Feng Liuab,
Ning Hu*bcd,
Jianyu Zhang*c,
Satoshi Atobee,
Shayuan Wengc,
Huiming Ningc,
Yaolu Liuc,
Liangke Wuc,
Youxuan Zhaoc,
Fuhao Moab,
Shaoyun Fuc,
Chaohe Xuc,
Alamusif and
Weifeng Yuanf
aThe State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body, Hunan University, Changsha, 410082, China
bCollege of Mechanical and Vehicle Engineering, Hunan University, Changsha, 410082, China. E-mail: ninghu@cqu.edu.cn; huning888@hotmail.com; Fax: +86-23-65102421; Tel: +86-23-65102527
cCollege of Aerospace Engineering, Chongqing University, Chongqing, 400044, China. E-mail: jyzhang@cqu.edu.cn
dThe State Key Laboratory of Mechanical Transmissions, Chongqing University, Chongqing, 400044, China
eDepartment of Aerospace Engineering, Tohoku University, Sendai, 980-8579, Japan
fSchool of Manufacturing Science and Engineering, Southwest University of Science and Technology, Mianyang, 621010, China
First published on 7th July 2016
The interfacial mechanical properties between graphene (GR) and a polymer matrix play a key role in load transfer capability for GR/polymer nanocomposites. Grafting of polymer molecular chains on GR can improve the dispersion of the GR in a polymer matrix and change the interfacial mechanical properties between the GR and the polymer matrix. In this work, we investigated the interfacial mechanical properties between GR functionalized with polymer molecular chains and a polyethylene (PE) matrix using molecular dynamics simulations. The influences of grafting density and chain length on the interfacial mechanical properties were analyzed. The results show that grafting of short PE molecular chains on GR can significantly improve the interfacial shear strength and interfacial Mode-II fracture toughness in functionalized GR/PE nanocomposites.
It is difficult to homogeneously disperse pristine GR in organic polymers, especially for a higher volume fraction of GR due to strong van der Waals (vdW) forces between graphene sheets (GRs).12 In general, surface modification is an efficient way to improve the compatibility between GR and polymer matrix since functional groups on GRs may effectively prevent GR aggregations in polymer matrices. For instance, Ramanathan et al.13 found that functionalized GRs can be dispersed very well and interact intimately with polar polymers. Moreover, some researchers found that the aggregations of nanostructured materials functionalized with polymer chains can be effectively alleviated, e.g., single-walled carbon nanotubes grafted with polystyrene (PS) in the polymer matrix as identified by Chadwick et al.14 With various grafted polymer chains, it was found that various carbon nanofillers can work much better as a reinforcement phase for improving the mechanical properties of various polymer matrices. For instance, multiwall carbon nanotubes (MWCNTs) grafted with poly(caprolactone) significantly improved the Young's modulus, tensile strength, toughness and ductility of poly(vinyl chloride) composites.15 Similar results were also obtained for polypropylene composites by the addition of GR grafted with alkyl chains16 and for PS composites reinforced by GR grafted with PS chains.17 Comparing with graphene oxide, GR grafted with diglycidyl ether of bisphenol-A significantly enhanced the Young's modulus, tensile strength and fracture toughness of epoxy composites.18
Although there have been some pioneer experimental works as described above, there is no systematical theoretical or numerical investigation on the improvement mechanisms of reinforcement carbon nanofillers grafted by polymer chain in polymer matrix. It is well-known that the performance of composites is critically controlled by the interfacial characteristics between the reinforcement phase and the matrix.19 In the present work, we conducted molecular dynamics simulations to investigate the interfacial mechanical properties between GR grafted with polyethylene (PE) chains and PE matrix. The effects of grafting density and chain length on the interfacial mechanical properties were analyzed. It was found that GR grafted with short PE molecular chains can significantly enhance the interfacial shear strength and Mode-II fracture toughness of GR/PE nanocomposites.
All the simulation models were relaxed in isothermal–isobaric (NPT) ensemble at temperature of 300 K and pressure of 1 atm for 500 ps using 1 fs time step. The canonical (NVT) ensemble was followed by using the same temperature and time step for another 500 ps. Additionally, 100 ps NVT ensemble was further run with the same conditions and 5 configurations were obtained every 20 ps for calculating the average interfacial mechanical properties. The cut-off distance for both the vdW and Coulomb interactions was 1.0 nm in all simulations.
