Local composition activity coefficient models for mixed-gas and liquid adsorption equilibria

Adsorption as a separation technique has gained a lot of momentum in the past century, thanks to the invention of newly improved and cost-effective adsorbents. The demand for adsorptive processes as a promising separation technique is likely to increase in the future. Adsorption has been widely used in various applications in the past several decades. Some of these applications are air separation, gas purification, separation of polyaromatic hydrocarbons (PAH’s) from gaseous effluents, removal of organic contaminants such as aromatic compounds, pesticides and dyes and inorganic contaminants such as heavy metals from polluted groundwater and many others. An efficient design and dynamics of a gas and liquid adsorptive process first requires a thorough and rigorous thermodynamic knowledge of the multicomponent equilibria. The focus of this dissertation is to provide an overall picture of the new advancements in the development of theoretical equilibrium models for representing mixed-gas and liquid adsorption behavior. The benchmark model most widely used in the industry is the Ideal Adsorbed Solution Theory (IAST). Although IAST has a strong theoretical background and is a powerful model for accurately predicting mixed-gas or liquid adsorption equilibria for similar adsorbate species, the model fails to give an accurate prediction of an azeotrope behavior or when adsorbate species are chemically or structurally dissimilar. Therefore, the lack of an accurate multicomponent adsorption equilibrium model has been the motivation to develop the adsorption Non-Random Two-Liquid (aNRTL) model. The aNRTL model has been developed from the fundamentals of the original NRTL model and takes into consideration the critical adsorbate-adsorbent interactions when two or more adsorbate species compete to occupy the same adsorption sites. The model accurately calculates the departure from ideality by predicting negative deviations typically observed in mixed-gas adsorption equilibria. The model has been comprehensively tested on various binary gas mixtures adsorbed on different types of adsorbents and can be easily extended to predict multicomponent adsorption behavior without requiring any additional binary interaction parameters. In addition to this, the model can also accurately predict azeotrope behavior which most of the models have failed to do in the past. While exploring the applicability of the newly developed aNRTL model on liquid adsorption mixtures, it has been observed that some additional crucial interactions are required in order to develop a rigorous thermodynamic framework for representing liquid adsorption behavior. It has been investigated that in liquid mixtures, the factors such as adsorbate size, shape, functional groups and polarizability might lead to critical adsorbate-adsorbate (or lateral) interactions which have been previously ignored while developing aNRTL model for mixed-gas adsorption equilibria. This significant insight has led to the development of extended aNRTL model which takes into account both the adsorbate-adsorbent (“competitive”) interactions and adsorbate-adsorbate (“lateral”) interactions. Interestingly, both positive and negative deviations from ideality has been observed in the liquid adsorption mixtures as opposed to only the negative deviations captured in mixed-gas adsorption equilibria. The extended aNRTL model has been thoroughly tried and tested on various liquid mixtures of aromatic compounds and dyes adsorbed on different types of adsorbents. The model by properly taking into account two types of interactions in a binary liquid mixture can be readily extended to predict multicomponent adsorption behavior. It is demonstrated for the first time, that the newly developed aNRTL and extended aNRTL models for mixed-gas and liquid adsorption equilibria respectively are capable of providing precise knowledge about the thermodynamics involved when various adsorbate species compete for the same adsorption sites. The binary interaction parameters in both the models have a physical significance, thereby, making these models powerful thermodynamic tools as opposed to the semi-empirical models available in the literature. These models can therefore be utilized as strong foundations for the efficient design, modeling and simulation of optimum mixed-gas and liquid adsorption process units.