Construction and Building Materials 153 (2017) 622–629

Contents lists available at ScienceDirect

Construction and Building Materials

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

A comprehensive study on the production of autoclaved aerated

concrete: Effects of silica-lime-cement composition and autoclaving
conditions
Ying-Liang Chen a,⇑, Juu-En Chang a, Yi-Chieh Lai b, Mei-In Melissa Chou c
a
Department of Environmental Engineering/Sustainable Environment Research Laboratories, National Cheng Kung University, No. 1, University Rd., Tainan 70101, Taiwan
b
Technology Education Division, National Science and Technology Museum, No. 720, Jiouru 1st Rd., Kaohsiung 80765, Taiwan
c
Illinois State Geological Survey/Department of Natural Resources and Environmental Sciences, University of Illinois, Urbana-Champaign, 615 E. Peabody Dr., Champaign, IL 61820,
USA

h i g h l i g h t s g r a p h i c a l a b s t r a c t

A comprehensive composition design 16

14
5h
8h
12 h
(a)
BE ABLE TO REDUCE
for the production of autoclaved
Compressive strength (MPa)

12

Silica
THE COSTS OF
10
0
100 OPERATION
aerated concrete was studied. 30
20
10
90

80

70
8

The effect of water-solids ratio on

40
60
50 4
50
60
40
2
70

bulk density was more notable than

30
80
20 0
90 12 15 18
10
100 Steam pressure (atm)
0

that of foaming agent addition.

Lime 0 10 20 30 40 50 60 70 80 90 100 Cement

High steam pressure curing

Comprehensive design of

High steam pressure curing can silica-lime-cement system Short reaction time
for autoclaved aerated High compressive strength
significantly reduce the reaction time concrete (AAC)

and obtain high compressive strength 16

14
12 h
16 h
24 h
(b)

products.
Compressive strength (MPa)

12

10

Prolonging curing time was helpful to 8

Tobermorite formed and filled the pores in AAC
⇓
6

the development of compressive 4

2
Enhance the compressive strength

strength under low steam pressure AAC

0
6 9
Steam pressure (atm)
12

conditions. Porous & Lightweight Low steam pressure curing MAY COST LESS
FOR THE

Tobermorite formation and pores Long reaction time EQUIPMENT
Low temperature demand
filling are curial to the compressive
strength of autoclaved aerated
concrete.

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

Article history: This study describes the effects of raw mix composition, amount of foaming agent, water-solids ratio,
Received 21 March 2017 steam pressure, and curing time on the characteristics of autoclaved aerated concrete (AAC). The bulk
Received in revised form 10 July 2017 density, compressive strength, and microstructures of AAC specimens were examined, and X-ray diffrac-
Accepted 12 July 2017
tion, mercury intrusion porosimetry, and thermal analyses were employed. The bulk density of AAC was
Available online 24 July 2017
affected by the amount of aluminum powder added and water-solids ratio, and especially the latter. An
increase in the amount of cement was not beneficial to the development of compressive strength. In con-
Keywords:
trast, autoclave curing greatly enhanced the compressive strength, and the AAC specimen autoclaved at
Autoclaved aerated concrete
Composition design
12 atm for 16 h had the highest compressive strength of 13.3 MPa. Increasing steam pressure can notably
Autoclaving conditions reduce the curing time, based on similar requirements of compressive strength. In contrast, prolonging
Bulk density curing time was helpful to the development of compressive strength under low steam pressure
Compressive strength conditions.
Ó 2017 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.
E-mail address: roy.yl.chen@gmail.com (Y.-L. Chen).