Model | Number of atom for PE layer | Thickness of PE layer (Å) | Maximum pull-out force (nN) | Interfacial cohesive stress (MPa) |
---|---|---|---|---|
1 | 5566 | 25 | 10.57 | 657.6 |
2 | 7018 | 30 | 11.70 | 727.9 |
3 | 8712 | 35 | 11.34 | 705.5 |
4 | 11132 | 40 | 11.87 | 738.4 |
Fig. 2 and Table 1 show that the maximum pull-out force is 11.70 nN and the corresponding interfacial cohesive stress is 727.9 MPa, respectively, which can be calculated from the following equation.
(1) |
This interfacial cohesive stress (i.e., 727.9 MPa) between the GR and the polymer matrix is much larger compared with that in carbon nanotube (CNT) based polymer nanocomposites, e.g., 479 MPa,27 indicating that the GR has a better reinforcement effect than CNT in polymer matrix.
For the transversal interfacial mechanical properties, i.e., Mode-II, to the best of our knowledge, most of previous investigations of the interfacial mechanical properties between the rectangle GR and the polymer matrix were based on the pull-out simulations in the transversal direction (x-axis or y-axis shown in Fig. 1).22,23 The applied pull-out direction may somehow have a minor impact on the results. In this work, we investigated the transversal interfacial mechanical properties by using sliding simulations, i.e., the GRs were pulled out from the PE matrix in transversal direction (y-axis). For this purpose, both the z- and the y-axes of simulation box were applied by non-PBCs. The relaxation processes of systems were the same as that in Section 2.2. The transversal pull-out velocity of 0.01 Å fs−1 in y-axis was applied to the carbon atoms of GR while the atoms of PE layer at the top 5 Å were constrained (see Fig. 1). Fig. 3 shows that a typical curve of pull-out force versus displacement, and its tendency is similar to that in our previous study for the GR pulled out from PE matrix in the x-axis direction.22 As shown in Fig. 3, the pull-out force increases sharply first and then drops dramatically. Afterwards, the pull-out force fluctuates at an average value, and then gradually decreases to zero as the GR is completely pulled out from PE matrix.
Using eqn (2) as shown in the following, the maximum interfacial shear stress of the y-axis direction was evaluated as 127.1 MPa, compared with that in x-axis in our previous work (121 MPa),22 the pull-out direction has only a small impact on the interfacial mechanical properties.
(2) |
Based on the obtained results, we can find that the interfacial shear strength is much lower than the interfacial cohesive strength, implying that the Model-II failure may be much easier compared with the Model-I.
The grafting density can be defined as14
(3) |
For the case of Mode-I in opening simulations, the interfacial cohesive stress decreases as the grafting density and the chain length increase (see in Fig. 5). Compared with pristine GR, the weak interfacial cohesive strength between the functionalized GR and the PE matrix indicates that the interatomic interactions between a pristine GR and PE chains are stronger than those among the PE chains.30 Therefore, larger grafting density and longer chain length of graft PE molecular will result in weaker interfacial cohesive strength, due to a larger and thicker PE layer grafted onto the GR plane will be formed, which can effectively separate the GR from the PE matrix. Consequently, the PE matrix chains cannot directly interact with GR atoms, leading to weaker interfacial cohesive strength.
In contrast with the above opening simulations, as shown in Fig. 6, the interfacial shear strength obtained by using sliding simulations can be improved significantly as the grafting density increases, especially for the case of shorter graft PE chains (5 repeat units). In this case, the interfacial shear strength of functionalized GR/PE nanocomposites increases by 83.1% for the small grafting density of 0.0025 and by 330.7% for the large grafting density of 0.01. For the case of long graft chains (chain length of 10 repeat units), the interfacial shear strength increases only slightly when the grafting density is larger than 0.0025, and then it decreases when the grafting density exceeds 0.075. For the longest graft PE chains (15 repeat units), from the grafting density of 0.0025, the interfacial shear strength fluctuates around the average value of ∼440 MPa (see the blue dashed line in Fig. 6). Fig. 7 shows the configurations of graft PE molecules after sufficient relaxation and the green dashed lines represent the horizontal base line. For long graft PE chains, as shown in Fig. 7(b) and (c), most of the graft PE atoms are on the horizontal base line, due to the stiffness of PE molecular chains is low, which makes the PE molecules be easily adsorbed to the GR and parallel to the GR plane. The effective enhancement of interfacial shear strength by the GR grafted with PE chains are attributed to interlocking between the graft PE chains and the chains of PE matrix. Especially for short graft PE chains, due to the high stiffness of the PE molecular chains, the GR plane becomes rougher, which can be seen in Fig. 7(a) that a large portion of the graft PE atoms significantly deviate from the horizontal base line. Fig. 8 shows that the short graft PE molecular chains (see Fig. 8(a)) penetrate more deeply into the PE matrix compared with the long grafted PE molecular chains (see Fig. 8(b) and (c)), which means that short PE molecular chains can be interlocked more effectively with surrounding PE matrix molecules. On the contrary, the long graft PE chains cannot effectively interlock with PE matrix molecular chains, which results in relatively weak interfacial shear strength. Comparing with pristine GR, the functionalized GR is of significantly high interfacial shear strength, implying that the interlocking between the graft PE chains and the chains of PE matrix dominates the interfacial shear strength, instead of the vdW interactions.
