Construction and Building Materials 87 (2015) 100–112

Contents lists available at ScienceDirect

Construction and Building Materials

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

Development of ultra-lightweight cement composites with low thermal

conductivity and high specific strength for energy efficient buildings
Yunpeng Wu a, Jun-Yan Wang b, Paulo J.M. Monteiro c, Min-Hong Zhang a,⇑
a
Department of Civil & Environmental Engineering, National University of Singapore, 1 Engineering Drive 2, Singapore 117576, Singapore
b
Key Laboratory of Advanced Civil Engineering Materials, Tongji University, Ministry of Education, Shanghai 201804, China
c
Department of Civil & Environmental Engineering, University of California, Berkeley, 725 Davis Hall, Berkeley, CA 94720, USA

h i g h l i g h t s

 Ultra-lightweight cement composites (ULCCs) are developed for energy efficiency.

3
 ULCCs with densities <1471 kg/m are developed using hollow cenospheres.
 ULCC has specific compressive strength similar to that of concrete of 110 MPa.
 ULCC has 80% lower thermal conductivity than concrete of similar 28-day strength.
 Thermal conductivity of ULCCs may be estimated using Hashin–Shtrikman lower bound.

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

Article history: Energy efficient building is defined as achieving satisfactory internal environment and service with mini-
Received 3 March 2015 mum energy consumption. One of the most important parameters that affect the heat transfer through
Received in revised form 2 April 2015 the building envelope is thermal conductivity. The thermal conductivity of lightweight concrete is
Accepted 7 April 2015
generally lower than that of normal-weight concrete due to the lower thermal conductivity of air.
Available online 20 April 2015
Although introducing voids in concrete reduces its thermal conductivity and increases its insulation
capacity, the mechanical properties are generally compromised.
Keywords:
This study focuses on the development of ultra-lightweight cement composites (ULCCs) with low ther-
Cement composite
Cenospheres
mal conductivity but high specific strength so that they can be used for structural applications. The light-
Lightweight weight is achieved by incorporating hollow cenospheres from fly ash generated in thermal power plants.
Mechanical property The ULCCs had 1-day densities ranged from 1154 to 1471 kg/m3 and 28-day compressive strengths ran-
Thermal conductivity ged from 33.0 to 69.4 MPa. The properties of the ULCCs were compared with those of cement pastes with
comparable water/binder and those of a concrete with 28-day compressive strength of 67.6 MPa.
Results indicate that the compressive strength, flexural tensile strength, and elastic modulus of the
ULCCs were reduced with the decrease in density. However, compressive and flexural tensile strength
of 69.4 and 7.3 MPa were achieved for the ULCC, respectively, similar to the cement paste with w/b of
0.35 and concrete. With similar 28-day compressive strength, the thermal conductivity of the ULCC
was 54% and 80% lower than that of the cement paste and concrete, respectively. The low thermal con-
ductivity of the ULCC is due to the incorporation of hollow cenospheres as micro-aggregate which effec-
tively introduce voids and decrease density of the ULCC. The high specific strength of the ULCC may be
attributed to (1) the presence of hard and stiff shell in the cenospheres, (2) the ‘‘control’’ of void sizes
in the cenospheres, and (3) the creation of strong cement paste matrices that provide ‘‘three dimensional’’
confinement to the cenospheres.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction primary energy consumption in the world is for building sector

Energy consumption is growing rapidly due to economic devel-
opment and population growth [1]. While over 40% of the total
⇑ Corresponding author.
[2], a significant part of that energy is consumed when buildings
are in use through heating, ventilation, and air conditioning
(HVAC) systems. In developed countries, 20% of the total energy
consumption is by HVAC systems [1]. An energy efficient building
is defined as achieving satisfactory internal environment and

http://dx.doi.org/10.1016/j.conbuildmat.2015.04.004
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112 101

service with minimum energy consumption [3]. The energy
required to achieve satisfactory indoor temperatures correlates
mainly with the heat transfer through building envelope, heat
convection on building surface, and heat generated by indoor
equipment and human activities.
One of the most important parameters that determine the heat
transfer through building envelope is the thermal conductivity. It is
a measure of the rate at which heat is transferred through a mate-
rial, and is defined as the ratio of heat flux to temperature gradient.
Concrete has low thermal conductivity compared with other build-
ing materials, and can be further reduced by introducing air voids
to resist heat flow between indoor and outdoor environment.
There are several ways to introduce air voids into concrete: (1)
voids in aggregates of various sizes, (2) voids in cement paste, (3)
elimination of sand in the concrete mixture leading to voids in
between the coarse aggregate, and (4) combinations of the above.
Among them, the first one is typically used to achieve higher speci-
fic strength and low permeability.
Thermal conductivity of concrete is affected primarily by the
thermal conductivity of the raw materials used, mixture propor-
tion, void content, and moisture condition of the concrete. The ther-
mal conductivity of most natural rock aggregates is much higher
than that of the cement pastes, and it varies depending on the rock
type. In general, rocks with crystalline structure show higher ther-
mal conductivity than amorphous and vitreous rocks of the same
composition [4,5]. The thermal conductivity of cement pastes
increases with the reduction of water/binder (w/b) and capillary
porosity. Due to the lower thermal conductivity of air (in the order
of 0.03 W/m K), the thermal conductivity of lightweight concrete
(LWC) is generally lower than that of normal weight concrete. For
a given type of lightweight aggregate (LWA), the reduction in the
concrete density results in a decrease in the thermal conductivity.
Moisture content is another major factor that influences the ther-
mal conductivity of concrete because the thermal conductivity of
water (0.5 W/m K) is higher than that of air.
In recent years, structural concrete mixtures with low thermal
conductivities have been developed using aggregates other than
some commonly used manufactured LWA or industrial by-product
such as expanded clay, expanded shale, or foamed slag etc.
Examples of the LWAs include pumice [6,7], perlite [8,9], ceno-
spheres [10,11], polyurethane foam [12], diatomite [7], expanded
glass [13], aerogel [14], and high-impact polystyrene [15]. A sum-
mary of the density, compressive strength, thermal conductivity,
and raw materials used for such lightweight-aggregate concretes
(LWAC) reported in literature are presented in Table 1. A summary
of the thermal conductivity data of the LWAC made with the tradi-
tional LWAs such as expanded clay and shale among others, can be
found in the FIP manual of lightweight-aggregate concrete [16].
However, no information is available on their corresponding
strengths. Thus, they are not included in Table 1.
Although introducing voids in concrete reduces its thermal con-
ductivity and increases its insulation capacity, the mechanical
properties, such as compressive strength and elastic modulus, are
generally compromised. This study focuses on the development
of structural materials which have low thermal conductivity for
energy efficiency purpose but sufficient strength to be used for
structural applications.
In this study, cenospheres were used to introduce voids into the
ultra-lightweight cement composites (ULCCs). Cenospheres are
hollow spheres typically produced as a byproduct of coal combus-
tion at thermal power plants [17,18] (Fig. 1). The process of burn-
ing coal in thermal power plants produces fly ash which includes
mainly solid particles and a small amount of hollow particles
(cenospheres). Cenosphere particles have spherical shapes with
sizes typically ranging from 10 to 400 lm [19]. They generally have
a hollow interior that is covered by a thin shell with thicknesses in
the order of 5–10% of its diameter [20]. Due to its hollow structure,
cenospheres have low particle densities typically ranging from 400
to 900 kg/m3. Because of their low particle densities, cenospheres
have been used for making ultra-lightweight cement composites
in recent years [10,11,21–26]. Ultra-lightweight cement compos-
ites with a density of 1430 kg/m3 and 28-day compressive strength
of about 60 MPa was developed and reported by Chia et al. [23].
The objectives of this study are to investigate the thermal con-
ductivity of ULCCs in comparison to that of cement pastes with
comparable water/binder ratios (w/b) of 0.35 and 0.45 and a

Table 1
Summary of relevant information in literature.

