Construction and Building Materials 167 (2018) 403–413

Contents lists available at ScienceDirect

Construction and Building Materials

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

Dispersion and stability of graphene nanoplatelet in water and its

influence on cement composites
Hongjian Du 1, Sze Dai Pang ⇑
Department of Civil & Environmental Engineering, National University of Singapore 117576, Singapore

h i g h l i g h t s

 A facile approach was proposed to disperse and stabilize graphene in water.

 The optimum combination of superplasticizer and sonication was determined.
 The stabilization mechanism of graphene was illustrated.
 1% graphene can effectively increase the hardness and refine the pore structures.

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

Article history: Cement nanocomposites were fabricated by slight modification of existing casting method to introduce
Received 28 August 2017 graphene nanoplatelet (GNP) into cement paste. Dispersant dosage and ultrasonication time were varied
Received in revised form 5 February 2018 to seek the optimum dispersion effectiveness. The stability of sonicated GNPs in suspension was assessed
Accepted 9 February 2018
quantitatively by light absorbance over time and results suggested that 1 h of sonication and 15% of sur-
Available online 22 February 2018
factant was sufficient to produce a 1% GNP suspension that remains stable for 6 h. Particle strengthening
effect was obtained through the strong physical bond at the graphene interface, which was reflected by
Keywords:
the doubling of the hardness. Mercury intrusion testing indicated that GNP could reduce the effective
Graphene
Nanocomposites
porosity and refine the critical pore diameter, particularly at the concentration of 1% GNP where reduc-
Microstructure tions of 37% and 30% on the effective porosity and critical pore diameter were found respectively.
Porosity Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction to mention that graphene used in previous studies vary in diameter

and thickness, depending on the raw materials and processing
Graphene, a carbon sheet of one-atom thickness, is believed to approaches. The essential features of those materials were
be the strongest material ever known to mankind and it also pos- carbon-based in chemistry (pristine or reduced graphene oxide)
sesses excellent thermal and electrical conductivities [1]. Introduc- and two-dimensional (thickness  diameter) in geometry.
ing graphene into composite materials is one possible route to The feasibility and potential advantages of using graphene or
harness its unusual properties for potential applications [2–4]. graphite nanoplatelet (GNP) in cement composites has been
Globally, concrete is the single most widely used manmade demonstrated by some previous studies. Du et al. [5] proposed
composite material. It is worthy extending the exploitation of the concept of introducing GNP as impermeable barriers in cement
graphene in cement-based composites, enabling such materials mortar against the transport of water and chloride ions. A reduc-
with higher mechanical strength, toughness, durability and tion in the water absorption and chloride migration coefficient
multi-functionalities for the next generation buildings and was observed for GNP incorporated mortar. More recently, Du
infrastructures. Some pioneering studies have been published and Pang [6] and Du et al. [7] carried out comprehensive investiga-
recently to demonstrate the promising advantages of graphene- tions into the transport properties of cement mortar and concrete
based materials in cement and concrete [4–16]. It is worthwhile containing varying amount of GNP respectively. Remarkable bar-
rier effect was reported against the ingress of aggressive ions.
⇑ Corresponding author. The 0.6 vol% GNP in mortar could cause significant decrease of
E-mail addresses: [email protected] (H. Du), [email protected] 64%, 70% and 31% for water penetration depth, chloride diffusion
(S.D. Pang). coefficient and chloride migration coefficient respectively. Accord-
1
Present address: Department of Civil and Construction Engineering, Swinburne ing to Du and Pang [6], the enhanced barrier effect and refined pore
University of Technology, PO Box 218, Hawthorn, Victoria, Australia.

https://doi.org/10.1016/j.conbuildmat.2018.02.046
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
404 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413

structure in GNP-mortar are the two main reasons for the better the influence of GNP stability on pore size distribution will be
durability performances. Peyvandi et al. [8,9] managed to apply assessed. Lastly, the influence of GNP content of pore structure
0.05 vol% GNP into industrial-scale concrete pipes, dry-casted in and hardness of paste were investigated.
manufacturing plant. Compared to the 70% loss in flexural strength
for plain concrete under acid attack, concrete with GNP exhibited
2. Materials and methods
only 10% loss after the acid attack. A 30% decrease in water absorp-
tion was noticed as well for GNP incorporated concrete. Tong et al.
2.1. Materials
[10] found that the resistance of mortar against acid corrosion was
enhanced by 0.1 wt% GNP, owning to the reshaped pore structures.
Ordinary Portland cement (OPC) CEM I 52.5 was used for the
Another important application of GNP in cement composites is the
cement paste samples throughout this study. Table 1 shows the
multi-functionalities such as piezoresistivity-based strain sensing,
oxide composition of OPC. GNP A3775 surface enhanced flake gra-
thermal diffusion and electrical conductivity [11–16]. The correla-
phite from Asbury Carbons was used to prepare the GNP cement
tion between mechanical loadings and electrical responses of GNP-
composites. The enhanced flake graphite consists of multiple
cement composites has been demonstrated [11–13], which could
stacked layers of graphene. This GNP was exfoliated from normal
enable the composites suitable for structural health monitoring.
flake graphite to increase the number of graphene nanoplatelets,
Le et al. [14] derived an analytical model to predict the damage
each with fewer graphene layers but a higher surface area-
extent for the cement composites while the addition of GNP could
volume ratio. The average properties of GNP are provided by the
increase the electrical conductivity and hence increase the sensi-
supplier and listed in Table 2. Fig. 1a shows the morphology of
tivity of sensing. Cui et al. [15] recently found that the addition
GNP as received. It is noted that the GNP particles are not perfectly
of 5 vol% GNP in cementitious composites could lead to a 1.5-fold
round and varying in thickness and diameter. A polycarboxylate
increase in the hardness.
superplasticizer (SP) (ADVA-181N, Grace Construction, 1.125 g/cc)
In spite of the above mentioned advantages, large-scale applica-
was used as GNP dispersant.
tion of graphene products in cement and concrete is still facing
challenges such as its uniform dispersion in cement matrix and
understanding of the influence of GNP on the pore system of 2.2. Processing of GNP and casting
cement composites. In order for the GNP to manifest its superior
properties in cement composites, it has to be well dispersed in To achieve better dispersion and distribution of GNP in the
the mixing water and subsequently in cement matrix, though some cement paste, chemical and mechanical means of dispersion
previous study used silica fume to disperse nano-materials [17]. are adopted during the pre-processing of the GNP suspension via
For instance, Liu et al. [16] suggested that the advantages of GNP the addition of a dispersant and sonication of the GNP respectively.
in reinforcing cement composites can only be realized with good The combination of these two methods has demonstrated its effec-
dispersion and distribution in the composites. Dispersing graphene tiveness in obtaining a better dispersion of nanoparticles [30–34].
in water is particularly challenging due to the hydrophobic nature A BioLogics ultrasonic homogenizer model 300VT with a titanium
of the sheets, but surfactant can mitigate this problem and help the solid tip of 19.1 mm in diameter was used in this study. Ultrasonic
separated sheets to remain suspended rather than flocculated [18]. processing was performed using a power output of 300 W and a
According to Guardia et al. [19], the presence of acid groups and probe frequency of 20 kHz. A pulse mode of 70% was set so that
ether linkage interacting with water make non-ionic surfactant no cavitation was induced. During the sonication, ice bath was
more effective in dispersing graphene in water than ionic surfac- placed under the glass beaker to ensure the temperature of the
tant. In 2012, Du et al. [5] pioneered the use of a cement- solution was in the range of 20 and 35 °C. The morphology of the
compatible superplasticizer with the application of sonication sonicated GNP was observed using freeze drying technique, as
energy to disperse graphene in water and since then, there has shown in Fig. 1b. After sonication of 2 h, GNP surface area does
been growing interest in the use of this method to introduce gra- not increase (Fig. 2), which means that the main function of
phene into cement composites for superior performance sonication is to deflocculate GNP clusters instead of exfoliating
[10,15,16,20–29]. As found by [28,29], polycarboxylate based
superplasticizers could be more effective in promoting the GNP
Table 1
uniform distribution in cement matrix than other widely used Chemical Composition of OPC.
types of superplasticizers. This is consistent with previous findings
on carbon nanotube or nanofiber [30–33]. However, in previous Chemical composition %

