Construction and Building Materials 26 (2012) 79–89

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Assessing drying shrinkage and water permeability of reactive powder

concrete produced in Hong Kong
C.M. Tam a, Vivian W.Y. Tam b,⇑, K.M. Ng a
a
Department of Building & Construction, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong
b
School of Engineering, University of Western Sydney, Locked Bag 1797, Penrith South DC, NSW 1797, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Reactive powder concrete (RPC) is coarse aggregate-free which differs from that of the ordinary concrete.
Received 8 November 2010 Instead, fine powders such as quartz sand and crushed quartz, with particle sizes ranging from 45 to
Received in revised form 30 March 2011 600 lm are used. In fact, it is rather a mortar than a concrete mixture because of the lack of coarse aggre-
Accepted 23 May 2011
gate. The mechanical property performance of RPC has been investigated by the previous researchers.
However, the performance on drying shrinkage and water permeability is of paucity in the literatures.
This paper examines the influences of water-to-binder ratio and superplasticizer dosage on the drying
Keywords:
shrinkage and water permeability of RPC. Recommendations for reducing drying shrinkage and water
Reactive powder concrete
Durability
permeability of RPC are also discussed.
Drying shrinkage Ó 2011 Elsevier Ltd. All rights reserved.
Water permeability
Hong Kong

1. Introduction reduction is good in producing more slender transportation

structures, reducing overall costs and increasing usable floor
Exciting new developments have been recorded in the area of space in high-rise buildings [1–3];
cementitious materials such as the recent innovations in high (ii) superior ductility and energy absorption (typical values of
strength concrete (HSC), high performance concrete (HPC) and 300 times greater than that of HPC and is comparable to
ultra-high strength (performance) concrete (UHSC or UHPC). With those of some metals) provide greater structure reliability
the development of these concrete technologies, these high-tech even under overload conditions or earthquakes [2–4]; the
materials provide great application potentials to the construction high energy absorption characteristics of RPC may also allow
field. Reactive powder concrete (RPC) is coarse aggregate-free improved post-elastic response of columns, beam-column
concrete, which has limited applications so far recorded in the joints and shear walls [5];
construction industry. RPC was originally developed by Rhodia, (iii) superior durability which leads to long service life with
Lafarge and Bouygues, the parent company of VSL, and is a regis- reduced maintenance. RPC is nearly impermeable, allowing
tered trademark under the name of Ductal [1], which reported that almost no carbonation or penetration of chlorides and sul-
the durability properties of RPC were in multiple folds of magni- phates. The enhanced abrasion resistance provides extended
tude better than the current high performance concrete. life for bridge decks and industrial floors [3]; while the
The ultra-high performance of RPC provides many advantages enhanced corrosion resistance provides protection to areas
compared to the conventional concrete; such as: with bad or harsh climate conditions such as concentrated
rain and snow, and heavy sandstorms [6];
(i) superior strengths with very high compressive strength of (iv) elimination of steel reinforcement bars reduces high labour
200 MPa (approximately four times the strengths of conven- costs and provides greater architectural freedom, allowing
tional concrete) result in significant reduction in dead load. nearly limitless structural member shapes and forms for
RPC structures may weigh only one-third or one-half of architects and designers [2–4];
the corresponding conventional concrete structures. Weight (v) reduction of thickness of concrete elements results in mate-
rial and cost savings [7];
(vi) a significant amount of unhydrated cement in the finished
product provides a self-healing potential under cracking
conditions [3]; and
⇑ Corresponding author. Tel.: +61 02 4736 0105; fax: +61 02 4736 0833. (vii) the fineness of the RPC product allows high quality surface
E-mail address: [email protected] (V.W.Y. Tam). finishes [3].

0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2011.05.006
80 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89

Though RPC possesses many outstanding properties, they also shrinkage increases with decreasing water to cement ratios and
have certain weaknesses; for example: with increasing silica fume content [14–17]. Autogenous shrinkage
becomes more significant as the strength of concrete is increased,
(i) cement content as high as 800–1000 kg/m3 not only affects and it is analogous to drying shrinkage [10]. The major differences
the production costs, but also has negative effects on the between drying shrinkage and autogenous shrinkage are listed in
heat of hydration, causing shrinkage problems [7]; and Table 1.
(ii) RPC is generally costly and cannot replace the conventional
concrete in most applications where the conventional mixes 2.3. Chemical shrinkage
can economically meet the performance criteria [3].
Chemical shrinkage is a phenomenon in which the absolute vol-
This paper aims to achieve the following objectives: ume (sum of volumes of solid and liquid phases) of hydration prod-
ucts is less than the total volume of unhydrated cement and water
 to investigate factors affecting the drying shrinkage and water
before hydration, and the consequence is a decrease in the absolute
permeability behaviour of RPC;
volume of the hydrated cement paste during its hydration (autog-
 to examine the influences of water-to-binder ratio and superp-
enous shrinkage) [18]. In other words, chemical shrinkage is the
lasticizer dosage on drying shrinkage and water permeability of
driving mechanism of autogenous shrinkage. Chemical shrinkage
RPC; and
is caused by hydration which depends on the porosity and on the
 to recommend methods for improving the behaviour of RPC.
rigidity of the material. It was confirmed that apparent volume
variations coincide with absolute volume variations until the min-
2. Theoretical background on shrinkage of concrete
eral skeleton is sufficiently rigid to restrain these deformations
[19,20].
Shrinkage of concrete is caused by loss of water due to evapora-
tion or by chemical change resulted from hydration of cement, as
well as by carbonation. In general, shrinkage of concrete is believed 3. Theoretical background on permeability of concrete
to take place in the cement paste matrix. There are many types of
shrinkages, in which drying shrinkage, autogenous shrinkage and Permeability refers to flow of fluids under a pressure differential
chemical shrinkage are the most common ones. Although some through a porous medium [21]. It is used for assessing the overall
other types of shrinkage, such as carbonation shrinkage, are also movement of fluids into and through concrete. Owing to the exis-
present, they contribute very little to the overall shrinkage and tence of different kinds of pores, some of which contribute to per-
may be neglected [8]. Shrinkage is the major cause of cracking in meability and some do not, it is thus important to distinguish
concrete, so it should be carefully monitored. between porosity and permeability.

