NEW SUPRAMOLECULAR ASSEMBLIES OF TOXIC METAL COORDINATION COMPLEXES by TIMOTHY GLEN CARTER A DISSERTATION Presented to the Department of Chemistry and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy March 2010 11 University of Oregon Graduate School Confirmation of Approval and Acceptance of Dissertation prepared by: Timothy Carter Title: "New Supramolecular Assemblies of Toxic Metal Coordination Complexes" This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Chemistry by: Michael Haley, Chairperson, Chemistry Darren Johnson, Member, Chemistry Shih-Yuan Liu, Member, Chemistry James Hutchison, Member, Chemistry Eric Johnson, Outside Member, Biology and Richard Linton, Vice President for Research and Graduate Studies/Dean of the Graduate School for the University of Oregon. March 20,2010 Original approval signatures are on file with the Graduate School and the University of Oregon Libraries. in the Department of Chemistry Timothy Glen Carter An Abstract of the Dissertation of for the degree of Doctor of Philosophy to be taken March 2010 111 Title: NEW SUPRAMOLECULAR ASSEMBLIES OF TOXIC METAL COORDINAnON COMPLEXES Approved: _ Professor Darren W. Johnson Supramolecular chemistry is a relatively new and exciting field offering chemists simplistic approaches to generating complex assemblies through strategically designed ligands. Much like the many spectacular examples of supramolecular assemblies in nature, so too are chemists able to construct large, elegant assemblies with carefully designed ligands which bind preferentially to target metal ions of choice. An important concept of supramolecular chemistry, often subtle and overlooked, is secondary bonding interactions (SBIs) which in some cases, act as the glue to hold supramolecular assemblies together. This dissertation examines SBIs in a number of systems involving the pnictogen elements of arsenic and antimony as well as aromatic interactions in self-assembled monolayers. The first half of this dissertation is an introduction to the concepts of supramolecular chemistry and secondary bonding interactions and how they are used in the self-assembly process in the Darren Johnson IV laboratory. Chapter I describes how secondary bonding interactions between arsenic and aryl ring systems and antimony and aryl ring systems assist with the assembly process. Chapter II is a continuation of the discussion of SBls but focuses on the interactions between arsenic and heteroatoms. The second half of this dissertation will describe work performed in collaboration with Pacific Northwest National Laboratory (PNNL) in Richland, WA This work was performed under the guidance of Dr. R. Shane Addleman in conjunction with Professor Darren W. Johnson of the University of Oregon. This portion describes novel systems for use in heavy metal ion remediation from natural and unnatural water sources. Chapters III-V describe functionalized mesoporous silica for use in heavy metal uptake from contaminated water sources. Chapter V describes a new technology invented during this internship at PNNL which utilizes weak bonding interactions between aryl ring systems to produce regenerable green materials for toxic metal binding. This work is ongoing in the Darren Johnson lab. This dissertation includes my previously published and co-authored material. CURRICULUM VITAE NAME OF AUTHOR: Timothy Glen Carter PLACE OF BIRTH: Beloit WI DATE OF BIRTH: January 13th, 1976 GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon San Francisco State University University of Wisconsin Stevens Point DEGREES AWARDED: Doctor of Philosophy in Chemistry, 2010, University of Oregon Bachelor of Science in Chemistry, 1998, University of Wisconsin Stevens Point AREAS OF SPECIAL INTEREST: Supramolecular Self-assemblies with Toxic Metal Ions Fluorescent Labeling of Amino Acids and Small Peptides Monomer and Polymer Synthesis PROFESSIONAL EXPERIENCE: Graduate Research Assistant, Department of Chemistry, University of Oregon, Eugene Oregon, 2005-2010. Graduate Teaching Assistant, Department of Chemistry, University of Oregon, Eugene, OR, 2004-2005. Chemist, Biosearch Technologies, Inc., Novato, CA, 1999-2004. Chemist, Axys Pharmaceuticals, South San Francisco, CA, 1999. v vi GRANTS, AWARDS AND HONORS: National Science Foundation IGERT Fellowship, 2006-2009 PUBLICATIONS: Lindquist, N. R; Carter, T. G.; Cangelosi, V. M.; Zakharov, L. N.; Johnson, D. W. Accepted, 2010. Cangelosi, V. M.; Carter, T. G.; Zakharov, L. N.; Johnson, D. W. Chem. Commun. 2009, 5606-5608. Addleman, R. S.; Bayes, J. T; Carter, T G.; Fontenot, S. A; Fryxell, G. E.; Johnson, D. W. 2009 U.S. Pat. Appl. # 611120,321. . Fontenot, S.A; Carter, T.G.; Johnson, D.W.; Fryxell, G.; Addleman, R.S.; Warner, M.C.; Yantasee, W. In Trace analysis with Nanomaterials; Pierce, D.T, Zhao, lX., Eds.; Wiley-VCH: 2010. Carter, T.G.; Yantasee, W.; Sangvanich, T.; Fryxell, G.E.; Johnson, D.W.; Addleman, RS. Chem. Commun. 2008,43,5583-5585. Yantasee, W.; Warner, C. L.; Sangvanich, T; Addleman, R S.; Carter, T G.; Wiacek, R. J.; Fryxell, G. E.; Timchalk, c.; Warner, M. G. Environ. Sci. Technol. 2007,41, 5114-5119. Carter, T. G.; Vickaryous, W. l; Cangelosi, V. M.; Johnson, D. W. Comments Inorg. Chern. 2007,28,97-122.. Carter, T. G.; Healey, E. R; Pitt, M. A; Johnson, D. W. Inorg. Chern. 2005,44,9634- 9636. Carter, T; Reddington, M. 2005 U.S. Pat. Appl. #2005-51666. Lyttle, M.H., Walton, TA, Dick, DJ., Carter, T.G., Beckman J.H. and Cook, RM. Bioconjugate Chem., 2002 13(5), 1146-1154. Lyttle, M. H.; Carter, T G.; Cook, R. M. Org. Proc. Res. & Dev., 2001,5,45-49. Lyttle, M. H.; Carter, T. G.; Dick, D. l; Cook, R. M. J Org. Chem. 200065,9033- 9038. Vll ACKNOWLEDGMENTS I would like to sincerely thank my research advisor, Professor Darren W. Johnson, for continued support and thought provoking conversation throughout my graduate career. I would also like to thank my committee chair, Professor Michael M. Haley, and the other members, Professor James E. Hutchison, Professor Shih-Yuan Liu and Professor Eric A. Johnson for practical advice and valuable input toward my research. I would like to acknowledge Dr. R. Shane Addleman of Pacific Northwest National Laboratory for providing me the internship opportunity in his research group and Dr. Wassana Yantasee of Oregon Health and Science University for her advice and guidance during my internship at PNNL. Dr. Rather and Dr. Zakharov for their crystallographic expertise. Ginny Cangelosi and Zack Mensinger for providing valuable feedback and editorial assistance for this dissertation. Monica Thilges for her editorial suggestions on nearly all my published and unpublished material. My fellow 2004 classmates, Adam Marwitz, Justin Crossland, Matt Carillo and Pat Blower for their late night stress relieving sessions. Ginger Shultz for her frequent and necessary lunch meetings. I would also like to thank every person that has assisted me with my studies by providing advice or lending ideas to help my research move forward. Finally, I would like to thank the National Science Foundation (NSF) IGERT for three years of funding. This dissertation is dedicated to my father, Orland Doak Carter, who instilled the necessary skill set to succeed. V111 IX TABLE OF CONTENTS Chapter Page I. SUPRAMOLECULAR ARSENIC COORDINATION CHEMISTRy.................. 1 General Overview.................................................................................................. 1 Introduction to Supramolecular Metal-ligand Self-assembly................................ 2 Overview of Research........................................................................................... 5 Motivation of Arsenic Research............................................................................ 7 Ligand Design Strategy for Supramolecular Arsenic Complexes......................... 8 Secondary Bonding Interactions Stabilize Supramolecular Structures................. 9 Seld-assembly of Discrete Dinuclear Assemblies (M2L3 Complexes).................. 11 Syn- and Anti- AS212Ch Macrocyles 16 Controlling Diastereomeric Excess in As2L2Ch Macrocyles 21 Supramolecular Sb2L2Ch Assemblies..... 25 Conclusion............................................................................................................. 26 Bridge to Chapter II..... 27 II. MONONUCLEAR ARSENIC COMPLEXES: PROBING SECONDARY BONDING INTERACTIONS (SBIs) IN ARSINE COMPLEXES................... 28 General Overview.................................................................................................. 28 Introduction........................ 29 Results and Discussion.......................................................................................... 32 Variable Temperature NMR Studies..................................................................... 37 UV-Visible Spectroscopy of As-HI Complexes 41 Substituted Triphenylethylene Systems for SBI Quantification 47 Conclusion 52 xChapter Page Experimental.......................................................................................................... 53 Crystallographic Data...................... 54 Bridge to Chapter III 59 III. ION UPTAKE FROM NATURAL AQUEOUS MATRICES BY FUNCTIONALIZED SELF-ASSEMBLED MONOLAYERS ON MESOPOROUS SILICA (SAMMS) 60 General Overview............... 60 Introduction........................................................................................................... 61 Thiol SAMMS....................................................................................................... 64 Results and Discussion of Thiol SAMMS............................................................. 69 pH-Dependent Binding Mechanism of Thiol-Based Ligands 71 Conclusion of pH-Dependency Study................................................................... 75 Dimercaptosuccinic Acid Fe304 Nanoparticles..................................................... 76 DMSA NP Results and Discussion 76 Microwave Digestion of Thiol SAMMS and Thiol Containing Fe Nanoparticles 81 EPA Standardized Digestion Results and Discussion.. 83 Experimental.......................................................................................................... 86 Bridge to Chapter IV 88 IV. METAL ANION CAPTURE THROUGH CATIONIC METAL CENTERS...... 89 General Overview.................................................................................................. 89 Introduction........................................................................................................... 89 Results and Discussion. 92 pH Studies ofCu2+- and Fe3+-EDA SAMMS 93 Anion Uptake in Three Natural Water Types........................................................ 98 Xl Chapter Page Conclusion of Anion Binding 100 ExperimentaL......................................................................................................... 101 Bridge to Chapter V............................................................................................... 103 V. NEW FUNCTIONAL MATERIALS FOR HEAVY METAL SORPTION: "SUPRAMOLECULAR" ATTACHMENT OF THIOLS TO MESOPOROUS SILICA SUBSTRATES 104 General Overview.................................................................................................. 104 Introduction........................................................................................................... 105 Results and Discussion.......................................................................................... 108 BMMB SAMMS Characterization 108 Solution Phase Uptake Studies.................. 113 pH-Dependent Uptake Studies 118 Probing pKa Effects ofThiol Ligand 120 Saturation Studies........................................... 121 Binding Isotherm................................................................................................... 124 Lead Contaminant Leaching Studies................................ 126 Conclusion............................................................................................................. 128 ExperimentaL............... 130 APPENDIX: CRYSTALLOGRAPHIC DATA 134 BIBLIOGRAPHY 136 XlI LIST OF FIGURES Figure Page Chapter I 1. Cartoon representation of discrete supramolecular complexes 3 2. Typical transition metal coordination geometries. . 3 3. Examples of self-assembled supramolecular complexes 4 4. High yielding self-assembled tetrahedron 5 5. Crystal structure examples of As(III) 8 6. Secondary bonding interactions 10 7. Hz1 and Asz13 self-assembled complex 11 8. Crystal structure of the Asz13 assembly looking down the As-As axis................. 13 9. Interconversion between two conforrnations......................................................... 15 10.Examples of twisted assemblies in the solid state from corresponding ligands..... 16 11. ORTEP representation of anti-AszlzClz...................................................... .......... 17 12. ORTEP representation of anti-AszlzClz............................................................ .... 20 13. ORTEP representation of syn-AszlzClz................................................................. 21 14. Isomeric naphthalene-based dithiolligands 22 15. Partial ball and stick of Asz5zClz showing steric interactions............................... 23 16. Crystal structures of anti-Asz5zClz, anti-Asz6zClz and syn-Asz7zClz 24 17. ORTEP representation of Sb213 and Sb2lzClz 25 Chapter II 1. Secondary bonding interactions between E and adjacent cr* 31 2. Crystal structure of the ligand HI. 33 xiii Figure Page 3. Crystal structures of [AsbCI] and [Asl3] 34 4. Packing diagram of Asl2CI................................................................................... 35 5. Packing diagram of Asl3 36 6. Crystal structure of C3 symmetric [Asl3] 37 7. Variable temperature NMR spectrum of C3 symmetric [Asl3] 40 8. UV spectra ofHl, land [Asb] 43 9. HOMO and LUMO ofperylene bisimides............................................................ 44 10. computational calculations of the HOMO and LUMO of 1,8-naphthalimide....... 45 11. Proposed structures corresponding to observed UV-Vis data............................... 46 12. DFT model of As-O secondary bonding interactions............................................ 47 13. Crystal structure of [As223] assembly.................................................................... 48 14. CACHe model of dimethyl arsenic sitting over phenyl ring 49 15. VT-NMR of dimethylarenic mercaptostilbene complex....................................... 52 Chapter III 1. Cationic ammonium templating to form soft body... 62 2. Representation of calcined pore. 63 3. pH-dependent trend observed with GT73 70 4. pH-dependent adsorption for Thiol SAMMS........................................................ 70 5. DMSA NP uptake in HN03 spiked water 78 6. Plot of log Kd values for DMSA NP in Columbia River water 79 7. DMSA kinetics study in Pb spiked buffer 80 Chapter IV 1. Graphical depiction of EDA SAMMS 91 2. pH-dependency of Cu2+-EDA SAMMS. 93 Figure 3. pH-dependency of Fe3+-EDA SAMMS .. 4 S .. d' f 5+. peclatlon lagram 0 As .. 5. Plot of log Kd values for Cr6+ .. 6. Plot of log Kd values for As5+ .. XIV Page 94 96 99 100 Chapter V 1. Graphical representation ofBM, 1,3-BMMB and 1,4-BMMB 107 2. TEM micrograph of l,4-BMMB SAMMS 109 3. TGA of l,4-BMMB on MCM-41 110 4. EI mass spectra of evolved gasses 111 5. FT-IR ofC-H and S-H stretching 112 6. Log Kd values of Thiol-SAMMS 115 7. Comparison of 1,3- and l,4-BMMB 116 8. 1,3- and 1,4-BMMB intercalated into phenyl monolayer 116 9. pH-dependent log Kd plot of Hg2+and Pb2+ 119 10. Ligand selection with varying pKa values 120 11. pH-dependent comparison of uptake for Hg2+and Pb2+ 121 12. Saturation studies of three 1,4-BMMB/Phenylloadings 123 13. Saturation data of low-loaded phenyl SAMMS 124 14. Binding uptake isotherm 125 15. Quantification of Pb2+contamination.................................................................... 127 16. Leaching studies of active thiollayer 130 ------_.. _--- ------- xv LIST OF SCHEMES Scheme Page Chapter II 1. HI synthesis........................................................................................................... 32 2. Addition of arsenic trichloride to Hl..................................................................... 33 3. Synthesis of substituted phenyl stilbenes 50 4. Synthesis of dimethylchloro arsenite.................. 51 Chapter III 1. Mercaptopropylsilane monolayer network............................................................ 64 2. pH-dependent metal hydrolysis under neutral conditions 71 3. pH-dependent metal hydrolysis under acid conditions......................................... 72 4. Hydrolysis mechanism of an aqua ligand.............................................................. 72 5. Mechanism for the deprotonation ofthiolate ligand 73 6. pH-dependent metal-thiolate equilibrium.............................................................. 73 XVI LIST OF TABLES Tables Page Chapter III 1. Comparison of Thiol SAMMS with GT73 at near neutral pH.. 67 2. Comparison of Thiol SAMMS with GT73 and pH adjusted to 2 with HN03 68 3. pKII of common metal ions................................................................................... 74 4. Percent uptake and recovery for Thiol SAMMS 84 5. Percent uptake and recovery for DMSA NP 85 Chapter IV 1. Acid dissociation values for arsenate and chromate. 95 XVll LIST OF EQUATIONS Equations Page Chapter III 1. Distribution coefficient Kd..................................................................................... 65 Chapter IV 1. Equilibrium equation of arsenate speciation.. 96 Chapter V 1. Langmuir isotherm.. 126 1CHAPTER I SUPRAMOLECULAR ARSENIC COORDINATION CHEMISTRY Some of this work has been previously published and is reproduced with permission from: Carter, T. G.; Vickaryous, W. 1.; Cangelosi, V. M.; Johnson, D. W. Comments Inorg. Chern. 2007,28,97-122. General Overview This dissertation describes the investigation of weak molecular interactions in the generation of supramolecular assemblies for binding toxic metal ions. This work can be separated into two categories: solution phase organothiolate arsenic self-assemblies and solid supported 'regenerable' sorbent materials for heavy metal uptake. Although seemingly different with respect to the type of chemistry from a cursory inspection (organic metal-ligand versus inorganic materials chemistries), the use of weak forces to achieve efficient binding and the mechanism of binding for both categories is very similar and will be discussed in detail. The concepts of supramolecular chemistry will be introduced first, followed by my research relating to the use of organothiolate ligands to bind arsenic with a discussion of Secondary. Bonding Interactions (SBIs) and their relevance in the self-assembly process. The remainder of this dissertation will cover work relating to silica-based sorbent materials for use in toxic metal ion capture from 2native water sources, including the novel material developed during a collaboration with scientists at Pacific Northwest National Laboratory (PNNL) in Richland, WA. Chapter I surveys our approach to developing design strategies to prepare self- assembled nanoscale supramolecular complexes containing main group ions, with a particular emphasis on supramolecular arsenic(III) coordination chemistry. The majority of material the originated in a publication in Comments on Inorganic Chemistry (2007, 28,97-122, © Taylor & Francis Group, LLC). This article was coauthored with W. Jake Vickaryous and Virginia M. Cangelosi who provided content, including experimental data, results and conclusions from their research for this manuscript. Professor Darren W. Johnson, also listed as a coauthor, provided intellectual and editorial contributions to this publication. Introduction to Supramolecular Metal-ligand Self-assembly Supramolecular chemistry has been described in 1969 by Jean-Marie Lehn- Nobel Prize recipient for chemistry in 1987 for his pioneering work in the field-as the "the chemistry of the intermolecular bond" or simply put, the linking of molecules by intermolecular interactions much like atoms are linked covalently in traditional synthesis. 1 This can be demonstrated graphically by the simplified representation presented in Figure 1 where n number of difunctionalligands (double arrow lines) cooperatively interact with m number ofmetal ions (spheres) to form a discrete complex, or in some examples, complexes held together by non-covalent contacts. 3+ gives or ( ( ) ) Figure 1 Cartoon representation of bisfunctional ligands (arrows) and metal ions (spheres) undergoing self-assembly to form discrete supramolccular complexes of various shapes. The supramolecular self assembly process for metal-ligand systems is often high yielding and can be influenced by numerous factors including solvent interactions, the presence of guest molecules or by varying concentrations of either component. Traditionally, self-assembled supramolecular complexes utilize either classic d-block transition metal centers with tetrahedral, octahedral or square planar coordination geometries,f-block metal-centers with expanded coordination geometry, or in some cases, a combination of the two2,J (figure 2). a b c d Figure 2 Typical transition metal coordination geometries (a) tetraheclral, (b) octahedral and (c) square planar. Main- group elements, specifically pnictoges with an oxidation stale of +3, prefer (d) trigonal pyramidal coordination geometries. 