Advanced searchCitation search
Skip main navigation
Close Drawer MenuOpen Drawer Menu
Previous article
Free access
Technical Papers
Feb 18, 2013

Materials Genome for Graphene-Cement Nanocomposites

Authors: Hunain Alkhateb [email protected], Ahmed Al-Ostaz [email protected], Alexander H.-D. Cheng [email protected], and Xiaobing Li [email protected]Author Affiliations
Publication: Journal of Nanomechanics and Micromechanics
Volume 3, Issue 3
https://doi.org/10.1061/(ASCE)NM.2153-5477.0000055
PDF

Abstract

Graphene nanoplatelets have unique mechanical, thermal, and electrical properties that render them ideal reinforcing materials. The attractive properties of graphene have led to intensive research on graphene-polymer nanocomposites. However, very little work has been reported on using graphene in manufacturing multifunctional cement-based nanocomposites. This paper attempts to bridge recent findings of science-based discovery and nanoscience to the ancient and challenging technology of cement. Utilizing a holistic approach (i.e., integrating modeling, synthesis, and analysis of cement) is a challenge that cannot be fully addressed in one paper or one research experiment. Therefore, this paper presents a general framework for using a system approach to study cement-based materials. The paper highlights primary findings in manufacturing and characterizing graphene-cement nanocomposites (GCNCs). A bottom-up approach is used to correlate the atomic assembly of GCNCs with their macroscopic properties. At the atomic level, X-ray diffraction is used to predict the chemical composition and crystallography of GCNCs. At a nanoscale level, atomic force microscopy (AFM) is used to examine the physical and chemical properties of GCNCs. Molecular dynamics (MD) analysis is conducted to estimate the interfacial strength between calcium silicate hydrate (C-S-H) and the graphene nanoplatelets functionalized with different chemical groups. At a microscale level, scanning electron microscopy (SEM) is used to obtain information about surface topography and the composition of GCNCs. At a mesoscale level, mechanical properties are measured using resonant ultrasound spectroscopy (RUS). This multiscale evaluation showed a strong correlation between the morphology and performance of GCNCs. Functionalizing graphene nanoplatelets tends to improve interfacial strength, which tends to improve the overall mechanical properties.

Introduction

During the last several decades, the need for high-performance structural materials and components has led to the rapid development of new classes of materials. The revolutionary Feynman vision of nanotechnology states that “the essence of nanotechnology is the ability to work at the molecular level, atom by atom, to create large structures with fundamentally new molecular organization.” The Feynman vision of a powerful and general nanotechnology opened the door for more research in utilizing nanomaterials for a variety of applications. New classes of nanomaterials such as carbon nanotubes, nanofibers, nanowires, and quantum dots are being assembled atom by atom with various high-tech applications in mind, e.g., electronics, biomedicine, energy, and the environment. For applications in civil infrastructure (e.g., bridges, dams, and buildings), these materials are still very expensive and can only be produced in relatively small quantities, which limits their applications.
Recent research revealed that traditional infrastructure materials (e.g., concrete) may be characterized by closely examining constituents that contain intricate nanostructures. In these cases, the materials’ macroscopic properties can be drastically altered by manipulating the nanostructure during their manufacturing process without the need to add nanoparticles. For example, the long-term creep (decades to a century) of cement is controlled by the packing density of its basic constituent, C-S-H (calcium-silicate-hydrate), at the nanoscale (Bazant 1972; Vandamme and Ulm 2009). However, adding a small amount of nanoadditives may also enhance the properties of concrete dramatically. To understand the effect of adding these nanoadditives to cement, a multiscale evaluation needs to be adapted. This paper employs a multiscale approach in evaluating the performance of graphene-cement nanocomposites (GCNCs) with a special emphasis on evaluating the mechanical properties. The following sections will review recent findings with respect to GCNCs.

