Hypercrosslinked polymers (HCPs) are porous polymers of interest for gas adsorption and sensors applications, among others. They are prepared by intensive post-crosslinking of solvent-swollen polymer beads, which results in a microporous structure due to the high rigidity inherent of the crosslinked network. HCPs offer unique advantages over other materials, such as metal-organic frameworks, namely light constituents, synthetic diversity, and chemical and thermal stability. Unfortunately, the amorphous structure of HCPs can be challenging to characterize experimentally. Molecular simulations can offer an additional perspective by providing a better understanding of the materials at the molecular level.
In this work, we present a computational algorithm for generating initial structures of HCPs, which account for crosslinking, using Materials Studio 5.0. The Amorphous Cell module is utilized to pack the simulation box with monomers, followed by an energy minimization. Then, several cycles of crosslinking and minimization steps are repeated until a predetermined degree of crosslinking is reached. Within each cycle, the closest reactive pair is identified and the crosslinking bond is formed. To improve the speed of the algorithm, energy minimizations are performed every ten bonds. At the end of the crosslinking procedure, an equilibration process is performed, consisting of a series of NVT and NPT simulations.
This procedure was performed with dichloroxylene (DCX) as the initial monomer to create poly(dichloroxylene) network structures. An unrestricted crosslinking algorithm was performed, allowing any number of substituents to be formed per ring, to obtain final structures with 66, 90, and 122% crosslinking. Additionally, a restricted structure was created to represent what can be considered an “ideal” infinite network. Here, crosslinking was restricted so that only one additional substituent can be formed on each aromatic ring, giving a total of three substituents per ring. In this case, one structure was obtained with the maximum 66% crosslinking.
To validate our computationally generated structures, we compared their properties to available experimental characterization data, such as elemental analysis, surface area, pore size distribution, and pore volume. Connolly and accessible surface areas, for example, were calculated to be around 2200 and 700 m2/g, respectively, which are similar to experimental BET surface areas reported for polyDCX, ranging from 600 to 1400 m2/g . Finally, adsorption isotherms were calculated using grand canonical Monte Carlo (GCMC) simulations for hydrogen at 77 K, 0-1.2 bar, and methane at 298 K, 0-20 bar. Results show reasonable agreement with the available experimental data. Future work includes the validation of the proposed methodology for building the initial structures with other HCPs and polymer networks.
 Wood, C.D.; Tan, B.; Trewin, A.; Niu, H.; Bradshaw, D.; Rosseinsky, M.J.; Khimyak, Y.Z.; Campbell, N.L.; Kirk, R.; Stöckel, E.; Cooper, A.I. Chem. Mater. 2007, 19, 2034-2048.
See more of this Group/Topical: Computational Molecular Science and Engineering Forum