MAIN GROUP SUPRAMOLECULAR COORDINATION CHEMISTRY: DESIGN STRATEGIES AND DYNAMIC ASSEMBLIES by MELANIE A. PITT A DISSERTATION Presented to the Department of Chemistry and the Graduate School ofthe University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy June 2009 11 University of Oregon Graduate School Confirmation of Approval and Acceptance of Dissertation prepared by: Melanie Pitt Title: "Main Group Supramolecular Coordination Chemistry: Design Strategies and Dynamic Assemblies" 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: Kenneth Doxsee, Chairperson, Chemistry Darren Johnson, Advisor, Chemistry David Tyler, Member, Chemistry Victoria DeRose, Member, Chemistry Stephen Remington, Outside Member, Physics and Richard Linton, Vice President for Research and Graduate Studies/Dean ofthe Graduate School for the University of Oregon. June 13, 2009 Original approval signatures are on file with the Graduate School and the University of Oregon Libraries. © 2009 Melanie A. Pitt iii iv An Abstract ofthe Dissertation of Melanie A. Pitt in the Department of Chemistry for the degree of to be taken Doctor of Philosophy June 2009 Title: MAIN GROUP SUPRAMOLECULAR COORDINATION CHEMISTRY: DESIGN STRATEGIES AND DYNAMIC ASSEMBLIES Approved: _ Darren W. Johnson Main group supramolecular chemistry is a rapidly expanding field that combines the tools of coordination chemistry with the unusual and frequently unexpected coordination preferences exhibited by the main group elements. Application of established supramolecular design principles to those elements provides access to novel structure types and the possibility of new functionality introduced by the rich chemistry of the main group. Chapter I is a general review ofthe field of main group supramolecular chemistry, focusing in particular on the aspects of coordination chemistry and rational design strategies that have been thus far used to prepare polynuclear "metal"-ligand assemblies. Chapter II is a discussion of work toward supramolecular assemblies based on the coordination preferences oflead(II), in vparticular focusing on the 2-mercaptoacetamide and arylthiolate functionalities to target four-coordinate and three-coordinate geometries, respectively. Several possible avenues for further pursuing this research are suggested, with designs for ligands that may provide a more fruitful approach to the coordination oflead(II). Chapter III deals with the preparation of AszL3 assemblies based on flexible ligand scaffolds. These assemblies exhibit structural changes in response to temperature and solvent, which may provide some insight into the subtle shape requirements involved in supramolecular guest binding. Chapter IV continues this work with an examination of how ligand structure affects mechanical coupling of stereochemistry between metal centers when the chelate ring is completed by a secondary bonding interaction such as the As-n contact. Finally, Chapter V presents a crystallographic and synthetic study of the nature of the interaction between pnictogens and arene rings. This interaction is ubiquitous in the coordination chemistry performed in the Johnson laboratory; understanding the role these interactions play in determining the final structure of supramolecular assemblies is vital to the preparation of more complex structures. Chapter VI presents a set of conclusions and outlook for future work on lead(II) supramolecular assemblies and the dynamic assemblies prepared from flexible organic scaffolds. This dissertation contains previously published and coauthored material. CURRICULUM VITAE NAME OF AUTHOR: Melanie A. Pitt PLACE OF BIRTH: Indianapolis, Indiana DATE OF BIRTH: April 1980 GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene Indiana University, Bloomington DEGREES AWARDED: Doctor of Philosophy, Chemistry, 2009, University of Oregon Master of Science, Chemistry, 2005, University of Oregon Bachelor of Science, Chemistry, 2003, Indiana University AREAS OF SPECIAL INTEREST: Main group coordination chemistry Supramolecular chemistry PROFESSIONAL EXPERIENCE: Graduate Research Assistant, University of Oregon, 2003-2008 Graduate Teaching Fellow, University of Oregon, 2003-2007 Graduate Research Fellow, Pacific Northwest National Laboratory, 2004 GRANTS, AWARDS AND HONORS: National Science Foundation IGERT Fellowship, 2004-2007 Pacific Northwest National Laboratory Doctoral Fellowship, 2004 vi vii PUBLICATIONS: Pitt, Melanie A, Zakharov, Lev N., Thompson, Ward H., Vanka Kumar, Laird, Brian B., and Johnson, Darren W., "Multiple Weak Supramolecular Interactions Stabilize A Surprisingly "Twisted" As2L 3 Assembly," Chemical Communications, 2008, 1O.1039/b806958a. Cangelosi, Virginia C., Fontenot, Sean F., Pitt, Melanie A., and Johnson, Darren W. "Host-Guest Interactions In A Series Of Self-Assembled As2 L 2 C12 Macrocyc1es," Journal o/the Chemical Society, Dalton Transactions. 2008, 3447· Pitt, Melanie A and Johnson, Darren W. "Main Group Supramolecular Chemistry," Chemical Society Reviews, 2007,36, 1441. Johnson, Darren W., Pitt, Melanie A., and Harris, James M., "Adsorbent With Multiple Layers," US Patent Application 2°°7°181502. Carter, Timothy G., Rather-Healey, Elizabeth, Pitt, Melanie A and Johnson, Darren W., "Secondary Bonding Interactions Observed In Two Arsenic Thiolate Complexes," Inorganic Chemistry, 2005 viii ACKNOWLEDGMENTS My sincere thanks go out to my advisor, Darren W. Johnson, for the past several years of interesting chemistry, insightful discussions, and barbecue-related debates. I also thank my committee chair, Professor Kenneth N. Doxsee, as well as my other committee members, Professors David R. Tyler, Victoria J. DeRose, and James R. Remington for insight and guidance. The funding and equipment provided by the University of Oregon is gratefully acknowledged, as is financial support from the National Science Foundation's IGERT fellowship. The other members of the Johnson laboratory have provided me with what I can only describe as priceless. In addition to their scientific and intellectual contributions and despite the inevitable ups and downs that come with research, there is very real friendship and camaraderie in the group. There have been too many past and present members to thank each individually, but I must single out Tim Carter, Zack Mensinger, and Dr. Randall Hicks for their support and assistance at a particularly trying time in my life. It will not be forgotten. Finally, I thank those on the outside: my friends, my family, especially my parents Jane and Randy Pitt, and the Wolfe family for welcoming me as one of their own. Especially Michelle. Most of alL however, I thank my wonderful partner Christopher Wolfe for more love, support, and friendship than any person could ever hope for. For everyone who helped me on my way here. ix Chapter TABLE OF CONTENTS Page x I. INTRODUCTION: MAIN GROUP SUPRAMOLECULAR CHEMISTRY 1 Introduction 1 Scope of Review 2 Coordination Geometries of Main Group Elements 3 Group 12 - Zinc, Cadmium, and Mercury 5 Group 13 - Aluminum and Gallium 12 Group 14 - Germanium, Tin, and Lead 19 Group 15 - Arsenic, Antimony, and Bismuth 27 Groups 16 and 17 - Chalcogens and Halides 33 Host-Guest Chemistry of Supramolecular Main Group Complexes. 