Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The purification of contaminated air streams via recative separation techniques
(USC Thesis Other)
The purification of contaminated air streams via recative separation techniques
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE PURIFICATION OF CONTAMINATED AIR STREAMS VIA RECATIVE SEPARATION TECHNIQUES by Majid Monji _____________________________________________________________________ A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (CHEMICAL ENGINEERING) May 2015 Copyright 2015 Majid Monji ii Dedication Anyone who has lost his/her life and health by chemical attacks especially my friendly Iranian people who were intoxicated during the Iran Iraq imposed war (1980-1988). iii Acknowledgments I would like to express my deep and sincere gratitude to my advisors, Professors Theodore T. Tsotsis and Fokion N. Egolfopoulos, for giving me the opportunity to work in their group. Their high level of knowledge and logical way of thinking have been of great value and motivation for me, and their personal guidance has provided a good basis for the present thesis. They are kind and smart scientists, with excellent sense of humor, who have given me considerable advice whenever I faced hard problems in my work. They not only provided me with help in my research, but have also given me their time generously to encourage, advice and support me. They were all the time available and open for any question and discussion that I had. I am grateful to my Ph.D. committee members, Professors Iraj Ershaghi, Katherine S. Shing, and Dr. Malancha Gupta for their valuable comments and suggestions and the time that they spent to read my thesis and giving me their instructions. In particular, Professor Iraj Ershaghi who was my defense committee member as well. I wish to thank Dr Richard W Baker, a role model for me, Professors Yi Hua Ma, Dr Haiqing Lin and Florian Stadler, and Farmarz Afshar Taromi for their kind and helpful support thoughout my research. I would like to thank Dr. Paul Liu, Dr. Rich Ciora, and Mr. Doug Parsley, our industrial collaborators, from Media and Process Technology, Inc. (MPT) for their continued help and support during this project. It would be impossible for me to complete this research without their support. iv I would also like to acknowledge the great support of Mrs. Tina Silva and Mr. Shokry Bastorous who always helped me in my experiments by providing me with any assistance that I needed. I also thank the other staff members in our Department, namely Mr. Martin Olekszyk, iv Mr. Andy Chen, Ms. Karen Woo, Ms. Laura Carlos, Ms. Heather Alexander, Ms. Angeline Fugelso, Ms. Idania Takimoto, and Mr. Jason Ordonez. Their valuable help throughout my graduate studies at USC is greatly appreciated. I appreciate the help of Sasan Dabir, a great friend and a classy colleague in the Tsotsis Group, who was available all the time to help me, I really thank him for all his help during my research. I wish to warmly thank my former and present graduate student colleagues in the Department, Dr. Wangxue Deng and Dr. Xiaojie Yan, Ms. Basabdatta Roychaudhuri, Hamed Barghi, Mr. Alireza Divsalar, Mr. Ashkan Garshasbi, and Mr. Devang Dasani for the friendly environment that they created in the department, and off course their support during my PhD studies. I want to thank Mr. Yuzheng Zhang for his technical help with the SEM images, and to express my warmest thank to Mr/ Abdullah Alwabel who helped me with the operation of the GPC and BET instruments. It is my great pleasure furthermore to acknowledge my parents, for their love and support, my brothers Ali Asghar and Ali Akbar, and my lovely sisters Leila, Soghra, Asieh and Zahra and my brothers in law, Jafar Najafi and Ardavan Khajeh and my sister in law Marzieh Khani for their moral support. My nephews Hamid Reza Khajeh and Abolfazl Najfi and my niece, Fatemeh Khajeh, who all members of our family. v Table of contents Dedication ii Acknowledgments iii List of Tables vi List of Figures vii Chapter 1: Introduction 1 1.1 Background 1 1.2 Adsorption 6 1.2.1 Adsorption Fundamentals 6 1.2.2 Transport of Adsorbed Gases 10 1.2.3 Surface Diffusion 14 1.2.4 Capillary Condensation from Mixed Vapors 26 References 27 Chapter 2: Carbon Molecular Sieve (CMS) Membranes 33 2.1 Introduction 33 2.2 CMS Membrane Studies 38 2.2.1 Membrane Preparation 38 2.2.2 Transport Measurements 42 2.2.3 BET analysis 45 2.3 Steam Activation as a Post-treatment Step 49 2.3 Conclusions 54 References 55 Chapter 3: Development of Multi-Tubular, Flow-Through Catalytic Membrane Reactor (MFTCR) 60 3.1 Introduction 60 3.2 Experimental Section 64 3.3 Results and Discussion 72 3.4 Conclusions 80 References 82 Chapter 4: Future Work 85 References 87 Bibliography 88 vi List of Tables Table 1.1 Comparison between physical adsorption and chemisorption 7 Table 1.2 Different categories of pores based on IUPAC 10 Table 1.3 Transport regimes in porous media 11 Table 1.4 Surface diffusion models 16 Table 1.5 typical works done on gas condensation phenomena 24 Table 2.1 Types of membranes 33 Table 2.2 Select papers on CMS membranes 36 Table 2.3 Material used in different layers of M&PT ceramic membranes used in this research, as reported by the vendor 39 Table 2.4 Some physical properties of M&P ceramic membranes used in this research, as reported by the vendor 40 Table 2.5 He and Ar single gas properties of membrane 44 Table 2.6 BET surface area of the substrate and CMS membranes 47 Table 2.7 Select papers on the use of steam activation in the preparation of microporous carbons 50 vii List of Figures Figure 1.1 Chemical structures of Sarin (a) and its simulant, DMMP (b) 5 Figure 1.2 Six types of adsorption isotherms 8 Figure 1.3 Adsorption isotherms for systems with incomplete wetting proposed by Adamson 9 Figure 1.4 Adsorption isotherms for ethylene on carbon black at different temperatures 10 Figure 1.5 Condensation of gas molecules inside a capillary and formation of meniscus, known as capillary condensation a) before capillary condensation; b) after capillary condensation. 20 Figure 1.6 A schematic plot of the permeability versus the relative pressure in case of multilayer diffusion and capillary condensation. The maximum gives the point of capillary condensation 21 Figure 1.7 six modes (F1-F6) flow of condensable vapor in micropores 23 Figure 2. 1 A schematic representation of membrane separation mechanism 35 Figure 2.2 Schematic representation of an asymmetric composite membrane, 1 is polymer layer coated on the substrate, 2 is the active layer, 3 is the intermediate layer and, 4is porous substrate 38 Figure 2.3 Chemical structure of PEI 40 Figure 2.4 samples of substrate (right) and membrane (left) 41 Figure 2.5 Schematic of experimental apparatus for measuring permeance and ideal selectivity Table 2.5 Permeances for single gases Ar/He 44 Figure 2.6 Pore size distribution of both substrate A and membranes B and C 48 viii Figure 2.7 Nitrogen adsorption isotherm (T<Tcritical=126.15 K) of the CMS membranes B and C 48 Figure 2.8 Permeance of membranes as a function of temperature and time of steam activation: (top) He permeance; (bottom) Ar permeance and ideal selectivity (membrane I 54 Fig. 3.1. The multi-tubular membrane bundle before (left), and after catalytic impregnation(right) 64 Fig. 3.2. Various views of the membrane module 65 Fig. 3.3. A schematic of the MFTCMR system 67 Fig. 3.4. The air permeation properties of bundle #1 before the DMMP decompositionexperiment 70 Fig. 3.5. Propane/air light-off curve for bundle#1 before exposure to DMMP 71 Fig. 3.6. DMMP thermocatalytic conversion of bundle #1; MMP feed concentration (ppmv)=1000±5% 72 Fig. 3.7. The air permeation properties of bundle #1 after exposure to the DMMP in the decomposition experiment 74 Fig. 3.8. Propane/air light-off curve for bundle#1 after exposure to DMMP 75 Fig. 3.9. Propane/air light-off curve for bundle#2 before exposure to DMMP 76 Fig. 3.10. DMMP thermocatalytic conversion of bundle #2; DMMP feed concentration(ppmv)=941±5 77 Fig. 3.11. Propane/air light-off curve for bundle#2 after exposure to DMMP 77 Fig. 3.12. Propane/air light-off curve for bundle#3 before exposure to DMMP 79 Fig. 3.13. DMMP thermocatalytic conversion of bundle #3; DMMP feed concentration (ppmv)=997±3% 79 ix Fig. 3.14. Propane/air light-off curve for bundle#3 after exposure to DMMP 80 1 Chapter 1: Introduction 1.1 Background Separation of condensable vapors, both those existing in the gas phase at room temperature and pressure conditions, such as Freon, carbon dioxide and ammonia, and those co-existing both in the liquid and vapor phases, such as water, dimethyl methylphosphonate (DMMP), etc., has long been investigated by researchers around the world. The physical properties of these compounds make their behavior different from that of non-condensable gases such as helium and argon. Experiments have confirmed, for example, that they behave differently during transport through porous media and in separation experiments through membranes. For example, in contrast to non-condensable species which show either a linear (when transport occurs by combined molecular and Knudsen diffusion) dependence of permeability on average transmembrane pressure or no dependence at all (when Knudsen transport prevails), these molecules show a non- monotonic permeability behavior (maximum) with respect to pressure (Lin & Freeman, 2004; Naito, Kamiya, Terada, Mizoguchi, & Wang, 1996; Stern, Fang, & Frisch, 1972; Tzevelekos, Kikkinides, Stubos, Kainourgiakis, & Kanellopoulos, 1998; Tzevelekos, Romanos, Kikkinides, Kanellopoulos, & Kaselouri, 1999). Under the right set of conditions, when separating condensable vapors from their mixtures with non-condensable gases with smaller molecular size (e.g., air or nitrogen), one observes (Pinnau & Toy, 1996; Shiflett & Foley, 1999; Suda & Haraya, 1995) a high separation selectivity of the membrane towards the condensable (and heavier) component, a phenomenon also known as “reverse selectivity”. All these interesting types of behavior observed when these species and their mixtures come in contact with porous media, in general, and microporous membranes in particular are due to condensation that occurs 2 within the pore structure as a result of Van der Waals bonds forming with the surface of solids (physical adsorption) and among the species themselves (liquefaction). As noted above, due to its technical significance the complexity of flow of adsorbable gases through microporous media has received increased scrutiny and attention. The condensable vapor of interest in this study is DMMP. Though it finds also a number of industrial uses such as flame retardant, additive for gasoline, anti-foaming agent, plasticizer, stabilizer, textile conditioner, antistatic agent, and as an additive for solvents and low-temperature hydraulic fluids, our interest in DMMP is primarily because it is a common simulant for Sarin, a potent nerve agent classified as a chemical warfare agent (CWA) stimulant (see Figure 1.1 for the structure of both compounds). DMMP is frequently studied by researchers because of the danger to work with CWA’s, directly (Grandcolas, Louvet, Keller, & Keller, 2009; Mera et al., 2010; Segal, Cao, Suib, Tang, & Satyapal, 2001; Trubitsyn & Vorontsov, 2005). The threat from CWA’s has increased in recent years. Potential attacks both on the battlefield as well as in civilian areas are likely to result in high casualties, thus the need for efficient protection. The current methods utilized involve adsorption of the CWA by a variety of adsorbents or the use of impermeable barriers (clothing). Adsorption is effective, but it presents the drawback that it does not lead to the destruction of the toxic agents, thus the need of the eventual disposal of the sorbent medium. In addition, common adsorbents are not specific enough towards the CWA and thus get saturated easily by the other contaminants present. Catalytic destruction of the CWA, as a result, has attracted recent attention, having the advantage of the complete destruction of the CWA. A novel catalytic technology using the concept of flow- through catalytic membrane reactor (FTCMR), has been shown to be very effective towards the destruction of CWA [Monji et al, 2014] and is the focus of this Thesis. 3 Catalytic destruction technologies themselves face the challenge that the oxidation of the CWA results in the formation of inorganic by-products (e.g., HCl, NH 3 , P 2 O 5 , etc.) which may interact with the catalyst thus diminishing the “protection life”, defined as the time during which the catalytic oxidizer provides complete protection towards the CWA. In our studies we have found (as did others) that protection life of a given device is dependent on the level of CWA challenge (i.e., its concentration) with higher challenges resulting in lower protection times, as expected. And though this may not be a serious problem for individual protection (IP) systems, it presents a challenge for the collective protection (CP) systems, which are supposed to be functional over extended periods of time. The successful use of a Multi-tubular flow-through catalytic membrane reactor (MFTCMR) as a protection system against a chemical warfare agent (CWA) is the primary focus (Monji; Monji & Tsotsis; Motamedhashemi, Monji, Egolfopoulos, & Tsotsis, 2014; Tsotsis et al., 2014) of this Thesis [Monji M., et al 2014]. This MFTCMR, employing a multi-tubular catalytically active mesoporous membrane, is applied for the destruction of dimethyl methylphosphonate, which as noted above is known as a chemical precursor (and used to simulate its characteristics) for Sarin (GB), a toxic CWA. In chapter 3, results are reported of efforts to scale-up a single-tube, lab-scale FTCMR into a MFTCMR towards a practical collective protection (CP) application. The MFTCMR has been shown quite effective, providing extended protection periods against this CWA simulant. A key part of the research effort focused on improving the preparation of the catalytically active membranes to prepare membrane bundles with reproducible transport characteristics and reactivity. A key challenge for the large-scale production of such devices, furthermore, remains the development of a simple non-destructive test that assures that the produced parts are appropriate for the proposed application and that they 4 continue to remain active during their shelf-life prior to their use. As part of this research a simple propane-in-air light-off test was studied, which was shown capable to track the reactivity and performance of the MFTCMR. We have also studied in our research hybrid device that combines the use of membrane separation (as stage I) followed by the catalytic oxidizer (FTCMR) as stage II. Stage I provides for the bulk removal of the CWA while stage II provides for its complete destruction by catalytic oxidation. Combining both technologies (membrane separation and catalytic oxidation) provides for significant synergy. Membranes are not by themselves capable to provide complete protection towards CWA which the catalytic oxidizer provides. However, the membrane, by removing a large fraction of the CWA helps to significantly prolong the protection time offered by the oxidizer. The purpose of this research, presented in this Thesis, is to investigate the characteristics of physical capture step as it is combined with the thermocatalytic destruction step using FTCMR system. A key focus in this Thesis is on the study of the membrane separation step, and eventually of the combined device. The physical capture step is accomplished via “inverse selectivity” or otherwise known as surface flow membranes (SFM). In our studies we make use of carbon molecular sieve (CMS) membranes with a high DMMP removal rate capability. The preparation method of these membranes is described in Chapter 2, while the studies on their use in DMMP separation is presented in Chapter 4. A simple continuum-type model to describe the separation mechanism for these membranes is also provided in Chapter 4. A comprehensive model describing the transport of gas mixtures containing condensable vapors through porous membranes is developed and presented chapter 5. The 5 Figure 1.1 Chemical structures of Sarin (a) and its simulant, DMMP (b) (Quenneville, Taylor, & van Duin, 2010) model accounts for the fact that depending on the conditions on either side of the membrane and its structural characteristics, e.g., average pore size and pore size distribution (PSD), vapors may transport through the pores either as liquid or gas species or both for a sufficiently broad PSD. For conditions under which the vapor completely condenses within the membrane structure and blocks the transport of the lighter component(s) in the mixture, the membrane exhibits behavior which is known in the technical literature as “reverse selectivity”. The present model provides qualitative insight into the factors that determine the transition from the conventional behavior, whereby the lighter component transports preferentially through the membrane, to that of “reverse selectivity” behavior. The model is applied here to describe the removal of dimethyl methylphosphonate from contaminated air streams. The ability of the membrane to effectively remove the CWA vapor is shown to be highly dependent on the membrane structure as well as the other operating conditions. In this paper, the effects of various parameters and experimental conditions on the membrane’s separation characteristics towards DMMP and the potential air leakage through the membrane are described. We conclude in Chapter 6 by describing future experimental and modeling work. In the remainder of this introductory Chapter we provide some background information on the adsorption and transport of gas molecules through microporous media and membranes to assist 6 our reader to better understand some of the discussion presented in the remaining Chapters of this Thesis. 1.2 Adsorption 1.2.1 Adsorption Fundamentals Molecular adsorption on a solid surface can be classified either as physical or chemical depending on the magnitude of the heat of adsorption. For chemisorption, the common view held is that there is a chemical bond formed between the adsorbate and the surface, whereas for physical adsorption, no direct chemical bond is formed, but rather, the adsorbate is held by physical (i.e., van der Waals and electrostatic) forces. Chemisorption is confined to a monolayer, while physisorption can be either in the form of single layer or multilayers, particularly at relatively high pressures. The energy of chemisorption is of the same order of magnitude as the energy change in a comparable chemical reaction. Physisorption is always exothermic, and the energy involved is not much larger than the energy of condensation of the adsorbate. Table 1.1 compares the properties of physical adsorption and chemisorption. Physical adsorption and desorption isotherms, particularly of nitrogen at its liquid temperature, known as the BET data, are useful in characterizing the surface area, the mesopore and micropore volumes and the pore size distribution of various porous materials, including membranes. Chemisorption analysis (typically using CO and hydrogen as probe gases), on the other hand, are useful to characterize the active areas, those capable of forming a chemical bond with the adsorptive species. Adsorbed molecules can be classified as either mobile or immobile depending on whether the molecule can move around while adsorbed on the surface, or remains in its adsorbed location 7 until it desorbs and returns to the fluid phase. As previously noted, physisorption can occur in the form of a monolayer or multilayers (or both) depending on the magnitude of the relative pressure (P/P 0 ), where P and P 0 are the partial vapor pressure of a component and its saturation vapor pressure at the same temperature, respectively. Table 1.1 Comparison between physical adsorption and chemisorption Physical adsorption Chemisorption The forces operating are weak van der-Waal’s forces. The forces operating are similar to those of a chemical bond. The heat of adsorption is low, i.e. about 20 – 40 kJ mol -1 The heat of adsorption is high, i.e. about 40 – 400 kJ mol -1 No compound formation takes place. Surface compounds are formed. The process is reversible, i.e. desorption of the gas occurs by increasing the temperature or decreasing the pressure. The process is sometimes irreversible. Efforts to free the adsorbed gas may result in a different compound. It does not require any activation energy. The process is activated. Adsorbed amount decreases with increase in temperature. Rate of adsorption first increases with increase of temperature. It is not specific in nature, i.e., all gases are adsorbed on all solids to some extent. It is specific in nature and occurs only when there is some possibility of compound formation between the gas being adsorbed and the solid adsorbent. The amount of the gas adsorbed relates to the ease of liquefaction of the gas. No such correlation exists. It can form multilayers. It forms only monolayers. Adsorption begins in the form of a monolayer, multilayers form as the pressure increases, eventually the pores being filled with a liquid-like phase, the phenomenon known as capillary condensation. The amount adsorbed q (mol) at equilibrium, by a given mass of solid m (g), is dependent on the pressure, P, the temperature T, and the properties of the gas-solid system: 0 ( , , , ) q f P P T system m 1.1 IUPAC classifies adsorption isotherms into six different types (Sing, 1994) as shown in Figure 1.2. Type I is often referred to as the Langmuir type and is used to describe monolayer 8 adsorption isotherms (Brunauer, 1943) in microporous adsorbents. Types II and III describe adsorption on macroporous adsorbents with strong and weak adsorbate-adsorbent interactions, respectively (Sing, 1994). Type II is the common S-shaped isotherm and the uptake at the top of the rise in the curve represents the completion of the monolayer and the beginning of the formation of the multilayer. Figure 1.2 Six types of adsorption isotherms (Braunauer, 1945; Donohue & Aranovich, 1998, 1999; Gilliland, Baddour, & Russell, 1958; Sing, 1994) Types IV and V represent mono-and multilayer adsorption plus capillary condensation ((E. A. Flood, 1967), p85) on highly porous adsorbents and the flattening of the isotherms at near- saturation pressure is ascribed to complete filling of all the pores. Type VI, which was not included in the original Brunauer classification (Brunauer, 1945), illustrates the fact that the adsorption isotherms can have one or more steps (A. W. Adamson & Gast, 1967). There are also a number of systems whose isotherms do not match any of the IUPAC types, and are not intermediate between the IUPAC cases. For example, systems with incomplete wetting have isotherms where adsorption tends to a finite limit as pressure goes to the saturation vapor pressure, even for macroporous adsorbents. Adamson suggested (A. W. Adamson & Gast, 1967) adding two new types of adsorption isotherms describing this behavior shown in Figure 1.3. 