Construction and Building Materials 24 (2010) 2664–2671

Contents lists available at ScienceDirect

Construction and Building Materials

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

Optimum mix design of enhanced permeable concrete – An

experimental investigation
C. Lian a,*, Y. Zhuge b
a
School of Natural and Built Environments, University of South Australia, Adelaide, South Australia, Australia
b
Faculty of Engineering and Surveying, University of Southern Queensland, Brisbane, Queensland, Australia

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

Article history: Permeable pavement, due to its high porosity and permeability, is considered as an alternative to tradi-
Received 27 May 2009 tional impervious hard pavements for controlling stormwater in an economical and friendly environmen-
Received in revised form 23 March 2010 tal way. Permeable concrete normally made of single-sized aggregate bound together by Portland
Accepted 2 April 2010
cement, using restrictedly as a pavement material, because of its insufficient structural strength. Aimed
Available online 23 May 2010
at developing a new type of permeable concrete with enhanced structural strength, various mix designs
were attempted and their effects on the compressive strength and permeability of permeable concrete
Keywords:
were investigated in this research. The optimum aggregate and mix components design were conse-
Aggregate
Compressive strength
quently recommended for enhanced permeable concrete.
Failure Ó 2010 Elsevier Ltd. All rights reserved.
Porous concrete
Permeability

1. Introduction cerns about their long-term performance. Consequently there has

1.1. Permeable pavement
Presently natural resources are increasingly consumed due to
rapid urbanization and thereafter human construction activities,
so that various strategies are being investigated by engineers to
protect and restore natural ecosystems in the world. Permeable
(porous/pervious) pavement is termed as comprising materials
that facilitate stormwater infiltrate and transfer to the underlying
subsoil [1]. With sub-structure which stores water underground
temporarily, it is called permeable pavement system. Instead of
installing rainfall detention ponds or soakaways, this new system
is more cost effective compared to the traditional impervious pave-
ment. Meanwhile, it has been acknowledged by many researchers
that permeable pavement system is capable of reducing the
sediments and contaminants for lessening the pollutant loads on
stormwater, thus it is considered as an economic and environmen-
tal-friendly construction as a part of city drainage system.
Ferguson identifies nine categories of permeable pavement [2]:
porous aggregate, porous turf, plastic geocells, open-jointed paving
blocks, open-celled paving grids, porous concrete, porous asphalt,
soft paving materials, and decks. Because of their many environ-
mental benefits there has been an increase in the use of permeable
pavements over the past few years. However, there are still con-
* Corresponding author. Tel.: +61 8 83025108; fax: +61 8 83025721.
E-mail address: Chunqi.Lian@unisa.edu.au (C. Lian).
been considerable research on the durability and effectiveness of
these technologies [3].
In Australia, permeable pavement has been utilized as a poten-
tial tool of water sensitive urban design (WSUD) to manage natural
water. From 1994 the University of New South Wales (UNSW)
started to research into permeable concrete paving and more re-
cently the University of South Australia (UniSA) is also involved.
However, the previous studies conducted both in UNSW and UniSA
mainly concentrated on water quality and pollution control
through permeable pavements and, only the properties of base
course materials in permeable pavement system and segmental
paving have been studied. There is still a gap of optimizing the sur-
face materials for permeable pavements.
1.2. Permeable concrete pavement
Permeable concrete is relatively porous, providing by the omis-
sion of fine aggregates and filled most of volume with coarse
aggregate [4], thus, porous concrete obtains more voids in the
structure leading to higher water infiltration and air exchange
rates compare to conventional concrete [5,6], but the structural
strength of it is compromised. As Larrard and Belloc pointed out
[7], the strength of concrete was indeed determined by the proper-
ties of mortar, coarse aggregate and the interface. Since cement
paste in permeable concrete is very thin to bond coarse aggregate
together, porous concrete tends to fail at the binder interface be-
tween the aggregates and results in the low compressive strength