35 Conclusions 40 II. LIGAND DESIGN FOR MAIN GROUP SUPRAMOLECULAR ASSEMBLIES 42 Health and Environmental Concerns Associated with the Main Group Elements 42 Supramolecular Design Principles and Application to Main Group Chemistry 44 Secondary Bonding Interactions ..46 Ligand Design for Pb(II) and Hg(II) ..47 Design and Synthesis of 2-Mercaptoacetamide Ligands .50 Chapter xi Page Trigonal Pyramidal Coordination of Pb(II) 54 Future Directions for Research in Supramolecular Lead(II) Chemistry 60 Experimental Details 65 General Information 65 Synthesis of 2-Mercaptoacetamides and Derivatives 65 Synthesis of Thiophenolate Complexes of Pb(II) 69 Synthesis of Bifunctional Thiophenol Derivatives 70 Crystallographic Details 73 III. MULTIPLE WEAK SUPRAMOLECULAR INTERACTIONS STABILIZE SURPRISINGLY TWISTED ASzL3 ASSEMBLIES: SOLID STATE DISTORTION AND TEMPERATURE SENSITIVE CONFORMATIONAL CHANGES IN SOLUTION 74 Introduction 74 Diphenylmethane in Supramolecular Chemistry 76 Synthesis of an AszL3 Assembly Bridged by Diphenylmethane 79 DFT Calculations Support Distorted Structure 83 AszLz3and Solution Stability 84 Conformational Equilibrium Based on Intramolecular CH."rr Interactions 85 Edge-to-Face Aromatic Interactions 87 Open-Closed Equilibrium Based on Edge-to-Face Interactions 90 Chapter xii Page Summary and Conclusions 92 Experimental Details 94 General Information 94 Ligand Syntheses 94 Synthesis ofAszL3 Assemblies 96 Crystallographic Data 97 2-Dimensional NMR Spectra 98 Variable Temperature NMR Spectra 100 Thermodynamic Parameters 101 DFT Calculations 106 IV. ARSENIC-TI INTERACTIONS INFLUENCE MECHANICAL COUPLING IN A SERIES OF SUPRAMOLECULARASSEMBLIES 110 Introduction: Mechanical Coupling in Supramolecular Chemistry 110 Stereochemical Descriptions ofAszL3 and Other Supramolecular Assemblies 113 Ligand Scaffolds and Effects on Helicity 116 Structural Effects ofPnictogen Arene Interactions 118 Conclusions: The Arene-CHzSH Group as a Chelating Moiety 120 Experimental Details 121 General Information 121 Chapter xiii Page General Method for Synthesis of AszL3 Assemblies 121 Crystallographic Data 122 V. DESIGN CONSIDERATIONS FOR THE GROUP 15 ELEMENTS: THE PNICTOGEN."rr INTERACTION AS A COMPLEMENTARY COMPONENT IN SUPRAMOLECULARASSEMBLY DESIGN 124 Introduction 124 Pnictogen Arene Interactions Observed in Mononuclear Complexes 125 Pnictogen"·Arene Interactions in a Supramolecular Context 127 Factors Contributing to Pnictogen•••Arene Interactions 129 Charge Transfer Model for the Pnictogen Arene Interaction 129 Structural Survey 130 Trends Observed in Pnictogen··.Arene Interactions 131 Are Interaction Distances and Angles Correlated? 136 How Does the Pnictogen"·rr Interaction Influence Self-Assembly? 138 Effects of Competing rr-Stacking Interactions 138 Macrobicyclic Effects and Steric Strain 140 Geometric Consequences of Chelate Ring Size 140 Conclusions 142 Experimental Details 144 Chapter xiv Page General Information 144 Ligand and Complex Syntheses 144 Crystal Structures Included in Survey : 146 VI. CONCLUSIONS AND OUTLOOK FOR FUTURE WORK 147 Supramolecular Design for Lead (II) and Mercury(II) 147 Extended Supramolecular Assemblies of As(III) 148 Mechanical Coupling and Design Strategies for Higher-Order AsrnLn Assemblies 149 Proposed Design Requirements for Higher-Order AsrnLn Assemblies 150 Pnictogen".rr Interactions as Supramolecular Design Elements 150 APPENDIX: SPREADSHEET FOR THERMODYNAMIC PARAMETERS EXTRACTED FROM VARIABLE TEMPERATURE NMR DATA. 152 REFERENCES 160 xv LIST OF FIGURES Figure Page 1.1. Coordination geometries commonly observed in main group coordination complexes 4 1.2 HgzLzClz macrocycle where L=adenine-Nl oxide 8 1.3 Formation of an HgzLzClz macrocycle whose structure depends strongly on the geometry of an intramolecular hydrogen bond 8 1.4 Expanded dinuclear Hg(II) macrocycle with diphenylfhiorene-derived spacer 9 1.5 A planar, trinuclear Hg(II) macrocycle bridged by ortho-substituted perfluorophenylligands 11 1.6 Synthesis of mercuracarborands with and without halide templates 11 1.7 Al4L4 molecular square with tetrahedral aluminum(III) centers bridged by 2-hydroxybenzoxazole ligands 14 1.8 Aluminum-based helicate prepared from a hydroxypyridinone ligand 15 1.9 Guest-dependent formation of helix or tetrahedron from a bis-catecholamide ligand 17 1.10 Phthalocyaninato ligand and its dinuclear TI(l) complex. 18 1.11 Abifunctional ligand containing catecholate and phosphine donors assembles with tin(IV) to form a C3h symmetric heterometallic structure 21 1.12 A pyrazine carboxylate ligand forms a trinuclear macrocycle with tin(IV) ....23 1.13 A pyridine dicarboxylate ligand assembles with organotin precursors to form a trinuclear macrocycle with a repeating capsular structure 24 1.14 A pyridine/pyrimidine Pb16Ls grid assembly 25 Figure xvi Page 1.15 Unexpected structures formed with pentatopic and tritopic pyridine/pyrimidine ligands 26 1.16. Formation of AszL3 mesocate and AszLzClz macrocycles 29 1.17 Antimony atoms are interspersed with alternating Na(I) cations and coordinated THF molecules 32 1.18 Supramolecular triangle based on coordination of dicarboxylates to organotellurane centers 34 1.19 A molecular square based on coordination of biphenyl bridges to cis-square planar iodinium cations .35 1.20 Mechanism for guest exchange from M4L6 host structures showing host distortion 37 1.21 Partial encapsulation ofan amphilic guest molecule .38 1.22 Halide sandwich complexes ofhexamethyl-[9]mercuracarborand-3 39 2.1 Examples of low coordination number