Sylvie Neyertz

Twenty-four molecular models for polyphosphazene networks were created via an in silico polymerization of phloroglucinol Phg (C₆H₆O₃) and hexachlorocyclotriphosphazene HCCP (N₃P₃Cl₆) mixtures at different Phg:HCCP ratios. A series of monomer mixtures at Phg-to-HCCP stoichiometric ratios ranging from 1:1 to 8:1 were created using molecular dynamics (MD) simulations. Alternating phases of reactions followed by relaxation steps led to the progressive formation of percolating polyphosphazene networks. The actual ratios of Phg to HCCP rings incorporated in the network polymers remained close to those in the mixtures for initial ratios up to 2:1. Above 2:1, there was a gradual divergence toward lower values in the networks as the limits to the number of possible bonds for each monomer started to take effect. The details of the structures were found to be very complex in terms of the probability distributions of links per Phg or HCCP ring. The highest degrees of connectivity and ring packing densities were found in the networks formed from the initial mixtures having Phg-to-HCCP ratios of around 2:1. Mechanical tests were carried out in order to ascertain the resistance of the model polyphosphazene networks to compression/decompression. There again, the networks obtained from the 2:1 initial mixture were found to have the highest Young's modulus and to display the most elasticity as they recovered their initial shape once the compression was removed. The influence of trapped excess monomers in the percolating networks was only noticeable at the highest mixture ratios. The most resistant Phg-HCCP networks are thus obtained from Phg-to-HCCP mixture ratios of around 2:1, with or without trapped excess monomers.

Two existing generic force fields have been augmented with partial charges and tuned in order to give intercompatible all-atom empirical potentials that can satisfactorily represent the known crystal phase structures of the organic phloroglucinol (Phg) (C6H6O3) and inorganic hexachlorocyclotriphosphazene (HCCP) (N3P3Cl6) molecules at several temperatures. It has been proposed that HCCP-Phg network polymers could act as efficient H-2 barrier layers in hydrogen storage tanks for cars. However, essential requirements for modeling such networks are adequate representations of both monomers in their pure dense phases using a common form of force field. Tests of their ability to maintain stable crystal structures have been made using classical molecular dynamics (MD) simulations on large 800-molecule supercells. The force fields have been optimized to match the densities calculated from the experimental unit cell dimensions at ambient conditions as well as the intermolecular potential energy, as estimated from experimental enthalpies of sublimation. For Phg, the crystal structure is stabilized by a network of hydrogen bonds and the Coulombic interactions contribute to over 55% of the total intermolecular potential energy. In contrast, the crystal structure of HCCP is intrinsically stabilized by the van der Waals terms. Both optimized force fields reproduce very well the orthorhombic symmetry of their respective crystals under constant-pressure NPT conditions. The model parameters tuned at ambient temperature also give reasonable agreement with crystallographic data at lower temperatures.

Fluorinated polyimides incorporated with triptycene units have gained growing attention over the last decade since they present potentially interesting selectivities and a higher free volume with respect to their triptycene-free counterparts. This work examines the transport of single-gas and mixed-gas N₂ and CH₄ in the triptycene-based 6FDA-BAPT homopolyimide and in a block 15,000 g mol⁻¹/15,000 g mol⁻¹ 6FDA-mPDA/BAPT copolyimide by using molecular dynamics (MD) simulations. The void-space analyses reveal that, while the free volume consists of small-to-medium holes in the 6FDA-BAPT homopolyimide, there are more medium-to-large holes in the 6FDA-mPDA/BAPT copolyimide. The single-gas sorption isotherms for N₂ and CH₄ over the 0–70 bar range at 338.5 K show that both gases are more soluble in the block copolyimide, with a higher affinity for methane. CH₄ favours sites with the most favourable energetic interactions, while N₂ probes more sites in the matrices. The volume swellings remain limited since neither N₂ nor CH₄ plasticise penetrants. The transport of a binary-gas 2:1 CH₄/N₂ mixture is also examined in both polyimides under operating conditions similar to those used in current natural gas processing, i.e., at 65.5 bar and 338.5 K. In the mixed-gas simulations, the solubility selectivities in favour of CH₄ are enhanced similarly in both matrices. Although diffusion is higher in 6FDA-BAPT/6FDA-mPDA, the diffusion selectivities are also close. Both triptycene-based polyimides under study favour, to a similar extent, the transport of methane over that of nitrogen under the conditions studied.

