Cement and Concrete Research 30 (2000) 1367 1374

Effect of admixtures on properties of alkali-activated slag concrete

T. Bakhareva,*, J.G. Sanjayana, Y.-B. Chengb
a

Department of Civil Engineering, Monash University, Wellington Road, Clayton, Victoria 3168, Australia
b
Department of Materials Engineering, Monash University, Clayton, Victoria 3168, Australia
Received 28 May 1999; accepted 21 June 2000

Abstract
This paper reports the results of an investigation on concrete that incorporated alkali-activated slag (AAS) as the only binder. The
activators were liquid sodium silicates (4 7% Na, mass of slag) and a multi-compound activator (NaOH + Na2CO3) (8% Na, mass of slag).
AAS utilizes industrial by-products and develops high early strength. However, some of its properties such as high shrinkage and poor
workability impede its practical application. Admixtures used for ordinary portland cement (OPC) were tested to improve these properties of
AAS concrete. Superplasticiser based on modified naphthalene formaldehyde polymers (S), air-entraining agent (AEA), water-reducing
(WRRe), shrinkage-reducing (SHR) admixtures at dosages of 6 10 ml/kg, and gypsum (G) (6% of slag weight) were used. The paper
presents the study of workability in the fresh state, shrinkage and compressive strength of AAS concrete, and the effect of admixtures and
type of activator on these properties. Concrete activated by liquid sodium silicate had the best mechanical properties. AEA, SHR, and G
significantly reduced its shrinkage. AEA also improved workability and had no negative effect on compressive strength. On the basis of this
investigation, AEA was recommended for use in AAS concrete. D 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Ground-granulated blast furnace slag; Alkali-activated cement; Admixtures; Workability; Shrinkage; Strength properties

1. Introduction
Alkali-activated slag (AAS) is a new type of binder used
in concrete manufacturing. Its production utilises industrial
by-products, requires less energy than ordinary portland
cement (OPC) and is associated with low CO2 emission.
AAS concrete has been found to have some superior properties as compared to OPC concrete, namely, low heat of
hydration, high early strength, and excellent durability in
aggressive environment [1 4]. Utilisation of a new material
in the cement industry requires practical experience in
solving processing problems such as rapid set and high
shrinkage, without impeding strength development. Previous investigations of AAS have shown that the strength
of AAS pastes is dependent on the concentration and type of
activator [5,6]. It has also been revealed that the material has
a workability problem, and that the shrinkage in most cases
exceeds that of OPC concrete [5,7,8]. Set-regulating admixtures developed for OPC concrete can be used to overcome

these problems. A previous investigation by Collins and

Sanjayan [9] of slag activated by sodium carbonate and
sodium hydroxide utilized such admixtures. The study
included mini-slump testing and was conducted on pastes,
but it did not consider the effect of admixtures on shrinkage
and compressive strength; also these tests were conducted on
pastes, whereas for concrete results could be different. Our
earlier study on the early hydration of alkali-activated slag
has shown that considerable differences exist in the mechanism of hydration when compared with OPC [10]. Thus, it
can be expected that admixtures could have different effects
on AAS than OPC. The focus of the current investigation is
on the study of the effects of admixtures on workability,
shrinkage, and compressive strength of AAS concrete activated by sodium silicate glass and multi-compound activator
in order to improve these properties of AAS concrete.
2. Materials
2.1. Slag

* Corresponding author. Tel.: +61-3-9905-4976; fax: +61-3-99054944.

E-mail address: [email protected] (T. Bakharev).

The chemical composition of the slag supplied by

SteelCement (Port Melbourne, Australia) is summarised

0008-8846/00/$ see front matter D 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 8 - 8 8 4 6 ( 0 0 ) 0 0 3 4 9 - 5

1368

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

Table 1
Composition of slag
Oxides

SiO2

Al2O3

Fe2O3

CaO

MgO

K2O

Na2O

TiO2

P2O5

MnO

SO3

Sulfide
(sulphur as S2 )

Cl

Loss on
ignition

Percentage

35.04

13.91

0.29

39.43

6.13

0.39

0.34

0.42

< 0.1

0.43

2.43

0.44

80 ppm

1.45

Table 2
Mix proportions of slag concrete
Mix no.

