Construction and Building Materials 251 (2020) 118960

Contents lists available at ScienceDirect

Construction and Building Materials

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

Thermo-mechanical and moisture absorption properties of fly ash-based

lightweight geopolymer concrete reinforced by polypropylene fibers
Yijiang Wang a,b,⇑, Tongda Zheng a, Xiaofeng Zheng b, Yifei Liu c, Jo Darkwa b, Guoqing Zhou a,⇑
a
State Key Laboratory for Geomechanics and Deep Underground Engineering, School of Mechanics and Civil Engineering, China University of Mining and Technology, Xuzhou
221116, China
b
Faculty of Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, UK
c
Shanghai Architectural Design and Research Institute, China State Construction Engineering Corporation, Shanghai 200062, China

h i g h l i g h t s

NaOH within mass ratio of 0–10% can improve thermo-mechanical properties.

Fibers increase thermal conductivity, compressive strength, moisture absorption.

Coarse aggregates reduce thermal conductivity and enhance moisture absorption.

Moisture absorption can be weakened by surface waterproofing treatment.

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

Article history: An experimental investigation on the thermo-mechanical and moisture absorption properties of light-
Received 12 December 2019 weight geopolymer concrete prepared with fly ash, NaOH, sodium silicate and Polypropylene Fibers
Received in revised form 29 February 2020 (PF) is presented in this study. The effects of dry density, NaOH, PF, aggregates and hydrophobic agent
Accepted 31 March 2020
on the compressive strength, thermal properties and moisture absorption were studied. Results indicate
Available online 8 April 2020
that thermo-mechanical properties of Fly ash-based Lightweight Geopolymer Concrete (FLGC) strongly
depend on the dry density, NaOH, PF and aggregates contents. The increase in dry density and fine aggre-
Keywords:
gate contents resulted in higher compressive strength and thermal conductivity. NaOH within mass ratio
Lightweight geopolymer concrete
Thermal insulation
of 0–10% is able to enhance thermo-mechanical properties. The optimal compressive strength was
Fly ash achieved when the length and content of the PF was 12 mm and 0.5% respectively. Meanwhile, PF in
Fiber reinforcement the range of 0–1% can also increase thermal conductivity and enhance moisture absorption. The increase
Thermo-mechanical properties in coarse aggregate ranging from 0 to 15% led to reduced dry density and thermal conductivity and
Moisture absorption enhanced moisture absorption, but did not affect compressive strength. Interestingly, the decrease in fine
aggregate with the same content had the opposite impact to the moisture absorption in comparison to
the coarse aggregate. However, the moisture absorption can be considerably weakened by surface water-
proofing treatment which makes the enhanced thermal performance durable. Therefore, the FLGC rein-
forced by PF has excellent thermo-mechanical properties and can also be engineered to be an
environmentally friendly and durable thermal insulation material with the assistance of waterproofing
treatment.
Ó 2020 Elsevier Ltd. All rights reserved.

1. Introduction also responsible for 1/3 of total greenhouse gases emission [1].
Hence, the usage of energy saving and efficient materials has
As the impact of climate change is causing great concern glob- attracted increasing attention in the construction industry due to
ally, there has been an increasing pressure on cutting carbon emis- the need for a decarbonized building sector. The optimization in
sions across all sections. Currently almost 40% of the total world’s thermal properties of construction materials has been proved to
energy consumption is contributed by the building sector, which is be a crucial way to improve the durability and energy efficiency
of buildings [2]. In addition, researchers have shown increased
interest in construction materials with reduced environmental
⇑ Corresponding authors. impact at both fabricating and operational stages of the material
E-mail addresses: [email protected] (Y. Wang), [email protected] (G. Zhou).

https://doi.org/10.1016/j.conbuildmat.2020.118960
0950-0618/Ó 2020 Elsevier Ltd. All rights reserved.
2 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960

