Sei sulla pagina 1di 8

Construction

and Building

Construction and Building Materials 20 (2006) 858–865


MATERIALS
www.elsevier.com/locate/conbuildmat

Compressive strength, chloride diffusivity and pore structure of


high performance metakaolin and silica fume concrete
C.S. Poon *, S.C. Kou, L. Lam
Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Hung Hom, Hong Kong, China

Received 14 August 2001; received in revised form 30 June 2005; accepted 4 July 2005
Available online 15 August 2005

Abstract

This study is to relate the mechanical and durability properties of high performance metakaolin (MK) and silica fume concretes
to their microstructure characteristics. The compressive strength and chloride penetrability of the control and the concretes incor-
porated with MK or silica fume (SF) at water-to-binder (w/b) ratios of 0.3 and 0.5 are determined. The pore size distribution and
porosity of the concretes are also measured. The effect of MK and SF on the interfacial porosity is discussed based on test results. It
is found that MK concrete has superior strength development and similar chloride resistance to SF concrete, and the MK concrete
at a w/b of 0.3 has a lower porosity and smaller pore sizes than the control (plain) concrete. The resistance of the concretes to chlo-
ride ion penetration correlates better with the measured concrete porosity than with the paste porosity. The differences between the
measured and calculated concrete porosity is smaller for MK and SF incorporated concrete than for the control concrete, indicating
an improvement in the interfacial microstructure with the incorporation of the pozzolanas. This difference is found to be related to
the strength and chloride penetrability of concrete to some degree.
 2005 Elsevier Ltd. All rights reserved.

Keywords: High performance concrete; Metakaolin; Silica fume

1. Introduction Metakaolin is a thermally activated alumino-silicate


material obtained by calcining kaolin clay within the tem-
It has been well recognized that the use of pozzolanic perature range 650–800 C [6]. It contains typically 50–
materials such as silica fume (SF) and fly ash (FA) is nec- 55% SiO2 and 40–45% Al2O3 and is highly reactive. An
essary for producing high performance concrete. These important difference between MK and natural pozzolans
materials, when used as mineral admixtures in high perfor- or other types of artificial pozzolans is that MK is a pri-
mance concrete, can improve either or both the strength mary product, while SF and FA are secondary products
and durability properties of the concrete. In recent years, or by-products. Thus, MK can be produced with a con-
there has been a growing interest in the use of MK as a trolled process to achieve the desired properties. It has
mineral admixture for a similar purpose [1–5]. A compre- been reported that the concrete incorporating 10% MK
hensive review of the studies on the use of metakaolin as a had a higher compressive strength than the control Port-
partial pozzolanic replacement for cement in mortar and land cement concrete at all ages up to 180 days [3,4].
concrete has recently been presented by Sabir et al. [6]. When compared with SF concrete at the same level of
replacement, MK concrete showed a faster strength
development at early ages, but had similar strength after
28 days [4]. With respect to the durability aspects, it has
*
Corresponding author. Tel.: +852 27666024; fax: +852 2334 6389. been reported that the resistance of MK concrete to chlo-
E-mail address: cecspoon@polyu.edu.hk (C.S. Poon). ride ion penetration was significantly higher than the

0950-0618/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2005.07.001
C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865 859