2.1. Drying shrinkage 3.1. Relationship between porosity and permeability

Drying shrinkage of concrete is the reduction in volume caused Porosity is a measure of the proportion of the total volume of
by the withdrawal of water from hardened concrete stored in concrete occupied by pores, and is usually expressed in percent
unsaturated air to the surroundings [8]. It is also defined as the [21]. If the porosity is high and the pores are interconnected, they
time-dependent strain measured at constant temperature in an contribute to the transport of fluids through concrete and thus its
unloaded and unrestrained specimen [9]. Shrinkage presents the permeability is high. As stated from the above, the permeability of
greatest value at the member surfaces exposed to drying and de- concrete is not a simple function of its porosity, but also depends
creases towards the interior of the member [9]. If drying conditions on the size, distribution, shape and continuity of the pores. For
are the same at both top and bottom surfaces of the specimen, the example, though cement gel has a porosity of about 28%, its perme-
total strain would be uniform over the depth of the specimen; ability is only 7  1016 m/s, due to the extremely fine texture of
however, if drying occurs at a different rate from the top and bot- hardened cement paste in which the pores and the solid particles
tom surfaces, the total strain distribution becomes inclined and are very small and numerous [22].
warping of the member would result. Drying shrinkage induces
tension and results in the formation of cracks, leading to service- 3.2. Factors affecting permeability of concrete
ability and durability problems. Factors that affect drying shrink-
age of concrete include constituents, porosity, age of concrete, 3.2.1. Progress of hydration
curing temperature, relative humidity, moisture content and types Permeability of cement paste varies with the progress of hydra-
of admixtures used [10,11]. tion [21]. With the progress of hydration, the permeability rapidly
decreases because the gross volume of gel (including the gel pores)
2.2. Autogenous shrinkage is approximately 2.1 times the volume of unhydrated cement, so
that the gel gradually fills some of the original water-filled space.
Autogenous shrinkage is the apparent volume (sum of volumes In a mature paste, permeability depends on size, shape, concentra-
of its different components such as solid, liquid or gaseous phases) tion of gel particles and on whether the capillaries have become
reduction of cementitious materials by cement hydration [12,13]. discontinuous or not [23].
However, it does not include volume change due to loss or ingress
of substances, temperature variation, application of an external 3.2.2. Water-to-cement ratio
force and restraint [14]. When there is no additional water beyond The lower the water-to-cement ratio of the cement paste or
that added during mixing is provided, the concrete would begin to concrete, the lower is its permeability. For cement paste, a reduc-
dry out even if no moisture is lost to the surroundings. Both water- tion of water-to-cement ratios from 0.7 to 0.3 will lower the coef-
to-cement ratio and the incorporation of silica fume have signifi- ficient of permeability by 3 orders of magnitude [24]. Similarly, for
cantly effects on the autogenous shrinkage of concrete [15], which concrete, a reduction of water-to-cement ratios from 0.75 to 0.26
is due to the internal consumption of water during hydration, will substantially decrease the value of the coefficient of perme-
which leads to autogenous shrinkage [11]. The autogenous ability up to 4 orders of magnitude [25].
 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89 81

Table 1
Comparison of drying shrinkage and autogenous shrinkage.

Drying shrinkage Autogenous shrinkage

Mass loss U X
Mechanism From drying surface to the core of concrete Isotropically within concrete
Relative humidity gradient U X
developed
Time of occurrence Develop slowly at surface when hardened concrete is exposed to a dry Start in a matter of hours after casting of
environment concrete

Table 3
3.2.3. Properties of cement Descriptions of SP.
Permeability of concrete is also affected by the properties of ce-
Type of SP Sikament ViscoCrete – 1250NT
ment [24]. With the same water-to-cement ratios, coarse cement Basis Polycarboxylic polyether type polymer
tends to produce a hardened cement paste with a higher porosity Appearance Yellowish brown liquid
than finer cement. Density 1.08 kg/l
In order to study the performance of RPC in terms of drying pH value 7
Recommended dosage For soft plastic concrete: 0.25–1.0% by
shrinkage and water permeability, a series of experimental tests weight of cementitious
have been designed, which are described in the following sections. For flowing and self compacting concrete:
1.0–2.5% by weight of cementitious

4. Experimental design

4.1. Materials
Table 4
In this paper, constituents used in preparing the RPC mixtures are different from Details of RPC.
the conventional concrete mixtures. The materials include Ordinary Portland Ce- Type of concrete mix RPC
ment, silica fume, quartz sand, crushed quartz, superplasticizer (SP) and water. De- Water-to-binder ratio 0.2
tails of each constituent are recapitulated as follows. Cement (kg/m3) 761
Silica fume (kg/m3) 247
4.1.1. Ordinary Portland Cement 150–600 lm quartz sand (kg/m3) 1090
The cement used throughout the experiments is Ordinary Portland Cement 10–45 lm crushed quartz (kg/m3) 226
(OPC) which complies with British Standards [26,27] and has a 28-day mortar Water (kg/m3) 202
compressive strength of 58 MPa. The density is 3120 kg/m3 and the fineness is SP (kg/m3) 19 (2.5% by weight of cement)
3390 cm2/g. The initial and final setting times are 140 min and 165 min respec-
tively. The chemical composition of the cement is given in Table 2.

