Silylated amines, also known as reversible ionic liquids (RevILs), have been designed and structurally modified by our group for potential use as solvents for CO₂ capture from flue gas. An ideal CO₂ capture ionic liquid should be able to selectively and reversibly capture CO₂ and have tolerance for other components in flue gas, including SO₂, NO₂, and O₂. In this project, we study the reactivity, selectivity, uptake capacity, and reversibility of RevILs in the presence of pure SO₂ and mixed gas streams tosimulate flue gas compositions. Tripropylsilylamine (TPSA), a candidate CO₂ capture RevIL, reacts with pure SO₂ to form an ionic liquid consisting of an ammonium group and a salfamate group, supported by IR and NMR results. The resulting IL with pure SO₂ partially reverses when heated to temperatures of upto 500 C in the TGA. TGA analysis of the ionic liquid formed from a 4 vol% SO₂ in CO₂ mixture indicates a possible reversal temperature in the 86-163 C range.

A new group of nanoporous materials, metal organic frameworks (MOFs), have emerged as a fascinating alternative to more traditional nanoporous materials for membrane based gas separations. Although hundreds of different MOF structures have been synthesized in powder forms, very little is currently known about the potential performance of MOFs as membranes since fabrication and testing of membranes from new materials require a large amount of time and resources. The purpose of this thesis is to predict the macroscopic flux of multi-component gas mixtures through MOF-based membranes with information obtained from detailed atomistic simulations. First, atomically detailed simulations of gas adsorption and diffusion in MOFs combined with a continuum description of a membrane are introduced to predict the performance of MOF membranes. These results are compared with the only available experimental data for a MOF membrane. An efficient approximate method based on limited information from molecular simulations to accelerate the modeling of MOF membranes is then introduced. The accuracy and computational efficiency of different modeling approaches are discussed. A robust screening strategy is proposed to screen numerous MOF materials to identify the ones with the high membrane selectivity and to direct experimental efforts to the most promising of many possible MOF materials. This study provides the first predictions of any kind about the potential of MOFs as membranes and demonstrates that using molecular modeling for this purpose can be a useful means of identifying the phenomena that control the performance of MOFs as membranes.

Compared to the conventional amine adsorption process to separate CO₂ from natural gas, the membrane separation technology has exhibited advantages in easy operation and lower capital cost. However, the high CO₂ partial pressure in natural gas can plasticize the membranes, which can lead to the loss of CH₄ and low CO₂/CH₄ separation efficiency. Crosslinking of polymer membranes have been proven effective to increase the CO₂ induced plasticization resistance by controlling the degree of swelling and segmental chain mobility in the polymer. This thesis focuses on extending the success of crosslinking to more productive asymmetric hollow fibers. In this work, the productivity of asymmetric hollow fibers was optimized by reducing the effective selective skin layer thickness. Thermal crosslinking and catalyst assisted crosslinking were performed on the defect-free thin skin hollow fibers to stabilize the fibers against plasticization. The natural gas separation performance of hollow fibers was evaluated by feeding CO₂/CH₄ gas mixture with high CO₂ content and pressure.

