ALKALINE ACTIVATION OF BLASTFURNACE SLAG FOR

THE PRODUCTION OF SPECIAL BINDERS

D. P. Dias, J. M. Désir, M. C. R.Pessanha

Laboratório de Engenharia Civil (LECIV), Centro de Ciência e Tecnologia (CCT),

Universidade Estadual do Norte Fluminense Darcy Ribeiro (UENF),
28013-600, Campos dos Goytacazes-RJ, Brasil

ABSTRACT

Since the decade of 40, the alkaline activation of materials made from Portland cement,
blastfurnace slag and pozzolans, has been accomplished throughout the world. The first
studies of alkaline activation of blastfurnace slag had begun by the discovery that the
reactivity increased considerably when the slag presented high amount of sodium and
potassium. Since then, quite a few studies had been carried out until the decade of 50, with
exception of those carried out in the former Soviet-Union and the Scandinavian countries, as
well. The development of alkali-activated cements based on blastfurnace slag generated
several alternative products in those countries, like Trief-cements and F-cements. These
cements served as support for the development of another binder system alkali-activated in
the USA, the Pyrament cement (a derivation of pozzolanic cement). Therefore, the main
objective of this study was to create and to develop a knowledge net, in Brazil, regarding the
achievement of special binders made from the alkaline activation of blastfurnace slag. An
experimental program was developed aiming to characterize certain physical, chemical and
mechanical properties of this material. The results showed that the potentiality of the use of
the alkalis opens new horizons for obtaining of special binders, with properties different from
those presented by the Portland slag cement.

KEYWORDS

Blastfurnace slag, alkaline activation, binders, cementitious materials.

INTRODUCTION

It is well know that blastfurnace slag has become a widely used and important ingredient as a
supplementary binder in ordinary Portland cement (OPC), because the major components of it
(SiO2, CaO, MgO and Al2O3) are common in commercial silicate glasses. In spite of that,
blastfurnace slag shows little or no cementing properties. However, alkali-slag cements can
develop very high strength in the presence of a proper alkaline activator (Krizan & Zivanovic,
2002).

Due to significant interest in this new type of binder, research on alkali-activated slag (AAS)
has been conducted in many countries including China, Australia, Russia, United States,
England, and Canada. The benefits of the use of AAS as a construction material include
utilisation of industrial by-products, low energy consumption associated with low CO2
emissions (Bakharev et al., 2001).

According to Palomo & López de la Fuente (2003), the alkaline activation of materials is a
chemical process that provides the transformation of determined structures, partial or totally
amorphous and/or metastables in compact cemented frameworks; and the alkaline activation
process are composed by three stages (Krivenco, 1994): 1) dissolution of soluble species in
the alkaline phase; 2) reorganisation and diffusion of dissolved ions with the formation of
small coagulated structures; and 3) growth and condensation of these structures to form the
hydrated products.

As can be seen, the alkaline activation depends on the chemical and phase compositions of
slag as well as the type and concentration of activators. Based on these assumptions, the focus
of this preliminary investigation is on the study of the compressive strength, at 3, 7 and
28 days after casting, of Brazilian AAS.

MATERIALS AND METHODS

Blastfurnace Slag (BFS)

The chemical composition and physical properties of blastfurnace slag (BFS), supplied by
Companhia Siderúrgica de Tubarão (CST) and grinding by Cimento Mizu, with a particle size
range of 1-80µm (Figure 1), are detailed in Table 1. Aiming to verify the fineness effect on
the compressive strength, the BFS was submitted to an additional grinding in balls mill during
half hour (BFSg).

Portland Slag Cement (PSC)

The Portland slag cement (PSC) Mauá CPIIE-32 type was used as a binder reference. The
chemical composition and physical properties of PSC are also detailed in Table 1.

Alkaline Activator

A blended alkali-containing activator composed by 70wt.% Na2SiO3⋅9H2O (liquid sodium

silicate – water glass) plus 30wt.% KOH (potassium hydroxide) was used for activation. The
dosage of alkali used was about 10wt.% of binder.
 FIGURE 1. SEM IMAGE OF BFS (X500)

Fine Aggregate

The fine aggregate was drying at 110oC during 24 hours and consisted of washed river sand
with 2.4 mm maximum size and specific gravity of 2.62 kg/m3.

TABLE 1. CHEMICAL COMPOSITION AND PHYSICAL PROPERTIES OF BFS AND PSC

Item BFS1 PSC2

SiO2 (%) 34.0 20.1
Al2O3 (%) 11.5 5.0
Fe2O3 (%) 0.3 3.7
CaO (%) 42.0 59.5
TiO2 (%) 0.5 -
SO3 (%) 1.2 3.0
K2O (%) - 0.1
MgO (%) 8.5 2.3
MnO (%) 0.5 -
Loss on ignition (%) - 4.8
Specific gravity (kg/m3)3 2.94 3.15
Fineness (m2/kg) 1673.04 / 1990.05 362.3
1
Data supplied by CST.
2
Data supplied by Cimento Lafarge.
3
Values obtained by sedimentation test.
4
BFS without additional grinding.
5
BFS with additional grinding in balls mill during half hour, marked as BFSg.