4.2. Mix proportions and curing conditions

4.1.2. Silica fume
The silica fume used is called Sikafume from Sika (Hong Kong) Limited which
In this study, a RPC mix is designed. Table 4 provides the details of the RPC mix
complies with ASTM standard [28]. The silica fume is extremely fine with particle
design which is based on some published recommended compositions [30–33]. The
size of 0.1 lm, which exists in grey powder form and contains latently reactive sil-
RPC mixes are produced using a high-speed mixer of volume 0.008 m3 with a max-
icon dioxide and no chloride or other potentially corrosive substances. The dry bulk
imum speed of about 1800 rpm.
density is 0.65 + 0.1 kg. The recommended dosage ranges from 2% to 10% of cement
The objective of this paper is to investigate drying shrinkage and water perme-
content by weight. For optimum results in concrete, it is suggested to use in con-
ability for RPC. Two RPC mix series with different water-to-binder ratios and SP
junction with Sikament SP, a superplasticizer.
dosages are produced for drying shrinkage tests and water penetration tests. Differ-
ent water-to-binder ratios, including 0.17, 0.2, 0.23, and 0.4 are used by keeping the
4.1.3. Quartz sand SP dosage constant at 2.5%; whereas to study the effect of SP dosage on drying
The majority of mixes are produced using quartz sand which replaces coarse shrinkage and water permeability for RPC, SP dosages are varied, including 2%,
aggregate from the conventional concrete. The quartz sand is supplied by Fantasy 2.5%, 3%, and 3.5% by keeping the water-to-binder ratio constant at 0.2. Drying
Quartz Sand Supply Company in Hong Kong. It is in a form of yellowish-white high shrinkage strains are measured up to 130 days; and water permeability coefficients
purity silica sand. are measured up to 105 days. The details are tabulated in Table 5.
A total of eight batches of RPC with different water-to-binder ratios and SP dos-
ages are experimented. Twenty specimens are cast in this study.
4.1.4. Crushed quartz
For drying shrinkage tests, 75  75  250 mm prisms are cast for the determi-
The crushed quartz used in the experiments is in a form of white powdered
nation of drying shrinkage for RPC. After demoulding, all specimens are placed in
quartz flour which replaces fine aggregate from the conventional concrete. The
a water tank at 27 °C and cured for 28 days. Then, all specimens are stored in a con-
quartz flour is also supplied by Fantasy Quartz Sand Supply Company in Hong Kong.
trolled environment of 25 + 2 °C and 50 + 5% relative humidity for subsequent mea-
The particle size used ranges from 10 to 45 lm.
surement. Tests are performed on two specimens each time due to the limited
volume (0.005 m3) of the high speed mixer. The strain readings are taken at 1, 7,
4.1.5. Superplasticizer (SP) 14, 28, 42, 56, 70, 90, 100, 110, 120, and 130 days.
The very low water-to-binder ratio used in RPC is only possible with the use of For water penetration test, 100  100  100 mm cubes are used for studying
SP for obtaining its workability. In this paper, the 3rd generation of SP called Sika- the water permeability of RPC. After demoulding, all specimens are placed in a
ment ViscoCrete – 1250NT from Sika (Hong Kong) Limited is used. It is an extremely water tank at 27 °C and cured for specific days for testing. Tests are performed on
high water-reducing agent that meets the requirements for SPs according to British three specimens each time. The readings are taken at 1, 7, 14, 28, 42, 56, 63, 70,
Standard [29]. Descriptions are provided in Table 3. 77, 84, 90, 98, and 105 days.

Table 2
Chemical composition of Ordinary Portland Cement.

Chemical composition SiO2 (%) Fe2O3 (%) Al2O3 (%) CaO (%) MgO (%) SO3 (%) Alkalis (%)
Ordinary Portland Cement 22.52 3.52 5.80 62.08 1.55 2.54 0.74
82 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89

Table 5
Experimental details.

Mix series Water-to-binder ratio SP dosage (% cement) Test and testing day
Effect of water-to-binder ratios 0.17 2.5% Drying shrinkage test at days 1, 7, 14, 28,
42, 56, 70, 90, 100, 110, 120, and 130.
0.20
0.23
0.40
Effect of SP dosages 0.20 2% Water penetration test at days 1, 7, 14, 28,
42, 56, 63, 70, 77, 84, 90, 98, and 105.
2.5%
3%
3.5%

4.3. Mixing sequence

An important factor for studying these new cementitious materials is the mix-
ing procedure [34]. This influence is often neglected and might be a source of error
when analyzing experimental results. Since RPC is composed of very fine materials,
the conventional mixing method is not appropriate and the mixing method cannot
be the same. The following sequence in mixing RPC is based on the previous studies
[5,32,35–38], as well as trial-and-error approaches from the research team:

1. dry mixing powders (including cement, quartz sand, crushed quartz and silica
fume) for about 1 min at a constant speed of 1800 rpm;
2. addition of half volume of water containing half amount of SP;
3. mixing for about 3 min with a speed of about 285 rpm;
4. addition of the remaining water and SP; and
5. mixing for about 8 min at a constant speed of 1800 rpm.

The entire mixing process takes about 12–16 min. When RPC mix is ready, it is
poured into the required moulds which have been sprayed with mould oil to reduce
the friction at the interface between the moulds and RPC mix. The RPC mix is com-
pacted using a vibrating table and hand tamping using a square tamping rod. The cast
moulds are covered by plastic sheets until demoulding to prevent moisture in the
concrete from losing to the surroundings. The specimens are demoulded at least
24 h after casting because of the high SP dosage which requires longer setting time.

4.4. Testing

4.4.1. Drying shrinkage

The shrinkage measuring apparatus used to determine the shrinkage strain for
RPC is the concrete length comparator as shown in Fig. 1. The comparator is com-
prised of an upright support of 1400 (356 mm) tall attached to a metal base and
including a 600 (152 mm) reference bar. It is designed to quickly and accurately mea-
sure the height of concrete samples to within ±0.1% of the total height. A digital
indicator to an accuracy of within 0.0001 in. (0.002 mm) with 0 to 100 (0 to Fig. 1. Concrete length comparator.
25 mm) total range is used. The shrinkage strain is calculated according to the
ASTM standard [39].

4.4.2. Water permeability

5.1.1. Effect of water-to-binder ratio on shrinkage behaviour of RPC
To assess the water permeability of the concrete, Germann’s Water Permeabil- There are various factors influencing the drying shrinkage of
ity testing device GWT – 4000 Kit developed by Germann Instruments A/S is used. hardened concrete. Water content is probably the largest single
The test is carried out in accordance with the Manual [40]. The GWT is used for test- factor influencing the shrinkage of paste and concrete [41,42].
ing micro-cracking and porosities of the concrete surface, the ‘‘skin concrete’’ by
Therefore, water-to-binder ratios would have a significant effect
applying water pressure to the surface and observe how the water flows. The differ-
ence in the gauge position over a given time (e.g. 5 min) is taken as a measure of the on shrinkage behaviour of RPC. Fig. 3 shows the drying shrinkage
water permeability for the water pressure. The GWT allows testing up to a maxi- development of RPC with different water-to-binder ratios within
mum of 6 bar pressure. Testing is performed on a horizontal face. The set up is illus- four months of about 130 days.
trated in Fig. 2. The attachment of the GWT unit is by means of two clamping pliers From Fig. 3, higher shrinkage rates are observed from day 1 to
anchored to the surface.
day 14 for all RPC mixes in general. The shrinkage rate increases
4.4.3. Microstructure and chemical composition
steadily from day 14 to day 100, and then gradually reduces from
Microstructure and chemical composition of RPC are examined using a JSM – day 100 to day 130. It has been reported that a saturated cement
820 Scanning Electronic Microscope fitted with an Energy Dispersive X-ray analyz- paste will not remain dimensionally stable when exposed to ambi-
ing facility. ent humidity that is below saturation. This is mainly because of the
loss of physically absorbed water from Calcium Silicate Hydrate
5. Results and discussions [CaOSiO2H2O or CSH] [11,21,43,44], resulting in a shrinkage
strain.
5.1. Drying shrinkage It is demonstrated that the higher the water-to-binder ratios,
the higher the shrinkage strains of RPC will be. The ultimate
Drying shrinkages of RPC with different water-to-binder ratios shrinkage strain of RPC with the water-to-binder ratio as high as
and SP dosages are measured at specific days. The result is summa- 0.4 is measured to be 1033 microstrain, which is significantly high-
rized in Table 6. er when compared to other lower water-to-binder ratios, with 454,
 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89 83

(a)

Water filling cup

with L-joint

Anchoring tools

(G-clamp)

(b)
Top knob of plier
Pressure chamber

Micrometer gauge

Pressure meter Tested Concrete

Fig. 2. (a) Set up of GWT – 4000 Kit; and (b) close up of the set up of GWT – 4000 Kit.

