Accepted Manuscript

Experimental study of the effects of graphene oxide on microstructure and

properties of cement paste composite

Haibin Yang, Manuel Monasterio, Hongzhi Cui, Ningxu Han

PII: S1359-835X(17)30280-4
DOI: http://dx.doi.org/10.1016/j.compositesa.2017.07.022
Reference: JCOMA 4742

To appear in: Composites: Part A

Received Date: 3 April 2017

Revised Date: 3 July 2017
Accepted Date: 21 July 2017

Please cite this article as: Yang, H., Monasterio, M., Cui, H., Han, N., Experimental study of the effects of graphene
oxide on microstructure and properties of cement paste composite, Composites: Part A (2017), doi: http://dx.doi.org/
10.1016/j.compositesa.2017.07.022

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
 Experimental study of the effects of graphene oxide on
microstructure and properties of cement paste composite

Haibin Yang+, Manuel Monasterio +, Hongzhi Cui*, Ningxu Han

Guangdong Provincial Key Laboratory of Durability for Marine Civil Engineering,

College of Civil Engineering, Shenzhen University, Shenzhen 518060, China

+
The authors contributed equally to this paper.
* Corresponding author: email: h.z.cui@szu.edu.cn; Tel: +86-755-2691 7849.

Abstract
Graphene oxide (GO) has been utilized to strengthen composite materials. In this
study, the effects of GO on hydration degrees, macro-mechanical strength and
calcium-silicate-hydrate (C-S-H) structure of cement based composites were
investigated through comprehensive experimental tests. In addition, the aggregation
mechanism of GO was verified by alkaline solution simulations, using Ca(OH)2 and
NH3·H2O. Based on the experimental results, it was found that the 3-day and 7-day
compressive strengths of cement based composites with 0.2 wt% of GO were
increased by 35.7% and 42.3%, respectively as compared to the control. Moreover,
the C-S-H structure of cement paste with GO was not observed to have undergone any
change via qualitative and quantitative analyses combined with FT-IR, XRD and
29 29
Si-NMR. Besides, the test results of TGA, DTG and Si-NMR showed that the
hydrated degree of cement paste increased to 10.4% at 28 days when incorporating
with 0.1% of GO.

Keywords: Graphene oxide (GO); Cement based composites; Hydration and

microstructures; calcium-silicate-hydrate (C-S-H).

Abbreviations
GO graphene oxide
C-S-H calcium-silicate-hydrate

1
OPC ordinary Portland cement
PC polycarboxylate superplasticizer
MCL mean chain length

1. Introduction

Traditional cement based composites are still widely used in infrastructure systems
of all types, such as building, roads, railways, bridges, tunnels, and harbors. However,
with development of society, the use of traditional cement based composite is difficult
to meet the requirements of infrastructure systems for durability and mechanical
properties, because the materials possess low resistance to fracture or crack
propagation. Cement based composites applying at civil engineering will be easier
and popular if these issues can be successfully mitigated. While additives such as steel
bar, carbon fiber, steel fiber or polypropylene fiber have been introduced to modify
the cement based composites, the benefits from these changes are realized on the
macro scale and have not yet been shown to decrease porosity or prevent the
formation of micro-cracks.
Nowadays, there are many different nanomaterials, that have been investigated as
additives to cement for use in civil engineering applications, for example, nano-silica,
nano-fiber, especially nanotube, graphene and their derivates [1–4]. These
nanomaterials are not just added as cement fillers; they have important applications in
others topics, for instance, as paste cement damage sensors [5–7]. Based on previous
studies, these nanomaterials present unique advantages and certain disadvantages:
zero-dimensional materials (nano-particles) including nano-silica and titanium
dioxide, lack the ability to arrest the growth of cracks, and act only as filling-materials,
Pozzolanic materials or nucleation seeds. The one-dimensional materials (nanowires
or carbon nanotubes) have a strong Van der Waal’s attractive force among them,
which provoke that they cannot play the intended roles without good dispersion.
GO is a graphene derivate and a kind of two-dimensional material. GO shares the
same intrinsic properties with graphene, including a unique atom-thick sp2 bond 2D
2
structure, super-high specific surface (~2600 m2g-1), high thermal conductivity
(~3000W m-1 K-1), excellent Young’s modulus (~1100 GPa), high tensile strength
(~125 GPa), tremendous mechanical strength (~1060 GPa) and stiffness [8–10].
Furthermore, GO has a wide range of oxygen functional groups such as epoxides,
ketone carbonyls, hydroxyl and carboxylic group on both basal planes and edges. The
functional group has been shown to improve the dispersion of GO in water [11,12].
After oxidation, GO becomes a non-conductive hydrophilic carbon material, which
leads to an increase in graphite interlayer spacing from 0.335 nm to 0.625 nm.
Additionally, GO achieves better dispersion behavior for above-mentioned changes
[13,14]. As shown in the research of Lin et al. [15], one of the most important roles of
GO in cement paste is that the functional groups can provide growing points and
planes for cement hydration products, i.e. C-S-H and Ca(OH)2. Moreover, Pan and
Wang suggested that GO has a strong covalent bond either between GO and C-S-H or
between GO and Ca(OH)2, through the carboxylic acid group [16,17].
In recent years, many publications have reported the effects of GO on properties of
cementitious materials. Most of these studies focus on mechanical strength, hydration,
bond strength, dispersion effect and durability. Wang et al. [18] reported that GO may
take part in the cement hydration, accelerating the nucleation, growth and phase
separation of hydrated products, thus promoting the hydration process and improving
the crystal order, which resulted in an enhancement of 90.5% and 52.4% in flexural
and compressive strength of the cement pastes, respectively. Lu et al. [19,20]
combined GO with carbon nanotubes (CNTs) to obtain excellent dispersion and to
reinforce the mechanical strength of cement based materials. Mohammed et al. [21]
proposed that the cement pore structure could be refined by GO, which increased
durability greatly, and the experimental results also showed that GO has the potential
to modify the cement matrix to exhibit high freeze-thaw resistance.
It has been shown that the main component providing strength in hardened Portland
cements is the amorphous C-S-H phase [22]. However, the current research about GO,
is more focused on the chemical reactions between the carboxylic acid group in the
functional groups of GO and the C-S-H gel. Works about the possible changes
3
produced in the microstructure of the C-S-H by GO have not been revealed. Despite
much research on cement based materials with GO, there is a lack of qualitative and
quantitative investigation into the effect of GO on the hydration process of cement
based materials.
In order to fill the above-mentioned knowledge gaps, this study systematically
investigates the effect of GO on the C-S-H structure and the hydration process of
29
cement paste with different GO dosage. The study was done by using Si NMR and
other advance characterization techniques, including thermogravimetric analysis
(TGA), X-ray diffraction (XRD), scanning electron microscopy (SEM), and Fourier
transform Infrared spectroscopy (FT-IR). The factors for enhancing the mechanical
strength of cement paste are also explained in this research.