4The result is typically a high symmetry coordination complex containing metal centers with predictable coordination geometries which utilize directing ligands to assist in the self-assembly process.4-6 By targeting a metal ion's preferred bonding geometry with a well designed rigid ligand, supramolecular chemists have generated spectacular examples of self-assemblies such as Hopfl's tin triangle? and Kieltyka's platinum square8 (Figure 3). Figure 3 Examples of self-assembled supramolecular complexes. Hopfl and coworkers coordinated three tin centers with three pyridinedicarboxylate ligands to form a triangle (left) and Kieltyka and coworkers demonstrated platinum's ability to form a square-like structure with four dipyridylligands. Supramolecular chemistry also provides an alternative approach to generating structures with high metal to ligand ratio which are often unattainable or if attainable, are doomed by a much more inefficient pathway using traditional synthetic approaches. For example, Raymond's group has synthesized tetrahedral structures by the self-assembly process starting from a simple naphthyl precursor and adding dicatechol functionality via amide linkages (Figure 4). 5SeJ:fAssembly of even more potent Fe chelator: o CIOOMe1: 2 I hOMe 76% 2: BBr3, 95% Figure 4 High yielding two step synthesis generates dicatecholligand capable of binding two metal centers. When combined with six ligands and four mctals, a self-assembled tetrahedron is formed. The high yielding, two-step synthesis combined with the self-assembly process often makes the supramolecular chemistry approach to metal chelation a greener alternative to traditional metal chelators. Supramolecular chemistry provides a powerful pathway to achieve high order, often symmetrical stlUctures through the careful design of organic ligands to target a whole host of metal ions based on their preferred coordination geometries. Ove.·view of' Resea.·ch A flurry of research activity has emerged in recent years resulting in reliable strategies for the formation of spectacular self-assembled metal-ligand clusters and capsules. Main group ions have not shared in this burst of activity. In fact, ions in this part of the periodic table have largely been overlooked for use as directing elements in self-assembly reactions, despite the need for improved chelators for main group ions for a variety of applications. Over the last two decades, there has been increasing interest in 6self-assembled nanoscale coordination complexes for use in a variety of applications including nanofabrication, molecular switches, host-guest chemistry and nanoscale chemical reactors.9-15 The incorporation of toxic metals and main group metalloid ions such as lead and arsenic, respectively, into self-assembled supramolecular complexes has received less scrutiny. Only a few examples can be found in the literature of complexes that incorporate main group elements in the self-assembly process, and of those, 16-19 only a few contain the highly toxic metalloid arsenic.2o-27 The overlying focus of this chapter is the discussion of design, synthesis and analysis of interactions between organothiol- based ligands and arsenic(III), antimony(III), bismuth and other main group metals and metalloids as a means to improve the understanding of their coordination chemistry, specifically, and main group supramolecular chemistry as a whole. Our supramolecular approach to metal chelation stems from the hypothesis that enhanced metal-ion specificity can be achieved by targeting the unusual coordination geometries of main group ions. Furthermore, the thermodynamic driving force provided by metal-ligand self-assembly reactions results in robust products. Typically, the self- assembly process of discrete supramolecular molecules leads to high and even quantitative yields as a result of this stabilization. Additionally, other weak forces such as secondary bonding interactions further amplify the thermodynamic stability of these self-assembled nanoscale complexes.28 Our approach to arsenic chelation focuses on the use of rigid, multidentate organothiolligands which target the unusual, but predictable, trigonal pyramidal coordination geometry of arsenic(III). The reversibility of As-thiolate bond formation 7allows for the self-assembly of discrete compounds to occur. We have successfully used this approach to synthesize dinuclear AS2L3 assemblies as well as a tetranuclear As4L229 assembly and a variety of As2L2Ch macrocycles and AS2LCh complexes. This chapter also reviews the diastereoselectivity in the self-assembly of these macrocycles and discusses the use of secondary bonding interactions as a means to bolster complex formation. For a recent review of the broader area of main group supramolecular chemistry see Pitt, et al.30 Motivation of Arsenic Research We have selected arsenic as the primary target for nanoscale coordination complex formation for three main reasons: 1) there are few chelators optimized for the preferred coordination geometry of arsenic, specifically, and the Group 15 ions in general 2) the coordination geometry of arsenic with thiolate ligands is predictable (trigonal pyramidal) and 3) As-S bonds are sufficiently labile to allow for self-assembly to occur.31,32 Arsenic, a semi-metal element, is best known for its toxicity toward humans, with the (+3) and (+5) oxidation states the most prevalent species found in the environment.33 Arsenic occurs naturally, and is found in ores of both common and coinage metals resulting in an environmental hazard associated with mining and metal smelting.34,35 Naturally contaminated well water has reached catastrophic proportions in Bangladesh, exposing tens of millions of people to arsenic resulting in numerous types of cancers and skin afflictions.36-38 Locally, a survey conducted of the Willamette Basin in western Oregon, USA concluded that more than 20 percent of wells tested have levels 8above the current EPA limit of lOllg/L. 39 With the ever-expanding population growth in the Willamette Basin (and the world as a whole), the likelihood of human exposure increases greatly, thus making research geared toward the study of arsenic and other toxic ions essential. Ligand Design Strategy for Supramolecular Arsenic Complexes Although the (+5) oxidation state of arsenic is the most prevalent form found in surface water,40 it is the (+3) state that is more toxic to humans as well as a more challenging target for remediation.41 Arsenic(III) species have a high affinity for thiol containing biological structures such as cysteine residues in proteins and enzymes.42,43 When coordinated with organothiolate ligands, As(III) typically prefers a trigonal- pyramidal geometry with a stereochemically active lone pair.44-46 In rare instances, arsenic(III) can adopt a distorted octahedral47 or even tetrahedral geometry, typically the result of weakly coordinating sulfur, oxygen or nitrogen atoms located in close proximity to the arsenic thiolate center (Figure 5).19,48.49 Figure 5 Stereo chemically active lone pair (left) and distorted ocwhedral (right) crystal structure examples of As(lll). 9The Johnson group has utilized this strategy to synthesize numerous examples of As(III), as well as other pnictogens-based, self-assembled supramolecular structures using rigid mono- and dithiolate ligands. Secondary Bonding Interactions Stabilize Supramolecular Structures Secondary bonding interactions (SBI's) have the potential to assist in the formation of self-assembled complexes containing organothiol-based ligands and main group elements. SBI's can occur between main group metals and aromatic systems, heteroatoms such as 0, N, S and the halogens.41 ,42,50-54 Numerous studies have been published in which supramolecular chemists are utilizing SBIs as a design criterion to aid in the self-assembly process. In doing so, supramolecular chemists are expanding the forces that drive self-assembly reactions.9,55 The most comprehensive study to date of secondary bonding interactions with arsenic describes the interactions between As(III) and either thiocarboxylic or dithiocarboxylic acid ligands. 56 Utilizing crystallographic and computational data, Tani and coworkers successfully measured close-contact distances with a number of substituted arsenic complexes and neighboring thiocarboxylato or dithiocarboxylato ligands. They then compared their findings to compounds devoid of secondary bonding interactions and discovered that often, in the solid state, ligands were twisted out ofplane to maximize close-contact interactions between the arsenic metal center and either the oxygen or sulfur of the carboxylate group. Additionally, bond elongation was observed suggesting that the interaction occurs between the nonbonding lone pairs of either the 10 ligand oxygen or sulfur atom and the 0'. orbital of an As-S bond (Figure 6). In some instances, As-S bond lengthening of as much as 0.19 A was observed, consistent with the population of an As-S 0'. orbital caused by a charge transfer from the heteroatom lone pair to the antibonding orbital of arsenic. UVNis spectroscopy provided confirmation of this charge transfer interaction by the observation of hypsochromic, or higher energy, peak migrations. 0, H ..(/-'" As~ \d- "'CI Figure 6 Secondary bonding interactions between the lone pair of E (E = 0 or S) and an adjacent cr· orbital of a metal center (M=As) resulting in bond elongation (left). Model system for computational determination of AsO SBI strength (center). Qualitative diagram depicting the approximate positions of the As-L cr* orbitals. Each cr* orbital is located diametrically opposite an As-L bond. Despite this extensive experimental work, quantification ofthe strength of the interaction between arsenic and the adjacent heteroatoms remains elusive. However, computational calculations by Tani and coworkers of stabilization energies based on phosphorous-oxygen and phosphorous-sulfur model interactions concluded that arsenic has a higher SBI stabilization energy than phosphorous. This was demonstrated experimentally by a increase in the measured SBI's between arsenic and the heteroatoms compared to phosphorus despite the larger atomic radius of arsenic. 11 Self-assembly of Discrete Dinuclear Assemblies (M2L3 Complexes) 1,4-bis(mercaptomethyl)benzene (Hz1, Figure 7) has the appropriate functionality and geometry to act as a bridging ligand between multiple arsenic ions. In the presence ofKOH in methanol and tetrahydrofuran, 1,4-bis(mercaptomethyl)benzene (Hz1) and Ase!) assemble into a dinuclear AsZ13 complex.zo Slow diffusion of pentane into a solution of AsZ13 in chloroform yields crystals suitable for X-ray diffraction. The solid state structure is shown in Figure 7. c c Figure 7 l,4-Bis(mercaptomethyl)benzene H21 (left), As213 self-assembled complex with both arsenic lone pairs pointing into the cavity of the complex due to lone pair-1{ interaction~ (center) and ORTEP diagram of the As213 crystal structure (left). The cocrystallized CHCh solvent is omitted for clarity. In this assembly, there are several close contacts between the arsenic ions and the aromatic rings. Each arsenic ion makes close contacts with two carbons of each aromatic ring-in effect each arsenic ion is involved in a 1"]z-secondary bonding interaction with each of three aromatic rings. Off-center arsenic-arene interactions such as these have previously been observed in the packing of separate discrete molecules. For example, Schmidbaur and coworkers crystallized a cyclophane adduct of arsenic trichloride with 1"]1 and 1"]z-secondary arsenic-arene interactions.zo,5? The AsZ13 assembly (Figure 7) exhibits 12 multiple low-hapticity arsenic-arene interactions in an intramolecular and multinuclear fashion. The exact nature of the arsenic-arene interaction warrants examination in the context of the direction of electron flow between the As(III) center and the arene. The n- system of an aromatic ring may act as either an electron donor or acceptor. For example, the cation-n58 and anion-n59 interactions are well-known examples of arenes acting as Lewis bases or Lewis acids, respectively. Similarly, arsenic(III) may act as either a Lewis base or a Lewis acid. Arsenic(III), particularly in arsines, is easily recognized as a Lewis donor because it is in the same group as nitrogen and phosphorus and similarly often exhibits a stereochemically active lone pair which, in some cases, may participate in coordinative bonding. Trialkyl arsine complexes are well known to coordinate metal ions through the lone pair on arsenic. Therefore, this lone pair cannot entirely be considered to be inert. As a representative example, triphenylarsines participate in dative bonding to platinum(II) in the complexes Pth(AsPh3) and Pth(AsPh3)pyr as described by Kuznik and coworkers.49 Proceeding down the Group 15 elements, from nitrogen to bismuth, the Lewis basicity decreases due to increasing localization of the lone-pair electrons in an s-orbital. Contrarily, the Lewis acidity of the elements increases on going down the group. The acceptor orbitals responsible for the Lewis acidity may be regarded as three a* orbitals oriented 1800 opposite the three full bonds of arsenic in a trigonal pyramidal coordination geometry (Figure 6).60 Arsenic occupies an intermediate position in that neither its Lewis basicity nor its Lewis acidity dominates, and either reactivity pattern may occur. 13 Two lines of evidence suggest that the arsenic-arene interaction involves electron donation from the x-system of the aromatic ring to the arsenic(lII) ion. First, the interaction is primarily observed between arsenic and electron-rich arenes.61 There is also a corresponding dearth of examples of arsenic interacting with electron-poor arenes. Second, the aromatic ring is often significantly tilted with respect to the three-fold axis of trigonal pyramidal arsenic(III) so that one of the 0-* orbitals is perpendicular to the plane of the aromatic ring. 62 This may be regarded as an orientation that maximizes orbital overlap between the arene-x system and one of the 0-* orbitals. The directionality of the arsenic-arene interaction with respect to the 0-* orbitals is exemplified in the structure of the AS213 assembly. Looking down the As-As axis, each aromatic ring is turned inward so that one side of the aromatic ring is closer to each arsenic ion (Figure 8). The shorter arsenic-arene distances are oriented nearly opposite of each As-S bond, in the expected position of the As-S 0-* orbitals. Figure 8 The crystal structure of the As213 assembly looking down the As-As axis. Superimposed on this structure are arrows representing electron flow into the estimated positions ofthe cr* orbitals. 14 The crystallization of the As213 assembly is diastereoselective. There is a chiral axis that runs through each arsenic ion: the three As-S-C bonds around each arsenic ion are bent and tilted like the blades of a propeller. Because each of the arsenic ions has its own chiral axis, the overall chirality of the assembly could in theory be ~,~; ~,A; or A,A (where ~ denotes a clockwise and A designates a counterclockwise twist along the As-As axis). In the crystalline state, only the meso-~,A diasteromer-which has a plane of symmetry perpendicular to the arsenic-arsenic axis-is observed. In solution the complex is fluxional. The IH NMR spectrum of the As213 assembly shows only one singlet in the aromatic region and one singlet in the methylene region. There are significantly fewer signals than expected for the solid state structure; all of the methylene protons in the static solid state structure are diastereotopic. A dynamic process must be interconverting the diastereomeric protons in solution: the axial chirality at each arsenic ion is rapidly switching. Given the high barrier to pyramidal inversion of arsines, the most likely mechanism of interconversion involves reversing the twist of each arsenic-sulfur-carbon bond. This interconversion process presumably involves a transition state structure where the C-S-As angle is intermediate between a clockwise and a counterclockwise twist. This process is illustrated in Figure 9. 15 • • Figure 9 Interconversion between the two conformations of the "meso" As213 complex likely proceeds via a torsional twist about the As-S-C angle. Although this dynamic process involves reversing the axial chirality at each arsenic center, there is no observed formation of the homoconfigurational fj".,fj". and A,A diastereomers. The stereochemical inversion at one arsenic ion exhibits mechanical coupling to the stereochemical inversion at the other arsenic ion. The transition state geometry for the interconversion process, therefore, may require all the sulfur and methylene carbons to be in the same plane as the arsenic-arsenic axis. Alternatively, the interconversion may proceed in two steps with each individual metal center inverting separately, albeit with the concentration of the transient intermediate too low to measure. 63 Examples from the Johnson research group do exist as chiral assemblies in the solid state and are believed to be the result of As-arene interactions.64 In each example, the proximity of the aryl group to the arsenic centers results in a stabilized twisted structure (Figure 10). However, in solution, these complexes interconvert as indicated by the sharp singlet in the IH NMR for the methylene portons. 16 SH H2 II H3 HS SH H4 Figure 10 Examples of twisted assemblies in the solid stale (top) from lhe corresponding ligands below. As2L) assemblies exils as IJ.,IJ. along lhe melal cenlers with ligand pilCh indicated by the directing of the spiraled arrow. SVIl- and Anti- Asz!2Ch Macrocyles In the absence of base, l,4-bis(mercaptomethyl)benzene (H21) and AsCh assemble into a mixture of anti- and syn-As212Ch macrocycles (Figure 11 and 13). The macrocycles exist as an equilibrium mixture of syn- and anti-diastereomers in solution, although the individual isomers can be crystallized selectively. Pentane diffusion into a CHCl3 solution of AsCh and H21 under different conditions of concentration and stoichiometry allowed selective crystallization of the individual isomers. 21 In the 17 presence of excess AsCh, the anti-macrocyc1e selectively crystallizes as an AsCh solvate. A mixture of H21 and AsCh at higher concentrations produces crystals containing exclusively the syn-diastereomer. In the crystal structures of the anti- and syn-macrocyc1es (Figures 11 and 13, respectively), the arsenic-n attraction is again evident in these dinuc1ear As(III) complexes. Arsenic-arene distances are as short as 3.165 A, which is less than the sum of the van der Waals radii of arsenic and carbon. Therefore, each arsenic ion is participating in arsenic-arene secondary bonding interactions with two aromatic rings. Figure 11 ORTEP representation of the crystal structure of the anti-As212C12 macrocycle. Cocrystallized AsCI} is omitted from the diagram. However, given the approximations inherent in van der Waals and ionic radii, definite assignment of a secondary interaction is strengthened by additional lines of evidence. That is, a short interatom distance, less than the sum ofthe van der Waals radii, does not necessarily mean the two atoms are participating in a secondary bonding interaction. The values given in tables of van der Waals radii involve several assumptions and are averaged for elements in many different compounds. The authors of 18 commonly cited tables of van der Waals radii themselves caution in the very papers so cited that their values are averages involving many approximations (see Bondi65 and Shannon66). For example, tables of van der Waals radii begin with the assumption that the atoms are spherical. Furthermore, the radius of a given atom is assumed to be invariant regardless of factors including: different substituents, different numbers of bonds, different oxidation states, different phases and different orientations. An atom is assigned the same radius regardless of these factors, even though this is inaccurate. An example famously given to illustrate the variation of van der Waals radius with orientation involves carbon tetrachloride. The chlorine atoms in carbon tetrachloride are 2.87 Aapart, significantly shorter than the sum of their van der Waals radii (3.6 A), yet Pauling observed that CC14 does not show any of the properties that would be associated with the great strain resulting from the repulsion between such close chlorine atoms.67 The van der Waals radius in directions close to the bond (C-Cl bond) is less in this case. Possibly, the strongest additional evidence for a given secondary interaction is spectroscopic observation of association in solution. The well-known infrared spectroscopic signature of hydrogen bonding is an archetypal example. In the case of the arsenic-x interaction, there is some solution NMR evidence and some limited solid-state infrared spectroscopy supporting the existence of the arsenic-x interaction. In many cases, however, it is necessary and convenient to assign secondary bonding interactions solely on atomic coordinates derived from the material in the solid state-both to understand the forces influencing crystal packing and in cases where competitive solvation makes measurement in solution difficult. 