lifecycle [3]. For instance, one of the important implementations is Fly ash-based Lightweight Geopolymer Concrete (FLGC) reinforced
the production of thermal insulation materials from solid wastes. by polypropylene fibers, which is a further advancement of the
Coal fly ash is a type of fine solid particulate residue driven out previous researches on the lightweight geopolymer concrete. In
of the boiler with the flue gases in coal-fired power plants. The dis- this study, the lightweight geopolymer concrete was prepared with
posal of fly ash is a major economic and environmental burden due fly ash, alkali solution (sodium hydroxide and sodium silicate),
to its abundance and release of toxic metals. The production of coal foaming agent, aggregates (ceramsite and standard sand) and rein-
fly ash in China had been gradually increased from 5.32  108 t forced by polypropylene fibers. Foam-stabilizing agent was also
(2013) to 6.20  108 t (2015) and 6.86  108 t (2017). However, utilized to improve the foam stabilization and reduce bubble
fly ash is a potential source of rare earth metals, which can be used cracking. The impacts of dry density, NaOH, polypropylene fibers
for making geopolymer that resemble a cement-like product [4]. and aggregates on the compressive strength and thermal conduc-
The geopolymer technology provides an alternative good solution tivity were reported. In addition, the effects of fiber length, aggre-
to the utilization of fly ash with minor negative impact on the envi- gates and hydrophobic agent on the moisture absorption
ronment [5]. McLellan et al [6] presents an examination of the life- properties were also investigated.
cycle cost and carbon impacts of ordinary Portland cement and
geopolymers in Australian. Results show that the geopolymer con- 2. Methodology
crete mixes based on typical feedstocks indicate potential for a 44–
64% reduction in greenhouse gas emissions. 2.1. Experimental materials
A considerable amount of research has been undertaken to
make use of the industrial wastes such as fly ash to produce ther- 2.1.1. Foaming agent
mally and mechanically enhanced construction materials to reduce Owing to the excellent performance of the surface activity and
the negative environmental impact of the industries using various the surface tension of liquid, foaming agent generated by animal
approaches associated with different ingredients and preparation protein was employed in this study. The performance of bubbles
processes. The compressive strength and thermal conductivity of caused by foaming agent strongly depends on foam expansion, sta-
lightweight geopolymer concrete were observed to be reduced by bility and hydrophobic property. According to the Chinese Stan-
the reduction of the density [7]. Palm shell [8,9], fly ash [3,10– dard JG/T266-2011 for foamed concrete, these three parameters
14] and wood fiber [15] have been used to produce the lightweight should accord with the following requirements shown in Table 1.
geopolymer concrete, which showed better thermo-mechanical The actual parameters of foaming agent in this study were also
properties comparing with the normal Portland cement foam con- presented in Table 1, from which the technical index of the foam-
crete with the same density. The thermo-mechanical properties of ing agent used in the experiments can be observed to completely
the product are also affected by ingredients and manufacturing meet the recommended technology parameters.
process. Huang et al [16] concluded that the optimal ratio for ther-
mal properties of foam, glass and sodium hydrate was 35:35:5, and
2.1.2. Fly ash
curing temperature was 55 °C. The contents of fine aggregate can
According to the Chinse Standard of GB/T1596-2017, fly ash of
significantly reduce the compressive strength, which was not
Class II from a coal mine company was activated by using alkaline
desirably affected by the increase in molarity [17].
activator solution in this experiment. X-Ray Diffraction (XRD) was
Apart from the aforementioned methods, fine glass particles,
employed to test the minerals of fly ash. It can be seen from Fig. 1
expanded polystyrene and other solid waste such as bottom ash,
that the main mineral components of fly ash are mullite, silliman-
crumb rubber, clay brick and pumice aggregates have also been
ite, quartz. As is generally known, the activity of fly ash strongly
utilized in attempts to enhance the mechanical and thermal per-
depends on the contents of the aluminum-silicon materials such
formance of lightweight concrete products and the researching
as mullite, sillimanite and quartz. The X-ray fluorescence (XRF)
findings showed it is easy to reduce thermal conductivity but
was used to accurately determine the detailed chemical composi-
mechanical strength can be compromised sometimes, which is
tions of fly ash and its contents. As illustrated in Table 2, the main
similar with the effect of aforementioned methods using fly ash.
chemical compositions of fly ash are Al2O3 and SiO2 with mass con-
Specimen prepared by recycled fine glass particles was observed
tents of 26.2% and 55.2% respectively. In addition, owing to the
to show better strength and thermal performance than that with
lower contents of calcium, the degree of polymerization of the fly
sand aggregate [18]. Colangelo et al [19] found that the compres-
ash is very high and therefore alkaline solution is necessary for
sive strength and bending strength decreased with the increased
the structural disintegration.
content of polystyrene, and the thermal conductivity was signifi-
As shown in Fig. 1 (a) and (b), the microstructure of the original
cantly reduced in comparison to that of the samples containing
fly ash was examined by the Scanning Electron Microscope (SEM),
normal weight micro-silica sand. Mechanical and thermal proper-
from which it can be seen that fly ash consists of spherical particles
ties of lightweight geopolymer mortar with locally available waste
in various sizes. These particles, with radius smaller than 10 lm,
materials including bottom ash, crumb rubber, clay brick and
are usually hollow and may contain smaller particles in their inte-
pumice aggregates were investigated by Wongsa et al [20–22],
rior spatial structure [26]. The surface of fly ash particles appears to
which exhibited better thermal insulation and fire resistance than
be smooth, and some vitreous and quartz particles can also be
that of normal aggregates. The production process can also be chal-
observed [24].
lenging, for instance, Sanjayan et al [23] indicated that the foaming
reaction is too fast to prevent complete alkali activation of
geopolymers and therefore many unreacted fly ash particles Table 1
remains in highly aerated specimens leading to poor mechanical Recommended and actual parameters of the foaming agent.
strength. In order to overcome that issue, Abdullah et al [24,25] Parameter Foam expansion Stability (1 h) Hydrophobic
proposed that the geopolymer lightweight concrete samples cured property (1 h)
at 60 °C produced the maximum compressive strength. Interest- Definition Ratio of the volume The descent The volume of
ingly, the water absorption and porosity were reduced by 6.78% of foam with volume length of the the liquid from
and 1.22% after 28 days. of liquid agent foam column cracked bubbles
This paper presents an experimental investigation on the Recommended >20 10 mm 80 mL
Actual 30 8.8 mm 70 mL
thermo-mechanical and moisture absorption characterization of
 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960 3

lized to improve the foam stabilization, reduce bubble cracking

and enhance the waterproof capacity.

2.2. Experimental scheme

2.2.1. Experimental programme

According to the pre-experimental results, the mass ratio of
total water to fly ash was set as 0.4. It should be noted that the
above-mentioned total water includes the water in alkaline solu-
tions and additional portion added into the concrete. In addition,
the mass ratio of fly ash to sodium silicate was set as 4. The mass
ratio of NaOH to fly ash in the range of 5%-20% was made to inves-
tigate the impact of the amount of NaOH on the thermo-
mechanical properties of lightweight geopolymer concrete. The
amount of aggregate and PF is based on the mass of geopolymer
concrete. The detailed experimental programme is presented in
Table 6.

Fig. 1. XRD and SEM of fly ash. 2.2.2. Experimental procedures

The experimental procedures for producing the fiber-reinforced
FLGC include five steps: (1) raw material preparation, (2) foaming,
2.1.3. Alkaline solution
(3) stirring, (4) casting mould/curing and (5) demoulding. More
The mesh structures of silicon and aluminum materials in the
details about the procedures are listed in Table 7.
fly ash need to be disintegrated and activated by strong alkaline
solution to produce the geopolymer concrete. The sodium hydrox-
ide and sodium silicate solutions with mass concentration of 27.5% 2.2.3. Testing method and data processing
and modulus of 2.3 were prepared as the alkaline activator in this The wet density, apparent density and dry density of samples
study. were calculated using the ratio of the samples’ mass to the mould’s
Sodium silicate also called water glass which is a compound volume. The 28 days compressive strength of samples was evalu-
containing sodium oxide (Na2O) and silicon dioxide (SiO2) that ated according to GB/T11969-2008. The two ends of all the sam-
forms a glassy solid with the very useful property of being readily ples were grinded to be in parallel and normal to the height to
soluble in water. The higher the ratio of SiO2 to Na2O and the avoid the stiffening effect that could be caused by uneven end sur-
higher the concentration of both ingredients, the more viscous the faces of the sample during the initial stage. The results reported
solution. The various grades of sodium silicate are characterized by herein are the average of the three measurements of the strength.
their SiO2:Na2O molar ratio. Grades with molar ratio below 2.85:1 The compressive strength was tested by an electro-hydraulic servo
are generally termed alkaline. In this study, the molar ratio of testing machine with 2000 kN capacity.
sodium silicate solution is 2.3 so the sodium silicate solution can The microstructural characterization was examined with the
be used as alkaline solution. aid of SEM and 3D Digital Microstructure (3D-DM) respectively.
XRD and XRF were used to characterize the detailed minerals
and chemical compositions of samples. The diffraction patterns
2.1.4. Polypropylene fiber were analyzed and identified with the assistance of the corre-
Due to the high performance of abrasion resistance, corrosion sponding software program. Thermal conductivities of samples
resistance and fire resistance, PF has been usually chosen to rein- were measured by a thermal constants analyser based on the the-
force the mechanical strength of lightweight concrete such as com- ory of the transient plane source. In addition, the thermal conduc-
pressive and tensile strengths but without obviously increasing the tivity, thermal diffusivity as well as specific heat per unit volume
corresponding thermal conductivity. The mechanical parameters can be obtained from one single transient recording and the agree-
of the PF are illustrated in Table 3, which shows that the tensile ment was considered exceptionally good compared with any other
strength of the PF is much larger than the compressive strength experimental technique [27].
of lightweight concrete. Consequently, the PF is an excellent mate- Water content tests were conducted through weighting the
rial to reinforce the mechanical properties of lightweight concrete. mass of samples with different moisture according to JGJT70-
2009. The mass ratio between hydrophobic agent and water was
2.1.5. Aggregate and additive agents 1:10. All the demoulded samples were put into an oven with tem-
Standard sand and ceramsite were served as the fine and coarse perature of 65 ± 2 °C for 24 h, and then dried in the oven with tem-
aggregates, and the corresponding physical properties are listed in perature of 105 ± 5 °C for another 24 h. To avoid damage to the
Tables 4 and 5 respectively. surface waterproofing layer, the dried samples were conducted
In addition, modified polyethoxylated silicone and silicone- by surface waterproofing and then naturally dried for 24 h. The
based hydrophobic agents were used as the foam-stabilizing agent dried samples with and without surface waterproofing were put
and hydrophobic agent in this study. Silicone-based hydrophobic into an environment chamber with constant temperature and rel-
agents can attach very readily to surfaces and have a very high ative humidity. When the moisture content test was conducted,
spreadability. Foam-stabilizing and hydrophobic agent were uti- the samples should be removed from the chamber and weighted