Genome of C-S-H

The understanding of the fundamental building blocks of materials is termed material genome (National Science and Technology Council 2011). In this section the writers examine the genome of C-S-H.
The strength of concrete originates from hydration products. The major portion of the hydration products is usually in the form of a rigid gel termed C-S-H. In other words, C-S-H gel is responsible for the strength and cohesion of concrete structures. The research of Pellenq and Van Damme (2004) provides preliminary understanding regarding the factors that control the setting and hardening of cement paste. Short-range and medium-range attractive electrostatic forces are the paradigmatic components of the cohesion of C-S-H gel. More in-depth information about the functionality of C-S-H requires more detailed investigations of the C-S-H atomic structure (Pellenq and Van Damme 2004).
C-S-H gel is not a perfect crystalline material, composed of elements and pores of various sizes. Whereas some studies have revealed an amorphous structure (Van Damme and Gmira 2006), others have proposed crystalline structures of C-S-H (Manzano et al. 2007; Richardson and Groves 1992; Richardson 2004; Taylor 1986). More than 30 crystalline C-S-H structures have been identified. Most C-S-H structure knowledge has been acquired from its crystalline phases, primarily represented by the 14-Å tobermorite and jennite. The amorphous structure is extremely challenging to model and very little information is available in the literature.
Typically, C-S-H is defined by the calcium-to-silicon (Ca/Si) ratio of the gel, which ranges between 0.6 and 2.3 (Selvam et al. 2009). This value is usually identified from the compositional standpoint. However, the stoichiometry of C-S-H is not fixed and the chemical composition changes from point to point within the cement paste. To characterize the nanostructure of C-S-H accurately, a detailed description needs to be provided, e.g., the variations of the Ca/Si ratio, silicate structure, and contents of Si-OH and Ca-OH (Chen et al. 2004).
Tobermorite has a molecular structure characterized by a Ca/Si ratio of 0.83 and a density of 2.18  g/cm3. Jennite has a Ca/Si ratio of 1.5 and a density of 2.27  g/cm3. Richardson (1999) suggested a two-fold classification to clarify C-S-H chemistry. This classification is termed tobermorite/jennite (T/J) models or tobermorite/calcium hydroxyl (T/CH) models. The latter models are composed of solid tobermorite layers sandwiched between calcium hydroxyl, which allows them to achieve a higher Ca/Si ratio than in tobermorite alone. The T/J model is assembled of tobermorite regions, followed by jennite domains. The outcome of the research of Richardson (1999) suggests that the C-S-H gel structure may contain both glass-like short-range order and crystalline properties of the mineral tobermorite. Their suggested atomic structures are based on the Taylor (1997) models for two essential C-S-H minerals; i.e., tobermorite, 14-Å Ca5Si6O16(OH)2·7H2O, and jennite, Ca9(Si6O18)(OH)6·8H2O.
Evaluation of C-S-H using a tunneling electron microscope (TEM) revealed the shape of C-S-H as disks of 5-nm thickness. During hydration these C-S-H particles multiply and aggregate to form low-density (LD) and high-density (HD) C-S-H gels, which have different mechanical properties. The proportion of the high-density and low-density C-S-H depends on the mixing design of cement paste. Jennings (2000) proposed a colloidal structure for C-S-H with different packing densities. As the packing density increases, the number of contacts between the particles increases. Therefore, the HD C-S-H results in higher stiffness and hardness compared with LD C-S-H because of the existence of a higher contact number of points. However, the Jennings (2000) model does not quantify the bond between the C-S-H particles.
By combining small-angle neutron scattering measurements and X-ray scattering data, and by exploiting the hydrogen/deuterium neutron isotope effect both in water and methanol, for C-S-H the Ca/Si ratio is 1.7 and the density is 2.6  g/cm3 (Allen et al. 2007). The C-S-H structure is comprised of a layered structure at small length scales (1–5 nm). The layers stack to form a compact domain on the order of a few nanometers, which is the same order of the interlayer distance of tobermorite or jennite crystals. At larger length scales (5–100 nm), these ordered stacks form three-dimensional (3D) structures termed C-S-H particles.
For the work presented in this paper, the C-S-H molecular structure suggested by Pellenq et al. (2009), from the Massachusetts Institute of Technology (MIT), is used for molecular dynamic (MD) simulations. In the molecular structure of C-S-H (see Fig. 1), short silica chains are distributed to facilitate the interaction between CaO, SiO2, and H2O to provide realistic values of C/Si ratios and densities. These chains take the form of monomers, dimers, and pentamers. Using this atomic structure, other elementary structural characteristics and fundamental physical properties of GCNCs can be predicted.
Fig. 1. Atomic model for C-S-H suggested by Pellenq et al. (2009) (CaO)1.65(SiO2)(H2O)1.75

Graphene Nanoplatelets

The nanoscience of carbon materials has been widely investigated in the last 2–3 decades. More recently, a renewed interest in graphene has emerged. The Nobel Prize for physics in 2010 was awarded to Andre Geim and Konstantin Novolesov for groundbreaking experiments extracting monolithic layers of graphene (Royal 2011).
Graphene is a single-layer sp2-bonded carbon sheet that forms a honeycombed crystal lattice. It is first introduced by Mouras et al. (1987) as the two-dimensional (2D) form of graphite. Novoselov et al. (2004, 2005) reported that these 2D carbon materials formed gigantic flat fullerene molecules, and this is the first report to describe their electronic properties. Lee et al. (2008) reported a Young’s modulus of 1.0 TPa and an intrinsic strength of 130 GPa measured by nanoindentation atomic force microscopy (AFM) for the monolayer graphene sheet. By those measurements, Lee et al. (2008) categorized the graphene monolayer sheet as the strongest material ever measured; consequently, graphene nanocomposites are expected to greatly enhance mechanical properties.
Exfoliated graphene nanoplatelets (xGnP) have the same chemical structures as carbon nanotubes (CNT). Their edges can be easily modified chemically for dispersion enhancement in polymeric composites. These nanoplatelets are typically less than 5-nm thick and can be synthesized with lateral dimensions ranging from less than 1 μm up to 100 μm (see Fig. 2). The use of exfoliated graphite flakes opens up many new applications in which electromagnetic shielding, electrical conductivity, high thermal conductivity, gas-barrier resistance, high fracture toughness, and low flammability are required.
Fig. 2. Nanoplatelets: (a) SEM morphology of xGnP showing 1-μm to 2-μm nanoplatelets; (b) TEM micrograph of individual xGnP platelets; (c) TEM micrograph showing an edge view of two nanoplatelets
An understanding of properties of exfoliated graphene nanoplatelets mixed with cement is still in its infancy. Therefore, there is a need to determine the roles of exfoliated and functionalized graphene (FG) nanoplatelets at the nanoscale to predict more precisely the behavior and performance of GCNCs.