References Aggregate information w/b Wet density Oven dry Compressive Moisture condition Thermal
type/maximum size (kg/m3) density strength (MPa) at which TC was conductivity
(kg/m3) determined (W/m K)
Blanco et al. [10] Cenosphere/4 mm 0.3 1090–1415 1050–1350* 5.0–30.1 Moist cured 0.46–0.60
Dried at 100 °C 0.36–0.46
Uysal et al. [6] Pumice/16 mm NA 2270–1329 NA NA 0% 1.46–0.78
Gül et al. [8] Perlite/16 mm 0.7 1773–1984 1590–1800* 11.3–25.1 NA 0.82–1.23
Topçu et al. [7] Diatomite/4 mm 0.2 NA 900 6 0% 0.13
Pumice/4 mm 1500 9 0.44
Mounanga et al. [12] Polyurethane foam waste/ 0.6–0.7 1261–2165 900–1980* 23.0–3.2 NA 1.44–0.31
10 mm Saturated 2.34–0.68
Tandiroglu [9] Perlite/15 mm 0.4–0.65 NA 1798–1883 60–80 0% 1.47–1.76**
Sengul. et al. [52] Expanded perlite/2–4 mm 0.55 392–1937 354–1833 0.1–28.8 0% 0.6–0.13
Kim et al. [54] Fine bottom ash/0.6 mm & 0.47 1800–1553 1500–1210** 22–8 0% 0.54–0.36
expanded shale/19 mm
Wang & Meyer [15] High impact polystyrene/ 0.55 NA 1980–1560 37–19 NA 0.61–0.27
2.36 mm
Huang et al. [11] Cenosphere/600 lm & Iron ore 0.26 1649–2001 1483–1890*** 44.3–48.1 NA 0.29–0.37
tailings/300 lm
*
Yu et al. [13] Expanded glass/4.0 mm 0.38–0.59 1280–1460 1100–1380 23–30 0% 0.49–0.85
Yun et al. [53] Glass Bubble/(Median 0.55 2011–2370 1900–2260* 43.9–24.6 NA 2.25–1.41
diameter: 65 lm)
Gao et al. [14] Aerogel/4 mm 0.38–0.39 1000–2050 900–1950* 8.3–60** NA 0.26–1.9**

NA – not available.
*
estimated based on ASTM C567 [51] where mix proportions are available.
**
obtained from figures in literature.
***
air dry density was used here because it’s lower than estimated oven dry density using ASTM C 567 [51].
102 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112
Fig. 1. Image of cenospheres (Exlite).
concrete with 28-day compressive strength of 67.6 MPa. The ULCCs
had densities of 1154–1471 kg/m3 after demold and compressive
strengths of 33.0–69.4 MPa after 28 days. Experiments on the ther-
mal conductivity and properties such as density, compressive and
flexural tensile strength, and elastic modulus of the materials were
conducted. Effect of the cenospheres, chemical admixtures, and
fibers on the performance of the ULCCs is discussed. The experi-
mental results of the thermal conductivity of the ULCCs are
compared with those estimated from composite models and an
empirical model.
2. Experimental details
2.1. Materials
ASTM Type I Portland cement (also EN ‘CEM I 52.5N’) and undensified silica
fume (SF) were used for most mixtures including ULCCs and cement pastes except
for the concrete in which only the Portland cement was used.
Two types of cenospheres (QK300 and Exlite) with average particle densities of
approximately 908 and 615 kg/m3, respectively, were used in ULCCs as micro-ag-
gregates. Particle size distributions of the cenospheres are given in Fig. 2, and the
results show that most of the particles had sizes from 10 to 300 lm. A previous
study [27] showed that the cenospheres had low content of CaO but high combined
content (approximately 90%) of SiO2 and Al2O3. X-ray diffraction analysis of the
cenospheres indicated that they contained a large amount of amorphous material
and small amounts of quartz and mullite crystals. ASTM C227 [28] and C1260 test
[29] results indicate that the cenospheres are not potentially deleterious in an
alkali-silica reaction when used in ULCCs [27].
Crushed granite with a nominal maximum size of 10 mm and natural sand with
a fineness modulus of 2.66 were used in the concrete mixture. Densities of the
coarse and fine aggregates were 2650 and 2630 kg/m3, respectively.
A polycarboxylate based superplasticizer (SP) was used for most of the mixtures
except for the concrete in which a naphthalene based SP was used. The reason for
the use of polycarboxylate based superplasticizer with higher water reducing
capacity in the ULCCs and cement paste was because of their large amount of fine
Fig. 2. Particle size distribution of the two types of cenospheres used.
materials (cenospheres and cementitious materials). However, such super-
plasticizer is dosage sensitive to the stability of the mixtures. Thus, naphthalene
based superplasticizer was used for the concrete mixture.
Shrinkage reducing admixture (SRA) was used for some mixtures. A previous
study [30] indicates that the SRA reduces the surface tension of deionized water
and synthetic pore solution. As a result, it can reduce entrapped air content in
the mixtures and improve the performance of the ULCCs [25].
In addition, a viscosity modifying agent (VMA) which is often used to increases
plastic viscosity and produces more cohesive concrete was used in some mixtures
to adjust workability. Silane oligomers (3-glycidyloxypropyl trimethoxysilane
(C9H20O5Si)) and a de-foaming admixture (Tributyl phosphate) were used in a
ULCC mixture reinforced with polyethylene (PE) fibers (properties given in
Table 2) to evaluate their effects. The silane oligomers were prepared by heating
the silane at 120 °C for 4 h [31,32].
2.2. Mix proportions and specimen preparation
Nine ULCCs, two cement pastes, and one concrete samples were included in this
study, and mix proportions are given in Table 3. The water/binder ratios of the
ULCCs ranged from 0.35 to 0.56. For comparison, two cement pastes with w/b of
0.35 and 0.45 were also prepared. However, the cement paste sample with w/b of
0.56 was not successful due to excessive bleeding. The ULCCs and cement pastes
contained 8% silica fume by mass of cementitious materials. Mix ULCC-7 and Mix
ULCC-8 contained 0.5% of PE fibers by volume of the ULCCs to reduce their brittle-
ness. A concrete with a target 28-day compressive strength of 65 MPa was prepared
for comparison.
The ULCCs and cement paste mixtures had good flowability determined accord-
ing to BS EN 1015-3 test [33], and no segregation was observed. For most of the
mixtures, the flow value was within a range of 190 to 215 mm with only a few
exceptions: due to larger amounts of the cenospheres used in Mixes ULCC-4 and
ULCC-6 in order to achieve lower densities, higher flow values of 240 to 250 mm
were used to properly cast of specimens.
The specimens shown in Table 4 were made for each mixture. They were cov-
ered with plastic sheets, and demolded within 48 h after casting. The specimens
were then cured in a fog room at temperatures of 28–30 °C (simulating tropical
environment) until testing at 7 and 28 days, or other testing ages with drying after
28-day moist curing.
2.3. Experimental methods
The test methods, relevant standards, testing ages, specimen sizes, and speci-
men numbers are summarized in Table 4.
2.3.1. Density and mechanical properties
The density of all the cube specimens was determined, after demolding, using
the water displacement method. The air content in the cement pastes or mortar
matrices (excluding the voids in the cenospheres) was estimated according to
ASTM C138/C138M-13a [34]. Compressive strength was determined using
100-mm cubes according to BS EN 12390-3:2009 [35]. Flexural performance was
determined using 75  100  400-mm prisms with third-point loading (4-point
bending, span length 300 mm) [36]. An Instron closed-loop, servo-controlled test-
ing system was used for the flexural test with a loading rate of 0.05 mm/min.
For measuring the elastic modulus, two strain gauges were pasted at mid height
of each specimen on two opposite sides [37]. A Denison 3000 kN compression test-
ing equipment was used with a loading rate of 0.01 mm/min. For Mixes ULCC-1–
ULCC-4 and CP 0.35, the specimens were tested at around 70 days after casting
(due to a problem of the equipment). During the period from 28 to 70 days, these
specimens were covered with plastic sheets to prevent moisture loss.
2.4. Thermal conductivity
Thermal conductivity of the ULCCs specimens was determined according to
ASTM C518-04 [38] by a heat flow meter. The applicable range of the equipment
was 0.002–0.500 W/m K and the results reported were the average from two
specimens.
The thermal conductivities of the cement pastes and concrete were determined
following ASTM C177-04 [39] using a guarded hot plate equipment. The applicable
testing range of the equipment was 0.02–2.00 W/m K and two 300  300  50-mm
specimens were required for each test.
The thermal conductivity of the concrete was also determined by a hot box
apparatus according to ASTM C1363 [40] because the result obtained using the
guarded hot plate equipment was beyond the testing range of the equipment.
The 300  300  50-mm specimens used were first moist cured for 28 days, and
then dried in laboratory air for two weeks followed by drying in an oven at 105 °C
until constant weight was achieved. In general, about two months were needed to
achieve a constant weight.
 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112 103