19 In addition to short contact distances between atoms, secondary bonding interactions are corroborated in the solid state by considering the orientations of reactive orbitals on the interacting species (Figure 6). The strongest arguments for secondary interactions in the solid state involve species in orientations such that their orbitals are clearly aligned to allow for an interaction. The case is considerably bolstered when it can be shown that the participating atoms are distorted out of place in order to maximize the interaction. The orientation of reactive orbitals should be appropriate and, depending on the proposed strength of the interaction, there should be structural distortion consistent with that interaction. The crystal structure of anti-As212Ch is shown in Figures 11 and 12 with an arsenic-arsenic distance of 5.02 Aand each arsenic ion makes close contact with two carbons of each aromatic ring of the ligand. Considering the nature of the arene-arsenic donor-acceptor interaction, one might expect each aromatic ring to be tilted slightly inward to make better contact with the presumed acceptor cr* orbitals on each arsenic ion. Unfortunately, considerable disorder of the positions of the arene carbons makes it difficult to draw conclusions regarding whether the interaction is influencing the lay of the aromatic rings (Figure 11). Regardless, the nearly parallel orientation of the two arene rings and their proximity to the As(III) ions suggest that the necessary orbitals are in an appropriate geometry for an ,,2-secondary interaction. 20 Figure 12 ORTEP representation of the crystal structure of anti-As212Ch macrocycle looking down the As-As axis. Cocrystallized AsCl3 is omitted from the diagram. The crystal structure of the syn-macrocycle shown in Figure 13 has an arsenic- arsenic distance considerably shorter than the anti diastereomer (4.65 A). Each arsenic ion makes close contact with two carbon atoms of each aromatic ring of the ligand. The two aromatic rings of the macrocycle are again nearly parallel. Presumably, any movement of one aromatic ring to make better contact with one a* orbital would weaken contact with another a* orbital on the other arsenic ion. Still, the closest As(III)-Cortho distances do occur opposite an As-S bond, in the expected vicinity of an acceptor As-S a* orbital. Despite the lack of dramatic structural distortions, the necessary orbitals are again in an orientation appropriate for 112-secondary bonding interactions. 21 Figure 13 ORTEP representation of the crystal structure of the syn-Aszl zClz macrocycle looking down the As-As axis. Rotational disorder of the left phenyl ring has been modeled as partial occupancy in two different orientations. The presence of multiple arsenic-arene interactions enforces unexpectedly short arsenic-arsenic distances in each macrocycle. CAChe molecular mechanics minimizations (MM2, MM3)-which do not take arsenic-arene secondary interactions into account-suggest that the arsenic-arsenic distance would be no less than ca. 6 A. The crystal structure shows the influence ofthe arsenic-arene interactions: the arsenic atoms are significantly drawn into the center of the macrocyclic cavity with an arsenic- arsenic distance of only 5.02 A in the syn-macrocycle and only 4.65 A in the anti-isomer. Controlling Diastereomeric Excess in As~kCI~Macrocyles Another aspect of our laboratory's research is the careful control of solid state structures by either varying crystallization conditions (e.g. solvent choice, rate and crystallization techniques) or through ligand modification. Similar to the 1,4- bis(mercaptomethyl)benzene system described above, three isomeric naphthalene-based ligands, 2,6-bis(mercaptomethy1)naphthalene (H22), 1,5-bis(mercaptomethyl)naphthalene (H23), and 1,4-bis(mercaptomethyl)naphthalene (H24) (Figure 14), were each reacted 22 with AsCh to form self-assembled equilibrium mixtures of syn- and anti-macrocycles. By using these ligands which differ only by the ring positions of the mercaptomethyl groups, we were able to access different ratios of the two macrocyclic isomers including mostly syn, mostly anti, and an almost statistical mixture of the two.26 Figure 14 Isomeric naphthalene-based dithiolligands. Within these macrocycles, the arsenic-n interaction causes the arsenic atom and its coordination sphere to be pulled toward the ligand backbone. Some steric congestion is present around the chlorine and sulfur atoms (Figure 15) and the diastereomer with the least amount of unfavorable steric strain forms in excess (giving rise to a diastereomeric excess, (de)). 23 Figure 15 Partial ball and stick models of As252Cl2 showing the steric repulsions between chlorine and sulfur atoms when the chlorine atom is pointing away from (a) and toward (b) the hydrocarbon backbone ofthe ligand. The de of these macrocycles in solution has been measured using IH NMR spectroscopy and found to be only 9% for AS252Ch (it is not known which isomer is in excess). Neither the crystal structure nor the computer models indicate any steric strain in either isomer, so the small de is not surprising. Crystals were obtained of anti- As252Ch (Figure 15a) and the structure reveals a clear arsenic-1t interaction within a cavity that is too small to accommodate any guests screened to date (metal cations, H+ or small organic molecules). Anti-As262Ch exists in 85% de in solution as determined by IH NMR spectroscopy. The observed excess seen for this macrocycle can be attributed to the large difference in steric interactions between the chlorine atom and the naphthalene backbone ofthe ligand when the Cl is pointing into the cavity (toward C-H) or pointing out (toward H). Again, crystals were obtained and found to be 100% anti (Figure 15h). (a) (b) (c) 24 Figure 16 Representations of single-crystal X-ray structures of ASzLzClz macrocycles derived from napthalene-based ligands. Space filling and wireframe representations of (a) anti-Asz5zClz, (b) anti-Asz6zCIz (c) syn-Asz7zClz macrocycles. In the case of AS272Ch macrocycles a 90% de of the syn macrocycles exists in solution. In the solid state, however, the syn isomer crystallizes exclusively (Figure 1Sc). Diastereocontrol is important in our design scheme to utilize macrocycles as synthons for larger assemblies. For instance, syn-macrocycles are preorganized to form larger discrete assemblies allowing us to study host-guest chemistry, while anti- macrocycles can be functionalized to provide extended structures. We are currently pursuing this goal, as well as trying to design ligands that will form arsenic-containing macrocycles with improved diastereocontrol.26 25 Supramolecular Sb2~Ch Assemblies Antimony is also able to participate in self-assembly with dithiolligands. In the absence of base, I,4-bis(mercaptomethyl)benzene (H21) and antimony(III) form a dinuclear Sb2bCh complex (Figure 11). This macrocycle also exhibits pnictogen-arene interactions: each antimony(IIl) ion participates in an '112-secondary interaction with one aromatic ring of the macrocycle and an '113-secondary interaction with the opposite ring.68 The shortest Sb-C interactions are again opposite an antimony-sulfur bond consistent with the view that the n-cloud of the arene ring is donating into the Sb-S a* orbital (Figure 16b). For example, the Sb-Cortho distance opposite the Sb-S bond is 3.50A versus 3.6IA for the other ortho-carbon atom. Since the antimony-chlorine bonds point away from, and bisect the two As-S bonds, regardless of the way the two arene rings are canted, there will always be two ortho-carbon atoms positioned closer to the antimony ions than the other two. (a) (b) Figure 17 ORTEP representation ofthe single crystal X-ray structure of (a) Sb213 and (b) Sb212C12. 26 Similarly to the case for arsenic, the deprotonated thiolate ligands assemble with Sbe!) to form an Sb213 complex. Interestingly, this assembly is chiral and posesses a helical twist, crystallizing as a racemic mixture of enantiomers (Figure 17a). The Sb-Sb distance in this complex is only 4.30 A, due to stronger antimony-7t interactions. To achieve this proximity between antimony atoms, the ligand must adopt a helical twist to "compress" the complex. As in the case for As213, a dynamic torsional rotation inverts the helical and meso conformations in solution. The crystal structure of Sb213 readily demonstrates the presence of many intramolecular Sb-arene secondary interactions. Each Sb(IlI) ion makes close contacts with two carbons of each of the three aromatic rings. The orientations of these close contacts are consistent with the expected positions of available reactive orbitals. Furthermore, the assembly bears further structural evidence consistent with electronic donation from the 7t-system into the Sb-S cr* orbital. Each aromatic ring is twisted inward in a manner that maximizes contact between the ring carbons and the expected positions of the Sb-S cr* orbitals Conclusion Our work continues to focus on the design and synthesis of organothiolligands for use in supramolecular coordination chemistry of arsenic and other main group elements. The lack of efficient chelators for toxic ions such as arsenic drives this work forward and the use of supramolecular chemistry provides for a different approach to arsenic sequestering. Metal binding specificity and interesting emergent spectral 27 properties comprise two encouraging results we have observed thus far in this line of research. We are optimistic that rich host-guest chemistry and new structures types are soon to follow on the basic science end of this research. Applications in sensing stemming from a supramolecular approach to metal chelation will be investigated, and we are currently exploring the use of these ligands to provide nanostructured materials for use as sorbents for water purification.69,7o Bridge to Chapter II Chapter II extends the discussion of secondary bonding interactions to include two mononuclear arsenic complexes and is based in part on published and unpublished results. In the first half of Chapter II, SBIs between arsenic and the ~-mercapto imido oxygen of a naphthalimide ligand are investigated by NMR, UV-Vis and X-ray chrystallography. In the second half of Chapter II, I-mercapto-2-phenyl stilbene was reacted with dimethylarsenic iodide and investigated by variable temperature NMR and computational analysis. Chapter III contains unpublished results pertaining to toxic metal ion uptake from natural water sources. Chapter IV contains unpublished results and focus on the use of cationic sorbent material to bind oxyanions from natural water sources. Chapter V contains published and unpublished material and describes a recent collaboration between myself and scientists at Pacific Northwest National Laboratory. 28 CHAPTER II MONONUCLEAR ARSENIC COMPLEXES: PROBING SECONDARY BONDING INTERACTIONS (SBIS) IN ARSINE COMPLEXES Some of this work has been previously published and is reproduced with permission from: Carter, T.G.; Healey, E.R; Pitt, M.A.; Johnson, D.W. Inorg. Chern. 2005,44, 9634-9636 General Overview This chapter discusses the concept of secondary bonding interactions (SBls) at both a theoretical and experimental level. A number of systems were explored in an attempt to further elucidate the nature of SBls with respect to both arene and heteroatom arsenic containing molecules. The diimide work outlined early in this chapter stems partially from the publication 'Secondary Bonding Interactions Observed in Two Arsenic Thiolate Complexes' in Inorganic Chemistry (2005, 44 (26),9634-9636, © American Chemical Society).! This article was co-authored with Elisabeth Rather Healey who performed X-ray crystallographic analysis, Melanie A. Pitt who performed experimental and editorial 29 assistance and Professor Darren W. Johnson who provided intellectual input and editorial assistance. Additionally, computational studies will be included along with the published results as well as a brief literature review on the topic of SBIs, unpublished spectroscopic studies and an unpublished structure with three-fold symmetry imparted by secondary bonding interactions between arsenic and intramolecular imido oxygens. The remaining content of this chapter covers the design and synthesis of arene-containing molecules for use in both UV-Vis and NMR spectroscopic studies and includes conclusions based off of experimental and computational studies. A more detailed description of SBIs can be found in chapter I of this dissertation. Introduction Supramolecular approaches to arsenic(ill) chelation represent a relatively emerging field,2-7 due in part to the peculiar coordination geometry of this highly toxic ion. Our research group has produced many spectacular examples of self-assembled supramolecular complexes by concomitant utilization of computational modeling and synthetic approaches. The unusual trigonal-pyramidal coordination geometry of As(ill) features a stereochemically active lone pair when coordinated by sulfur-based ligandsA and is predictable enough to be exploited as a target for specific ligand design.8-10 The characteristic coordination of As(ill) by sulfur-containing biological molecules such as glutathione or cysteine has recently been reported in the context of developing a better understanding of arsenic toxicity.9,11 However, there are relatively few known A See Chapter I Figure 6 for examples of the preferred binding geometry for arsenic(III)-thiolate and stereo chemically active lone pair. 30 structures of arsenic thiolate complexes: a search of the Cambridge Structure Database (CSD)12 reveals only 83 examples of an As(III) ion coordinated by one or more organothiolate ligands.B The use ofthiolate ligands optimized for the specific coordination geometry of As(III) that also possess additional functional groups capable of exhibiting secondary bonding interactions is relevant toward designing specific chelators and sensors for this toxic main group element. We incorporate weak attractive forces known as secondary bonding interactions (SBIs) into our design strategy for forming self-assembled complexes containing organothiolate-based ligands and main group elements. SBIs can occur between main group metals and an aromatic system, heteroatoms such as 0, N, S and the halogens. 13,14 SBIs can be predictable, occurring within the sum ofthe van der Waals radii of two or more atoms,15,C and can offer a potentially powerful method for ligand design to optimize chelation of main-group metals. 16 More and more examples are appearing in the literature where supramolecular chemists are utilizing SBIs as design criteria to aid in the self- assembly process. In doing so, supramolecular chemists are expanding beyond traditional self-assembly forces such as metal coordination or hydrogen bonding. 17,18 As mentioned in Chapter 1 of this dissertation, the most comprehensive study of secondary bonding interactions for arsenic is limited to As(III) and either thiocarboxylic or dithiocarboxylic acid derivatives. 19 The authors utilized both crystallographic and computational data to successfully measure close contact distances with a number of B A CSD (Conquest Version 1.10,2008 Release) search ofcompounds that possessed at least one As-S bond (with S also bonded to a C) revealed 83 structures of which only 63 contained an As ion not also bonded to a carbon or transition metal atom. C The mean van der Waals radii as calculated by Bondi are: As-1.85 A, 0-1.52 A, N-1.52 Aand S-1.80 A. 31 substituted arsenic complexes and neighboring thiocarboxylato or dithiocarboxylato ligands Figure 1. a b 0, HR~-'" As~ \d- "'CI Figure 1 Secondary bonding interactions (dashed line and overlapping white filled orbital depictions) between the lone pair of electrons on E (oxygen or sulfur) and an adjacent a* orbital of a metal center (M = As) resulting in bond elongation (solid line and overlapping black and white filled orbital depictions) (a) and model system for computational determination ofAs··O SBI strength (b). As-S bond elongation was also observed suggesting that the interaction involves the participation of the electron density of the nonbonding lone pairs of either oxygen or sulfur with the cr· orbitals of adjacent As-S bonded ligands (Figure la). UV-Vis spectroscopy provided confirmation of charge transfer interaction by the observation of hypsochromic, or higher energy, peak migrations. Tani and coworkers have suggested that the strength of SBls between heteroatoms and main group elements are similar to that of a weak hydrogen bond?O Computational studies of the pnictogens forming vinylic five-member ring systems with SBls to oxygen have an assigned value around 2 kcal/mol for arsenic, depending on the substituents off the arsenic center (Figure Ib).21 Despite the extensive experimental work performed by the authors, they were unsuccessful in quantifying the strength of the interaction between arsenic and the adjacent heteroatoms. The use of SBls13,14 between As(III) and heteroatoms of 32 appropriate ligands offers an additional and complementary tool for designing ligands specific for this ion. Results and Discussion This section of Chapter II reports the synthesis and crystal structure of a new first- generation sulfur-based ligand N-(2-mercaptoethy1)-1,8-naphthalimide, Hl-capable of binding As(III) via a thiolate group with a complementary SBI between the imido oxygen and arsenic. The naphthalimide core was chosen as a model for future supramolecular chelators due to the proximity of its imido oxygens atoms to the thiol and for its well known electronic absorption and emission properties. The complexes [AsI2Cl] and [Ash] form by treatment of HI with AsCh. Thiol ligand HI was synthesized (Scheme 1) in ca. 85% yield22 and formed light brown single crystals suitable for diffraction analysis from diffusion of acetonitrile into a CHCh solution of HI. °eo o °I ":: ":: .0 .0 + Triethylamine .. Pyridine 135°C SH ~aeoN° I ":: ~ .0 .0 H1 Scheme I HI synthesis by combining cysteamine HCI with triethylamine in pyridine and stirred for 20 minutes under a nitrogen atmosphere. I ,8-naphthalic anhydride was added and the solution heated to 135°C overnight. 33 The single crystal X-ray structure of HI consisted of two sets of layers of ligands interacting via aromatic face-to-face 11:-11: stacking with an intermolecular distance between adjacent layers of 3.45 A (Figure 2, also see Appendix A for crystallographic values). a b Figure 2 Crystal structure of the ligand HI. (a) ORTEP representation with 50% thermal ellipsoids and (b) view of thc crystal packing down [10 I] showing the non-parallcl arrangcmcnt bctween tile planes of aromatic stacking formcd by HI. The planes formed by these two sets of layers are twisted at an angle of ca. 20°. This non-parallel arrangement is sustained by weak interactions between aromatic CH and oxygen atoms of the imide with d(C··O) =3.21 A. FUlthermore, in the crystalline state the mercapto ethylene moiety adopts the expected anti conformation with a dihedral angle between the sulfur atom and the nitrogen atom of the imide of 177°. H1 1/3 AsCI3 .. 1mol KOH + + Scheme 2 Addition of arsenic trichloride to 1-11 in a 1:3 ratio in the presence of base formed Asl mCl n where m=I-3 and n=0-2 as verified by NMR spectroscopy. 34 Treatment of HI with AsCl3 and a stoichiometric amount of KOH yields a mixture of [AslChJ, [Asl2ClJ and [Asl3J complexes according to IH NMR spectroscopy (Scheme 2). Pale yellow crystals suitable for X-ray analysis were obtained for both the [AshCJ] and [Asl)] complexes (Figure 3, see also Appendix A). Figure 3 X-ray crystal slructures or the two complexes [Ast 2el] (top) and [Asl,] (bollom) showing the gal/che conformation adopted by the mercapo-erhylene moieties allowing for the secondary As···O bonding interactions (dashed lines) within the complexes. The structure of [AsI2Cl]·CHCh reveals that the complex is stabilized by As···O secondary bonding interactions (d(As"'O) =2.91 A) resulting from one ligand adopting a higher energy gauche conformation about the mercapto-ethylene moiety with a dihedral angle between the sulfur and nitrogen atoms of ca. 51 0 (Fig I'e 4). This secondary bonding interaction is within the range of previously observed As···O secondary interactions,D.'9 and it suggests that in the crystalline state crystal packing forces and/or the As···O secondary interaction is at least strong enough to compensate for one unfavorable gauche interaction. The other ligand exists in the expected anti conformation with a corresponding dihedral D A CSD search revealed that the few As"'O secondary bonding interactions previously observed are in the range of 2.71-2.94A. For a representative example see K. Tani et al. 35 angle of ca. 170°. Presumably, the rate of crystallization of the complex prevents additional As···Q interactions from forming. The crystal packing of [AsbCllCHCh is sustained by aromatic stacking similar to that observed for HI (Figure 2b), with distances between the centroids of naphthalimide moieties of adjacent arsenic complexes ranging from 3.40 to 3.65 A. Figure 4 Packing diagram of As1 2C1 in the crystalline state generated using the program Mercury. All non-hydrogen atoms are shown in wireframe with arsenic purple, oxygen red, sulfur yellow, chlorine green, nitrogen light blue and carbon gray. Measured bond lengths of 2.25 A for the As-CI opposite of the weakly bound oxygen in AsbCI (Figure 3 top) are within calculated values of 2.2 A. 15 However, a search of the CSD for structures containing AsClx with observed secondary bonding interactions produced a wider range of bond elongations. The measured value of AsbCI falls between reported lengths for As-CI bonds that do not experience weak SBIs (2.19 A) and for As-CI bonds opposite of a weakly bound heteroatom (2.38 A). 