0950-0618/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2010.04.057
 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671
[8]. Therefore, currently the permeable concrete pavement is only
capable of tolerating the light traffic loadings, with the biggest use
in carparks, footpaths and bicycle trails [2]. This study aims to im-
prove the compressive strength of porous concrete without losing
permeability so that it could be adoptable for supporting higher
traffic volume. As it is noticed that not only the size of aggregate,
but also the gradation and amount of aggregate will affect the com-
pressive strength and static modulus of elasticity on porous port-
land cement concrete [9], this research will firstly investigate the
effect of various types of aggregate to establish the best local re-
source and then proceed to the design of optimal mix with various
additives.
2. Experimental study
2.1. Materials
At the first stage, three different kinds of coarse aggregate were
used without fine aggregate and other admixtures. Sands and silica
fume were applied to enhance the strength of porous concrete at
the second stage based on the results of stage one.
2.1.1. Cement
The ordinary Portland cement from local supplier was used in
each mix, comply with the specification given in AS (Australian
Standard) 3972-1997.
2.1.2. Aggregate
Coarse aggregate was used as a primary ingredient in making
the permeable concrete. According to Krezel [10], crushed igneous
rocks are more preferable as coarse aggregate for concrete due to
their higher strength. However, since the availability of igneous
rock in Australia is becoming scarce [10], this research diverted
to the crushed sedimentary and metamorphic rocks. Three types
of coarse aggregate were obtained from local quarry: quartzite,
dolomite and limestone. Dolomite was a sedimentary carbonate
rock, composed of the mineral dolomite, also contained impurities
such as calcite, quartz and feldspar. Dolomite formed in groups of
rhombohedral crystals with curved, saddle-like faces. Limestone
was also sedimentary rock. Although some limestones were nearly
pure calcite, there were often varying amounts of clay, silt and
sand. Quartzite was a dense, hard metamorphic rock. The quartz-
ites obtained from local quarry were red due to a large amount
of iron oxide. In order to explore the optimum aggregate for mak-
ing porous concrete, these three types of coarse aggregate were
investigated and compared at the first stage. The geological and
mechanical properties of aggregate were tested and the results
were given in Table 1.
In addition, considering the smaller size aggregate will result in
the increase of the specific surface and the binding area between
cement and aggregate, which is beneficial to the strength
and durability of concrete, fine aggregate was used at the second
stage.
Table 1
Engineering properties of aggregates.
Aggregate Flakiness Mean water Los Angeles Dry strength
Index (%) absorption (%) Abrasion value (%) (KN)
Type A 21 2.8 27 163
Type B 35 0.8 15 225
Type C 15 0.3 38 74
Type A: quartzite; type B: dolomite; type C: limestone.
2665
2.1.3. Admixtures
The results of previous research [11–13] indicated that mineral
additives could lead to the improvement of concrete properties
such as mechanical strength and concrete durability, since the
mineral composite reduced the thickness of the interfacial transi-
tion zone (ITZ) between the aggregate and the cement matrix.
For example, the results given by [14,15] showed the ITZ thickness
of concrete was less than 10 lm with silica fume compared to
50 lm without silica fume. Therefore, silica fume, namely Microsil-
ica 920-u, was tried to seek adequate strength of porous concrete
at the second stage of testing.
Besides, a new generation superplasticiser was incorporated as
the chemical intensifier in this study. It is based on a unique car-
boxylic ether polymer with long lateral chains, which greatly im-
proves the cement dispersion. It is called hyperplasticiser.
2.2. Samples preparation
2.2.1. Sieving
The preparation of standard concrete test specimens is based on
Australian Standards and Guidelines. All of the raw 10 mm aggre-
gates from quarries were sieved and separated into different
groups using standard sieves. Specific gradations were then ob-
tained by recombining small fractions of separated aggregates.
2.2.2. Washing
The coarse aggregate sourced from quarry usually contains a
high amount of adhere clay and other impurities. Before the mix-
ing, washing is a necessary procedure, the aggregate surface coat-
ings interfere with the development of bond between aggregate
and cement paste, which plays an important role in providing sat-
isfactory strength of porous concrete. Hence, the coarse aggregates
were washed using tap water and dried in oven for one day to re-
move the silt or crusher dust.
2.2.3. Mixing
The mixing procedure complied with Australia Standard:
AS1012.2-1994.
2.2.4. Casting and compaction
The compaction method for making porous concrete is one of
the most influential factors in the sample preparation. Two com-
paction methods have been assessed in our previous research
[16], one was using compaction hammer and the other was using
vibration table. Although the hammer compaction packed the
aggregate particles together more tightly, the density of porous
concrete samples increased with the loss of permeability. As the
impaction strength of a falling hammer was so strong to crush
the weak aggregate and create weak layers, the vibration method
seemed to be more suitable for majority of aggregates, such as
limestone and dolomite. However, for the sake of achieving the
maximum cohesion between aggregate particles, a combined com-
