Effects of colloidal nanosilica on rheological and mechanical properties of fly
ash–cement mortar
Peng-kun Hou
a,b,
⇑
, Shiho Kawashima
b
, Ke-jin Wang
c
, David J. Corr
d
, Jue-shi Qian
a
, Surendra P. Shah
b
a
College of Materials Science and Engineering, Chongqing University, Chongqing 400045, China
b
Department of Civil and Environmental Engineering, Northwestern University, Evanston, IL 60208, USA
c
Department of Civil, Construction and Environmental Engineering, Iowa State University, Ames, IA 50011, USA
d
Department of Civil and Environmental Engineering and Infrastructure Technology Institute, Northwestern University, Evanston, IL 60208, USA
article info
Article history:
Received 16 March 2012
Received in revised form 22 August 2012
Accepted 23 August 2012
Available online 11 September 2012
Keywords:
Fly ash
Nanosilica
Rheology
Fluidity
Strength gain
abstract
The present study is aimed at investigating the combined effects of colloidal nanosilica (CNS) and fly ash
on the properties of cement-based materials. The fresh and hardened properties of mixtures with CNS of
10 nm size and two Class F fly ashes were evaluated. Results revealed that CNS accelerates the setting of
fly ash–cement systems by accelerating cement hydration, while fly ash can offset the reduction in flu-
idity caused by CNS. The early-age strength gain (before 7 d) of fly ash–cement systems was improved
by CNS. However, the strength gain of mixtures with CNS diminished at later ages (after 28 d), where
strength was eventually comparable to or exceeded by mixtures without CNS. Results showed that lack
of Ca(OH)
2
, which results from the high pozzolanic reactivity of CNS at early ages, and the hydration hin-
drance effect of CNS on cement at later ages can be the critical reasons.
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
Fly ash is the most widely used industrial waste product in the
cement and concrete industry. It introduces many advantages,
including lower cost, increased flowability [1], increased durabil-
ity, and improved strength gain at later ages. However, delays in
early-age strength gain and setting times are considered to be ma-
jor drawbacks [2–4]. Although these delays are desirable for some
applications, such as in mass concrete [5], in most cases this delay
is undesirable. To compensate for this shortcoming, many methods
have been explored to accelerate the early-age hydration of fly
ash–cement systems, including mechanical grinding [6], chemical
activation [7], mechanochemical treatment [8] and hydrothermal
treatment [9,10].
In recent years, nanosilica has been introduced into cement and
concrete research. Multiple studies have shown that even at small
dosages, nanosilica can improve the mechanical properties of
cementitious materials [11–13]. According to Nazari and Riahi, a
70% compressive strength improvement of concrete can be
achieved with an addition of 4% nanosilica by mass of cement
[11]. In a separate study, Shih et al. showed that a 0.6% colloidal
nanosilica addition can improve the compressive strength of
cement paste by 43.8% [12]. And Li et al. found that when 3% and
5% nanosilica were added to plain cement mortar, compressive
strength increased by 13.8% and 17.5% at 28 d, respectively [13].
For the high pozzolan replacement cementitious materials, Zhang
et al. [14,15] found that the physical and mechanical properties
of high-volume fly ash/slag systems can be greatly enhanced. Li’s
results [16] showed that nanosilica improved the compressive
strength of 50% fly ash replacement concrete by 7% after 2 years
hydration.
There are two important problems to be considered regarding
using nanosilica powder. One important consideration is disper-
sion. Intensive mechanical/ultrasonic dispersion and/or surface
treatment were applied in the aforementioned studies. Failing to
achieve proper dispersion will cause a negative effect on the
strength evolution process [11]. Another problem that should be
taken into consideration in using nanosilica is the decrease in flu-
idity, due to its high surface area to volume ratio and increased
water demand. In considering the two constituents (fly ash and
nanosilica), the benefits of each material can help counteract the
shortcomings of the other: reactive nanosilica can improve the
early-age mechanical properties while fly ash can improve flow-
ability. Regarding the issue of dispersion, water-based colloidal
nanosilica (CNS) can be used in place of nanosilica powder.