studies, the GNP dispersing efficiency was determined in an indi- SiO2 20.8
rect means of measuring the electrical resistivity of hardened Al2O3 4.6
Fe2O3 2.8
cement mortar. The effect of polycarboxylate superplasticizer CaO 65.4
dosage and sonication time on the stability of GNP in water is MgO 1.3
unclear yet. SO3 2.2
Literature review has shown that the stability of GNP in water Na2O 0.31
K2O 0.44
by the combined use of sonication and polycarboxylate-based sur-
factant has not been studied. The objective of this study is to quan-
titatively evaluate the stability of GNP suspension as well as to
characterize the pore structures and hardness of GNP-cement com- Table 2
Physical Properties of GNP A3775.
posites. To make a fair comparison, GNP-cement composites
should have similar flowability, which is mainly controlled by Properties Values
the surfactant dosage. Hence, the first step is to determine the min- 3
Density (g/cm ) 2.26
imum surfactant dosage (at a constant GNP content and sonication Surface area (m2/g) 24
time) which can make GNP cement composite as flowable as the Diameter (lm) 8.0
plain paste. With this surfactant dosage, the sonication time is var- Thickness (nm) 37
Aspect ratio 210
ied to examine the stability of GNP in suspension, which affects the Purity (%) 98.0
distribution of GNP in the hardened cement matrix. At this stage,
 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413 405

ond, sonication time was studied at 0, 15, 30, 60 and 120 min, for a
constant 1% of GNP and 15% of SP. Third, GNP content of 0, 0.25%,
0.5%, 1%, 2% and 4% was used to prepare the cement paste to com-
pare the performances while the DSP dosage and sonication time
were kept at 15% of GNP and 1 h, respectively.

2.4. Test methods

To obtain a quantitative measure of stability, the absorbance

spectra of GNP aqueous suspension was measured by UV–vis
spectrophotometer (HACH DR5000). The sonicated GNP suspen-
sion was diluted to 0.2 g/L and measured in the wavelength range
of 200 and 500 nm. It is known that the optical absorbance of sus-
pension exhibits a linear relationship with the concentration of
particles in the solution [35], which was confirmed in this study
as shown in Fig. 3. Over time, it is expected that the suspended
GNP particles will aggregate and settle down at the bottom. By
measuring the change in the optical density of samples, the con-
centration of particles in the solution could be obtained over time.
After mixing, mini slump test was performed to determine the
flowability of the fresh cement paste. The conical mold has a base
diameter of 100 mm, top diameter of 70 mm, and height of 60 mm.
This mold was filled with fresh paste and vertically pulled upwards
so that the paste would flow from underneath the mold, as shown
in Fig. 4. The diameter of the stilled paste was measured across 4
different diagonals and the average reading was taken as the slump
diameter.
Fresh paste mixture were cast in steel molds after the mini
slump test. All specimens were demolded the next day and store
in water until 28 days. Pore size distribution (PSD) was determined
Fig. 1. Morphology of (a) as-received GNP and (b) sonicated GNP.
using Micromeritics Autopore III porosimeter with a maximum
pressure of 415 MPa. Samples (0.5–1 cm3 in volume) were col-
lected from the core of one paste sample. Those MIP samples were
slowly dried for 3 days in a vacuum desiccator at 40 °C. There are
several major limitations for the results obtained by MIP: (1) pores
in cement paste are irregular in shape instead of cylindrical; (b)
ink-bottle effect; and (c) minimum pore diameter that could be
detected is 3.8 nm. Nevertheless, as pointed out by Ye [36] and
confirmed by Ma and Li [37] and Quercia et al. [38], the effective
porosity, which governs the permeability in porous materials,
could be extracted from MIP results, as illustrated in Fig. 5.
Macroscopic hardness was measured to infer the influence of
the GNP content on the surface hardness of the cement paste. After

Fig. 2. BET surface area measurement.

graphene layers. After sonication for a specified time, the GNP

aqueous suspension was mixed with cement in a N50 Hobart
mixer for 1 min at low speed followed by 1 min at medium speed
and again 1 min at low speed.