paction method was attempted, that was, not only applied the
standard rodding compaction method, but also incorporated a sta-
tic compactor in the consequent vibrating procedure. This compac-
tion effort allowed most of the coarse aggregate not to deform
under compaction whilst increase the contact surface and align-
ment of aggregate particles, which was believed a substantial as-
pect to increase the strength of porous concrete.
2.3. Testing procedures
The cast cylinders were demoulded after 24 h, labeled and
weighted for various testing. Then the samples were cured in a
lime bath at 23 ± 2 °C, according to AS1012.8.1-2000. For each
batch, two samples were prepared for permeability testing and
2666 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671
others were for compression, three tested at 7 days and 28 days,
respectively. The results showing up below were all average
values.
2.3.1. Uniaxial compressive strength and flexural strength
The unconfined compressive strengths (UCS) testing of concrete
specimens were carried out in the lab according to AS1012.9-1999.
Prior to loading process, caps were placed on the ends of samples.
Type of capping used depended on surface condition of the con-
crete samples. Rubber capping was usually used for conventional
concrete with smooth top and bottom surface; and sulphur cap-
ping was used for samples with rough surface like porous concrete.
The study by Harber [17] showed that sulphur capping eliminated
the problem of tilting and failing which was caused by the dislodg-
ing edge aggregate. It also concluded that the compressive strength
of the porous concrete would increase dramatically through the
use of the sulphur capping, as this capping restrained the aggre-
gates on the top effectively (see Figs. 1 and 2). Thus, sulphur cap-
ping was adopted for all samples in this study.
The flexural strength testing of concrete specimens were con-
ducted in the lab according to AS1012.11-2000. The samples were
prepared in the 100  100  470 mm mould. The moduli of rup-
ture of permeable concrete made with different aggregates were
determined.
2.3.2. Water permeability
Permeability as a unique ability for water to penetrate through
porous concrete was expressed in millimetres per second (mm/s).
Since porous concrete generally owns a much higher permeability
compared to the normal dense concrete, the permeability test
method for the latter one was not suitable for testing porous con-
crete. As there is no Australian Standards for such testing, a testing
method which was similar to the falling head test method for soil
(AS1289.6.7.2-2001) was adopted in this research.
The testing apparatus has been gradually improved from our
previous research [16]. Instead of using a rigid perspex tube as pre-
vious testing, the cylindrical plastic pipe was used in this test. With
inline steel wire and adjustable steel tie, the pipe was tight to in-
hibit water leakage along the sides of the sample (see Fig. 2). More-
over, the tiny gap between the specimen and the pipe at the
bottom was sealed with processed plasticene to prevent water
infiltration through the edge of pipe, which will affect the accuracy
of the permeability coefficient. Subsequently, the water permeabil-
ity rate of porous concrete was calculated by Eq. (1).
aL h1
K¼  ln
At h2
Fig. 1. Uniaxial compressive strength testing rig.
Fig. 2. Permeability testing rig.
where k is the permeability coefficient (mm/s), a is the area of the
cylindrical pipe (mm2), A is the area of specimen (mm2), L is the
length of specimen (mm), t is the time for water to pass from level
h1 to h2 (s) through the pipe.
2.3.3. Porosity
The porosity test was carried out at 28 day of age. The open
porosity was measured as the percentage of pore volume or void
space within the concrete that can contain water. The sample
was oven dried at 110 °C firstly and was left to cool for measure-
ment. The dimensions of the sample were measured in dry condi-
tion and the total volume of sample (VT) including the solid and
void component was determined. Then the sample was sunk into
a bucket filled with sufficient water to cover the whole sample
and the water level was marked. After 24 h, the sample was moved
out from the bucket and the water was refilled up to the marked
level. The weight of water added was read by the scale and the
magnitude of this reading was equal to the changed volume (VC),
using the concept of 1 g = 1cm3 for water. The open porosity of
the concrete sample was calculated with Eq. (2):
VT  VC
Pð%Þ ¼  100% ð2Þ
VT
Where P is the open porosity (%), VT is the total volume of spec-
imen (mm3), ‘‘VT  VC” is the volume of void space (mm3).
ð1Þ
3. Test results and discussions
3.1. Aggregate investigation
The effect of coarse aggregate on the compressive strength and
flexural strength of porous concrete was investigated. Three types
of aggregate, namely quartzite, dolomite and limestone were used
plus the size and gradation of aggregates were restricted as shown
in Table 2. The testing results on compressive and flexural strength
at 7 and 28 days are illustrated in Tables 3 and 4.
3.1.1. Effect of aggregate properties
As it is shown in Table 3, for single-sized aggregate (type 1),
Dolomite B1 yielded the highest compressive strength at both 7
and 28 days, followed by limestone, quartzite achieved the lowest
strength. In Table 4 it is also showed that the samples made with
dolomite had a better flexural strength. This indicated that the type
of coarse aggregate affects the strength of porous concrete even
though the aggregates were in the same size and gradation. This
may be attributed to the difference of dry strength, particle shapes
 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671 2667