36 The crystal structure of [Asl:Jl reveals that two of the ligands adopt the expected anti conformation ((S-C-C-N) = 180° and 170°), while the third thiolate again adopts a gauche conformation to allow for one weak As· ..Q interaction (d =3.21 A) with a corresponding dihedral angle of ca. 60° (Figure 3). Crystal packing is similar to that of H (Figure 2b) and [As12Cl] (Figure 4) with the two sets of planes sustained by aromatic stacking (d(centroid-centroid) =3.50 to 3.56 A) and twisted at an angle of ca. 37° (Figure 5). Figure 5 Packing diagram of AsI, in lhc crystalline stale generaled using the program Mercury. All non-hydrogen aloms are shown in wireframe wilh arsenic purple, oxygen red, sulfur yellow, chlorine green, nilrogen Iighl blue and carbon gray. A different polymorph of the original [As13] complex was obtained by slow diffusion of AsCh into a solution containing deprotonated ligand 1 (Figure 6). 37 Figure 6 Crystal structure of C, symmetric arsenic complex [Asl,]. Representative secondary bonding interactions are denoted by dashcd line. As"'O distance of 3.16 Afor the C, symmetric polymorph versus an As"'O distance of 3.21 Afor the non-symmetric structure. The resulting complex was found to be C3-symmetric down the arsenic center with one imido oxygen from each ligand involved in an SBl with an As"'O distance of 3.16 A which is less than that of [As12el] and the non C3 symmetric [As1)] complex of 3.21 A. As mentioned above, the alkyl linkage of each ligand is forced into the higher energy gauche conformation, accounting for a 2-3 kcallmol increase in energy as well as nonparallel n-stacking crystal planes. A slight bond elongation was observed in the C) symmetric complex with As-S bonds measuring 2.25 A versus 2.22 A for the non-SBl As-S bond in the unsymmetrical [As13]. Variable Temperature NMR Studies Based on the new C3-symmetric complex and the potential for such a structure to be locked in conformation in solution due to the SBls, attempts to quantify the As"'O secondary bonding interactions were performed by solution-phase variable temperature 38 NMR (VT-NMR) studies. Crystals, of the new C3-symmetric polymorph were grown as described in the experimental section. To assure the proper polymorph of the crystalline material, the space group was verified by X-ray analysis and compared to the known C3 symmetric structure for preliminary studies. The crystals were then dissolved in CDCb with no further purification, yielding promising initial room temperature NMR data. The appearance of a complex spectrum consisting of splitting patterns consistent with a chiral C3-symetric complex containing diastereotopic protons was observed both in the aromatic and aliphatic regions. It was rationalized that if the complex was indeed locked in conformation at room temperature in solution on the NMR time scale, then heating the sample by VT-NMR should impart enough energy to cause a departure from the locked conformation. A freely rotating structure should cause the diastereotopic protons to become equivalent and the split peaks, particularly Hd and Hd', to coalesce into an averaged triplet from the coupling with He,e'. Crystals were once again obtained as described above and the unit cell verified to match the known C3 symmetric complex. The crystals were washed with chloroform, methanol, 4: 1 methanol/water mixture and again with methanol before being dried under high vacuum. The As13 complex is only sparingly soluble in chloroform and methanol, therefore, the above washes assure pure product by removing unreacted starting material and KCI byproduct. The washed and dried crystals were dissolved in deuterated 1,1 ,2,2-tetrachloroethane and heated to 100°C from room temperature by 10 °C intervals with a fifteen-minute equilibration time between runs. The result of this 39 experiment did not provide any peak migration; mainly due to the absence of the complex splitting patterns with this solvent system. It would appear that the strength of the As"'O SBI is solvent dependent due to the observation of splitting in CDCh but not in C2D2C14. An alternative approach involved cooling the sample and checking for the appearance of a new peak as a result of limiting the available energy for rotation of the ligands. The C2D2Cl4 sample was chilled to -60 °C from room temperature again at 10°C intervals and a fifteen-minute equilibration time. The NMR spectra of the prepared crystals in tetrachloroethane were devoid of any secondary structure observed in the preliminary experiment using chloroform as the solvent. In an attempt to recreate the preliminary NMR studies, crystals from the above purification were dissolved in CDCh and the sample was slowly cooled to -40°C from room temperature as described above. Initially, a well-defined spectrum was obtained at room temperature with no indication of complex splitting. However, at temperatures as low as 0 °C, peaks observed in the preliminary study appeared to grow from the baseline. As the temperature was decreased, this time at 5 °C intervals, the peaks appeared to grow in intensity, suggesting that the structure was indeed becoming locked in conformation in solution as the temperature decreased Figure 7. 40 1:1\5He~ Hd ..H~Hd' H 4.0000 .-+-----------------:~.-.: ......'-----------___j o -J 2.0000 -1---------------------------1 1.0000 -1------- • Ag Cd ....... Hg _Fb _11 0.0000---·------,---- -1.2 -0.2 0.8 1.8 2.8 pH 3.8 4.8 5.8 6.8 Figure 4 Datil illustrating pH dependent 11.1 13.36 lard/Soft H S S(B) S(B) B S B S S Tabie 3 pK II of common metal ions along wIth correspondlllg hardness/softness designations. Values obtained from Lange's Handbook of Chemistry. Blue arrows indicate M 2+ soft ions which follow the expected pH trend. Red arrows indicate M 1+ soft iOlls which deviate from the expected pI-! trend. I-! =hard, S =soft and 8 =both. The observed experimental pH-dependent trend is apparent for the M2+ ions of mercury, lead and cadmium (blue arrows in Table 3) as well as several other ions but deviates greatly for the M 1+ ions of silver and thallium (red arrows in Table 3). Both Ag 1+ and TI 1+ have large pK, values yet are unaffected by pH, however silver binds approximately 3 orders of magnitude greater than thallium in our original study. This difference in the binding capacity of Thiol SAMMS is not in agreement with other thiol- based sorbents.27 Unfortunately, only a few publications are available for thallium-thiol systems and Jacobson et al demonstrated near equal binding capacity for both Ag 1+ and Tl 1+ ions.27 Both metals ions are considered to be 'soft' by Pearson's standards28 and 75 both have relatively large ionic radii, with thallium having one of the largest of the elements. A possible explanation for the high binding affinity for Ag1+to Thiolligands is due to its relatively slow water exchange rate kex of 106 S-I versus Pb2+at 108 S-I and Hg2+ at 109 s-I.23 Because there is a direct relationship between the rate at which water exchanges and complex formation with other ligands, the slow exchange rate of silver often leads to the formation of extremely stable complexes.23 The formation of stable complexes plus the high affinity silver(l) has for thiolligands is in good agreement with the observed trend and accounts for the high binding affinity with thiol SAMMS. Conclusion of pH-Dependency Study It appears that the observed pH-dependent capture of metal ions can be primarily attributed to a metal ion's capacity to undergo hydrolysis and, secondarily, to an ion's water exchange rate or ease of forming a new, stable complex. Another factor of pH- dependent uptake may indeed be the difference in pKa between aryl thiols (GT73) and alkyl thiols (Thiol SAMMS). This would not explain why some metal ions are greatly affected by pH while others are not, but would explain why GT73 experienced very little change in Kd values under acidic conditions while Thiol SAMMS did. The lower pKa of the aryl thiol may facilitate binding to metal ions even in high [H+] environments. It is the current belief that the combination of factors described above dictates the effectiveness of thiol-based sorbent materials toward soft metal cations. 76 Dimercaptosuccinic Acid Fe304 Nanoparticles In keeping with the theme ofthiol-based sorbent materials for toxic metal cation remediation, dimercaptosuccinic acid coated iron oxide nanoparticles (DMSA NP) were investigated.29 Iron oxide nanoparticles (Fe NP) possess many desirable properties making them an interesting material for use in aqueous systems. Fe NP can be treated with a number of ligands for use in metal ion remediation or even metal ion sensing from natural water sources and bodily fluids?O Typically, Fe NP are first synthesized by the stabilization of the nanoparticles with a lipophilic outer ligand shell. Exchange reactions of the outer shell can impart both functionality and in this case, water solubility. Examples of exchanged ligand shells include ethylenediaminetetraacetic acid (EDTA),31 humic acids,32 and, in this work, dimercaptosuccinic acid, all of which have good affinities toward toxic metal cations. Fe NP are superparamagnetic meaning they are attracted to a magnetic field and re-disperse when the field is removed. This is an attractive property due the potential of 'harvesting' the DMSA NP sorbent after achieving maximum uptake. These properties along with the ability of Fe NP to be easily dispersed in aqueous media make them ideal for heavy metal capture. Similar tests described above for Thiol SAMMS were also performed with DMSA NP to determine their feasibility as a sorbent material for toxic metal cation capture and are described below. DMSA NP Results and Discussion DMSA NP were tested in either filtered natural water sources or in-house generated Nanopure water spiked with target metal cations. In some experiments, pH 77 was adjusted by the addition ofHN03. DMSA NP were allowed to soak in metal ion spiked test solutions for two hours unless otherwise noted. After the two hour soak, a measured volume of test sample was removed, filtered with a 0.2 Ilm syringe filter and diluted appropriately for ICP-MS analysis. Values are reported in log Kd unless otherwise noted. In this study, Nanopure water was spiked with Ag1+, Cd2+, Hg2+, Pb2+and TI1+ and the pH was adjusted with HJ\T03. The initial and final ion concentrations were determined by ICP-MS with each sample run in duplicate. Kd values were averaged and the log value graphed in Figure 5. From the data, it became apparent that the pH- dependent binding to Pb2+and Cd2+was not present with DMSA NP. The likely explanation for this observation is the added benefit of the chelation effect with a.-fJ thiol head groups on the DMSA ligand. The chelation effect essentially states that when DMSA is bound to a cation, the thiols, ethyl backbone and metal center form a highly stable, five membered ring with optimized geometry, imparting favorable binding which is thus less perturbed by the effects of change in [H+].33,34 Another explanation to the lack of observed pH-dependent behavior warranting mentioning is the catechol-salicylate mode of binding. Cohen and coworkers discovered that low pH, protonation of a catechol alcohol resulted in the metal center 'hopping' over to a carbonyl oxygen of2,3- dihydroxybenzoyl amide ligand.35 This too may occur with DMSA, where the protonation of one thiol may result in the metal center binding with the neighboring carbonyl oxygen thus no significant change in binding will occur as pH decreases. 78 In addition to being unaffected by pH due to the formation of a stable ring, DMSA NP experience very little leaching. Overall, DMSA NP perform quite well from fairly acidic to neutral pH. They do however suffer from lower Kd values due to their decreased surface area, as compared to Thiol SAMMS. 6 5 o 1 2 3 pH 4 5 Figure 5 DMSA NP uptake or Ag l -', Cd 2+, Hg2+, Pb2> and 1'1 1+ in I-I NO) spiked Nanopure water at indicated pH values_ Samples were soaked in the test solutions for two hrs and ion concentrations determined by ICP-MS. DMSA NP were also tested for ion capture efficiency in metal cation spiked natural water sources to examine the effects that water chemistry played on ion uptake. As described above with Thiol SAMMS, natural water sources were first filtered then spiked with target ions. After a two hour soak, remaining ion concentrations were 79 quantified by rep-MS. An additional test matrix (bovine serum albumin, BSA) was also examined to determine the feasibility of DMSA NP for use in in vivo metal ion capture therapy Figure 6. 7 6 4 ; .Ag ~ • Cd tul 0 Hg-' • Pb TI a CRW HGW SEA BSA Figure 6 Plot of log Kd values for DMSA NP in Columbia River water (CRW), Hanford ground water (HGW), Pacific Ocean water (SEA) and bovine serum albumin (BSA). DMSA NP performed fairly consistently in all four test matrices for all cations. Kd values above 101 are considered acceptable and values above 104 exceptional with DMSA NP achieving values above 104 for the fresh water sources.16 DMSA NP performance is slightly lower than Thiol SAMMS under similar conditions, presumably due to the reduced number of binding sites as a result of decreased surface area. 80 However, this material does perform consistently better than the commercially available resin GT73 under most conditions. With the exception of Cd2+ binding, DMSA NP also perform quite impressively in BSA, especially considering the complex nature and potential for competitive binding with metal ions in the mixture. A number of experiments have been performed by PN1\TL staff and others to demonstrate the ability of DMSA NP to work in a number of biological fluids for both uptake and detection of toxic ions?9,30 Pb adsorption on DMSA-Fe304 vs. time (LIS =10000) 18 o 50 2 16 4 14 6 12 Cb10 ~SQ. 8 :s .c a. 40 ~ Pb concentration -111- Pb Uptake 20 30 Time (min) 10 )I-_.-.-.....----..---.-.....----lIIl---.....---.-..------.-( 1800 1600 -.cQ. 1400Q. -C 0 1200+:i :s 0 t/) 1000 c c 8000 +:i f! - 600c Cb(J c 4000 (J .c a. 200 0 0 Figure 7 DMSA kinetics study in Pb spiked 0.025M sodium acetate buffer. LIS of 10000 was used to minimize the effect of sampling during the experiment. 81 Lastly, kinetic studies were performed with DMSA NP in Pb2+ spiked 0.025M sodium acetate buffer at pH 6.5. A 100 mL test solution spiked with Pb2+ and a liquid-to- solid ratio (LIS) of 10000 (1 OOmLlO.O 1g) was used so that the removal of test solution during the course of the experiment would have a minimal effect on ion capture. Test samples were taken at the following points: 0, 1,5, 10,30,60, 120,480 and 1440 minutes and the results were plotted in Figure 7. Inspection of the results from the first 50 minutes reveals near complete uptake by t = 30 min. No leaching was evident during the duration of the full test period of 1440 minutes, likely a result of the tight bonds between DMSA and the metal cation and DMSA and the Fe NP. To be an effective sorbent, especially for in vivo treatments, the material must be able to remain bound to the captured toxic ion. Microwave Digestion of Thiol SAMMS and Thiol Containing Fe Nanoparticles There is currently no portable technology capable of sampling water sources in remote locations while providing accurate quantification of trace metals. Thiol SAMMS offers a unique opportunity to provide accurate analysis of minimally accessible water formations by pre-concentration of toxic ions followed by ICP-MS analysis in a certified laboratory. This can be achieved by passing a known amount of test sample over Thiol SAMMS, then taking the material back to a laboratory for acid digestion. The metal content of the test source can then be quantified by ICP-MS. This section explores the concept of utilizing microwave-based acid digestion of Thiol SAMMS and 2,3- dimercaptosuccinic acid coated iron nanoparticles (DMSA NP) following the 82 Environmental Protection Agency (EPA) guidelines from method 3052 in report # SW- 846. This data demonstrates that Thiol SAMMS and DMSA NP materials can be loaded with a quantified amount of toxic metal cation, digested as outlined by the EPA as a certifiable laboratory technique and metal content ofthe resulting solution quantified by ICP-MS to give high returns of metal ions. Thiol SAMMS were synthesized as described in this chapter and Feng et al9 and DMSA NP as outlined by Sun et al and Yantasee et al.29,37 Stock solutions were prepared by dilution of metal standards in either filtered (0.45 !lm) Columbia River water (CRW) or filtered Hanford ground water (HGW) with a target concentration of 500 ppb for each metal. Final concentrations were determined by ICP-MS. Test solutions contained approximately 0.05 g of silica sorbent material or 0.01 g iron nanoparticles suspended in 50 or 10 mL metal stock solution, respectively. Solutions were mixed on a platform shaker for two hours, after which time, a small aliquot was removed from each test solution for metal uptake quantification. The remaining sorbent-containing solution was allowed to sit for 30 minutes followed by centrifugation for 30 minutes. The resulting pellet was rinsed 3 times with approximately 1 mL Nanopure water and stored hydrated until digestion. Microwave digestion was carried out as described by EPA method 3052 in report # SW-846. Briefly, the sorbent pellet was resuspended in its storage solution followed by addition of acid digestion solution (l0 mL) and quickly transferred to the fluorinated polymer digester. A final temperature of 180°C was reached over a five minute ramp followed by an isothermal period of 4.5 minutes. After this, the digesters and solutions contained therein were allowed to cool to 30°C before 83 dilution. Digestion solutions were transferred separately to 50 mL volumetric flasks and water added to reach the final volume. Samples containing HF were first quenched with boric acid before dilution by the addition of 3 mL saturated boric acid. Solutions were stored until metal ion quantification by ICP-MS was performed. The following acid solutions were used for the digestion process: 1. 10 mL, 95% HJ\TO} / 5% HCl 2. 10 mL, 90% RNO}/ 10% HCl 3. 10 mL, 75% RNO} / 25% HCl 4. 10 mL, 50% Nanopure water / 47.5% RNO} /2.5% HCl 5. 10 mL, 95% RNO} /5% HF 6. 10 mL, cone. RNO} Thiol SAMMS not loaded with metal was wetted with MeOH and digested with solution 1 as a control. DMSA NP were digested using 10 mL of concentrated RNO} (solution 6). EPA Standardized Digestion Results and Discussion This study was performed to demonstrate that small filter-like cartridges containing Thiol SAMMS or DMSA NP can be used to pre-concentrate metal ions from remotely located test areas and accurately quantify toxic metal ion concentrations. By adhering to strict guidelines established by the EPA for infield water sampling, acid 84 digestion of test samples and metal ion quantification by ICP-MS, this technique can be implemented as a validated testing protocol. Table 4 is a compilation of data from the microwave digestion of Thiol SAMMS sorbent material loaded with select metal ions. Two test matrices were used, Columbia River water and Hanford site well water. Metal uptake was nearly 100% for each ion and matrix tested, which is in good agreement with previously acquired data. Digest Solution % Uptake Sorbent Matrix (from above list) Ag1+ Cd2+ Hg2+ Pb2+ Filtered 1 100 100 100 100 Well 2 100 100 100 100 Water 3 100 100 100 100 Filtered 4 100 99.9 99.9 99.9 River 5 100 100 99.6 100 Water 6 100 100 99.7 100 Thiol SAMMS % Recovery Filtered 1 23 87 >100 95.5 Well 2 77.8 93 >100 >100 Water 3 75 82.6 >100 90.9 Filtered 4 53.3 86.7 78.9 86.7 River 5 8.9 86.4 81.6 86.4 Water 6 35.6 80 74.4 84.4 Table 4 Percent uptake and recovery for select metal cation for Thiol SAMMS in filtered Hanford site well water and Columbia River water in Richland, WA. Recovery percentages exceeded 100% in some runs most likely due to calibration error of the ICP-MS resulting in drift. 