control concrete but similar to the SF concrete [4]. Over- Table 1


all, as concluded by Sabir et al. [6] after a review of exist- Properties of Portland cement, metakaolin and silica fume
ing studies, MK is a very effective pozzolan. When used Parameters Cement Metakaolin Silica fume
as a partial replacement in concrete, it results in enhanced SiO2 (%) 21.0 53.2 85–96
early strength with no detrimental effect to the long-term Al2O3 (%) 5.9 43.9 –
strength, and greatly improves the resistance to the trans- Fe2O3 (%) 3.4 0.38 –
CaO (%) 64.7 0.02 –
portation of water and diffusion of harmful ions. MgO (%) 0.9 0.05 –
A number of studies have been conducted on the Na2O (%) – 0.17 –
hydration process and microstructure changes of cement K2O (%) – 0.10 –
pastes containing MK [7–12]. These studies show that at TiO2 (%) – 1.68 –
early ages the rate of pozzolanic reaction is higher in SO3 (%) 2.6 – 0.3–0.7
Loss on ignition (%) 1.2 0.50 3.5
MK pastes than in SF pastes, but after prolonged curing Specific gravity 3.15 2.62 2.22
it becomes slower in MK pastes. When used in cement Fineness (>45 lm, %) – – 3–5
pastes with a higher water-to-binder (w/b) ratio (e.g. Specific surface (cm2/g) 3520 12,680 –
w/b = 0.55), MK results in smaller pore sizes but higher
total porosity [8]. However, when used in high perfor-
mance cement pastes (low w/b ratio), it reduces both mixes with 5%, 10%, and 20% metakaolin and two SF
the pore sizes and total porosity [12]. mixes with 5% and 10% silica fume; and a control mix
The present study is to relate the mechanical and without any mineral admixture. MK or SF was used
durability properties of high performance MK concrete as a replacement of cement on a weight to weight basis.
to their microstructure properties. Two series of con- The maximum nominal size of the coarse aggregate was
crete mixtures are prepared at the water-to-binder ratios 10 mm for the concrete with a w/b of 0.3 and 20 mm for
of 0.3 and 0.5. The compressive strength and chloride the concrete with a w/b of 0.5. For the concrete at a w/b
penetrability of MK blended concretes are determined of 0.3, the superplasticizer was used to achieve a slump
and compared with those of SF concretes. The porosity of not less than 100 mm. The details of the concrete mix-
and pore size distribution of high performance MK and tures are given in Table 2.
SF concrete mixes prepared with a w/b of 0.3 are deter-
mined using mercury intrusion porosimetry (MIP). The 2.3. Specimen preparation, curing and compression test
effect of MK and SF on the porosity of the interfaces
between the cement matrix and aggregates is evaluated The concrete mixtures were prepared in the labora-
by comparing the data of the present study with the data tory using a pan mixer. Cubes of 100 · 100 · 100 mm
on the porosity of cement pastes reported in a separated in dimension and cylinders of 100 mm in diameter and
paper [12]. The correlation between the pore structure 200 mm in height were cast in steel mould and com-
characteristics and the strength and durability properties pacted on a vibrating table. The cubes and cylinders
are then discussed. were removed from the moulds 24 h after casting, and
were allowed to cure in water at 27 C. The 100 mm
cubes were used for the determination of compressive
2. Experimental details strength at 3, 7, 28 and 90 days. The compression load
was applied using a Denison compression machine with
2.1. Materials 3000-kN capacity, at the rate of 0.3 N/mm2 s. The 100
by 200 mm cylinders were used for the determination
The cementitious materials used in this study were of the chloride penetrability and pore size distribution
Portland cement (PC) equivalent to ASTM Type I, of the concrete at the above ages. From each concrete
metakaolin (MK) named MetaStar 450 obtained from cylinder, two 50-mm thick sections and one 20-mm thick
Imerys Minerals; and condensed silica fume (SF) named sections were obtained for these tests using a diamond
Force 10,000D microsilica obtained from W. Grace. The saw.
chemical and physical properties of these materials are gi-
ven in Table 1. The aggregates used were crushed granite 2.4. Determination of chloride penetrability
and natural river sand. A naphthalene-based superplast-
icizer with a solid content of 38.6% by weight was used The chloride penetrability of concrete was determined
for the mixes prepared at a water-to-binder ratio of 0.3. following the procedures of ASTM C1202-94 using the
50-mm thick portions as sample obtained from the 100
2.2. Concrete mixtures by 200 mm concrete cylinders. The resistance of concrete
against chloride ion penetration is expressed as an elec-
Two series of concrete mixes were prepared at the w/b trical indication: the total charge passed in coulombs
ratios of 0.3 and 0.5. Each series included three MK during a test period of 6 h.
860 C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865

Table 2
Proportions of concrete mixtures prepared
Series w/b Mix Water Binder (kg/m3) Fine aggregates Coarse aggregates Superplasticizer
(kg/m3) (kg/m3) (kg/m3) (l/m3)
Cement MK SF
1 0.30 Control 150 500 0 0 725 1087 0.5
0.30 5% MK 150 475 25 0 721 1087 0.6
0.30 10% MK 150 450 50 0 717 1087 0.8
0.30 20% MK 150 400 100 0 708 1087 1.0
0.30 5% SF 150 475 0 25 716 1087 0.6
0.30 10% SF 150 450 0 50 707 1087 0.8
2 0.50 Control 205 410 0 0 662 1081 –
0.50 5% MK 205 390 20.5 0 659 1081 –
0.50 10% MK 205 369 41 0 656 1081 –
0.50 20% MK 205 328 82 0 653 1081 –
0.50 5% SF 205 390 0 20.5 655 1081 –
0.50 10% SF 205 369 0 41 648 1081 –