2. Materials and methods

2.1 Materials

Type I ordinary Portland cement (OPC) was used in this study; its chemical
composition is shown in Table 1. The material was provided by China building
materials academy. Industrial graphene oxide (TNIGO) was supplied by Chengdu
Organic Chemicals CO. Ltd., and its physical properties are shown in Table 2.
Polycarboxylate superplasticizer (PC), provided by Sika (Switzerland), served as a
surfactant to disperse the GO within the cement paste. In order to eliminate the
influence of air bubbles in the cement paste, an anti-foaming agent was added during
the mixing of cement pastes.

Table 1. Chemical composition of OPC

SiO2 Al203 Fe2O3 CaO MgO SO3 Na2O f-CaO Cl- Loss
21.80 4.55 3.45 64.40 2.90 2.45 0.532 0.93 0.011 1.27

Table 2. Physical properties of the GO (TNIGO)

Properties Purity Size Thickness Layers Selling State
Values > 97 wt% 2-10 μm 0.55-3.0 nm <5 Brown Powder

4
2.2 Characterization techniques

The strengths of the cement paste composites were tested according to

GB/T17671-2007 (Method of testing cements – Determination of strength). The
loading speed of compressive strength was 2.4 kN/second.
Characterization of the as-received GO was performed by Atomic Force
Microscope using a Bruker ICON-PT-PKG in tapping mode. The XPS system was a
ULVAC-PHI VPII model and the measurement was performed using 4 scans per eV.
Transmission electron microscopy (TEM) images of the GO morphology were
obtained using a JEM-1230. A Bruker AVANVE IIITM 600MHz NMR spectrometer
equipped with a 7 mm MAS broadband probe was employed in the NMR tests of the
samples. The tests operated at 119.28 MHz for 29Si, and the MAS spin rate was 5 KHz.
Scanning electron microscope (FEI, QUANTA FEG 250), and energy dispersive
spectrometer (EDS) were used to determine the size of GO, hybrid morphology and
integration of GO with hydration products. The Ca(OH)2 content was determined by
thermogravimetric analysis (TGA) using a NETZSCH STA 409 PC at 5 K/min,
starting from room temperature (28 oC) to 1000 oC. The phase composition of the
cement matrix, with different ages and external dosages of GO, was measured with a
Bruker D8 ADVANCE (XRD), which uses Cu tube radiation (λ=1.5418 Å) and 2θ in
the range of 2.5˚ to 40˚. The measurements at wide angle (5˚ to 80˚) were performed
at 0.3 seconds per scan and the measurements at narrow angle (3˚ to 30˚) were
performed at 1 second per scan with the addition of a screen to reduce the background
noise signal.
The Raman spectrum was performed at room temperature in a Renishaw inVia (UK)
model. The spot size was approx. 1µm, using 50x objective, 5% of the laser power, 10
seconds of exposure and 3 accumulative measurements. The laser wavelength was
532 nm.
The Fourier transform Infrared spectra of the cement based composite were
obtained using a Perkin Elmer Spectrum 100, and tested from 500 to 4000 cm-1 using
32 scans per measurement.
5
2.3 Characterization of as-received GO

The GO in its as-received state was measured by several methods to obtain its
characteristic parameters. Firstly, the GO as-received was analyzed by spectroscopy
techniques, specifically TEM, AFM and Raman spectrum devices. Figure 1 shows the
results from TEM and AFM, which were used to evaluate the morphology and size of
the GO as received. The measured thickness was between 2 and 3 nm.

a)

b) 2
1

Figure 1. GO as received morphology and size characterization. a) TEM image of the GO; and b)
AFM image where is plotted two different scan lines for determining the GO heights.

Characterization of the purity and the number of GO layers were performed by

XPS and XRD measurements respectively, as shown in Figure 2. XPS results
indicated a very pure GO, finding only the peaks related with carbon and oxygen
atoms. The carbon peak was fitted with Lorenz curves for a larger binding analysis. A
greater amount of C-C and C-O bindings was found, similar to that reported in the
literature [23,24]. The elemental analysis determined 67.4% of C and 31.2% of O, so
the C/O ratio is 2.16, slightly lower than the ratio reported by other authors [25,26].
6
With respect to the number of layers, the XRD results showed a unique and intense
peak at 10.7˚. Through the Bragg’s equation, it is well-known that this degree value
corresponds with 0.824 nm, which is the interlayer distance between GO sheets.
The Raman spectrum measurement showed 3 different peaks, 2 of them at 1350
and 1580 cm-1, designed D and G peaks, respectively, and three small peaks, at 2500 –
3250 cm-1, designed 2D peaks. D peak is due to the breathing modes of six-atom rings,
and G peak corresponds to the E2g phonon at the Brillouin zone center [27–29]. The
presence of G band is owing to the stretching of the C-C bond in graphitic materials
[30]. The 2D peaks are the second order of the D peaks. It is a single peak when the
graphene is a single-layer sample. When the graphene presents a larger number of
layer, as GO, this peak splits in four peaks, reflecting the evolution of the electron
band structure [31,32]. The ratio was 0.92.

