Molecular dynamics study of the interfacial mechanical properties of the graphene–collagen biological nanocomposite

https://doi.org/10.1016/j.commatsci.2012.11.030Get rights and content

Abstract

The reinforcement of biological polymers, such as collagen, with nanoscale fillers for use in tissue engineering and regenerative medicine is a promising research area that is just beginning to be explored. We study the mechanical properties of Type I collagen embedded with graphene nanoribbons (GNRs) as an important template employed in tissue repair. The graphene–collagen interfacial bonding properties play a significant role during load transfer in the nanocomposite and, hence, as the primary goal of this work, a set of pull-out studies, using molecular dynamics (MD) simulation, are performed to characterise the interfacial mechanical properties. The dependence of the pull-out force on the presence of ripples on the surface of the GNRs, arising due to thermal fluctuations, is also examined, as well as the influence of the pull-out velocity.

Highlights

• Mechanical properties of Type I collagen embedded with GNRs is examined using MD. • Pull-out studies are performed to characterize the interfacial properties of these biological nanocomposites. • Contributions of GNR ripples and pull-out velocity are numerically computed. • Presence of ripples causes an increase in the shear force of graphene–collagen nanocomposite. • By increasing the pull-out rate, the interfacial force is enhanced.

Introduction

An important field in medical nanotechnology is the production of efficient scaffolds for use in various areas of tissue engineering and stem cell research. Current research is focused on the development of biocompatible materials that possess suitable mechanical and biological properties. Collagen is an abundant body protein. It is an excellent candidate for applications in such areas as bone tissue engineering and enjoys such physical and biochemical properties that are required in the extracellular matrix of many tissues [1]. It is an important constituent protein of skin, bone, tendon, cartilage, blood vessels and teeth. Collagen molecules are composed of three polypeptide chains, formed as staggered self-assembled collagen fibers, each of length 300 nm and diameter of 1.5 nm [2], a segment of which is shown in Fig. 1.

The outstanding properties of collagen, together with its low toxicity, make this natural polymer a suitable matrix in regenerative medicine. In tissue engineering, collagen scaffold can provide support for damaged tissue, or for mending cavities. The performance of this biological polymer has been, however, restricted by its poor mechanical properties. Recent results have demonstrated that purified collagen does not exhibit the necessary mechanical strength required for in vivo applications [3]. For this reason, many studies have been devoted to enhance its mechanical properties by adding a second phase as a reinforcement material. Recent advances in nanotechnology have made possible the enhancement of the mechanical characteristics of these biological polymers through incorporation of nanofillers, resulting in fiber-reinforced nanocomposite biomaterials.

Previously, several groups have attempted to investigate new hybrid biomaterials, for use in tissue engineering, combining a nanofiller and collagen. Cunniffe et al. [4] developed and characterised a collagen nano-hydroxyapatite composite scaffold for bone tissue engineering. Significant increases in the Young modulus of collagen were achieved in their investigations. In a similar work, it was found that the thickness of hydroxyapatite nanocrystal controls the mechanical properties of the collagen−hydroxyapatite interface [5]. Also, Heinemann et al. [6] investigated experimentally the effect of silica and hydroxyapatite mineralisation on the mechanical properties and the biocompatibility of the nanocomposite collagen scaffold. Their results demonstrated that both silica and hydroxyapatite offer comparable possibilities to tailor mechanical properties of collagen-based scaffolds without being detrimental to biocompatibility.

