ABSTRACT Title of Dissertation: ADSORPTION AND MECHANISTIC STUDIES OF DIMETHYL METHYLPHOSPHONATE FOR CWA DEFEAT Kim L. Huynh, Doctor of Philosophy, 2018 Dissertation Directed By: Professor Bryan W. Eichhorn, Department of Chemistry and Biochemistry This thesis work systematically investigated carefully designed ASZM- TEDA deconstruction–reconstruction experiments. Our objective is to understand the role of each impregnant, the factors that influence filter performance, and filter lifetime. Chemical warfare agent simulant, dimethyl methylphosphonate (DMMP), is used to evaluate the adsorption properties and reactivity of different materials using a temperature programmed desorption (TPD)/quantitative 1H-NMR method, combined with inductively coupled plasma-atomic emission spectroscopy (ICP- AES) bulk analysis. Our investigation begins with a full analysis of the current commercial gas mask filter, ASZM-TEDA. In Chapter 2, studies with ordered mesoporous carbons (OMCs) and a disordered microporous carbon (BPL) reveal that surface area and pore volume dictate total adsorption loading. It was also discovered that an ordered pore network would lead to higher DMMP desorption energies by 30–40%, confirming stronger binding to the carbon surface in pores ≤2.4 nm. In Chapter 3, our investigation of an ASZM-TEDA deconstructed analog, carbon/CuO, further our understanding of CuO functionality within the adsorbent. Our study reveals that CuO impregnation increases adsorption capacity (up to 64%) in comparison to native carbon adsorbents. The CuO surface activity (reported as DMMP area capacity) was over 3.5 times higher in comparison to the OMCs (0.075–0.078 m2 g-1), and 17 times higher than BPL (0.016 m2 g-1). In addition, the extent of DMMP decomposition is greater for nano sized endo-pore CuO relative to larger exo-pore CuO particles. In Chapter 4, ordered mesoporous metal oxides (OMMs) are described with more reactive sites and designed to presumably have better mass transfer into the pores in comparison to impregnated carbons. DMMP chemisorption to the metal oxide reactive sites allow for greater decomposition (reported as the decomposed DMMP capacity), which is 2 to 4 times higher for the OMMs in comparison to OMC. TPD measurements for meso-Al2O3 and meso-Fe2O3 suggest that DMMP conversion to methanol is roughly one-to-one, whereas meso-TiO2 undergoes greater decomposition with the cleavage of a second methoxy group. It was also discovered that meso-Al2O3 achieved the highest DMMP total volume capacity in comparison to the other materials, making it the most efficient adsorbent studied for CWA defeat. ADSORPTION AND MECHANISTIC STUDIES OF DIMETHYL METHYLPHOSPHONATE FOR CWA DEFEAT By Kim Lien Huynh Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2018 Advisory Committee: Professor Bryan Eichhorn, Chair Professor Liangbing Hu, Dean’s Representative Professor Zhihong Nie Professor Efrain Rodriguez Professor Michael Zachariah © Copyright by Kim Lien Huynh 2018 Dedication I would like to dedicate my dissertation to my family, especially my parents, who immigrated to America so I could have this opportunity. This is one of the many ways I intend to honor their sacrifices. ii Acknowledgements I am deeply thankful for everyone who helped make this dissertation a possibility. I would like to first acknowledge Dr. Bryan Eichhorn for his guidance, support, and patience throughout my studies. I would also like to thank my current, former, and honorary lab mates, in particular order, Dr. Junkai Hu, Scott Holdren, Miles Rehwoldt, Dr. Aaron Geller, Luning Wang, Dr. Samantha DeCarlo, Dr. Will Gibbons, Dr. Yi Wang, and Lauren Stevens, as well as my friends, Amanda Lee, Steven Wolf, Marcus Carter, Shweta Ganapati, Sitara Chauhan, Tessy Thomas, Andrea Tobochnik, Andrea Castillo, Jessica Long, Rebecca Monterusso, Rachael Riley, Tara Michael, Kyle Webb, and many others, for our discussions and your friendship. In addition, thank you to my many collaborators at the University of Maryland (Dr. Maija Kuklja, Dr. Michael Zachariah), Johns Hopkins University (Dr. Kit Bowen), Naval Research Lab (Dr. Jeffrey Owrutsky, Dr. Jeffrey Long), Lawrence Berkley National Lab (Dr. Hendrik Bluhm), and their students. This work would not have been possible without funding from the Defense Threat Reduction Agency. Finally, I would like to acknowledge my advisory committee, Professor Michael Zachariah, Professor Efrain Rodriguez, Professor Zhihong Nie, and Professor Liangbing Hu. And last but not least, I would like to thank my dog, Copper (someone will have to read this to him). He contributed the most man hours (metaphorically) in helping me complete my thesis with his endless love and support. ii i Table of Contents Dedication ………………………………………………………………..ii Acknowledgments …………………..…………………………………..iii Table of Contents ……………………...……………………………..….iv List of Tables ………………...……………………………………….....vii List of Figures ...……………………………………………………..…...x Abbreviations ...……………………………………………..…..…..…xvi Chapter 1 Introduction and Background ………………...……...……..1 1.1 Sarin (GB) and its Simulant, Dimethyl Methylphosphonate (DMMP) …………………………………………………........2 1.2 DMMP Adsorption and Decomposition Mechanisms ….……..3 1.3 Porous Materials and their Characterization …..........................9 1.4 Adsorption Properties of Porous Materials ……………...…...18 1.5 Historical Attempts at Respirator Development ………..……19 1.6 ASZM-TEDA Design and Considerations …..……………...24 1.7 ASZM-TEDA Literature Background ………………...…….28 1.8 Motivation …………………...………………………………31 Chapter 2 Dimethyl Methylphosphonate Adsorption Behavior and Desorption Energies on Ordered Mesoporous Carbons........................32 2.1 Introduction……………………………..……………………32 2.2 Experimental Section ……..………………………………....35 2.2.1 Materials …………………………………………..…35 2.2.2 Microporous Carbons …………………….………….35 2.2.2.1 Synthesis of a Sucrose Based Microporous Carbon (SMC) ………………….................…35 2.2.2.2 Synthesis of a Zeolite Y-template Microporous Carbon (ZY) …………….….………..………36 2.2.3 Ordered Mesoporous Carbons (OMCs)………………37 2.2.3.1 Synthesis of 1D Cylindrical Ordered Mesoporous Carbon FDU-15................................................37 2.2.3.2 Synthesis of 3D Hexagonal Ordered Mesoporous Carbon CMK-3 ………….....………………38 2.2.3.3 Synthesis of 3D Bicontinuous Ordered Mesoporous Carbon CMK-8 ……..…….........39 2.2.4 Characterization of Carbon Materials ……………..…39 2.2.5 Exposure of Carbon Samples with Dimethyl Methylphosphonate (DMMP) ……………..…..…….40 2.2.5.1 Sample Preparation for 1H-NMR Study ……...40 2.2.5.2 Sample Preparation for Thermogravimetric Analysis …………………………………...…41 2.2.6 Fixed-Bed Reactor Setup ………………..…………..42 iv 2.2.7 Methods ……………………………………………...43 2.2.7.1 1H-NMR Study ………………………………43 2.2.7.2 Thermogravimetc Analysis (TGA) ………..…43 2.3 Results and Discussion ………………………………………45 2.3.1 Characterization of Ordered Mesoporous Carbons ..…45 2.3.2 DMMP Adsorption on Different Carbon Architectures ………………………………………………………..49 2.3.3 Temperature Programmed Desorption Studies of Different Carbon Architectures …………………...…51 2.3.4 Desorption Energies of DMMP ………………….…..54 2.4 Conclusions ………………………………………………….62 Chapter 3 Endo-pore and Exo-pore Copper Oxide Deposition in Ordered Mesoporous Carbons: Synthesis, Structures, and Adsorption/Decomposition Properties ………………………………..64 3.1 Introduction ………………………………………….………64 3.2 Experimental Section ……..……………………………...….66 3.2.1 Materials …………………………………………..…66 3.2.2 CuO Deposition on BPL ………………….……….…66 3.2.3 CuO Deposition for OMCs ………………..………....67 3.2.3.1 CuO Endo-pore and/or Exo-pore Impregnation …………………....…………………………..67 3.2.3.2 Selective Removal of Exo-pore CuO ……...…68 3.2.4 Characterization of Adsorbents ……………………...69 3.2.5 Exposure of Adsorbents to DMMP ………………….70 3.2.6 Fixed-Bed Reactor Setup …………………………….71 3.2.7 Methods ………………………………………...……72 3.2.7.1 Quantitative 1H-NMR .…………………….…72 3.2.7.2 ICP-AES Analysis …..…………………….…72 3.3 Results and Discussion ………………………………………73 3.3.1 Characterization of CuO Impregnated Carbons ……..73 3.3.1.1 CuO Particle Size, Distribution, and Morphology ……………………………………………..…73 3.3.1.2 Structural Properties of Carbon/CuO Composites ………………………………………….…….77 3.3.1.3 Determination of Cu Oxidation State and CuO Loading ……………….……………………..80 3.3.2 DMMP Adsorption Properties …………..…………..82 3.3.3 DMMP Decomposition Activity ……...……………..85 3.3.3.1 Determination of DMMP Conversion to Methanol (MeOH) ……….…………………..86 3.3.3.2 Thermal Decomposition of DMMP …….……88 3.3.3.3 Surface Reactivity Comparison for CuO and Carbon ………….……………………………92 3.3.3.4 Effect of CuO Size and Location on DMMP Decomposition ………………………..……..94 v 3.3.3.5 Significance of Structural Effects ……………95 3.4 Conclusions ………………………………………………….97 Chapter 4 Dimethyl Methylphosphonate Adsorption Properties and Decomposition Activity on Mesoporous Metal Oxides ………...…….98 4.1 Introduction ………………………………….………………98 4.2 Experimental Section ……..……………………..…………100 4.2.1 Materials ……………………………………………100 4.2.2 Mesoporous Metal Oxides …………………….……100 4.2.2.1 Synthesis of Mesoporous Alumina (meso-Al2O3) …………………………….……………….. 101 4.2.2.2 Synthesis of Mesoporous Titania (meso-TiO2) …………………………………………...… 101 4.2.2.3 Synthesis of Mesoporous Ceria (meso-CeO2) and Iron Oxide (meso-Fe2O3) ……..…………….102 4.2.2.4 Mesoporous Iron Oxide Aerogel (aerogel-Fe2O3) …………………………………………..…. 103 4.2.3 Characterization of Mesoporous Metal Oxides ……..104 4.2.4 Exposure of Mesoporous Metal Oxides to DMMP ...104 4.2.5 Fixed-Bed Reactor Setup ..………...…………..……105 4.2.6 Methods ……………………………………….……106 4.2.6.1 1H-NMR Study ……………………………..106 4.2.6.2 ICP-AES Study …….……………………….106 4.2.7.3 FTIR Spectroscopy …………………………107 4.2.7.4 DFT Calculations …………………………...107 4.3 Results and Discussion ……………………………………..107 4.3.1 Characterization of Mesoporous Metal Oxides …..…107 4.3.2 DMMP Decomposition Activity .…………………..117 4.3.2.1 Determination of DMMP Conversion to Methanol ………………………...…………119 4.3.3 Evaluation of DMMP Adsorption/Decomposition Performance ……….…….…………………………121 4.4 Conclusions …………………………………………..…… 129 Chapter 5 Future Work and Outlook ………….………………..……131 Publications ……………………………………………………………134 References …………………………………………………..…………135 v i List of Tables Chapter 1: Table 1.1. Typical Absorptive Values of Different Charcoals Against Various Gases. Taken from Reference 64 ……………...……21 Table 1.2. Gas Mask Requirements and ASZM-TEDA Design ………..26 Table 1.3. ASZM-TEDA Impregnants and their Function …………….31 Chapter 2: Table 2.1. N2 Adsorption –Desorption Properties of OMCs and BPL Carbon ………………………………………………………49 Table 2.2. R2 Coefficients Measured Based on the Linear Fit of DMMP Remaining on the Different Carbons ………………………..56 Chapter 3: Table 3.1. N2 Adsorption–Desorption Properties of Unimpregnated and CuO Loaded Carbons ………………………………………..79 Table 3.2. Weight % Loading of CuO Impregnated Carbons ……….…80 Table 3.3. 70 h Mass Capacity of Decomposed DMMP, Desorbed DMMP, and Total Mass Capacity for Unimpregnated Carbons and Carbon/CuO Composites………………….……………84 Table 3.4. Phosphorus Concentration of Cycled Samples ……….…….87 Table 3.5. 70 h Mass Capacity of Decomposed DMMP Not Including Methanol Production from Thermal Decomposition, Desorbed DMMP, and Total Mass Capacity for Unimpregnated Carbons and Carbon/CuO Composites ………………………………90 Table 3.6. 5 h Area Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for the Carbons and CuO Nanoparticles ………………………………………………93 Table 3.7. DMMP Decomposition Capacity (mg g-1) per g CuO (wt%) ………………………………………………………………95 vi i Chapter 4: Table 4.1. N2 Adsorption-Desorption Properties of Adsorbents…….…116 Table 4.2. Phosphorus Concentration of Cycled Mesoporous Metal Oxides …………………………………………..…………………..120 Table 4.3. 5 h Mass Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for Various Adsorbents ……..…..123 Table 4.4. 5 h Area Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for Various Adsorbents …….…..126 Table 4.5. 5 h Volume Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for Various Adsorbents ……………………………………………………………...128 Chapter 5: Table 5.1. List of CWA simulants, sulfur mustard (HD/HT), sarin (GB), and VX………………………………...……………………133 vi ii List of Figures Chapter 1: Figure 1.1. Chemical warfare agent, (a) GB and its simulant (b) DMMP. Figure 1.2. Proposed pathways for DMMP decomposition on CuO surface. Taken from Reference 31. Figure 1.3. Illustration of the gas sorption process. Figure 1.4. Classification of physisorption isotherms. Taken from Reference 49. Isotherm types I-VI are described in the text. Figure 1.5. Illustration of the capillary condensation and evaporation processes in mesopores. Figure 1.6. Classification of hysteresis loops. Taken from Reference 49. See text for definitions. Figure 1.7. Photomicrographs of carbonized charcoal before and after 31 minutes of steam activation. “Magnified 732 Diameters.” Taken from Reference 64. Figure 1.8. Modern-day gas mask with an ASZM-TEDA canister. Figure 1.9. Schematic of a frustum layered canister with different sized ASZM-TEDA granules in a carbon bed. Adapted from Reference 85. Chapter 2: Figure 2.1. (a) Dimethyl methylphosphonate (DMMP), (b) Sarin (GB), and (c) Soman (GD). Figure 2.2. Schematic of the synthetic processes for the OMCs. x Figure 2.3. TEM images of (a) ordered mesoporous FDU-15, (b) ordered mesoporous CMK-3, (c) ordered mesoporous CMK-8 and (d) disordered microporous BPL. The inset cartoons show their general framework structures. Figure 2.4. (a) N2 adsorption-desorption isotherms and (b) BJH pore size distributions for the different carbon architectures, FDU-15, CMK-3, CMK- 8, and BPL carbon. Figure 2.5. Plot of DMMP uptake as a function of time for the ordered mesoporous carbons and microporous BPL carbon. The inset shows initial uptake of DMMP in the first 60 h before the plateau region. The dotted lines represent extrapolations from the 60 h data to the fully saturated systems. The time-to-saturation was not measured in these systems. Figure 2.6. The adsorption capacity after 800 h of DMMP exposure for different carbons (red dots) with different total pore volumes. A fitted linear curve of adsorption capacity versus pore volume (blue line) was added. Figure 2.7. Adsorption capacity after (a) 70 h and (b) 5 h of DMMP exposure in relationship to the total surface area of the different carbon structures. Figure 2.8. Adsorption capacity after 1 h of DMMP exposure in relationship to the mesopore surface area of the different carbon structures. Figure 2.9. DMMP uptake versus cumulative pore volume ≤ 2.4 nm. Figure 2.10. (a) TGA curves of DMMP exposed carbon samples at 10 °C min-1. TGA curves were normalized to 100 wt% after a 60 min Ar purge. The amount of DMMP desorbed follows the same trend as the cumulative x pore volume of pores ≤ 2.4 nm (b) NLDFT N2 adsorption pore volume measurements for different carbons tested. Figure 2.11. TGA of SMC (red) and DMMP exposed SMC (black). There is a slightly larger weight loss for the DMMP exposed SMC sample. Figure 2.12. (a) TGA curves of DMMP desorption from BPL carbon at different heating rates. The TGA curves were normalized to 100 wt% after the second purge (60 min under 100 ml min-1 of Ar). (b) Representative desorption energy determination of DMMP desorption from BPL carbon at 25% DMMP desorbed. The error bars in plot (b) are smaller than the symbols. Figure 2.13. DMMP desorption energies from carbon samples. Chapter 3: Figure 3.1. Schematic of the impregnation process for copper oxide into/onto the OMC pores. Figure 3.2. Schematic of the selective removal of exo-pre CuO from FDU- 15/CuO. Figure 3.3. TEM and SEM images of the carbon/CuO composites: (a) BPL/CuO (exo-pore), (b) FDU-15/CuO (endo and exo-pore), (c) FDU- 15/CuO (endo-pore), (d) CMK-3/CuO (endo-pore), and (e) CMK-8/CuO (endo-pore). Figure 3.4. SEM image of (a) BPL/CuO with (b) Cu (red) and C (light green) phase maps. (d) is a zoomed in SEM image of (a) to compare to (e) mixed BPL and commercial CuO nanoparticles. Figure 3.5. Cu wt% calculated using ICP-AES and XPS measurements for OMC/CuO composites. x i Figure 3.6. BJH pore size distributions to compare the bare carbons and the carbon/CuO composites: (a) BPL, BPL/CuO, (b) FDU-15, FDU-15/CuO, (c) CMK-3, CMK-3/CuO, and (d) CMK-8, CMK-8/CuO. Figure 3.7. (a) XRD diffraction patterns and (b) BJH pore size distributions of endo and exo-pore (red) and endo-pore only CuO impregnated FDU-15 (blue). Figure 3.8. C2-XRD diffraction pattern of FDU-15/CuO (black), CMK- 3/CuO (red), and CMK-8/CuO (blue) composites. The peaks correspond to carbon and CuO Tenorite (PDF 01-073-6023). Figure 3.9. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after TPD experiments with the unimpregnated carbons (green) and carbon/CuO composites (blue) after 70 h of DMMP exposure. Figure 3.10. Proposed pathways for DMMP decomposition on CuO surface. Taken from Reference 31. Figure 3.11. Illustration of DMMP decomposition to produce gaseous methanol and surface-bound methyl methylphosphonate (MMP). Figure 3.12. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after multiple TPD experiments with FDU-15 after 5 h, 25 h, 70 h, and 2 weeks of DMMP exposure. Figure 3.13. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after a TPD experiment with the different unimpregnated (green) and CuO loaded carbons (blue) after approximately 70 h of DMMP exposure. xi i Figure 3.14. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after a TPD experiment with BPL, FDU-15, CMK-3, CMK-8, and CuO nanoparticles after 5 h of DMMP exposure. Figure 3.15. Representation of the limited diffusion of DMMP into the CuO loaded 1D pores of FDU-15, in comparison to the more accessible 3D pores of CMK-3 and CMK-8. Chapter 4: Figure 4.1. Schematic of ordered meso-Al2O3/meso-TiO2 synthesis using a soft template method. Figure 4.2. Schematic of ordered meso-CeO2/ meso-Fe2O3 synthesis using an ordered mesoporous silica template, SBA-15. Figure 4.3. Schematic of aerogel-Fe2O3 synthesis using a sol-gel method. Figure 4.4. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of amorphous meso-Al2O3. Figure 4.5. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline anatase and rutile phase meso-TiO2. Figure 4.6. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline cerianite phase meso-CeO2. Figure 4.7. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline hematite phase meso-Fe2O3. xi ii Figure 4.8. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline hematite phase aerogel-Fe2O3. Figure 4.9. BJH pore size distributions for (a) meso-Al2O3 (pink), meso- TiO2 (blue), and FDU-15 (gray), as well as (b) meso-CeO2 (green), meso- Fe2O3 (brown), and CMK-3 (gray). Figure 4.10. BJH pore size distributions for the diverse mesoporous metal oxides, meso-Al2O3 (pink), meso-TiO2 (blue), meso-CeO2 (green), meso- Fe2O3 (orange), and aerogel-Fe2O3 (red). Figure 4.11. Graphical representation of the surface area and pore volume of the OMMs, meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and aerogel-Fe2O3. Figure 4.12. General schematic of DMMP decomposition on hydroxylated metal oxides, and the possible surface and gaseous byproducts. Figure 4.13. DRIFTS spectrum of gas phase DMMP (black line) in comparison to commercial DMMP adsorbed on meso-TiO2 (blue line) and meso-CeO2 (red line). Figure 4.14. DMMP chemisorption on (a) rutile TiO2 (110) surface and (b) α-Al2O3 (0001) surface. Preliminary data by Dr. Roman Tsyshevsky. Figure 4.15. Graphical representation of the DMMP mass capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, and BPL. Figure 4.16. Graphical representation of the DMMP area capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for xi v four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, and BPL. Figure 4.17. Graphical representation of the DMMP volume capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, and BPL. Figure 4.18. Graphical representation of the DMMP (a) mass, (b) area, and (c) volume capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, BPL, and ASZM- TEDA. Chapter 5: Figure 5.1. Schematic of a frustum layered canister with different sized ASZM-TEDA granules (to the left)85 and a new canister with a mixture of ASZM-TEDA and meso-TiO2/CuO (to the right). Figure 5.2. (a) Dimethyl methylphosphonate (DMMP), (b, c) Deuterium labelled DMMP, with possible bond cleavages are indicated with a dotted line. x v Abbreviations AC: hydrogen cyanide AES: auger electron spectroscopy APXPS: ambient pressure X-ray photoelectron spectroscopy ASC: copper (A), silver (S), chromium (C) impregnated carbon AsH3: arsine ASZ: copper (A), silver (S), zinc (Z) impregnated carbon ASZM-TEDA: copper (A), silver (S), zinc (Z), molybdenum (M), and triethylene diamine (TEDA) impregnated carbon BET: Brunauer-Emmett-Teller method BJH: Barrett-Joyner-Halenda method BPL: bituminous coal-based activated carbon CD3CN: deuterated acetonitrile/acetonitrile-d3 CG: phosgene CH3I: methyl iodide CH3OCH3: dimethyl ether CH3OH/MeOH: methanol CHCOOH: formic acid Chemisorption: chemical adsorption/reaction CHO: formaldehyde CK: cyanogen chloride CO: carbon monoxide CO2: carbon dioxide CWA: chemical warfare agent xv i DFT: density functional theory DMMP: dimethyl methylphosphonate DRIFTS: diffuse reflectance infrared Fourier transform spectroscopy EC: ethylene carbonate EDX: energy-dispersive X-ray spectroscopy Endo-pore: inside the porous network Exo-pore: outside the porous network GB: Sarin, also known as (RS)-Propan-2-yl methylphosphono- fluoridate GCMS: grand canonical Monte Carlo simulations 1H-NMR: proton nuclear magnetic resonance H2S: hydrogen sulfide HCl: hydrochloric acid HCN: hydrogen cyanide HNO3: nitric acid ΔHvap: heat of vaporization ICP-AES: inductively coupled plasma-atomic emission spectroscopy Macropores: pore diameter > 50 nm Mesopores: 50 nm > pore diameter > 2 nm Micropores: pore diameter < 2 nm MMP: methyl methylphosphonate MP: methylphosphonate NLDFT: nonlocal density functional theory Olat: lattice oxygen xv ii OMC: ordered mesoporous carbon OMM: ordered mesoporous metal oxide P: equilibrium pressure P/P0: partial pressure P0: saturation pressure PAW: projector augmented-wave PH3: phosphine Physisorption: physical adsorption/interaction Pluronic F127: triblock copolymer, EO106PO70EO106, where EO = ethylene oxide, PO = propylene oxide Pluronic P123: diblock copolymer, EO20PO70EO20, where EO = ethylene oxide, PO = propylene oxide PSD: pore size distribution SA: arsine SCCM: standard cubic centimeters per minute SMC: disordered sucrose-based microporous carbon SO2: sulfur dioxide TEDA: triethylenediamine, also known as 1,4-diazabicyclo-2,2,2- octane TEOS: tetraethyl orthosilicate TGA: thermogravimetric analysis TPD: temperature programmed desorption UHV: ultra-high vacuum Whetlerite: type of activated carbon ZY: zeolite Y-template microporous carbon xv iii Chapter 1: Introduction and Background Recent terrorist attacks and crime wars have increased the demand for long-lasting, highly reactive, cost efficient gas masks for protection against chemical warfare agents (CWAs). Our efforts are directed at removing a specific CWA, sarin (GB, also known as (RS)-Propan-2-yl methylphosphono-fluoridate), which is a lethal, toxic nerve agent that poses a current threat. For safety reasons, the GB simulant, dimethyl methylphosphonate (DMMP) will be used in our studies (Section 1.1). Prior investigations of DMMP activity on model systems provide a general understanding of DMMP adsorption properties and decomposition mechanisms (Section 1.2). However, these studies relied on surface analytical techniques that require ultra-high vacuum (UHV), which does not mimic operating conditions with realistic materials. We will investigate DMMP interactions on porous materials, a typical filter support, under ambient pressure. Our studies begin with a thorough characterization of the materials’ pore properties (Section 1.3), which play a significant role in our understanding of the adsorption process (Section 1.4). Continuous research via an Edisonian approach has led to the invention of ASZM-TEDA, the current filter used in gas masks (Section 1.5). However, due to 1 numerous disadvantages including heavy weight, low breathability, and deactivation after surface poisoning (Section 1.6), we seek to improve ASZM- TEDA. Previous studies about ASZM-TEDA (Section 1.7) will help guide us in our investigation to develop, new superior filters to aid in the removal of CWAs (Section 1.8). 