Multiscale model to investigate the effect of graphene on the fracture characteristics of graphene/polymer nanocomposites
© Parashar and Mertiny; licensee Springer. 2012
Received: 30 August 2012
Accepted: 16 October 2012
Published: 26 October 2012
In this theoretical research work, the fracture characteristics of graphene-modified polymer nanocomposites were studied. A three-dimensional representative volume element-based multiscale model was developed in a finite element environment. Graphene sheets were modeled in an atomistic state, whereas the polymer matrix was modeled as a continuum. Van der Waals interactions between the matrix and graphene sheets were simulated employing truss elements. Fracture characteristics of graphene/polymer nanocomposites were investigated in conjunction with the virtual crack closure technique. The results demonstrate that fracture characteristics in terms of the strain energy release rate were affected for a crack lying in a polymer reinforced with graphene. A shielding effect from the crack driving forces is considered to be the reason for enhanced fracture resistance in graphene-modified polymer nanocomposites.
KeywordsNanocomposite Polymers Fracture Finite element analysis
Nanocomposites composed of nanofiller reinforcement and a polymer matrix are currently subject to intense research due to possible improvements in physical, mechanical, and/or electrical properties compared to neat polymer. Graphene is emerging as a potential candidate for nanoscale reinforcement of polymer nanocomposites. Stankovich et al.  introduced a novel technique for mass-producing graphene at comparatively low cost, which provides the opportunity of using graphene for a variety of conventional purposes and applications, such as improving the characteristics of adhesively bonded joints.
In their experimental works, Rafiee et al. [2, 3] observed increased fracture toughness of graphene-modified epoxy nanocomposites. They reported that the addition of graphene sheets into the epoxy matrix resulted in a distinct increase in fracture toughness, i.e., fracture toughness of epoxy was increased by up to 65% with an inclusion of 0.125% weight fraction of graphene . It was also reported in their experimental work that a uniform distribution of graphene sheets in an epoxy matrix remains a challenging undertaking, which currently limits the full understanding of the mechanisms behind the property improvements. Hence, modeling is considered a viable alternative to explore the effects of nanofiller dispersion on the fracture properties of polymer nanocomposites.
The number of theoretical/numerical works published on graphene-based nanocomposites has so far been limited. Existing works in this field are, for example, the molecular dynamics-based simulation techniques employed by Awasthi et al. , who studied the load transfer mechanisms between polyethylene and a graphene sheet. Cho et al.  employed a Mori-Tanaka approach to study the elastic constants of nanocomposites with randomly distributed graphene sheets. Most recently, Montazeri and Tabar  developed a multiscale finite element model to study the elastic constants of a graphene-based polymer nanocomposite. Parashar and Mertiny  also proposed a multiscale model using finite elements to characterize the buckling phenomenon in graphene/polymer nanocomposites.
To the present authors' knowledge, no theoretical studies are available in the technical literature on the fracture behavior of graphene/polymer nanocomposites. In the present paper, an attempt has been made to develop a multiscale model to investigate the fracture characteristics of graphene-modified epoxy nanocomposites. The proposed multiscale modeling technique was developed in the ANSYS (version 13) finite element software environment (ANSYS Inc., Canonsburg, PA, USA) in conjunction with the virtual crack closure technique (VCCT). A multiscale approach employing the VCCT provides an efficient numerical analysis scheme in terms of the involved degrees of freedom. As a consequence, the analysis can be performed with widely available computational systems.
In the present work, considering the polymer matrix and graphene nanofiller as a continuum and atomistic phase, respectively, a multiscale model in conjunction with finite element analysis was developed. The bond interaction between carbon atoms in graphene was simulated with the help of beam elements. In the current finite element simulation, the modified Morse potential was employed to model the bonded interaction between C-C bonds. The Morse potential has already been applied in a number of research works [8, 9], where structures were subjected to large strain values.
Modified Morse potential variables and parameters
Length of C-C bond
0.1421 × 10−9
Equilibrium C-C bond distance in graphene
(r − ro) / ro
Strain in C-C bond
6.03105 × 10−19
2.625 × 1010
Constant controlling the ‘width’ of the potential
Current angle of the adjacent bond
Initial angle of the adjacent bond
k θ (Nm/rad2)
0.9 × 10−18
Force constant for bond bending
Constant in bending term of potential
The most significant part of the proposed multiscale model is the interphase region between the atomistic graphene model and the continuum polymer representation (see Figure 2). A number of approaches have been considered to account for the interfacial properties and thickness. These depend on the type of bonding, i.e., functionalized or non-functionalized bonding, as well as on the load transfer mechanism through the interface. Hence, the interfacial properties have not unambiguously been defined yet. Hu et al.  in their work on polystyrene and carbon nanotubes (CNT) considered van der Waals interactions to be responsible for maintaining interfacial strength. They assumed 0.2851 to 0.5445 nm as the interface thickness. Meguid et al.  simulated the interaction between the CNT and polymer chains with an interfacial thickness of 0.3816 nm. In the present work, the interfacial thickness was set to 0.172 nm, which is consistent with that in the numerical model proposed by Li and Chou .