Table 6
Results of drying shrinkage for RPC.

Water-to-binder ratios SP dosages (%) Drying shrinkage (106)

Day 1 Day 7 Day 14 Day 28 Day 42 Day 56 Day 70 Day 90 Day 100 Day 110 Day 120 Day 130
0.17 2.5 43 67 99 111 141 163 231 342 363 407 439 454
0.2 167 212 247 261 300 353 426 504 532 557 561 572
0.23 203 287 313 330 350 397 448 539 569 600 611 619
0.4 241 324 495 606 684 751 810 920 955 1007 1018 1033
0.2 2 33 67 110 157 189 244 315 364 402 441 488 501
2.5 167 212 247 261 300 353 426 504 532 557 561 572
3 202 283 350 394 433 499 562 695 753 807 842 863
3.5 230 397 441 519 574 663 768 834 920 1018 1110 1148

572 and 619 microstrains for water-to-binder ratios of 0.17, 0.20 particular at the early age. Another reason may be attributed to
and 0.23 respectively. It is also noted that the shrinkage rate of the surface tension that develops in small pores as water evapo-
RPC at the early age (first 14 days) increases with increasing rates. When cement reacts with water, C–S–H forms in water-filled
water-to-binder ratios. Particularly for RPC with high water-to- spaces [45]. These spaces are not completely filled by the C–S–H
binder ratio of 0.4 which shows rapid increase in initial shrinkage. and so a capillary pore network develops. As water evaporates, a
The high drying shrinkage strain and shrinkage rate for RPC with meniscus will be formed among the pores. The surface tension of
high water-to-binder ratios are mainly due to disequilibrium in the water pulls the pore walls inward causing the concrete to
the relative humidity between concrete and its environment. More shrink. Therefore, the high drying shrinkage strain for RPC with
water is lost due to evaporation and thus concrete shrinks more, high water-to-binder ratio maybe due to more pores formed at
84 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89

Taking SP dosage of 2.5%, its drying shrinkage varies between

167 and 572 microstrain within 130 days. Many investigations
Drying Shrinkage (10-6)

showed that significant SP dosage is needed to decrease the

amount of water required without loss of workability [52,53].
The action of these high-range water reducers mainly derives from
a better particle dispersion as admixture is adsorbed on the cement
particles and hence, reduction of the surface tension of water re-
sults in lower capillary pressure which consequently decreases
the shrinkage strain [54].
The ultimate shrinkage strains of RPC with 3% and 3.5% SP dos-
ages are 863 and 1148 microstrains respectively, which are signif-
icantly higher when compared with other two SP dosages, which
are 501 and 572 microstrains for SP dosages of 2% and 2.5%
Age (Days)
respectively.
Fig. 3. Drying shrinkage of RPC with different water-to-binder ratios under various The high drying shrinkage strain and shrinkage rate are found
curing ages. for RPC with higher SP dosages. The high SP dosage would cause
changes in the pore structure which influence the state of the
evaporable water (or physical bound water) and consequently
higher water-to-binder ratios which develop higher surface ten- the shrinkage behaviour. Mixes with retardation or high SP dosage,
sion in pores. both autogenous shrinkage and evaporation increased, drying
It is interesting to point out that although the drying shrinkage shrinkage is increased and thus the crack tendency [50]. High SP
is lower for the lower water-to-binder mixture, the total shrinkage dosage will lead to increase in mass loss because of the continuous
could be higher at early and medium ages due to larger autogenous increase of water loss or evaporation [55].
shrinkage in low water-to-binder mixture [46]. It is known that
high strength concrete with low water-to-binder ratios can induce 5.1.3. Methods for reducing shrinkage
autogenous shrinkage because curing water cannot reach the It has been reported that concrete with low water-to-binder ra-
internal small pores and thus results in internal drying which is tio, high cement content, addition of superplasticizer and incorpo-
called self-desiccation [47]. However, for concrete with higher ration of silica fume, which are the characteristics of RPC, creates a
water-to-binder ratios, autogenous shrinkage is usually small be- water starved mixture and results in an extremely dense micro-
cause the concrete has enough water supplied into pores during structure within a few days or less [56]. This rapid development
curing period for hydration [48,49]. Concrete with a low water- of a very fine pore network within the cement paste of RPC would
to-binder ratio, autogenous shrinkage would be the dominant create an impermeable medium and effectively seal the concrete
cause whereas for concrete with a high water-to-binder ratio, off from its environment. As a result, concrete is unable to obtain
evaporation is the governing cause [50]. Therefore, it can be con- the externally available curing water and thus consumes water
cluded that low water-to-binder ratios for RPC has low drying from its own capillary pores. This self-desiccating action leads to
shrinkage but high autogenous shrinkage while high water-to-bin- microstructural stresses during drying, which may increase crack-
der ratios for RPC has high drying shrinkage but low autogenous ing of RPC. This can be demonstrated by SEM image in Fig. 5, which
shrinkage. shows microcracks present in RPC with water-to-binder ratio of 0.2
and SP dosage of 2.5% at 14 days. More cracks can be found for RPC
with the water-to-binder ratio of 0.4 at 14 days as shown in Fig. 6.
5.1.2. Effect of superplasticizer dosage on shrinkage behaviour for RPC
The combination of autogenous and drying shrinkage may lead
Superplasticizers are chemical admixtures that are incorporated
to cracks in concrete structures. In construction, shrinkage cracking
into the concrete to increase the workability in its fresh state. The
of concrete is one of the main factors causing deterioration of con-
incorporation of chemical admixtures is known to affect the poros-
crete durability. Cracking may cause excessive deflections which
ity of the hardened cement paste, both the total porosity and the
may limit the serviceability of structures. In addition, cracks in li-
pore size distribution [51], which consequently affects the shrink-
quid container can lead to leakage that limits the effectiveness of
age behaviour of the concrete. Fig. 4 shows the drying shrinkage
the structure [46]. Therefore, reduction of cracking is one of the
development of RPC with different SP dosages under constant
key design criteria. One method is to incorporate shrinkage reduc-
water-to-binder ratio of 0.2 within 4 months of about 130 days.
ing admixture (SRA) into the concrete. It was found that the addi-
tion of SRA leads to a significant reduction in the drying shrinkage
due to the decrease of surface tension in the water at the pores,
which results in lower capillary stresses. It has been reported that
the 28-day concrete shrinkage is reduced by 50–80% and the ulti-
Drying Shrinkage (10-6)

mate shrinkage is reduced by 25–50% [57,58]. However, these

chemical admixtures may have potential negative side effects.
The primary negative side effects include the chance of strength
loss early in the concrete’s life, a slight delay in setting time and
potential interaction with other admixtures such as air entraining
agent or superplasticizer [59].

