Cement and Concrete Research 32 (2002) 1181 – 1188

Effects of dosage and modulus of water glass on early hydration of

alkali–slag cements
Darko Krizana, Branislav Zivanovicb,*
a
Institute for Testing Materials, Bulevar vojvode Misica 43, 11000 Belgrade, FR, Yugoslavia
b
Material Science Department, Centre for Multidisciplinary Studies, University of Belgrade, Kneza Viseslava 1a, 11000 Belgrade, FR, Yugoslavia
Received 30 April 2001; accepted 22 October 2001

Abstract

This paper examines the early hydration of alkali – slag cements activated with water glass with different n moduli and sodium metasilicate
(Na2SiO3
5H2O) in solution at 25 °C. The early hydration of alkali-activated blast furnace slag cements has been studied using isothermal
conduction calorimetry. The cumulative heat of hydration increases by increasing the n modulus as well as the dosage of water glass, but is
still lower than that of Portland cement. The compressive strength of normal-cured water glass slag cements is higher than Portland cement
mortars. Drying shrinkage of alkali – slag cements is considerably higher than that of Portland cement. Consequently, industrial use of alkali –
slag cement needs better understanding of the hardening mechanism and requires further research based on presented observations and
results. D 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Alkali-activated cement; Calorimetry; Hydration; Compressive strength

1. Introduction Notwithstanding the fact that the Portland cement will,

probably, remain the basic determining binding substance in
The goal of sustainable development of the cement and the nearest discernable future, Glukhowsky et al. [2] set a
concrete industry is very important, and can be reached if task to develop new highly active mineral binding materials
we make a serious effort for the complete utilization of the that would substantially enhance the strength of concrete,
cementitious and pozzolanic by-products produced in ther- their frost and corrosion resistance, and other qualities. The
mal power plants and metallurgical industries [1]. essence of idea of creating such binders according to
Finely ground granulated slag has become a widely Glukhowsky et al. [2] consists of the exclusion from the
used and important ingredient as a supplementary binder cement composition of highly basic minerals (C3A, C3S,
in ordinary Portland cement (OPC). Since slag itself is and C4AF), whose hydration products either have no essen-
nothing more than latent hydraulic binder, it must be tial effect on the synthesis of hydrated cement strength
activated in practice to react and provide strength. It is (C3AH6) or enhance its reaction capacity with corrosion
recognized that alkali additions can activate pozzolanic mediums (C3AH6, C3AH13, C2SH2, Ca(OH)2) and in the
materials to set and harden in their own right, and there introduction instead of alkali activators ensures hydration of
exists considerable scope for the development of OPC-free the alumosilicate substance of the cement. Slag is usually
cementitious binders that may have improved mechanical activated chemically but this kind of activation can be
and durability properties over OPC-based binders. Con- improved if it is combined with mechanical (more finely
sidering the energy-intensive manufacturing process of ground) or thermal (increased temperature) activation. A
OPC, such binders could eventually have an important two-component system, alkali– slag cements, consist of a
role to play. cementitious component such as blast furnace slag and an
alkaline activator. In spite of that blast furnace slag may
show little or no cementing properties, alkali– slag cements
* Corresponding author. Tel.: +381-11-766-154; fax: +381-11-555-258. can develop very high strength in the presence of a proper
E-mail address: bane@ibiss.bg.ac.yu (B. Zivanovic). alkaline activator.

0008-8846/02/$ – see front matter D 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 8 - 8 8 4 6 ( 0 1 ) 0 0 7 1 7 - 7
1182 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188