Mortar Mixes

Five types of mixes were used for preparation of mortar specimens (Table 2). The water to
binder (w/b) ratio of 0.38 was used. According to Wang et al. (1994) the optimum range w/b
ratio is within 0.38-0.45. A fixed sand-to-cementitious binder ratio 3:1 was used to avoid the
influence of factors related to the mortar mix design.
 TABLE 2. TYPE OF MORTAR MIXES

Specimen Composition
S1* 100% PSC
S2 100% BFS
S3 100% BFSg
S4 70% BFS + 30% PSC
S5 70% BFSg + 30% PSC
* Reference mortar – not alkali-activated.

Testing Procedures

The size of the mortar cylinders for compressive strength testing was 50 mm in diameter and
100 mm in length. The cylindrical specimens were demoulded 24 hours after casting and were
stored in drying room where temperature and relative humidity were maintained at
21oC ± 2oC and 60 ± 5%, respectively.

Compressive strength measurements of mortars were performed on a Versa Tester machine in

a load control regime with a loading rate of 0.5 MPa/s. Four cylinders were tested for each
data point.

EXPERIMENTAL RESULTS AND DISCUSSION

Compressive Strength

The Figure 2 shows compressive strength development of mortars with time. Each
compressive strength result is an average of four specimens, and the highest standard
deviation was about 2.3%. The S3 mortar presented rapid early compressive strength
development and the highest 28-day compressive strength (32 MPa). These preliminary
results showed that 70wt.% Na2SiO3⋅9H2O + 30wt.% KOH alkaline activator was most
effective to this type of mortar. The effects of alkaline activator on compressive strength
development of the S2, S4 and S5 mortars were not significant, especially as compared with
S1 mortar.

Except the 3-day compressive strength of S2 mortar, S2 and S4 mortars exhibited lower
compressive strengths than S1 (Figure 2). Nevertheless, partial or totally replacement of PSC
by BFS and AAS in making pastes, mortars and concretes are satisfactory since they are often
readily available in regions where the use of more conventional materials such as PSC may
not be economically viable, and environmentally friendly materials, consuming less energy
and generating less waste in cement production.

Compressive strength about 55.0 MPa was achieved in previous works (Silva, 2000;
Dias, 2001) with same alkaline activator. However, the mortars were manufactured with
metakaolin, subjected to bath curing in lime-saturated water and lower dosage of alkali was
used. Thus, it can be related the minor compressive strength achieved in this study (32 MPa),
as compared with previous works, mainly to:
 • Curing method: alkali-activated specimens that were air-cured presents higher drying
shrinkage than that of corresponding PSC specimens, leading to a higher incidence of
cracking and, consequently, lower mechanical strength (Talling, 1989; Krizan &
Zivanovic, 2002). Therefore, it was concluded that is not desirable to use air-curing
method for AAS systems.

• Alkali Dosage: several reports on this aspect have been made (Isozaki et al., 1986;
Andersson et al., 1988; Talling, 1989; Bakharev et al., 1999). The high alkali dosage
(10%wt. by binder weight) used in this work caused efflorescence because of the
effect of more free alkali in the specimens. Then, trying to increase the mechanical
strength by elevation of alkali dosage is not recommended in this case; and according
to Bakharev et al. (1999), mortars with high alkali dosage can also exhibit high
shrinkage rate, which results in microcracking, and it may lead to a lower mechanical
strength.

It can also be seen in the Figure 2 that the S3 and S5 mortars presented higher compressive
strength than S2 and S4 mortars, respectively. This behaviour can be attributed to:

• Slag fineness: increasing the fineness of slag improves the mechanical strength,
regardless of activator type (Andersson et al., 1988). As can be seen in Table 1, the
slag used to manufactured S3 and S5 mortars (BFSg) presented approximately 19%
higher fineness than S2 and S4 mortars (BFS). Moreover, there is additional evidence
that a more uniform microstructure would be expected for the more finely ground raw
material and that this will lead to higher mechanical strengths (Hewlett, 1998).

