(684e) How the Shape and Surface Energy of Nanopores Affects the Adsorption of Confined Fluids

It is expected that the unique and specific details of the interactions between the molecules of a fluid have a significant impact on their adsorption and diffusion under confinement in nanopores. However, in addition to the fluid–fluid interactions, the detailed characteristics of the adsorbate material play an equally important role on the ultimate behaviour of the confined fluid.

In general, theoretical and semiempirical models used to predict the adsorption isotherms of gases require assumptions to be made about the pore geometry and the topology of the existing pore network. The most common models are based on simplified scenarios, typically either cylindrical or slit-like pore geometries. However, in contrast with these idealized conformations, a detailed analysis of porous materials (e.g., montmorillonite, silica, carbons, clays, etc.) through techniques such as atomic force microscopy, scanning electron microscopy, and transmission electron microscopy (TEM) reveal the widespread presence of steps, angles and wedges. In particular, the presence of angular pores (or wedges) in materials can affect the behaviour of confined fluids as evidenced by changes in filling transitions, the rise of a liquid and bubbles in a tube, unexpected adsorption of liquids, formation of solid phases, and the nucleation kinetics due to orientational order effects. Following this idea, the presence of irregular pores seems to play a role in the behaviour of confined fluids, although we are not aware of any systematic study elucidating the structure—adsorption relationship. In spite of this, most researchers exclusively employ the cylindrical pore (or a slit pore) as an approximation of the more realistic random-shaped pores found in natural amorphous porous materials. There is an implicit presumption that the general shape and overall characteristics of the adsorption isotherm will not be affected by the shape of the pore, as long as the pore surface and volume are retained. The main objective of this work is to question if this is a valid assumption.

Grand Canonical Monte Carlo simulations of model fluids in well-defined pores are employed to study the effect that the pore geometry has on the adsorption isotherms. The chosen pores have regular structures with different cross-sectional shapes but identical pore cross-sectional areas. The proposed pores have triangle, square, pentagon, hexagon, octagon, decagon, and circular cross- sectional area of different sizes (probing different surface/bulk fluid ratios) and surface energies to quantify the effect that these energetical and morphological heterogeneities have on adsorbed fluid behaviour. Three different pore sizes commensurate with the molecular diameters along with three different values of fluid–solid energy interactions are chosen to perform the GCMC simulations at a subcritical temperature.

For the weak wall interaction all isotherms have a Type III shape. The adsorption into the triangular pore shows the highest uptake in the gas region. The isotherms for the square, pentagon, hexagon, octagon, and decagon continue the trend gradually converging to curves that seem indistinguishable from those of the cylindrical pore. On the other hand, at higher wall energies a Type IV isotherm is observed, where the triangular-shaped pore sees a higher uptake across the whole pressure range. As the number of vertices increase, the curves tend to converge to a single result coinciding with the cylindrical pore. This behaviour can be rationalized with help from the energy maps which show that for the polygonal pores it is possible to observe a concentration of higher energy sites at the vertices. As the number of vertices increases, the strength of this concentrated energy in the corners decreases and effectively smears out, with a limiting case in the circular pore section, where the entire wall has a homogeneous energy distribution.

For pore sizes larger than 4 nm the effects of the specific pore shape start to become irrelevant. This, in itself, is a particular interesting result which explains the rationale behind the success of simple models (e.g., slit pores, cylindrical pores) in being able to represent commonly encountered natural and synthetic heterogeneous nanoporous materials, such as catalysts, mem- branes, shale rocks, etc. where pore size distributions are biased toward the micropore regions.