Materials and Structures (2007) 40:1161–1171

DOI 10.1617/s11527-007-9258-5

ORIGINAL ARTICLE

Comparative study on strength, sorptivity, and chloride

ingress characteristics of air-cured and water-cured
concretes modified with metakaolin
Erhan Güneyisi Æ Kasım Mermerdaş

Received: 29 November 2006 / Accepted: 2 May 2007 / Published online: 25 May 2007

RILEM 2007

Abstract This paper reports an investigation in 1 Introduction

which the performance of plain and metakaolin
(MK)-modified concretes were studied under two Concrete is the most important element of the
different curing regimes. The purpose of this study is infrastructure and well-designed concrete can be a
to evaluate the effectiveness of MK in enhancing the durable construction material. However, the environ-
strength and permeation properties of concrete. MK mental aspects of Portland cement are a growing
was used to replace 0–20% of Portland cement by concern, as cement manufacturing is responsible for
weight in concrete with two water-binder (w/b) ratios about 2.5% of total worldwide emissions from
of 0.35 and 0.55. The change in compressive strength, industrial sources. One effective way to diminish
sorptivity, and chloride ingress with age at all cement the environmental impact is to use natural pozzolans
replacement levels under both air and water curing and/or supplementary cementing materials, as a
are compared with those of the control concrete. The partial cement replacement. This strategy will have
results indicated that the inclusion of MK greatly the potential to reduce costs, conserve energy, and
reduced sorptivity and chloride permeability of reduced waste volumes [1]. The cementing materials
concrete in varying magnitudes, depending mainly that are widely used, concrete constituents, are fly
on replacement level of MK, w/b ratio, curing ash, granulated blast furnace slag, and silica fume [2].
condition, and chloride exposure period. It was found Metakaolin (MK), produced by controlled thermal
that under the inadequate or poor curing, MK- treatment of kaolin, is the most recent mineral
modified concretes suffered a more severe loss of admixture to be commercially introduced to the
compressive strength and permeability-related dura- concrete construction industry. The utilization of
bility than the plain concretes. calcined clay in the form of high-reactivity MK as
pozzolans for concrete has received considerable
interest in recent years. This interest has been focused
Keywords Chloride ingress
Concrete
Curing on the consumption of calcium hydroxide (CH)
regime
Metakaolin
Sorptivity produced by cement hydration which is associated
with poor durability. Thus, the use of MK improves
long-term strength and durability. In addition, it is
also possible to obtain early strength enhancement
E. Güneyisi (&)
K. Mermerdaş
through the filling effect [3].
Department of Civil Engineering, Gaziantep University,
Gaziantep 27310, Turkey According to the literature, it has been claimed
e-mail: [email protected] that concrete incorporated with MK exhibits pre-
1162 Materials and Structures (2007) 40:1161–1171

mium-level engineering properties comparable to diffusion testing showed a continued improvement in

