Multiscale modeling of asphalt systems using quantum mechanics and molecular simulation
Asphalt is widely used as a binder in constructing road pavements. Premature failure in asphalt roads are still a major problem due to a lack of comprehensive understanding in the materials properties and mechanisms. The chemical compositions of asphalts are complex heterogeneous mixtures of hydrocarbons and contain hundred of thousands of different organic molecules. These chemical compositions are known to affect their rheological and mechanical properties which have direct or active role on asphalt road performance. Computational models can be used to help understand the relationship between asphalt chemistry and mechanical properties. Quantum mechanics, molecular simulation and computational methods were used to study the properties and behaviors of model asphalt systems. Molecule choices were based on literature studies with geochemistry as the source for many of the compounds. Based on the Hansen solubility parameters, size and functional groups, the molecules were classified into saturates, naphthene aromatics, polar aromatics and asphaltenes. New compositions were suggested for computational models of AAA-1, AAK-1 and AAM-1 to overcome limitations in density and viscosity of prior model asphalt systems. Molecular structures proposed in the literature as representative of asphaltenes attempt to encompass the results of numerous experiments and quantum mechanics calculations. This work demonstrates that certain features of chemical bonding in proposed structures lead directly to high energies and thus low relative occurrence probabilities. Hartree-Fock, quantum density functional theory and classical force field calculations indicate features such as non-planar aromatic rings that occur in structures proposed recently in the literature. Energy differences for altered pendant group locations were compared using smaller molecules as a reference.Small changes to side chain positions preserve the overall architecture of the proposed asphaltene molecules while decreasing repulsive forces and restoring planar aromatic rings. The rotational relaxation times and diffusion coefficients of molecules in the revised model asphalt AAA-1 system were calculated. The modified Kohlrausch-Williams-Watts equation was used to calculate the average rotational relaxation time. The diffusion coefficients of molecules were calculated based on center of mass displacement. The diffusion followed an Arrhenius dependence as a function of temperature. The largest size molecule asphaltene-pyrrole diffused an order of magnitude slower than the smallest size molecule benzobisbenzothiophene. In addition, the rotational relaxation rate of molecules decreased significantly with increasing size and temperature. In one comparison, the relaxation rate of benzo-bisbenzothiophene is over 50 times faster than asphaltene-phenol. The product of diffusion coefficient and rotational relaxation time increased with decreasing temperature which means that the constant value is consistent with equal contribution to viscosity. The Green-Kubo and Einstein methods were used to calculate the viscosity at high temperatures. The viscosity at lower temperatures was estimated using the Debye-Stokes-Einstein theory. The new model asphalt has significantly higher viscosity compared to the previous asphalt model. The viscosity of new model asphalt agrees much better to the S.H.R.P. experimental results than the older model. The new model asphalt has higher density values compared the previous model but it is slightly lower than the experimental data. These studies and improvements provided better understandings of the relationship between asphalt chemistry and mechanical properties.
Derek D Li,
"Multiscale modeling of asphalt systems using quantum mechanics and molecular simulation"
Dissertations and Master's Theses (Campus Access).