Table 2 Table 5
Aggregate size distribution. Permeability of porous concrete made with different aggregates at 28 day curing
time.
Sieve size (mm) 16 13.2 9.5 6.7 4.75 2.36 1.18
Mix number Passing percentage by mass (%) Permeability (mm/s)

Type A Quartzite Dolomite Limestone

A1 100 100 100 0 0 0 0
A1 A2 B1 B2 B3 C1 C2
A2 100 100 100 30 0 0 0
A3 100 100 90 30 0 0 0 27.47 13.67 19.87 8.51 14.78 13.27 15.99

Type B
B1 100 100 100 0 0 0 0
B2 100 100 100 30 0 0 0
B3 100 100 90 30 0 0 0
was widen. It can be seen from Table 3 that quartzite and dolomite
concrete behaved similarly. Take dolomite as an example, when
Type C
C1 100 100 100 0 0 0 0
changing from a single sized grading (B1) to a grading varying from
C2 100 100 100 30 0 0 0 9.5 mm to 4.75 mm (B2), the compressive strength of porous con-
C3 100 100 90 30 0 0 0 crete increased from 15.0 MPa to 16.0 MPa at 7 days and from
15.8 MPa to 19.0 MPa at 28 days. The immerged proportion of
smaller size aggregate produced higher strength of porous con-
crete. But when larger sized aggregate was mixed (B3 vs. B2),
Table 3
although it showed a better gradation, the compressive strength
Compressive strength using different aggregates at different age.
and flexural strength of porous concrete were both decreased.
Curing time (days) Compressive strength (MPa) Besides, the permeability measurement was also conducted
Quartzite Dolomite Limestone after 28 days curing time. The permeability coefficients of porous
A1 A2 B1 B2 B3 C1 C2 concrete specimens were given in Table 5. It can be found that
the smaller aggregate size will lead to a lower permeability of por-
7 11.6 13.0 15.0 16.0 14.3 14.3 13.5
28 11.8 15.5 15.8 19.0 15.5 15.5 14.0 ous concrete except for that made with limestone. However, three
types of aggregates all showed a good permeability so some filler
materials could be used to further enhance the strength of porous
concrete based on this gradation.
Table 4
Flexural strength using different aggregates at different age. 3.2. Optimal mix
Curing time (days) Flexural strength (MPa)
The second stage of this research involved using chemical addi-
Quartzite Dolomite Limestone
tives and fine aggregates to improve the strength of porous con-
A3 B2 B3 C3 crete. The compressive strength will be taken as main criterion.
7 1.5 2.9 1.7 1.5 Dolomite was collected as coarse aggregate based on the testing re-
28 1.6 3.0 1.9 1.5 sults at stage one.

and textures of aggregate itself. In light of the highest dry strength
among three types, dolomite produced the highest compressive
strength. However, limestone, which had the lowest dry strength,
did not produce the lowest compressive strength. This was possi-
bly caused by two factors. Firstly, the shape of aggregate was
judged according to AS1141.15-1999 rather than only by vision.
The flakiness index of aggregate was examined to distinguish dif-
ferent shapes. The result (see Table 1) presented limestone was
more rounded than quartzite, the flaky quartzite particles were
more likely to be oriented in one plane under compaction force,
not handling the loaded strength identically in three dimensions.
Therefore, it was brittle to resist higher compressive strength. Sec-
ondly, as shown in Table 1, quartzite particles absorbed more
water compared to limestone in mixtures, which would make
the cement paste around it less viscous to develop as high adhesive
strength as around limestone. Thus, the quartzite showed the
worst compressive strength rather than limestone. But comparing
with dolomite, although the flaky index of dolomite was also high-
er than that of limestone, the advantage arouse by dry strength
could not be cancelled out by this drawbacks. In addition, its water
absorption rate was not as high as quartzite. It still can yield the
biggest compressive strength for porous concrete. Hence, dolomite
would be regarded as the best aggregate for making permeable
concrete.
3.1.2. Effect of aggregate size and gradation
Firstly, single-sized aggregate was used for all three kinds of
aggregate (A1, B1, C1) as reference, and then aggregate gradation
3.2.1. Admixture investigation
Table 6 shows that samples made with additives (B4, B5, B6 and
B7) exhibited higher strengths than the one without (B2). Silica
fume exerted positive influence on compressive strength of porous
concrete as it functioned on normal concrete. Technically speaking,
when the silica fume is added, more water is demanded for wetting
the large specific surface area of silica fume particles in a concrete
mixture to keep its workability. Thus, if the same water/cement ra-
tio was used for samples with and without silica fume, the one
with silica fume normally experienced problem. As it was observed
during the testing, some silica fume particles concentrated over a
small region where the sediment and segregation were easily seen.
Therefore, the benefit of using silica fume was not achieved with-
out other chemical admixtures.
Through a series of trial and error exercises, it was found that by
adding a small amount of superplasticiser to the mixtures contain-
ing silica fume, both the workmanship and the compressive
strength of the samples were improved extensively. This was pro-
ven by the delicate change of slump of fresh porous concrete. As
shown in Table 6, the compressive strength of B4 (10% of silica
fume only) was just slightly higher than that of B2. However, using
additional 0.8% superplasticiser, sample B6 performed much better
without losing the permeability.
The reason is that the inclusion of silica fume will intensify the
flocculation of cement paste and thereby give rise to more floccu-
lating colloids [18], which is porous and will go against the
strength increase of porous concrete unexpected. However, since
superplasticizer normally has long molecules of high molecular
mass, these long molecules will wrap themselves around the ce-
2668 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671