Table 2
Properties of polyethylene fibers.

Tensile strength (MPa) Elastic modulus (GPa) Length (mm) Diameter (lm) Aspect ratio Density (kg/m3)
Polyethylene fibers 2610 79 12 39 308 970

Table 3
Mix proportions and workability of ultra-lightweight cement composites (ULCCs), cement pastes, and concrete.

Mix ID Variables w/b Cenosphere Water Coarse Sand Binder* Cenosphere SP SRA VMA Silane** PE Flow/
type aggr fibers slump
(kg/m3)  (kg/ Vol (%) in (L/m3)  (kg/m3)  (mm)
m3)  ULCCs
Concrete Max aggr = 10 mm 0.42 – 172 946 810 410 – – 5.4 0 0 0 0 95  
ULCC-1 w/c, cenosphere type 0.35 QK300 302 – – 909 348 38.3 4.9 10.5 0 0 0 200
ULCC-3 and density 0.37 QK300 282 – – 796 402 44.3 5.2 9.8 0 0 0 210
ULCC-5 0.37 Exlite 287 – – 795 268 43.6 5.9 8.9 0.202 0 0 215
ULCC-4 0.45 QK300 282 – – 660 442 48.7 5.6 9.8 0 0 0 240
ULCC-6 0.56 Exlite 290 – – 542 317 51.6 6.6 9.1 0.202 0 0 250
ULCC-2 SRA 0.35 QK300 305 – – 920 352 38.7 5.2 0 0 0 0 195
ULCC-1 0.35 QK300 302 – – 909 348 38.3 4.9 10.5 0 0 0 200
ULCC-4 VMA, control 0.45 QK300 282 – – 660 442 48.7 5.6 9.8 0 0 0 240
ULCC-4 workability 0.45 QK300 282 – – 659 442 48.6 6.7 9.7 0.176 0 0 195
(VMA)
ULCC-2 Fiber and 0.35 QK300 305 – – 920 352 38.7 5.2 0 0 0 0 195
ULCC-7 silane + defoamer 0.35 QK300 304 – – 842 350 38.6 3.6 0 0 4.2 5.3 200
ULCC-8 0.35 QK300 301 – – 904 346 38.1 4.3 0 0 0 5.7 190
CP 0.35 Cement pastes, w/c 0.35 – 499 – – 1473 – – 1.3 14.9 0 0.0 0.0 200
CP 0.45 0.45 – 561 – – 1282 – – 0.0 16.9 0.912 0.0 0.0 160
*
Binder contained 8% silica fume (SF) by mass of cementitious materials except for concrete which contains no SF.
**
5% (by weight of silane) of tributyl phosphate (TBP) was added together with silane to avoid excessive foaming.
 
For 1 m3 of all materials added.
  
Slump.

Table 4
Test methods and specimen information.

Property Test standard Testing age Specimen

Type and size No.
Workability (flow table) (ULCCs and BS EN 1015-3:1999 [33] Right after – –
cement pastes) mixing
Workability (Slump test) (concrete) ASTM C143: 2003 [55] Right after – –
mixing
Density of hardened specimens after BS EN 12390-7:2009 [56] 1–2 day Cube, 6
demold 100  100  100 mm
Estimation of porosity ASTM C138: 2013 [34] – – –
Compressive strength BS EN 12390-3:2009 [35] 7 & 28 days Cube, 3
100  100  100 mm
Flexural performance ASTM C1609: 2007 [36] 28 days Prism,75  100  400 mm 3
Elastic modulus ASTM C469 (2 strain gauges were glued at mid height of each cylinder 28 days Cylinder, Ø100  200 mm 3
specimen) [37]
Thermal conductivity of ULCCs ASTM C518-10: 2010 [38] About Slab, 300  300  50 mm 2
Thermal conductivity of cement pastes ASTM C177-13: 2013 [39] 100 days*
Thermal Conductivity of concrete ASTM C177-13: 2013 and ASTM C1363-11: 2011 [39,40]
*
28 days moist curing followed by drying in lab air for 14 days and then drying in oven at 105 °C to constant weight.

3. Results and discussion
The results for density, mechanical properties, and thermal
conductivity of the ULCCs, cement pastes, and concrete are
summarized in Table 5.
3.1. Compressive strength, flexural tensile strength, and elastic
modulus
Fig. 3 clearly shows that the density of the ULCCs and concrete
is affected by the density of the aggregates embedded in the
cement paste matrices. The ULCCs had significantly lower density
(1154 to 1471 kg/m3 after demold) due to the incorporation of hol-
low cenospheres (particle densities = 615 and 908 kg/m3) as micro
aggregates compared with the concrete (2341 kg/m3 after demold)
which included granite coarse aggregate and natural sand with
densities of 2650 and 2630 kg/m3, respectively.
The compressive strength of the ULCCs, cement pastes, and con-
crete samples was affected by their density. The compressive
strength of the ULCCs and cement pastes was reduced with the
decrease in density (Fig. 3). Nevertheless, 28-day compressive
strength of 69.4 MPA was achieved for the mixture ULCC-1, similar
 104
Table 5
Density, mechanical properties, and thermal conductivity.