Activator, % Na, and Ms

1
2
3
4
5
6
7
8

Na silicate, 4% Na, Ms = 0.75

Na silicate, 7% Na, Ms = 1.25

Na2CO3, 1.7% Na; NaOH, 6.3% Na

Admixture,
ml/kg slag

Na hydroxide
solution, kg/m3

Activator, kg/m3

Added water,
kg/m3

WRRe, 10
S, 10
AAE, 6
G, 6%
SHR, 6
WRRe, 10
WRRe, 10

30.14

Na silicate, D, 36.72

147.4

32.47
65.99

Na silicate, D, 128.43
Na2CO3, 14.4

95.22
154.86

in Table 1. The blast furnace slag is a granulated product

ground to fineness of about 460 m2/kg, with the particle
size range of 1 10 mm and is neutral with the basicity
coefficient Kb=(CaO + MgO)/(SiO 2 + Al 2O 3), which is
equal to 0.93. The slag is supplied with 2% blended
gypsum (G).
2.2. Activators
The activators investigated were sodium silicate glass,1
sodium hydroxide solution,2 and sodium carbonate.3 Liquid sodium silicate and sodium hydroxide were blended
providing the modulus in solution (mass ratio of SiO2 to
Na2O), Ms, equal to 0.75 and 1.25. A multi-compound
activator, composed of sodium hydroxide (6.3% Na)4 and
sodium carbonate (1.7% Na) with a total Na content of
8% in the mixture was also used in the investigation.
Previous investigations of Collins and Sanjayan [7,8] of
slag activated by powdered sodium silicate utilized activator in dry powder state, together with hydrated lime.
The mechanism of activation and, consequently, properties
were rather different.
2.3. Chemical admixtures
The chemical admixtures5 investigated in this work
were as follows: S = short-life superplasticiser based on
modified naphthalene formaldehyde polymers, which com1

PQ Australia, sodium silicate solution grade D, weight ratio SiO2/

Na2O = 2, %Na2O = 14.7, %SiO2 = 29.4, technical grade.
2
Ajax Chemicals, 60% w/v water solution, technical grade.
3
Ajax Chemicals, technical grade.
4
Everywhere in the paper, percentage of Na refers to percent by mass
of slag.
5
W.R. Grace, Melbourne, Australia.

plies with ASTM C-494 Type A and F; WRRe = waterreducing and set-retarding admixture based on lignosulphonates, which complies with ASTM C 494 Types B and
D; AEA = air-entraining admixture with a soluble salt of an
alkyl aryl sulphonate, complying with ASTM C260, AS
1478-Type AEA; SHR = shrinkage-reducing admixture,
nonstandard in ASTM. In addition, G supplied by SteelCement was also used. G = gypsum, 80% CaSO4
2H2O, 15%
CaSO4
0.5H2O, and 5% CaSO4 ground to a fineness of
550 m2/kg.

3. Procedures
3.1. Concrete mixes
The main concrete mix design comprised slag (360),
coarse aggregates (1130), fine aggregates (830), and total
water (180) kg/m3. Table 2 lists the various mixes that
were made using different admixtures. The water to
binder (w/b) ratio was fixed at 0.5 to enable direct
comparison and reasonable concrete workability. Mixing
of concrete was performed in a 70-l mixer. The sequence
of mixing was as follows: mix for 2 min, rest for 2 min,
followed by remixing for 2 min. Activators were added in
water, the chemical admixtures were added to concrete
during mixing.
3.2. Test procedures
Table 3 shows the experimental programme including the
summary of specimens and tests. For AAS concrete, the
effect of type of activator on compressive and flexural
strength was elucidated first. Then the effect of admixtures
on workability, compressive strength, and shrinkage was

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

1369

Table 3
Summary of experimental programme
Mix no. Type of specimen
1
2
3
4
5
6
7
8
9

