.15 23.0 32.5 44.6 42.0
PG6 0.45 0.4 0.15 25.4 32.6 45.0 46.9
PG7 0.35 0.5 0.15 21.7 30.7 37.9 40.7
PG8 0.25 0.6 0.15 5.6 15.8 24.9 32.8
PG: PSA + gypsum [1 :1(mass ratio)]
(a) 3-day strength, MPa
(b) 7-day strength, MPa
(c) 28-day strength, MPa
(d) 56-day strength, MPa
Fig. 4 Composition for optimum compressive strengths of OPC-GGBFS-gypsum cements: strengths after (a) 3, (b) 7, (c) 28, and (d) 56
days.
initially lower than that of OPC but became higher after 7
days. With 15.0% PSA clinker, mortars expanded excessively
with negligible development of compressive strength.
However, both 7.5 and 15% of the mixed activator of PSA
clinker and gypsum formed GGBFS blended cement with
compressive strengths higher than those of OPC at all curing
times for GGBFS contents of

214 H.-S. Kim, J.-W. Park, Y.-J. An, J.-S. Bae and C. Han
(a) 3-day strength, MPa
(b) 7-day strength, MPa
(c) 28-day strength, MPa
(d) 56-day strength, MPa
Fig. 5
Composition for optimum compressive strengths of OPC-GGBFS-PSA cements: strengths after (a) 3, (b) 7, (c) 28, and (d) 56 days.
When only gypsum was used as activator, there was much
varation of compressive strength of GGBFS blended cement
with different mixtures of OPC, GGBFS and gypsum.
Furthermore, good early compressive strengths of GGBFS
blended cement were not achieved, with all the tested blends
being initially weaker than OPC. However, GGBFS blended
cement had improved long term strength when less than 5%
gypsum was used.
When only 4–7% PSA clinker was used, the compressive
strengths of GGBFS blended cement after 3 and 7 days were
higher than those of OPC. However, its increase from 7.5
to 15% reduced significantly the compressive strength of
GGBFS blended cement. However, after 28 days, cement
blended with 30–60% GGBFS with activator content 7:5% activator; 7 days with 40–55%
GGBFS and 7.5–15% activator; after 28 days with 40–55%
GGBFS and 3–10% activator.
3.2 Hydration property
The hydration characteristics of PSA clinker as activator
were tested by mixing with calcium hydroxide (Ca(OH) 2 )
and gypsum (CaSO 4 2H 2 O) at various molar ratios ranging
from 1:13:1to 1:13:5, and reaction with water for 28
days. The XRD results of the hydration are shown in Fig. 7.
With increasing gypsum content, ettringite peaks increased.
At the mole ratio of 1:13:5, only ettringite was
formed, indicating that PSA had undergone hydration by the
following reaction:
2KAl(SO 4 ) 2 þ 2Al 2 O 3 þ 13Ca(OH) 2
þ 5CaSO 4 2H 2 O þ 73H 2 O
! 3(3CaOAl 2 O 3 3CaSO 4 32H 2 O) þ 2KOH ð3Þ
Ettringite has been detested as ‘a cement-bacillus’ which
causes expansive fracture of concrete. In 1936, Lossier 17)
began a study to produce chemically prestressed concrete and
thereafter Lafuma 18) and Klein 19) took over the study to build
up a foundation of the expansive cement.
ACI standards (proposal) in the USA 20) include three types
of expansive cement as follows:
(1) K-type: Portland cement mixed with anhydrous hauyne
(3CaO3Al 2 O 3 CaSO 4 ), gypsum (CaSO 4 ) and quick
lime (CaO).
(2) M-type: Portland cement mixed with alumina cement
and gypsum (CaSO 4 ) at a reasonable ratio.
(3) S-type: normal Portland cement mixed with larger
amount of tricalcium aluminate (C 3 A) and gypsum
(CaSO 4 2H 2 O)
In general, K-type calcium sulfoaluminate (3CaO3Al 2 O 3
CaSO 4 ), reacts with water to form calcium monosulfatoaluminate
hydrate, 3CaOAl 2 O 3 CaSO 4 12H 2 O (AFm), and
hydrates of Al 2 O 3 , without forming ettringite. 21) However,
when at least 2 mol of CaSO 4 is mixed with 1 mol of

Activation of GGBFS Cement by Calcined Alunite 215
(a) 3-day strength, MPa
(b) 7-day strength, MPa
(c) 28-day strength, MPa
(d) 56-day strength, MPa
Fig. 6
Composition for optimum compressive strengths of OPC-GGBFS-PSA cements: strengths after (a) 3, (b) 7, (c) 28, and (d) 56 days.