Mix ID Variable w/b Aggr Other 1 day Oven dry Estimated air content in 7 days comp 28 days Thermal
type information density density paste matrix, expressed as strength (MPa) conductivity

Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112

Comp strength Flexural strength E
(kg/m3) (kg/m3) % in 1 m3 (W/m K)
(MPa) (MPa) (GPa)
Concrete Max aggr = 10 mm 0.42 Granite – 2341 2251 2.9 59.2 67.6 6.5 35.5 1.98
ULCC-1 w/c, cenosphere type 0.35 QK300 – 1471 1303 6.6 57.8 69.4 7.3 17.0 0.40
ULCC-3 and density 0.37 QK300 – 1357 1198 9.2 40.4 56.9 5.4 14.8 0.36
ULCC-5 0.37 Exlite – 1355 1178 1.2 42.8 55.9 4.7 13.9 0.35
ULCC-4 0.45 QK300 – 1240 1087 11.3 38.7 49.8 4.4 13.0 0.33
ULCC-6 0.56 Exlite – 1154 966 1.0 24.5 33.0 3.6 10.4 0.28
ULCC-2 SRA 0.35 QK300 No SRA 1457 1298 7.8 57.7 66.1 7.0 17.0 0.39
ULCC-1 0.35 QK300 With SRA 1471 1303 6.6 57.8 69.4 7.3 17.0 0.40
ULCC-4 VMA, control 0.45 QK300 No VMA 1240 1087 11.3 38.7 49.8 4.4 13.0 0.33
ULCC-4 workability 0.45 QK300 With VMA 1196 1042 14.4 32.0 40.9 4.1 12.0 0.31
(VMA)
ULCC-2 Fiber and 0.35 QK300 No fiber 1457 1298 7.8 57.7 66.1 7.0 17.0 0.39
ULCC-7 silane + defoamer 0.35 QK300 With fiber & 1455 1300 7.8 57.2 66.5 6.3 16.0 0.43
silane
ULCC-8 0.35 QK300 With fiber no 1361 1219 13.4 43.6 54.4 5.2 14.3 0.39
silane
CP 0.35 Cement pastes, w/c 0.35 – Cement paste 2000 1715 0.0 52.5 63.1 6.7 21.6 0.84
CP 0.45 0.45 – CP, with VMA 1854 1479 0.4 37.8 51.3 3.5 16.5 0.80
 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112
Fig. 3. 28-day compression strength vs wet density after demold.
to that of the cement paste with w/b of 0.35 and the concrete. The
specific strength (strength/density) of the ULCC-1 is 0.047 MPa/kg/
m3, which is equivalent to a normal weight concrete with compres-
sive strength of about 110 MPa.
Flexural tensile strength of the ULCCs, cement pastes, and con-
crete increased with the increase in the compressive strength
(Fig. 4a). The flexural tensile strength of the ULCCs was about
8.4–10.9% of their 28-day compressive strength, which is com-
parable to that of the cement pastes and concrete. The flexural
tensile strength of the ULCCs and cement pastes also decreased
as the density decreased (Fig. 4b).
The elastic modulus of cement composites and concrete was
generally affected by their compressive strength and density. As
shown in Fig. 5a, the elastic modulus of the ULCCs increased
with an increase in the compressive strength. With the 28-
day compressive strength within a range from 33.0 to
69.4 MPa, the elastic modulus of the ULCCs, cement pastes,
and concrete decreased with the density (Fig. 5b). The elastic
modulus of the ULCCs and concrete were also affected by the
density of cenospheres and aggregates incorporated in the
cement paste matrices.
Fig. 4. (a) Flexural tensile strength vs compressive strength; (b) flexural tensile strength vs wet density after demold.
105
Eq. (1), given by ACI 318 [41], estimates the elastic modulus of
concrete Ec based on the compressive strength fc and density qc of
concrete. The equation is applicable to concrete with compressive
strengths <41 MPa and densities from 1440 to 2480 kg/m3:
qffiffiffiffi
0
Ec ¼ 0:043qc1:5 fc ð1Þ
The experimental values and the calculated values of the elastic
modulus of the ULCCs, together with their difference, are shown in
Table 6. The difference were plotted against 1-day density and 28-
days compressive strength and shown in Fig. 6a and b, respectively.
From Fig. 6, It is clear that the equation overestimates the elastic
modulus of the ULCCs when the density >1230 kg/m3 and com-
pressive strength >43 MPa but underestimates the elastic modu-
lus when the density <1230 kg/m3 and compressive strength
<43 MPa. The difference between the calculated and experimental
elastic moduli of the ULCCs seems to increase when the density and
compressive strength are increased beyond 1230 kg/m3 and
43 MPa, respectively. It should be noted that although the
ACI-318 equation is applicable to concrete, the equation may be
used to estimate the elastic modulus of the ULCCs with reasonable
accuracy based on the results.
3.2. Thermal conductivity
Thermal conductivity of the ULCCs, cement pastes, and concrete
reduced with the decrease in oven dry density (Fig. 7). The ULCCs
had significantly lower thermal conductivity than the cements
pastes and concrete due to the incorporation of the hollow ceno-
spheres. Comparing the mixture ULCC-2, cement paste (CP0.35),
and concrete of similar 28-day compressive strength, the thermal
conductivity of the ULCC-2 (0.39 W/m K) was 54% lower than that
of the cement paste (0.84 W/m K) and 80% lower than that of the
concrete (1.98 W/m K) (Table 5). With the decrease of oven dry
density to 966 kg/m3, the thermal conductivity of the ULCC was
further reduced to 0.28 W/m K (Table 5).
It should be mentioned that the thermal conductivity of
2.11 W/m K was obtained for the concrete from the guard hot plate
test. Although the result was beyond the testing range of the
equipment (0.02–2.00 W/m K), this method generally gives more
accurate result than the hot-box method. The thermal conductivity
of the concrete was also affected by the source of the aggregate.
Thermal conductivities of 2.6–2.7 W/m K and 3.1 W/m K for
106 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112

Fig. 5. (a) Elastic modulus vs compressive strength, (b) elastic modulus vs wet density after demold.

Table 6
Comparison of experimental and calculated elastic modulus of the ULCCs.

Mix ID w/b Aggr 1 day density 28 days compressive Experimental elastic Calculated elastic Difference between experimental and
type (kg/m3) strength (MPa) modulus (GPa) modulus (GPa) calculated E (%)
ULCC-1 0.35 QK300 1471 69.4 17.0 20.2 18.9
ULCC-2 0.35 QK300 1457 66.1 17.0 19.4 14.4
ULCC-3 0.37 QK300 1357 56.9 14.8 16.2 9.6
ULCC-4 0.45 QK300 1240 49.8 13.0 13.2 1.9
ULCC-4 0.45 QK300 1196 40.9 12.0 11.4 5.2
(VMA)
ULCC-5 0.37 Exlite 1355 55.9 13.9 16.0 15.4
ULCC-6 0.56 Exlite 1154 33.0 10.4 9.7 6.9
ULCC-7 0.35 QK300 1455 66.5 16.0 19.5 21.6
ULCC-8 0.35 QK300 1361 54.4 14.3 15.9 11.4

