Date of Award
2015
Degree Type
Dissertation
Degree Name
Doctor of Philosophy in Chemical Engineering
Department
Chemical Engineering
First Advisor
Michael L. Greenfield
Abstract
Asphalt is an amorphous material whose mechanical performance relies on viscoelastic responses to applied strain or stress. High quality pavement can enhance the commute experience and also can reduce the risk of accidents as well as preventing air pollution. Furthermore, it also can reduce the need for expensive reconstruction and maintenance of roads. Having a broad understanding of asphalt chemistry is the key to understand and improve mechanical properties of this material. Although asphalt is a widely seen material, our knowledge about its molecular structure and properties is limited. Molecular simulations of asphalts can be exploited to infer how the actions of individual molecules contribute to the nanoscale mechanical behavior of a model system.
In this work, chemical composition and its effect on the viscoelastic properties of asphalts have been investigated by computing complex modulus from molecular dynamics simulation results for two different model asphalts (Zhang, L., \& Greenfield, M. L. (2008). Energy Fuels, 22(5), 3363–3375 [ZG08] and Li, D. D., \& Greenfield, M. L. (2014). J. Chem. Phys., 140(3), 034507 [LG14]) whose compositions each resemble the Strategic Highway Research Program AAA-1 asphalt in different ways. Results from equilibrium molecular dynamics simulations these have been interpreted by converting the stress relaxation modulus G(t) to the complex modulus and phase angle delta. Complex modulus at different temperatures have been calculated using fast Fourier transform to study the effect of temperature on viscoelastic properties of asphalt models. Because of inherent noise, any comparison was challenging. To remove the noise, signal processing techniques have been exploited. Signal processing techniques enhanced the clarity of the results and enabled removing the noise from the modulus results. The LG14 system contains larger molecules, and its results have shown a very good agreement with the low and high frequency scaling limits of the Maxwell model within the frequency ranges spanned by the molecular dynamics simulations, while results for the ZG08 model asphalt only follow the high frequency scaling limits of the Maxwell model. A Black plot or van Gurp-Palman plot of complex modulus vs. phase angle for the LG14 system suggests some overlap among results at different temperatures for less high frequencies, with an interdependence consistent with the empirical Christensen-Anderson-Marasteanu model. Both model asphalts are thermorheologically complex at very high frequencies, where they show a loss peak that appears to be independent of temperature and density.
While molecular simulation can provide relatively accurate results, it is computationally expensive. To provide a faster approach to determine the relationship between the asphalt chemistry and macroscale properties, an equation of state (EOS) can be used. Most equations of state need critical properties which are not always available for all compounds. Since most molecules within the ZG08 and LG14 model asphalts are hypothetical molecules and no experimental data are available for them, a group contribution method by Nannoolal et al. (Nannoolal, Y., et al. (2004). Fluid Phase Equilib., 226, 45–63; Nannoolal, Y., et al. (2007). Fluid Phase Equilib., 252(1-2), 1–27; Nannoolal, Y., et al. (2008). Fluid Phase Equilib., 269(1-2), 117–133) has been applied for all molecules in both model asphalts to determine their critical properties and acentric factors. These provide the required properties to use in a common cubic EOS, e.g., Peng-Robinson (PR), to predict phase behavior of multicomponent systems. The results for densities and thermal expansion have been compared to experimental results. Since cubic equations of state do not predict liquid density precisely, COSTALD volume translation has been applied to improve density predictions. The results from PR EOS were quite good but the results from molecular dynamics were better. This fact should not devalue the EOS results because an EOS calculation is much faster and takes a fraction of one step of molecular simulation time. The result from PR without COSTALD volume translation for thermal expansion was even better than results from molecular dynamics. By knowing asphalt phase behavior, one can predict its performance in the real world. For example, rutting of asphalt is usually related to the waxy molecules within asphalt mixture. Precipitation of waxes can cause the cracking of asphalt pavement as well as increase asphalt mixture viscosity during compaction of asphalt pavement. To predict this phenomenon, precipitation of squalane was investigated. The chemical potential of pure squalane (as a waxy compound) and squalane in the multicomponent system have been estimated. To calculate the Gibbs free energy required to study phase behavior of asphalt, the critical properties of molecules were calculated using the Nannoolal et al. correlations. The chemical potential of pure squalane somewhat below its melting point was lower than its chemical potential in the multicomponent system, which suggests that squalane will precipitate below its melting point. The temperature that the wax precipitates can help to choose the right asphalt binder for a specific location.
In total, the approaches used here provide ways to model how differences in asphalt can affect thermodynamic and mechanical properties. Signal processing can help remove the noise from data which leads to a better understanding and interpretation of the results. Having equation of state parameters for a system will help to predict phase stability of the asphalt systems at different conditions.
Recommended Citation
Masoori, Mohammad, "RHEOLOGICAL AND THERMODYNAMIC PROPERTIES OF MODEL ASPHALT USING FFT AND GROUP CONTRIBUTION METHODS" (2015). Open Access Dissertations. Paper 345.
https://digitalcommons.uri.edu/oa_diss/345
Terms of Use
All rights reserved under copyright.