5.2. Water permeability

Water permeability of RPC with different water-to-binder ratios

Age (Days) and SP dosages are measured at specific days. The result is summa-
rized in Table 7. The typical permeability coefficient of RPC at
Fig. 4. Drying shrinkage for RPC with different SP dosages under various ages. around 98 days is about 0.0005; whereas the typical permeability
 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89 85

coefficient of normal strength concrete at 98 days is about 0.0015

[60], which differs by an order of magnitude. This may be attrib-
uted to the homogenous and dense microstructure of RPC, which
minimizes the porosity by filler effects of crushed quartz and silica
fume that leads to discontinuous and disconnected pores and re-
sults in a very low permeability.

5.2.1. Effect of water-to-binder ratio on water permeability for RPC

The permeability of concrete plays an important role in durabil-
ity as it controls the rate of entry of moisture that may contain
aggressive chemicals and the movement of water during heating
or freezing [11]. The single parameter that has the largest influence
on durability is the water-to-binder ratios, which strongly affect
the capillary porosity of concrete. As the water-to-cement ratio is
decreased, the porosity of the paste is decreased and the concrete
becomes more impermeable, which can increase the concrete
strength and improve the resistance to cracking from the internal
stresses that may be generated by adverse reactions. Fig. 7 shows
the permeability coefficient of RPC with different water-to-binder
(a) RPC shown at lower magnification ratios within three and a half months of about 105 days.
In a hydrated cement paste, the size and continuity of the pores
at any point during the hydration process would control the per-
meability coefficient [21,44]. Therefore, the low permeability of
RPC with low water-to-binder ratios may be explained by the small
and discontinuous pores that present in the very homogenous,
compacted and dense paste of RPC. As hydration proceeds, the cap-
illary network becomes increasingly tortuous as interconnected
pores are blocked by the formation of C–S–H which leads to
continuous decrease in permeability coefficient. Moreover, for
RPC with low water-to-binder ratios, the pozzolanic reaction is
continuous after 28 days and additional C–S–H is formed in the
matrix which occupies all the spaces available, reinforces the bind-
ing of the composite. Therefore, the increase in thickness of C–S–H
between capillaries forms an impermeable region in RPC which
leads to a very low water permeability. On the other hand, the
higher the strength of the hardened cement paste, the lower its
permeability will be because strength is a function of the relative
volume of gel in the space available [24]. Therefore, it can be con-
cluded that the lower the water-to-binder ratios (which can still
reach high workability), the smaller the pores will be, gross volume
of C–S–H gel and crystals will be formed, which disconnect the
(b) Close up of cracks capillary pores, the higher the strength is achieved and thus a low-
er water permeability.
Fig. 5. Scanning electron micrograph of cracks found in RPC at 14 days. Therefore, the high water permeability for high water-to-binder
ratios for RPC can be explained by the higher amount of air void
found for high water-to-binder ratios, which is difficult to achieve
complete discontinuity of capillary pores.

5.2.2. Effect of superplasticizer dosage on water permeability for RPC

Superplasticizers reduce the macroporosity as they make
concrete more flowable and compaction easier. Moreover, they
considerably reduce the capillary porosity by reducing the water-
to-binder ratios [61]. The reduction in porosity causes reduction
in concrete permeability and therefore increases durability.
Though the use of superplasticizers has positive effects on work-
ability and thus the permeability of RPC, they have to be used with
optimum dosages. Overdosage of superplasticizer can lead to
chemical incompatibility problems [62] which may in turn affect
the water permeability of concrete, since the addition of superp-
lasticizer may alter the pore structure of RPC. Fig. 8 shows the
permeability coefficient of RPC with different SP dosages within
three and a half months of about 105 days.
It has been reported that the permeability coefficient decreases
with the increase in SP dosage [63]. However, in this investigation,
it is found that there exists an optimal SP dosage which results in
Fig. 6. Cracks found in RPC with the water-to-binder ratio of 0.4 at 14 days. the lowest permeability. Too low or too high SP dosages cannot
86 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89

0.001569
0.010242
0.001295

0.001646
0.001019
0.000552

0.000552
Day 105

Permeability Coefficient, Kcp

0.00098

(mm2 / s.BAR)
0.012014
0.000862
0.000583
0.001609

0.001301
0.000583
0.001013
0.001064
Day 98

0.001574
0.009896

0.001662
0.000772
0.000551

0.000551
0.001052
0.00153
Day 90

0.001681
0.010131
0.001562

0.001332
0.000862
0.000642

0.000642

0.002805
Age (Days)
Day 84

Fig. 7. Water permeability for RPC with different water-to-binder ratios under
various ages.
0.001156

0.001652
0.008911
0.001651

0.001156
0.000431

0.000431

0.003084
Day 77

Permeability Coefficient, Kcp

0.001128

0.010428

0.001218
0.003424
0.000447
0.001704

0.001704
0.000447
Day 70

(mm2 / s.BAR)
0.001265

0.001792
0.010618
0.001792

0.001333
0.000326

0.000326

0.003702
Day 63

0.011312
0.001915

0.001472
0.004467
0.000811
0.000354

0.000354
0.00169
Day 56

Age (Days)
0.015775
0.001848

0.003362
0.005962
0.000871
0.000586

0.000586
0.00197
Day 42

Fig. 8. Water permeability of RPC with different SP dosages under various ages.

result in low permeability of RPC. From Fig. 8, it is observed that

0.017954
0.001722

0.004623

0.004412
0.006712
0.001808
0.000507

0.000507
Day 28

the addition of 2.5% SP gives the lowest permeability coefficient

of RPC. It is also noted that in general the permeability coefficient
of all RPC mixes significantly drop in the first 28 days (especially
with a SP dosage of 2%) and the curves turn flat afterwards. Com-
0.039256
0.014521
0.001575
0.001411

0.001411
0.007271
0.007271
0.002078
Water permeability (in mm2/s bar)
Day 14

paring with Fig. 7, it is found that the effect of SP dosage is not

as significant as that of water-to-binder ratios because the values
of permeability coefficient of the various SP dosages become closer
at the later age.
0.011265
0.045724
0.003188
0.003912

0.003912

0.009674
0.02065

0.00983

Addition of 2% SP dosage results in drastic drop in water perme-

Day 7

ability of RPC in the first 28 days. This may be attributed to the

insufficient SP dosage which makes compaction difficult and leads
to high level of porosity. It has been reported that at high levels of
0.022139
0.057938
0.033659

0.017199
0.015041
0.009767

0.009767

0.019407

porosity, interconnected pores will predominate under normal

Day 1

conditions [64] which would result in high water permeability

for RPC. These have significantly influence on matrix durability,
providing tortuous channels through which material transport
SP dosage (%)

can occur [64]. Moisture movement and action of chemical species

Results of water permeability for RPC.

from the environment such as sulphate, chloride, and carbon diox-

ide can lead to chemical degradation from within the matrix as
2.5

2.5

3.5
2

well as on the external faces of the concrete.