Carbon-Cement Nanocomposites

Multiscale research and its emphasis at the nanoscale have the potential to resolve the 3D packing of hydrated cement phases and their interactions with nanoparticles. The addition of nanoparticles to cement will enhance the nanoscale and microscale structure. The behavior of cement mixed with CNTs and carbon nanofibers (CNFs) has been studied by several researchers, e.g., Abu Al-Rub et al. (2011), Cwirzen et al. (2008), and Larson et al. (1990). However, only very limited work has been reported on the behavior of cement mixed with exfoliated pristine graphene nanoplatelets and functionalized exfoliated graphene nanoplatelets. Hence, an urgent need emerges to decode the thermal, mechanical, and physical behaviors of GCNC constituents.
Addition of small amounts of graphene nanoplatelets to cement may cause an increase in the composite material’s toughness, especially tensile, flexural, and impact strengths. To understand how cement properties are improved by graphene nanoplatelets and to understand the level of adhesion and interfacial failure modes, which are necessary to obtain optimum nanoreinforced cement properties, various functionalized graphene nanoplatelets are embedded in C-S-H. Their effects on the graphene-cement adhesion and the composite material properties may be determined using fiber pull-out MD simulations.

Specimen Preparation

Graphene Nanoplatelet Dispersions

A 0.5% by weight solution of 50-μm exfoliated pristine graphene nanoplatelets are mixed with cement to manufacture GCNCs. To confirm a homogenous distribution of the nanoparticles, graphene nanoplatelets are suspended in water. A solution of 0.162 g of exfoliated graphene nanoplatelets in 8.955 g of water is sonicated for 3 min.

Graphene Nanoplatelet Functionalization

By sonicating different specimens with graphene (water ratio of 1∶55 for a duration of 3 min), functionalized graphene nanoplatelets are dispersed in water. The procedure used for functionalizing graphene is similar to the work proposed by Rosca et al. (2005). For example, 3 g of pristine graphene and 300 mL of nitric acid (HNO3) are mixed in a 500-mL glass flask and sonicated for 15 min. The mixture is stirred and refluxed for 24 h before being connected at the top of the flask to a condenser (Fig. 3). After cooling, the mixture is purified by filtration using distilled water. Prior to each filtration, graphene-distilled water solution is sonicated for 5 min. The latter step is repeated 6×.
Fig. 3. Process of functionalizing xGnP
Thermal gravimetric analysis (TGA) testing is performed by heating specimens in nitrogen from room temperature to 700°C at 10°C/min. Fig. 4 shows typical TGA results for functionalized graphene. Fig. 4 shows that functionalized graphene has less weight loss compared with that of the pristine material, which could be attributable to the removal of impurities during the functionalization process using nitric acid.
Fig. 4. TGA results for xGnP and functionalized xGnP

Graphene-Cement Composites

The graphene platelets are dispersed by sonication in water. A 0.162-g sample of functionalized graphene is added to 8.955 g of distilled water in a small glass bottle. The mixture is then sonicated for about 3 min using a bath sonicator (FS30D, Fisher Scientific, Massachusetts). Sufficient portland cement Type I/II is added to produce a 0.5% by weight graphene/cement ratio. After the sonication process, a manual beater is used to mix the cement with graphene. The mixing time ranges from 15–30 min depending on the workability of the mixture.
The mixture is cast in 2×2×2-cm molds and cured at room temperature for 7 days. After curing the test specimens, they are cut and prepared for several multiscale tests. The tests include, on the atomic level, AFM and X-ray diffraction, and on the microscale and mesoscale, scanning electron microscopy (SEM) and resonance ultrasound spectroscopy (RUS), respectively.

Molecular Dynamics Simulations

The writers performed MD simulations using Materials Studio 4.2. Prior to molecular mechanics calculation, each molecular configuration is subjected to energy minimization, followed by molecular dynamics simulation to ensure thermodynamic equilibration. For geometry optimization, the target minimum derivative is set to be equal to 0.1  kcal/Å. The initial optimized geometry is then subjected to two-step temperature relaxation. In this study, a constant number of atoms, constant volume, and constant temperature (NVT) ensemble, controlled by a Hoover thermostat (Frenkel and Smit 2002; Hoover 1985) is employed. The total trajectory time monitored for the molecular dynamics simulations is set for 400 ps on each structure. The first stage of the temperature relaxation is set at 300 K for 200-ps dynamic time with a 0.10-ps time step. The second stage is set at 300 K for 200 ps with a 10-ps time step. The dynamic trajectories for the total relaxation time of 400 ps are compared with selected longer time evolutions (e.g., 600 ps) to validate thermodynamic equilibrium. In this paper, the atom based summation method, with a cutoff point of 8.5 Å, is adopted for the van der Waals interactions. For long-range Coulombic interactions, the Ewald summation method is implied.
Molecular dynamics simulations rely on the chosen empirical force field, which accounts for intermolecular and intramolecular interactions for simulated atomic structures. Adapting a well-parameterized force field is a challenging task. This is the focus of ongoing research by the writers of this paper. For simplicity, in this paper, the condensed—phase optimized molecular potentials for atomistic simulation studies (COMPASS) force field is adopted. COMPASS enables relatively accurate and simultaneous predictions of structural, conformational, vibrational, and thermophysical properties for a broad range of molecules in isolated, condensed phases, and under a wide range of temperature and pressure conditions. It accounts for two function category valence terms, including the diagonal and off-diagonal cross-coupling terms and the nonbond interaction terms. In the COMPASS force field, the valence terms represent internal coordinates of bonds, angles, torsion angles, and out-of-plane angles; the cross-coupling terms include two or three internal coordinate potentials. The nonbond interaction terms include van der Waals interactions for the Lennard-Jones function and a Coulombic function for electrostatic interactions (Sun 1998).
Molecular dynamics simulations are performed to study the atomic interfacial strength between C-S-H and functionalized graphitic structures. Three functionalized graphitic structures are simulated, as follows: (1) hydroxyl (−OH), (2) amine (−NH2), and (3) carboxyl (−COOH). Additionally, the interfacial strength of C-S-H pristine graphene is simulated.