1.1 Sarin (GB) and its Simulant, Dimethyl Methylphosphonate (DMMP) GB is a toxic nerve agent that is lethal at low concentrations. Despite being banned by the Chemical Weapons Convention in 1993,1 GB was still used in multiple terrorist attacks (1994: Matsumoto, 1995: Tokyo)2-4 and crime wars (1980-1988: Iran-Iraq war, 2011-present: Syrian civil war), killing thousands of soldiers and civilians. Since GB is relatively easy to synthesize and readily accessible, protection against GB will continue to be a persistent problem for years to come.5 Therefore, we need a long-lasting, highly reactive, and cost efficient gas mask for protection against chemical warfare. Our research objective is to study the adsorption and decomposition activity of a GB simulant to reveal crucial factors needed to rationally design new, more effective gas masks for CWA defeat. For safety reasons, DMMP is commonly studied due to its comparable functional groups, shown in Figure 1.1.6- 22 Due to similar physical and chemical properties, DMMP is used to study sorption properties and decomposition mechanisms on carbon and metal oxide 2 surfaces.23-25 O O H3C P O H3C P OCH3 F OCH3 (a) (b) Figure 1.1. Chemical warfare agent, (a) GB and its simulant (b) DMMP. DMMP and GB possess similar functional groups that interact with the adsorbent through the phosphoryl and/or alkoxide groups; therefore, DMMP is an excellent adsorption simulant.26-29 However, the decomposition of DMMP compared to GB is predicted to be more difficult to ascertain.30 Due to the absence of the highly reactive fluorine atom, it was unclear whether DMMP decomposition on any particular surface accurately mimics GB decomposition.31, 32 However, theoretical studies have shown that GB undergoes similar decomposition pathways proposed for DMMP.27-29 1.2 DMMP Adsorption and Decomposition Mechanisms Current systems remove DMMP from the airstream via two mechanisms, physical (molecular) adsorption and chemical (dissociative) reactions with dry/hydroxylated metal oxide surfaces (Scheme 1.1).33 The chemisorption between DMMP and the adsorbent surface can occur at two sites, an uncoordinated metal site (Scheme 1a) or a surface hydroxyl site (Scheme 1b).17, 34 The dissociative reaction with the Lewis acid metal site can result in DMMP 3 decomposition to form methoxy and methyl methylphosphonate (MMP) surface groups (Scheme 1c). DMMP decomposition has been observed on many dry and hydroxylated metal oxide surfaces, such as TiO ,2, 24, 26, 30 MgO,8, 16, 35-37 LaO,162 Al2O3,16, 17, 23, 25, 34, 38, 39 MnO ,22, 40 CuO,31 Fe O ,14, 163 2 3 and Ni, Fe, Cu, and V oxide supported on γ- Al2O3, SiO , and TiO 252 2. Under certain conditions, DMMP can breakdown to surface phosphate, alkoxide, and alkyl groups, while subsequently forming gaseous byproducts, such as carbon dioxide (CO2), carbon monoxide (CO), dimethyl ether (CH3OCH3), formaldehyde (CHO), formic acid (CHCOOH), and more commonly, methanol (CH OH).2, 14-17, 23, 24, 26, 30, 34, 37, 383 H3COCH3 CH3 H3CO OCH3 P CH P 3 OCH3 H3CO OCH3 O O P + H O O M O M O M M O M O M O M O M O (a) (b) H3CO CH3 P O OCH3 O M O M O M (c) Scheme 1.1. Structures (a) and (b) are two possible molecular chemisorption modes and (c) is the dissociative chemisorption reaction, proposed by Templeton et al. The latter surface structure forms through the cleavage of the phosphorus- oxygen bond. Adapted from Reference 17 and 26. 4 Templeton et al. studied DMMP decomposition on Al2O3 using inelastic electron tunneling spectroscopy.16, 17, 34 These investigators reported initial binding of the phosphoryl oxygen at a coordinately unsaturated aluminum atom (Lewis acid site), followed by nucleophilic substitution at phosphorus by a surface hydroxyl group (Brønsted-Lowry acid site). The methoxy group (the better leaving group) was cleaved, which leads to the formation of gaseous CH3OH, leaving behind a bridging phosphonate byproduct by 200 °C (Scheme 1.2). By 400 °C, a surface oxygen atom performs a second nucleophilic substitution at the remaining methoxy carbon with a simultaneous protonation (from a surface hydroxyl group) to form another molecule of methanol and surface-bound methylphosphonate (MP). This decomposition behavior for DMMP was also observed on MgO and La O surfaces.16, 17, 342 3 CH3 CH CH CH CHO 3 3 O CH3 3 3 O H3C P O H3C P O H3C P O H O O H O O H O O O M O M O M O M O M O M O M O M O M Scheme 1.2. Proposed nucleophilic attack of the Al2O3 surface-bound hydroxyl group on the electrophilic phosphoryl group of DMMP. Adapted from Reference 16. Mitchell et al. studied DMMP adsorption on Fe2O3 using diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS). These investigators observed DMMP decomposition on Fe2O3 through the formation of a different intermediate, when compared to the other metal oxides (Al2O3, MgO, 5 and La2O3). By 100 °C under UHV conditions, a surface hydroxyl group attacks the phosphorus atom of the phosphoryl group. Subsequent reformation of the phosphoryl group results in a gaseous CH3OH byproduct, illustrated in Scheme 1.3. O O H3C CH3 H C CH3 O CH3 3H3C P O P O P O H3C H H3C H H3CO O O O O HO O M O M O M O M O M O M O M O M O M Scheme 1.3. Proposed nucleophilic attack of the Fe2O3 surface-bound hydroxyl group on the electrophilic phosphoryl group of DMMP. Adapted from Reference 16. By 300 °C under UHV conditions, all carbon containing substituents are lost with no selectivity over leaving group (−OCH3 or −CH3) priority. Due to the multiple oxidation states of iron, the lattice oxygen (Olat) from Fe2O3 oxidizes the P−CH3 bond through the Mars-Van Krevelen mechanism.14, 16, 41-44 In this mechanism, the surface is an active part of the reaction to form gaseous CH3OH, which results in an oxygen vacancy on the Fe2O3 surface.43 Henderson et al. also observed P−CH3 cleavage on 𝛼 −Fe2O3 using temperature programmed desorption (TPD) and auger electron spectroscopy (AES), presented in Scheme 1.4. DMMP adsorption began at 170 K and oxidation of the P−CH3 bond was observed upon heating to 250 K. By 600 K, decomposition and desorption of all carbon-containing species (CH3OH, CO2, 6 CO, and HCOOH) were observed, some of which were oxidized by lattice oxygens. The phosphorous byproduct then migrated into the bulk leaving the surface available for additional adsorption and decomposition. Additional cycles showed lower DMMP decomposition (~20% of the original level), most likely a result of some phosphorus species deactivating the surface.14 Scheme 1.4. Proposed stepwise decomposition of DMMP on Fe2O3 surface. Taken from Reference 14. 7 Trotochaud et al. proposed several pathways for DMMP decomposition on CuO surfaces using in situ ambient pressure X-ray photoelectron spectroscopy (APXPS) and DRIFTS, coupled with DFT calculations. These investigators observed DMMP cleavage at the P−CH3, P−OCH3, and PO−CH3 bonds, with the incorporation of lattice oxygens in the successive formation of Olat−CH3, Olat−P, and Cu−OCH3 groups, as shown in C4−C7 of Figure 1.2. 31 This study revealed several pathways for room temperature DMMP decomposition on CuO, leading to the formation of surface methoxy groups. Figure 1.2. Proposed pathways for DMMP decomposition on CuO surface. Taken from Reference 31. 8 Many prior investigations of DMMP sorption properties on model systems have relied on surface analytical techniques that require UHV conditions, which has proven to be a long-standing problem when translating to commercial filter systems. Specifically, major drawbacks include both material and pressure gaps.45, 46 Due to the material gap, the mechanistic details for DMMP on model systems such as, bulk single crystal, metal oxide foils, and clusters, will most likely differ from the real filter materials (metal/metal oxide impregnated porous carbons). Additionally, the pressure gap does not provide a realistic understanding of DMMP surface chemistry. Adsorption experiments performed under UHV (𝑝 ≈ 10'( 𝑏𝑎𝑟) have a pressure that is nine orders of magnitude higher than ambient pressure, contributing additional surface energy to each gas molecule. Despite these disadvantages, studies performed on model systems under UHV did contribute to our fundamental understanding of DMMP sorption properties on an atomic level. However, additional studies investigating how the surface behaves under operating conditions are needed to develop a comprehensive understanding of how DMMP adsorbs and decomposes on porous adsorbents. 1.3 Porous Materials and their Characterization Porous systems are of scientific interest due to their ability to interact with ions and molecules not only on their surface, but also throughout the bulk of the material. Porous materials, such as activated carbons, zeolites, and mesoporous metals, have been used for ion exchange, adsorption, and catalysis.47 9 Constitutionally, a porous system is comprised of a solid skeletal frame that contains voids, also known as pores.48 Pores are historically categorized into three classes: macropores (pore diameter > 50 nm), mesopores (50 nm > pore diameter > 2 nm), and micropores (pore diameter < 2 nm); however, materials containing different sized pores, defined as multiscale porous networks, have also been demonstrated. Examples of these materials are macroporous polymers, mesoporous carbons, and microporous ceramics. The porosity of these systems is calculated by the volume of the pores divided by the total volume occupied by the solid (eq. 1.1):49, 50 𝑃 = 0(23456) (1.1) 0(737) where P is porosity, V(pores) is the volume of the pores, and V(tot) is the total volume. A detailed characterization of a porous material is necessary to understand how their pore properties affect their performance as adsorbents. Pore size, volume, and surface area are determined from adsorption-desorption isotherms. These properties are measured after pre-treating the sample (by heat and vacuum) to remove all adsorbed contaminants, such as water or CO2, while avoiding irreversible changes to the surface or solid structure. Next, adsorption measurements are taken as a gas (N2, Ar, Kr, CO2, or H2) is dosed in controlled increments, with pressure allowed to reach equilibrium after each dose (Figure 10 1.3a).51 As the partial pressure (P/P0, where P = equilibrium pressure and P0 = saturation pressure) increases, gas molecules generally adsorb in three stages. First, adsorption occurs in the smallest dimension pores and proceeds until the monolayer capacity (𝑛:9), which is the amount of adsorbate required to cover the surface layer of an adsorbent, is achieved (Figure 1.3b). Continuous, sequential pore filling of the progressively larger pores results in multilayer adsorption, where multiple layers of adsorbed molecules form, which are not all in direct contact with the adsorbent surface (Figure 1.3c). Then, pressure continues to increase until the saturation pressure (P/P0 = 1), when adsorption ceases because the pores are completely filled, resulting in pore condensation (Figure 1.3d).52 (a) Pore Filling Begins Micropore <2nm Mesopore2−50 nm (b) Monolayer Formation (d) Pore Filling Completed (c) Continuous Pore Filling Figure 1.3. Illustration of the gas sorption process. Adapted from the Particle Technology Labs Website. 11 The relationship between the amount of gas adsorbed and P/P0, at a constant temperature, is the adsorption isotherm.49, 52 The Brunauer-Emmett- Teller (BET) method transforms the physisorption isotherm into a plot, which is applied to the BET equation (eq. 1.2): 2 2° = = + ?'= A2 (1.2) < =' < ? <2° > >? A° where 𝑛 is the amount adsorbed at a relative pressure, A , 𝐶 is the energy of the A° monolayer adsorption, and 𝑛9 is the monolayer capacity. A convenient way to 2 derive 𝑛9, is plotting the linear relationship between 2° 2 and A , which is used < =' A°2° to calculate the surface area (𝐴D) in eq. 1.3: 𝐴 = 𝑛:D 9 ∙ L ∙ 𝜎9 (1.3) where L is Avogadro constant, and 𝜎9 is the area occupied by the adsorbate. 49 The type of adsorption isotherm can also characterize the pore size and distribution of the material, which can largely be divided into six major classes (Figure 1.4): • Type I: characteristic of microporous solids. The steep uptake at very low P/P0 is a result of micropore filling. Materials with narrow micropores (< 1 nm) have Type I(a) isotherms, whereas materials with a broad pore size distribution (< 2.5 nm) have Type I(b) isotherms. 12 • Type II: characteristic of nonporous or macroporous adsorbents. If Point B in Figure 1.4 is steep, which occurs when the 𝐶 parameter in the BET range is ≥80, the extent of adsorption corresponds to monolayer coverage. A more gradual curvature, where B cannot be identified (𝐶 ≤ 50), indicates a mixture of monolayer and multilayer adsorption. • Type III: characteristic of nonporous or macroporous adsorbents. There is no identifiable monolayer formed. • Type IV: characteristic of mesoporous adsorbents. The two plateau regions represent monolayer and multilayer formation, followed by pore condensation. Materials with cylindrical pores (>4 nm) have Type IV(a) isotherms with hysteresis. Adsorbents with smaller conical and cylindrical mesopores have completely reversible Type IV(b) isotherms. • Type V: characteristic of water adsorption on hydrophobic microporous and mesoporous adsorbents. Molecular clustering is followed by pore filling. • Type VI: characteristic of layer-by-layer adsorption for highly uniform nonporous adsorbents. The step height represents the capacity of each adsorbed layer. 13 Figure 1.4. Classification of physisorption isotherms. Taken from Reference 49. Isotherm types I-VI are described in the text. 14 The desorption process begins after the pores are fully saturated. The gas pressure is reduced in increments to desorb the condensed gas from the system. If there is a difference in P/P0 between the recorded adsorption and desorption isotherms (Figure 1.4 IV(a) and V), hysteresis is observed.49 Hysteresis arises during the capillary condensation and evaporation process in the mesopores, illustrated in Figure 1.5.53 As molecules adsorb on a surface, they add layer by layer to fill high energy sites near the pore walls followed by low energy sites farther away (Figure 1.5a-c). When gas molecules accumulate on two opposing walls and area is minimized, the ensuing convex meniscus collapses into a thermodynamically lower energy state, resulting in capillary condensation (Figure 1.5d).52 During desorption, capillary evaporation occurs through a receding hemispherical (concave) meniscus at a pressure that is lower than the pore condensation pressure, resulting in hysteresis (Figure 1.5e). The delay observed in the vapor liquid transition is due to metastable adsorption layers and hindered nucleation of liquid bridges, which only occurs on the adsorption branch (Figure 1.5c-d).54-57 When the pore is filled by a condensate, the liquid-vapor interface is already present, and evaporation occurs without nucleation at lower P/P0 values (Figure 1.5e). The hysteresis ends when the adsorbed multilayer is in equilibrium with the vapor in the pore and the bulk gas phase (Figure 1.5f). 15 (a) (b) (c) convex concave (d) (e) (f) Figure 1.5. Illustration of the capillary condensation and evaporation processes in mesopores. 16 The type of hysteresis loop can characterize the pore size, shape, and network of the material, resulting in five main types observed in adsorption- desorption isotherms for porous materials (Figure 1.6):49, 50, 56-63 • Type H1: characteristic of uniform mesopores and non-connected cylindrical pores with narrow pore distribution. Also found in networks of ink-bottle pores. • Type H2: characteristic of complex pore structures with ill-defined shapes and wide pore distribution. The very steep desorption branch in the H2(a) loops can be attributed to pore- blocking/percolation, whereas the H2(b) is associated with pore blocking for a wider neck size. • Type H3: characteristic of plate-like particles giving rise to macropores, which are not completely filled with the adsorbate. • Type H4: characteristic of aggregated crystal zeolites, mesoporous zeolites, and micro-mesoporous carbon. • Type H5: characteristic of both open and partially blocked mesopores, such as plugged hexagonal templated silica. 17 Figure 1.6. Classification of hysteresis loops. Taken from Reference 49. Hysteresis loops (H1-H5) are described in the text. 1.4 Adsorption Properties of Porous Materials For porous materials, adsorption can occur on the external surface and within the pores. The adsorption behavior in micropores is dominated by interactions between the fluid molecules and the pore walls. The continuous adsorption process results in micropore filling, followed by multilayer adsorption on the small external area.49, 50, 57 Multilayer adsorption in mesopores depends on fluid-wall attraction and attractive interactions between the fluid molecules of the multilayers. The stability of the adsorbed multilayers is determined by long-range van der Waals interactions, surface tension, and curvature of the liquid-vapor interface.50, 62 18 Adsorption can be characterized as a physical interaction (physisorption) or a chemical reaction (chemisorption).49 Physisorption has been observed through weak intermolecular forces (e.g. van der Waals forces and long-range London dispersion forces)50 between the adsorbate and the adsorbent. Chemisorption involves the formation of covalent (or polar covalent) bonds, which can result in additional chemical reactions (discussed previously in Section 1.1). 1.5 Historical Attempts at Respirator Development “There is only one thing, and one thing only, that can save us our present and future health, the health of our descendants, and in many cases, our own lives… Never be without your Soldier’s Friend. The Soldier’s Friend is his Small Box Respirator.” - Taken from an unsigned document during World War I Protection against CWAs have been a persistent problem for military personnel since World War I, where chlorine gas (the first CWA used in warfare), killed over 6000 soldiers during the Battle of Gravenstafel Ridge on April 22, 1915.1, 64 Consequently, the demand for protection gave rise to the development of cost-effective and readily available respirators. Early designs consisted of a cotton pad/cloth soaked in a solution of sodium carbonate, sodium thiosulfate, and water.5 Unfortunately, these respirators only worked for a few minutes against a normal concentration of chlorine, which would later be improved and serve as inspiration for early gas mask designs.65 19 The first gas mask contained two features, namely a tight seal around the face, and a snout covering the nose and mouth that directed breathing from a singular hole in the face piece, which protected the user from toxic gases.5 The air flowing through the hole needs to be supplied or filtered. Since it is difficult to supply clean air on the field, the military utilized filters to clean the air before the user inhaled it.65 The first gas mask filter was a combination of charcoal carbon and soda lime.5, 64 Carbon based filters were commonly used because of its sizable network of channels and pores, which can adsorb large quantities of organic gases.66-70 The different types of charcoal carbons and their protection time against a variety of gases are listed in Table 1.1. Soda lime (hydrated lime, cement, kieselguhr, sodium hydroxide, and water) was added as an additional defense against inorganic contaminants.64, 65 The combination of charcoal carbon and soda lime was used in a small box respirator, which performed better than each individual component. The mixture decreased the negative effects of high temperature and humidity on the charcoal adsorption capacity, while increasing the soda lime reactivity.5 However, fine particulates found in toxic smoke/fumes still required additional layers of protection. Consequently, felt was incorporated to remove the fumes, but this was found to increase breathing resistance. To combat this, later efforts included carbon-impregnated/asbestos-impregnated filter paper; however later work demonstrated that prolonged inhalation of asbestos fibers led to fatal 20 illnesses, such as lung cancer, mesothelioma, and asbestosis, leading to its eventual ban in the 1970s.65 Table 1.1. Typical Absorptive Values of Different Charcoals Against Various Gases. Taken from Reference 64. Service Time, Minutes Standard Conditions No. Charcoal Nation 1 Poor cocoanut U.S.A 0 10 120 175 20 18 55 50 270 2 Medium cocoanut U.S.A 0 30 350 260 25 25 65 65 370 3 Good cocoanut U.S.A 0 60 620 310 27 30 75 70 420 4 Wet No. 2 U.S.A 12 18 320 330 35 16 35 95 5 Impregnated No. 2 U.S.A 0 35 400 700 70 400 70 190 510 6 Wood French 0 2.5 25 75 9 0 1 20 7 Wood British 0 6 70 90 18 4 5 30 8 Peach stone British 0 16 190 135 30 25 65 60 9 Treated wood German 0 42 230 105 20 20 22 25 10 Impregnated No. 9 German 30 9 90 320 16 1 110 120 Standard Conditions of Tests Mesh of absorbent……………………………………………............. 8-14 Depth of absorbent layer …………………………………………... 10 cm. Rate of flow per sq. cm. per min………………………………..…. 500 cc. Concentration of toxic gas …………………………….……… 0.1 percent Relative humidity ………………………………...……………. 50 percent Temperature …………………………………………………..…….. 20 °C Results expressed in minutes to the 99 percent efficiency points. Results corrected to uniform concentrations and size of particles. By the end of World War I, J.C. Whetzel and E. W. Fuller invented whetlerite, an activated carbon impregnated with copper, which was investigated as a suitable asbestos alternative. In this system, a process of heat and steam was 21 H2O Content, % Accel. Chloropicrin, Service Time, Min. Chloropicrin Phosgene Hydrocyanic Acid Arsine Cyanogen Chloride Trichloromethyl- chloroformate Chlorine used to “activate” the charcoal carbon5 by eliminating volatile compounds blocking the pores, increasing porosity, and the number of active sites (Figure 1.7).24, 64, 65 The copper impregnated activated carbon provided twice the protection of regular charcoal against phosgene (CG), triple the protection against hydrogen cyanide (AC), and ten times the protection against arsine (SA).68, 71-75 By World War II, whetlerite was used in the standard filter material known as whetlerite AS (80% whetlerite and 20% soda-lime). By 1943, whetlerite AS was further improved through the replacement of soda lime with other impregnants to make whetlerite ASC (copper (A), silver (S), and hexavalent chromium (C)), which provided greater protection against SA, AC, and cyanogen chloride (CK). 66, 68, 76-78 Before After Figure 1.7. Photomicrographs of carbonized charcoal before and after 31 minutes of steam activation. “Magnified 732 Diameters.” Taken from Reference 64. 22 ASC filters were widely used until the 1980s, when they were discontinued due to the cancer-producing effects of the Cr(VI) species, which also made disposal of used carbon fibers difficult and expensive.71, 77, 79-83 Consequently, chromium was replaced with a combination of zinc, molybdenum, and triethylenediamine (TEDA, also known as 1,4-diazabicyclo-2,2,2-octane),66, 77, 84 in the development of an upgraded filter. Calgon Carbon designed ASZM-TEDA (copper (A), silver metal (S), zinc (Z), molybdenum (M), and TEDA) to be a chromium-free replacement of whetlerite ASC. ASZM-TEDA is still widely used in gas mask canisters to remove toxic gases and/or vapors in military, industrial, and other applications (Figure 1.8). However, due to the sensitive nature of the material, there is very little published in academia. While the literature lists concentrations of ASZM- TEDA impregnants, it lacks vital “sensitive” data, such as the properties of the impregnants (oxidation state, size, location, etc.), which is necessary to have a comprehensive understanding of the surface chemistry. While government labs may have access to this sensitive information, their research primarily focuses on performance. Most of the published data on ASZM-TEDA is disclosed in patents and government documents, which discuss design and performance, such as bed height vs. pressure drop, layering and particle size, axial vs. radial flow, gas life challenge/breakthrough, and filter lifetimes.85-90 23 Figure 1.8. Modern-day gas mask with an ASZM-TEDA canister. 1.6 ASZM-TEDA Design and Considerations The logic behind the design and performance of gas mask filters is important to recognize before considering a replacement. The current military mask composition, ASZM-TEDA, is a compromise, embodying an optimum balance among many requirements listed in Table 1.2.5 One of the requirements is maximum protection, which is directly related to the amount of material used, and the capacity of the mechanical filter. However, the more material used, the larger and heavier the canister, which impairs the user’s vision and not practical for field use. 24 Higher packing density and deeper bed height can be applied to increase filter protection time. However, this would also increase the pressure drop (breathing resistance). Although this pressure drop can be lowered by increasing the diameter of the bed, this will also increase the size of the canister, restricting the movement and vision for the user. By 1997, Newton et al. designed a frustum layered canister, utilizing different sized ASZM-TEDA granules in a carbon bed. This new design demonstrated a lower pressure drop in comparison to the cylinder canister, and increased performance for CK removal (Figure 1.9).85 A bed consisting of a 0.4 cm layer of 12x30 mesh carbon followed by a 0.6 cm layer of 30x40 mesh carbon offered a 25% increase in CK protection time and a 14% reduction in airflow resistance, when compared to a 1.0 cm bed depth of 30x40 mesh carbon.87 The higher protection time was achieved by a greater mass transfer rate, which is the movement of adsorbate (toxic molecule) into the adsorbent. Separation of ASZM- TEDA particles by mesh size decreased the packing density and in turn, increased the mass transfer rate. 25 Table 1.2. Gas Mask Requirements and ASZM-TEDA Design Gas Mask Requirements ASZM-TEDA Ref. High-level filtration/protection Removes cyanogen chloride (CK), against a broad spectrum of methyl iodide (CH3I), hydrogen 23, 71, CWAs sulfide (H2S), sulfur dioxide (SO2), 73, 76, hydrogen cyanide (HCN), chlorine 81, 84, (Cl2), arsine (AsH3), phosphine 91 (PH3), organophosphonate esters, and acid-gas producers Changed canister shape (cylinder to Low breathing resistance frustrum), decreased packing density, layered different size granules in the carbon bed for a 85 lower pressure drop Does not interfere with vision, Delicate balance of chemicals for small, light-weight, comfortable the best protection in a reasonable 5 canister size Able to withstand field Degrades from prolonged exposure conditions with a service life for to humid air, even worse with low several months, and long shelf- levels of airborne contaminants, 92 life (several years) such as, SOx, NOx, and fuel vapors, which block active sites Easy to manufacture and cost- effective Made in Kg quantities, $9/pound 93 26 Particulate filter/ Water repellant glass fiber paper Mesh ASZM-T (large particles) Mesh ASZM-T (small particles) Dust screen/ Water repellant glass fiber paper Figure 1.9. Schematic of a frustum layered canister with different sized ASZM- TEDA granules in a carbon bed. Adapted from Reference 85. When freshly prepared, ASZM-TEDA has a high capacity for the removal of CWAs and selected TICs. However, ASZM-TEDA degrades from prolonged exposure to humid air, and is further hastened when exposed to humid air containing low levels of airborne contaminants, such as, SOx, NOx, and fuel vapors. Degradation of ASZM-TEDA results in a decreased protection capability of the filter and costly filter change-outs.91 Although much work was performed to develop the current filters via an Edisonian approach, a more fundamental approach towards understanding the mechanisms of CWA removal is needed for a rational development of new, more superior filters. 