For simulating the van der Waals interactions at the graphene/polymer interphase, a truss model (LINK8) was employed as illustrated in Figure 2. The activation of a truss element is determined by the distance between an atom in the graphene sheet and a node in the continuum state model, that is, a truss element is activated when the distance between an atom/node in the graphene and a node in the continuum mesh for the polymer material is less than or equal to 0.65 nm.
Results and discussion
Here, VF represents the graphene volume fraction, and the subscripts Gr and RVE designate the dimensions of the graphene sheet and the representative volume element, respectively, as illustrated in Figure 3. The graphene considered in the multiscale models had 19 cells fixed along the length of the RVE, whereas a different number of cells was employed along the width of the RVE according to the graphene volume fraction. Three cells correspond to a volume fraction of 2.125%, and hence, 6 and 12 cells were used to model volume fractions of 4.25% and 8.5%, respectively.
Significance of graphene volume fraction
Initially, SERR values were estimated for a crack propagating in neat polymer with boundary conditions as shown in Figure 3a. Resulting SERR values were considered as reference values (GIR) for subsequent investigations.
Observed improvements in fracture toughness of graphene nanocomposites were attributed to the change in stress distribution in the (continuum) polymer epoxy phase due to the graphene inclusion in the vicinity of the crack. It is postulated that graphene with its space frame structure and high stiffness bears most of the applied load and shields the crack tip from opening loads or crack-driving forces, whereas higher SERR G1R values in neat epoxy can be attributed to the absence of this shielding effect.
Shielding effects in graphene/polymer nanocomposites
It can be inferred from the SERR values plotted in Figure 8 that a significant improvement in fracture characteristics was obtained for the nanocomposite reinforced with a graphene sheet on both sides of the crack plane. For this configuration, GI values were reduced by up to 24% (for crack propagation near the graphene sheet) compared with the GIR values. Modeling results therefore indicate that the crack tip shielding effect from crack driving forces is more pronounced when graphene is present next to the crack on both sides of the crack plane. It shall be mentioned at this point that an improvement in SERR also occurs when a ‘Single graphene’ model with a graphene volume fraction identical to that of the ‘Twin graphene’ case is considered, which is shown in the subsequent section.
Graphene dispersion and dimensional effects
Earlier experimental works [2, 3] showed that graphene/polymer nanocomposites have enhanced fracture toughness compared to neat polymer. Due to challenges associated with the mixing of graphene in polymer, none of the studies have thus far been devoted to evaluating the effect of graphene dispersion and its aspect ratio on nanocomposite fracture toughness. Experimental results mostly dealt with graphene volume fraction and its impact on fracture toughness. This section of numerical analysis was conducted to study the effect of graphene aspect ratio and its distribution in the polymer matrix on fracture toughness of the developed nanocomposite.
First, modeling was performed for a graphene volume fraction of 8.5%. Corresponding SERR values are plotted with respect to crack length in Figure 9, where legend entries of ‘Single graphene’ and ‘Twin graphene’ represent graphene with an aspect ratio (A.R.) of 1.4 and 2.8, respectively. In Figure 9, the maximum improvement in fracture toughness (crack passing the graphene sheet) compared to the neat polymer was approximately 18% for the ‘Single graphene’ case, whereas the improvement was about 24% for the ‘Twin graphene’ model, which is a difference of 6 percentage points. These results indicate that nanocomposite fracture toughness improves with increasing graphene aspect ratio as well as for nanofillers being uniformly distributed in the matrix.
To corroborate the above findings, modeling was performed next for a graphene volume fraction of 4.25%, and results are shown in Figure 10. Here, legend entries of ‘Single graphene’ and ‘Twin graphene’ correspond to graphene aspect ratios of 2.8 and 5.6, respectively. Figure 10 again shows a superior performance for the ‘Twin graphene’ configuration, yet improvements in SERR were lower in absolute terms than those for the previous case with a higher graphene volume fraction. Notably, for this lower graphene volume fraction as well as the higher graphene aspect ratios, the fracture toughness improvement for the ‘Twin graphene’ configuration compared to the ‘Single graphene’ model was greater, which is now 10 percentage points (compared to 6 percentage points in the previous analysis with a graphene volume fraction of 8.5%). Based on these results, it is postulated that the greatest enhancement of fracture toughness can be expected from a nanographene filler that is uniformly distributed in the polymer matrix and has high-aspect-ratio graphene sheets.
