Alkali-Activated Slag concrete development and their practical use

V.Bilek 1, M. Urbanova2, J. Brus2, D. Kolousek3

1
ZPSV a.s., Brno, Czech Republic; 2Institute of Macromolecular Chemistry, Prague,
3
Czech Republic; Institute of Chemical Technology, Prague, Czech Republic

Abstract

New inorganic binders are searched which are able to substitute Portland
cement in some branches of the building industry. The reason is the
lowering of emissions of carbon dioxide and use of some secondary
materials. Alkali-activated slag (AAS) concretes - relatively well-known
composites - are one of the perspective binders. They show good
mechanical properties and good resistance to the attack of a variety of
chemical media but for practical application various different problems
must be solved; the optimum content of alkaline activator and its nature;
the composition of the activator for optimum setting and hardening time,
good workability, smallest volume exchanges and maximum strength and
high durability. These problems are discussed in the present paper. The
aim of the paper is to design an appropriate mixture for producing AAS
concrete and for performing actual pilot tests in a concrete plant.

1. Introduction

Alkali-Activated Slag (AAS) composites are composites prepared from

Ground Granulated Blast furnace Slag (GGBFS) activated through
appropriate alkali solution. Usually it is water glass and/or sodium
hydroxide sometimes also potassium hydroxide. Especially different forms
of CSH gel arise as main binding phases [1,2]. Sometimes also some
geopolymers occur. Some authors [3] recommend to use the term
„geopolymers“ for AAS composites. In any case, they show good
mechanical properties, good resistance to the attack of a variety of
chemical media and some other suitable properties but for practical
purposes it is necessary to solve any problems [ 4], especially the setting
of the AAS concretes and high shrinkage of AAS concretes which results
in cracking, especially at the surfaces of elements.

Some authors [2] published results about quickness of setting and

hardening in dependence on the nature of the activator. A very high
strength was obtained for water glass activated composites, lower for
NaOH or Na2CO3 activated composites. The different strengths are a
consequence of different form of CSH gel; w hich is confirmed by NMR
analysis. The properties of CSH gel and binding of alkalis were studied for
example in [2,5].
The shrinkage is a topic of interest of many researchers. Occurrence of
cracks is affected by the presence of silica in alkali-silicate activator [ 6].
The larger amount of silica, the bigger autogenous shrinkage. The
influence of storage condition is also very important for the cracks
occurrence and strength development [7] . S ignificantly lower strengths
were obtained for specimens exposed to laboratory conditions (23 oC and
50% r.h.) than for sealing or bath specimens.

The results may be affected by properties of slag. For this reason it is not
possible to adopt the results without any study with particular slag and
also alkali activator. Some important results were shown in other paper s
[4,8]. They show that minimum 6% of Na2O is necessary for a reasonable
course of setting and hardening for our specific condition. As was shown
in [8] the setting and hardening is strongly affected by the ratio Na2O/SiO2
in activator (water glass and sodium hydroxide). During hardening the
significant efflorescence take place on the surface of specimen. It is
necessary to avoid the efflorescence. The use of K + ions instead Na+ ions
in activator can be an appropriate way.

During the investigation three main ways were kept:

1. designed AAS concretes must to be easy to produce – without
problems with setting and hardening, with number of compounds
(minimum new containers for materials); the formulae should be as
simple as possible.
2. designed AAS concretes should have workability similar to Self-
Compacting Concrete. The AAS concretes can be useful in
competition with SCC concretes because alkalis and SiO2 bring
good workability.
3. AAS concretes should be cheaper or as expensive as usual
concrete