27 1.7 ASZM-TEDA Literature Background ASZM-TEDA is a microporous, activated carbon with high surface area (785−850 m2 g-1)89 and a pore volume of 0.386 cm3 g-1.92 ASZM-TEDA contains 4−6 wt% copper (A), 0.03−0.1 wt% silver metal (S), 4−6 wt% zinc (Z), 1−3 wt% molybdenum (M), presumably as oxides, and 2−4 wt% TEDA,33, 78, 86, 89, 92, 93 to protect against a broad spectrum of toxic gases. These carefully selected metals/metal oxides were impregnated into/onto the carbon support to remove CWAs without forming toxic byproducts or suffering from side reactions with the other components.76, 94 However, it is still unclear what specific processes are involved and the cooperative effects of the impregnants in the removal of the toxic gases and/or vapors. The carbon support used in ASZM-TEDA is BPL, a well-studied bituminous coal based, activated carbon, with a high surface area >1000 m2 g-1. BPL is predominantly microporous and well known for its capacity to remove toxic gases from the airstream and contaminants from water.69, 95, 96 Many research groups have studied the adsorption of chemical warfare agents/toxic industrial chemicals (sarin, NH3, SO2, HCN, Cl2, C6H6, CH2Cl2, and/or ethylene oxide),97-109 air pollutants/small molecules (H2S, COS, N2, CH4, CO, CO2, and/or H O),110-1302 and organic compounds (n-alkane, aromatics, alcohols, and/or haloginated hydrocarbons)101, 102, 131-170 with BPL carbon. The mesopores provide good transport properties within the granule, while the micropores have relatively 28 strong potential wells in which gases can be adsorbed, making it a sufficient adsorbent for filtration purposes. Due to the significance of the porous properties of BPL, different gas equilibrium models (N2, Ar, Kr, CO2, or H2) have been used to determine the pore size distribution (PSD). Traditionally, PSD analysis of porous materials are evaluated with N2 at 77 K because it is simple and inert.171 However, for predominantly microporous systems like BPL, temperature diffusion of N2 molecules into the micropores <0.7 nm is very slow, which can influence the adsorption measurements. Therefore, to mitigate these problems, multiple research groups employed different gas adsorbents (CO2 or H2) or methods such as, nonlocal density functional theory (NLDFT) and grand canonical Monte Carlo simulations (GCMS), to determine the PSD of BPL. These studies showed pore sizes to be 0.59 – 2.5 nm.51, 57, 171, 172 Although BPL is a high ash-loading material (maximum 8 wt%), containing mineral matters primarily of Si (~32%), Al (~26%), S (~24%), Fe (~9%), Ca (~2.9%), Ti (~2.2%), and K (~1.8%),93 additional impregnants (CuOx, Ag, ZnO, MoOx) were added to increase the binding/chemisorption of toxic gases.70, 76, 94 The success of ASZM-TEDA is a result of the intentional physical distribution and chemical interplay of the multiple metals/metal oxides that are incorporated into/onto BPL. Rossin et al. reported a cooperative mechanism between Cu and Zn impregnants for the removal of hydrogen cyanide by forming 29 Zn(CN)2, as opposed to cyanogen byproducts that are generated with Cu only.81 Similarly, Nickolov et al. described the use of Ag to remove arsine and phosphine by catalytic oxididation,71 and Smentkowski et al. reported the heterogeneous catalytic oxidation of organophosphonate esters on Mo without forming the undesired accumulation of carbonaceous or phosphorus species on the Mo surface.23 ASZM-TEDA impregnants can remove a variety of CWAs such as, cyanogen chloride (CK), methyl iodide (CH3I), hydrogen sulfide (H2S), sulfur dioxide (SO2), hydrogen cyanide (HCN), chlorine (Cl2), arsine (AsH3), phosphine (PH3), organophosphonate esters, and acid-gas producers. 23, 71, 73, 76, 81, 84, 173 A summary of the function of each ASZM-TEDA impregnant is listed in Table 1.3. While the high surface area and reactive impregnants of ASZM-TEDA provide initial protection against toxic gases, the adsorption capacity of this filter degrades over time. As toxic gases and/or water condensation adsorb into the pores and onto the surface, the pore properties change. Active sites are poisoned, ultimately leading to the failure of these filters. 30 Table 1.3. ASZM-TEDA Impregnants and their Function Impregnant Function Reference Removal of HCN, CK, Cl2, AsH3, CuOx PH3, HCl, HF, H2S, and acid-gas 71, 81, 84, 173 producers such as phosgene Ag Catalytic oxidation of AsH3, PH3 71, 84 ZnO Removal of HCN, CK, NH 71, 81, 17371, 81, 1733 MoO Catalytic oxidation of x organophosphonate esters 23 TEDA Removal of CK, CH3Br, CH3I, HCN, H2S, and SO2 73, 76, 173 1.8 Motivation ASZM-TEDA characterization in the public literature lists metal concentrations, but the nature of the impregnants (oxidation state, size, location, etc.) is unknown. A full understanding of the impregnants and their role in CWA removal can guide us in the strategic development of superior filtration materials. Beginning with a full analytical characterization of ASZM-TEDA, we can identify crucial information to aide us in the synthesis of deconstructed analogs of ASZM-TEDA. Studying these single variable models can identify the functions and limitations of each component (Chapter 2 and 3), so we can tailor our synthesis for new materials (Chapter 4), with catalytic performance, selective adsorption, and higher capacities for CWA defeat (Chapter 5). 31 Chapter 2: Dimethyl Methylphosphonate Adsorption Behavior and Desorption Energies on Ordered Mesoporous Carbons The work presented in this chapter was adapted from Huynh, K.; Holdren, S.; Hu, J.; Wang, L.; Zachariah, M. R.; Eichhorn, B. W., Dimethyl Methylphosphonate Adsorption Capacities and Desorption Energies on Ordered Mesoporous Carbons. ACS Appl. Mater. Interfaces, 2017, 9, 40638-40644. 2.1 INTRODUCTION ASZM-TEDA is used in nearly all United States nuclear, biological, and chemical filters for the removal of chemical warfare agent (CWA) vapors. The effectiveness of these filters largely depends on the limited space available for physical adsorption in the pores and the amount of reactive metal oxides contained in/on the carbon.33 While it is accepted that the adsorbent pore properties directly influence filter performance,174 the significance of the pore parameters (i.e. the size, volume, and surface area) remains unknown. The current carbon support used in ASZM-TEDA is BPL, a bituminous coal based, steam activated, granular carbon, with a high surface area >1000 m2 g- 1. While this material is largely microporous, there are significant variations among the pore diameter distributions. Russel et al. used five different gas equilibrium models to determine the pore width to be between 0.59−1.8 nm.172 In 32 addition, BPL is a high ash-loading material (maximum 8 wt%), containing mineral matters primarily of Si (~32%), Al (~26%), S (~24%), Fe (~9%), Ca (~2.9%), Ti (~2.2%), and K (~1.8%).93 While BPL has been established as an effective adsorbent for CWAs, very little information is available for the dynamic effects occurring at the molecular level. Further investigations are needed to understand the mechanistic details of where and how CWAs adsorb and desorb from the carbon adsorbent materials. When evaluating the effectiveness of gas mask filters, studies are performed to test their ability to remove CWA vapors, determine their adsorption capacities, and filter lifetimes. For safety reasons, the CWA simulant, DMMP is commonly used to investigate the sorption properties of G-series nerve agents, such as Sarin (GB) and Soman (GD), which possesses similar functional groups as illustrated in Figure 2.1.175 DMMP is used primarily as an adsorption simulant because it lacks the highly reactive P−F group common to all G-agents, which is a highly reactive functionality.69 Many theoretical and experimental studies show that DMMP adsorption occurs through the phosphoryl group or O−P−O moieties, onto carbon nanotubes, active carbons, and metal oxides.23, 92, 176, 15 (a) (b) (c) O O O H3C P OCH3 H3C P O H3C P O OCH3 F F Figure 2.1. (a) Dimethyl methylphosphonate (DMMP), (b) Sarin (GB), and (c) Soman (GD). 33 Harris et al. monitored DMMP adsorption to BPL and ASZM using 31P magic angle spinning NMR. DMMP adsorbs into adsorbent micropores at low loadings, then ultimately forming multilayers of weakly bound DMMP molecules on the surface.92 However, desorption energies and pore structure effects were not investigated. Desorption energies of organic compounds from various carbon substrates have been investigated with TPD experiments using techniques such as thermogravimetric analysis (TGA) and mass spectrometry.177-181 For example, Li. et al. performed numerous studies of naphthalene, acenaphthene, and phenanthrene desorption from various carbon supports and mesoporous silicas.177, 179, 180 However, to the best of our knowledge, the determination of DMMP desorption energies from various carbon supports have not been thoroughly investigated. In this study, we compared the micropore (pore diameter <2 nm) and mesopore (2< pore diameter <50 nm) structural effects by investigating DMMP adsorption in three different ordered mesoporous carbons (OMCs), a zeolite- templated ordered microporous carbon (ZY), a disordered microporous carbon (SMC), and the microporous ASZM support, BPL. DMMP adsorption was evaluated using quantitative 1H-NMR to measure monolayer/low surface coverage (1 h exposure), multilayer coverage (5 and 70 h exposure), and liquefaction/total pore filling (800 h exposure) for BPL and the OMCs. TGA was 34 employed to determine desorption energies of DMMP pre-saturated on all the carbon materials. These studies show that the surface area and pore volume dictate total adsorption loading with differences in the desorption energies as a direct result of the carbon pore networks. 2.2 EXPERIMENTAL SECTION 2.2.1 Materials AR grade dimethyl methylphosphonate (≥97.0%) (DMMP) was purchased from Sigma Aldrich (USA) and used without further purification. Sucrose and Zeolite Y were purchased from Fischer Chemical and Alfa Aesar, respectively. The commercial activated carbon, BPL (12x30 mesh), was supplied by the U.S. Army Edgewood Chemical and Biological Center. 2.2.2 Microporous Carbons The microporous carbons, sucrose based microporous carbon (SMC) and zeolite Y-template microporous carbon (ZY), were prepared by Mr. Luning Wang using slight modifications of methods as described elsewhere. 182, 183 2.2.2.1 Synthesis of a Sucrose Based Microporous Carbon (SMC) The microporous carbon was synthesized using sucrose as the carbon source described in Reference 182. In a typical reaction, 20 g sucrose was dissolved in 6 M sulfuric acid to form 5% sucrose solution, which was refluxed in 35 a 1 L round bottom flask at 120 °C overnight. The resulting black suspension was filtered and washed with distilled water several times. The product was dried at 100 °C in an oven for 24 h, and underwent carbonization at 1000 °C for 3 h with a heating rate of 5 °C min-1 under argon/5% H2 atmosphere. 2.2.2.2 Synthesis of a Zeolite Y-template Microporous Carbon (ZY) The synthesis of this microporous carbon using zeolite Y as a template was described in Reference 183. In a typical synthesis, 1 g zeolite Y (Hydrogen, 80:1 molar ratio of SiO2:Al2O3) was exchanged with ~ 50 ml 1 M NH4Cl for three days. The ammonium ions were decomposed at 500 °C for 1 h in air to produce Brønsted-Lowry acid sites in the zeolite. The obtained zeolite Y powder was heated at ~150 °C under vacuum for 5 h in a round bottle flask on a Schlenk line to fully activate the pores. Then the zeolite Y was transferred into a drybox and 0.26 g phenol was added. Afterwards, the mixture was incubated at 65 °C under reduced pressure overnight. Excess paraformaldehyde, ~0.4 g was heated to 120 °C to liberate monomeric formaldehyde, which was transferred as a gas to the phenol/zeolite composite to react at room temperature. The white zeolite/phenol composite turned into a pink/red color, which confirmed the polymerization process. The product was calcined in N2 for 5 h at 125 °C (ramp rate 1 °C min-1) to cross-link the polymer, followed by a further pyrolysis process at 900 °C for 14 h (ramp rate 5 °C min-1). After the carbonization process, the dark sample was added to excess concentrated hydrofluoric acid (50%) and stirred overnight. Then 36 the sample was filtered and washed by deionized water thoroughly and dried at 105 °C under vacuum overnight. 2.2.3 Ordered Mesoporous Carbons (OMCs) The OMCs, FDU-15, CMK-3, and CMK-8, were prepared by Dr. Junkai Hu using slight modifications of methods as described elsewhere. 184-186 A general schematic of these synthetic processes are described in Figure 2.2. RBesloocl/ks uCrofapcotlaynmt emr iFce1l2le7s Surfactant Self-assembly removal Carbon Carbonization source Ordered Carbon mesoporous silica impregnation Silica removal Figure 2.2. Schematic of the synthetic processes for the OMCs. 2.2.3.1 Synthesis of 1D Cylindrical Ordered Mesoporous Carbon FDU-15 FDU-15 was synthesized by a soft template method described in Reference 184. Briefly, a resol precursor was prepared from a low-molecular- weight polymer derived from phenol and formaldehyde using a polymerization 37 FDU-15 CMK-3 CMK-8 Soft template Hard template method. A triblock copolymer Pluronic F127 (EO106PO70EO106, where EO = ethylene oxide, PO = propylene oxide) was used as a structure directing reagent, while urea and formaldehyde were used as the carbon source. In a typical reaction, 1 g F127 was dissolved in 8.0 g ethanol. Then 5 g of the 20% resol precursor was added to the solution and the mixture was stirred for 1 h. The mixture was transferred to a petri dish for 5−8 h at room temperature and then 24 h at 100 °C to evaporate ethanol. Then the gel was thermopolymerized for 12 h in an oven at 180 °C. The product was then ground to a fine powder, calcined under Ar for 3 h at 600 °C, and another 2 h at 900 °C. The heating rates were 1 °C min-1 to 600 °C and 5 °C min-1 to 900 °C. 2.2.3.2 Synthesis of 3D Hexagonal Ordered Mesoporous Carbon CMK-3 CMK-3 was synthesized by a hard template method described in Reference 185. In a typical reaction, 1.2 g SBA-15 was added to a solution of 1.5 g sucrose, 0.17 g sulfuric acid, and 6 g H2O. The mixture was dried for 6 h in an oven at 100 °C, followed by another 6 h at 160 °C. The silica sample was treated again with a mixture of 0.96 g sucrose, 0.11 g sulfuric acid, and 6 g H2O, followed by the same drying process for 6 h in an oven at 100 °C, then another 6 h at 160 °C. The sample was then carbonized at 600 °C in flowing Ar for 3 h in a tube furnace, followed by another 2 h in flowing Ar at 900 °C. The heating rates were 1 °C min-1 to 600 °C and 5 °C min-1 to 900 °C. The silica was removed by dissolution with 3M NaOH solution at 50 °C. The carbon was collected by 38 centrifugation and washed with deionized water. The extraction/washing process was done in repetition three times. Two batches of CMK-3 were synthesized for our investigation. They will be referred to as CMK-3 and CMK-3*, due to their slight differences in pore properties, which are described in Table 2.1. 2.2.3.3 Synthesis of 3D Bicontinuous Ordered Mesoporous Carbon CMK-8 CMK-8 was synthesized by a hard template method described in Reference 186. KIT-6 was used as a template, with the molar ratio of 0.017 P123/1.67 tetraethyl orthosilicate (TEOS)/1.67 n-butanol/1.83 hydrochloric acid (HCl)/195 H2O, while sucrose was used as the carbon source. The carbon nanocasting process was similar to CMK-3, with adjustments to the sucrose and sulfuric acid amounts to match the pore volume of KIT-6. In a typical reaction, 1.2 g KIT-6 was added to a solution of 1.17 g sucrose, 0.13 g sulfuric acid, and 4.7 g H2O. The mixture was dried for 6 h in an oven at 100 °C, then another 6 h at 160 °C. The sample was treated again with a mixture of 0.62 g sucrose, 0.07 g sulfuric acid, and 3.85 g H2O, followed by the same drying process for 6 h in an oven at 100 °C, then another 6 h at 160 °C. Annealing and silica removal were as described for CMK-3. 2.2.4 Characterization of Carbon Materials Pore size, volume, and surface area were determined from nitrogen (N2) adsorption-desorption isotherms recorded with a Micromeritics ASAP 2020 39 Porosimeter Test Station. Samples were degassed in vacuum at 100 °C for 12 h prior to characterization. The specific surface areas were calculated using the Brunauer-Emmett-Teller (BET) method from the N2 adsorption data in the relative pressure range (P/P0) of 0.05−0.20. The general porosity distribution was calculated from the adsorption branch using the Barrett-Joyner-Halenda (BJH) equation. The micropore size distribution was calculated from the adsorption branch based on nonlocal density functional theory (NLDFT),151, 187, 188 with a relative pressure beginning at 0.001. Samples were imaged with a JEM 2100 Field Emission transmission electron microscope operating at 200 kV. The carbons were dispersed in methanol and a 10 µL aliquot of the resulting dispersion was drop cast on the TEM grids. The TEM grids used were carbon-coated Cu grids (CF200-Cu, Electron Microscopy Sciences). 2.2.5 Exposure of Carbon Samples to DMMP 2.2.5.1 Sample Preparation for 1H-NMR Study Carbons were pre-treated at 100 °C for ≥6 h under static vacuum to remove any adsorbed contaminants before exposure to DMMP using the vial-in- vial method, modified from the literature.189 Approximately 25 mg of carbon was transferred to small glass vials. Each small glass vial was placed inside a larger vial that contained a sufficient amount of liquid DMMP (200 𝜇L) to saturate the head space of the vial assembly. The larger vial was then capped, and stored in a 40 desiccator. Typically, room temperature vapor pressure of DMMP (~1.0 Torr) is enough to accomplish adsorption of the vapor into/onto the carbon samples,16 however, due to detection limits with 1H-NMR for the 1 h exposure tests, the carbon and small glass vials were taken directly from the oven and placed inside larger vials with liquid DMMP, then capped, before stored in a desiccator. For all samples, the exposure process was conducted under atmospheric pressure at room temperature (24−26 °C). The samples were then removed after the desired exposure time (800, 70, 5, or 1 h) and analyzed using a fixed-bed reactor, which is described in Section 2.2.6. 2.2.5.2 Sample Preparation for Thermogravimetric Analysis Samples used for TGA measurements were prepared by mixing approximately 25−50 mg of carbon with 70−90 uL of liquid DMMP and subsequently sonicated for 5 min to ensure that DMMP penetrated the pores. This impregnation method completely saturates the internal pore structures with liquid DMMP. The samples were then subjected to flowing Ar for sufficient time (>24 h) to remove excess liquid DMMP and desorb the weakly bound DMMP from the carbon. This exposure method allowed us to separate the differences between weakly bound DMMP and strongly bound DMMP (i.e. DMMP that remains adsorbed on the carbons and needs to be removed by heating) in a fast, reproducible fashion. After the 24 h purge, TGA experiments showed minimal 41 weight loss after flowing 100 mL min-1 of Ar over the sample for 60 min at room temperature. 2.2.6 Fixed-Bed Reactor Setup TPD studies on the carbon samples were conducted using a temperature controlled fixed-bed reactor system. Each carbon sample (25 mg) was loaded into a 10 mm ID quartz reactor. The reactor was composed of a quartz frit with ~10 mg quartz wool layered atop the frit. The sample was placed on top of the wool. After loading the sample into the reactor, the system was sealed with Swagelok fittings, and Ar (Standard, Airgas) was flowed at a constant flow rate of 50 SCCM (standard cubic centimeters per minute). The mass flow was controlled with a Brooks 5850 series mass flow controller. A Eurotherm 91 P PID controller and K- type thermocouple controlled the sample temperature. All lines downstream of the reactor were maintained at a constant 150 °C before and during testing to minimize DMMP adsorption in effluent lines. Heating was initiated after 10 min of Ar purging at 30 °C to remove weakly physisorbed molecules, which was not collected for analysis. Then the effective adsorption capacity measurements were programmed to include the following temperature regimes: (1) A linear temperature increase of 10 °C min-1, from 30 °C to 450 °C (2) Hold at 450 °C for 2 min (3) A fast cooling to 30 °C 42 The reactor effluent was directed through a bubbler filled with a mixed solution of deuterated acetonitrile (CD3CN) and ethylene carbonate (EC) that sufficiently submerged the bubbler stem. The CD3CN/EC solution was maintained at 0 °C via an ice-water bath. The sample was analyzed using 1H-NMR, which is described in Section 2.2.7.1. 2.2.7 Methods 2.2.7.1 1H-NMR Study NMR spectra were recorded with a Bruker AV-400 MHz Spectrometer. For each spectrum, 128 transients were collected at 298 K with an acquisition time of 2.6 s, d1 relaxation time of 4 s. The DMMP-acetonitrile-d3 solutions were spiked with precise amounts of EC to quantify the amount of DMMP desorbed. 2.2.7.2 Thermogravimetric Analysis (TGA) All TGA related studies were performed by Mr. Scott Holdren in the Zachariah Lab. TGA at various heating rates (5, 10, 25, and 50 °C min-1) was used to determine the effective desorption energy of DMMP from the OMCs and microporous carbons using a modified Flynn-Wall-Ozawa isoconversional method (eq. 2.1).190 𝑙𝑛 LO.QR = −1.0008 T + 𝐶V (eq. 2.1) MN UMN where β is the heating rate, Tf is the temperature at a fixed conversion of the reaction (i.e. in this work it is defined as percent of DMMP desorbed), R is the 43 ideal gas constant, and E is the activation energy (desorption energy) for the process. An isoconversional method was chosen to analyze the DMMP desorption energies because these methods can be easily applied to a variety of reaction systems simply by knowing the Tf for various heating rates at a fixed point in the reaction process. An isoconversional method was chosen to analyze the DMMP desorption energies because these methods can be easily applied to a variety of reaction systems simply by knowing the Tf for various heating rates at a fixed point in the reaction process. These isoconversional methods (e.g. Kissinger, Ozawa) apply approximations to the temperature integral and a wide variety of approximations exist, which leads to numerous types of isoconversional methods available. The modified method was chosen because the Flynn-Wall-Ozawa isoconversional model can be inaccurate and lead to deviations in activation energy more than 10%.191, 192 TGA measurements were made using a TA Instruments SDT-Q600. For these experiments, pre-exposed DMMP carbons were purged with Ar >24 h, then placed in the TGA, purged for an additional 60 min under 100 ml min-1 of Ar, then heated to 450 °C at different heating rates. The desorption energy of DMMP was calculated at different DMMP coverages, where 0% DMMP desorbed was defined as the point at which heating begins for the kinetic analysis. 44 2.3 RESULTS AND DISCUSSION 2.3.1 Characterization of Ordered Mesoporous Carbons In this investigation, we studied three different mesoporous carbons; the 1D cylindrical FDU-15, 3D hexagonal CMK-3, and 3D bicontinuous CMK-8, with well-defined pore size, volume, and surface area. Three microporous carbons were also investigated, disordered microporous BPL carbon, the zeolite-Y templated ordered microporous carbon (ZY), and sucrose derived disordered microporous carbon (SMC). The structures of the mesoporous and microporous carbons were first investigated using TEM. Figure 2.3 compares the TEM images of the 1D cylindrical mesoporous FDU-15, 3D hexagonal mesoporous CMK-3, 3D bicontinuous mesoporous CMK-8, and disordered microporous BPL carbon. The mesoporous carbons show uniform pore structure and pore size distributions of approximately 3 nm, in comparison to BPL, which shows disordered pores with pore sizes less than 2 nm. 45 Figure 2.3. TEM images of (a) ordered mesoporous FDU-15, (b) ordered mesoporous CMK-3, (c) ordered mesoporous CMK-8 and (d) disordered microporous BPL. The inset cartoons show their general framework structures. 46 Figure 2.4. (a) N2 adsorption-desorption isotherms and (b) BJH pore size distributions for the different carbon architectures, FDU-15, CMK-3, CMK-8, and BPL carbon. 