Intensity, arb.unit
E
Fig. 7
E
10
E
E
E E
E E
E E
E Q E
20
Q
Q
Q
Q
30
E: Ettringite
Q: Quartz
Calcined alunite:Ca(OH) 2
:Gypsum
E
E
(molar ratio)
E
1 : 13 : 5
40
Diffraction angle, 2θ /degree
1 : 13 : 4
1 : 13 : 3
1 : 13 : 2
1 : 13 : 1
3CaO3Al 2 O 3 CaSO 4 , ettringite is produced by the following
reaction: 22)
3CaO3Al 2 O 3 CaSO 4 þ 2CaSO 4 2H 2 O þ 36H 2 O
! 3CaOAl 2 O 3 3CaSO 4 32H 2 O þ 2Al 2 O 3 ð4Þ
Also, total conversion of 3CaO3Al 2 O 3 CaSO 4 to ettringite
requires additional CaO and CaSO 4 for the following
reaction: 23)
3CaO3Al 2 O 3 CaSO 4 þ 6Ca(OH) 2 þ 8CaSO 4 þ 74H 2 O
! 3(3CaOAl 2 O 3 3CaSO 4 32H 2 O) ð5Þ
50
Hydration products of the PSA clinker by curing time.
60
70
80
The microstructure of ettringite is strongly dependent on
the presence of lime. 24) Ettringite formed in absence of lime
by the reaction in eq. (4) has been reported nonexpansive and
can develop high early strengths of cement. 25) However,
when formed in the presence of lime by the reaction in
eq. (5), ettringite is expansive which can be exploited in
special applications such as shrinkage-resistant and self
stressing cement. 22)
In order to examine the activation effects of gypsum and
PSA, mixtures of GGBFS and with either gypsum or PSA
were prepared and subjected to the hydration reaction for 28
days. The XRD results from the hydrates are shown in Figs. 8
and 9.
Figure 8 shows that as gypsum reacts with GGBFS to form
ettringite, it strongly influences the activation of GGBFS.
Due to this reason, gypsum is generally used as an activator
for GGBFS blended cement. However, Tables 5 and 6 show
that its use reduced the early strength of GGBFS blended
cement below that of OPC.
Figure 9 shows that as the amount of PSA clinker
increases from 5 to 20%, the gypsum diffraction peaks
increased, and also ettringite was formed. This indicates that
the PSA clinker reacted with GGBFS to form gypsum, which
reacted with GGBFS to form ettringite.
Figure 10 displays the XRD results of the formed hydrates
when GGBFS, calcium hydroxide and PSA clinker were
mixed at a mass ratio of 10:4:2and reacted with water for
28 days.

216 H.-S. Kim, J.-W. Park, Y.-J. An, J.-S. Bae and C. Han
Intensity, arb. unit
G
E
E
10
G
E E E
E E E
20
G
G
EG
E
30
E E
40
Diffraction angle, 2θ /degree
E: Ettringite
G: Gypsum
Slag : Gypsum(wt. ratio)
70 : 30
80 : 20
85 : 15
90 : 10
95 : 5
Fig. 8 XRD patterns of hydrates formed at various mixing ratios of
GGBFS and gypsum.
Intensity, arb. unit
E
10
G
E
G
20
G
Q
G
G
After 3 days, diffraction peaks of ettringite, C 4 AH 13 and
unreacted calcium hydroxide were observable. As the
hydration progressed, the diffraction peaks of calcium
hydroxide reduced. After 28 days, calcium hydroxide was
undetectable but hydrates such as ettringite, C 4 AH 13 , and
C-S-H were observed.