85 Recovery percentages after digestion, however, seemed to vary greatly between metal ions and acid solutions in this experiment. Recovery rates greater than 100% indicate two possible errors in the analysis technique: 1) introduction of metals from external sources such as test containers, SAMMS materials or acid sources or 2) improperly calibrated ICP-MS resulting in drift during the analysis ofthe digest solutions. Scenario 2 is most likely the cause of error due to the lack of any obvious anomalous trends in metal concentration between samples. Similar digestion experiments were carried out with DMSA NP. As previously demonstrated, DMSA NP were found to achieve high uptake percentages from metal ion spiked solutions as measured by ICP-MS (Table 5). % Uptake Sorbent Matrix A~l+ Cd2+ H~2+ Pb2+ Filtered 99.9 97 97.5 99.9DMSA NP Ground % RecoveryWater >100 >100 >100 >100 Table 5 Percent uptake and recovery for selected metal cations for DMSA NP in filtered ground water from a well located on the Hanford site (Richland, WA). Recovery percentages exceeded 100%. Runs were performed in duplicate. The DMSA NP digest solutions were analyzed by ICP-MS at the same time as the above Thiol SAMMS thus resulting in the same introduced error. Both experiments are worth revisiting in the future based on the promising data given above. In some digest test solutions, both Thiol SAMMS and DMSA NP solids were completely dissolved. This 86 was verified both visually and by filtration. The fact that no solids remain is a clear indicator that all metal ions must have been released into solution. Therefore, assuming that the above error in recovery can be isolated and corrected, this procedure should be a viable technique for trace metal detection for natural water sources. Experimental General Procedures Water sources where taken either from the Columbia River (Richland, WA), well water located on the Hanford site, (Richland WA), Pacific Ocean, (Sequim, WA) or Nanopure (18.2 mn) water produced in-house where noted. All natural water sources were passed through a 0.45 /lm filter before use and stored in the dark. BVA was used as received. Glassware and plastic bottles were rinsed with a 10-20% nitric acid (low metal ICP-MS grade) solution followed by copious rinses with Nanopure water. All SAMMS materials were weighed out on a bench top analytical balance and reported to the nearest tenths of a mg. Digital pipettes (manual and electronic) were used to deliver small volumes of solutions (less than 100 mL). Large volumes were measured on a bench top analytical balance to the nearest tenth of a gram with the assumption of density equal to 1 for all solutions. Natural water samples were passed through a 0.2 /lm filter by vacuum and stored in the dark until used. Test samples were filtered with 0.2 !!m syringe filters before ICP-MS analysis. Samples were diluted as needed to range between 50 to 100 ppb to best match metal calibration curves. ICP-MS was performed on a Agilent Technologies 7500ce with calibration standards checked periodically to eliminate drift. All experiments were performed in duplicate or triplicate and values averaged before plotting unless otherwise noted. 87 Uptake Experiments SAMMS materials (approximately 10 mg) was weighed into 20 mL scintillation vials and 5-10 /lL ofMeOH was added to wet the material. To this, 10 mL oftest solution was added and the vials agitated for 2 hrs on a platform shaker set to 1 Hz. After which, a small aliquot was removed and passed through a 0.2 /lm syringe filter and diluted with a 5% nitric acid solution for rCP-MS. Thiol SAMMS Thiol SAMMS were made in-house as described by Feng et al by means of 3-mercaptopropyltrimethoxysilane.9 GT73 Resin was purchased from Aldrich and used as received. Samples were kept sealed until used to prevent evaporation of as shipped swelling solvent. Kinetics Study 10 mg Thiol SAMMS was placed in a 250 mL Erlenmeyer flask and wetted with 5-10 /lL MeOH. To this, 100 mL of test solution was added and the timing experiment started. The flask was continuously agitated with a platform shaker set to 1 Hz. A small aliquot was removed at the designated time and passed through a 0.2 /lm syringe filter followed by dilution with 5% RN03 solution for rCP-MS. Microwave Digest ofSAMMS and DMSA NP Materials Unless otherwise noted, all solutions were prepared using Nanopure water (18.2 MQ-cm). Glassware, digestion vessels and hardware, as well as high-density polyethylene (HDPE) bottles, were initially washed with an aqueous 10-20% RN03 solution to remove trace metals followed by a rinse with Nanopure water prior to use. Digestion vessels and hardware were air dried overnight before use. Pipette tips, Falcon tubes and glass scintillation vials were used as received without acid wash. 88 Metal stock solutions containing approximately 500 parts per billion (ppb) of Cd, Hg, Pb and Ag were prepared in either filtered (0.45f..lm) Columbia River water (CRW) or Hanford ground water (HGW). Approximately 0.0500 g of Thiol-SAMMS material was weighed and transferred into screw top Falcon tubes along with 50 mL of either CRW or HGW metal stock solutions (in triplicate totaling six samples). The mixtures were shaken horizontally on a platform shaker for 2.5 hours at 1Hz and allowed to rest vertically for 30 minutes to help with settling of solids. Then, sample tubes were centrifuged for 30 minutes at 7000 RPMs. After centrifugation, solutions were decanted off into new Falcon tubes for metal ion uptake quantification by rCP-MS. The resulting pellet was washed three times with 1 mL Nanopure water while being careful not to disturb the pellet. Nanopure water (l mL) was added to each tube and placed on a vortex mixer to re-suspend the solids. Each tube was then charged with 10 mL of acid solution and quickly transferred to a Teflon microwave reaction vessel and individually heated per the EPA guidelines outlined in method 3052 in report # SW-846. Bridge to Chapter IV Chapter rv investigates cationic functionalized mesoporous silica materials for the uptake of arsenate and chromate oxyanions. This chapter is a continuation into the research pertaining to SAMMS sorbent materials first introduced in Chapter III. Studies include the effect of pH-dependent capture of target anions. Natural water sources are also tested to investigate the effects water chemistry has on binding. 89 CHAPTER IV METAL ANION CAPTURE THROUGH CATIONIC METAL CENTERS General Overview Chapter IV includes data from research performed at Pacific Northwest National Laboratory in Richland, Washington as a continuation of ongoing work with sorbent technology under the guidance of Dr. R. Shane Addleman. This chapter contains unpublished work with the intention of combining this data with existing pertechnetate data from PNNL for publication in either Analysis or Analytical Chemistry. Coauthorship will include Wassana Yantasee who provided experimental oversight and assistance, Thanapan Sangvanich who performed mass spectrometry analysis, Glen E. Fryxell who provided experimental oversight, Matthew O'Hara who performed pertechnetate studies and Darren W. Johnson and R. Shane Addleman who provided intellectual contributions and editorial input. Introduction Oxyanions such as chromate, arsenate and pertechnetate pose serious environmental hazards yet are challenging remediation targets due to their tetrahedral ionic structure, weak basisity and diffuse electron density.l 90 Flocculation,2 coprecipitation3 and reverse osmosis4 are typical techniques employed to remove toxic oxyanions from drinking water sources, none of which are selective toward oxyanions over common anions found in water such as Cl l -. However, large, specialized equipment, reagents or multiple treatments are needed to achieve efficient scrubbing and are not practical in rural communities with limited resources. A more practical remediation approach is the capture of targeted oxyanions with sorbent materials housed in filter devices affixed to well heads or point-of-use water supplies. Examples of sorbent materials capable of sequestering oxyanions include ionic exchange resins,s biosorbents,6 inorganic layered double hydroxides (LDHs),7,8 quaternary ammonium containing materials9 and ethylenediamine (EDA) functionalized mesoporous silicas complexed with either Cu2+or Fe3+. This section is a continuation of the original study of Cu2+-EDA and Fe3+-EDA functionalized mesoporous silica with non-native water sources spiked with chromate (Cr6l, arsenate (Assl and pertechnetate (Tcs+) oxyanions to simulate waste streams.IO,11 The relevance of this work pertains to the expansion of test matrices into native water sources versus sterile laboratory prepared water samples. In both studies, test matrices are spiked with known amounts of metal anion and in some experiments, the pH is adjusted using HN03. Results from this work are compared to a commercially available ion exchange resin and a quaternary ammonium salt SAMMS material to simulate the functionality found in most anion specific exchange resins. 91 Fe 2+ or Cu 2+ A S5+ Cr6+ Tc7+ [----] [----] (",---,,] Figure 1 Graphical dcpiction of ethylenediamine (EDA) functionalized mesoporous silica without ClI or Fe metal center (left) with ClI or Fe metal ccnter (center) and with metal centcr and bound oxyanion (right). Notc that Fc2+ is used in solution but is quickly oxidized to Fe'+ when bound as indicated by the appearance of a deep red color. MCM-41 is reacted with 1-(2-aminoethyl)-3-aminopropyltrimethoxysilane to impmt ethylenediamine (EOA) chelation functionality in a relatively dense monolayer after condensation of the silane monomers by azeotropic removal of water. 12,13 Cu2+ or Fe3+ is incorporated by simple mixing of the EOA-SAMMS material and the corresponding metal(II) acetate or chloride salt in water or isopropyl alcohol, respectively, for two hours at 25 °C. I•IO Upon binding to the EOA ligand, iron undergoes oxidation from Fe2+ to Fe3+ as indicated by the formation of a deep red color. The result is a dense layer of cationic metal centers with a high affinity and specificity for chromate, arsenate and pertechnetate oxyanions over other anions typically present in surface waters (Figure 1). 92 Results and Discussion As discussed thoroughly in Chapter IV of this dissertation, pH has an overwhelming effect on metal ion binding and the efficiency of sorbent materials. To further study the pH-dependence of sorbents, water from the Columbia River (Richland, WA) was passed through a 0.2 flm filter followed by metal ion spiking and pH adjustment with HN03. River water was chosen over ion doped Nanopure water for the purpose of simulating real world applications by testing the effectiveness of the sorbent with competitive natural ions and buffering components. It is acknowledged that the filtering step does remove large particulates and bioorganisms which may affect the efficiency of the sorbent material but is necessary to prevent spoilage of the test sample over the period of the study. Due to the low concentration of metals, data is reported as the distribution coefficient (Kd) or the mass-weighted partition coefficient of the sorbent material and test solution unless otherwise stated. The distribution coefficient can be determined by Equation 1 as stated on page 65 in chapter III by knowing the initial and final concentration, Co and C/, respectively, of the target ions which is determined by ICP-MS, the volume, V, in milliliters and the mass, M, in grams of the sorbent material. Capacity values which compare the amount of target ion to amount of sorbent material used would not be ideal for these experiments do to the potential of influencing the uptake equilibrium. 93 pH Studies ofCu2+- and Fe3+-EDA SAMMS The positive charge imparted on the Cu or Fe metal center when chelated to the surrounding amines provides for a high affinity of both Cr6+and As5+oxyanions from a pH range of3 to 7 (Figures 2 and 3, respectively). Below pH 3, Cu2+-EDA SAMMS quickly loses its ability to bind to HxAs04(x-3) ion but only exhibits a modest loss in binding to CrO/- ion (Figure 2). Fe3+-EDA SAMMS also suffers a decrease in binding affinity with HxAs04(x-3) below pH 3 but to a much lesser extent than Cu2+-EDA. However, Fe3+-EDA SAMMS has an overall lower affinity for HCr04 I- than Cu2+-EDA. With the exception of a complete loss of binding to HxAs04(x-3) at pH 1 for Cu2+-EDA, both materials appear to maintain good uptake of the target ions over 7 pH units. Cu 2+-EDASAMMS in HN03 spiked river water 6 .... 5 ~ ~/ ....4 ~r ~u~ ~Cr04(9 3 / /a .....As04...J 2 I1 I0 0 2 4 6 8 pH Figure 2 pH-dependency ofCu2+-EDA SAMMS material in Cr and As5+spiked Columbia River water. Low metal HN03 was used to adjust pH. Metal uptake was determined by ICP-MS. 94 Fe3+-EDA SAMMS in HN 0 3 spiked river water 6 5 K .....--...... ~ ~4 ?~"0~ -+-Cr04CJ 3 (/0 I ooooo&-As04-l 2 / 1 0 0 2 4 6 8 pH Figure 3 pH dependency ofFe3+-EDA SAMMS material in Cr6+and As5+ spiked Columbia River water. Low metal HN03 was used to adjust pH. Metal uptake was determined by ICP-MS. The observed decrease in binding can possibly be attributed to at least two causes: 1) protonation of the ethylenediamine nitrogen atoms, or, 2) change in speciation of the target oxyanions. To address the observed trends of both MX+-EDA SAMMS materials, it is important to discuss each above scenario separately, starting with hypothesis 1 (amine protonation). Yoshitake and coworkers have demonstrated that soaking Fe3+-EDA SAMMS in 1M Hel after metal uptake for 10 hours at room temperature results in almost complete desorption of both the captured oxyanions and the iron metal center due to protonation of the diamine nitrogen atoms.! Some Fe3+and oxyanion did remain, however, so complete stripping was not achieved but the treatment did not degrade the covalently attached EDA monolayer. Test batches of acid stripped EDA SAMMS were capable of undergoing regeneration by the simple loading of Fe3+back onto the material. Considering the extreme conditions necessary to strip the metal ions from the EDA 95 SAMMS, it is reasonable to conclude that the amines are not undergoing protonation to any appreciable extent and therefore not releasing the bound Fe3+ or Cu2+ metal centers (Table 1). Even with pKas around 10.3 and 7.5 for N-n-propylethylenediamine (a similar compound to the attached monolayer of EDA SAMMS material),14 as in the case of organothiols when bound to metals, the pKa can drop many orders in magnitude to resist proton addition when acting as a chelate ligand. IS Table of Acid Dissociation Values H2AsO/- 2.24 Arsenate HAsol- 9.20 Asol- 20.7 Chromate HCr04J- -0.86 CrOl- 6.51 N-n- R-NH2J+-R' 7.5 propylethylenediamine R"-NH3J+ 10.3 Table 1 Acid dissociation values for arsenate and chromate oxyanions and N-n-propylethylenediamine (a similar diamine to 1-(2-aminoethyl)-3-aminopropyltrimethoxysilane used in EDA SAMMS). Additional evidence suggests that the nitrogen centers are not being protonated due to the lack of a conserved trend in decreasing Kd between materials and oxyanions. There is no indication of clearly defined titration curve as a result of protonation of the amines as would be expected, however, one could argue that some protonation does occur resulting in the decrease in Kd values around pH 3.5 for both materials and oxyamons. ------- 96 The second argument is due to a change in speciation of the target oxyanion as the pH decreases. This hypothesis does appear to fit in the case of Cu2+-EDA binding HxAs04(d) ion. At around pH 2.2, arsenate undergoes a loss of a proton (or gain depending on the direction of the reaction) resulting in the following equilibrium expression (Equation 1): IS Equation 1 Equilihriul1l equation of arsenate speciation under acid conditions. The pK" for this process is 2.2 which is in good agrecmcnl with the observed trend in Figure 2 rcsulting in a total loss of arsenate uptakc. As(V) Speciation 100 --,..==-------------::::====::::::::-----------, 80 ~ Hv\sO, H2As04 1• 0 ...- 60~ m]1 6 ~ 40 HAsO<\2- ~0 20 °T----.----.----r--...,.---...,.-----.----.---,' o 2 4 pH 6 8 Figure 4 Speciation diagram of As1+ oxyanion from pH 0 to 8. Plot generatcd using HySS2006, part of the HyperQuad suite. 97 This is in good agreement with the observed data for As04 binding with Cu2+-EDA (Figure 2). The inflection point does appear to be close to pH 2.2 or equal to the pKal of H3As04, a neutral species. 16 The formation of the neutral arsenate species below pH 2 results in the loss of any affinity toward the charged cationic Cu2+ center Figure 4. Albeit reasonable in explaining the trend for Cu2+, it does not explain why Fe3+_ EDA SAMMS is unaffected by pH for HxAs04(x-3) other than the drop in two log units typical of all EDA SAMMS tested. The observation with Fe3+-EDA SAMMS can be explained by the formation of a tightly bound Fe(III) H2As041- complex which alters the pKal of the arsenate oxyanion. 17 The slight lowing of pKal of the arsenate oxyanion allows for modest binding even at low pH. This complex does not form with copper therefore protonation of the oxyanion occurs as expected. However, further research into the mechanism for the observed behavior of cationic EDA SAMMS materials is needed to conclusively address the pH-dependence of this class of sorbents. Chromium(VI) on the other hand does not experience any appreciable protonation between 0-6 and most likely remains unaffected by the drop in pH. 18,19 Above pH 6, HCrO/ loses a proton to become CrOl- which would still likely bind to the cationic metal center of EDA SAMMS. Below pH 6, Cr20l forms which constitutes about 20% of the chromate species in solution and may attribute to reduced uptake at lower pH values. This is in good agreement with the observed data for both Cu2+_ and Fe3+-EDA SAMMS which exhibit similar pH binding dependencies (Figures 2 and 3, respectively). There is clearly a reduction in binding as the pH decreases from 4 to 0, again requiring further investigation into the pH-dependant mechanism. 98 Anion Uptake in Three Natural Water Types To test the affects of water chemistry on binding ofCr6+ and As5+, three natural sources were investigated at near neutral (or as collected) pH. As described previously, water sources were passed through a 0.2 ).lm filter followed by the addition of chromate and arsenate ion. Sorbents were soaked for two hours and the amount of metal anion quantified (initial/final) by ICP-MS. All runs where in duplicate and reported as the average Kd with the exception of the Bio-Rad resin and quaternary ammonium containing SAMMS data which were single runs and thus inherently contain some degree of error but are in agreement with past, in-house observations. The inclusion of a quaternary ammonium-based resin from Bio-Rad as well as a N-propyl-N,N,N-trimethyl quaternary ammonium (Quat Salt) SAMMS in these studies provides a side-by-side comparison of the current technology for targeting these oxyanions. In the first graph (Figure 5), log Kd values of Cr6+ is compared with all four materials in spiked Columbia River water, Hanford site well water and Pacific Ocean sea water at near neutral pH. As expected, all materials perform particularly well in fresh water sources but are greatly affected by the high ion concentration in sea water. 99 6 5 4 • CU(II)-EDA Fe(III)-EDA Bio-Rad Resin • Quat Salt 2 o River Water Ground Water Sea Water Figure 5 Plot of log Xu values for C/\+ oxyanion ill neutral Columbia River Watcr (Icft), Hanford site ground water (ccnter) and Pacific Ocean sea water (right). The high ion concenlrillion of the sea water causes reduced binding of the oxynnion with all four testmaterinls due 10 compcting rcactions. Both EDA-based SAMMS perform about the same as the commercially available resin from Bio-Rad for arsenate ion in fresh water but suffer a similar reduction in binding in sea water (Figure 6). Fe3+-EDA SAMMS again is the overall better pelforming material for both arsenate and chromate ion uptake from native water sources in each experiment. 