47 The pore size and volume of the different carbons were compared by N2 adsorption-isotherms. The detailed structure parameters are presented in Table 2.1. Based on the isotherms in Figure 2.4a, all three OMCs show typical Type-IV isotherms with hysteresis, in good agreement with their mesoporosity.50 The different pore structure, network, and size distributions are evident by the type of hysteresis loops in Figure 2.5a and further illustrated in Figure 2.4b. FDU-15 and CMK-3 have Type H1 hysteresis loops, indicating cylindrical pore geometry and a narrow pore distribution with peak maxima at 3.5 and 2.8 nm, respectively. CMK-8 has Type H4 hysteresis, which is associated with a wide pore distribution represented by the bimodal porosity in Figure 2.4b, with a clear maximum at 3.6 nm. In contrast, BPL shows a Type-I isotherm with relative P/P0 change below 0.2, which indicates its microporosity with relatively small mesopore/external surfaces.193 The OMCs, FDU-15, CMK-3, and CMK-8, have similar mesopore sizes of 2.8−3.6 nm, in comparison with microporous BPL with pore sizes less than 1.7 nm. These data are in good agreement with the pore sizes measured from the TEM studies and previous studies in the literature.19-21 Table 2.1 also compares the surface area and pore volume of the different carbons. In general, FDU-15 shows the smallest surface area of 600 m2 g-1, while the surface area of CMK-3 (976 m2 g-1), CMK-8 (1075 m2 g-1), and BPL (1229 m2 g-1) are similar. The total pore volume of the carbons follows FDU-15 < BPL < CMK-3 < CMK-8. The pore volume of CMK-8 is 1.136 cm3 g-1, which is more 48 than double that of the other carbons. The different pore properties will influence the adsorption and desorption properties of DMMP, which will be discussed in detail below. Table 2.1. N2 Adsorption –Desorption Properties of OMCs and BPL Carbon Carbon Pore Mesopore Sample Diameter Total Pore Volume Total Surface (cm3 g-1)b Area (m2 g-1)c Surface Area (nm)a (m2 g-1)a FDU-15 3.6 0.325 600 223 CMK-3 2.8 0.454 976 462 CMK-3* 2.7 0.669 1280 728 CMK-8 3.5 1.136 1075 774 BPL <1.7 0.391 1229 72 aCalculated using the Barrett-Joyner-Halenda (BJH) model. bThe total pore volume is based on the sum of the mesopore and macropore volume from the BJH model and the micropore volume from the t-plot method. cCalculated using the multipoint Brunauer-Emmett-Teller (BET) method. 2.3.2 DMMP Adsorption on Different Carbon Architectures Plots of DMMP uptake versus time (t) up to 800 h are shown in Figure 2.5. The adsorption curves are similar for all carbons, with an initial rapid increase in the first 80 h, followed by a saturation region, typical of adsorption processes. After 50 h, the 3D bicontinuous CMK-8 has the highest DMMP uptake, followed by CMK-3, FDU-15, and BPL, respectively. The slower adsorption rates for the other carbons may be due to slower diffusion of DMMP 49 into the less accessible interior channels after the surface sites become saturated with adsorbed DMMP. A similar saturation curve seen in Figure 2.5 was observed by Khanday et al.194 and Saxena et al.69 in their studies of DMMP adsorption in different zeolites from 2−18 h and activated carbons from 0−600 h, respectively. The adsorption of DMMP in FDU-15 saturates at approximately 0.3 g DMMP/m2 carbon at 400 h, which represents filling of the total pore volume of the material. While the time-to-saturation was not measured for the other the three samples, their 800 h capacities represent filling of the total pore volumes and can be considered the saturation limit, as discussed below. Figure 2.5. Plot of DMMP uptake as a function of time for the ordered mesoporous carbons and microporous BPL carbon. The inset shows initial uptake of DMMP in the first 60 h before the plateau region. The dotted lines represent extrapolations from the 60 h data to the fully saturated systems. The time-to- saturation was not measured in these systems. 50 2.3.3 TPD Studies of DMMP from Different Carbon Architectures TPD experiments were conducted to determine the adsorption capacity and kinetics of DMMP desorption. After 800 h of DMMP exposure, the pores are fully saturated with liquid DMMP,174 as evidenced by the linear relationship between adsorption capacity and total pore volume, shown in Figure 2.6. Figure 2.6. The adsorption capacity after 800 h of DMMP exposure for all carbons (red dots) with different total pore volumes. A fitted linear curve of adsorption capacity versus pore volume (blue line) was added. To probe the adsorption regime between full surface coverage and liquid saturation, we evaluated samples after 70 h and 5 h of DMMP exposure. Excluding BPL, there is a general linear correlation between DMMP uptake and the total surface area of the OMCs (Figure 2.7a and b). This relationship between DMMP adsorption and the surface area is similar to that observed by Khanday et al. when comparing DMMP adsorption in four different zeolite samples (Zeolite 51 X, Erionite, MCM-22, and Zeolite-A),6 where liquefaction has not occurred. BPL presumably does not follow this trend since the majority of its porosity originates from micropores and competing processes, such as the onset of liquefaction and diffusion through micropores versus mesopores, which may be confounding effects. After longer exposure times (70 h), the BPL mass adsorption deviates less from the linear trend (Figure 2.7a), which we believe is a result of micropore filling and less competing processes. 0.4 (a) 70 h CMK-8 0.3 CMK-3 0.2 BPL FDU-15 0.1 0 0 400 800 1200 Total Surface Area (m2/g) 0.12 (b) 5 h CMK-3* CMK-8 0.08 FDU-15 0.04 BPL 0 0 400 800 1200 Total Surface Area (m2/g) Figure 2.7. Adsorption capacity after (a) 70 h and (b) 5 h of DMMP exposure in relationship to the total surface area of the different carbon structures. 52 Mass Adsorption Capacity Mass Adsorption Capacity (g DMMP/g carbon) (g DMMP/g carbon) To further investigate DMMP adsorption properties before liquefaction, we evaluated samples after 1 h of DMMP exposure. Previous studies state that initial adsorption occurs in the micropores, which is more thermodynamically stable.50, 92 It was experimentally difficult to prove the adsorption process within the micropores as discrete steps (for example, first surface coverage then micropore filling) for a few reasons. First, pore sizes <0.6 nm are inaccessible to DMMP due to sterical reasons (which will further be described in Section 2.4.4); therefore, those pores will remain vacant during DMMP exposure. Second, it is uncommon to use terms like “surface area” and “pore volume” separately to describe a measurement of available adsorption sites for pore diameters <2 nm, since the entire pore represents a space where adsorption may occur. For instance, given the size of DMMP (~0.6 nm),195 monolayer adsorption on the opposite walls of the micropore could be close enough that it leads to pore filling.50, 174, 193 Therefore, after 1 h of DMMP exposure, we speculate that the accessible micropores (0.6–2 nm) are filled and the continuous adsorption of DMMP occurs as a monolayer across the surface of the mesopores. This notion is supported by a linear correlation between DMMP mass adsorption and the mesopore surface area for all four carbons, as shown in Figure 2.8. 53 0.04 CMK-8 CMK-3* 0.03 0.02 BPL 0.01 FDU-15 0 0 200 400 600 800 Mesopore Surface Area (m2/g) Figure 2.8. Adsorption capacity after 1 h of DMMP exposure in relationship to the mesopore surface area of the different carbon structures. 2.3.4 Desorption Energies of DMMP DMMP desorption energies for the microporous carbons and the OMCs were determined at various surface coverages using the modified Flynn-Wall- Ozawa isoconversional method, described in the experimental section 2.6.2. For these experiments, the carbons were saturated with liquid DMMP and purged with Ar for >24 h before the analysis to remove loosely-bound DMMP on the surface, and in the majority of the mesopores. After the purge, the remaining DMMP for all carbons resides predominately in the pores that are ≤2.4 nm in diameter, as evidenced by the linear relationship between DMMP content and cumulative pore volume shown in Figure 2.9. This relationship was empirically derived from the TGA DMMP weight loss and NLDFT N2 adsorption pore volume measurements 54 Mass Adsorption Capacity (g DMMP/g carbon) (Figures 2.10a and b). Based on Figures 2.10a and b, cumulative pore volumes of other pore sizes were investigated (1.8 to 2.6 nm) and fitted to the observed TGA DMMP weight loss (Table 2.2). We found that 2.0−2.4 nm range provided the best linear fit of the data. Although the remaining strongly bound DMMP resides in pores ≤2.4 nm for all the carbons tested, the cumulative volume for these pores are all different due to the carbons unique internal structures. As such, the desorption energies measured in the study represent DMMP desorption from carbon pores ≤2.4 nm in diameter. Figure 2.9. DMMP uptake versus cumulative pore volume ≤ 2.4 nm. 55 Table 2.2. R2 Coefficients Measured Based on the Linear Fit of DMMP Remaining on the Different Carbons NLDFT Pore Size Range (nm)a R2 b 0−2.6 0.9577 0−2.4 0.9938 0−2.2 0.9933 0−2.0 0.9954 0−1.8 0.9549 aCalculated from the adsorption branch based on NLDFT, with a relative pressure beginning at 0.001. bCalculated from the linear fit of DMMP uptake with respect to the pore volume from different pore size ranges. 56 Figure 2.10. (a) TGA curves of DMMP exposed carbon samples at 10 °C min-1. TGA curves were normalized to 100 wt% after a 60 min Ar purge. The amount of DMMP desorbed follows the same trend as the cumulative pore volume of pores ≤2.4 nm (b) NLDFT N2 adsorption pore volume measurements for different carbons tested. 57 Figure 2.11. TGA of SMC (red) and DMMP exposed SMC (black). There is a slightly larger weight loss for the DMMP exposed SMC sample. In addition to the four carbons described previously, two microporous carbons were also tested to obtain a better understanding of how the structure and porosity affects the experimentally determined DMMP desorption energies. One of these carbons is a sucrose derived microporous carbon (SMC) detailed in Xu et al. and the other is a microporous zeolite Y template (ZY) derived carbon described by Johnson et al.182, 183 All six of these carbons were exposed to liquid DMMP, as described in the previous Section 2.2.5.2. When SMC was exposed to DMMP, TGA showed that very little to no DMMP was absorbed (Figure 2.11). According to Xu et al., the average pore size of SMC is 0.5 nm, which is too small for DMMP to penetrate the pores.17 This aligns well with previous estimates of the size of a DMMP molecule, which is ~0.6 nm.195 Therefore, no DMMP 58 desorption energies could be determined for SMC. TGA curves of DMMP desorption from BPL carbon at different heating rates are presented in Figure 2.12a. Figure 2.12. (a) TGA curves of DMMP desorption from BPL carbon at different heating rates. The TGA curves were normalized to 100 wt% after the second purge (60 min under 100 ml min-1 of Ar). (b) Representative desorption energy determination of DMMP desorption from BPL carbon at 25% DMMP desorbed. The error bars in plot (b) are smaller than the symbols. 59 The temperature at various DMMP desorption percentages (α) for each TGA curve were fitted to the Flynn-Wall-Ozawa isoconversional method, shown in Figure 2.12b, to extract a DMMP desorption energy. These DMMP desorption energies for the different carbons were calculated as a function of % DMMP desorbed (α), and shown in Figure 2.13. Figure 2.13. DMMP desorption energies from carbon samples. The DMMP desorption energies from the five different carbon samples are shown in Figure 2.13 as a function of the % DMMP desorbed (α). At α = 25%, the desorption energies for all carbon structures are similar, ranging from 68−74 kJ mol-1. With an increase in fraction desorbed, the apparent activation energy also increases, with the exception of BPL, which is relatively constant. At α = 80%, the desorption energies for the synthetic OMCs are 95−103 kJ mol-1, a 60 30−40% increased. All of these desorption energies are above the DMMP heat of vaporization (ΔH -1vap), which is 52 kJ mol , but are significantly lower than covalent bonds,196 implying that DMMP-carbon interactions are best characterized as physisorption. The origin of the increase in desorption energy in our systems relative to the ΔHvap is not entirely clear but we propose two possible explanations. At α = 25%, most of the DMMP presumably resides on the pore walls, which could facilitate weak Van der Waals type interactions (i.e. weak physisorption) that could increase the desorption energy relative to ΔHvap. This scenario is similar to the hydrogen bonding interactions of DMMP to silica proposed by Wilmsmeyer et al. and Henderson.14, 197 Alternatively, the increase from the ΔHvap value may arise from the DMMP vapor pressure suppression due to capillary condensation.50, 193 The carbon porosity allows liquid DMMP to become entrapped in the micropores, which suppresses the vapor pressure of DMMP requiring additional energy (16−22 kJ mol-1). This proposal is reminiscent of the entrapment model advanced by Ferguson-McPherson et al.30 The origin of the 30−40% increase in OMCs when α = 80% desorption energies relative to the α = 25% values is also unclear. Since the BPL sample did not show the same increase, we evaluated ZY, given that both BPL and ZY have large microporosity, and similar DMMP adsorption capacities (Figure 2.10a). However, ZY has an ordered microstructure imparted by the zeolite template, 61 whereas BPL has a random, disordered microporous network. As shown in Figure 2.13, the desorption energy at α = 80% for ZY increases by 15% relative to the α = 25%, in contrast to the behavior of BPL. While the increase is less than that of the OMCs, the data suggests that the increases at lower coverage of the templated carbons originates from their ordered network structures. Li et al. has reported similar trends in the desorption of naphthalene and acenaphthene from a bituminous coal activated carbon (similar to BPL). These investigators measured desorption energies of approximately 60 kJ mol-1, which remained constant throughout the entire desorption process.151 In a subsequent study, Li et al. measured a naphthalene desorption energy of 74 kJ mol-1 from the mesoporous carbon CMK-3.181 These investigators observed a 23% increase in desorption energy for naphthalene and CMK-3, relative to BPL, which is similar to our findings, and further suggests that the ordered nature of the carbon networks increases desorption energies at low surface coverages. 2.4 CONCLUSIONS We have demonstrated from various DMMP exposure times: 1, 5, 70, and 800 h, that adsorption of DMMP is initially influenced by the mesopore surface area, then total surface area, and the pore volume after liquefaction in each OMC. In addition, the amount of energy required to desorb low coverage DMMP from the OMCs is dependent on the pore properties of each carbon, and the weakly bound DMMP on the carbon surface or in pores >2.4 nm will readily desorb at 62 room temperature, and no DMMP will absorb in pores smaller than 0.5 nm. The desorption energies of the remaining physisorbed DMMP are primarily a measurement of the heat of vaporization and excess energy due to either weak Van der Waals interactions or vapor pressure suppression from capillary condensed DMMP in pores ≤2.4 nm. At lower surface coverages (α = 80), the DMMP desorption energies are 20−30 kJ mol-1 higher for the ordered micro- and mesoporous carbons but remains relatively constant for BPL. These data suggest that diffusion barriers and transport through an ordered network increases the effective desorption energies. A few important generalizations can be made for designing new adsorption materials. First is that maximizing the mesopore volume of a carbon leads to a larger adsorption capacity. Second, the most tightly-bound DMMP molecules reside in the pores between 0.6−2.4 nm. Third, ordered architectures give rise to higher apparent desorption energies that may be associated with transport barriers. 63 Chapter 3: Endo-pore and Exo-pore Copper Oxide Deposition in Ordered Mesoporous Carbons: Synthesis, Structures, and Adsorption/Decomposition Properties 3.1 INTRODUCTION The adsorption process is commonly used for purification purposes with gas mask filters employing two types of adsorption, physisorption and chemisorption, to remove chemical warfare agents (CWAs) from the airstream.24, 33, 49, 50 Dimethyl methylphosphonate (DMMP), a CWA simulant, is widely studied to investigate the adsorption process involved in CWA removal. In Chapter 2, our studies revealed that DMMP binding to unimpregnated carbons can be characterized as physisorption, a product of Van der Waal’s interactions. The desorption energy calculated for the tightly bound DMMP molecules at low surface coverages (80% DMMP desorbed relative to capacity), ranged from 68−103 kJ mol-1 for microporous (BPL, ZY) and mesoporous (FDU-15, CMK-3, and CMK-8) carbons, which is similar to the heat of vaporization (52 kJ mol-1). Given that DMMP is not strongly bound to carbon, the adsorptive properties are enhanced by metal oxide impregnation. It is well-known that DMMP chemisorbs to metal oxides (consequently increasing the enthalpy of adsorption), and/or alternatively decomposes DMMP into nontoxic byproducts.70, 76, 198 Additional reasons to incorporate metal oxide nanoparticles include, gas adsorption of 64 molecules that do not bind to carbon (broader target spectrum), increased adsorption capacity and reactive sites, and reduction of the amount of adsorbent needed in a canister by optimizing the surface-to-volume ratio. Specifically, this will result in smaller, lighter, and more portable gas-masks for field use. Previous studies have shown that the impregnation of activated carbons with metals such as, Mn, Co, Ni, Fe, Cu, Zn, Ag, Cr, Mo, and V, can increase the removal of toxic gases not typically removed by carbon.76, 93, 199, 200 Particularly, Cu/CuO has proven to be promising due to an increased efficiency in adsorbing and decomposing CWA simulants and TICs, such as DMMP and ammonia, as demonstrated by Ji et al.,201 Petit et al.,100 and our collaborators at the University of Maryland (Dr. Maija Kuklja, Dr. Michael Zachariah), Naval Research Lab (Dr. Jeffrey Owrutsky, Dr. Jeffrey Long), and Lawrence Berkley National Lab (Dr. Hendrik Bluhm).31 The carbon support and the CuO properties are expected to influence the DMMP adsorption and decomposition properties, based on our observations in Chapter 2. In this chapter, we investigated CuO impregnated carbon systems, which represents one of the deconstructed analogs of ASZM-TEDA. A diverse group of model systems with strategically placed impregnants inside (endo) and outside the porous network (exo-pore) were synthesized. The model compounds include, BPL with exo-pore CuO, OMCs (FDU-15, CMK-3, and CMK-8) with endo-pore CuO, and FDU-15 with endo and exo-pore CuO. Our studies with these synthetic 65 CuO endo and/or exo-pore carbons further our understanding of CuO functionality within the adsorbent and investigate the sources of ASZM-TEDA deactivation, such as impregnant migration and poisoning of the surface with phosphorus species202 (previously discussed in Chapter 1). Additionally, it was discovered that CuO impregnation increases adsorption capacity and reactivity, in comparison to carbon adsorbents. The extent of DMMP decomposition is higher for nano sized endo-pore CuO, relative to larger exo-pore CuO particles. The DMMP decomposition product observed is gaseous methanol, which presumably creates a methyl methylphosphonate (MMP) byproduct on the surface, and consequently, deactivates the adsorbent. 3.2 EXPERIMENTAL SECTION 3.2.1 Materials AR grade dimethyl methylphosphonate (≥97.0%) (DMMP) was purchased from Sigma Aldrich (USA) and used without further purification. The commercial activated carbon, BPL (12x30 mesh), was supplied by the U.S. Army Edgewood Chemical and Biological Center. 3.2.2 CuO Deposition on BPL In a typical reaction, 0.1 g BPL was washed with deionized water (2 mL), and collected by vacuum filtration to remove any impurities. Then, BPL was fully saturated with a Cu(NO3)2 aqueous solution (8 wt%) for 4 h. The BPL filled with 66 Cu(NO3)2 aqueous solution was collected by centrifugation (14000 rpm, 1 min). The saturation/collection process was repeated three times. Next, the excess solution was removed by vacuum filtration. The BPL/Cu(NO3)2 solid was dried in a box furnace at 90 °C for 6 h. The final product was annealed at 380 °C for 5 h under Ar flow, with a heating rate of 1 °C min-1. The resulting solid was annealed a second time to 250 °C in air, with a heating rate of 1 °C min-1 (Figure 3.1). The CuO impregnated BPL was collected and characterized. 3.2.3 CuO Impregnation for OMCs The OMCs, FDU-15, CMK-3, and CMK-8, were prepared using slight modifications of methods as described in Chapter 2, Section 2.2.3.1−2.2.3.3, and in the literature.184-186 These OMCs were later impregnated with CuO nanoparticles by Dr. Junkai Hu. 3.2.3.1 CuO Endo-pore and/or Exo-pore Impregnation In a typical reaction, 0.1 g OMC (FDU-15, CMK-3, CMK-8) was washed with deionized water (2 mL), and collected by vacuum filtration to remove any impurities. Then, OMC was fully saturated with a Cu(NO3)2 aqueous solution (8 wt% for FDU-15, 3.5 wt% for CMK-3 and CMK-8) for 4 h. The OMC filled with Cu(NO3)2 aqueous solution was collected by centrifugation (14000 rpm, 1 min). The saturation/collection process was repeated three times. Next, the excess solution was removed by vacuum filtration. The OMC/Cu(NO3)2 solid was dried 67 in a box furnace at 100 °C for 6 h. The final product was annealed at 380 °C for 5 h under Ar flow, with a heating rate of 1 °C min-1. The resulting solid was annealed a second time to 250 °C in air, with a heating rate of 1 °C min-1 (Figure 3.1). The CuO impregnated carbons were collected and characterized. OMC Soaking Filtering 1. 100 °C, Solvent 2% - 8% Cu(NO3)2 in H O Evaporation 2 2. 380 °C in Ar, Salt decomposition CuCOuO 250 °C in air, Oxidation of Cu(I) to Cu(II) Cu2O Figure 3.1. Schematic of the impregnation process for copper oxide into and/or onto the OMC pores. 3.2.3.2 Selective Removal of Exo-pore CuO FDU-15 was impregnated with CuO as described in Section 3.2.3.1, except FDU-15 was fully saturated with a higher concentration Cu(NO3)2 aqueous solution (40 wt%). In a typical reaction, 0.1 g FDU-15/CuO (11.7 CuO wt%) was fully saturated by nonane (2 mL) for 1 h to protect the endo-pore CuO nanoparticles. The FDU-15/CuO filled with nonane was collected by centrifugation (14000 rpm, 1 min). Next, 1M H2SO4 (2 mL) was added, stirred for 0.5 min, then immediately washed with deionized water (2 mL), and collected by 68 centrifugation (14000 rpm, 1 min). The endo-pore CuO impregnated FDU-15 was dried in a box furnace at 160 oC overnight (Figure 3.2). Nonane H2SO4 wash Nonane removal Figure 3.2. Schematic of the selective removal of exo-pre CuO from FDU-15 originally with endo and exo-pore CuO nanoparticles. 3.2.4 Characterization of Adsorbents Pore size, volume, and surface area were determined from N2 adsorption- desorption isotherms recorded with a Micromeritics ASAP 2020 Porosimeter Test Station. Samples were degassed in vacuum at 100 °C for 12 h prior to characterization. The specific surface areas were calculated using the BET method from the N2 adsorption data in the relative pressure range (P/P0) of 0.05−0.20. The general porosity distribution was calculated from the adsorption branch using the Barrett-Joyner-Halenda (BJH) equation. Structural morphologies of the samples were imaged with a scanning electron microscope (Hitachi SU-70 SEM, operated at an acceleration voltage of 10 kV) and a transmission electron microscope (JEM 2100 Field Emission operated at 200 kV). The samples were dispersed in methanol and a 10 µL aliquot of the resulting dispersion was drop cast on the TEM grids. The TEM grids used were carbon-coated Cu grids (CF200-Cu, Electron Microscopy Sciences). X-ray diffraction (XRD) patterns were recorded by a Bruker Smart1000 (Bruker AXS 69 Inc., USA) using CuKα radiation. The copper content of the carbon/CuO composites was determined using inductively coupled plasma-atomic emission spectroscopy (ICP-AES) characterization performed on a Perkin Elmer ICP Optima 4700. The samples were dissolved in pure trace-metal grade nitric acid (Sigma-Aldrich) and diluted with milli-Q water to a known volume (100 mL) with a final concentration of 2 wt% nitric acid (HNO3), before administered to the plasma. Intensities were measured at 327.396 nm for Cu. A linear calibration curve (R2 = 0.9999) was collected with five standards, a blank (2 wt% HNO3), 0.1, 0.2, 0.5, and 1.0 ppm Cu (High Purity Standard, 1000 𝜇g/g Cu in 2 wt% HNO3). 3.2.5 Exposure of Adsorbents to DMMP Adsorbent materials were pre-treated at 100 °C for ≥6 h under static vacuum to remove any adsorbed contaminants before exposure to DMMP using the vial-in-vial method, modified from the literature.189 Approximately 35−100 mg of material was transferred to small glass vials. Each small glass vial was placed inside a larger vial that contained a sufficient amount of liquid DMMP (200 𝜇L) to saturate the head space of the vial assembly. The larger vial was then capped, and stored in a desiccator. The exposure process was conducted under ambient pressure at room temperature (24−26 °C). The samples were then 70 removed after the desired exposure time (5 or 70 h) and analyzed using a fixed- bed reactor, which is described in Section 3.2.6. 