Membrane separation technology has recently emerged as a potential alternative technique for removing higher hydrocarbons (C₃₊) from natural gas. For economic reasons, membranes for this application should be organic vapor selective materials such as poly(dimethylsiloxane) (PDMS) or poly(1-trimethylsilyl-1-propyne) (PTMSP). These polymers, often called solubility selective polymers, sieve penetrant molecules based largely on relative penetrant solubility in the polymer. The sorption and transport properties in such polymers have been reported previously. However, most studies present only pure gas sorption and transport properties. Mixture properties, which are important for estimating membrane separation performance, are less often reported. In addition, mixed gas sorption and diffusion data in such polymers, to the best of our knowledge, have never been investigated before. This research work provides a fundamental database of mixture sorption, diffusion, and permeation data in solubility selective polymers. Two solubility selective polymers were studied: poly(dimethylsiloxane) (PDMS) and poly(1-trimethylsilyl-1-propyne) (PTMSP). The vapor/gas mixture was n-C4H10/CH4. CH4 partial pressures ranged from 1.1 to 16 atm, and [subscript n-]C₄H₁₀ partial pressures ranged from 0.02 to 1.7 atm. Temperatures studied ranged from -20 to 50 oC. The pure and mixed gas [subscript n-]C₄H₁₀ and CH₄ permeability and solubility coefficients in PDMS and PTMSP were determined experimentally using devices constructed specifically for these measurements. The pure and mixed gas diffusion coefficients were calculated from permeability and solubility data. In rubbery PDMS, the presence of [subscript n-]C₄H₁₀ increases CH₄ permeability, solubility, and diffusivity. On the other hand, the presence of CH₄ does not measurably influence [subscript n-]C₄H₁₀ sorption and transport properties. The [subscript n-]C₄H₁₀/CH₄ mixed gas permeability selectivities are lower than those estimated from pure gas measurements. This difference is due to both lower solubility and diffusivity selectivities in mixtures relative to those in pure gas. Plasticization of PDMS by [subscript n-]C₄H₁₀ does little to n-C4H10/CH₄ mixed gas diffusivity selectivity. Increases in mixed gas permeability selectivity with increasing [subscript n-]C₄H₁₀ activity and decreasing temperature were mainly due to increases in solubility selectivity. Unlike PDMS, the presence of [subscript n-]C₄H₁₀ decreases CH₄ permeability, solubility, and diffusivity in PTMSP. The competitive sorption and the blocking effects significantly reduce CH₄ solubility and diffusion coefficients in the polymer, respectively. However, similar to PDMS, the presence of CH₄ has no measurable influence on [subscript n-]C₄H₁₀ sorption and transport properties. [subscript n-]C₄H₁₀ /CH₄ mixed gas permeability selectivities in PTMSP are higher than those determined from the pure gas measurements. This deviation is a result of higher solubility and diffusivity selectivities in mixtures relative to the pure gas values. Mixed gas permeability, solubility, and diffusivity selectivities in PTMSP increased with increasing [subscript n-]C₄H₁₀ activity and decreasing temperature. The partial molar volumes of [subscript n-]C₄H₁₀ and CH₄ in the polymers were determined from sorption and dilation data. The dilation isotherms of PDMS and PTMSP in mixtures agree with estimates based on pure gas sorption and dilation measurements. The partial molar volumes of n-C4H10 and CH4 in PDMS are similar to those in liquids. In contrast, the partial molar volumes of [subscript n-]C₄H₁₀ and CH₄ in glassy PTMSP are substantially lower than those in liquids. Several models were used to fit the experimental data. For instance, the FFV model, the activated diffusion model, and the Maxwell-Stefan model were employed to describe the mixture permeability data in PDMS. Based on the Maxwell-Stefan analysis, the influence of coupling effects on permeation properties in PDMS were negligible. The dual mode sorption and permeation models were used to describe the mixed gas data in PTMSP. The dual mode permeability model must be modified to account for [subscript n-]C₄H₁₀ -induced reductions in CH₄ diffusion coefficients (i.e., the blocking effect). The FFV model provides poor correlations in PTMSP. There seems to be other factors, besides FFV per se, contributing to the temperature and concentration dependence of diffusion coefficients in PTMSP.

The removal of higher hydrocarbons from natural gas streams is an important separation that has been identified as a growth area for polymer membranes. An ideal membrane material for this separation would be more permeable to higher hydrocarbons (i.e., C3+ compounds) than to CH₄. This allows the CH₄ rich permeate to be retained at or near feed pressure, thus minimizing the requirement for repressurization followingmembrane separation. A polymer which demonstrates the ability to separate vapor from gases with high efficiency is poly [1-(trimethylsilyl)-1-propyne] (PTMSP). PTMSP is a stiff chain, high free volume glassy polymer well known for its very high gas permeability and outstanding vapor/gas selectivity. However, PTMSP is soluble in many organic compounds, leading to potential dissolution of the membrane in process streams where its separation properties are of greatest interest. PTMSP also undergoes significant physical aging, which is the gradual relaxation of non-equilibrium excess free volume in glassy polymers. Crosslinking PTMSP with bis(azide)s was undertaken in an attempt to increase the solvent resistance and physical stability of the polymer. A fundamental investigation into crosslinking PTMSP with a bis(azide) crosslinker was the focus of this thesis. Pure gas transport measurements were conducted with N₂, O₂, CH₄, C₂H6, C₃H₈, and n-C₄H₁₀ over temperatures raging from -20°C to 35°C and pressures ranging from 0 to 20 atm. Mixed gas permeation experiments were conducted using a 98 mol % CH₄, and 2 mol % n-C₄H₁₀ mixture. The mixed gas permeation experiments were conducted at temperatures ranging from -20°C to 35°C, and pressures ranging from 4 to 18 atm. Inorganic nanoparticles such as fumed silica (FS) were added to uncrosslinked and crosslinked PTMSP, and the effects of their addition on the transport properties were investigated. Crosslinking PTMSP with bis(azide)s increases its solvent resistance, and crosslinked films are insoluble in common PTMSP solvents such as toluene. At all temperatures, the initial pure and mixed gas permeabilities of crosslinked PTMSP films are less than those of uncrosslinked PTMSP. This decrease in permeability is consistent with the fractional free volume (FFV) decrease that accompanies crosslinking. Pure gas solubility coefficients are relatively unaffected by the crosslinking process, so the decrease in permeability is caused by decreases in diffusivity. The addition of FS nanoparticles increases the initial pure and mixed gas permeabilities of uncrosslinked and crosslinked PTMSP. The pure gas permeabilities and solubilities of all PTMSP films increase when the temperature decreases, while the diffusivities decrease. The rates of change in pure gas transport properties with temperature is similar for all films, so the temperature dependence of pure gas transport properties of PTMSP is unaffected by the addition of crosslinks or FS. The aging of uncrosslinked and crosslinked PTMSP films was investigated by monitoring N₂, O₂ and CH₄ permeabilities and FFV over time. The FFV and permeabilities of crosslinked films decreased over time, so crosslinking did not arrest the physical aging of PTMSP, as has been previously reported, and these differences in aging observations are likely to be a consequence of differences in post film casting thermaltreatments. The addition of 10 wt % polysiloxysilsesquioxanes (POSS) nanoparticles decreases the permeabilities of uncrosslinked and crosslinked PTMSP by approximately 70 %, and the permeability and FFV values of the resulting nanocomposite films were stable over the course of 200 days. In all PTMSP films, the mixed gas permeabilities of n-C₄H₁₀ increase with decreasing temperature, while the mixed gas CH₄ permeabilities decrease with decreasing temperature. As a result, the mixed gas n-C₄H₁₀/CH₄ permeability selectivities increase with decreasing temperatures. The addition of crosslinks and FS nanoparticles to PTMSP decreases the mixed gas n-C₄H₁₀/CH₄ permeability selectivities, and changes in the free volume characteristics of PTMSP caused by crosslinking and FS nanoparticles are thought to reduce the blocking of CH₄ permeation by n-C₄H₁₀. / text

