MOLECULAR MODELING OF INTERACTIONS BETWEEN CATALYTIC NANOPARTICLES AND POLYMER CARRIERS

Nowadays, the scientific and technological field is at a very high level, which is related, among other things, to the large production of various products, and this unfortunately brings a high degree of environmental threat. For example, highly hazardous gases might be released into the environment accidentally, therefore a solution to these dangerous situations needs to be found. A research is conducted into cerium(IV) oxide, platinum, and palladium nanoparticles (NPs), which have been found to have hazardous gases degrading properties. This work complements this field by studying the interactions between NPs and their potential polymer nanofiber carriers, polyamide 6 (PA) and β-polyvinylidene fluoride (PVDF). In order to determine the suitability of these carriers, molecular modeling (geometry optimization using COMPASS forcefield) was involved. Interaction energies (E int ) between the most occupied (hkl) planes of NPs ((100), (110), (111), and (311)) and different (hkl) planes of PA and PVDF surfaces ((001), (100) and (010), (001), (00-1), respectively) were determined. Although all types of NPs showed attractive interactions with all types of surfaces, preferred orientations are clearly identifiable. Attention was focused on the possible influence of the NPs interlayer distance d hkl on the resulting E int .


INTRODUCTION
Today, the scientific and technical field is at a very high level, which is associated not only with new advanced knowledge of the biological, chemical and physical nature of environment or large number of high quality products, but also with enormous environmental impact [1]. In this regard, attention should be focused, among other things, on the development of systems and methods that eliminate pollutants from the gas and liquid phases. The use of nanoparticles and nanofibers has proven to be effective due to their interactivity based on their specific properties. In the case of nanofibers, primarily small pore sizes are used and the principle of filtration thus consists in the impermeability of undesirable particles. Nanoparticles work on the principle of chemical interaction with pollutants, which either capture or degrade [2]. Nowadays, cerium(IV) oxide, palladium and platinum nanoparticles are frequently studied due to its catalytic and antioxidant properties with the potential to degrade hazardous gases in the air [3,4]. Interactions between nanofiber layers formed of polyamide 6 (PA) or poly(vinylidene fluoride) (PVDF) and incorporated CeO2 (CeO2), palladium (Pd) and platinum (Pt) nanoparticles (NP) are investigated using molecular modeling.
Area of (0 0 1), (1 0 0) PA surface and (0 1 0), (0 0 1), (0 Models were prepared by placing single NP in the center of PA or PVDF surfaces oriented with the base parallel to the polymer surfaces and different rotations to polymer chains of the surfaces. 9 models were prepared for each NP type on each polymer surface type.

Modeling conditions
Geometry optimization of each model was performed in MS/Forcite module. Atoms in models were parameterized and their charges were assigned by COMPASS force field [9].
Smart algorithm as implemented in MS with 5·10 5 steps was used. Convergence thresholds for the maximum energy and maximum force changes were 1·10 −4 kcal/mol and 5·10 −3 kcal/mol/Å, respectively [10]. For the Coulombic and the van der Waals forces, atom based summation method as implemented in MS was used. The cell parameters were not optimized. For each model, interaction energy (Eint) was calculated from potential energies (Ep) using the following equation (all energies are in kcal/mol): where:

Ep1 -Ep of whole model
Ep2 -Ep of PA or PVDF surface without the NP Ep3 -Ep of the NP without the PA or PVDF surface

