Materials and Design 92 (2016) 949–959

Contents lists available at ScienceDirect

Materials and Design

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

Study on the preparation and properties of high-porosity foamed

concretes based on ordinary Portland cement
Jun Jiang a,c, Zhongyuan Lu a,⁎, Yunhui Niu a,b,⁎, Jun Li a, Yuping Zhang a
a
State Key Laboratory Cultivation Base for Nonmetal Composites and Functional Materials, Southwest University of Science and Technology, Mianyang 621010, PR China
b
China Academy of Engineering Physics, Mianyang 621010, PR China
c
Sichuan Huashi Green Homeland Building Materials Co., Ltd., Chengdu 610000, PR China

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

Article history: Fire-resistant, low-cost, high-porosity foamed concretes, which have desirable thermal-insulation properties and
Received 19 May 2015 can be cast in-place, were prepared by adding preformed foam to ordinary Portland-cement slurries. The air-void
Received in revised form 5 December 2015 structure, hardened properties, curing method, and stability of the fresh foamed concrete were investigated. The
Accepted 14 December 2015
results indicated that foamed concretes with porosities of 88.5%–95.4%, compressive strengths of 0.12 MPa–
Available online 17 December 2015
0.75 MPa, and low thermal conductivities of 0.036 W m−1 K−1 to 0.063 W m−1 K−1 could be obtained with
Keywords:
the addition of preformed foam. Moreover, the addition of an accelerator and low dosages of a superplasticizer,
High-porosity foamed concrete as well as a reduction in the water-to-cement ratio prevented the collapse of the fresh foamed concrete. Small
Portland cement air-void sizes and narrow distributions of these sizes were obtained by increasing the apparent density, and by
Air-void structure adding an accelerator and a suitable superplasticizer. In addition, humidity-insulated curing enhanced the
Stability strength of the foamed concrete.
Hardened property © 2015 Elsevier Ltd. All rights reserved.

1. Introduction that have low thermal conductivity, is difficult. Aerated concrete is pre-
pared by adding aluminum powder or hydrogen peroxide to the cement
Building insulation is essential to reducing energy consumption, slurry; foamed concrete is obtained by adding preformed foam or
especially in developing countries, where large-scale urbanization is a foaming agent to the slurry. However, the method of injecting
ongoing [1]. Factors such as fire-resistance, low cost, and durability, preformed foam into the slurry has better thermal-isolation properties
have spurred many researchers toward the development of various than other lightweight concretes [6–14], and this kind of foamed con-
types of lightweight concrete, in order to reduce energy consumption crete can be used at construction sites, and therefore offers more conve-
[1–4]. However, the relatively high density (300–1800 kg m− 3) of nience, and provides better control of the density and pore structure,
these materials, compared with that of other porous organic materials than its aerated counterpart [1,5,15].
such as polyurethane board and expanded polystyrene, has limited However, the high energy associated with the gas–liquid interface
their use as thermal insulators. The thermal insulation performance of and its thermodynamically unstable nature, render the foam unstable
lightweight concrete can be improved by increasing the porosity of [16]. The energy and total interfacial area of the foam may be reduced
these materials [5]. via coalescence and disproportionation, and this reduction leads to in-
Lightweight concrete may be obtained via various methods. These creased air-void sizes and eventual collapse of the structure. Therefore,
include: a) Substitution of lightweight aggregates with low specific the desired density and properties are unattainable. Foamed concrete is
gravity for the cement paste or the normal-weight aggregate. This con- typically subjected to an accelerated hardening process in order to pre-
crete is referred to as lightweight aggregate concrete. b) Inducing the vent the aforementioned coarsening and collapse. Special cements (alu-
formation of air voids in the cement paste, concrete or mortar mass. minate cement, sulphoaluminate cement, and other quick-hardening
This type of concrete is referred to as aerated and foamed concrete. cementitious materials) undergo rapid setting and hence high strengths
c) Eliminating fine aggregates from the mix, such that only coarse can be attained in the initial stages of curing; therefore, these cements
aggregates of ordinary weight remain. This concrete is referred to as are typically used to prepare foamed concrete [17]. However, a shortage
no-fines concrete [6]. Owing to their low porosity and pore structure, of the raw materials required for these cements, high cost, and low du-
realization of lightweight aggregate concrete and no-fines concrete rability have stymied the development of foamed concrete [4].
Therefore, in this work, high-porosity foamed concrete (HPFC) was
prepared by adding preformed foam to a Portland-cement slurry, and
⁎ Corresponding authors. the stability of the HPCF slurry, during the hardening process, was in-
E-mail addresses: jiangjunicl@163.com (J. Jiang), niuyunhui@swust.edu.cn (Y. Niu). vestigated. The air-void structure, compressive strength, thermal-

http://dx.doi.org/10.1016/j.matdes.2015.12.068
0264-1275/© 2015 Elsevier Ltd. All rights reserved.
950 J. Jiang et al. / Materials and Design 92 (2016) 949–959