Pb(II) complexes with oxygen and sulfur donors 50 2.2 2-Mercaptoacetamide ligands designed for Pb(I!) coordination 52 2.3 Crystal structure of dibenzyl-2-mercaptoacetamide 4 .53 2.4 Example of incorrect assignment of lead(II) coordination sphere in the solid state 55 2.5 Infinite chain structure obtained from biphasic reaction ofthiophenol and Pb(OAc)z 57 2.6 Tris-pyrazyl Pb(II) complex model of ALAD active site 58 2.7 Proposed designs for the next-generation ligands for Pb(II) 61 Figure xvii Page 2.8 Aryldithiolate ligand design with substituents to block thiolate bridging interactions 62 2.9 Diphenylmethane spaced 2-mercaptoacetamide ligand 63 3.1 Examples of supramolecular hosts containing diphenylmethane 78 3.2 Predicted three-dimensional structure of AszLz3based on molecular mechanics (MM3) calculations 80 3.3 Three-dimensional structure of AszLz3 82 3.4 Overlays of DFT optimized geometry for AszLz3mesocate with crystal structure and optimized helicate geometries 84 3.5 VT-NMR of AszLz3in CDzClz 87 3.6 Examples of edge-to-face aromatic interactions 89 3.7 gHMBC of AszLz3 (CDzClz, 500 MHz, -20 °C) 98 3.8 gHMQC of AszLz3 (CDzClz, 500 MHz, -20 °C) 99 3.9 Variable temperature lH NMR of AszLz3in CDCb 100 3.10 Variable temperature lH NMR of AszLz3in CDClzCDClz l00 3.11 Chemical shiftofH3 in AszLz3as a function of temperature in CDzClz l0l 3.12 Observed and calculated chemical shifts for H3 of ASZL23 in CDCb 102 3.13 Observed and calculated chemical shifts for H3 of AszLz3in CDClzCDClz ..... l03 4.1 The new dome of the German Reichstag building in Berlin displays a prominent pair of right-handed helices 111 4.2 Crystal structures of new AszL3 assemblies shown as the /j./j. isomer 118 Figure xviii Page 5.1 Examples of structures from the CSD with particularly short pnictogen•••arene distances 127 5.2 Examples of supramolecular assemblies featuring pnictogen"·arene interactions 128 5.3 A charge transfer model for the pnictogen arene interaction in supramolecular assemblies 130 5.4 Schematic for measurement of pnictogen•• •arene interaction distances and interaction angles 131 5.5 Scatter plot of interaction distance/angle pairs 137 5.6 Macrocyclic effects and pnictogen."arene interactions 139 5.7 Effect of increased "chelate" ring size on an anti-AszLzCIz macrocycle 141 xix LIST OF TABLES Table Page 3.1 Thermodynamic parameters for the closed-open equilibrium of AszLZ3 in CDzClz, CDCb, and CDClzCDClz based on the chemical shift change associated with H3 92 A 3.2 Observed chemical shift of H3 and derived values for Pc, D, and K for AszLz3in CDzClz 104 A 3.3 Observed chemical shift of H3 and derived values for Pc, D, and K for AszLz3in CDCI3 104 3.4 Observed chemical shift of H3 and derived values for Pc, D, and K for AszLz3in CDClzCDClz 105 3.5 Atomic coordinates for (.1,.1)-AszLZ3 DFT geometry 106 3.6 Atomic coordinates for (.1,A)-AszLZ3 DFT geometry 108 4.1 Stereochemical outcomes for AszL3 assemblies 119 5.1 Mean interaction distances and angles for pnictogen."n interactions 134 5.2 Arsenic-containing crystal structures included in structural survey 146 5.3 Antimony-containing crystal structures included in structural survey 146 5.4 Bismuth-containing crystal structures included in structural survey 146 xx LIST OF CHARTS Chart Page 2.1 Frequency of coordination numbers found in Pb(II) complexes 50 5.1 Mean pnictogen."n distances observed in the CSD 133 5.2 Mean angles for pnictogen".n interactions observed in the CSD 134 5.3 Angular preferences for pnictogen".arene interactions as observed in the CSD 135 xxi LIST OF SCHEMES Scheme Page 2.1 Two methods for synthesis of2-mercaptoacetamide ligands 52 2.2 Synthesis of Pb(SPh)3 salts 56 2.3 Synthesis of aryldithiolate ligands .59 3.1 Synthesis of AszLz3 79 4.1 General self-assembly reaction forming AszL3 helicates and mesocates from bis(mercaptomethyl)arene ligands and illustration oflocal stereochemistry of arsenic(IlI) centers 114 1CHAPTER I MAIN GROUP SUPRAMOLECULAR CHEMISTRY This chapter is a general survey of the field of main group supramolecular chemistry, with a focus on the structural diversity that can be attained when main group elements are incorporated into supramolecular designs. The material in this chapter has been published in Chemical Society Reviews1! and written by myself with editorial input from D.W. Johnson. Introduction The field of supramolecular chemistry has produced numerous examples of chemically interesting and aesthetically pleasing self-assembled structures using metals as directing elements.2-6 The self-assembly process that guides the formation of these thermodynamically stable architectures is thought to have many advantages over traditional stepwise synthesis in accessing large and ordered structures.z Most of these assemblies leverage the predictable and well- characterized coordination preferences of the transition metals - typically octahedral, square planar or tetrahedral. A growing area of interest, however, lies in * Dedicated to Professor Kenneth N. Raymond on the occasion of his 65 th birthday. 2the exploration of main group metals as directing elements in order to access structure types previously unattainable through traditional means.6 The main group elements possess unique coordination preferences and electronic properties observed rarely in the rest of the periodic table, presenting unique opportunities for the preparation of novel structures with new and interesting characteristics.3-s This tutorial review highlights examples of self-assembled supramolecular structures comprising main group elements with a focus on the unusual coordination geometries often observed in these assemblies. Scope of Review The main group elements are typically defined as the set of sand p block elements, plus zinc, cadmium, and mercury.4 For the purposes of this review, we mainly consider the heavier p block elements, and we focus on supramolecular structures containing main group metal centers bearing unusual coordination geometries. Some notable examples of main group supramolecular complexes found in tetrahedral or octahedral coordination environments are also included. To limit the scope of this tutorial review, we highlight key representative examples of the complexes found from Groups 12-17. Complexes in Group 12 are particularly challenging to separate into "main group" versus transition metal-type complexes, as zinc tends to behave more as a transition metal in terms of reactivity and coordination and is often covered thoroughly in general supramolecular reviews. 