According to Glukhowsky [3], caustic alkalis (ROH), Table 1

Chemical composition (%) and physical properties of GBFS and OPC
nonsilicate salts of weak acid (R2CO3, R2SO3, R2S, RF,
etc.), silicate salts type (R2O(0.5 – 2.5)SiO2), aluminates Item GBFS OPC
(M2OnAl2O3), aluminosilicates (M2OAl2O3(2 – 6)SiO2), SiO2 38.26 18.67
and nonsilicate strong acid salts (M2SO4) can be used as Al2O3 9.32 5.42
Fe2O3 0.60 2.44
alkali components of the slag – alkaline cements. The early
CaO 40.49 60.72
hydration and reaction kinetics of alkali-activated binders MgO 8.70 2.22
have normally been studied using isothermal conduction SO3 0.47 2.86
calorimetry [4,5]. McCarter and Ezirim [6] and McCarter Na2O 0.40 0.52
et al. [7] focus on monitoring the change of electrical K2 O 0.39 0.31
S 0.42 –
properties (conductance and capacitance) in the study of
MnO 0.81 0.05
the early hydration reaction kinetics of alkali-activated slag FeO – 0.24
cements. Zhou et al. [8] concluded that the hydration P2O5 – 0.05
process of alkali – slag cements could be classified into IL 0.42 3.12
five periods based on the heat evolution curves (initial, IR – 3.00
Total 99.86 99.62
induction, acceleration, deceleration, and period of slow,
Density (kg/m3) 2920 3150
continued reaction), which is similar to that of OPC, but Blaine fineness (m2/kg) 428 351
the mechanisms are different in both cases. Shi and Li [9]
and Shi and Day [10 – 12] found that all caustic alkalis and
alkali compounds whose anions or anion groups can react
with Ca2 + to produce Ca compounds that are less soluble metasilicate was 29.26% Na2O and 27.98% SiO2. It was
than Ca(OH)2 can act as activators of slags. They have found that the mass of Na2O, instead of total mass of
also found that due to the variation of activator, the activators, correlated best with the physical properties of
hydration of alkali –slag cements can be described by three alkali – slag cements [3]. Accordingly, all activators were
models: Type I — one initial peak occurs during the first added as 3% and 4% by mass of Na2O for the physicome-
few minutes and no more peaks appear thereafter; Type chanical property examinations and 3%, 4%, and 6% by mass
II — one initial peak appears before the induction period of Na2O for the heat of hydration. Quartz sand (0– 2 mm) was
and one accelerated hydration peak appears after the used as aggregate in the mortar specimens.
induction period; and Type III — two peaks (one initial
and one addition initial) appear before the induction period 2.2. Measurement of heat evolution rates
and one accelerated hydration peak appears after the
induction period. The evolution of heat of hydration was monitored using a
This paper examines the effect of activator dosage on Tonical Isothermal Conduction Calorimeter. The mass of
early hydration of alkali– slag cement activated with a third slag used was 5 g and a water-to-slag ratio of 0.43 was used.
group of alkalis according to Glukhowsky [3], using a OPC was used as the binder reference with water-to-cement
conduction calorimeter. Water glass with different n moduli ratio of 0.5. Measurements of heat of evolution were
and sodium metasilicate was used as alkaline activator. The performed at a constant temperature of 25 °C. When thermal
strength development and shrinkage of alkali-activated slag equilibrium was achieved, the slag and the solution con-
were investigated too. taining the alkaline activator were mixed by injecting the
solution into slag.

2. Experimental 2.3. Mortar mix proportions, handling, and curing

2.1. Materials used Alkali-activated slag mortar was manufactured according

to Yugoslavian standard JUS EN 196-1, with a water-to-
Granulated blast furnace slag (GBFS; SARTID, Smeder- binder ratio of 0.43 while the OPC mortar was manufac-
evo) with chemical composition in accordance with Table 1 tured with a water-to-cement ratio of 0.5. The activator was
has been used in our experiments. The OPC from Beocin added by dissolving in mixing water. The weight ratio
has been used as a binder reference. The slag was activated between binder and aggregate was 1:3. Mortar prisms with
with water glass [Na2O(n)SiO2] and sodium metasilicate a dimension of 40  40  160 mm were manufactured for
(Na2SiO3
5H2O) in solution. The n modulus of liquid water determining other physical properties. After curing, the
glass (the mass ratio between SiO2 and Na2O) was 3.01 with compressive strengths of the specimens were tested at the
9.62% Na2O and 29.0% SiO2. different ages up to 90 days. The mortar specimens of
NaOH was added to change the n modulus, and water alkali– slag cements were kept in a climate room with a
glass with n moduli of 0.6, 0.9, 1.2, and 1.5 was used in temperature of 20 °C (68 °F) and relative humidity (RH)
experiments. The chemical composition of solid sodium of 90% directly after manufacture and up to testing. The
 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188 1183

Fig. 1. Heat evolution rate of Portland cement, OPC-CEM I 42.5.