2.3. Mix proportions

In this study, the reference cement paste has a water-cement

(w/c) ratio of 0.5. The influence of DSP dosage, sonication time,
and GNP content was determined. First, the DSP dosage effect
was assessed at varying dosage of 0, 7.5%, 11.25%, 15%, 18.75%, Fig. 3. Absorbance spectra of aqueous suspension at different concentrations. Inset:
22.5% and 30% by weight of GNP. GNP content remains constant Calibration of the absorbance peak (260 nm) and the concentration, and Photos of
at 1% by weight of cement and sonication time was kept at 1 h. Sec- GNP aqueous suspension corresponding to each concentration.
406 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413

Fig. 4. Flowability determination of fresh cement paste containing 1% GNP (a) Vertical removal of mold and (b) Measurement of the diameter.

3. Results and discussion

3.1. Influence of dispersant dosage

3.1.1. Flowability of cement paste

The slump diameters for all the GNP-cement pastes are listed in
Table 3. The reference plain cement paste had a slump diameter of
248 mm in our study. When 1% of GNP was added and mechani-
cally stirred in the cement paste, the GNP clusters hindered the
movement of the cement clinkers which compromised the flowa-
bility of the cement paste. However, with the application of sonica-
tion, those clusters were deflocculated into small sizes and
distributed more evenly in the fresh paste to maintain the flowa-
bility of the cement paste.
When the GNP content and the sonication time were main-
Fig. 5. Extracted results from MIP.
tained at 1% and 60 min respectively, the slump diameter
increased almost linearly as the DSP dosage was increased from
0 to 30%. The mechanism for dispersing GNP with the use of super-
curing, the samples were saturated-surface-dried before the hard- plasticizer is attributed primarily to the steric repulsion associated
ness test. First the surface was ground using a 600 grit sandpaper with the long lateral ether chains and secondarily to the electro-
to ensure surface smoothness. Second, a 10 mm grid was drawn static repulsion owing to carboxylate groups [30,32]. Once the sur-
on the cube to create 16 sampling points for the measurement of faces of the GNPs and cement particles became saturated with DSP,
hardness as shown in Fig. 6a. Third, indentation of paste sample flocculation will be broken so that the water locked inside can be
was carried out by a Zwick ZHR Rockwell hardness tester and a released. As such, the shear yield strength of the paste would
1.5875 mm steel ball indentor, as shown in Fig. 6b. Hardness was decrease, leading to a higher flowability [39]. In this study, a DSP
determined using the Superficial Rockwell methods with a total dosage of 15% could make the GNP-cement paste as flowable as
load of 147.1 N, and expressed in the HR15T scale. X-ray diffraction the reference paste. Hence, a DSP/GNP ratio of 15% was recom-
patterns were tested using Shimadzu LabX XRD-6000. The scan- mended for the type of DSP and GNP used in this study to provide
ning range varied from 15° to 40° at a constant rate of 2°/min. the desired flowability.
Before the XRD testing, samples were oven-dried at 105 °C for There are research works that explored the efficiency of various
18 h and grounded to a size finer than 150 mm. dispersion routes for carbon nanomaterials [30–32,40–43]. In

Fig. 6. (a) Cement paste surface with 10-mm grid measurements, and (b) The residual indent after hardness test.
 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413 407

Table 3
Influence of SP dosage on cement paste flowability.

Mix ID GNP,% SP,% Sonication, minute Flowability, mm

GNP0-SP0-T0 0 0 0 248
GNP1-SP0-T60 1 0 60 175
GNP1-SP7.5-T60 1 7.5 60 208
GNP1-SP11.25-T60 1 11.25 60 221
GNP1-SP15-T60 1 15 60 233
GNP1-SP18.75-T60 1 18.75 60 261
GNP1-SP22.5-T60 1 22.5 60 296
GNP1-SP30-T60 1 30 60 352

Table 4
Comparison of the optimum dispersant for different carbon nanoparticles.

DSP nano-C SFC/nano-C nano-C/water,% w/c A,m2/g DSP/nano-C/A, mg/m2 Matrix Study
N.A CNT 4.0 0.16 0.5 110, 230–330 12.1–17.4, 36.4 Paste [32]
PC CNT 1.6 0.7 0.35 40–300 5.3–40 Paste [30]
SDS CNT 1–1.5 0.1–0.5 0.5 110 9.1–13.6 Paste [43]
PEL CNT 1–1.5 0.1–0.5 0.5 110 9.1–13.6 Paste [43]
PL CNT 4.0 0.2, 0.4 0.5 300–380, 350–420 9.5–13.3 Mortar [42]
N.A CNF 4.0 0.096 0.5 20–50 80–200 Paste [41]
N.A CNF 4.0 0.26 0.3 N.A N.A Paste [40]
PC CNF 1.5 1.14 0.35 50–60 25–30 Paste [31]
PC GNP 1–2 0.2 0.5 120–150 6.7–16.7 Paste [28]
PC GNP 0.15 2 0.5 24 6.3 Paste This study

DSP: Dispersant.
N.A: Not available.
PC: Polycarbonate.
SDS: Sodium dodecyl sulfate.
PEL: Polyoxyethylene lauryl ether.
PL: Pluronic.
A: Surface area.

comparison with dry mixing with cement, or mechanical stirring in

water, ultrasonication with the addition of surfactant showed the
highest efficiency in dispersing and stabilizing the carbon nanoma-
terials in aqueous suspension. Table 4 summarizes the dosage of
dispersant required to achieve the optimum stability of nanoparti-
cles in cementitious matrix. A carbon nanomaterial with a larger
surface area per unit weight will require a higher dosage of disper-
sant in order to surround itself with the surfactant molecules to
repel each other and disperse more evenly in the suspension. Com-
pared to previous works [30–32,40–43], the amount of surfactant
used in this study is significantly lower due to its smaller surface
area, but when the amount of dispersant is normalized by the sur-
face area of the carbon nanomaterials as shown in Table 4, the
amount of surfactant used is comparable to the previous works,
albeit closer to the lower bound value.

