A stochastic method for asphaltene structure formulation from experimental data: avoidance of implausible structures

Abstract

This work presents a stochastic procedure designed to formulate a discrete set of molecular structures that, as a whole, adjust properly to experimental asphaltene data. This algorithm incorporates the pentane effect concept and Clar's sextet rule to the formulation process. The set of viable structures was constructed based on probability distribution functions obtained from experimental information and an isomer database containing all plausible configurations for a given number of rings, avoiding high-energy structures. This procedure was applied to a collection of experimental data from the literature. Ten sets, consisting of 5000 structures each, were obtained. Each set was then optimized. For the most accurate representation, four molecules were sufficient to properly reproduce the experimental input. The asphaltene system obtained is consistent with the reported molecular weight, number of aromatic rings and heteroatom content. Molecular dynamic simulations showed that the asphaltene representation adequately reproduced asphaltene aggregation behavior in toluene and n-heptane. In toluene, a single three-molecule aggregate was observed, and the majority of asphaltene molecules remained in a monomeric state. In n-heptane, aggregates containing up to four molecules were observed; both porous and compact aggregates were found. The asphaltene molecular representation obtained, which allows researchers to avoid inappropriate torsions in the molecule, is able to reproduce interplanar distances between aromatic cores of 4 Å or less for the aggregation state, as supported by experimental results.

Catalogue of Plausible Molecular Models for the Molecular Dynamics of Asphaltenes and Resins Obtained from Quantitative Molecular Representation

“…At the start of the century, Mullins et al 38 influence in the way the community depicted asphaltene molecules and many subsequent works have either employed the models directly 41 , 42 , 43 or have been inspired by them, sometimes correcting some of the irregularities in the original models. 44,45 The "island" nature of the Mullins' models influenced, but did not dominate, the propositions and models for years to come. The asphaltene model shown in Figure 11, built based on samples characterized by elemental, molecular weight and NMR ( 1 H and 13 C) analyses to assist in the investigation of the toluene-insoluble and toluene-soluble asphaltene fractions, is an example 46 .…”

Asphaltene Molecular Representation: Impact on Aggregation Evaluation

“…The structure of asphaltene molecules used in this study was constructed by an stochastic algorithm that generates structures from experimental data of asphaltene samples [17]. This algorithm specifically avoids the construction of molecules with the pentane effect [18] and violation of the Clar's sextet rule for the amount and location of disjoint aromatic π-sextets in the core of asphaltene molecules [19].…”