OPC specimens were kept in the water. The shrinkage was and C3A [13]. The heat evolution curves then decline to an
continuously followed up to 90 days. Immediately after induction period lasting from 1 to 2 h. In this stage, the
stripping (after 1 day), the mortar prisms of alkali – slag hydration of all the clinker minerals progresses very slowly.
cements were kept in climate room and the Portland The concentration of calcium hydroxide in the liquid phase
cement prism in the water. The shrinkage that took place reaches its maximum and starts to decline. This is followed
before stripping has not been studied. The length of prisms by an accelerated hydration peak that results from the
was then determined after 4, 7, 14, 21, 28, and 90 days of hydration of C3S and formation of C-S-H phase and
storage at 20 °C and 55% RH. portlandite. The calcium sulphate becomes completely dis-
solved and the concentration of SO42  in the liquid phase
start to decline due to the formation of the AFt phase during
3. Experimental results and discussion the deceleration period. In the postacceleration period, the
hydration rate slows down gradually and the rate of the
3.1. Effect of activator dosage on heat evolution rate hydration process becomes diffusion-controlled.

3.1.1. Portland cement +water 3.1.2. Slag activated with different-moduli water glass as
The heat evolution curves of Portland cement (Fig. 1) well sodium metasilicate
show that initial peak appears within first few minutes. Fig. 2 illustrates heat evolution rates of alkali – slag
Preinduction period signifies that immediately upon contact cements activated by different amounts of water glass with
of cement with water, a rapid dissolution of ionic species n = 0.6 at 25 °C.
into the liquid phase and the formation of hydrate phases The slag hydration process must be considered with regard
occur. The rather intense liberation of heat with a maximum to the oxygen forms present in slag, double-fixed in the
within a few minutes is due to the rapid hydration of C3S QSiUOUSiQ bond, single-fixed in QSiUOUMeU bonds,

Fig. 2. Effect of water glass with n = 0.6 on heat evolution rate of alkali – slag cement.
1184 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188