Type of activator

Na content,% weight slag Admixture Compressive strength

AAS concrete, w/b = 0.5 Liquid sodium silicate

4% Na, Ms = 0.75

AAS concrete, w/b = 0.5 Liquid sodium silicate

AAS concrete, w/b = 0.5 Na2CO3; NaOH
OPC Gr40

7% Na, Ms = 1.25
1.7% Na; 6.3% Na

WRRe
S
AEA
G
SHR
WRRe
WRRe

Slump Shrinkage

1, 3, 7, 28, 100, 200, 350 days X

1, 3, 7, 28 days
X
X
X

1, 3, 7, 28, 100, 200, 350 days X

1, 3, 7, 28, 100, 200, 350 days X
1, 3, 7, 28, 100, 200, 350 days X

X
X
X
X
X
X
X
X

Table 4
Properties of fresh concretes
Mix no.

Admixture,
ml/kg slag

Na/slag

Slump,
mm

Unit weight,
kg/m3

Air content,
%

Concrete
temperature, C

1
2
3
4
5
6
7
8
9

WRRe, 10
S, 10
AEA, 6
G, 6%
SHR, 6
WRRe, 10
WRRe, 10

4
4
4
4
4
4
7
9.4

55
200
200
200
60
95
200
250
100

2462
2460
2480
2367
2445
2523
2512
2435
2470

2
1.5
2
6
2
1
2
2.5
0.5

22
20
21
22
22
23
21
23
22

investigated mainly for concrete with sodium silicate activator. OPC concrete samples were used for comparison.
Compressive strength testing was conducted on cylinders
(100200 mm), and a total of three cylinders were tested for
each data point. Workability of concrete was assessed by
slump loss test. Shrinkage measurement was performed
using concrete shrinkage prisms 7575285 mm. A minimum of two shrinkage prisms was used in each test. The
first reading was taken immediately after demoulding (24 h
after casting). The prisms for drying shrinkage determination were stored in drying room where temperature was
maintained at 21 2C and relative humidity at 50 5%.
Shrinkage prisms used for measurements of autogenous
shrinkage were stored in 20-l plastic containers (four prisms
in each container) with sealed lids that prevents exchange of
moisture with atmosphere.

4. Results
4.1. Slump
The results of measurements of properties of the
freshly mixed concrete, i.e., temperature, slump, unit
weight, and air content, are presented in detail Table 4.
Fig. 1(a b) shows the effect of admixtures on workability
of AAS concrete prepared with multi-compound activator
and sodium silicate glass. Initially, slump was measured
for AAS concrete without any admixture. Sodium-silicateactivated slag concrete, with 4% Na, had a slump of 55
mm immediately after mixing, and remained workable
during half an hour, although slump continued to decrease. The same sodium-silicate-activated slag concrete
with WRRe admixture of 6 ml/kg had a slump of 80 mm

Fig. 1. Effect of admixtures on slump loss for AAS concrete samples with various activators: (a) multi-compound activator; (b) sodium silicate activator.

1370

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

Fig. 2. Compressive strength of AAS concrete.

and had better workability than concrete without admixture, but it also had some retardation in the initial set.
WRRe admixture addition of 10 ml/kg in the same
concrete mix produced a slump of 200 mm, which was
reduced to zero after 40 min. AEA at a dosage of 6 ml/
kg produced a considerable improvement in workability.
The mixes had a good workability during 30 40 min.
The sodium-silicate-activated slag concrete mix (4% Na),
with the addition of superplasticiser was initially very fluid,
having a slump above 200 mm. However, after 10 min, the
concrete lost its fluidity completely, and it appeared that the
superplasticiser was acting as set accelerating. Addition of
6% G did not improve workability.
The AAS concrete activated by sodium carbonate and
sodium hydroxide and containing the WRRe (10 ml/kg) had
a slump exceeding 200 mm after mixing. Later, slump
remained high for a long period of time. Thus, interaction
of AAS concrete with admixture depended on the activator.
WRRe and AEA were the most effective admixtures for
improving workability. In all cases, no adverse effects on
drying shrinkage have been observed, although the early
strength of AAS concrete for WRRe was relatively impaired
(see Fig. 3).
4.2. Strength
It was found that the type of activator has a significant
effect on strength of AAS concrete. Fig. 2 shows the results
of strength measurements for concretes activated by different