2. Materials

The used GGBFS is produced near Ostrava in the Czech Republic and
milled in Stramberk cement plant to specific surface 350 or 420 m2/kg. The
finer fraction, 420 m2/kg, was used in this paper. The RDA pattern of the
slag (Fig. 1) shows that the slag is glassy, only merwinite peaks are
recorded. The chemical composition of the slag is also shown in figure 1.
It is possible to use different kinds of water glasses which differ by silicate
modulus Ms. In literature [2,6,7] authors usually use water glass with Ms =
1.5. But this water glass is very sensitive to low temperatures; the solid
sodium silicate can arise. For this reason the water glass with Ms = 1.8
and dry mass content about 45% was used. For the modification of Ms of
water glass during mixing the solution of NaOH (concentration 50%) was
formerly used [8,9]. The NaOH solution is relatively cheap but this brings
some problems – especially efflorescence on the surfaces of beams. For
this reason, the replacement of solution of NaOH by solution of KOH
(concentration 50%) was tested.

Grounded limestone was sometimes used with fineness similar to GGBFS

fineness. Originally, lowering of concrete mixture price was the reason for
this but later also some better properties of these mixtures were recorded.
Drinking water and usual sand and aggregates were also used to prepare
mortars and concretes.

SiO2 39.8
Fe2O3 0.3
Al 2O3 8.3
CaO 37.3
MgO 11.8
S 0.6
Na2O 0.4
K2O 0.5

Blaine 428 kg/m2

Figure 1: RDA pattern and chemical composition of used slag

3. Searching for best activator composition

A number of possibilities as to retarding the setting of AAS composites is

known in the theoretical studies. The setting and hardening of the pastes
made from slag, sodium water glass (M s=1.8), 50% KOH and drinking
water (water/slag = 0.35) was investigated. The amount of dry mass of
activator to mass of slag and limestone (SiO2 + Na 2O + K 2O)/ slag +
limestone) is 6% or 10%. The reason of replacement of limestone is
discussed below. The setting (start of setting and time of setting) was
measured by Vicat apparatus.

It is evident that courses of setting differ for different amounts of alkali

activator, for NaOH or KOH and for the replacement of slag by limestone.
There are two possibilities to mix AAS mortars and concretes with
convenient time of setting. It is possible to use ratio
(Na2O+K2O)/SiO2 between 50/50 or 60/40 or use only KOH as activator
respectively – 100% K2O. As the content of Na2O must to be minimum
6%, see [4], the content Na2O + SiO2 in activator (total dry mass in
activator) must to be minimum 10% (see peak of grey dashed curve in
figure 2). For the optimum combination of sodium water glass and solution
of KOH, 10% of solid dry mass of activator represents 5% of alkalis (K2O +
Na2O) and 5% of SiO2.
A very important result is the retarding of setting by limestone which
replaces slag. The start of setting is twice as long for pastes with
replacement of 1/3 of amount of slag by limestone.

slag + 6% aa
slag + 10% aa
140 (slag + limestone) + 10% aa
120 with NaOH
start of setting [min]

100
80
60
40
20
0
40 50 60 70 80 90 100
% (Na 2 O + K 2 O)

Figure 2: Start of setting of AAS pastes in dependence on alkali activator

(aa) composition. The dry mass content Na2O+K2O+SiO2 is 6 or 10%

6.4 % SiO2 5% SiO2 4% SiO2

3% SiO2 2% SiO2 0% SiO2

Figure 3: Cracking of samples after measurement of setting through Vicat

apparatus. The dry mass content in activator (Na2O+K2O+SiO2) is 10%
There is some basic insight into cracking of AAS composites in figure 3.
The samples after Vicat test are conspicuously cracked if SiO2 content
increases. No cracking is observed for only K2O activated pastes. But it is
important to note the cracking is not the result of shrinkage only, but also
tensile strength and modulus of elasticity.