and ‘‘free’’ in UMeUOUMeU bond and the conditions of The small peak appears just after the addition of the
hydrate phase formation [14]. The reaction in aqueous activator solution, which is followed by a very diffuse peak
medium proceed according to the following equations: during the preinduction period [10]. When the activator
dosage is 3%, the initial peak and additional initial peak
QSiUOUSiQ þ H
OH ! 2QSiUOH
combine into one and occurs after 6 min. With increasing
water glass dosage, initial peaks appear earlier (5 and 3 min),
QSiUOUMeUOUSiQ þ 2 H
OH ! 2QSiUOH ¼ additional initial peaks are higher and more diffuse than
2+ initial peaks and appear after 30 and 40 min, respectively.
þ Me þ OH 
The first destruction step under high concentration of OH 
can be described as the breaking of MeUO (MeQCa,Mg),
QSiUOUMe: : : O: : : MeUOUSiQ SiUOUSi, AlUOUAl, and AlUOUSi of the slag grain.
2+ Since the MeUO bonds are much weaker than AlUO and
þ 3H
OH ! 2 QSiUOH þ 2 Me þ 4 OH 
SiUO bonds, more Ca and Mg enter into water than Al and
The presence of alkali cation, which at the stage of Si, and the Si/Al-enriched layer forms quickly on the surface
destruction –coagulation structure formation, first acts as of slag [19,20]. The phase structure of the slag provided
destruction catalyst and then, with subsequent development channels for OH  ions to enter the bulk to complete the
of condensation structure alkali content decreases as a hydration process. Besides, the dispersed medium has a high
result of its bonding into new formations, which suggest concentration of [SiO4]4  from water glass, and the first step
that the alkaline component affects essentially the structure may correspond to the hydration and hardening of the water
formation processes according to the following equation: glass or reaction between [SiO4]4  and Ca2 + dissolved from
the surface of slag grains. The initial peak corresponds to the
QSiUOH +NaOH!QSiUONa + H
OH wetting and dissolution of slag grains and adsorption of some
ions onto the surface of slag grains, and the second peak to the
These alkaline silicates may undergo reactions of ion formation of precipitate layer at the slag particle surface [10].
interchange with divalent ions, forming QSiUOU Hydration products have formed in the space between the
CaUOH-type complexes [15]. Aluminia is present in slag grains, originally occupied by alkaline solution, indic-
similar formations, hydroxyl groups affect in the same ating their formation by dissolution and precipitation mech-
manner the AlUSiUO bond, giving rise to soluble anism at early ages [21]. The higher the water glass dosage is,
aluminate anion Al(OH)4  hydroxocomplex and surface the higher are the combined and accelerated hydration peaks.
groups QAlUOH [16]. The accelerated hydration peaks for 3%, 4%, and 6% acti-
Alkaline activation of blast furnace slag is affected due vated slag were higher than the corresponding combined
to the high concentration of OH  ions in the mix [17,18]. preinduction peaks. The precipitation of C-S-H gel results in
Due to the present of alkali (Na + or K + ), even in small a very long induction effect. All curves show an appreciable
amount, the pH value in Portland cement is usually more induction period and 22 h is almost on the top of the
than 13.0. Destruction of the dispersed phase of alkaline accelerated hydration peak for alkali–slag cement activated
cement, in contrast to that of Portland cement, has its own with 3% water glass. The longer induction period, varying
peculiarities caused by the lower basicity of the initial from about 5 to 14 h, is observed before the appearance of the
minerals and the presence of high concentration of OH  accelerated hydration peak. The higher the activator dosage,
ions in the dispersion medium. the shorter is the induction period. When the slag grains were

Fig. 3. Effect of water glass with n = 0.9 on heat evolution rate of alkali – slag cement.
 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188 1185

Fig. 4. Effect of sodium metasilicate on heat evolution rate of alkali – slag cement.

coated by a hydrate layer, the hydration kinetics was con- period, varying from about 6 to 27 h, is observed before the
trolled by a diffusion process [4] until the completion of the appearance of the accelerated hydration peak. The higher the
reaction, the diffusion process being affected by the degree of activator dosage, the shorter is the induction period. In
structural defectiveness [16]. contrast to slag paste activated with water glass with
The magnitude of accelerated hydration peak of the n = 0.6, the accelerated peak with n = 0.9 is lower than the
cement with 6% of water glass is higher than that with additional initial peak during the preinduction period.
3% and 4% but still lower than the accelerated peak of Fig. 4 illustrates heat evolution rates of alkali – slag
Portland cement. cements activated by different amounts of sodium metasili-
Fig. 3 illustrates heat evolution rates of alkali – slag cate Na2SiO3x5H2O at 25 °C. Compared with slag paste
cements activated by different amounts of water glass with activated with n = 0.9, very similar features of heat evolution
n = 0.9 at 25 °C. rates can be observed. Slightly higher effectivity in sodium
The small peak appears just after the addition of the metasilicate-activated paste during preinduction is visible.
activator solution, which is followed by a diffuse peak during All curves continued to have a trend of lower induction
the preinduction period. When activator dosage is 6%, the period peaks compared with additional initial peaks during
initial peak appears first and occurs after 2 or 3 min. With preinduction period.
increasing water glass dosage, the initial peak appears earlier Figs. 5 and 6 illustrate heat evolution rates of alkali –slag
and additional initial peaks were higher and less diffuse. The cements activated by different amounts of water glass with
higher the water glass dosage, the higher are the combined n = 1.2 and n = 1.5, respectively, at 25 °C. From the Figs. 5
and accelerated hydration peaks. The precipitation of and 6, the trend of magnifying the additional initial peaks by
C – S– H gel results in a very long induction effect. All curves increasing the modulus as well as the dosage of water glass
show an appreciable induction period and 35 h is almost on is observed. The higher the water glass dosage, the higher
the top of the accelerated hydration peak for alkali – slag are the combined and accelerated hydration peaks. The
cement activated with 3% water glass. The longer induction precipitation of C-S-H gel results in a very long induction