3.2.6 Fixed-Bed Reactor Setup TPD studies were conducted using a temperature controlled fixed-bed reactor system. Each sample was loaded into a 10 mm ID quartz reactor. The reactor was composed of a quartz frit with ~10 mg quartz wool layered atop the frit. The sample was placed on top of the wool. The system was sealed with Swagelok fittings, and Ar (Standard, Airgas) was flowed at a constant flow rate of 50 SCCM (standard cubic centimeters per minute). The mass flow was controlled with a Brooks 5850 series mass flow controller. A Eurotherm 91 P PID controller and K-type thermocouple controlled the sample temperature. All lines downstream of the reactor were maintained at a constant 150 °C before and during testing to minimize DMMP adsorption in effluent lines. Heating was initiated after 10 min of Ar purging at 30 °C to remove weakly physisorbed molecules, which was not collected for analysis. Then the effective adsorption capacity measurements were programmed to include the following temperature regimes: (4) A linear temperature increase of 10 °C min-1, from 30 °C to 450 °C (5) Hold at 450 °C for 2 min (6) A fast cooling to 30 °C 71 The reactor effluent was directed through a bubbler filled with a mixed solution of deuterated acetonitrile (CD3CN) and ethylene carbonate (EC) that sufficiently submerged the bubbler stem. The CD3CN/EC solution was maintained at 0 °C via an ice-water bath. The sample was analyzed using 1H- NMR, which is described in Section 3.2.7.1. 3.2.7 Methods 3.2.7.1 Quantitative 1H-NMR NMR spectra were recorded using a Bruker AV-400 MHz Spectrometer. For each spectrum, 128 transients were collected at 298 K with an acquisition time of 2.6 s, d1 relaxation time of 4 s. The DMMP-acetonitrile-d3 solutions were spiked with precise amounts of EC that was added as a quantitative calibrant to measure the amount of DMMP and byproducts desorbed. 3.2.7.2 ICP-AES Analysis The phosphorus content of the cycled adsorbent materials were determined using ICP-AES characterization performed on a Perkin Elmer ICP Optima 4700. The samples were dissolved in pure trace-metal grade nitric acid (Sigma-Aldrich) and diluted with milli-Q water to a known volume (100 mL) with a final concentration of 2 wt% nitric acid (HNO3), before administered to the plasma. Intensities were measured at 177.5 and 178.3 nm for P. A linear calibration curve (R2 = 0.9999) was collected with five standards, a blank (2 wt% 72 HNO3), 0.1, 0.2, 0.5, and 1.0 ppm P (High Purity Standard, 1000 𝜇g/g P in 2 wt% HNO3). 3.3 RESULTS AND DISCUSSION 3.3.1 Characterization of CuO Impregnated Carbons Five samples of carbon/CuO composites: BPL with exo-pore CuO, OMCs (FDU-15, CMK-3, and CMK-8) with endo-pore CuO, and FDU-15 with CuO endo and exo-pore, were prepared and characterized. The diverse group of model systems are used to probe the influence of CuO on DMMP adsorption and decomposition, which will further be discussed in Section 3.3.2−3.3.3. 3.3.1.1 CuO Particle Size, Distribution, and Morphology The particle size, distribution, and morphology of the CuO impregnants were characterized using electron microscopy, TEM and SEM. TEM images of the spherical CuO impregnants revealed 2−4 nm particles evenly distributed inside the pores and 30−50 nm exo-pore particles/large aggregates on the surface, which was further confirmed with SEM (Figure 3.3). Aggregates of CuO particles were formed on the external surface of BPL after CuO impregnation process. We speculate that this aggregation process is caused by the inability of the Cu(NO3)2 precursor solution to enter the pores of the activated carbon.199 In Figure 3.4a-d, a micron-sized CuO aggregate on the BPL 73 surface was confirmed using SEM/EDX, which appears similar in size and morphology to unsupported commercial CuO mixed with BPL (Figure 3.4e). TEM SEM (a) 100 nm 500 nm (b) 100 nm 200 nm (c) 100 nm 200 nm (d) 100 nm 500 nm (e) 100 nm 200 nm Figure 3.3. TEM and SEM images of the carbon/CuO composites: (a) BPL/CuO (exo-pore), (b) FDU-15/CuO (endo and exo-pore), (c) FDU-15/CuO (endo-pore), (d) CMK-3/CuO (endo-pore), and (e) CMK-8/CuO (endo-pore). 74 (a) (b) (c) 1"m 1"m (d) (e) 500 nm 500 nm Figure 3.4. SEM image of (a) BPL/CuO with (b) Cu (red) and C (light green) phase maps. (d) is a zoomed in SEM image of (a) to compare to (e) mixed BPL and commercial CuO nanoparticles. The OMCs (1D cylindrical FDU-15, 3D hexagonal CMK-3, and 3D bicontinuous CMK-8) were impregnated with CuO (Figure 3.3b-e). The Cu(NO3)2 precursor solution was able to access the mesopores and form endo- pore CuO nanoparticles since the OMCs have uniform pore size distributions of ~ 3 nm (Table 3.1). CuO impregnation is evidenced by the 2−4 nm particles in alignment with the ordered pore structure, and the absence of CuO particles on the OMC surface (Figure 3.3b-e). The absence of exo-pore particles and the co- alignment of the CuO particles with the carbon pore structure provide strong evidence that the metal oxide is exclusively in the endo-pore. Furthermore, FDU- 15 with endo and exo-pore CuO has 2−4 nm endo-pore nanoparticles and 30−50 nm particles on the surface, as shown in the SEM images in Figure 3.3b. 75 Total based on ICP-AES Surface based on XPS FDU-15/CuO FDU-15/CuO CMK-3/CuO CMK-8/CuO endo & exo-pore endo-pore only Figure 3.5. Cu wt% calculated using ICP-AES and XPS measurements for OMC/CuO composites. The CuO location for the OMC/CuO composites was further investigated using XPS to compare with the bulk concentration determined using ICP-AES. Total Cu concentration measured using ICP-AES was 5.2–8.5 wt% for all OMC/CuO composites (Figure 3.5, blue bars). Discrepancies between Cu wt% using XPS and ICP-AES, arises from the surface sensitivity of XPS, which only probes ~10 nm of the sample surface. FDU-15 with endo and exo-pore CuO has 16.2 Cu wt%, which is significantly higher (more than 2x greater) than the bulk analysis (6.6%). The higher Cu wt% is a result of a larger Cu to carbon ratio for XPS measurements (which only measures the carbon surface up to 10 nm depth). 76 The Cu metal detected (0.5–2.5 wt%) using XPS for carbons with endo-pore CuO (FDU-15/CuO, CMK-3/CuO, and CMK-8/CuO) is significantly lower than the bulk analysis (5.2–8.7 Cu wt%), indicating that the Cu concentration is more localized in the center of the carbon support beyond the XPS range. The Cu wt% reported using XPS (0.5–2.5) is presumably located in the pores within 10 nm of the surface (Figure 3.5, red bars). These results are in good agreement with the TEM/SEM/SEX studies. 3.3.1.2 Structural Properties of Carbon/CuO Composites The pore volume and surface area of the carbons were affected by CuO impregnation, however the pore sizes remained unchanged, apart from FDU-15 (Figure 3.6 and 3.7b). In general, the pore volume decreased due to CuO endo- pore impregnation, while the surface area increased from exo-pore deposition (Table 3.1). The measured OMCs, FDU-15, CMK-3, and CMK-8, have similar mesopore sizes of 2.8–3.6 nm, in comparison to microporous BPL with pore sizes less than 1.7 nm. The detailed porosity parameters for CuO impregnated BPL, FDU-15, CMK-3, CMK-8, and their respective unimpregnated counterparts, are presented in Table 3.1. These data are in good agreement with the pore sizes observed from the TEM images in Chapter 2, and previous studies in the literature.19-21 TEM and BET analysis showed that the pore sizes remained unchanged with CuO impregnation (apart from FDU-15), which is consistent with 77 the BJH pore size distribution curves shown in Figure 3.6 and 3.7b. The pore volume either remained the same or decreased after impregnation, which agrees with the literature.70, 94, 199 The pore volume decreased by 16% for 1D cylindrical FDU-15 with endo-pore CuO, most likely due to decreased access to the internal pores after impregnation. The pore volume remained unchanged for the more accessible 3D bicontinuous CMK-8 architecture after CuO loading. The pore volume of BPL/CuO also did not change relative to native BPL, which we attribute to the absence of CuO impregnants within the micropores. However, the BPL/CuO surface area increases by 8%, from 1229 m2 g-1 to 1300 m2 g-1, after the addition of exo-pore CuO particles. (a) (b) (c) (d) Figure 3.6. BJH pore size distributions to compare the bare carbons and the carbon/CuO composites: (a) BPL, BPL/CuO, (b) FDU-15, FDU-15/CuO, (c) CMK-3, CMK-3/CuO, and (d) CMK-8, CMK-8/CuO. 78 Table 3.1. N2 Adsorption–Desorption Properties of Unimpregnated and CuO Loaded Carbons Carbon Sample Pore Diameter Total pore Total Surface Area (nm)a volume (cm3 g-1)b (m2 g-1)c BPL <1.7 0.39 1229 BPL/CuO d (exo-pore) <1.7 0.39 1330 FDU-15 3.6 0.32 600 FDU-15/CuO d d (endo and exo-pore) 3.3 0.32 760 FDU-15/CuO (endo-pore) 3.3 0.27 d 591d CMK-3 2.8 0.45 976 CMK-3/CuO (endo-pore) 2.8 0.38 d 935d CMK-8 3.5 1.14 1075 CMK-8/CuO d d (endo-pore) 3.5 1.15 1246 CuO nanoparticles 82 aCalculated using the Barrett-Joyner-Halenda (BJH) model. bThe total pore volume is based on the sum of the mesopore and macropore volume from the BJH model and the micropore volume from the t-plot method. cCalculated using the multipoint Brunauer-Emmett-Teller (BET) method. dValues were corrected for the additional weight due to CuO loading (wt% listed in Table 4.2) by normalizing measurements to carbon weight only. 79 3.3.1.3 Determination of Cu Oxidation State and CuO Loading ICP-AES and XRD indicate a 6.4−10 wt% loading of Cu(II) oxide on the carbon/CuO samples. For endo and exo-pore impregnated CuO on FDU-15, 55% of the CuO nanoparticles are located endo-pore, while 45% is formed exo-pore (listed in Table 3.2). ICP-AES measurements were taken before and after the selective removal of exo-pore CuO, previously described in Section 3.2.3.2. Quantitative elemental analysis for all carbon/CuO composites were determined using ICP-AES, and presented in Table 3.2. X-ray diffraction showed that the CuO is crystalline and adopts a monoclinic structure (Figure 3.8 and 3.9). In addition, the wider diffraction peaks observed in the XRD spectra are consistent with 2−4 nm CuO nanoparticles observed in the TEM studies. Table 3.2. Weight % Loading of CuO Impregnated Carbons CuO Weight % BPL/CuO (exo-pore) 10.0 8.3 (total) FDU-15/CuO (endo and exo-pore) 4.6 (endo-pore) 3.7 (exo-pore) FDU-15/CuO (endo-pore) 6.4 CMK-3/CuO (endo-pore) 10.6 CMK-8/CuO (endo-pore) 9.8 Intensities were measured using ICP-AES at 327.396 nm for Cu. 80 (a) endo & exo-pore (b) endo & exo-pore endo-pore endo-pore Figure 3.7. (a) XRD diffraction patterns and (b) BJH pore size distributions of endo and exo-pore (red) and endo-pore only CuO impregnated FDU-15 (blue). Figure 3.8. C2-XRD diffraction pattern of FDU-15/CuO (black), CMK-3/CuO (red), and CMK-8/CuO (blue) composites. The peaks correspond to carbon and CuO Tenorite (PDF 01-073-6023). 81 3.3.2 DMMP Adsorption Properties TPD experiments were conducted to determine the DMMP adsorption and decomposition activity on carbon/CuO composites. Samples were evaluated after 70 h of DMMP exposure to probe the adsorption regime between full surface coverage and liquid saturation. It is well-known that a linear relationship exists between effective adsorption capacities and total surface area for unimpregnated cabons6 in the absence of chemisorption processes.203, 204 DMMP adsorption on carbon/CuO composites deviate from strictly surface area dependence, due to the additional effects of CuO located endo and/or exo-pore. Our studies suggest that CuO nanoparticles provide additional active sites for chemisorption, which results in greater DMMP adsorption capacity and decomposition. Generally, the carbon/CuO composites have higher (up to 64%) decomposed and desorbed DMMP capacity, in comparison to their unimpregnated carbon counterparts (Figure 3.9 and Table 3.3). Multiple factors contribute to this observation and will be discussed in depth in the subsequent sections. 82 Figure 3.9. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after TPD experiments with the unimpregnated carbons (green) and carbon/CuO composites (blue) after 70 h of DMMP exposure. 83 Table 3.3. 70 h Mass Capacity of Decomposed DMMP, Desorbed DMMP, and Total Mass Capacity for Unimpregnated Carbons and Carbon/CuO Composites Desorbed Total Mass Carbon Sample Decomposed DMMP (g/g)a,b DMMP Capacity (g/g)b (g/g)b BPL 0.0067 0.2237 0.2304 BPL/CuO (exo-pore) 0.0079 0.2838 0.2917 FDU-15 0.0079 0.1697 0.1776 FDU-15/CuO (endo & exo-pore) 0.0188 0.1572 0.1760 FDU-15/CuO (endo-pore) 0.0180 0.2134 0.2314 CMK-3 0.0051 0.2607 0.2658 CMK-3/CuO (endo-pore) 0.0133 0.3977 0.4110 CMK-8 0.0065 0.4489 0.4554 CMK-8/CuO (endo-pore) 0.0279 0.7201 0.7480 CuO Nanoparticles 0.0005 0.0078 0.0083 aDecomposed DMMP was calculated by applying a conversion factor for the mole ratio of DMMP to methanol from Table 3.4. bMeasurements reported for the carbon/CuO composites were calculated after disregarding the CuO weight for a more accurate comparison to the unimpregnated carbon samples. 84 3.3.3 DMMP Decomposition Activity A few research groups have observed extensive decomposition of DMMP on copper surfaces.31, 205 Trotochaud et al. used in situ ambient pressure X-ray photoelectron (APXPS) and infrared spectroscopies, to observe DMMP bond cleavages of PO-CH3, P-OCH3, and P-CH3 at room temperature, consequently forming methoxy groups on the CuO surface (Figure 4.10).31 Ma et al. used mass spectrometry and XPS to observe gaseous H2, methane, methyl, formaldehyde, methanol, and molecular DMMP, as well as, atomic carbon, phosphorus, and POx, on Cu films and clusters surfaces, after DMMP thermal decomposition.205 These studies and additional investigations of DMMP decomposition on other metal oxide surfaces (TiO ,2, 24, 26, 30 MgO,8, 16, 35-37 LaO,16 Al O ,16, 17, 23, 25, 34, 38, 392 2 3 Fe O ,14, 162 3 ), suggest that DMMP decomposes to surface phosphate, alkoxide, and alkyl groups, while forming gaseous byproducts, such as carbon dioxide, carbon monoxide, dimethyl ether, formaldehyde, formic acid, and more commonly, methanol. 85 Figure 3.10. Proposed pathways for DMMP decomposition on CuO surface. Taken from Reference 31. 3.3.3.1 Determination of DMMP Conversion to Methanol (MeOH) Decomposition of DMMP on carbon/CuO substrates generates one mole of MeOH for each mole of DMMP that decomposes. This conclusion was determined from a combination of quantitative 1H-NMR and ICP-AES measurements. The evolution of gas-phase products, predominantly methanol with trace amounts of formaldehyde, was measured using a method described in Section 3.2.6−3.2.7. No phosphorus-containing products other than molecular DMMP were detected, which agrees with previous observations in the literature.24 The phosphorus content remaining on the carbon/CuO adsorbents after DMMP 86 decomposition were quantitatively evaluated using ICP-AES after the TPD experiments. These experiments indicate that the amount of phosphorus (mg P/g material) observed agrees with the expected amount of phosphorus, which was calculated with the assumption that each mole of DMMP produces a mole of gaseous MeOH (Table 3.4). CH3 H3CO OCH3 P H3C O CH3 H3C O CH3 H3C P H3C P O OCHO O 3 O O + CH3OH Figure 3.11. Illustration of DMMP decomposition to produce gaseous methanol (CH3OH) and surface-bound methyl methylphosphonate (MMP). Table 3.4. Phosphorus Concentration of Cycled Samples FDU-15/CuO FDU-15 (endo & exo-pore) Observeda,b 2.22 3.13 mg P/g material Equivalent Amount of DMMP 1.02 0.98 aObserved value for FDU-15 was measured using ICP-AES at 177.5 nm for P. aObserved value for FDU-15/CuO (endo & exo-pore) was measured using ICP- AES at 178.2 nm for P. 87 3.3.3.2 Thermal Decomposition of DMMP As described above, the methanol liberated in the TPD experiments can be used as a quantitative indicator of DMMP decomposition capacity. We speculate that the methanol produced on the inert carbon surface is caused by thermal decomposition, due to high operating temperatures (up to 450 °C) in our TPD experiments. Methanol production from unimpregnated FDU-15 was measured at multiple exposure times, and compared to FDU/CuO composites to differentiate between the amount of methanol produced from thermal decomposition or reactive decomposition with the CuO surface. The results showed that the decomposed DMMP capacity for bare FDU-15 was consistently between 0.007−0.008 g DMMP/g FDU-15, regardless of the length of DMMP exposure (Figure 3.12). Specifically, the extent of decomposition (i.e. decomposed DMMP capacity) for the bare FDU-15 after 70 h of DMMP exposure is 0.008 g DMMP/g FDU-15, whereas both FDU-15/CuO composites have a decomposed DMMP capacity over 125% higher (0.018−0.019 g DMMP/g FDU-15). We speculate that the 0.010−0.011 g DMMP/g FDU-15 difference is due to the DMMP decomposition activity on the CuO surface. The decomposed DMMP capacities listed in Table 3.5 neglect the methanol produced from thermal decomposition to more accurately assess CuO effects on DMMP adsorption and decomposition, in comparison to the other samples shown graphically in Figure 3.13. 88 Figure 3.12. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after multiple TPD experiments with FDU- 15 after 5 h, 25 h, 70 h, and 2 weeks of DMMP exposure. 89 Table 3.5. 70 h Mass Capacity of Decomposed DMMP Not Including Methanol Production from Thermal Decomposition, Desorbed DMMP, and Total Mass Capacity for Unimpregnated Carbons and Carbon/CuO Composites Carbon/CuO Decomposed Desorbed Total Mass DMMP (g/g)a,b DMMP (g/g)b Capacity (g/g)b BPL/CuO (exo-pore) 0.0012 0.2838 0.2850 FDU-15/CuO (endo & exo-pore) 0.0109 0.1572 0.1681 FDU-15/CuO (endo-pore) 0.0101 0.2134 0.2235 CMK-3/CuO (endo-pore) 0.0082 0.3977 0.4059 CMK-8/CuO (endo-pore) 0.0214 0.7201 0.7415 aDecomposed DMMP was calculated by applying a conversion factor for the mole ratio of DMMP to methanol from Table 3.4. bMeasurements reported for the carbon/CuO composites were calculated after disregarding the CuO weight for a more accurate comparison to the unimpregnated carbon samples. Figure 3.13 summarizes the DMMP desorbed and decomposed capacity for OMCs and the CuO impregnated OMCs. In general, the carbon/CuO composites (blue bars) achieve higher desorbed and decomposed capacity than their unimpregnated carbon counterparts (green bars), as shown in Figure 3.12. While 90 their pore properties are relatively similar (Table 3.1), the CuO impregnants improve DMMP adsorption and decomposition due to stronger binding and higher surface reactivity. Figure 3.13. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after a TPD experiment with the different unimpregnated (green) and CuO loaded carbons (blue) after approximately 70 h of DMMP exposure. 91 3.3.3.3 Surface Reactivity Comparison for CuO and Carbon The experiments above show that the CuO surface reactivity is significantly higher in comparison to the carbons, resulting in greater DMMP adsorption and decomposition. The adsorption properties of CuO nanoparticles, BPL, FDU-15, CMK-3, and CMK-8, were studied at low surface coverages (i.e. after 5 h of DMMP exposure) to reveal the differences in surface reactivity. Desorbed and decomposed DMMP capacities were measured after 5 h of DMMP exposure, and normalized by their surface area (Figure 3.13, Table 3.6). The CuO surface activity (total DMMP area capacity) was over 3.5 times higher in comparison to the OMCs (0.075−0.078 m2 g-1), and 17 times higher than BPL (0.016 m2 g-1). The significantly higher CuO reactivity observed could explain the increase in decomposed and desorbed capacities for the carbon/CuO composites, in comparison to their unimpregnated counterparts. The chemisorbed DMMP molecules can react with the CuO impregnants, ultimately resulting in the observed increase in adsorption potential and decomposition, which agrees with previous observations made by Saxena et al.69 92 Figure 3.14. Graphical representation of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) after a TPD experiment with BPL, FDU-15, CMK-3, CMK-8, and CuO nanoparticles after 5 h of DMMP exposure. Table 3.6. 5 h Area Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for the Carbons and CuO Nanoparticles Carbon Sample Decomposed Desorbed Total Area Capacity DMMP (g m-2)a DMMP (g m-2) (mg m-2) BPL 0.016 FDU-15 0.076 CMK-3 0.074 CMK-8 0.078 CuO nanoparticles 0.007 0.269 0.276 93 3.3.3.4 Effect of CuO Size and Location on DMMP Decomposition The size and location of CuO nanoparticles affect the adsorption and decomposition of DMMP on carbon/CuO composites with similar loading levels. The highest decomposition activity was observed for the smaller endo-pore CuO nanoparticles with larger surface area, and more reactive sites. During the impregnation process, pore constrictions resulted in smaller CuO nanoparticles formed inside the pores. The DMMP decomposition capacity per g CuO (wt%) for endo-pore and exo-pore CuO were compared to probe the effects of CuO size and location on DMMP decomposition. The smaller CuO nanoparticles (2−4 nm) located within the endo-pores showed the highest DMMP decomposition capacity per g CuO (1.58 mg DMMP/g material). The DMMP decomposition capacity for the exo-pore CuO nanoparticles (30−50 nm) supported on FDU-15, per CuO wt% decreases by 38% (0.98 mg DMMP/g material). For the larger CuO particles, such as the aggregates on the BPL surface, the DMMP decomposition capacity was significantly lower, 0.12 mg DMMP/g material per g CuO (Table 3.7). These results suggest that the smaller endo-pore CuO nanoparticles are more reactive, resulting in higher DMMP decomposition in comparison to larger exo-pore CuO particles. 94 Table 3.7. DMMP Decomposition Capacity (mg g-1) per g CuO (wt%) Endo-pore CuO Exo-pore CuO Decomposition Activity Decomposition Activity (mg g-1) (mg g-1) FDU-15/CuO (endo-pore) 1.58 0.00 FDU-15/CuO (endo & exo-pore) 1.58 0.98 BPL/CuO (exo-pore) 0.00 0.12 3.3.3.5 Significance of Structural Effects The 3D architecture of CMK-3 and CMK-8 allows for higher DMMP decomposition and adsorption into the more accessible structure in comparison to 1D FDU-15, after CuO impregnation. The higher DMMP decomposition is a result of the more accessible CuO reactive sites in CMK-3 and CMK-8. These studies suggest that all the carbon/CuO samples show higher DMMP adsorption, apart from the CuO endo and exo-pore impregnated FDU-15 (Figure 3.13). The CuO exo-pore impregnants blocked access into the internal pores (Figure 3.15). This blockage limits DMMP diffusion through the 1D architecture of the FDU-15 pores, resulting in lower DMMP adsorption after 70 h of DMMP exposure. 95 Blocked by CuO NPs Accessible by other directions 1D pores (FDU-15) 3D pores (CMK-3) Figure 3.15. Representation of the limited diffusion of DMMP into the CuO loaded 1D pores of FDU-15, in comparison to the more accessible 3D pores of CMK-3 and CMK-8. CMK-8/CuO achieved the highest DMMP adsorption and decomposition in comparison to the other carbon/CuO samples (Table 3.5). Given that the CuO loading amount and impregnation process was similar for all the samples, the difference in adsorption/decomposition activity is linked to the differences in the CuO size, location, and 3D architecture of CMK-8. The more reactive, smaller, endo-pore CuO nanoparticles are more accessible within CMK-8, in comparison to the other carbon/CuO systems. While CMK-3 also possesses 3D pores, its CuO impregnants are larger than the CuO nanoparticles that reside in CMK-8 (Figure 3.4). The smaller CuO nanoparticles in CMK-8 have higher surface area for greater reactivity. In addition to the advantages of a 3D pore architecture and nano-sized CuO impregnants, CMK-8/CuO has the largest pore volume, allowing for higher DMMP uptake after extended periods of exposure time (Table 3.5). These factors result in an increase in the performance of CMK-8/CuO for DMMP removal relative to the other systems. 96 3.4 CONCLUSIONS At 70 h of DMMP exposure, our studies show that CMK-8/CuO is the best system for DMMP removal. The nano sized endo-pore CuO in CMK-8 was the most reactive material for DMMP decomposition, in comparison to the other carbon/CuO composites with similar CuO loading. The 3D pore architecture and large pore volume also allowed for easy accessibility and higher DMMP adsorption capacity. These findings are important factors that influence the effectiveness of DMMP removal from the airstream. During our investigation, we observed multiple limitations to the inferior performing carbon/CuO composite, FDU-15 with endo and exo-pore CuO. Our studies suggest that FDU-15/CuO exo-pore impregnants occupy the mesopores and block internal surface sites, consequently decreasing the adsorption potential. Due to these complications, there is a need for a new, more effective absorbents with high surface area, and abundant reactive surfaces. In our next chapter, we will explore mesoporous metal oxides as more effective adsorbents for CWA defeat. 97 Chapter 4: Dimethyl Methylphosphonate Adsorption Properties and Decomposition Activity on Mesoporous Metal Oxides 4.1 INTRODUCTION Advancements in gas mask filtration devices heavily rely on our understanding of CWA adsorption and decomposition mechanisms (described in Chapter 1), and how adsorbent properties affect their performance (revealed in Chapter 2 and 3). The investigation of ASZM-TEDA revealed the advantages of key design features and the disadvantages that hinder their functionality. These findings can guide us in the synthesis of new, superior filters. Our research efforts have revealed the significance of the carbon adsorbents in Chapter 2. The large surface area (600−1300 m2 g-1) and pore volumes (0.2−1.2 cm3 g-1) allow for high DMMP adsorption capacities (up to 1.2 cm3 DMMP/g carbon). After increasing the pore uniformity from the disordered BPL carbon to the ordered mesoporous carbon, stronger binding to the carbon surface is achieved, as demonstrated by the higher DMMP desorption energies (30−40% increase). However, this binding is still characterized as physisorption and is relatively weak in comparison to interactions with metal oxide impregnants. Therefore, due to the weak bonding and inert carbon surface, there is little to no DMMP decomposition observed. 