Characterization of the GO as-received showed high purity with a well-defined

interlayer distance and small number of layers.

7
Figure 2. GO as-received measurements by XPS and XRD devices. a) XPS result where it can be
found O1s and C1s peaks. The inset shows the fitting curves of carbon peak, using
Gaussian curves. Different carbon bindings are indicated. b) XRD results where only
one peak is obtained at 10.7˚. It is known that this degree coresspond to 0.824 nm, the
interlayer distance in GO. c) Raman spectrum results. Bands D, G and 2D are indicated.

2.4 GO dispersion simulation test

To study the GO dispersion behavior in fresh cement paste (alkali solution), the
saturated Ca(OH)2 solution and NH3·H2O (pH=13.5) were used to simulate the pore
solution environment of cement paste. GO suspension was prepared by the following
procedures: firstly, GO was mixed with water and PC, and the mix solution was
stirred by magnetic stirring apparatus for 15 mins at 300 rpm. Then, the mix solution
was exposed to violent ultrasonic agitation (Xinzhi, JY92-IIN, China) for 30 mins at
390W power with pulses of 2 seconds. Two different GO suspensions were prepared.

8
One containing 0.2 wt% GO and the other one was a hybrid solution containing 0.2%
of GO and 0.28% of PC (GO@PC). During the simulation test, the saturated Ca(OH)2
solution and NH3·H2O was dropped into two GO suspensions, respectively. Then, GO
dispersion behavior was observed visually.

2.5 Preparation of GO/cement paste composite

GO can reduce the fluidity of cement paste for its super large specific surface [18].
In order to achieve the same workability and acquire a reasonable mix proportion, PC
was used to improve the fluidity of cement paste in accordance with the method of
GB/T8077-2000 (Method for testing uniformity of concrete admixture). The mix
proportions are listed in Table 3.
In this study, 40×40×40 mm samples were used for compressive strength test
according to the requirements of China national standard of GB/T17617-2007 (Test
methods of strength of cement mortar). For sample preparation, GO suspension and
polycarboxylate superplasticizer were mixed into the cement paste with a
water/cement ratio of 0.4. In the process of mixing the GO/cement paste; the water,
superplasticizer and GO suspension were added into the cement and mixed for 3
minutes at low-speed and then another 1 minute at high-speed. After the mixing was
completed, the GO/cement paste was dropped into the 40×40×40 mm steel molds.
The samples were cast in two layers and vibrated for 30s for each layer. In order to
prevent water loss, all samples were covered with plastic films. After 24 hours, the

samples were removed from the mold and then were kept in a curing room (20  1 oC
and 98% RH) for 1, 3, 7, 14 or 28 days before testing.

Table 3. Mix proportions of GO/cement paste composite

Sample Cement Water GO SP Antifoam Fluidity
denomination (Normalized) (Normalized) (wt%) (wt%) (wt%) Diameter(mm)
GC 1 0.4 0 0.04 0.12 183
G10 1 0.4 0.10 0.16 0.12 185
G15 1 0.4 0.15 0.22 0.12 186
G20 1 0.4 0.20 0.28 0.12 195
9
3. Results and discussion
3.1 GO agglomerate tests in alkaline solutions

Two comparison groups were prepared and shown in Figure 3. It can be clearly
observed by the naked eyes that GO agglomerated in the Ca(OH)2 solution. No GO
agglomeration was found in the GO@PC solution. A similar dispersion mechanism
was suggested by Babak et al. They proposed that the strong steric impediment effect
on PC based superplasticizers generated a repulsion between the cement particles [33].
The carboxylic acid group in PC would attach onto the GO surface, increasing the
hydrophilic groups. The van der Waals force between the GO sheets would,
consequently, be reduced and the energy required to pull them from the cement matrix
would be increased. There are two clear advantages of adding PC into the cement
matrix: a moderate amount of PC could contribute to an appropriate fluidity without
affecting the final cement strength, and the GO is able to disperse sufficiently within
the cement matrix.
Posterior experiments validate the fact that Ca2+ caused the GO to form aggregates.
In contrast, the addition of NH3·H2O in the alkaline solution presented a good
dispersion even without the addition of PC. This phenomenon confirmed that the
concentration of hydroxyls has not been affected by the GO addition and it further
revealed that the replacement reactions of carboxylic acid groups by Ca2+ and H+ led
to GO agglomeration[34,35].

a) b)
Figure 3. Tests for GO agglomerate behaviour. a). GO content generated agglomerates in the
Ca(OH)2 (left) and displayed better dispersion in the NH3·H2O disolutions (Right); b).
PC@GO had better dispersion in both Ca(OH)2 (Left) and NH3·H2O dissolution (Right).

10
3.2 Macro-mechanical strength of cement paste

The compressive strength of different samples is presented in Table 4 and Figure 4.

It is clearly shown that GO additions improved compressive strength, especially in
early ages. In a quantitative comparison, the compressive strengths of G15 and G20
were increased by 30.7% and 42.3% after 3 days, compared with the control sample,
GC. These GO effects are in agreement with previous research [33,34,36,37]. From
Table 4, the maximum improvement in compressive strength compared with the
control sample was reached after 3 days for all the additive groups. This implies GO
played a major role in improving compressive strength during their nucleation seed
effect during the early stages of curing. After 3 days, the compressive strength
improvements gradually decreased with the curing age, reaching 4.8%, 7.4% and 11.2%
after 28 days for G10, G15 and G20 respectively. Figure 4b) presents the strength
improvements in function of GO content. It can be shown from the figure that the
strength of sample GC increased by 51.9% between day 1 and day 3. For the same
period, the compressive strength of samples G10, G15 and G20 increased by 56.4%,
69.1% and 81.8% respectively, when compared to the control. These strength
improvements substantiated the conclusion that GO can significantly affect the
hydration process. From Figure 4, it can be known that, the higher the GO content
into the cement is; the larger the improvement is. In engineering application, the user
can choose different GO dosage based on the corresponding compressive strength
requirement.
In order to confirm the improvement and reveal the mechanism of reinforcement,
TGA was used to determine the effect of hydration after 1, 3, 7, 14 and 28 days. SEM
was used to investigate the micro-effect of bonding and agglomeration. XRD, FT-IR
and NMR were employed to illustrate the structure of C-S-H at 28 days.
Table 4. Compressive strength of the specimens at 1, 3, 7, 14 and 28 days
Compressive strength (Increase percentage)
Age (days)
GC G10 G15 G20
1 18.5 (100) 21.4 (115.7) 21.7 (117.3) 22.0 (119.2)
3 28.1 (100) 35.0 (124.7) 36.7 (130.7) 40.0 (142.3)

