University of Ioannina
Modeling sorption and diffusion of alkanes, alkenes in porous materials using molecular simulations and artificial neural networks
Abstract
dc:descriptionMetal organic frameworks (MOFs) constitute a versatile class of porous materials with high structural tunability, offering significant potential across gas storage, separation, catalysis, and pharmaceutical applications. The rapid growth in the number of available MOF structures renders exhaustive experimental evaluation impractical, motivating the increasing adoption of molecular simulations and data-driven methodologies. In this thesis, adsorption and diffusion phenomena of light alkanes, alkenes, and their mixtures in representative MOFs are investigated through Grand Canonical Monte Carlo (GCMC) and Molecular Dynamics (MD) simulations. Sorption isotherms, selectivities, density profiles, and self-diffusion coefficients are computed for different MOFs over wide temperature and pressure ranges, and the resulting predictions are compared with available experimental measurements. Different MOFs have been studied: the Zeolitic Imidazolate Framework (ZIF-8) and the NIIC-20-Bu novel mesoporous material, where NIIC stands for Nikolaev Institute of Inorganic Chemistry, in which the NIIC-20-Bu’s synthesis took place. Beyond conventional simulation approaches, artificial neural networks (ANNs) models are developed and trained on combined simulation and experimental datasets to accurately predict adsorption and diffusion behavior with minimal computational cost. This integrated simulation, data approach provides a valuable tool to utilize experimental and simulation data for fast and accurate predictions. Specifically, the sorption and diffusion behavior of small hydrocarbons in ZIF-8 was investigated using GCMC and canonical ensemble (NVT) molecular dynamics simulations. These studies focused on methane, ethane, and ethene, both as pure components and in ternary mixtures across various temperatures. Additional simulations were performed for propane, propene, butane, and butene. To enhance predictive capabilities, ANNs were trained on both experimental and simulation data to estimate sorption isotherms and self-diffusion coefficients. For NIIC-20-Bu, adsorption characteristics of methane, ethane, and their equimolar mixture were explored through GCMC simulations and ANN predictive modeling. Adsorption isotherms were obtained at different temperatures offering insights into the of pure and mixed components within NIIC-20-Bu. Finally, a comparison was conducted to evaluate the performance of NIIC-20-Bu in relation to other popular porous materials, which demonstrated its sorptive superiority.
Degree
thesis:*- Grantor dc:publisher
- University of Ioannina
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Grourras, Arsenios
- Γκούρρας, Αρσένιος
Subjects
dc:subject × 8Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/61087