In recent years, carbon nanostructures have attracted considerable interest in materials science and nanotechnology research programmes because of their superior physical and mechanical properties. Carbon-based nanostructures can be assembled to form three dimensional porous structures with outstanding mechanical properties [7]. Previous studies have shown that carbon nanotubes (CNTs) can be noncovalently functionalised by proteins, carbohydrates and nucleic acids [8], [9]. Nanostructured 3-D collagen/nanotube biocomposites have been demonstrated to work as excellent scaffolds in future bone regeneration [10], [11]. MacDonald et al. [1] showed that single-walled carbon nanotubes (SWCNTs) functionalised with carboxyl groups can be easily incorporated into Type I collagen scaffolds, and that the presence of SWCNTs in the matrix does not significantly affect gel compaction, cell viability, or cell proliferation. It is worth noting that this process could drastically enhance the mechanical properties of collagen [3]. Interaction of collagen triple-helix with SWCNTs has also been studied experimentally by Kuboki et al. [12]. However, the detailed mechanism of the interactions between the two substances has not been described. Therefore, in addition to the experiments, few atomistic simulations have been carried out to explore the noticeable role of nanostructures in improving the mechanical strength of collagen. In this regard, computational techniques such as classical molecular dynamics (MD) simulation have been implemented by Gopalakrishnan et al. [13] to gain insight into the adsorption of collagen-like peptides onto SWCNTs in an aqueous environment. Also, Ghanbari and Naghdabadi [14] have presented a hierarchical multiscale modelling method for the analysis of cortical bone material by considering its nanostructure.

Since single layer graphene sheets (SLGSs) were successfully isolated from bulk graphite by Novoselov et al. [15], research on graphene has been the focus of much attention to explore its unusual electronic and mechanical properties. The advanced properties of graphene together with its high specific surface area and its strong nanofiller-matrix adhesion have made this material an obvious candidate for use in high performance polymer-based composites. Recently, it was demonstrated experimentally and computationally that polymer-based nanocomposites with exfoliated graphene sheets significantly out-performed the CNTs [16], [17], [18], [19]. Thus, the inclusion of graphene in a collagen polymer matrix promises to improve markedly the mechanical properties of this biomaterial. To date, the interactions between collagen molecules and graphene layers have not been well characterised. Cazorla [20] presented a theoretical study to investigate the adsorption of glycine, proline and hydroxyproline amino acids, the major constituents of Type I collagen protein, onto graphene, graphane and Ca-doped graphene, using density functional theory calculations and ab initio molecular dynamics (AIMD) simulations. The results presented in that work provided fundamental insights into the quantum interactions of collagen protein components with graphene-based nanostructures, and could be useful for developments in biomaterial fields. Despite this effort, studies dealing with the interaction between collagen molecules and graphene monolayers have been scarce. To fully understand the collagen–graphene nanocomposites and to address major unresolved issues concerning the mechanical properties of these future scaffolds, this paper aims to study the interfacial mechanical characteristics of these materials via MD simulations.

In the last decade, it has been demonstrated that the enhanced mechanical performance of the polymer nanocomposites not only depends on the inherent properties of the nanofiller, but also, more importantly, depends on the nature of the bonding at the interface and the mechanical load transfer capability from the matrix to the nanofiber. In order to fully harness the outstanding mechanical properties of nanofillers it is, therefore, crucial to understand at the atomistic level the interactions between the polymer matrix and its embedded nano-inclusions. Several techniques for measurement of interfacial strength have been reported so far. Among these the single-fiber pull-out test is the preferred technique for interface characterisation, because it provides a direct estimate of the interfacial properties.

Recently, some groups have simulated numerically the pull-out problem in order to investigate the interfacial properties of CNT-reinforced polymer composites. Frankland and Harik [21] utilised the interfacial friction model for the pull-out of SWCNTs from a polyethylene (PE) matrix. Liao and Li [22] reported the study on the interfacial characteristics of a CNT-polystyrene (PS) composite system employing molecular mechanics (MM) simulations and elasticity calculations. They found that in the absence of chemical bonding between the inclusion and the matrix, the non-bonding interactions consist of electrostatic and van der Waals (vdW) forces, as well as induced stresses arising from the mismatch between the coefficients of thermal expansion of the inclusion and the matrix, which may be more important in affecting the interfacial characteristics of the CNT/PS system. In a similar manner, a series of pull-out simulations have been carried out by Li et al. [23] to investigate the interfacial properties between CNT and polymer matrix, employing only vdW interaction. It was seen that the pull-out force was independent of the nanotube length, but was proportional to its diameter. Using MD simulations, Chowdhury and Okabe [24] simulated the CNT pull-out from the polyethylene matrix to investigate the effects of the density of the matrix, chemical cross-links at the interface, and geometrical defects in the CNT, on the interfacial shear strength (ISS) of the reinforced composite. In addition, Nishikawa et al. [25] investigated the effect of cross-link switching on the pull-out of CNT from amorphous polymer by the same method. MD simulations were also performed by Namilae and Chandra [26] with the view to characterise the interfacial region of CNT-based composites. It is noted that the evaluation of interfacial characteristics for the CNT–polymer systems using experimental methods is also available, but limited in number, due to difficulties in accessing individual interfaces [27], [28].