concretes with granite aggregates are mentioned in textbooks by
Mehta & Monteiro [42] and Mindess et al. [43], respectively.
The thermal conductivities of the cement pastes with w/b of
0.35 and 0.45 with oven dry densities of 1715 and 1479 kg/m3
were 0.84 and 0.80, respectively, which are higher than those
reported by Baghban et al. [44]. They reported a good correlation
between the thermal conductivity and oven dry density ranged
from about 1050 to 1730 kg/m3 using experimental results of
cement pastes with water/cement (w/c) from 0.36 to 0.84.
According to their relationship, the thermal conductivities of the
cement pastes with densities comparable to those in our study
should be about 0.73 and 0.60. The higher thermal conductivity
values of the cement pastes obtained in this study may be due to
the incorporation of 8% silica fume in the cement pastes which pro-
duces denser microstructure.
The relationship between thermal conductivity and dry density
of concrete has been explored and reported in literature
[7,8,45,46]. The thermal conductivity may be predicted by dry den-
sity of concrete, however, the relationship is affected by the type of
aggregate used in concrete [47].
Based on the results of the ULCCs from this study, a regres-
sion equation (Eq. (2)) was obtained with R2 of 0.92 (plotted
in Fig. 7). Eq. (2) is in similar form to an equation proposed
by Valore et al. [45] (Eq. (3)) for predicting the thermal conduc-
tivity of concrete. The thermal conductivity of the ULCCs can be
estimated by using Valore’s equation with a multiplication factor
of about 1.2.
k ¼ 0:11e^ ðd=1000Þ
k ¼ 0:072e^ ð1:25  d=1000Þ ½45 ð3Þ
where, k is the thermal conductivity (W/m K) and d is the oven dry
density (kg/m3).
3.3. Estimation of thermal conductivity of cenospheres and ULCCs
Considering ULCC as a discrete two-phase composite material
with the cenosphere particles as micro aggregate embedded in
cement paste matrix, its thermal conductivity may be estimated
if the thermal conductivity of each component (cement paste and
cenospheres) and their relative proportions are known. Some com-
monly used models for estimating the thermal conductivity of
ultra-lightweight cement composites include Series and Parallel
models, Hashin–Shtrikman (H–S) models [48] and cubic model
[47].
For fine particulate materials such as cenospheres, loosely
packed samples are often used to determine their thermal conduc-
tivities. The thermal conductivity values of the samples obtained in
such tests depend on the thermal conductivity of the particulate
materials and void space among the particles as shown in Fig. 8.
As the test sample includes particles and air voids among various
particles, the values may underestimate the actual thermal con-
ductivities of the particles because of the low thermal conductivity
of air.
The thermal conductivity of the cenospheres QK300 was, there-
fore, estimated using parallel model (Eq. (4)), series model (Eq. (5)),
Hashin–Shtrikman models (Eqs. (6) and (7)) and cubic model (Eq.
ð2Þ (8)). The calculation and analyses are based on the assumptions
 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112 107

Fig. 6. Difference between experimental and calculated elastic modulus vs (a) wet density at 1 day and (b) 28 days compression strength.

that (1) ULCC is a two-phase composite of cement paste and ceno-

spheres, and (2) volume sum of the cement paste and cenospheres
equal to 1, and (3) volume of the SRA, SP, VMA and air voids is con-
sidered as a part of the cement paste.
Parallel model:

kULCC ¼ kcp V cp þ kceno V ceno ð4Þ

Series model:

1 V cp V ceno
¼ þ ð5Þ
kULCC kcp kceno
Hashin–Shtrikman (H–S) model (upper bound):

V ceno
kULCC ¼ kcp þ V
½48 ð6Þ
1
kceno kcp
þ 3kcpcp

Hashin–Shtrikman (H–S) model (lower bound):

V cp
kULCC ¼ kceno þ 1 V ceno
½48 ð7Þ
Fig. 7. Thermal conductivity vs oven-dry density. kcp kceno
þ 3k ceno

Cubic model:
2 3
6 7
6 2=3 7
6 V ceno 7
kULCC ¼ kcp 6
6 0 17
7 ½47 ð8Þ
6 2=3 7
4V @ 
V ceno A5
ceno  V ceno þ k
2=3
ceno V ceno 2=3
kcp
þ1V ceno

where kLLC, kcp, kceno – thermal conductivity of ULCCs, cement paste,

Fig. 8. Sketch of loosely packed cenosphere sample for thermal conductivity test
(not to scale).
and cenospheres. Vcp and Vceno – volume ratio of cement paste and
cenospheres in ULCC.
The ULCC mixtures 1, 2, and 4 and corresponding cement pastes
(CP 0.35 and CP 0.45) were used for estimating the thermal con-
ductivity of the cenospheres QK300, and results are shown in
Table 7. The ULCC Mixtures 4(VMA), 7, and 8 were not used in
the estimation due to possible influence of VMA, fibers, and silane
on the thermal conductivity. The ULCC-3 was also not included
because of different w/b. However, these four mixtures were used
to verify if the estimated thermal conductivity of the cenospheres
QK300 can be used to estimate the thermal conductivity of the
ULCCs.
108 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112

Table 7
Estimation of thermal conductivity of the cenospheres by series, parallel, cubic and Hashin–Shtrikman (H–S) models.

Mix ID w/b kULCCs (W/m K) kcp (W/m K) Vceno Thermal conductivity kceno (W/m K), estimated by using
Parallel model Series model Cubic model H–S upper bound H–S lower bound
ULCC-1 0.35 0.398 0.835 0.383 0.306 0.216 0.118 0.056 0.137
ULCC-2 0.35 0.388 0.835 0.387 0.320 0.210 0.128 0.066 0.127
ULCC-4 0.45 0.327 0.804 0.487 0.175 0.201 0.040 0.006 0.136

Table 8
Comparison of experiment and estimated thermal conductivity of the ULCCs.

Mix ID w/b kcp (W/m K) Vceno Thermal Estimated thermal conductivity of ULCCs
conductivity
From H–S lower bound From series model
of ULCCs
(experiment) Estimated thermal Error of Estimated thermal Error of
(W/m K) conductivity of estimation (%) conductivity of estimation (%)
cenospheres = 0.133 cenospheres = 0.209
W/m K W/m K
ULCC-1 0.35 0.835 0.383 0.398 0.392 1.6 0.389 2.3
ULCC-2 0.35 0.835 0.387 0.388 0.389 0.3 0.387 0.3
ULCC-4 0.45 0.804 0.487 0.327 0.322 1.5 0.337 3.0
ULCC-4 0.45 0.804 0.486 0.308 0.323 4.8 0.337 9.5
(VMA)
ULCC-3 0.37 0.829* 0.443 0.355 0.352 0.9 0.358 0.9
ULCC-7 0.35 0.835 0.386 0.425 0.390 8.3 0.387 8.9
ULCC-8 0.35 0.835 0.381 0.39 0.393 0.8 0.390 0.0
*
Calculated by linear interpolation.