9 Figure 1.3 Adsorption isotherms for systems with incomplete wetting proposed by Adamson (A. W. Adamson & Gast, 1967). The adsorption of fluids under supercritical conditions, i.e., when the temperature of the experiment exceeds the critical temperature of the compound, does not fall under any of the IUPAC types. Figure 1.4, for example, shows adsorption isotherms of ethylene on carbon black for temperatures from 263 to 323 K (Findenegg, Korner, Fischer, & Bohn, 1983). The critical temperature of ethylene is about 282 K, and Figure 1.4 shows the change in shape of the adsorption isotherm as the temperature condition changes from subcritical (curves 1 and 2) to supercritical (curves 3-9). The adsorption isotherms for subcritical ethylene are of Type II in the IUPAC classification, but for supercritical ethylene, the adsorption isotherms show maxima that range from being very pronounced (curves 3-7) to rather weak (curves 8 and 9). Similar behavior has been observed for supercritical methane, ethane, argon, neon, krypton, nitrogen, carbon oxide, carbon dioxide, and nitrogen oxide on different adsorbents (Bose, Chahine, Marchildon, & St ‐Arnaud, 1987; Michels, Menon, & Ten Seldam, 1961; MOFFAT & Weale, 1955; Payne, Sturdevant, & Leland, 1968; Vidal et al., 1990; Wakasugi, Ozawa, & Ogino, 1981) 10 Figure 1.4 Adsorption isotherms for ethylene on carbon black at different temperatures (Findenegg et al., 1983). 1.2.2 Transport of Adsorbed Gases Mass transport of fluids in pores depends on their pore size. IUPAC classifies (Table 1.2) pores into three different types on the basis of their size (diameter, d): Table 1.2 Different categories of pores based on IUPAC (Choi, Do, & Do, 2001) Type of porosity Average pore size range, APS (Å) Micro APS < 20 Meso 20< APS < 500 Macro APS > 500 During transport in micropores, the overlapping surface forces by the opposing pore walls dominate. In mesopores, surface forces are still important and capillary forces contribute as well. For the macropores, the surface contributes little to transport. In the absence of capillary condensation and monolayer or multilayer adsorption (Choi et al., 2001; Keizer, Uhlhorn, & Burggraaf, 1988), transport in pores can occur by five different mechanisms as Table 1.3 indicates, depending on the Knudsen number, Kn.N, defined as λ/d where λ is the molecular mean free path, i.e., the average distance travelled by a moving molecule between successive collisions and d the diameter of the pore. 11 Table 1.3 Transport regimes in porous media (Gad-el-Hak, 2001) Flow mode Knudsen number (Kn.N) Pore diameter (ºA) Viscous Kn.N << 1 >200 Knudsen Kn.N >> 1 20-1000 Transition Kn.N = 1 NA Molecular diffusion NA >100 Micropore or configurational diffusion NA <15 Continuum while no molecular diffusion Kn.N →0 NA Continuum with molecular diffusion Kn.N ≤10 -3 NA Continuum transition 10 -3 ≤ Kn.N ≤10 -1 NA Transition region 10 -1 ≤ Kn.N NA Free molecular 10≤ Kn.N NA In the presence of substantial adsorption surface flow also occurs in parallel with viscous flow and/or diffusion. When capillary condensation occurs inside the pore, the transport is thought to occur by Poiseuille flow of a liquid (Sidhu & Cussler, 2001). Knudsen diffusion occurs in small pores, whose diameter (d) is smaller than the molecular mean free path λ; under such conditions the gas molecules collide with the pore walls more frequently than with each other. Knudsen derived an expression for the molecular flow through a single cylindrical capillary (Knudsen, 1909). Using a simple model to describe the porous medium consisting of a bundle of nonintersecting parallel capillaries the Knudsen theory can be used to describe the flow rate (mole.s -1 ) as 3 1 8 3 2 2 P g AG n r dP dP F dx dx MRT MRT 1.2 Where 1 8 3 r G or using Equation 1.3 (Holt et al., 2006); 3 28 3 g P F r A MRT L 1.3 12 where, A p (the cross-sectional area of a single pore), r( pore radius), n (number of pores), M (molecular weight), R (gas constant), T(absolute temperature), V m is the molar volume, ∆P is the pressure drop, L is the thickness of the membrane, σ is the areal pore density, and A is the total area of the membrane. The permeability of a porous material is commonly defined as (Knudsen, 1909; Lee & Hwang, 1986): 1 2 () g g P P F G Q P MRT A L 1.4 Or using Equation 1.3, 33 2 8 16 1 () 33 2 () g g F Q r r P MRT MRT A L 1.5 where L p is the thickness of the porous medium and ∆P is the pressure difference. In the above equations the presence of an adsorbed layer is ignored. When such layer is present and its presence impacts the cross-sectional area of the pore available for transport the equations may have to be appropriately modified. When the mean free path of the gas molecules is comparable with the pore diameter, transition flow occurs by the combined effects of both the viscous and Knudsen mechanisms. For single gases, the total flux J t can be written as: t Visc Knud J J J 1.6 22 11 22 11 2 2 28 83 28 83 pp pp pp pp r P dp r dp RT dz MRT dz r dp r dp P RT dz MRT dz 1.7 13 From the above equation the permeability Q g =J t /(∆P/L) is a linear function of average transmembrane pressure, with a positive intercept, which is the Knudsen permeability, whereas the slope is governed by the gaseous viscous flow. The increase of the permeability with pressure is an indication that gaseous viscous flow may be responsible for mass transfer. However, as the surface diffusion permeability may also exhibit an increase with pressure (see further discussion below) to distinguish between these two mechanisms, one needs to conduct an experiment with a non-adsorbing gas and to obtain the viscous flow permeability of such a gas. Because viscous flow is inversely proportional to viscosity, the viscous flow permeability for the adsorbing species can be then estimated easily. If the rate of increase of the experimental permeability versus pressure is greater than what is estimated, this additional transport can be assumed to be contributed by the surface diffusion. Experimentally, surface diffusion is expected to dominate at low pressures while viscous flow is expected to dominate at much higher pressures. As mentioned above, when conditions are such that adsorption on the pore walls is significant and the adsorbed species are mobile, literature data have shown that the additional surface flow should be added to the gaseous flow. Because this transport mechanism is hard to confirm directly, it is usually inferred by comparing the real flux observed with that expected (calculated) based from measurements with helium, which is believed never to undergo surface transport (Sidhu & Cussler, 2001). Surface diffusion is thought to contribute significantly to mass transport in porous media, especially in separation through membranes, when adsorption and reaction phenomena are also present. For example, it has been reported that, for microporous catalysts, surface diffusion can account for more than 50% of the total mass flow rate (Butt & Reed Jr, 1971; Choi et al., 2001; Schneider & Smith, 1968). 14 1.2.3 Surface Diffusion Diffusion of adsorbed molecules on solid surfaces is thought to be an activated process involving molecules jumping from a site to neighboring sites. Several important variables can influence such a process, including surface concentration, operating temperature, the type of gas species, and the pore structure. A number of theoretical models have been proposed in the literature. Carman and co-workers, for example, found that surface diffusivity increases rapidly as monolayer coverage is approached (Carman & Raal, 1951). Beyond monolayer coverage, the diffusivity showed a decline and then increased again rapidly near the capillary condensation region. The models are divided into two distinct groups according to the two different approaches followed: (i) "The site hopping models" are typically used for low levels of adsorption (Tamon, Okazaki, & Toei, 1981). These models consider the flux to take place by (non-correlated) jumps of adsorbed molecules from site to site, where the sites are not necessarily non-occupied (R. Uhlhorn, Keizer, & Burggraaf, 1992; R. J. R. Uhlhorn, 1990).The surface flux is commonly described by a Fickian equation (Crank, 1975). S SS dC JD dx 1.8 where J S is the surface diffusion flux (mole.m -2 .s -1 ) , C S is the concentration of adsorbed layer (mole.m -3 ), x is the length of pores (m) and D s is the surface diffusivity (m 2 .s -1 ) , which is thought to have an Arrhenius dependence on temperature. Since in this model the diffusion process is assumed to occur by jumping motions from one site to another, the activation energy for diffusion is related to the heat of adsorption. This suggests that strongly adsorbed molecules are less mobile than weakly adsorbed ones. For very low pressure, where gaseous viscous flow is 15 negligible and the adsorption isotherm is linear (C s = KP), the total flux is the sum of the Knudsen flux and the surface diffusion flux: 28 3 ks s r dP dP J D K RTM dx dx 1.9 Where J ks is the Knudsen flux (mole.m -2 .s -1 ). K is a constant with the unit of mole.m -3 .pa -1 .. The permeability (Q ks , mole.m.m -2 .pa -1 .s -1 ) for this combined Knudsen and surface diffusions is 28 3 ks s r Q D K RTM 1.10 Experimentally (Bouwmeester, Burgraaf, Burgraaf, & Cot), and as predicted by the above equation, the decrease of the surface diffusivity with temperature is much faster than that of Knudsen diffusion. This result arises because the heat of adsorption is greater than the activation energy for surface diffusion. Therefore, if surface diffusion is to be eliminated, experiments must be carried out at high temperatures. Unfortunately literature data have shown that the surface diffusivity is a strong function of both the surface coverage and the operating temperature. A number of theoretical models describe the dependence of the surface diffusivity on the surface concentration (Table 1.4) which are reviewed in detail elsewhere (Do, 1996). Higashi et al. [1963] studied the dependence of surface diffusivity on concentration using propane on silica glass. Their model was based on Hill’s “hopping model” (Hill, 1960), in which the adsorbed molecules migrate from adsorption site to adsorption site via hopping movements. Their model further assumes that, when a molecule encounters a site occupied by another molecule, it is not bound to this site, but rather immediately bounces off after colliding with the occupying molecule. Then, the molecule continues to jump until it finds an unoccupied site on 16 Table 1.4 Surface diffusion models (Chen & Yang, 1991; Do, 1996; Higashi, Ito, & Oishi, 1964; Suzuki & Fujii, 1982; R. T. Yang, Fenn, & Haller, 1973) Referens Model (Higashi et al., 1964) 0 (1 ) S S D D (R. T. Yang et al., 1973) 12 ( )/ 1 0 2 1 (1 ) [ ] S E E RT S D V D e V (Suzuki & Fujii, 1982) 0 0 1/ 1 ( . ) ; ln( ); 11 n S s S n D a q Q Q aq q D KC KC (Chen & Yang, 1991) 2 2 0 1 ( ) (2 ) [ (1 )](1 )( ) 22 (1 / 2) S S H D D (Do, 1996) 0 1 1/ () (1 ) s d d s tt dd D D which to rest. As the surface concentration increases, the number of jumps increases. Higashi et al. [1963] report the following equation for the surface diffusivity 0 exp( ) (1 ) S S D E D RT 1.11 Where ( , ' ) 1 ( , ' ) b T adsorptions strenght P b T adsorptions strenght P where D S0 (m 2 .s -1 ) is the surface diffusivity of the gas at zero surface coverage (ɵ=0), ɵ is the surface coverage of monolayer ranging from 0 to 1, and E is the activation energy. R is the universal gas constant and T is the absolute temperature. This expression is the same with the one proposed by Darken (Darken, 1948) when the equilibrium relationship between the two phases was assumed to take the form of the Langmuir equation [Grant, 1998]. As can be seen by the equation above, the model of Higashi et al. [1963] fails at near monolayer coverage because as ө→1, the surface diffusivity D s →∞. Yang et al. (R. T. Yang et al., 1973) modified the theory of Higashi et al. [1963] by proposing a model which also considers second-layer adsorption (R. T. Yang et al., 1973), and obtained the following equation for surface diffusivity 17 12 ( )/ 1 0 2 1 (1 ) [ ] S E E RT S D V D e V 1.12 where V 1 and V 2 are the vibration frequencies of the vacant and occupied site, respectively, and ∆E 1 and ∆E 2 are the activation energies for surface diffusion on the first and second layers, respectively. Suzuki, 1982, investigated the surface diffusion of propionic acid in activated carbon and suggested the following equation for the surface concentration dependence of the surface diffusivity. 0 ( . ) n S S D aq D 1.13 where a and n are constants such that the isosteric heat of adsorption Q s and the isotherm q(C) are described by 0 ln( ) s Q Q aq 1.14 1/ 1 11 n q KC KC 1.15 where Q 0 is a constant. This model correlated well with their experimental data; however, the trends in the variation of surface diffusivity with respect to the amount adsorbed are different from the models of Higashi et al. [1963] and Yang et al. [1973]. As monolayer coverage is approached, these models predict a rapid increase in diffusivity, whereas the model of Suzuki and Fujii (Suzuki & Fujii, 1982) exhibits a slow increase. Chen and Yang [1991] proposed the following model (based on transition state theory) of surface diffusivity as a function of surface coverage 2 2 0 1 ( ) (2 ) [ (1 )](1 )( ) 22 (1 / 2) S S H D D 1.16 18 In the above equation, H is the Heaviside step function and λ is a parameter that indicates the degree of blockage of the diffusing molecule by other molecules on already occupied sites. 1; 0 () 0; 0 H For no blockage (λ=0), equation 1.16 reduces to the Higashi et al. [1963] equation 1.11. However, for non-zero values of λ the surface diffusivity shows different dependence on surface concentration. It increases with increasing surface concentration for small λ, whereas it decreases with increasing concentration for large λ. (ii) The hydrodynamic models are applied for high levels of adsorption, and are well developed quantitatively (E. Flood & Huber, 1955) (Gilliland et al., 1958). These models consider the surface transport as the result of a two-dimensional fluid slipping over the surface (R. J. R. Uhlhorn, 1990). According to hydrodynamic model: 1 2 2 2 2 2 p app P sg p R t p g app P R t p m A RT x F dp C S L p A RT p x C S L p 1.17 Where 12 1 2 ln m pp p p p In the above equations, F S is the surface flow rate (m 3 .s -1 ), C R is a coefficient of resistance, S t is specific surface area of the porous medium (m 2 ), ρ app is the apparent density of the porous material (kg.m -3 ), and x is the amount of adsorbed gas (mol/g). It is possible to integrate Equation 1.17 using the adsorption isotherm, and when the experimentally determined permeability is plotted against the integral, a value of C R is obtained by Lee (Lee & Hwang, 1986) for the range of low ∆P/P m values. 19 2 2 SP s P app R t m FL Q AP RT x C S p 1.18 As noted above, capillary condensation occurs when the pore space becomes filled with condensed liquid from the vapor phase (in fact, in porous materials with heterogeneous surfaces, capillary condensation may start to occur on some parts of the surface while multi layering is still occurring on others, leading to the coexistence of the multilayer-adsorbed and capillary- condensed phases in the pores (Choi et al., 2001)). The unique aspect of capillary condensation is that vapor condensation occurs below the saturation vapor pressure, P sat , of the pure liquid (Hunter, 2001). This is due to enhanced van der Waals interaction between the vapor phase molecules inside the confined capillary space (Figure 1.5). Once condensation occurs, a meniscus forms at the liquid-vapor interface which allows for equilibration of forces in between the vapor and liquid phases (see Figure 1.5). The hydrostatic pressure drop across a meniscus interface is dependent on the surface tension of the liquid and the shape of the capillary, and is described by the Young-Laplace equation below. 1.19 The Kelvin equation (equation 1.20 below) provides a relationship between the vapor pressure P t and the dimensions of the largest capillary filled with liquid in equilibrium with that vapor phase. 0 2 cos ln( ) t m P RT V P r 1.20 2 cos C P r 20 where P 0 is the saturated vapor pressure for a planar interface, ρ is the density of the condensate, σ is the interfacial tension, M is the molecular weight, r is the radius of cylindrical capillary, and ɵ is the contact angle. Figure 1.5 Condensation of gas molecules inside a capillary and formation of meniscus, known as capillary condensation a) before capillary condensation; b) after capillary condensation. (Choi et al., 2001) Barrer and co-workers (R Ash, Barrer, Clint, Dolphin, & Murray, 1973; R Ash, Barrer, & Pope, 1963; Richard Ash, Barrer, & Lowson, 1973) were among the first to observe the effect of capillary condensation experimentally when studying transport through membranes. They described a blocking effect, whereby the condensed phase in the membrane blocked the transport of the non-condensed species. There are widely different experimental results and theoretical interpretations with respect to the flow of adsorbable gases during capillary condensation (Shahraeeni & Or, 2012), and in some cases they appear to contradict each other. High separation factors are usually attained (Tsujikawa, Osawa, & Inoue, 1985) and substances as pure as 99.999% can be produced . The reason that capillary condensation results in high selectivity (R. Uhlhorn et al., 1992) is first that the multilayer diffusional flux is much 21 larger than the gas phase flux (e.g., up to 20 times larger in the case of Freon in Vycor glass). The second important reason is that due to the capillary condensation, as also noted above, the pore is blocked by the condensate, which thus prevents the transport of other components in the gas/vapor mixture. The behavior is rather complicated however. Since multilayer diffusion is much higher than liquid flow (Gilliland et al., 1958; Lee & Hwang, 1986; Tamon et al., 1981; R. J. R. Uhlhorn, 1990), the permeability will drop as soon as all pores are completely filled with liquid, see Figure 1.6. As Uhlhorn [1992] notes, the advantage of using capillary condensation as a separation mechanism is the high achievable separation factors combined with very high fluxes. Disadvantages of this mechanism is that a condensable component is necessary, so always temperature and/or pressure must be chosen in such a way, that condensation of at least one component of a mixture inside the pores of the membrane is possible. At higher temperatures this would require higher pressures. If only one single pore size existed, the separation factor would decrease sharply at one specific relative pressure. The existence of a pore size distribution in the γ-alumina film makes this decrease more gradual. Figure 1.6 A schematic plot of the permeability versus the relative pressure in case of multilayer diffusion and capillary condensation. The maximum gives the point of capillary condensation. Taken from reference (R. Uhlhorn et al., 1992) For a pore filled with condensed liquid in equilibrium on either side with a vapor above its condensation pressure but below its saturation pressure Uchytil et al. [2005] developed the following equation to describe the transport in the presence of an external pressure gradient. 22 1.21 where A is the membrane surface area, ɛ is porosity and τ indicates the tortuosity, ρ the density of condensate, and P m is calculated using equation 1.22. 