100 5 4 3 Cu(II)-EDA Fe(III)-EDA 2 l' I 0 River Water Ground Water Sea Water Bio-Rad Resin • Quat Salt Figure 6 PIOl of Log K" values for As~+ oxyanion in neutral Columbia River Water (left), Hanford site ground water (center) and Pacific Ocean sea water (right). The high ion concentration of the sea water causes reduced binding of the oxyanion with all four test materials due to completing reactions, Conclusion of Anion Binding From the provided data, it is apparent that Cu2+- and Fe3+-EDA SAMMS are effective sorbents for arsenate and chromate oxyanions in natural water sources from neutral to highly acidic conditions. Fe3+-EDA SAMMS is a better overall sorbent for both oxyanions at low to neutral pH although Cu2+-EDA SAMMS exhibits a similar binding trend for Cr6+ through this range. Water chemistry clearly affects metal binding in sea water greatly reducing the efficiency of both materials. Because of the complex makeup of the natural water sources, it is extremely difficult to claim conclusively which 101 species are present within each water source and how binding is affected by the presence of multiple ions. A number of organic acids present in natural water sources could be affecting binding through competitive binding routes. As a whole, Cu2+- and Fe3+-EDA SAMMS are easy to generate, perform equal to or better than commercially available resins and work over a wide pH range for the uptake of arsenate and chromate. Experimental General Procedures Water samples were taken from the Columbia River (Richland, WA), a well water located on the Hanford site (Richland WA) the Pacific Ocean (Sequim, WA) and in-house Nanopure (18.2 mn) water sources where noted. Glassware and plastic bottles were rinsed with a 5% nitric acid (low metal ICP-MS grade) solution followed by copious rinses with Nanopure water. All SAMMS materials were weighed out on a bench top analytical balance and reported to the nearest tenths of a mg. Digital pipettes (manual and electronic) were used to deliver small volumes of solutions (less than 100 mL). Large volumes were measured on a bench top analytical balance to the nearest tenth of a gram with the assumed density equal to 1 for all solutions including sea water. Variation in the actual water sample densities results in a slightly higher or lower final metal concentration which is measured accurately by ICP- MS and therefore accounted for in each experiment. Natural water samples were passed through a 0.2 /lm filter by vacuum and stored in the dark until used. Test samples were filtered with 0.2 /lm syringe filters before ICP-MS analysis. Samples were diluted as needed to range between 50 to 100 ppb to best match metal calibration curves. ICP-MS 102 was performed on an Agilent Technologies 7500ce with calibration standards checked periodically to eliminate drift. All experiments were performed in duplicate or triplicate and values averaged before plotting unless otherwise noted. MCM-41 was generated in house with a pore size of approximately 35A as determined by BET analysis and surface area of 761.1 m2/g. Ethylenediamine SAMMS (EDA SAMMS) were made in house as detailed by Fryxell et al. I0 Quat Salt SAMMS Combined 5.102 g MCM-41 with toluene (200 mL) in a 500 mL round bottom flask under nitrogen while stirring. Nanopure water (1.8 mL) was added and the solution stirred for approximately 4 hours. Added N-trimethoxysilylpropyl- N)V,N-trimethylammonium chloride (50% solution in MeOH, 5.053 g, 9.8 mmol) by syringe and heated to reflux. Continued stirring overnight for a total time of 16 hrs. Cooled slightly and affixed a Dean Stark trap between the flask and condenser, returned to reflux. Continued for two hours (until obvious signs of water collection in the trap) then cooled slightly. Added 100 mL of isopropyl alcohol (lPA), returned to reflux until ~4 mL total water was collected in the trap (approximately 1.5 hrs). Cooled to room temperature and added 200mL IPA while stirring. Vacuum filtered with a sintered glass Buchner funnel and washed with copious amounts of acetone, slurring each time to assure complete rinsing of SAMMS material. Dried in vacuum oven at 30°C and -25 inches Hg overnight. Total mass gain 2.067 g for a loading of 1.621 molecules per nrn2 (target 1.5 molec./nrn2). 103 2+ :Cu -EDA SAMMS Combined copper (II) acetate monohydrate (5.836 g, 29.2 mmol), and Nanopure water (250 mL) and stirred until dissolved. Added EDA-SAMMS (5.010 g, 761.1 m2/g, 3.5 molec./nm2) and swirled gently. Solution turned from dark blue to light blue while continuing to stir on a platform shaker at 1 Hz for 1 hr. The material was filtered then re-slurred with IPA. Dried in vacuum oven overnight at 30°C and -25 in. Hg. Transferred to 500 mL round bottom t1ask and added toluene (200 mL) under N2. Heated to reflux for 3.5 hrs then cooled slightly. Affixed a Dean Stark trap atop the flask and returned to reflux for 2 hrs. Cooled overnight and filtered the next morning. The solid was washed with 3 x 100 mL acetone, slurring each time. Dried in vacuum oven at 25°C over the weekend. Final weight 5.055g (0.22 mmols Cu2+ addition). Fe3+-EDA SAMMS This material was prepared as described above using FeCb as described by Yokoi and coworkers.! Bridge to Chapter V Chapter V discusses a new class of sorbent material which utilizes weak intermolecular interactions to stabilize aryl ring containing ligands to an aryl monolayer on mesoporous silica. The use of benzyl and dibenzyl thiolligands non-covalently bound to the silica substrate has proven to be quite effective in toxic metal ion capture. In many respects, this material performed as good as the covalently bound material previously studied in Chapter III. Studies include the effects of pH and water chemistry on binding cations in both HN03 spiked Nanopure and native water sources. 104 CHAPTER V NEW FUNCTIONAL MATERIALS FOR HEAVY METAL SORPTION: "SUPRAMOLECULAR" ATTACHMENT OF THIOLS TO MESOPOROUS SILICA SUBSTRATES Some of this work has been previously published and is reproduced with permission from: Carter, T.G.; Yantasee, W.; Sangvanich, T.; Fryxell, G.E.; Johnson, D.W.; Addleman, R.S. Chem. Commun. 2008,43, 5583-5585. General Overview Chapter V describes a research project conceived at the University of Oregon and reduced to practice at Pacific Northwest National Laboratory in Richland, Washington under the guidance of Dr. R. Shane Addleman. This chapter contains work published in Chemical Communications © by the Royal Chemical Society and coauthored with Wassana Yantasee who provided experimental oversight, Thanapan Sangvanich performed mass spectrometry analysis, Glen E. Fryxell provided experimental oversight, Darren W. Johnson and R. Shane Addleman provided intellectual contributions to the project and editorial input. Leaching data acquired by Sean A. Fontenot and Brian Theobald is included in the conclusion section. This work is ongoing at both the UO and PNNL by Sean A. Fontenot toward his dissertation. 105 Introduction Access to sustainable, clean drinking water is an increasing concern as the Earth's human population continues its steady growth. l Degrading water quality in both industrialized and non-industrialized nations has the potential to cause great economic strain on the world's governing bodies.2 At the same time, this offers a challenge to chemists to discover new, functional, designer materials that have high loading capacities and selectivity for environmental contaminants.3 Regenerable reusable functional materials have the added benefit of providing a sustainable approach to clean drinking water by reducing waste and increasing the lifetime of the product. Although water contamination from natural sources does occur, such as the devastating results of high arsenic levels in Bangladesh,4 a significant level can be attributed to human activities. The need to develop inexpensive and efficient water purification media is a high priority.a Current filter media typically consist of granular activated carbon (GAC) or a hybrid material that combines an inorganic oxide such as gamma alumina to achieve satisfactory water purification. This type of filter medium, although common and inexpensive, offers limited uptake potential. Additionally, spent media must be disposed of properly to prevent it from conceivably becoming a source of contamination due to leaching over time. An alternative to "one-shot" filters is the implementation of renewable media capable of many purification cycles before fouling. One such approach is described in this chapter utilizing the technique of non-covalent absorption of • In February 2005, the National Academy of Engineering (NAE) announced a $1 million prize for practical technologies capable of sequestering arsenic from drinking water. Abul Hussam of George Mason University, Fairfax, VA SaNa household filter system was the recipent of the prize in 2007. 106 organothiolligands onto self-assembled monolayers on mesoporous supports (SAMMSTM). Both the Darren Johnson laboratory and Pacific Northwest National Laboratory have programs designed to understand how toxic metal ions, specifically main group and transition metal ions, interact with thiolate ligands by studying their binding preferences using a supramolecular approach or through monitoring hazardous ion uptake from contaminated waters using functionalized materials.5-9 Functionalized mesoporous supports have been found to be excellent sorbent materials, capable of being chemically modified with reactive head groups for toxic metal, metalloid and oxyanion uptake as well as radioactive species. 10-12 One particular silica-based support, MCM-41, has garnered much attention due to its controllable honeycomb-like porosity, structural integrity, chemical resistivity and high surface area, approaching 1000 m2 g-!.13,!4 A variety of commercially available and synthetically accessible functionalized organosilanes can be affixed inside the pores of the silica support as self-assembled monolayers. The result is a dense population of chelating sites which can achieve exceptionally high uptake levels of target toxic ion species. Chapter N and V describe in detail a number of covalently functionalized SAMMS materials for uptake of a variety of toxic metal cations and oxyanions, respectively, for use in natural aqueous media. Thiol-SAMMS derived from covalent attachment of an alkylthiolsilane (tris(methoxy)mercaptopropylsilane, TMMPS, for example) have demonstrated excellent uptake levels of soft and moderately soft metal ions such as Hg2+, Pb2+, Cd2+, and Ag\see chapter IV).!! We have extended the breadth of thiol-SAMMS style sorbent 107 materials with this research to include noncovalently bound thiol and dithiolligands attached by relatively weak, reversible 1t-stacking interactions. It was recognized that aryl monolayers (such as a phenyl-based system), in which the aromatic rings were rigidly held upright and perpendicular to the surface, were well suited, both sterically and electronically, to serve as hosts to other functionalized arenes. This approach was envisioned not only to be a very versatile and easy functionalization strategy, but also one that would readily provide for 'refreshable' functionalization. This chapter describes the functionalization of MCM-41 with phenyl monolayers at various densities to provide a hydrophobic scaffolding for noncovalently bound benzylmercaptan (BM), 1,3- and 1A- bis(mercaptomethyl)benzenes (1,3- and IA-BMMB, respectively) for use in heavy metal cation uptake from native waters (Figure 1). SH 6SHI~~ ~SH~SH SH SH SH SH SH Figure 1 Graphical representation of organothiolligands benzyl mercaptan (BM), 1,3-bis(mercaptomethyl)benzene (1 ,3-BMMB), 1A-bis(mercaptomethyl)benzene (I A-BMMB) (top). Graphical representation of functionalized mesoporous silica monolayer with 1'(-1'( interactions between functionalized support and chemisorbed ligand (bottom). Shown is an idealized offset stacking but the actual arrangement may be in a herringbone-like manner as charecterized in other phenyl modified surface characterization data. 108 Results and Discussion Remarkably, 1,3- and 1,4-BMMB SAMMS exhibit comparable metal ion uptake levels to their covalently bound analogs. BMMB SAMMS were prepared by functionalizing MCM-41 at different loading levels by first hydrating the silica surface with a toluene/water mixture (equal to two monolayers of water based on total surface area) followed by addition oftrichlorophenylsilane and an overnight stir at room temperature. This was achieved by first dispersing the MCM-41 in toluene by stirring. Water was added, resulting in flocculation of the silica immediately followed by dispersing over 2-3 hours of stirring. Trichlorophenylsilane was then added and stirred overnight. The resulting solution was translucent in appearance. Isopropyl alcohol (IPA) was added and the solid filtered, washed with IPA and vacuum dried in an oven at 40°C. Phenyl coverage ranged from 0.01 molecules nm-2 (sparsely covered surface) to 3.1 molecules nm-2 (near maximum achievable loading accessible by this technique) by varying the trichlorophenylsilane and water concentrations. BMMB SAMMS Characterization Transmission electron microscopy (TEM) of phenyl loaded MCM-41 shows an increase in electron density in the pore structure of l,4-BMMB SAMMS with saturated Pb2+ ion loading (Figure 2). The left hand image shows the honey comb hexagonal arrangement of the pores with thin SiOx walls circled in red. This structure is not visible in the Pb2+ loaded material (right) presumably due to the high electron density of the lead versus the SiOx walls. Energy-dispersive X-ray spectroscopy (EDX) comparison of the 109 bright versus dark areas yields a high concentration of lead and sulfur in the bright areas and silica and oxygen in the dark regions. The lack of hexagonal porous structure could be due to the misalignment of the pores with the beam resulting in viewing down the length of the pore wall causing the effect of bright spots versus resolved pores. This is apparent in the non-lead containing TEM image as well. Nonetheless, the distance measured center to center between bright spots is in good agreement with the BET measurement of 40 Apores. Figure 2 TEM micrograph of 1,4-BMMB SAMMS without Pb2+ (left) and with Pb2+ right. Hexagonal arrangement indicated by red hexagon on right image. Total phenyl loadings were determined by thermogravimetric analysis (TGA). TG analysis was collected on a TA Instruments 2950 coupled to an electron impact (El) Balzers Thermostar mass spectrometer at a ramp rate of 2°C per minute up to 600°C and capillary temperature of 100°C to assure complete bum off of organics and provide clean 110 transition temperatures. A comparison of 1,4-BMMB physisorbed onto native MCM-41 versus the phenyl modified support (Phenyl-SAMMS) containing chemisorbed 1,4- BMMB revealed a significantly different burn off rate by TGA after the initial water desorption. Two different BMMB ligand loadings (equal to 2: 1 or I: 1 phenyl moieties to BMMB molecules for the phenyl functionalized silica, low loading and high loading, respectively), were generated by adding Phenyl-SAMMS to a solution ofBM, 1,3- BMMB or l,4-BMMB dissolved in dichloromethane in a sealed vessel and placed on an orbital shaker and mixed overnight. After mixing, the vessel was uncapped and the solvent was evaporated followed by overnight vacuum drying at 40 DC. TGA ofthe native silica support shows a relatively rapid weight loss starting around 235 DC and ending near 255 DC (Figure 3, dashed lines). 100 -- - Low onMGM-4·1 .... '""'.... HiF onMiCM·~41 ---Low onPhenyl"S.A.MMS -,~=_.= Hi@! onPhenyl~S~MS ..... -~ ... ~~~. - _. '_.- 300 gO laO 200 400 T elUp4!~nture °C 500 Figure 3 TGA comparison of various chemisorbed loading densities of 1,4-BMMB on phenyl functionalized MCM-41 (solid lines) versus physisorbed 1,4-BMMB on native MCM-41 silica (dashed lines). 111 1,4-BMMB ligand desorption from the phenyl modified support occurs at a slightly elevated temperature compared to the non-phenyl monolayer stabilized silica and continues over roughly 200°C to around 350 °C, verified by the continued detection of SO+ (47.9 m/z) and S02+ (63.8 m/z) by mass spectrometry. Degradation and oxidation of 1,4-BMMB occurred due to trace levels of oxygen contamination present in the TGA purge gas resulting in the observation of a large carbon dioxide peak and SOx peaks seen in Figure 4. ::::IA 502+ 5~2+ I.ODE"1j'~~~~ 01.00E'1l+-) -~-_--~-_--~- a a g ~ ~ ~ 15 mg ligand per 100 mg phenyl SAMMS 502+ I.30E·1O 11 1.10E·l0 I9.ODE· 1I i 7.5 mg ligand per 100 mg phenyl SAMMS 1.51);·10 1.3OE·10 1.10E·l0 9.00E·11 7.00E-11 5.00E·11 I 3.00E·l1 1.00E·l1 -lOOE-l1 43 502+ 4' 53 58 63 6B Figure 4 EI mass spectra of evolved gasses during the burn off process. Collected between 200 and 300°C. The loss of the phenyl monolayer was observed above 350 °C and continued to 600 °e, also monitored by EI-MS with ions detected at 49.9,50.8,51.8 and 78.2 m/z which correspond to C6H6+fragmentation typical of this type of mass analyzer. The extended burn off range of 1,4-BMMB can be accounted for by an increase in stabilization of the chemisorbed arylthiolligands afforded by the phenyl monolayer. Benzylmercaptan (BM) was also found to be stabilized by the phenyl monolayer as 112 indicated by TGA (data not shown) however, a strong, thiol odor emanated from this material making it unsuitable for use in filtration applications. Additionally, uptake levels for BM were less than those of 1,3- and 1,4-BMMB. In addition to thermogravimetric analysis, Fourier Transform Infrared spectroscopy (FT- IR) and Powder X-ray Diffraction (PXRD) spectroscopy was used to qualitatively visualize loading of organic material onto the silica substrate. FT-IR provided useful analysis of the surface makeup by comparing relative intensities of prominent peaks between samples (Figure 5). . - , 70 60 so -MCM-41 -Phenyl SAMMS -2:1 Loading 30 "~_,~ 1: Hoading 20 10 3,900 3,400 2,900 2,400 1,900 1,400 900 400 Figure 5 FT-IR of native silica (blue trace), phenyl monolayer (red trace), 2: I loading (purple trace) and I: I loading (green trace). Note the increase in C-H (aryl and alkyl) stretching as ligand loading increases as well as S-H stretching around 2545 wave numbers. 113 Changes in transmittance comparing to the native silica (MCM-41), phenyl monolayer a 2:1 loading and a 1:1 loading of phenyl base layer to 1,4-BMMB provides a qualitative spectroscopic handle for ligand loading. Thiol S-H and C-S stretching, 2545 and 669 cm- I , respectively, was normalized to the aryl C-H (2926 and 698) and C=C (1595,1512 and 1431 cm- I ) stretching to verify loading of both monolayer and adsorbed ligand. Powder XRD revealed a dominant (100) peak at 2.11 0 but lacked higher angle peaks for all substrates. It has been reported by others in the field that a decrease in peak intensity is directly related to the extent of modification of the pore with organics. We observed a similar decrease in the (100) peak intensity consistent with the chemisorptions of 1,4-BMMB onto phenyl-SAMMS. 15 Solution Phase Uptake Studies Initial metal uptake studies were carried out using well water from the Hanford site in Richland, Washington. Natural water sources such as that from a well provides a 'real world' test matrix containing typical ions and buffering agents. All aqueous solutions are from native sources from either a well located on the Hanford site, the Columbia River or the Pacific Ocean. Matrices were pretreated by passing through a 0.221lm filter followed by pH adjustments with HN03 and metal ion addition with target levels of 500 ppb per ion species in a typical experiment although some tests used 100 ppb ion concentration. Sorbent/matrix contact times were typically two hours with the sorbent material preconditioned with a few microliters ofmethanol to enable effective 114 surface wetting and facilitate metal ion uptake. This step is necessary due to the increased hydrophobicity imparted by the phenyl monolayer. Mass transfer of toxic ions from the aqueous test matrices to the functionalized sorbent material does not occur without the initial wetting-much like other neutral, organic modified silica sorbent materials. Uptake values are reported as the distribution coefficient (Kd), which is the mass- weighted partition coefficient between the sorbent material and matrix. The distribution coefficient can be determined by Equation 1 as stated on page 65 in chapter III by knowing the initial and final concentration, Co and CI, respectively, of the target ions which is quantified by ICP-MS, the volume, V, in milliliters as measured accurately by volumetric pipettes and the mass, M, in grams of the sorbent material measured to the nearest 10th of a milligram. Under trace level analysis such as this work, Kd is a more relevant value to gauge ion capture rather than mg/g (amount of metal ion captured to grams of sorbent used), which is typically used under saturation conditions. A plot oflog Kd values for Phenyl-SAMMS loaded at 3.1 molecules nm-2 containing chemisorbed 1,4-BMMB loaded at either 2:1 (low) or 1:1 (high) (monolayer phenyl moiety to aryl dithiolligand) shows similar uptake capacities with the covalently attached Thiol-SAMMS in Hanford well water matrix spiked with 500 ppb Hg2+, Pb2+, Cd2+, and Ag+ ions (Figure 6). 115 8.00 "'Cl ~ OIJ 6.00 o ~ 4.00 Thiol-SAMMS High Loading • Low Loading Ag Cd Hg Pb Figure 6 Log Kti values ofThiol-SAMMS, versus 1-4-BMMB at two different loading levels. A slight preference for Hg2+ with Thiol-SAMMS over the chemisorbed materials is observed possibly due to neighboring BMMB ligands binding with one Hg2+ ion thus reducing the actual available binding sites for target ions. Likewise, l,3-BMMB and l,4-BMMB exhibited similar uptake levels at near equivalent loadings (Figure 7). This is unexpected due to the assumed preferred orientation of the chelating thiol groups, with 1,3-BMMB positioned presumably with both chelating thiols above the plane of the monolayer whereas l,4-BMMB could conceivably have one or both thiol group buried in the monolayer making it inaccessible to the metal ions. This argument assumes that the aromatic rings of the phenyl monolayer and BMMB are orientated in an edge-to-face herringbone-like manner. 