With sufficient SP added, water content and thus the amount of
Water-to-binder ratio

pores are reduced, and material particles are better dispersed. The
RPC mixture becomes highly workable which results in a very
dense and compacted paste with low porosity. The reduction of
volume porosity not only increases mechanical performance of
concrete but it also tends to change the nature of the pores and
Table 7

0.17

0.23
0.2

0.4
0.2

consequently concrete durability [64]. Therefore, the low perme-

ability of RPC with optimal SP dosage may be due to the lower
 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89 87

volume of porosities, in which the emphasis shifts towards isolated

Compressive Strength, fc (MPa)

pores that are not connected to the outer boundaries of the RPC.

Permeability Coefficient, Kcp

However, viscosity is found to significantly decrease with increas-
ing SP dosage. When excess SP dosage is added, the RPC would be

(mm2 / s.BAR)
associated with large slump loss and the problem of segregation
would occur. More air would be trapped during mixing which
may lead to higher porosity and consequently higher water
permeability.
It is found that the influence of superplasticizers on water per-
0
meability of RPC is more significant at an early age while the influ-
ence is not significant in the long term. It is also noted that the
curve turns flat at the later age with increasing SP dosage, regard-
less of the optimal SP dosage of 2.5% in which the permeability
coefficient remains constant at 28 days. The permeability coeffi- w/b ratio
cient for RPC with SP dosage of 2% is almost constant at 42 days,
whereas for RPC with SP dosages of 3% and 3.5%, it takes 56 days Fig. 10. Relationship between compressive strength and water permeability of RPC
and 90 days respectively for the permeability coefficient to become with different water-to-binder ratios.

constant. This may be due to the delayed initial hydration of ce-

ment when SP dosage is increased which results from excessive
retardation of the setting time. Superpalsticizers would delay the in the volume of gel and hydration products and hence decreases
initial hydration of cement but they hardly influence the hydration the voids in RPC.
reaction at the later stage [65]. It is believed that the delayed It is known that both compressive strength and water perme-
hydration of cement would cause delay for the formation of suffi- ability are largely influenced by water-to-binder ratios. Fig. 10
cient C–S–H or other hydration products to block the capillary shows the relationship between compressive strength and water
pores. Therefore, it takes longer time for the permeability coeffi- permeability of RPC with different water-to-binder ratios.
cient of RPC to become steady when higher SP dosage is From Fig. 10, it is found that the water-to-binder ratio of 0.2
incorporated. gives the maximum compressive strength and minimum water
permeability of RPC. Beyond this optimal water-to-binder ratio,
5.2.3. Relationship between strength and permeability of RPC increasing the water-to-binder ratio decreases the compressive
It has been reported that strength and permeability are inver- strength and increases the water permeability of RPC due to high
sely proportional to the volume of large pores in the hydrated ce- amount of large air voids found in high water content mixture;
ment paste, the considerable pore refinement leads to significant decreasing the water-to-binder ratio also decreases the compres-
increase in impermeability of concrete [66]. The relationship be- sive strength and increases the water permeability of RPC. Since
tween compressive strength and water permeability of RPC under the mixture is too dry and not completely compacted, the unwork-
various ages is shown in Fig. 9. able RPC would contain large percentage of voids and hence low
From Fig. 9, it is observed that water permeability of RPC de- strength with high permeability. Therefore, it can be concluded
creases as compressive strength increases across the age. It is also that RPC mix containing minimum amount of water would develop
noted from Fig. 9 that the longer the curing period, the greater the maximum attainable strength and minimum permeability at any
ultimate compressive strength and the lower the corresponding age, provided that the concrete is fully compacted.
water permeability are obtained. The compressive strength of Fig. 11 shows the relationship between the permeability coeffi-
RPC grows over a lengthy period. The most rapid increase can be cient and corresponding compressive strength for all RPC mixes.
observed during the initial period of hardening up to 28 days and The trend shows that the increase of compressive strength is
then the rate gradually decreases with age. On the other hand, accompanied with a decrease in water permeability of RPC.
the permeability of RPC is adversely affected by its age. The perme- From the above results, a minimum permeability and a high
ability coefficient after 3 days of curing equals about 11 times the strength of RPC can be achieved by: (i) increasing the curing peri-
permeability coefficient after 90 days of curing. This is attributed od; (ii) proper use of the water-to-binder ratio and SP dosage; and
to the completion of hydration process which leads to the increase (iii) good mixing and compaction.
Compressive Strength, fc (MPa)

Permeability Coefficient, Kcp

Permeability Coefficient, Kcp

(mm2 / s.BAR)

(mm2 / s.BAR)

Age (Days) Compressive Strength, fc (MPa)

Fig. 9. Relationship between compressive strength and water permeability for RPC Fig. 11. Relationship between the permeability coefficient and corresponding
under various ages for the water-to-binder ratio 0.4 and SP dosage 2.5%. compressive strength for all RPC mixes.
88 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89