silica fume concrete [4, 5]. This results from the chloride resistance for all of the mixtures. In the study
characteristics of its chemical composition, fineness, of Gruber et al. [21], it was also concluded that the
and poorly crystalline nature [6]. The research work use of 8% and 12% MK significantly lowered the
about the use of MK deals with two main areas. In the chloride ion diffusion coefficient of concrete. The
first one, the heat treatment parameters, the dehaydr- level of reduction compared to control specimens was
oxylation/amorphization process of the kaolinite, and on average 50% and 60% for mixes with 8% and 12%
the formation of the metakaolinite have been studied MK, respectively. Furthermore, Bai et al. [20]
[6–9]. The second one concerns the filling and showed that significant reductions in chloride pene-
pozzolanic behavior of MK and its influence on tration depths occurred when Portland cement was
cement and concrete properties [4, 10–24]. Research partially replaced with pulverized fuel ash and MK.
related to the effect of MK on the performance of These reductions increased with both increasing total
concrete is in progress due to the fact that MK is a replacement level and increasing exposure time.
relatively new mineral admixture for concrete indus- According to Ding and Li [16], the incorporation of
try. MK is also considered as an alternative supple- both silica fume and MK in concrete greatly reduced
mentary cementing material. It is comparable to silica the chloride diffusion rate. It was found that 15% MK
fume in pozzolanic reactivity, but is lower in price had a significant improvement while 5% MK had also
[16]. some improvement on the chloride resistance.
It is generally agreed that ingress of aggressive In this study, MK was used as a replacement for
species into concrete leads, in many structures, to Portland cement and the effect of MK replacement,
long-term deterioration. The concrete performance ranging from 0% to 20% by weight upon material
depends on mainly on the environmental conditions, properties was examined. For this, six different
the microstructures, and the chemistry of the con- concrete mixtures with two w/b ratios of 0.35 and
crete. Here, the continuation of hydration reactions in 0.55 were designed. The compressive strength, sorp-
portland cement is essential to improve the potential tivity, and chloride ingress characteristics of the MK-
strength and durability of concrete. This continuation modified concretes were determined at different ages
depends on the type (chemical and mineralogical up to 90 days. The study also investigates the effect
composition) and fineness of cement, the type and the of different curing regimes on the permeability-
amount of supplementary material present, the water- related durability of concrete incorporating MK as a
binder (w/b) ratio and the curing condition, especially supplementary cementing material. Based on the test
at early ages [25]. It is obvious that those factors are results, the effects of MK, w/b ratio, age, and curing
essential to obtain the durable concrete structures. procedure upon concrete properties were discussed.
Curing is necessary for concrete to fully realize its
potential properties [26, 27]. The curing becomes
more important for concrete incorporating pozzolans, 2 Experimental program
especially in hot and dry environments [28].
Recent works have shown that MK is effective on 2.1 Materials
enhancing the durability performance of concrete, for
example, alkali–silica reaction [21] and resistance to The materials used in this investigation were Portland
sodium sulfate attack [17]. In particular, concerning cement, MK, fine aggregate, coarse aggregate and
resistance to chloride ingress of MK concrete, few high range water-reducing admixture (HRWA). Port-
studies have been reported. Boddy et al. [19] studied land cement (CEM I 42.5R) conforming to the
long-term testing of the chloride penetration resis- Turkish standard TS EN 197–1 (which mainly based
tance of concrete containing high reactivity MK. on the European EN 197–1) and commercially
Concrete mixtures were cast with 0, 8, or 12% by available MK were utilized as cementitious materials.
mass replacement of Portland cement with MK at w/b MK is produced by thermal activation of high purity
ratios of 0.30 and 0.40. They found that resistance to kaolin clay, which originated from India, within a
chloride migration increased with increasing MK specific temperature range. The MK used in this
content and decreasing w/b. The long-term bulk study is a white powder with a Hunter L whiteness
Materials and Structures (2007) 40:1161–1171 1163

value greater than 90 (on a scale from ‘‘0-black’’ to 20% MK (by weight) for both series. Thus, totally six
‘‘100-maximum whiteness’’). It has a specific gravity different mixtures were prepared in this study. Details
of about 2.60, a specific surface area of 8600 m2/kg, of the mixtures are given in Table 3. Grading of the
and an average particle size less than 2 mm. The aggregate mixture was kept constant for all concretes.
chemical compositions and the physical properties of The mixtures given in Table 3 were designed to have
Portland cement and MK are given in Table 1. The slump values of 140 ± 20 mm and 180 ± 20 mm for
fine aggregate was a mix of river sand and crushed w/b ratios of 0.35 and 0.55, respectively for the ease
sand whereas the coarse aggregate was river gravel of handling, placing, and consolidation. The high
with a maximum particle size of 16 mm. Both range water-reducing admixture was added at the
aggregates were obtained from local sources. Prop- time of mixing to achieve the specified slump at each
erties of the aggregates are presented in Table 2. A w/b ratio. All concretes were mixed in accordance
commercially available sulphonated naphthalene with ASTM C192 standard in a power-driven
formaldehyde-based high range water-reducing revolving pan mixer. For each mixture, twelve
admixture was used to give a consistent workability. 100 · 100 · 100 mm cube and eight
Its specific gravity was approximately 1.22. 100 · 100 · 500 mm prism specimens were cast
and compacted by a vibrating table. After casting, the
2.2 Details of mixtures, specimens, and curing moulded specimens were covered with a plastic sheet
regimes and left in the casting room for 24 h. They were then
demoulded and divided into two equal groups and
Two series of control mixtures with w/b ratios of 0.35 cured under the given conditions below. Thereafter,
(Series 1) and 0.55 (Series 2) were designed. To they were tested in accordance with the specified test
develop the MK-modified concrete mixtures, the methods.
Portland cement was partially replaced with 10% and
2.3 Air curing (AC)

Table 1 Chemical compositions and physical properties of Specimens were left to cure continuously in the
cementitious materials used
laboratory under ambient condition until the test age.
Item Portland Metakaolin The temperature was moderately controlled at 20oC,
cement
but humidity was uncontrolled and generally ranged
SiO2 (%) 19.73 51.8 between 50% and 80%. Specimens were designated as
Al2O3 (%) 5.09 45.8 AC and are called air-cured at ambient temperature.
Fe2O3 (%) 3.99 0.35
CaO (%) 62.86 0.01 2.4 Water curing (WC)
MgO (%) 1.61 0.03
SO3 (%) 2.62 – Specimens were left to cure continuously in water at
Na2O (%) 0.18 0.13 20 ± 2oC until the test age. Specimens were
K2O (%) 0.80 0.06 designated as WC and are called water-cured at
Cl (%) 0.01 – controlled temperature.
Insoluble residue (%) 0.24 –
Loss on ignition (%) 1.90 0.91 2.5 Test methods
Free lime (%) 0.57 –
Specific gravity (g/cm3) 3.14 2.60 2.5.1 Compressive strength
Setting time, Vicat needle Initial/ 2-46/3-44 –
Final (h-min) The concrete cubes (100 · 100 · 100 mm) were used
Expansion, Le Chatelier apparatus 1 – for the compressive strength test at 28 and 90 days.
(mm)
The compression test was carried out on the speci-
Specific surface area (m2/kg) 327 8600
mens by a 3000 kN capacity testing machine. The test
Color Gray White
procedure followed during the test was in conformity
1164 Materials and Structures (2007) 40:1161–1171