Table 1 which contained dissolved SA, was then added to the slurry and
Chemical composition of the cement. mixed for 60 s until well-blended pastes were formed. Preformed
Composition (%) SiO2 Al2O3 Fe2O3 CaO MgO SO3 Loss foam (in the amounts listed in Table 2) was then immediately added
Cement 21.06 6.10 3.08 57.98 2.74 2.40 4.07
to the cement slurry and mixed for 60 s, until uniformly distributed
therein. The resulting HPFC was poured into 70.7 × 70.7 × 70.7 mm3
and 300 × 300 × 30 mm3 molds, covered with a clinging film, and
insulation properties and curing method were also determined. The re- cured for 72 h at 20 ± 2 °C and a relative humidity (RH) of 90%, in a
sults obtained in this work may offer insight into the low-cost prepara- chamber. The specimens were removed from the molds and cured for
tion and application of foamed concrete with better thermal-insulation 7 days, 28 days, and 56 days, via standard curing, humidity-insulated
properties than other lightweight concretes. curing, and natural curing. The compressive strength of the specimens
was determined at the end of each curing period. Standard curing was
2. Material and methods performed at a temperature and humidity level of 20 ± 2 °C and N95%
RH, respectively. Humidity-insulated curing was also conducted at
2.1. Materials and preparation 20 ± 2 °C, but the samples were insulated from the curing environment
by using a plastic film. Furthermore, the samples that underwent natu-
HPFC mixtures were prepared from ordinary Portland cement, a set- ral curing were placed outside, where they were exposed to environ-
accelerator (SA), polycarboxylate-based superplasticizer (PC), foaming mental temperatures and humidity levels ranging from 18 °C to 35 °C
agent, and water. The cement (P.O 42.5R, in accordance with Chinese and 30% RH to 95% RH, respectively.
standard: GB175-2007) was produced at the Lafarge Shuangma cement
plant, Jiangyou, Sichuan Province, China. This cement was used to stabi- 2.2. Test methods
lize the foam structure, using initial and final setting times of 206 min
and 318 min, respectively, in accordance with Chinese standard GB/T Achieving a suitable match between the foam breakage and the
1346-2011. The chemical composition of the cement is shown in hardening process is essential for the preparation of stable foamed con-
Table 1; the corresponding particle size distribution (Fig. 1) was deter- crete. In fact, foamed concrete is hardened, owing to the generation and
mined by using a Mastersizer 2000 (Malvern, England). A set- flocculation of the hydration products [18]. The rheometer (Fig. 2,
accelerator, which was composed primarily of aluminum sulfate, was Viskomat NT, Schlebinger, Germany) consists of internal and external
used to accelerate the hardening reaction. In addition, the rheological steel cylinders, paddle blades, a data processing system, and a
behavior of the cement pastes was modified by using a constant-temperature water bath. This rheometer was used to deter-
polycarboxylic-type superplasticizer (KS-JS50Z; Sichuan SikaKeshuai mine the torque required for the cement slurry to resist destruction by
Construction Material Co., Ltd., Sichuan Province, China). The corre- the paddle blades, immediately after preparation. The torque was then
sponding preformed foam, which had a density of 35 kg m−3, was pre- used to determine the extent of flocculation of the hydration products.
pared by using a foam generator. Using this generator, compressed air Furthermore, the air bubbles in the slurry undergo changes only after
and a commercial animal-protein-based foaming agent solution (Huatai overcoming the resistance of the slurry. A slurry with a strongly floccu-
Building Materials Co., Ltd., Nanyang, China, dilution ratio of 1:15 by lating structure exhibits strong resistance to changes in the foam; the
weight,) were pumped into a mixing chamber. The mixture was then torque of the slurry reflects its ability to resist these changes. Discontin-
subjected to a series of high-density restrictions, which resulted in the uous measurements were performed in order to achieve a balance be-
formation of the foam. tween the torque development of various slurries and the
The mixing procedure has a significant effect on the porosity, as well measurement range of the instrument. As Fig. 3 shows, the measure-
as the size and distribution of the pores, and therefore, a suitable speci- ments were performed by increasing the rotational speed from 0 to
men preparation procedure is crucial. In this study, the HPFC specimens 180 rpm in 20 s, maintaining this speed for 250 s, and then decelerating
were all prepared at a temperature of 20 ± 2 °C, in the laboratory, by to 0 rpm in 30 s. This sequence was repeated for a total testing time of
mixing the proportions shown in Table 2. During the preparation, 1 h, and the resulting torque curve was recorded by a data processing
about 80% of water was poured into a vertical mixer (along with PC), system.
to which Portland cement was then added. The water and cement To determine the distribution of air-void sizes, 69-mm2 images of
were mixed for 90 s until a homogeneous cement slurry (without each sample were captured at a magnification of 20 ×, by using an
lumps) of undispersed cement was obtained. The remaining water, industrial digital microscope; each pixel in each image represents
15 μm. Six representative images of each composition were obtained
from the cubic sample that was used for the compressive strength
test, and each image was obtained from the center of each of the six
cut surfaces. The image processing step consisted of five operations
namely, dilation, erosion, opening, closing, and hole-filling. Using this
set of operations, each image was then digitized and converted to binary
form, with air voids and the surrounding matrix denoted as white and
black, respectively. A typical image and its binary form are shown in
Fig. 4. The diameters of the artificial air voids in the foamed concrete
were typically larger than 50 μm [15]. After the image processing
steps were completed, the voids in a given region were identified and
the corresponding mean diameter was determined by using an image
analysis software; these diameters were stored in Excel. The data
was then analyzed by using Origin and the cumulative frequency of
the air-void size was plotted. The D10, D50, and D90, which were derived
from this cumulative frequency, were used to quantify and compare the
distributions corresponding to each of the mixes; D50 corresponds to
the median void diameter, whereas D10 and D90 represent the diameters
that correspond to cumulative frequency values of 10% and 90%, respec-
Fig. 1. Particle size distribution of the cement. tively [15,19].
 J. Jiang et al. / Materials and Design 92 (2016) 949–959 951