98 For carbon/CuO composites investigated in Chapter 3, the DMMP mass adsorption capacity was increased up to 64% in comparison to their unimpregnated counterparts, apart from 1D cylindrical FDU-15 with CuO endo and exo-pore. The exo-pore CuO is believed to decrease the adsorbent performance by blocking internal adsorption sites after metal oxide deposition. In addition, the endo-pore CuO decreased the pore volume up to 18%, which consequently, lowers the adsorption capacity for extended periods of DMMP exposure. Due to these disadvantages and additional shortcomings, such as low breathability and deactivation from surface poisoning on the carbon-based adsorbents,88, 206, 207 there is a need for more effective, catalytic adsorbents with high surface areas, and more reactive adsorption sites.203, 208, 209 In this chapter, ordered mesoporous metal oxides (OMM) are synthesized with 1D cylindrical and 3D hexagonal architectures with more reactive sites, and designed to presumably have better mass transfer into the pores in comparison to impregnated carbons. Additionally, the OMMs have a higher surface area and pore volume in comparison to metal oxide nanoparticles, for greater adsorption and decomposition of DMMP. We describe the extent of DMMP removal for the OMMs: alumina (Al2O3), titania (TiO2), ceria (CeO2), iron oxide (Fe2O3), and aerogel-Fe2O3 in comparison to FDU-15/CuO (endo and exo-pore), FDU-15, and BPL, using TPD/quantitative 1H-NMR and ICP-AES bulk analysis. Additional 99 surface chemistry insight was provided by our collaborators using infrared spectroscopy and a theoretical approach. 4.2 EXPERIMENTAL SECTION 4.2.1 Materials AR grade DMMP (≥97.0%) and metal precursors (Al(NO3)3·9H2O, Ti(OBu)4, Ce(NO3)2, Fe(NO3)3) were purchased from Sigma Aldrich (USA) and used without further purification. 4.2.2 Mesoporous Metal Oxides The mesoporous metal oxides, Al2O3, TiO2, CeO2, Fe2O3, and aerogel Fe2O3 were successfully synthesized by a soft template (Figure 4.1), hard template (Figure 4.2), and sol-gel method (Figure 4.3). The OMMs, Al2O3 and TiO2, were prepared by myself and Dr. Junkai Hu, respectively, using slight modifications of methods described in the literature.210-212 The OMMs, CeO2 and Fe2O3, were synthesized by a different method developed by Dr. Junkai Hu. Fe2O3 aerogels were prepared by our collaborators at the Naval Research Laboratory (NRL), Dr. Jeffrey Long, using slight modifications of methods previously published in the literature.213-215 100 4.2.2.1 Synthesis of Mesoporous Alumina (meso-Al2O3) Meso-Al2O3 was synthesized by a soft template method described by Wang et al.210 In a typical reaction, 5.6 g of Al(NO3)3·9H2O was dissolved in a solution of anhydrous ethanol and deionized water (15 mL), with the ratio of 0W7XY = \. Then the mixture was transferred to a petri dish without cover for the 0Z[754 = solvothermal pre-hydrolysis treatment, where the sample was dried at 80 °C for 5 h. The Al-OH solid was slowly added to a solution containing 1.8 g of P123 (diblock copolymer, EO20PO70EO20, where EO = ethylene oxide, PO = propylene oxide) and 0.6 g of citric acid, dissolved in anhydrous ethanol (30 mL). After vigorous stirring for 24 h at 30 °C, the resultant mixture was transferred to a petri dish, and dried at 45 °C for 48 h. The resulting solid was then dried at 100 °C for 24 h in a box furnace. The final product was calcined at 400 °C under O2 for 5 h to remove the template and the citric acid. 4.2.2.2 Synthesis of Mesoporous Titania (meso-TiO2) Meso-TiO2 was synthesized by a soft template method described by Fan et al.211 and Brinker et al.212 In a typical reaction, 10 mmol of Ti(OBu)4, 40 mmol of acetic acid, 24 mmol of HCl, and 1.6 g of F127 (triblock copolymer, EO96PO70EO96, where EO = ethylene oxide, PO = propylene oxide) were dissolved in ethanol (30 mL). After vigorous stirring for 1 h at 30 °C, the resultant mixture was transferred to a petri dish, and dried at 40 °C for 12 h. The resulting 101 solid was then heated at 65 °C for 24 h in a box furnace. The final product was calcined at 350 °C under O2 for 5 h to remove the template. Surfactant Self-assembly removal Ordered Mesoporous Metal salt/ Metal Oxide surfactant micelles (OMM) Figure 4.1. Schematic of ordered meso-Al2O3/meso-TiO2 synthesis using a soft template method. 4.2.2.3 Synthesis of Mesoporous Ceria (meso-CeO2) and Iron Oxide (meso-Fe2O3) Meso-CeO2 and meso-Fe2O3 were synthesized by a hard template method developed by Dr. Junkai Hu. In a typical reaction, 0.2 g of SBA-15 was fully saturated with a metal nitrate solution (50 wt% for Ce(NO3)2 and 45 wt% for Fe(NO3)3). Then the excess solution was removed by vacuum filtration. The SBA-15 filled with metal nitrate solution was collected, and dried at 90 oC overnight, then annealed at 170 °C or 350 °C for Fe2O3 and CeO2, respectively. Next, samples were annealed for 3 h under Ar flow with a heating rate of 1°C min-1. The impregnation process was repeated again, except the 2nd annealing step is 400 °C for 3 h for both Fe2O3 and CeO2. The silica was removed by dissolution with 3M NaOH solution at 50 °C for 1 h, and this process was repeated twice 102 until the SiO2 was removed, and confirmed by SEM/EDX. The meso-CeO2 or meso-Fe2O3 was collected by centrifugation and washed with deionized water. The extraction/washing process was repeated three times. MNO3 impregnation Silica removal Solvent evaporation SiO2 + 2NaOH = Na2SiO3 +H2O Ordered & salt decomposition Ordered Mesoporous Mesoporous Metal Oxide Silica (OMM) Figure 4.2. Schematic of ordered meso-CeO2/ meso-Fe2O3 synthesis using an ordered mesoporous silica template, SBA-15. 4.2.2.4 Mesoporous Iron Oxide Aerogel (aerogel-Fe2O3) The Fe2O3 aerogel was synthesized by Dr. Jeff Long (NRL) according to a published sol-gel method described by Long et al.213 and Gash et al.214, 215 Supercritical evaporation Iron Oxide Aerogel Sol-gel Iron Oxide Figure 4.3. Schematic of aerogel-Fe2O3 synthesis using a sol-gel method. 103 4.2.3 Characterization of Mesoporous Metal Oxides Pore size, volume, and surface area of the mesoporous metal oxides were determined from N2 adsorption-desorption isotherms recorded with a Micromeritics ASAP 2020 Porosimeter Test Station. Samples were degassed in vacuum at 100 °C for 12 h prior to characterization. The specific surface areas were calculated using the Brunauer-Emmett-Teller (BET) method from the N2 adsorption data in the relative pressure range (P/P0) of 0.05−0.20. The general porosity distribution was calculated from the adsorption branch using the Barrett- Joyner-Halenda (BJH) equation. Structural morphologies of the samples were imaged with a scanning electron microscope (Hitachi SU-70 SEM, operated at an acceleration voltage of 10 kV), and a transmission electron microscope (JEM 2100 Field Emission operated at 200 kV). The samples were dispersed in methanol and a 10 µL aliquot of the resulting dispersion was drop cast on the TEM grids. The TEM grids used were carbon-coated Cu grids (CF200-Cu, Electron Microscopy Sciences). X-ray diffraction (XRD) patterns were recorded by a Bruker Smart1000 (Bruker AXS Inc., USA) using CuKα radiation. 4.2.4 Exposure of Mesoporous Metal Oxides to DMMP Mesoporous metal oxides were pre-treated at 100 °C for ≥ 6 h under static vacuum to remove any adsorbed contaminants before exposure to DMMP using the vial-in-vial method, modified from the literature.189 Approximately 75 mg of 104 material was transferred to small glass vials. Each small glass vial was placed inside a larger vial that contained a sufficient amount of liquid DMMP (200 𝜇L), to saturate the head space of the vial assembly. The larger vial was then capped, and stored in a desiccator. The exposure process was conducted under ambient pressure at room temperature (24−26 °C). The samples were removed after 5 h, and analyzed using a fixed-bed reactor, which is described in Section 4.2.5. 4.2.5 Fixed-Bed Reactor Setup TPD studies on the samples were conducted using a temperature controlled fixed-bed reactor system. Each sample (75 mg) was loaded into a 10 mm ID quartz reactor. The reactor was composed of a quartz frit with ~10 mg quartz wool layered atop the frit. The sample was placed on top of the wool. After loading the sample into the reactor, the system was sealed with Swagelok fittings, and Ar (Standard, Airgas) was flowed at a constant flow rate of 50 SCCM (standard cubic centimeters per minute). The mass flow was controlled with a Brooks 5850 series mass flow controller. A Eurotherm 91 P PID controller and K- type thermocouple controlled the sample temperature. All lines downstream of the reactor were maintained at a constant 150 °C before and during testing to minimize DMMP adsorption in effluent lines. The effective adsorption capacity measurements were programmed to include the following temperature regimes: (7) A linear temperature increase of 10 °C min-1, from 30 °C to 450 °C (8) Hold at 450 °C for 2 min 105 (9) A fast cooling to 30 °C The reactor effluent was directed through a bubbler filled with a mixed solution of deuterated acetonitrile (CD3CN) and ethylene carbonate (EC) that sufficiently submerged the bubbler stem. The CD3CN/EC solution was maintained at -42 °C via a dry ice/acetonitrile bath. The sample was analyzed using 1H-NMR, which is described in Section 4.2.6.1. 4.2.6 Methods 4.2.6.1 1H-NMR Study NMR spectra were recorded using a Bruker AV-400 MHz Spectrometer. For each spectrum, 128 transients were collected at 298 K with an acquisition time of 2.6 s, d1 relaxation time of 4 s. The DMMP-acetonitrile-d3 solutions were spiked with precise amounts of EC that was added as a quantitative calibrant to measure the amount of DMMP and byproducts desorbed. 4.2.6.2 ICP-AES Study The phosphorus content of the cycled adsorbent materials were determined using ICP-AES characterization performed on a Perkin Elmer ICP Optima 4700. The samples were dissolved in pure trace-metal grade nitric acid (Sigma-Aldrich) and diluted with milli-Q water to a known volume (100 mL) with a final concentration of 2 wt% nitric acid (HNO3), before administered to the plasma. Intensities were measured at 178.3 nm for P. A linear calibration curve (R2 = 0.9999) was collected with five standards, a blank (2 wt% HNO3), 0.1, 0.2, 0.5, and 1.0 ppm P (High Purity Standard, 1000 𝜇g/g P in 2 wt% HNO3). 106 4.2.7.3. FTIR Spectroscopy Mr. Scott Holdren collected FTIR spectra on a Nicolet cooled MCT-A detectors. A Harrick Scientific Praying Mantis DRA optical accessory was used with an associated Harrick Scientific high temperature reaction chamber HVCDRP-5 for the DRIFTS measurements. 4.2.7.4. DFT Calculations Dr. Roman Tsyshevsky calculated the adsorption of DMMP on metal oxide surfaces by means of DFT-based modeling. Solid state periodic calculations were performed using hybrid functionals, which include corrections for weak van der Waals interactions and projector augmented-wave (PAW) pseudo-potentials, as implemented in the VASP code as described elsewhere.216- 218 4.3 RESULTS AND DISCUSSION 4.3.1 Characterization of Mesoporous Metal Oxides In this investigation, we studied five mesoporous metal oxides, which include four OMMs: Al2O3, TiO2, CeO2, and Fe2O3 with well-defined pore size, volume, and surface area. Disordered mesoporous aerogel Fe2O3 was also investigated for comparison. The structures of the OMMs were first investigated by electron microscopy (TEM and SEM) and X-ray diffraction. These characterization techniques verified the ordered pore structure, size, and 107 crystallinity (or lack thereof) for the OMMs. Figure 4.4 and 4.5, reveals the 1D cylindrical structure of amorphous meso-Al2O3 and crystalline rutile/anatase phase meso-TiO2, respectively. Figure 4.6 and 4.7 shows the 3D hexagonal structure of crystalline cerianite phase meso-CeO2 and hematite phase meso- Fe2O3, respectively. Aerogel-Fe2O3 is also hematite phase, however the structure is more disordered with varying pore sizes that are difficult to distinguish (Figure 4.8). (a) (b) 200 nm 100 nm (c) Amorphous Figure 4.4. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of amorphous meso-Al2O3. 108 (a) (b) 200 nm 100 nm (c) Figure 4.5. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline anatase and rutile phase meso-TiO2. 109 (a) (b) 200 nm 100 nm (c) Figure 4.6. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline cerianite phase meso-CeO2. 110 (a) (b) 200 nm 100 nm (c) Figure 4.7. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline hematite phase meso-Fe2O3. 111 (a) (b) 200 nm 100 nm (c) Figure 4.8. (a) SEM image, (b) TEM image, and (c) X-ray diffraction pattern of crystalline hematite phase aerogel-Fe2O3. The OMMs and OMCs with similar architectures have different pore properties, such as the surface area, pore size, and volume, which are compared in Figure 4.9. Meso-Al2O3 and meso-TiO2 were synthesized by a soft template method, similar to the 1-D cylindrical FDU-15. The comparisons between meso- Al2O3, meso-TiO2, and FDU-15, reveal larger pore sizes (7.2 and 5.1 nm, respectively) in comparison to FDU-15 (3.6 nm) (Figure 4.9a). Meso-CeO2 and meso-Fe2O3 were synthesized by a hard template method, similar to 3D hexagonal 112 CMK-3. The comparisons between their pore size distributions are shown in Figure 4.9b, revealing a larger average pore size of 4.4 nm for meso-Fe2O3, but similar sizes for meso-CeO2 and CMK-3 (2.5 and 2.7 nm, respectively). On the other hand, aerogel-Fe2O3 has a wide range of pore sizes, which are much larger than any of the OMMs (Figure 4.10). The detailed structure parameters are listed in Table 4.1. These data are in good agreement with our TEM results and previous studies in the literature.184-186, 210-215 113 (a) 0.08 3.6 nm 0.07 7.2 nm Al2O3 5.1 nm TiO2 0.06 FDU-15 0.05 0.04 0.03 0.02 0.01 0.00 0 2 4 6 8 10 12 14 16 18 20 Pore Size (nm) (b) 0.0035 2.7 nm CeO2 Fe2O3 CMK-3 0.0030 0.03 CeO 0.0025 2 Fe2O3 0.02 0.0020 0.0015 0.01 0.0010 0.00 0 5 10 15 20 0.0005 2.5 nm 4.4 nm Pore Size (nm) 0.0000 0 2 4 6 8 10 12 14 16 18 20 Pore Size (nm) Figure 4.9. BJH pore size distributions for (a) meso-Al2O3 (pink), meso-TiO2 (blue), and FDU-15 (gray), as well as (b) meso-CeO2 (green), meso-Fe2O3 (brown), and CMK-3 (gray). 114 dV/dD (cm3 g-1 nm-1) dV/dD (cm3 g-1 nm-1) dV/dD (cm3 g-1 nm-1) 0.07 Meso-Al2O3 0.06 Meso-TiO2 Meso-CeO 0.05 2 Meso-Fe2O3 0.04 Aerogel-Fe2O3 0.03 0.02 0.01 0.00 0 5 10 15 20 25 30 35 40 45 50 Pore Size (nm) Figure 4.10. BJH pore size distributions for the diverse mesoporous metal oxides, meso-Al2O3 (pink), meso-TiO2 (blue), meso-CeO2 (green), meso-Fe2O3 (orange), and aerogel-Fe2O3 (red). Table 4.1 lists the surface area, pore size, and volume of the all the mesoporous metal oxides, as well as FDU-15/CuO (endo and exo-pore), FDU-15, and BPL, for comparison. Of all the mesoporous metal oxides, TiO2 has the highest surface area of 239 m2 g-1, while Al2O3 (175 m2 g-1), CeO 2 -12 (123 m g ), Fe2O3 (188 m2 g-1), and aerogel-Fe2O3 (131 m2 g-1) are relatively similar. Aerogel- Fe2O3 has the highest pore volume (0.87 cm3 g-1), over two times greater than Al2O3 (0.36 cm3 g-1), CeO2 (0.28 cm3 g-1), Fe2O3 (0.30 cm3 g-1), and TiO2 (0.23 cm3 g-1). These differences in pore properties (represented graphically in Figure 4.11) are significant, as we have seen from our previous studies how they influence DMMP adsorption and decomposition. 115 dV/dD (cm3 g-1 nm-1) Table 4.1. N2 Adsorption –Desorption Properties of Adsorbents Samples Pore Diameter Total Pore Volume Total Surface (nm)a (cm3 g-1)b Area (m2 g-1)c Meso-Al2O3 7.2 0.36 175 Meso-TiO2 5.1 0.23 239 Meso-CeO2 2.5 0.28 123 Meso-Fe2O3 4.4 0.30 188 Fe2O3 aerogel wide range 0.87 131 FDU-15/CuO endo and exo-pore 3.3 0.32 760 FDU-15 3.6 0.32 600 BPL <1.7 0.39 1229 aCalculated using the Barrett-Joyner-Halenda (BJH) model. bThe total pore volume is based on the sum of the mesopore and macropore volume from the BJH model and the micropore volume from the t-plot method. cCalculated using the multipoint Brunauer-Emmett-Teller (BET) method. 22 2 Figure 4.11. Graphical representation of the surface area and pore volume of the OMMs, meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and aerogel-Fe2O3. 116 4.3.2 DMMP Decomposition Activity DMMP decomposition has been observed on many dry and hydroxylated metal oxide surfaces, such as CuO,31 MoO ,22, 40 TiO ,2, 24, 26, 30 MgO,8, 16, 35-37 3 2 LaO,16 Al O ,16, 17, 23, 25, 34, 38, 392 3 Fe 14, 162O3, and Ni, Fe, Cu, and V oxide supported on γ-Al2O3, SiO 252, TiO2. These studies revealed that DMMP decomposition results in the formation of surface phosphate, alkoxide, and alkyl groups, while subsequently forming gaseous byproducts, such as carbon dioxide, carbon monoxide, dimethyl ether, formaldehyde, formic acid, and most commonly, methanol (Figure 4.12).2, 14-17, 23, 24, 26, 30, 34, 37, 38 The different mechanistic pathways were previously discussed in Chapter 1, Section 1.2. Solids Gases CO , CO P CH3OH, 2 O M O M O M O H H3C CH3 H3CO P OCH3 + O CH3 O CH3 M O M O M O M O M O M H H O CH3 M O M O M H OH Figure 4.12. General schematic of DMMP decomposition on hydroxylated metal oxides, and the possible surface and gaseous byproducts. DRIFTS studies were performed to monitor DMMP adsorption and decomposition on meso-TiO2 and meso-CeO2. The DRIFTS spectrum of DMMP 117 shows vibrational modes in good agreement with gas phase DMMP vibrational frequencies and modes in the literature.26 The spectra shows that DMMP adsorbs and decomposes on meso-TiO2 and meso-CeO2 surfaces from the presence of the peak at 1098 cm-1, which corresponds to a surface methoxy group, denoted as M- OCH3 in Figure 4.13. This identification is based on prior work that investigated methanol adsorption on ThO2, CeO2, TiO2, and CuO.26, 219, 220 In addition, the peak corresponding to the phosphoryl (P=O) bond of adsorbed DMMP is shifted to a lower frequency with respect to the DMMP gas phase spectrum (1276 to 1242 cm-1), which is a common feature for adsorbed DMMP on metal oxides.26, 221 The lower frequency observed from the shift of the P=O is most likely due to a strong interaction of the phosphoryl’s oxygen with a metal Lewis acid site or surface hydroxyl. The decrease in absorbance of the surface OH groups further supports DMMP interaction with the hydroxyls on the surface of these metal oxides. These hydroxyl groups can promote binding and decomposition, which leads to methanol formation at low temperatures. 118 Mesoporous TiO nOCH2 3 Mesoporous CeO H 3 H 32 PC OC Gas Phase DMMP nnPCH n3 M-OCH3 Surface OH groups 3600 3300 3000 2700 Wavenumbers (cm-1) nOCH nP=O M-OCH 33 rOCH3 dPCH3 dPCH3 nPO2 1400 1300 1200 1100 1000 900 800 700 Wavenumbers (cm-1) Figure 4.13. DRIFTS spectrum of gas phase DMMP (black line) in comparison to commercial DMMP adsorbed on meso-TiO2 (blue line) and meso-CeO2 (red line). 4.3.2.1 Determination of DMMP Conversion to Methanol The extent of DMMP decomposition on the mesoporous metal oxides was measured by monitoring the evolution of gas-phase products by quantitative 1H- NMR. The products were predominantly methanol with trace amounts of 119 Absorbance (a.u.) Absorbance (a.u.) formaldehyde. However, no phosphorus-containing products other than DMMP were detected in the volatile decomposition products, which agrees with literature results.24 The phosphorus concentration of the potential surface-bound phosphate byproduct(s) was measured using ICP-AES after the TPD experiments. We observed three scenarios for DMMP conversion to methanol (Table 4.2): 1) 1-to-1 conversion on meso-Fe2O3 suggesting that each mole of DMMP produces a mole of gaseous CH3OH, leaving behind a methyl methylphosphonate (MMP) surface group.16, 39 2) 1-to-1.6 conversion on meso-TiO2 suggesting that each mole of DMMP may produce up to two moles of gaseous CH3OH, leaving behind a methylphosphonate (MP) surface group.2, 24, 26, 30 3) 1-to-0.9 conversion on meso-Al2O3 suggesting that each mole of DMMP may produce a mole of gaseous CH3OH, which partially undergoes oxidation to CO2, CO, and H 2, 14, 2222, thereby explaining the mass balance discrepancy. Table 4.2. Phosphorus Concentration of Cycled Mesoporous Metal Oxides Meso-Al2O3 Meso-TiO2 Meso-Fe2O3 Observeda 14 6.1 5.5 mg P/g material Equivalent Amount of DMMP 0.9 1.6 1.0 aObserved values were calculated using ICP-AES to measure the P concentration at 178.3 nm. 120 The OMMs, Al2O3, Fe2O3, and TiO2, have different adsorption sites allowing for the phosphoryl’s oxygen to bind to the metal acid sites and surface hydroxyl groups through different interactions.16 Unlike the other OMMs, meso- TiO2 undergoes greater decomposition with the cleavage of a second methoxy group. This may be influenced by the rutile TiO2 structure that provides the ideal coordination to facilitate increased DMMP decomposition.223 Isotopic labelling experiments can provide more insight to this observation (further discussion in Chapter 5). 4.3.3 Evaluation of DMMP Adsorption/Decomposition Performance The nature of the DMMP binding is stronger for the OMMs in comparison to carbon. DFT-calculations for DMMP adsorption on rutile TiO2 (110) and α- Al2O3 (0001) surfaces were identified as chemisorption, which is evident by the high binding energies, 200.7 and 190.1 kJ mol-1, respectively. These values represent the formation of a strong chemical bond between the phosphoryl oxygen and an under-coordinated surface metal atom (Figure 4.14). DMMP chemisorption to the highly reactive surface of Al2O3 and TiO2, allows for 73 and 96% adsorbed DMMP decomposition, respectively. This differs from FDU-15, where 85% of the adsorbed DMMP remains fully intact on the inert carbon surface (Figure 4.15). 121 (a) (b) Ead = 200.7 kJ mol-1 Ead = 190.1 kJ mol-1 Figure 4.14. DMMP chemisorption on (a) rutile TiO2 (110) surface and (b) α- Al2O3 (0001) surface. Preliminary data by Dr. Roman Tsyshevsky. 0.10 Mass Decomposed Capacity Mass Desorbed Capacity 0.08 0.06 0.04 0.02 0.00 O 3 iO 2 2 3 3 Al 2 -T Ce O e O O2 e 2 Cu O - F / U- 15 PL - F B so oes so o - s el - -15 FD Me M Me gMe ero UA FD Figure 4.15. Graphical representation of the DMMP mass capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, and BPL. 122 DMMP Mass Capacity (g DMMP/g material) Table 4.3. 5 h Mass Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for Various Adsorbents Adsorbents Decomposed Desorbed DMMP Total Mass DMMP (g/g)a (g/g) Capacity (g/g) Meso-Al2O3 0.0528 0.0198 0.0726 Meso-TiO2 0.0270 0.0011 0.0281 Meso-CeO2 0.0325 0.0083 0.0408 Meso-Fe2O3 0.0237 0.0206 0.0443 Fe2O3 aerogel 0.0174 0.0225 0.0399 FDU-15/CuO 0.0304b 0.0349b 0.0653b FDU-15 0.0070 0.0391 0.0461 BPL 0.0067 0.0231 0.0298 aDecomposed DMMP was calculated by applying a conversion factor for the mole ratio of DMMP to methanol from ICP-AES results (Table 4.2). The decomposed DMMP for CeO2 was calculated by assuming a one-to-one conversion of DMMP to methanol. bMeasurements reported for the FDU-15/CuO composites were calculated after disregarding the CuO weight for a more accurate comparison. DMMP decomposition, reported as the decomposed DMMP capacity, is 2 to 4 times higher for the OMMs than the carbon adsorbents, FDU-15 and BPL (Table 4.3). Illustrated in Figure 4.15, the total DMMP mass capacity decreases from meso-Al2O3 (175 m2 g-1) > meso-Fe2O3 (188 m2 g-1) > meso-CeO 22 (123 m g-1) > meso-TiO2 (239 m2 g-1). It is evident that the mass capacity does not scale with the total surface area, as seen previously with carbon adsorbents in Chapter 2, and previously in the literature.71, 76, 224 This deviation from surface area 123 dependence implies a shift from a physical adsorption mechanism to a chemical one, which agrees with our theoretical studies (Figure 4.14). The desorbed (molecular) DMMP capacity for FDU-15 and FDU-15/CuO (endo and exo-pore) is higher than the mesoporous metal oxides, which we attribute to the larger surface area of the carbon-based adsorbents (600−760 m2 g- 1) in comparison to the mesoporous metal oxides (123−239 m2 g-1). After normalizing the mass capacity to the surface area of each adsorbent, this allows for a comparison between DMMP adsorption and decomposition per m2 of each material (Figure 4.16). After applying the surface area normalization, the area DMMP capacity for the mesoporous metal oxides are greater (>170%) than the carbon-based adsorbents, apart from meso-TiO2. While meso-TiO2 shows similar DMMP area capacity to FDU-15 and FDU-15/CuO, it is predominantly decomposed DMMP (Figure 4.16, Table 4.4). Our results imply that metal oxide surfaces have a higher reactivity in comparison to the carbon-based adsorbents. 124 0.6 Area Decomposed Capacity 0.5 Area Desorbed Capacity 0.4 0.3 0.2 0.1 0.0 O 3l TiO 2 2 3 3 eO e O e O uO 5 L -A 2 - 1 o so o-C 2 2 - P s o- F s el- F B e s -15 /C DU Me M e g F M Me ero DUA F Figure 4.16. Graphical representation of the DMMP area capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, and BPL. 125 DMMP Area Capacity (mg DMMP/m 2 material) Table 4.4. 