Table 2
Chemical compositions of fly ash.

Molecular formula Al2O3 SiO2 Fe2O3 K2O CaO MgO K2O Na2O
Contents (%) 26.2 55.2 3.106 1.879 1.73 1.45 2.857 0.35
4 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960

Table 3
Parameters of polypropylene fiber.

Density (g/cm3) Diameter (mm) Tensile strength (MPa) Modulus of elasticity (MPa) Elongation at break (%)
0.91 0.017 461 4987 19

Table 4
Physical properties of standard sand.

Grain density (kg/m3) Dry density (kg/m3) Mean grain size (mm) Void ratio Nonuniform coefficient
2643 1430–1740 0.39 0.52–0.85 1.542

Table 5
Physical properties of coarse aggregate.

Particle size (mm) Bulk density (kg/m3) Apparent density (kg/m3) Water absorption-1 h (%) Cylinder compressive strength (MPa)
5–10 520–540 780–800 8.7 2.8

Table 6
Experimental programme.

No. Fly ash (g) Na2SiO3 (g) NaOH (%) Aggregate (%) Foam (L) PF (%) Fiber length (mm)
A 6000 1500 10 – 0–12 0.5 3
B 6000 1500 5/10/15/20 – 6 0.5 3
C 6000 1500 10 – 6 0.5 0/3/6/9/12/19
D 6000 1500 10 – 6 0/0.5/1.0/1.5 3
E 6000 1500 10 – 6 0/0.5/1.0/1.5 12
F 6000 1500 10 0/5/10/15 (fine) 6 0.5 3
G 6000 1500 10 0/5/10/15 (coarse) 6 0.5 3
H 6000 1500 10 – 6 0.5 3/6/12/19
I 6000 1500 10 – 6 0.5/1.0/1.5/2.0 3

Table 7 Table 8
Details of experimental procedures. Accuracies and uncertainties of testing variables.

Steps Detailed procedures Variables Thermal conductivity Compressive strength Mass

(a) Raw material The mass of fly ash and sodium silicate were kept Testing accuracy ±2% ±1% ±0.01 g
preparation constant at 6000 g and 1500 g in this study Uncertainty ±5% ±7% –
respectively. The sodium hydroxide was mixed with
water and then cooled to room temperature. Aggregate
and PF are then measured by an electronic scale. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
(b) Foaming The foam agent was fed into the foaming machine with uv ¼ D2v þ r2v ð1Þ
optimal foaming pressure of 0.5 MPa and foam
expansion ratio of 1:40 to produce the foam which was
where uv is uncertainty of directed variables, Dv is the test accuracy
measured by a measuring glass.
(c) Stirring The pre-measured raw materials including fly ash, of the variables, rv is the Bessel equation of standard deviation and
sodium hydroxide solution, sodium silicate solution, PF the equation is shown:
and foam were orderly added into the blender to stir vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uN  
for about 8–9 min. uP  2
u
(d) Casting mould The slurry was poured into
t i xi  x
and curing 100 mm  100 mm  100 mm plastic molds. The wet
rv ¼ ð2Þ
density of samples was tested. The samples with plastic N1
molds were moved to a curing chamber with relative
humidity of no less than 90% and temperature of where xi and x are individual testing values and the mean value of
20 ± 1 °C for 24 h. individual testing values, N is the number of testing items.
(e) Demoulding After the curing, these samples were demoulded and
The measurement uncertainties of thermal conductivity and
then placed at room temperature to wait for the
experimental tests. compressive strength are ±5% and ±7% based on calculated data
respectively, which can ensure the accuracy of the experimental
set. In addition, the standard deviation is also calculated to quan-
by an electronic balance with accuracy of ±0.01 g. In order to tify the divergence of experimental results.
reduce the influence of indoor environment on the moisture con-
tents, each specimen was tested only once. 3. Results and discussions
To ensure the accuracy of the experimental results, the testing
accuracy and uncertainties of the experimental setup are analyzed 3.1. Thermal conductivity and compressive strength
and presented in Table 8. The uncertainties were computed
through the testing accuracy and the Bessel equation of standard 3.1.1. Dry density
deviation. The equations of uncertainties for directed variables Table 9 shows the variation of dry density with foam contents
are presented by [28]: (group A in Table 6). It is clearly that the dry density of the light-
weight geopolymer concrete is nonlinearly decreased by the
 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960 5

Table 9
Dry density versus foam volumes.