• Blended mix: alkali-activated slag-cement blend usually presented low mechanical

strength due to the effect of alkaline activator on PSC hydration, which results in
formation of a significant amount of calcium hydroxide (CH) and Na-substituted Ca in
calcium silicate hydrates (N-C-S-H) at early stages of hydration, which was reported
to be less dense and has lower mechanical strength than hydration product of cement
and water (Way & Shayan, 1989).
 34
32
Compressive Strength (MPa) 30
28 S1
S2
26
S3
24
S4
22 S5
20
18
16
0 5 10 15 20 25 30
Time (days)

FIGURE 2. COMPRESSIVE STRENGTHS OF MORTAR SPECIMENS, AT 3, 7 AND 28 DAYS

Alkali Modulus (AM)

The alkali modulus (AM) can be defined as wt.% of SiO2 to wt.% of alkalis ratio. The AM
value in solution of bi-compound alkaline activator used in this study was equal to 0.85, and
was calculated as follows:

%SiO 2
AM = (1)
% Na 2 O + % K 2 O

Several investigations (Glukhovsky et al., 1980, Andersson et al., 1988, Pu et al., 1988,
Talling, 1989, Wang et al., 1994) using water glass as alkaline activator revealed that AM
values of 1.0 to 1.5 resulted on the best mechanical strengths for basic slags (see item
Chemical Moduli and Hydration Modulus), regardless of curing conditions, type or fineness
of binder. However, it was found that these parameters had great influence on compressive
strengths (Figure 2) for AM = 0.85.

The AM value also has a significant relationship with the pH value of the alkaline activator.
Decreasing the AM value means increase in alkaline activator pH value (Chang, 2003).
Therefore, after activator preparation, the pH value of the alkaline activator in solution was
then measured when the temperature of the activator lowered to 25oC. The pH value equal to
14.4 ± 0.1 is in accordance to Fraay & Bejen (1989), who consider that the pH value of the
alkaline activator in solution must be higher than 13.3 to ensure sound activation and,
consequently, high mechanical strength. In despite of that, not too high compressive strengths
were achieved (Figure 2) because, as mentioned before, the air-curing method and the high
alkali dosage adopted. It can then be concluded from this evidence that there is strong
interaction (interdependence) of factors affecting the mechanical strength in AAS systems. By
correct combination of these conditions, compressive strength of 60 to 150 MPa without heat
treatment or special additives has been made in China (Shi, 1987, Pu et al., 1988).

Chemical Moduli and Hydration Modulus

Several reports have been published (Wang, 1992, Wang et al., 1994, Chang, 2003) trying to
find a method to estimate a relationship between alkaline activity of slag and mechanical
strength. According to Wang (1992), the hydraulic activity of slag can be measured by
chemical moduli, which is composed by two parameters:

• Basicity coefficient (Kb)

%CaO + % MgO
Kb = (2)
%SiO 2 + % Al 2 O3

• Quality coefficient (Kq)

%CaO + % MgO + % Al 2 O 3
Kq = (3)
%SiO 2 + %TiO 2 + (% MnO + % P2 O5 + % F)

And, Chang (2003) suggest that the hydration capacity of slag can be measured by the
hydration modulus (HM), where HM is defined as:

% CaO + % MgO + % Al 2 O 3
HM = (4)
%SiO 2

The resume of chemical characteristics of the binders used to manufacture the alkali-activated
mortars is shown in Table 3. The kb values indicates that both binders are basics, which
means, related to the type of alkaline activator employed, higher readiness to alkaline
activation than acid or neutral binders. The quality of the binders to alkaline activation can
also be estimated by kq values. The higher Kq values the higher hydration capacity of the
binders. Finally, the HM values obtained in this study were higher than 1.4 suggesting,
according to Chang (2003), favourable conditions to alkaline activation. Based on these
assumptions, would be expected higher compressive strengths to S4 and S5 mortars than S2
and S3. It can then be concluded that the correlation between these chemical parameters (kb,
kq and HM) and compressive strength (Figure 1) is inconsistent, because do not take into
consideration important binder characteristics such as fineness of binder and multi-compound
binder matrixes.

TABLE 3. RESUME OF CHEMICAL CHARACTERISTICS OF ALKALI-ACTIVATED MORTARS

Specimen kb kq HM
S2 and S3 1.11 1.77 1.82
S4 and S5 1.37 2.08 2.13
CONCLUSIONS

The results of this preliminary investigation indicate that:

The alkali-activated mortar manufactured with blastfurnace slag with additional grinding
(BFSg) presented rapid early compressive strength development and the highest compressive
strengths, at 3, 7 and 28 days, especially the mortar made with 100% BFSg.

The effects of alkaline activation on compressive strength development of blended-binder

mortars (70% BFS + 30% PSC) were not significant, especially as compared as the mortar
made with 100% PSC (reference mortar – not alkali-activated).

The curing conditions of specimens, the alkali dosage of activator, and the type and fineness
of binder can limit the mechanical strength development of AAS systems.

The correlation between kb, kq and HM parameters and compressive strength of AAS systems
is inconsistent, because do not take into consideration important characteristics of binder.