Table 2 Sieve analysis and

Sieve size (mm) Fine aggregate Coarse aggregate
physical properties of
aggregates River sand Crushed sand No I No II

16.0 100 100 100 100

8.0 100 100 31.5 1.9
4.0 86.6 95.4 1.0 1.1
2.0 56.7 63.3 0.5 1.0
1.0 37.7 39.1 0.5 0.9
0.50 25.7 28.4 0.5 0.9
0.25 6.7 16.4 0.4 0.8
Fineness modulus 2.87 2.57 5.66 5.93
Specific gravity 2.66 2.45 2.72 2.73
Absorption, % 0.55 0.92 0.45 0.42

Table 3 Mixture proportions of the concrete

Concrete w/b MK Cement MK (kg/ Water Fine aggregate Coarse aggregate HRWAa
series ratio (%) (kg/m3) m3) (kg/m3) (kg/m3)
River sand Crushed sand No I (kg/ No II (kg/
(kg/m3) (kg/m3) m3) m3)

1 0.35 0 450 0 158 724 233 611 241 7.9

0.35 10 405 45 158 720 232 608 240 10.1
0.35 20 360 90 158 717 231 605 239 12.4
2 0.55 0 350 0 193 726 234 612 242 3.5
0.55 10 315 35 193 723 233 610 241 4.4
0.55 20 280 70 193 721 232 608 240 6.1
a
HRWA: High-range water-reducing admixture

with ASTM C39. Three specimens from each mixture total surface area of water within the tray should not
were tested at each testing age. be less than 10 times that of the specimen cross-
sectional area [29]. The specimens were removed
2.5.2 Sorptivity from the tray and weighed at different time intervals
up to 1 h to evaluate mass gain. The volume of water
The sorptivity test measures the rate at which water is absorbed was calculated by dividing the mass gained
drawn into the pores of concrete. For this, three test by the nominal surface area of the specimen and by
specimens having a dimension of the density of water. These values were plotted
100 · 100 · 100 mm cut from 100 · 100 · 500 prism
specimens were employed. The specimens were dried concrete sample
in an oven at about 50 ± 5oC until constant mass and
then allowed to cool to the ambient temperature in a
sealed container. Afterwards, the sides of the spec-
paraffin wax
imens were coated by paraffin wax and as shown in
Fig. 1, the sorptivity test was carried out by placing
the specimens on glass rods in a tray such that their water
bottom surface up to a height of 5 mm is in contact
with water. This procedure was considered to allow
free water movement through the bottom surface. The Fig. 1 Measurement of concrete sorptivity
Materials and Structures (2007) 40:1161–1171 1165

against the square root of time. The slope of the line made close to the corners of the section and ahead of
of the best fit was defined as the sorptivity coefficient the coarse aggregates were ignored. The test method
of concrete. For each test, measurements were followed in the present investigation to determine the
obtained from three specimens and the average depth of chloride penetration is reliable and accurate,
values were reported. and has also been recommended by other researchers
[30–33].
2.5.3 Chloride ion penetration