Abstract not available

Metal-organic frameworks are a new class of porous materials that have potential applications in gas storage, separations, catalysis, sensors, non-linear optics, displays and electroluminescent devices. They are synthesized in a "building-block" approach by self-assembly of metal or metal-oxide vertices interconnected by rigid linker molecules. The highly ordered nature of MOF materials and the ability to tailor the framework's chemical functionality by modifying the organic ligands give the materials great potential for high efficiency adsorbents. In particular, MOFs that selectively adsorb CO₂ over N₂, and CH₄ are very important because they have the potential to reduce carbon emissions from coal-fired power plants and substantially diminish the cost of natural gas production. Despite their importance, MOFs that show high selective gas adsorption behavior are not so common. Development of MOFs for gas adsorption applications has been hindered by the lack of fundamental understanding of the interactions between the host-guest systems. Knowledge of how adsorbates bind to the material, and if so where and through which interaction, as well as how different species in adsorbed mixture compete and interact with the adsorption sites is a prerequisite for considering MOFs for adsorptive gas separation applications. In this work, we seek to understand the role of structural features (such as pore sizes, open metal site, functionalized ligands, pore volume, electrostatics) on the adsorptive separation of CO₂, CO and N₂ in prototype MOFs with the help of molecular modeling studies (GCMC simulations). Our simulation results suggest that the suitable MOFs for CO₂ adsorption and separation should have small size, open metal site, or large pore volume with functionalized groups. Some of the experimental challenges in the MOF based adsorbents for CO₂ capture include designing MOFs with smaller pores with/without open metal sites. Constructing such type of porous MOFs can lead to greater CO₂ capacities and adsorption selectivities over mixtures of CH₄ or N₂. Therefore, in the second project, we focused on design and development of small pore MOFs with/without open metal sites for adsorptive separation of carbon dioxide from binary mixtures of methane and nitrogen. We have synthesized and characterized several new MOFs (single ligand and mixed ligand MOFs) using different characterization techniques like single-crystal X-ray diffraction, powder X-ray diffraction, TGA, BET, gravimetric adsorption and examined their applicability in CO₂/N₂ and CO₂/CH₄ mixture separations. Our findings from this study suggest that further, rational development of new MOF compounds for CO₂ capture applications should focus on enriching open metal sites, increasing the pore volume, and minimizing the size of large pores. Flue gas streams and natural gas streams containing CO₂ are often saturated by water and its presence greatly reduces the CO₂ adsorption capacities and selectivities. So, in the third project, we investigated the structural stability of the developed MOFs by measuring water vapor adsorption isotherms on them at different humid conditions to understand which type of coordination environment in MOFs can resist humid environments. The results of this study suggest that MOFs connected through nitrogen-bearing ligands show greater water stability than materials constructed solely through carboxylic acid groups.