1.22 Where P 1 and P 2 are the pressure at the beginning and the end of the pores respectively; this equation was developed on the assumption that Poiseuille flow prevails and that the hydrostatic pressure across the meniscus on either side of the pore is described by the Young-Laplace equation. The effective pressure drop across the capillary is given by the following equation ** 12 12 12 12 12 12 2 cos 2 cos 2 cos 2 cos ( ) [ ] ( )(1 ) m P P P PP rr PP rr RT PP MP 1.23 From the above equation one can estimate that the effective pressure drop ΔP driving the liquid flow across the pore is one to two orders of magnitude greater than the pressure drop in the vapor phase (P 1 −P 2 ). The different modes of vapor permeation through a cylindrical capillary have been discussed by Rhim and Hwang (Rhim & Hwang, 1975) and by Lee and Hwang (Lee & Hwang, 1986). They identified several different modes of transport (shown schematically in Figure 1.7) depending on the upstream and downstream vapor pressures and the saturation pressure for the vapor at the conditions of the experiment. These modes of transport and the continuum-type mathematical equations that describe them are discussed further in Chapter 4. A true porous 2 12 ( )(1 ) 8 m r RT F A P P Mz MP 12 2 m PP P 23 medium, of course, is described by many pores and all the different vapor permeation modes may occur simultaneously. When considering the porous medium to consist of a bundle of parallel pores following a pore size distribution the total flow rate is described by the following equation: 0 () tr F f r F dr 1.24 where F r is the flow rate (m 3 .s -1 ) for a single pore of radius r, and f(r) is the pore size distribution function (a function showing pore size distribution). Of course, the “bundle of parallel pores” model is a substantial oversimplification of the true nature of porous materials including the microporous membranes under study here; as part of our future research, therefore, we propose to utilize more realistic network models to study vapor permeation of condensable gases, including DMMP, through such membranes, see further discussion in Chapter 4. Figure 1.7 six modes (F 1 -F 6 ) flow of condensable vapor in micropores (Lee & Hwang, 1986) 24 Table 1.5 typical works done on gas condensation phenomena Gas (s) Comments Reference (s) N 2 / CH 4 Elkamel et al. used a Local Density Approximation (LDA) for estimating the condensation; then they concluded with a better approximation of this model for selectivity and flux compared to that of Kelvin equation. (Elkamel & Noble, 1992) C 2 H 6 , n-C 4 H 10 , and CO 2 through Vycor glass Rhim and Hwang using domain theory proposed six modes for flow of gases; experimentally they reached a maximum of permeability at lower temperatures by changing the relative pressure for these gases; but at higher temperature, the change was low. Their models are capable to predict a monotonic increase of permeability as well. (Rhim & Hwang, 1975) Freon-113 and H,O through Vycor glass Lee and Hwang later modified their models by using as basis the Kelvin equation. They fitted the theory with experimental data very well. Their results depended on the pressure of both sides of the pore. With correction of the Kelvin equation to account for the thickness of adsorbed layer, their result fits well with the experimental data. (Lee & Hwang, 1986) C 3 H 6 /N 2 Ulhorn et al. observed that the permeation of C 3 H 6 through γ- Al 2 O 3 membrane (25 nm) is higher than that of propane. (R. J. R. Uhlhorn, 1990) H 2 /CO 2 Ulhorn et al. modified the surface of alumina with silver and measured the surface flux of carbon dioxide; they observed a reduction in surface flux by increasing the magnesia as a surface modifier. (Keizer et al., 1988) Light alcohols from water The flow of condensate through a silica/alumina membrane (3 nm pores) was investigated. (ASAEDA & DU, 1986) CH 2 OH/H 2 The H 2 permeability decreased by three orders of magnitude when capillary condensation of CH 2 OH occurred The CH 2 OH permeation rate actually increased slightly after capillary condensation occurred Measurements were taken up to 473 K; an alumina membrane with approximately 2.5nm diameter pores was utilized. (Sperry, Falconer, & Noble, 1987) Propylene on supported γ- alumina films Multilayer diffusion in the range of (P/P 0 =0.4-0.8) strongly increased the permeability to 6 times the Knudsen permeability value, yielding permeabilities of 3.2*10 -5 mol/(m 2 .sec.pa). The occurrence of a maximum in the permeability coincides with blocking of the pore by adsorbate (capillary condensation). Separation factors of N 2 -C 3 H 6 = 27 through supported γ- alumina films, with C 3 H 6 the preferentially permeating component. S.F= 85 after modification of the system with magnesia by the reservoir method. However, the permeability of propylene decreased by a factor of 20 to 1.6*10 -6 mol/m 2 .sec.pa. (R. Uhlhorn et al., 1992) Alcohol/water Asaeda used a mixture of alcohol/water and observed the separation factor S.F=60 with water being the preferentially permeating component at temperature 90-100°C. (ASAEDA & DU, 1986) Nitrogen For cylindrical pores (0.01 mm) at low temperature, Sidhu and Cussler, observed a combination of Knudsen diffusion and capillary flow. (Sidhu & Cussler, 2001) 25 For the transport of non-absorbable gases through small pore-size membranes through which Knudsen transport prevails the permeability does not change over a range of pressures. But permeability of condensable gases through the same membranes shows non-monotonic behavior, passing through a maximum with the change in pressure, as noted previously. The experimental data with such systems are much more scattered than those of non-condensable gases and in addition to the non-monotonic behavior described above they show other types of interesting behavior, including hysteresis phenomena. Table 1.5 summarizes some of the key studies on vapor permeation through membranes. Hexane and N 2 The effect of pore morphology on capillary condensation and evaporation in nanoporous silicon was studied. Condensation in an isolated pore (cavity or constriction) with a high amount of disorder is nucleated at a lower pressure than that predicted by the Kelvin equation because of local pore heterogeneities that catalyze condensation. (Casanova , Chiang, Ruminski, Sailor, & Schuller, 2012) Water Theoretical and experimental investigation of the transport mechanisms leading to the formation of water menisci in thermodynamic equilibrium with the water vapors in air at a single-asperity nanocontact. It was observed that an instantaneous formation of a water meniscus by coalescence of the water layers adsorbed on the AFM tip and sample surfaces, followed by a time evolution of meniscus toward a stationary state corresponding to thermodynamic equilibrium. (Sirghi, 2012) Nitrogen Condensation of N 2 on multiwall carbon nanotubes (Inoue, Ichikuni, Suzuki, Uematsu, & Kaneko, 1998) He, N 2 , Ar, CO 2 and C 3 H 8 Using glass membrane with narrow pores (2.3-4.2 nm), Markovi et al measured the permeability for the aforementioned gases; they observed Knudsen diffusion and viscous flow for gas phase flow plus surface diffusion when an adsorbed phase is involved. (Marković , Stoltenber g, Enke, Schlünder, & Seidel- Morgenste rn, 2009) n-C 4 H 10 and CO 2 Sing et al. modified a butane selective glass membrane with an organic compound (organosilane) and created a carbon dioxide selective membrane. (Singh, Way, & McCarley, 2004) 26 1.2.4 Capillary Condensation from Mixed Vapors So far, more attention has been devoted to the condensation of a specific gas from a mixture of non-condensable vapors (A. Adamson, 1997; Gelb, Gubbins, Radhakrishnan, & Sliwinska- Bartkowiak, 1999; Gregg & Sing, 1983) compared to the mixtures of condensable vapors. In a mixture of condensable gases, the adsorbability of different gases is different due to their physical properties; which raises the complexity of such systems. Using the basics of the Kelvin equation for one components, the partial pressure of each component in binary components is derived using the Equation 1.25, , ,0 () () ln ; 1,2 () m j cond i cond i cond VX pX i p X r RT 1.25 where r indicates the mean radius of curvature of the meniscus, X cond shows the composition of the condensate, p i is the partial pressure of component i, p i,0 (X cond ) is the partial pressure of component i over a bulk solution (with planar interface) of composition X cond ; and V m,i (X cond ) and γ(X cond ) are, respectively, the composition-dependent partial molar volume of component i in liquid phase and surface tension of the liquid mixture. Shapiro [1997] used the Kelvin equation for binary mixtures to model a mixture of hydrocarbons in oil-gas reservoirs. The separation of liquid mixtures using capillary distillation is studied by Yeh, Abu and Shin (R. Yang, Chen, & Yeh, 1991) following the work of Shapiro (Shapiro & Stenby, 2001). Equation 1.25 is based on the chemical potential of each component and never predicts the priority to adsorption of the components. Laaksonen and Kipling (Kipling, 1965; Laaksonen & Stilbs, 1991), for example, also studied the selectivity but they also fail to predict the condensation of components. 27 References Adamson, A. (1997). Adsorption of gases and vapors on solids. Physical Chemistry of Surfaces. 6th ed. New York, NY: John Wiley and Sons, 605. Adamson, A. W., & Gast, A. P. (1967). Physical chemistry of surfaces. ASAEDA, M., & DU, L. D. (1986). Separation of alcohol/water gaseous mixtures by thin ceramic membrane. Journal of Chemical Engineering of Japan, 19(1), 72-77. Ash, R., Barrer, R., Clint, J., Dolphin, R., & Murray, C. (1973). Isothermal and thermo-osmotic transport of sorbable gases in microporous carbon membranes. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 275(1249), 255-307. Ash, R., Barrer, R., & Pope, C. (1963). Flow of adsorbable gases and vapours in a microporous medium. II. Binary mixtures. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 271(1344), 19-33. Ash, R., Barrer, R. M., & Lowson, R. T. (1973). Transport of single gases and of binary gas mixtures in a microporous carbon membrane. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 69, 2166-2178. Bose, T., Chahine, R., Marchildon, L., & St ‐Arnaud, J. (1987). New dielectric method for the measurement of physical adsorption of gases at high pressure. Review of scientific instruments, 58(12), 2279-2283. Bouwmeester, H., Burgraaf, A., Burgraaf, A., & Cot, L. Fundamentals of Inorganic Membrane Science and Technology, 1996, 435: Elsevier, Amsterdam. Braunauer, S. (1945). The adsorption of gases and vapours. The Adsorption of Gases and Vapours. Brunauer, S. (1943). Adsorption of gases and vapors. Brunauer, S. (1945). The Absorption of Gases and Vapors (Vol. 1): Princeton University Press. Butt, J. B., & Reed Jr, E. (1971). Surface diffusion of single sorbates at low and intermediate surface coverage. The Journal of Physical Chemistry, 75(1), 133-141. Carman, P., & Raal, F. (1951). Diffusion and flow of gases and vapours through micropores. III. Surface diffusion coefficients and activation energies. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 209(1096), 38-58. 28 Casanova, F. l., Chiang, C. E., Ruminski, A. M., Sailor, M. J., & Schuller, I. K. (2012). Controlling the role of nanopore morphology in capillary condensation. Langmuir, 28(17), 6832-6838. Chen, Y., & Yang, R. (1991). Concentration dependence of surface diffusion and zeolitic diffusion. AIChE Journal, 37(10), 1579-1582. Choi, J.-G., Do, D., & Do, H. (2001). Surface diffusion of adsorbed molecules in porous media: Monolayer, multilayer, and capillary condensation regimes. Industrial & Engineering Chemistry Research, 40(19), 4005-4031. Crank, J. (1975). The mathematics of diffusion. Darken, L. S. (1948). Diffusion, mobility and their interrelation through free energy in binary metallic systems. Trans. Aime, 175(184), 41. Do, D. D. (1996). A model for surface diffusion of ethane and propane in activated carbon. Chemical Engineering Science, 51(17), 4145-4158. Donohue, M., & Aranovich, G. (1998). Classification of Gibbs adsorption isotherms. Advances in colloid and interface science, 76, 137-152. Donohue, M., & Aranovich, G. (1999). A new classification of isotherms for Gibbs adsorption of gases on solids. Fluid phase equilibria, 158, 557-563. Elkamel, A., & Noble, R. D. (1992). A statistical mechanics approach to the separation of methane and nitrogen using capillary condensation in a microporous membrane. Journal of Membrane Science, 65(1), 163-172. Findenegg, G., Korner, B., Fischer, J., & Bohn, M. (1983). Supercritical Gas Adsorption in Porous Materials, I. Storage of Krypton in Carbon Molecular Sieves. Ger. Chem. Eng, 6, 80-84. Flood, E., & Huber, M. (1955). THERMODYNAMIC CONSIDERATIONS OF SURFACE REGIONS: ADSORBATE PRESSURES, ADSORBATE MOBILITY, AND SURFACE TENSION. Canadian Journal of Chemistry, 33(2), 203-214. Flood, E. A. (1967). The solid-gas interface (Vol. 2): M. Dekker New York. Gad-el-Hak, M. (2001). The MEMS handbook: CRC press. Gelb, L. D., Gubbins, K., Radhakrishnan, R., & Sliwinska-Bartkowiak, M. (1999). Phase separation in confined systems. Reports on Progress in Physics, 62(12), 1573. Gilliland, E., Baddour, R., & Russell, J. (1958). Rates of flow through microporous solids. AIChE Journal, 4(1), 90-96. 29 Grandcolas, M., Louvet, A., Keller, N., & Keller, V. (2009). Layer ‐by ‐Layer Deposited Titanate ‐Based Nanotubes for Solar Photocatalytic Removal of Chemical Warfare Agents from Textiles. Angewandte Chemie, 121(1), 167-170. Gregg, S., & Sing, K. S. (1983). Adsorption, Surface Area, and Porosity. Higashi, K., Ito, H., & Oishi, J. (1964). Surface Diffusion Phenomena in Gaseous Diffusion,(II) Separation of Binary Gas-mixtures. Journal of Nuclear Science and Technology, 1(8), 298-304. Hill, T. L. (1960). Statiscal-Thermodynamics: Addison-Wesley. Holt, J. K., Park, H. G., Wang, Y., Stadermann, M., Artyukhin, A. B., Grigoropoulos, C. P., . . . Bakajin, O. (2006). Fast mass transport through sub-2-nanometer carbon nanotubes. Science, 312(5776), 1034-1037. Hunter, R. J. (2001). Foundations of colloid science. Inoue, S., Ichikuni, N., Suzuki, T., Uematsu, T., & Kaneko, K. (1998). Capillary condensation of N2 on multiwall carbon nanotubes. The Journal of Physical Chemistry B, 102(24), 4689- 4692. Keizer, K., Uhlhorn, R., & Burggraaf, A. (1988). Gas separation mechanisms in microporous modified γ-al< sub> 2</sub> o< sub> 3</sub> membranes. Journal of Membrane Science, 39(3), 285-300. Kipling, J. J. (1965). Adsorption from Solutions of Non-electrolytes. Knudsen, M. (1909). Effusion and the molecular flow of gases through openings. Ann. Phys, 29, 179. Laaksonen, A., & Stilbs, P. (1991). Molecular dynamics and NMR study of methane-water systems. Molecular Physics, 74(4), 747-764. Lee, K.-H., & Hwang, S.-T. (1986). The transport of condensible vapors through a microporous Vycor glass membrane. Journal of Colloid and Interface Science, 110(2), 544-555. Lin, H., & Freeman, B. D. (2004). Gas solubility, diffusivity and permeability in poly (ethylene oxide). Journal of Membrane Science, 239(1), 105-117. Marković, A., Stoltenberg, D., Enke, D., Schlünder, E.-U., & Seidel-Morgenstern, A. (2009). Gas permeation through porous glass membranes: Part I. Mesoporous glasses—effect of pore diameter and surface properties. Journal of Membrane Science, 336(1), 17-31. 30 Mera, N., Hirakawa, T., Sano, T., Takeuchi, K., Seto, Y., & Negishi, N. (2010). Removal of high concentration dimethyl methylphosphonate in the gas phase by repeated-batch reactions using TiO< sub> 2</sub>. Journal of hazardous materials, 177(1), 274-280. Michels, A., Menon, P., & Ten Seldam, C. (1961). Adsorption of nitrogen on alumina at high pressure. Recueil des Travaux Chimiques des Pays-Bas, 80(5), 483-501. MOFFAT, D. H., & Weale, K. (1955). Sorption by coal of methane at high pressures. Fuel, 34(4), 449-462. Monji, M. 2014 Annual Meeting October 28-November 2, 2012. Monji, M., & Tsotsis, T. 2013 North American Membrane Society Meeting October 28- November 2, 2012. Motamedhashemi, M. Y., Monji, M., Egolfopoulos, F., & Tsotsis, T. (2014). A Hybrid catalytic membrane reactor for destruction of A chemical warfare simulant. Journal of Membrane Science. Naito, Y., Kamiya, Y., Terada, K., Mizoguchi, K., & Wang, J. S. (1996). Pressure dependence of gas permeability in a rubbery polymer. Journal of Applied Polymer Science, 61(6), 945- 950. Payne, H., Sturdevant, G., & Leland, T. (1968). Improved two-dimensional equation of state to predict adsorption of pure and mixed hydrocarbons. Industrial & Engineering Chemistry Fundamentals, 7(3), 363-374. Pinnau, I., & Toy, L. G. (1996). Gas and vapor transport properties of amorphous perfluorinated copolymer membranes based on 2, 2-bistrifluoromethyl-4, 5-difluoro-1, 3- dioxole/tetrafluoroethylene. Journal of Membrane Science, 109(1), 125-133. Quenneville, J., Taylor, R. S., & van Duin, A. C. (2010). Reactive molecular dynamics studies of DMMP adsorption and reactivity on amorphous silica surfaces. The Journal of Physical Chemistry C, 114(44), 18894-18902. Rhim, H., & Hwang, S.-T. (1975). Transport of capillary condensate. Journal of Colloid and Interface Science, 52(1), 174-181. Schneider, P., & Smith, J. (1968). Adsorption rate constants from chromatography. AIChE Journal, 14(5), 762-771. Segal, S. R., Cao, L., Suib, S. L., Tang, X., & Satyapal, S. (2001). Thermal decomposition of dimethyl methylphosphonate over manganese oxide catalysts. Journal of Catalysis, 198(1), 66-76. 31 Shahraeeni, E., & Or, D. (2012). Pore scale mechanisms for enhanced vapor transport through partially saturated porous media. Water Resources Research, 48(5). Shapiro, A. A., & Stenby, E. H. (2001). Thermodynamics of the multicomponent vapor–liquid equilibrium under capillary pressure difference. Fluid phase equilibria, 178(1), 17-32. Shiflett, M. B., & Foley, H. C. (1999). Ultrasonic deposition of high-selectivity nanoporous carbon membranes. Science, 285(5435), 1902-1905. Sidhu, P. S., & Cussler, E. (2001). Diffusion and capillary flow in track-etched membranes. Journal of Membrane Science, 182(1), 91-101. Sing, K. S. (1994). Physisorption of gases by carbon blacks. Carbon, 32(7), 1311-1317. Singh, R. P., Way, J. D., & McCarley, K. C. (2004). Development of a model surface flow membrane by modification of porous vycor glass with a fluorosilane. Industrial & Engineering Chemistry Research, 43(12), 3033-3040. Sirghi, L. (2012). Transport Mechanisms in Capillary Condensation of Water at a Single- Asperity Nanoscopic Contact. Langmuir, 28(5), 2558-2566. Sperry, D. P., Falconer, J. L., & Noble, R. D. (1987). Methanol—hydrogen separation by capillary condensation in inorganic membranes. Journal of Membrane Science, 60(2), 185-193. Stern, S., Fang, S. M., & Frisch, H. (1972). Effect of pressure on gas permeability coefficients. A new application of “free volume” theory. Journal of Polymer Science Part A ‐2: Polymer Physics, 10(2), 201-219. Suda, H., & Haraya, K. (1995). Molecular sieving effect of carbonized Kapton polyimide membrane. J. Chem. Soc., Chem. Commun.(11), 1179-1180. Suzuki, M., & Fujii, T. (1982). Concentration dependence of surface diffusion coefficient of propionic acid in activated carbon particles. AIChE Journal, 28(3), 380-385. Tamon, H., Okazaki, M., & Toei, R. (1981). Flow mechanism of adsorbate through porous media in presence of capillary condensation. AIChE Journal, 27(2), 271-277. Trubitsyn, D. A., & Vorontsov, A. V. (2005). Experimental study of dimethyl methylphosphonate decomposition over anatase TiO2. The Journal of Physical Chemistry B, 109(46), 21884-21892. Tsotsis, T. T., Egolfopoulos, F., Nair, N., Prosser, R., Ren, J.-Y., Liu, P., . . . Monji, M. (2014). Catalytic removal of gas phase contaminants: Google Patents. 32 Tsujikawa, H., Osawa, T., & Inoue, H. (1985). Separation of benzene and nitrogen by permeation through porous Vycor glass. Kagaku Kogaku Ronbunshu, 11(5), 534-541. Tzevelekos, K., Kikkinides, E., Stubos, A., Kainourgiakis, M., & Kanellopoulos, N. (1998). On the possibility of characterising mesoporous materials by permeability measurements of condensable vapours: theory and experiments. Advances in colloid and interface science, 76, 373-388. Tzevelekos, K., Romanos, G., Kikkinides, E., Kanellopoulos, N., & Kaselouri, V. (1999). Experimental investigation on separations of condensable from non-condensable vapors using mesoporous membranes. Microporous and Mesoporous Materials, 31(1), 151-162. Uhlhorn, R., Keizer, K., & Burggraaf, A. (1992). Gas transport and separation with ceramic membranes. Part I. Multilayer diffusion and capillary condensation. Journal of Membrane Science, 66(2), 259-269. Uhlhorn, R. J. R. (1990). Ceramic membranes for gas separation: synthesis and transport properties: Universiteit Twente. Vidal, D., Malbrunot, P., Guengant, L., Vermesse, J., Bose, T., & Chahine, R. (1990). Measurement of physical adsorption of gases at high pressure. Review of scientific instruments, 61(4), 1314-1318. Wakasugi, Y., Ozawa, S., & Ogino, Y. (1981). Physical adsorption of gases at high pressure: V. An extension of a generalized adsorption equation to systems with polar adsorbents. Journal of Colloid and Interface Science, 79(2), 399-409. Yang, R., Chen, Y., & Yeh, Y. (1991). Prediction of cross-term coefficients in binary diffusion: diffusion in zeolite. Chemical Engineering Science, 46(12), 3089-3099. Yang, R. T., Fenn, J. B., & Haller, G. L. (1973). Modification to the Higashi model for surface diffusion. AIChE Journal, 19(5), 1052-1053. 