A Molecular Dynamics Study on Polycaprolactone -Metal Oxide Interactions

2020 In order to realize the macroscopic features of a number of chemically bonded multi-layer dielectric and composite materials, interactions of metal oxide surfaces, polymer surface atoms, and near-surface atoms are very beneficial. The simulation study of polymer-metal oxides interfaces is of great importance in investigating the adhesion and miscibility features of these systems that are inherently challenging to obtain experimentally or for which there is no experimental data, even if some low data exist, they are not reliable. Polycaprolactone (PCL) is biocompatible, biodegradable, non-toxic and hydrophobic polyester that has been used in tissue engineering, such as a bioactive implant. Hence, the molecular dynamics simulations of PCL are carried out to investigate its surface interaction with metal oxides as ZnO, CuO, Fe 2 O 3 , NiO and SiO 2 . The force field of COMPASS is applied to simulations in order to compute interfacial and solubility parameters. Molecular dynamics approach to investigate the interaction and adsorption manner of PCL with metal oxides. Whereas investigations are useful in exploring polymer composites. Much better adhesion is achieved by the calculations between the PCL oligomer and the metal oxides under investigation. The negative values of interaction energy have to be forecasted despite the presence of acid-base or hydrogen-bonding


Introduction
Mixtures of metallic particles and polymers have many uses, e.g., in sensing, catalytic and medical applications 1 , as well as manufacturing methods 2 . The metallic interface of polymers can vary effectively from its bulk manner. Whereas polymers have a surface excellence density, the action of the substances at the nano drastically results in the characteristics of the system at the macro 3 . Computational methods, such as Monte Carlo(MC) 4 , molecular dynamics(MD) 5 , and dissipative particle dynamics(DPD) 6 , have engrossed much attention in this field. MD simulations have been applied by Fernandes et al. 7 and Negreiros et al. 8 to investigate the interaction between water with iron and iron oxide surfaces, respectively. Using MD simulations, some other materials, including iron surfaces in ionic liquids 9 , together with solid-liquid metals 10 have been investiagated.
MD and MC simulations have been applied by Mansfield and Theodorou 11 to investigate the adhesion and interaction of atactic polypropylene with graphite. Matsuda et al. 12 investigated n-alkane melt's MD simulations at temperatures of 300 K and 400 K in attractive generic and neutral, as well as crystalline surfaces. Gooneie et al. 13 offered a modeling method to simulate the interface among a narrow film of polyethylene melt based on a semiinfinite graphite stratum. Kacar et al. 14 presented surface MD models that were large-grained into DPD models. MD simulations have been performed to investigate the interaction of poly (n-butyl acrylate-co-acrylic acid) and (n-butyl acrylate) with α-quartz, α-ferrite, and α-ferric oxide by Anastassiou and Mavrantzas 15 . Liu et al. 16 , using the MD method, showed that the surface of CdS 2 improved in the exposure of mercapto propyl tri methoxy silane. Ta et al. 17 carried out MD simulations to study the adsorption energies and structural properties of alkanes on iron and its oxide surfaces.
Polycaprolactone(PCL) is biocompatible, biodegradable, biocompatible, and also easy to process from aqueous solutions. It may be applied in implants 18 , controlled drug delivery systems 19 , scaffolds in tissue engineering 20 , food packaging 21 , medical devices, and bio packaging 22

