Professional Documents
Culture Documents
and Building
Received 24 June 2004; received in revised form 9 December 2004; accepted 31 January 2005
Available online 23 March 2005
Abstract
Clay brick generated from construction sites is usually delivered to landfills for disposal. With the limited landfill space in Hong
Kong, there is an immediate need to explore the possible use of crushed clay brick as a new civil engineering material. In Hong
Kong, government specifications promote the use of paving blocks made with 70–100% recycled concrete aggregate in order to facil-
itate the recycling of construction and demolition (C&D) waste. This paper presents a recent study at the Hong Kong Polytechnic
University on the investigation of blending recycled concrete aggregate and crushed clay brick as aggregates in the production of
paving blocks.
The results indicated that the incorporation of crushed clay brick reduced the density, compressive strength and tensile strength
of the paving blocks. Due to the high water absorption of crushed clay brick particles, the water absorption of the resulting paving
blocks were higher than that of the paving blocks that did not incorporate crushed clay brick. Although it was found that crushed
clay brick impaired the quality of the resulting paving blocks to a certain extent, the paving blocks using 50% crushed clay brick met
the minimum requirements specified by AS/NZS 4455 and ETWB of Hong Kong (Grade B) for pedestrian areas. Furthermore, it
was feasible to produce paving blocks prepared with 25% crushed clay brick that satisfied the compressive strength requirement for
paving blocks (Grade B) prescribed by ETWB of Hong Kong for trafficked area.
Ó 2005 Elsevier Ltd. All rights reserved.
Keywords: Construction and demolition waste; Recycled concrete aggregates; Crushed clay brick; Paving blocks
0950-0618/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2005.01.044
570 C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577
burnt) brick as aggregates in concrete. The unit weight, sonry waste to entirely replace sand in the mixtures of
bulk specific gravity and water absorption of the clinker calcium silicate products. However, there is currently
brick particles were between 1013 and 1161 kg/m3, 1.75 very limited amount of data on the use of crushed clay
and 1.98 and 1.7% and 4.1%, respectively. The resulting brick in the production of concrete paving blocks. Most
concrete had an unit weight of about 2100 kg/m3. The research was carried out for the paving blocks using
average compressive, tensile and flexural strengths of recycled concrete aggregate. Poon et al. [11] conducted
brick concrete were 7%, +2% and +15% of those of a thorough investigation to study the feasibility of using
normal concrete. Padmini et al. [3], who used a frac- recycled concrete aggregate in the production of paving
tional factorial experimental design method, studied blocks. They concluded that the replacement of coarse
the relative influence of different parameters on the and fine natural aggregates by recycled concrete aggre-
strength of concrete using low-strength bricks as aggre- gates can reach 50% without registering a reduction in
gates. The brick aggregates were derived from bricks the compressive strength of the paving blocks. The dura-
with compressive strengths between 6 and 13 MPa. bility performances of the resulting paving blocks using
The water absorption of the brick aggregate ranged recycled concrete aggregate were also satisfactory.
from 17% and 19% and the 10% fines values (crushing This paper presents a recent study at the Hong Kong
strength) were between 18 and 20 kN. It was found that Polytechnic University on the feasibility of blending
the strength of brick concrete was most influenced by recycled concrete aggregate and crushed clay brick as
the cement content, the aggregate conditions (i.e., pre- coarse and fine aggregates in the manufacture of paving
wet or dry before mixing) and the strength of brick from blocks. The major objective of this investigation was to
which the aggregates were derived. Since the strength of provide a viable option for the use of crushed clay brick
bricks varied considerably, it is therefore very difficult to derived from C&D waste in Hong Kong. The density,
quantify the quality of the resulting brick concrete. Kib- compressive strength, tensile splitting strength, trans-
riya and Speare [4] used three different types of brick verse breaking load, skid and abrasion resistances and
aggregates to assess their impacts on the strengths and water absorption were determined for the block speci-
the long-term durability of concrete. The brick aggre- mens. These values were compared to the standard
gates had specific gravities (SSD), 10% fines values and requirements prescribed by AS/NZS, BSI, ASTM and
water absorption values between 1.86 and 2.33, 40.8 the Environment, Transport and Works Bureau
and 144.3 kN, and 2.8% and 19.6%, respectively. The (ETWB) of Hong Kong.