activators. Table 5 summarises the strength measurements at

1 month. As reported earlier [5], the best results were
obtained for sodium silicate glass. Two concretes were
prepared using sodium silicates at 4% Na and 7% Na by
mass of slag, and they exhibited compressive strength of 40
and 58 MPa at 28 days, respectively. The higher percentage
of sodium in the activator produced a higher strength.
Workability in the fresh state was decreasing rapidly with
concentration of activator. Thus, WRRe had to be introduced
in the second mix. The modulus of rupture determined for
these concretes by the four point bending method (AS 1012Part 11) was 6.5 and 6.4 MPa, respectively. Introduction of
5% ultrafine slag (UFS) with fineness of 1200 m2/kg
produced improvement in strength and workability. At 28
days, strength for the sodium-silicate-activated concrete with
UFS was increased to 46 MPa (comparing to 40 MPa), and
continued to rise at later stages. The concrete with the multicompound activator did not produce high strength; it was 20
MPa at 28 days with modulus of rupture of 3.2 MPa.
Fig. 3 shows the effect of admixtures at dosages of 6 10
ml/kg slag (mix proportions in Table 2) on compressive
strength of AAS concrete for up to 28 days. Superplasticiser
caused 25% loss of 28 day strength, and use of water
reducing admixture resulted in reduction of early strength,
up to 14 days. AEA had some effect on early strength up to
7 days, after that strength development was similar to AAS
concrete without admixture.
4.3. Shrinkage
Table 6 shows the summary of drying shrinkage test
results. The highest drying shrinkage was observed for the
AAS concrete with superplasticiser based on modified
naphthalene polymers (see Fig. 4). The drying shrinkage
in this case considerably exceeded the autogenous shrinkage. It was concluded that it is not desirable to use this type
of admixture in AAS concrete, although it may give good
results for OPC concrete.
Fig. 5 shows shrinkage of the sodium-silicate-activated
slag concrete prepared with lignosulphonate based WRRe,
at a dosage of 10 ml/kg slag, with 4% Na in the mix. This
concrete mix had shrinkage values considerably lower than
those for the concrete with superplasticiser, but the shrinkage was still higher than that of OPC concrete.

Table 5
Compressive and flexural strength of AAS concretes
Specimen type
Sodium-silicate-activated concrete, 4% Na, Ms = 0.75, w/b = 0.5
Sodium-silicate-activated concrete, 4% Na, Ms = 0.75 with 5%
ultrafine slag, w/b = 0.5
Sodium-silicate-activated concrete, 7% Na, Ms = 1.25, w/b = 0.5
AAS concrete with compound activator, 8% Na, w/b = 0.5
OPC Gr40, w/b = 0.5

Compressive strength
at 1 month, MPa

Flexural strength
at 1 month, MPa

40
46

6.5
7.0

58
20
45

6.4
3.2
5.6

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

1371

Fig. 3. Compressive strength of AAS concrete prepared using different

admixtures (dosage of admixtures in Table 2).

Fig. 4. Autogenous and drying shrinkage of sodium-silicate-activated slag

concrete, 4% Na with superplasticiser S at a dosage of 10 ml/kg slag.

Fig. 6 shows shrinkage characteristics for the concrete

with multi-compound activator and WRRe at a dosage of 10
ml/kg slag. The autogenous and drying shrinkage exceed the
shrinkage of OPC concrete.
Shrinkage curves for the AAS concretes activated by
sodium silicate (4% Na) and containing the air entraining
and shrinkage reducing admixtures, 6 ml/kg slag, are shown
in Figs. 7 and 8, respectively. The drying shrinkage exceeded the autogenous one, but in both cases, it was not
more than 400 microstrain at 56 days, and 450 microstrain
at 130 days, lower than for OPC concrete, though shrinkage
continued to grow. Shrinkage reducing admixture was found
to be insoluble in alkaline solution, so its interaction with
the concrete can give variable results.
Shrinkage curves for the AAS concretes activated by
sodium silicate (4% Na) and containing 6% G are given in
Fig. 9. The autogenous shrinkage exceeded the drying
shrinkage during the first 125 days. The drying shrinkage
was below 150 microstrain at 56 days. The autogenous
shrinkage was about 500 microstrain at 56 days. XRD traces
of AAS pastes with G (4 10%) at 3 days after start of
hydration are presented in Fig. 10. The traces show that
formation of Aft and AFm phases is high at 6% G, while at