5 h Area Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for Various Adsorbents Adsorbents Decomposed Desorbed DMMP T otal Area Capacity DMMP (mg/m2)a (mg/m2) (mg/m2)c Meso-Al2O3 0.3017 0.1131 0.4148 Meso-TiO2 0.1144 0.0046 0.1190 Meso-CeO2 0.2642 0.0675 0.3317 Meso-Fe2O3 0.1261 0.1096 0.2357 Fe2O3 aerogel 0.1328 0.1718 0.3046 FDU-15/CuO b b endo and exo-pore 0.0400 0.0459 0.0859 b FDU-15 0.0117 0.0652 0.0769 BPL 0.0055 0.0188 0.0243 aDecomposed DMMP was calculated by applying a conversion factor for the mole ratio of DMMP to methanol from ICP-AES results (Table 4.2). The decomposed DMMP for CeO2 was calculated by assuming a one-to-one conversion of DMMP to methanol. bMeasurements reported for the FDU-15/CuO composites were calculated after disregarding the CuO weight for a more accurate comparison. cTotal area capacity was calculated by the mass capacity (Table 4.3) divided by the total surface area reported in Table 4.1. Due to the large differences between metal oxide densities (3.95, 4.23, 7.22, 5.24, and 6.31 g cm-1, for Al2O3, TiO2, CeO2, Fe2O3, and CuO, respectively), the DMMP mass capacity was further normalized by the density and pore volume of each material to determine the DMMP volume capacity. This allows for a comparison between DMMP adsorption and decomposition per cm3 of each material. It was revealed that meso-Al2O3 achieved the greatest DMMP 126 total volume capacity in comparison to the other adsorbents. Figure 4.17 shows all OMMs (apart from meso-TiO2) having a higher DMMP volume capacity than FDU-15 impregnated with endo and exo-pore CuO. However, meso-TiO2 achieved the greatest DMMP decomposition by converting 96% of adsorbed DMMP to gaseous methanol. The lower surface area and disordered structure of aerogel Fe2O3 limits its ability to perform as well as the OMMs, yet it is still achieved greater DMMP decomposition in comparison to FDU-15 and BPL (Table 4.5). 0.16 Volume Decomposed Capacity 0.14 Volume Desorbed Capacity 0.12 0.10 0.08 0.06 0.04 0.02 0.00 l O 3 2 TiO 2 2 3 3 eO O O O 15 PL o-A - so o-C F e 2 e 2 u -- Bs e s so el -F /C U 15 FD Me M Me Me ero g U- A FD Figure 4.17. Graphical representation of the DMMP volume capacity of desorbed DMMP (solid bars) and decomposed DMMP (crisscross bars) for four different OMMs: meso-Al2O3, meso-TiO2, meso-CeO2, meso-Fe2O3, and disordered aerogel-Fe2O3, as well as carbon adsorbents: FDU-15/CuO (endo and exo-pore), FDU-15, and BPL. 127 DMMP Volume Capacity (g DMMP/cm 3 material) Table 4.5. 5 h Volume Capacity of Decomposed DMMP, Desorbed DMMP, and Total Area Capacity for Various Adsorbents Adsorbents Decomposed Desorbed DMMP T otal Volume Capacity DMMP (g/cm3)a (g/cm3) (g/cm3)c Meso-Al2O3 0.0862 0.0323 0.1185 Meso-TiO2 0.0579 0.0024 0.0603 Meso-CeO2 0.0777 0.0198 0.0975 Meso-Fe2O3 0.0483 0.0420 0.0903 Fe2O3 aerogel 0.0164 0.0212 0.0376 FDU-15/CuO endo and exo- 0.0398b 0.0458b 0.0856b pore FDU-15 0.0010 0.0513 0.0523 BPL 0.0013 0.0479 0.0492 aDecomposed DMMP was calculated by applying a conversion factor for the mole ratio of DMMP to methanol from ICP-AES results (Table 4.2). The decomposed DMMP for CeO2 was calculated by assuming a one-to-one conversion of DMMP to methanol. bMeasurements reported for the FDU-15/CuO composites were calculated after disregarding the CuO weight for a more accurate comparison. cTotal volume capacity calculated using the mass capacity (reported in Table 4.3) divided by the sum of the theoretical density and pore volume (Table 4.1). The theoretical densities used were 3.95, 4.23, 2.26, 7.22, 5.24, and 6.31 g cm-1, for Al2O3, TiO2, carbon, CeO2, Fe2O3, and CuO, respectively. 128 4.4 CONCLUSIONS In this investigation, we studied four different OMMs, Al2O3, TiO2, CeO2, and Fe2O3, with well-defined pore size, volume, and surface area. Disordered aerogel-Fe2O3 was also investigated for comparison. DRIFTS spectra of meso- TiO2 and CeO2 showed a strong interaction between the phosphoryl oxygen with a metal Lewis acid site or surface hydroxyl, characterized as chemisorption by DFT-based calculations. The decrease in absorbance of the surface OH groups further supports DMMP interaction with the hydroxyls on the surface of these metal oxides. These hydroxyl groups can promote binding and decomposition, which leads to methanol formation at low temperatures. The extent of DMMP decomposition on the mesoporous metal oxides were measured by the evolution of gas-phase methanol. Our TPD measurements for meso-Al2O3 and meso-Fe2O3 suggest that DMMP conversion to methanol is roughly one-to-one, whereas meso-TiO2 undergoes greater decomposition with the cleavage of a second methoxy group. This may be influenced by the rutile TiO2 structure that provides the ideal coordination to facilitate increased DMMP decomposition. The lower surface area and disordered structure limits the ability of aerogel Fe2O3 to adsorb high concentrations of DMMP, supported by the measured volume DMMP capacity (decomposed and desorbed), which was significantly lower than the OMMs. Meso-Al2O3 achieved the highest DMMP 129 total volume capacity in comparison to the other adsorbents. Conversely, meso- TiO2 achieved the greatest DMMP conversion to methanol by decomposing 97% of adsorbed DMMP. Additional isotope labelling experiments are necessary to reveal mechanistic details in the decomposition of DMMP. 130 Chapter 5: Future Work and Outlook This dissertation focused on understanding the adsorption properties and decomposition mechanisms of DMMP on different adsorbents to improve gas mask functionality. The effectiveness of these filters largely depends on the limited space available for physical adsorption in the pores, and the amount of reactive metal/metal oxides for chemical binding to the surface and/or decomposition. Our results revealed the significance of high surface area, large pore volume, and ordered structure for higher adsorption capacities, as well as abundant metal/metal oxide reactive sites for increased decomposition. The TPD/1H-NMR methodology, coupled with ICP-AES bulk analysis was developed as a quantitative but generic tool that may be used to evaluate adsorption and/or decomposition of other simulants and agents (listed in Table 5.1). The mesoporous metal oxides developed can be utilized by others to create catalytic materials resistant to phosphorus species deactivating the surface, which is one of the biggest limitations to ASZM-TEDA.206, 207 Its been reported that TiO2 is inert to P2O5 poisoning, 25 making it an ideal candidate for a new, more superior adsorbent; for example, mesoporous TiO2 impregnated with CuO endo-pore nanoparticles. Meso-TiO2/CuO may not replace ASZM-TEDA entirely, however, exchanging the second layer of ASZM-TEDA with highly reactive OMM/CuO would increase protection time, while using less material to make a smaller, lighter, and more portable gas mask, which is greatly desired for 131 field use (Figure 5.1). The impact of this work is not limited to gas mask filtration devices, but can also be expanded into fibers for protective clothing, and tarps for safe housing, in areas where CWAs are a persistent problem. Longer protection time Particulate filter/ Water repellant glass fiber paper Longer filter lifetime Smaller, lighter, more portable Mesh ASZM-T (large particles) Mesh Replace 2nd layer of ASZM- TEDA with meso-TiO2/CuO ASZM-T (small particles) Meso-TiO2/CuO Dust screen/ Water repellant glass fiber paper Figure 5.1. Schematic of a frustum layered canister with different sized ASZM- TEDA granules (to the left)85 and a new canister with a mixture of ASZM-TEDA and meso-TiO2/CuO (to the right). There are still unanswered questions regarding the surface chemistry during DMMP decomposition. Trotochaud et al. proposed multiple mechanistic pathways for the formation of Cu−OCH3 and O−CH3 bonds on the CuO surface, which was observed using APXPS. The participation of Olat in the formation of Cu−Olat−P and Olat−CH3 bonds is also proposed, based on DFT calculations. Further investigation with deuterium labelled DMMP (Figure 5.2) and18O- labelled metal oxides can differentiate between the bond cleavages of P−OCH3, PO−CH3, and P−CH3, and expose the role of lattice oxygens (Olat) during the 132 decomposition of DMMP. Quantitative TPD experiments can reveal which pathways are more favorable and contribute additional insight on the surface chemistry of these metal oxides, which can further be used to rationally design superior gas mask devices. O O O CD3 H3C P OCH3 D3C P OCH3 H3C P O OCH3 OCH3 OCD3 (a) (b) (c) Figure 5.2. (a) Dimethyl methylphosphonate (DMMP), (b, c) Deuterium labelled DMMP, with possible bond cleavages are indicated with a dotted line. Table 5.1. List of CWA simulants, sulfur mustard (HD/HT), sarin (GB), and VX. HD/HT GB VX 2-chloroethyl ethyl diethyl methylphosphonate diethyl methylphosphonate sulfide (CEES) (DEMP)
 (DEMP)
 2-chloroethyl phenyl diisopropyl p-nitrophenyl sulfide (CEPS) methylphosphonate diethylphosphate (PNPDEP (DIMP)
 or paraoxon) di-n-butyl sulfide diisopropyl tri-n-butyl phosphate
 phosphorofluoridate (DFP)
 methyl phenyl sulfide p-nitrophenyl O,S-diethyl (thioanisole) diethylphosphate methylphosphonothioate* (PNPDEP or paraoxon) ethyl phenyl sulfide p-nitrophenyl O,S-diethyl diphenylphosphate phenylphosphonothioate (PNPDPP) 133 Publications Huynh, K.; Holdren, S.; Hu, J.; Wang, L.; Zachariah, M. R.; Eichhorn, B. W., Dimethyl Methylphosphonate Adsorption Capacities and Desorption Energies on Ordered Mesoporous Carbons. ACS Appl. Mater. Interfaces, 2017, 9, 40638- 40644. Long, J.; Wallace, J.; Peterson, G.; Huynh, K. Manganese Oxide Nanoarchitectures as Broad-Spectrum Sorbents for Toxic Gases. ACS Appl. Mater. Interfaces, 2016, 8 (2), 1184-1193. 134 References 1. Lenhart, M. K.; Tuorinsky, S. D., Textbook of Military Medicine. US Army Medical Department Center and School, Fort Sam Houston, TX, 2008. 2. Mera, N.; Hirakawa, T.; Sano, T.; Takeuchi, K.; Seto, Y.; Negishi, N., Removal of High Concentration Dimethyl Methylphosphonate in the Gas Phase by Repeated-Batch Reactions using TiO2. J. Hazard. Mater. 2010, 177, 274-280. 3. Morita, H.; Yanagisawa, N.; Nakajima, T.; Shimizu, M.; Hirabayashi, H.; Okudera, H.; Nohara, M.; Midorikawa, Y.; Mimura, S., Sarin Poisoning in Matsumoto, Japan. Lancet 1995, 346, 290-293. 4. Suzuki, T.; Morita, H.; Ono, K.; Maekawa, K.; Nagai, R.; Yazaki, Y., Sarin Poisoning in Tokyo Subway. Lancet 1995, 345, 980-980. 5. Prentiss, A. M., Chemicals in War. A Treatise on Chemical Warfare. McGraw-Hill Book Company, Inc., Fort Sam Houston, TX, 1937; Vol. 2. 6. Khanday, W. A.; Majid, S. A.; Shekar, S. C.; Tomar, R., Synthesis and Characterization of Various Zeolites and Study of Dynamic Adsorption of Dimethyl Methylphosphate Over Them. Mater. Res. Bull. 2013, 48, 4679-4686. 7. Bartelt-Hunt, S. L.; Knappe, D. R. U.; Barlaz, M. A., A Review of Chemical Warfare Agent Simulants for the Study of Environmental Behavior. Crit. Rev. Env. Sci. Technol. 2008, 38, 112-136. 8. Li, Y. X.; Klabunde, K. J., Nanoscale Metal-Oxide Particles as Chemical Reagents - Destructive Adsorption of a Chemical-agent Simulant, Dimethyl Methylphosphonate, on Heat-treated Magnesium-Oxide. Langmuir 1991, 7, 1388- 1393. 9. Graven, W. M.; Weller, S. W.; Peters, D. L., Catalytic Conversion of an Organophosphate Vapor Over Platinum-Alumina. Ind. Eng. Chem. Process Des. Dev. 1966, 5, 183. 10. Graven, W. M.; Paton, J. D.; Weller, S. W., Sorption of Dimethyl Methylphosphonate Vapor on an Amberlyst Cation Exchange Resin. Ind. Eng. Chem. Prod. Res. Dev. 1966, 5, 34. 11. Tzou, T. Z.; Weller, S. W., Catalytic-Oxidation of Dimethyl Methylphosphonate. J. Catal. 1994, 146, 370-374. 12. Oshea, K. E.; Beightol, S.; Garcia, I.; Aguilar, M.; Kalen, D. V.; Cooper, W. J., Photocatalytic Decomposition of Organophosphonates in Irradiated TiO2 Suspensions. J. Photochem. Photobiol., A 1997, 107, 221-226. 13. Smentkowski, V. S.; Hagans, P.; Yates, J. T., Study of the Catalytic Destruction of Dimethyl Methylphosphonate-Oxidation Over MO(110). J. Phys. Chem. 1988, 92, 6351-6357. 14. Henderson, M. A.; Jin, T.; White, J. M., A TPD/AES Study of the Interaction of Dimethyl Methylphosphonate with alpha-Fe2O3 and SiO2. J. Phys. Chem. 1986, 90, 4607-4611. 135 15. Ekerdt, J. G.; Klabunde, K. J.; Shapley, J. R.; White, J. M.; Yates Jr., J. T., Surface Chemistry of Organophosphorus Compounds. J. Phys. Chem. 1988, 92, 6182-6188. 16. Mitchell, M. B.; Sheinker, V. N.; Mintz, E. A., Adsorption and Decomposition of Dimethyl Methylphosphonate on Metal Oxides. J. Phys. Chem. B 1997, 101, 11192-11203. 17. Templeton, M. K.; Weinberg, W. H., Adsorption and Decomposition of Dimethyl Methylphosphonate on an Aluminum-Oxide Surface. J. Amer. Chem. Soc. 1985, 107, 97-108. 18. Lee, K. Y.; Houalla, M.; Hercules, D. M.; Hall, W. K., Catalytic Oxidative Decomposition of Dimethyl Methylphosphonate Over Cu-Substituted Hydroxyapatite. J. Catal. 1994, 145, 223-231. 19. Henderson, M. A.; White, J. M., Adsorption and Decomposition of Dimethyl Methylphosphonate on Platinum(111). J. Am. Chem. Soc. 1988, 110, 6939-6947. 20. Guo, X.; Yoshinobu, J.; Yates, J. T., Decomposition of an Organophosphonate Compound (Dimethyl Methylphosphonate) on the Ni(111) and Pd(111) Surfaces. J. Phys. Chem. 1990, 94, 6839-6842. 21. Hegde, R. I.; White, J. M., Interaction of Dimethyl Methylphosphonate with Oxidized Iron and the Effect of Coadsorbed Water. Appl. Surf. Sci. 1987, 28, 1-10. 22. Head, A. R.; Tsyshevsky, R.; Trotochaud, L.; Yu, Y.; Kyhl, L.; Karslioglu, O.; Kuklja, M. M.; Bluhm, H., Adsorption of Dimethyl Methylphosphonate on MoO3: The Role of Oxygen Vacancies. J. Phys. Chem. C 2016, 120, 29077-29088. 23. Cao, L. X.; Suib, S. L.; Tang, X.; Satyapal, S., Thermocatalytic Decomposition of Dimethyl Methylphosphonate on Activated Carbon. J. Catal. 2001, 197, 236-243. 24. Moss, J. A.; Szczepankiewicz, S. H.; Park, E.; Hoffmann, M. R., Adsorption and Photodegradation of Dimethyl Methylphosphonate Vapor at TiO2 Surfaces. J. Phys. Chem. B 2005, 109, 19779-19785. 25. Cao, L. X.; Segal, S. R.; Suib, S. L.; Tang, X.; Satyapal, S., Thermocatalytic Oxidation of Dimethyl Methylphosphonate on Supported Metal Oxides. J. Catal. 2000, 194, 61-70. 26. Rusu, C. N.; Yates, J. T., Adsorption and Decomposition of Dimethyl Methylphosphonate on TiO2. J. Phys. Chem. B 2000, 104, 12292-12298. 27. Troya, D., Reaction Mechanism of Nerve-Agent Decomposition with Zr- Based Metal Organic Frameworks. J. Phys. Chem. C 2016, 120, 29312-29323. 28. O'Brien, C. J.; Greathouse, J. A.; Tenney, C. M., Dissociation of Sarin on a Cement Analogue Surface: Effects of Humidity and Confined Geometry. J. Phys. Chem. C 2016, 120, 28100-28109. 29. Vaiss, V. S.; Borges, I.; Leitao, A. A., Sarin Degradation Using Brucite. J. Phys. Chem. C 2011, 115, 24937-24944. 136 30. Obee, T. N.; Satyapal, S., Photocatalytic Decomposition of DMMP on Titania. J. Photochem. Photobiol. A 1998, 118, 45-51. 31. Trotochaud, L.; Tsyshevsky, R.; Holdren, S.; Fears, K.; Head, A. R.; Yu, Y.; Karshoglu, O.; Pletincx, S.; Eichhorn, B.; Owrutsky, J.; Long, J.; Zachariah, M.; Kuklja, M. M.; Bluhm, H., Spectroscopic and Computational Investigation of Room-Temperature Decomposition of a Chemical Warfare Agent Simulant on Polycrystalline Cupric Oxide. Chem. Mater. 2017, 29, 7483-7496. 32. DeCoste, J. B.; Peterson, G. W., Metal-Organic Frameworks for Air Purification of Toxic Chemicals. Chem. Rev. 2014, 114, 5695-5727. 33. Morrison, R. W., NBC Filter Performance. 2001. 34. Templeton, M. K.; Weinberg, W. H., Decomposition of Phosphonate Esters Adsorbed on Aluminum-Oxide. J. Amer. Chem. Soc. 1985, 107, 774-779. 35. Li, Y. X.; Koper, O.; Atteya, M.; Klabunde, K. J., Adsorption and Decomposition of Organophosphorus Compounds on Nanoscale Metal-Oxide Particles - in situ GC-MS Studies of Pulsed Microreactions Over Magnesium- Oxide. Chem. Mater. 1992, 4, 323-330. 36. Lin, S. T.; Klabunde, K. J., Thermally Activated Magnesium-Oxide Surface-Chemistry - Adsorption and Decomposition of Phosphorus-Compounds. Langmuir 1985, 1, 600-605. 37. Segal, S. R.; Suib, S. L.; Tang, X.; Satyapal, S., Photoassisted Decomposition of Dimethyl Methylphosphonate Over Amorphous Manganese Oxide Catalysts. Chem. Mater. 1999, 11, 1687-1695. 38. Mitchell, M. B.; Sheinker, V. N.; Tesfamichael, A. B.; Gatimu, E. N.; Nunley, M., Decomposition of Dimethyl Methylphosphonate (DMMP) on Supported Cerium and Iron Co-impregnated Oxides at Room Temperature. J. Phys. Chem. B 2003, 107, 580-586. 39. Sheinker, V. N.; Mitchell, M. B., Quantitative Study of the Decomposition of Dimethyl Methylphosphonate (DMMP) on Metal Oxides at Room Temperature and Above. Chem. Mater. 2002, 14, 1257-1268. 40. Head, A.; Tang, X.; Hicks, Z.; Wang, L.; Bleuel, H.; Holdren, S.; Trotochaud, L.; Yu, Y.; Kyhl, L.; Karsioglu, O.; Fears, K.; Owrutsky, J.; Zachariah, M.; Bowen, K.; Bluhm, H., Thermal Desorption of Dimethyl Methylphosphonate from MoO3. Catal. Struct. React. 2017, 3, 112-118. 41. Doornkamp, C.; Ponec, V., The Universal Character of the Mars and Van Krevelen Mechanism. J. Mol. Catal. Chem. 2000, 162, 19-32. 42. Grootendorst, E. J.; Verbeek, Y.; Ponec, V., The Role of the Mars and van Krevelen Mechanism in the Selective Oxidation of Nitrosobenzene and the Deoxygenation of Nitrobenzene on Oxidic Catalysts. J. Catal. 1995, 157, 706- 712. 43. Zhang, Z.; Yates Jr., J. T., Defects at Oxide Surfaces. Springer International Publishing: Switzerland, 2015. 137 44. Mars, P.; van Krevelen, D. W., Oxidation Carried Out by Means of Vanadium Oxide Catalysts. In Conference on Oxidation Processes, Amsterdam, 1954. 45. Freund, H. J.; Kuhlenbeck, H.; Libuda, J.; Rupprechter, G.; Baumer, M.; Hamann, H., Bridging the Pressure and Materials Gaps Between Catalysis and Surface Science: Clean and Modified Oxide Surfaces. Top. Catal. 2001, 15, 201- 209. 46. Oosterbeek, H., Bridging the Pressure and Material Gap in Heterogeneous Catalysis: Cobalt Fischer-Tropsch Catalysts from Surface Science to Industrial Application. Phys. Chem. Chem. Phys. 2007, 9, 3570-3576. 47. Davis, M. E., Ordered Porous Materials for Emerging Applications. Nature 2002, 417, 813-821. 48. Hierarchically Structured Porous Materials: From Nanoscience to Catalysis, Separation, Optics, Energy, and Life Science. Hierarchically Structured Porous Materials: from Nanoscience to Catalysis, Separation, Optics, Energy, and Life Science 2012, 1-651. 49. Thommes, M.; Kaneko, K.; Neimark, A. V.; Olivier, J. P.; Rodriguez- Reinoso, F.; Rouquerol, J.; Sing, K. S. W., Physisorption of Gases, with Special Reference to the Evaluation of Surface Area and Pore Size Distribution (IUPAC Technical Report). Pure Appl. Chem. 2015, 87, 1051-1069. 50. Thommes, M., Physical Adsorption Characterization of Nanoporous Materials. Chem. Ing. Tech. 2010, 82, 1059-1073. 51. Jagiello, J.; Thommes, M., Comparison of DFT Characterization Methods Based on N2, Ar, CO2, and H2 Adsorption Applied to Carbons with Various Pore Size Distributions. Carbon 2004, 42, 1227-1232. 52. Hedden, R. C.; Lee, H. J.; Soles, C. L.; Bauer, B. J., Characterization of Pore Structure in a Nanoporous Low-Dielectric-Constant Thin Film by Neutron Porosimetry and X-ray Porosimetry. Langmuir 2004, 20, 6658-6667. 53. Vincent, O.; Marguet, B.; Stroock, A. D., Imbibition Triggered by Capillary Condensation in Nanopores. Langmuir 2017, 33, 1655-1661. 54. Everett, D. H., Solid-Gas Interface. Chem. Br. 1967, 3, 341. 55. Monson, P. A., Contact Angles, Pore Condensation, and Hysteresis: Insights from a Simple Molecular Model. Langmuir 2008, 24, 12295-12302. 56. Ball, P. C.; Evans, R., Temperature-Dependence of Gas-Adsorption on a Mesoporous Solid - Capillary Critically and Hysteresis. Langmuir 1989, 5, 714- 723. 57. Ravikovitch, P. I.; Vishnyakov, A.; Russo, R.; Neimark, A. V., Unified Approach to Pore Size Characterization of Microporous Carbonaceous Materials from N2, Ar, and CO2 Adsorption Isotherms. Langmuir 2000, 16, 2311-2320. 58. Grosman, A.; Ortega, C., Capillary Condensation in Porous Materials. Hysteresis and Interaction Mechanism Without Pore Blocking/Percolation Process. Langmuir 2008, 24, 3977-3986. 138 59. Neimark, A. V.; Ravikovitch, P. I., Density Functional Theory of Adsorption Hysteresis and Nanopore Characterization. In Characterization of Porous Solids V, Unger, K. K.; Kreysa, G.; Baselt, J. P., Eds. 2000; 128, 51-60. 60. Neimark, A. V.; Ravikovitch, P. I.; Vishnyakov, A., Adsorption Hysteresis in Nanopores. Phys. Rev. E 2000, 62, R1493-R1496. 61. Valiullin, R.; Naumov, S.; Galvosas, P.; Karger, J.; Woo, H. J.; Porcheron, F.; Monson, P. A., Exploration of Molecular Dynamics During Transient Sorption of Fluids in Mesoporous Materials. Nature 2006, 443, 965-968. 62. Thommes, M.; Kohn, R.; Froba, M., Sorption and Pore Condensation Behavior of Nitrogen, Argon, and Krypton in Mesoporous: MCM-48 Silica Materials. J. Phys. Chem. B 2000, 104, 7932-7943. 63. Nicolas, R.; Berube, F.; Kim, T. W.; Thommes, M.; Kleitz, F., Tailoring Mesoporosity and Intrawall Porosity in Large Pore Silicas: Synthesis and Nitrogen Sorption Behavior. Zeolites and Related Materials: Trends, Targets and Challenges, Proceedings of the 4th International Feza Conference 2008, 174, 141-148. 64. Fries, A. A.; West, C. J., Chemical Warfare. McGraw-Hill Book Company: New York, 1921. 65. Walk, R. D. The History of Military Mask Filters. 66. Deitz, V.; Puhala, R. J.; Stroup, D. B.; Dickey, G. F., Influence of Atmospheric Weathering on the Performance of Whetlerite. 1982. 67. Romero, J. V.; Smith, J. W. H.; White, C. L.; Trussler, S.; Croll, L. M.; Dahn, J. R., A Combinatorial Approach to Screening Carbon Based Materials for Respiratory Protection. J. Hazard. Mater. 2010, 183, 677-687. 68. Smith, J. W. H.; Romero, J. V.; Dahn, T. R.; Dunphy, K.; Croll, L. M.; Dahn, J. R., The Effect of Co-Impregnated Acids on the Performance of Zn-based Broad Spectrum Respirator Carbons. J. Hazard. Mater. 2012, 235, 279-285. 69. Saxena, A.; Singh, B.; Sharma, A.; Dubey, V.; Semwal, R. P.; Suryanarayana, M. V. S.; Rao, V. K.; Sekhar, K., Adsorption of Dimethyl Methylphosphonate on Metal Impregnated Carbons Under Static Conditions. J. Hazard. Mater. 2006, 134, 104-111. 70. Fortier, H.; Westreich, P.; Selig, S.; Zelenietz, C.; Dahn, J. R., Ammonia, Cyclohexane, Nitrogen and Water Adsorption Capacities of an Activated Carbon Impregnated with Increasing Amounts of ZnCl2, and Designed to Chemisorb Gaseous NH3 from an Air Stream. J. Colloid Interface Sci. 2008, 320, 423-435. 71. Nickolov, R. N.; Mehandjiev, D. R., Comparative Study on Removal Efficiency of Impregnated Carbons for Hydrogen Cyanide Vapors in Air Depending on their Phase Composition and Porous Textures. J. Colloid Interface Sci. 2004, 273, 87-94. 72. Liang, S. H.; Harrison, B. H.; Pagotto, J. G., Determination of the Impregnant Concentrations of ASC Type Charcoal. A Magnetic Susceptibility Study. Defence Research Establishment Ottawa: 1987; Report No. 973. 139 73. Mahle, J. J.; Peterson, G. W.; Schindler, B. J.; Smith, P. B.; Rossin, J. A.; Wagner, G. W., Role of TEDA as an Activated Carbon Impregnant for the Removal of Cyanogen Chloride from Air Streams: Synergistic Effect with Cu(II). J. Phys. Chem. C 2010, 114, 20083-20090. 74. Lee, Y. W.; Kim, H. J.; Park, J. W.; Choi, B. U.; Choi, D. K., Adsorption and Reaction Behavior for the Simultaneous Adsorption of NO-NO2 and SO2 on Activated Carbon Impregnated with KOH. Carbon 2003, 41, 1881-1888. 75. Park, S. W.; Park, H. S.; Lee, W. K.; Moon, H., Effect of Water-Vapor on Adsorption of Methyl-Iodide to Triethylenediamine-Impregnated Activated Carbons. Sep. Technol. 1995, 5, 35-44. 76. Wu, L. C.; Chang, T. H.; Chung, Y. C., Removal of Hydrogen Sulfide and Sulfur Dioxide By Carbons Impregnated with Triethylenediamine. J. Air Waste Manag. Assoc. 2007, 57, 1461-1468. 77. Rossin, J.; Petersen, E.; Tevault, D.; Lamontagne, R.; Isaacson, L., Effects of Environmental Weathering on the Properties of ASC-Whetlerite. Carbon 1991, 29, 197-205. 78. Park, S. H.; McClain, S.; Tian, Z. R.; Suib, S. L.; Karwacki, C., Surface and Bulk Measurements of Metals Deposited on Activated Carbon. Chem. Mater. 1997, 9, 176-183. 79. Hammarstrom, J. L.; Sacco, A., Investigation of Hydrogen Reactivity and its Use as a Surface Probe on High Surface-Area Copper Chromium Silver Impregnated Charcoal. J. Catal. 1986, 100, 293-304. 80. Hammarstrom, J. L.; Sacco Jr., A., Investigation of Deactivation Mechanisms of ASC Whetlerite Charcoal. J. Catal. 1988, 112, 267-281. 81. Rossin, J. A.; Morrison, R. W., Spectroscopic Analysis and Performance of an Experimental Copper-Zinc Impregnated, Activated Carbon. Carbon 1991, 29, 887-892. 82. Bac, N.; Hammarstrom, J. L.; Sacco, A., Thermal-Decomposition Studies on Copper-Chromium-Silver Impregnated Activated-Charcoal. Carbon 1987, 25, 545-549. 83. Ehrburger, P.; Lahaye, J.; Dziedzinl, P.; Fangeat, R., Effect of Aging on the Behavior of Copper-Chromium Compounds Supported on Activated Carbon. Carbon 1991, 29, 297-303. 84. Doughty, D.; Groose, J. Chromium-Free Impregnated Activated Carbon for Adsorption of Toxic Gases and/or Vapors. 1991. 85. Newton, R. A. Frustrum Layered Canister. 1997. 86. Sewell, T.; Balboa, A.; Peterson, G. W.; Karwacki, C. J. Chemisorption of Arsine on Activated, Impregnated Carbon Adsorbents; US Army Edgewood Chemical and Biological Center, Aberdeen Proving Ground, MD. 87. Newton, R. A.; Jones, P.; Morrison, R. W. Layering ASZM-TEDA Carbon Beds of Varying Particle Sizes to Optimize Filter Performance; Soldier Biological Chemical Command Aberdeen Proving Ground, MD, 2003. 140 88. Ramaseshan, R.; Ying-Jun, L.; Sundarrajan, S.; Ramakrishna, S., Can the Activated Carbon That is Currently Used in NBC Protective Wear Be Replaced. Advanced Structural and Functional Materials for Protection 2008, 136, 1-21. 89. Karwacki, C. J.; Morrison, R. W., Adsorptive Retention of Volatile Vapors for Nondestructive Filter Leak Testing. Ind. Eng. Chem. Res. 1998, 37, 3470-3480. 