The prismatic test specimens (100 · 100 · 500 mm) 3 Test results and discussion
after being subjected to 28 days of initial curing
procedures were soaked continuously in 4% NaCl 3.1 Compressive strength
solution for 90 days. After the 90-day chloride
exposure period, the test specimens were withdrawn The data concerning the variation of compressive
from the soaking tank at 7, 14, 28, 56, and finally strength with w/b ratio, concrete age, and curing
90 days to determine the depth of chloride penetration condition for concretes incorporated with 0, 10, and
into plain and MK-modified concretes. For this 20% MK are plotted in Fig. 3. The strength values for
purpose, the prisms were first split at mid point and the plain and MK-modified concretes ranged from
the freshly split surfaces were soon sprayed with 0.1 N 36.1 to 72.0 MPa and from 37.4 to 93.6 MPa,
silver nitrate (AgNO3) solution. The AgNO3 solution respectively. The effect of MK on compressive
preferentially reacts with the free chloride present in strength of concrete is well observed in Fig. 3. The
the harden matrix to form a white precipitate of silver figure indicated that there was a systematic increase in
chloride (AgCl); whereas at greater depths, where free compressive strength with the increase in MK content.
chlorides are absent, AgNO3 reacts with the hydrox- This is more pronounced for concretes subjected to
ides to form a brown precipitate of silver oxide (AgO). water curing (WC). The degree of strength enhance-
Thus, the depth of chloride penetration is clearly ment (due to MK) is in direct relation to the
indicated as the boundary of color change. View of the parameters of age, curing condition, and the amount
boundary of color change after spraying AgNO3 of MK present. MK not only increases the early
solution for the plain and MK-modified concrete (28 days) strength of concrete but also increases the
specimens is illustrated in Fig. 2. The white coloring later strength (90 days). For example, the increase in
(formation of silver chloride) occurs wherever the compressive strength with increasing age was ob-
concentration of free chloride ion is greater than served to increase up to approximately 11% and 20%
0.15% by weight of cement. Measuring the depth of for the plain and MK-modified concretes, respec-
color change was performed from the four sides of the tively. It was also noted that for both w/b ratios,
split section at intervals of 20 mm. Measurements concretes with MK had remarkably higher compres-

Fig. 2 Photographs
showing the typical
boundary of color change
on the specimens after the
28 day chloride exposure
period (a) plain concrete
(MK0) and (b) MK
concrete (MK20)
1166 Materials and Structures (2007) 40:1161–1171

1.20
100 28 days
Compressive strength (MPa)

90 days

Normalized compressive strength, fc,air/fc,water

90
80
70 1.00
60
50
40
0.80
30
20
10
0 0.60
MK0

MK10

MK20

MK0

MK10

MK20

MK0

MK10

MK20

MK0

MK10

MK20
water curing air curing water curing air curing
0.40
w/b=0.35 w/b=0.55 28 day air-cured MK concrete
Concrete type 90 day air-cured MK concrete
0.20 28 day air-cured plain concrete
Fig. 3 Variation in compressive strength of plain and MK- 90 day air-cured plain concrete
modified concretes subjected to different curing regimes Water curing
0.00
0 20 40 60 80 100 120
Compressive strength of concrete subjected to water
sive strength than the plain ones. The percent increase curing (MPa)
in the strength with the use of MK ranged from
approximately 10–30% for the water-cured concretes Fig. 4 Influence of curing condition on 28 and 90 day
compressive strengths of concretes incorporated with MK
but from 2% to 14% for the air-cured concretes,
depending mainly on the variation in w/b ratio, age,
and replacement level of MK. The addition of MK 3.2 Sorptivity
into the matrix improves the bond between the cement
paste and aggregate particles as well as increasing the The change in sorptivity with w/b ratio, concrete age,
density of the cement paste which in turn significantly and curing condition for the plain and MK-modified
improves the compressive strength of the concretes. concretes are given in Fig. 5. It is apparent that
According to the literature, the main factors that affect sorptivity decreases systematically with an increase
the contribution of MK in strength are (a) the filling in curing period (from 28 to 90 days), and the
effect, (b) the dilution effect, and (c) the pozzolanic gradients of the sorptivity tends to decrease with
reaction of MK with CH [12]. increase in the replacement level of MK. At early age
The effect of curing condition on the 28- and 90- (28 days), the sorptivity values are clearly reflected in
day compressive strength of the plain and MK- the strength values. Thus, the concretes with the
modified concretes is illustrated in Fig. 4. For both lowest sorptivities have the highest strengths, espe-
concrete types, air curing (AC) resulted in signifi- cially for those subjected to water curing regime.
cantly lower 28- and 90-day strengths as compared to Increasing the MK content reduced both the 28- and
water curing (WC). It was observed that the ratios for 90 day sorptivities of the concrete. This is particularly
the plain and MK-modified concretes subjected to AC for the high w/b ratio water-cured concretes. The
did not fall within 10% of equality and deviated up to
approximately 24% and 34% from those cured
under WC, respectively. This implies that the reduc- 0.18 28 days
0.16 90 days
tion in strength is particularly large for concretes with
Sorptivity (mm/min )
0.5

0.14
MK, especially at later age (90 days). For inadequate 0.12
0.10
curing (air curing, AC), the absence of moist curing 0.08
0.06
resulted in lower strengths where the concrete 0.04
containing MK suffered more at both testing ages 0.02
0.00
(28 and 90 days). This can be attributed to the lack of
MK0