8% and 10% G was reacted and no traces of G and AFt or

AFm type phases are evident.
5. Discussion
5.1. Workability
An increase in workability was achieved with the lignosulphonate water-reducing and -retarding admixture for
the activators used, namely, sodium silicate and combination
of sodium hydroxide and sodium carbonate, whereas the
time of initial set was retarded in both cases. With superplasticiser, the increase in workability was only at the initial
stage, and then a quick set occurred, which appeared to have
been accelerated by the admixture. Therefore, the behaviours of the pastes with these two admixtures were considerably different.
The lignosulphonate admixture had a prolonged positive
effect on workability as compared with the superplasticiser,
possibly due to its essentially nonpolar molecule and low
coulomb attraction of the slag particles with adsorbed nonpolar molecules. The superplasticiser has a polar molecule,

Table 6
Results of drying shrinkage test
Drying shrinkage, microstrain
Mix no.

Admixture

7 Days

14 Days

28 Days

112 Days

224 Days

400 Days

1
2
3
4
5
6
7
8
9

WRRe
S
AEA
G
SHR
WRRe
WRRe

580
200
1300
100
22
82
350
170
86

980
350
1500
140
30
100
550
240
356

1015
420
1800
230
56
116
780
470
600

1400
1000
2300
390
350
260
1235
1612
800

1814
1250
2500
440
550
600
1350
1970
850

1865
1300
2600

1450
1990
910

1372

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

Fig. 5. Autogenous and drying shrinkage of sodium-silicate-activated slag

concrete, 4% Na with lignosulphonate WRRe admixture at a dosage of 10
ml/kg slag.

Fig. 8. Shrinkage in AAS concrete with SHR admixture at a dosage of

6 ml/kg slag.

Fig. 9. Shrinkage in AAS concrete with 6% G.

Fig. 6. Autogenous and drying shrinkage of alkali-activated slag concrete
with compound activator with WRRe at a dosage of 10 ml/kg slag.

which can be absorbed rapidly on charged particles. This can

increase the zeta-potential of the hydrating particles and
promote quick set due to coulomb attraction between positively and negatively charged particles. A hypothesis is that
admixtures with nonpolar molecules are better suited to work
in a media of strongly charged particles of AAS and activator.

Fig. 7. Shrinkage of slag concrete activated by sodium silicate, 4% Na, with

AEA at a dosage of 6 ml/kg slag.

Collins and Sanjayan [9] indicated that slag activated by

NaOH + Na2CO3 had a reasonable response to retarders, but
the loss of workability beyond 30 min was significant, and
the 1-day strength was reduced. These results are close to
our findings. The effect of lignosulphonates is similar to the
action on portland cement concrete, except the time of
improved workability is much shorter, and dosage to obtain

Fig. 10. XRD of AAS paste with different content of G, 3 days after start of
hydration; C * = overlapping C S H and calcite, G = gypsum, CaSO4
2H2O; E = AFt phase, C3A
3CaSO4
32H2O; A = AFm phase, type
C3A
CaSO4
H11 13.

