Comparison of Experimental and Predicted Adsorption Isotherms of Mixtures

Comparison of experimental and predicted adsorption isotherms of mixtures

Julian Butz, Sabine Enders

Karlsruhe Institute of Technology, Institute of Technical Thermodynamics and Refrigeration Engineering, Engler-Bunte-Ring 21, D-76131 Karlsruhe, Germany.

Email: julian.butz@kit.edu; sabine.enders@kit.edu

Key words: adsorption isotherms of mixtures, prediction

Abstract:

Multicomponent substances who build close-boiling system are a huge challenge for the separation through distillation. In such systems adsorption often is a lucrative separation alternative. The foundation for a good estimation of the outcome of the adsorption process is the adsorption equilibrium which can be described via adsorption isotherms. Unfortunately, measurements of adsorption isotherm for mixtures are still seldom while the measurement of pure component adsorption isotherms are a common but tedious method. Because of that it is suitable to predict the mixtures adsorption isotherms from only one adsorption isotherm for each pure component. The density functional approach [1] in combination with an convenient equation of state gives an excellent opportunity to do so. The solid-liquid interactions are modelled with a Steele 10-4-3 potential [2] in a slit pore while the liquid-liquid interactions are modelled with the PC-SAFT equation of state [3]. While the parameters of the PC-SAFT EOS for a lot pure components are listed in the literature [3] the parameters of the energy parameters of the Steele potential and the geometry parameters of the slit pore must be extracted from all pure component adsorption isotherms at one temperature. In terms of parameters this means one interaction solid-fluid energy parameter for every component ε_sf,i one interaction distance for the adsorbent σ_ss, the mean pore width H as well as the mean pore volume V_p. For a pure component system that are 4 parameters while a ternary system needs 6 parameters. Having these parameters fitted to the pure components it is now possible to examine broad parameter studies. For the case of a varying feed composition in the adsorption process it is possible to pretend the adsorption isotherms for the possible feed configurations. Furthermore, a prediction of adsorption for temperatures yields a good opportunity to choose a good process window. It was found that a comparison to experimental multicomponent adsorption data [4] at different temperatures yields a very good compliance to the calculated predictions.

[1] C. Lastoskie, K.E. Gubbins, N. Quirke, J. Phys. Chem. 97 (1993) 4786.

[2] W.A. Steele, Surf. Sci. 36 (1973) 317.

[3] J. Gross, G. Sadowski, Ind. Eng. Chem. Res 40 (2001) 1244.

[4] R. Reich, W. T. Ziegler, K.A. Rogers, Ind. Eng. Chem. Process Des. Dev. 19 (1980) 336.