This work compares eight classical H₂ molecular models in the gas phase taken from the existing literature. All models are based on Lennard-Jones (LJ) 12-6 terms for the van der Waals interactions and hence easier to transfer to multiphase molecular simulations than more sophisticated potentials. The H₂ potentials tested include one-site, two-site, three-site, and five-site models, with the sites being either the H atoms, the center-of-mass of the H₂ molecule, or massless sites. For the multisite models, high-frequency H-H stretching modes can lead to poor equipartition of the kinetic energy, and the timestep for molecular dynamics (MD) simulations should be reduced to maintain a stable numerical integration of the equations of motion. As such, only those models with rigid bonds are considered. In the present case, 600 MD simulations of H₂ gas were carried out over a large range of temperatures (-50 to +90 °C) and at densities corresponding to a pressure range of 50 to 2000 bar, which include the operating conditions of on-board storage tanks in hydrogen-fueled vehicles. Most of the models under study were found to reproduce reasonably well the experimental pVT phase diagram as well as the solubility. Discrepancies only became significant at the highest densities tested, and these could be used to rank the different models. All model diffusion coefficients were essentially indistinguishable from experimental results, and as such, kinetically dominated dynamic properties could not be used as a criterion for the choice of model. Among the eight models tested, two of them, i.e., the two-site model of Yang and Zhong and the one-site model derived from Buch performed very well over the range of conditions tested. They represent a good compromise between realism, simplicity, and computational efficiency.

High-performance polymers with polybenzoxazole (PBO) structures, formed via thermal rearrangement (TR) of aromatic polyimide precursors, have been developed for gas separation applications.

This work illustrates the potential of using atomistic molecular dynamics (MD) and grand-canonical Monte Carlo (GCMC) simulations prior to experiments in order to pre-screen candidate membrane structures for gas separation, under harsh conditions of temperature and pressure. It compares at 300 °C and 400 °C the CO₂/CH₄ and CO₂/N₂ sieving properties of a series of hybrid networks based on inorganic silsesquioxanes hyper-cross-linked with small organic PMDA or 6FDA imides. The inorganic precursors are the octa(aminopropyl)silsesquioxane (POSS), which degrades above 300 °C, and the octa(aminophenyl)silsesquioxane (OAPS), which has three possible meta, para or ortho isomers and is expected to resist well above 400 °C. As such, the polyPOSS-imide networks were tested at 300 °C only, while the polyOAPS-imide networks were tested at both 300 °C and 400 °C. The feed gas pressure was set to 60 bar in all the simulations. The morphologies and densities of the pure model networks at 300 °C and 400 °C are strongly dependent on their precursors, with the amount of significant free volume ranging from ~2% to ~20%. Since measurements at high temperatures and pressures are difficult to carry out in a laboratory, six isomer-specific polyOAPS-imides and two polyPOSS-imides were simulated in order to assess their N₂, CH₄ and CO₂ permselectivities under such harsh conditions. The models were first analyzed under single-gas conditions, but to be closer to the real processes, the networks that maintained CO₂/CH₄ and CO₂/N₂ ideal permselectivities above 2 were also tested with binary-gas 90%/10% CH₄/CO₂ and N₂/CO₂ feeds. At very high temperatures, the single-gas solubility coefficients vary in the same order as their critical temperatures, but the differences between the penetrants are attenuated and the plasticizing effect of CO₂ is strongly reduced. The single-gas diffusion coefficients correlate well with the amount of available free volume in the matrices. Some OAPS-based networks exhibit a nanoporous behavior, while the others are less permeable and show higher ideal permselectivities. Four of the networks were further tested under mixed-gas conditions. The solubility coefficient improved for CO₂, while the diffusion selectivity remained similar for the CO₂/CH₄ pair and disappeared for the CO₂/N₂ pair. The real separation factor is, thus, mostly governed by the solubility. Two polyOAPS-imide networks, i.e., the polyorthoOAPS-PMDA and the polymetaOAPS-6FDA, seem to be able to maintain their CO₂/CH₄ and CO₂/N₂ sieving abilities above 2 at 400 °C. These are outstanding performances for polymer-based membranes, and consequently, it is important to be able to produce isomer-specific polyOAPS-imides for use as gas separation membranes under harsh conditions.