11
 7 33.9 (100) 38.5 (113.6) 39.5 (116.4) 46.0 (135.7)
14 42.1 (100) 47.9 (113.7) 49.1 (116.7) 51.1 (121.3)
28 49.3 (100) 51.7 (104.8) 53.0 (107.4) 54.9 (111.2)
Note: The compressive strength is expressed in MPa. In parentheses are the normalized values with percentages in
function of pure OPC.

Figure 4. Compressive strength development of cement paste. a) Effect of curing time on

compressive strength of cement paste; and b) Effect of GO content on compressive
strength of cement paste.

3.3 Effect of GO on hydration characteristics of cement paste

In order to prove the relationship between the hydration process and mechanical
strength, the hydration degrees were quantitatively analyzed by TGA and DTG as
shown in Figure 5. In this figure, the DTG presented obvious inflection points, which
showed a subtle change in the hydration process when the temperature increased.
TGA provided accurately the Ca(OH)2 content, which reflected hydration degree.
According to several previous studies [38–41], different steps of the reaction process
were summarized as follow: 30-105 oC: the evaporable water and part of the bound
water escapes; 110 - 300 oC: the loss of bound water from the decomposition of the
C-S-H; 450 - 550 oC: dehydroxylation of the Ca(OH)2; 650-800 oC: decarbonation of
calcium carbonate.
Figure 5b) shows the Ca(OH)2 content as a function of GO. In this figure, it is
shown that the Ca(OH)2 amount gradually increased with curing time except for 1 day
of curing time. At this early hydration step, the GO could absorb a certain amount of

12
water molecules, due to its large surface area. This absorption reduced the free water
content required for full hydration, leading to a reduction in the Ca(OH)2 content for
larger doses [33,42]. Another factor related with the reduction of Ca(OH) 2 content is
the adsorption, during the initial hydration process, of Ca2+ ions upon GO surface
[43,44]. This is in contradiction with the compressive strength values and their
respective hydration degree, shown previously in Figure 4b. However, the variation in
compressive strength with GO content at this early stage of curing is small and this
conclusion may be affected by experimental error. Furthermore, at this stage of curing
it is likely that the effect of adding GO dominated when compared with the effects of
hydration. Once hydrated phases found nucleation seed points, the hydration
increased rapidly, which facilitated the formation of the interfacial bonding between
the active sites of GO and hydrated phases of cement, with both mechanisms leading
to improved mechanical strength. Additionally, the water molecules on the GO
surface, may act as a water reservoir and transport channels for further hydration of
the cement [15].

Figure 5. Effect of GO on cement paste composites. a) TGA/DTG test results; and b) Relationship
of GO content and Ca(OH)2 concentration.

3.4 Integration of GO sheets on the cement pastes

The interaction between GO sheets and cement paste were analyzed by SEM and
XRD measurements. Figure 6a) shows SEM images of G20 samples after 28 days of
curing time. Through EDS analysis, a hybrid materials of GO and hydration products

13
were found, because elements of hydration products were found as well. Observation
of a fractured sample clearly demonstrated that the GO settled on the hydration
products. Figure 6b) and c) show the EDS analysis of spot 1 and 2, indicated in Figure
6a). These EDS measurements confirmed the presence of GO sheets in the cement
paste samples and their complete integration into the cement. Similar results have
been reported by Li et al., who reported the chemical reactions between the carboxylic
acid of GO and the C-S-H or Ca(OH)2 of the cement matrix [45]. The GO addition
distributed the stress of external loadings and reduced the pressure over C-S-H due to
a better interaction with cement matrix.
Figure 7 shows an agglomeration of GO that was observed in the sample. Although
the aggregations rarely formed, their existence in these samples indicates that the
mechanical properties could have been further improved with a better dispersion of
GO. It may be caused by a soaring Van der Waals force on cement sample and a
broken electrostatic equilibrium. This effect was not found in dissolution samples.
Uddin et al. [46–48] discovered different solvent to affect the dispersion of GO in
function of their different Van der Waals force. Moreover, needle-like crystals formed
in the cement based composite, which are presented in Figure 7. The needle-like
crystallites may be induced by the GO aggregations, because the hydration products
are, in general, an interpenetrating network formed by polyhedral forms at 28 days of
curing time [49].

14
 a)

EDS Spot 1 EDS Spot 2

Figure 6. Microstructural analysis of GO modified cement paste. a) SEM images sample G20 of
hardened cement paste of bonding relationship between GO with C-S-H at 28 days of
hydration; and b) EDS for hybrid GO and hydration products; and c) EDS for pure
hydration products.

Figure 7. Aggregation of GO inside cement paste composite in G20 sample at 28 days of curing.

3.5 Effect of GO on C-S-H structure

Analysis of the interactions on C-S-H structure by GO were performed by several

15
techniques, which are explained below.