33 Chapter 2: Carbon Molecular Sieve (CMS) Membranes 2.1 Introduction 1 Membrane separation can be a very efficient and economical way of separating mixtures of gases, and as a result the field of membrane-based gas separations, has experienced good progress during the last two decades. The membrane is a physical barrier (Figure 2.1) that allows certain compounds in the mixture to pass through, while preventing others from doing so, depending on their physical and/or chemical properties. Membranes are typically asymmetric and commonly consist of a porous support layer with a thin dense or nanoporous layer on the top that forms the actual membrane. The permeability and ideal selectivity and/or the separation factor are the most important characteristics of a membrane. For porous membranes, they are typically determined by the thickness, average pore size and surface porosity of the membrane, while for a dense membranes, the mechanism for permeation and separation is more complex (Li, 2007). Table 2.1 shows a classification of asymmetric ceramic membranes and their various uses depending on their structure. Depending on the material they are made from, membranes are categorized as polymeric and a : IUPAC classification 1 The discussion and some of the data of this chapter are taken from an already published paper Lee, H.-C., Monji, M., Parsley, D., Sahimi, M., Liu, P., Egolfopoulos, F., & Tsotsis, T. (2012). Use of steam activation as a post-treatment technique in the preparation of carbon molecular sieve membranes. Industrial & Engineering Chemistry Research, 52(3), 1122- 1132. Table 2.1 Types of membranes (Li, 2007) Type a Pore size (nm) Mechanism Application Macroporous > 50 Sieving US, MF Mesoporous 2-50 Knudsen Diffusion US, NF, Gas Separation Microporous < 2 Micropore Diffusion Gas Separation Dense - Diffusion Gas Separation, Reaction 34 ceramic membranes. Currently, most of the commercial membrane-based gas separation applications use polymeric membranes (Lin, Van Wagner, Freeman, Toy, & Gupta, 2006). These are easily fabricated in various configurations (e.g., flat sheet, hollow fiber, etc.) and have a modest cost. On the other hand, polymeric membranes have, in general, modest separation characteristics (dictated by permeability vs. selectivity trade-off relationship (Anshu Singh & Koros, 1996) known as the Robeson plot) and are not, typically, intended for use under high temperatures and pressures. Different types of materials have, so far, been used to prepare ceramic membranes such as Al 2 O 3 , TiO 2 , ZrO 2 , SiO 2 , etc. Compared to polymeric membranes, ceramic membranes have the advantage of being able to function at high temperatures, and pressures and in aggressive environments. Ceramic membranes also offer an additional advantage that, when subjected to fouling, they can recover their properties by treatment at high temperatures, for example, via steam treatment which is also used for sterilizing such membranes for application in food processing and for medical applications (Julbe, Farrusseng, & Guizard, 2001). Ceramic membranes, in particular, consist of several layers of one or several different types of ceramic materials with different pore size. The first layer is a macroporous layer, coated with one or two mesoporous layers. The last (top-most) layer usually is the permselective layer and is either dense or microporous. As shown in Figure 2.2, the bottom layer provides mechanical support, while the middle layers bridge the pore size difference between the support layer and the top layer where the actual separation takes place (Mikuláŝek, Doleček, Šedá, & Cakl, 1994). The focus of this research is supported carbon molecular sieve membranes. These are made by coating a thin polymer layer on a ceramic substrate, and by pyrolyzing it at high 35 temperatures in a controlled atmosphere (e.g., vacuum or inert gas), to convert it into a nanoporous carbon layer. Figure 2. 1 A schematic representation of membrane separation mechanism CMS membranes exhibit separation performance which often lies above the Robeson plot of competitive polymeric membranes and, in addition, show resistance to high temperatures and pressures. During pyrolysis to prepare CMS membranes, most of the heteroatoms present in the precursor polymeric macromolecules are progressively removed, with only a cross-linked and stiff, primarily carbon skeleton structure remaining. (Lagorsse, Magalhaes, & Mendes, 2004). The pore structure of the CMS membranes is thought to be non-homogeneous and to consist (J. Koresh & Soffer, 1980) of both larger-pore regions, with sizes in the range of 10−20 Å (which explains the large gas permeation rates of such materials), separated by nanoporous constrictions (a few Å in size) that are, principally, responsible for their molecular sieving characteristics (Centeno & Fuertes, 1999). It is such a hybrid pore structure that explains the ability of CMS membranes to perform molecular sieving type separations while still maintaining the high-flux character of carbon materials. Our group and others have produced high-quality CMS membranes in the past few years. 36 Ismail and David (Ismail & David, 2001) reviewed some of the earlier work, and Table 2.2 lists some of the key studies. The development of CMS membranes with the appropriate pore size characteristics has been accomplished mainly by choosing the appropriate polymeric precursor and/or by varying the carbonization conditions (e.g., the atmosphere, temperature and duration of pyrolysis). Often, however, the technique does not succeed [9] to consistently provide the tight pore size control required for the separation of important gas pairs (e.g., O 2 /N 2 and CO 2 /CH 4 ), and thus, additional post-treatment steps such as steam activation (see further discussion to follow) are needed to fine-tune the pore structure of the CMS membranes. Table 2.2 Select papers on CMS membranes Precursor Configuration Pyrolysis Temperature ( o C) Pyrolysis Atmosphere Permeation Properties Ref. Acrylonitrile Hollow fiber 600 - 1200 N 2 Not reported (Nishihara & Yoneyama, 1992) PAN/PMMA Hollow fiber 600 and 950 N 2 Not reported (Linkov, Sanderson, & Jacobs, 1994) Cellulose Hollow fiber supported film 500-800 Ar I.S (He/O 2 )=2.8 He * =1.78*10 -4 O 2 *= 6.24*10 -5 (J. E. Koresh, Saggy, & Soffer, 1987) Cellulose Hollow fiber 400 - 1200 Ar N 2 ** =0.508 He ** =0.726 I.S (He/N 2 )=1.43 Data at 100 °C (Gilron & Soffer, 2002) Kapton polyimide Film 1000 Vacuum I.S (He/Ar)=15.7 Ar * =9.3*10 -8 He * =1.2*10 -6 (Suda & Haraya, 1997) Phenolic resin Supported film 900 N 2 H 2 ** =3.086*10 -3 MeOH ** =4.049*10 -4 Propane ** =3.283*10 -4 Conditions: T=100 °C, ΔP=1.72 bar (Clint, Lear, Oliver, & Tennison, 1992) Phenolic resin Supported film 500-1000 Vacuum He ** =0.064 at 25 o C N 2 ** =8.064 *10 -4 at 25 o C O 2 ** =1.2 *10 -2 at 25 o C (Centeno & Fuertes, 1999) PVDC-AC Supported film 600 N 2 Not reported (Thaeron et al., 1999) 37 PEI Supported film 800 Vacuum He*=1.505 *10-7 O2*=8.208*10-8 N2*=1.505*10-9 At 298 K (Fuertes & Centeno, 1998) PEI Supported film 600-750 Ar I.S (H 2 /Ar)=19 H 2 ** =1.67 Ar ** =0.09 at 393 K (Sedigh, Jahangiri, Liu, Sahimi, & Tsotsis, 2000; Sedigh, Xu, Tsotsis, & Sahimi, 1999) PFA Supported film 450 He I.S (O 2 /N 2 )=7.4 (Shiflett & Foley, 2001) Sulfonated phenolic resin Supported film 250-800 N 2 H 2 ** =5.26 CO 2 ** =2.16 O 2 ** =0.65 I.S (H 2 /CH 4 )=65 I.S (CO 2 /CH 4 )=27 I.S (O 2 /N 2 )= 5.2 (Chen & Yang, 1994) PFA Macroporous graphite disc 500 H 2 No permeability or permeance data reported (Chen & Yang, 1994) Phenolic formaldehyde resin (PFR) Thin film 800-950 N 2 I.S (H 2 / O 2 )=2.2 I.S (H 2 / N 2 )=23.6 H 2 * =1.4*10 -5 O 2 * =6.29*10 -6 N 2 * =5.91*10 -7 (Shusen, Meiyun, & Zhizhong, 1996) PFR Tubular shape 800 Air H 2 ** =1.094*10 -2 H 2 /N 2 =2.45 (Wei, Qin, Hu, You, & Chen, 2007) 6FDA/BPDA Flat sheet 550 Vacuum inert I.S (O 2 /N 2 )= 6.4-7.5 O 2 * =(1.47-1.72) *10 -6 (Kiyono, Williams, & Koros, 2010) Poly(phthalazinone ether sulfone ketone) (PPESK) Flat sheet 460 Air I.S(H 2 /N 2 )= 41.4-73.8 (Wang et al., 2009) Sulfonated poly(aryl ether ketone) (SPAEK) ion-exchanged with different counter ions (H+, Na+ and Ag+) Flat sheet 800 Vacuum N 2 * =5.33*10 -11 - 7.6*10 -9 H 2 * =9.30 *10 -9 - 1.69*10 -6 (Xiao et al., 2010) Polyimide/Cs 2 CO 3 Supported on porous alumina 600 N 2 CO 2 ** = 8.75*10 -3 I.S (CO 2 /N 2 )= 11 (Kai, Kazama, & Fujioka, 2009) 38 PI/PVP Flat sheet 550-700 Ar I.S (O 2 /N 2 )= 10 (Kim, Park, & Lee, 2004) PI/PVP Flat sheet 700 - O 2 * =(1.532-2.216)*10 - -6 I.S (O 2 /N 2 )=7-10 (Kim, Park, & Lee, 2005) Phenol formaldehyde novolac resin (PFNR) containing CMS. Tubular shape 800 Inert CO 2 H 2 ** =0.139 O 2 ** =5.76*10 -2 N 2 ** =6.65*10 -2 (Zhang, Hu, Zhu, & Zhu, 2006) * Permeability, m 3 *m/(m 2 *bar*h); ** Permeance, m 3 /(m 2 *bar*h); I.S indicates ideal selectivity 2.2 CMS Membrane Studies 2.2.1 Membrane Preparation Separation of condensable vapors (in this research DMMP) using the surface flow technique requires membranes with narrow pore size distribution with pores in the range of a few Å (a bit larger than the diameter of the DMMP molecule). This requires being able during their preparation to tailor-make the properties of such membranes, and in some instances to modify their properties by additional modification steps. The methods by which we prepare and characterize such membranes are described below. We mostly use as substrates, during preparation, mesoporous ceramic membranes with an average pore size of 40 and 100 Å available to us by our industrial collaborators in this research project, Media and Process Technology, Inc. (M&PT). Figure 2.2 Schematic representation of an asymmetric composite membrane, 1 is polymer layer coated on the substrate, 2 is the active layer, 3 is the intermediate layer and, 4 is porous substrate (Briceño, Garcia ‐Valls, & Montané, 2010) 39 There are five different types of ceramic membranes that are provided for our research by our industrial collaborator (M&P) whose characteristics are described in Tables 2.3 and 2.4. Four of these are ceramic membranes which in our group we use as substrates to prepare catalytically active materials for our catalytic oxidation (FTCMR) studies. Two of these tubular membranes (B and C) are used as substrates in our research when we prepare CMS membranes. M&P researchers themselves are preparing CMS membranes which we also use in our SFM studies in tandem with the membranes we prepare ourselves. All of the ceramic membranes, other than membrane type S, can be categorized as “mesoporous” materials. For these, the macroporous substrate (Figure 2.2) acts as a mechanical support for the intermediate layer in order to provide strength to the final membrane. The intermediate layer serves the purpose to mask the surface irregularities of the substrate, so that an ultra-thin, high-flux surface layer can be deposited on it. In the following sections we will discuss deposition of additional layers on the membrane top layer in order to prepare CMS permselective membranes. Table 2.3 Material used in different layers of M&PT ceramic membranes used in this research, as reported by the vendor Membrane module Support Material First Separation Layer Second Separation Layer Third separation Layer S-type α-Al 2 O 3 - - - A-type α-Al 2 O 3 α-Al 2 O 3 - - B-type α-Al 2 O 3 α-Al 2 O 3 γ-Al 2 O 3 - C-type α-Al 2 O 3 α-Al 2 O 3 γ-Al 2 O 3 - CMS α-Al 2 O 3 α-Al 2 O 3 γ-Al 2 O 3 γ Carbon Layer Our group has been preparing CMS membranes for over ten years (Abdollahi, Yu, Liu, et al., 2010; Liu, Sahimi, & Tsotsis, 2012; Sedigh et al., 2000; Sedigh et al., 1998; Sedigh et al., 1999) 40 through the controlled-atmosphere carbonization of a variety of polymeric precursors. The resulting membranes have good separation characteristics and excellent resistance to high- temperature and pressure environments (Abdollahi, Yu, Liu, et al., 2010; Lee et al., 2012; Liu et al., 2012). In this research, polyetherimide (PEI) has been used as a polymeric precursor in the preparation of the CMS membranes, which our group has also previously used to prepare such membranes for fundamental studies of transport and separation of gas mixtures (Lee et al., 2012; Sedigh et al., 2000; Sedigh et al., 1998; Sedigh et al., 1999). Table 2.4 Some physical properties of M&P ceramic membranes used in this research, as reported by the vendor Membrane module Support Thickness (µm)/ Pore diameter (Å) / Porosity (%) First Separation Layer Thickness (µm)/ Pore diameter (Å) / Porosity (%) Second Separation Layer Thickness (µm)/ Pore diameter (Å) / Porosity (%) S-type 1100 /2000-4000 /20-25 - - A-type 1100 /2000-4000 /20-25 10 – 20 /500 /N.A - B-type 1100 /2000-4000 /20-25 10 – 20 /500 /N.A 2 – 3 µm /100 Å/25-35 C-type 1100 /2000-4000 /20-25 10 – 20 /500 /N.A 2 – 3 µm /40 Å/N.A The PEI used in this work is Ultem1000, supplied by General Electric, and its structure is shown in Figure 2.3. It is a good model precursor to use as it is inexpensive, easily available, and dissolves readily in a variety of common polar solvents, a key requirement for the preparation of supported CMS membranes. Figure 2.3 Chemical structure of PEI 41 It prepares, in addition, CMS membranes with good permeation characteristics (high throughput and selectivity) and thermal stability that compare favorably with membranes prepared by much more expensive and difficult to process polymeric precursors. To prepare the supported CMS membranes in this study, as noted above, we use ceramic M&P tubular supports (3.5 mm ID and 6 mm OD). Prior to coating with the polymeric precursor (see discussion below), the ends (1 cm each end, see Figure 2.4) of each substrate were glazed (using a Duncan GL Ultraclear glaze) to ensure that they are impermeable to gas transport. The quality of the resulting substrates was validated by permeation tests with He and Ar. Only substrates that exhibited good performance, with He/Ar ideal selectivities close to the Knudsen value, were selected for the preparation of CMS membranes. To prepare the coating solution, the PEI resin was dissolved in 1, 2- dicholoroethane (DCE) while continuously stirring for 24 h, with a slight heating applied in the early stages of dissolution. Dip-coating solutions of different concentrations were prepared, with the more concentrated solution applied in the initial coatings, and the less concentrated solutions in the subsequent coatings (for most of the membranes in this paper use was made of two different solutions with PEI concentrations of 6 and 2 wt %, respectively). Figure 2.4 samples of substrate (right) and membrane (left). 42 For the substrates utilized here, the γ-alumina layer is deposited on the inside surface of the tube. To prepare CMS membranes using these substrates, their outer surface was first wrapped with Teflon tape. The substrate was then immersed in the 6 wt % PEI/DCE solution for 3 min, and was subsequently withdrawn (“pulled-out”) from the solution at a constant rate of 2 cm/min, in order to coat a PEI layer with a uniform thickness on the inside (γ-alumina) surface of the tube. After coating of the PEI film, the membrane was stored in an incubator to dry for 24 h. The membrane was subsequently carbonized in a cylindrical furnace in flowing ultrahigh pure (UHP) Ar, in order to remove all the gases evolved during the pyrolysis process. The carbonization protocol involved first heating (1 °C/min) the membrane to 350 °C and holding it at that temperature for 1 h. The temperature was then raised (1 °C/min) to 650 °C, and the membrane was held there for an additional 4 h. Subsequently, the membrane was cooled down to 180 °C (at 2 °C/min) and further down to room temperature at a rate of 5 °C/min. The coating/carbonization procedure was then repeated a number of times (typically 3− 4), in order to achieve the desired membrane performance (for successive coatings, as noted earlier, the 2 wt % PEI/DCE solution was utilized). The resulting CMS membranes are then ready to be subjected to additional post- treatment via steam activation. 2.2.2 Transport Measurements Using the permeation measurement apparatus shown in Figure 2.5, the transport characteristics of the membranes were determined by measuring the flux through the membrane of a number of test gases, namely He and Ar. In the experiments, ultra high pure gases with purity of 99.999% were utilized, and purifier columns of dried molecular sieves were used, in addition, in order to remove any water vapor, hydrocarbons and other impurities, that may be present. Measurement of the permeation of He and Ar provides a good measure of the changes in the pore structure of 43 the membranes, since these two gases are noncondensable, nonadsorbing and inert, and are thus thought to interact minimally with the membrane pore surface. During this study, only single-gas permeation tests were routinely performed. In the past, our group has shown (Abdollahi, Yu, Hwang, et al., 2010; Abdollahi, Yu, Liu, et al., 2010) that single- and mixed-gas permeances for noncondensable gases (e.g., He, Ar) are close to each other; and while single-gas permeation experiments are convenient to perform, as they only require the measurement of the permeate- side flow, mixed-gas permeabilities, on the other hand, they are notoriously time-consuming and difficult to measure with these high-flux nanoporous membranes and, in addition, are prone to experimental artifacts, including, but not limited to, concentration polarization effects. For the single-gas permeation experiments, the membrane were sealed with two appropriate rubber o- rings inside the permeation apparatus (made of a stainless tube of the same length with appropriate end-fittings) via silicone O-rings wrapped around the glazed ends of the membrane tube. The gas was fed through the inside of the membrane tube, the pressure (30 psi) of which was controlled by a needle valve on the tube-side outlet. The pressure on the permeate side was maintained at 1 atm. The flow rate of the outlet gas from the permeate side was measured using a soap-bubble flow meter, see Fig. 2.5. On the basis of these measurements, the membrane permeances (Pj) and ideal selectivities (Sij) which characterize the ability of membrane to separate gas (i) from gas (j) for cylindrical membranes were calculated by the following equations: 2.1 2.2 0 0 2 m j m VT P P RLT P P i ij j P S P 44 where V is the volumetric flow rate (m 3 /h) of the gas across the membrane, P m is the pressure (bar) at which the volumetric flow of permeate side was measured, T m is the temperature (K) where the measurements were taken, L is the membrane length (m), R is the membrane inner radius (m), P o and T o are the pressure (bar) and temperature (K) at standard conditions, and ΔP (bar) is the pressure difference across the tube. Figure 2.5 Schematic of experimental apparatus for measuring permeance and ideal selectivity (Lee et al., 2012) Table 2.5 shows the He and Ar single gas permeances and ideal selectivities for one of the CMS membranes we prepared via the pyrolysis of PEI. Specifically, to prepare this Table 2.5. He and Ar single gas properties of membrane Gas Permeance mol/(pa.s.m 2 ) Ideal selectivity (He/Ar ) He 2.33E-07 90.1 Ar 2.59E-09 membrane the dip-coated support tube was heated-up to 450 °C (with a heating rate of 1 °C/min) and held there for 4 h. Its temperature was then raised to 700 °C (0.5 °C/min), held there for 8 h, and subsequently, the membrane was cooled down to 25 °C (1 °C/min). This membrane is microporous and has a high He permeance. The same membrane is used below in the post- treatment procedure using steam activation. 