Fig. 2. Rheometer.

Table 2
Mix proportions of HPFCs.

Cement slurry Foam (wt.% of cement)

Water-to-cement ratio SA (wt.% of cement) PC(wt.% of cement)

0.8 1, 2, 3, 4, 5, 6 0.6 20
0.8 4 0, 0.3, 0.6, 0.9, 1.2, 1.5 20
0.5, 0.6, 0.7, 0.8, 0.9, 1.0 4 0.6 20
0.8 4 0.6 13.3, 20.0, 26.6, 40.0
952 J. Jiang et al. / Materials and Design 92 (2016) 949–959

Fig. 5. HPFCs with different dosages of SA.

of the cement paste [20]. Large amounts of preformed foam must be

Fig. 3. Method used for the torque test.
added to the cement slurry, in order to form the HPFC. If the slurry has
poor rheological properties, then the foam will fracture when mixed
After curing, the specimens were placed for 24 h in a room whose with the slurry. The rheological behavior is typically modified by adding
respective temperature and humidity were 20 °C and 60% RH. The a superplasticizer and water to the slurry. However, these may influence
properties of the hardened state of these specimens were evaluated the stability of the HPFC mixture, owing to the negative effect of the
via compressive strength tests. Three samples hardening process of the cement. An understanding of the effect of
(70.7 × 70.7 × 70.7 mm3) of each composition and at each curing age, these parameters on the stability of fresh mixtures is therefore crucial
were tested at a loading rate of 500 N/s, by using a SANS CMT5105 ma- for the preparation of a stable HPFC.
chine. Moreover, the thermal conductivity of the 28-day-cured foamed
concrete was measured in accordance with Chinese standard GB/T 3.1.1. Effect of a set-accelerator on the stability of fresh foamed concrete
10294-2008, by using a heat-flow conductometer (JW-III, Beijing). Fig. 5 shows samples of HPFC, which have different dosages of set-
These measurements were performed by using an oven-dried accelerator. Significant and slight collapses occurred at dosages of 1%
300 × 300 × 30 mm3 slab. In addition, the mineral phases of the samples and 2%, respectively, whereas stable HPFCs were obtained for dosages
were identified via X-ray diffraction (XRD) measurements (PANalytical N2%. This difference in stability is attributed to the extent of flocculation
X'PertPro, copper target, step size: 0.03°). The microstructure of each of the hydration products. As Fig. 6 shows, the torque of the slurries
foam was examined by using a scanning electron microscope (SEM; increased with increasing accelerator dosage during 1 h of testing. The
Hitachi TM 1000). Prior to the SEM and XRD measurements, the sam- torque increased initially and remained almost constant thereafter, for
ples were dried in a vacuum oven at 45 °C, until a constant mass was dosages N 2%. The cement slurry with 1% accelerator had the lowest
obtained. torque, and its 2%-dosed counterpart had a torque of 1.6 N mm; these
torques remained approximately constant during testing.
3. Results and discussion However, the torque increased when the flocculation of the hydra-
tion products decomposed at a slower rate than that at which rebuilding
3.1. Stability of fresh foamed concrete occurred. Low dosages (1% and 2%) of the accelerator resulted in slow
production and low flocculation rates of the hydration products. There-
Due to their thermodynamically unstable nature, coalescence and fore, the cement slurry was easily destroyed by the paddle blades and
disproportionation of the air bubbles in fresh foamed concrete are hence the torque remained relatively constant. The torque increases
unavoidable. This coalescence and disproportionation lead, in general, however, once the external force of destruction is overcome. The initial
to collapse of the HPFC during the hardening process, and in turn to increase in the torque indicates that the flocculated structure forms
an undesired density and poor properties. An accelerated hardening rapidly at high dosages (≥3%). As a result, the extent of flocculation of
process is essential for preventing this collapse and improving the the hydration products improves, and the resistance of the foam to
stability of foamed concrete [18], and is typically achieved by using a changes in the HPFC increases. This leads, in turn, to an increase in the
set-accelerator. However, the hydration products crystallize rapidly, stability of the HPFC, and the rapid coalescence and breakage (Fig. 7)
owing to the accelerator, thereby leading to poor rheological properties of the preformed foam is thereby prevented. From a thermodynamic

Fig. 4. Typical initial image and processed image of the foamed concrete.
 J. Jiang et al. / Materials and Design 92 (2016) 949–959 953

Fig. 8. HPFCs with different dosages of PC.