90. Peterson, G. W.; Karwacki, C. J.; Rossin, J. A. Novel Collective Protection Filters for Emerging TIC Requirements: Axial and Radial-Flow Filter Designs; Edgewood Chemical Biological Center. U.S. Army Research, Development and Engineering Command, 2005. 92. Knapke, M. J.; Rossin, J. A. Processes for Filtering Chemicals from Air Streams. 2017. 94. Harris, R. K.; Thompson, T. V.; Norman, P. R.; Pottage, C., Phosphorus- 31 NMR Studies of Adsorption Onto Activated Carbon. Carbon 1999, 37, 1425- 1430. 95. Suzin, Y.; Buettner, L. C.; LeDuc, C. A., Behavior of Impregnated Activated Carbons heated to the Point of Oxidation. Carbon 1998, 36, 1557-1566. 96. van der Merwe, M. M.; Bandosz, T. J., A Study of Ignition of Metal Impregnated Carbons: the Influence of Oxygen Content in the Activated Carbon Matrix. J. Colloid Interface Sci. 2005, 282, 102-108. 97. Tsunoda, R., Asdorption of Organic Vapors on Active Carbons. J. Colloid Interface Sci. 1989, 130, 60-68. 98. Wood, G. O., Activtaed Carbon Adsorption Capacities for Vapors. Carbon 1992, 30, 593-599. 99. Furtado, A. M. B.; Wang, Y.; LeVan, M. D., Carbon Silica Composites for Sulfur Dioxide and Ammonia Adsorption. Microporous and Mesoporous Mater. 2013, 165, 48-54. 100. Sullivan, P.; Moate, J.; Stone, B.; Atkinson, J. D.; Hashisho, Z.; Rood, M. J., Physical and Chemical Properties of PAN-Derived Electrospun Activated Carbon Nanofibers and their Potential for Use as an Adsorbent for Toxic Industrial Chemicals. Adsorption. 2012, 18, 265-274. 101. Britt, D.; Tranchemontagne, D.; Yaghi, O. M., Metal-Organic Frameworks with High Capacity and Selectivity for Harmful Gases. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 11623-11627. 102. Petit, C.; Karwacki, C.; Peterson, G.; Bandosz, T. J., Interactions of Ammonia with the Surface of Microporous Carbon Impregnated with Transition Metal Chlorides. J. Phys. Chem. C 2007, 111, 12705-12714. 103. Kaplan, D.; Shmueli, L.; Nir, I.; Waysbort, D.; Columbus, I., Degradation of Adsorbed Sarin on Activated Carbons: A P-31-MAS-NMR Study. Clean-Soil Air Water 2007, 35, 172-177. 104. Ding, L. P.; Bhatia, S. K., Vacancy Solution Theory for Binary Adsorption Equilibria in Heterogeneous Carbon. AlChE J. 2002, 48, 1938-1956. 141 105. Keaney, A.; Alder, J. F., Mechanistic Study of the Effect of Sulphur Dioxide and Nitrogen Dioxide Exposure of ASC/T, Whetlerised BPL and Whetlerised Chlorinated BPL Activated Aarbons on the Chemisorption of Hydrogen Cyanide from Air. Adsorpt. Sci. Technol. 1998, 16, 101-115. 106. Lee, J. K.; Hudgins, R. R.; Silveston, P. L., SO2 Oxidation in a Periodically Operated Trickle Bed: Comparison of Activated Carbon Catalysts. Environ. Prog. 1996, 15, 239-244. 107. Barton, S. S.; Evans, M. J. B.; Liang, S.; McDonald, J. A. F., The Influence of Surface Modification of BPL Carbons on Aging. Carbon 1996, 34, 975-982. 108. Tsai, W. T.; Chang, C. Y., Surface Characterization and Thermodynamics of Adsorption of Methylene-Chloride on Activated Carbons. J. Environ. Sci. Health A Tox. Hazard Subst. Environ. Eng. 1995, 30, 525-535. 109. Tsai, W. T.; Chang, C. Y., Surface-Chemistry of Activated Carbons and its Relevance for Effects of Relative-Humidity on Adsorption of Chlorinated Organic Vapors. Chemosphere 1994, 29, 2507-2515. 110. Tsai, W. T.; Chang, C. Y., Adsorption of Methylene-Chloride Vapor on Activated Carbons. J. Chem. Technol. Biotechnol. 1994, 61, 145-151. 111. Eissmann, R. N.; Levan, M. D., Coadsorption of Organic-Compounds and Water-Vapor on BPL Activated Carbon 2,1,1,2-Trichloro-1,2,2-Trifluoroethane and Dichloromethane. Ind. Eng. Chem. Res. 1993, 32, 2752-2757. 112. Brea, P.; Delgado, J. A.; Agueda, V. I.; Uguina, M. A., Modeling of Breakthrough Curves of N2, CH4, CO, CO2 and a SMR Type Off-Gas Mixture on a Fixed Bed of BPL Activated Carbon. Sep. Purif. Technol. 2017, 179, 61-71. 113. Fan, W.; Chakraborty, A.; Kayal, S., Adsorption Cooling Cycles: Insights Into Carbon Dioxide Adsorption on Activated Carbons. Energy 2016, 102, 491- 501. 114. Alvarez-Gutierrez, N.; Gil, M. V.; Martinez, M.; Rubiera, F.; Pevida, C., Phenol-Formaldehyde Resin-Based Carbons for CO2 Separation at Sub- Atmospheric Pressures. Energies 2016, 9. 115. Delgado, J. A.; Agueda, V. I.; Uguina, M. A.; Sotelo, J. L.; Brea, P.; Grande, C. A., Adsorption and Diffusion of H2/CO, CH4, and CO2 in BPL Activated Carbon and 13X Zeolite: Evaluation of Performance in Pressure Swing Adsorption Hydrogen Purification by Simulation. Ind. Eng. Chem. Res. 2014, 53, 15414-15426. 116. McEwen, J.; Hayman, J. D.; Yazaydin, A. O., A Comparative Study of CO2, CH4 and N2 Adsorption in ZIF-8, Zeolite-13X and BPL Activated Carbon. Chemical Physics 2013, 412, 72-76. 117. Paez, C. A.; Contreras, M. S.; Leonard, A.; Blacher, S.; Olivera-Fuentes, C. G.; Pirard, J. P.; Job, N., Effect of CO2 Activation of Carbon Xerogels on the Adsorption of Methylene Blue. Adsorption. 2012, 18, 199-211. 142 118. Liu, Z. S.; Wang, Q.; Zou, Z. S.; Tan, G. L., Arrhenius Parameters Determination in Non-Isothermal Conditions for the Uncatalyzed Gasification of Carbon By Carbon Dioxide. Thermochim. Acta 2011, 512, 1-4. 119. Martin, C. F.; Stockel, E.; Clowes, R.; Adams, D. J.; Cooper, A. I.; Pis, J. J.; Rubiera, F.; Pevida, C., Hypercrosslinked Organic Polymer Networks as Potential Adsorbents for Pre-combustion CO2 Capture. J. Mater. Chem. 2011, 21, 5475-5483. 120. Banerjee, R.; Furukawa, H.; Britt, D.; Knobler, C.; O'Keeffe, M.; Yaghi, O. M., Control of Pore Size and Functionality in Isoreticular Zeolitic Imidazolate Frameworks and their Carbon Dioxide Selective Capture Properties. J. Amer. Chem. Soc. 2009, 131, 3875. 121. Nguyen, T. X.; Bhatia, S. K., Kinetic Restriction of Simple Gases in Porous Carbons: Transition-State Theory Study. Langmuir 2008, 24, 146-154. 122. Sullivan, P. D.; Stone, B. R.; Hashisho, Z.; Rood, M. J., Water Adsorption with Hysteresis Effect onto Microporous Activated Carbon Fabrics. Adsorption 2007, 13, 173-189. 123. Liu, J. C.; Monson, P. A., Monte Carlo Simulation Study of Water Adsorption in Activated Carbon. Ind. Eng. Chem. Res. 2006, 45, 5649-5656. 124. Sattler, M. L.; Rosenberk, R. S., Removal of Carbonyl Sulfide Using Activated Carbon Adsorption. J. Air Waste Manag. Assoc. 2006, 56, 219-224. 125. Nan, Q.; LeVan, M. D., Adsorption Equilibrium Modeling for Water on Activated Carbons. Carbon 2005, 43, 2258-2263. 126. Nguyen, T. X.; Bhatia, S. K.; Nicholson, D., Prediction of High-Pressure Adsorption Equilibrium of Supercritical Gases Using Density Functional Theory. Langmuir 2005, 21, 3187-3197. 127. Belmabkhout, Y.; De Weireld, G.; Frere, M., High-Pressure Adsorption Isotherms of N2, CH4, O2, and Ar on Different Carbonaceous Adsorbents. J. Chem. Eng. Data 2004, 49, 1379-1391. 128. Sward, B. K.; LeVan, M. D., Frequency Response Method for Measuring Mass Transfer Rates in Adsorbents via Pressure Perturbation. Adsorption 2003, 9, 37-54. 129. Sun, J.; Chen, S.; Rood, M. J.; Rostam-Abadi, M., Correlating N2 and CH4 Adsorption on Microporous Carbon Using a New Analytical Model. Energy Fuels 1998, 12, 1071-1078. 130. Tanida, K.; Sato, Y.; Masuoka, H., Measurement and Correlation of Supercritical Carbon Dioxide Adsorption on Activated Carbon. Kagaku Kogaku Ronbun. 1996, 22, 385-391. 131. Shen, C. M.; Worek, W. M., Cosorption Characteristics of Solid Adsorbents. Int. J. Heat Mass Transfer 1994, 37, 2123-2129. 132. Hassan, N. M.; Ghosh, T. K.; Hines, A. L.; Loyalka, S. K., Adsorption of Water-Vapor on BPL Activated Carbon. Carbon 1991, 29, 681-683. 143 133. Nastaj, J.; Witkiewicz, K.; Chybowska, M., Modeling of Multicomponent and Multitemperature Adsorption Equilibria of Water Vapor and Organic Compounds on Activated Carbons. Adsorpt. Sci. Technol. 2016, 34, 144-175. 134. Delgado, J. A.; Agueda, V. I.; Uguina, M. A.; Sotelo, J. L.; Garcia-Sanz, A.; Garcia, A., Separation of Ethanol-Water Liquid Mixtures By Adsorption on BPL Activated Carbon with Air Regeneration. Sep. Purif. Technol. 2015, 149, 370-380. 135. Mitchell, L. A.; Schindler, B. J.; Das, G.; dos Ramos, M. C.; McCabe, C.; Cummings, P. T.; LeVan, M. D., Prediction of n-Alkane Adsorption on Activated Carbon Using the SAFT-FMT-DFT Approach. J. Phys. Chem. C 2015, 119, 1457-1463. 136. Wang, Y.; Mahle, J. J.; Furtado, A. M. B.; Glover, T. G.; Buchanan, J. H.; Peterson, G. W.; Levan, M. D., Mass Transfer and Adsorption Equilibrium for Low Volatility Alkanes in BPL Activated Carbon. Langmuir 2013, 29, 2935- 2945. 137. Glover, T. G.; Peterson, G. W.; Schindler, B. J.; Britt, D.; Yaghi, O., MOF-74 Building Unit has a Direct Impact on Toxic Gas Adsorption. Chem. Eng. Sci. 2011, 66, 163-170. 138. Nabatilan, M. M.; Moe, W. M., Effects of Water Vapor on Activated Carbon Load Equalization of Gas Phase Toluene. Water Res. 2010, 44, 3924- 3934. 139. Nabatilan, M. M.; Harhad, A.; Wolenski, P. R.; Moe, W. M., Activated Carbon Load Equalization of Transient Concentrations of Gas-Phase Toluene: Effect of Gas Flow Rate During Pollutant Non-Loading Intervals. Chem. Eng. J. 2010, 157, 339-347. 140. Glover, T. G.; Wang, Y.; Levan, M. D., Diffusion of Condensable Vapors in Single Adsorbent Particles Measured via Concentration-Swing Frequency Response. Langmuir 2008, 24, 13406-13413. 141. Schindler, B. J.; Buettner, L. C.; Levan, M. D., Transition to Henry's law in Ultra-Low Concentration Adsorption Equilibrium for n-Pentane on BPL Activated Carbon. Carbon 2008, 46, 1285-1293. 142. Wang, Y.; Levan, M. D., Nanopore Diffusion Rates for Adsorption Determined by Pressure-Swing and Concentration-Swing Frequency Response and Comparison with Darken's Equation. Ind. Eng. Chem. Res. 2008, 47, 3121- 3128. 143. Glover, T. G.; LeVan, M. D., Sensitivity Analysis of Adsorption Bed Behavior: Examination of Pulse Inputs and Layered-bed Optimization. Chem. Eng. Sci. 2008, 63, 2086-2098. 144. Moe, W. M.; Collins, K. L.; Rhodes, J. D., Activated Carbon Load Equalization of Gas-Phase Toluene: Effect of Cycle Length and Fraction of Time in Loading. Environ. Sci. Technol. 2007, 41, 5478-5484. 144 145. Hung, H. W.; Lin, T. F., Prediction of the Adsorption Capacity for Volatile Organic Compounds onto Activated Carbons by the Dubinin- Radushkevich-Langmuir Model. J. Air Waste Manag. Assoc. 2007, 57, 497-506. 146. Walton, K. S.; Cavalcante, C. L.; Levan, M. D., Adsorption of Light Alkanes on Coconut Nanoporous Activated Carbon. Braz. J. Chem. Eng. 2006, 23, 555-561. 147. Crespo, D.; Yang, R. T., Adsorption of Organic Vapors on Single-Walled Carbon Nanotubes. Ind. Eng. Chem. Res. 2006, 45, 5524-5530. 148. Kim, D.; Cai, Z. L.; Sorial, G. A., Determination of Gas Phase Adsorption Isotherms - A Simple Constant Volume Method. Chemosphere 2006, 64, 1362- 1368. 149. Walton, K. S.; Levan, M. D., Natural Gas Storage Cycles: Influence of Nonisothermal Effects and Heavy Alkanes. Adsorption 2006, 12, 227-235. 150. Chiang, H. L.; Lin, K. H.; Chen, C. Y.; Choa, C. G.; Hwu, C. S.; Lai, N., Adsorption Characteristics of Benzene on Biosolid Adsorbent and Commercial Activated Carbons. J. Air Waste Manag. Assoc. 2006, 56, 591-600. 151. Li, X.; Guo, H. F.; Chen, H. M., Determination of the Activation Energy for Desorption by Derivative Thermogravimetric Analysis. Adsorpt. Sci. Technol. 2006, 24, 907-914. 152. He, Y. F.; Seaton, N. A., Monte Carlo Simulation and Pore-Size Distribution Analysis of the Isosteric Heat of Adsorption of Methane in Activated Carbon. Langmuir 2005, 21, 8297-8301. 153. Seaton, N. A.; Walton, J.; Quirke, N., A New Analysis Method for the Determination of the Pore-Size Distribution of Porous Carbons from Nitrogen Adsorption Measurements. Carbon 1989, 27, 853-861. 154. Himeno, S.; Okubo, K.; Fujita, S., Prediction of Non-Ideal Binary Gas Adsorption Equilibria of Volatile Organic Compounds by the Thermodynamic Non-Ideal Adsorbed Solution Model. Kagaku Kogaku Ronbun. 2006, 32, 79-87. 155. Moe, W. M.; Li, C. N., A Design Methodology for Activated Carbon Load Equalization Systems Applied to Biofilters Treating Intermittent Toluene Loading. Chem. Eng. J. 2005, 113, 175-185. 156. Qi, N.; LeVan, M. D., Coadsorption of Organic Compounds and Water Vapor on BPL Activated Carbon. 5-Methylethylketone, Methylisobutylketone, Toluene, and Modeling. Ind. Eng. Chem. Res. 2005, 44, 3733-3741. 157. Walton, K. S.; Pigorini, G.; LeVan, M. D., Simple Group Contribution Theory for Adsorption of Alkanes in Nanoporous Carbons. Chem. Eng. Sci. 2004, 59, 4425-4432. 158. He, Y. F.; Yun, J. H.; Seaton, N. A., Adsorption Equilibrium of Binary Methane/Ethane Mixtures in BPL Activated Carbon: Isotherms and Calorimetric Heats of Adsorption. Langmuir 2004, 20, 6668-6678. 159. Jorge, M.; Seaton, N. A., Predicting Adsorption of Water/Organic Mixtures Using Molecular Simulation. AlChE J. 2003, 49, 2059-2070. 145 160. Ye, X. H.; Qi, N.; Ding, Y. Q.; LeVan, M. D., Prediction of Adsorption Equilibrium Using a Modified D-R Equation: Pure Organic Compounds on BPL Carbon. Carbon 2003, 41, 681-686. 161. Ye, X. H.; Qi, N.; LeVan, M. D., Prediction of Adsorption Equilibrium Using a Modified D-R Equation: Organic-Water Vapor Mixtures on BPL Carbon. Carbon 2003, 41 (13), 2519-2525. 162. MacDonald, J. A. F.; Evans, M. J. B., Adsorption and enthalpy of phenol on BPL carbon. Carbon 2002, 40 (5), 703-707. 163. LeVan, M. D.; Pigorini, G., Group-contribution theory for coadsorption of gases and vapors on solid surfaces. Foundations of Molecular Modeling and Simulation 2001, 97 (325), 296-299. 164. Davies, G. M.; Seaton, N. A.; Vassiliadis, V. S., Calculation of pore size distributions of activated carbons from adsorption isotherms. Langmuir 1999, 15 (23), 8235-8245. 165. Tsai, W. T.; Chang, C. Y.; Ho, C. Y.; Chen, L. Y., Adsorption Properties and Breakthrough Model of 1,1-Dichloro-1-fluoroethane on Activated Carbons. J. Hazard. Mater. 1999, 69, 53-66. 166. Taqvi, S. M.; Appel, W. S.; LeVan, M. D., Coadsorption of Organic Compounds and Water Vapor on BPL Activated Carbon. 4. Methanol, Ethanol, Propanol, Butanol, and Modeling. Ind. Eng. Chem. Res. 1999, 38, 240-250. 167. Zhao, X. S.; Ma, Q.; Lu, G. Q. M., VOC Removal: Comparison of MCM- 41 with Hydrophobic Zeolites and Activated Carbon. Energy Fuels 1998, 12, 1051-1054. 168. Lehmann, C. M. B.; Rostam-Abadi, M.; Rood, M. J.; Sun, J., Reprocessing and Reuse of Waste Tire Rubber to Solve Air-Quality Related Problems. Energy Fuels 1998, 12, 1095-1099. 169. Russell, B. P.; LeVan, M. D., Coadsorption of Organic Compounds and Water Vapor on BPL Activated Carbon. 3. Ethane, Propane, and Mixing Rules. Ind. Eng. Chem. Res. 1997, 36, 2380-2389. 170. Sun, J.; Rood, M. J.; RostamAbadi, M.; Lizzio, A. A., Natural Gas Storage with Activated Carbon from a Bituminous Coal. Gas Sep. Purif. 1996, 10, 91-96. 171. Shojibara, H.; Sato, Y.; Takishima, S.; Masuoka, H., Adsorption Equilibria of Benzene on Activated Carbon in the Presence of Supercritical Carbon-Dioxide. J. Chem. Eng. Jpn. 1995, 28, 245-249. 172. Rudisill, E. N.; Hacskaylo, J. J.; Levan, M. D., Coadsorption of Hydrocarbons and Water on BPL Activated Carbon. Ind. Eng. Chem. Res. 1992, 31, 1122-1130. 173. Herrera, L. F.; Fan, C. Y.; Nguyen, V.; Do, D. D.; Horikawa, T.; Nicholson, D., A Self-Consistent Method to Determine Accessible Volume, Area and Pore Size Distribution (APSD) of BPL, Norit and AX-21 Activated Carbon. Carbon 2012, 50, 500-509. 174. Russell, B. P.; LeVan, M. D., Pore Size Distribution of BPL Activated Carbon Determined by Different Methods. Carbon, 32. 146 175. Smith, S. J.; Hern, J. A. Broad Spectrum Filter System for Filtering Contaminants from Air or Other Gases. 2002. 189. Bering, B. P.; Dubinin, M. M.; Serpinsky, V. V., Theory of Volume Filling for Vapor Adsorption. J. Colloid Interface Sci. 1966, 21, 378-393. 190. Yang, L.; Shroll, R. M.; Zhang, J. X.; Lourderaj, U.; Hase, W. L., Theoretical Investigation of Mechanisms for the Gas-Phase Unimolecular Decomposition of DMMP. J. Phys. Chem. A 2009, 113, 13762-13771. 191. Ahmadian, N.; Ganji, M. D.; Laffafchy, M., Theoretical Investigation of Nerve Agent DMMP Adsorption onto Stone-Wales Defected Single-Walled Carbon Nanotube. Mater. Chem. Phys. 2012, 135, 569-574. 192. Li, Z. Y.; Liu, Y. S.; Yang, X.; Xing, Y.; Wang, Z. Y.; Yang, Q.; Yang, R. T., Desorption Kinetics of Naphthalene and Acenaphthene over Two Activated Carbons via Thermogravimetric Analysis. Energy Fuels 2015, 29, 5303-5310. 193. Zacharia, R.; Ulbricht, H.; Hertel, T., Interlayer Cohesive Energy of Graphite from Thermal Desorption of Polyaromatic Hydrocarbons. Phys. Rev. B 2004, 69, 7. 194. Li, Z. Y.; Liu, Y. S.; Yang, X.; Xing, Y.; Tsai, C. J.; Wang, Z. Y.; Yang, Q.; Yang, R. T., Desorption of Polycyclic Aromatic Hydrocarbons on Mesoporous Sorbents: Thermogravimetric Experiments and Kinetics Study. Ind. Eng. Chem. Res. 2016, 55, 1183-1191. 195. Li, Z. Y.; Liu, Y. S.; Yang, X.; Xing, Y.; Tsai, C. J.; Meng, M. M.; Yang, R. T., Performance of Mesoporous Silicas and Carbon in Adsorptive Removal of Phenanthrene as a Typical Gaseous Polycyclic Aromatic Hydrocarbon. Microporous Mesoporous Mater. 2017, 239, 9-18. 196. Liu, Y. S.; Li, Z. Y.; Yang, X.; Xing, Y.; Tsai, C. J.; Yang, Q.; Wang, Z. Y.; Yang, R. T., Performance of Mesoporous Silicas (MCM-41 and SBA-15) and Carbon (CMK-3) in the Removal of Gas-Phase Naphthalene: Adsorption Capacity, Rate and Regenerability. RSC Adv. 2016, 6, 21193-21203. 197. Zhang, B.; Qin, X.; Li, G. R.; Gao, X. P., Enhancement of Long Stability of Sulfur Cathode by Encapsulating Sulfur into Micropores of Carbon Spheres. Energy Environ. Sci. 2010, 3, 1531-1537. 198. Johnson, S. A.; Brigham, E. S.; Ollivier, P. J.; Mallouk, T. E., Effect of Micropore Topology on the Structure and Properties of Zeolite Polymer Replicas. Chem. Mater. 1997, 9, 2448-2458. 199. Meng, Y.; Gu, D.; Zhang, F. Q.; Shi, Y. F.; Yang, H. F.; Li, Z.; Yu, C. Z.; Tu, B.; Zhao, D. Y., Ordered Mesoporous Polymers and Homologous Carbon Frameworks: Amphiphilic Surfactant Templating and Direct Transformation. Angew. Chem. Int. Ed. 2005, 44, 7053-7059. 200. Jun, S.; Joo, S. H.; Ryoo, R.; Kruk, M.; Jaroniec, M.; Liu, Z.; Ohsuna, T.; Terasaki, O., Synthesis of New, Nanoporous Carbon with Hexagonally Ordered Mesostructure. J. Amer. Chem. Soc. 2000, 122, 10712-10713. 201. Kim, T. W.; Kleitz, F.; Paul, B.; Ryoo, R., MCM-48-like Large Mesoporous Silicas with Tailored Pore Structure: Facile Synthesis Domain in a 147 Ternary Triblock Copolymer-Butanol-Water System. J. Amer. Chem. Soc. 2005, 127, 7601-7610. 202. Lastoskie, C.; Gubbins, K. E.; Quirke, N., Pore-Size Heterogeneity and the Carbon Slit Pore - A Density-Functional Theory Model. Langmuir 1993, 9, 2693- 2702. 203. Lastoskie, C.; Gubbins, K. E.; Quirke, N., Pore-Size Distribution Analysis of Microporous Carbons - A Density-Functional Theory Approach. J. Phys. Chem. 1993, 97, 4786-4796. 203. McGarvey, W. R.; Creasy, W. R. Headspace Gas Chromatography Method for Studies of Reaction and Permeation of Volatile Agents with Solid Materials; Edgewood Chemical Biological Center. U.S. Army Research, Development and Engineering Command. Aberdeen Proving Ground, MD 21010- 5424., 2005; 1-34. 204. Starink, M., The Determination of Activation Energy from Linear Heating Rate Experiments: A Comparison of the Accuracy of Isoconversion Methods. Thermochim. Acta 2003, 404, 163-176. 205. Starink, M. J., A New Method for the Derivation of Activation Energies from Experiments Performed at Constant Heating Rate. Thermochim. Acta 1996, 288, 97-104. 206. Graydon, J. W.; Thorpe, S. J.; Kirk, D. W., Interpretation of Activation- Energies Calculated from Nonisothermal Transformation of Amorphous Metals. Acta Metall. Mater. 1994, 42, 3163-3166. 206. Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T., Reporting Physisorption Data for Gas Solid Systems with Special Reference to the Determination of Surface Area and Porosity. (Recommendation 1984). Pure Appl. Chem. 1985, 57, 603-619. 207. Khanday, W. A.; Majid, S. A.; Shekar, S. C.; Tomar, R., Dynamic Adsorption of DMMP over Synthetic Zeolite-Alpha. Arabian J. Chem. 2014, 7, 115-123. 208. Kanan, S. M.; Waghe, A.; Jensen, B. L.; Tripp, C. P., Dual WO3 Based Sensors to Selectively Detect DMMP in the Presence of Alcohols. Talanta 2007, 72, 401-407. 209. Delafuente, G. F.; Huheey, J. E., The Synthesis and Characterization of Mixed (Alkylamino) and (Arylamino)phenylphosphonium Compounds. Phosphorus, Sulfur Silicon Relat. Elem. 1993, 78, 23-36. 210. Wilmsmeyer, A. R.; Uzarski, J.; Barrie, P. J.; Morris, J. R., Interactions and Binding Energies of Dimethyl Methylphosphonate and Dimethyl Chlorophosphate with Amorphous Silica. Langmuir 2012, 28, 10962-10967. 211. Dabrowski, A.; Podkoscielny, P.; Hubicki, Z.; Barczak, M., Adsorption of Phenolic Compounds by Activated Carbon - a Critical Review. Chemosphere 2005, 58, 1049-1070. 212. Fortier, H.; Zelenietz, C.; Dahn, T. R.; Westreich, P.; Stevens, D. A.; Dahn, J. R., SO2 Adsorption Capacity of K2CO3-impregnated Activated Carbon as 148 a Function of K2CO3 Content Loaded by Soaking and Incipient Wetness. Appl. Surf. Sci. 2007, 253, 3201-3207. 213. Westreich, P.; Fortier, H.; Flynn, S.; Foster, S.; Dahn, J. R., Exclusion of Salt Solutions from Activated Carbon Pores and the Relationship to Contact Angle on Graphite. J. Phys. Chem. C 2007, 111, 3680-3684. 214. Ji, X. M.; Yao, W.; Peng, J.; Ren, N.; Zhou, J.; Huang, Y. P., Evaluation of Cu-ZSM-5 Zeolites as QCM Sensor Coatings for DMMP Detection. Sens. Actuators, B 2012, 166, 50-55. 215. Peterson, G. W.; Friday, D.; Shrewsbury, M. Residual Life Indicator for Physical Adsorption Capacity of NBC Filters. Part I. Acetone Vapor Pulses and the Effect of Moisture Content on Retention Characteristics; U.S. Army Research, Development and Engineering Command: Hunter Applied Research Center, 2008. 216. Atteya, M.; Klabunde, K. J., Nanoscale Metal-Oxide Particles as Chemical Reagents - Heats of Adsorption of Heteroatom-Containing Organics on Heat- Treated Magnesium Oxide Samples of Varying Surface-Areas. Chem. Mater. 1991, 3, 182-187. 217. Kagel, R. O.; Greenler, R. G., Infrared Study of Adsorption of Methanol and Ethanol on Magnesium Oxide. J. Chem. Phys. 1968, 49, 1638. 218. Ma, S.; Zhou, J.; Kang, Y. C.; Reddic, J. E.; Chen, D. A., Dimethyl Methylphosphonate Decomposition on Cu Surfaces: Supported Cu Nanoclusters and Films on TiO2(110). Langmuir 2004, 20, 9686-9694. 219. Smiechowski, M. F.; Feaver, W. B., AC Impedance Monitoring of Activated Carbon Filter Contamination. Impedance Techniques: Diagnostics and Sensing Applications 2012, 41, 57-67. 220. Bao, J. J.; Giurgiutiu, V.; Rubel, G. O.; Peterson, G. W.; Ball, T. M. In Active Carbon Filter Health Condition Detection with Piezoelectric Wafer Active Sensors, Conference on Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2011, San Diego, CA, Mar 07-10; San Diego, CA, 2011. 221. Li, Y. X.; Schlup, J. R.; Klabunde, K. J., Fourier-Transform Infrared Photoacoustic-Spectroscopy Study of the Adsorption of Organophosphorus Compounds on Heat-Treated Magnesium-Oxide. Langmuir 1991, 7, 1394-1399. 222. Utamapanya, S.; Klabunde, K. J.; Schlup, J. R., Nanoscale Metal-Oxide Particles Clusters as Chemical Reagents - Synthesis and Properties of Ultrahigh Surface-Area Manesium Hydroxide and Magnesium-Oxide. Chem. Mater. 1991, 3, 175-181. 223. Wang, X.; Pan, D. H.; Xu, Q.; He, M.; Chen, S. W.; Yu, F.; Li, R. F., Synthesis of Ordered Mesoporous Alumina with High Thermal Stability using Aluminum Nitrate as Precursor. Mater. Lett. 2014, 135, 35-38. 224. Fan, J.; Boettcher, S. W.; Stucky, G. D., Nanoparticle Assembly of Ordered Multicomponent Mesostructured Metal Oxides via a Versatile Sol-Gel Process. Chem. Mater. 2006, 18, 6391-6396. 149 225. Brinker, C. J.; Lu, Y. F.; Sellinger, A.; Fan, H. Y., Evaporation-Induced Self-Assembly: Nanostructures Made Easy. Adv. Mater. 1999, 11, 579. 226. Long, J. W.; Logan, M. S.; Rhodes, C. P.; Carpenter, E. E.; Stroud, R. M.; Rolison, D. R., Nanocrystalline Iron Oxide Aerogels as Mesoporous Magnetic Architectures. J. Amer. Chem. Soc. 2004, 126, 16879-16889. 227. Gash, A. E.; Tillotson, T. M.; Satcher, J. H.; Hrubesh, L. W.; Simpson, R. L., New Sol-Gel Synthetic Route to Transition and Main-Group Metal Oxide Aerogels Using Inorganic Salt Precursors. J. Non-Cryst. Solids 2001, 285, 22-28. 228. Gash, A. E.; Tillotson, T. M.; Satcher, J. H.; Poco, J. F.; Hrubesh, L. W.; Simpson, R. L., Use of Epoxides in the Sol-Gel Synthesis of Porous Iron(III) Oxide Monoliths from Fe(III) Salts. Chem. Mater. 2001, 13, 999-1007. 229. Blochl, P. E., Pprojector Augmented-Wave Method. Phys. Rev. B 1994, 50, 17953-17979. 230. Kresse, G.; Hafner, J., Ab Initio Molecular-Dynamics for Liquid-Metals. Phys. Rev. B 1993, 47, 558-561. 231. Kresse, G.; Furthmuller, J., Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169-11186. 232. Badri, A.; Binet, C.; Lavalley, J. C., Use of Methanol as an IR Molecular Probe to Study the Surface of Polycrystalline Ceria. J. Chem. Soc. Faraday Trans 1997, 93, 1159-1168. 233. Lamotte, J.; Moravek, V.; Bensitel, M.; Lavalley, J. C., FT-IR Study of the Structure and Reactivity of Methoxy Species on ThO2 and CeO2. React. Kinet. Catal. Lett. 1988, 36, 113-118. 234. Aurianblajeni, B.; Boucher, M. M., Interaction of Dimethyl Methylphosphonate with Metal-Oxides. Langmuir 1989, 5, 170-174. 235. Novakova, J.; Jiru, P.; Zavadil, V., Oxidation of Methanol and Formaldehyde on Fe2O3 in Comparison with MoO3 and a Mixed Mo6+Fe3+O Catalyst. J. Catal. 1971, 21, 143. 236. Morgan, B. J.; Watson, G. W., A DFT+U Description of Oxygen Vacancies at the TiO2 Rutile (110) Surface. Surf. Sci. 2007, 601, 5034-5041. 237. Bandosz, T. J., On the Adsorption/Oxidation of Hydrogen Sulfide on Activated Carbons at Ambient Temperatures. J. Colloid Interface Sci. 2002, 246, 1-20. 150