Although it has been well-documented that nanosilica can sig-
nificantly improve the early-age mechanical properties of cementi-
tious materials and the improvements are attributed to the
0958-9465/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.cemconcomp.2012.08.027
⇑
Corresponding author at: College of Materials Science and Engineering,
Chongqing University, Chongqing 400045, China. Tel.: +86 23 65126109.
E-mail address: pkhou@163.com (P.-k. Hou).
Cement & Concrete Composites 35 (2013) 12–22
Contents lists available at SciVerse ScienceDirect
Cement & Concrete Composites
journal homepage: www.elsevier.com/locate/cemconcomp
nucleation seeding effect and pozzolanic reactivity of the nanopar-
ticle [17], there are some results that show nanoparticles can have
an adverse effect on later-age strength gain [11,18,19]. Very
recently, Berra et al. [20] investigated the influence of mixing
procedures on the rheological and mechanical properties of nano-
silica-added cement-based materials and stated that the instanta-
neous formation of gel upon nanosilica’s meeting cement solution
and superplasticizer would adversely affect its influence on the
property evolution of cement-based materials. However, very few
studies regarding the combined effects of CNS and fly ash on the
fresh properties and the effects of nanoparticles on later-age
mechanical properties have been completed [21,22].
In this work, the effects of CNS on the fresh and hardened prop-
erties (early and long-term) of fly ash cementitious materials were
studied. Factors affecting these properties, such as fly ash class/
replacement ratio and CNS dosage were investigated. Moreover,
reasons governing the mechanical property evolution characteris-
tics of CNS-added high-fly ash replacement cementitious materials
were studied.
2. Experimental
2.1. Raw materials
The CNS is produced by the sol–gel technique. In this technique,
the nanoparticles are formed by the condensation and polymeriza-
tion of SiO
2
monomers that form through hydrolysis of trimethy-
lethoxysilane or tetraethoxysilane – the commonly used
precursors for synthesis of nanosilica. Eq. (1) can be used to sum-
marize the nanosilica synthesis process. When drying is applied to
the produced gel, a nanosilica powder can be obtained [19]. In this
study, a water-based and sodium-stabilized CNS with an average
particle size of 10 nm (CNS-10) was used. The basic properties, as
provided by the manufacturers of the CNS, are listed in Table 1
[23]. Transmission electron microscopy (TEM) images of CNS-10
in Fig. 1 show that although most of the CNS particles are evenly
distributed, some are still agglomerated, which may be due to
the sample preparation technique
nSiðOC
2
H
5
Þ
4
þ 2nH
2
O !
C
2
H
5
OH;NH
3
nSiO
2
þ 4nC
2
H
5
OH ð1Þ
A type I Portland cement with a Blaine fineness of 385 m
2
/kg
and a 28 d compressive strength of 45.1 MPa was used in this
study. Two fly ashes (Table 2), which vary greatly in calcium oxide
content, were used: 8.4% for fly ash 1 (FA1) and 1.46% for fly ash 2
(FA2). Although according to ASTM C618 [24] both fly ashes fall
under class F, the difference in aluminosilicate/ferrite content
(SiO
2
%+Al
2
O
3
%+Fe
2
O
3
%) is significant, indicating that the vitreous
constituents that are the source of pozzolanic reactivity are differ-
ent. This assumption was verified by X-ray diffraction (XRD) re-
sults, shown in Fig. 2. It can be seen that the diffusive diffraction
background of FA2 is greater than that of FA1 in the two theta
range of 20–35°, in which the main phases are aluminosilicates.
This indicates that the aluminosilicates amorphous phase content
of FA2 is greater. This difference should be reflected by the pozzo-
lanic reactivity of the fly ashes [4]. Fig. 2 also indicates that there
are more crystalline phases in FA1 than in FA2. This may be due
to the difference in CaO content, which can greatly influence the
pozzolanic reactivity [25].
The water demand ratios of FA1 and FA2 are 93% and 98%,
respectively. It is shown in the scanning electron microscopy
(SEM) images of the fly ashes (Fig. 3) that FA1 are round in shape
and contain very few impurities, while FA2 contains some porous
particles. From the LOI values listed in Table 2, as well as the mor-
phological characteristics of the fly ash components, it is deduced
that the porous substance is carbon [26]. The water demand ratio
and SEM images of the fly ashes suggest that FA1 has a higher
water reducing effect than FA2. Other parameters of the raw mate-
rials are also shown in Table 2.