Fig. 5. Effect of water glass with n = 1.2 on heat evolution rate of alkali – slag cement.
1186 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188

Fig. 6. Effect of water glass with n = 1.5 on heat evolution rate of alkali – slag cement.

period and the top of the accelerated hydration peak with with water glass or sodium metasilicate, the contribution of
3% of water glass with moduli of 1.2 and 1.5 is about 40 heat of hydration depends on the nature of slag, the modulus,
and 43 h, respectively. and the dosage of activators used. According to results ob-
The longer induction period, varying from about 5 h, for tained, some authors [8,10] reported that alkali – slag cements
slag activated with 6% of water glass with moduli of 1, 2, cannot be regarded as low-heat cements. Our experiments did
and 1.5, to 36 h, is observed before the appearance of the not include dosage over 6% Na2O by the mass of slag.
accelerated hydration peak. The higher the activator dosage,
the shorter is the induction period. 3.3. Effect of modulus and activator dosage on
physicomechanical properties of alkali– slag cements
3.2. Effect of activator dosage on cumulative heat
of hydration The strength development of alkali – slag cement mortars
activated with different moduli (n) and dosage of water glass
The effect of activator dosage on cumulative heat of and sodium metasilicate was studied using Portland cement
hydration of alkali– slag cement at 25 °C is illustrated in as control. Considering all factors affecting strength devel-
Figs. 7 –9. Increasing the modulus of water glass as well as opment of Portland cement, the nature and the dosage of
dosage shows an increase in cumulative heat of hydration. activators distinguish itself playing the crucial role in deter-
During the first 48 h, water glass with different moduli as mining the strength of alkali –slag cements. Many results
well sodium metasilicate-activated slag releases less heat than confirm the effect an alkaline activator has on strength
Portland cement at a given isothermal temperature for all development and that it may be different for slag of different
activator dosages. Thus, for alkali – slag cements activated origins [22,23]. The growth in compressive strength between

Fig. 7. Effect of water glasses with different n moduli and sodium Fig. 8. Effect of water glasses with different n moduli and sodium
metasilicate in the dosage of 3% Na2O of slag weight on cumulative heat of metasilicate in the dosage of 4% Na2O of slag weight on cumulative heat of
hydration of alkali – slag cement compared to Portland cement at 25 °C. hydration of alkali – slag cement compared to Portland cement at 25 °C.
 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188 1187

Fig. 11. Effect of water glasses with different n moduli and sodium
Fig. 9. Effect of water glasses with different n moduli and sodium metasilicate in the dosage of 4% Na2O of slag weight on compressive
metasilicate in the dosage of 6% Na2O of slag weight on cumulative heat of strength of mortar compared to Portland cement.
hydration of alkali – slag cement compared to Portland cement at 25 °C.