3.1.2. Pore structures of cement paste

The influence of DSP content on the pore structures is displayed
in Fig. 7 for cement paste with constant GNP content of 1% and son-
ication time of 1 h. The reference plain cement paste showed the
highest intrusion volume and the coarsest PSD. After 1 h of sonica-
tion, the GNP clusters are broken and dispersed in the liquid while
their stability would be improved through the use of DSP to resist
the attractive van der Waals force between GNPs, as discussed ear-
lier. A reduction in the effective intrusion volume could be
observed even with the absence of DSP. This reduction became
more significant with higher DSP dosage of up to 15%. This is
attributed to the larger number of GNPs retained in the aqueous
suspension. However, the benefit of GNP in reducing the effective Fig. 7. Influence of SP dosage on MIP results.
intrusion volume was compromised at higher DSP dosage of 30%.
It is well known that over-dosage of DSP would reduce the cement to [45], carbon nanotubes would re-agglomerate via surfactant
hydration rate and increase the probability of bleeding and conse- molecules if the surfactant concentration is so high that the CNT
quently increase the porosity of the hardened paste [44]. According were covered by multi-layers of surfactant molecules and the
408 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413

outermost layer could attract the neighboring CNT. In this study,

with the introduction of sonication, the agglomeration of GNP
could be broken and thus DSP would have access to be adsorbed
on GNP surfaces. Below the optimal dosage, more DSP molecules
could generate higher repulsive forces between GNP particles, sta-
bilizing the graphene liquid. However, once the GNP surfaces were
saturated with DSP molecules, increasing DSP concentration would
lead to the formation of multiple DSP adsorption layers on GNP
interface. Based on this result, the DSP/GNP ratio is kept at 15%
in the following sections.

3.2. Influence of sonication time

3.2.1. Stability of suspension

Fig. 9. GNP retained in suspensions over time.
The sedimentations of GNP in aqueous suspension (6 g GNP in
300 g water) after 1 h and 24 h are shown in Fig. 8a and b respec-
tively. For the suspension without sonication (indicated by 0 in remained stable with negligible change in the concentration of sus-
Fig. 7), GNPs tend to agglomerate and precipitate to the bottom pended GNPs in the fluid. With an increase in sonication energy, a
after 1 h, resulting in a distinct separation of an almost transparent higher degree of deflocculation of the GNPs could be achieved to
liquid above the black precipitation. However, with sonication of allow more DSP to be adsorbed on their surface, which could then
15 min or more, the GNPs remained suspended in the aqueous sus- repel each other, as illustrated in Fig. 10. For the samples with 60
pensions after 1 h, as reflected by the uniform opaqueness in min or longer sonication, the GNPs remained stable in the suspen-
Fig. 7a. The extent of sedimentation after 24 h is shown in sion for up to 6 h without noticeable change in the concentration of
Fig. 7b, and the sedimentation of black precipitates is obvious for GNP retained in the suspension, as shown in Fig. 9.
samples of 30 min of sonication or less. This observation supports There are two repulsive forces existing between GNP particles
sonication period of at least 60 min to enhance the deflocculation as provided by polycarboxylate SP: steric stabilization and electro-
of GNP agglomerates and stabilization of GNPs in suspensions for static forces. Zeta potential (f) is a measure of electrostatic repul-
at least 1 day. sion between nanoparticles and is useful for characterizing the
The stability of each GNP-suspension over time is shown in dispersion efficiency [46,47]. The measured zeta potential is shown
Fig. 9. The quantitative assessment matches well with the visual in Fig. 11. For electrostatic reason, colloidal particles with |f| > 15
observation in Fig. 9. For the samples without sonication, the con- mV are expected to be stable [48]. In this study, the zeta potential
centration of GNPs suspended in fluid dropped to 26% of its original for plain GNP in water is below 15 mV, indicating that GNPs have a
concentration after a sedimentation time of 1 h. With an increase high tendency to agglomerate. With the addition of surfactant, |f|
in sonication time to 60 min or longer, almost all the GNPs of GNP increased to 20–40 mV, depending on the sonication time.
remained suspended in the fluid after 1 h. Over time, the GNPs In addition, the application of sonication could break the GNP clus-
were expected to settle to the bottom due to the gravity but the ters and thus more surfaces of GNP were available to absorb the
suspensions which have been sonicated for more than 1 h surfactant, leading to even higher |f|. Repulsive interactions are
required to stabilize the dispersed GNPs, therefore, the number
of GNPs suspended in the solution is expected to increase with
higher f-potential.
For fresh cement mixture, the stability of GNP-suspension
within the first few hours is vital since cement would not reach
the initial setting until 2–4 h [49]. From this study, 60 min of son-
ication seems to yield sufficiently stable GNP aqueous solution to
allow the cement to hydrate before sedimentation of the GNP in
the cement mixtures. If the GNP suspension is sonicated for 120
min, more than half of the GNPs can be retained in the suspension
after 168 h (7 days) of settling, which demonstrated the efficiency
of this sonication assisted liquid-phase dispersion and in good
agreement with previous studies [50,51].

3.2.2. Flowability of cement paste

When the GNP content and the DSP were maintained at 1% and
0.15 respectively, the slump diameter of pastes at different sonica-
tion times are shown in Table 5. It is noted that the use of DSP
alone would not yield paste as flowable as the reference paste. In
liquid suspension, GNP particles agglomerate to form clusters
which will only allow the DSP molecules to be adsorbed on the
outer surface of GNP clusters, rather than penetrating to the inner
graphene layers. As such, these flocs would hinder the movement
of cement particles in the fresh paste by locking water inside. With
Fig. 8. Photographs of GNP aqueous suspension after (a) 1 h and (b) 24 h.
the application of sonication, even as short as 15 min, those flocs
Sonication time varies from 0 (extreme left) to 15 min, 30 min, 60 min and 120 were forced to break up and the locked water was released. Conse-
min (extreme right). quently, the flowability of paste with sonicated suspension was
 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413 409

Fig. 10. Illustration of (a) Polycarboxylate SP molecule, (b) Agglomeration of GNPs, and (c) Dispersion of GNPs.