MK10

MK20

MK0

MK10

MK20

MK0

MK10

MK20

MK0

MK10

MK20

development of hydration and pozzolanic reactions to water curing air curing water curing air curing
produce a dense microstructure as well as the w/b=0.35 w/b=0.55
extensive shrinkage cracking, which may have Concrete type

developed due to continuous air curing which has Fig. 5 Variation in sorptivity of plain and MK-modified
also been indicated by other researchers [28, 34]. concretes to different curing regimes
Materials and Structures (2007) 40:1161–1171 1167

sortivity values of the concrete containing MK were This implies that MK concretes are very sensitive to
approximately from 2% to 36% and from 8% to 60% curing method adopted. Figure 6 also revealed that
lower than that of the plain concretes at 28 and there is a marked difference in sorptivity values for
90 days, respectively, depending on w/b ratio, the plain and MK-modified concretes at 90 days as a
amount of MK used, and curing regime. Generally, result of pozzolanic reaction which effectively
MK concrete performed better than the control reduces the volume of pore space.
concrete and marked improvements in terms of lower
rate of water penetration through capillary suction 3.3 Resistance to chloride ion penetration
were apparent, particularly under water curing con-
dition. This reduced sorptivity reflects a finer pore Figure 7 demonstrates the chloride penetration depth
structure that would, for example, inhibit ingress of measured periodically (at 7, 14, 28, 56, and 90 days)
aggressive elements into the pore system [35]. through the plain and MK-modified concretes sub-
Taşdemir [36] have also shown that type and fineness jected to two different initial curing followed by
of mineral admixtures (such as fly ash, limestone immersion in salt solution. Depths of chloride
filler, sandstone filler, and silica fume) together with penetration for plain concretes subjected to the initial
curing condition affect greatly the pore structure of curing conditions of AC and WC were in the range of
concrete and it was reported that a significant about 8–23 mm and 6–17 mm, respectively, depend-
reduction in sorptivity for the blended mixtures was ing mainly upon w/b ratio and immersion period in
obtained when the proper curing was applied.
A comparison of air-cured and water-cured con-
crete sorptivity with age for the plain and MK- 25
(a)
modified concretes is given in Fig. 6. The figure
Depth of chloride penetration (mm)

distinguishes the prevailing effect of water curing on 20

the MK specimens, especially tested at later age

(90 days). It was noted that air curing caused 15

significant increase in sorptivity of both plain and

especially MK concretes as compared to water 10
MK0-WC
curing. Results showed that ratios for plain and MK MK10-WC

concrete specimens under air curing condition lied 5 MK20-WC

MK0-AC
within a range of 1.2–1.4 and 1.4–2.5, respectively. MK10-AC
MK20-AC
0
0 2 4 6 8 10 12 14
3.00 Immersion period (weeks)
28 day air-cured MK concrete
90 day air-cured MK concrete 25
2.50 28 day air-cured plain concrete (b)
Depth of chloride penetration (mm)
Normalized sorptivity, Sair/Swater

90 day air-cured plain concrete

Water curing 20
2.00

15
1.50

10
1.00 MK0-WC
MK10-WC
MK20-WC
5
MK0-AC
0.50 MK10-AC
MK20-AC
0
0.00 0 2 4 6 8 10 12 14
0 0.05 0.1 0.15 0.2 Immersion period (weeks)
Sorptivity of concrete subjected to water curing
(mm/min0.5) Fig. 7 Variation in chloride penetration depth with immersion
period for plain and MK-modified concretes subjected to
Fig. 6 Influence of curing condition on 28 and 90 day different curing regimes (a) w/b ratio: 0.35 and (b) w/b ratio:
sorptivity coefficients of concretes incorporated with MK 0.55
1168 Materials and Structures (2007) 40:1161–1171