http://dx.doi.org/10.1016/j.conbuildmat.2017.07.116
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629
1. Introduction
Lightweight concrete has attracted greater attention in the last
couple of decades, since urban buildings are now built higher and
higher and there is a need to reduce their inherent weight. The use
of lightweight concrete can also improve the seismic capability of a
building [1], and moreover lowers the costs and difficulties in
transportation and construction [2–4]. Lightweight concrete can
generally be divided into two groups, namely lightweight aggre-
gate concrete and aerated concrete. Lightweight aggregate con-
crete, as the name suggests, is a kind of concrete made by
replacing normal aggregates with lightweight ones, thus reducing
the bulk specific gravity. The use of lightweight aggregate concrete
is basically identical to that of normal concrete, so that it can be
used for structural units of a building and most of other concrete
components [5,6]. However, the costs of manufacturing light-
weight aggregate concrete are generally high, because lightweight
aggregates are produced from a high-temperature sintering pro-
cess [7]. Furthermore, the problem of uneven distribution of aggre-
gates is very common, and this thus makes lightweight aggregate
concrete unpopular in practice [8,9].
On the other hand, aerated concrete contains no coarse aggre-
gates and is relatively homogeneous compared with lightweight
aggregate concrete [10]. Aerated concrete, also known as cellular
concrete, foam concrete, porous concrete, and so on, is produced
with a cement or lime mortar, in which air voids are generated
by a gas-forming or foaming process and then entrapped, and this
therefore makes the material lightweight. However, the air voids in
the aerated concrete diminish the mechanical strength [11], and an
autoclave curing process is usually employed to remedy this
defect. Under high-pressure steam conditions for curing, the gel-
type calcium silicate hydrates (CSH) will become crystalline tober-
morite, which then enhances the material’s mechanical strength.
The resulting material is called autoclaved aerated concrete
(AAC), autoclaved lightweight concrete, or autoclaved cellular
concrete.
The basic raw materials of AAC include silica sand, lime powder,
and cement. In recent years, some substitute materials and indus-
trial byproducts, such as coal-fired fly and bottom ashes [12,13],
silica fume [14], mine tailing, and iron slag [15], have been studied,
and there exists a wide variation in the raw mix compositions sug-
Fig. 1. Silica-lime-cement ternary diagram for raw mix compositions of AAC.
623
gested by various authors [12,16–20]. The other significant vari-
ables are the particle size of raw materials, the amounts of
foaming agent and water added, and the autoclave curing condi-
tions. Some studies [21,22] indicate that the particle size of quartz
is an important factor in AAC production. Finer quartz is more reac-
tive than coarser quartz, but it also promotes the formation of
gyrolite, an undesirable mineral in AAC. Zhao et al. [23] noted that
the tobermorite gradually transformed to xonotlite with continu-
ous autoclaving, a finding which suggests that immoderate auto-
clave curing may harm the properties of AAC.
While many studies have been done on AAC production, and
indicated the effects of different process conditions, some of the
conclusions are inconsistent and few works have made a compre-
hensive study under fully controlled conditions. Accordingly, the
purpose of this study was to investigate the effects on the AAC
characteristics by widely testing a set of variables, including raw
mix composition, amount of foaming agent, water-solids ratio,
steam pressure, and curing time. The bulk density and compressive
strength of the AAC products were determined, and the
microstructures, including crystalline compositions, hydration
products, and pore size distribution, were analyzed in this work.
By eliminating the influence of particle size, the effects of silica-
lime-cement composition on the properties of AAC products
should be revealed. Moreover, increasing the steam pressure is
theoretically beneficial to the reduction in reaction time, but it is
needed to further examine the appropriate combinations of steam
pressure and reaction time.
2. Materials and methods
2.1. Raw materials and composition design of AAC
The silica-lime-cement ternary diagram of the raw mix compositions of AAC is
shown in Fig. 1, based on a review of several studies on AAC production from the
1980s to the present [12,16–20]. The figure shows that the differences in the
amounts of lime and cement are greater than the differences in the amounts of sil-
ica. Lime and cement both varied between 5 and 30 wt%, while silica varied
between 60 and 70 wt%. The experimental design of the AAC raw mixes in this
study was based on the results of literature survey, and the exact proportions of
the raw materials are given in Table 1.
In order to minimize the influence of particle size, the particle size distribution
of raw materials was determined and controlled at the same level for all the sam-
ples. Fig. 2 shows the particle size distribution of silica, lime, and cement used for
AAC production. The high-purity silica powder (99.5%) was purchased from Alfa
624 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629