Fig. 11. Effect of sonication time on the zeta potential of GNP particles in
suspension. (Zeta potential of the colloidal GNP was determined using a potential
analyzer, ZetaPALS Brookhaven Instruments Corporation. Note: the color of the
suspension qualitatively reflects the higher GNP dispersion efficient with longer
sonication time.)

much increased. Beyond 15 min, the duration of sonication will

alter the flowability of fresh cement paste slightly.

3.2.3. Pore structures of cement paste

The influence of sonication time on the pore size distribution of
cement paste with constant 1% of GNP and 15% DSP is shown in
Fig. 12. The PSD curve was remarkably shifted towards left, indicat-
ing that the cement paste was refined with the inclusion of GNP Fig. 12. Influence of sonication time on MIP results.
even with only mechanical stirring. Ultrasonication appears to
slightly affect the pore size distribution. The effective intrusion
was significantly reduced with longer sonication period, especially longer than 30 min. Consequently, the potential advantages that
within the first hour. As expected, the longer sonication period can be realized with the addition of GNP, such as barrier effect,
could result in higher dispersion and stability for GNP in liquid. would be more evident at longer sonication period. It is also inter-
It is noted that the critical pore diameter of paste also decreased esting to note that a small increase in the effective intrusion was
from 67.7 nm to 54.2 nm with the application of ultrasonication observed for mechanically stirred GNP suspension. This result

Table 5
Influence of sonication time on cement paste flowability.

Mix ID GNP,% SP,% Sonication, minute Flowability, mm

GNP1-SP15-T0 1 15 0 215
GNP1-SP15-T15 1 15 15 253
GNP1-SP15-T30 1 15 30 241
GNP1-SP15-T60 1 15 60 233
GNP1-SP15-T120 1 15 120 238
GNP1-SP15-T180 1 15 180 248
410 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413

Table 6
Influence of GNP content on cement paste flowability.

Mix ID GNP,% SP,% Sonication, minute Flowability, mm

GNP0.25-SP15-T60 0.25 15 60 249
GNP0.5-SP15-T60 0.5 15 60 244
GNP1-SP15-T60 1 15 60 233
GNP2-SP15-T60 2 15 60 254
GNP4-SP15-T60 4 15 60 264

(a)

GNP

(b)

GNP

Fig. 13. Influence of GNP content on MIP results.

(c)

GNP

Fig. 15. SEM images of cement paste with 1% GNP.

Fig. 14. Hardness of graphene-cement nanocomposites. Error bar indicates the
standard deviation and the dots are the experimental data. Inset: Illustration of the
ball indentor and the reaction of the GNP-cement composites as well as the higher
resistance to shear band motion caused by the addition of GNP. microstructure while further sonication beyond the first hour did
not result in significant improvement.
revealed the importance of dispersion of nano-particles by ultra-
sonication. Otherwise, the microstructures of composites with 3.3. Influence of GNP content
non-uniformed nanoparticles might be compromised due to the
existence of GNP clusters [8]. From the results, a sonication of 1 From the previous studies, we have determined that sonication
h would be recommended to produce a denser cement paste time of 1 h at 210 W per 6 g of GNP and DSP dosage of 15% would
 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413 411

Fig. 16. SEM images of cement paste with 4% GNP. (a) Dispersion of GNPs, (b) Close-up view of selected interface, (c) Dispersion and clustering of GNP, and (d) Close-up view
of selected interface.

3.3.2. Hardness of cement paste

The enhancement of graphene on the mechanical properties has
been well documented for composites with various matrix, such as
polymer, ceramic, metal, geopolymer and cement [6,52–57],
though some studies fail to reproduce this enhancement due to
dispersion issues [58]. The extension of graphene to building mate-
rials application have been mainly focused on the enhancement in
strength, modulus and toughness under bending or tension for
cement paste [8,59–63]. Measurement of hardness is a fast indica-
tion of the effects of GNP on the mechanical properties of cement
paste since it is directly related to the strength of a material [64].
Fig. 17. XRD patterns for GNP and GNP-cement nanocomposites. More specifically, according to [65], macroscopic hardness of the
concrete is in direct proportion as its compressive strength.
Fig. 14 displayed the superficial Rockwell hardness for cement
produce stable GNP suspension. This section will focus on the pastes of varying GNP contents. From the results, the reference
influence of GNP content on cement paste properties. The flowabil- plain cement paste showed an average hardness of 5.7 GPa. With
ity of cement paste was maintained in the range of 244–264 mm, the addition of GNP, all cement pastes exhibited increasing hard-
slightly affected by the variation of added GNP, up to 4%, as shown ness due to the benefits of mechanical reinforcement by GNP. At
in Table 6. GNP content of 1%, the average hardness could be as high as
11.2 GPa, almost doubled in comparison with the reference paste.
Beyond 1% addition, the hardness showed however a reduction
3.3.1. Pore structure of cement paste to 8.8 GPa for 2% GNP and recovered to 11.6 GPa for 4% GNP.
For cement paste with constant sonication time of 1 h and DSP/ SEM photos clearly show evidence of good distribution of GNP
GNP ratio at 15%, the influence of GNP content on cement paste in the hardened cement matrix, as shown in Fig. 15 for 1% GNP
PSD is plotted in Fig. 13. Both the effective intrusion and critical and Fig. 15a and b for 4% GNP. The presence of good mechanical
pore size decreased continuously with more GNP addition of up interlocking and the presence of obstacles to the motion of shear
to 1%. However, no obvious change could be noticed beyond this bands contributes to the increased hardness of GNP-cement com-
GNP content. For up to 1% GNP, the input sonication energy and posites, as illustrated by Fig. 14 inset. However, at GNP content
DSP dosage were sufficient to disperse and stabilize the GNPs in of 2% and 4%, the GNPs might not be well dispersed and partially
the suspension to achieve well distributed GNPs in the hardened remained as clusters. Fig. 16c and d show GNP clustering where
cement paste. For 2% and 4% of GNP, a proportional increase in several GNPs overlapped and no cement hydration product was
energy was required to produce better GNPs distribution and to formed between those parallel sheets. A possible consequence aris-
promote lower porosity. At these two GNP contents with only 1 ing from the presence of these clusters is that these GNPs could
h of sonication, clusters and voids were observed which compro- readily separate or slide against each other when subject to tensile
mised the benefits of GNP on pore refinement. or shear stress around the cluster.
412 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413