3Mercury, however, tends to display the unusual coordination geometries common to the main group elements and will be discussed in ((Group 12 - Zinc, Cadmium, and Mercury." Finally, structures containing Group 1 and 2 elements are not covered, and the reader is referred to recent reviews,?-9 The types of structures reviewed include discrete, polynuclear complexes without bonds between the main group metal centers. We also exclude organometallic main group complexes and those in which the multiple main group elements are bridged by simple halogen or chalcogen ions to form dimers or larger complexes. This review focuses on discrete self-assembled metal-ligand complexes where the main group ion is a directing element for the self-assembly reaction. Throughout the review, we are careful to note where the properties and coordination geometries ofthe main group elements depart from those expressed by the transition metals. Given the unusual coordination preferences of the main group, we first classify these coordination spheres, followed by a more detailed discussion of the individual supramolecular structures. Coordination Geometries of Main Group Elements The set of coordination geometries most commonly observed with transition metal complexes is relatively small and limited compared with the main group metals. Ligand distributions far removed from the typical tetrahedral and octahedral arrangements are frequently observed in complexes containing elements from Groups 13-15; these unusual structures are intimately linked with the unique electronic properties of these elements.3 The heavier main group elements such as lead(II) and thallium(I) are stabilized in unexpectedly low oxidation states due to relativistic stabilization of s orbitals. The presence of this "inert pair" of electrons often results in some of the more unusual hemidirected coordination spheres observed in main group complexes. 4 d b f Figure 1.1. Coordination geometries commonly observed in main group coordination complexes. The bond angles shown correspond to the idealized coordination geometries. Depending on the nature of the bonding in each case, the precise angles may vary. This is quite pronounced in the trigonal pyramidal geometryf, where increasing p character in the bonding can reduce the bond angles to near 90°. 5Tetrahedral and octahedral ligand distributions are frequently observed with the trivalent Group 13 elements, such as aluminum, gallium, and indium. A selection ofthe coordination geometries of particular relevance is illustrated in Figure 1.1, in which a and b indicate the more common tetrahedral and octahedral coordination spheres, respectively. Less common are the trigonal bipyramidal (c), disphenoidal (d), square pyramidal (e), and trigonal pyramidal (f) coordination geometries. Coordination geometries common to the heavier p-block elements often exhibit stereoactive lone pairs, but these preferences are less predictable. Particularly common are trigonal pyramidal (observed often with the pnictogens), square pyramidal, and disphenoidal (observed frequently in Pb(II) complexes) coordination geometries about the metal center. Group 12 - Zinc, Cadmium, and Mercury The Group 12 elements are a somewhat special case when considering the chemistry of the main group elements. Zinc in particular resembles a transition metal in both reactivity and coordination preferences and has been used extensively in the preparation of supramolecular structures. The literature on this topic is certainly too broad to be considered here and has been discussed in many general supramolecular reviews.10,1l,2 Cadmium displays similar properties and coordination preferences to zinc, although it does possess a preference for softer Lewis base donors, similar to its softer main group brethren. In particular, trigonal 6planar coordination of three cysteine thiolates is used to assemble tripeptide bundles about a Cd(lI) center.1Z Mercury(II)-containing supramolecular structures, however, tend to have much more in common with main group elements, in part because of their propensity to form structures with low-coordinate linear or other distorted coordination geometries. This fact has been used to direct the formation of a variety of macrocyclic structures, some of which exhibit novel guest binding properties. Mercury(II) is not often included in the design of well-ordered, discrete supramolecular complexes. In the few examples of supramolecular mercury complexes, the Hg(lI) ion typically adopts either two-coordinate, linear geometries or a distorted tetrahedral, nearly disphenoidal type of geometry. An example of the disphenoidal geometry occurs when mercury(lI) is coordinated to adenine-N1- oxide in an HgzLz macrocyclic structure (Figure 1.2).13 While distorted from ideality, one can clearly 0 bserve the nearly linear (155°) arrangement of the Cl-Hg-Cl triad, as well as a much s.harper angle (96.7°) between the oxygen and nitrogen donors. An elegant example of mercury(II)-directed self-assembly combines both the lability of Hg-N bonds as well as the directional preferences of hydrogen bonds to influence the structure.l4 The twofold symmetric dipyridylligands shown in Figure 1.3 contain both a metal-coordination site and a pair of hydrogen bond donorjacceptors. The nature of the structure formed depends exquisitely on the 7presence or absence of one methylene group between the pyridine and the amide linkage or between the central phenylene ring and the amide linkage - without these methylene groups, an HgzLzXz (X = CI, Br, I) macrocycle forms, while with the methylene groups, polymers and sheets form. In the structure lacking the methylene linkages, two intramolecular hydrogen bonds stabilize the macrocycle. Both tetrahydrofuran (THF) and dichloroethane (DCE) have been found in the cavity. The Hg(II) atoms in the structure are separated by 12.77 Awhile the distance between the aromatic rings of the two ligands measures approximately 7.68 Awhen DCE is encapsulated. Another pyridine-containing Hg macrocycle has been prepared with a much larger ester-linked diphenylfluorene backbone and bridging two meta-substituted pyridine rings (Figure 1.4).15 The cavity formed in this maerocycle is significantly larger than the previously described macrocycle; the Hg atoms are separated by 16.68 Aand the distance between the two central, quaternary carbon centers is 15.04 A. The crystal structure of this macrocycle reveals several disordered water molecules in the cavity. 