Foam volumes (L) 0 2 4 6 8 10 12

Dry density (kg/m3) 1722 1503 1299 858 662 554 498

increased volume of foam added into the FLGC. The amount of tigations which are shown in Fig. 2 (a, b, c & d). It can be seen from
added foam should be carefully selected in order to acquire the the microstructural images of four samples that most air voids are
lightweight geopolymer concrete with excellent compressive approximately spherical in shape. These air voids are not of uni-
strength and thermal conductivity. form size and the diameter of the largest air voids was observed
Thermal conductivity and compressive strength, which are the to be increased by the decreased dry density. The air voids in four
two most important physical properties of thermal insulation samples with different dry densities consist of two types of air
materials, strongly depend on the dry density of the samples, bubbles which are called closed and merged air voids respectively.
which results from the volume of foam added into the FLGC. The The closed air voids are those that are fully covered with the binder
thermal conductivity and compressive strength of samples with paste, and the merged air bubbles are those resulted from the com-
different dry densities are illustrated in Fig. 2, in which it can be bination of more than two adjacent air voids [29]. As illustrated in
seen that the compressive strength and thermal conductivity non- Table 9, the dry density of samples was decreased when the vol-
linearly reduced as the dry density decreased. It was also observed ume of the foam added into the FLGC increased, which suggests
that the compressive strength and thermal conductivity signifi- that the porosity of samples has a strongly positive correlation
cantly decreased when the dry density was reduced in the range with the dry density. As the dry density of samples decreases,
of 858 kg/m3–1722 kg/m3. For example, the compressive strength the distances between two or more than two air voids decreased,
and thermal conductivity of the sample with dry density of while the number of those adjacent air bubbles increased. Conse-
1722 kg/m3 was 7.20 MPa and 1.14 W/(mK), which was 4.1 and quently, the probability of merging adjacent air bubbles to form
5.4 times larger than those of the one with dry density of air voids of larger diameter was considerably increased. As shown
858 kg/m3 respectively. The reason is that the connection between in Fig. 2(a) and (d), the diameter of the largest air bubble in the
FLGC gels was broken by the air bubbles as the foam was added sample with dry density of 554 kg/m3 can be up to 1300 lm which
into the FLGC. The compressive strength of FLGC samples was is about six times larger than that formed with a dry density of
shown to be apparently decreased by the decreased dry density. 1299 kg/m3.
The air bubbles in the samples were full of air with lower thermal Fig. 3 demonstrates the distribution of the diameter of air bub-
conductivity than that of FLGC. Hence, the effective thermal con- bles in the samples with different dry densities shown in Fig. 2 (a,
ductivity of samples also considerably decreased when the dry b, c & d) which was counted using the software of 3D-DM. As can
density decreased. The decreasing rate of thermal conductivity be seen from Fig. 3, the air voids with diameter in the range of 300–
and compressive strength of FLGC samples were not so significant 500 lm account for the largest proportion, and decrease with
when the dry density was lower than 858 kg/m3. It was observed decreasing dry density. The air bubbles with diameter larger than
that the compressive strength and thermal conductivity respec- 1000 lm were only observed in samples with dry density lower
tively decreased from 1.75 MPa and 0.21 W/(mK) to 0.95 MPa than 858 kg/m3. The reason was that the adjacent air bubbles with
and 0.095 W/(mK) when the dry density was decreased from smaller diameters merged to generate air voids with larger diam-
858 kg/m3 to 498 kg/m3. Both approached the lower limits when eter. The air bubbles with larger diameter could be therefore
the dry density decreased to around 498 kg/m3, at which point it observed in the samples with lower dry density. Although the per-
is not feasible to further reduce the thermal conductivity by centage of the air bubbles with larger diameter was obviously less
increasing the volume of the foam which leads to decreased dry than that with smaller diameter, the area or volume occupied by
density and compressive strength. air voids with larger diameter were hundreds of thousands of
3D-DM was conducted to investigate the pore structures of the times larger than those of smaller air bubbles.
FLGC samples. Four cubic samples with density of 554 kg/m3, According to the variations of compressive strength and ther-
662 kg/m3, 858 kg/m3 and 1299 kg/m3 were selected for the inves- mal conductivity of FLGC shown in Fig. 2, the samples with dry
density ranging in 500–600 kg/m3 were prepared to investigate

60
1299 kg/m3
52.2
858 kg/m3
50 47.7
662 kg/m3
44.2
554 kg/m3
40 39.6
percentage (%)

32.631.8
30.2
30
27.1

20.8
20
16.3
13.6
10.9
10
6.3 6.3
4.7 4.5 4.7
2.2 2.3 2.2
0
200-300 300-500 500-700 700-900 >900
pore size (μm)
Fig. 2. Variations of thermal conductivity and compressive strength of FLGC with
dry density. Fig. 3. Pore size distribution with different dry densities.
6 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960