Fig. 6. Torque of cement slurries with different dosages of SA. 3.1.3. Effect of the water-to-cement ratio on the stability of fresh foamed
concrete
Samples having different water-to-cement ratios (w/c) are compared
viewpoint, the gas–liquid interfacial energy of the air bubbles decreases in Fig. 10. During the hardening process, the HPFC remained stable but
with increasing amount of hydration products formed at the surface of collapsed slightly at w/c values of less than 1.0 and equal to 1.0, respec-
the bubbles, and the stability of the foam increases [16]. The flocculated tively. A comparison of the ability of the slurries to resist collapse
structure deteriorated with continuous testing, and an inadequate (Fig. 11) reveals that the torque decreases with increasing w/c. For exam-
amount of rebuilding after ~20 min of testing, resulted in almost con- ple, a w/c of 1.0 resulted in a very low torque of ~2.40 N mm, which
stant torques, owing to the lack of an accelerator. Although the slurry remained almost constant during the 1-h testing period.
with an acceleration dosage of 3% had a low torque, the initial, continu- Increasing values of the w/c lead to a low concentration of particles
ously increasing torque contributed significantly to maintaining the in the slurry, increased difficulty associated with the overlapping of hy-
stability of the HPFC. dration products, and reduction of the torque associated with the slurry.
The low torque obtained at a w/c of 1.0 resulted in a decrease in the sta-
bility, and eventual collapse, of the HPFC. The low torque value obtained
3.1.2. Effect of the superplasticizer on the stability of fresh foamed concrete at a w/c of 0.9 was also indicative of this low stability. However, a slight
The HPFCs exhibited poor stability, for superplasticizer dosages increase in the torque with testing times of up to 20 min at a w/c of 0.9,
N0.9% (Fig. 8). These amounts resulted in either low torque or slight prevented this collapse. The increase in the volume of the slurry with in-
changes thereof, during 30 min of testing (Fig. 9). In contrast, the torque creasing w/c, also had a positive effect on the coating and resulted in
of the slurry without the superplasticizer, increased initially and even- minimal changes to the air bubbles; this, in turn, resulted in the forma-
tually decreased to values lower than those of their 0.3%- tion of a stable HPFC.
superplasticizer-containing counterpart (Fig. 9). This was attributed
mainly to the rapid consumption of the accelerator (without the
superplasticizer-induced retardation of hydration), and hence rapid re- 3.2. Air-void structure and hardened properties
action, which prevented the subsequent rebuilding of the flocculated
structure from being rebuilt. In addition, the foam exhibited (Fig. 7) se- 3.2.1. Effect of the water-to-cement ratio on the air-void structure and
vere coalescence and breakage at an early stage (i.e., within 30 min). hardened properties
The ability to increase the torque, extent of flocculation of the structure The w/c exerts significant influence on the strength of the foamed
or the ability to resist the growth of air bubbles at an early stage, is concrete. The air voids may also be influenced by the w/c, due to the
therefore crucial. Moreover, due to the low and approximately constant change in the rheological properties and the ability of the foam to resist
torque, HPFCs containing N 0.9% of the superplasticizer exhibited poor collapse. Therefore, the effect of w/c on the properties and air-void
stability and underwent significant collapse. structure was investigated.

Fig. 7. Foam at 0 min and 30 min.

954 J. Jiang et al. / Materials and Design 92 (2016) 949–959

Fig. 9. Torque of cement slurries with different dosages of PC.

Fig. 11. Torque of cement slurries with different w/c values.