8Figure 1.2. HgzLzClz macrocycle where L = adenine-N1 oxide. Hg(I1) atoms are separated by 7.43 A. The Cl-Hg-CI triad is nearly linear (155°), while the N-Hg-O triad lies at a sharper angle (96.7°). In most cases, spheres represent the main group clement directing assembly formation. Unless otherwise noted, gray = carbon, white = hydrogen, blue =nitrogen, red =oxygen, and green = chlorine. Gray spheres = Hg(Il). G - HF. DCE '}-(-/> /lt...- f I ) t1.( ) N- / Figure 1.3. Formation of an HgzLzClz macrocycle whose structure depends strongly on the geometry ofan intramolecular hydrogen bond. The hydrogen bonds are represented as dashed lines. Hg(II) cations are separated by 12.77 A, while the two pyridyl rings interacting with the guest are separated by 7.68 A. The Cl-Hg-Cl triad is again nearly linear, while a more bent arrangement is observed for the N-Hg-N triad. 9Figure 1.4. Expanded dinuclear Hg(II) macrocycle with diphenylfluorene-derived spacer. The larger fluorene based ligands provide a much larger cavity; disordered water molecules have been removed for clarity. The Hg(I1) centers are separated by 16.68 A, while the distance between the quaternary fluorene carbons is 15.04 A. The tendency of Hg(II) to form linear coordination complexes has been used to prepare a series of trinuclear macrocycles.l 6 While these structures fall into the realm of organometallic chemistry by virtue of the Hg-carbon bonds that connect the components, these complexes are a striking example of main group elements as integral parts of a macromolecular assembly. In this case, the ligand is an ortho substituted tetrafluorophenyl ring, with Hg(II)-C bonds occupying these two ortho positions. The linearly coordinated Hg(II) ions form the edges of an equilateral 10 triangle capped by the ortho substituted ligands, with an average Hg·..Hg distance of 3.63 A. The Hg3L3 macrocycle has been successfully cocrystallized with a wide variety of organic compounds that generally do not penetrate the cavity of the macrocycle, but rather are complexed to the exterior of the macrocycles. A representative example of an Hg3L3 structure is shown in Figure 1.5. Icosahedral carboranes such as closo-1,2-CzBl0H1Z (ortho) and closo-l,7- CZBlOH1Z (meta) have also been used to great effect in the preparation oftri- and tetranuclear macrocycles of mercury(ll), also referred to as mercuracarborands or "anti-crown" reagents (due to their anion binding properties),17 The formation of tri- versus tetranuclear structures is dependent on the presence of an anion template - when halide salts of Hg(lI) are employed, a planar, tetranuclear macrocycle [[12]mercuracarborand-4) is obtained. A variety of structures form based on the size of the halide template. For example, when HgCIz is the Hg(lI) source, the chloride is coordinated to all four mercury atoms in a square-planar fashion inside the macrocycle. Bromide and iodide, however, are both too large to fit inside the macrocycle and are therefore coordinated to the Hg(II) ions out of the plane of the tetranuclear macrocycle. If Hg(lI) salts such as acetate are used instead, the template effect is lost and a trinuclear macrocycle forms (Figure 1.6). These triangles exhibit a rich host-guest chemistry of their own, including the formation of 11 Figure 1.5. A planar, trinuclear Hg(II) macrocycle bridged by orthO-Sllbstitllted perfluorophenylligands. The average Hg(II) ...Hg(II) distance is 3.63 Aand the Hg(II) centers form an equilateral triangle. Light blue =fluorine. [12]Mercuracarborand-4 [9] Mercuracarborand-3 Figure 1.6. Synthesis of mercuracarborands with and without halide templates. A tetranuclear macrocycle (left, [12J-mercuracarborand-4) forms in the presence of a halide template. Without the template (right, [9J-mercuracarborand-3), a trinuclear macrocycle is formed. Vertices of polyhedra represent boron, while solid dots represent carbon. Hydrogens are omitted for clarity. 12 sandwich-type structures with both iodide and benzene in the solid state. (See "Host-Guest Chemistry of Supramolecular Main Group Complexes lJ). Hg(II) also forms a variety of supramolecular complexes with amino acid derived ligands. Cysteine, for example, assembles with Hg(II) to form an M4L4 supramolecular square where each terminal thiolate group bridges two metal cations. iS While Hg(II) often adopts the aforementioned linear coordination sphere, a trigonal tris-thiolato structure can be enforced by use of a de novo designed peptide sequence to assemble tripeptide bundles similar to those reported with Cd(II).12 Group 13 - Aluminum and Gallium The Group 13 metals have recently emerged as important building blocks in the fabrication of self-assembled supramolecular structures. Aluminum, gallium, and indium are generally found as trivalent cations and consequently have a preference for harder donor ligands based on oxygen and nitrogen and tend to take on predictable tetrahedral and octahedral coordination geometries. Multibranched chelating ligands based on ~-diketonates, catecholates, hydroxamates, and other such groups have been used to great effect in the rational synthesis of supramolecular structures. 13 An interesting example of an Al4L4 square structure has been prepared from 2-hydroxybenzoxazole and trimethylaluminum, where each Al(III) loses one methyl group and is coordinated by two ligands.l9 The resulting tetrahedral aluminum center comprises each corner of the square, and it is bridged by coordination of a hydroxyl oxygen on one ligand and an oxazole nitrogen on a second ligand (Figure 1.7). Most supramolecular structures with a square topology such as this are formed from cis-square planar coordination to metal centers - in this case, the tetrahedral geometry around the corner units induces a bent structure analogous to cyclobutane. A representative example of the many metallohelicates prepared from trivalent main group metals is shown in Figure 1.8. The bis(hydroxipyridinone) ligand shown on the left forms an AlzL3 helicate as a racemic mixture of tJ.,tJ. and A,A isomers. This molecule encapsulates one guest water molecule in the center of its cavity in the crystalline