resulting concrete had comparable compressive, tensile
and flexural strengths to those of normal concrete but
the modulus of elasticity was drastically reduced. Fur- 2. Materials
thermore, it was found that the use of brick aggregate
significantly increased the shrinkage of concrete at 90 2.1. Cementitious materials
days, and the creep tested at one year was slightly in-
creased as well. Moreover, the initial surface absorption Ordinary Portland cement and fly ash were used as
(ISA) and the chloride diffusion were found to be signif- the cementitious materials. Both cement and fly ash,
icantly higher for concrete prepared with brick aggre- commercially available in Hong Kong, comply with
gates than those of normal concrete. BS 12 and BS 3892 [12,13], respectively. The properties
Due to the potential problems of using crushed clay of cement and fly ash are shown in Table 1.
brick in concrete, many international agencies restrict
the amount of crushed brick that can be used in concrete 2.2. Recycled concrete aggregate (RCA)
[5–7], which hinders the recycling of this masonry waste.
However, the shortcomings of using crushed brick can Recycled concrete aggregate (RCA) was produced in
be avoided in the production of non-structural concrete a recycling facility located in Tuen Mun in Hong Kong.
products such as molded paving blocks for pavement
construction projects. In some climatic conditions Table 1
Chemical properties of cement and fly ash
(freezing and thawing do not occur in tropical or sub-
tropical places like Hong Kong), strength is the only Properties Cement Fly ash
concern for the paving blocks [8]. The concrete mixtures SiO2 (%) 19.61 56.79
for producing the paving blocks not only can tolerate a Fe2O3(%) 3.32 5.31
Al2O3(%) 7.33 28.21
greater amount of crushed brick but also maintain qual-
CaO (%) 63.15 <3
ities at acceptable standards. MgO (%) 2.54 5.21
Many investigations have already shown that it is fea- SO3(%) 2.13 0.68
sible to use crushed brick in mantle and storage block Loss on ignition (%) 2.97 3.90
productions [9]. Furthermore, Schuur [10] proved that Specific gravity 3.16 2.31
Specific surface area (cm2/g) 3520 3960
it is feasible to use crushed clay brick derived from ma-
C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577 571
The physical properties and the grading curves of the was crushed manually using a hammer to produce both
recycled concrete aggregates are shown in Table 2 and 10 and <5 mm aggregates and they are referred to as
Fig. 1, respectively. crushed clay brick (CB) in this study. The crushed clay
brick mainly contained brick rubbles and also a high
2.3. Crushed clay brick (CB) amount of adhered mortar and other impurities such
as tile, wood and dust. These impurities were not re-
A clay brick partition wall (Fig. 2) sourced from a moved prior to the experiment. The physical properties
demolition site in North Point in Hong Kong was deliv- and the grading curves of crushed clay brick are listed in
ered to our laboratory for this study. The partition wall Table 3 and Fig. 1, respectively.
Fig. 1. Grading curves for the recycled concrete aggregates and crushed clay brick.
Fig. 2. Photograph of a clay brick partition wall derived from the demolition site in North Point, Hong Kong.
572 C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577
Table 3 Table 5
Properties of crushed clay brick Ten percent fines values of blended aggregates
Properties Aggregate size Test method Ten percent 75%RCA 50%RCA 25%RCA
fines 25%CB 50%CB 75%CB
10 mm <5 mm
Dry (kN) 107 86 72
Density-SSD (kg/m3) 2147 2042 BS 812 Part 2
Density-oven-dry (kg/m3) 1797 1560 Soaked (kN) 77 60 48
Water absorption (%) 19.5 30.9
Indeed, the high amount of adhered mortar attached to Two types of concrete paving blocks are usually used.