Although various experimental and modelling studies have been carried out to determine the nanotube-polymer interactions at the interface, studies on the interfacial properties of graphene-based polymer composites are limited. Awasthi et al. [29] analysed the interfacial mechanical behaviour of graphene-PE composites by MD simulations. Separation in sliding mode together with normal loading was investigated in this work. Also, size dependence studies were conducted to obtain the limiting behaviour in the force–displacement responses. To evaluate the effect of chemisorption on the interfacial bonding characteristics of graphene polymer composites, pull-out tests have been modelled by Lv et al. [30] using atomistic MD and MM simulations. These pull-out tests provided the data on variation of shear stress, and interfacial bonding energy, with displacement. It was shown that these parameters increase with the increase in the concentration of functional groups. They concluded that the attachment of some suitable chemical groups with a reasonable concentration to the graphene surface may be an effective way to significantly improve the load transfer properties between the embedded graphene sheets and polymers when graphene is used to produce nanocomposites.

In this paper we have modelled the interaction between the collagen biopolymer and graphene nanoribbons (GNRs) at the atomistic level. For this purpose, graphene pull-out from Type I collagen has been studied via MD simulation to evaluate the key factors responsible for load transfer in the interfacial region. To our knowledge, no numerical modelling of the pull-out of graphene from biological polymers and proteins has been reported so far. The main motivation of this work is to fill this gap. In particular, we focus here on two fundamental issues; (i) the effect of pull-out velocity on the interaction of collagen with graphene at the atomistic level, and (ii) how these interactions are affected by intrinsic ripples present on the surface of GNRs due to thermal fluctuations. In Section 2, the geometrical models and simulation details are briefly presented. Subsequently, results and discussions are given in Section 3. Finally, some concluding remarks are summarised.

Section snippets

Non-bonding inter-atomic potential

In our modelling, we have described the energetics of the non-bonding interactions between all pairs of collagen and graphene atoms via the Lennard–Jones potential:U(rij)=0,rijrcut4εijσijrij12-σijrij6,rij<rcutwhere rij is the separation distance, ɛij is the equilibrium energy, σij is the equilibrium distance between atoms i and j, and rcutt is the cut-off distance. The corresponding potential parameters are listed in Table 1.

Collagen inter-atomic potentials

For the bonding interactions within collagen, the bond-stretch and

Results and discussions

To investigate the interfacial bonding characteristics, we used a series of MD pull-out simulations. In each simulation, the system was composed of a GNR totally embedded inside the collagen fibers. Each initial configuration consisted of two collagen molecular chains surrounding the graphene sheet. Experimental investigations had shown that the collagen density is in the range of 0.23–0.91 g/cm3 in various tissues, with an average value of 0.51 g/cm3 [41]. The density of collagen–graphene

Conclusions

The interfacial bonding characteristics of graphene/collagen nanocomposite is studied via a set of MD-based pull-out simulations. Several possible factors that can affect the bonding energy and shear stress between the collagen component and the embedded GNR, among them the contributions of ripples on the surface of the GNRs, and the pull-out velocity, were numerically computed in this study. In order to ascertain the effect of the presence of a real rough surface on graphene, firstly MD

References (47)

  • G.M. Spinks et al.

    Sens. Actuators, B

    (2006)
  • E. Hirata et al.

    Carbon

    (2011)
  • J. Ghanbari et al.

    J. Biomech.

    (2009)
  • A. Montazeri et al.

    Phys. Lett. A

    (2011)
  • S. Ebrahimi et al.

    Comp. Mater. Sci.

    (2012)
  • C. Cazorla

    Thin Solid Films

    (2010)
  • S.J.V. Frankland et al.