Functionalized Graphene Molecular Structure

An infinite graphite plane with oxygen-containing and nitrogen-containing functionalities has been constructed. For an infinite network, atoms are made visible in accordance with the center of geometry of connected sets of atoms, so that only complete molecules are displayed. This representation should be suitable for most types of structures in which the visibility is controlled by the atomic coordinates. The periodic graphene unit cell platelet is constructed with lattice parameters of a=b=2.46  Å, c=3.40  Å, α=β=90°, and γ=120°. The unit cell is then replicated 5× along the a-axis, and 11× along the b-axis; lattice angles α=92.65°, β=80.01°, and γ=121.68° are adjusted accordingly to match the C-S-H atomic structure. The pristine graphene structure is constructed of 110 atoms. The functionalized graphene nanoplatelets molecular structures are constructed similarly. Figs. 5–8 display the atomic structures for the simulated graphene nanoplatelets. The functionalities considered for this study are pristine graphene, −OH, −NH2, and −COOH. Each of the functionalized graphene nanoplatelets has 10 functional groups along the b-axis and is capped with hydrogen atoms along the a-axis. Hence, this functionalization will lead to 18.18% coverage of the total area of the graphene sheet.
Fig. 5. Atomic model for pristine graphene G()
Fig. 6. Atomic model for hydroxyl-functionalized graphene G(OH)
Fig. 7. Atomic model for amine-functionalized graphene G(NH2)
Fig. 8. Atomic model for carboxyl-functionalized graphene G(COOH)
The MD simulations noted previously allow the constructed computational cells to evolve in accordance with Newton’s law of motion and the predefined COMPASS force field.

Hybrid C-S-H-G()-C-S-H Composite

Fig. 9 displays a typical GCNC hybrid composite for the simulated structures. The three-layered sandwich layout of C-S-H-G()-C-S-H is constructed. A 30-Å vacuum is introduced at the top of the simulated structure using Material Studio 4.2. The 30-Å vacuum slab is added to ensure that there are no interactions with the periodic image of the computational cell. In other words, the added thickness of the vacuum slab specifies the repeating distance of the periodic image of the C-S-H-G()-C-S-H molecular structure. Table 1 shows molecular structure geometry properties of the simulated hybrid GCNCs.
Fig. 9. Typical atomic model for (C-S-H)-G()-C-S-H hybrid nanocomposite interfacial pullout
Table 1. Molecular Computational Cell Properties for GCNCs
CompositeNumber of atomsρ (g/cm3)Volume (Å3)
(C-S-H)-G()-C-S-H1,4901.524,807.2
(C-S-H)-G(OH)-C-S-H1,5501.5824,893.4
(C-S-H)-G(NH2)-C-S-H1,5701.5625,217.3
(C-S-H)-G(COOH)-C-S-H1,5841.5925,252.4

Interfacial Strength

Generally, the bonding strength between the fiber and the matrix can be evaluated by using the interfacial bonding energy. Typically, this energy can be estimated from the energy difference, ΔE, between the energy of the hybrid composite and the sum of the individual energies of the composite constituents:
ΔE=Etotal−(Efiber+Ematrix)
(1)
where Etotal = total energy of the hybrid composite; Efiber = energy of the reinforcing fiber without the matrix; and Ematrix = energy of the embedding matrix without the fiber. In other words, the interaction energy, estimated by the energy difference described previously, can be calculated by the difference between the minimum energy of the hybrid system and the energy at an infinite separation of the graphene platelet and C-S-H. If the total energy difference between the hybrid composite system and the individual components is negative, this indicates a favorable configuration. However, an increase in the total energy is attributable to an increase in the potential energy of the formation of unfavorable new interfacial surfaces. Furthermore, the total energy difference ΔE is twice the interfacial bonding energy γ scaled by the contact area A (Al-Ostaz and Pal 2008; Gou et al. 2004; Lordi and Yao 2000; Sanchez and Zhang 2008):
γ=ΔE2A
(2)
where the contact area for the graphene platelet is represented by the parameter of the sheet multiplied by the thickness.
According to Gou et al. (2004), the pullout energy Epullout is defined as the energy difference between the fully embedded configuration of the fiber and the complete pullout configuration. Furthermore, the energy difference is divided into three parts, which include the energy change of the fiber, matrix, and their interactions:
Epull out=E2−E1=(Efiber 2−Efiber 1)+(Ematrix 2−Ematrix 1)+(ΔE2−ΔE1)
(3)
where Efiber = energy of the fiber; Ematrix = energy of the matrix; and ΔE = interaction energy between the fiber and the matrix. Epull out can be related to the interfacial shear strength τi
Epull out=∫0L2(W+t)·(L−x)τidx=τi(W+t)L2
(4)
τi=Epull out(W+t)L2
(5)
where W, L, and t = width, length, and thickness of the graphene plate, respectively; and x = displacement of the graphene sheet.
Prior to the energy calculations, the periodic boundary conditions imposed on the structure are removed. Furthermore, if the simulated system is subjected to periodic boundary conditions, there will be no energy difference between the simulated scenarios shown in Fig. 9. When a graphene platelet is pulled from one end, another graphene platelet enters from its neighboring periodic image. As a result, the total energy of the system remains almost the same.
If the total energy difference between the hybrid composite system and individual components is negative, this indicates a favorable configuration. In contrast, the increase in the total energy is attributable to an increase in the potential energy of the formation of unfavorable new interfacial surfaces. However, the energy calculations obtained from the simulations indicated negative values for the total energy difference. The negative energy difference is assumed to be the attractive forces between the graphene platelet and C-S-H. The interaction energy primarily originated from the van der Waals and electrostatic interactions between the graphene and C-S-H. In this case, the magnitude of the energy difference is primarily influenced by the electrostatic energy interactions between the C-S-H and polarity of the functional groups of the graphene platelets. Table 2 shows the calculated interfacial strength for GCNCs.
Table 2. Interfacial Strength for GCNCs
CompositeInterfacial strength (GPa)
(C-S-H)-G()-C-S-H1.2
(C-S-H)-G(OH)-C-S-H13.5
(C-S-H)-G(NH2)-C-S-H6.1
(C-S-H)-G(COOH)-C-S-H11.8