Table 6
Properties of porous concrete made with and without additives.

Batch No. Aggregate Water to Fine aggregate SF (%) SP (%) Density 7 days compressive 28 day compressive Porosity Permeability
size cement ratio (kg/m3) strength (Mpa) strength (Mpa) (%) (mm/s)
B2 9.5–4.75 0.36 0 0 0 1926 16.0 19.0 16.6 8.51
B4 9.5–4.75 0.36 0 10 0 2012 17.0 22.0 13.2 6.13
B5 9.5–4.75 0.28 0 7 0.8 2079 22.0 24.3 16.0 12.64
B6 9.5–4.75 0.32 S1 7 0.8 2140 28.5 30.0 9.0 5.39
B7 9.5–4.75 0.36 S2 7 0.8 2248 25.8 33.2 7.50 3.98

S1: fine dolomite particles; S2: quarry sand.

ment particles, resulting in the deflocculation of cement paste and

the dispersion of silica fume [19]. Then the fine particles of silica
fume is capable of being distributed uniformly and then fill in
the capillary or gel pores of cement paste [20] so that the compres-
sive strength of porous concrete could be distinctly improved, such
as B5.
Furthermore, with assistance of silica fume and superplasticizer
simultaneously, fine aggregate could be utilized to achieve a higher
strength. As shown in Table 6, both B6 and B7 achieved higher
compressive strength than B5. In the mean time, after 28 days cur-
ing, B7 with quarry sand performed better than B6, which is made
with fine dolomite particles of 4.75–2.36 mm. The quarry sand
could promote the development of cement hydration product,
which would reduce the capillary pores in cement matrix during
the 28 days curing and then achieved a dense microstructure,
Fig. 3. Relationship between w/c ratio and compressive strength.
showing a higher compressive strength. In contrast, the smaller
sized dolomite particles could not bridge the crystallized hydrated
cement to form more paste to increase the bonding strength.
pressive strength of 20.3 MPa and 23.3 MPa at 7 days and 28 days,
Therefore, the use of quarry sand was more effective than that of
respectively.
fine dolomite particles. As is shown, the strength was increased
Meanwhile, the change of permeability based on different w/c
and the permeability was maintained as an acceptable level in
ratio was given in Fig. 4. It can be found that the trend of perme-
mix B7. So this combination will be adopted.
ability inverse to that of compressive strength for porous concrete.
The minimum point was also taken place at w/c ratio of 0.34 (B8),
where the permeability is down to 1.22 mm/s. Taking the value of
3.2.2. Optimum water to cement ratio 0.34 as a threshold, once the amount of water overran this thresh-
Taking into account the significant influence of water propor-
tion on the properties of concrete, the water content was adjusted
gently to explore the optimum mix design of porous concrete. The
variations of compressive strength with respect to water/cement
ratio are shown in Table 7. The water to cement (w/c) ratio ranging
from 0.30 to 0.38 was used while other compositions were kept
identical.
It can be clearly seen from Fig. 3 that the mix B8 with water to
cement ratio of 0.34 yielded the highest compressive strength after
both 7 and 28 days curing. It reached 36.8 Mpa and 46.2 Mpa,
respectively. In addition, the relationship between w/c ratio and
compressive strength of porous concrete did not go through a sim-
ple linear path. The presence of the turning point at 0.34 (B8) di-
vided the whole trendline into two different stages. When the w/
c ratio was less than 0.34, the compressive strength was slightly in-
creased along with the increase of w/c ratio; when it was more
than 0.34, the compressive strength steeply declined. The mix
B11 with water to cement ratio of 0.38 produced the lowest com- Fig. 4. Relationship between w/c ratio and permeability coefficient.