50
60
70
G:Gypsum
Q:Quartz
E:Ettringite
Slag : Calcined alunite
(wt. ratio)
30 40 50 60 70
Diffraction angle, 2θ /degree
95 : 5
80
80 : 20
85 : 15
90 : 10
Fig. 9 XRD patterns of hydrates formed at various mixing ratios of
GGBFS and calcined alunite.
Intensity, arb. unit
A
S
E
A
E
A
10
E E
L
L
Q
Q
S
C
E
Q
A
S
E
L
A
A L
E
20 30 40 50 60
Diffraction angle, 2θ /degree
S
A
A
80
E: Ettingite
L: Ca(OH) 2
C: Calcite
A: C 4 AH 13
Q: Quartz
S: CSH
28 days
70
7 days
3 days
Fig. 10 XRD patterns of hydrates formed in a system comprising GGBFS,
Ca(OH) 2 , and calcined alunite.
80
The exact reaction mechanism, which explains the setting
and hardening of alkali-activated binders, is not yet quite
understood, although it is thought to be dependent on the
prime material as well as on the alkaline activator. 13)
The hydration products of alkali-activated slag cements
have been investigated and reported. It is generally agreed
that its main hydration product is C-S-H. There is no doubt
that the minor hydration products of alkali-activated slag
cement will change with the nature of the slag and
activator. 14) The chemical composition of the slag varies
with the type of iron being made and the type of ore being
used. There is no doubt that the chemical composition of
GGBFS has a significant effect on the hydration process,
hydration product and properties of hardened alkali-activated
slag cements. In many cases, the MgO content of GGBFS is
low and the slag can be described by the CaO-SiO 2 -Al 2 O 3
system. The phase diagram of the CaO-SiO 2 -Al 2 O 3 -H 2 O
system, indicates that five different products, such as C-S-H,
Ca(OH) 2 ,C 4 AH 13 ,C 2 ASH 8 and CS 2 H could appear in this
system, while calcium hydroxide and gehlenite hydrate can
not co-exist at equilibrium. Also, ettringite is one of the
main components of expansive, shrinkage-resistant, rapid
hardening, high early strength and low energy cements. 26–28)
Therefore, PSA clinker was shown to have a significant effect
on the activation of GGBFS.
As we can see in Fig. 7, for the alumina and sulfate
components of PSA clinker to form ettringite and gypsum,
additional calcium hydroxide and gypsum are required.
Therefore, Figs. 11 and 12 shows XRD and DTA data of
hydrates formed at different curing times when 8% gypsum,
PSA clinker or a mixture of both were blended with OPC and
GGBFS, which had a mass ratio of 35 : 65.
When gypsum was used as activator (Fig. 11(a)) on day 1,
initially the diffraction peaks of ettringite, Ca(OH) 2 and
gypsum were mainly observed. After 3 days, the diffraction
peaks of ettringite and Ca(OH) 2 but not gypsum were
observed. Since the crystallinities of the C-S-H hydrates
were very low, their diffraction peaks were difficult to
observe. The DTA results suggest that C-S-H was formed
as the existence of an endothermic peak by the dehydration
of C-S-H appearing at around 100 C, although this peak
of C-S-H overlapped with that of ettringite (approx. 90–
110 C).
The endothermic peak of gypsum dehydration was initially
observed at around 135 C. However, after 3 days, it had
disappeared, indicating that all the gypsum had reacted with
the GGBFS blended cement to form ettringite. Therefore,
when gypsum was used as activator, the main products of
the hydration reaction were C-S-H, ettringite and Ca(OH) 2 ,
which likely develop the compressive strength of GGBFS
blended cement.
When PSA clinker was used as activator, as shown in
Fig. 11(b), ettringite was immediately formed as hydrates.
However, diffraction peaks of gypsum and Ca(OH) 2 , that
could have been formed through the hydration of PSA and
OPC, were not observed.
As shown in Fig. 12(b), endothermic peak due to the
dehydration of C-S-H hydrates and ettringite were observed
at around 100 C but those of Ca(OH) 2 and gypsum was not
observed.