followed by water glass with lower n modulus and OPC,

different alkali-activated slag mortars and OPC is compared respectively, as shown in Fig. 12. Some authors reported that
in Figs. 10 and 11. alkali-activated slag concretes based on water glass often
All alkali– slag cements show higher ultimate strengths undergo greater drying shrinkage due to formation of silica
than Portland cement, but only the cement activated with or silica-rich gel during hydration [24 – 26]. According to
4% of water glass by weight of Na2O (n = 0.6) show higher Glukhowsky at al. [27], the initial phases formed in alkali-
strength at early age — after 1 day. The cements with higher activated slag are rich in silica hydrosilicates and silica acid.
dosage of water glass show rapid strength gain after 3 days. The silica acid can be polymerized into silica gel. The ease of
The lower initial strengths of alkali –slag cements can be the silicic acid to condense is increased at alkaline pH where
partly attributed due to length of induction period as it is slightly dissociated or in molecular state and occurs
mentioned before. More work is needed to establish rela- within only a few minutes. A silica gel has a high water
tionships between the strength of alkali – slag cements and content and begins to shrink, expelling water from the mass.
various factors. The shrinkage begins with this syneresis process, when free
The shrinkage in the water glass as well sodium meta- liquid enmeshed in the gel is spontaneously expelled. Syne-
silicate-activated slag mortars that were air-cured is consid- resis comes to an end when the water content of the gel is still
erably higher than that of corresponding prisms with OPC. high, about 90%. On further drying, the structure of the gel
Water glass with a high n modulus gives greatest shrinkage, continues to shrink and number of particle-to-particle bonds
in the structure gradually break.

Fig. 10. Effect of water glasses with different n moduli and sodium Fig. 12. Effect of water glasses with different n moduli and sodium
metasilicate in the dosage of 3% Na2O of slag weight on compressive metasilicate at the dosage of Na2O by the mass of slag on shrinkage of
strength of mortar compared to Portland cement. alkali – slag cement compared to Portland cement.
1188 D. Krizan, B. Zivanovic / Cement and Concrete Research 32 (2002) 1181–1188