45 2.2.3 BET analysis Studying the transport properties of carbon molecular sieve membranes prior to (as well as following the steam activation post-treatment step – see further discussion below) provides interesting information regarding the separation properties of these systems and the various phenomena that occur during the steam post-treatment procedure. In this work, the transport investigations have been complimented with parallel ones of the sorption characteristics of the same materials. For the samples with pores larger than 3.5 Å, adsorption and desorption measurements using probe N 2 at its liquid temperature, provide a good insight into the pore structure characteristics, including the membrane internal surface area and the pore volume. It should be noted here, that though transport behavior and sorption measurements provide reasonable information about the membrane pore space, they each provide information about unique aspects of the pore space as well. For example, sorption experiments are thought to access the “totality” of the pore space; transport measurements, on the other hand, “sample” only the flow-through porosity. The nitrogen sorption experiments were carried out at 77 K (the liquid temperature of nitrogen) in a static mode using a commercial system (ASAP 2010 Micropore Analyzer from Micromeritics, Inc.). This system is equipped with two UHV pumps that were used to degas the membrane sample (and the overall sorption system) down to 10 −5 Torr (the degassing step, carried out at 250 °C, lasted typically about 24 h). As noted above, though measuring the amount of adsorbed and desorbed nitrogen from the membrane sample and calculating the total surface area and pore volume are rather straightforward tasks, extracting additional structural information (e.g., average pore size and PSD) is much dependent upon selecting the appropriate geometric (or molecular) model of the pore space. The latter step is rather complicated and involves elaborate simulations that go beyond the scope of the present investigation. Here 46 instead, we used two common techniques (the so-called BJH method for the mesoporous region, and the Horvath−Kawazoe (HK) method for the microporous region) which were utilized to extract additional structural information from the nitrogen adsorption data. Since these two techniques make use of substantial simplifying assumptions (e.g., the HK method assumes the material’s pore structure to consist of slit-like pores), it is the qualitative trends in the various structural parameters (e.g., average pore diameters) that are more important rather than their exact values. A good “base-line” measurement, when working with supported CMS membranes, is that of the starting support material itself on which the permselective membrane layers are deposited. Figure 2.6 presents the PSD data of both substrates and membranes, based on experimental sorption data and calculated using the BJH method. For these ceramic substrates, with an asymmetric pore structure, gas adsorption mainly takes place in the γ-alumina top layer, whose weight is only a small fraction of the total weight of the support substrate; (Sedigh et al., 1999) as a result, the values reported for the cumulative and differential pore volumes are low, due to the fact that the total weight of the support tube is used in the calculation. The PSD analysis indicates an average pore diameter for the mesoporous layer between 4 and 5 nm, which is consistent with what the literature typically reports for sol−gel γ-alumina layers, and also the average pore diameter this group measures from transport data with the same support tubes. Table 2.6 summarizes the BET results for ceramic substrates and CMSs membranes prepared from these substrates. As shown in the table, the BET surface area increases notably for the CMS membranes when compared to the substrate they are prepared from. 47 Figure 2.6 shows the PSD of both a substrate (A), and of two CMS membranes B and C dip- coated with one layer (in the case of membrane C) or three layers (in the case of membrane B) of polymer and pyrolyzed to prepare the top CMS layer. Notice that the presence of the carbon layer introduces a nanoporous region in the resulting membranes. The method of preparation also has an impact in the final properties of the resulting membranes. For example, membrane C was prepared after the substrate and the polymer precursor solution were both evacuated overnight. By this method, any gases inside the polymer solution and water molecules and other impurities in the pores of the substrate are removed. The presence of these molecules during pyrolysis may result in the formation of pinholes and other defects. This is visible in Fig. 2.6 where membrane B shows a mesopore region with a broader pore size distribution and larger average pore size. Figure 2.7 also shows nitrogen adsorption (black line) and desorption (red line) versus relative pressure of nitrogen from P/P 0 =0 to 1.0 for membranes B and C. The adsorption curve for both membranes is of Type II in the IUPAC classification (Chapter 1). Table 2.6. BET surface area of the substrate and CMS membranes Samples Name BET Surface Area (m²/g) Substrate 0.0351 3-Layer carbon membrane (B) 2.0339 1-Layer carbon membrane (C) 1.2500 48 Figure 2.6 Pore size distribution of both substrate A and membranes B and C 49 Figure 2.7 Nitrogen adsorption isotherm (T<T critical =126.15 K) of the CMS membranes B and C 2.3 Steam Activation as a Post-treatment Step Steam activation is a technique used often for preparing activated carbons (AC) and other microporous materials with desired pore structures; see Table 2.7 for a listing of some of the key studies. During the procedure, steam reacts with carbon atoms on the internal pore surface area of the solid, and this creates additional microporosity. (Bansal, Donnet, & Stoeckli, 1988) At some point, as the activation process continues, pore walls begin to collapse, and the number of micropores and the corresponding micropore surface area go through a maximum, and then start decreasing. (Walker Jr, 1996) 50 Table 2.7. Select papers on the use of steam activation in the preparation of microporous carbons Material Gas Atmosphere Comments Ref. Pitch precursor Steam treatment following thermo- oxidation (H 2 SO 4 or HNO 3 ) and subsequent low- pressure foaming without stabilization H 2 SO 4 favors the formation of carbon foams with a dense and ordered structure, while nitric acid favors the insertion of heteroatoms in the carbon skeleton. These differences play an important role in the development of the porosity when these foams are activated with steam. (Tsyntsarski et al., 2012) Fly-ash carbons Steam Not all fly-ash carbon samples are suitable for activation. The concentration of active sites influences carbon gasification reactions. (Lu, Maroto-Valer, & Schobert, 2008) Pyrolytic tire carbon back Steam Kinetic parameters for the steam activation are reported. (Aranda, Murillo, García, Callén, & Mastral, 2007) Olive stone chars Steam Low activation temperature (800 °C) favors the widening of narrow micropores (<0.7 nm); when the activation temperature is increased to 850 °C porosity increases, but then decreases for higher temperatures. (M. González, Molina-Sabio, & Rodriguez- Reinoso, 1994) Extracted rockrose CO 2 and steam Steam activation, especially at 700 °C, yields AC with larger total pore volumes than those prepared by CO 2 activation. (Pastor-Villegas & Duran-Valle, 2002) Carbon- silica sorbent (carbosil) prepared by thermal decompositi on of dichloromet hane on silica gel grains Steam Steam activation causes substantial changes in the surface structural parameters of carbosil sorbents. (Gierak, Czechowski, & Leboda, 1994) Eucalyptus globules and peach stones Steam Steam activation causes substantial changes in the porosity and concentration of oxygen surface groups in the chars.. (Arriagada, Garcia, Molina-Sabio, & Rodriguez- Reinoso, 1997) Olive stone Steam and CO 2 CO 2 generates narrow pores, which get wider upon prolonged exposure. Steam helps widen existing micropores from the early stages of activation, the resulting AC exhibiting lower micropore volume. (Menendez-Diaz & Martin-Gullon, 2006) AC Steam and CO 2 CO 2 and steam produce different porous structures. CO 2 activation mainly creates new micropores. Steam, on the other hand, widens the micropore PSD. (Rodriguez- Reinoso, Molina- Sabio, & Gonzalez, 1995) AC Steam and CO 2 Steam activation widens the microporosity from the early stages of the activation process, the resulting AC exhibiting a lower micropore volume. (Molina-Sabio, Gonzalez, Rodriguez- Reinoso, & Sepúlveda- Escribano, 1996) 51 Granular AC Steam and CO 2 Steam activation, especially at high temperature, yields AC with total pore volume larger than those prepared by CO 2 activation. (Pastor-Villegas & Duran-Valle, 2002) AC prepared from walnut shell Steam and CO 2 Steam is more reactive and produces AC with greater N 2 adsorption capacities. The increase in the fraction of mesopores is much larger for steam. (J. F. González, Román, González- García, Nabais, & Ortiz, 2009) AC Steam and CO 2 The AC produced from CO 2 activation show higher BET surface areas and pore volumes when compared to AC from steam activation. SEM shows that the former AC have smoother surface and better microstructures as compared to AC from the steam activation process. (Arjun Singh & Lal, 2010) Carbon foam Steam Carbon foam prepared by pyrolysis of olive stones under steam (Rios, Martínez- Escandell, Molina- Sabio, & Rodríguez- Reinoso, 2006) At elevated temperatures, the actual process is described by the chemical reaction, R1, below: (Menendez-Diaz & Martin-Gullon, 2006) 2 2 1 () 132 / C H O CO H R H kJ mole Reaction R1 has been extensively studied, not only because of its importance in the preparation of a variety of AC, but also because of its role during coal gasification. In addition to R1, other reactions that, likely, participate during the steam activation of carbons include the water−gas shift reaction R2, and the direct hydrogenation of carbon to generate methane R3. 2 2 2 2 () 41.5 / CO H O CO H R H kJ mole 2 4 3 () 87.5 / C H CH R H kJ mole However, it is R1 that is thought to be, primarily, responsible for the changes in the carbon pore structure during steam activation, the other reactions, likely, making minor contributions. As noted above, the phenomena taking place during steam activation are complex. The change in porosity, during the early stages of activation, is thought to result from the opening (widening) of 52 the pore constrictions and from the development of a more interconnected pore structure. Continued reaction (“burn-off”) causes further widening of existing micropores; for CMS membranes this may originally be beneficial, as it increases the permeability, but eventually further widening of the pores, if left unchecked, will negatively impact the membrane selectivity. To the best of our knowledge (prior to our own publication (Lee et al., 2012) the results of which are presented here), there were no published studies dealing exclusively with the impact of steam activation on the properties of CMS membranes; there are numerous studies, on the other hand, dealing with steam treatment as a method to prepare AC. Rodriguez- Reinoso et al., (Rodriguez-Reinoso et al., 1995) for example, prepared AC from carbonized olive stones via steam gasification and studied the impact on porosity development of conditions like the partial steam pressure and activation temperature. They reported that activation with steam resulted in the widening of the existing micropores. Rodrigues-Reinoso et al., (Rodriguez-Reinoso et al., 1995) reported that CO and H 2 , produced via R 1 , act as inhibitors for R 1 itself. The lower concentration of these inhibitors found in the exterior of the particles (due to diffusional limitations) favors the reaction of steam at the external particle surface. They reported that increasing the temperature from 750 to 800 °C increased the rate of R1 but also made the inhibiting effect of CO and H 2 less severe, the net result being a faster rate of porosity formation, with the maximum in porosity being displaced toward the larger burn-offs (50%). In summary, the effect of steam activation on the properties of AC adsorbents has been investigated by a number of groups. The characteristics of the resulting materials have been shown to depend on the type of carbon source, the temperature and time of activation, and the burn-off level. In this research, the focus is on the use of steam activation as a post-treatment 53 technique in order to modify the pore size characteristics and transport and separation properties of CMS membranes. The impacts of the operating conditions, such as time of activation, on the properties of the resulting membranes such as ideal selectivity, permeation and removal rate of DMMP vapor by condensation are systematically investigated. For the steam activation step, the as prepared CMS membranes with a number of carbon layers were placed in a stainless steel tubular furnace (with ID=5.1 cm and L=60.6 cm equipped with three heating elements) and heated (1 °C/min) in the presence of UHP Ar (99.999%) to a predetermined temperature; when this temperature was reached, the Ar was replaced by a flowing steam/Ar stream with a molar ratio of 2:1. The membrane was kept in the steam/Ar flowing mixture for a predetermined period before being slowly cooled down (2 °C/min) to room temperature. Figure 2.8 shows the impact of steam activation at two temperatures and various time periods on the permeation characteristics. As the results show, temperature and duration of the activation treatment are the key operating parameters. Note also, that steam activation provides for a delicate control of the pore structure and permeation properties of the treated membranes. For the (He/Ar) gas pair, steam treatment results in a monotonic decrease in the ideal selectivity. This is not always the case, however, with other gas pairs, which show different behavior (Lee et al., 2012); this is a manifestation of the complex phenomena that govern membrane transport through such nanoporous membranes. 54 Figure 2.8 Permeance of membranes as a function of temperature and time of steam activation: (top) He permeance; (bottom) Ar permeance and ideal selectivity (membrane I). (Lee et al., 2012) 2.3 Conclusions CMS membranes show good potential for important gas separations. Activation of the CMS membranes with steam provides additional flexibility to adjust the pore structure characteristics. This post-treatment is particularly suitable for preparing membranes with reverse selectivity for removing DMMP from air streams, which is the main focus of this research, see further discussion in Chapter 4. 55 References Abdollahi, M., Yu, J., Hwang, H. T., Liu, P. K., Ciora, R., Sahimi, M., & Tsotsis, T. T. (2010). Process intensification in hydrogen production from biomass-derived syngas. Industrial & Engineering Chemistry Research, 49(21), 10986-10993. Abdollahi, M., Yu, J., Liu, P. K., Ciora, R., Sahimi, M., & Tsotsis, T. T. (2010). Hydrogen production from coal-derived syngas using a catalytic membrane reactor based process. Journal of Membrane Science, 363(1), 160-169. Aranda, A., Murillo, R., García, T., Callén, M., & Mastral, A. (2007). Steam activation of tyre pyrolytic carbon black: Kinetic study in a thermobalance. Chemical Engineering Journal, 126(2), 79-85. Arriagada, R., Garcia, R., Molina-Sabio, M., & Rodriguez-Reinoso, F. (1997). Effect of steam activation on the porosity and chemical nature of activated carbons from< i> Eucalyptus globulus</i> and peach stones. Microporous Materials, 8(3), 123-130. Bansal, R. C., Donnet, J.-B., & Stoeckli, F. (1988). Active carbon: M. Dekker. Briceño, K., Garcia ‐Valls, R., & Montané, D. (2010). State of the art of carbon molecular sieves supported on tubular ceramics for gas separation applications. Asia ‐Pacific Journal of Chemical Engineering, 5(1), 169-178. Centeno, T. A., & Fuertes, A. B. (1999). Supported carbon molecular sieve membranes based on a phenolic resin. Journal of Membrane Science, 160(2), 201-211. Chen, Y., & Yang, R. (1994). Preparation of carbon molecular sieve membrane and diffusion of binary mixtures in the membrane. Industrial & Engineering Chemistry Research, 33(12), 3146-3153. Clint, J., Lear, A., Oliver, L., & Tennison, S. (1992). Membranes. EP Patent 0,474,424. Fuertes, A., & Centeno, T. (1998). Carbon molecular sieve membranes from polyetherimide. Microporous and Mesoporous Materials, 26(1), 23-26. Gierak, A., Czechowski, F., & Leboda, R. (1994). Improvement of carbon-silica sorbent (carbosil) surface properties upon steam activation at 1073 K. Materials chemistry and physics, 36(3), 264-270. Gilron, J., & Soffer, A. (2002). Knudsen diffusion in microporous carbon membranes with molecular sieving character. Journal of Membrane Science, 209(2), 339-352. González, J. F., Román, S., González-García, C. M., Nabais, J. V., & Ortiz, A. L. (2009). Porosity development in activated carbons prepared from walnut shells by carbon dioxide or steam activation. Industrial & Engineering Chemistry Research, 48(16), 7474-7481. 56 González, M., Molina-Sabio, M., & Rodriguez-Reinoso, F. (1994). Steam activation of olive stone chars, development of porosity. Carbon, 32(8), 1407-1413. Ismail, A. F., & David, L. (2001). A review on the latest development of carbon membranes for gas separation. Journal of Membrane Science, 193(1), 1-18. Julbe, A., Farrusseng, D., & Guizard, C. (2001). Porous ceramic membranes for catalytic reactors—overview and new ideas. Journal of Membrane Science, 181(1), 3-20. Kai, T., Kazama, S., & Fujioka, Y. (2009). Development of cesium-incorporated carbon membranes for CO< sub> 2</sub> separation under humid conditions. Journal of Membrane Science, 342(1), 14-21. Kim, Y. K., Park, H. B., & Lee, Y. M. (2004). Carbon molecular sieve membranes derived from thermally labile polymer containing blend polymers and their gas separation properties. Journal of Membrane Science, 243(1), 9-17. Kim, Y. K., Park, H. B., & Lee, Y. M. (2005). Gas separation properties of carbon molecular sieve membranes derived from polyimide/polyvinylpyrrolidone blends: effect of the molecular weight of polyvinylpyrrolidone. Journal of Membrane Science, 251(1), 159- 167. Kiyono, M., Williams, P. J., & Koros, W. J. (2010). Effect of pyrolysis atmosphere on separation performance of carbon molecular sieve membranes. Journal of Membrane Science, 359(1), 2-10. Koresh, J., & Soffer, A. (1980). Study of molecular sieve carbons. Part 1.—Pore structure, gradual pore opening and mechanism of molecular sieving. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 76, 2457- 2471. Koresh, J. E., Saggy, S., & Soffer, A. (1987). Separation device: Google Patents. Lagorsse, S., Magalhaes, F., & Mendes, A. (2004). Carbon molecular sieve membranes: sorption, kinetic and structural characterization. Journal of Membrane Science, 241(2), 275-287. Lee, H.-C., Monji, M., Parsley, D., Sahimi, M., Liu, P., Egolfopoulos, F., & Tsotsis, T. (2012). Use of steam activation as a post-treatment technique in the preparation of carbon molecular sieve membranes. Industrial & Engineering Chemistry Research, 52(3), 1122- 1132. Li, K. (2007). Ceramic membranes for separation and reaction: John Wiley & Sons. 57 Lin, H., Van Wagner, E., Freeman, B. D., Toy, L. G., & Gupta, R. P. (2006). Plasticization- enhanced hydrogen purification using polymeric membranes. Science, 311(5761), 639- 642. Linkov, V., Sanderson, R., & Jacobs, E. (1994). Highly asymmetrical carbon membranes. Journal of Membrane Science, 95(1), 93-99. Liu, P. K., Sahimi, M., & Tsotsis, T. T. (2012). Process intensification in hydrogen production from coal and biomass via the use of membrane-based reactive separations. Current Opinion in Chemical Engineering, 1(3), 342-351. Lu, Z., Maroto-Valer, M. M., & Schobert, H. H. (2008). Role of active sites in the steam activation of high unburned carbon fly ashes. Fuel, 87(12), 2598-2605. Menendez-Diaz, J., & Martin-Gullon, I. (2006). Types of carbon adsorbents and their production. Interface Science and Technology, 7, 1-47. Mikuláŝek, P., Doleček, P., Šedá, H., & Cakl, J. (1994). Alumina ‐Ceramic Microfiltration Membranes: Preparation, Characterization and Some Properties. Developments in Chemical Engineering and Mineral Processing, 2(2‐3), 115-123. Molina-Sabio, M., Gonzalez, M., Rodriguez-Reinoso, F., & Sepúlveda-Escribano, A. (1996). Effect of steam and carbon dioxide activation in the micropore size distribution of activated carbon. Carbon, 34(4), 505-509. Nishihara, Y., & Yoneyama, H. (1992). Carbon based porous hollow fiber membrane and method for producing same: Google Patents. Pastor-Villegas, J., & Duran-Valle, C. (2002). Pore structure of activated carbons prepared by carbon dioxide and steam activation at different temperatures from extracted rockrose. Carbon, 40(3), 397-402. Rios, R., Martínez-Escandell, M., Molina-Sabio, M., & Rodríguez-Reinoso, F. (2006). Carbon foam prepared by pyrolysis of olive stones under steam. Carbon, 44(8), 1448-1454. Rodriguez-Reinoso, F., Molina-Sabio, M., & Gonzalez, M. (1995). The use of steam and CO< sub> 2</sub> as activating agents in the preparation of activated carbons. Carbon, 33(1), 15-23. Sedigh, M. G., Jahangiri, M., Liu, P. K., Sahimi, M., & Tsotsis, T. T. (2000). Structural characterization of polyetherimide ‐based carbon molecular sieve membranes. AIChE Journal, 46(11), 2245-2255. Sedigh, M. G., Onstot, W. J., Xu, L., Peng, W. L., Tsotsis, T. T., & Sahimi, M. (1998). Experiments and simulation of transport and separation of gas mixtures in carbon molecular sieve membranes. The Journal of Physical Chemistry A, 102(44), 8580-8589. 58 Sedigh, M. G., Xu, L., Tsotsis, T. T., & Sahimi, M. (1999). Transport and morphological characteristics of polyetherimide-based carbon molecular sieve membranes. Industrial & Engineering Chemistry Research, 38(9), 3367-3380. Shiflett, M. B., & Foley, H. C. (2001). Reproducible production of nanoporous carbon membranes. Carbon, 39(9), 1421-1425. Shusen, W., Meiyun, Z., & Zhizhong, W. (1996). Asymmetric molecular sieve carbon membranes. Journal of Membrane Science, 109(2), 267-270. Singh, A., & Koros, W. (1996). Significance of entropic selectivity for advanced gas separation membranes. Industrial & Engineering Chemistry Research, 35(4), 1231-1234. Singh, A., & Lal, D. (2010). Preparation and characterization of activated carbon spheres from polystyrene sulphonate beads by steam and carbon dioxide activation. Journal of Applied Polymer Science, 115(4), 2409-2415. Suda, H., & Haraya, K. (1997). Alkene/alkane permselectivities of a carbon molecular sievemembrane. Chem. Commun.