Wetting Simulations of High-Performance Polymer Resins on Carbon Surfaces as a Function of Temperature Using Molecular Dynamics

Micromechanical Dilution of PLA/PETG-Glass/Iron Nanocomposites: A More Efficient Molecular Dynamics Approach

Polylactic acid (PLA) and poly(ethylene terephthalate glycol) (PETG) are popular thermoplastics used in additive manufacturing applications. The mechanical properties of PLA and PETG can be significantly improved by introducing fillers, such as glass and iron nanoparticles (NPs), into the polymer matrix. Molecular dynamics (MD) simulations with the reactive INTERFACE force field were used to predict the mechanical responses of neat PLA/PETG and PLA-glass/iron and PETG-glass/iron nanocomposites with relatively high loadings of glass/iron NPs. We found that the iron and glass NPs significantly increased the elastic moduli of the PLA matrix, while the PETG matrix exhibited modest increases in elastic moduli. This difference in reinforcement ability may be due to the slightly greater attraction between the glass/iron NP and PLA matrix. The NASA Multiscale Analysis Tool was used to predict the mechanical response across a range of volume percent glass/iron filler by using only the neat and highly loaded MD predictions as input. This provides a faster and more efficient approach than creating multiple MD models per volume percent per polymer/filler combination. To validate the micromechanics predictions, experimental samples incorporating hollow glass microspheres (MS) and carbonyl iron particles (CIP) into PLA/PETG were developed and tested for elastic modulus. The CIP produced a larger reinforcement in elastic modulus than the MS, with similar increases in elastic modulus between PLA/CIP and PETG/CIP at 7.77 vol % CIP. The micromechanics-based mechanical predictions compare excellently with the experimental values, validating the integrated micromechanical/MD simulation-based approach.

Scalable High Tensile Modulus Composite Laminates Using Continuous Carbon Nanotube Yarns for Aerospace Applications

An approach is established for fabricating high-strength and high-stiffness composite laminates with continuous carbon nanotube (CNT) yarns for scaled-up mechanical tests and potential aerospace structure applications. Continuous CNT yarns with up to 80% degree of nanotube alignment and a unique self-assembled graphitic CNT packing result in their specific tensile strengths of 1.77 ± 0.07 N/tex and an apparent specific modulus of 92.6 ± 3.2 N/tex. Unidirectional CNT yarn reinforced composite laminates with a CNT concentration of greater than 80 wt % and minimal microscale voids are fabricated using filament winding and aerospace-grade resin matrices. A specific tensile strength of up to 1.71 GPa/(g cm-3) and specific modulus of 256 GPa/(g cm-3) are realized; the specific modulus exceeds current state-of-the-art unidirectional carbon fiber composite laminates. The specific modulus of the laminates is 2.76 times greater than the specific modulus of the constituent CNT yarns, a phenomenon not observed in carbon fiber reinforced composites. The results demonstrate an effective approach for fabricating high-strength CNT yarns into composites for applications that require specific tensile modulus properties that are significantly beyond state-of-the-art carbon fiber composites and potentially open an unexplored performance region in the Ashby chart for composite material applications.

A Molecular Dynamics Study of Monomer Melt Properties of Cyanate Ester Monomer Melt Properties

The objective of this work was to computationally predict the melting temperature and melt properties of thermosetting monomers used in aerospace applications. In this study, we applied an existing voids method by Solca. to examine four cyanate ester monomers with a wide range of melting temperatures. Voids were introduced into some simulations by removal of molecules from lattice positions to lower the free-energy barrier to melting to directly simulate the transition from a stable crystal to amorphous solid and capture the melting temperature. We validated model predictions by comparing melting temperature against previously reported literature values. Additionally, the torsion and orientational order parameters were used to examine the monomers’ freedom of motion to investigate structure–property relationships. Ultimately, the voids method provided reasonable estimates of melting temperature while the torsion and order parameter analysis provided insight into sources of the differing melt properties between the thermosetting monomers. As a whole, the results shed light on how freedom of molecular motions in the monomer melt state may affect melting temperature and can be utilized to inspire the development of thermosetting monomers with optimal monomer melt properties for demanding applications.

Atomistic simulation of volumetric properties of epoxy networks: effect of monomer length

Properties of epoxy thermosets can be varied broadly to suit design requirements by altering the chemistry of the component agents. Atomistically-detailed molecular dynamics simulations are well-suited for molecular insight into the structure-property relationship for a rational tailoring of the chemistry. Since the macroscopic properties of interest for applications emerge hierarchically from molecular-scale chemical interactions, seamless integration of experiment, computation, and theory is of great interest. Recently, a Specific Volume-Cooling Rate analysis protocol was successfully developed to quantitatively compare the volumetric properties of an epoxy network model with experimental results in the literature, in spite of the nine orders of magnitude mismatch in the accessible time-scales. Here, we extend the application of the method for two epoxy networks in the same class of chemistry but whose monomers have a higher number of repeating units compared to the previous one for validating the generality of our approach. We observed that atomistic simulations are able to predict the experimental temperature trend of the specific volume within 0.4% for both these networks. Using the William-Landel-Ferry equation to account for rate differences, we also see good agreement between the computational and experimental values of the glass transition temperature.