As Fig. 12 shows, the small pores increased in size with increasing w/ strong ability of the slurry to limit changes to the bubbles gave rise to
c of 0.5–0.9. These small pores disappeared at a w/c of 0.9, and the large a high percentage of small pores; therefore, the pore size corresponding
pores became rounded. However, irregular-shaped, highly connected to D90 increased only slightly. The minimum D90 occurred at a w/c of 0.8
pores formed at w/c values of b0.8 and became rounded at values owing to the low plastic viscosity of the slurry, which resulted in a de-
N0.8. In addition, D10 (Fig. 13) changed only slightly and D50 increased crease in the number of large irregular pores. The D90 increased at a
from 0.147 mm to 0.376 mm, with increasing w/c of 0.5–0.9. The w/c of 0.9, although the lowest plastic viscosity was obtained at this
value of D90 increased for w/c values of 0.5–0.7, decreased slightly (to value; this increase is attributed primarily to a decrease in the ability
0.667 mm) at a w/c of 0.8, and increased thereafter; D50, remained ap- to limit the growth of the air bubbles.
proximately constant for w/c of up to 0.7 and increased thereafter. The densification of the solid phase in the HPFC decreased with
Moreover, as the cumulative frequency distribution of the air-void size increasing w/c, whereas its strength at 7 days, 28 days, and 56 days
(Fig. 14) shows, the range of pore sizes broadens with increasing w/c. increased significantly, as shown in Fig. 15. In fact, 56-day compressive
The foams all exhibited porosities of ~ 92%, despite having different strengths of 0.51 MPa and 0.35 MPa were achieved at w/c values of 0.9
pore structures. and 0.8, respectively. The number of capillary pores increased with in-
The plastic viscosity, number of broken air bubbles, and number of creasing w/c values. In fact, for the same porosity, the increasing fraction
irregular pores all decreased, and the flow of air to the outside was facil- of capillary pores leads to a decreasing number of air-voids. Air voids are
itated, with increasing w/c. At high w/c values, the ability to limit the significantly larger than capillary pores, and hence the strength in-
growth of air bubbles decreased, and hence the pores increased in creases with decreasing density of these voids. In addition, for the
size. The small pores also disappeared gradually with increasing w/c. same porosity, samples consisting of small air voids have a larger total
However, growth of the pores was prevented, at w/c values of b 0.8, surface area of voids than their counterparts, which have larger voids.
owing to the extremely limited change in the air bubbles; as such, a The large air voids formed in the HPFC, at high w/c values, contributed
relatively high percentage of small pores persisted, thereby resulting to the coating of the foam and hence a decrease in the connectivity of
in only a slight change of D50 with increasing w/c (0.5–0.7). These the pores, and an increase in the strength. High w/c values aid in lower-
small pores disappeared with further increases in the w/c, due to the ing the plastic viscosity and therefore lead to a decrease in the density of
low ability of the slurry to resist changes in the bubbles and as a result, large irregular pores, and an increase in the strength.
D50 increased significantly. At w/c values ranging from 0.5–0.7, the
3.2.2. Effect of the set-accelerator on the air-void structure and hardened
properties
Set-accelerators promote the hydration reaction and give rise to fine
hydration products, which influence the air-void structure and have an
adverse effect on the rheological properties. However, these accelera-
tors result in the formation of stable HPFCs and therefore, constitute a
vital component of the slurry.
The pore size decreased significantly with increasing amount
(3%–5%) of accelerator (Fig. 16), but further increases in the accelerator
had no effect on these sizes. The pore size corresponding to D10
remained approximately constant (Fig. 17) with accelerator dosages
ranging from 3% to 6%. However, those corresponding to D50 and D90
decreased initially from 0.521 mm to 0.236 mm and 0.962 mm to
0.667 mm, respectively, and varied only slightly thereafter. The porosity
(~92%) also varied only slightly with increasing dosage (Fig. 17). How-
ever, the pore size distribution narrowed (Fig. 18), and hence the uni-
formity of the pore size increased, with increasing dosage. Increasing
dosage levels also resulted in increased ability of the slurry to limit air
Fig. 10. HPFCs with different w/c values. bubbles and hence the pores decreased in size. At dosages of 5% and
 J. Jiang et al. / Materials and Design 92 (2016) 949–959 955

Fig. 12. Micrographs of HPFCs with different w/c values.

6%, the hydration reactions progressed rapidly, thereby preventing the

growth of the bubbles, which in turn resulted in only slight changes in
the pore size.
The air-void structure (and therefore the strength) of the HPFC can
be modified by adjusting the accelerator dosage. As Fig. 19 shows, the
strength of the HPFC at 7 days, 28 days, and 56 days increased with dos-
ages ranging from 3% to 4%, but decreased thereafter. The 56-day-cured
HPFC dosed with 4% of the accelerator, exhibited the maximum
compressive strength (0.35 MPa). The strength of the foamed concrete
increased with progressive narrowing of the distribution of the air-
void sizes [5]. Moreover, for the same porosity, small air-void sizes
gave rise to narrow distributions and hence, a high surface area of air
bubbles. These corresponded, in turn, to high levels of pore connectivity,
thin thickness of solid (distributed among the pores) and hence, a de-
crease in the strength of the HPFC. However, the most suitable air-
void size distribution, and therefore the highest strength, were achieved
at an accelerator dosage of 4%.

Fig. 13. Effect of w/c on D10, D50, D90, and the porosity. 3.2.3. Effect of the superplasticizer on the air-void structure and hardened
properties
The workability of the fresh HPFC improved with the addition of the
superplasticizer. Due to this improved workability, foamed concrete

Fig. 14. Effect of w/c on the cumulative frequency distribution of the air-void size. Fig. 15. Effect of w/c on the compressive strength.
956 J. Jiang et al. / Materials and Design 92 (2016) 949–959

Fig. 16. Micrographs of HPFCs with different dosages of SA.