(1), 93-94. Thaeron, C., Parrillo, D., Sircar, S., Clarke, P., Paranjape, M., & Pruden, B. (1999). Separation of hydrogen sulfide–methane mixtures by selective surface flow membrane. Separation and purification technology, 15(2), 121-129. Tsyntsarski, B., Petrova, B., Budinova, T., Petrov, N., Velasco, L. F., Parra, J. B., & Ania, C. O. (2012). Porosity development during steam activation of carbon foams from chemically modified pitch. Microporous and Mesoporous Materials, 154, 56-61. Walker Jr, P. (1996). Production of activated carbons: use of CO< sub> 2</sub> versus H< sub> 2</sub> O as activating agent. Carbon, 34(10), 1297-1299. Wang, T., Zhang, B., Qiu, J., Wu, Y., Zhang, S., & Cao, Y. (2009). Effects of sulfone/ketone in poly (phthalazinone ether sulfone ketone) on the gas permeation of their derived carbon membranes. Journal of Membrane Science, 330(1), 319-325. Wei, W., Qin, G., Hu, H., You, L., & Chen, G. (2007). Preparation of supported carbon molecular sieve membrane from novolac phenol–formaldehyde resin. Journal of Membrane Science, 303(1), 80-85. Xiao, Y., Chng, M. L., Chung, T.-S., Toriida, M., Tamai, S., Chen, H., & Jean, Y. (2010). Asymmetric structure and enhanced gas separation performance induced by in situ growth of silver nanoparticles in carbon membranes. Carbon, 48(2), 408-416. 59 Zhang, X., Hu, H., Zhu, Y., & Zhu, S. (2006). Effect of carbon molecular sieve on phenol formaldehyde novolac resin based carbon membranes. Separation and purification technology, 52(2), 261-265. 60 Chapter 3: Development of Multi-Tubular, Flow-Through Catalytic Membrane Reactor (MFTCR) 3.1 Introduction 1 Chemical warfare agents (CWA) have been employed, unfortunately, on more than one occasion since the end of World War II. For example, Sarin (GB) a toxic chemical warfare agent was utilized during urban terrorist attacks in the early nineties in Japan resulting in the injury and death of several people (Mera et al., 2010; Ohbu et al., 1997; Yanagisawa et al., 1995). In addition, cleaning-up the contamination resulting from the Sarin attack proved quite a daunting challenge as well, requiring the use of copious amounts of water, thus resulting in secondary water pollution. The most recent incident involving CWA is their reported use during last year in the internal conflict in Syria that, by some accounts, killed or injured scores of people. The use of chemical weapons was also reported during the regional war between Iran and Iraq a couple of decades or so ago, but also in internal conflicts in Iraq and Syria during the same time period. Because of the immense threat CWA present to human and animal life as well as the environment, there have been significant technical efforts through the years to deal with their presence worldwide and their potential use. The efforts have focused primarily on three areas: (i) The complete and effective destruction of CWA stored by various nations, the most recent example here being the current international effort involving the removal and complete elimination of the Syrian government’s stockpile of chemical weapons; (ii) the decontamination of contaminated soil, buildings and equipment (e.g., civilian and military vehicles) that is likely 1 The discussion and data of this chapter are taken from a paper in press in JMS. Monji M., Ciora R., Liu P., Parsley D., Egolfopoulos F., Tsotsis T., Thermocatalytic Decomposition of Dimethyl Methylphosphonate (DMMP) in a Multi-Tubular, Flow-Through Catalytic Membrane Reactor, journal of membrane science, in press (2015). 61 to result from a CWA attack; and (iii) the protection of human and animal life during a potential employment of such weapons – these are classified as individual protection (IP) or collective protection (CP) applications, depending on the number of individuals involved. For the destruction of CWA stockpiles various approaches have been investigated, but the technique of choice to date has been thermal incineration (Korobeinichev, Ilyin, Bolshova, Shvartsberg, & Chernov, 2000; Werner & Cool, 1999), since it has proven in the past to be an efficient and relatively inexpensive method to destruct toxic hazardous materials including CWA. The method involves completely burning the compound(s) in question in a very hot flame environment created from the combustion of a sacrificial fuel, for example methane or hydrogen (Korobeinichev et al., 2000; Werner & Cool, 1999). Decontamination of affected surfaces, following a potential CWA attack, presents significant challenges as well (Mera et al., 2010). Typically, it involves the use a variety of cleaning solutions that are strong oxidizing or hydrolysis agents (Kiselev, Mattson, Andersson, Palmqvist, & Österlund, 2006; Obee & Satyapal, 1998; Rusu & Yates, 2000; Trubitsyn & Vorontsov, 2005). Due to increased concerns regarding terrorist attacks involving CWA, IP and CP systems have been studied as well in recent years with solid adsorbents (Henderson & White, 1988; K. Y. Lee, Houalla, Hercules, & Hall, 1994; Mitchell, Sheinker, & Mintz, 1997) and catalytic oxidizing agents (Cao, Segal, Suib, Tang, & Satyapal, 2000; Kiselev et al., 2006; Mitchell, Sheinker, Tesfamichael, Gatimu, & Nunley, 2003; Moss, Szczepankiewicz, Park, & Hoffmann, 2005; Obee & Satyapal, 1998; Rusu & Yates, 2000; Segal, Cao, Suib, Tang, & Satyapal, 2001; Trubitsyn & Vorontsov, 2005; Vorontsov et al., 2002) having garnered ample attention. The current technology in use involves adsorption with activated carbon (AC) beds. There are technical challenges associated with using conventional adsorption, however. CWA removal on 62 AC is via physisorption which is highly concentration-dependent, and during device operation desorption can take place when concentration spikes occur or other volatile organic compounds (VOC) are present in the contaminated air stream. Furthermore, the actual AC saturation capacity is often rather low (vs. predictions) due to pore-plugging by other contaminants (e.g., VOC) limiting access to the AC micropores. In addition, adsorption is a discontinuous operation with spent-bed disposal needed, because the CWA are not destroyed but, instead, accumulate in the bed; disposal is potentially hazardous, with exposure to CWA during such operations being a distinct possibility. As a result of the challenges common adsorption faces, other separation technologies, including the use of membranes, are attracting their fair share of attention today for both IP and CP applications. To date, both conventional (Jung, 2010; H.-C. Lee et al., 2012) and reactive (Motamedhashemi, Egolfopoulos, & Tsotsis, 2011) membrane-based separations have been investigated. The former make use of polymeric or mixed-matrix membranes which are tailor- made to block the toxic CWA from going through while allowing the air and water vapor to permeate, or alternately reverse-selectivity carbon membranes (H.-C. Lee et al., 2012) that allow the CWA to go through while blocking the air flow. Membrane-based reactive separations, recently employed by this group (Monji, Abedi, Pourmahdian, & Taromi, 2009), on the other hand, use reactive inorganic membranes. Operating as flow-through catalytic membrane reactors (FTCMR), these devices are able to completely oxidize the toxic CWA. In a recent study, this group utilized (Monji et al., 2009) a FTCMR with a single asymmetric mesoporous alumina membrane to investigate the destruction of dimethyl methylphosphonate (DMMP) in air. (DMMP is a stimulant for the nerve agent Sarin and, as a result, has received quite a bit of attention in the past for both its adsorption (Mitchell et al., 2003), and reaction (Bailin, Sibert, 63 Jonas, & Bell, 1975; Cao et al., 2000; Chen, Vorontsov, & Smirniotis, 2003; Henderson & White, 1988; Kiselev et al., 2006; K. Y. Lee et al., 1994; Mitchell et al., 1997; Moss et al., 2005; Obee & Satyapal, 1998; Rusu & Yates, 2000; Segal et al., 2001; Smentkowski, Hagans, & Yates Jr, 1988; Trubitsyn & Vorontsov, 2005; Uhm, Cho, Hong, Park, & Park, 2008; Vorontsov et al., 2002) characteristics). The experimental results with the FTCMR were quite promising indicating that complete DMMP conversions could be obtained with this reactive separation system. An experimentally-validated model for the FTCMR was also developed (Motamedhashemi et al., 2011), which was utilized to compare its performance, on an equitable basis, with other competitive reactor configurations (e.g., monolith reactors). In this chapter, results are reported of continued efforts to scale-up this lab-scale FTCMR towards practical IP and CP applications. A multi-tubular FTCMR (MFTCMR) has been developed and tested towards the destruction of DMMP. The MFTCMR has been shown quite effective, providing extended protection periods, as described in the Sec. 3 below. A key part of the research effort, furthermore, focused on efforts (described in Sec. 2 below) for improving the preparation methods of the catalytically active membranes to prepare membrane bundles with fairly reproducible transport characteristics and reactivity. A key challenge for the large-scale production of such devices, furthermore, remains the development of a simple non-destructive test that assures that the produced parts are appropriate for the proposed application, and that they continue to remain active during their shelf-life prior to their use. As part of this effort, a simple “propane-in-air” light-off test was studied which has been shown capable to track the reactivity and performance of the MFTCMR, see further discussion in Sec. 3 below. The multi- tubular FTCMR under development and described here is slated for field-testing at a US Government-sponsored facility in the near future. 64 3.2 Experimental Section For the experiments reported here, ceramic membrane bundles were prepared consisting Fig. 3.1. The multi-tubular membrane bundle before (left), and after catalytic impregnation (right). typically of twenty tubular asymmetric α-Al 2 O 3 membranes. These consist of an α-Al 2 O 3 support tube ~1.25 mm thick with an average pore diameter of 0.4 µm. On the top (outside surface) of this support layer a thin (~20 µm) α-Al 2 O 3 layer with an average pore diameter of ~50 nm is deposited via slip-casting. One end of these membrane tubes is potted in a non-porous ceramic collar, while the other end of the tube is free-standing but is completely plugged via a glass and ceramic tip. One such membrane bundle is shown in Fig. 3.1 (left). (For the bundle in Fig. 3.1 the exposed length of the membranes is approximately ~ 7 cm). The membranes are rendered catalytic via a wet-impregnation technique. For this step, hexa-chloroplatinic acid (H 2 Cl 6 Pt) solution (8 wt.%), which was purchased from Sigma Aldrich, was utilized to prepare a dilute solution (0.08 wt. %) in distilled water. Then, the membrane bundle was immersed in that solution, while it was constantly stirred via a magnetic stirrer, for a 65 period of 96 h. Subsequently, the membrane bundle was removed from the solution and was allowed to dry in room air for ~ 2 h. It was then placed in a box-furnace and calcined in air at a temperature of 350 o C for 5 h. The membrane bundle in Fig. 1 (left) after the impregnation and calcination steps is shown in Fig. 3.1 (right). After the impregnation and calcination steps, the membrane bundle is installed in its module, which is shown in Fig. 3.2. The module, made out of carbon steel, consists of two components, specifically: (i) the module body to house the membrane element, and (ii) a cap with a packing compression gland to seal the membrane in the housing. Fig. 3.2. Various views of the membrane module. This is accomplished via graphite tape (Palmetto, USA) that is compressed by the cap compression gland onto the module body against the membrane bundle’s ceramic collar to form a gas-tight seal suitable for high-temperature, leak-free operation. The module is equipped with input and output tubing and a thermo-well that accommodates a K-type thermocouple. Before the actual reaction runs are initiated, the catalytic membrane bundle is activated by reduction in a flowing H 2 stream for 4 h at 400 o C. The bundle is then first purged with He at a temperature of 66 250 o C for 4 h, followed by purging with air at the same temperature for 1 h, and then allowed to cool to room temperature for the permeation experiment to begin, see below. A schematic of the overall experimental system is shown in Fig. 3.3. The experimental system consists of a bubbler system (further details to be provided below), a pressure gauge to measure the feed pressure (Omega DPG-4000, 0-30 psi (206.84 KPa)), a pressure gauge to measure the permeate pressure (Omega DPG-4000; 0-100 psi (689.48 KPa)), a differential pressure transducer (Omega PX2300-25BDI, 25 psi ( 173.5 KPa)) to accurately measure the pressure difference across the two sides, and a GC/MS instrument (further details to be provided below) for the measurement of the DMMP concentrations in the feed and permeate streams. Heating tapes (HTS/Amptek AWH-101-020DM) connected to temperature controllers (Powerstat; Type 3PN1168) are used to heat-trace the lines connecting the bubbler to the reactor inlet, and the reactor outlet to the GC/MS in order to avoid adsorption and/or condensation of the DMMP. To further minimize the sorption of DMMP on the tubing walls, polytetrafluoroethylene (PTFE) tubing was used throughout the system. The system is connected to tanks containing different gases (He, H 2 , air, air/propane mixture, and Ar) whose flow rates are controlled by MFC (Brooks Series 5850, 0-500 sccm; Cole Parmer model 32907-67, 0.01 – 1 slpm). 67 Fig. 3.3. A schematic of the MFTCMR system. The DMMP-containing air feed stream for the reaction experiments was prepared by passing the air through a bubbler system consisting of two glass vessels (saturators) placed in series and filled with DMMP. The second saturator was also loaded with glass beads in order to diminish the formation of air bubbles and to increase the contact surface between the liquid and the flowing gas. To control the liquid temperature, the saturators were placed in a 5 gal water cooler filled with circulating water coming from a temperature-controlled thermal bath (Lauda Type RE-220). High purity DMMP (>97% pure) was purchased from Sigma Aldrich and it was used in the experiments without any further treatment or purification. For the experiments described here, the feed (air/DMMP or air/propane) is fed through the exterior side of the bundle, transports through (and reacts inside) the membrane and leaves from the membrane’s permeate side. The feed and permeate exit concentrations of DMMP (in the experiments reported here the reactor operates in a total flow-through mode, i.e., 100% stage- cut) were analyzed with a HP-6890 gas chromatograph (GC) coupled to a HP-5793 mass- 68 spectrometric detector (MSD) and equipped with a pneumatically-driven 6x-port valve. The GC/MS instrument contains a capillary column (J&W HP-5 MS), which was shown to be able to provide repeatable results for the analysis of DMMP in the feed and permeate gas streams. The GC/MS instrument was calibrated using standard samples of DMMP in methanol (prepared by dissolving precisely measured weights of DMMP in known volumes of methanol), and the corresponding peak areas obtained by the GC/MS were correlated via a linear equation to the moles of DMMP passing through the column. The GC/MS was equipped with a 5 ml sample- loop, heated at 80 o C in order to minimize the possible reaction/condensation/adsorption of DMMP. As noted above, to minimize the potential for DMMP condensation, the temperature of all tubing and fittings used in the reactor and connecting it to the GC/MS was kept above the ambient temperature by wrapping them with heating tapes (HTS/Amptek AWH-101-020DM). For the propane light-off experiments a certified gas mixture containing 5000 ppm v of propane in air, which was purchased from the Gilmore Air Gas company (www.gilmoreliquidair.com) was used as the feed. During the experiments, the feed and permeate stream compositions were measured using a GC (VARIAN, CP-3800) equipped with a TCD detector and a capillary column (WCOT Fused Silica 50MX0.53 MMID Coating CP-Sili 5CB). We have utilized two different instruments and columns for the DMMP and propane composition measurements in order to be able to independently optimize the analysis procedure. For the calibration of the VARIAN GC we have used standard propane-in-air mixtures prepared from the aforementioned certified gas mixture via dilution with air. The MFTCMR experiments were run at a fixed air molar flow rate and reactor temperature. During the experiments, the pressure on the permeate side was set at atmospheric conditions and the transmembrane pressure difference was constantly monitored via a 69 differential pressure transducer. This, then, allows one to estimate on line the membrane bundle’s instantaneous permeance (mol/m 2 sPa), and to be able to detect any changes in the membrane structure and permeation characteristics that may be brought upon by the deposition of any reaction by-products. For each membrane bundle tested here the following experimental protocol was established: 1. Treat the membrane bundle in flowing He (60 sccm) at 250 o C overnight in order to desorb any potential impurities (e.g., H 2 O and hydrocarbons) that may still be adsorbed on the bundle. 2. Turn the He flow rate off, turn on the air mixture flow, and lower the temperature to room temperature to carry out a permeation experiment. This involves keeping the permeate side at atmospheric pressure and increasing step-wise the feed pressure while measuring the flow through the membrane via a bubble-flow meter. For the fresh bundle#1, for example, the experimental permeation data are shown in Fig. 3.4, and exhibit a fairly linear dependence on transmembrane pressure drop, indicative of a predominantly Knudsen flow through the catalytically impregnated membrane bundle. The slope of the line passing through the experimental data is taken as the average permeance of the freshly-impregnated membrane bundle in this region of pressures (for bundle#1, 2.66*10 -6 mol/m 2 sPa). 70 Fig. 3.4. The air permeation properties of bundle #1 before the DMMP decomposition experiment. 3. Turn-off the pure air flow and turn-on the flow of the propane/air mixture. Set the air flow to a certain value, e.g., 0.117 mol/h (43.8 sccm) for all experiments with bundle#1 and bundle#2 (the total flow rate of air during the propane light-off experiments is kept the same with that used in the DMMP/air experiments). Measure the feed composition and the permeate exit composition as well as the pressure drop across the membrane bundle. Once the measurements stabilize, as indicated by three consecutive composition and pressure measurements being within 5% from each other, the module temperature is increased to 50 o C and the experimental procedure is repeated, i.e., record the permeate exit composition and the pressure drop and wait till all readings stabilize. Repeat the experiments at 100 o C, 150 o C, 200 o C, and 250 o C. The results with the propane/air light-off experiment with bundle#1 are shown in Fig. 3.5. Slope = 2.66*10 -6 (mole/m 2 *pa*s) 0.E+00 1.E-03 2.E-03 3.E-03 4.E-03 5.E-03 6.E-03 7.E-03 0 500 1000 1500 2000 2500 3000 Molar Flux (mole/m 2 *s) Pressure drop, dp (Pa) 71 Fig. 3.5. Propane/air light-off curve for bundle#1 before exposure to DMMP. 4. Upon completion of the propane/air light-off experiment, the flow of the propane/air mixture is turned-off, and the flow of He is turned-on (60 sccm) for 4 h at 200 o C in order to purge the membrane bundle and/or module of any propane or reaction products that may remain adsorbed. Subsequently, the temperature is raised under flowing He to 250 o C. When the module temperature stabilizes at this value, the He flow is turned-off and the DMMP/air flow is turned- on at 0.117 mol/h (43.8 sscm) for all experiments with bundle#1 and bundle#2, and at 0.238 mol/h (88.8 sccm) in the experiments with bundle#3, intended to investigate the effect of feed air flow rate. During the experiments, the DMMP concentration in the permeate stream and the pressure drop across the membrane bundle are being constantly monitored. The experiment is, typically, terminated when the conversion dips below 99.9%, with the last measured conversion 0 500 1000 1500 2000 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 Pressure drop, dP (Pa) Propane convertion (%) Temperature (ºC) 72 measured prior to termination reported on the Figures. The results with the DMMP/air mixture for membrane bundle#1 are shown in Fig. 3.6. 5. Upon the termination of the DMMP/air reaction experiment, the flow of the feed mixture is terminated and is replaced with a He flow (60 sccm) at 250 o C overnight in order to purge the module of any DMMP that may remain. Subsequently, the oven heater is turned-off and the module temperature is lowered to room temperature under flowing He. The permeation experiment (as described in Step 2 above) and the propane/air light-off experiment (as described in Step 3 above) are repeated, aiming to identify the impact on the membrane transport and reaction characteristics resulting from the DMMP/air reactor experiments. Fig. 3.6. DMMP thermocatalytic conversion of bundle #1; DMMP feed concentration (ppm v )=1000±5%. 