state; in solution, the complex converts slowly from the chiral enantiomers to an achiral tJ.,A meso structure. When this ligand is treated with Ga(lll), however, the tJ.,A mesocate is observed in the crystalline state. In solution, the structure rapidly interconverts between the achiral and chiral forms as a result of a fast Bailar twist at the metal centers. This rather uncommon example of the same ligand driving the formation of two different isomers is thought to be caused by the relatively small size of Al (III) compared with Ga(III). The small Al(III) 14 Figure 1.7. AI4 L" molecular square with tetrahedral aluminum(III) centers bridged by 2- hydroxybenzoxazole ligands. The unusual tetrahedral coordination mode observed in the tetranuclear square complex induces a distorted shape akin to that observed in cyclobutane. Light pink =Al(IlI). center should make the trigonal prismatic intermediate in a Bailar twist much less stable as a result of interligand repulsions, and thus the meso and helical isomers interconvert slowly.2o Following the pioneering work of Saalfrank in preparing the first M4L6 coordination cluster,2l many groups have presented spectacular examples of related tetrahedral assemblies. The use of gallium (III] has played a crucial role in this research, as it maintains a predictable octahedral coordination geometry with hard donor ligands (such as catecholates] and the complexes are diamagnetic, enabling study by NMR spectroscopy. In this case, it is clear that main group elements can serve as effective models to study supramolecular systems when transition metals 15 AI(III) ~~OH HN ~ (fOH / 0 • Figure 1.8. Aluminum-based helicate prepared from a hydroxypyridinone ligand. The hydroxypyridinone ligand (left) forms the homochiral AIzL3 helicate (right). The Al(IJI) centers are separated by 7.13 A. Upon encapsulation of a water molecule, the helicate converts to a heterochiral meso-helicate. have properties (such as paramagnetism in iron catecholates) that are not conducive to the use of certain spectroscopic or characterization techniques. Rather than survey the entire field of supramolecular gallium coordination clusters, the reader is directed to a thorough recent review on this topic.2z A few representative examples based on catecholate and ~-diketonateligands are reviewed here. It should be noted that many structures analogous to those prepared with Ga(III) have also been observed using In (III) instead. 16 In particular, dicatecholamide ligands linked with an anthracene backbone are known to form either an MzL3 helicate or an M4L6 tetrahedron depending on the presence or absence of a suitable cationic guest (Figure 1.9). In this case, the use of the more labile Ga(III) cluster rather than a substitutionally-inert Ti(IV) analog was the key to allow study of this interconversion process on a reasonable timescale. Phenyl, naphthyl, and pyrenyl groups have also been used to great effect as ligand spacers in preparation of these tetrahedral structure types. A large series of coordination tetrahedra based on chelation of catecholamide donors to aluminum, gallium and other main group ions has also been recently reported by Raymond and coworkers.z3 These certainly fall under the purview of main group supramolecular chemistry; however, all the coordination modes in these structures are octahedral, so we defer discussion of these structures to "Host-Guest Chemistry of Supramolecular Main Group Complexes," where their rich host-guest chemistry is discussed in depth. 17 Figure 1.9. Guest-dependent formation of helix or tetrahedron from bis(catecholamide) ligand (H4 L). Addition of an appropriate guest (in this case, tetramethylammonium) to the helicate drives the assembly to a tetrahedral structure. Dark red =Ga(ll!). Thallium diverges sharply from the other Group 13 elements in that it prefers a monovalent oxidation state and exhibits a much wider variety of coordination modes. Structural studies have indicated that many TI(I) complexes display a stereochemically active lone pair which leads to hemidirected coordination geometries such as trigonal pyramidal. TI(I) participates in a wide variety of secondary bonding interactions in the solid state, including at least one example of a TI(I) ...TI(l) interactions. 18 a b Figure 1.10. Phthalocyaninato ligand and its dinuclear TI(I) complex. Each TI(I) is coordinated to four nitrogen donors to the phthalocyaninato ligand (a) in a square pyramidal fashion; a weak TI(I) ...TI(I) interaction (dashed line) is observed in the complex (b) where the distance between the two metal centers is 3.69 A. Red-brown =TI(I). A dinuclear complex containing two TI(I) centers coordinated to a phthalocyaninato macrocycle contains several interesting features (Figure 1.10).24.25 Each TI(I) center is coordinated to four ring nitrogens and has a stereoactive lone pair, leading to a very uncommon square pyramidal coordination geometry. Furthermore, a weak TI(I) ...TI(I) interaction (dashed line) is observed through the center of the macrocycle, where the two TI(I) centers are separated by 3.69 A. This is slightly less than the sum of the van der Waals radii of two Tl atoms (3.92 A). It is apparent from this and other structures that these weak interactions between main group ions may have important implications for the design and synthesis of main group supramolecular complexes. Their inclusion may allow for fine tuning of 19 desirable electronic and optical properties, and these interactions alone comprise a supramolecular interaction that can be exploited as a self-assembly motif.26 Group 14 - Germanium. Tin, and Lead The heavier elements of Group 14 (Ge, Sn, and Pb) tend toward the formation of complexes where the central atom is in the (II) or (IV) oxidation state and bears a combination of 0, N, and S-donor ligands. The coordination geometries found in these complexes tend to be variable and complex. While tetrahedral and octahedral complexes are certainly observed frequently, coordination numbers can range from two up to ten, and even twelve in rare cases.27 A deeper understanding of the coordination chemistry of these elements is desirable, especially given their well- known toxic effects and health hazards. These elements also provide interesting opportunities for ligand design, due to the frequently observed stereochemically active lone pairs that cause hemidirected coordination geometries. Germanium is generally found in coordination complexes as Ge(IV) and is known to coordinate strongly to hard oxygen donors such as catecholates. In contrast to the structures observed with chalcogen donors and metals such as AI(III), Ga(III), or In(I1I), Ge(IV) tends toward the formation of cage-type structures of the form [Ge4X10]4- with X being a chalcogen donor such as sulfur or selenium. 