4. NMR study

Setting results evokes some question concerning the nature of hydration

products. It is also interesting to know if it is possible to consider
investigated AAS composites as geopolymers.
27
Al MAS NMR 29
Si MAS NMR
Al(IV)

Al(VI)

A A

B B

C
C

D D

E E
100 90 80 70 60 50 40 30 20 10 0 ppm

Figure 4: A - 10% K 2O; B - 6% K 2O; C – 5%(K2O+Na2O)+5%SiO2;

D - 3%(K2O+Na2O)+3%SiO2; E – CSH (white cement and water)

It was found from 27Al MAS NMR analysis (Fig. 4) that tetrahedral groups
exist in AAS pastes and their amount increase with increasing alkali
content (paste A,B). These groups are dominant compounds of
geopolymers. But in AAS pastes the resonance of Al IV groups occurs at 72
ppm and not at 58-60 ppm as it is obvious for geopolymers. The signal at
72 ppm indicates tabular structures with periodically alternating Al IV and
Al IV units. It is not possible to consider these structures as geopolymers;
they are characterised by 3-dimensional network. If water glass is a part of
activator (pastes C and D) the 27Al MAS NMR analysis shows increase of
Al IV inordinate units, see signals around 68 ppm. It indicates tetrahedrally
coordinated Al-atoms in amorphous phase which can correspond to
Q4(3Al) groups of geopolymeric mass. The water glass (SiO2 presence)
facilitates the creation of 3-dimensional geopolymer similar network. 29Si
MAS NMR analysis confirms the presence of Q 4(3Al) in pastes activated
by KOH and water glass and the absence of Q4(3Al) in only alkali (mix of
NaOH and KOH) activated pastes.

Detected signals correspond to Q 0, Q1 and Q 3 structural units which are

also present in usual CSH gel (see sample E). Nevertheless it is evident
that alkali activation induces polymerization of Q 0 and Q 1 and the origin of
short polymeric Si-O-Si chains. Generally, the observed results are in
accordance with other results [10,11].

5. Workability of mortars

On the basis of optimum course of setting and hardening some mortars

were prepared with a different replacement of Na2O by K2O.

250

(K+N)/S = 6/4
200
cone flow [mm]

(K+N)/S = 5/5
6%N + 0%S
150
6%K + 0%S
(a)
100

50
0 2 4 6 8
[% K 2O]

300

(K+N)/S = 6/4
250
cone flow [mm]

(K+N)/S = 5/5
6%N + 0%S
200
(b) 6%K + 0%S

150

100
0 2 4 6 8
[% K 2O]

Figure 5: Cone flow (a) and cone flow after 15 blows of table (b).
The ratio (K2O+Na2O)/SiO2 was kept 6/4, the ratio
(K2O+Na2O+SiO2)/(slag+ limestone) was 10%. Also mixtures activated by
only Na2O and K2O were prepared and mixtures with (K2O+Na2O)/SiO2
ratio 5/5. Water/(slag + limestone) ratio was 0.6.

It follows from the figure 5 that alkalis enhances workability, see also [4].
It is evident that workability is significantly better for the increasing of
amount of K2O in activator. This is true of (K2O+Na2O)/SiO2 ratio 6/4
(advantageous for Na2O in activator) and also of the ratio 5/5
(advantageous for mix of Na 2O and K 2O in activator). Also SiO 2 content
probably play a role because mixtures with 5% of SiO 2 show better
workability than mixture with 4% SiO2. These results are very important
because they permit to lowering of water/slag ratio.

6. Strengths development

With respect to early strengths it can be seen in figure 6 the minimum

content of alkali activator should be 10%. The content of slag has not got
some any mayor effect. However, for self-compacting concrete some
minimum amount of fine particles should be present in the mixture. The
alkali and water glass presence brings good plasticity and the amount of
fine particles content can be reduced but not below 450 kg/m 3 for used
aggregates. For this reason the replacement of slag by limestone was
tested.