the variation of thermo-mechanical and moisture absorbing prop- that the patterns of FLGC are similar to that of fly ash, which indi-
erties hereafter. cates that the degree of amorphous and crystallization of fly ash
was not noticeably changed by geo-polymerization. The amount
3.1.2. NaOH contents of crystal of quartz and mullite from geopolymer which mainly
As the main alkaline activated materials, the contents of NaOH consisted of amorphous aluminosilicate products was similar or
is commonly believed to be one of the most significant factors very slightly increased to those from fly ash [30]. There are two
affecting the thermo-mechanical properties of lightweight main differences between XRD patterns of lightweight geopolymer
geopolymer concrete [30,31]. Fig. 4 demonstrates the influence of concrete and fly ash. The first one is the shift of amorphous silli-
the mass ratio of NaOH to fly ash on the compressive strength manite peak from around 25-28° for FLGC and fly ash, which indi-
and thermal conductivity of the lightweight geopolymer concrete cates that the silicate glass phase in FLGC was highly disordered.
(group B in Table 6). An increase in the thermal conductivity of Another point to be aware of is graphics of FLGC at 5-8° curved
FLGC samples was observed when NaOH contents increased. It slightly upwards compared to that of fly ash. This could be the for-
appears that the effects of NaOH contents on compressive strength mation of meso-materials of poorly crystalline nature [34].
in this study differed from previous observation, which suggested
there is a monotonic increasing or decreasing relationship between 3.1.3. Polypropylene fiber
compressive strength and NaOH concentration [32]. However, the Fiber reinforcement in lightweight concrete was widely
results herein revealed the compressive strength was increased believed as one of the most effective way to improve its properties,
firstly and then decreased as the contents of NaOH increased from such as flexural capacity, toughness, post-failure ductility and
5% to 20%. In details, the maximum compressive strength 1.38 MPa crack control [35]. Fibers can be categorized as metallic, glass,
was achieved when the NaOH content was 10%. When the contents polymeric, carbon, mineral and asbestos. Among the various types
of NaOH were larger than 10%, the compressive strength of the of fibers, polypropylene fiber is the most commonly used for ther-
samples was observed to be decreased by increased NaOH contents mal insulation purposes. Other reasons for the greater usage of PF
which agrees with the experimental results in previous literatures also include economical advantage and excellent resistance to
[33]. environmental aggressiveness.
Lower NaOH contents resulted in the insufficient geo- Fig. 7 presents the impact of PF length on the improvement of
polymerization between the fly ash and alkali solutions. As shown compressive strength of lightweight geopolymer concrete on
in Fig. 5 (a), the fly ash was clearly seen not to be completely dis- 28 days (groups C/D/E in Table 6). It is seen from the figure that
solved by alkali solutions, and a large number of micro-cracks the compressive strength of fiber reinforced FLGC with fiber
which could limit the strength of geopolymer were also found in lengths of 3 mm, 6 mm, 9 mm, 12 mm and 19 mm was increased
the matrix. Owing to the insufficient geo-polymerization, the com- by 57%, 46%, 57%, 71% and 6% respectively comparing to the sam-
pressive strength of final products with NaOH contents of 5% was ples without PF. Hence, it suggests that the compressive strength
therefore not high. Dissolution of fly ash was accelerated when strongly depends on the fiber length. The high performance in
the NaOH was sufficient, which in this case study is 10%. Fly ash compressive strength of fiber reinforced lightweight geopolymer
particles were not easily noticed as most particles were fully dis- concrete might be achieved by the PF that mechanically interacted
solved and covered with geopolymer gel, as shown in Fig. 5 (b). with the FLGC. When the uniaxial load applied on the sample
It formed a continuous mass of gel which resulted in a relatively exceeded its peak stress, cracks were observed on the surface
dense geopolymer paste with higher compressive strength [30]. and interior of the cubic specimen. As typical observation from
As shown in Fig. 5 (c, d), the excess hydroxide ion concentration fiber reinforced FLGC specimens shown in Fig. 8, crack tips in FLGC
from higher NaOH contents led to aluminosilicate gel precipitation samples would be concatenated by fibers, resulting in decreasing
at the very early stages [32,33]. The leaching of silicon and alu- the number of cracks and blocking the propagation of existing
minum in fly ash was hindered. The geo-polymerization was con- cracks [36]. The compressive strength of lightweight geopolymer
sequently obstructed which resulted in negative impact on the concrete was therefore increased by PF.
compressive strength. Residual compressive strength of cracked fiber reinforced FLGC
The comparison of XRD results between the FLGC samples with provides a clearer comparison of the post cracking behaviors. As
different NaOH contents and fly ash is presented in Fig. 6. It is clear seen from Fig. 8, the specimen without fibers had the lowest resid-
ual compressive strength owing to the deficiency of fiber-concrete
bond. The post-failure compressive strength of specimen without
PF was decreased to 25% of the peak strength with the axial dis-
placement of 5 mm. Meanwhile, the post-failure compressive
strength of samples with PF of 12 mm for the second and third load
strength reduced to 90%, 91% with axial displacement of 1.4 mm
and 1.9 mm respectively, which is about 1.55 times larger than
the peak strength of specimen without fibers. This might be attrib-
uted to that the bridging effect of fibers at the crack face is capable
of preventing cracks from propagating [37].
In addition, the failure forms of fiber reinforced specimens and
samples without fibers significantly varied, as shown in Fig. 8. Ver-
tical cracks appeared first around the mid-height of the cubic FLGC
specimen without fibers, and diagonally propagated to four corners
with a further increasing uniaxial loads. Obvious spalling frag-
ments of FLGC without fibers were observed under the peak com-
pressive loads. This kind of failure configuration of FLGC without
fibers was truncated pyramids or brittle failure [38]. It is seen that
there was no obvious spalling concrete in fiber reinforced FLGC
compared with those samples without fibers. This may be related
Fig. 4. Thermal conductivity and compressive strength versus NaOH contents. to the bridge action of fibers, indicating that the lateral deforma-
 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960 7

Fig. 5. SEM of FLGC versus NaOH contents.

Fig. 6. The XRD of FLGC versus NaOH contents.

Fig. 7. Effects of fiber length on compressive strength.

tion could be constrained by the addition of fibers. The data in

Fig. 7 and Fig. 8 verified that the addition of fibers not only in Fig. 9. The compressive strength of fiber reinforced FLGC first
improved the compressive strength of FLGC, but also changes its
increased and then decreased when the PF contents increased from
failure configuration [39]. zero to 2%, with a critical point for the PF contents of 0.5%, where
The variations of compressive strength and thermal conductiv-
the internal structure of FLGC was considerably improved. Similar
ity of fiber reinforced FLGC with various PF contents are presented
8 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960

of additive. Increase in the contents of coarse aggregate from 0% to

15% resulted in the reduction of the thermal conductivity and dry
density from 0.17 W/(mK) to 0.086 W/(mK) and 650 kg/m3 to
480 kg/m3 respectively. However, the opposite effect was given
by fine aggregate. As shown in Fig. 10 (b), both the thermal con-
ductivity and dry density were increased by the increase of con-
tents of fine aggregate. For instance, increase in the contents of
fine aggregate from 0% to 15% resulted in the increases of the ther-
mal conductivity and dry density from 0.17 W/(mK) to 0.21 W/
(mK) and 550 kg/m3 to 660 kg/m3 respectively.
Fig. 11 demonstrates the variation of compressive strength with
the contents of coarse and fine aggregates. It is clear that the com-
pressive strength of samples strongly depended on the amount of
standard sand (fine aggregate). Increase in the contents of standard
sand from 0% to 15% resulted in the increases of the compressive
strength from 1.38 MPa to 2.6 MPa. However, the increase in the
contents of ceramsite (coarse aggregate) from 0% to 15% did not
significantly change the compressive strength.
The bulk density of ceramiste and standard sand was about
400 kg/m3 and 1400 kg/m3 respectively. Considering the dry den-
Fig. 8. Stress-displacement curves for different fiber lengths. sity of the FLGC lies between the bulk densities of ceramsite and
standard sand, adding the coarse or fine aggregates into the
geopolymer concrete would therefore produce samples with quite
variation for thermal conductivity with PF contents is also
different dry densities. Increase in the contents of coarse aggregate
observed in Fig. 9 (b). The main reason for the decrease in the com-
resulted in decrease of dry density, whilst increase of fine aggre-
pressive strength when the PF contents is greater than 0.5% is that
gate resulted in increase of dry density. As plotted in Fig. 2, it
the dispersion of fiber, especially in high volume fractions is very
can be seen that dry density of the samples has a direct impact
difficult and consequently causes poor workability and incomplete
on the compressive strength, while the effects of coarse and fine
compaction [39].
aggregates on the compressive strength are essentially different.
Besides, the compressive strength and thermal conductivity of
Similar explanation could also be applied to illustrate the variation
fiber reinforced FLGC with fiber length of 12 mm were observed
of thermal conductivity with the contents of coarse and fine
to be slightly higher than that of 3 mm. The contact area between
aggregates.
fibers and geopolymer concrete increased with increasing fiber
length, which resulted in larger frictional force and preferable per-
formance of bridge effects of longer fibers. More air voids could be
linked through longer fibers compared with shorter fibers. The
3.2. Moisture absorption
thermal conductivity of samples with longer fibers was therefore
increased.
One of the most important motivations for investigating the
variation of moisture absorption is that the thermal conductivity
3.1.4. Aggregates of FLGC samples strongly depends on the moisture contents. Com-
Ceramsite fabricated by domestic sludge and standard sand in pared to the normal weight concrete, the lightweight concrete had
the city of Xiamen, was used as the coarse and fine aggregates higher moisture absorption owing to the lower dry density.
respectively. The change of dry density and thermal conductivity According to our previous testing results, the thermal conductivity
of FLGC with ceramsite and standard sand is illustrated in Fig. 10 of FLGC specimen with mass water contents of 0%, 6% and 8% was
(groups F/G in Table 6). It is obvious from Fig. 10 (a) that both gradually increased from 0.17 W/(mK) to 0.28 W/(mK) and
the thermal conductivity and dry density of the samples with 0.39 W/(mK) respectively. Therefore, it is necessary to study the
coarse aggregate were decreased steadily by the increased amount factors that affect the moisture contents.