could be fabricated at construction sites. However, the air-void struc- ed, owing to the addition of the superplasticizer. In fact, the
ture, and ultimately the HPFC, collapsed with excessive additions of superplasticizer-induced reduction in the plastic viscosity of the slurry
the superplasticizer. resulted in a decrease in the number of broken air bubbles and the per-
The pore size decreased with increasing dosage of up to 0.6% of the centage of large air voids. Excessive amounts of superplasticizer result-
superplasticizer (Fig. 20), and increased thereafter. As Fig. 21 shows, ed, however, in retardation of the hydration reaction, a decrease in the
the pore size corresponding to D10 remained approximately constant, ability to resist the growth of air bubbles, and hence an increase in the
whereas those corresponding to D90 and D50 decreased initially with in- pore size.
creasing (0%–0.6%) dosage levels and increased thereafter. The mini- The strength of the HPFC improved with a suitable distribution of
mum pore sizes associated with D50 and D90 were obtained at a air-void sizes. As previously stated, the HPFC exhibited the narrowest
dosage of 0.6%. Fig. 21 also shows that the porosities vary only slightly distribution of air-void sizes and the maximum compressive strength
(i.e., from 91.6% to 92.5%) with increasing dosage. Moreover, the cumu- (0.35 MPa) at 56 days, and a superplasticizer dosage level of 0.6% (Fig.
lative frequency distribution of the air-void sizes narrowed with in- 23). The maximum strength at a dosage level of 0.3% occurred at
creasing dosage of up to 0.6%, and broadened thereafter (Fig. 22). The 7 days of curing (Fig. 23). The solid phase of the HPCFs had extremely
rheological properties improved and the hydration reaction was retard- low strengths after short curing times. However, the large pores formed
thick pore walls, which provided adequate resistance to external loads.
Moderate increases in the pore size gave rise to correspondingly moder-
ate reductions in the strength, and therefore the 7-day-cured HPFC
dosed with 0.3% of the superplasticizer, exhibited the highest strength.

3.2.4. Effect of the dry density on the air-void structure and hardened
properties
The mass of the foam was varied in order to prepare HPFC (Fig. 24) of
various apparent densities. Apparent densities of 272.8 kg m− 3 to
109.1 kg m−3 were obtained with increasing mass of 13.3%–40.0% and
the air voids increased in size. The D50 and D90 values decreased from
0.582 mm to 0.220 mm and 1.345 mm to 0.575 mm, respectively,
with increasing apparent density, whereas the sizes corresponding to
D10 remained approximately constant (Fig. 25). In addition, the porosity
decreased significantly from 95.4% to 88.5% and the distribution of air-
void sizes (Fig. 26) narrowed. The amount of slurry contained in the
HPFC decreased with increasing amount of foam used in the prepara-
Fig. 17. Effect of SA on D10, D50, D90, and the porosity.
tion. This resulted in a decrease in the apparent density and hence an in-
crease in the porosity. A reduction in the amount of slurry used, leads to
an increase in the difficulty of coating the air bubbles. As a result, the

Fig. 18. Effect of SA on the cumulative frequency distribution of the air-void size. Fig. 19. Effect of SA on the compressive strength.
 J. Jiang et al. / Materials and Design 92 (2016) 949–959 957

Fig. 20. Micrographs of HPFCs with different dosages of PC.

number of large air voids increases and the air-void-size distributions at an apparent density of 272.8 kg m−3. However, the corresponding
are broadened. HPFC was thermally insulating and exhibited acceptable strength, com-
The mass of the cement paste, and therefore the strength of the parable to that reported in other studies [15,21–30]. The HPFC also sat-
HPFC, decrease with increasing mass of the foam. In fact, the com- isfied the criteria of lightweight concrete used for insulating purposes
pressive strength of the HPFC at 7 days, 28 days, and 56 days of curing [31,32].
increased linearly (Fig. 27) with increasing apparent density. Apparent
densities of 272.8 kg m−3 and 109.1 kg m−3 resulted in compressive 3.3. Effect of the curing method on the mechanical properties
strengths of 0.75 MPa and 0.12 MPa, respectively, after 56 days of cur-
ing. In addition, the initially low thermal conductivity of the HPFC in- Due to the high porosity and the low amount of cement slurry used
creased with increasing apparent density and ranged from a minimum in the HPFC, some open air voids persist. The mechanical properties may
of 0.036 W m−1 K−1 to a maximum of 0.063 W m−1 K−1. The highest be affected by environmental factors, owing to the potential connectiv-
thermal conductivity (0.063 W m−1 K−1) of all the samples, occurred ity of these voids. Choosing suitable curing method becomes important.
As Fig. 28 shows, the HPFC cured in a natural environment had the low-
est strength irrespective of the curing time (7 days, 28 days, and 56
days), and the strength decreased with increasing curing time. In
contrast, the strength of the samples that underwent standard curing
and humidity-insulated curing increased with increasing curing age.
The 56-day-compressive strength resulting from the standard curing
is 10.8% lower than that (0.39 MPa) resulting from humidity-insulated
curing.
The XRD patterns (Fig. 29) of these HPFCs revealed that a severe
carbonation reaction occurred during curing. The pattern obtained
from the 28-day naturally cured HPFC exhibits strong and weak peaks
corresponding to CaCO3 and Ca(OH)2, respectively. However, owing to
the carbonation reaction, the former was strengthened whereas the
latter was no longer observed, after 56 days. The permeability of the
sample may decrease with carbonation [33]. This permeability may
only have a limited effect on the strength, owing to the low humidity
of the curing environment and the ease with which moisture escapes
from the sample. The loss of moisture started almost immediately
Fig. 21. Effect of PC on D10, D50, D90, and the porosity. after the demolded samples were placed in the natural environment.
Therefore, the hydration process may have suffered negative influence,
the formation of strength-imparting phases (such as C-S-H phases) may
be suppressed, and hence the strength of the sample, may decrease. The