the crushed clay brick particle led to a decrease in parti- The first type is used in pedestrian areas and the other
cle density and an increase in water absorption value. type is used in trafficked areas. Table 6 summaries the
The strengths of the aggregate particles were tested in specifications stipulated by AS/NZS 4455 and ETWB
terms of the 10% fines values in dry and soaked condi- of Hong Kong for paving blocks used in pedestrian
tions in accordance with BS 812 Part 111 [15]. This test areas [17,18]. The two standards have different require-
provides a measure of the resistance of the aggregate to ments. AS/NZS 4455 only requires a minimum value
crushing under a gradually applied compressive load. A of transverse breaking load while ETWB of Hong Kong
higher 10% fines value indicates a better resistance to requires a minimum compressive strength for Grade B
crushing. The results (Table 4) indicated that the recy- paving blocks and some additional durability require-
cled concrete aggregate had a better resistance to crush- ments for Grade A paving blocks. On the other hand,
ing compared to the crushed clay brick. Furthermore, it the strength requirements for the paving blocks in traf-
is worth noting that the strengths of the recycled con- ficked areas are more stringent as shown in Table 7.
crete aggregate and crushed clay brick after soaking BS 6717 requires a minimum tensile splitting strength
decreased from 146 to 109 kN and from 49 to 35 kN, while ASTM and ETWB of Hong Kong require a min-
respectively. The corresponding reductions were imum compressive strength [18–20]. Furthermore, some
approximately 25% and 28% for recycled concrete performance requirements such as skid resistance, abra-
aggregate and crushed clay brick, respectively. Further- sion resistances and cold water absorption are also pre-
more, it is shown in that the 10 and <5 mm aggregates of scribed by BSI, ASTM and ETWB of Hong Kong.
both materials conformed to the grading limits for the
single-sized aggregate and the fine aggregate per BS 2.6. Mixing ratios
882 [16].
In this study, a portion of recycled concrete aggre- A total of two series of mixtures was initially pre-
gate was replaced by crushed clay brick in the produc- pared. The first series was prepared using coarse and fine
tion of the paving blocks. The replacement levels (by aggregates, cement and water. The resulting paving
weight) were 25%, 50% and 75% for the paving blocks. blocks aimed to meet the requirements stipulated by
Since both recycled concrete aggregate and crushed AS/NZS and ETWB of Hong Kong for pedestrian
clay brick had distinctive strength characteristics, the areas. On the other hand, the second series was prepared
10% fines values were different after blending. The using coarse and fine aggregates, cement, fly ash and
10% fines values of the blended aggregates are pre- water. The resulting paving blocks aimed to meet the
sented in Table 5. A comparison between Tables 4 requirements stipulated by BSI, ASTM and ETWB of
and 5 shows that the increasing use of crushed clay
brick reduced the 10% fines values in both the dry Table 6
and soaked conditions. As the crushed clay brick con- Requirements of paving blocks used in pedestrian areas prescribed by
tent increased from 0% to 75%, the 10% fines values re- AS/NZS and ETWB of Hong Kong [17,18]
duced by 51% and 56% in the dry and soaked Requirement AS/NZS ETWB
conditions, respectively. 4455 (Hong Kong)
Grade A Grade B
Compressive strength (MPa)a – P30 P30
Table 4 Transverse breaking load (kN)a P2.0 – –
Ten percent fines values of recycled concrete aggregate and crushed Skid resistance (BPN)b – P45 –
clay brick Abrasion resistance (mm)b – 623 –
Cold water absorption (%) – 66 –
Ten percent Recycled concrete Crushed Test method
a
fines aggregate (RCA) brick (CB) For a batch of paving blocks, the mean strength value of the paving
blocks should be greater than the strength specified in the above Table
Dry (kN) 146 49 BS 812 Part 111
plus 1.65 times the unbiased standard deviation.
Soaked (kN) 109 35 b
Measured in accordance with BS 6717.
C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577 573
Table 7
Requirements for paving blocks used in trafficked areas prescribed by BSI, ASTM and ETWB of Hong Kong [18–20]
Requirement BS 6717 ASTM C936 ETWB (Hong Kong)
Grade A Grade B
Compressive strength (MPa) – P55.2 P45a P45a
Tensile splitting strength (MPa) P3.9 – – –
Skid resistance (BPN) P45 – P45 –
Abrasion resistanceb 623 mm 615 cm3/50 cm2 623 mm –
Cold water absorption (%) – 65 66 –
a
For a batch of paving blocks, the mean strength value of the paving blocks should be greater than the strength specified in the above Table plus
1.65 times the unbiased standard deviation.
b
Refer to BS 6717 and ASTM C 418 for methods for measuring abrasion resistance.