Table 1 2.3. Material testing and analysis methods

The compositions of AAC raw mixes investigated in the study.
The bulk density and water absorption of the AAC specimens were determined
Sample name Weight (%)
first, and the value of apparent porosity was calculated using its relationship with
Silica Lime Cement water absorption. Afterwards the compressive strength test was carried out, and
the crushed AAC fragments were collected for the following microstructure analy-
S60-L05-C35 60 5 35
ses. The pore size distribution of AAC specimens was examined using a mercury
S60-L15-C25 60 15 25
intrusion porosimeter (MIP, Thermo Scientific, Pascal 240). A few of the AAC frag-
S60-L25-C15 60 25 15
ments were directly analyzed with MIP, while some of them were ground to a fine
S65-L05-C30 65 5 30
powder with a particle size less than 75 lm (No. 200 mesh, USA standard testing
S65-L10-C25 65 10 25
sieve) for X-ray diffraction (XRD) and thermal analyses. The AAC powder was ana-
S65-L20-C15 65 20 15
lyzed using an XRD spectrometer (Bruker, D8 Advance) with Cu Ka radiation to
S65-L30-C05 65 30 5
acquire the mineralogical composition of the AAC specimens. The XRD step-scan
S70-L05-C25 70 5 25
mode was carried out at 0.03°2h per step in 5–55°2h with a data collection rate
S70-L15-C15 70 15 15
of 2 s/step. Thermal analysis was conducted to study the hydration products in
S70-L25-C05 70 25 5
the AAC specimens, and a simultaneous differential scanning calorimetry and ther-
mogravimetric analyzer (DSC-TGA, TA Instrument, SDT 2960) was employed.
Aesar, and its mean particle diameter (dm) was 3.05 lm. The lime powder Approximately 20–50 mg of AAC powder (<75 lm) was loaded into an alumina cru-
(dm = 39.90 lm) was prepared by crushing and grinding the calcium oxide lumps cible for analysis, and the furnace temperature was programmed to rise from the
(Merck, 97%). The cement used in this work was a locally produced ordinary Port- ambient temperature to 1000 °C at a heating rate of 10 °C/min. This thermal anal-
land cement equivalent to ASTM Type I, and the dm value of the cement particles ysis was performed under a dynamic nitrogen atmosphere (100 mL/min).
was 14.83 lm.

2.2. Preparation of AAC specimens 3. Results and discussion

The preparation of AAC specimens can be divided into three steps: (1) raw
material mixing, (2) foaming and pre-curing, and (3) autoclave curing, described
as follows:

At first, silica, lime, and cement powders were precisely weighed in accordance
with the experimental design (shown in Table 1) and then put into a steel
blending bowl. Aluminum powder (Alfa Aesar, 99.5%) was employed as a foam-
ing agent, and its addition amount was 0.25–1.00 wt%. The raw materials were
mixed with the foaming agent by using an electric blender until the mix had
become homogeneous.

Secondly, water was subsequently added into a raw mix at a specific water-
solids ratio (0.6, 0.7, 0.8, 0.9, or 1.0 L/kg) and then blended to a mortar as soon
as possible. The mortar was poured into cast-iron molds (50-mm cube in shape)
where it was allowed to expand. After standing for 30 min, the expansion of
mortar nearly finished, and then the parts bulging out of molds were scraped
off. The specimens with molds were immediately put in a moist closet where
the relative humidity was about 95%, and the temperature was controlled at
23 ± 2 °C. After pre-curing in the moist closet for 24 h, the hardened specimens
were removed from the molds.

Finally, the hardened specimens were cured by using an autoclave with a pro
portional-integral-derivative controller in order to complete the hydration reac-
tions, and thus AAC specimens were obtained. The autoclave curing conditions
investigated in this study were temperature, steam pressure, and curing time.
The temperature was controlled at 160–208 °C, while the corresponding steam
pressure varied between 6 and 18 atm. The curing time was set from 5 to 24 h.
To find out the appropriate combinations of autoclave curing conditions, the
lower steam pressure was accompanied by a longer curing time, whereas the
higher steam pressure was accompanied by a shorter curing time.
3.1. Bulk density and changes in characteristics
The porosity is very important to AAC, because it provides
excellent physical properties, including thermal and sound insula-
tion, fire resistance, and light weight. During AAC production, the
foaming process fundamentally decides the porous characteristic,
so the factors related to foaming should be examined. In this work,
the effects of the amount of aluminum powder added and water-
solids ratio on the bulk density of AAC were studied, and the
results are shown in Fig. 3. It is known that hydrogen gas is gener-
ated when aluminum reacts with alkaline substances, e.g. calcium
hydroxide, and this makes the AAC mortars expand and conse-
quently reduces the bulk density. The representative chemical
reaction of aluminum and calcium hydroxide is given in Eq. (1).
2Al þ 3CaðOHÞ2 þ 6H2 O ! 3CaO  Al2 O3  6H2 O þ 3H2 " ð1Þ
As the amount of aluminum powder varied from 0.25 to 1.0 wt
%, the AAC with different silica-lime-cement compositions (S60-
L25-C15, S65-L20-C15, and S70-L30-C15) showed similar results.
The bulk density of AAC gradually decreased with the increase in
the addition of aluminum powder, but the rate of this decrease
between 0.25 and 0.5 wt% of aluminum powder was higher than
that above 0.5 wt%. This shows that although more foaming agent

Fig. 2. Particle size distribution of the raw materials for AAC.