From Table 7, the thermal conductivity of the cenospheres
QK300 estimated from the Series model and H–S lower bound give
reasonably consistent values from 0.201 to 0.216 W/m K and 0.127
to 0.137 W/m K, respectively, whereas those estimated from
Parallel, H–S upper bound, and cubic models give negative values.
The negative values indicate that the Parallel, H–S upper bound,
and cubic models may not be suitable in this case. The estimated
average thermal conductivity of the cenospheres is 0.209 W/m K
from Series model and 0.133 W/m K from H–S lower bound.
These estimated thermal conductivity values of the cenospheres
QK300 are significantly higher than that provided by manufacturer
(0.08 W/m K) which is likely due to the effect of void space among
the cenosphere particles mentioned earlier.
Based on the estimated thermal conductivity of the cenospheres
above, the thermal conductivities of the ULCCs are calculated using
the Series model and H–S lower bound, which are further
compared with corresponding experimental values. Since the ther-
mal conductivity of the cement paste is not affected significantly
by the w/b between 0.35 and 0.45, the thermal conductivity of
the cement paste for ULCC-3 (w/b = 0.37) was estimated to be
0.83 W/m K using linear interpolation. Comparison of the
calculated and experimental values of the thermal conductivities
of ULCCs is given in Table 8. The error of estimation using the
H–S lower bound and Series model are both less than 10%.
Overall, the H–S lower bound gives lower error of estimation than
the Series model in estimating the thermal conductivity of the
ULCCs.
3.4. Effect of some ingredients in the ULCCs on mechanical properties
and thermal conductivity
3.4.1. Effect of cenospheres density and amount
A scanning electron microscope image of a typical ULCC is
shown in Fig. 9. The incorporation of hollow cenospheres in the
ULCCs reduced the density and thermal conductivity substantially
but maintained the compressive strength compared to the cement
paste of similar w/b although the elastic modulus was reduced. For
example, with w/b of 0.35 (compare ULCC-2 and CP0.35), the
incorporation of the cenospheres reduced the density and thermal
conductivity by 27% and 54%, respectively; with w/b equal to 0.45
(compare ULCC-4 and CP0.45), the incorporation of the ceno-
spheres reduced the density and thermal conductivity by 33%
and 59%, respectively. For both w/b the elastic modulus was
reduced by about 20%.
The good strength properties of the ULCCs in comparison to that
of the cement pastes may be attributed to (1) the presence of hard
and stiff shell in the cenospheres, (2) ‘‘controlled’’ void sizes in the
cenospheres, and (3) the creation of strong cement paste
matrices that provide ‘‘three dimensional’’ confinement to the
cenospheres.
For a given type of cenospheres used (Mixes ULCC-1, -3, and -4
or Mixes ULCC-5 and -6) (Table 5), the increase in the amount of
the cenospheres and w/b leads to a reduction in the thermal con-
ductivity and mechanical properties. For a given w/b, the ULCC-5
had lower porosity in the cement paste matrix but higher porosity
in the cenospheres (Exlite) compared with ULCC-3 (Tables 5). As
the two ULCCs had similar density, their mechanical properties
and thermal conductivity were similar.
3.4.2. Effect of shrinkage reducing admixture and viscosity modifying
admixture
Although the shrinkage reducing admixture (SRA) reduces the
surface tension of deionized water and synthetic pore solution of
cement paste [30] and, consequently, reduce the air content of
the ULCCs [25], the results shown in Table 5 indicate that the
SRA did not have significant influence on the density, mechanical
properties, and thermal conductivity of the ULCC-1 and ULCC-2.
This may be due to the different superplasticizer used in this study.
Mix ULCC-4 had a higher flow value of 240 mm in comparison
to other ULCCs mixtures, therefore the VMA was used in Mix
ULCC-4VMA (flow value = 195 mm) to determine if the flow value
and VMA can affect the density, thermal conductivity, and
mechanical properties of the ULCC. The Results shown in Table 5
indicate that the density and mechanical properties were some-
what reduced in the mixture with VMA; however, the thermal con-
ductivity was not affected.
 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112
Fig. 9. Scanning electron microscope image of a ULCC with cenospheres.
Fig. 10. Flexural stress – deflection curves of fiber-reinforced ULCCs (ULCC 7 and 8)
in comparison to that without fibers (ULCC 2).
3.4.3. Effect of fibers and silane
Fibers were incorporated to reduce the brittleness of the ULCCs.
Fig. 10 shows flexural stress vs deflection curves of the fiber-rein-
forced ULCC-7 and ULCC-8 in comparison to that of the plain ULCC-
2. The absence of post-peak curve of the ULCC-2 under bending is
due to limitations in the experimental setup. When the flexural
strength was reached, each specimen of the plain ULCC was broken
into two pieces without further load carrying capability, whereas
the specimens of the fiber-reinforced ULCCs were still able to carry
bending load as shown in the Figure. This indicates reduced brittle-
ness of the fiber-reinforced ULCCs.
However, the incorporation of the polyethylene fibers (Mix 8)
reduced the density, strengths, and elastic modulus of the ULCC,
even though it did not affect the thermal conductivity significantly.
To determine the effect of the fibers at given density of the ULCC,
the silane and defoaming agent described earlier were used in
Mix ULCC-7. The fiber-reinforced ULCC-7 had density, strengths,
and elastic modulus similar to those of the plain ULCC-2 without
fibers. However, the thermal conductivity of the former was higher
than the latter. The higher thermal conductivity of the fiber-
reinforced ULCC-7 in comparison to the plain ULCC-2 was
explained by the higher thermal conductivity of the polyethylene
fibers (0.46–0.52 W/m K [49]).
It should be noted that the silane may also affect the thermal
conductivity. Baghban et al. [44] reported decreased thermal
109
conductivity for oven dried cement paste with alkyl alkoxysilane
silane, whereas Xu and Chung [50] reported increased thermal
conductivity of cement paste with amino vinyl silane in
comparison to the control paste with similar density.
3.5. Balance of thermal conductivity and mechanical properties
A decrease in the density of concrete generally results in
reduced thermal conductivity but also decreases the mechanical
properties, such as strengths and elastic modulus. The former is
beneficial to reduce heat transfer through building envelopes but,
on the other hand, sufficient mechanical properties are essential
for structural applications.
Since the thermal conductivity of concrete is significantly
affected by moisture content, the thermal conductivity is generally
determined based on oven dry specimens. In the following discus-
sion to compare results from this study to those reported in litera-
ture, if the oven dry density is not reported, it is then estimated
based on mix proportions provided according to ASTM C 567 [51].
In this study high compressive and flexural tensile strengths of
69.4 MPa and 7.3 MPa, respectively, were achieved for the ULCC-1
sample, comparable to those of the concrete sample. The ULCC-1
also had a thermal conductivity of 0.4 W/m K, substantially lower
than that of the concrete. The low thermal conductivity of the
ULCCs is due to hollow cenospheres which effectively introduce
voids and decrease density of the ULCCs, whereas the high
strengths obtained may be attributed to the hard and rigid ceno-
spheres under confinement of dense cement paste matrices and
controlled void sizes in the cenospheres as mentioned earlier.
The combination of low thermal conductivity, lightweight, and
high specific strength of the ULCCs is significant improvement
compared with those reported in literature (Fig. 11a and Table 1).
As shown in Fig. 11a, most low thermal conductive concrete
(thermal conductivity <0.4 W/m K) have compressive strength less
than 30 MPa. Although the composites reported by Huang et al. [11]
have low thermal conductivity of <0.4 W/m K, the composites have
lower compressive strength (<50 MPa) and higher estimated oven
dry density (P1480 kg/m3) compared with the ULCCs obtained in
this study. The lightweight concretes reported by Tandiroglu [9]
have high compressive strengths of 60–80 MPa but also higher
thermal conductivity (>1.4 W/m K) and oven dry density
(>1798 kg/m3).
Blanco et al. [10] also used cenospheres to make lightweight
concretes and obtained low densities comparable to those of the
ULCCs in this study and low thermal conductivities of 0.36–
0.46 W/m K. However, the compressive strength of most of their
lightweight concretes were only 5 to 30 MPa (Table 1), much lower
than that of the ULCCs obtained in this study. Another study incor-
porated cenospheres is Huang et al. [11] which is discussed above.
For comparison between various mix proportions, it is desirable
to define a merit number M = compressive strength/(thermal con-
ductivity ⁄ density). Higher values of the merit number indicate
better optimization between mechanical and thermal properties.
The histogram shown in Fig. 11b indicates that the merit number
of the ULCC is basically one order of magnitude higher than most
of the results from Sengul et al. [52], Mounanga et al. [12], Gül
et al. [8], and Yun et al. [53].
Overall, the combination of low thermal conductivity, light-
weight, and high specific strength of the ULCCs developed in this
study may be used to improve the energy efficiency of buildings
which will be discussed in a future paper.
4. Summary and conclusions
Mechanical properties and thermal conductivity of ultra-
lightweight cement composites (ULCCs) with 1-day density from
110 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112