Fig. 22. Effect of PC on the cumulative frequency distribution of the air-void size. Fig. 23. Effect of PC on the compressive strength.
958 J. Jiang et al. / Materials and Design 92 (2016) 949–959

Fig. 24. Micrographs of HPFCs with different dry densities (foam 40%:109.1 kg m−3, foam 26.6%:148.7 kg m−3, foam 20%:175.2 kg m−3, foam 13.3%:272.8 kg m−3).

carbonation reaction may also influence the shrinkage of the sample;

this shrinkage may result in the formation of microcracks and contrib-
ute to reductions in the strength [33]. Fig. 29 shows that the peak corre-
sponding to Ca(OH)2 in the 28-day standard-cured sample was stronger
than its counterpart in the 56-day-cured sample. Conversely, in the case
of humidity-insulated curing, the peaks corresponding to Ca(OH)2 and
other hydration products vary only slightly. This indicates that
humidity-insulated curing prevents carbonation, thereby leading to in-
creased strength. Fig. 30 shows the microstructures of samples obtained
after 56 days of curing via various methods. The naturally cured samples
consist of loose microstructural features, which resulted possibly from
continuous carbonation. Samples that underwent humidity-insulated

Fig. 27. Effect of the dry density on the compressive strength and thermal conductivity.

Fig. 25. Effect of the dry density on D10, D50, D90, and the porosity.

Fig. 28. Effect of the curing method on the compressive strength.

Fig. 26. Effect of the dry density on the cumulative frequency distribution of the air-void
size. Fig. 29. XRD patterns of HPFCs obtained via different curing methods.
 J. Jiang et al. / Materials and Design 92 (2016) 949–959 959

Fig. 30. SEM micrographs of HPFCs obtained from various curing methods.

