The Formation of Comprehensive Micro-Structural Characteristics of Highly-Fidelity Carbon Fiber and the Effects on its Mechanical Properties Via Molecular Dynamics

Abstract

The preparation of high-performance carbon fiber (CF) is always essential, but challenging to achieve from an only experimental perspective due to the complex formation process and numerous influencing factors. Alternatively, molecular dynamics simulation methods address such gaps by intuitively constructing heterogeneous structure and calculating corresponding properties.The high fidelity models with different orientations along the axis (from 0o to 26o), pores (content, pore orientation, size), amorphous (content, graphitization degree, etc.) and integrated structures with each other are generated, based on the large-scale MD simulation and built-in script in this study. The microstructures are further characterized in terms of the density, hybridization of carbon atom, orientation, crystalline size, interlayer distance, by the method of structural analysis tools like X-ray diffraction and Herman’s Orientation Factor, etc. The properties including strength, modulus, and elongation as a function of multiple structures are also explored and numerically fitted, so that which structure specifically manipulated which performance is accurately identified and recognized different series intervals carbon fiber is obtained. Specially, the predictions of modulus in the range of 450-1200 GPa highly agree with the reported experimental data, the analysis of which indicates that the modulus is mainly controlled by the fiber action area and uniformity as well as sp2C ratio, regardless of whether the structure arises from structural orientation, porosity, or amorphous. The strength is monitored by the combined effect of sp2C and sp/sp3C, most of which originates from amorphous. This data will provide beneficial references and basis for high-performance targeted fiber design.

Keywords: carbon fiber, molecular dynamics, high- fidelity model, numerical fitting, characterization and prediction