3.5.1 Analysis based on FT-IR tests

The FT-IR transmittance spectra of pure GO, GC, G10, G15 and G20 at 28 days of
curing are presented in Figure 8a). The spectrum of GO illustrates C-O stretching
vibration at 1042 cm-1, C-O-C strength vibrations at 1154 cm-1, C=C at 1620 cm-1
assigns to skeletal vibrations of unoxidized graphitic dominates, C=O in carboxylic
acid and carbonyl moieties at 1722 cm-1, phenolic C-OH stretching at 1343 and 3300
cm-1, Phenolic C-COOH stretching at 3300 cm-1 and the bond from 2265 to 2420
indicates H-O-H bending band of the absorbed H2O molecules [9,20,33,37,42,50].
The intense band at 955 cm-1 is attributed to the stretching vibrations of Si-O bonds of
C-S-H gel with a tetrahedral shape [15,51,52]. Both peaks at 820 to 830 and 1100 to
1200 cm-1 belong to the characteristic sulfates absorption bands caused by the v3
vibration of the SO-24 group in sulfates. From Figure 8a), it becomes obvious that
there are no changes for cement based composites in presence of GO, except for a
small CO3 content produced probably by different storage conditions. Figure 8b)
shows a comparison between G20 and pure GO signals, where the only peak present
in both samples is related to CO3, however, with differing signal intensity. The
absence of changes in the measurements is due to two reasons; the amount of GO
used in the cement pastes was not enough to be detected by the FT-IR sensitivity
range. Others studies have indicated that a larger proportion of GO content is
necessary to be visible by FT-IR, with percentages of even 3% of GO [9,15]. The
other reason that the addition of GO may not change the C-S-H structure. As further
discussed for NMR-MAS results in Section 3.5.3, the compressive stress
improvement due to GO additions are more related to the interaction of GO with
hydrated products than changes in the microstructure of the C-S-H gel.

16
Figure 8. FT-IR transmittance spectra of a) different cement samples; and b) pure GO and GC
samples.

3.5.2 Analysis based on XRD tests

On the way to judge the crystal phase of hydration crystals over different dosages,
XRD tests between 5˚ and 80˚ degrees were performed, and the results are shown in
Figure 9a). Tricalcium silicate (C3S), dicalcium silicate (C2S) and Ca(OH)2 were
detected. The results of hydration phase crystals exhibited by XRD were the same as
the results obtained by TGA, which further verified the observed results. It must be
mentioned that the amount of Ca(OH)2 crystalline phase increased with the GO
concentration, but the pattern of crystalline phase was similar. Although, XRD could
not detect the C-S-H gel, due to its amorphous structure, the C-S-H quantity can be
estimated by Ca(OH)2 [36]. Within these parameters exposed above and despite the
limitations of XRD measurements, the results demonstrated that GO would not
change the structure of C-S-H but only accelerated the hydration process.

17
 The measurements performed at small angle and using higher time per scan, which
are shown in Figure 9b), display the interlayer distance of GO. There is an obvious
change in the signal intensity between pure GO and other samples due to the different
proportions of GO content. The most important result is the interlayer distance
obtained in the samples G10, G15 and G20. The interlayer GO distance is strongly
related to the oxidation and hydration degree of GO where a higher hydration or
oxidation level means longer distance [53,54]. So, the samples G10, G15 and G20
showed a larger amount of water or oxygen molecules between layers owing to the
cement hydration process, because some of these water or oxygen molecules have
been trapped in the GO interlayer. These adsorption properties affect the fluidity of
the cement pastes and may slow down the cement hydration process during the early
steps. However, these trapped molecules improved the hydration degree in posterior
steps. This result is in agreement with TGA and mechanical results and with other
studies [15,33,42].

a) b)

Figure 9. Results of XRD test at 28 days of curing time. a) The XRD patterns of the hardened
cement pastes at different dosages of GO added. Positions marked 1 are peaks produced
by Ca(OH)2. Positions marked 2 are produced by C3S and C2S phases; and b) the
interlayer GO distance for different samples.

3.5.3 Analysis based on NMR tests

29
Si MAS NMR spectroscopy was used as a method to identify the C-S-H
structure. It is generally accepted that the C-S-H phase has a tobermorite-like
18
structure, which includes CaO2 layers with seven coordinated Ca2+ ions, where the
oxygen atoms are shared with Si4+ in chains of SiO4 tetrahedra [55]. Additionally, it is
also known that the spectrum of the silicate anion is represented in the range of -70
and -76 ppm for Q0, -78 for Q1, -82.8 for Q2p and -84.5 for Q2b as it is shown in Figure
10. Qn sites (n = 0 - 4) are usually observed in relatively separate ranges, which allow
for the assignment of the silicate connectivity [56]. The Q0 sites represent C3S content
or dehydrated cement, Q1 species correspond to the end-chain silicate tetrahedral, and
Q2 represents the middle-chain silicate tetrahedral, with two neighboring Q1 sites
where the middle-chain silicate can be bound to a bridging positions of a dreierketten
structure (Q2b)[57,58] or to protons (Q2p)[59]. The deconvolution of the experimental
data of each silicate anion was fitted with a Gaussian function [51]. Richardson’s
equation was used to calculate the mean chain length (MCL) in the C-S-H [60]:

Eq. 1

29
The Si-NMR spectrums of samples after 14 and 28 days of curing time are
shown in Figure 11, and all the information of the calculated deconvolution are
displayed in Table 5. The hydration degrees were quantitatively obtained using the Q0
area. These values were 53.0 %, 55.3 %, 59.7 % and 59.4 % after 14 days, and 56.9 %,
67.3 %, 67.3 % and 65.4 % after 28 days for GC, G10, G15 and G20 samples,
respectively. The hydration degree was higher for cement samples with GO additions
than the pure cement sample. These results are in agreement with TGA and XRD
measurements. Furthermore, they prove definitively that the improvement in the
hydration degree exists with respect to the control sample GC. The MCL values,
calculated through Eq. 1, showed a similar C-S-H long value between 3.1 to 3.3
silicon tetrahedrals. This variance in the values was within the error of measurement
and reflected a similar C-S-H structure for all samples. The result is in agreement with
FT-IR and XRD measurements. Considering the discussed conclusions regarding
bonding with hydration degree, the added GO in OPC only reacted chemically with

19
the carboxylic acid and it could not change the structure of C-S-H.

Figure 10. Schematic illustration of 29Si-NMR spectrums. Black line represented experimental
values. Green, yellow, blue and purple colors lines represented the Gaussian
deconvolution of Q0, Q1 Q2p and Q2b peaks, respectively. Finally, red dashed line
indicated the fitting values of the convolutions.