    Surf. Sci.

    (2003)
  • Y. Li et al.

    Comput. Mater. Sci.

    (2011)
  • S.C. Chowdhury et al.

    Composites Part A

    (2007)
  • N. Lachman et al.

    Composites Part A

    (2010)
  • A. Adnan et al.

    Compos. Sci. Technol.

    (2007)
  • S.J.V. Frankland et al.

    Compos. Sci. Technol.

    (2003)
  • K. Iyakutti et al.

    Comput. Mater. Sci.

    (2012)
  • R.A. MacDonald et al.

    J. Biomed. Mater. Res. Part A

    (2005)
  • A. Rich et al.

    J. Mol. Biol.

    (1961)
  • C.M. Homenick et al.

    J. Mater. Chem.

    (2010)
  • G.M. Cunniffe et al.

    J. Mater. Sci.: Mater. Med.

    (2010)
  • Z. Qin et al.

    Langmuir

    (2012)
  • S. Heinemann et al.

    Appl. Mater. Interf.

    (2011)
  • N. Park et al.

    J. Am. Chem. Soc.

    (2007)
  • S. Chatterjee et al.

    Carbon

    (2009)
  • E.E. da Silva et al.

    Nano Res.

    (2009)
  • Y. Kuboki et al.

    Biomed. Mater. Eng.

    (2009)
  • Cited by (47)

    • Modeling and simulation of graphene-based composites

      2022, Innovations in Graphene-Based Polymer Composites
    • Synergistic effects of graphene quantum dots and carbodiimide in promoting resin–dentin bond durability

      2021, Dental Materials
      Citation Excerpt :

      GQDs present a similar layered structure to graphene, but their carbon sheets are heavily oxygenated with hydroxyl and carboxylic functional groups on the basal plane [27]. Carbonyl/carboxylic acids facilitate the combination of oxygen-containing groups with specific functional groups on biomolecules or other polymers [28], and this process can be achieved either through covalent [29] (e.g., nucleophilic substitution) or noncovalent [30] (e.g., Van der Waals forces, electrostatic interaction, hydrogen bonding) methods and be used for modifying materials. Therefore, carboxyl groups in GQDs may be combined with amino groups in collagen fibers.

    • Crack pathway analysis in graphene-like BC<inf>3</inf> nanosheets: Towards a deeper understanding

      2021, Journal of Molecular Graphics and Modelling
      Citation Excerpt :

      As the first two-dimensional carbon-based nanostructure, graphene is synthesized in 2004 via mechanical exfoliation of graphite [1–3]. Graphene is a monolayer hexagonal carbon sheet, which is a gapless semimetal with the Dirac-like electronic structure [4,5]. In parallel with ever-increasing attention towards the use of graphene for diverse applications, such as nanocomposite development [6–8], water treatment [9,10], biomedicine [11,12], and electronics [13,14], the analysis of mechanical properties (fracture, toughness, and strength) of graphene and graphene-like structures has become a matter of scientific challenges [15,16].

    • Design of tissue engineering scaffold by means of mathematical modeling

      2021, Encyclopedia of Materials: Technical Ceramics and Glasses
    • Graphene nanoribbons: A state-of-the-art in health care

      2021, International Journal of Pharmaceutics
      Citation Excerpt :

      Therefore, GNRs are used in electrochemical enantiomeric bio-sensing of several amino acids like tyramine, methionine, etc. (Martín et al., 2015). A study conducted on the interfacial bonding characteristics on GNRs functionalized with collagen using molecular dynamics revealed the collagen-graphene nanocomposites density to be 0.5 g/cm3, that is very close to the density of tissues in the body (0.51 g/cm3) (Ebrahimi et al., 2013). The adsorption of four serum proteins like bovine fibrinogen (BFG), immunoglobulin (Ig), transferrin (Tf) and bovine serum albumin (BSA) was demonstrated on the rGO and GO using molecular dynamics wherein it illustrates the binding affinities in the order of BFG > Ig > Tf > BSA, as shown in Fig. 3.

    View all citing articles on Scopus
    View full text