Hong Kong for trafficked areas. For all the mixtures, aggregate, fine aggregate, cement, fly ash and added
the aggregates were blended in a ratio of 25% coarse water in Series II. The aggregate-to-cement (A/C) and
aggregate to 75% fine aggregate. the aggregate-to-cementitious materials ratios (A/CM)
Series I included four mixtures. A control mixture were 4.1:1 and 2.4:1, respectively.
was prepared with only recycled concrete aggregate. Finally, a single mixture (BL-25-Ff) was prepared
The second, third and fourth mixtures were prepared using only fine aggregate, fly ash, cement and water.
by replacing 25%, 50% and 75% by weight of recycled The mix ratio of this mixture is shown in Table 10. This
concrete aggregate with crushed clay brick, respectively. mixture had the same A/C ratio, A/CM ratio and
For Series I, the mixing ratio between coarse aggregate, crushed clay brick replacement level as those of mixture
fine aggregate, cement and added water is summarized BL-25-F in Series II. Due to the limited availability of
in Table 8 which shows that an aggregate-to-cement ra- crushed clay brick for the experiment, only one replace-
tio (A/C) of 4.8:1 was used. ment level was prepared. However, the difference
Series II slightly differed from Series I. Likewise, Ser- between the results of this mixture and BL-25-F can pro-
ies II contained four mixtures in which a control mixture vide some indications of the possible effect of the aggre-
was prepared with recycled concrete aggregate. The sec- gate size on the properties of the paving blocks.
ond, third and fourth mixtures were prepared by replac-
ing 25%, 50% and 75% by weight of recycled concrete 2.7. Mixing procedure
aggregate with crushed clay brick, respectively. The only
difference between mixtures in Series I and II was that Initially, coarse and fine aggregates, cement and fly
the mixtures in Series II used fly ash to replace about ash were mixed in a pan mixer for approximately
15% by weight of both coarse and fine aggregates in Ser- 3 min. After mixing for 3 min, water was added to the
ies I. Table 9 shows the mixing ratio between coarse materials. The procedure of mixing and adding water
Table 8
Mixing ratio of Series I
Name Mixing ratio (by weight)
10 mm <5 mm Total aggregate Cement Added water
RCA CB RCA CB
BL-control 1.20 0 3.60 0 4.80 1.0 0.59
BL-25 0.90 0.30 2.70 0.90 4.80 1.0 0.65
BL-50 0.60 0.60 1.80 1.80 4.80 1.0 0.75
BL-75 0.30 0.90 0.90 2.70 4.80 1.0 0.96
Table 9
Mixing ratio of Series II
Name Mixing ratio (by weight)
10 mm <5 mm Total aggregate Cement Fly ash Added water
RCA CB RCA CB
BL-control-F 1.03 0 3.07 0 4.10 1.0 0.70 0.62
BL-25-F 0.77 0.26 2.30 0.77 4.10 1.0 0.70 0.80
BL-50-F 0.52 0.52 1.53 1.53 4.10 1.0 0.70 0.85
BL-75-F 0.26 0.77 0.77 2.30 4.10 1.0 0.70 1.00
574 C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577
Table 10
Mixing ratio of BL-25-Ff
Mixing ratio (by weight)
<5 mm Total aggregate Cement Fly ash Added water
RCA CB
BL-25-Ff 3.07 1.03 4.10 1.0 0.70 0.74
was iterated until the desired moisture content for these of the same specimen. Cold water absorption was mea-
dry mixtures was obtained. Normally, workability is not sured by immersing the block specimen in cold water at
important for molded blocks and only a small amount room temperature for 24 h while hot water absorption
of water is required. However, in the case of the inclu- was calculated by immersing the same specimen in hot
sion of crushed clay brick (with a high water absorption water at a temperature of 100 °C for 5 h.
value) in the mixture, the amount of added water used in The compressive strength was determined using a
this study was relatively higher. For producing paving compressive testing machine with a maximum capacity
blocks, the added water varied from 12% to 25% by of 3000 kN. The load was applied to the nominal area
weight of the total aggregate. of paving blocks. Prior to the loading test, the paving
blocks were soft capped with two pieces of plywood.