Figure 11. 29Si-NMR spectrums of specimens at 14 and 28 days of curing.

20
 Table 5. Percentages of silicate chemical shift from 29Si-NMR spectra.
Curing time 14 days 28 days
0 1 2 0 1
Sample Q (%) Q (%) Q (%) MCL Q (%) Q (%) Q2 (%) MCL
GC 47.0 31.4 21.5 3.374 43.1 48.6 29.1 3.199
G10 44.7 35.0 20.2 3.153 32.7 42.3 25.0 3.181
G15 40.3 37.7 21.9 3.165 32.7 40.6 26.6 3.311
G20 40.6 37.2 22.1 3.188 34.6 40.2 25.3 3.260

4. Conclusions

1) The Ca2+ ions were the primary reason for the GO aggregation in the cement
based composites. PC acted as a good dispersant that induced a better GO dispersion
into Ca(OH)2 solution through mixing and sonication. The dispersion effect was
produced by a strong steric hindrance effect. This was notable for the dispersion in
G20 sample, where the compressive strength values at 3 and 7 days were 42.3 % and
35.7 %, respectively, higher than that of the control.
2) SEM measurements showed GO surfaces were coved by hydration products,
which illustrated how GO sheets bonded with OPC hydration products for further
improvements in mechanical strength results. The SEM results also demonstrated that
GO aggregation occurred in some samples. This fact revealed that the mechanical
strength could be further enhanced if better dispersion was in place. Results of XRD
and TGA proved that there was an enhancement in the Ca(OH)2 content, which is
29
strongly related to the hydration degree. The Si-NMR measurements showed that
the addition of GO improved the hydration degree values.
3) The higher interlayer GO distance found in G10, G15 and G20 samples in
comparison with pure GO could explain the poor amount of Ca(OH)2 at early stages
of curing. During the initial stage of hydration, GO absorbed water, Ca2+ ions and
oxygen molecules, which restricted the growth of Ca(OH)2. However, this absorption
effect improved the hydration degree at the end of the curing process as shown in
TGA results.
4) The results from 29Si-NMR, FT-IR and XRD indicated that GO has no influence
21
on C-S-H structure. The results from 29Si-NMR showed a similar MCL values for all
the measured samples. Despite of the possible resolution issue, the MCL differences
between the samples were very small. So, it can be considered that GO did not change
the C-S-H microstructure and the improvements in mechanical properties mainly
came from the hydration degree acceleration and chemical reactions.

Acknowledgments
The work described in this paper was fully supported by grants from Natural
Science Foundation of China (51678366) and Shenzhen Fundamental Research
Funding (JCYJ20160422092836654).