Modeling Details
Molecular simulations are carried out by Material Studio (6.0) software 33 . The simulation approach consists of molecular dynamics and mechanics calculations utilizing the Forcite module 34 . The force field of COMPASS (condensed-phase optimized molecular potentials for atomistic simulation studies) is used for MD simulation, being one of the first ab initio force field methods validated and parameterized by the condensed-phase features 35 . In all the simulations, the temperature is equilibrated with the Andersen algorithm 36 . The velocity Verlet algorithm 37 was then applied to integrate the equations of motion. The non-bonded interactions have been calculated using group-based method with vidid atom sums being calculated to 9.5 A˚ . The tail correction was used to non-bonded interactions during the MD run.
The oligomer chain of polycaprolactone is made with 10 monomer units with C 66 H 112 O 22 formula.
Polycaprolactone is optimized, and its energy is minimized. According to the relative densities of the selected oligomers, amorphous cells are generated. The method used in constructing the amorphous cell module of MS modeling was the combined use of an algorithm expanded by Theodorou and Suter 38 and the scanning method of Meirovitch 39 .
Generated amorphous cells are optimized to a convergence level of 0.01 kcal/mol/A° employing the former technique. Applying the Forcite modules, constant volume and temperature ensemble are chosen for MD simulations. The optimized structures with 3D grating are steadied in the NVT ensemble at 298 K. MD run for 100 ps are carried out for eliminating the unsuitable local minima.
This make sures that a relatively thin layer would feel the effective pressure equivalent to that in the bulk. Because the system includes a vacuum space, both the PCL and metal oxids systems are free to expand even though the ensemble is at a constant volume. The unit cell structure of metal oxides comes to exist with the Material studio Software. Whereof they are crystal structures provided experimentally, the lattice parameters are the experimental ones. Symmetry, space group, and lattice parameters of studied metal oxides are given in Table 1. The most stable surface of Hematite (001), SiO2 (100), CuO(101), NiO(111) and ZnO wurtzite structure (001) were used as a surface against which PCL was brought in contact during the simulation step.
As regards the bonds between oxygen atoms and metal being ionic, in essence, the parameters do not form covalent bonds among them; therefore, the bonds have to be put away in optimization and minimization. Using the surface builder module of MS modeling, metal oxide surfaces are prepared by the preferred cleave planes that provide the surface fractional depth to be more than the non-bonded cut-off distance of 9.5 A°.The build vacuum slab crystal is applied for building a crystal from a surface. The crystal is constructed by repeating the surface in a given direction using a repeat distance which is greater than the surface thickness. This introduces a region of vacuum between the surface units. The crystal surface of the metal oxide slab, which is used in the simulation box to examine the adhesion calculation, is designed in MS modeling by the crystal builder facility.
The PCL is assembled in the simulation box with the metal oxide surface, and c-dimension of the box is extended to 30A°. Then, the MD simulation is performed 30,000 steps with a time step of 1 fs at 298 K. Surface atoms all are inhibited in NVT dynamics as, in the earlier step, the metal oxide surface is minimized. Figure 1 illustrates the final structures the PCL simulated structure with metal oxides after MD simulations. The interaction energy or adhesion energy is computed based on Equation (1). Firstly, the energy (E total ) for the simulation box with both PCL and surface atoms is computed, and then without the surface, the PCL energy (E PCL ) is calculated. Eventually, the surface atoms are retained, and PCL is removed to calculate the surface energy (E surface ). Next, the interaction energy or adhesion energy of PCL and the surface is computed as: where V is the molar volume of PCL.