2.8. Fabrication of paving blocks The compressive strength was calculated by dividing
the failure load by the loading area of the paving blocks.
The paving blocks were fabricated in steel moulds The transverse breaking load of the paving blocks
with internal dimensions of 200 mm long, 100 mm in was determined by a three-point bending test with a
wide, and 60 mm high. The mixed materials were laid span length of 175 mm. Furthermore, the tensile split-
into each mould in three layers of about equal thickness. ting strength was determined in accordance with BS
After each of the first two layers was laid, compaction 6717. The test was carried out along the longest splitting
was applied manually using a hammer and a wood stem. section (i.e., the length) of the block specimen. Prior to
After the third layer was laid, a compressive force the test, the block specimen was concentrically packed
increased at a rate of 600 kN/min was applied for about with two steel packing pieces on the top and bottom
50 s to mechanically compact the materials within the faces in contact with the platens of the loading machine.
mould. Excessive materials were then removed with a A load was then gradually applied and the test was ter-
trowel in order to provide a good surface texture of minated when the specimen split into two halves. The
the resulting blocks. After that, a final compaction force failure load was recorded and the tensile splitting
was applied at the same rate for 60 s. The paving blocks strength was calculated based on the failure load.
were then demoulded one day after casting and were The skid resistance of paving blocks was determined
cured in water until the day of testing. using a British Pendulum Skid Resistance Tester and it
The amount of mixed materials used for a paving was expressed as the measured British Pendulum Num-
block specimen was between 3.0 and 3.2 kg. The weight ber (BPN) as per BS 6717. The abrasion resistance was
of the materials used in each specimen decreased as the determined by abrading the surface of the block speci-
amount of crushed clay brick in the mixture increased. men with an abrasive material under controlled condi-
This was attributed to the lower particle density of the tions as specified by BS 6717. The dimension of the
crushed clay brick. groove as a result of the abrasive action was used to mea-
sure the abrasion resistance of paving blocks where a
2.9. Test methods smaller groove indicated a better resistance to abrasion.
Various tests were carried out to determine the den- 2.10. Test results and discussions
sity, cold and hot water absorption, 7-day and 28-day
compressive strengths, tensile transverse/splitting The test results of Series I and II are summarized in
strengths, abrasion resistance and skid resistance of Tables 11 and 12. Each presented value is an average
the block specimens. of two measurements. The results indicated that the
The density of paving blocks was determined using a incorporation of crushed clay brick decreased the den-
water displacement method as per BS 1881 Part 114 for sity of the paving blocks (Fig. 3). As the crushed clay
hardened concrete [21]. The cold and hot water absorp- brick replacement level increased from 0% to 75%, the
tion values of the prepared blocks were determined and densities of the paving blocks in Series I and II reduced
were expressed as a ratio of the mass of the absorbed by about 10%. The decrease in density was expected be-
water of an immersed specimen to the oven dried mass cause crushed clay brick had a lower particle density
C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577 575
80
28-day compressive strength (MPa)
2350
Series I 70 Series II
2300
y = -2.691x + 2265.6 y = 0.4799x + 4.0892
60
Density (kg/m^3)
2
2250 R = 0.9647 Series I R2 = 0.9989
Series II 50
2200 Series I
40
y = 0.5083x - 12.53
2150 R2 = 0.9915
Series II 30
2100 y = -2.8133x + 2286.6
2 20
R = 0.9643 Series I
2050 10
Series II
2000 0
0 25 50 75 100 50 70 90 110 130 150 170
Crushed clay brick replacement level (%) Ten Percent Fines Value (kN)
Fig. 3. Relationship between densities of paving blocks and crushed Fig. 5. Relationship between 28-day compressive strengths (MPa) and
clay brick contents. ten percent fines values of blended aggregates (kN).