Reference
[1] Silvestre J, Silvestre N, de Brito J. Review on concrete nanotechnology. Eur J Environ Civ Eng
2015;8189:1–31. doi:10.1080/19648189.2015.1042070.
[2] Stephens C, Brown L, Sanchez F. Quantification of the re-agglomeration of carbon nanofiber
aqueous dispersion in cement pastes and effect on the early age flexural response. Carbon N Y
2016;107:482–500. doi:10.1016/j.carbon.2016.05.076.
[3] Hou D, Lu Z, Li X, Ma H, Li Z. Reactive molecular dynamics and experimental study of
graphene-cement composites: Structure, dynamics and reinforcement mechanisms. Carbon N Y
2017;115:188–208. doi:10.1016/j.carbon.2017.01.013.
[4] Singh AP, Gupta BK, Mishra M, Govind, Chandra A, Mathur RB, et al. Multiwalled carbon
nanotube/cement composites with exceptional electromagnetic interference shielding properties.
Carbon N Y 2013;56:86–96. doi:10.1016/j.carbon.2012.12.081.
[5] Azhari F, Banthia N. Cement-based sensors with carbon fibers and carbon nanotubes for
piezoresistive sensing. Cem Concr Compos 2012;34:866–73.
doi:10.1016/j.cemconcomp.2012.04.007.
[6] Materazzi AL, Ubertini F, D’Alessandro A. Carbon nanotube cement-based transducers for
dynamic sensing of strain. Cem Concr Compos 2013;37:2–11.
doi:10.1016/j.cemconcomp.2012.12.013.
[7] Konsta-Gdoutos MS, Aza CA. Self sensing carbon nanotube (CNT) and nanofiber (CNF)
cementitious composites for real time damage assessment in smart structures. Cem Concr
Compos 2014;53:162–9. doi:10.1016/j.cemconcomp.2014.07.003.
[8] Dong L long, Chen W ge, Deng N, Zheng C hao. A novel fabrication of graphene by chemical
reaction with a green reductant. Chem Eng J 2016;306:754–62. doi:10.1016/j.cej.2016.08.027.
[9] Horszczaruk E, Mijowska E, Kalenczuk RJ, Aleksandrzak M, Mijowska S. Nanocomposite of
cement/graphene oxide - Impact on hydration kinetics and Young’s modulus. Constr Build
Mater 2015;78:234–42. doi:10.1016/j.conbuildmat.2014.12.009.
[10] Kuilla T, Bhadra S, Yao D, Kim NH, Bose S, Lee JH. Recent advances in graphene based
polymer composites. Prog Polym Sci 2010;35:1350–75.
doi:10.1016/j.progpolymsci.2010.07.005.
22
[11] Geim AK, Novoselov KS. The rise of graphene. Nat Mater 2007;6:183–91.
doi:10.1038/nmat1849.
[12] Zhu Y, Murali S, Cai W, Li X, Suk JW, Potts JR, et al. Graphene and graphene oxide: Synthesis,
properties, and applications. Adv Mater 2010;22:3906–24. doi:10.1002/adma.201001068.
[13] Parades JI, Villar-Rodil S, Martínez-Alonso A, Tascón JMD. Graphene oxide dispersions in
organic solvents. Langmuir 2008;24:10560–4. doi:10.1021/la801744a.
[14] Daniel R. Dreyer, Sungjin Park CWBRSR. The chemistry of graphene oxide. ChemSocRev
2009;39:228–40. doi:10.1039/b917103g.
[15] Lin C, Wei W, Hu YH. Catalytic behavior of graphene oxide for cement hydration process. J
Phys Chem Solids 2016;89:128–33. doi:10.1016/j.jpcs.2015.11.002.
[16] Wang M, Wang R, Yao H, Farhan S, Zheng S, Du C. Study on the three dimensional
mechanism of graphene oxide nanosheets modified cement. Constr Build Mater
2016;126:730–9. doi:10.1016/j.conbuildmat.2016.09.092.
[17] Pan Z, He L, Qiu L, Korayem AH, Li G, Zhu JW, et al. Mechanical properties and
microstructure of a graphene oxide-cement composite. Cem Concr Compos 2015;58:140–7.
doi:10.1016/j.cemconcomp.2015.02.001.
[18] Wang Q, Wang J, Lu C-X, Liu B-W, Zhang K, Li C-Z. Influence of graphene oxide additions
on the microstructure and mechanical strength of cement. New Carbon Mater 2015;30:349–56.
doi:10.1016/S1872-5805(15)60194-9.
[19] Sun G, Liang R, Lu Z, Zhang J, Li Z. Mechanism of cement/carbon nanotube composites with
enhanced mechanical properties achieved by interfacial strengthening. Constr Build Mater
2016;115:87–92. doi:10.1016/j.conbuildmat.2016.04.034.
[20] Li X, Wei W, Qin H, Hang Hu Y. Co-effects of graphene oxide sheets and single wall carbon
nanotubes on mechanical properties of cement. J Phys Chem Solids 2015;85:39–43.
doi:10.1016/j.jpcs.2015.04.018.
[21] Mohammed A, Sanjayan JG, Duan WH, Nazari A. Graphene Oxide Impact on Hardened
Cement Expressed in Enhanced Freeze – Thaw Resistance 2016;28:2–7.
doi:10.1061/(ASCE)MT.1943-5533.0001586.
[22] Hansen MR, Jakobsen HJ, Skibsted J. 29Si chemical shift anisotropies in calcium silicates from
high-field 29Si MAS NMR spectroscopy. Inorg Chem 2003;42:2368–77.
doi:10.1021/ic020647f.
[23] Fan Z, Wang K, Wei T, Yan J, Song L, Shao B. An environmentally friendly and efficient route
for the reduction of graphene oxide by aluminum powder. Carbon N Y 2010;48:1686–9.
doi:10.1016/j.carbon.2009.12.063.
[24] Park S, An J, Piner RD, Jung I, Velamakanni A, Nguyen ST, et al. Aqueous Suspension and
Characterization of Chemically Modified Graphene Sheets Chemically Modified Graphene
Sheets. Chem Mater 2008;20:6592–4. doi:10.1021/cm801932u.
[25] Stankovich S, Dikin DA, Piner RD, Kohlhaas KA, Kleinhammes A, Jia Y, et al. Synthesis of
graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon N Y
2007;45:1558–65. doi:10.1016/j.carbon.2007.02.034.
[26] Gao W, Alemany LB, Ci L, Ajayan PM. New insights into the structure and reduction of
graphite oxide. Nat Chem 2009;1:403–8.
[27] Tuinstra F, Koenig JL. Raman Spectrum of Graphite. J Chem Phys 1970;53:1126–30.
doi:10.1063/1.1674108.

23
[28] Ferrari AC, Robertson J. Interpretation of Raman spectra of disordered and amorphous carbon.
Phys Rev B 2000;61:95–107.
[29] Ferrari AC, Robertson J. Resonant Raman spectroscopy of disordered , amorphous , and
diamondlike carbon. Phys Rev B 2001;64:1–13. doi:10.1103/PhysRevB.64.075414.
[30] Some S, Kim Y, Yoon Y, Yoo H, Lee S, Park Y, et al. High-Quality Reduced Graphene Oxide
by a Dual-Function Chemical Reduction and Healing Process. Sci Rep 2013;3:1–5.
doi:10.1038/srep01929.
[31] Ferrari AC, Meyer JC, Scardaci V, Casiraghi C, Lazzeri M, Mauri F, et al. Raman Spectrum of
Graphene and Graphene Layers. Phys Rev Lett 2006;97:1–4.
doi:10.1103/PhysRevLett.97.187401.
[32] Cançado LG, Reina A, Kong J, Dresselhaus MS. Geometrical approach for the study of G’
band in the Raman spectrum of monolayer graphene , bilayer graphene , and bulk graphite.
Phys Rev B 2008;77:1–9. doi:10.1103/PhysRevB.77.245408.
[33] Babak F, Abolfazl H, Alimorad R, Parviz G. Preparation and mechanical properties of graphene
oxide: Cement nanocomposites. Sci World J 2014;2014. doi:10.1155/2014/276323.
[34] Wang M, Wang R, Yao H, Wang Z, Zheng S. Adsorption characteristics of graphene oxide
nanosheets on cement. RSC Adv 2016;6:63365–72. doi:10.1039/C6RA10902K.
[35] Chowdhury I, Mansukhani ND, Guiney LM, Hersam MC, Bouchard D. Aggregation and
Stability of Reduced Graphene Oxide: Complex Roles of Divalent Cations, pH, and Natural
Organic Matter. Environ Sci Technol 2015;49:10886–93. doi:10.1021/acs.est.5b01866.
[36] Sharma S, Kothiyal NC. Influence of graphene oxide as dispersed phase in cement mortar
matrix in defining the crystal patterns of cement hydrates and its effect on mechanical,
microstructural and crystallization properties. RSC Adv 2015;5:52642–57.
doi:10.1039/c5ra08078a.
[37] Zhao L, Guo X, Ge C, Li Q, Guo L, Shu X, et al. Investigation of the effectiveness of PC@GO
on the reinforcement for cement composites. Constr Build Mater 2016;113:470–8.
doi:10.1016/j.conbuildmat.2016.03.090.
[38] Zajac M, Rossberg A, Le Saout G, Lothenbach B. Influence of limestone and anhydrite on the
hydration of Portland cements. Cem Concr Compos 2014;46:99–108.
doi:10.1016/j.cemconcomp.2013.11.007.
[39] Taylor HFW. Cement Chemistry:2nd edition. 1990.
[40] Hewlett P. Lea’s Chemistry of Cement and Concrete. 2004.
[41] Villain G, Thiery M, Platret G. Measurement methods of carbonation profiles in concrete:
Thermogravimetry, chemical analysis and gammadensimetry. Cem Concr Res
2007;37:1182–92. doi:10.1016/j.cemconres.2007.04.015.
[42] Shang Y, Zhang D, Yang C, Liu Y, Liu Y. Effect of graphene oxide on the rheological properties
of cement pastes. Constr Build Mater 2015;96:20–8. doi:10.1016/j.conbuildmat.2015.07.181.
[43] Li X, Korayem AH, Li C, Liu Y, He H, Sanjayan JG, et al. Incorporation of graphene oxide and
silica fume into cement paste: A study of dispersion and compressive strength. Constr Build
Mater 2016;123:327–35. doi:10.1016/j.conbuildmat.2016.07.022.
[44] Park S, Lee K-S, Bozoklu G, Cai W, Nguyen ST, Ruoff RS. Graphene oxide papers modified
by divalent ions-enhancing mechanical properties via chemical cross-linking. ACS Nano
2008;2:572–8. doi:10.1021/nn700349a.
[45] Lu Z, Hou D, Meng L, Sun G, Lu C, Li Z. Mechanism of cement paste reinforced by graphene