Table 1
Properties of colloidal nanosilica.
CNS Average
particle
size (nm)
Solid
content
(wt.%)
Viscosity
(Pa S)
pH Solid
density
(g/cm
3
)
CNS-10 10 30 8 10.5 2.39
Fig. 1. TEM images of CNS-10.
Table 2
Physiochemical properties of fly ash and cement.
Materials Type I cement Fly ash 1 Fly ash 2
SiO
2
20.2 46 53.5
Al
2
O
3
4.7 17.8 25.2
Fe
2
O
3
3.3 18.2 7.2
SO
3
3.3 2.59 0.32
CaO 62.9 8.4 1.46
MgO 2.7 0.95 1.43
Na
2
O / 0.59 0.49
K
2
O / 2.16 3.51
LOI 1.1 1.49 4.87
Total 98.2 98.2 98
Water requirement ratio H/% / 93 98
Density/(g/cm
3
) 3.10 2.30 2.26
Water requirement ratio H: water required (compared to that of plain mortar) for
30 wt.% fly ash replaced cement mortar to achieve the same fluidity with that of
plain mortar.
P.-k. Hou et al. / Cement & Concrete Composites 35 (2013) 12–22
13
2.2. Mixture proportions
Three dosages of CNS addition (0%, 2.25% and 5% (solid state) by
mass of binder) and three levels of cement replacement with fly
ash (20%, 40%, and 60% by mass) were used. Water to binder (w/
b) ratios of 0.41 and 0.35 by mass were used to investigate rheolog-
ical and setting properties, respectively. (When calculating the w/b
ratio, the water content of the CNS was appropriately considered.)
A w/b ratio of 0.5 and a binder-to-sand ratio of 1/3 were used for
all mortar mixtures. Raw materials were dry mixed for 1 min at
low speed to obtain a homogenous mixture, wet mixed at low
speed for another minute, and then mixed at medium speed for
3 min. As CNS was in a water base, it was hand-stirred in the mix-
ing water prior to adding the other materials, to achieve a homoge-
neous solution and the small strength variation of replicate
samples (shown in the results section) indicated that a homoge-
neous mixing has been achieved.
To sustain a constant water-to-binder ratio, the replacement of
fly ash and addition of CNS were on a mass basis, which, however,
may introduce differences in the water volume fractions in the
mixtures with the addition of CNS due to its density difference to
that of the other powders. As calculated in Ref. [22], the difference
in water volume fraction of mixtures with and without CNS at the
same fly ash replacement ratio is very slight; the influence of water
volume fraction differences on the properties of mixtures of the
same fly ash replacement ratio with and without CNS is therefore
considered to be negligible. However, the difference in water vol-
ume fraction between the control mixture without fly ash and
those with fly ash is more significant, due to the specific gravity
of the cement being about 3.1, while those of the fly ashes are near
2.3.
2.3. Setting time
ASTM C191 [27] was followed and a manual Vicat apparatus
was used to determine the initial and final setting time of pastes.
Setting times were determined from the time-Vicat needle pene-
tration depth curve.
2.4. Shear rheology and fluidity
A temperature-controlled rheometer with a coaxial cylinder
geometry was used to measure the shear rheological properties
of pastes. The inner and outer cylinders were 10 mm and
10.85 mm in diameter, respectively. Although the ‘wall slip’ effect
of the coaxial cylinder rheometer will affect the accuracy of the
absolute rheological values [28], the comparison of the relative
rheological data can still adequately reflect the effect of CNS on
the rheological properties of the system. Plastic viscosities and
yield stresses were obtained using the Bingham model:
s
¼
s
0
þ
l
p
c
ð2Þ
where
s
is the shear stress,
s
0
is the yield stress,
l
p
is the plastic vis-
cosity and
c
is the shear rate. In the previous study, it was found
that after mixing for 8 min, cementitious pastes would acquire a
stable shear strain–stress curve. So in the real tests, the raw mate-
rials were first dry-mixed for 1 min and then wet-mixed for 8 min.