Experimental Multiscale Approach

Nanoimaging Using Atomic Force Microscopy

A Digital Instruments Nanoscope IIIa multimode scanning probe microscope is used. The AFM, which is based on sensing interactions between the atomically sharp tip and surface, has become the instrument of choice for characterization of surface nanomechanical properties. The nanoscope is operated at ambient temperature in tapping mode using TESP7 silicon cantilever tips. The tips possess a 363–432-kHz resonant frequency. The tapping force is varied by controlling the set point for each scan and is varied depending on the specimen conditions. For scans shown in Figs. 10–12, the resolution is kept at 512×512  pixels, the scan rate is kept constant at 0.50 Hz, and the scan size is 5×5  μm.
Fig. 10. AFM images (5×5  μm) of hydrated portland cement type I/II, 0.45  w/c ratio: (a) height image; (b) phase image
Fig. 11. AFM images (5×5  μm) of hydrated portland cement type I/II mixed with pristine graphene nanoplatelets, 0.45  w/c ratio: (a) height image; (b) phase image
Fig. 12. AFM images (5×5  μm) of hydrated portland cement type I/II mixed with functionalized graphene nanoplatelets, 0.45  w/c ratio: (a) height image; (b) phase image
Force deflection curves are also obtained for the different graphene cementitious nanocomposite specimens to study the adhesion stiffness variation and interpret the AFM images. However, because of the statistical analysis required, the writers intend to publish the results in a separate expanded paper.
From Fig. 10, the high-density and low-density C-S-H structures can be identified. The 3D phase image provides an encoding for the stiffness variation of the scanned specimen. The coloring scheme reflects the height variation as between 0 and 1 μm. Dark brown reflects a deep topography and bright pink reflects a high topography.
From Figs. 11 and 12, the AFM images reveal the high-density and low-density C-S-H structures. However, the phase topography differs from Fig. 10. The variation in the low-density and high-density C-S-H distribution and packing could be disturbed because of the presence of the graphene nanoplatelets. For Fig. 11(a), a graphene platelet at the top-right corner is identified. The 3D phase image in Fig. 11(b) confirms the graphene platelet clearly correlated with the high-stiffness phase topography.

Microimaging Using Scanning Electron Microscopy

Scanning electron microscopy can characterize the topography, morphology, composition, and crystallographic information for cementitious nanocomposites. An understanding of the phase composition of hydraulic cementitious nanocomposites should provide a new understanding of early-age hydration characteristics and graphene nanoplatelet-cement interactions, which will help in the creation of a new generation of materials.
Fig. 13 shows several magnification images captured for the hydrated portland cement. Fig. 13(b) shows that calcium hydroxide (C-H) is dominating, which is expected per the clinker hydration process at early stages.
Fig. 13. SEM image of portland cement type I/II
In Fig. 14, for the exfoliated pristine graphene-cement specimens, SEM imaging has distinguished a graphene platelet in the presence of a HD C-S-H, which confirms the writers’ MD interfacial strength simulations, showing a high interfacial strength on the order of 1.00 GPa between the C-S-H and exfoliated pristine graphene platelet. Fig. 15 shows SEM images for the functionalized graphene-cementitious composites. Scattered C-H and LD C-S-H domination are identified. No evidence of crack bridging has been observed.
Fig. 14. SEM image of portland cement type I/II mixed with pristine graphene
Fig. 15. SEM image of portland cement type I/II mixed with functionalized graphene