salt solution. However, the concretes incorporating The chloride penetration coefficient, K, is used in
MK had comparatively lower chloride penetration the present investigation as an indicator of the
depths which ranged from 6 to 20 mm for the initial resistance of concrete against chloride penetration.
curing of AC and from 2 to 15 mm for the initial Several authors previously used the similar test
curing of WC. It was also observed that the chloride method to calculate this parameter for comparing
penetration depths of all concretes decreased with the resistance to chloride ion penetration of different
decreasing w/b ratio from 0.55 to 0.35, irrespective of mortars and concretes [30, 32].
initial curing conditions. The effect of w/b ratio was Figure 8 shows the chloride penetration coeffi-
well observed at the end the 7-day chloride exposure cients (K) of all concretes. The figure indicated that
period for all initial curing conditions. The plain for the plain and MK-modified concrete specimens
concretes with high w/b ratio gave about 2.0 and 1.8 pre-cured for a similar condition, an increase in w/b
times greater chloride penetration depths than those ratio leads to a larger K or faster rate of chloride
with low w/b ratio for the initial curing conditions of penetration. Moreover, as expected, for the concretes
AC and WC, respectively. However, these values for of all w/b ratios, K marginally decreased with
the MK-modified concretes ranged from 1.7 to 2.3, improving pre-curing condition. For a given pre-
depending on replacement level of MK and initial curing condition and w/b ratio of the mix, the
curing period. The order of the increase reduced with magnitudes of K for the plain concrete specimens
the increasing chloride immersion period (from 7 to were also relatively higher compared to the corre-
90 days) for both concretes. In addition, the test sponding K for the MK-modified concrete specimens.
results exhibited that MK-modified concretes had It was also noted that increasing the replacement
noticeably lower chloride penetration depth in com- level of MK decreased considerably the chloride
parison to the plain concretes during the 90-day diffusivity. Moreover, graphs in Fig. 8 showed that
chloride exposure period, particularly under the the concretes made with 20% MK had the lowest
initial water curing condition. For example, at the chloride penetration coefficient, irrespective of w/b
end of the 90-day immersion period, the depth of ratio and pre-curing condition.
chloride penetration for high and low w/b ratio water- The effect of initial curing condition on the
cured concretes containing 10% MK were approxi- chloride penetration coefficient of the plain and
mately 7% and 10% lower than that of the control MK-modified concretes is illustrated in Fig. 9. Lack
concrete at similar condition, respectively. However, of proper curing produced similar effect on the
those for concretes with 20% MK were about 17% chloride ingress to that on the sorptivity of concrete.
and 21%. This result agrees well with the results from Air curing leads to higher chloride penetration
a previous study by Ding and Li [16]. They demon- coefficient for both plain and especially MK con-
strated that the incorporation of MK in concrete cretes. It was observed that the ratio of the chloride
reduced the chloride diffusion rate significantly. It
was concluded that 5% MK had some improvement
on the chloride diffusivity while 15% MK has a 8.00
WC
7.44
7.04
Chloride penetration coefficient, K

significant improvement. 7.00 AC

6.43

The chloride penetration depths measured for both 6.00 5.71

(mm per sqr t week)

5.42
4.90 4.85
plain and MK-modified concrete specimens during 5.00
4.17
4.52
4.16
the 90-day chloride exposure period were then used 4.00 3.71
3.14
to compute the chloride penetration coefficient, K 3.00

(mm per square root of week). The magnitudes of K 2.00

were estimated by using the empirical relationship 1.00

(3.1) between the depth of chloride penetration (X) 0.00

MK0 MK10 MK20 MK0 MK10 MK20
and duration of immersion in salt solution (t) as w/b=0.35 w/b=0.55
Concrete type
proposed by Bakker [37].
pffi Fig. 8 Variation in chloride penetration coefficient of plain and
X ¼ aþK t ð3:1Þ MK-modified concretes subjected to different curing regimes
Materials and Structures (2007) 40:1161–1171 1169

2.00 of the strength for the MK concretes was remarkably

Normalized chloride penetration coefficient, Kair/Kwater

Air-cured MK concrete
Air-cured plain concrete higher than that for the plain concretes, especially
Water curing
under water curing. In the case of the permeability
1.50
related-durability tests, the range of sorptivity coeffi-
cients for the MK concretes was comparatively lower
1.00
than that for the plain cement concretes. The ratio of
sorptivity coefficient of the plain concretes to that of
MK concretes ranged from 1.01 to 2.03, depending
0.50 mainly on the curing regimes. The behavior observed
in the sorptivity of the concretes was quite similar to
that in depth and coefficient of chloride penetration.
0.00 The concretes containing MK had lower chloride
0 1 2 3 4 5 6 7 8
Chloride penetration coefficient of concrete subjected to water
permeability than the plain concretes. The difference
curing (mm/sqrt week) between both concrete types reached as high as 33%
with improving the initial curing regime.
Fig. 9 Influence of curing condition on chloride penetration
coefficient of concretes incorporated with MK

4 Conclusions
penetration coefficient of the specimens pre-cured
under air curing to those pre-cured under water curing From the results presented in this paper, the following
for the plain concrete deviated up to 30%. However, conclusions can be drawn:
this ratio for concretes containing MK lied within a
range of 32–52%. This indicates that the chloride 1. This study indicated that the inclusion of MK
penetration of the concretes containing MK is more into concrete significantly enhanced the strength
sensitive to the lack of curing. This is due to the and especially permeability-related durability
change in hydration kinetics of total cementitious characteristics of the concrete in varying magni-
material and the hydration development at the curing tudes. Concrete containing MK shows higher
cessation time [38]. strength than that of the plain concrete but
Table 4 summarized the comparison of the order of marginally lower chloride penetration depths and
the variation in the concrete properties (compressive sorptivities. The order of the magnitude is
strength, sorptivity, depth of chloride penetration at depended mainly on replacement level of MK,
90 days, and chloride penetration coefficient). As it is w/b ratio, concrete age, and curing condition.
seen from the table, the four measured concrete 2. Curing played a critical role in realizing the full
properties had similar tendency for both plain and potential of concrete. It is necessary to pay
MK-modified concretes in that the latter had better careful attention when using MK in concrete due
performance in comparison to the former. The range to the fact that the performance properties of the