6. Conclusion [3] Dauriac C. Special concrete may give steel stiff competition. The Seattle Daily
Journal of Commerce; 1997. <http://www.djc.com/special/concrete97/
10024304.htm>.
This paper has investigated the drying shrinkage behaviour and [4] Dowd W. Reactive powder concrete: ultra-high performance cement based
water permeability of RPC. Different water-to-binder ratios and SP composite. United States: Construction Innovation Forum; 1999.
[5] Lee NP, Chisholm DH. Study report reactive powder concrete. BRANZ
dosages were experimented. It was found that RPC generally
2005;1(146).
exhibits relatively low dying shrinkage. This may be partly due [6] Ji WY, An MZ, Yan GP, Wang JM. Study on reactive powder concrete used in the
to the reduction of the pore size and connectivity of the voids side walk system of the qinghai–tibet railway bridge. Schmitz Ready Mix Inc.;
2008.
which results from the high packing density of RPC and partly
[7] Yazici H, Yardimci MY, Aydin S, Karabulut AS. Mechanical properties of
due to the prestressing effect of silica fume which slows down reactive powder concrete containing mineral admixtures under different
the water evaporation and restrains the shrinkage. The higher curing regimes. Constr Build Mater 2009;23(3):1223–31.
the water-to-binder ratios, the higher the shrinkage strains of [8] Wong ACL, Childs PA, Berndt R, Macken T, Peng GD, Gowripalan N.
Simultaneous measurement of shrinkage and temperature of reactive
RPC will be. This may be due to the greater disequilibrium in the powder concrete at early-age using fibre Bragg grating sensors. Cement
relative humidity between concrete and the external drying envi- Concr Compos 2007;29(7):490–7.
ronment which leads to higher evaporation and more pores formed [9] Gilbert RI. Time effect in concrete structure. United States: Elsevier Science
Publishing; 1988.
at higher water-to-binder ratio which develops higher surface ten- [10] Barr BIG, Hoseinian SB, Beygi MA. Shrinkage of concrete stored in natural
sion in pores that causes concrete to shrink. Higher SP dosages give environment. Cement Concr Compos 2003;25(1):19–29.
higher drying shrinkage strains and shrinkage rate development. [11] Mindess S, Young F, Darwin D. Concrete. Upper Saddle River, NJ: Prentice Hall;
2003.
This is because high SP dosage would cause changes in the pore [12] Boivin S, Acker P, Rigaud S, Clavaud B. Experimental assessment of chemical
structure which influences the state of the evaporable water and shrinkage of hydrating cement pastes. Concrete Department, Lafarge
consequently the shrinkage behaviour. Microcracks in RPC were Laboratoire Centrel de Recherche, St. Quentin, Fallavier, France; 1999.
[13] Buil M. Contribution a letude du retait de la pate de ciment durcissante.
observed under SEM which is possibly due to the self-desiccating
Rapport de Recherche LPC No. 92; 1979.
action. Incorporation of shrinkage reducing admixture can help re- [14] Tazawa E, Miyazawa S. Autogenous shrinkage of concrete and its importance
duce cracks. in concrete technology. In: Creep and shrinkage of concrete proceedings of the
5th international RILEM symposium, Barcelona, Spain; 1993. p. 159–68.
It was also found that RPC generally exhibits very low water
[15] Zhang MH, Tam CT, Leow MP. Effect of water-to-cementitious materials ratio
permeability, with the permeability coefficient typically lower and silica fume on the autogenous shrinkage of concrete. Cem Concr Res
than that of the normal concrete by an order of magnitude. This 2003;33(10):1687–94.
may be attributed to the homogenous and dense microstructure [16] Brooks JJ, Cabrera JG, Megat Johari MA. Factors affecting the autogenous
shrinkage of silica fume high-strength concrete. In: Proceedings of
of RPC which minimizes the porosity by filler effects of silica fume international workshop on autogenous shrinkage of concrete, Hiroshima,
and crushed quartz that leads to disconnected pores and results in Japan; 1998. p. 195–201.
a very low permeability. Water permeability of RPC decreases as [17] Mak SL, Ritchie D, Taylor A, Diggins R. Temperature effects on early age
autogenous shrinkage in high performance concretes. In: Proceedings of
the water-to-binder ratio decreases. The permeability coefficient international workshop on autogenous shrinkage of concrete, Hiroshima,
lowers by 2 orders of magnitude when water-to-binder ratio de- Japan; 1998. p. 155–65.
creases from 0.4 to 0.2 at 105 days. This may be attributed to the [18] Aitcin PC. Autogenous shrinkage measurement. In: Tazawa Eiichi, editor.
Autogenous shrinkage of concrete. London, United Kingdom; 1999. p.
small pores found with low water-to-binder ratio and gross vol- 257–67.
ume of C–S–H gel and crystals formed which block and disconnect [19] Acker P. Comportment mecanique du beton: apports de I’approache physic-
the capillary pores. The effect of SP dosage is not as significant as chimique. Rapport de Recherche LPC; 1998.
[20] Justnes H, Van Gemert A, Verboven F, Sellevold EJ. Total and external chemical
that of water-to-binder ratio on water permeability of RPC. How-
shrinkage of low w/c ratio cement pastes. Adv Cem Res 1996;8(31):121–6.
ever, an optimal SP dosage can reduce the water content and [21] Neville AM. Properties of concrete. Burnt Mill, Harlow, Essex, New
amount of pores, and help disperse the material particles better York: Longman; 1995.
[22] Powers TC. Structure and physical properties of hardened Portland cement
which results in a very dense and compacted paste with low poros-
paste. J Am Ceram Soc 1958;41(1):1–6.
ity. Insufficient SP dosage would make compaction difficult and [23] Powers TC, Copeland LE, Mann HM. Capillary continuity or discontinuity in
lead to high level of porosity and interconnected pores predomi- cement pastes. J Portland Cem Assoc Res Dev Lab 1959;1(2):38–48.
nate. Excessive SP dosage can lead to chemical incompatibility [24] Powers TC, Copeland LE, Hayes JC, Mann HM. Permeability of Portland cement
paste. J Am Concr Inst 1954;51(3):285–98.
problems and segregation may occur, resulting in higher porosity [25] Whiting D. Permeability of selected concretes. ACI Mater J 1988;108:195–222.
and consequently higher water permeability. Influence of SP on [26] BS 12. Specification of Portland cement. London (United Kingdom): British
water permeability of RPC is more significant at the early age than Standards Institution; 1996.
[27] BS EN 197-1. Cement: composition, specifications and conformity criteria for
in the long term. This is because superpalsticizers would delay the low heat common cements. London (United Kingdom): British Standards
initial hydration of cement which delays the formation of sufficient Institution; 2000.
C–S–H or other hydration products to block the capillary pores; but [28] ASTM C1240. Standard specification for silica fume used in cementitious
mixtures. United States: American Society for Testing and Materials
they hardly influence the hydration reaction at the later stage. International; 2005.
There is an inverse relationship between strength and permeability [29] BS 5075-3. Concrete admixtures: specification for superplasticizing
of RPC. It was found that water permeability of RPC decreases as admixtures. London (United Kingdom): British Standard Institution; 1985.
[30] Cheyrezy M, Maret V, Frouin L. Microstructural analysis of reactive powder
compressive strength increases. concrete. Cem Concr Res 1995;25(7):1491–500.
[31] Richard P, Cheyrezy M. Composition of reactive powder concretes. Cem Concr
Res 1995;25(7):1501–11.
Acknowledgement [32] Shaleen E, Shrive NG. Optimization of mechanical properties and durability of
reactive powder concrete. ACI Mater J 2006;103(6):444–51.
[33] Washer G, Fuchs P, Graybeal BA, Hartmann JL. Ultrasonic testing of reactive
The work described in this paper was fully supported by a stra- powder concrete. IEEE Trans Ultrason Ferroelectr Freq Control 2004;51(2):
tegic research grant from the City University of Hong Kong. 193–201.
[34] Geiker M, Ekstrand JP, Hansen R. Effect of mixing on properties of SCC. In:
Proceedings of the fifth international RILEM symposium. Paris, France; 2007. p.
References 231–8.
[35] Bonneau O, Lachemi M, Dallaire E, Dugat J, Aitcin PC. Mechanical properties
and durability of two industrial reactive powder concretes. ACI Mater J
[1] Rebentrost M, Cavill B. Reactive powder concrete bridges. In: Austroads 6th
1997;94(4):286–90.
bridge conference: bridging the gap. Perth, Western Australia, Australia; 2006.
[36] Feylessoufi A, Cohen Tenoudji F, Morin V, Richard P. Early ages shrinkage
p. 1–11.
mechanisms of ultra-high-performance cement-based materials. Cem Concr
[2] Blais PY, Couture M. Precast, prestressed pedestrian bridge – world’s first
Res 2001;31(11):1573–9.
reactive powder concrete structure. PCI J 1999;44(55):60–71.
 C.M. Tam et al. / Construction and Building Materials 26 (2012) 79–89 89

