Structure-property relationships from molecular simulation of polymeric matrix composites

Molecular dynamics;Polymeric matrix composites;Covalent adaptable networks;ReaxFF;Carbon fiber

Lohse, Alexander Mark ; Jacob, Karl Chow, Edmond Materials Science and Engineering Kalidindi, Surya Qi, Jerry Yao, Donggang ; Jacob, Karl

University:Georgia Institute of Technology

Department:Materials Science and Engineering

关键词: Molecular dynamics;    Polymeric matrix composites;    Covalent adaptable networks;    ReaxFF;    Carbon fiber;   

Others  :  https://smartech.gatech.edu/bitstream/1853/61664/1/LOHSE-DISSERTATION-2018.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

The use of composite materials has risen significantly over the past several decades as many industries begin to take advantage of the high specific strength, specific modulus, and tailorable mechanical properties that polymeric matrix composites can provide. However, the number of filler types, surface functionalities, and filler orientations in composites reinforced with nano and macroscale fillers leads to a complicated and lengthy trial and error experimental approach. Computational researchers have been trying to find a new approach to narrow down promising combinations of fillers and matrix materials using their known chemical and mechanical attributes. This work is directed toward developing structure-property linkages in a range of composite materials using chemical reaction modeling in molecular dynamics simulation. This research analyzed the (1) interphase atomistic structure of an epoxy/graphene system with different surface functionalities, (2) dissolution (recycling) mechanism and thermomechanical properties of covalent adaptable networks (CAN), and (3) reaction mechanism pathways and structure development for carbonization of polyacrylonitrile (PAN).The first research area characterized epoxy interphase structure when graphene had non-covalent and covalent surface functionalities using 2-point spatial statistics and/or radial distribution functions. Principal component analysis was used to reduce the statistical representations of the interphase structure in epoxy to its most salient features and ordinary least-squares regression related the principal components to temperature changes during cooling through glass transition and stress values during simulated tensile strain. The models were able to predict bulk system temperature and stress each with over 90% prediction accuracy. The second research area developed a classical molecular dynamics simulation methodology for bond exchange reactions and dissolution in covalent adaptable networks interacting with ethylene glycol solvent. Increasing the rate of bond exchange reactions was found to increase both dangling end intradiffusion and ethylene glycol interdiffusion. Bond exchange reactions were found to improve the solvent diffusivity when degree of crosslinking was over the gelation point. Under the gelation point the enhanced chain mobility from bond exchange reactions is negligible compared to the chain mobility as a result of a lightly crosslinked network. Thermomechanical properties of CANs with different crosslink densities were correlated to overall free volume and individual void shape calculations. The final research area used ReaxFF simulations to study detailed reaction mechanisms and structure development resulting from carbonization of stabilized polyacrylonitrile (s-PAN). Higher oxygen content fibers promoted formation of CO and CO2 gases which lowered the final fiber carbon content. Regardless of the initial molecular orientation of s-PAN, the carbonization process shifted the final graphitic structures toward a random orientation relative to the fiber axis. Higher temperature was found to significantly improve the numbers of graphite rings in the final structure. Detailed reaction mechanism analysis showed that H2 and N2 elimination produce carbon radicals that form crosslinks with other s-PAN molecules and lead to ring formation and fiber densification.