16 116 8.00 ~ ~ 6.00 4.00 • 1,3-BMMB High 1,3-BMMB Low 1,4-BMMB High 1,4-BMMB Low Ag Cd Hg Pb Figure 7 Comparison of 1,3 and I,4-BMMB at similar loading Icvels (boltom) in Hanford well water. All analyscs performed in either duplicate or trirlieate with variances grcatcr than 10% discarded. SH SH SH ~ Figure 8 Graphical reprcsentation of 1,3- and 1,4-BMMB ligands intercalated into phenyl monolayer with both thiols accessible (right and middle) I,4-BMMB with one thiol buried in the monolayer. However, this assumed packing appears not to be the case due to the similar uptake levels between Thiol-SAMMS, which has an overall higher surface ligand density (5 molecules nm- 2), versus the chemisorbed 1,3- and 1,4-BMMB, which have lower monolayer density 117 of 3.1 molecules run-2 but slightly higher chelation site density due to the difunctionality of the two arylthiolligands (Figure 8). A possible explanation of this observation is that the BMMB ligands are only partially intercalated into the phenyl monolayer in an offset stacking17 resulting in sufficient accessibility by the metal ions to the bulk ofthe thiol head groups (Figure 8). One interesting observation is that as l,4-BMMB loading increases, uptake decreases for Hg2+, Cd2+, and Ag+ but not for Pb2+. This may be due to the thiol sites becoming buried in the monolayer as loading densities increase. It has been shown that the densely packed thiol groups of Thiol-SAMMS material are shared by the same metal ion resulting in MLn species where n > 1 in previous studies.9 This desire to maximize metal-thiol contacts may manipulate the weakly bound chemisorbed ligands to adopt a more ideal geometry for binding at the various loadings investigated. In either case, the metal affinity levels of the chemisorbed BMMB Phenyl-SAMMS is near equal to that of the covalently bound Thiol-SAMMS which has been shown to have affinity levels for heavy metal ions one to three orders of magnitude greater than commercially available thiol-based resins such as GT-73. 8 The effects of density ofthe covalently bound phenyl layer were also probed by varying the loading from 0.01 phenyl molecules run-2 and as high as 3.1 molecules run-2 while maintaining 1,3- and 1,4-BMMB loadings levels equal to previous tests. Surprisingly, the sparsely populated Phenyl-SAMMS with chemisorbed BMMB performed equal to that of material of higher phenyl density for Hg2+ ion uptake. The exact nature of the interaction between the covalently attached phenyl ring and that of 118 chemisorbed BMMB is not well understood at this time, but the bound phenyl ring appears to be capable of acting as a nucleation site resulting in BMMB anchoring to the surface to provide an area rich in chelation sites capable of metal ion uptake-a feature lacking with the native silica and phenyl modified support. There is however an apparent reduction in stabilization of the BMMB active layer by the low phenyl base layer loading as evident in leaching studies (vide infra). pH-Dependent Uptake Studies The pH-dependent metal binding trend first described in Chapter IV of this dissertation is conserved for chemisorbed 1,3- and 1,4-BMMB SAMMS material. Experiments were carried out using metal ion spiked Columbia river water (Co2+, Cu2+, As5+, Ag1+, Cd2+, Hg2+, TI1+, Pb2+ at roughly 100ppb initial concentration) with a US ratio of 5000 and pH ranging from 0-8 in 2 unit increments (measured both before and after the two hour soak to account for drift in pH). Experiments were performed in triplicate with average values plotted. To simplify the discussion, data will be limited to Pb2+ and Hg2+ and are plotted in Figure 9 to demonstrate two typical pH-dependent binding extremes. 119 pH dependant log plot of Kd Hg (1 ,4-mercaploxylene) .......... Hg (1.3-mercaploxylene) ....... Pb (1 ,4-mercaploxylene) -- Pb (1.3-mercaploxylene) 8.00 j 7.00 --------------------.:=;;;;:::-.----- 5.00 t-----------------:r----------"-...:------, 2.00 ~"""""~----~'-------------------------, 6.00 .~---------------r"_.:---------- 3.00 -1-------_<---+------------------. u ~ 4.00 t----------~_/_-------------'---- o -J 1.00 +------------------------------j 9.000.007.006005004003.00200100 0.00 -I----,----,----,-------,r-----,-----,------,------,----; 000 pH Figure 9 pH-depcndcnt log K" pial or Hg2+ and Pb2+ rrom a mixed metal ion spiked Columbia river water matrix. From Figure 9, it is apparent that both 1,3- and l,4-BMMB chemisorb SAMMS follow similar trends for Hg2+ and Pb2+ ion binding with Thiol SAMMS (see Chapter IV). The exact nature of the dip observed for lead at pH 2 is unknown but was verified in triplicate as well as Pb2+ uptake experiments for benzoic acid-based and 1,3- MMB chemisorbed ligands performed on different days (vide infra). 120 Probing pKa Effects of Thiol Ligand In an attempt to try to demonstrate the above observation of pH-dependent uptake of metal ions, an experiment was devised which used thiol-based ligands with different pKa values Figure 10. SH ~ HX SH SH YJ Figure 10 Ligand selection with varying pKa. From left, 4-mercaptobenzoic acid, 4-(mercaptomethyl) benzoic acid and 1,3-BMMB. A rough estimate of the pKa values for 4-mercapto and 4-(mercaptomethyl) benzoic acid would lie around 10 and 15, respectively, based off of similar molecules in Bordwell's table measured in DMSO. It is reasonable to assume that these values will be closer to 6 or 7 for 4-mercapto and 10 to 12 for 4-(mercaptomethyl) benzoic acid in aqueous solutions. Figure 11 shows very little effect of metal uptake with varying pKa's of the thiol ligand. For all intents and purpose, these ligands behave similarly to each other within error and therefore no conclusions can be drawn from this data. Most likely, stronger electron donating and withdrawing groups will be necessary to properly verify the effect ofpKa on metal binding with thiolligands. It appears that the observed pH-dependent adsorption of metal ions can be primarily attributed to a metal ion's capacity to undergo 121 hydrolysis and secondly from an ion's water exchange rate or ease of forming a new, stable complex (see Chapter IV for a complete binding mechanism discussion). ...... Hg (4-mercapto) -#- Pb (4-mercapto) +-----1 ...... Hg (mercaptomethyl) "'*"" Pb (mercaplomelhyl) Hg (1 ,3-bis(mercaplomelhyl)) +---..., ...... Pb (1 ,3-bis(mercaplomethyl))3.00 7.00 ] 6.00 L,=====:::;;;,!'";;::.~;;;;;::;:;;;==~===:::;:::::;::::::::;::::::;:--------- 4.00 5.00 2.00 1.00 0.00 0.00 2.00 4.00 6.00 8.00 10.00 Figure 11 pH-dependent comparison of thiol benzoic acids and I,3-BMMB uptake for Hg2+ and Pb2+ over pH 0-8. Saturation Studies This study was peIformed to determine the maximum uptake capacity of one metal ion (Hg2+) which can be effectively absorbed by the chemisorbed sorbent material. Studies were carried out using pH-adjusted filtered Columbia River water doped with Hg(N03h for a final concentration of 500 parts per million (500 mg/L or ppm) and a working pH of 2. Liquid to solid ratios of 5000 (LIS, 100 mLlO.020 g) were used to 122 assure a favored equilibrium to the metal bound versus unbound ligand. Sample aliquots were removed after two hours, during which samples were gently agitated on a table-top mixer. Samples were filtered as described above to remove suspended sorbent material and diluted to an appropriate concentration for ICP-MS (target concentration of 50 to 200ppb is ideal to fit the calibration curve for most experiments). Base layers consisting of 3.1, 1.6 and 0.01 phenyl molecules per nanometer of silica were used with corresponding BMMB ratios (i.e. 1:1 or 2:1 binding of base phenyl rings to BMMB active layer). Under saturated conditions, 1,4-BMMB out performs 1,3-BMMB for every loading combination tested (Figure 12). This suggests that the thiol functional groups of 1,3-BMMB are in close proximity for one mercury ion to bind with two neighboring ligands thus effectively reducing the potential binding sites. This also suggest that the phenyl binding mechanism for both 1,3- and 1,4-BMMB are similar-with the BMMB phenyl rings intercalated into the phenyl monolayer in an offset stacking with both thiol functional groups available for binding. If this were not the case, and 1,4-BMMB had a buried thiol, a near equal or even slight preference for 1,3-BMMB would be observed. This is merely a hypothesis which requires advanced surface characterization techniques to elucidate absolute binding between the phenyl rings but many examples of solution data appear to support this binding motif. 123 8000 700.0 600.0 5000 ~ '"§- 400.0 -a; E 300.0 200.0 100.0 0.0 " I 01 ,4-bls(mercaptomethyl) ,----I---- I 01.3·bis(mercaptomethyl) I--- I Trt----- 1 I t----- -I '----- - '----- - - j---- 3.1 molec/nm2 (1 :1) 1.6 molec/nm2 (1: 1) 0.01 molec/nm2 (1:1) Figure 12 Saturation studies of three IA-BMMB/Phenyl loadings with Hg+2 doped Columbia river water. All samples run in triplicate. In a similar experiment, low loaded phenyl SAMMS (l:2 BMMB active layer to phenyl base layer) was used under saturation conditions. This data supports the above hypothesis that one mercury ion is binding to two neighboring 1,3-BMMB ligands. As was assumed to be the case, the 1,3-BMMB ligands were more dispersed in the phenyl base layer in the lower loaded material thus distributing the thiol functional groups around the sorbent and preventing multiple ligand binding to one ion. This is apparent from Fig re 13 with the near similar uptake for both 1,3- and l,4-BMMB, unlike the higher loaded material from Figure 12. At this lower loading, a slight improvement in the Hg2+ uptake by l,4-BMMB was also observed which would be expected as well. 124 1:2 ligand/monolayer 900.0 800.0 700.0 600.0 500.0 400.0 300.0 200.0 100.0 0.0 3.11:2p 1.61:2p 0.011:2p 3.11:2m 1.61:2m 0.011:2m Figure 13 Saturation data oflow-loaded phenyl SAMMS with 1,4-(designated by p) and l,3-BMMB (designated by m). In this experiment, the spacing of the thiol functional groups prohibits the shared binding of two ligands to one mercury ion. Binding Isotherm An experiment was performed to test the sorbent capacity toward a single ion by varying the total concentration of the target while keeping all other parameters fixed. 18 This data is helpful when attempting to determine the amount of sorbent material necessary to achieve a desired binding density of toxic target ions. Due to the wide range of Hg2+ concentrations for this experiment, a high LIS ratio (505000, 10mLll.98E-05g) was necessary. In this study, the uptake capacity of 1:1 loaded 1,4-BMMB phenyl SAMMS was quantified with Hg(N03)2 at pH 2 in spiked Columbia River water. HN03 was used to adjust the pH to an acceptable value. 125 At such a low sorbent concentration, it is possible to plot a binding or uptake isotherm from both experimental and calculated values. This is possible due to the increased target ion concentration driving the equilibrium to the right, forcing binding of the ion with the sorbent material. This experiment was not optimized due to the lack of allotted time while at PNNL but it does give a general idea about the materials uptake capacity (Figure 14). The actual data (blue diamonds) matches the predicted data (pink line) to a reasonable extent with the exception of actual uptake amount. Hgisotherm 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.001 mg Hg/L 1.000 2.000 • • ~- ~ t • • Senes1 f...------,~ -ill- isotherm( ·--·-Log.(Senes1) / 'l" 0.0 0.000 100.0 200.0 ~ 400.0 ~ ::l ~ ~ 300.0 500.0 600.0 700.0 Figure 14 Binding uptake isotherm for l:lloaded l,4-BMMB phenyl SAMMS with Hg2+ ion. Red curve is calculated binding, blue line is log plot and blue diamonds are actual data. This experiment was not run in duplicate. ------------- ---_._-- 126 Despite the obvious erroneous data points due to single pass testing, this data is in good agreement with data obtained for thio1 SAMMS as well as data in Figure 12 for 1: 1 loaded l,4-BMMB SAMMS at 1.6 molecules/nm2. The predicted binding (pink line, Figure 14) was found by using the Langmuir isotherm equation (Equation 1): C 1 C -=-+-Q K b Equation 1 Langmuir isotherm equation. where C is the equilibrium concentration of mercury (mg/L), Q is the mercury equilibrium loading on SAMMS (mg/g), K is the Langmuir adsorption constant (giL), and b is the maximum bound ion to the l,4-BMMB SAMMS material. A plot of Qn over Cn where n is a value for each solution yields the theoretical plot (pink, Figure 14). Lead Contaminant Leaching Studies From the data above, it is apparent that thiol-based sorbent materials have a high affinity for soft metal ions such as lead, mercury, cadmium and silver. Unfortunately, this can result in batch contamination if trace metals, specifically Pb2+, are present on glassware, solvents or plastic labware. After a number of skewed data sets, it was determined by ICP-MS that four batches of BMMB (both 1,3- and 1,4-) where contaminated with lead at some point in the experiment, either during the synthesis of the material or during testing. The base MCM-41 and phenyl ligands were used in numerous experiments with no apparent sign of Pb2+ contamination. To quantify the Pb2+ 127 contamination, a competitive binding experiment was devised using Hg2+ to displace the Pb2+ ions. This was carried out at pH 2, a range known to inhibit Pb2+ binding to thiol- based ligands. Samples were prepared as described above with LIS ratio of 5000 and an initial Hg2+ concentration of 4500 ppb to assure preferred binding. The result is a lead contamination range of around 1 mg/g (lead to sorbent) to nearly 8 mg/g (Figure 15). Pb Leaching at pH 2 8.00 7.00 6.00 E 5.00 .. -e o VI 01 (;; 4.00 a. 'D .. Q) -'E3.00 2.00 1.00 0.00 ~ r-- , I , - ...... I ~ I I I ~ I - I r- I ~ f----- - IP- - - f----- - _r-- r--'---- - - f----- - - ---; I 30331 303311 3033111 30331V 3033V 3033VI Figure 15 Quantification of Pb2+ conlnmination by means of competitive binding by Hg2+ Hlue bnrs nre 1,4-BMMB nnd red nre 1,3-BMMB. All tests performed at pH 2to help with lead displncement. All dalll collected in duplicate and averaged. From this data, it appears that lead contamination occurred fairly consistently between batches of 1,3- and 1,4-BMMB suggesting that contamination occurred during the addition of the active layer. This is a reasonable assumption due to the experimental 128 technique where batches of 1,3- and l,4-BMMB are prepared together, using various phenyl base layers. Although this material could not be used for multiple ion uptake studies nor those specific to lead, this material showed no inhibition to Hg2+ binding at low pH when compared to other BMMB-based sorbents. Conclusion A new and versatile material utilizing weak interactions to reversibly bind aromatic molecules containing reactive head groups capable of selective capture of toxic metal ions from aqueous matrices at levels equal to covalently bound analogs has been described herein. Ultimately, this study will demonstrate the feasibility ofloading this material with toxic ions, rinsing the bound material off the phenyl base support and reloading with more thiol-containing active layer to produce a 'regenerable' green material. Initial studies indicate that the Hg-thiol complexes can be rinsed from the support with organic solvents regenerating pristine Phenyl-SAMMS as qualitatively measured by ICP-MS. l,4-BMMB SAMMS (lOmg) was placed in filter cartridges typically used in solid support flow-through synthesis containing sub-micron polypropylene filter disks. The samples were wetted first with methanol followed by exposure to Hg2+ ion in filtered Columbia River water. Nanopure water was used to rinse the samples followed by a small methanol rinse and vacuum oven drying at 40°C overnight. Samples where then rinsed with approximately lmL of the following: methanol, isopropanol, chloroform, hexanes, pentane and toluene. The effluent was evaporated and the residue was dissolved 129 in pH 2 Nanopure water. ICP-MS analysis of the acidified solution showed that chloroform and hexanes both contained reasonable levels of mercury with hexanes having a greater detectable amount. It is important to understand that this experiment was merely qualitatively performed to demonstrate proof of concept and will be further studied by Sean Fontenot in the upcoming months. However, the ability to refresh or replace the surfaces of highly engineered sorbent support structures, such as mesoporous silica, could significantly increase the range of viable applications for these materials. This material, although effective in toxic metal ion uptake, does suffer from ligand leaching and requires modification of either the base or active layers to address this problem. Studies by Sean Fontenot of the Johnson Laboratory have shown a direct correlation between base layer loading and percentage of active layer leaching (Figure 16). In his study, MCM-41 was functionalized with the phenyl base layer ranging from 2.4 to 0.06 molecules per nm2• As expected, leaching of the active layer occurred at a much higher percentage with the sparsely populated surface, presumably due to a lack of stabilization of the active layer with 1C-contacts with the base layer. From this data, MCM-41 populated with a phenyl base layer of 1.3 molecules per nm 2 performed the best due to its high active layer uptake and low leaching levels. Sean is currently working on modifying the base layer to include perfluorobenzene, naphthalene and cyclohexane. To be of any utility as a sorbent material, leaching of the captured ions must be addressed. 130 Base laver loading comparison 0.7 C 0.6 Q/ .Ll ~ ll50 '"E b ft4 ~ Q/ U. O.~, 0 :£ 02 "'0 E 0.1E 0 70 60 "tJ SO Q/oJ: u '" 4ft .2!... Q/ :>- 30 Jl!Q/ ~ 20 uc.( .. <> 10 0 2.2 1.6 1.3 .72 O.:cq 0.06 Bas layer denslt, (mole(ules/nm~) Figlln~ 16 Leaching studies of active thiollayer compared 10 phenyl base layer density. Blue bars indicate actual active layer loading in mmol (left axis) while red bars indicate percentage of active layer leaching (right axis). Data collected from Ellman's test and quanti lied by UV-Vis. Experimental General Procedures Water sources where taken either from the Columbia River (Richland, WA), well water located on the Hanford site, (Richland WA) or Nanopure (18.2 mn) water produced in-house where noted. All natural water sources were passed through a 0.45 ~tm filter before use and stored in the dark. Glassware and plastic bottles were rinsed with a 10-20% nitric acid (low metal ICP-MS grade) solution followed by copious rinses with Nanopure water. All SAMMS materials were weighed out on a bench top analytical balance and reported to the nearest tenths of a mg. Digital pipettes (manual and electronic) were used to deliver small volumes of solutions (less than 100 131 mL). Large volumes were measured on a bench top analytical balance to the nearest tenth of a gram with the assumption of density equal to 1 for all solutions. Natural water samples were passed through a 0.2 flm filter by vacuum and stored in the dark until used. Test samples were filtered with 0.2 J.!m syringe filters before ICP-MS analysis. Samples were diluted as needed to range between 50 to 100 ppb to best match metal calibration curves. ICP-MS was performed on a Agilent Technologies 7500ce with calibration standards checked periodically to eliminate drift. All experiments were performed in duplicate or triplicate and values averaged before plotting unless otherwise noted. TGA was performed at the University of Oregon on a TA 2950 model starting at ambient temperature and ramping at 2°C a minute to 600°C under N2 purge gas environment set to 90% balance, 10% furnace. A Balzers Thermostar residual gas analyzer (Electron Impact Mass Spectrometer) was hooked directly to the exit port of the TA 2950 furnace via stainless steel sheath to limit capillary damage. Scans were set in analog mode to targeted specific masses and ran the entire length of the TGA experiment. Phenyl SAMMS MCM-4l (1.l0lg, 761.1 m2jg) was combined with toluene (100 mL) in a 250 mL round bottom flask and stirred under N2. To this 0.3757 mL Nanopure water was added (approximately 3 monolayers worth based off of surface area) and stirred for 4 hours or until the silica material re-dispersed in the toluene. 