Date of Award


Document Type

Open Access Dissertation

Degree Name

Doctor of Philosophy in Mechanical Engineering-Engineering Mechanics (PhD)

Administrative Home Department

Department of Mechanical Engineering-Engineering Mechanics

Advisor 1

Gregory M. Odegard

Committee Member 1

Julia A. King

Committee Member 2

Ravindra Pandey

Committee Member 3

Gowtham S


The thermal property of epoxy as the binder in the Carbon Fiber (CF) composites, especially thermal conductivity is important to achieve the advance technology and to improve the performance of materials. Multiscale modeling including molecular dynamic (MD) modeling and micromechanical modeling is used to study the properties of neat Cycloaliphatic Epoxies (CE) and Graphene nanoplatelet (GNP)/CE with and without covalent functionalization.

The thermal properties (glass-transition temperature, thermal expansion coefficient, and thermal conductivity) and mechanical properties of CE system are investigated by MD modeling using OPLS-All Atom force field. A unique crosslinking technique is developed to achieve the cured CE models which has the complex curing mechanism. The thermal conductivity and elastic modulus of CF/CE models are further calculated by using micromechanical modeling. The results are validated with the experiments which are in good agreement.

GNP/CE nanocomposites models are established by MD with four different levels of GNP dispersion, namely, 1, 2, 3, and 4 layer(s) of graphene. The thermal conductivities of GNP/CE nanocomposites models are determined by Equilibrium MD (EMD) method. The thermal conductivities are randomized by arithmetic average and varied GNP volume fractions using micromechanics. The resultant thermal conductivities increase with the GNP volume fraction and the better dispersion which compared well with experiments. The 1-layered GNP/CE (perfectly dispersed) model gives the highest thermal conductivity.

The covalently functionalized GNP (fGNP)/CE models are created by functionalizing carboxyl groups onto the single-layered GNP surfaces by MD modeling. The similar method for the pristine GNP/CE models is applied to obtain the effective thermal conductivities fGNP/CE composites. The predicted values suggest that the thermal conductivity decreases with increased functionalization on GNP due to the GNP defect. However, the thermal conductivities of fGNP/CE models are higher than the thermal conductivities of 2, 3, and 4-layer GNP/CE models which the experiment found that the functionalization improves the dispersion.

The coefficients of linear thermal expansion (CLTE) of GNP/EPON862 system are studied with the similar work flow which the results show the improvement of CLTE regarding to GNP dispersion. Finally, the GNP aspect ratio is included to improve the micromechanical modeling for thermal conductivity.