Fig. 9. Compressive strength and thermal conductivity versus fiber contents.

 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960 9

Fig. 10. Thermal conductivity and dry density versus aggregate contents.

fiber contents respectively. Consequently, the moisture absorption

capability of samples could be improved by mixing longer fibers
and larger amount of fibers into the FLGC.

3.2.2. Aggregates
Fig. 13 shows the impact of fine and coarse aggregates on the
moisture absorption capacity of FLGC samples (groups F/G in
Table 6). As shown in Fig. 13 (a), an increase in the contents of fine
aggregate reduced the moisture absorption of FLGC specimens,
which agrees with the finding shown in Fig. 10 (b), i.e. the dry den-
sity of FLGC samples was gradually increased when contents of fine
aggregate increased. The larger dry density was attributed to lower
volumes of pores and larger fraction of standard sand. The FLGC
specimens with high dry density contained lower amount of air
voids, and this defective porous structure consequently led to
poorer moisture absorbing capacity. However, Fig. 13 (b) shows
that increasing the amount of ceramsite resulted in higher mois-
ture absorption. As shown in Fig. 10 (a), the dry density was
decreased when the amount of coarse aggregate increased. The
FLGC samples with lower dry density contained higher volume of
Fig. 11. Compressive strength versus aggregate contents. coarse aggregates and air voids, which in consequence absorbed
more moisture owing to their highly porous structures [2]. In addi-
tion, another reason was due to the physical properties of the cer-
3.2.1. Polypropylene fiber amsite aggregates which have higher moisture absorption capacity
Fig. 12 demonstrates the effects of fiber length and fiber con- due to increased specific surface area. Similar conclusions were
tents on the moisture absorption of FLGC samples (groups H/I in drawn by Aslam et al [40] and Ahmmad et al [41] that water
Table 6). The FLGC samples with added fibers in different length absorption of the concrete increased gradually by increasing the
were put in the chamber with dry bulb temperature of 35 ± 0.3 ° added amount of ceramite in lightweight concrete.
C and relative humidity of 80 ± 1.5%, whilst the specimens mixed
with 3 mm fibers at various levels of contents were placed in the
chamber with dry bulb temperature of 35 ± 0.3 °C and relative 3.2.3. Hydrophobic agent
humidity of 90 ± 1.5%. Hydrophobic agent, which is surface protection materials and
As shown in Fig. 12, the moisture contents of FLGC sample in capable of increasing the angle of contact between the water dro-
the chamber with relative humidity of 90% is 3.14% at 19.1 h, com- plet and the concrete surface [42], can be used to reduce liquid
pared with moisture contents of 2.33% at 19.3 h with relative water penetration into the concrete and the thermal property of
humidity of 80%. Hence, the moisture contents of FLGC samples concrete is also not deteriorated. The hydrophobic agent with
was closely linked to the surrounding thermodynamic parameters diluted concentration by 10 times was homogeneously sprayed
including relative humidity. As shown in Fig. 12 (a), the moisture onto the surface of the FLGC samples, and all specimens were nat-
contents increased when the fiber length increased. In addition, urally dried at room temperature. The waterproofing specimens
the increase in the fiber contents resulted in higher moisture were then put into the chamber with dry bulb temperature of
absorption capacity, as presented in Fig. 12 (b). It is widely 35 ± 0.3 °C and relative humidity of 90 ± 1.5% to investigate the
accepted that the excellent performance of water absorption was effect of waterproofing treatment on moisture absorption of FLGC
achieved by the capillary effect of fibers in the FLGC. The fibers samples.
mixed into the FLGC samples were equivalent to the water- Fig. 14 presents the variation of moisture contents of water-
conducting channels. The moisture in the surrounding environ- proofing specimens with different fiber contents and the compar-
ment could be continuously diffused into the FLGC samples along ison of moisture contents of waterproofing sample with that of
the fibers. The length and number of the water conducting chan- original specimen with the same fiber length of 3 mm (group D
nels was believed to increase with the increased fiber length and in Table 6). It is observed that the moisture absorption of water-
proofing samples was considerably reduced compared with origi-
10 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960

Fig. 12. Moisture contents versus fiber length and fiber contents.

Fig. 13. Moisture contents versus aggregate contents.

moisture contents of waterproofing samples with fibers contents

of 1.0% and 1.5% with those of original specimen with same fiber
contents shown in Fig. 12 (b). The waterproofing treatment is
therefore the most effective method to reduce the moisture
absorption of FLGC samples.

4. Conclusions

An experimental investigation on the thermo-mechanical and

moisture absorption properties of lightweight geopolymer con-
crete prepared with fly ash, NaOH, sodium silicate solutions and
polypropylene fibers was presented in this study. The main find-
ings are as follows:

(1) The dry density of the lightweight geopolymer concrete non-

linearly decreased when the foam volume increased. The
compressive strength and thermal conductivity generally
nonlinearly reduced as the dry density decreased, which is
probably intuitive. But it was found out that it is not feasible
Fig. 14. Moisture contents of waterproofing samples versus fiber contents. to further reduce the thermal conductivity by increasing the
volume of the foam as it would lead to poor binding because
of excessively lower compressive strength when the dry
nal samples shown in Fig. 12 (b). In addition, the effect of fiber con-
density decreased to about 498 kg/m3.
tents on moisture absorption capacity of waterproofing samples
(2) The compressive strength increased firstly and then
also drastically decreased. For instance, the moisture content of
decreased when the contents of NaOH increased from 5%
waterproofing samples with fiber content of 0.5% was 0.93% at
to 20%, the turning point was found to be at 10%. The exces-
23.7 h, compared with the moisture content of 3.52% for the
sive hydroxide ion concentration from higher NaOH con-
untreated specimen with the same fiber contents at 23.7 h. Similar
tents led to aluminosilicate gel precipitation, which
conclusions could also be drawn through the comparison of the
 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960 11

consequently weakened the compressive strength. It was References:

observed that the thermal conductivity was consistently
increased by the increased contents of NaOH. Therefore, [1] R. Ruparathna, K. Hewage, R. Sadiq, Improving the energy efficiency of the
existing building stock: a critical review of commercial and institutional
NaOH can be used to enhance the compressive strength buildings, Renew. Sust. Energy Rev. 53 (2016) 1032–1045.
rather than thermal insulation performance. [2] M.R. Ahmad, B. Chen, S.F.A. Shah, Investigate the influence of expanded clay
(3) The compressive strength of fiber reinforced FLGC with fiber aggregate and silica fume on the properties of lightweight concrete, Constr.
Build. Mater. 220 (2019) 253–266.
lengths of 3 mm, 6 mm, 9 mm, 12 mm and 19 mm was [3] Z.H. Zhang, J.L. Provis, A. Reid, H. Wang, Geopolymer foam concrete: an
respectively increased by 57%, 46%, 57%, 71% and 6%. This emerging material for sustainable construction, Constr. Build. Mater. 56 (2014)
suggests the addition of fiber is able to enhance the com- 113–127.
[4] J. Davidovits, Geopolymers: inorganic polymeric new materials, J. Therm. Anal.
pressive strength of FLGC owing to the bridging effect of
Calorim. 37 (8) (1991) 1633–1656.
fibers at the crack face. In addition, the enhancement is less [5] X.Y. Zhuang, L. Chen, S. Komarneni, C.H. Zhou, D.S. Tong, H.M. Yang, W.H. Yu, H.
obvious when the fiber length exceeds 12 mm as the longer Wang, Fly ash-based geopolymer: clean production, properties and
polypropylene fibers were difficult to uniformly disperse in applications, J. Clean. Prod. 125 (2016) 253–267.
[6] B.C. McLellan, R.P. Williams, J. Lay, A. van Riessen, G.D. Corder, Costs and
the FLGC which resulted in the anisotropic sample. This sim- carbon emissions for geopolymer pastes in comparison to ordinary portland
ilar finding was also applicable to the thermal conductivity, cement, J. Clean. Prod. 19 (9–10) (2011) 1080–1090.
as both parameters first increased and then decreased when [7] R.A. Aguilar, O.B. Diaz, J.I.E. Garcia, Lightweight concretes of activated
metakaolin-fly ash binders, with blast furnace slag aggregates, Constr. Build.
the fiber contents increased from zero to 2%. Mater. 24 (7) (2010) 1166–1175.
(4) Increase in the contents of coarse aggregate from zero to 15% [8] M.Y.J. Liu, U.J. Alengaram, M.Z. Jumaat, K.H. Mo, Evaluation of thermal
resulted in the reduction of the thermal conductivity and dry conductivity, mechanical and transport properties of lightweight aggregate
foamed geopolymer concrete, Energy Build. 72 (2014) 238–245.
density. The opposite trend was observed with the fine [9] M.Y.J. Liu, U.J. Alengaram, M. Santhanam, M.Z. Jumaat, K.H. Mo,
aggregate, i.e. the thermal conductivity increased when the Microstructural investigations of palm oil fuel ash and fly ash based binders
contents of fine aggregate increased because the dry density in lightweight aggregate foamed geopolymer concrete, Constr. Build. Mater.
120 (2016) 112–122.
of samples also increased with the increase of fine aggregate. [10] Z.H. Zhang, J.L. Provis, A. Reid, H. Wang, Mechanical, thermal insulation,
In addition, increase in the content of fine aggregate from thermal resistance and acoustic absorption properties of geopolymer foam
zero to 15% resulted in the increases of the compressive concrete, Cem. Concr. Comp. 62 (2015) 97–105.
[11] P. Posi, C. Ridtirud, C. Ekvong, D. Chammanee, K. Janthowong, P. Chindaprasirt,
strength from 1.38 MPa to 2.6 MPa, whilst the increase in
Properties of lightweight high calcium fly ash geopolymer concretes
the content of coarse aggregate did not apparently affect containing recycled packaging foam, Constr. Build. Mater. 94 (2015) 408–413.
the compressive strength. Coarse aggregate is therefore a [12] D.M.A. Huiskes, A. Keulen, Q.L. Yu, H.J.H. Brouwers, Design and performance
better ingredient for producing thermal insulation material evaluation of ultra-lightweight geopolymer concrete, Mater. Design 89 (2016)
516–526.
owing to its lower thermal conductivity and relative higher [13] B. Nematollahi, R. Ranade, J. Sanjayan, S. Ramakrishnan, Thermal and
compressive strength. mechanical properties of sustainable lightweight strain hardening
(5) It was observed that the moisture content increased when geopolymer composites, Arch. Civ. Mech. Eng. 17 (1) (2017) 55–64.
[14] N.N. Shao, Y.B. Zhang, Z. Liu, D.M. Wang, Z.T. Zhang, Fabrication of hollow
the fiber length increased from 3 mm to 19 mm. Increasing microspheres filled fly ash based foam geopolymers with ultra-low thermal
the fiber contents from 0.5% to 2.0% in the FLGC samples conductivity and relative high strength, Constr. Build. Mater. 185 (2018) 567–
resulted in higher moisture absorption. Both of these two 573.
[15] R.T.T. Fongang, J. Pemndje, P.N. Lemougna, U.C. Melo, C.P. Nanseu, B. Nait-Ali,
variations can be explained by the capillary action which E. Kamseu, C. Leonelli, Cleaner production of the lightweight insulating
resulted from the polypropylene fibers. Increase in the con- composites: microstructure, pore network and thermal conductivity, Energy
tents of fine and coarse aggregates from 5% to 15% reduced Build. 107 (2015) 113–122.
[16] Y.J. Huang, L.L. Gong, L. Shi, W. Cao, Y.L. Pan, X.D. Cheng, Experimental
and increased the moisture absorption respectively, which investigation on the influencing factors of preparing porous fly ash-based
was resulted from the opposite variations of dry density geopolymer for insulation material, Energy Build. 168 (2018) 9–18.
with the aggregates. The moisture content decreased from [17] R.H. Kupaei, U.J. Alengaram, M.Z. Bin Jumaat, H. Nikraz, Mix design for fly ash
based oil palm shell geopolymer lightweight concrete, Constr. Build. Mater. 43
3.5% for sample without waterproofing to 0.94% for speci-
(2013) 490–496.
men with surface waterproofing treatment at 24 h in the [18] A. Hajimohammadi, T. Ngo, A. Kashani, Sustainable one-part geopolymer
chamber with the dry bulb temperature of 35 ± 0.3 °C and foams with glass fines versus sand as aggregates, Constr. Build. Mater. 171
the relative humidity of 90 ± 1.5%. The moisture absorption (2018) 223–231.
[19] F. Colangelo, G. Roviello, L. Ricciotti, V. Ferrandiz-Mas, F. Messina, C. Ferone, O.
can be considerably reduced by surface waterproofing treat- Tarallo, R. Cioffi, C.R. Cheeseman, Mechanical and thermal properties of
ment, and the difference of moisture contents for FLGC sam- lightweight geopolymer composites, Cem. Concr. Comp. 86 (2018) 266–272.
ples with different fiber contents also decreased. Therefore, [20] A. Wongsa, Y. Zaetang, V. Sata, P. Chindaprasirt, Properties of lightweight fly
ash geopolymer concrete containing bottom ash as aggregates, Constr. Build.
surface waterproofing treatment can be applied to the ther- Mater. 111 (2016) 637–643.
mal insulation material with high water absorption proper- [21] A. Wongsa, V. Sata, B. Nematollahi, J. Sanjayan, P. Chindaprasirt, Mechanical
ties which could result in decreased thermal insulating and thermal properties of lightweight geopolymer mortar incorporating
crumb rubber, J. Clean. Prod. 195 (2018) 1069–1080.
capacity. [22] A. Wongsa, V. Sata, P. Nuakiong, P. Chindaprasirt, Use of crushed clay brick and
pumice aggregates in lightweight geopolymer concrete, Constr. Build. Mater.
188 (2018) 1025–1034.
[23] J.G. Sanjayan, A. Nazari, L. Chen, G.H. Nguyen, Physical and mechanical
Declaration of Competing Interest properties of lightweight aerated geopolymer, Constr. Build. Mater. 79 (2015)
236–244.
[24] M.M.A. Abdullah, K. Hussin, M. Bnhussain, K.N. Ismail, Z. Yahya, R.A. Razak, Fly
The authors declare that they have no known competing finan- Ash-based geopolymer lightweight concrete using foaming agent, Int. J. Mol.
cial interests or personal relationships that could have appeared Sci. 13 (6) (2012) 7186–7198.
[25] M.M.A. Abdullah, M.F.M. Tahir, K. Hussin, M. Binhussain, I.G. Sandu, Z. Yahya,
to influence the work reported in this paper.
A.V. Sandu, Fly ash based lightweight geopolymer concrete using foaming
agent technology, Rev. Chim.-Bucharest 66 (7) (2015) 1001–1003.
[26] A. Fernandez-Jimenez, A. Palomo, M. Criado, Microstructure development of
Acknowledgements alkali-activated fly ash cement: a descriptive model, Cem. Concr. Res. 35 (6)
(2005) 1204–1209.
This study was financially supported by the National Natural [27] S.E. Gustafsson, Transient plane source techniques for thermal conductivity
and thermal diffusivity measurements of solid materials, Rev. Sci. Instrum. 61
Science Foundation of China (51978653, 51204170) and National (3) (1991) 797–804.
111 Project (B14021).
12 Y. Wang et al. / Construction and Building Materials 251 (2020) 118960