1.472 g ( 6.96 mmol, 1.2 mL) phenyl-trichlorosilane was added by syringe and the mixture stirred overnight (16 hours). Isopropyl alcohol (IPA, 100) mL was added to quench unreacted silane and stirred for 1 hour. The solution was filtered with a fine meshed sintered filter and rinsed with copious amounts of IPA then Acetone followed by drying in a vacuum 132 oven (-25 inches ofHg) at 40°C over night. Loading was determined gravimetrically using an analytical balance. Phenyl loading was adjusted as needed by altering the amount of silane. J,3-bis(mercaptomethyl) benzene (l,3-BMMB) Reprehensive experimental procedure applicable to 1,4-BMMB as well. 1,3-bis(chloromethyl) benzene (3.014g, 17.2 mmol) was dissolved in 70 mL of reagent grade acetone in a 250 mL round bottom flask while stirring. To this, thiourea (2.889g, 37.9 mmol) was added and the mixture heated to reflux under N2 for 3 hr. The mixture was allowed to cool. The white solid was filtered and rinsed with acetone followed by vacuum drying over night (3.77g, 67%). The thiouronium salt was placed back into a dry 250mL flask and to this, 50 mL of degassed 2M NaOH was added under a N2 atmosphere. The solution was heated to reflux for 1.5 hr and allowed to cool. Degassed 2M HCI was added to adjust the pH to ~2 followed by the addition of dichloromethane. The organic layer was separated from the aqueous and washed 2 x 50 mL Nanopure water and dried over Na2S04 for 15 minutes followed by filtration and evaporation in vacuo to produce a smelly clear liquid (1.694 g, 86%). BMMB SAMMS Typical loading scheme for 1,3- and 1,4-BMMB SAMMS involves dissolving the appropriate amount of ligand calculated from phenyl base layer coverage and target loading (l: 1 or 2: 1 base layer to active layer) 0.2 mL dichloromethane or chloroform in a 20 mL scintillation vial. To this, Phenyl SAMMS is added and the vial capped and mixed gently overnight on a platform shaker set to 1Hz. The cap is removed and the solvent is allowed to evaporate over a 24hr period. After 133 which, the material is placed in a vacuum over overnight at ambient temperature and -25 in Hg. Total active layer uptake was determined by gravimetric analysis with an analytical balance or total ligand burn-off by TGA. Leaching tests BMMB SAMMS material (l0 mg) was first wet with MeOH then soaked in 0.1 M phosphate buffer (pH 8) for 2hrs while gently agitating at 1 Hz with a platform shaker. The solution is then filtered through a 0.2 )Jm syringe filter. To this, 100 to 200 )JL of Ellman's reagent was added and the mixture shaken for 10-15 minutes. The solution was filtered with a 0.2 )Jm syringe filter and the absorbance was measured at 412 nm by an Agilent 8453 UV-Vis spectrometer. 134 APPENDIX CRYSTALLOGRAPHIC DATA HI: Monoclinic, P2(1)/c, a = 8.1026(11), b =12.9398(17), c =11.2969(16) A, P= 91.735(3t, V = 1183.9(3) A3, Z = 4, Dc = 1.444 g cm-3,.u = 0.265 mm- l , F(OOO) = 536, 20max =54.18° (-10 S h S 10, -16 S k S 17, -14 S 1S 14). Final residuals (163 parameters) R1 = 0.0614 for 1323 reflections with I > 20(1), and R1 = 0.1320, wR2 = 0.1827, GooF = 1.008 for all 2604 data. Residual electron density was 0.321 and -0.281 e.A3; [Ashel]: Triclinic, P-1, a =7.9782(7), b = 18.8417(16), C =20.3116(17) A, a = o 3 -397.856(2), P= 97.098(2), y = 91.123(2) 0, V = 2999.4(4) A ,Z = 4, Dc = 1.644 g cm ,.u = 1.669 mm- l , F(OOO) =1496, 20max =56.52° (-10 S h S 10, -24 S k S 24, -26 S 1S 26). Final residuals (757 parameters) R1 =0.0706 for 6238 reflections with I > 2a(1), and R1 =0.1780, wR2 = 0.1428, GooF =0.943 for all 13528 data. Residual electron density was 1.083 and -0.557 e.A-3; 135 [As1]]: Monoclinic, P2(1)/n, a = 14.8242(19), b = 10.6849(13), c = 24.910(3) A,!J = 103.239(4)°, V= 3840.8(8) A3, Z= 4, Dc = 1.570 g cm-3, Ji = 1.115 mm- I , F(OOO) = 1856, 2emax = 49.42° (-17 s.h S.17, -12 S. kS.12, -29 s.f S.29). Final residuals (508 parameters) Rl = 0.0931 for 2108 reflections with I> 2(J(I), and Rl = 0.2846, wR2 = 0.2456, GooF = 0.941 for a116518 data. Residual electron density was 0.458 and -0.490 e.A"3. 136 BIBLIOGRAPHY Chapter I (1) Desiraju, G. R. Nature 2001, 412, 397-400. (2) Spessard, G. 0.; Miessler, G. L. Organometallic Chemistry; Prentice-Hall Inc.: Upper Saddle River, 2000. (3) Piguet, C.; Edder, C.; Rigault, S.; Bernardinelli, G.; Bunzli, 1. G.; Hopfgartner, G. J Chem. Soc. I Dalton Trans. 2000,3999-4006. (4) Holliday, B. J.; Mirkin, C. A. Angew. Chem. 2001,40,2022-2043. (5) Orr, G. W.; Barbour, L. 1.; Atwood, 1. L. Science 1999, 285, 1049-1052. (6) Sun, X.; Johnson, D. W.; Raymond, K. N.; Wong, E. H.lnorg. Chern. 2001,40, 4504-4506. (7) Garcia-Zarracino, R.; Hopfl, H. JAm. Chem. Soc. 2005,127,3120-3130. (8) Kieltyka, R.; Englebienne, P.; Fakhoury, 1.; Autexier, C.; Moitessier, N.; Sleiman, H. F. JAm. Chem. Soc. 2008,130, 10040-10041. (9) Lehn, J.-M. Proc. Natl. Acad. Sci. U S. A. 2002,99,4763-4768. (10) Vignon, S. A.; Jarrosson, T.; Iijima, T.; Tseng, H.-R.; Sanders, J. K. M.; Stoddart, J. F. JAm. Chem. Soc. 2004,126,9884-9885. (11) Caulder, D. L.; Raymond, K. N. J Chem. Soc., Dalton Trans. 1999,1185-1200. (12) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N. Ace. Chem. Res. 2005,38, 349-358. (13) Hamilton, T. K.; MacGillivray, L. R. Cryst. Growth Des. 2004,4,419-430. 137 (14) Truner, D. R.; Pastor, A; Alajarin, M.; Steed, 1. W. Struct. Bonding 2004, 108, 97-168. (15) Hof, F.; Craig, S. L.; Nuckolls, C.; Rebek, J. Angew. Chem. Int. Ed. 2002,41, 1488-1508. (16) Radhakrishnan, D.; Stang, P. 1. J Org. Chem. 2003,68,9209-9213. (17) Paver, M. A.; Joy, 1. S.; Hursthouse, M. B. Chem. Commun. 2002, 2150-2151. (18) Garcia, A. M.; Romero-Salguero, F. 1.; Bassani, D. M.; Lehn, 1.-M.; Baum, G.; Fenske, D. Chem. Eur. J 1999,5, 1803-1808. (19) Marcovich, D.; Duesler, E. N.; Tapscott, R. E.; Them, T. F.Inorg. Chem. 1982, 21,3336-3341. (20) Vickaryous, W. J.; Herges, R.; Johnson, D. W. Angew. Chem., Int. Ed. 2004,43, 5831-5833. (21) Vickaryous, W. 1.; Rather Healey, E.; Berryman, O. B.; Johnson, D. W.Inorg. Chem. 2005,44,9247-9252. (22) Kerr, P. G.; Leung, P. H.; Wild, S. B. JAm. Chem. Soc. 1987,109,4321-4328. (23) Shaikh, T. A.; Bakus, R. C.; Parkin, S.; Atwood, D. A J Organ0 met. Chem. 2006,691,1825-1833. (24) Shaikh, T. A; Parkin, S.; Atwood, D. A J Organomet. Chem. 2006,691,4167- 4171. (25) Cangelosi, V. M.; Carter, T. G.; Zakharov, L. N.; Johnson, D. W. Chem. Commun. 2009, 5606-5608. (26) Cangelosi, V. M.; Sather, A. C.; Zakharov, L. N.; Berryman, O. B.; Johnson, D. W.Inorg. Chem. 2007,46,9278-9284. (27) Cangelosi, V. M.; Zakharov, L. N.; Fontenot, S. A.; Pitt, M. A.; Johnson, D. W. Dalton Trans. 2008,3447-3453. (28) Alcock, N. W. In Advances in Inorganic Chemistry and Radiochemistry; Emeleus, H. 1., Sharpe, A G., Eds.; Academic Press, Inc.: New York, 1972; Vol. 15, pI-58. 138 (29) Lindquist, N. R.; Carter, T. G.; Cangelosi, V. M.; Zakharov, 1. N.; Johnson, D. W. Accepted, 2010. (30) Pitt, M. A.; Johnson, D. W. Chem. Soc. Rev. 2007. (31) Aposhian, H. V.; Aposhian, M. M. Chem. Res. Toxicol. 2005, 18, 1287-1295. (32) Magalhaes, M. C. F. Pure Appl. Chem. 2002, 74, 1843-11850. (33) Hindmarsh, J. T.; Abernathy, C. 0.; Peters, G. R.; McCurdy, R F. In Heavy Metals in the Environment; Sarkar, B., Ed.; Marcel Dekker, Inc.: New York, 2002, p 217-229. (34) Bhattacharya, P.; Jacks, G.; Frisbie, S. H.; Smith, E.; Naidu, R.; Sarkar, B. Arsenic in the Environment: A Global Perspective; Marcel Dekker, Ink: New York,2002. (35) Roy, P.; Saha, A. Curro Sci. 2002,82,38-45. (36) Hossain, M. F. Agriculture Ecosystems & Environment 2006,113,1-16. (37) MandaI, B. K.; Suzuki, K. T. Talanta 2002,58,201-235. (38) Frisbie, S. H.; Ortega, R; Manynard, D. M.; Sarkar, B. Environ. Health Perspect. 2002,110, 1147-1153. (39) Hinkle, S. R.; Polette, D. J. Arsenic in Ground Water o/the Willamette Basin, Oregon, USGS, 1999. (40) Solis, A. R.; Mukopadhyay, R.; Rosen, B. P.; Stemmler, T. 1. Inorg. Chem. 2004,43,2954-2959. (41) Cullen, W. R.; Reimer, K. 1. Chem. Rev. 1989,89, 713-764. (42) Silver, S.; Phung, 1. T. Appl. Environ. Microbiol. 2005, 71, 599-608. (43) Adams, S. R; Campbell, R E.; Gross, 1. A.; Martin, B. R.; Walkup, G. K.; Yao, Y.; Llopis, J.; Tsien, R. Y. JAm. Chem. Soc. 2002, 124, 6063-6076. (44) Cruse, B. W. T.; James, M. N. G. Acta Crystallographica 1972, B28, 1325- 1331. (45) Farrer, B. T.; McClure, C. P.; Penner-Hahn, J. E.; Pecoraro, V. 1. Inorg. Chem. 2000,39,5422-5423. 139 (46) Pappalardo, G. c.; Chakravorty, R.; Irgolic, K. 1.; Meyers, E. A. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 1983, C39, 1618-1620. (47) Raston, C. L.; White, A. H. Journal ofthe Chemical Society-Dalton Transactions 1975, 2425-2429. (48) Garje, S. S.; Jain, V. K. Coord. Chem. Rev. 2003,236,35-56. (49) Kuznik, N.; Wendt, O. F. J. Chem. Soc., Dalton Trans. 2002, 15,3074-3078. (50) Alcock, N. W. Secondary Bonding to Nonmetallic Elements; Academic Press, Inc.: New York, 1972; Vol. 15. (51) Alcock, N. W. Bonding and Structure: structural principles in inorganic and organic chemistry; Ellis Horwood: New York, 1990. (52) Alkorta, 1.; Rozas, 1.; Elguero, 1. JAm. Chem. Soc. 2002, 124, 8593-8598. (53) Mascal, M.; Armstrong, A.; Bartberger, M. D. JAm. Chem. Soc. 2002,124, 6274-6276. (54) Quifionero, D.; Garau, C.; Rotger, c.; Frontera, A.; Ballester, P.; Costa, A.; Deya, P. M.Angew. Chem. Int. Ed. 2002,41,3389-3392. (55) Hof, F.; Rebek Jr., 1. Proc. Nat!. Acad. Sci. U S. A. 2002,99,4775-4777. (56) Tani, K.; Hanabusa, S.; Kato, S.; Mutoh, S.; Suzuki, S.; Ishida, M. J Chem. Soc., Dalton Trans. 2001,518-527. (57) Probst, T.; Steigelmann, 0.; Riede, 1.; Schmidbaur, H. Chem. Ber. 1991, 124, 1089-93. (58) Ma,1. c.; Dougherty, D. A. Chemical Reviews (Washington, D. C.) 1997,97, 1303-1324. (59) Berryman, O. B.; Bryantsev, V. S.; Stay, D. P.; Johnson, D. W.; Hay, B. P. J Am. Chem. Soc. 2007,129,48-58. (60) Carmalt, C. 1.; Norman, N. C. In Chemistry ofarsenic, antimony, and bismuth; Norman, N. C., Ed.; Blackie Academic & Professional: London, 1998, p 1-38. (61) Schmidbaur, H.; Nowak, R.; Steigelmann, 0.; Muller, G. Chem. Ber. 1990,123, 1221-1226. 140 (62) Camerman, A.; Trotter, J. J Chem. Soc. 1965, 730-8. (63) Kersting, 8.; Meyer, M.; Powers, R E.; Raymond, K. N. JAm. Chem. Soc. 1996,118,7221-7222. (64) Pitt, M. A.; Cangelosi, V. M.; Fontenot, S. A; Sather, A c.; Zakharov, 1. N.; Johnson, D. W. In Prep.. (65) Bondi, A. J Phys. Chem. 1964,68,441-451. (66) Shannon, R. D. Acta Crystallographica, Section A: Crystal Physics, Diffraction, Theoretical and General Crystallography 1976, A32, 751-67. (67) Pauling, 1. C. The Nature ofthe Chemical Bond and the Structure ofMolecules and Crystals. An Introduction to Modern Structural Chemistry. 3rd ed, 1960. (68) Vickaryous, W. J.; Zakharov, 1. N.; Johnson, D. W. Main Group Chem. 2006, 5,51-59. (69) Johnson, D. W.; Pitt, M. A; Harris, J. M.; USA: 2006; Vol. 111350,202. (70) Addleman, R. S.; Bayes, 1 T; Carter, T G.; Fontenot, S. A; Fryxell, G. E.; Johnson, D. W., 2009 U.S. Pat. Appl. # 611120,321. Chapter II (1) Carter, T. G.; Healey, E. R; Pitt, M. A.; Johnson, D. W. Inorg. Chem. 2005,44, 9634-9636. (2) Vickaryous, W. 1; Herges, R.; Johnson, D. W. Angew. Chem.) Int. Ed. Eng!. 2004,43,5831-5833. (3) Vickaryous, W. J.; Healey, E. R; Berryman, O. 8.; Johnson, D. W. Inorg. Chem. 2005,44,9247-9252. (4) Cangelosi, V. M.; Zakharov, 1. N.; Fontenot, S. A.; Pitt, M. A.; Johnson, D. W. Dalton Trans. 2008,3447-3453. (5) Cangelosi, V. M.; Carter, T. G.; Zakharov, 1. N.; Johnson, D. W. Chem. Commun. 2009, 5606-5608. (6) Cangelosi, V. M.; Sather, A c.; Zakharov, 1. N.; Berryman, O. B.; Johnson, D. W.Inorg. Chem. 2007,46,9278-9284. 141 (7) Carter, T. G.; Vickaryous, W. J.; Cangelosi, V. M.; Johnson, D. W. Comments Inorg. Chem. 2007,28,97-122. (8) Dhubhghaill, O. M. N.; Saddler, P. J. Struct. Bonding (Berlin) 1991, 78, 129. (9) Farrer, B. T.; McClure, C. P.; Penner-Hahn, J. E.; Pecoraro, V. L. Inorg. Chem. 2000,39,5422-5423. (10) Johnson, J. M.; Voegtlin, C. J Bio!. Chem. 1930,89,5422. (11) Matzapetakis, M.; Farrer, B. T.; Weng, T. C.; Hemmingsen, L.; Penner-Hahn, J. E.; Pecoraro, V. L. JAm. Chem. Soc. 2002,124,8042-8054. (12) Allen, F. H.; Kennard, O. Chem. Des. Autom. News 1993, 8. (13) Alcock, N. W. In Advances in Inorganic Chemistry and Radiochemistry; Emeleus, H. 1., Sharpe, A. G., Eds.; Academic Press, Inc.: New York, 1972; Vol. 15, pI-58. (14) Alcock, N. W. Bonding and Structure: structural principles in inorganic and organic chemistry; Ellis Horwood: New York, 1990. (15) Bondi, A. J Phys. Chem. 1964,68,441-451. (16) Srivastava, P. C. Phosphorus, Sulfur Silicon Relat. Elem. 2005, 180, 969-983. (17) Hof, F.; Rebek, J., Jr. Proceedings ofthe National Academy ofSciences ofthe United States ofAmerica 2002,99,4775-4777. (18) Lehn, J.-M. Proc. Nat!' Acad. Sci. Us. A. 2002,99,4763-4768. (19) Tani, K.; Hanabusa, S.; Kato, S.; Mutoh, S.; Suzuki, S.; Ishida, M. J Chern. Soc., Dalton Trans. 2001,518-527. (20) Cozzolino, A. F.; Vargas-Baca, I.; Mansour, S.; Mahmoudkhani, A. H. JAm. Chem. Soc. 2005, 127, 3184-3190. (21) Minkin, V. I.; Minyaev, R. M.; Milov, A. A.; Gribanova, T. N. Russ. Chern. Bull. 2001, 50, 2028~2045. (22) Jankowski, Z.; Stolarski, R. Po!' J Chem. 1981,55,555-563. 142 (23) Vickaryous, W. J.; Herges, R.; Johnson, D. W. Angewandte Chemie 2004,43, 5831-5833. (24) Vickaryous, W. J.; Rather Healey, E.; Berryman, O. B.; Johnson, D. W. Inorganic Chemistry 2005,44,9247-9252. (25) Schmidbaur, H.; Nowak, R.; Steigelmann, 0.; Muller, G. Chem. Ber. 1990,123, 1221-1226. (26) Rao, C. N. R. Ultra-violet and Visible Spectroscopy; 3 ed.; Butterworth & Co. Ltd.: London, 1975. (27) Perkampus, H. H. In UV- Vis Spectroscopy and Its Applications; Springer- Verlag: New York, 1992, p 149-158. (28) WOOhner, F. Chem. Commun. 2004, 1564-1579. (29) Rademacher, A; Markle, S.; Langha1s, H. L. Chem. Ber. 1982,115,2927-2934. (30) Pitt, M. A; Zakharov, L. N.; Vanka, K.; Thompson, W. H.; Laird, B. B.; Johnson, D. W. Chem. Commun. 2008,3936-3938. (31) Inamoto, K.; Arai, Y; Hiroya, K.; Doi, T. Chem. Commun. (Cambridge, U. K) 2008, 5529-5531. (32) Banks, C. K.; Morgan, J. F.; Clark, R. L.; Hatle1id, E. B.; Kahler, F. H.; paxton, H. W.; Cragoe, E. J.; Andres, R. 1.; E1pem, B.; coles, R. F.; Lawhead, 1.; Hamilton, C. S. J Am. Chern. Soc. 1947,69,927-930. (33) Van Der Ke1en, G. P. Bull. Soc. Chim Belg. 1958,65,343-349. (34) Dyke, W. J. C.; Jones, W. J. J Chem. Soc. 1930,2426-2430. (35) Quick, A J.; Adams, R. JAm. Chem. Soc. 1922,44,805-816. (36) Claeys, E. G. J Organornet. Chern. 1966,5,446-453. (37) Cline, D. J.; Thorpe, C.; Schneider, J. P. J Am. Chern. Soc. 2003,125,2923- 2929. 143 Chapter III (1) Nordberg, G. F.; Sandstrom, B.; Becking, G.; Goyer, R. A. In Heavy Metals in the Environment; Sarkar, B., Ed.; Marcel Dekker, Inc: New York, 2002. (2) Shannon, M. A.; Bohn, P. W.; Elimelech, M.; Georgiadis, J. G.; Marinas, B. 1.; Mayes, A. M. Nature 2008,452,301-310. (3) Briggs, D. British Medical Bulletin 2003, 68, 1-24. (4) lamp, L. Occupational and Environmental Medicine 2003,60,461-462. (5) Wildung, R. E.; Cataldo, D. A. Environ. Health Perspect. 1978,27,149-159. (6) Jamp, L. British Medical Bulletin 2003,68,167-182. (7) Smith, A. H.; Lingas, E. 0.; Rahman, M. Bulletin o/the World Health Organization 2000,78, 1093-1103. (8) Dermont, G.; Bergeron, M.; Mercier, G.; Richer-Lafleche, M. J Hazard. Mater. 2008, 152, 1-3 1. (9) Feng, X; Fryxell, G. E.; Wang, L.-Q.; Kim, A. Y; Liu, J.; Kemner, K. M. Science 1997,276,923-926. (10) Fryxell, G., E.; Wu, H.; Lin, Y; Shaw, W. 1.; Birnbaum, J. C.; Linehan, J. C.; Nie, Z.; Kemner, K.; Kelly, S. J Mater. Chem. 2004,14,3356-3363. (11) Fryxell, G. E.; Liu, J.; Hauser, T. A.; Nie, Z. M.; Ferris, K. F.; Mattigod, S.; Gong, M. L.; Hallen, R. T. Chem. Mater. 1999,11,2148-2154. (12) Fryxell, G. E.1norg. Chem. Commun. 2006,9,1141-1150. (13) Beck, J. S.; Vartuli, J. C.; Roth, W. 1.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. JAm. Chem. Soc. 1992,114,10834-10842. (14) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, 1. C.; Beck, 1. S. Nature 1992,359,710-712. (15) Han, A.; Qiao, Y. J Phys. D: Appl. Phys. 2007,40,5743-5746. 144 (16) Monnier, A; SchUth, F.; Huo, Q.; Kumar, D.; Margolese, D.; Maxwell, R. S.; Stucky, G. D.; Krishnamurty, M.; Petroff, P.; Firouzi, A; Janicke, M.; Chmelka, B. F. Science 1993,261, 1299-1303. (17) Gao, W.; Reven, L. Langmuir 1995,11,1860-1863. (18) Wasserman, S. R.; Whitesides, G. M.; Tidswell, 1. M.; Ocko, B. M.; Pershan, P. S.; Axe, 1. D. JAm. Chem. Soc. 1989, 111,5852-5861. (19) Linehan, J. c.; Stiff, C. M.; Fryxell, G. E.Inorg. Chem. Commun. 2006,9,239- 241. (20) Mattigod, S. V.; Feng, x.; Fryxell, G., E. Sep. Sci. Techno!. 1999,34,2329- 2345. (21) Lam, K F.; Yeung, K L.; McKay, G. Langmuir 2006,22,9632-9641. (22) Deratani, A; Sebille, B. Ana!. Chem. 1981,53, 1742-1746. (23) Richens, D. T. The Chemistry ofAqua Ions; John Wiley & Sons Ltd: West Sussex, 1997. (24) Leckie, J. 0.; James, R. O. In Aqueous-Environmental Chemistry ofMetals; Rubin, A J., Ed.; Ann Arbor Science Publishers Inc.: Ann Arbor, 1974, p 1-76. (25) Helm, L.; Merbach, A E. Coord. Chem. Rev. 1999, 187, 151-181. (26) Dean,1. A Lange's HandvookofChemistry; 15 ed.; McGraw-Hill, Inc: New York, 1999. (27) Jacobson, A. R.; Klitzke, S.; McBride, M. B.; Baveye, P.; Steenhuis, T. S. Water Air and Soil Pollution 2005, 160,41-54. (28) Pearson, R. G. JAm. Chem. Soc 1963, 85, 3533-3539. (29) Yantasee, W.; Warner, C. L.; Sangvanich, T.; Addleman, R. S.; Carter, T. G.; Wiacek, R. J.; Fryxell, G. E.; Timchalk, C.; Warner, M. G. Environ. Sci. Techno!. 2007,41,5114-5119. (30) Yantasee, W.; Hongsirikarn, K; Warner, C. L.; Choi, D.; Sangvanich, T.; Toloczko, M. B.; Warner, M. G.; FryxelI, G. E.; Addleman, R. S.; Timchalk, C. Analyst 2008, 133, 348-355. 145 (31) Koehler, F. M.; Rossier, M.; Waelle, M.; Athanassiou, E. K.; Limbach, L. K.; Grass, R. N.; Gunther, D.; Stark, W. 1. Chem. Commun. 2009,4862-4864. (32) Liu, J. F.; Zhao, Z. S.; Jiang, G. B. Environmental Science & Technology 2008, 42,6949-6954. (33) Hancock, R. D. J Chem. Educ. 1992,69,615-621. (34) Steed, J. W.; Atwood, J. L. Supramolecular Chemistry; 2 ed.; John Wiley and Sons, Ltd, 2009. (35) Cohen, S. M.; Raymond, K. N. Inorg. Chem. 2000,39,3624-3631. (36) Fryxell, G. E.; Lin, Y; Fiskum, S.; Birnbaum, 1. c.; Wu, H. Environ. Sci. Technol. 2005,39, 1324-1331. (37) Sun, S. H.; Zeng, H. JAm. Chem. Soc. 2002, 124,8204-8205. Chapter IV (1) Yokoi, T.; Tatsumi, T.; Yoshitake, H. J Colloid Interface Sci. 2004,274,451- 457. (2) De Carlo, E. H.; Thomas, D. M. Environmental Science & Technology 1985,19, 538-544. (3) Roundhill, D. M.; Koch, H. F. Chem. Soc. Rev. 2002, 31, 60~67. (4) Fox, K. R.; Sorg, T. J. Journal American Water Works Association 1987, 79,81- 84. (5) Helfferich, F. Ion exchange; Dover: Mineola, 1995. (6) Wang, J. L.; Chen, C. Biotechnol. Adv. 2009,27, 195-226. (7) Oliver, S. R. J. Chem. Soc. Rev. 2009,38, 1868-1881. (8) Goh, K. H.; Lim, T. T.; Dong, Z. Water Res. 2008,42, 1343-1368. (9) Atia, A. A. J Hazard. Mater. 2006, 137, 1049-1055. (10) Fryxell, G. E.; Liu, J.; Hauser, T. A; Nie, Z. M.; Ferris, K. F.; Mattigod, S.; Gong, M. L.; Hallen, R. T. Chem. Mater. 1999,11,2148-2154. 146 (11) Yoshitake, H.; Yokoi, T; Tatsumi, T Chern. Mater. 2003,15,1713-1721. (12) Wasserman, S. R.; Whitesides, G. M.; Tidswell, 1. M.; Ocko, B. M.; Pershan, P. S.; Axe, J. D. J Am. Chern. Soc. 1989, 111, 5852-5861. (13) Feng, X.; Fryxell, G. E.; Wang, L.-Q.; Kim, A Y; Liu, 1.; Kemner, K. M. Science 1997, 276, 923-926. (14) Sillen, L. G.; Martell, A E.; Bjerrum, 1. Stability constants ofmetal-ion complexes; Chemical Society: London, 1964. (15) Richens, D. T The Chemistry ofAqua Ions; John Wiley & Sons Ltd: West Sussex, 1997. (16) Godfrey, S. M.; McAuliffe, C. A.; Mackie, A. G.; Pritchard, R. G. In Chemistry ofArsenic, Antimony, and Bismuth; Norman, N. C., Ed.; Blackie Academic & Professional: London, 1998, p 94. (17) Raposo, J. C.; Olazabal, M. A; Madariaga, J. M. J Solution Chern. 2006,35, 79-94. (18) Richard, F.; Bourg, A. C. M. Water Res. 1991,25,807-816. (19) Palmer, C. D.; PuIs, R. W.; EPA, Ed.; Technology Innovation Office: Ada, 1994. Chapter V (1) Shannon, M. A; Bohn, P. W.; Elimelech, M.; Georgiadis, J. G.; Marinas, B. J.; Mayes, A. M. Nature 2008, 452,301-310. (2) Hightower, M.; Pierce, S. A. Nature 2008,452, 285-286. (3) Derrnont, G.; Bergeron, M.; Mercier, G.; Richer-Lafleche, M. J Hazard. Mater. 2008,152,1-31. (4) Smith, A. H.; Lingas, E. 0.; Rahman, M. Bull. World Health Organ. 2000, 78, 1093-1103. (5) Pitt, M. A; Johnson, D. W. Chem. Soc. Rev. 2007,36, 1441-1453. (6) Carter, T. G.; Vickaryous, W. J.; Cangelosi, V. M.; Johnson, D. W. Comments Inorg. Chern. 2007,28,97-122. 147 (7) Yantasee, W.; Hongsirikarn, K.; Warner, C. 1.; Choi, D.; Sangvanich, T.; Toloczko, M. B.; Warner, M. G.; Fryxell, G. E.; Addleman, R. S.; Timchalk, C. Analyst 2008, 133, 348-355. (8) Yantasee, W.; Warner, C. 1.; Sangvanich, T.; Addleman, R. S.; Carter, T. G.; Wiacek, R. J.; Fryxell, G. E.; Timchalk, c.; Warner, M. G. Environ. Sci. Technol. 2007,41,5114-5119. (9) Fryxell, G. E.; Mattigod, S. V.; Lin, Y. H.; Wu, H.; Fiskum, S.; Parker, K.; Zheng, F.; Yantasee, W.; Zemanian, T. S.; Addleman, R. S.; Liu, J.; Kemner, K.; Kelly, S.; Feng, X D. J Mater. Chem. 2007, 17, 2863-2874. (10) Fryxell, G. E.; Liu, J.; Hauser, T. A; Nie, Z.; Ferris, K. F.; Mattigod, S.; Gong, M.; Hallen, R. T. Chem. Mater. 1999,11,2148-2154. (11) Feng, X; Fryxell, G. E.; Wang, L.-Q.; Kim, A Y; Liu, 1.; Kemner, K. M. Science 1997,276,923-926. (12) Fryxell, G. E.; Lin, Y; Fiskum, S.; Birnbaum, J. C.; Wu, H. Environ. Sci. Techno!. 2005,39,1324-1331. (13) Kresge, C. T.; Leonowicz, M. E.; Roth, W. 1.; Vartuli, J. C.; Beck, J. S. Nature 1992,359,710-712. (14) Beck, J. S.; Vartuli, J. C.; Roth, W. 1.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. 1. JAm. Chem. Soc 1992, 114, 10834-10842. (15) Yoshitake, H.; Yokoi, T.; Tatsumi, T. Chem. Mater. 2002,14,4603-4610. (16) Chang, S.-C.; Chao, 1.; Tao, Y-T. JAm. Chem. Soc 1994, 116, 6792-6805. (17) Hunter, C. A.; Sanders, J. K. M. JAm. Chem. Soc 1990, 112, 5525-5534. (18) Feng, X; Liu, 1.; Fryxell, G. E.; Gong, M.; Wang, L.-Q.; Chen, X; Kurath, D. E.; Ghormley, C. S.; Klasson, K. T.; Kemner, K. M. SelfAssembled Mercaptan on Mesoporous Silica (SAMMS) Technology for Mercury Removal and Stabilization, Pacific Northwest National Laboratory, 1997.