Phase Composition Using X-Ray Diffraction

Although the complexity, heterogeneity, and physical structure of cement have not yet been fully revealed, identification of the phase composition is a key to understand the relationships between the mechanical and physical properties. A quantitative analysis of the crystalline and amorphous phases of hydrated cementitious nanocomposites is studied using the internal standard X-ray diffraction (XRD) technique.
Given that the amorphous phases do not produce additional visible reflexes, determination of the amorphous content in cement cannot be detected directly by XRD. However, by using a defined quantity of crystalline standard material, it is possible to determine the ratio of the crystalline material in the specimen to the crystalline standard and thus calculate the content of amorphous material in the specimen by the backward calculation principle (Walenta and Fullmann 2004).
Fig. 16 shows phase quantifications for portland hydrated cement and functionalized GCNCs, demonstrating the percentages of the hydrated cement crystalline and amorphous phases. Alumina is added in a weight ratio of approximately 10% to ensure the precision of the quantification. Figs. 16 (a and b) show similar compositions as expected. All of the tested specimens are consistent in their chemical phase composition. However, this consistency ensures that there is no correlation between the chemical composition of the cementitious nanocomposites and improvement in the chemical properties.
Fig. 16. Phases from powder speciments: (a) hydrated portland cement; (b) functionalized graphene-cement

Mesomechanical Properties Using Resonant Ultrasound Spectroscopy

The study presented in this paper is focused on the linear elastic properties obtained by RUS, which are represented by the Young’s elastic modulus and the shear modulus of hydrated cement paste and GCNCs.
A polished specimen, often in a regular parallelepiped, is placed between two acoustic transducers. The first transducer drives the specimen with a sinusoidal stress and the second transducer encounters the specimen response. As soon as the drive frequency couples the natural vibration frequency of the object, the surface displacement amplitudes increase by a quality factor (Q) of the resonance. The measured spectrum is an arranged list of center frequencies, which are extracted from the collected resonance peaks of the Lorentzian line shape (Wu et al. 2010).
The writers’ approach utilizes the fact that the mechanical vibration resonance spectrum depends on the geometry, mass density, and elastic tensor of the sample. This calculation is termed the forward problem. In contrast, the inverse problem performs the calculations required to determine all of the elastic constants from the measured spectrum and the known parameters of the density and specimen dimensions. After performing the forward calculation and estimating the natural frequencies for the tested specimen, an inverse measurement is conducted in which the calculated frequencies are compared with the measured natural frequencies. Parameters are adjusted accordingly through each iteration until a good agreement is obtained between the measured and estimated frequencies. The adjustable parameters include the initial frequency, final frequency, time step, and voltage amplitude (see Fig. 17).
Fig. 17. Forward and inverse calculation setup for RUS testing
For the RUS testing, all of the specimens are cut using a diamond saw in parallelepiped shapes. They are then polished to eliminate the shape distortion effect on the measurements. Table 3 lists the measured dimensions and densities. To eliminate the moisture content effect, all of the specimens are subjected to the same level of humidity. More specifically, 24 h after casting the specimens, the specimens are submerged in a water bath for 7 days at room temperature. The specimens are then oven-dried; the drying is conducted at 60°C for 7 h, followed by 75°C for 12 h. Mass and density measurements are taken prior to and after the drying process, and the measured weight losses are about 50%. Therefore, all of the specimens are subjected to the same curing and drying procedure and are all tested at the age of 10 days. During the test, the specimens are supported gently by transducers at diametrically opposite corners to ensure capturing the different modes. The specimen is excited by one transducer with a driving voltage of 5 V, whereas the other transducer collects the resonance response. An average of five measurements is recorded. A frequency sweep of a range 0.050–0.20 MHz is tested. Figs. 18–20 show frequency spectrums for the tested specimens.
Table 3. RUS Specimen Measurements
MaterialMass (g)x (cm)y (cm)z (cm)ρ (g/cm3)
Hydrated portland cement6.621.271.731.731.63
Cement-Ga1.520.931.020.921.73
Cement-FGb0.980.830.800.861.72
a
Pristine graphene.
b
Functionalized graphene.
Fig. 18. Spectrum for a parallelepiped-shaped hydrated portland cement specimen
Fig. 19. Spectrum for a parallelepiped-shaped hydrated graphene-cement specimen
Fig. 20. Spectrum for a parallelepiped-shaped hydrated functionalized graphene-cement specimen
The mechanical properties of the cementitious nanocomposites can be measured through a nondestructive technique with a high precision, calculated Young’s and shear moduli, as shown in Table 4. Although the percentage of nanoadditives is relatively small (0.5% by weight), there is a relatively large enhancement in the mechanical properties of higher percentages. Young’s and shear elastic constants, obtained from the RUS testing for the hydrated portland cement specimen, are in a good agreement with reported literature values at early ages. This technique is highly recommended, and it has proven the possibility of obtaining the mechanical properties of the hydrated cement products at different time schemes and different environmentally controlled conditions.
Table 4. Measured Mechanical Properties of GCNCs from RUS Testing
MaterialYoung’s modulus (GPa)Shear modulus (GPa)
Portland cement18.56.7
Cement-Ga19.78.1
Cement-FGb22.89.2
a
Pristine graphene.
b
Functionalized graphene.