ACKNOWLEDGMENTS

The authors wish to thank to the financial support provided by Universidade Estadual do
Norte Fluminense Darcy Ribeiro (UENF) and the efforts and assistance with the laboratory
work provided by Milton Soares Jr., Vanuzia Ferreira, and André Manhães
(LECIV/CCT/UENF). Thanks to Companhia Siderúrgica de Tubarão (CST) and Cimento
Mizu for the blastfurnace slag supply and Cimento Lafarge for the chemical Portland slag
cement data supply, and finally to Laboratório de Ciências Químicas (LCQUI/CCT/UENF)
for the BET analysis and Laboratório de Biologia Celular e Tecidual (LBCT/CBB/UENF) for
the SEM image.

REFERENCES

Andersson, R.; Gram, H.-E.; Malolepszy, J.; Deja, J. (1988). Alkali-Activated Slag. Swedish
Cement and Concrete Research Institute, Stockholm, p.104.

Bakharev, T.; Sanjayan, J. G.; Cheng, Y.-B. (1999). Alkali Activation of Australian Slag
Cements. Cement and Concrete Research 29, p.113-120. ISSN: 0008-8846.

Bakharev, T.; Sanjayan, J. G.; Cheng, Y.-B. (2001). Resistance of Alkali-Activated Slag
Concrete to Carbonation. Cement and Concrete Research 31, p.1277-1283. ISSN: 0008-8846.

Chang, J. J. (2003). A Study on the Setting Characteristics of Sodium Silicate-Activated Slag

Pastes. Cement and Concrete Research 33, p.1005-1011. ISSN: 0008-8846.

Dias, D. P. (2001). Cimentos Geopoliméricos: Estudo de Agentes Químicos Agressivos,

Aderência e Tenacidade à Fratura. Doct. Degree Thesis, Military Engineering Institute, p.216.

Fraay, A. L. A.; Bejen, J. M. (1989). The Reaction of Fly Ash in Concrete: a Critical
Examination. Cement and Concrete Research 19, p.235-246. ISSN: 0008-8846.
Glukhovsky, V. D., Rostovskaja, G. S.; Rumyna, G. V. (1980). High Strength Slag-Alkaline
Cements. 7th International Congress on Chemistry of Cement, Paris, vol. 3, p.164-168.

Hewlett, P. C. (1998). Lea’s Chemistry of Cement and Concrete, Arnold Publishers, London-
UK. ISBN: 0 340 56589 6.

Isozaki, K.; Iwamoto, S.; Nakagawa, K. (1986). Some Properties of Alkali-Activated Slag
Cement. 8th International Congress on Chemistry of Cement, Rio de Janeiro/Brazil, vol. 6,
p.395-399.

Krivenco, P. V. (1994). Alkaline Cements. Proceedings of the 1st International Conference,

VIPOL Stock Co., p.11-130.

Krizan, D.; Zivanovic, B. (2002). Effects of Dosage and Modulus of Water Glass on Early
Hydration of Alkali-Slag Cements. Cement and Concrete Research 32, p.1181-1188. ISSN:
0008-8846.

Palomo, A.; López de la Fuente, J. I. (2003). Alkali-Activated Cementitious Materials:

Alternative Matrices for the Immobilisation of Hazardous of Boron. Cement and Concrete
Research 33, p.281-288. ISSN: 0008-8846.

Pu, X. C.; Gan, C. C.; Wang, S. D.; Yang, C. H. (1988). Summary Reports of Research on
Alkali-Activated Slag Cement and Concrete. Chongqing Institute of Architecture and
Engineering, vol. 1-6.

Ramachandran, V. S. (1995). Concrete Admixtures Handbook: Properties, Science, and

Technology. Noyes Publications. ASIN: 0815509812.

Shi, C. (1987). An Investigation on the Activation of Granulated Phosphorous Slag. M. Sc.

Degree Thesis, Nanging Institute of Technology.

Silva, F. J. (2000). Reforço e Fratura em Compósitos de Matriz Álcali-Ativada. Doct. Degree

Thesis, Military Engineering Institute, p.271.

Talling, B. (1989). Effect of Curing Conditions on Alkali-Activated Slags. 3rd International

Conference on Fly-Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, Norway,
vol. 2, p.1485-1500.

Wang, S.-D. (1992). A New Method to Estimate the Hydraulic Activity of Slag: the
Reactivity Method. Journal of the Chinese Ceramic Society 20(2), p.196-200.

Wang, S.-D.; Scrivener, K. L.; Pratt, P. L. (1994). Factors Affecting the Strength of Alkali-
Activated Slag. Cement and Concrete Research 24, p.1033-1043. ISSN: 0008-8846.

Way, S. J.; Shayan, A. (1989). Early Hydration of a Portland Cement in Water and Sodium
Hydroxide Solutions: Composition of Solutions and Nature of Solid Phases. Cement and
Concrete Research 19, p.759-769. ISSN: 0008-8846.