Faculty Advisor or Committee Member

Nikolaos K. Kazantzis, Advisor


Robert W Thompson




The primary goal of the research studies conducted was to apply statistical mechanical and computer simulation methods to describe the equilibrium behavior of hazardous dipolar/quadru-polar single-gases and mixtures confined in micro porous adsorbents. Statistical mechanical models capable of handling the energetic heterogeneity by complex electrostatic interactions between adsorbate-adsorbent and adsorbate-adsorbate electrostatic interactions were developed and studied. The heterogeneous pore shape and size of different adsorbents were taken into account by two different approaches described in the following paragraphs. Under certain conditions, the use of Mean Field Perturbation Theories (MFPTs) is more attractive than Monte-Carlo (MC) simulations because of the enhanced physical insights that they offer, as well as very low computational times required. Existing literature shows that the applications of MFPTs for studying adsorption of polar molecules were limited due to the orientation dependency of the intermolecular potentials for electrostatic interactions, that in turn poses the challenging problem of seeking analytical expressions for the various thermodynamic functions involved. Furthermore, other existing approaches of accounting for complex electrostatic interactions through hydrogen bonding have limitations due to the requirement of parameter estimation related to radial distribution functions and the critical orientation values of molecules for hydrogen bonds, which are generally obtained through MC simulations and X-ray scattering techniques. In the first stage of research efforts, an attempt was made to express angle-dependent intermolecular potentials in the form of angle-independent intermolecular potential terms by employing statistical averaging methods. In particular, the permanent dipole-dipole and permanent dipole-induced dipole intermolecular potentials were expressed as angle-averaged intermolecular potentials. Then, angle-averaged intermolecular potentials were used to predict water isotherms in nano-slit pores. Furthermore, the angle-averaged intermolecular potentials were used for a binary mixture of polar molecules (water-methanol) to predict the adsorption behavior in nano-slit pores. However, significant limitations of MFPTs arise when they are used for the study of adsorption in zeolites that exhibit irregular shaped cavities with surface heterogeneities. The latter certainly represent a future meaningful research direction. It should be pointed out, that the mean field approach allows us to predict equilibrium sorption properties in homogeneous adsorbents like graphitic carbon (slit), carbon nano tubes (cylinder) and highly siliceous faujasites (spherical) as they have regular shaped cavities. The applications of such kinds of theory remained limited due to the (generally) unknown distribution of functional sites on adsorbents of interests (mainly activated carbons and zeolites) and their locations in the adsorbent framework. The second stage of research efforts focused on models capable of incorporating surface heterogeneities and addressing complex pore geometries. The models developed relied on Grand Canonical Monte-Carlo (GCMC) simulations. In particular, two types of GCMC simulations were carried out, namely molecular and atomistic MC simulations. Both techniques were applied to simulate sorption isotherms on zeolites and activated carbon to remove mercury chloride (quadrupole), hydrogen cyanide (HCN, dipole) and methyl ethyl ketone (MEK, dipole) from air. The molecular based MC technique utilized molecular properties of the molecules namely dipole, quadrupole moments, molecular polarizability and molecule size (kinetic diameter). The molecule was considered to be a spherical shaped particle. The dispersion interactions were calculated using Vaan der Waals equation and electrostatic interactions were quantified through the multi-pole expansion method. This approach was used to simulate adsorption of HgCl2, HCN and MEK in zeolite NaX and activated carbon with functional sites namely carbonyl, hydroxyl and carboxyls. Simulation results indicated that HgCl2 sorption could be attributed to charge-induced dipole interactions for activated carbon, suggesting that sorbents with more number surface charges can be useful except for the case of carbonyls in which quadrupole moments plays a crucial role in reducing sorbent capacities, in turn implying that relative positions of positively and negatively charged cations are indeed important. However, for zeolite NaX, performance characteristics were primarily attributed to charge-quadrupole interactions and dispersion interactions. Moreover, zeolite-NaX performance characteristics for capturing HCN and MEK were attributed to dipole-Na interactions due to the relatively large dipole moments of the molecules under consideration. In the case of activated carbon, HCN sorption was governed by mainly charge-dipole and charge-induced dipole interactions, and hence, carbons with carboxyls seemed to perform better than hydroxyls and carbonyls. MEK sorption was influenced by dispersion interactions (due to the large polarizability of MEK) and charge-dipole interactions, which makes carbon with carbonyls more efficient rather than carbons with hydroxyls having the same charge densities. However, application of the aforementioned molecular approaches was limited to sorbents with regular shape cavities having some surface heterogeneity such as activated carbons. Finally, in order to account for sorbents with irregular shaped cavities, such as silicalite and mordenite, one needs to use atomistic MC simulations. The atomistic MC technique utilizes appropriate atomic sizes and charges for the molecules under consideration to quantify intermolecular forces among the adsorbate molecules and the atoms of the zeolite framework as well as activated carbon. The dispersion interactions were calculated using the Van-Der Waals equation and electrostatic interactions were quantified through a standard Coulombic equation. The bond distances among atoms were kept fixed but variations in angular movement and dihedral/torsional movements were considered, and appropriate harmonic potentials were used to account for angle bending and torsional effects. The sorption performance was evaluated for mordenite, silicalite and zeolite beta for a Si/Al ratio of 47-197 for both an HCN and MEK system. The results of HCN/MEK sorption suggested that silicalite has greater capacity than that of mordenites .In the case of MEK Zeolite beta with sodium cations, performance was better than that of mordenites and silicalites. Sorption of HCN in silicalite was observed in straight and zigzag channels, and mainly attributable to hydrogen bonding among HCN molecules. The increase in sodium cations however decreases the capacity of silicalite, zeolite beta and mordenite slightly. The sorption of MEK in mordenite was mainly observed in an 12- and 8-member ring channel. It was found that an increase in sodium cations did not increase the sorption capacity of mordenite significantly as most of the cations in mordenite were located in an 8-member ring channel where MEK molecules can not be accommodated properly due to steric effects. However, the sorption of MEK in zeolite beta seemed to be influenced by the presence of sodium cations as most of the cations are at the intersection of two 12 member rings which provide sufficient space to orient MEK molecules at the intersection and maximize electrostatic interactions. The sorption of MEK in silicalite exhibited similar trends as in the case of mordenite, as all cations were at the intersection of straight and zigzag channels . Finally, in the last Section of the Thesis, a comparative assessment was made of all three approaches in terms of their significance in applications and the ease in applying them.


Worcester Polytechnic Institute

Degree Name



Chemical Engineering

Project Type


Date Accepted





Activated carbons, hydrogen cyanide, Methyl ethyl ketone, Adsorption, mercury, Monte-Carlo, solvents, Molecular simulations, Zeolites, water, methanol, nanopores, Adsorption, Air, Pollution, Statistical mechanics, Molecules, Computer simulation