4. Conclusions [5] A. Fernández-Jiménez, F. Puertas, A. Artega, J. Therm. Anal. 52

(1998) 945 – 955.
[6] W.J. McCarter, H.C. Ezirim, Monitoring the early hydration of poz-
Hydration process of alkali – slag cements and OPC zolan – Ca(OH)2 mixtures using electrical methods, Adv. Cem. Res. 10
are similar to each other, but mechanisms of hydration (4) (1998) 161 – 198.
are different. [7] W.J. McCarter, T.M. Chrisp, G. Starrs, The early hydration of alkali-
activated slag: Developments in monitoring techniques, Cem. Concr.
 Water glass-activated slag displays two heat evolution Compos. 21 (1999) 277 – 283.
[8] H. Zhou, X. Wu, Z. Xu, M. Tang, Kinetic study on hydration of alkali-
peaks during preinduction period, and by increasing activated slag, Cem. Concr. Res. 23 (1993) 1253 – 1258.
the modulus of water glass, the value of additional [9] C. Shi, Y. Li, Investigation on some factors affecting the characte-
initial peak increases and induction period lasting ristics of alkali-phosphorus slag cement, Cem. Concr. Res. 19 (1989)
increases as well. 527 – 533.
By increasing the dosage of activator, the induction [10] C. Shi, R.L. Day, A calorimetric study of early hydration of alkali –
slag cements, Cem. Concr. Res. 25 (6) (1995) 1333 – 1346.
period decreases. [11] C. Shi, R.L. Day, Some factors affecting early hydration of alkali –
 Cumulative heat of hydration increases by increasing slag cements, Cem. Concr. Res. 26 (3) (1996) 439 – 447.
the n modulus as well as dosage of water glass, but is [12] C. Shi, R.L. Day, Selectivity of alkaline activators for the activation of
still lower than that of Portland cement. slags, Cem. Concr. Aggregates 18 (1) (1996) 8 – 14.
 Alkali – slag cements activated with water glass with [13] P.C. Hewlett (Ed.), Lea’s Chemistry of Cement and Concrete, fourth
ed., Arnold, London, 1998.
moduli between 0.6 and 1.5 in appropriate dosage [14] Y.Sh. Shkolnick, Physicochemical principles of the hydraulic activity
show higher ultimate strengths than Portland cement. of blast furnace slags, 8th Int. Congr. Chem. Cem., Rio de Janeiro,
 Drying shrinkage of alkali – slag cements is consid- Theme 3, 4 (1986) 133 – 136.
erably higher than that of Portland cement and in- [15] P.V. Krivenko, Alkaline cements, 9th Int. Congr. Chem. Cem., New
creases by increasing the modulus as well as the Delhi 3 (1992) 482 – 488.
[16] B. Talling, J. Brandstetr, Clinker-free concrete based on alkali acti-
dosage of water glass. vated slag, Prog. Cem. Concr. Miner. Admixtures Proc., ABI Books,
 The alkali –slag cement, as low-cost product and with 1993, pp. 263 – 341.
the possibility of using by-products and wastes in [17] A.A. Novopashin, Theoretical principles of slag – glasses hydrolysis,
alkali-activated slag mixes, provides a solution to the Tr. Nauchno-Tekh. Konf., Izuch. Prim. Slatsezol’nye Tsem., (1971)
35 – 39.
present ecological problem.
[18] R. Dron, Setting mechanism of an alkaline activated granular slag,
Rapp. Rech.-Lab. Cent. Ponts. Chaussees 38 (1974) 128.
[19] K. Sato, E. Konishi, F. Fukaya, Hydration of blast furnace particle, 8th
Acknowledgments Int. Congr. Chem. Cem., Rio de Janeiro, Theme 3, 4 (1986) 98 – 103.
[20] Z. Rajaokarivony-Andriambololona, P. Ballif, J.H. Thomassin, J.C.
The authors thank the Republic of Serbia Ministry of Touray, Experimental hydration of two synthetic glassy blast furnace
slags in water and alkaline solutions at 40 °C: Structure, compo-
Science, Development and Technology for funding this sition and origin of the hydrated layer, J. Mater. Sci. 25 (5) (1990)
research through Project No. 02 E 41. Thanks to Dr. M. 2399 – 2410.
Komljenovic for useful discussion during paper preparation. [21] S.-D. Wang, K. Scrivener, Hydration products of alkali activated slag
Thanks to Mr. Z. Slovic and Mrs. E. Ostojic for the blast cements, Cem. Concr. Res. 25 (3) (1996) 561 – 571.
furnace slag supply and Mr. S. Dakovic and N. Dacic for [22] J. Malolepszy, Activation of synthetic melilite slags by alkalies, 8th
Int. Congr. Chem. Cem., Rio de Janeiro, Theme 3, 4 (1986) 104 – 107.
water glass and sodium metasilicate supply. [23] J. Malolepszy, M. Petri, High strength slag – alkaline binders, 8th Int.
Congr. Chem. Cem., Rio de Janeiro, Theme 3, 4 (1986) 108 – 111.
[24] C. Shi, R.L. Day, X. Wu, M. Tang, Comparison of the microstructure
References and performance of alkali – slag and Portland cement pastes, 9th Int.
Congr. Chem. Cem., New Delhi 3 (1992) 482 – 488.
[1] K.P. Mehta, Sixth CANMET/ACI (ICFSSNT) Int. Conf. Fly Ash, [25] S.-D. Wang, X.-C. Pu, K. Scrivener, P.L. Pratt, Alkali-activated slag
Silica Fume, Slag Nat. Pozzolans Concr., Bangkok, Thailand 1 cement and concrete: A review of properties and problems, Adv. Cem.
(1998) 1 – 21. Res. 7 (27) (1995) 93 – 107.
[26] T. Kutti, Hydration products of alkali activated slag, 9th Int. Congr.
[2] V.D. Glukhowsky, G.S. Rostovskaja, G.V. Rumyna, 7th Int. Congr.
Chem. Cem., Paris III (1980) V164 – V168. Chem. Cem., New Delhi 3 (1992) 468 – 474.
[3] V.D. Glukhowsky, (Slag – Alkali Concretes Produced From Fine- [27] V. Glukhovskij, Yy. Zaitsev, V. Pakhomov, Slag – alkaline cements
Grained Aggregates), Vishcha Shkola, Kiev, USSR, 1981, p. 223 and concretes — structure, properties, technological and economical
aspects of the use, Silicate Industries 10 (1983) 197 – 200.
(in Russian).
[4] A. Fernández-Jiménez, F. Puertas, Alkali-activated slag cements:
Kinetic studies, Cem. Concr. Res. 27 (3) (1997) 359 – 368.