Surface diffusion and effective pore diffusion in reversed-phase liquid chromatography studied by molecular dynamics simulations

点击翻译按钮获取中文摘要

The project goal is to establish a consistent, molecular-level picture of surface diffusion and effective pore diffusion in reversed-phase liquid chromatography (RPLC), the most popular liquid chromatographic separation technique. Diffusion in and near the stationary phase is spatiotemporally resolved using molecular dynamics (MD) simulations in a model RPLC system. Pore-level diffusion is subsequently traced up to the macroscopic fixed-bed level using a hierarchical diffusion model, to allow for critical comparisons with chromatographic data. Small aromatic hydrocarbons like ethylbenzene and benzyl alcohol are used as representative analytes, which can be distinguished regarding their polarity and thus concentration, molecular orientation, and mobility in and near the RPLC stationary phase. They also partition to a varying degree into the hydrophobic alkyl chains of the chromatographic interface. This partitioning into the alkyl chains (i.e., into the bonded stationary phase), together with analyte adsorption onto the chains, forms the molecular basis for effective pore diffusion, analyte retention, and selectivity, as well as for nonlinear effects due to concentration overloading. These aspects are elaborated, while analyte (size, polarity) and mobile phase properties (composition of the water-acetonitrile mixture) are adjusted. Additionally investigated aspects are the polarity of the stationary phase (C18 vs. C8 chains as surface modification) and the pore geometry (planar vs. cylindrical). The MD simulations provide key data about local concentrations, residence times, diffusive mobilities, as well as data on a differential partitioning and adsorption on the pore level. They can rationalize the macroscopic, thus experimentally accessible transport dynamics and separation effects like retention and selectivity as well as the shape of adsorption isotherms, or even allow to predict these characteristics of (non)linear RPLC. The MD simulation-based effective pore diffusion coefficient (single-mesopore level) is combined with available physical reconstructions of mesopore and macropore spaces from hierarchically structured chromatographic beds. It allows us to analyze (through mass transport simulations based on a random-walk approach) effective diffusion in the interconnected mesopore space and, subsequently, effective diffusion at the macroscopic level of a macroporous-mesoporous chromatographic bed (e.g., a packing of mesoporous particles or a monolith). This hierarchical diffusion approach guarantees that relevant molecular details (stationary phase and analyte properties) and resulting phenomena (e.g., diffusion in and near the stationary phase; adsorption and partitioning under nonlinear conditions) are properly accounted for on the pore level, in the MD simulations, and can be morphologically traced up to the macroscopic field level, where the resulting data can challenge chromatographic experiments (and vice versa).