2.5. Mercury intrusion porosimetry


Table 3
Compressive strength of control and blended concretes
Three small concrete cores of 21 mm in diameter were
obtained from each 20-mm thick section using a dia- Series w/b Mix Compressive strength (MPa)
mond drill. The small concrete cores were soaked in ace- 3 days 7 days 28 days 90 days
tone for stopping the hydration and were dried in a 1 0.3 Control 68.5 81.1 96.5 102.5
vacuum oven at 60 C for 48 h before testing. A ‘‘Pore 5% MK 73.0 88.2 103.6 112.9
Sizer 9320’’ mercury intrusion porosimeter (MIP) with 10% MK 85.9 99.8 116.8 120.3
20% MK 70.8 87.6 99.6 113.8
a maximum mercury intrusion pressure of 210 MPa
5% SF 67.0 79.3 106.5 110.2
was used for the determination. Cylindrical pore geom- 10% SF 63.2 76.9 107.9 115.6
etry and a contact angle h of 140 were assumed [13,14].
2 0.5 Control 28.6 41.2 52.1 60.4
The mercury intruded pore diameter dp at an intrusion
5% MK 32.6 45.9 57.1 66.5
pressure of PIn was calculated by dp = 4c cosh/PIn, 10% MK 40.4 55.2 66.2 71.6
where c = 0.483 N m1, the surface tension of mercury 20% MK 30.0 43.2 58.4 69.1
[14]. 5% SF 27.4 47.0 54.3 67.5
10% SF 25.8 47.4 58.4 69.1

3. Results and discussion

3.1. Compressive strength paper, the present authors showed [12] that at a 10%
replacement level, the degree of metakaolin reaction is
The results of the compressive strength test are shown 15% at 3 days while the degree of silica fume reaction
in Table 3, where each value is averaged from the results is 10% at this age.
of three cubes. The results show that the metakaolin It can also be seen from Table 3 that at the age of 3
used in this study is superior to silica fume in terms of days, the metakaolin replacement resulted in a higher
the strength enhancement of concrete. Among different strength increase for concretes with a higher w/b ratio
replacement levels, the use of metakaolin at the replace- than with a lower w/b ratio. This observation is contrary
ment level of 10% performed the best, which resulted in to the observation that fly ash contributes better to the
the highest strength increase over the control concretes early strength development at a lower w/b ratio than a
at all the test ages, particularly at the age of 3 days. higher w/b ratio [15]. This difference can be attributed
However, the incorporation of silica fume did not result to the difference between the strengthening mechanisms
in any strength increase for the concrete at 3 days, of metakaolin and fly ash in concrete. In MK concretes,
although it increased the strength at the ages of or after metakaolin contributes to the strength of concrete at
7 days. early ages mainly by the fast pozzolanic reaction [12].
The higher rate of strength development at early ages But in fly ash concrete, pozzolanic reaction is not the
for metakaolin concrete than for silica fume concrete predominant strengthening mechanism at early ages.
has also been observed by Zhang and Malhotra [4]. Fly ash contributes to concrete strength mainly by the
They attributed this phenomenon to the higher rate of pore filling effect and the enhancement of cement hydra-
hydration in the metakaolin concrete. In a different tion that are more significant at lower w/b ratios [15].
C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865 861