Activation of GGBFS Cement by Calcined Alunite 217
E: Ettringite, A: C 3 S
B: C 2S, C:CSH, G: Gypsum
28days
P: Ca(OH) 2
28days
Intensity, arb. unit
E
E
G
E PE
E
G G
E PE E
A
C
A B
PE
A A,B
A
B
A
B
P
P
14days
7days
3days
1day
DTA(uV)
14days
7days
3days
10
20
30
40
50
60
1day
Diffraction angle, 2θ /degree
(a) Gypsum
373
473
573
673
773
873
E: Ettringite
A: C 3 S,B:C 2 S, C:CSH
Temperature, T/K
(a) Gypsum
Intensity, arb. unit
E
E E
E
P: C 4 AH 18
B
E
E E
E B
P 28days
B
14days
B
7days
A
A 3days
E
E
E
EE
E A A,B A,B A 1day
10 20 30 40 50 60
Diffraction angle, 2θ /degree
2θ
DTA(uV)
28days
14days
7days
3days
DTA(uV)
28days
14days
7days
3days
(b) PSA clinker
1day
1day
Intensity, arb. unit
E
C M
E
M
E E
PE
E P E
A E
E: Ettringite, M: Monosulfate
A: C 3 S,B:C 2 S, C:CSH, D:C 4 AH 13
G: Gypsum, P: Ca(OH) 2
D B
B
PE
E
P 28days
B
E
14days
B B
E
7days
A,B
A,B
A E
Diffraction angle, 2θ /degree
When the mixture of PSA clinker and gypsum was used as
activator, as shown in Fig. 11(c), ettringite was stably
produced in the early stages of hydration. Its diffraction
peaks became smaller and those of monosulfate increased
after 14 days. The presence of ettringite, monosulfate,
C 4 AH 13 , and Ca(OH) 2 were confirmed after 28 days.
Figure 12(c) shows endothermic peaks after 14 days at
around 180 C caused by the dehydration of the monosulfate.
The very small endothermic peak due to the dehydration of
Ca(OH) 2 formed by the hydration of OPC emerged at a
temperature of around 460 C. The formation of monosulfate
showed that a small amount of ettringite transformed to
monosulfate due to the lack of gypsum but did not strongly
influence the compressive strength of GGBFS blended
cement as shown in Table 8 and Fig. 6.
A,B
(c) PSA clinker + gypsum
P
3days
1day
10 20 30 40 50 60
Fig. 11 XRD patterns of hydrates formed in OPC-GGBFS-activator
systems.
373
473
573
Therefore when the mixed PSA clinker and gypsum
activator was used, KAl(SO 4 ) 2 in PSA clinker reacted with
Ca(OH) 2 , a hydrate of OPC, to form CaSO 4 2H 2 O. The
newly formed gypsum reacted with C 3 A in OPC and the
Al 2 O 3 component of PSA clinker or GGBFS to form the
hydration products of alkali-activated slag cements such
as ettringite, C-S-H, Ca(OH) 2 and C 4 AH 13 . This series of
hydrations is thought to improve stably both the early and
long-term strengths of GGBFS-OPC blended cement.
4. Conclusions
673
Temperature, T/K
(b) PSA clinker
773
873
The hydration properties and the compressive strength of
OPC and GGBFS mixed cement were investigated by using
calcined alunite, consisting of KAl(SO 4 ) 2 and amorphous
Al 2 O 3 , as activator. To transform PSA clinker into ettringite,
additional gypsum and calcium hydroxide is required. PSA
clinker appears to have undergone hydration by the following
reaction: 2KAl(SO 4 ) 2 þ 2A1 2 O 3 þ 13Ca(OH) 2 þ 5CaSO 4
2H 2 O þ 73H 2 O ! 3(3CaOA1 2 O 3 3CaSO 4 32H 2 O) þ 2KOH.
When >15 w% PSA clinker was mixed into blended cement
consisting of OPC and GGBFS, the compressive strength of
the blended GGBFS cement decreased due to expansion by
the formation of an excessive amount of ettringite in the
373
473
573
673
Temperature, T/K
773
(c) PSA clinker + gypsum
Fig. 12 DTA analysis of hydrates formed in OPC-GGBFS-activator
systems.
