Computational modeling of structured polymers for advanced gas separation

  1. Computational modeling of structured polymers for advanced gas separation

    28197 / Model and software development
    Promotor(en): L. Vanduyfhuys, V. Van Speybroeck / Begeleider(s): E. Van den Broeck

    Background and problem

    Polymers (or polymeric chains) are chain-like macromolecules composed of a well-defined monomeric subunits and are the main building blocks of many ubiquitous materials, which can be both natural (e.g. cellulose, latex) or synthetic (e.g. polyvinyl chloride ‘PVC’, polyethylene terephthalate ‘PET’). Furthermore, these materials have many applications, not only in packaging, but also in drug delivery and gas separation. Due to their intermolecular interactions, polymer chains of different size and range interweave and form amorphous structures whose complexity depends on different parameters, such as the mutual linkage pattern (cross-linking) and the rigidity/softness of their monomeric units.

    Figure 1. Molecular models of soft (e.g. PVA, left) and rigid (e.g. 6FDA-DAM) polymeric structures coming and their respective monomeric units.

    To be able to intelligently design polymers for new applications, it is of great importance to understand their microstructure. Molecular modeling and simulation tools are of particular relevance for this end since they grant a molecular-scale insights. However, one of the main relates to the multiple possibilities and complex networks in which polymers may arrange themselves. Several computational approaches have already been proposed in literature to tackle this challenge, including the stochastic distribution and linkage of monomers[1], and an iterative procedure of successive linkage and energy minimization.[2] Irrespective of the approach used, one always has a need for material properties that can be used to characterize the structure and validate with experiment. An important example of such property is the glass transition temperature, defined by the temperature in which a polymer undergoes a transition from a hard and brittle state, i.e. a so-called glass state, to a viscous rubbery state. In the context of thermodynamic phase transitions, the glass transition is considered to be of a second order, meaning that it is characterized by a discontinuity of second-order derivatives of the thermodynamic free energy, such as heat capacity, bulk modulus and thermal expansion coefficient and is a key engineering feature determining the application range of various polymeric materials. Other polymeric material properties of interest are the void fraction and pore size distribution, which characterize the porosity of the polymer and hence its capacity for adsorbing gas molecules. This is of particular importance for gas separation applications, where a mixture of gases of different sizes permeate the polymer phase and is separated according to the inherent polymer porosity.


    In this thesis, the student will apply various molecular modeling techniques to obtain representative polymeric structures and compute their corresponding properties such as density, pore size distribution and glass transition temperature. Two polymers will be studied: PVA, a widely-used synthetic polymer with a simple monomer structure[3], as well as PIM-1, a more complex polymer that is of great interest in the context of separation membranes.[4] More specifically, the student will apply a procedure consisting of successive linkage (i.e. adding additional monomeric units to the polymeric chains) and molecular dynamics (i.e. solving Newtons equations of motion to track the dynamic evolution of the system) cycles resulting in an ensemble of structures that are representative of the polydisperse nature of experimental polymers.[2] These structures will serve as input for a series of additional molecular dynamics simulations at varying temperatures and pressures in order to achieve a thermodynamic equilibration of the polymer. From these simulations, properties such as the density, void fraction and pore size distribution will be computed and serve as a first validation. Glass transition temperatures will be estimated using molecular dynamics simulations with gradual cooling rates. These computationally-obtained values will be compared with the corresponding experimental measurements.

    Figure 2. In silico polymerization and validation with respect to experimental values.


    The topics addressed in this thesis are under the scope of the MOONRISE moonshot project (, an initiative regrouping both academic and industrial researching institutions devoted to the development of highly performing mixed matrix membranes for targeted gas separations.

  1. Study programme
    Master of Science in Engineering Physics [EMPHYS]
    Molecular dynamics, Statistical physics, polymer structure, density, pore size distribution, glass transition temperature

    [1] J. Genzer, Macromolecules 2006, 39, 7157–7169

    [2] L.J. Abbott, K.E. Hart, C.M. Colina, Theor. Chem. Acc. 2013, 132, 1–19

    [3] M.I. Baker et al., J. Biomed. Mater. Res. Part B Appl. Biomater. 2012, 100B, 1451–1457

    [4] M.D. Guiver et al., Macromolecules 2020, 53, 8951–8959


Louis Vanduyfhuys
Veronique Van Speybroeck