20 M4L4 and M4L6 tetrahedral coordination assemblies analogous to those formed with Ga(III) (Figure 1.9) and Al (I II) have been reported using Ge(IV) instead; these were prepared in order to elucidate the mechanism of guest exchange in the Group 13 assemblies, as Ge(IV) is much more inert to ligand substitution than its trivalent counterparts. Ge4L6 was characterized by NMR and high-resolution mass spectral techniques rather than crystallographically due to the poor quality of crystals obtained. Had guest exchange occurred via a partial ligand dissociation mechanism, the rate of guest exchange should have slowed relative to the Ga(III) structure upon substitution of a tetravalent metal cation such as Ge(IV). As the rate of guest exchange was unaffected, the Ge(IV) structure helped to show that the exchange mechanism involved deformation of the host rather than by partial ligand dissociation. Tin(IV) is also well known to form tris-chelate complexes with catecholate donors, leading to the design and synthesis of C3h symmetric mesocates (achiral structures with both ..1 and A centers) containing two tin ions bound to symmetry equivalent coordination sites and a second metal such as silver or palladium coordinated to the softer phosphine donors of the ligand (Figure 1.11). When the mesocate is prepared using silver (I) , a Cs(I) guest appears to mediate the structure's formation - this guest is included in the cluster's cavity and is necessary for formation (the supramolecular structure does not form in its absence).28 This 21 a b Figure 1.11. A bifunctional ligand containing catecholate and phosphine donors assembles with tin (IV) to form a C3h symmetric heterometallic structure. The phosphine catecholate ligand (0) is represented as a heavy line. The assembly forms with either silver(l) or paliadium(II) (b). Silver(I)jpaliadium(II) are represented as smaller blue spheres, while tin (IV) is shown as larger gray spheres. approach takes into account the hard/soft coordination preferences of both metals and allows the formation of heterometallic supramolecular structures with great site specificity. A variety of tin-containing macro cyclic triangles derived from organotin precursors have been reported using bifunctional bridging ligands. A more thorough review ofthese multinuclear organotin structures has been prepared by Haiduc.26 Pyrazine carboxylic acids have also been shown by Ma and coworkers to drive the self-assembly of a trinuclear Sn(IV) macrocycle (Figure 1.12).3 These structures bear tin (IV) in a distorted octahedral geometry where the metal center is coordinated by a chelate ring containing both an oxygen and nitrogen on one ligand, 22 as well as a carboxylate oxygen on a second ligand. Each chelating portion of the ligand is oriented at a 60° angle to the chelating group of the next ligand, which supports the formation of an equilateral triangle. It is interesting to note that the R groups on the organotin precursor have a strong effect on the crystal packing of the triangle: when R is a methyl group, all the triangles exist in an extremely close- packed structure, whereas when R is the much bulkier di-n-butyl group, the triangles spread out and segregate into overlapping regions of alkyl and aromatic groups in the crystalline lattice. Many other examples of tin-containing supramolecular triangles have been prepared; a few representative examples are discussed below. The contributions of Hopfl and coworkers illustrate the effects of solvent on the formation of larger tin- containing architectures,29 especially in the formation of a spherical hydrogen- bonded capsule which forms an extended three-dimensional structure in the solid state. In this structure (Figure 1.13), a pyridine dicarboxylate ligand (a) assembles with organotin precursors to form a triangle (b), which then forms a repeating capsular structure in the solid state (c), where each triangular face is shared to form a large, porous structure sustained by the presence of 36 hydrogen bonds per repeating unit. 23 8 b Figure 1.12. A pyrazine carboxylate ligand forms a trinuclear macrocycle with tin(IV). The pyrazine carboxylate ligand with chelate vectors oriented at 60° (0) assembles with an organic tin(IV) precursor (SnR4) to form an equilateral triangle (b). Sn(IV) centers are separated by 5.35 A. Dark gray =Sn(IV). Lead(ll) has appeared in several supramolecular assemblies, presumably a result of its reasonably predictable coordination preferences with certain types of ligand donors, especially those containing pyridyl nitrogens. The large size of the Pb(lI) ion also provides opportunities for the synthesis of expanded structures with less steric strain than an analogous transition metal complex. Pb(l!) is generally observed in distorted tetrahedral or octahedral coordination environments with a prominent stereochemically active lone pair. 24 ..::'Lf~~O ci~-~0 s b c Figure 1.13. A pyridine dicarboxylate ligand assembles with organotin precursors to form a trinuclear macrocycle with a repeating capsular structure. The pyridine carboxylate ligand (0) assembles with organotin precursors to form a trinuclear macrocycle (b). The extended solid-state structure of this assembly forms a repea ting capsular structure (c). Lehn and coworkers have reported a series of grid-type architectures self- assembled from linear pyridine/pyrimidine ligands and Pb(II) ions. This strategy is well illustrated in a spectacular example of a self-assembled grid-type structure containing eight tetratopic, tridentate pyridine based ligands and sixteen Pb(I1) ions, constituting a [4 x 4] grid, shown in top-down and side-on views in Figure 1.14.30 The self-assembly of this twenty four component system containing 96 coordinate bonds overcomes enormous energetic barriers and effectively highlights the power of self-assembly in the fabrication of grid-type architectures. The structure also contains 16 closely coordinated triflate anions and 8 