Lime - CaO
Portlandite - Ca(OH)2
Anhydrite - CaSO4
Quartz low - SiO2
Mullite - Al6Si2O13
Hematite - Fe2O3
Maghemite-C - Fe2O3
20 30 40 50 60 70
2-Theta (°)
Quartz - SiO2
Mullite - Al6Si2O13
Hematite - Fe2O3
20 30 40 50 60 70
2-Theta (°)
20 30 40 50 60
0
1000
2000
3000
4000
5000
FA1 FA2
FA2
FA1
ff
70
2-Theta (°)
Fig. 2. XRD patterns of fly ashes.
Fig. 3. SEM images of fly ashes.
14 P.-k. Hou et al. / Cement & Concrete Composites 35 (2013) 12–22
The CNS was first hand-stirred in the mixing water to facilitate dis-
persion. The shear rheological protocol and typical flow curve ob-
tained are given in Figs. 4 and 5. The coefficients of variation of
the plastic viscosities and yield stress of three replicate samples
were about 10–15% and 20%, respectively.
Following ASTM C230 [29], the fluidity of mortar was measured
using a flow table.
2.5. Compressive strength
ASTM C109 [30] was followed to measure the compressive
strength of mortar by using a 4448 kN MTS hydraulic test machine.
For each mixture at each age, three
U
5.08 cm 10.16 cm cylindri-
cal samples were tested and the average value was taken to be the
representative strength. To evaluate the effect of CNS on the
strength evolution of fly ash–cement systems, the compressive
strength ratio at each curing age was calculated using the follow-
ing equation:
Rð%Þ¼100 f
i
=f
c
ð3Þ
where R is compressive strength ratio, %, f
i
is compressive strength
of fly ash mortar with various dosages of CNS, and f
c
is compressive
strength of plain cement mortar.
An R value greater than 100% indicates a higher strength than
that of plain cement mortar. The slope of the R curve shows the
strength evolution rate compared to that of plain mortar. By com-
paring R, the influence of fly ash and CNS on strength evolution can
be determined. All samples were water-cured at room temperature
beginning at 24 h after casting.
2.6. Hydration degree of cement at later ages
Image analysis of graphs obtained from backscattered electron
microscopy technique (BSE, Hitachi S-3400) was used to evaluate
the hydration degree of cement paste at later ages. The intensity
of the BSE signal is mainly a function of the average atomic number
of the local area of the sample. During hydration, water is incorpo-
rated into unhydrated cement particles and lowers the average
atomic number of the hydrates, and thus strong imaging contrast
of unreacted (anhydrous) and reacted component (hydrates) can
be obtained [31] and the hydration degree can be evaluated
through image analysis.
Before testing, a thin sample section of approximately 5 mm
was cut out of the specimen cast in a 2 cm 2cm 8 cm mold
and mounted on a metal sample holder for polishing. Samples
were polished using silicon carbide paper of gradations 22
l
m,
14.5
l
m, and 6.5
l
m, and the polishing time of each step was
5 min. In the final step, the polished samples were ultrasonically
cleaned in water for 1 min using a bath sonicator to remove polish-
ing debris from the sample surface. Then, the sample was soaked in
acetone for 1 d before being vacuum-dried at 50 °C for 12 h. Sam-
ples were coated with 20 nm of gold to provide a conductive sur-
face. To accurately evaluate the hydration degree, five BSE
images were taken to do image analysis and they were averaged
as the representative value of each sample. At the same time, the
smallest magnification of 100X was used to obtain an image with
the largest practical field of view.
2.7. Ca(OH)
2
content
Thermogravimetric analysis (TGA, TGA/sDTA 851) was carried
out to measure the Ca(OH)
2
content of samples. The weight loss
between 440 °C and 510 °C was considered to be due to the decom-
position of Ca(OH)
2
. Before measuring, powder samples were
oven-dried at 105 °C for 4 h.