Fig. 11. (a) Density, compressive strength, and thermal conductivity of the ULCCs in comparison to concrete with various lightweight aggregates reported in literature, (b)
merit number defined as M = compressive strength/(thermal conductivity * density) (MPa/(W/m K) * (kg/m3) = 104 K m3/W s2) of various lightweight concretes mentioned
in (a).

1154 to 1471 kg/m3 and 28-day compressive strength from 33.0
to 69.4 MPa were studied. The low densities of the ULCCs were
achieved by incorporating hollow cenospheres from fly ash
generated in thermal power plants. The properties of the ULCCs
were compared with those of cement pastes with w/b of 0.35
and 0.45 and those of a concrete with 28-day compressive
strength of 67.6 MPa. Based on the experimental results,
numerical calculation, and discussion, following conclusions can
be drawn:
1. Compressive strength, flexural tensile strength, and elastic
modulus of the ULCCs were reduced with decrease in density.
However, the compressive and flexural tensile strength of 69.4
and 7.3 MPa were achieved for the ULCC-1, respectively, simi-
lar to those of the cement paste with w/b of 0.35 and the con-
crete. The specific strength of the ULCC-1 is 0.047 MPa/kg/m3,
equivalent to that of a normal weight concrete with compres-
sive strength of about 110 MPa. The density and elastic modu-
lus of the ULCCs and concrete were affected by the density of
aggregates embedded in the cement paste matrices. The ratio
of the flexural tensile strength/compressive strength of the
ULCCs is comparable to that of the cement pastes and
concrete.
2. Thermal conductivity of the ULCCs was reduced with decrease
in the oven-dry density and it was significantly lower than that
of the cement pastes and concrete due to the incorporation of
hollow cenospheres. With similar 28-day compressive strength,
the thermal conductivity of the ULCC was 54% and 80% lower
than that of the cement paste and concrete, respectively.
3. Compared to the properties from various lightweight concretes
reported in literature, the ULCCs have an optimum combination
of low thermal conductivity, lightweight, and high specific
strength. The low thermal conductivity of the ULCCs is due to
the incorporation of hollow cenospheres as micro-aggregate
which effectively introduce voids and decrease density of the
ULCCs. The high specific strength of the ULCCs may be attribu-
ted to (1) the presence of hard and stiff shell in the cenospheres,
(2) the ‘‘control’’ of void sizes in the cenospheres, and (3) the
creation of strong cement paste matrices that provide ‘‘three
dimensional’’ confinement to the cenospheres.
 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112
4. The incorporation of polyethylene fibers reduced the brittleness
observed in the ULCC samples, but did not have significant
effect on the strength and elastic modulus if the density and
w/b were kept constant. However, the fiber reinforced ULCC
had higher thermal conductivity than that without fibers, which
may be due to the higher thermal conductivity of the polyethy-
lene fibers. The shrinkage reducing admixture and viscosity
modifying agent did not have significant effect on the mechani-
cal properties and thermal conductivity of the ULCCs.
5. The thermal conductivity of the ULCC can be estimated by using
Valore’s equation (1980) with a multiplication factor of about 1.2.
6. The estimated average thermal conductivity values of the ceno-
spheres based on H–S lower bound and Series model are signifi-
cantly higher than that provided by manufacturer due to the
effect of void space among the cenosphere particles.
7. Based on the estimated thermal conductivity of cenospheres,
the thermal conductivity of the ULCCs can be estimated using
H–S lower bound and Series model with error of estimation less
than 10%. Overall, the H–S lower bound gives lower error of
estimation than the Series model.
Acknowledgements
The first author acknowledges the scholarship from National
University of Singapore (NUS). This research is funded by the
Republic of Singapore’s National Research Foundation through a
grant to the Berkeley Education Alliance for Research in Singapore
(BEARS) for the Singapore-Berkeley Building Efficiency and
Sustainability in the Tropics (SinBerBEST) Program.
Acknowledgement is also made to Air Conditioning Laboratory,
Mechanical Engineering Department, NUS, and Dr. Chen Fangzhi,
Solar Energy Research Institute of Singapore, for assistance in deter-
mining the thermal conductivity of cement pastes and concrete.
References
[1] Pérez-Lombard L, Ortiz J, Pout C. A review on buildings energy consumption
information. Energy Build 2008;40(3):394–8.
[2] International Energy Agency. Promoting energy efficiency investments: case
studies in the residential sector. Paris: OECD Publishing; 2008.
[3] Chartered Institution of Building Services Engineers. Energy efficiency in
buildings: CIBSE guide F. London: CIBSE; 2004.
[4] Zoldners NG. Thermal properties of concrete under sustained elevated
temperatures. ACI Publication SP-25. 1971:1–31.
[5] Harmathy TZ. Thermal properties of concrete at elevated temperatures. J Mater
1970:47–74.
[6] Uysal H, Demirboğa R, Sßahin R, Gül R. The effects of different cement dosages,
slumps, and pumice aggregate ratios on the thermal conductivity and density
of concrete. Cem Concr Res 2004;34(5):845–8.
_
[7] Topçu IB, Uygunoğlu T. Properties of autoclaved lightweight aggregate
concrete. Build Environ 2007;42(12):4108–16.
[8] Gül R, Okuyucu E, Türkmen I, _ Aydin AC. Thermo-mechanical properties of fiber
reinforced raw perlite concrete. Mater Lett 2007;61(29):5145–9.
[9] Tandiroglu A. Temperature-dependent thermal conductivity of high strength
lightweight raw perlite aggregate concrete. Int J Thermophys
2010;31(6):1195–211.
[10] Blanco F, Garcia P, Mateos P, Ayala J. Characteristics and properties of
lightweight concrete manufactured with cenospheres. Cem Concr Res
2000;30:1715–22.
[11] Huang X, Ranade R, Zhang Q, Ni W, Li VC. Mechanical and thermal properties
of green lightweight engineered cementitious composites. Constr Build Mater
2013;48:954–60.
[12] Mounanga P, Gbongbon W, Poullain P, Turcry P. Proportioning and
characterization of lightweight concrete mixtures made with rigid
polyurethane foam wastes. Cem Concr Comp 2008;30(9):806–14.
[13] Yu QL, Spiesz P, Brouwers HJH. Development of cement-based lightweight
composites – Part 1: mix design methodology and hardened properties. Cem
Concr Comp 2013;44:17–29.
[14] Gao T, Jelle BP, Gustavsen A, Jacobsen S. Aerogel-incorporated concrete: an