RESULTS AND DISCUSSION
Interaction energies between the NPs and PA and PVDF surfaces of nanofibers do not depend on the rotation of the NP bases on the polymer surfaces (except of Pd_311 and Pt_311, see below).The main factor determining average interaction energies of models (Eint) was the size of the NP base (S) adjacent to the polymer surfaces. The larger the S, the higher the number of interatomic interactions between the NPs and the surfaces, resulting in lower Eint. Among models with PA and PVDF surfaces, the lowest Eint was obtained for Ce_311 lying on PA_001 (Eint = -13 240 kcal/mol) and PVDF_001 (Eint = -18 924 kcal/mol), respectively. In order to determine interaction energies of models irrespective of the size of the NP base S, Eint of the models were divided by S and interaction energies (Eint/S) were thus related to one square nanometer (kcal/mol/nm 2 ) (Figures 3a and 3b).
Eint/S is related to the distance between atomic layers (dhkl) of the NPs. The larger the dhkl, the lower Eint/S it shows with PA or PVDF surfaces (Figures 3a and 3b). The largest dhkl (Ce_111, O_111, Pd_111 and Pt_111) provides the greatest specificity of each atomic layer charge, which also refers to the longest, and therefore weakest, atomic bonds between the atomic layers thanks to which the NPs are able to make stronger noncovalent intermolecular attractive interactions with the polymer surfaces. Interactions of CeO2 NPs lying on PVDF surfaces are 34 % stronger than on PA surfaces, however Pd and Pt NPs interact more strongly by 60 % with PA surfaces than with PVDF surfaces (Figures 3a and 3b, Tables 1  and 2). Since CeO2 is a binary compound (Figure 1), partial charges and thus polarity are shown on atoms, which provides stronger interactions with polar chains of the surface. Pd and Pt are elements (Figure 1), so there is no polarity. PA chains are characterized by polar (peptide bonds -COHN-) and nonpolar (aliphatic) parts alternation (Figure 2), while the polarity direction of PVDF chains is perpendicular to the chain axis (one side of the chain forms bonds with F atoms, the other one with H atoms; Figure 2). Therefore, stronger polar interactions with PVDF surfaces occur in the case of CeO2, while Pd and Pt NPs interact more strongly with PA surfaces due to the presence of nonpolar aliphatic parts of PA chains (Table1).  In the case of CeO2 on PVDF surfaces, Eint/S trends are different for each surface type (Figure 3a). Models with CeO2 NP on PVDF_010 and PVDF_001 have comparable Eint/S,avg ( Table 2), which is 22 % lower than on PVDF_00-1. This is due to the fact, that the upper layer of PVDF_010 and PVDF_001 contains F atoms providing strong hydrogen bonds with Ce atoms of the NP, because of the highest difference in electronegativities (Ce -F = 2.86 [11]) within the possible interacting atomic pairs. For comparison according to atomic pair denotation "NPpolymer surface": Ce -H = 1.08, O -F = 0.54 and O -H = 1.24. All chains in the upper layer of PVDF_010 interact with the NP, but only a half of the interacting atoms are F (Figure 2), which results in more interacting atoms but less F atoms per nm 2 . In the case of PVDF_001, all atoms in the upper layer are F, but only a half of the chains are available for interactions with NP (Figure 2), which is less interacting atoms but more F atoms per nm 2 . In the case of PVDF surfaces, the model containing Ce_111 lying on PVDF_001 exhibits the lowest Eint/S (-3 073 kcal/mol/nm 2 ) due to strong interactions between positive Ce atoms (of Ce_111) and negative F atoms (of PVDF_001). Models containing CeO2 NP lying on PVDF_00-1 exhibits different Eint/S trends (Figure 3a) due to the H atoms forming upper surface layer (Figure 2). O_200 lying on PVDF_00-1 shows the lowest Eint/S. Regarding to PVDF_00-1 having formed the upper layer from H atoms (Figure 2), previous result corresponds to making stronger interactions with O atoms of the NP instead of Ce atoms according to influence of electronegativity difference (O:H = 1.24 > Ce:H = 1.08). O atoms number of the NP base has another impact on Eint/S besides atomic layer distance dhkl. O_200 has the second highest dhkl (O_111 has the first) and the base contains 13.7 O atoms, while O_111 base contains 7.9 O atoms per nm 2 ( Table 2), which provides more attractive polar interactions and consequently lower Eint/S. Model with the lowest Eint/S (-2 467 kcal/mol/nm 2 ) contains O_200 lying on PVDF_00-1.
Eint/S of models containing Pd and Pt NP lying on the PA and PVDF surfaces is on average 16× higher (i.e. weaker) than in the case of CeO2 NP (Figures 3a and 3b). Pt NPs exhibit 23 % lower Eint/S,avg than Pd NP on all PA and PVDF surfaces, which corresponds to the difference in atomic layer distances dhkl based on the lattice parameters (3.891 Å for Pd, 3.924 Å for Pt). The lower lattice parameter indicates shorter and thus stronger bonds between atoms and consequently lower tendency to interact with other molecules. Eint/S trends of models containing Pd and Pt NPs are dependent on dhkl, as in cases of models with CeO2 NPs. The higher the dhkl, the lower the Eint/S (Figures 3a and 3b). Exceptions are models containing Pd_311 and Pt_311 (denoted as Pd_311_0 and Pt_311_0, respectively) on PA or PVDF surfaces, where the gaps between the lines of Ce atoms (of NP bases; Figure 1a) are parallel to the PA or PVDF polymer chains. This position allows the polymer chains of the surfaces to penetrate the NP bases gaps (Figure 1b), leading to increased attractive interactions between the NPs and PA and PVDF surfaces and consequently lower Eint/S. While models containing Pd_311 and Pt_311 NPs show the highest Eint/S, models with Pd_311_0 and Pt_311_0 show the lowest Eint/S ( Table 1).

CONCLUSION
The modeling focused on CeO2, Pd and Pt NPs on the polymer surfaces of PA and PVDF. The strongest interactions are shown by CeO2 NP with the highest dhkl Ce_111 on PVDF_001, in the case of Pd and Pt NPs it is Pd_311_0 on PA_001 and Pt_311_0 on PA_100, respectively. CeO2 NPs interact more strongly with PVDF surfaces than with PA surfaces due to the polar interactions, while Pd and Pt NPs interact more strongly with aliphatic parts of PA chains in PA_001 and PA_100 surfaces. The lowest average Eint/S,avg of all NPs having bases cleaved along (111) plane corresponded to the lowest atomic layers distance dhkl and the highest number of atoms per nm 2 . Within the types of surfaces, it showed on average the strongest interactions PA_001, PVDF_010 and PVDF_100 (comparable), in the case of NP types, it was Ce_111, Pd_111 and Pd_311_0 (comparable), Pt_111 and Pt_311_0 (comparable). Further research will clarify degrading effects against hazardous gases of NPs incorporated on the carriers.