576 C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577
breaking load and tensile splitting strength tended to in- 4455. All paving blocks achieved a transverse breaking
crease with an increase in the compressive strength. load value greater than 2.0 kN. Also, paving blocks
It was found that the skid resistance of the paving BL-control, BL-25, and BL-50 satisfied the minimum
blocks in Series I was higher than that of the paving compressive strength prescribed by ETWB of Hong
blocks in Series II. This was reasonable since the use Kong. Although the skid and abrasion resistances of
of fly ash as a partial replacement of aggregate produced the paving blocks in Series I met the requirements pre-
a more homogenous mix with a smoother surface tex- scribed by ETWB of Hong Kong for Grade A paving
ture, thus reduced the skid resistance. However, the blocks, the cold water absorption of the paving blocks
abrasion resistance of the paving blocks in Series I and exceeded the limit of 6%.
II were found to be similar. Unlike the results for the On the other hand, most paving blocks in Series II
skid resistance, the use of fly ash did not impair the did not meet the strength requirements prescribed by
abrasion resistance of the paving blocks. BSI, ASTM and ETWB of Hong Kong. Only BL-con-
Because of the high water absorption of the crushed trol-F (control mix) satisfied all the requirements stipu-
clay brick particles, the cold and hot water absorption lated by these three standards. All mixtures containing
values increased with an increase in the crushed clay crushed clay brick failed to meet the tensile splitting
brick content as shown in Fig. 7. Although the water strength requirement of 3.9 MPa imposed by BSI and
absorption values of the paving blocks in Series II in- only the mixture with 25% of recycled concrete aggre-
creased with an increase in the crushed clay brick con- gate replaced by crushed clay brick satisfied the com-
tents (Fig. 7), the magnitude was reduced when pressive strength requirements of 55 and 45 MPa
compared to the results of Series I. This was attributed prescribed by ASTM and ETWB of Hong Kong, respec-
to the effect of the replacement of the coarse and fine tively. Furthermore, all paving blocks containing
aggregates by the fine fly ash particles which filled the crushed clay brick failed to achieve a cold water absorp-
voids within the mixture. tion value of less than 5% and 6% as recommended by
The paving blocks in Series I met the transverse ASTM and ETWB of Hong Kong for Grade A paving
breaking load requirement prescribed by AS/NZS blocks, respectively.
The effect of using only fine crushed brick is shown in
Table 13. By comparing the mixtures BL-25-F and BL-
4.0 12
Tensile splitting strength (MPa)
Series I
y = 0.1326x + 2.0492
3.5 2
10 Using only <5 mm aggregate reduced the density of
R = 0.846
8 the paving block since the density (SSD) of the 10 mm
3.0 aggregate was more than 10% higher than that of
6
2.5 Series II
<5 mm aggregate. Although the compressive strengths
y = 0.0233x + 1.5422 4 were similar between the two mixtures, a significant
2.0 Series II Series I R2 = 0.9573 improvement of the tensile splitting strength was
2
observed for BL-25-Ff in which no 10 mm aggregate
1.5 0
0 10 20 30 40 50 0 0 0
was used. A possible reason can be attributed to the
increase in the bonding area between the aggregate
28-day compressive strength (MPa)
and the cement paste. The corresponding increases in
Fig. 6. Relationship between 28-day compressive strengths (MPa) and cold and hot water absorption values were 20% and
tensile splitting strengths (MPa)/transverse breaking loads (kN). 30%, respectively. This was mainly attributed to the rel-
atively high water absorption of the <5 mm aggregate
24
Series I (Cold) Series I (Hot)
Series I (Cold)
21 y = 0.1603x + 8.7709
y = 0.1277x + 6.8716
Series I (Hot) R2 = 0.9671
Water absorption (%)
R2 = 0.9094
18 Series II (Cold)
Table 13
15 Series II (Hot) Test results of mixture BL-25-Ff
12 Series II (Hot) Name BL-25-Ff
y = 0.111x + 5.556
R2 = 0.7628
3
9 Density (kg/m ) 2160
Compressive strength (MPa)
6 SeriesII (Cold)
y = 0.1061x + 4.0128 7-day 37.9
3 R2 = 0.881 28-day 52.9
0 Tensile splitting (MPa) 3.7
0 25 50 75 100 Skid resistance (BPN) 97
Abrasion resistance (mm) 20.0
Crushed clay brick replacement level(%) Water absorption (%)
Fig. 7. Relationship between water absorption values and crushed clay Cold 7.3
brick contents. Hot 9.6
C.S. Poon, D. Chan / Construction and Building Materials 20 (2006) 569–577 577