 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629 625

can generate a larger amount of gas, and thus should promote the
(a) expansion of AAC mortars, not all of the gas is entrapped and the
excess escapes from the mortars. On the other hand, in comparison
with the addition of aluminum powder, the water-solids ratio had
more noticeable effects on the bulk density of AAC. The bulk den-
sity sharply decreased from about 1000 to 300 kg/m3 when the
water-solids ratio was increased from 0.6 to 1.0 L/kg. A possible
explanation for this may lie in the fact that the mortar with a
higher water-solids ratio is less viscous and therefore facilitates
the formation of air voids, further reducing the bulk density of
AAC. Once again, the silica-lime-cement composition of AAC seems
insensitive to the effect of water-solids ratio on the bulk density.
The above findings suggest that the water-solids ratio is more
important with regard to the bulk density of AAC than the amount
of aluminum powder added. In sum, a thin mortar is beneficial to
the formation of air voids, but a dense one easily entraps them,
so that an appropriate water-solids ratio is a critical factor which
needs to be identified in the initial stage of AAC production.
After foaming and pre-curing, the mortars turned into hardened
specimens, which will become AAC by further autoclave curing.
Table 2 presents the bulk density and compressive strength of
the hardened and autoclaved specimens. The bulk density of hard-
ened specimens was almost identical to that of autoclaved speci-
mens. This indicates that the bulk density of AAC is already
(b) decided in the foaming and pre-curing processes, and the autoclave
curing has little influence. However, with respect to compressive
strength, the hardened and autoclaved specimens were completely
different. The average compressive strength of the hardened spec-
imens was only 45.0–103.2 kPa, while that of autoclaved speci-
mens increased to 5.7–13.3 MPa. The results demonstrate that
autoclave curing can markedly increase the compressive strength
of AAC, even by more than one hundred times.
Fig. 4 shows the XRD patterns of the AAC raw mix, hardened
specimen, and autoclaved specimen (S60-L25-C15). In the raw
mix, quartz (SiO2) and lime (CaO) were the dominant mineral
phases, which account for 60 wt% and 25 wt%, respectively. By
mixing raw mix with water and pre-curing, the hardened specimen
was obtained and there were some changes in the mineral compo-
sition. The diffraction peaks of lime almost disappeared and port-
landite, Ca(OH)2, substituted for lime. After autoclave curing, the
portlandite vanished and a crystalline hydrate, tobermorite (Ca5(-
OH)2Si6O164H2O), formed in the specimen. In addition, the inten-
sity of the diffraction peaks of quartz slightly decreased, which
implies that the quantity of quartz partially diminished. Under a
high-pressure steam condition, some studies [24,25] suggested
Fig. 3. Bulk density of the AAC specimens produced with different amounts of
that tobermorite was produced from amorphous CSH whereas
aluminum powder added and water-solids ratios. another [26] reported that tobermorite can be formed by the reac-
tion of Ca(OH)2 and SiO2. The results of this study appear to be con-
sistent with the latter, and in this work there was very little

Table 2
Bulk density and compressive strength of the hardened and autoclaved specimens.

Sample name Hardened specimen Autoclaved specimena

3
Bulk density (kg/m ) Compressive strength (kPa) Bulk density (kg/m3) Compressive strength (MPa)
S60-L05-C35 597 ± 0.5 45.0 ± 4.6 595 ± 5.3 5.7 ± 0.2
S60-L15-C25 620 ± 9.1 84.6 ± 4.6 598 ± 6.0 7.3 ± 0.4
S60-L25-C15 656 ± 13.6 95.2 ± 7.8 643 ± 11.5 10.1 ± 0.7
S65-L05-C30 581 ± 6.5 50.3 ± 2.5 596 ± 2.4 5.7 ± 0.2
S65-L10-C25 612 ± 6.7 66.1 ± 4.6 611 ± 5.6 6.8 ± 0.1
S65-L20-C15 626 ± 9.2 89.9 ± 3.2 621 ± 3.7 8.0 ± 0.8
S65-L30-C05 664 ± 5.8 103.2 ± 0.1 663 ± 6.4 11.4 ± 0.2
S70-L05-C25 618 ± 6.0 60.8 ± 3.6 636 ± 6.8 7.0 ± 0.2
S70-L15-C15 655 ± 2.4 89.9 ± 4.8 661 ± 4.5 8.9 ± 0.2
S70-L25-C05 674 ± 3.7 100.5 ± 4.5 673 ± 5.2 13.3 ± 0.4
a
Steam pressure = 12 atm; curing time = 16 h.
626 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629