70 slag 450 g, aa 6%
compressive strength [MPa]

60 slag 450 g, aa 8%
slag 450 g, aa 10%
50
slag 400 g, aa 10%
40
30
20
10
0
24 hours 28 days - auto 28 days - dry

Figure 6: Influence of amount of slag and amount of alkali activator on

compressive strength of mortars (beams 160x40x40 mm)

Compressive strength development for mortars with different replacement

of slag by limestone is shown in figure 7 for two different amounts of alkali
activator dry mass – (K 2O + Na 2O + SiO 2)/(slag + limestone) is 10% and
15% respectively. Ratio (K2O + Na 2O)/SiO2 is 5/5 or 7.5/7.5 respectively.
It is interesting that mortar with the replacement of 100 kg of slag by
limestone shows higher strengths than that made from slag only. Probably
the limestone affects the strength only a little and the effect of higher
actual ratio (K2O+Na2O+SiO2)/slag plays important role for strength
development. While this ratio is only 10% or 15% for mixture with only
slag, it increases to 12.8% or 15% respectively for mixture where 100 kg
or 150 kg of slag were substituted by limestone. The effect of storage
conditions (covering by PE-foil or dry storing in t ≈ 20 oC and r.h. ≈ 60%).

Na2O+K 2O+SiO 2 = 10% s 450 g, l 0 g

70
compressive strength [MPa]

s 350 g, l 100 g
60
s 300 g, l 150 g
50
40
30
20
10
0
24 hours 28 days - foil 28 days - dry

Na2 O+K 2O+SiO 2 = 15% s 450 g, l 0 g

70
compressive strength [MPa]

s 350 g, l 100 g
60
s 300 g, l 150 g
50
40
30
20
10
0
24 hours 28 days - foil 28 days - dry

Figure 7: Development of compressive strength of AAS mortars with

replacement GGBFS by limestone (beams 160x40x40 mm)

Strength development for mixtures tested for workability is shown in figure

8. The combine effect of K2O + Na 2O mixture is evident; the mixture of
these brings higher strengths than either one of them used separately.
The composition for 1.6% of K2O and 4.4% of Na 2O is convenient
because it is cheaper than the others. This ratio is one of the ratios used
for concrete preparation.
10
compressive strength [MPa]
8 K+N/S = 6/4
K+N/S = 5/5
6 6%N + 0%S
6%K + 0%S
4

0
0 2 4 6 8
[% K 2O]
Figure 8: 24-hour’s strength development of mortars used for workability
measurement (beams 160x40x40 mm)

7. AAS concretes

On the basis of previous experiments the concretes were prepared in a

laboratory mixer and some tests were performed. The 1m3 composition
and some results are shown in table 1.

Table 1: Composition and some properties of AAS concretes designed on

basis of the above mentioned tests

NK6/4 NK5/5-1 NK5/5-2 NK5/5-3

Na2O+K2O+SiO2 10% 7.8 % 10% 10%
(Na2O+K2O) /SiO2 6/5 5/5 5/5 5/5
Water/(slag+limestone) 0.50 0.55 0.52 0.52
GGBFS 420 m2/kg 450 350 350 300
Limestone (450 m2/kg) 0 100 100 150
Na-water glass Ms 1.8 62 60 78 78
50% solution KOH 60 28 36 36
Water 150 175 167 167
Sand 0/4 mm 790 790 775 775
Crushed aggr. 4/8 mm 240 240 260 260
Crushed aggr. 8/16 mm 550 550 515 515
workability - cone flow 585x580 620x600 580x600 580x580
Compressive str. 24 h. 5.6 3.6 4.2 3.7
Bending strength 24 h. 1.8 0.9 1.1 1.0
Compressive str. 28 d. 73.5 54.5 59.6 58.0
Workability was measured as Abram’s cone flow; bending strength was
measured on beams 400x100x100mm, compression strength was
measured on 100 mm cubes. Specimens were stored in wet condition
(r.h. ≈ 95%, t ≈ 20 oC).

All of the concretes show low early strength which is the biggest
disadvantage of these materials. Workability is good as well as 28-days
strengths. Cracking can occur (dry cured samples) but this is not a
problem if the concrete is wet cured.