[28] T.J. Sullivan, Introduction to Uncertainty Quantification, Springer, Switzerland, [36] H.K. Lee, S.Y. Song, Performance characteristics of lightweight aggregate
2015. cellular concrete containing polypropylene fibers, J. Reinf. Plast. Comp. 29 (6)
[29] T.T. Nguyen, H.H. Bui, T.D. Ngo, G.D. Nguyen, Experimental and numerical (2010) 883–898.
investigation of influence of air-voids on the compressive behaviour of foamed [37] M.K. Yew, H. Bin Mahmud, P. Shafigh, B.C. Ang, M.C. Yew, Effects of
concrete, Mater. Design 130 (2017) 103–119. polypropylene twisted bundle fibers on the mechanical properties of high-
[30] U. Rattanasak, P. Chindaprasirt, Influence of NaOH solution on the synthesis of strength oil palm shell lightweight concrete, Mater. Struct. 49 (4) (2016)
fly ash geopolymer, Miner. Eng. 22 (12) (2009) 1073–1078. 1221–1233.
[31] K. Somna, C. Jaturapitakkul, P. Kajitvichyanukul, P. Chindaprasirt, NaOH- [38] T. Wu, X. Yang, H. Wei, X. Liu, Mechanical properties and microstructure of
activated ground fly ash geopolymer cured at ambient temperature, Fuel 90 lightweight aggregate concrete with and without fibers, Constr. Build. Mater.
(6) (2011) 2118–2124. 199 (2019) 526–539.
[32] D. Hardjito, S.E. Wallah, D.M.J. Sumajouw, B.V. Rangan, On the development of [39] G.L. Xue, E. Yilmaz, W. Song, E. Yilmaz, Influence of fiber reinforcement on
fly ash-based geopolymer concrete, Aci. Mater. J. 101 (6) (2004) 467–472. mechanical behavior and microstructural properties of cemented tailings
[33] J. He, Y.X. Jie, J.H. Zhang, Y.Z. Yu, G.P. Zhang, Synthesis and characterization of backfill, Constr. Build. Mater. 213 (2019) 275–285.
red mud and rice husk ash-based geopolymer composites, Cem. Concr. Comp. [40] M. Aslam, P. Shafigh, M.A. Nomeli, M.Z. Jumaat, Manufacturing of high-
37 (2013) 108–118. strength lightweight aggregate concrete using blended coarse lightweight
[34] E. Alvarez-Ayuso, X. Querol, F. Plana, A. Alastuey, N. Moreno, M. Izquierdo, O. aggregates, J. Build. Eng. 13 (2017) 53–62.
Font, T. Moreno, S. Diez, E. Vazquez, M. Barra, Environmental, physical and [41] R. Ahmmad, U.J. Alengaram, M.Z. Jumaat, N.H.R. Sulong, M.O. Yusuf, M.A.
structural characterisation of geopolymer matrixes synthesised from coal (co- Rehman, Feasibility study on the use of high volume palm oil clinker waste in
)combustion fly ashes, J. Hazard. Mater. 154 (1–3) (2008) 175–183. environmental friendly lightweight concrete, Constr. Build. Mater. 135 (2017)
[35] M. Hassanpour, P. Shafigh, H. Bin Mahmud, Lightweight aggregate concrete 94–103.
fiber reinforcement – a review, Constr. Build. Mater. 37 (2012) 452–461. [42] M. Medeiros, P. Helene, Efficacy of surface hydrophobic agents in reducing
water and chloride ion penetration in concrete, Mater. Struct. 41 (1) (2008)
59–71.