L25-C05, had a slight reduction in compressive strength at a curing

time of 24 h. The above results show that the hydration reactions
progress with the curing time under a high-pressure steam condi-
tion, but there is an undesirable conversion occurring if the curing
time is excessive. Some studies [21,27] reported that gyrolite
(2CaO3SiO22.5H2O) formed from tobermorite after longer periods
of autoclaving, and this may explain the reduction in compressive
strength. The raw mix composition of the AAC specimens also had
some effects on the compressive strength. Although cement was
employed as a binder, the AAC specimens with a higher cement
content had lower compressive strength. This may suggest that
the development of compressive strength of AAC cannot be
ascribed to the hydration of the cement, and the reaction of lime
and silica are more important with regard to this. In this work,
the AAC specimen containing 70 wt% of silica, 25 wt% of lime,
and 5 wt% of cement (S70-L25-C05) had the highest compressive
strength of 13.3 MPa after being autoclaved at 12 atm for 16 h.
As mentioned above, the high-pressure steam conditions can
promote the hydration reactions and thus increase the compres-
sive strength of AAC. However, little research is available on the
Fig. 4. XRD patterns of the AAC raw mix, hardened specimen, and autoclaved effects of the level of steam pressure. Accordingly, this study fur-
specimen (S60-L25-C15).
ther examines the different combinations of steam pressure and
curing time by using high stream pressures accompanied by short
unreacted portlandite persisting in the autoclaved specimen, a curing times, and low stream pressures accompanied by long cur-
finding which shows that the reaction of Ca(OH)2 with SiO2 seems ing times. Fig. 6 presents the compressive strength of the AAC
complete. specimens autoclaved at different steam pressures and curing
times. When increasing the steam pressure from 12 to 18 atm
(Fig.6a), the development of compressive strength became faster.
3.2. Development of compressive strength
The compressive strength of the specimen autoclaved at 18 atm
for 5 h was similar to that seen when autoclaved for 8 or 12 h,
From the above results, it is known that both the amount of alu-
and was higher than that seen for the specimens autoclaved at
minum powder added and the water-solids ratio greatly affect the
12 or 15 atm for 12 h. In contrast, when decreasing the steam pres-
bulk density of AAC, while the silica-lime-cement composition
sure from 12 to 6 atm (Fig.6b), the compressive strength of the AAC
does not have significant effects. The subsequent experiments
specimens developed more slowly. The AAC specimen autoclaved
were conducted with a specific setting of aluminum powder addi-
at a relatively lower steam pressure required a longer curing time
tion (0.5 wt%) and water-solids ratio (0.7 L/kg), thus controlling the
to enhance its compressive strength. What these findings imply is
bulk density of AAC between 600 and 700 kg/m3, which are close to
that using a high steam pressure should be able to reduce the time
the median of the bulk density requirements in the AAC specifica-
costs, while selecting a low steam pressure system may cost less
tions. In addition to bulk density, the compressive strength of AAC
for the equipment, but require more time for operation.
is another important physical property, and an AAC product can be
classified into three classes according to this, i.e. AAC-2, AAC-4, and
AAC-6. Table 3 gives the physical requirements including bulk den-
3.3. Microstructure characterization
sity and compressive strength of the AAC listed in ASTM C1693. As
is apparent from the data, the AAC-2, AAC-4, and AAC-6 classes
Fig. 7 shows the correlation between apparent porosity and
require at least 2, 4, and 6 MPa of compressive strength,
bulk density of the AAC specimens. The apparent porosity was
respectively.
determined by immersing the AAC specimens in water until the
Fig. 5 shows the compressive strength of AAC specimens with
voids were saturated and then measuring the total amount of the
different silica-lime-cement compositions after being autoclaved
absorbed water for the calculation. In Fig. 7, it is noted that the
at 12 atm for 5–24 h. It is obvious that the average compressive
apparent porosity of the AAC specimens was negatively correlated
strength of the AAC specimens gradually increased with curing
with the bulk density. This shows that most of the air voids in the
time, and the development became nearly constant after 12–
AAC specimens were not closed, and thus were mostly connected,
16 h. However, some specimens, such as S65-L30-C05 and S70-
and a fluid, e.g. water or air, could thus pass through them. This
feature may provide the breathable properties of AAC products.
Table 3 The XRD patterns of the S70-L25-C05 specimens autoclaved at
Physical requirements of AAC (ASTM C1693). 12 atm for 5, 12, and 16 h are shown in Fig. 8. There were three
Strength Compressive Nominal dry bulk Density limits
major mineral phases observed in the AAC specimens, i.e. quartz,
class strength (MPa) density (kg/m3) (kg/m3) portlandite, and tobermorite. Quartz was ubiquitously present in
AAC-2 2.0 400 350–450
the specimens, and its diffraction intensity slightly decreased with
500 450–550 the curing time. The diffraction peaks of portlandite were clear in
AAC-4 4.0 500 450–550
the AAC specimen autoclaved for 5 h, but became more insignifi-
600 550–650 cant when the curing time was increased to 12 and 16 h. Tober-
700 650–750 morite can be found in all the autoclaved specimens, and the
800 750–850 diffraction intensity gradually increased with the curing time.
AAC-6 6.0 600 550–650 The results from the XRD analysis support the claim reported
700 650–750 above that the reaction between lime and silica to form tober-
800 750–850
morite was promoted as the curing time increased from 5 to 16 h.
 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629 627