experimental study. Constr Build Mater 2014;52:130–6.
[15] Wang R, Meyer C. Performance of cement mortar made with recycled high
impact polystyrene. Cem Concr Comp 2012;34(9):975–81.
[16] Federation Internationale de la Precontrainte (FIP). FIP manual of lightweight
aggregate concrete. New York, Glasgow: Surrey University Press; 1983.
111
[17] Wandell T. Cenospheres: from waste to profits. The American Ceramic Bulletin
1996;75(6):79–81.
[18] Montgomery D, Diamond S. The influence of fly ash cenospheres on the details
of cracking in flyash-bearing cement pastes. Cem Concr Res 1984;14:767–75.
[19] Parameswaran V, Shukla A. Processing and characterization of a model
functionally gradient material. J Mater Sci 2000;35(1):21–9.
[20] Cardoso RJ, Shukla A, Bose A. Effect of particle size and surface treatment on
constitutive properties of polyester–cenosphere composites. J Mater Sci
2002;37:603–13.
[21] Tiwari V, Shukla A, Bose A. Acoustic properties of cenosphere reinforced
cement and asphalt concrete. Appl Acoust 2004;65(3):263–75.
[22] Mcbride SP, Shukla A. Processing and characterization of a lightweight
concrete using cenospheres. J Mater Sci 2002;37:4217–25.
[23] Chia KS, Zhang MH, Liew JYR. High-strength ultra lightweight cement
composite - material properties. 9th International Symposium on High
Performance Concrete design, Verification and Utilization. Aug 2011.
[24] Wang J-Y, Yang Y, Liew J-YR, Zhang M-H. Method to determine mixture
proportions of workable ultra lightweight cement composites to achieve
target unit weights. Cem Concr Comp 2014;53:178–86.
[25] Wang J-Y, Chia K-S, Liew J-YR, Zhang M-H. Flexural performance of fiber-
reinforced ultra lightweight cement composites with low fiber content. Cem
Concr Comp 2013;43:39–47.
[26] Chia KS, Liu X, Liew JYR, Zhang MH. Experimental study on creep and
shrinkage of high performance ultra lightweight cement composite of 60 MPa.
Struct Eng Mech 2014;50(5):635–52.
[27] Wang J-Y, Zhang M-H, Li W, Chia K-S, Liew RJY. Stability of cenospheres in
lightweight cement composites in terms of alkali–silica reaction. Cem Concr
Res 2012;42(5):721–7.
[28] ASTM C227-10. Standard test method for potential alkali reactivity of cement-
aggregate combinations (Mortar-Bar method). American Society of Testing
and Materials; 2010.
[29] ASTM C 1260-14. Standard test method for potential alkali reactivity of
aggregates (Mortar-Bar method). American Society of Testing and Materials;
2014.
[30] Wang J-Y, Banthia N, Zhang M-H. Effect of shrinkage reducing admixture on
flexural behaviors of fiber reinforced cementitious composites. Cem Concr
Comp 2012;34(4):443–50.
[31] Monticelli F, Toledano M, Osorio R, Ferrari M. Effect of temperature on the
silane coupling agents when bonding core resin to quartz fiber posts. Dent
Mater 2006;22(11):1024–8.
[32] Ishida F. The interfacial interactions in polymeric composites. Kluwer
Academic Publisher; 1993.
[33] BS EN 1015-3. Methods of test for mortar for masonry – Part 3: determination
of consistence of fresh mortar (by flow table). British Standard; 1999.
[34] ASTM C138/C138M. Standard test method for density (unit weight), yield, and
air content (gravimetric) of concrete. American Society of Testing and
Materials; 2013.
[35] BS EN 12390-3. Testing hardened concrete Part 3: compressive strength of test
specimens. British Standard; 2009.
[36] ASTM C1609M-07. Standard test method for flexural performance of fiber-
reinforced concrete (using beam with third-point loading). American Society
of Testing and Materials; 2007.
[37] ASTM C469/C469M. Standard test method for static modulus of elasticity and
poisson’s ratio of concrete in compression. American Society for Testing and
Materials; 2014.
[38] ASTM C518-10. Standard test method for steady-state thermal transmission
properties by means of the heat flow meter apparatus. American Society of
Testing and Materials; 2010.
[39] ASTM C-177. Standard test method for steady-state heat flux measurements
and thermal transmission properties by means of the guarded-hot-plate
apparatus. American Society of Testing and Materials; 2013.
[40] ASTM C1363-11. Standard test method for thermal performance of building
materials and envelope assemblies by means of a hot box apparatus. American
Society of Testing and Materials; 2011.
[41] ACI Committee 318. Building code requirements for structural concrete and
commentary, part 3. ACI Manual of Concrete Practice; 2010.
[42] Mehta PK, Monteiro PJM. Concrete: microstructure, properties, and materials.
3rd ed. New York: McGraw-Hill; 2014.
[43] Mindess S, Young JF, Darwin D. Concrete. Upper Saddle River, NJ: Prentice Hall;
2003.
[44] Hajmohammadian Baghban M, Hovde PJ, Jacobsen S. Analytical and
experimental study on thermal conductivity of hardened cement pastes.
Mater Struct 2013;46(9):1537–46.
[45] Valore Jr RC. Calculation of U-values of hollow concrete masonry. Concr Int
1980;2(2):40–63.
[46] Demirboğa R, Gül R. The effects of expanded perlite aggregate, silica fume and
fly ash on the thermal conductivity of lightweight concrete. Cem Concr Res
2003;33(5):723–7.
[47] ACI-121R. Guide to thermal properties of concrete and masonry
systems. American Concrete Institute; 2002.
[48] Hashin Z, Shtrikman S. A variational approach to the theory of the effective
magnetic permeability of multiphase materials. J Appl Phys 1962;33(10):
3125–31.
[49] Mark JE. Physical properties of polymers handbook. 2nd ed. Springer Science;
2007.
112 Y. Wu et al. / Construction and Building Materials 87 (2015) 100–112
[50] Xu Y, Chung DDL. Cement of high specific heat and high thermal conductivity,
obtained by using silane and silica fume as admixtures. Cem Concr Res
2000;30:1175–8.
[51] ASTM C567. Standard test method for determining density of
structural lightweight concrete. American Society of Testing and
Materials; 2000.
[52] Sengul O, Azizi S, Karaosmanoglu F, Tasdemir MA. Effect of expanded perlite
on the mechanical properties and thermal conductivity of lightweight
concrete. Energy Build 2011;43(2):671–6.
[53] Yun TS, Jeong YJ, Han T-S, Youm K-S. Evaluation of thermal conductivity for
thermally insulated concretes. Energy Build 2013;61:125–32.
[54] Kim HK, Jeon JH, Lee HK. Workability, and mechanical, acoustic and thermal
properties of lightweight aggregate concrete with a high volume of entrained
air. Constr Build Mater 2012;29:193–200.
[55] ASTM C 143M. Standard test method for slump of hydraulic-cement
concrete. American Society of Testing and Materials; 2003.
[56] BS EN 12390-7. Testing hardened concrete Part 7: density of hardened
concrete. British Standard; 2009.