Table 4 Table 5
Chloride penetrability of control and blended concretes MIP measured total porosity of series 1 concrete (w/b = 0.3)
Series w/b Mix Total charge passed (C) Mix MIP measured total porosity (% v/v)
3 days 7 days 28 days 90 days 3 days 7 days 28 days 90 days
1 0.30 Control 2461 2151 1035 931 Control 8.69 ± 0.11 8.44 ± 0.13 7.92 ± 0.12 6.97 ± 0.28
5% MK 1327 1244 862 646 5% MK 7.22 ± 0.13 7.01 ± 0.15 6.40 ± 0.10 N.A.
10% MK 417 347 199 135 10% MK 6.87 ± 0.14 5.38 ± 0.12 4.75 ± 0.09 4.48a
20% MK 406 395 240 124 20% MK 6.59 ± 0.08 5.32 ± 0.10 4.66 ± 0.12 N.A.
5% SF 1060 945 665 426 5% SF 7.53 ± 0.16 7.24 ± 0.14 6.31 ± 0.11 5.85 ± 0.02
10% SF 567 445 360 336 10% SF 7.64 ± 0.15 6.14 ± 0.13 5.66 ± 0.12 5.11 ± 0.41
2 0.5 Control 5312 4054 2971 2789 N.A., not available.
a
5% MK 4215 3765 2079 1065 Only one measurement.
10% MK 1580 1247 918 752
20% MK 751 740 640 580
5% SF 3156 2047 1641 1235
10% SF 3140 1877 1523 1053 3.3. MIP measured concrete porosity and pore size
distribution

3.2. Chloride penetrability The porosity and average pore diameter of the con-
crete prepared at a w/b of 0.3 were determined using
The results of the chloride penetration test are shown mercury intrusion porosimetry (MIP) and the results
in Table 4, where each value is the average of two mea- are summarized in Tables 5 and 6, respectively, where
surements. Both the MK and SF concretes showed low- each value is the average of two measurements. Typi-
er total charges passed than the control. At the w/b of cal pore size distributions of the control concrete and
0.3 the concrete with a 10% MK replacement showed the concrete containing MK or SF are shown in
the best performance, while at the w/b of 0.5 the con- Fig. 1.
crete with a 20% replacement was the best. This was The results of the porosity and pore size distribution
slightly different from the results of the compressive of the cement pastes at a w/b ratio of 0.3, with the
strength tests, which showed that the concretes with a replacement of cement by MK, SF and FA at different
20% MK replacement had a lower compressive strength replacement levels had been presented in a separate
than with a 10% MK replacement at both w/b ratios. paper [12]. It was found that the incorporation of MK
When comparing the MK and SF concretes, it can be in the cement pastes resulted in a very dense microstruc-
seen that a 5% SF replacement in concrete resulted in ture of the paste, with a lower total porosity and finer
a lower total charge passed than a 5% MK replacement, pore size distribution compared with the plain Portland
but a 10% MK replacement resulted in a lower total cement pastes and SF blended cement pastes.
charge passed than a 10% SF replacement, at both The results of the present study show that the MK
w/b = 0.3 and 0.5. The difference in the total charge blended high performance concretes have lower MIP
passed between the MK concrete and the SF concrete porosity and smaller pore diameters than the corre-
with a 10% replacement level is particularly significant sponding SF concretes and the control concrete. This
when they were prepared at the w/b of 0.5. This seems trend is similar to that of the pastes. This can also be
to indicate that at a higher w/b ratio and percentage of seen in Fig. 1, where the cumulative mercury intrusion
replacement the use of MK is more effective than SF curves of the MK and SF concretes, particularly that
in improving the resistance of concretes to chloride ion of the concrete with a 20% MK replacement, are below
penetration. that of the control concrete at all studied ages.

Table 6
MIP measured average pore diameters of series 1 concrete (w/b = 0.3)
Mix Average pore diameter (lm)
3 days 7 days 28 days 90 days
Control 0.0370 ± 0.0012 0.0366 ± 0.0010 0.0322 ± 0.0009 0.0401 ± 0.0009
5% MK 0.0322 ± 0.0008 0.0314 ± 0.0012 0.0289 ± 0.0011 N.A.
10% MK 0.0276 ± 0.0011 0.0262 ± 0.0010 0.0259 ± 0.0009 0.0239a
20% MK 0.0269 ± 0.0014 0.0215 ± 0.0016 0.0186 ± 0.0015 N.A.
5% SF 0.0350 ± 0.0013 0.0351 ± 0.0014 0.0307 ± 0.0016 0.0394 ± 0.0004
10% SF 0.0345 ± 0.0015 0.0329 ± 0.0012 0.0295 ± 0.0011 0.027 ± 0.002
N.A., not available.
a
Only one measurement.
862 C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865

0.04
3 days 0.035
0.035 Control Control 28 days
Cumulative intrusion volume

5% MK 0.03

Cumulative intrusion volume


5% MK
0.03
10% MK 10% MK
0.025
0.025 20% MK 20% MK
0.02 5% SF
(ml/g)