Fig. 5. Compressive strength of the AAC specimens autoclaved at 12 atm for 5–24 h.

Fig. 7. Correlation between apparent porosity and bulk density of AAC specimens.

Fig. 6. Compressive strength of the AAC specimens (S70-L25-C05) autoclaved at Fig. 8. XRD patterns of the S70-L25-C05 specimens autoclaved at 12 atm for 5, 12,
different steam pressures and curing times. and16 h.
628 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629
Fig. 9 presents the TG and DSC curves of the S70-L25-C05 spec-
imens autoclaved at 12 atm for 5, 12, and 16 h. Alarcon-Ruiz et al.
[28] used the thermal analysis to examine cement pastes and
described the reactions that occur with increasing temperatures
in cement pastes and concrete materials. In the temperature range
of 180–300 °C, CSH was decomposed and lost the bound water. In
this study, the weight loss related to CSH is between 1.67 and
2.23 wt%, a finding which shows that only a small amount of CSH
was present after autoclave curing. In addition to the decomposi-
tion of CSH, it is found that there are three obvious peaks in a
Fig. 9. TG and DSC curves of the S70-L25-C05 specimens autoclaved at 12 atm for 5,
12,and 16 h.
Fig. 10. Pore size distribution of the hardened and autoclaved specimens (S70-L25-C05).
DSC curve of an AAC specimen, i.e. two endothermic peaks accom-
panied by the loss of weight and one exothermic peak. The
endothermic peak at 400–500 °C is related to the decomposition
of Ca(OH)2, and the other endothermic peak located around
700 °C is attributed to the decarbonation of CaCO3. Both Ca(OH)2
and CaCO3 can be regarded as unreacted calcium compounds per-
sisting in the AAC specimens. As the autoclave curing time
increased from 5 to 16 h, the peaks related to Ca(OH)2 and CaCO3
became smaller and smaller, a finding which means that the
hydration reactions in the AAC specimens tend to be complete.
The decrease in Ca(OH)2 is in accordance with the observation in
the XRD analysis (Fig. 8). On the other hand, a large exothermic
peak was observed in a DSC curve at about 850 °C, and there was
no significant variation in weight present near the temperature.
Connan et al. [29] used thermal analysis to characterize an auto-
claved cement and indicated that an exothermic reaction occurring
at about 800–900 °C was related to the crystallization of beta-
wollastonite (b-CaSiO3) from CSH and tobermorite. Since the CSH
was insignificant in the AAC specimens, the exothermic reaction
was mainly attributed to tobermorite. Moreover, Connan et al.
[29] indicated that the integrated area and height of the exother-
mic peak correlated closely with the level of compressive strength.
In Fig. 9, the exothermic peak related to beta-wollastonite became
more significant when the curing time was increased. The results
suggest that the amount of tobermorite in the AAC specimens
increases when prolonging the curing time from 5 to 16 h, and this
should be an explanation for the enhancement of compressive
strength mentioned above.
The pore size distribution of the hardened and autoclaved spec-
imens was examined by means of the MIP analysis to understand
the variations in microstructure ascribed to the autoclave curing,
and the analysis results are given in Fig. 10. It is noted that the pore
size in the hardened specimen was mainly distributed in the range
from 101 to 100 lm, while that in the autoclaved specimens was
mostly below 101 lm. Some studies also reported a similar pore
size distribution of materials after autoclaving [30,31]. The results
show that the pores which are initially formed are further filled
with the hydration products generated during the autoclave cur-
ing. In addition, the amount of fine pores between 102 and
2  102 lm was enhanced when the curing time increased from
12 to 16 h, and this may account in part for the improvement in
the compressive strength of AAC due to the prolongation of curing
time. Nambiar and Ramamurthy [11] indicated that a negative cor-
relation exists between the compressive strength and pore size of
