Pore size estimation using image segmentation in silica aerogels

Kurzfassung

The structure-property relations in silica aerogels are have recently been shown to be dependent on their pore structure. Exemplarily, classical silica aerogels are considered to be mesoporous structures. To characterise their pore sizes, nitrogen porosimetry is applied by accounting for the Barrett-Joyner-Halenda (BJH) model. To this end, pore sizes in the range of 2 to 50 nm can ideally be characterised. The BJH model is based on the Kelvin equation which assumes a cylindrical pore shape, which is far from reality in the case of aerogels. Thus, there exists a need for an effective measurement technique for measuring the pore-size distribution in silica aerogels. It is notoriously difficult to obtain tomographic scans of such aerogels in the mesoporous range. This drives the necessity to exploit computational models that adhere to the structural correlations as observed in aerogels. This contribution involves implementing a diffusion-limited cluster-cluster aggregation (DLCA) algorithm to model and subsequently use image segmentation approaches to analyse the pore size distribution in silica aerogels. DLCA has been shown to effectively model the silica aerogel's morphological and material properties [1]. In this work, computational micro-structures are generated, which have the same morphological features and the solid fraction as silica aerogels synthesised in the laboratory. To analyse the pore-size distribution and the pore networks in the generated models, watershed image segmentation is implemented. The calculated pore size distribution through pore segmentation are compared with the experimental results from BJH analysis, so as to analyse the differences between the computational model and the experimental results. Thus, this work provides an alternative computationally driven approach for solving the problem of effective measurement of pore sizes based on simulation and image segmentation techniques. References: [1] Abdusalamov R., Scherdel C., Itskov M., Milow B., Reichenauer G., and Rege A., J. Phys. Chem. B, 125, 1944-1950, 2021.