873

218 H.-S. Kim, J.-W. Park, Y.-J. An, J.-S. Bae and C. Han
initial stage of the hydration. However, when a mixed
activator of equal masses of PSA clinker and gypsum was
used, the compressive strengths of the GGBFS blended
cement tested after 3 and 7 days were raised above that of
OPC in proportion to the amount of the activator and GGBFS
used and the hydration products of alkali-activated slag
cements such as ettringite, C-S-H, Ca(OH) 2 and C 4 AH 13
were formed. Therefore, PSA clinker containing KAl(SO 4 ) 2
and amorphous Al 2 O 3 can be used as an activator to improve
the early and long-term strengths of blended cement
consisting of OPC and GGBFS.
Acknowledgments
This research was conducted by Korea Evaluation Institute
of Industrial Technology, Rep. of Korea.
REFERENCES
1) ACI Committee 226: ACI manual of Concrete Practice, (1989) pp. 1–
16.
2) H. Uchigawa: 8th Int. Congr. on the Chemistry of Cement 3 (1986)
249–280.
3) P. K. Mehta: 3rd Int. Conf. on Fly Ash, Silica Fume, and Natural
Pozzolans in Concrete (1989) pp. 1–43.
4) T. Hakkinen: Cem. Concr. Res. 23 (1993) 407–421.
5) T. Hakkinen: Cem. Concr. Res. 23 (1993) 518–530.
6) M. Regourd, J. H. Thomassin, P. Ballif and J. C. Touray: Cem. Concr.
Res. 13 (1983) 549–556.
7) J. Deja: Cem. Concr. Composites 25 (2003) 357–361.
8) G. J. Osborne: Cem. Concr. Composites 21 (1999) 11–21.
9) K. M. Lee and P. J. Park: Conservation Recycl. 44 (2005) 139–151.
10) ACI Committee 233: ACI Mater. J. 92 (1995) 321–322.
11) A. Fernandez-Jimenez, J. G. Palomo and F. Puertas: Cem. Concr. Res.
29 (1999) 1313–1321.
12) A. Palomo, M. W. Grutzeck and M. T. Blanco: Cem. Concr. Res. 29
(1999) 1323–1329.
13) P. T. Fernando, C. G. Joao and J. Said: Cons. Build Mater. 22 (2008)
1305–1314.
14) S. Caijun, K. P. V and R. Della: Alkali-activated cements concretes.
1st ed. USA, (Taylor and Francis, 2006).
15) H. S. Kim, J. W. Ahn, J. Y. Hyun and D. S. Cho: Mater. Sci. Forum 439
(2003) 106–114.
16) H. S. Kim, G. C. Han, J. W. Ahn, K. H. Cho and H. C. Cho: Sensors 9
(2009) 5059–5067.
17) H. Lossier and A. Caguol: Genie Civil (Paris) 121 (1944) 61–65.
18) H. Lafuma: Expansive cements, Proc. 3rd Int. Symp. on the Chemistry
of Cement, Cement and Concrete Association, London, (1954) pp. 581–
596.
19) A. Klein and G. E. Troxell: ASTM 58 (1958) 986–1008.
20) AC1 Committee 223: ACI J. (1970) 583–610.
21) F. P. Glasser and L. Zhang: J. Chin. Ceram. Soc. 25 (2000) 340–347.
22) I. Odler, A. Bentur and S. Mindess (Eds.): Special inorganic cements,
(E & FN Spon, London & New York, 2000) pp. 69–87.
23) M. Su, W. Kurdowski and F. Sorrentino: Proc. 9th Int. Congr. on the
Chemistry of Cement 1 (1992) pp. 317–354.
24) P. K. Mehta: Cem. Concr. Res. 3 (1973) 1–6.
25) J. Beretka, N. Sherman, M. Marroccoli, A. Pompo and G. L. Valenti:
Proc. 10th Int. Congr. Chem. Cement 2 (1997) p. 8.
26) W. Kurdowski, C. M. George and F. P. Sorrentino: Proc. 8th Int. Congr.
Chem. Cement 1 (1986) pp. 292–318.
27) S. Muzhen, W. Kurdowski and F. Sorrentino: Proc. 9th Int. Congr.
Chem. Cement 1 (1992) pp. 317–354.
28) J. A. Deng, W. M. Ge, M. Z. Su and X. Y. Li: Proc. 7th Int. Congr.
Chem. Cement 1 (1980) pp. 381–386.