Author

Jason Hicks

Date of Award

January 2013

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Chemistry

First Advisor

Jerome Delhommelle

Second Advisor

Mark R. Hoffmann

Abstract

Anthropogenic pollution has greatly increased since the industrial revolution and continues to increase as more of the world becomes dependent upon fossil fuels for important applications like transportation and power production. In a general case, whenever a fossil fuel is consumed, a primary product of a complete combustion reaction is carbon dioxide. In a more specific case, the collection, processing and combustion of coal for power production are one of the primary ways by which trace elements, such as arsenic and selenium, are released into the environment. All of these pollutants are known to have harmful effects, whether on the environment, human health or power production itself. Because of this there has been an increasing interest in studies related to combating these pollutants.

Concerning CO2 emissions, recently there has been a significant amount of work related to CO2 capture. A promising method involves the encapsulation of CO2 into isoreticular metal-organic frameworks (IRMOFs). The effectiveness of IMROFs greatly depends on the choice of both metal and organic parts. Molecular simulations have been used in the past to aid in the design and characterization of new MOFs, in particular by generating an adsorption isotherm. However, these traditional simulation methods have several drawbacks. The method used in this thesis, namely expanded Wang-Landau, not only overcomes these drawbacks but provides access to all the thermodynamic properties relevant to the adsorption process through a solution thermodynamics approach. This is greatly beneficial, since an excellent way to characterize the performance of various MOFs is by comparing their desorption free energy, i.e., the energy it takes to regenerate a saturated MOF to prepare it for the next adsorption cycle. Expanded WL was used in the study of CO2 adsorption into IRMOF-1, 8 and 10 at eight temperatures, spanning both the subcritical and supercritical regimes and the following were obtained: adsorption isotherms, Gibbs free energy, enthalpy, entropy and desorption free energy. It was found that, when the maximum loading was compared to the regeneration costs, IRMOF-10 had the best performance, followed by IRMOF-8 then IRMOF-1.

During the combustion of coal, not only is CO2 produced, but also the trace elements of arsenic and selenium escape into the environment though this process. Both arsenic and selenium are known to have a harmful effects on the environment and human health. Arsenic is also known to poison the catalytic converter used in selective catalytic reduction of NOx. These trace elements have been found on fly ash or in the hot flue gases released into the atmosphere. In flue gases they most often exist as oxides. There have been many experimental and a few theoretical studies on the monomeric oxides, AsOx and SeOx, where x = 1, 2, or 3. However, little is known concerning the corresponding dimeric oxides, As2Ox and Se2Ox, where x = 3 or 5, though these compounds are expected from their similarities to nitrogen and sulfur chemistry, respectively. From an experimental perspective, they are very challenging to study due to the high temperatures, complex environments and low concentrations needed for a direct study of the form and structures of the dimeric oxides. From a theoretical perspective, they can be challenging to study due to their multireference character and the need for both dynamic and non-dynamic electron correlation due to bonds forming and breaking during isomerization. However, high level multireference ab initio methods which account for both dynamic and non-dynamic electron correlation can be used. In the work contained in this thesis, GVVPT2 and CR-CC(2,3) were used to study the relative stabilities of all relevant isomers and transition states of As2Ox and Se2Ox. The structures used where generated through DFT using the B3LYP functional. Not only were plausible stationary points located for all species, it was further confirmed that GVVPT2, though with lower computational cost than CR-CC(2,3), can accurately predict such complex surfaces.

Share

COinS