Phase analysis was carried out by XRD, as displayed in Fig. 17. [4] B. Han, S. Sun, S. Ding, L. Zhang, X. Yu, J. Ou, Review of nanocarbon-engineered
multifunctional cementitious composites, Compos. A 70 (2015) 69–81.
With increasing GNP content of up to 4%, the intensity of peak cor-
[5] H. Du, S.D. Pang, S.T. Quek, Transport properties of cement mortar with
responding to graphene (2h = 26.7°, 0.335 nm) in the paste powder graphite nanoplatelet, in: 20th International Conference on Composites/Nano
became more evident while the other peaks corresponding cement Engineering. Beijing, July, 2012.
hydration products remained unaltered. Under mechanical load- [6] H. Du, S.D. Pang, Enhancement of barrier properties of cement mortar with
graphene nanoplatelet, Cem. Concr. Res. 76 (2015) 10–19.
ing, the load could transfer to graphene, which is much stronger [7] H. Du, J.H. Gao, S.D. Pang, Improvement in concrete resistance against water
than cement matrix, via the interaction at the interface between and chloride ingress by adding graphene nanoplatelet, Cem. Concr. Res. 83
graphene and cement paste. Therefore, improved resistance (2016) 114–123.
[8] A. Peyvandi, P. Soroushian, A.M. Balachandra, K. Sobolev, Enhancement of the
against deformation and fracture can be expected as long as a good durability characteristics of concrete nanocomposites pipes with modified
interface could be achieved. This observation is consistent with a graphite nanoplatelets, Constr. Build. Mater. 47 (2013) 111–117.
recent study [23]. [9] A. Peyvandi, P. Soroushian, Structural performance of dry-cast concrete
nanocomposites pipes, Mater. Struct. 48 (1) (2015) 175–186.
[10] T. Tong, Z. Fan, Q. Liu, S. Wang, S. Tan, Q. Yu, Investigation of the effects of
4. Conclusions graphene and graphene oxide nanoplatelets on the micro- and macro-
properties of cementitious materials, Constr. Build. Mater. 106 (2015) 102–114.
[11] H. Du, S.T. Quek, S.D. Pang, Smart multifunctional cement mortar containing
The dispersion method of GNP and its influences on the flowa- graphite nanoplatelet, in: Proceedings of SPIE 8692, San Diego, March, 2013,
bility and pore structures of cement paste were investigated exper- 869223.
imentally. A polycarboxylate based superplasticizer was found to [12] S.D. Pang, J.H. Gao, C. Xu, S.T. Quek, H. Du, Strain and damage self-sensing
cement composites with conductive graphene nanoplatelet, in: Proceedings of
be effective in dispersing the GNP, and 15% of this surfactant, by SPIE 9061, San Diego, March, 2014, 9061126.
weight of GNP, was found to be optimum in dispersion, without [13] M. Saafi, L. Tang, J. Fung, M. Rahman, F. Sillars, J. Liggat, Z. Zhou, Graphene/fly
affecting the workability of the cement paste. The separation of ash geopolymeric composites as self-sensing structural materials, Smart
Mater. Struct. 23 (2014) 065006.
the GNP was achieved via sonication of the aqueous suspension [14] J.-L. Le, H. Du, S.D. Pang, Use of 2-D graphene nanoplatelets (GNP) in cement
at 210 W. A sonication time of 1 h per 6 g of GNP (this coincides composites for structural health evaluation, Compos. B 67 (2014) 555–563.
with 1% GNP in this study) could retain all the GNPs in suspension [15] X. Cui, S. Sun, B. Han, X. Yu, J. Ouyang, S. Zeng, J. Ou, Mechanical, thermal and
electromagnetic properties of nanographite platelets modified cementitious
after 6 h. This provides a facile approach to make stable GNP liquid, composites, Compos. A 93 (2017) 49–58.
which can be used in many application like cement and concrete, [16] Q. Liu, Q. Xu, Q. Yu, R. Gao, T. Tong, Experimental investigation on mechanical
geopolymer and other composites, battery, etc. It should be noted and piezoresistive properties of cementitious materials containing graphene
and graphene oxide nanoplatelets, Constr. Build. Mater. 127 (2016) 565–576.
that the liquid in fresh cement paste is more complex, which con-
[17] Y. Cai, P. Hou, X. Cheng, P. Du, Z. Ye, The effects of nanoSiO2 on the properties
tains alkalis and hydroxide released from cement grains during the of fresh and hardened cement-based materials through its dispersion with
continuous hydration. Hence, the stability of GNP might not be silica fume, Constr. Build. Mater. 148 (2017) 770–780.
[18] M. Cai, D. Thorpe, D.H. Adamson, H.C. Schniepp, Methods of graphite
identical to what has been observed in the pure water. More
exfoliation, J. Mater. Chem. 22 (2012) 24992.
research is needed to study the stability of GNP in simulated pore [19] L. Guardia, M.J. Fernandezámerino, J.I. Paredes, P. Solis-Fernandez, S. Villar-
solution (rich in Ca2+, Na+, K+, OH , and SO24 ). Rodil, A. Martinze-Alonso, J.M.D. Tascon, High-throughput production of