Conclusions

This paper sets a framework for evaluating the performance of cementitious materials reinforced with functionalized graphene nanoplatelets. Graphene cementitious nanocomposites show promising results in improving the overall properties of cementitious nanocomposites. Molecular dynamics results indicate that functionalizing graphene improves interfacial strength and hence the overall response of cementitious nanocomposites. In addition, MD simulations showed that the interfacial strength for the FG C-S-H nanocomposites is dependent on the electrostatic forces of the functional group.
Phase imaging for the different cementitious specimens demonstrates the effect of the graphene platelet additive on the phase formation and surface roughness. This has been associated with improvement of the elastic properties of the GCNCs.
The outcomes of this paper can be used as a benchmark for further investigation of utilizing GCNCs for future infrastructure applications.

Acknowledgments

The writers thank Dr. A. M. Rajendran for discussion of this paper; James Talbot for performing the semiquantitative XRD analysis of the hydrated cement specimens; Dr. Josh Gladden and his graduate student, Rasheed Adebisi, Department of Physics and Astronomy at the University of Mississippi, for their expedited help performing the RUS to measure the elastic properties of the cementitious nanocomposites; and Dr. Ellen Lackey and Dr. James Vaughan, Department of Mechanical Engineering at the University of Mississippi, for permission to access their SEM instrument and manufacturing laboratories in addition to training and helping in imaging the specimens. The writers acknowledge Dr. John O’Haver and his graduate student Poh Lee Cheah, Department of Chemical Engineering at the University of Mississippi, for unlimited access to the AFM instrument and training; and Dr. Roland Pelleneq, Department of Civil and Environmental Engineering at MIT, for sharing C-S-H molecular structure information. This paper was partially supported by funding received under a subcontract from the Department of Homeland Security-Sponsored Southeast Region Research Initiative (SERRI) at the DOE’s Oak Ridge National Laboratory.