5% SF

(ml/g)
0.02
10% S F 10% SF
0.015
0.015

0.01 0.01

0.005 0.005

0 0
1000 100 10 1 0.1 0.01 0.001 1000 10 0 10 1 0.1 0.01 0.001
a c

0.04 0.03
Con trol 7 days 90 days

Cumulative intrusion volume


0.03 5
0.025 Control
Cumulative intrusion volume

5% MK
0.03 10% MK 10% MK
20% MK 0.02 5% SF
0.02 5
10% SF

(ml/g)
5% SF
(ml/g)

0.02 0.015
10% SF
0.01 5 0.01
0.01
0.005
0.00 5
0
0
1000 10 0 10 1 0.1 0.01 0.001
1000 10 0 10 1 0.1 0.01 0.001
d
b

Fig. 1. Pore size distribution of MK, SF concretes and the control with a w/b of 0.3. (a) At 3 days. (b) At 7 days. (c) At 28 days. (d) At 90 days.

3.4. Calculated concrete porosity sand and coarse aggregate; VPa, VS and VA are the vol-
ume fractions of paste, sand and coarse aggregates in the
A comparison has been made by Kayyali [16] concrete. The calculated porosities of the control and
between the measured concrete porosity and calculated the MK and SF concretes are given in Table 7, where
porosity based on the measured porosities of the cement the porosities of pastes are extracted from [12], the
paste and aggregates. Poon et al. [17] demonstrated that porosities of sand and coarse aggregate are 2.33% and
for Portland cement mortars, the measured porosity is 1.89%, respectively, measured using MIP. The differ-
higher than the calculated porosity and attributed this ences between the measured and calculated porosities,
difference to the interfacial porosity, defined as the vol- which is given by PCmea  PCcal, are given in Table 8,
ume of interfacial pores per unit volume of mortar. They where PCmea is the MIP measured total porosity of con-
also found that the differences between the measured crete.
and calculated porosities for mortars blended with fly It can be observed from Table 8 that the measured
ash or both fly ash and silica fume are small and even porosity of the control concrete sample is greater than
negative, indicating the improvement effects of these the calculated porosity at all the ages. This is because
materials to the interfacial microstructure. In the present
study, a comparison is also made between the measured
and calculated porosities for the control and the MK Table 7
and SF concretes. The effect of MK and SF on the inter- Calculated porosity of series 1 concrete (w/b = 0.3)
facial porosity of concrete, defined as the volume of Mix Calculated total porosity PCcal (% v/v)
interfacial pores per unit volume of concrete, is then dis- 3 days 7 days 28 days 90 days
cussed. For a given volume of concrete, the calculated Control 7.64 6.99 6.24 4.85
porosity of concrete is given by 5% MK 7.07 6.19 5.72 4.41
P Ccal ¼ P Pa V Pa þ P S V S þ P A V A ; ð1Þ 10% MK 6.68 6.16 5.28 4.16
20% MK 6.55 5.47 4.80 3.45
where PCcal is calculated porosity of concrete; PPa, PS, 5% SF 7.27 6.68 5.96 4.75
and PA are the measured porosities of cement pastes, 10% SF 6.77 6.30 5.90 4.59
C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865 863

Table 8 140

Concrete compressive strength (MPa)


Difference between measured and calculated porosities 130
Mix PCmea  PCcal (% v/v) 120
y = 897.37x– 0.8492
3 days 7 days 28 days 90 days 110 2
R = 0.6145
100
Control 1.05 1.45 1.68 2.12
5% MK 0.15 0.82 0.68 90
10% MK 0.19 0.78 0.53 0.32 80
20% MK 0.04 0.15 0.14 70
5% SF 0.26 0.56 0.35 1.11
60
10% SF 0.87 0.16 0.24 0.52
50
40
5 10 15 20 25
in the calculation, the existence of interfacial pores has
a Paste porosity (%)
not been taken into account. It is found that the differ-
ences between PCmea and PCcal are smaller for the con- 140
crete with MK and SF than for the control. For the 130

Compressive strength (MPa)