Extended sonication beyond that only results in marginal pristine graphene in an aqueous dispersion assisted by non-ionic surfactant,
Carbon 49 (2011) 1653–1662.
change in the physical properties of the GNP. The cement paste [20] W. Meng, K.H. Khayat, Mechanical properties of ultra-high-performance
with 1% GNP had doubled in hardness while the effective porosity concrete enhanced with graphite nanoplatelets and carbon nanofibers,
and the critical pore diameter were reduced by 37% and 30%, Compos. B 107 (2016) 113–122.
[21] B. Han, Q. Zheng, S. Sun, S. Dong, L. Zhang, X. Yu, J. Ou, Enhancing mechanisms
respectively. At GNP content of 2% and 4%, the nanoplatelets were of multi-layer graphenes to cementitious composites, Compos. A 101 (2017)
not well separated, which resulted in weak zones formed by clus- 143–150.
ters and GNP-paste interface, thereby compromising the gains in [22] S. Sun, S. Ding, B. Han, S. Dong, X. Yu, D. Zhou, J. Ou, Multi-layer graphene-
engineered cementitious composites with multifunctionality/intelligence,
hardness and the refinement of the pores. Compos. B 129 (2017) 221–232.
[23] G. Jing, Z. Ye, X. Lu, P. Hou, Effect of graphene nanoplatelets on hydration
Acknowledgements behaviour of Portland cement by thermal analysis, Adv. Cem. Res. 29 (2017)
63–70.
[24] B. Wang, R. Jiang, Z. Wu, Investigation of the mechanical properties and
Mr. Rob Fleerakkers, from Eindhoven University of Technology, microstructure of graphene nanoplatelet-cement composite, Nanomaterials 6
is acknowledged for his great help with GNP processing and paste (2016) 200.
[25] B. Wang, R. Jiang, W. Song, H. Liu, Controlling dispersion of graphene
characterizations during his internship in NUS. The authors wish to
nanoplatelets in aqueous solution by ultrasonic technique, Russ. J. Phys.
acknowledge the support of A⁄Star SERC (grant number R-302- Chem. A 91 (2017) 1517–1526.
000-034-305). The authors would also like to thank Asbury Carbon [26] B. Wang, R. Jiang, R. Zhao, W. Song, H. Liu, Q. Jun, Investigation of a novel
in USA for providing GNP materials. The support of the Structural surfactant to modify the surface of graphene nanoplatelets with ultrasonic,
Nanosci. Nanotechnol. Lett. 9 (2017) 134–143.
Engineering Laboratory at National University of Singapore [27] X. Sun, Q. Wu, J. Zhang, Y. Qing, Y. Wu, S. Lee, Rheology, curing temperature
towards the experiments is acknowledged. Also, the authors are and mechanical performance of oil well cement: combined effect of cellulose
grateful to Prof. Hu Jiangyong and Dr. Cai Qinqing for their help nanofibers and graphene nano-platelets, Mater. Des. 114 (2017) 92–101.
[28] E. Wotring, P. Mondal, C. Marsh, Characterizing the dispersion of graphene
with the UV–vis spectrophotometer operation in Environmental nanoplatelets in water with water reducing admixtures, in: Nanotechnology in
Engineering Laboratory, NUS. The discussions with Dr. Dou Qingq- Construction, Switzerland: Springer, 2015, pp. 141–148.
ing inspired the authors to think deep of the dispersing mechanism [29] Z.S. Metaxa, Polycarboxylate based superplasticizers as dispersant agents for
exfoliated graphene nanoplatelets reinforcing cement based materials, J. Eng.
of surfactant. Sci. Technol. Rev. 8 (5) (2015) 1–5.
[30] F. Collins, J. Lambert, W.H. Duan, The influences of admixtures on the
References dispersion, workability and strength of carbon nanotube-OPC paste mixtures,
Cem. Concr. Compos. 34 (2) (2012) 201–207.
[31] A. Yazdanbakhsh, Z. Grasley, B. Tyson, R.K. Abu Al-Rub, Distribution of carbon
[1] K.S. Novoselov, A.K. Geim, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V.
nanofibers and nanotubes in cementitious composites, Transp. Res. Rec. 142
Grigorieva, A.A. Firsov, Electric field effect in atomically thin carbon films,
(2010) 89–95.
Science 306 (5696) (2004) 666–669.
[32] M.S. Konsta-Gdoutos, Z.S. Metaxa, S. Shah, Highly dispersed carbon nanotube
[2] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlhaas, E.J. Zimney, E.A.
reinforced cement based materials, Cem. Concr. Res. 40 (7) (2010) 1052–1059.
Stach, R.D. Piner, S.T. Nguyen, R.S. Ruoff, Graphene-based composite materials,
[33] M. Liebscher, A. Lange, C. Schrofl, R. Fuge, V. Mechtcherine, J. Plank, A.
Nature 442 (2006) 282–286.
Leonhardt, Impact of the molecular architecture of polycarboxylate
[3] O.C. Compton, S. Kim, C. Pierre, J.M. Torkelson, S.T. Nguyen, Crumpled
superplasticizers on the dispersion of multi-walled carbon nanotubes in
graphene nanosheets as highly effective barrier property enhancers, Adv.
aqueous phase, J. Mater. Sci. 52 (2016) 2296–2307.
Mater. 22 (42) (2010) 4759–4763.
 H. Du, S.D. Pang / Construction and Building Materials 167 (2018) 403–413 413