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

desirable effect is higher. AEA considerably improved

workability of AAS concrete possibly due to the physical
effects of entrained air.
5.2. Strength
Fig. 2 shows that AAS has the benefit of early strength
development as compared to OPC concrete. The strength
properties of AAS concrete measured are close to those
reported earlier by Collins and Sanjayan [7]. The effect of
the state of activator: powdered sodium silicates (Collins
and Sanjayan) and liquid sodium silicate glass (present
study) disappeared after the first 7 days. Liquid sodium
silicates provide a higher early strength, but concrete has a
lower workability due to immediate reaction and rapid
hardening of AAS [5,10]. However, powdered sodium
silicates are reliable for use only in a dry climate due to
the hygroscopic properties of dry sodium silicates.
The admixtures had different effects on the strength of
AAS concrete. Lignosulphonates retarded strength development. Similar to the effect on OPC, they increased the
workability and delayed the set of AAS concrete. Although
the superplasticiser caused an accelerated set, the compressive strength at later ages was lower than that of AAS
concrete without any admixture. Together with its high
shrinkage effect, this makes the superplasticiser unfavourable for use in AAS concrete. AEA had the least effect on
strength, but reduction in drying shrinkage and improvement in workability was significant.
5.3. Shrinkage
The reported values for shrinkage of AAS concrete vary
significantly [5,7,8,11,12], possibly due to different testing
procedures used. Similar to our work, Collins and Sanjayan
[7] and Kutti et al. [11] found that drying shrinkage in AAS
concrete activated by sodium silicates for RH>50% is 2.3
times drying shrinkage in OPC concrete, but Kutti et al. [11]
also show that for RH>70%, the drying shrinkage is similar
to OPC concrete. Collins and Sanjayan [8] reported that
shrinkage of AAS concrete with fully saturated blast furnace
slag coarse aggregate was 19% higher than that of OPC
concrete. Possibly, the saturated aggregates gradually release
the moisture, thus reducing the shrinkage of AAS concrete.
According to Australian Standard AS 1012, Part 13, standard
shrinkage prisms after demoulding have 7 days bath curing
followed by curing at drying conditions 23C and
RH = 50 5%. It is well known that shrinkage rate is at
maximum during first 7 days, thus, the measured shrinkage
would be less for bath cured prisms than for those exposed to
drying conditions from the first day after demoulding.
Collins and Sanjayan [8,12] observed some expansion during first 3 days for bath cured AAS concrete. It was not
observed for prisms exposed to drying conditions immediately after demoulding. However, the reported drying shrinkage for AAS concrete is, in most cases, much higher than for

1373

OPC concrete and reflects the high sensitivity of AAS

concrete to curing conditions.
Admixtures have a significant influence on volume stability of AAS concrete. The superplasticiser increased
shrinkage considerably, while the lignosulphonate admixture
did not increase it, but caused a slight decrease in shrinkage.
A low-drying shrinkage, comparable to that of OPC concrete
was observed for the AAS concrete prepared with AEA,
shrinkage reducing admixtures and G (6%). According to the
XRD results (Fig. 10), the observed low shrinkage was due
to the compensating effect of AFt and AFm phases which
formed in significant amounts at 6% G addition and had a
volume expansion when they were formed. Possibly, at 8%
and 10%, G was incorporated in a low-Ca C S H that
formed in this system [13]. The binding properties of C S
H increase with a decrease in the C/S ratio.
AAS concrete with AEA, G, and SHR had also small
differences between the autogenous and drying shrinkage.
AAS concrete develops large differences in strain inside the
member and on the surface due to large differences between
autogenous and drying shrinkage (see Figs. 4 6). Inside the
member the conditions are effectively autogenous, while
exterior surface is exposed to drying conditions. The measurements of in situ strain at the center and at the faces of a
column with the cross-section 800  800 mm2 made by
Collins and Sanjayan [12] showed that the differential strain
was 149 microstrain at 91 days. It is not high, but it may
cause cracks in OPC concrete, and, possibly, was a reason
for surface crazing in AAS concrete as it was reported in
Ref. [12]. For AAS concretes with AEA, SHR, and G, the
difference between autogenous and dying shrinkage was
less than 300 microstrain after 250 days, compared to 600
microstrain for OPC concrete. The drying shrinkage was
close to or less than for OPC, although it continued to grow
after 250 days. Thus, AEA, SHR, and G were effective in
minimising shrinkage cracking.
On the basis of our investigation, we can conclude that
AEA admixture is the most suitable for use in AAS concrete
activated by sodium silicate glass, as it improves workability, reduces shrinkage, and does not deteriorate later
strength. However, admixtures for OPC often have to be
used in large dosage, and may have significant side effects
(increased shrinkage, strength loss) when applied for AAS.
Effect of admixtures on AAS depends on the type of
activator, and can be opposite to the effect on OPC.
6. Conclusions
(a) An increase in workability was achieved with the
lignosulphonate admixture for all types of activators tested.
(b) With the naphthalene formaldehyde superplasticiser,
the increase in workability was only at the initial stage
and then a quick set occurred, which was considered
to have been accelerated by the admixture.