curing and standard curing consist of air voids whose inner surfaces are [7] P.G. Collishaw, J.R.G. Evans, An assessment of expressions for calculating the appar-
ent thermal conductivity of cellular materials, J. Mater. Sci. 29 (1994) 2261–2273.
smooth and compact. However, weaker carbonation reactions in the [8] S.J. Powell, J.R.G. Evans, The structure of ceramic foams prepared from polyurethane
former treatment, compared to those in the latter, prevented the de- ceramic suspensions, Mater. Manuf. Process. 10 (1995) 757–771.
composition of hydration products and resulted in denser, higher- [9] D.M. Liu, Fabrication of hydroxyapatite ceramic with controlled porosity, J. Mater.
Sci. Mater. Med. 8 (1997) 227–232.
strength HPFCs. [10] C. Tuck, J.R.G. Evans, Porous ceramics prepared from protein foams, J. Mater. Sci. Lett.
18 (1999) 1003–1005.
4. Conclusions [11] C.M.S. Ranito, F.A.C. Oliveira, J.P. Borges, Hydroxyapatite foams for bone replace-
ment, Key Eng. Mater. 284–286 (2005) 341–344.
[12] A.M. Williams, C.P. Garner, J.G.P. Binner, P. I. Mech. Eng. D. J. Aut. 222 (2008)
This study revealed that HPFC (88.5%–95.4%) based on ordinary 2235–2247.
Portland cement can be prepared by using the pre-foaming method. [13] P. Colombo, M. Griffoni, M. Modesti, Ceramic foams from a preceramic polymer and
polyurethanes, J. Sol-Gel Sci. Technol. 13 (1998) 195–199.
The following conclusions can be drawn from the results of this work:
[14] J. Newman, B.S. Choo, Advanced Concrete Technology Processes, Butterworth-
Heinemann, Burlington, 2003.
(1) The stability of fresh HPFC can be increased by enhancing the
[15] E.K.K. Nambiar, K. Ramamurthy, Air-void characterisation of foam concrete, Cem.
extent of flocculation of the hydration products. The addition of Concr. Res. 37 (2007) 221–230.
an accelerator and a suitable superplasticizer to the HPFC, as [16] U.T. Gonzenbach, A.R. Studart, E. Tervoort, L.J. Gauckler, Ultrastable particle-
well as the reduction of the w/c can be used to enhance this ex- stabilized foams, Angew. Chem. 118 (2006) 3606–3610.
[17] J. Jun, Z. Lu, N. Yunhui, J. Yong, Effect of composite set-accelerator on the properties
tent, and thereby prevent the collapse of the HPFC. of ultra-lightweight foamed concrete, J. Funct. Mater. 46 (2015) 12116–12121.
(2) The pre-foaming method can be used to prepare HPFC that has [18] Z. Qi, Preparation and Properties of Lightweight Foam Concrete in Cold Regions
stable porosity. [Dissertation], Harbin Institute of Technology, Harbin, 2014.
[19] F. Krauss Juillerat, U.T. Gonzenbach, L.J. Gauckler, Tailoring the hierarchical pore
(3) The air-void size decreased and narrow air-void size distri- structures in self-setting particle-stabilized foams made from calcium aluminate
butions were obtained by increasing the apparent density and cement, Mater. Lett. 70 (2012) 152–154.
by using an accelerator and a suitable superplasticizer. [20] P. Kumar Mehta, P.J.M. Monteiro, Concrete-Microstructure, Properties and Materials.
3rd ed, McGraw-Hill Education, New York, 2006.
(4) Owing to their mechanical and thermal-insulation properties, [21] S. Herrero, P. Mayor, F. Hernández-Olivares, Influence of proportion and particle size
the HPFCs obtained in this study have significant potential for gradation of rubber from end-of-life tires on mechanical, thermal and acoustic
use as thermal insulation materials of buildings. properties of plaster–rubber mortars, Mater. Des. 47 (2013) 633–642.
[22] V. Corinaldesi, A. Mazzoli, G. Moriconi, Mechanical behaviour and thermal conduc-
(5) Natural and humidity-insulated curing had an adverse effect on
tivity of mortars containing waste rubber particles, Mater. Des. 32 (2011)
and increased the strength, respectively, of the HPFCs. 1646–1650.
[23] J.N. Eiras, F. Segovia, M.V. Borrachero, J. Monzó, M. Bonilla, J. Payá, Physical and me-
chanical properties of foamed Portland cement composite containing crumb rubber
from worn tires, Mater. Des. 59 (2014) 550–557.
Acknowledgment [24] N. Narayanan, K. Ramamurthy, Structure and properties of aerated concrete: a
review, Cem. Concr. Compos. 22 (2000) 321–329.
[25] A. Laukaitis, R. Zurauskas, J. Kerien, The effect of foam polystyrene granules on
The authors want to acknowledge the financial support of the Fund cement composite properties, Cem. Concr. Compos. 27 (2005) 41–47.
of Science and Technology on Surface Physics and Chemistry Laboratory [26] F.K. Akthar, J.R.G. Evans, High porosity (N90%) cementitious foams, Cem. Concr. Res.
(No. SPCXT201303). 40 (2010) 352–358.
[27] U. Johnson Alengaram, B.A.A. Muhit, M.Z. bin Jumaat, M.L.Y. Jing, A comparison
of the thermal conductivity of oil palm shell foamed concrete with conventional
References materials, Mater. Des. 51 (2013) 522–529.
[28] J. Feng, R. Zhang, L. Gong, Y. Li, W. Cao, X. Cheng, Development of porous fly ash-
[1] S. Wei, C. Yiqiang, Z. Yunsheng, M.R. Jones, Characterization and simulation of mi- based geopolymer with low thermal conductivity, Mater. Des. 65 (2015) 529–533.
crostructure and thermal properties of foamed concrete, Constr. Build. Mater. 47 [29] V. Medria, E. Papaa, M. Mazzocchia, L. Laghib, M. Morgantib, J. Francisconib, E.
(2013) 1278–1291. Landia, Production and characterization of lightweight vermiculite/geopolymer-
[2] Z. Zhang, J.L. Provis, A. Reid, H. Wang, Geopolymer foam concrete: an emerging based panels, Mater. Des. 85 (2015) 266–274.
material for sustainable construction, Constr. Build. Mater. 56 (2014) 113–127. [30] F. Iucolano, B. Liguori, D. Caputo, F. Colangelob, R. Cioffib, Recycled plastic aggregate
[3] F. Pacheco-Torgal, Eco-efficient construction and building materials research under in mortars composition: effect on physical and mechanical properties, Mater. Des.
the EU Framework Programme Horizon 2020, Constr. Build. Mater. 51 (2014) 52 (2013) 916–922.
151–162. [31] RILEM LC2, Functional Classification of Lightweight Concretes, 1974.
[4] Z. Huang, T. Zhang, Z. Wen, Proportioning and characterization of Portland cement- [32] JC/T 2200-2013, Foamed Cement Insulation Panel, 2013.
based ultra-lightweight foam concretes, Constr. Build. Mater. 79 (2015) 390–396. [33] A.M. Neville, J.J. Brooks, Concrete Technology, second ed. Longman, Harlow Essex,
[5] K. Ramamurthy, E.K. Kunhanandan Nambiar, G. Indu Siva Ranjani, A classification of 2010.
studies on properties of foam concrete, Cem. Concr. Compos. 31 (2009) 388–396.
[6] J.H. Mohammed, A.J. Hamad, A classification of lightweight concrete: materials,
properties and application review, Int. J. Adv. Eng. Appl. 7 (2014) 52–57.