3.3 Results and Discussion Fig. 3.5 shows the propane light-off test for membrane bundle#1. It is evident that the MFTCMR is effective in decomposing the propane whose conversion is 9.4% even at room temperature. As 100 99.99 99 0 500 1000 1500 2000 0 10 20 30 40 50 60 70 80 90 100 0 100 200 300 400 500 600 700 800 Pressure drop, dP (Pa) DMMP conversion (%) Time (min) 73 the temperature is increased to 250 o C, the conversion of propane becomes 71.3%. Monitoring of the transmembrane pressure drop indicates an increase from 527 Pa (2.1 in of H 2 O) to 868 Pa (3.46 in of H 2 O), which corresponds to a ratio of the room temperature permeance of air to that at 250 o C of 1.65. Based on the expected dependence of Knudsen transport that ratio should be 1.32, indicative that propane catalytic oxidation does not substantially impact the membrane’s structure – see further discussion below. Fig. 3.6 shows the results of the DMMP thermocatalytic decomposition test with this membrane bundle after the completion of the propane/air light-off test. The multi-tubular FTCMR device is quite effective in decomposing the DMMP with more than 99.99% conversion attained over a 9.67 hr (9 h and 40 min) period. The room temperature permeation test for this membrane bundle after exposure to DMMP is shown in Fig. 3.7. Transport through the membrane continues to be dominated by Knudsen transport with a fairly linear dependence of gas flow on the transmembrane pressure gradient. The average permeance is 2.57*10 -6 mol/m 2 sPa which is ~ 3.50% less than the permeance of the freshly impregnated membrane of 2.66*10 -6 mol/m 2 sPa. This slight decrease in permeance is potentially due to the deposition of phosphorous-containing mineral acid by-products (Sahimi & Tsotsis, 1985) that coat the membrane surface and clog its pore structure (within the detection limit of the GC/MS, no gas- phase products were observed, however, other than CO 2 and H 2 O). 74 Fig. 3.7. The air permeation properties of bundle #1 after exposure to the DMMP in the decomposition experiment. Fig. 3.8 shows the results of the propane/air light-off test with the same membrane bundle after it was used in the DMMP thermocatalytic decomposition tests. When compared with the propane light-off tests with the fresh membrane, these results indicate a less active membrane. For example, the propane conversion at 250 o C is 59.4%, which is 16.69% less of the conversion of 71.3% attained with the fresh membrane at 250 o C. This indicates the impact that oxidative decomposition of DMMP has on the catalytic properties of membrane bundle#1, and also the fact that the propane/air light-off test is a sensitive indicator of the reactivity characteristics of the membrane bundle#1. As a reminder, at the termination (after 12.67 h from the start) of the experiments shown in Fig. 3.6 the multi-tubular FTCMR was still capable to deliver ~ 99% conversion of the DMMP. Slope = 2.57*10 -6 (mole/m 2 *pa*s) 0.E+00 1.E-03 2.E-03 3.E-03 4.E-03 5.E-03 6.E-03 7.E-03 0 500 1000 1500 2000 2500 3000 Molar Flux (mole/m 2 *s) Pressure drop, dP (Pa) 75 Fig. 3.8. Propane/air light-off curve for bundle#1 after exposure to DMMP. Figs. 3.9-3.11 show the results with a different membrane (bundle#2). The average permeance of this membrane after catalytic impregnation was measured to be 2.47*10 -6 mol/m 2 sPa (see supplementary materials for the permeation data plot, Figure A.1), which is 7.14% less than the permeance of 2.66*10 -6 mol/m 2 sPa of membrane bundle#1. This demonstrates our ability to prepare membrane bundles with fairly reproducible transport characteristics. The propane/air light-off results with membrane bundle#2 are shown in Fig. 3.9 and indicate that the membrane bundle#2 is also quite active in catalytically oxidizing the propane in air (a conversion of 7.9% at room temperature rising to 69.3% at 250 o C). Again, as with bundle#1, the light-off test has no discernible impact on the membrane’s transport properties with the differences in air permeance as the temperature increases being in line (the transmembrane pressure drop increases from 527 Pa to 895 Pa corresponding to a ratio of the room temperature permeance of air to that at 250 o C of 1.7), within the uncertainty of the experimental measurements, with the expected dependence of Knudsen diffusivity on temperature. Fig. 3.10 shows the results of the thermocatalytic 0 500 1000 1500 2000 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 Pressure drop, dP (Pa) Propane convertion (%) Temperature (ºC) 76 oxidation of the DMMP/air mixture. Bundle#2 is again quite active with a protection time (the period during which the conversion stays above 99.99%) of ~10.3 h, which is relatively close to the protection time of ~ 9.67 h (see Fig. 3.6) offered by membrane bundle#1; this again indicates the ability to prepare membrane bundles with fairly reproducible reactivity. The permeance of membrane bundle#2 after the DMMP test was measured to be (see supplementary materials section, Figure A.2) 2.38*10 -6 mol/m 2 sPa, which is 3.64 % less than the average permeance of the freshly impregnated membrane, indicating once more a slight clogging of the membrane pore structure. The results with the propane/air light-off test, carried out after the DMMP experiments, are shown in Fig. 11 and indicate once more (as with bundle#1) that the DMMP test has a substantial impact on reactivity and also that the propane test is a sensitive indicator of these changes (the propane conversion at 250 o C is 60.11% which is 13.2% less of the conversion of 69.3% attained with the fresh membrane at the same temperature). Fig. 3.9. Propane/air light-off curve for bundle#2 before exposure to DMMP. 0 500 1000 1500 2000 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure drop, dP(Pa) Propane converstion (%) Temperature (°C) 77 Fig. 3.10. DMMP thermocatalytic conversion of bundle #2; DMMP feed concentration (ppm v )=941±5%. Fig. 3.11. Propane/air light-off curve for bundle#2 after exposure to DMMP. Figs. 3.12-3.14 show the results with a different membrane (bundle#3). The average permeance of this membrane after catalytic impregnation was measured to be 2.49*10 -6 mol/m 2 sPa (see 100 99.98 99.01 0 500 1000 1500 2000 0 20 40 60 80 100 0 100 200 300 400 500 600 700 800 Pressure drop, dP (Pa) DMMP conversion (%) Time (min) 0 500 1000 1500 2000 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure drop, dP (Pa) Propane converstion (%) Temperature (°C) 78 supplementary materials for the permeation data plot, Fig. A.3), which is ~6.39% less than the permeance of membrane bundle#1, and 0.8% more than the permeance of membrane bundle#2, indicating once more our ability to prepare such membrane bundles with fairly reproducible transport characteristics. For this membrane bundle, the air flow for both the propane/air light-off experiments as well as the DMMP/air reaction experiments is set at 0.238 mol/h (88.8 sccm) in order to investigate the impact on performance of the membrane throughput (for this case the initial pressure drop at room temperature is ~ 4 in H 2 O but rises above that at higher temperatures due to the impact of temperature on membrane permeance). The propane/air light- off results with membrane bundle#3 are shown in Fig. 3.12, and indicate that the membrane bundle#3 is also quite active in catalytically oxidizing the propane in air. The light-off test also, as with bundle#1, has no substantial impact on the membrane’s transport properties. Fig. 3.13 shows the results of the thermocatalytic oxidation of the DMMP/air mixture. Bundle#3 is again quite active with a protection time (the period during which the conversion stays above 99.99%) of 6.20 h which is 35.88% less than the protection time of ~ 9.67 h offered by membrane bundle#1 (see Fig. 6) and 39.8% less than the protection time of ~ 10.3 h offered by membrane bundle#2 (see Fig. 10). This then indicates the significant impact that the air flow has on the operation of the device. Similar observations were previously made with the single-tube FTCMR operation [21]. The permeance of membrane bundle#3 after the DMMP test was measured to be (see supplementary materials section, Fig. A.4) 2.41*10 -6 mol/m 2 sPa, which is 3.3% less than the average permeance of the freshly impregnated membrane, indicating once more a slight clogging of the membrane pore structure. The results with the propane/air light-off test, carried out after the DMMP experiments, are shown in Fig. 3.14 and indicate once more (as with bundle#1) that 79 the DMMP test has a substantial impact on reactivity, and also that the propane test is quite a sensitive indicator of these changes. Fig. 3.12. Propane/air light-off curve for bundle#3 before exposure to DMMP. Fig. 3.13. DMMP thermocatalytic conversion of bundle #3; DMMP feed concentration (ppm v )=997±3%. 0 500 1000 1500 2000 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure drop, dP (Pa) Propane conversion (%) Temperature (°C) 100 99.99 99 0 500 1000 1500 2000 0 20 40 60 80 100 0 100 200 300 400 500 600 700 Pressure drop, dP(Pa) DMMP conversion (%) Time (min) 80 Fig. 3.14. Propane/air light-off curve for bundle#3 after exposure to DMMP. 3.4 Conclusions The successful use of a FTCMR as a protection system against DMMP, a CWA simulant, was reported recently by this group [21]. This FTCMR employed a single reactive mesoporous membrane and was shown effective for the destruction of the CWA. In this chapter, results are reported of efforts to scale-up this lab-scale FTCMR towards practical IP and CP applications. A multi-tubular FTCMR has been developed and tested towards the destruction of DMMP. The MFTCMR has been shown quite effective, providing extended protection periods against this CWA simulant. A key part of the research effort focused on improving the preparation of the catalytically active membranes to prepare membrane bundles with reproducible transport characteristics and reactivity. Three different membrane bundles were tested, which showed fairly similar reactivity and permeance. The impact of contaminated air flow through the device was also investigated and it was shown to be quite significant, as one may have expected. As part of this effort a simple propane-in-air light-off test was studied, which has been shown capable to 0 500 1000 1500 2000 0 20 40 60 80 100 0 50 100 150 200 250 300 Pressure drop, dP(Pa) Propane conversion (%) Temperature (°C) 81 track the reactivity and performance of the MFTCMR. This is important, as the development of such a simple non-destructive test assures that the produced parts are appropriate for the proposed application, and that they continue to remain active during their shelf-life prior to their use. 82 References Bailin, L. J., Sibert, M. E., Jonas, L. A., & Bell, A. T. (1975). Microwave decomposition of toxic vapor simulants. Environmental Science & Technology, 9(3), 254-258. Cao, L., Segal, S. R., Suib, S. L., Tang, X., & Satyapal, S. (2000). Thermocatalytic oxidation of dimethyl methylphosphonate on supported metal oxides. Journal of Catalysis, 194(1), 61-70. Chen, Y.-C., Vorontsov, A. V., & Smirniotis, P. G. (2003). Enhanced photocatalytic degradation of dimethyl methylphosphonate in the presence of low-frequency ultrasound. Photochem. Photobiol. Sci., 2(6), 694-698. Henderson, M. A., & White, J. (1988). Adsorption and decomposition of dimethyl methylphosphonate on platinum (111). Journal of the American Chemical Society, 110(21), 6939-6947. Jung, K. H. (2010). Nonwovens Containing Novel Polymer Fillers. Kiselev, A., Mattson, A., Andersson, M., Palmqvist, A., & Österlund, L. (2006). Adsorption and photocatalytic degradation of diisopropyl fluorophosphate and dimethyl methylphosphonate over dry and wet rutile TiO< sub> 2</sub>. Journal of Photochemistry and Photobiology A: Chemistry, 184(1), 125-134. Korobeinichev, O., Ilyin, S., Bolshova, T., Shvartsberg, V., & Chernov, A. (2000). The chemistry of the destruction of organophosphorus compounds in flames—III: The destruction of DMMP and TMP in a flame of hydrogen and oxygen. Combustion and flame, 121(4), 593-609. Lee, H.-C., Monji, M., Parsley, D., Sahimi, M., Liu, P., Egolfopoulos, F., & Tsotsis, T. (2012). Use of Steam Activation as a Post-treatment Technique in the Preparation of Carbon Molecular Sieve Membranes. Industrial & Engineering Chemistry Research, 52(3), 1122-1132. Lee, K. Y., Houalla, M., Hercules, D. M., & Hall, W. K. (1994). Catalytic oxidative decomposition of dimethyl methylphosphonate over Cu-substituted hydroxyapatite. Journal of Catalysis, 145(1), 223-231. Mera, N., Hirakawa, T., Sano, T., Takeuchi, K., Seto, Y., & Negishi, N. (2010). Removal of high concentration dimethyl methylphosphonate in the gas phase by repeated-batch reactions using TiO< sub> 2</sub>. Journal of hazardous materials, 177(1), 274-280. 83 Mitchell, M. B., Sheinker, V., & Mintz, E. A. (1997). Adsorption and decomposition of dimethyl methylphosphonate on metal oxides. The Journal of Physical Chemistry B, 101(51), 11192-11203. Mitchell, M. B., Sheinker, V. N., Tesfamichael, A. B., Gatimu, E. N., & Nunley, M. (2003). Decomposition of dimethyl methylphosphonate (DMMP) on supported cerium and iron co-impregnated oxides at room temperature. The Journal of Physical Chemistry B, 107(2), 580-586. Monji, M., Abedi, S., Pourmahdian, S., & Taromi, F. A. (2009). Effect of prepolymerization on propylene polymerization. Journal of Applied Polymer Science, 112(4), 1863-1867. Moss, J. A., Szczepankiewicz, S. H., Park, E., & Hoffmann, M. R. (2005). Adsorption and photodegradation of dimethyl methylphosphonate vapor at TiO2 surfaces. The Journal of Physical Chemistry B, 109(42), 19779-19785. Motamedhashemi, M., Egolfopoulos, F., & Tsotsis, T. (2011). Application of a flow-through catalytic membrane reactor (FTCMR) for the destruction of a chemical warfare simulant. Journal of Membrane Science, 376(1), 119-131. Obee, T. N., & Satyapal, S. (1998). Photocatalytic decomposition of DMMP on titania. Journal of Photochemistry and Photobiology A: Chemistry, 118(1), 45-51. Ohbu, S., Yamashina, A., Takasu, N., Yamaguchi, T., Murai, T., Nakano, K., . . . Hinohara, S. (1997). Sarin poisoning on Tokyo subway. Southern medical journal, 90(6), 587-593. Rusu, C. N., & Yates, J. T. (2000). Photooxidation of dimethyl methylphosphonate on TiO2 powder. The Journal of Physical Chemistry B, 104(51), 12299-12305. Sahimi, M., & Tsotsis, T. T. (1985). A percolation model of catalyst deactivation by site coverage and pore blockage. Journal of Catalysis, 96(2), 552-562. Segal, S. R., Cao, L., Suib, S. L., Tang, X., & Satyapal, S. (2001). Thermal decomposition of dimethyl methylphosphonate over manganese oxide catalysts. Journal of Catalysis, 198(1), 66-76. Smentkowski, V., Hagans, P., & Yates Jr, J. (1988). Study of the catalytic destruction of dimethyl methylphosphonate (DMMP): oxidation over molybdenum (110). The Journal of Physical Chemistry, 92(22), 6351-6357. Trubitsyn, D. A., & Vorontsov, A. V. (2005). Experimental study of dimethyl methylphosphonate decomposition over anatase TiO2. The Journal of Physical Chemistry B, 109(46), 21884-21892. 84 Uhm, H. S., Cho, S. C., Hong, Y. C., Park, Y. G., & Park, J. S. (2008). Destruction of dimethyl methylphosphonate using a microwave plasma torch. Applied Physics Letters, 92(7), 071503. Vorontsov, A. V., Davydov, L., Reddy, E. P., Lion, C., Savinov, E. N., & Smirniotis, P. G. (2002). Routes of photocatalytic destruction of chemical warfare agent simulants. New journal of chemistry, 26(6), 732-744. Werner, J. H., & Cool, T. A. (1999). Kinetic model for the decomposition of DMMP in a hydrogen/oxygen flame. Combustion and flame, 117(1), 78-98. Yanagisawa, N., Morita, H., Nakajima, T., Okudera, H., Shimizu, M., Hirabayashi, H., . . . Mimura, S. (1995). Sarin poisoning in Matsumoto, Japan. The Lancet, 346(8970), 290- 293. 85 Chapter 4: Future Work A computer programming model is currently developed and being finalized for the separation of condensable vapors through nanoporous material [Monji, M. et al]. In this model, for the first time we are taking to account the influence of surface diffusion for DMMP and even other gases which have tendency to the capillary condensation. The effects of different parameters such as membrane length, sweep and tube molar flow rate and pressure, membrane average pore size, membrane pore size distribution and total pressure and molar flow rate on the separation of DMMP and air loss in feed side were investigated and a journal paper is under preparation for that topics. So far, we have found out that the application of a forced-flow through catalytic reactor using catalytically impregnated alumina membranes for the oxidation of DMMP is a promising approach, but the observed pore blockage limits its application to: - Low catalyst loadings - Membranes having large pores - Low reaction temperatures In order to deepen our knowledge into the pore-blockage phenomena and to find possible methods to overcome it (or at least lessen its effects) in our experiments, in the near future, we will need to investigate the pore-blockage phenomenon using other support materials (in addition to alumina), including titania and zirconia. Applying various techniques, including NMR, we will need to investigate the presence of phosphoric and/or methyl phosphonic acid in the membrane structure as they have been frequently reported to be the source of catalyst pore blockage in the published literature. We will need also to focus our attention on completely 86 identifying all the reaction products (in addition to carbon dioxide), if any in the gas phase. Carefully tracking the concentration of these products is important not only for the interpretation of the kinetics data for oxidation of DMMP, but also in terms of the eventual application of the device as some of these products may be irritants or poisonous to humans. The other crucial aspect which our future studies need to focus on is developing a mathematical model appropriate for simulation and interpretation of the behavior of our reaction system. The model will properly account for the transport properties of the multilayer membranes, the reaction kinetics (which will be measured independently), as well as the effect of pore blockage on membrane transport properties and reactivity. Developing such a mathematical model is important not only in terms of developing optimum operating conditions for our reactor system, but also to properly compare its behavior with that of other conventional reactors (e.g., fixed-bed, monolith, etc.) Finally, one key aspect of our future studies will be improving our catalyst performance. Our approach here will involve the development of membranes with uniformly dispersed nano- sized Pt crystallites, impregnated mostly in the membrane separation layer. Our studies will involve the detailed surface characterization of these membranes and the careful evaluation of their transport characteristics and reactivity. 87 References Monji M., Egolfopoulos F., Tsotsis T. T., A comprehensive model for the capture of condensable vapors and transport through nanostructure membranes, under preparation , Journal of Membrane Science (2015). 88 Bibliography Abdollahi, M., Yu, J., Hwang, H. T., Liu, P. K., Ciora, R., Sahimi, M., & Tsotsis, T. T. (2010). Process intensification in hydrogen production from biomass-derived syngas. Industrial & Engineering Chemistry Research, 49(21), 10986-10993. Abdollahi, M., Yu, J., Liu, P. K., Ciora, R., Sahimi, M., & Tsotsis, T. T. (2010). Hydrogen production from coal-derived syngas using a catalytic membrane reactor based process. Journal of Membrane Science, 363(1), 160-169. Adamson, A. (1997). Adsorption of gases and vapors on solids. Physical Chemistry of Surfaces. 6th ed. New York, NY: John Wiley and Sons, 605. Adamson, A. W., & Gast, A. P. (1967). Physical chemistry of surfaces. Aranda, A., Murillo, R., García, T., Callén, M., & Mastral, A. (2007). Steam activation of tyre pyrolytic carbon black: Kinetic study in a thermobalance. Chemical Engineering Journal, 126(2), 79-85. Arriagada, R., Garcia, R., Molina-Sabio, M., & Rodriguez-Reinoso, F. (1997). Effect of steam activation on the porosity and chemical nature of activated carbons from< i> Eucalyptus globulus</i> and peach stones. Microporous Materials, 8(3), 123-130. Asada, M., & DU, L. D. (1986). Separation of alcohol/water gaseous mixtures by thin ceramic membrane. Journal of Chemical Engineering of Japan, 19(1), 72-77. Ash, R., Barrer, R., Clint, J., Dolphin, R., & Murray, C. (1973). Isothermal and thermo-osmotic transport of sorbable gases in microporous carbon membranes. Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, 275(1249), 255-307. Ash, R., Barrer, R., & Pope, C. (1963). Flow of adsorbable gases and vapours in a microporous medium. II. Binary mixtures. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 271(1344), 19-33. Ash, R., Barrer, R. M., & Lowson, R. T. (1973). Transport of single gases and of binary gas mixtures in a microporous carbon membrane. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 69, 2166-2178. Bailin, L. J., Sibert, M. E., Jonas, L. A., & Bell, A. T. (1975). Microwave decomposition of toxic vapor simulants. Environmental Science & Technology, 9(3), 254-258. Bansal, R. C., Donnet, J.-B., & Stoeckli, F. (1988). Active carbon: M. Dekker. 