MK and SF concretes, this difference even become neg- 120
ative at the ages of 7 and 28 days for the concrete with 110
10% and 20% replacements. This indicates a significant 100
improvement in the interfacial microstructure of the 90
concrete with MK and SF incorporations. 80
It should be noted that the difference between the 70
y = 346.99x–0.7288
measured and the calculated porosities does not reflect 60
R2 = 0.5038
the changes of interfacial microstructure with age, 50
because both the porosity of pastes and the concrete 40
change with ages. It just represents the effect of pozzola- 4 5 6 7 8 9
nic admixtures on the interfacial microstructure of con- b Concrete porosity(%)
crete at the given ages. So this difference may be smaller
140
Concrete compressive strength (Mpa)

at an earlier age but higher at a latter age, as can be seen


from Table 8. 120

100
3.5. Correlation of data
80
The relationships between the compressive strength 60
of concrete and paste porosity, concrete porosity and 3 days
the difference between measured and calculated porosi- 40 7 days
ties are shown in Fig. 2. It can be seen from Fig. 2(a) 20
28 days

and (b) that the compressive strength of concrete corre- 90 days

lated better with the paste porosity than with the con- 0
–1 –0.5 0 0.5 1 1.5 2 2.5
crete porosity. This might be due to that in the present
c P Cmea - P Cca l (%)
study the measurement of the concrete porosity is less
accurate than the measurement of the paste porosity Fig. 2. Relationship between compressive strength and porosities
because of the small concrete samples. In Fig. 2(c), the (w/b = 0.3). (a) Paste porosity. (b) Measured concrete porosity.
(c) Difference between measured and calculated concrete porosity.
compressive strength data are plotted against the differ-
ence between the measured and the calculated concrete
porosities. It is clear that for a given age, the strength
of concrete decreases with an increasing difference show that, the Coulomb data correlates better with the
between the measured and the calculated concrete concrete porosity than with the paste porosity. This is
porosity (PCmea  PCcal). This indicates that an consistent with our previous study on fly ash and SF
improvement in the interfacial microstructure will result blended concretes [17], showing the effect of interfacial
in an improvement in concrete strength. porosity on the chloride penetrability of concrete. In
In Fig. 3, the data of chloride penetration test, Fig. 3(c), the data are also plotted separately for differ-
expressed as the total charge passed in Coulombs, are ent ages. The data show clearly an increasing trend of
plotted against the paste and concrete porosities total charge passed with increasing difference in
(Fig. 3(a) and (b)), and the difference between the mea- (PCmea  PCcal). This confirms the dependence of chlo-
sured and the calculated concrete porosities ride ion penetrability of the interfacial properties as indi-
(PCmeaPCcal), respectively (Fig. 3(c)). Figs. 3(a) and (b) cated by others [18,19].
864 C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865

3000 determined with a w/b of 0.3. Based on the results and


discussion, the following conclusions can be drawn.
Total charge passed (Coulomb)

2500

2000
1. The performance of the MK used in this study was
superior to the SF in terms of the strength develop-
1500
y = 0.0648x3.3753
ment of concrete. But the performance of MK was
R2 = 0.6302 similar to SF in terms of the chloride resistance of
1000 concrete. MK concretes prepared at a w/b ratio of
0.5 showed higher compressive strength increases at
500
early ages than at a w/b ratio of 0.3, when compared
0 with the control concrete.
0 5 10 15 20 25 2. MK concrete prepared with a w/b of 0.3 had smaller
a Paste porosity (%) average pore diameters and lower MIP measured
total porosities at the ages of up to 28 days than
3000 the SF concrete and the control. The higher the
Total charge passed (Coulomb)

MK content, the lower was the porosity. This is con-


2500
sistent with our previous results on the porosity and
2000 pore sizes of MK blended cement pastes.
3.5366 3. The results of chloride penetration test, expressed as
y = 0.8552x
1500 2 the total charge passed in Coulombs, correlated bet-
R = 0.8274
ter with the MIP measured concrete porosity than
1000
with the paste porosity. This indicates a strong depen-
500 dence of the concrete transport properties on the
microstructure properties of both the cement matrix
0 and the matrix-aggregate interface.
2 4 6 8 10 4. The difference between the measured concrete porosity
b Concrete porosity (%) and the concrete porosity calculated based on the paste
3000
porosity was smaller for MK and SF blended concrete
3 days than for the control concrete, indicating the important
Total charge passed (Coulomb)

2500 7 days effect of MK and SF on the interfacial porosity. This


28 days difference was found to be related to the strength and
2000 90 days chloride penetrability of concrete to some degrees.
1500