Table 7
Properties of porous concrete made with different w/c ratio.

Batch No. Aggregate Water to Sand (%) SF (%) SP (%) Density 7 days compressive 28 day compressive Porosity (%) Permeability
size cement ratio (kg/m3) strength (Mpa) strength (Mpa) (mm/s)
B10 9.5–4.75 0.30 18 7 0.8 2266 35.2 40.3 8.0 1.61
B9 9.5–4.75 0.32 18 7 0.8 2243 36.8 40.5 6.5 2.42
B8 9.5–4.75 0.34 18 7 0.8 2325 36.8 46.2 3.5 1.22
B7 9.5–4.75 0.36 18 7 0.8 2248 25.8 33.2 7.5 3.98
B11 9.5–4.75 0.38 18 7 0.8 2092 20.3 23.3 16.5 8.42
 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671 2669

old, the permeability of porous concrete went up straight, reaching

8.42 mm/s by B11; while when it was below this threshold, the
permeability did not bounce back. As shown in Fig. 4, the perme-
ability of B9 was higher than that of B8, but B10 had a lower one
than B9, It was observed in the mixing of B10, when the water con-
tent reduced to a ratio of 0.30, the sand and silica fume particles
were very hard to combine with cement uniformly. After vibration,
some of these tiny particles within fresh mixture were even sepa-
rated to the surface of sample and thereafter different dense layer
was prone to form in the sample. More pores were probably gener-
ated in the middle part of the sample, but the surface had been
sealed with these particles, affecting the permeable coefficient ad-
versely. Hence, the water to cement ratio less than 0.30 is not
recommended.
Correlated with the variations of compressive strength for por-
Without admixtures With admixtures
ous concrete, it can be concluded that in terms of water to cement
ratio, relative higher permeability could be obtained in range of Fig. 5. Comparison of the failure mode between porous concrete samples made
0.34–0.38, but the large amount of water would intensify the with and without admixtures.
shrinkage of porous concrete during the curing days, resulting in
a great percentage of pore voids in the hardened concrete and then
the compressive strength of porous concrete will be decreased tre-
mendously. In contrast, higher strength could be gained in range of
0.30–0.34, but the water content is too low to hydrate sufficient ce-
ment. In this case, the cement mortar could not wrap all of the
aggregates. However, due to these changes that occurred gradually
during the adjustment, optimum water ratio could be collected. As
the acceptable flow rates for water through pervious concrete are
typically from 2 mm/s to 5.4 mm/s [2,21,22], B9 with w/c ratio of
0.32 seemed to perform best as shown in Figs. 3 and 4, reaching
a high compressive strength and a satisfied permeability as well.