[34] S. Kawashima, J. Seo, D. Corr, M. Hersam, S.P. Shah, Dispersion of CaCO3 electrochemical, chemical, and sonication-assisted techniques: a nanoscale
nanoparticles by sonication and surfactant treatment for application in fly ash- study, Adv. Funct. Mater. 23 (2013) 4684–4693.
cement systems, Mater. Struct. 47 (6) (2014) 1011–1013. [51] A. Ciesielski, P. Samori, Graphene via sonication assisted liquid-phase
[35] J.H. Potgieter, C.A. Strydom, Determination of the clay index of limestone with exfoliation, Chem. Soc. Rev. 43 (2014) 381–398.
methylene blue adsorption using a UV-VIS spectrophotometric method, Cem. [52] D. Cai, J. Jin, K. Yusoh, R. Rafiq, M. Song, High performance
Concr. Res. 29 (11) (1999) 1815–1817. polyurethane/functionalized graphene nanocomposites with improved
[36] G. Ye, Experimental Study and Numerical Simulation of the Development of mechanical and thermal properties, Compos. Sci. Technol. 72 (6) (2012)
the Microstructure and Permeability of Cementitious Materials PhD Thesis, 702–707.
Delft University of Technology, The Netherland, 2003. [53] C.L.P. Pavithra, B.V. Sarada, K.V. Rajulapati, T.N. Rao, G. Sundararajan, A new
[37] H. Ma, Z. Li, Realistic pore structure of Portland cement paste: experimental electromechanical approach for the synthesis of copper-graphene
study and numerical simulation, Comput. Concr. 11 (4) (2013) 317–336. nanocomposite foils with high hardness, Sci. Rep. 4 (2014) 4049.
[38] C. Quercia, P. Spiesz, G. Husken, H.J.H. Brouwers, SCC modification by use of [54] M.A. Raza, A.V.K. Westwood, A.P. Brown, C. Stirling, Texture, transport and
amorphous nano-silica, Cem. Concr. Compos. 45 (2014) 69–81. mechanical properties of graphite nanoplatelet/silicone composites produced
[39] J. Hot, H. Bessaie-Bey, C. Brumaud, M. Duc, C. Castella, N. Roussel, Absorbing by three roll mill, Compos. Sci. Technol. 72 (3) (2012) 467–475.
polymers and viscosity of cement pastes, Cem. Concr. Res. 63 (2014) 12–19. [55] M. Saafi, L. Tang, J. Fung, M. Rahman, J. Liggat, Enhanced properties of
[40] Z.S. Metaxa, M.S. Konsta-Gdoutos, S.P. Shah, Carbon nanofiber cementitious graphene/fly ash geopolymeric composite cement, Cem. Concr. Res. 67 (2015)
composites: effect of debulking procedure on dispersion and reinforcing 292–299.
efficiency, Cem. Concr. Compos. 36 (2013) 25–32. [56] L.S. Walker, V.R. Marotto, M.A. Rafiee, N. Koratkar, E.L. Corral, Toughening in
[41] Z.S. Metaxa, J.W.T. Seo, M.S. Konsta-Gdoutos, M.C. Hersam, S.P. Shah, Highly graphene ceramic composites, ACS Nano 5 (4) (2011) 3182–3190.
concentrated carbon nanotube admixture for nano-fiber reinforced [57] K.Y. Shin, J.Y. Hong, S. Lee, J. Jang, Evaluation of anti-scratch properties of
cementitious materials, Cem. Concr. Compos. 34 (5) (2012) 612–617. graphene oxide/polypropylene nanocomposites, J. Mater. Chem. 22 (16)
[42] S. Parveen, S. Rana, R. Fangueiro, M.C. Paiva, Microstructure and mechanical (2012) 7871–7879.
properties of carbon nanotube reinforced cementitious composites developed [58] H. Wang, G. Xie, Z. Zhu, Z. Ying, Y. Zeng, Enhanced tribological performance of
using novel dispersion technique, Cem. Concr. Res. 73 (2015) 215–227. the multi-layer graphene filled poly(vinyl chloride) composites, Compos. A 67
[43] A. Sobolkina, V. Mechtcherine, V. Khavrus, D. Maier, M. Mende, M. Ritschel, A. (2014) 268–273.
Leonhardt, Dispersion of carbon nanotubes and its influences on the [59] S. Chuah, Z. Pan, J.G. Sanjayan, C.M. Wang, W.H. Duan, Nano reinforced cement
mechanical properties of the cement matrix, Cem. Concr. Compos. 34 (10) and concrete composites and new perspective from graphene oxide, Constr.
(2012) 1104–1113. Build. Mater. 73 (2014) 113–124.
[44] R. Gagne, A. Boisvert, M. Pigeon, Effect of superplasticizer dosage on [60] Z. Pan, L. He, L. Qiu, A.H. Korayem, G. Li, J.W. Zhu, F. Collins, D. Li, W.H. Duan, C.
mechanical properties, permeability, and freeze-thaw durability of high- M. Wang, Mechanical properties and microstructure of a graphene oxide-
strength concrete with and without silica fume, ACI Mater. J. 93 (2) (1996) cement composite, Cem. Concr. Compos. 58 (2015) 140–147.
111–120. [61] S. Lv, Y. Ma, C. Qiu, T. Sun, J. Liu, Q. Zhou, Effect of graphene oxide nanosheets
[45] R. Rastogi, R. Kaushal, S.K. Tripathi, A.L. Sharma, I. Kaur, L.M. Bharadwaj, of microstructure and mechanical properties of cement composites, Constr.
Comparative study of carbon nanotube dispersion using surfactant, J. Colloid Build. Mater. 49 (2013) 121–127.
Interface Sci. 328 (2) (2008) 421–428. [62] Z. Li, Z. Lu, D. Hou, L. Meng, G. Sun, C. Lu, Mechanism of cement paste
[46] Z. Sun, V. Nicolosi, D. Rickard, S.D. Bergin, D. Aherne, J.N. Coleman, Quantitative reinforced by graphene oxide/carbon nanotube composites with enhanced
evaluation of surfactant-stabilized single-walled carbon nanotubes: dispersion mechanical properties, RSC Adv. 5 (2015) 100598–100605.
quality and its correlation with zeta potential, J. Phys. Chem. C 112 (9) (2008) [63] N. Ranjbar, M. Mehrali, M. Mehrali, U.J. Alengaram, M.Z. Jumaat, Graphene
10692–10699. nanoplatelet-fly ash based geopolymer composites, Cem. Concr. Res. 76 (2015)
[47] D. Li, M.B. Muller, S. Gilje, R.B. Kaner, G.G. Wallace, Processable aqueous 222–231.
dispersion of graphene nanosheets, Nat. Nanotechnol. 3 (2008) 101–105. [64] B. Das, K. Eswar Prasad, U. Ramamurty, C.N.R. Rao, Nano-indentation studies
[48] P.C. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, on polymer matrix composites reinforced by few-layer graphene, Nanotechnol
Marcel Dekker, New York, 1997. 20 (12) (2009) 125705.
[49] S. Mindess, J.F. Young, D. Darwin, Concrete, Prentice Hall, New Jersey, 2003. [65] D.N. Winslow, A Rockwell Hardness Test for Portland Cement Concrete. Joint
[50] Z.Y. Xia, S. Pezzini, E. Treossi, G. Giambastiani, F. Corticelli, V. Morandi, A. highway research project technical report, Purdue University, West Lafayett,
Zanelli, V. Bellani, V. Palermo, The exfoliation of graphene in liquids by USA, 1981.