A multiscale modeling technique in conjunction with a representative volume element approach was successfully employed in this investigation to study the fracture characteristics of graphene/polymer nanocomposites. Substantial improvements in the fracture characteristics of graphene/polymer nanocomposites were observed for higher volume fractions of graphene. It was concluded from this research that shielding of the crack tip from driving forces occurred in the vicinity of graphene sheets. In addition to graphene volume fraction, a uniform distribution of graphene in the polymer matrix as well as high aspect ratio of graphene sheets were found to enhance fracture toughness of graphene nanocomposites. Consequently, crack growth retardation can be achieved in graphene nanocomposites even without diverting or increasing the path of crack propagation.
The authors would like to acknowledge the financial support by the Alberta Ingenuity Graduate Scholarship in Nanotechnology.
- Stankovich S, Dmitriy A, Dikin DA, Dommett G, Kohlhaas K, Zimney E, Stach E, Piner R, Nguyen S, Ruoff R: Graphene based composite material. Nature 2006, 442: 282–286. 10.1038/nature04969View ArticleGoogle Scholar
- Rafiee MA, Rafiee J, Srivastave I, Wang Z, Song H, Yu ZZ, Koratkar N: Fracture and fatigue in graphene nanocomposites. Small 2010, 6: 179–183. 10.1002/smll.200901480View ArticleGoogle Scholar
- Rafiee MA, Rafiee J, Wang Z, Song H, Yu ZZ, Koratkar N: Enhanced mechanical properties of nanocomposites at low graphene content. ACS Nano 2003, 3: 3884–3890.View ArticleGoogle Scholar
- Awasthi AP, Lagoudas DC, Hammerand DC: Modeling of graphene polymer interfacial mechanical behavior using molecular dynamics. Modell Simul Mater Sci Eng 2009, 17: 31–37.View ArticleGoogle Scholar
- Cho J, Luo JJ, Daniel IM: Mechanical characterization of graphite/epoxy nanocomposites by multi-scale analysis. Compos Sci Technol 2007, 67: 2399–2407. 10.1016/j.compscitech.2007.01.006View ArticleGoogle Scholar
- Montazeri A, Tabar HR: Multiscale modeling of graphene and nanotube based reinforced polymer nanocomposites. Phys Lett A 2011, 375: 4034–4040. 10.1016/j.physleta.2011.08.073View ArticleGoogle Scholar
- Parashar A, Mertiny P: Representative volume element to estimate buckling behavior of graphene/polymer nanocomposite. Nanoscale Res Lett 2012, 7: 515. 10.1186/1556-276X-7-515View ArticleGoogle Scholar
- Tserpes KI, Papanikos P, Tsirkas SA: A progressive fracture model for carbon nanotubes. Composites Part B 2006, 37: 662–669.View ArticleGoogle Scholar
- Tserpes KI, Papanikos P, Labeas G, Pantelakis SG: Multiscale modeling of tensile behavior of carbon nanotube reinforced composites. Theor Appl Fract Mech 2008, 49: 51–60. 10.1016/j.tafmec.2007.10.004View ArticleGoogle Scholar
- Belytschko T, Xiao S, Schatz G, Ruoff R: Atomistic simulations of nanotube fracture. Phys Rev B: Condens Matter 2002, 65: 235430.View ArticleGoogle Scholar
- Meguid SA, Wernik JM, Cheng ZQ: Atomistic based continuum representation of the effective properties of nano-reinforced epoxies. Int J Solids Struct 2010, 47: 1723–1736. 10.1016/j.ijsolstr.2010.03.009View ArticleGoogle Scholar
- Hu N, Fukunaga H, Lu C, Kameyama M, Yan B: Prediction of elastic properties of carbon nanotube reinforced composites. Proc R Soc Lond A 2005, 461: 1685–1710. 10.1098/rspa.2004.1422View ArticleGoogle Scholar
- Li C, Chou TW: Multiscale modeling of compressive behavior of carbon nanotube/polymer composites. Compos Sci Technol 2006, 66: 2409–2414. 10.1016/j.compscitech.2006.01.013View ArticleGoogle Scholar
- Odegard GM, Gates TS, Wise KE, Park C, Siochi EJ: Constitutive modeling of nanotube reinforced polymer nanocomposites. Compo Sci Technol 2003, 63: 1671–1687. 10.1016/S0266-3538(03)00063-0View ArticleGoogle Scholar
- De Roeck G, Wahab MMA: Strain energy release rate formulae for 3D finite element. Eng Fract Mech 1995, 50: 569–580. 10.1016/0013-7944(94)00232-7View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.