Robust industrially relevant membranes for CO₂ removal from aggressive natural gas feed streams were developed and characterized. Asymmetric hollow fiber membranes with defect-free selective skin layers on an optimized porous support substructure were successfully spun and subsequently stabilized by covalent crosslinking within the economical membrane formation process. Thermal treatment conditions, which promote sufficient crosslinking without introducing defects or undesired substructure resistance, were identified. It was found that crosslinking improves membrane efficiency and plasticization resistance as well as mechanical strength of fibers. The capability to maintain attractive separation performance under realistic operating conditions and durability against deleterious impurities suggests that the crosslinked fibers have great potential for use in diverse aggressive applications, even beyond the CO₂/CH₄ example explored in this work.

A new concept of a "fiber sorbent" has been investigated. The fiber sorbent is produced as a pseudo-monolithic material comprising polymer (cellulose acetate, CA) and zeolite (NaY) by applying hollow fiber spinning technology. Phase separation of the polymer solution provides an appropriately porous structure throughout the fiber matrix. In addition, the zeolite crystals are homogeneously dispersed in the polymer matrix with high loading. The zeolite is the main contributor to sorption capacity of the fiber sorbent. Mass transfer processes in the fiber sorbent module are analyzed for hydrogen recovery and compared with results for an equivalent size packed bed with identical diameter and length. The model indicates advantageous cases for application of fiber sorbent module over packed bed technology that allows system downsizing and energy saving by changing the outer and bore diameters to maintain or even reduce the pressure drop. The CA-NaY fiber sorbent was spun successfully with highly porous structure and high CO2 sorption capacity. The fiber sorbent enables the shell-side void space for thermal moderation to heat of adsorption, while this cannot be applied to the packed bed. The poly(vinyl alcohol) coated CA-NaY demonstrated the thermal moderation with paraffin wax, which was carefully selected and melt at slightly above operating temperature, in the shell-side in a rapidly cycled pressure swing adsorption. So this new approach is attractive for some hydrogen recovery applications as an alternative to traditional zeolite pellets.

The goal of this project was to understand and model the performance of hybrid inorganic-organic membranes under realistic operating conditions for hydrocarbon gas/vapor separation, using butane isomers as the model vapors and a hybrid membrane of 6FDA-DAM-5A as an advanced separation system. To achieve the set goal, three objectives were laid out. The first objective was to determine the factors affecting separation performance in dense neat polymer. One main concern was plasticization. High temperature annealing has been reported as an effect means of suppressing plasticization. A study on the effect of annealing temperature was performed by analyzing data acquired via sorption and permeation measurements. Based on the findings from this study, a suitable annealing temperature was determined. Another factor studied was the effect of operating temperature. In deciding a suitable operating temperature, factors such as its possible effect on plasticization as well as reducing heating/cooling cost in industrial application were considered. Based on the knowledge that industrial applications of this membrane would involve mixture separation, the second objective was to understand and model the complexity of a mixed gas system. This was investigated via permeation measurements using three feed compositions. An interesting transport behavior was observed in the mixed gas system, which to the best of our knowledge, has not been observed in other mixed gas systems involving smaller penetrants. This mixed gas transport behavior presented a challenge in predictability using well-established transport models. Two hypotheses were made to explain the observed transport behavior, which led to the development of a new model termed the HHF model and the introduction of a fitting parameter termed the CAUFFV fit. Both the HHF model and CAUFFV fit showed better agreement with experimental data than the well-established mixed gas transport model. The final objective was to explore the use of mixed matrix membranes as a means of improving the separation performance of this system. A major challenge with the fabrication of good mixed matrix membranes was the adhesion of the zeolite particle with the polymer. This was addressed via sieve surface modification through a Grignard treatment process. Although a Grignard treatment procedure existed, there was a challenge of reproducibility of the treatment. This challenge was addressed by exploring the relationship between the sieves and the solvent used in the treatment, and taking advantage of this relationship in the Grignard treatment process. This study helped identify a suitable solvent, which allowed for successful and reproducible treatment of commercial LTA sieves; however, treatment of lab-made sieves continues to prove challenging. Based on improved understanding of the Grignard treatment reaction mechanism, modifications were made to the existing Grignard treatment procedure, resulting in the introduction of a "simplified" Grignard treatment procedure. The new procedure requires less control over the reaction process, thus making it more attractive for industrial application. Permeation measurements were made using mixed matrix membranes in both single and mixed gas systems. Selectivity enhancements were observed under both single and mixed gas systems using sieve loadings of 25 and 30wt%. The Maxwell model was used to make predictions of mixed matrix membrane performance. Although the experimental results were not in exact agreement with Maxwell predictions, the observed selectivity enhancement was very encouraging and shows potential for future application. Recommendations were made for future study of this system.