89 Bose, T., Chahine, R., Marchildon, L., & St ‐Arnaud, J. (1987). New dielectric method for the measurement of physical adsorption of gases at high pressure. Review of scientific instruments, 58(12), 2279-2283. Bouwmeester, H., Burgraaf, A., Burgraaf, A., & Cot, L. Fundamentals of Inorganic Membrane Science and Technology, 1996, 435: Elsevier, Amsterdam. Briceño, K., Garcia ‐Valls, R., & Montané, D. (2010). State of the art of carbon molecular sieves supported on tubular ceramics for gas separation applications. Asia ‐Pacific Journal of Chemical Engineering, 5(1), 169-178. Braunauer, S. (1945). The adsorption of gases and vapours. The Adsorption of Gases and Vapours. Brunauer, S. (1943). Adsorption of gases and vapors. Brunauer, S. (1945). The Absorption of Gases and Vapors (Vol. 1): Princeton University Press. Butt, J. B., & Reed Jr, E. (1971). Surface diffusion of single sorbates at low and intermediate surface coverage. The Journal of Physical Chemistry, 75(1), 133-141. Cao, L., Segal, S. R., Suib, S. L., Tang, X., & Satyapal, S. (2000). Thermocatalytic oxidation of dimethyl methylphosphonate on supported metal oxides. Journal of Catalysis, 194(1), 61-70. Carman, P., & Raal, F. (1951). Diffusion and flow of gases and vapours through micropores. III. Surface diffusion coefficients and activation energies. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 209(1096), 38-58. Casanova, F. l., Chiang, C. E., Ruminski, A. M., Sailor, M. J., & Schuller, I. K. (2012). Controlling the role of nanopore morphology in capillary condensation. Langmuir, 28(17), 6832-6838. Centeno, T. A., & Fuertes, A. B. (1999). Supported carbon molecular sieve membranes based on a phenolic resin. Journal of Membrane Science, 160(2), 201-211. Chen, Y., & Yang, R. (1991). Concentration dependence of surface diffusion and zeolitic diffusion. AIChE Journal, 37(10), 1579-1582. Chen, Y., & Yang, R. (1994). Preparation of carbon molecular sieve membrane and diffusion of binary mixtures in the membrane. Industrial & Engineering Chemistry Research, 33(12), 3146-3153. Chen, Y.-C., Vorontsov, A. V., & Smirniotis, P. G. (2003). Enhanced photocatalytic degradation of dimethyl methylphosphonate in the presence of low-frequency ultrasound. Photochem. Photobiol. Sci., 2(6), 694-698. 90 Choi, J.-G., Do, D., & Do, H. (2001). Surface diffusion of adsorbed molecules in porous media: Monolayer, multilayer, and capillary condensation regimes. Industrial & Engineering Chemistry Research, 40(19), 4005-4031. Clint, J., Lear, A., Oliver, L., & Tennison, S. (1992). Membranes. EP Patent 0,474,424. Crank, J. (1975). The mathematics of diffusion. Darken, L. S. (1948). Diffusion, mobility and their interrelation through free energy in binary metallic systems. Trans. Aime, 175(184), 41. Do, D. D. (1996). A model for surface diffusion of ethane and propane in activated carbon. Chemical Engineering Science, 51(17), 4145-4158. Donohue, M., & Aranovich, G. (1998). Classification of Gibbs adsorption isotherms. Advances in colloid and interface science, 76, 137-152. Donohue, M., & Aranovich, G. (1999). A new classification of isotherms for Gibbs adsorption of gases on solids. Fluid phase equilibria, 158, 557-563. Elkamel, A., & Noble, R. D. (1992). A statistical mechanics approach to the separation of methane and nitrogen using capillary condensation in a microporous membrane. Journal of Membrane Science, 65(1), 163-172. Findenegg, G., Korner, B., Fischer, J., & Bohn, M. (1983). Supercritical Gas Adsorption in Porous Materials, I. Storage of Krypton in Carbon Molecular Sieves. Ger. Chem. Eng, 6, 80-84. Flood, E., & Huber, M. (1955). THERMODYNAMIC CONSIDERATIONS OF SURFACE REGIONS: ADSORBATE PRESSURES, ADSORBATE MOBILITY, AND SURFACE TENSION. Canadian Journal of Chemistry, 33(2), 203-214. Flood, E. A. (1967). The solid-gas interface (Vol. 2): M. Dekker New York. Fuertes, A., & Centeno, T. (1998). Carbon molecular sieve membranes from polyetherimide. Microporous and Mesoporous Materials, 26(1), 23-26. Gad-el-Hak, M. (2001). The MEMS handbook: CRC press. Gelb, L. D., Gubbins, K., Radhakrishnan, R., & Sliwinska-Bartkowiak, M. (1999). Phase separation in confined systems. Reports on Progress in Physics, 62(12), 1573. 91 Gierak, A., Czechowski, F., & Leboda, R. (1994). Improvement of carbon-silica sorbent (carbosil) surface properties upon steam activation at 1073 K. Materials chemistry and physics, 36(3), 264-270. Gilliland, E., Baddour, R., & Russell, J. (1958). Rates of flow through microporous solids. AIChE Journal, 4(1), 90-96. Gilron, J., & Soffer, A. (2002). Knudsen diffusion in microporous carbon membranes with molecular sieving character. Journal of Membrane Science, 209(2), 339-352. González, J. F., Román, S., González-García, C. M., Nabais, J. V., & Ortiz, A. L. (2009). Porosity development in activated carbons prepared from walnut shells by carbon dioxide or steam activation. Industrial & Engineering Chemistry Research, 48(16), 7474-7481. González, M., Molina-Sabio, M., & Rodriguez-Reinoso, F. (1994). Steam activation of olive stone chars, development of porosity. Carbon, 32(8), 1407-1413. Grandcolas, M., Louvet, A., Keller, N., & Keller, V. (2009). Layer ‐by ‐Layer Deposited Titanate ‐Based Nanotubes for Solar Photocatalytic Removal of Chemical Warfare Agents from Textiles. Angewandte Chemie, 121(1), 167-170. Gregg, S., & Sing, K. S. (1983). Adsorption, Surface Area, and Porosity. Henderson, M. A., & White, J. (1988). Adsorption and decomposition of dimethyl methylphosphonate on platinum (111). Journal of the American Chemical Society, 110(21), 6939-6947. Higashi, K., Ito, H., & Oishi, J. (1964). Surface Diffusion Phenomena in Gaseous Diffusion,(II) Separation of Binary Gas-mixtures. Journal of Nuclear Science and Technology, 1(8), 298-304. Hill, T. L. (1960). Statiscal-Thermodynamics: Addison-Wesley. Holt, J. K., Park, H. G., Wang, Y., Stadermann, M., Artyukhin, A. B., Grigoropoulos, C. P., . . . Bakajin, O. (2006). Fast mass transport through sub-2-nanometer carbon nanotubes. Science, 312(5776), 1034-1037. Hunter, R. J. (2001). Foundations of colloid science. Ismail, A. F., & David, L. (2001). A review on the latest development of carbon membranes for gas separation. Journal of Membrane Science, 193(1), 1-18. Julbe, A., Farrusseng, D., & Guizard, C. (2001). Porous ceramic membranes for catalytic reactors—overview and new ideas. Journal of Membrane Science, 181(1), 3-20. 92 Jung, K. H. (2010). Nonwovens Containing Novel Polymer Fillers. Kai, T., Kazama, S., & Fujioka, Y. (2009). Development of cesium-incorporated carbon membranes for CO< sub> 2</sub> separation under humid conditions. Journal of Membrane Science, 342(1), 14-21. Kim, Y. K., Park, H. B., & Lee, Y. M. (2004). Carbon molecular sieve membranes derived from thermally labile polymer containing blend polymers and their gas separation properties. Journal of Membrane Science, 243(1), 9-17. Kim, Y. K., Park, H. B., & Lee, Y. M. (2005). Gas separation properties of carbon molecular sieve membranes derived from polyimide/polyvinylpyrrolidone blends: effect of the molecular weight of polyvinylpyrrolidone. Journal of Membrane Science, 251(1), 159- 167. Kiselev, A., Mattson, A., Andersson, M., Palmqvist, A., & Österlund, L. (2006). Adsorption and photocatalytic degradation of diisopropyl fluorophosphate and dimethyl methylphosphonate over dry and wet rutile TiO< sub> 2</sub>. Journal of Photochemistry and Photobiology A: Chemistry, 184(1), 125-134. Kiyono, M., Williams, P. J., & Koros, W. J. (2010). Effect of pyrolysis atmosphere on separation performance of carbon molecular sieve membranes. Journal of Membrane Science, 359(1), 2-10. Koresh, J. E., Saggy, S., & Soffer, A. (1987). Separation device: Google Patents. Koresh, J., & Soffer, A. (1980). Study of molecular sieve carbons. Part 1.—Pore structure, gradual pore opening and mechanism of molecular sieving. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 76, 2457- 2471. Korobeinichev, O., Ilyin, S., Bolshova, T., Shvartsberg, V., & Chernov, A. (2000). The chemistry of the destruction of organophosphorus compounds in flames—III: The destruction of DMMP and TMP in a flame of hydrogen and oxygen. Combustion and flame, 121(4), 593-609. Lagorsse, S., Magalhaes, F., & Mendes, A. (2004). Carbon molecular sieve membranes: sorption, kinetic and structural characterization. Journal of Membrane Science, 241(2), 275-287. Lee, H.-C., Monji, M., Parsley, D., Sahimi, M., Liu, P., Egolfopoulos, F., & Tsotsis, T. (2012). Use of steam activation as a post-treatment technique in the preparation of carbon molecular sieve membranes. Industrial & Engineering Chemistry Research, 52(3), 1122- 1132. 93 Lee, K. Y., Houalla, M., Hercules, D. M., & Hall, W. K. (1994). Catalytic oxidative decomposition of dimethyl methylphosphonate over Cu-substituted hydroxyapatite. Journal of Catalysis, 145(1), 223-231. Li, K. (2007). Ceramic membranes for separation and reaction: John Wiley & Sons. Lin, H., Van Wagner, E., Freeman, B. D., Toy, L. G., & Gupta, R. P. (2006). Plasticization- enhanced hydrogen purification using polymeric membranes. Science, 311(5761), 639- 642. Linkov, V., Sanderson, R., & Jacobs, E. (1994). Highly asymmetrical carbon membranes. Journal of Membrane Science, 95(1), 93-99. Liu, P. K., Sahimi, M., & Tsotsis, T. T. (2012). Process intensification in hydrogen production from coal and biomass via the use of membrane-based reactive separations. Current Opinion in Chemical Engineering, 1(3), 342-351. Lu, Z., Maroto-Valer, M. M., & Schobert, H. H. (2008). Role of active sites in the steam activation of high unburned carbon fly ashes. Fuel, 87(12), 2598-2605. Menendez-Diaz, J., & Martin-Gullon, I. (2006). Types of carbon adsorbents and their production. Interface Science and Technology, 7, 1-47. Mera, N., Hirakawa, T., Sano, T., Takeuchi, K., Seto, Y., & Negishi, N. (2010). Removal of high concentration dimethyl methylphosphonate in the gas phase by repeated-batch reactions using TiO< sub> 2</sub>. Journal of hazardous materials, 177(1), 274-280. Mikuláŝek, P., Doleček, P., Šedá, H., & Cakl, J. (1994). Alumina ‐Ceramic Microfiltration Membranes: Preparation, Characterization and Some Properties. Developments in Chemical Engineering and Mineral Processing, 2(2‐3), 115-123. Mitchell, M. B., Sheinker, V., & Mintz, E. A. (1997). Adsorption and decomposition of dimethyl methylphosphonate on metal oxides. The Journal of Physical Chemistry B, 101(51), 11192-11203. Mitchell, M. B., Sheinker, V. N., Tesfamichael, A. B., Gatimu, E. N., & Nunley, M. (2003). Decomposition of dimethyl methylphosphonate (DMMP) on supported cerium and iron co-impregnated oxides at room temperature. The Journal of Physical Chemistry B, 107(2), 580-586. Molina-Sabio, M., Gonzalez, M., Rodriguez-Reinoso, F., & Sepúlveda-Escribano, A. (1996). Effect of steam and carbon dioxide activation in the micropore size distribution of activated carbon. Carbon, 34(4), 505-509. 94 Monji, M., Abedi, S., Pourmahdian, S., & Taromi, F. A. (2009). Effect of prepolymerization on propylene polymerization. Journal of Applied Polymer Science, 112(4), 1863-1867. Monji M., Egolfopoulos F., Tsotsis T. T., A comprehensive model for the capture of condensable vapors and transport through nanostructure membranes, under preparation , Journal of Membrane Science (2015). Moss, J. A., Szczepankiewicz, S. H., Park, E., & Hoffmann, M. R. (2005). Adsorption and photodegradation of dimethyl methylphosphonate vapor at TiO2 surfaces. The Journal of Physical Chemistry B, 109(42), 19779-19785. Motamedhashemi, M., Egolfopoulos, F., & Tsotsis, T. (2011). Application of a flow-through catalytic membrane reactor (FTCMR) for the destruction of a chemical warfare simulant. Journal of Membrane Science, 376(1), 119-131. Nishihara, Y., & Yoneyama, H. (1992). Carbon based porous hollow fiber membrane and method for producing same: Google Patents. Obee, T. N., & Satyapal, S. (1998). Photocatalytic decomposition of DMMP on titania. Journal of Photochemistry and Photobiology A: Chemistry, 118(1), 45-51. Ohbu, S., Yamashina, A., Takasu, N., Yamaguchi, T., Murai, T., Nakano, K., . . . Hinohara, S. (1997). Sarin poisoning on Tokyo subway. Southern medical journal, 90(6), 587-593. Pastor-Villegas, J., & Duran-Valle, C. (2002). Pore structure of activated carbons prepared by carbon dioxide and steam activation at different temperatures from extracted rockrose. Carbon, 40(3), 397-402. Rios, R., Martínez-Escandell, M., Molina-Sabio, M., & Rodríguez-Reinoso, F. (2006). Carbon foam prepared by pyrolysis of olive stones under steam. Carbon, 44(8), 1448-1454. Rodriguez-Reinoso, F., Molina-Sabio, M., & Gonzalez, M. (1995). The use of steam and CO< sub> 2</sub> as activating agents in the preparation of activated carbons. Carbon, 33(1), 15-23. Rusu, C. N., & Yates, J. T. (2000). Photooxidation of dimethyl methylphosphonate on TiO2 powder. The Journal of Physical Chemistry B, 104(51), 12299-12305. Sahimi, M., & Tsotsis, T. T. (1985). A percolation model of catalyst deactivation by site coverage and pore blockage. Journal of Catalysis, 96(2), 552-562. Sedigh, M. G., Jahangiri, M., Liu, P. K., Sahimi, M., & Tsotsis, T. T. (2000). Structural characterization of polyetherimide ‐based carbon molecular sieve membranes. AIChE Journal, 46(11), 2245-2255. 95 Sedigh, M. G., Onstot, W. J., Xu, L., Peng, W. L., Tsotsis, T. T., & Sahimi, M. (1998). Experiments and simulation of transport and separation of gas mixtures in carbon molecular sieve membranes. The Journal of Physical Chemistry A, 102(44), 8580-8589. Sedigh, M. G., Xu, L., Tsotsis, T. T., & Sahimi, M. (1999). Transport and morphological characteristics of polyetherimide-based carbon molecular sieve membranes. Industrial & Engineering Chemistry Research, 38(9), 3367-3380. Segal, S. R., Cao, L., Suib, S. L., Tang, X., & Satyapal, S. (2001). Thermal decomposition of dimethyl methylphosphonate over manganese oxide catalysts. Journal of Catalysis, 198(1), 66-76. Shiflett, M. B., & Foley, H. C. (2001). Reproducible production of nanoporous carbon membranes. Carbon, 39(9), 1421-1425. Shusen, W., Meiyun, Z., & Zhizhong, W. (1996). Asymmetric molecular sieve carbon membranes. Journal of Membrane Science, 109(2), 267-270. Singh, A., & Koros, W. (1996). Significance of entropic selectivity for advanced gas separation membranes. Industrial & Engineering Chemistry Research, 35(4), 1231-1234. Singh, A., & Lal, D. (2010). Preparation and characterization of activated carbon spheres from polystyrene sulphonate beads by steam and carbon dioxide activation. Journal of Applied Polymer Science, 115(4), 2409-2415. Smentkowski, V., Hagans, P., & Yates Jr, J. (1988). Study of the catalytic destruction of dimethyl methylphosphonate (DMMP): oxidation over molybdenum (110). The Journal of Physical Chemistry, 92(22), 6351-6357. Suda, H., & Haraya, K. (1997). Alkene/alkane permselectivities of a carbon molecular sievemembrane. Chem. Commun.(1), 93-94. Thaeron, C., Parrillo, D., Sircar, S., Clarke, P., Paranjape, M., & Pruden, B. (1999). Separation of hydrogen sulfide–methane mixtures by selective surface flow membrane. Separation and purification technology, 15(2), 121-129. Trubitsyn, D. A., & Vorontsov, A. V. (2005). Experimental study of dimethyl methylphosphonate decomposition over anatase TiO2. The Journal of Physical Chemistry B, 109(46), 21884-21892. Tsyntsarski, B., Petrova, B., Budinova, T., Petrov, N., Velasco, L. F., Parra, J. B., & Ania, C. O. (2012). Porosity development during steam activation of carbon foams from chemically modified pitch. Microporous and Mesoporous Materials, 154, 56-61. 96 Uhm, H. S., Cho, S. C., Hong, Y. C., Park, Y. G., & Park, J. S. (2008). Destruction of dimethyl methylphosphonate using a microwave plasma torch. Applied Physics Letters, 92(7), 071503. Vorontsov, A. V., Davydov, L., Reddy, E. P., Lion, C., Savinov, E. N., & Smirniotis, P. G. (2002). Routes of photocatalytic destruction of chemical warfare agent simulants. New journal of chemistry, 26(6), 732-744. Walker Jr, P. (1996). Production of activated carbons: use of CO< sub> 2</sub> versus H< sub> 2</sub> O as activating agent. Carbon, 34(10), 1297-1299. Wang, T., Zhang, B., Qiu, J., Wu, Y., Zhang, S., & Cao, Y. (2009). Effects of sulfone/ketone in poly (phthalazinone ether sulfone ketone) on the gas permeation of their derived carbon membranes. Journal of Membrane Science, 330(1), 319-325. Wei, W., Qin, G., Hu, H., You, L., & Chen, G. (2007). Preparation of supported carbon molecular sieve membrane from novolac phenol–formaldehyde resin. Journal of Membrane Science, 303(1), 80-85. Werner, J. H., & Cool, T. A. (1999). Kinetic model for the decomposition of DMMP in a hydrogen/oxygen flame. Combustion and flame, 117(1), 78-98. Xiao, Y., Chng, M. L., Chung, T.-S., Toriida, M., Tamai, S., Chen, H., & Jean, Y. (2010). Asymmetric structure and enhanced gas separation performance induced by in situ growth of silver nanoparticles in carbon membranes. Carbon, 48(2), 408-416. Yanagisawa, N., Morita, H., Nakajima, T., Okudera, H., Shimizu, M., Hirabayashi, H., . . . Mimura, S. (1995). Sarin poisoning in Matsumoto, Japan. The Lancet, 346(8970), 290- 293. Zhang, X., Hu, H., Zhu, Y., & Zhu, S. (2006). Effect of carbon molecular sieve on phenol formaldehyde novolac resin based carbon membranes. Separation and purification technology, 52(2), 261-265.
Abstract (if available)
Abstract
The potential for the use of chemical weapons has increased in recent times. The successful use of a flow‐through catalytic membrane reactor (FTCMR) as a protection system against a chemical warfare agent (CWA) was recently reported by this group. This FTCMR, employing a single catalytically active mesoporous membrane, was applied for the destruction of dimethyl methylphosphonate (DMMP), which is known as a chemical precursor (and used to simulate its characteristics) for Sarin (GB), a toxic CWA. In this paper, results are reported of efforts to scale‐up this lab‐scale FTCMR towards practical individual protection (IP) and collective protection (CP) applications. A multi‐tubular FTCMR (MFTCMR) has been developed and tested towards the destruction of DMMP. The MFTCMR has been shown quite effective, providing extended protection periods against this CWA simulant. A key part of the research effort focused on improving the preparation of the catalytically active membranes to prepare membrane bundles with reproducible transport characteristics and reactivity. A key challenge for the large‐scale production of such devices, furthermore, remains the development of a simple non‐destructive test that assures that the produced parts are appropriate for the proposed application and that they continue to remain active during their shelf‐life prior to their use. As part of this effort a simple propane‐in‐air light‐off test was studied, which was shown capable to track the reactivity and performance of the MFTCMR.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
A flow-through membrane reactor for destruction of a chemical warfare simulant
PDF
The use of carbon molecule sieve and Pd membranes for conventional and reactive applications
PDF
A high efficiency, ultra-compact process for pre-combustion CO₂ capture
PDF
Lab-scale and field-scale study of siloxane contaminants removal from landfill gas
PDF
Process intensification in hydrogen production via membrane-based reactive separations
PDF
Biogas reforming: conventional and reactive separation processes and the preparation and characterization of related materials
PDF
A study of the application of membrane-based reactive separation to the carbon dioxide methanation
PDF
Novel methods for landfill gas and biogas clean-up
PDF
Mesoscale SOFC-based power generator system: modeling and experiments
PDF
Methanol synthesis in the membrane reactor
PDF
Flame ignition studies of conventional and alternative jet fuels and surrogate components
PDF
A process-based molecular model of nano-porous silicon carbide membranes
PDF
Methanol synthesis in a membrane reactor
PDF
CFD design of jet-stirred chambers for turbulent flame and chemical kinetics experiments
PDF
Fabrication of nanoporous silicon carbide membranes for gas separation applications
PDF
Fabrication of nanoporous silicon oxycarbide materials via a sacrificial template technique
PDF
Experimental studies and computer simulation of the preparation of nanoporous silicon-carbide membranes by chemical-vapor infiltration/chemical-vapor deposition techniques
PDF
Fabrication of silicon-based membranes via vapor-phase deposition and pyrolysis of organosilicon polymers
PDF
A hybrid adsorbent-membrane reactor (HAMR) system for hydrogen production
PDF
Preparation of polyetherimide nanoparticles by electrospray drying, and their use in the preparation of mixed-matrix carbon molecular-sieve (CMS) membranes
Asset Metadata
Creator
Monji, Majid
(author)
Core Title
The purification of contaminated air streams via recative separation techniques
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Chemical Engineering
Publication Date
02/23/2015
Defense Date
11/10/2014
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
chemical warfare agents,membrane,membrane reactor,multi‐tubular MFTCMR,OAI-PMH Harvest,reaction,separation
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Egolfopoulos, Fokion E. (
committee chair
), Tsotsis, Theodore T. (
committee chair
)
Creator Email
monji@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-535802
Unique identifier
UC11298387
Identifier
etd-MonjiMajid-3209.pdf (filename),usctheses-c3-535802 (legacy record id)
Legacy Identifier
etd-MonjiMajid-3209.pdf
Dmrecord
535802
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Monji, Majid
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
chemical warfare agents
membrane
membrane reactor
multi‐tubular MFTCMR