1000
Acknowledgements

500 The authors acknowledge the financial support of the


Hong Kong Polytechnic University.
0
–1 –0.5 0 0.5 1 1.5 2 2.5
c P Cmea - P Ccal
References
Fig. 3. Relationships between chloride penetrability and porosities
(w/b = 0.3). (a) Paste porosity. (b) Concrete porosity. (c) Difference [1] Kostuch JA, Walters V, Jones TR. High performance concretes
between measured and calculated porosities. incorporating metakaolin: a review. In: Dhir RK, Jones MR,
editors. Concrete 2000. E&FN Spon; 1993. p. 1799–811.
[2] Caldarone MA, Gruber KA, Burg RG. High-reactivity metaka-
4. Conclusion olin: a new generation mineral admixture. Concrete Int
1994;16(November):37–40.
[3] Sabir BB, Wild S, Khatib JM. On the workability and strength
The aim of this study is to relate the mechanical and development of metakaolin concrete. In: Dhir RK, Dyer TD,
durability properties of high performance metakaolin editors. Concrete for environmental enhancement and protec-
(MK) concretes to their microstructure characteristics. tion. E&FN Spon; 1996. p. 651–6.
The compressive strength and chloride penetrability of [4] Zhang MH, Malhotra VM. Characteristics of a thermally
activated alumino-silicate pozzolanic material and its use in
the control concrete and the concretes incorporated with
concrete. Cement Concrete Res 1995;25(8):1713–25.
MK or silica fume (SF) have been determined with the [5] Curcio F, Deangelis BA, Pagliolico S. Metakaolin as a pozzolanic
water-to-binder (w/b) ratios of 0.3 and 0.5. The pore size microfiller for high-performance mortars. Cement Concrete Res
distribution and porosity of the concretes have been 1998;28(6):803–9.
C.S. Poon et al. / Construction and Building Materials 20 (2006) 858–865 865

[6] Sabir BB, Wild S, Bai J. Metakaolin and calcined clay as pozzolans [13] Day L, Marsh BK. Measurement of porosity in blended cement
for concrete: a review. Cement Concrete Compos 2001;16:441–54. pastes. Cement Concrete Res 1988;18:63–73.
[7] Murat M. Hydration reaction and hardening of calcined clays and [14] Taylor HFW. Cement chemistry. London: Academic Press;
related materials, I. Preliminary investigation on metakaolinite. 1990.
Cement Concrete Res 1983;13:259–66. [15] Lam L, Wong YL, Poon CS. Degree of hydration and gel/space
[8] Khatib JM, Wild S. Pore size distribution of metakaolin paste. ratio of high-volume fly ash/cement systems. Cement Concrete
Cement Concrete Res 1990;27:1545–53. Res 2000;30:747–56.
[9] Wild S, Khatib JM. Portlandite consumption in metakaolin [16] Kayyali OA. Porosity of concrete in relation to the nature of
cement pastes and mortars. Cement Concrete Res paste-aggregate interfaces. Mater Struct 1987;20:19–26.
1997;27(1):137–46. [17] Poon CS, Lam L, Wong YL. Effect of fly ash and silica fume on
[10] Frias M, de Rojas MIS, Carbrera J. The effect that the pozzolanic interfacial porosity of concrete. J Mater Civil Eng ASCE
reaction of metakaolin has on the heat of evolution in metakaolin- 1999;11:197–205.
cement mortars. Cement Concrete Res 2000;30:209–16. [18] Otsuki N, Miyazato S-I, Yodsudjai W. Influence of recycled
[11] Frias M, Cabrera J. Pore size distribution and degree of hydration aggregate on interfacial transition zone, strength, chloride pene-
of metakaolin-cement pastes. Cement Concrete Res tration and carbonation of concrete. J Mater Civil Eng ASCE
2000;30(4):561–9. 2003;15:443–51.
[12] Poon CS, Lam L, Kou SC, Wong YL, Wong R. Rate of [19] Yang CC, Cho SW. Approximate migration coefficient of
pozzolanic reaction of metakaolin in high performance cement percolated interfacial transition zone by using accelerated chloride
pastes. Cement Concrete Res 2001;31:1301–6. migration test. Cement Concrete Res 2005;35:344–50.

Potrebbero piacerti anche