3.3. Failure analysis

Concrete as a three phase composite material at a microscopic

scale included mortar matrix, aggregate and the interfacial transi-
tion zone (ITZ) between the two. Although the interfacial transition
zone was smaller in proportion compared to mortar matrix and
aggregate, its characters influenced the mechanical behavior of
concrete significantly and it was normally regarded as the weakest
link in concrete [23,24]. For normal concrete, Zaitsev [25] pointed
out the separation crack occurred first due to the shrinkage of ce-
ment matrix and then along the interface of the aggregate and ce-
ment paste. But for porous concrete, more fractures developed in
the interfacial zone, it certified that without filler material, such
as sand and chemical admixture, the bond strength of aggregate
and cement in porous concrete was not adequate and thereafter Fig. 6. Fracture surface of the enhanced porous concrete.
it became a controlling factor of determining the compressive
strength.
It was observed from Fig. 5 that the majority of failures inten- were void pores. In the graph with additives (Fig. 7a), black bands
sively took place in the interface between cement and aggregate surrounding each yellow particle was darker, more even and con-
or the hardened cement paste of porous concrete without admix- tinuous when comparing to those without additives (Fig. 7b). It
tures involved. In comparison, the application of silica fume and presented that the additives-cooperated mortar wrapping the
superplasticizer indeed provided a good bond for cement covering aggregate particles was thicker and wider than that of plain ce-
the aggregate individuals. It can be seen from Fig. 5b that more ment paste, which meant the bonding area between two aggregate
fractures occurred right through the aggregate particles rather particles had been enlarged. Neville [26] reported the mechanism
than through the shell of aggregates. In addition, the broken sur- of this phenomenon for normal concrete. Without mineral addi-
face was smoother and trimmer as what usually happened in nor- tion, ITZ normally formed a wall effect which restrains the cement
mal concrete (see Fig. 6). This indicated that the aggregate or the particles from approaching the aggregate surface; but when silica
cement paste failure had taken place of failure from the interface fume is added, the wall effect will be broken, as the micro particles
zone between them, dominating the fracture development under of the silica fume possibly get close to the aggregate particles,
compression. In order to further reveal the failure mechanism of which in turns lower the porosity of ITZ and strengthened the
the interfacial zone, the microstructure of porous concrete was bonding force.
examined with a polarizing optical microscope (Olympus BH2- Furthermore, in Fig. 7b, it can be seen that some colour around
UMA, Japan). The typical micrographs were shown in Fig. 7. The yellow particles was gradually changed from cyan to black; and
dolomite aggregates were seen as the yellow objectives and the ce- some cyan void space directly bordered on yellow surfaces. This
ment paste were the black bands around; the blank or cyan spaces could be the failure development between aggregate and cement
2670 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671
(a) With additives
(b) Without additives
Fig. 7. Polarizing optical micrographs of the cut surfaces of porous concrete made
with and without additives.
paste without admixtures. It illustrated that a few of cement pastes
were pulled away from the surface of aggregates then left void
spaces between them; and a few did not cover the whole aggregate
surfaces or even not glued to the surface of aggregates directly.
This induced the weakest zone of porous concrete. In contrast,
Fig. 7a shows that the cement pastes were tightly sticked to the
aggregate surfaces, the bonding zone was integrated and coherent;
the cracks occurred inside the yellow aggregate individuals. This
demonstrated that the possible failure mode has been transferred
from the ITZ to the aggregate or the cement paste inside. Therefore,
the strength of porous concrete is obviously enhanced.
4. Conclusions
This paper presents the results of experimental investigations
on the influential factors on the strengths and permeability of por-
ous concrete. As two important characteristics, the strength is in-
versely proportional to the permeability, a series of experiments
have been conducted for exploring the optimum mix design. The
findings obtained through this study can be concluded as follows:
(1) The effect of three most common types of aggregate from
Australian local quarries was investigated and compared
via laboratory testing. In addition to achieve the higher com-
pressive strength, dolomite was also suggested more resis-
tant to abrasion for porous concrete, this character should
be considered when the porous concrete is expected to use
as a pavement material in road construction. Moreover,
although the quartzite showed a lower flakiness index and
a better permeability than dolomite, the clay contamination
and impurities such as a large amount of iron oxide covered
on the surface of quartzite cannot be omitted for gaining a
good bond in concrete mixtures. Therefore, dolomite is
believed as the best aggregates among them to make porous
concrete.
(2) The inclusion of silica fume did not appear to be very effec-
tive for improving the strength of porous concrete. Due to
the high porosity of the mixture, the fine particles of silica
fume tend to be segregated and deposited after the compac-
tion. Whereas the effects of Superplasticizer with regard to
assisting with the silica fume dispersing was outstanding
for porous concrete. As a dispersion agent, the application