1374

T. Bakharev et al. / Cement and Concrete Research 30 (2000) 13671374

(c) Lignosulphonates may cause retarded strength

development and reduced flexural strength of
AAS concrete.
(d) Lignosulphonate-based admixture caused a slight
reduction in shrinkage, while naphthalene-based
superplasticiser significantly increased the shrinkage
and reduced the strength of AAS concrete.
(e) Air-entraining and SHR admixtures, as well as G
(6%) were effective in reducing shrinkage in AAS
concrete; AEA also greatly improved workability.
Thus, it was considered that AEA is the most suitable
for use in AAS concrete.
(f) Replacement of the ordinary slag by UFS (at
the 5% replacement level) improved workability
and strength.
Acknowledgments
Financial support for this project was provided by
Independent Cement and Lime, Blue Circle Southern
Cement and Australian Steel Mill Services. Assistance with
the laboratory work provided by Jeff Doddrell, Roger
Doulis, and Peter Dunbar is also gratefully acknowledged.
References
[1] D.M. Roy, G.M. Idorn, Hydration, structure, and properties of blast
furnace slag cements, mortars, and concrete, ACI J 79 (12) (1982)
444 457.

[2] X.C. Pu, C.C. Gan, S.D. Wang, C.H. Yang, Summary Reports of
Research on Alkali-Activated Slag Cement and Concrete, Vols. 1
6, Chongqing Institute of Architecture and Engineering, Chongqing, 1988.
[3] V.D. Glukhovsky, Slag Alkali Concretes Produced from FineGrained Aggregate, Vishcha Shkola, Kiev, USSR, 1981 (in Russian).
[4] V.D. Glukhovsky, V.A. Pakhomov, Slag Alkali Cements and Concretes, Budivelnik Publishers, Kiev, 1978 (in Russian).
[5] T. Bakharev, J.G. Sanjayan, Y.-B. Cheng, Alkali activation of Australian slag cements, Cem Concr Res 29 (1999) 113 120.
[6] S. Wang, K.L. Scrivener, P.L. Pratt, Factors affecting the strength
of alkali-activated slag, Cem Concr Res 24 (6) (1994) 1033 1043.
[7] F.G. Collins, J.G. Sanjayan, Workability and mechanical properties
of alkali activated slag concrete, Cem Concr Res 29 (1999)
607 610.
[8] F.G. Collins, J.G. Sanjayan, Strength and shrinkage properties of alkali activated slag concrete containing porous coarse aggregate, Cem
Concr Res 29 (1999) 607 610.
[9] F.G. Collins, J.G. Sanjayan, Early age strength and workability of slag
pastes activated by NaOH and Na2CO3, Cem Concr Res 28 (5) (1998)
655 664.
[10] T. Bakharev, J.G. Sanjayan, Y.-B. Cheng, Hydration of slag activated
by alkalis, J Australas Ceram Soc 34 (2) (1998) 195 200.
[11] T. Kutti, L. Berntsson, S. Chandra, in: V.M. Malhotra (Ed.),
Shrinkage of cements with high content of blast-furnace slag, Proc.
Fourth CANMET/ACI Int. Conf. on Fly Ash, Slag, and Natural
Pozzolans in Concrete, Istanbul, Turkey, 1992, pp. 615 625 Supplementary Papers.
[12] F. Collins, J. Sanjayan, Strength and shrinkage properties of alkali
activated slag concrete placed into a large column, Cem Concr Res,
1998 (accepted for publication).
[13] T. Bakharev, J. Sanjayan, Y.-B. Cheng, Early hydration and microstructure of alkali-activated slag cements, Cem Concr Res, 1999 (submitted for publication).