Table 4 Comparison of the

Concrete property Curing Plain MK-modified Ratio (plain/
order of the variation in the
condition concretes concretes metakalolin
concrete properties with
concretes)
respect to curing conditions
and concrete types Range of compressive strength AC 36.1–60.1 37.4–62.2 0.96–0.97
(MPa) WC 45.3–72.0 50.0–93.6 0.76–0.90
Range of sorptivity coefficients AC 0.100–0.171 0.071–0.169 1.01–1.41
(mm/min0.5) WC 0.071–0.129 0.035–0.084 1.54–2.03
Range of chloride penetration AC 17.4–22.2 14.7–20.4 1.09–1.18
depths at 90 days (mm) WC 12.9–17.1 10.8–15.4 1.11–1.19
Range of chloride permeability AC 5.42–7.44 4.52–7.04 1.05–1.20
coefficients(mm/week0.5) WC 4.17–5.71 3.14–4.85 1.18–1.33
1170 Materials and Structures (2007) 40:1161–1171

MK-modified concretes are more sensitive to 5. Kristof E, Juhasz AZ, Vassanyi I (1993) The effect of
curing method adopted. mechanical treatment on the crystal structure and thermal
behavior of kaolinite. Clays Clay Miner 41:608–612
3. Irrespective of w/b ratio and replacement levels 6. Frias M, Cabrera J (2001) Influence of MK on the reaction
of MK, air-cured concretes exhibit lower strength kinetics in MK/lime and MK-blended cement systems at
and considerably higher permeability relative to 20oC. Cem Concr Res 31:519–527
equivalent concrete that is water cured. It was 7. Gamiz E, Melgosa M, Maranon MS, Garcia JMM, Delgado
R (2005) Relationships between chemico–mineralogical
observed that the strength of the plain and MK- composition and color properties in selected natural and
modified concretes subjected to air curing devi- calcined Spanish kaolins. Appl Clay Sci 28:269–282
ated up to 24% and 34% from those cured 8. Shvarzman A, Kovler K, Grader GS, Shter GE (2003) The
under water, respectively. The application of air effect of dehyroxylation/amorphization degree on pozzo-
lanic activity of kaolinite. Cem Concr Res 31:405–416
curing also resulted in marked differences with 9. Kakali G, Perraki T, Tsivilis S, Badogiannis E (2001)
respect to the water curing in terms of sorptivity Thermal treatment of kaolin: the effect of mineralogy on
and chloride ingress characteristics of the con- the pozzolanic activity. Appl Clay Sci 20:73–80
cretes, particularly for those made with MK. 10. Caldarone MA, Gruber KA, Burg RG (1994) High-reac-
tivity metakaolin: a new generation mineral admixture.
4. The comparison of the order of the variation in Concr Int 16(11):37–40
the concrete properties (strength, sorptivity coef- 11. Zhang MH, Malhotra VM (1995) Characteristics of a
ficient, depth of chloride penetration at 90 days, thermally activated alumino-silicate pozzolanic material
and chloride penetration coefficient) pointed out ant its use in concrete. Cem Concr Res 25:1713–1725
12. Wild S, Khatip JM, Jones A (1996) Relative strength,
that there is a strong correlation between them. pozzolanic activity and cement hydration in superplasti-
The four measured concrete properties had cized metakaolin concrete. Cem Concr Res 26:1537–1544
similar tendency for both plain and MK-modified 13. Khatib JM, Wild S (1996) Pore size distribution of me-
concretes in that the latter had better performance takaolin paste. Cem Concr Res 26:1545–1553
14. Brooks JJ, Johari MAM (2001) Effect of metakaolin on
in comparison to the former. creep and shrinkage of concrete. Cem Concr Compos
23:495–502
In this study, a colourimetric method of measuring
15. Sabir BB, Wild S, Bai J (2001) Metakaolin and calcined
the depth of chloride penetration into concrete by clays as pozzolans for concrete: a review. Cem Concr
spraying a 0.1N AgNO3 solution was employed to Compos 23:441–454
compare the chloride penetration resistance of the 16. Ding JT, Li Z (2002) Effects of metakaolin and silica fume
on properties of concretes. ACI Mater J 99(4):393–398
plain and MK-modified concretes. However, as a
17. Khatib JM, Wild S (1998) Pore size distribution of me-
further research, the use of some additional long term takaolin paste. Cem Concr Res 28(1):83–92
tests such as divided diffusion cell or ponding test 18. Bai J, Sabir BB, Wild S, Kinuthia J (2000) Strength
would be helpful to measure or predict the chloride development in concrete incorporating PFA and metakao-
lin. Mag Concr Res 52(3):153–162
content of such concretes for comparison purposes.
19. Boddy A, Hooton RA, Gruber KA (2001) Long-term
testing of the chloride-penetration resistance of concrete
Acknowledgements The authors would like to thank to Mr. containing high-reactivity metakaolin. Cem Concr Res
Vakıf Kiriş for his invaluable assistance during the laboratory 31:759–765
phase of the study. 20. Bai J, Wild S, Sabir BB (2003) Chloride ingress and
strength loss in concrete with different MK-PFA-PC binder
compositions exposed to synthetic seawater. Cem Concr
References Res 33:353–362
21. Gruber KA, Ramlochan T, Boddy A, Hooton RD, Thomas
1. Aldea CM, Young F, Wang K, Shah SP (2000) Effects of MDA (2001) Increasing concrete durability with high-
curing conditions on properties of concrete using slag reactivity metakaolin. Cem Concr Compos 23:479–484
replacement. Cem Concr Res 30:465–472 22. Lee ST, Moon HY, Hooton RD, Kim JP (2005) Effect of
2. Neville AM (1996) Properties of concrete. Addison Wes- sodium concentration and replacement levels of metakao-
ley Logman, 4th and final edn, England lin on the resistance of mortars exposed to magnesium
3. Cook DJ (1986) Calcined clay, shale, and other soils. In: sulfate solutions. Cem Concr Res 35:1545–1553
Swamy RN (ed) Cement replacement materials, Surrey 23. Batis G, Pantazopoulou P, Tsivilis S, Badogiannis E (2005)
University Press, pp. 40–72 The effect of metakaolin on the corrosion behavior of ce-
4. Dunster AM, Parsonage JR, Thomas MJK (1993) Pozzo- ment mortars. Cem Concr Compos 27:458–467
lanic reaction of metakaolinite and its effects on Portland 24. Revilla CF, Ramirez SM, Varela MTB (2006) Modelling
cement hydration. J Mater Sci 28:1345–1351 of slaked lime-metakaolin mortar engineering characteris-
Materials and Structures (2007) 40:1161–1171 1171