 Y.-L. Chen et al. / Construction and Building Materials 153 (2017) 622–629
foam concrete. These findings show that the enhanced compres-
sive strength of AAC results not only from the hydration reactions,
but also the pores filling with the hydration products. In terms of
the cumulative pore volume, the results of the analysis for the
hardened and autoclaved specimens are similar. The cumulative
pore volume in the hardened specimen was 545 mm3/g, and those
in the autoclaved specimens were 555 mm3/g (for 12 h) and
531 mm3/g (for 16 h). This implies that the autoclave curing does
not significantly affect the total volume of pores in the AAC, but
just alters the distribution of the pores.
4. Conclusions
Based on the results of this study, the following conclusions can
be drawn. Both the amount of aluminum powder added and the
water-solids ratio affected the bulk density of AAC, but the effect
of water-solids ratio was more notable. The bulk density of AAC
was already determined in the processes of foaming and pre-
curing, and there was no significant change in this due to autoclave
curing. In terms of compressive strength, the hardened specimens,
which were not autoclaved yet, had a very low compressive
strength, and increasing the amount of cement was not beneficial
to this. However, the compressive strength of the specimens was
greatly increased by means of autoclave curing, and increasing
the steam pressure was able to drastically reduce the curing time
to achieve required levels of compressive strength. In contrast,
low steam pressure accompanied by long curing time was another
option to obtain the AAC products with comparable compressive
strength. In the microstructure characterization, it was found that
most of the air voids in the AAC specimens should be connected,
thus providing the routes for gas or liquid to pass through. This
study also revealed that the development of compressive strength
of AAC was relevant to not only the progress of hydration reactions,
but the pores filling with the hydration products. The results of this
work can provide useful information for AAC manufacturing,
although additional research may be needed on the effects of
impurities when taking into consideration of the use of industrial
wastes or byproducts as substitute raw materials for AAC
production.
Acknowledgements
The authors gratefully acknowledge the Ministry of Science and
Technology, Taiwan, for its financial support of this study (Contract
No.: MOST 102-2221-E-006-005-MY3 and MOST 106-3113-E-007-
002).
References
[1] E. Yasar, C.D. Atis, A. Kilic, H. Gulsen, Strength properties of lightweight
concrete made with basaltic pumice and fly ash, Mater. Lett. 57 (15) (2003)
2267–2270.
[2] M.N. Haque, H. Al-Khaiat, O. Kayali, Strength and durability of lightweight
concrete, Cem. Concr. Compos. 26 (4) (2004) 307–314.
[3] H.K. Kim, J.H. Jeon, H.K. Lee, Workability, and mechanical, acoustic and thermal
properties of lightweight aggregate concrete with a high volume of entrained
air, Constr. Build. Mater. 29 (2012) 193–200.
[4] H.-J. Chen, C.-H. Huang, C.-W. Tang, Dynamic properties of lightweight
concrete beams made by sedimentary lightweight aggregate, J. Mater. Civil
Eng. 22 (6) (2010) 599–606.
629
[5] S. Chandra, L. Berntsson, Lightweight Aggregate Concrete, William Andrew
Publishing, Norwich, NY, 2002.
[6] N.U. Kockal, T. Ozturan, Strength and elastic properties of structural
lightweight concretes, Mater. Des. 32 (4) (2011) 2396–2403.
[7] M.W. Brocklesby, J.B. Davison, The environmental impacts of concrete design,
procurement and on-site use in structures, Constr. Build. Mater. 14 (4) (2000)
179–188.
[8] B. Chen, J. Liu, Contribution of hybrid fibers on the properties of the high-
strength lightweight concrete having good workability, Cem. Concr. Res. 35 (5)
(2005) 913–917.
[9] F.S. Barbosa, A.L. Beaucour, M.C.R. Farage, S. Ortola, Image processing applied
to the analysis of segregation in lightweight aggregate concretes, Constr. Build.