Fig. 7 Schematic of GR grafted with 4 PE molecular chains with different chain lengths after sufficiently relaxed (a) 5 repeat units, (b) 10 repeat units, (c) 15 repeat units. |
Fig. 8 Interfacial models after sufficiently relaxed, 4 graft PE molecular chains with different chain lengths (a) 5 repeat units, (b) 10 repeat units, (c) 15 repeat units. |
According to the theory of fracture mechanics, the energy release rate can be obtained as34
(4) |
U = Ue − Us | (5) |
(6) |
Moreover, A is assumed to be equal to the initial contact area between GR and PE layer, i.e., A = Ac.
For the opening simulations, the interfacial fracture toughness of pristine GR based nanocomposites is 0.187 J m−2 (shown by the blue dashed line in Fig. 9), which is larger than the interfacial cohesive energy of 0.107 J m−2 for CNT based nanocomposites computed by cohesive law.27 It suggests that the interfacial fracture toughness of GR/polymer nanocomposites is stronger than that of CNT/polymer nanocomposites. Using the sliding simulations, the interfacial fracture toughness of pristine GR/polymer nanocomposites is 0.180 J m−2, which is almost equal to the opening simulation result (Model-I).
Some previous experimental results show that the Mode-II interfacial fracture energy is within a range of 0.054–0.80 J m−2 for double-walled CNTs pulled out from poly(methyl methacrylate) (PMMA),35 and 0.05–0.25 J m−2 for pristine MWCNT and Epon 828 interface.36 It indicates that the present results are reasonable. One exception is that the interfacial fracture energy obtained by Barber et al.37 ranged from 4.0 J m−2 to 70.0 J m−2 for the interface between CNT and thermoplastic polymer matrix. This result is one order of magnitude higher than other results. One possible reason is that there are covalent bonds between CNTs and the polymer matrix.
For opening simulations, Fig. 9 shows that the Mode-I interfacial fracture toughness GI firstly increases and then decreases with the increase of grafting density. Also, the interfacial fracture toughness increases with the graft chain length. Compared with non-functionalized GR (see the blue dashed line in Fig. 9), GR grafted with long PE chains (chain length of 15 repeat units) is of significantly higher interfacial fracture toughness. Especially at the grafting density of 0.005, the interfacial fracture toughness increases by 42.2%. The reason may be that the atomic interactions between the graft PE chains and the matrix chains are dominated by vdW force instead of interlocking. Therefore, a longer separation or opening displacement is needed for the longer graft chain, i.e., larger system potential energy required.
Fig. 10 shows the results of the sliding simulations. In this figure, the Mode-II interfacial fracture toughness GII of functionalized GR/PE nanocomposites is much larger than that of non-functionalized GR/PE nanocomposites, i.e., 0.180 J m−2. Especially for the short graft PE chains (5 repeat units), the GII is improved by ∼258% at the grafting density of 0.005. The reason may be that many PE matrix chains were pulled out with the GR grafted short PE chains due to rough plane (see Fig. 7(a)) and strong interlocking effect (see Fig. 8(a)), leading to larger work consumed. However, for the case of the chain length of 15, the fracture toughness decreases sharply with the increase of grafting density and then fluctuates as shown in Fig. 10. This implies that the interlocking effect of the long graft chain may remarkably decrease when the grafting density increases. Moreover, comparing with Fig. 9, for functionalized GR based PE nanocomposites, the interfacial fracture toughness of Mode-II is much larger than that of Mode-I, i.e., the interfacial crack propagation of Model-I is much easier than that of Model-II. This is very different with the case of non-functionalized or pristine GR/PE nanocomposites, where the interfacial fracture toughness of Mode-II is almost equal to that of Mode-I.
This journal is © The Royal Society of Chemistry 2016 |