24
 oxide-carbon nanotubes composites with enhanced mechanical properties. RSC Adv
2015;5:100598–605. doi:10.1039/C5RA18602A.
[46] Gudarzi MM. Colloidal Stability of Graphene Oxide: Aggregation in Two Dimensions.
Langmuir 2016;32:5058–68. doi:10.1021/acs.langmuir.6b01012.
[47] Guardia L, Fernández-Merino MJ, Paredes JI, Solís-Fernández P, Villar-Rodil S,
Martínez-Alonso A, et al. High-throughput production of pristine graphene in an aqueous
dispersion assisted by non-ionic surfactants. Carbon N Y 2011;49:1653–62.
doi:10.1016/j.carbon.2010.12.049.
[48] Uddin ME, Kuila T, Nayak GC, Kim NH, Ku BC, Lee JH. Effects of various surfactants on the
dispersion stability and electrical conductivity of surface modified graphene. J Alloys Compd
2013;562:134–42. doi:10.1016/j.jallcom.2013.01.127.
[49] Lv S, Ting S, Liu J, Zhou Q. Use of graphene oxide nanosheets to regulate the microstructure
of hardened cement paste to increase its strength and toughness. CrystEngComm
2014;16:8508–16. doi:10.1039/C4CE00684D.
[50] Lu Z, Hou D, Ma H, Fan T, Li Z. Effects of graphene oxide on the properties and
microstructures of the magnesium potassium phosphate cement paste. Constr Build Mater
2016;119:107–12. doi:10.1016/j.conbuildmat.2016.05.060.
[51] Monasterio M, Gaitero JJ, Erkizia E, Guerrero Bustos AM, Miccio LA, Dolado JS, et al. Effect
of addition of silica- and amine functionalized silica-nanoparticles on the microstructure of
calcium silicate hydrate (C-S-H) gel. J Colloid Interface Sci 2015;450:109–18.
doi:10.1016/j.jcis.2015.02.066.
[52] Saafi M, Tang L, Fung J, Rahman M, Liggat J. Enhanced properties of graphene/fly ash
geopolymeric composite cement. Cem Concr Res 2015;67:292–9.
doi:10.1016/j.cemconres.2014.08.011.
[53] Cossio MLT, Giesen LF, Araya G, Pérez-Cotapos MLS, VERGARA RL, Manca M, et al.
Springer Handbook of Nanomaterials. 2013. doi:10.1007/978-3-642-20595-8.
[54] Cerveny S, Barroso-Bujans F, Alegría Á, Colmenero J. Dynamics of Water Intercalated in
Graphite Oxide. J Phys Chem C 2010;114:2604–12. doi:10.1021/jp907979v.
[55] Andersen MD, Jakobsen HJ, Skibsted J. Characterization of white Portland cement hydration
and the C-S-H structure in the presence of sodium aluminate by 27Al and 29Si MAS NMR
spectroscopy. Cem Concr Res 2004;34:857–68. doi:10.1016/j.cemconres.2003.10.009.
[56] Balachandra AM, Ph D. Monitoring of Sulfate Attack in Concrete by MAS NMR Spectroscopy
2013;27:1–10. doi:10.1061/(ASCE)MT.1943-5533.0001175.
[57] Klur I, Pollet B, Virlet J, Nonat A. CSH structure evolution with calcium content by
multinuclear NMR. Nucl Magn Reson Spectrosc Cem Mater 1998:119–141.
[58] Monasterio M, Gaitero JJ, Erkizia E, Guerrero Bustos AM, Miccio LA, Dolado JS, et al. Effect
of addition of silica- and amine functionalized silica-nanoparticles on the microstructure of
C-S-H gel. J Colloid Interface Sci 2015;450:109–18. doi:10.1016/j.jcis.2015.02.066.
[59] Matsushita F, Aono Y, Shibata S. Calcium silicate structure and carbonation shrinkage of a
tobermorite-based material 2004;34:1251–7. doi:10.1016/j.cemconres.2003.12.016.
[60] Richardson IG. Tobermorite/jennite- and tobermorite/calcium hydroxide-based models for the
structure of C-S-H: applicability to hardened pastes of tricalcium silicate, β-dicalcium silicate,
Portland cement, and blends of Portland cement with blast-furnace slag, metakaolin, or silica
fume. Cem Concr Res 2004;34:1733–77. doi:10.1016/j.cemconres.2004.05.034.

25