[37] Morin V, Choen-Tenoudji F, Feylessoufi A, Richard P. Evolution of the capillary [54] Shah SP, Karaguler ME, Sarigaphuti M. Effects of shrinkage reducing
network in a reactive powder concrete during hydration process. Cem Concr admixture on restrained shrinkage cracking of concrete. ACI Mater J
Res 2002;32(12):1907–14. 1992;89(3):88–90.
[38] Chan YW, Chu SH. Effect of silica fume on steel fiber bond characteristics in [55] Ribeiro MG, Fiho RP. Risk assessment of chemicals in foundries: the
reactive powder concrete. Cem Concr Res 2004;34(7):1167–72. international chemical toolkit pilot-project. J Hazard Mater 2006;136(3):
[39] ASTM C490/C490M. Standard practice for use of apparatus for the 432–7.
determination of length change of hardened cement paste, mortar and [56] Bentz DP, Jensen OM. Mitigation strategies for autogenous shrinkage cracking.
concrete. United States: American Society for Testing and Materials; 2008. Cement Concr Compos 2004;26(6):677–85.
[40] GWT 4000. Instruction and maintenance manual. Germany: GWA [57] Engineering Bulletin. Eclipse shrinkage reducing admixture. Grace
International Limited; 2002. Construction Products; 2002.
[41] Feldman RF, Hauang CY. Properties of Portland cement–silica-fume pastes I: [58] Engineering Bulletin. Eclipse plus shrinkage reducing admixture. Grace
porosity and surface properties. Cem Concr Res 1985;15(5):765–74. Construction Products; 2004.
[42] Feldman RF, Hauang CY. Properties of Portland cement–silica-fume pastes II: [59] Rodden RA, Lange DA. Feasibility of shrinkage reducing admixtures for
mechanical properties. Cem Concr Res 1985;15(6):943–52. concrete runway pavements. Urbana Champaign, United States: Center of
[43] Hewlett PC. Lea’s chemistry of cement and concrete. London: Arnold; 1998. Excellence for Airport Technology, University of Illinois; 2004.
[44] Mehta PK, Monteiro JM. Concrete: structure, properties, and materials; 1993. [60] Tam WYV. Recycled aggregate from concrete waste for higher grades of
[45] Balogh A. New admixture combats concrete shrinkage. Concr Constr concrete construction. Hong Kong, China: Department of Building and
1996;41(7):546–51. Construction, City University of Hong Kong; 2005.
[46] Tongaroonsri S, Tangtermsirikul S. Effect of mineral admixtures and curing [61] Monosi S, Collepardi M. Low porosity and early durability of concrete.
periods on shrinkage and cracking age under restrained condition. Constr In: Proceedings of the 1st international congress held by RILEM, in pore
Build Mater 2009;23(2):1050–6. structure and construction materials properties. Versailles, France; 1987.
[47] Altoubat SA, Lange DA. Tensile basic creep: measurements and behaviour at p. 1–8.
early age. ACI Mater J 2001;98(5):386–93. [62] Kwan AKH. Development of high performance concrete for Hong Kong.
[48] Holt E. Contribution of mixture design to chemical and autogenous shrinkage Materials science and technology in engineering conference: now, new and
of concrete at early ages. Cem Concr Res 2005;35(3):464–72. next. Hong Kong; 2003. p. 3.
[49] Holt E, Leivo M. Cracking risks associated with early age shrinkage. Cement [63] Anwar M, Soghair H, Ahmed M, Fujii M. New aspects affecting the mutual
Concr Compos 2004;26(5):521–30. relation between strength and permeability of concrete. J Ann Thesis Rep
[50] Esping O. Early age properties of self-compacting concrete: effects of fine 1993;15(1):1229–34.
aggregate and limestone filler. Sweden: Chalmers University of Technology; [64] Geslin NM, Israel D, Lachowshi EE, Macphee DE. Durability and microstructure
2007. of pore reduced cements. In: Symposium proceedings of materials research
[51] Attiogbe EK, Nmai CK, Gay FT. Air-void system parameters and freeze–thaw society. San Francisco, California, United States; 1995. p. 237–44.
durability of concrete containing superplasticizers. Concr Int 1992;14(7): [65] Sakai E, Kasuga T, Sugiyama T, Asaga K, Daimon M. Influence of
57–61. superplasticizers on the hydration of cement and the pore structure of
[52] Lane RO, Best JF. Laboratory studies on the effects of superplasticizers on the hardened cement. Cem Concr Res 2006;36(11):2049–53.
engineering properties of plain and fly ash concrete. PISSC 1978;1:295–323. [66] Chan RWM, Ho PNL, Chan EPW. Report on concrete admixtures for
[53] Qi M, Li ZJ, Ma BG. Shrinkage and cracking bahaviour of high performance waterproofing construction. Architectural Services Department, Hong Kong
concretes containing chemical admixtures. J Zhejiang Univ Sci 2001;3(2): Special Administration Region; 1999.
188–93.