2.8. Rate of hydration
The hydration temperature was measured by a semi-adiabatic
calorimeter to assess the effect of CNS on the hydration heat of ce-
ment pastes. Samples were prepared at a constant w/b ratio of 0.4
by mass, with 100 g of cement and 40 g of mixing water at a tem-
perature of 27 °C. Mixtures were cast in
U
5cm 10 cm plastic cyl-
inders within 3 min after initial cement and water contact. The
sample was then covered, placed in the calorimeter, and the tem-
perature of the sample was recorded every 3 min for 20 h.
3. Results and discussions
3.1. Setting time
Fig. 6 shows the influence of CNS on the hardening of fly ash–
cement pastes. It is apparent that CNS greatly shortened the initial
and final setting time of all pastes. When CNS was added, the initial
and final setting time of fly ash–cement pastes were significantly
shortened. Although fly ash delays the hardening of pastes, the
addition of CNS can greatly offset this effect. A great effect of
02468
0
200
400
600
800
Shear rate/S
-1
Time/min
Fig. 4. Shear rheological protocol.
0 100 200 300 400 500 600
0
50
100
150
200
250
Shear rate /S
-1
Shear stress /Pa
Slope=viscosity
Fig. 5. Typical hysteresis loop for cement paste and Bingham plastic viscosity
calculated using the slope of the down curve from 600 to 300 S
1
.
P.-k. Hou et al. / Cement & Concrete Composites 35 (2013) 12–22
15
nanosilica on the reduction of setting times of fly ash–cement sys-
tems was also obtained by other researchers [14,15].
The shortened setting time of cement–fly ash pastes with CNS
can be accounted for by the accelerating effect of the CNS on ce-
ment hydration [17]. It was found that the dissolution and pre-
cipitation processes of cement particles and hydrates were
accelerated by CNS at the beginning of reaction, and thus the
hydration and hardening of cementitious materials were en-
hanced [21]. The acceleration effect was demonstrated in Fig. 7
in that both the hydration temperature peak and hydration rate
(shown as the 1st Dev. of the temperature curve) were
increased.
3.2. Rheological properties
The viscosities of fly ash–cement pastes with and without
CNS-10 are shown in Fig. 8. For the plain cement paste without
CNS, the addition of fly ash increased viscosity, but it did not in-
crease proportionally with the replacement ratio. For all mixtures,
the addition of CNS led to an increase in viscosity. The increase is
more significant for plain cement pastes, where the viscosity of
the 5% CNS paste was 4.3 times that of the 0% CNS paste. But,
for fly ash–cement pastes, the viscosities of the 5% CNS pastes
were about 2.5 times greater than those of the 0% CNS pastes.
It is also shown that the viscosity of fly ash–cement paste with
CNS increased almost linearly with CNS dosage. For plain cement
paste, however, the increase in viscosity was more marked, being
between 2.25% and 5%. Other work [32] showed only a slight in-
crease in viscosity when CNS dosage was within 2.5 wt.%, as well.
Furthermore, it is shown in Fig. 8 that the viscosities of the 5%
CNS FA1 cement pastes were almost the same as that of the 5%
CNS cement paste. For FA2 cement pastes, these values were low-
er. To explore the reason for the difference of the effects of CNS
on the viscosity evolution of the pastes, the CNS adsorption
behaviors of the two fly ashes were studied. This behavior was
studied through measurement of the solid concentration of a
CNS solution of known concentration after it has been mixed, stir-
red and centrifuged within a fly ash solution with a fly ash-to-
water mass ratio of 0.5. The CNS adsorption capacity of fly ash
was determined as the CNS concentration loss after these pro-
cesses. It is shown in Fig. 9 that little CNS was adsorbed by
FA1, while more than 40% was absorbed by FA2. This may be
due to the differences of particle size distribution, morphology,
and mineralogy of the two fly ashes. The adsorption of CNS on
fly ash surfaces may lead to a more serious agglomeration of
0 2 4 6 8 101214161820
-4
0
4
8
30
40
Temperature/
o
C
dT/dt
time/h
0% CNS
5% CNS
Fig. 7. Effect of CNS-10 on cement hydration.
Fig. 6. Influence of CNS on the hardening of FA1 pastes.
16 P.-k. Hou et al. / Cement & Concrete Composites 35 (2013) 12–22