References

Abu Al-Rub, R. K., Tyson, B. M., Yazdanbakhsh, A., and Grasley, Z. (2012). “Mechanical properties of nanocomposite cement incorporating surface—Treated and untreated carbon nanotubes and carbon nanofibers.” J. Nanomech. Micromech., 2(1), 1–6.
Crossref
Google Scholar
Al-Ostaz, A., and Pal, G. (2008). “Molecular dynamics simulation of SWCNT-polymer nanocomposite and its constituents.” J. Mater. Sci., 43(1), 164–173.
Crossref
Google Scholar
Allen, A. J., Thomas, J. J., and Jennings, H. M. (2007). “Composition and density of nanoscale calcium-silicate-hydrate in cement.” Nature Mater., 6(4), 311–316.
Crossref
Google Scholar
Bazant, Z. P. (1972). “Thermodynamics of interacting continua with surfaces and creep analysis of concrete structures.” Nucl. Eng. Des., 20(2), 477–505.
Crossref
Google Scholar
Chen, J. J., Thomas, J. J., Taylor, H. F. W., and Jennings, H. M. (2004). “Solubility and structure of calcium silicate hydrate.” Cement Concr. Res., 34(9), 1499–1519.
Crossref
Google Scholar
Cwirzen, A., Habermehl-Cwirzen, K., and Penttala, V. (2008). “Surface decoration of carbon nanotubes and mechanical properties of cement/carbon nanotube composites.” Adv. Concrete Res., 20(2), 65–73.
Google Scholar
Frenkel, D., and Smit, B. (2002). Understanding molecular simulation from algorithms and applications, Elsevier, San Diego.
Google Scholar
Gou, J., Minaie, B., Wang, B., Liang, Z., and Zhang, C. (2004). “Computational and experimental study of interfacial bonding of single-walled nanotube reinforced composites.” Comput. Mater. Sci., 31(3), 225–236.
Crossref
Google Scholar
Hoover, W. G. (1985). “Canonical dynamics: Equilibrium phase-space distributions.” Phys. Rev. A, 31(3), 1695–1697.
Crossref
Google Scholar
Jennings, H. M. (2000). “A model for the microstructure of calcium silicate hydrate in cement paste.” Cement Concr. Res., 30(1), 101–116.
Crossref
Google Scholar
Larson, B. K., Drzal, L. T., and Sorousian, P. (1990). “Carbon fibre-cement adhesion in carbon fibre reinforced cement composites.” Composites, 21(3), 205–215.
Crossref
Google Scholar
Lee, C., Wei, X., Kysar, J. W., and Hone, J. (2008). “Measurement of the elastic properties and intrinsic strength of monolayer graphene.” Science, 321(5887), 385–388.
Crossref
Google Scholar
Lordi, V., and Yao, N. (2000). “Molecular mechanics of binding in carbon-nanotube–polymer composites.” J. Mater. Res., 15(12), 2770–2779.
Crossref
Google Scholar
Manzano, H., Dolado, J. S., Guerrero, A., and Ayuela, A. (2007). “Mechanical properties of crystalline calcium-silicate-hydrates: Comparison with cementitious C-S-H gels.” Phys. Status Solidi A, 204(6), 1775–1780.
Crossref
Google Scholar
Materials Studio 4.2 [Computer software]. San Diego, Accelrys.
Google Scholar
Mouras, S., Hamm, A., Djurado, D., and Cousseins, J.-C. (1987). “Synthesis of first stage graphite intercalation compounds with fluorides.” Rev. Chim. Miner., 24(5), 572–582.
Google Scholar
National Science and Technology Council. (2011). Material genome initiative for global competitiveness, Washington, DC.
Google Scholar
Novoselov, K. S., et al. (2004). “Electric field effect in atomically thin carbon films.” Science, 306(5696), 666–669.
Crossref
Google Scholar
Novoselov, K. S., et al. (2005). “Two-dimensional atomic crystals.” Proc. Natl. Acad. Sci. USA, 102(30), 10451–10453.
Crossref
Google Scholar
Pellenq, R. J.-M., et al. (2009). “A realistic molecular model of cement hydrates.” Proc. Natl. Acad. Sci. USA, 106(38), 16102–16107.
Crossref
Google Scholar
Pellenq, R. J.-M., and Van Damme, H. (2004). “Why does concrete set?: The nature of cohesion forces in hardened cement based materials.” MRS Bull., 29(5), 319–323.
Crossref
Google Scholar
Richardson, I. G. (1999). “The nature of C-S-H in hardened cements.” Cement Concr. Res., 29(8), 1131–1147.
Crossref
Google Scholar
Richardson, I. G. (2004). “Tobermorite/jennite and tobermorit/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 with blast-furnace slag, metakaolin, or silica fume.” Cement Concr. Res., 34(9), 1499–1519.
Crossref
Google Scholar
Richardson, I. G., and Groves, G. W. (1992). “Models for the composition and structure of calcium silicate hydrate (C-S-H) gel in hardened tricalcium silicate pastes.” Cement Concr. Res., 22(6), 1001–1010.
Crossref
Google Scholar
Rosca, I. D., Watari, F., Uo, M., and Akasaka, T. (2005). “Oxidation of multiwalled carbon nanotubes by nitric acid.” Carbon, 43(15), 3124–3131.
Crossref
Google Scholar
Royal Swedish Academy of Sciences. (2011). “The 2010 Nobel Prize in physics–Press release.” 〈http://www.nobelprize.org/nobel_prizes/physics/laureates/2010/press.html〉 (Nov. 1, 2011).
Google Scholar
Sanchez, F., and Zhang, L. (2008). “Molecular dynamics modeling of the interface between surface functionalized graphite structures and calcium-silicate-hydrate: Interaction energies, structure, and dynamics.” J. Colloid Interface Sci., 323(2), 349–358.
Crossref
Google Scholar
Selvam, R. P., Subramani, V. J., Murray, S., and Hall, K. (2009). “Potential application of nanotechnology on cement based materials.” Project Rep. Mack-Blackwell Rural Transportation Center (MBTC)/U.S. DOT 2095/3004, Univ. of Arkansas, Fayetteville, AR.
Google Scholar
Sun, H. (1998). “COMPASS: An ab initio force-field optimized for condensed-phase applications—Overview with details on alkane and benzene compounds.” J. Phys. Chem. B, 102(38), 7338–7364.
Crossref
Google Scholar
Taylor, H. F. W. (1986). “Proposed structure of calcium silicate hydrate gel.” J. Am. Ceram. Soc., 69(6), 464–467.
Crossref
Google Scholar
Taylor, H. F. W. (1997). Cement chemistry, Thomas Telford Publishing, London.
Crossref
Google Scholar
Van Damme, H., and Gmira, A. (2006). “Cement hydrates.” Handbook of clay science, Elsevier, Boulevard, UK, 1113–11127.
Google Scholar
Vandamme, H., and Ulm, F.-J. (2009). “Nanogranular origin of concrete.” Proc. Natl. Acad. Sci. USA, 106(26), 10552–10557.
Crossref
Google Scholar
Walenta, G., and Füllmann, T. (2004). “Advances in quantitative XRD analysis for clinker, cements, and cementitious additions.” Powder Diffr., 19(1), 40–44.
Crossref
Google Scholar
Wu, W., Al-Ostaz, A., Gladden, J., Cheng, A. H.-D., and Li, G. (2010). “Measurement of mechanical properties of hydrated cement paste using resonant ultrasound spectroscopy.” J. ASTM Int., 7(5), 1–8.
Google Scholar

Information & Authors

Information

Published In

Go to Journal of Nanomechanics and Micromechanics
Journal of Nanomechanics and Micromechanics
Volume 3 • Issue 3 • September 2013
Pages: 67 - 77

Copyright

© 2013 American Society of Civil Engineers.

History

Received: Nov 28, 2011
Accepted: Feb 15, 2013
Published online: Feb 18, 2013
Discussion open until: Jul 18, 2013
Published in print: Sep 1, 2013

Permissions

Request permissions for this article.

Authors

Affiliations

Hunain Alkhateb [email protected]
M.ASCE
Nano Infrastructure Research Group, Dept. of Civil Engineering, Univ. of Mississippi, University, MS 38677. E-mail: [email protected]
View all articles by this author
Ahmed Al-Ostaz [email protected]
M.ASCE
Nano Infrastructure Research Group, Dept. of Civil Engineering, Univ. of Mississippi, University, MS 38677 (corresponding author). E-mail: [email protected]
View all articles by this author
Alexander H.-D. Cheng [email protected]
M.ASCE
Nano Infrastructure Research Group, Dept. of Civil Engineering, Univ. of Mississippi, University, MS 38677. E-mail: [email protected]
View all articles by this author
Xiaobing Li [email protected]
Nano Infrastructure Research Group, Dept. of Civil Engineering, Univ. of Mississippi, University, MS 38677. E-mail: [email protected]
View all articles by this author

Metrics & Citations

Metrics

Citations

Download citation

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by

View Options

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share with email

Email a colleague

Share