Mater. 25 (8) (2011) 3375–3381.
[10] N. Narayanan, K. Ramamurthy, Structure and properties of aerated concrete: a
review, Cem. Concr. Compos. 22 (5) (2000) 321–329.
[11] E.K.K. Nambiar, K. Ramamurthy, Air-void characterisation of foam concrete,
Cem. Concr. Res. 37 (2) (2007) 221–230.
[12] H. Kurama, I.B. _ Topçu, C. Karakurt, Properties of the autoclaved aerated
concrete produced from coal bottom ash, J. Mater. Process. Technol. 209 (2)
(2009) 767–773.
[13] K. Ramamurthy, N. Narayanan, Factors influencing the density and
compressive strength of aerated concrete, Mag. Concr. Res. 52 (3) (2000)
163–168.
[14] W. Wongkeo, A. Chaipanich, Compressive strength, microstructure and
thermal analysis of autoclaved and air cured structural lightweight concrete
made with coal bottom ash and silica fume, Mater. Sci. Eng., A 527 (16–17)
(2010) 3676–3684.
[15] X.-Y. Huang, W. Ni, W.-H. Cui, Z.-J. Wang, L.-P. Zhu, Preparation of autoclaved
aerated concrete using copper tailings and blast furnace slag, Constr. Build.
Mater. 27 (1) (2012) 1–5.
[16] B.-G. Ma, L.-X. Cai, X.-G. Li, S.-W. Jian, Utilization of iron tailings as substitute
in autoclaved aerated concrete: physico-mechanical and microstructure of
hydration products, J. Cleaner Prod. 127 (2016) 162–171.
[17] T. Mitsuda, K. Sasaki, H. Ishida, Phase evolution during autoclaving process of
aerated concrete, J. Am. Ceram. Soc. 75 (7) (1992) 1858–1863.
[18] N.Y. Mostafa, Influence of air-cooled slag on physicochemical properties of
autoclaved aerated concrete, Cem. Concr. Res. 35 (7) (2005) 1349–1357.
[19] N. Narayanan, K. Ramamurthy, Microstructural investigations on aerated
concrete, Cem. Concr. Res. 30 (3) (2000) 457–464.
[20] F.H. Wittmann, Autoclaved Aerated Concrete, Moisture and Properties,
Elsevier Science Publishing Company Inc., New York, 1983.
[21] N. Isu, H. Ishida, T. Mitsuda, Influence of quartz particle size on the chemical
and mechanical properties of autoclaved aerated concrete (I) tobermorite
formation, Cem. Concr. Res. 25 (2) (1995) 243–248.
[22] N. Isu, S. Teramura, H. Ishida, T. Mitsuda, Influence of quartz particle size on
the chemical and mechanical properties of autoclaved aerated concrete (II)
fracture toughness, strength and micropore, Cem. Concr. Res. 25 (2) (1995)
249–254.
[23] Y. Zhao, Y. Zhang, T. Chen, Y. Chen, S. Bao, Preparation of high strength
autoclaved bricks from hematite tailings, Constr. Build. Mater. 28 (1) (2012)
450–455.
[24] J.R. Houston, R.S. Maxwell, S.A. Carroll, Transformation of meta-stable calcium
silicate hydrates to tobermorite: reaction kinetics and molecular structure
from XRD and NMR spectroscopy, Geochem. Trans. 10 (1) (2009) 1.
[25] K. Kunchariyakun, S. Asavapisit, K. Sombatsompop, Properties of autoclaved
aerated concrete incorporating rice husk ash as partial replacement for fine
aggregate, Cem. Concr. Compos. 55 (2015) 11–16.
[26] Z. Jing, F. Jin, T. Hashida, N. Yamasaki, H. Ishida, Hydrothermal solidification of
blast furnace slag by formation of tobermorite, J. Mater. Sci. 42 (19) (2007)
8236–8241.
[27] R. Siauciunas, K. Baltakys, Formation of gyrolite during hydrothermal
synthesis in the mixtures of CaO and amorphous SiO2 or quartz, Cem. Concr.
Res. 34 (11) (2004) 2029–2036.
[28] L. Alarcon-Ruiz, G. Platret, E. Massieu, A. Ehrlacher, The use of thermal analysis
in assessing the effect of temperature on a cement paste, Cem. Concr. Res. 35
(3) (2005) 609–613.
[29] H. Connan, D. Klimesch, A. Ray, P. Thomas, Thermal characterisation of
autoclaved cement made with alumina-silica rich industrialwaste, J. Therm.
Anal. Calorim. 84 (2) (2006) 521–525.
[30] R. Sarkar, S.K. Das, P.K. Mandal, H.S. Maiti, Phase and microstructure evolution
during hydrothermal solidification of clay–quartz mixture with marble dust
source of reactive lime, J. Eur. Ceram. Soc. 26 (3) (2006) 297–304.
[31] M. Jerman, M. Keppert, J. Výborný, R. Černý, Hygric, thermal and durability
properties of autoclaved aerated concrete, Constr. Build. Mater. 41 (2013)
352–359.