8. Conclusions

On the basis of experiments which were performed on the AAS pastes,

mortars and concretes some of these important conclusions can be
emphasized:

- Combination of sodium water glass and potassium hydroxide is a

convenient way for efflorescence avoidance
- Combination of sodium water glass and potassium hydroxide is a
convenient way for enhancement of workability
- The optimum (Na2O+K2O)/SiO2 ratio has been found for both -
appropriate course of setting and strengths development
- Concretes were designed which can be used in practice. Problems can
represent some early strength
- Cracks occurrence is minimalised by careful curing
- NMR study shows that term „geopolymers“ is not very convenient for
this type of AAS composites because only a little part of binding products
can be consider as geopolymer-like products

Despite some problems, some elements were made from the above
mentioned AAS concrete mixtures. The investigation continues to reach
better properties and to avoid mentioned problems.

Acknowledgement
This outcome has been achieved with the financial support of the Ministry
of Education, Youth and Sports, project No. 2B06120.
The authors would like to express their sincere thanks to Ms. Pehalova,
from ZPSV a.s. for her help with the preparation and testing of specimens.

Literature
[1] P.V. Krivenko, The evolution of alkaline cements: Theory and practice,
in V. Bilek, Z. Kersner (Eds), Non-traditional Cement and Concrete, Brno
University of Technology, ISBN 80-214-2130-4, Brno, 2002, pp. 18-28
[2] A. Fernández-Jiménez, F. Puertas, Effect of activator mix on the
hydration and strength behaviour of alkali-activated slag cements,
Advances in Cement Research, 2003, Vol.15, No.3, July, pp.129 – 136

[3] J. Davitovits, personal communication, July 2006

[4] V. Bilek, Some experience of using slag alkaline concrete in Czech

Republic, in V. Bilek, Z. Kersner (Eds), 2 nd Non -traditional Cement and
Concrete, Brno University of Technology, ISBN 80-214-2853-8, Brno,
2005, pp. 488 – 493

[5] C. Shi, On the state and role of alkalis during the activation of
alkali/activated slag cement, in G. Grieve and G. Owens (Eds),
Proceedings of 11th ICCC, Durban, South Africa, ISBN 0-9584085-8-0,
2003, pp.2097-2105

[6] M.A. Cincotto, A.A. Melo, W.L. Repette, Effect of differentactivators

type and dosages and ralation to autogenous shrinkage of activated blast
furnace slag cement, in G. Grieve and G.Owens (Eds), Proceedings of
11th ICCC, Durban, South Africa, ISBN 0-9584085-8-0, 2003, pp.1878-
1887

[7] F. Collins, J.G. Sanjayjan, Cracking tendency alkali activated slag

concrete subjected to restrained shrinkage, Cem Concr Res 30 (2000)
pp.791 – 798

[8] V. Bilek, Alkali-activated slag concrete - development for practical use,

in M.V.Malhotra (Eds.) Proceedings of CANMET/ACI conference on
Warsaw 2007, in print

[9] V. Bilek, P. Rovnanik, Resistance of alkali activated slag concretes to

freezing and thawing in different solutions, in V.M.Malhotra (Eds)
Proceedings. of 8 th CANMET/ACI Conference on Recent Advances in
Concrete Technology, Montreal, 2006, Supplementary papers, pp. 1-14

[10] A. Fernández-Jiménez, F. Puertas, I. Sobrados, J. Sanz, Structure of

Calcium Silicate Hydrates Formed in Alkali Activated Slag: Influence of
the Type of Alkaline Activator, J.Am.Ceram Soc., Vol. 86, (2003), No. 8,
pp.1389 – 94

[11] F. Puertas, A. Fernández-Jiménez, M.T.Blanco-Varela, Pore solution

in alkali activated slag cement pastes. Relation to the composition and
structure of calcium silicate hydrate, Cem Concr Res 34(2004),pp.139-148