tics in terms of process variables. Cem Concr Compos chloride penetration into the hardened cementitious matrix
28:458–467 materials. ACI Mater J 89(6):587–592
25. Neville AM, Brooks JJ (1990) Concrete technology. 32. Wee TH, Suryavanshi AK, Tin SS (1999) Influence of
Longman scientific and Technical, Singapore aggregate fraction in the mix on the reliability of the rapid
26. Khatri RP, Sirivivatnanon V, Yu LK (1997) Effect of chloride permeability test. Cem Concr Compos 21:59–72
curing on water permeability of concretes prepared with 33. Meck E, Sirivivatnanon V (2003) Field indicator of chlo-
normal Portland cement and with slag and silica fume. ride penetration depth. Cem Concr Res 33:1113–1117
Mag Concr Res 49(180):162–172 34. Toutanji HA, Bayasi Z (1999) Effect of curing procedures
27. Güneyisi E, Özturan T, Gesoğlu M (2005) A study on on properties of silica fume concrete. Cem Concr Res
reinforcement corrosion and related properties of plain and 29:497–501
blended cement concretes under different curing condi- 35. Bai J, Wild S, Sabir BB (2002) Sorptivity and strength of
tions. Cem Concr Compos 27(4):449–461 air-cured and water-cured PC-PFA-MK concrete and the
28. Ramezanianpour AA, Malhotra VM (1995) Effect of cur- influence of binder composition on carbonation depth. Cem
ing on the compressive strength, resistance to chloride ion Concr Res 32:1813–1821
penetration and porosity of concretes incorporating slag, 36. Taşdemir C (2003) Combined effects of mineral admix-
fly ash, or silica fume. Cem Concr Compos 17:125–133 tures and curing conditions on the sorptivity coefficient of
29. Razak HA, Chai HK, Wong HS (2004) Near surface concrete. Cem Concr Res 33:1637–1642
characteristics of concrete containing supplementary 37. Bakker RFM (1988) Chapter 3: initiation period, corrosion
cementing materials. Cem Concr Compos 26:883–889 of steel in concrete. In: Schiessl P (ed) RILEM Report of
30. Wee TH, Suryavanshi AK, Tin SS (2000) Evaluation of the Technical Committee 60-CSC. Chapman and Hall,
rapid chloride permeability test (RCPT) results for con- London, pp 22–55
crete containing mineral admixtures. ACI Mater J 38. Bonavetti V, Donza H, Rahhal V, Irassar E (2000) Influ-
97(2):221–232 ence of initial curing on the properties of concrete con-
31. Otsuki N, Nagataki S, Nakashita K (1992) Evaluation of taining limestone blended cement. Cem Concr Res 30:703–
AgNO3 nitrate solution spray method for measurement of 708