of Superplasticizer is necessary for making the high-strength
porous concrete. The bonding strength of ITZ was enhanced
as a result of using these mineral and chemical additives,
so that the porous concrete could be stronger under
compression.
(3) The use of quarry sands will increase the compressive
strength of porous concrete, which is better than use of a
dense gradation of aggregate from 9.5 mm to 2.36 mm.
(4) The water content is one of the paramount factors for the
compressive strength. The control of water proportion is
essential to produce the fresh cement paste with a good
workability and not clogging up all of the pores. With the
fine aggregates and additives, the optimum water to cement
ratio turns out to be 0.32, this could produce the compres-
sive strength of 40 MPa after 28 days and water permeability
above 2 mm/s. When the requirement for structural strength
is not very high or potential clogging problems are particu-
larly concerned for permeable concrete pavement during
its long service time, a higher water to cement ratio of
0.36 could be used.
(5) Due to different configuration of porous concrete from the
conventional concrete, it cannot be denied that the perfor-
mance of porous concrete will be very sensitive to these fac-
tors. Although some optical observation was conducted, the
complex pore structure at meso or micro level has not been
revealed thoroughly. The testing method of porosity in this
study was not accurate driven by human errors inevitably.
Therefore, further research is required to investigate the
relationship between pore structure and strength and per-
meability for permeable concrete.
Acknowledgements
The authors would like to express their special thanks to Mr.
Wong Chee Chung for his assistance during the experimental work,
as well as the LabSA in Adelaide, Australia, for their support given
to this research.
References
[1] ARMCANZ & ANZECC: Australian and New Zealand Environment and
Conservation Council and Agriculture and Resource Management Council of
Australia and New Zealand, 2000.
[2] Ferguson BK. Porous pavements. Boca Raton, Florida: CRC Press; 2005.
[3] Booth DB, Levitta J. Field evaluation of permeable pavement systems for
improved stormwater management. J Am Plan Assoc 1999;65(3):314.
[4] Croney P, Croney D. The design and performance of road
pavements. London: McGraw-Hill; 1997.
[5] Meininger C. No-fines pervious concrete for paving. Concr Int 1988;10(8):
20–27.
[6] Ghafoori N, Dutta S. Laboratory investigation of compact no-fines concrete for
paving materials. J Mater Civ Eng 1995;7(3):183–91.
[7] Larrard F, Belloc A. Influence of aggregate on the compressive strength of
normal and high-strength concrete. ACI Mater J 1997;94(5):26–7.
 C. Lian, Y. Zhuge / Construction and Building Materials 24 (2010) 2664–2671
[8] Yang J, Jiang G. Experimental study on properties of pervious concrete
pavement materials. Cem Concr Res 2003;33(3):381–6.
[9] Crouch PE, Pitt J, Hewitt R. Aggregate effects on pervious portland
cement concrete static modulus of elasticity. J Mater Civ Eng 2007;19(7):
561–8.
[10] Krezel ZA. Recycled aggregate concrete acoustic barrier. Swinburne:
University of Technology, <http://adt.lib.swin.edu.au/uploads/approved/
adtVSWT20060821.154340/p-ublication>, 2006.
[11] Maso JC. Influence of the interfacial transition zone on composite
mechanical properties. In: Maso JC, editor. Interfacial transition zone in
concrete. London: E&FN SPON; 1996. p. 103–16.
[12] Kobayashi K, Hattori A, Miyagawa T. Characters of interfacial transition zone
cement with admixtures. In: Katz A et al., editors. The interfacial transition
zone in cementitious composites. London: E&FN SPON; 1998. p. 311–7.
[13] Rossignolo JA. Interfacial interactions in concretes with silica fume and SBR
latex. Constr Build Mater 2009;23(2):817–21.
[14] Monterio PJM, Metha PK. Interaction between carbonate rock and cement
paste. Cem Concr Res 1986;16:127–32.
[15] Monterio PJM, Metha PK. Effect of aggregate, cement, and mineral admixtures
on the microstructure of the transition zone. Invited paper. Boston: MRS
Society; 1988.
[16] Zhuge Y. Comparing the performance of recycled and quarry aggregate and
their effect on the strength of permeable concrete. In: Aravinthan T et al.,
editors. Proceeding of the 20th Australasian conference on the mechanics of
structures and materials, Australia: Toowoomba, 2008. p. 215–221.
2671
[17] Harber PJ. Applicability of no-fines concrete as a road pavement. University of
Southern Queensland Faculty of Engineering and Surveying 2005; <http://
eprints.usq.edu.au/archive/00000472/01/PaulHARBER-2005.pdf>.
[18] Allan ML, Kukacka LE. Comparison between slag and silica fume modified
grouts. ACI Mater J 1996;93(6):1–10.
[19] Dastol, Magne. Concrete additive comprising a multicomponent admixture
containing silica fume, its method of manufacture and concrete produced
therewith. United Patents 5275652, 1994, <http://www.wikipatents.com/
5275652.html>.
[20] Hommer H. Interaction of polycarboxylate ether with silica fume. J Eur Ceram
Soc 2008. in press.
[21] Tennis PD, Leming ML, Akers DJ. Pervious concrete pavements. Skokie,
IL: Portland Cement Association (PCA); 2004.
[22] Haselbach LM, Valavala S, Montes F. Permeability predictions for sand-clogged
Portland cement pervious concrete pavement systems. J Environ Manage
2006;81(1):42–9.
[23] Prokopski G, Halbiniak J. Interfacial transition zone in cementitious materials.
Cem Concr Res 2000;30:579–83.
[24] Poon CS, Shui ZH, Lam L. Effect of microstructure of ITZ on compressive
strength of concrete prepared with recycled aggregates. Constr Build Mater
2004;18(6):461–8.
[25] Zaitsev YB. Crack propagation in a composite material. Frac Mec Con
1983;8:251–99.
[26] Neville AM. Properties of concrete. England: Longman Group; 1995. p. 666–
672.