Results and Discussion
The calculation of the interaction energy of metal oxides is vital to realize the metal oxides' physisorption with PCL. The metal oxide surface modeling is the core phase in computing interaction energy 40 . In the present work, the quartz crystalline structure of SiO 2 , ZnO, and Fe 2 O 3 with a hexagonal symmetry, CuO with Monoclinic, as well as NiO with Cubic symmetry, are considered. In the main step, adhesion raises with a decline in the energy of the surface. It should be pointed out that a direct concordance exists between the ideas of 'surface energy' and 'surface stability,' that is to say surfaces with lower surface energy are more stable and vice versa.
C=O and C-O groups are the leading sites of interaction with metal oxide surfaces in PCL. It is assumed that the C-O-C bond's intensity ratio drops compared with the C=O one. The outcome energy and orientation of PCL adsorbed on five diverse metal oxide surfaces are compared. A similar configuration is observed for PCL with other metal oxide surfaces, but not presented to withhold plenty here. In the case of Fe 2 O 3 , the energy needed for distinct segments of PCL from hematite (namely, the adhesive energy value) is calculated as -128.90kcal/mol. For NiO, similarly, the energy is found to be -41.67 kcal/mol. Also, for SiO2, CuO, and ZnO, the computed adhesive energies are -86.55, -111.93, and -10.88 kcal/mol, respectively. It is specified that in the metal oxide of Fe 2 O 3 , the interaction energy is remarkably greater than other metal oxides; however, CuO and silica display excellent adhesiveness that is definite from its adhesion energy value. The NiO interaction energy value is modest than those of other metal oxide surfaces, showing their lesser adherence to PCL. Also, the interaction energy value of ZnO is the lowest. The energy of the surface, polymer, as well as total and interaction energies of metal oxides with the PCL surface model are provided in Table 2.
Also , theoretical simulations of such interactions were calculated through vibrational modes. In this portion, have been computed the vibrational modes of PCL with chosen metal oxides via instantaneous normal mode analysis 41 . The vibrational modes were obtained by employing the CASTEP module. As the force constants are well defined, the vibrational absorptions can be calculated with a capable accuracy and have been examined the interaction between metal oxide and the C=O group of PCL from the intensities of vibrational modes. In PCL, the primary site of interaction with metal oxide surface is C=O and C-O groups 42 . From the spectroscopic data of PCL, vibrational modes of C=O is measured to be in the range of 1715-1730 cm -1 and C-O-C is measured to be in the range of 1163-1210 cm -1 .
The simulated frequencies for PCL/metal interfaces are given in Table 2, trend of the Frequency of C=O bond as compared to C=O bond of PCI. Thus, it is clear that the Frequency of C=O bond becomes less upon interaction with metal oxides; it also follows that Frequency of C=O bond weakens upon interaction with metal oxides increases in an order for metal oxides having less preference for interaction with PCL. Therefore, the calculation of adhesive character of the studied metal oxides with PCL through the intensity ratio of C=O bond by the vibrational modes assign credible approximations with the obtained interaction energies.
Charge density difference can be investigated taking the superposition of non-interacting atoms as reference and allows us to analyze atomic bonds but misses global redistributions of charge. This presedure makes it possible to realize global effects of interaction in surface redistribution due to the presence of the other set. Materials Studio provides the option of generating the electron density difference with respect to a linear combination of the densities of sets of atoms included   Figure 2 and Figure 3 have been illustrated. In this plot a loss of electrons is indicated in blue, while electron enrichment is indicated in red. White indicates regions with very little change in the electron density. The charge density near the vacuum surface is very important to the vacuum potential energy. The plots of total charge density and charge density difference are illustrated as an isosurface in Figure 2 and Figure 3. These plots confirm that the interaction between PCL and metal oxides are driven by electrostatics. Chiefly , the charge density difference plot clearly shows that there is redistribution of electrons between PCL and metal oxides. PCL polymer and metal oxides have acidic or basic sites which can interact to increase adsorption, chargetransfer, wettability and adhesion. These polar interactions are independent of dipole moments and occur only when one material has acid groups which can interact with basic groups of the other material.

Conclusion
This study provides a good insight into polymer-nonpolymer interfacial interactions by molecular modeling procurement. The negative values of interaction energy have to be forecasted despite the presence of strong acid-base or hydrogen-bonding interactions. Enhancement negative values of interaction energies excuse mutual solubility of the polymer pair increase. Although the interfacial study  of PCL and metal oxides considers the designation of PCL adhesion with five metal oxides (Fe 2 O 3 , CuO, NiO, SiO 2 , and ZnO), the calculations display a better adhesion among PCL oligomer and the metal oxides under investigation. The interfacial chemistry assessment among polymers and different metal oxides makes a base able to synthesize and sketch nanoclusters.

Acknowledgements
I gratefully thank Payame Noor University for financial support. The author would like to express their appreciations to department of chemistry this university for providing research facilities.