waters; there are a further 16 triflate anions and one additional water molecule located in secondary coordination with the structure. The ligands in the structure are bent such that the overall form 25 Figure 1.14. A pyridine/pyrimidine Pb16LlJ grid assembly. Note the "saddle"-type structure. Triflate anions and coordinated water molecules have been removed for clarity. Left: top view, right: side view. rr •••rr stacking distance between ligands is 3.62 ft.; average Pb(II) ••• Pb(II) distance is ~6.5 A. Dark purple = Pb(II). is that of a saddle. The short (3.62 A) Tt-stacking distances clearly contribute to the stability of this assembly. The lead(II) atoms are coordinated in a hemidirected fashion with between 7 and 9 donors depending on the number of anions associated with each. Similar pyridine/pyrimidine ligands have induced the formation of [3 x 3] and [3 x 2] grids. Attempts to prepare even larger grids from ligands containing five chelating sites on each ligand led to the surprising formation of a dinuclear helix, with one ligand wrapped around two Pb(II) centers separated by 3.94 A. In the absence of Pb(II), the ligand was observed to exist in an extended, linear conformation - the helix was observed only when Pb(II) was coordinated, shown in Figure 1.15. The specific folding and unfolding of an organic precursor in the 26 a b c d Figure 1.15. Unexpected structures formed with pentatopic and tritopic pyridine/pyrimidine ligands. A pentatopic pyridine/pyrimidine ligand (a) yielded the dinuclear Pb2L helix (b). The tritopic ligand (c) yields two strands of Pb3L bridged by four perchlorate anions (d). presence of a metal template could have promising implications for the design of functional materials that mimic biological systems. 3D An alternate approach to preparing supramolecular structures containing Pb(II) has focused on using ligands which bear a negative charge to balance the divalent lead(II) cation, reducing the presence of coordinating counterions. While the use of alkoxydiazine groups on the chelating site did not lead to the expected formation of a [3 x 3] grid structure, a rather fascinating hexanuclear lead(II) 27 structure formed where three lead ions were found coordinated to each of two ligand strands. (Figure 1.15, C and d.) The grids were then linked in the solid state by bridging perchlorate anions.31 Group 15 - Arsenic. Antimony. and Bismuth The heavier members of Group 15, also known as the pnictogens, are particularly interesting as design elements for supramolecular main group chemistry, owing to a rare preference for a tripodal trigonal pyramidal coordination geometry found infrequently in transition metal complexes. This coordination mode is particularly attractive as a design element, as the coordination vectors involved produce the vertex of a convex polyhedron. This convergent arrangement of ligands in these complexes should favor formation of discrete species without the requirement for blocking ligands often used in other supramolecular design strategies. When observed in their preferred coordination sphere, an additional stereochemically active lone pair of electrons is available: the pnictogens are well known to act as both Lewis acids and bases. In particular, the Lewis acidity of these elements increases markedly as one moves down the group: Bi(III) predominantly behaves as a Lewis acid (due to the lone pair primarily inhabiting a low energy s- orbital), while nitrogen behaves primarily as a Lewis base. This fact may contribute to the possibility for further reactivity of the pnictogen after being incorporated into a supramolecular structure. Finally, these elements are frequently involved in a 28 variety of the weak secondary interactions that form the basis of supramolecular chemistry.3z As the field of supramolecular pnictogen chemistry is extremely broad, we present several representative examples ofthe many complex types known in the literature, with relevant reviews noted as appropriate. The tripodal coordination mode found in many arsenic-thiolate complexes has led to the facile synthesis of both AszL3 triple mesocates33 and a pair of AszLzClz macrocycles34 from 1,4-bis(mercaptomethyl)benzene and As(III) (Figure 1.16). These structures are stabilized by interactions between the arsenic lone pair and the TI system of the central aromatic ring and are in fact so robust that extended heating under ambient atmosphere in the presence of competing metal ions, trifluoromethanesulfonic acid, and p-toluenesulfonic acid fail to cause dissociation of the assembly. Despite the remarkably robust nature of the assembly, proton NMR experiments indicate the rapid interconversion ofthe two macrocycles in solution. Analogous SbzL3 and an SbzLzClz macrocyc1e have also been reported using the same ligand. 35 Tartrate salts are well known to form a variety of interesting structures with both As(Ill) and Sb(III). A dinuclear, double stranded macrocyc1e forms with As(III) coordinated by two hydroxylate and two carboxylate groups from each ligand, leading to a disphenoidal 4-coordinate geometry around the arsenic center, with the stereoactive lone pair maintained.36 As tartaric acid bears two chiral carbon centers 29 SH YSCI3KOH HS Figure 1.16. Formation of AszL3 mesocate and AszLzCh macrocycles. 1,4- bis(mercaptomethyl)benzene leads to the formation of either an MzL3 mesocate in the presence of base or a mixture of syn and anti AszLzCIz macrocycles in the absence of base. Yellow = sulfur and light purple =As(III). (R,R, S,5, and R,5 are possible), any metal complexes arising from this ligand will have several possibilities for their overall stereochemistry. Steric arguments have been used to explain the much greater stability of assemblies containing enantiomerically pure tartrate ligands - the physical interactions between As (III) centers by way of the bridging ligand causes the chirality at one metal center to depend strongly on the other. Antimony tartrate salts are also quite well known. Potassium antimony(III) tartrate has a long history as a pharmaceutical and is in fact one of the few 30 supramolecular complexes in widespread medical use,37 mainly in the treatment of parasites such as schistosomiasis. Interestingly, this complex was long believed to exist as a monomer until modern X-ray crystallography techniques revealed that this species is a self-assembled dinuclear macrocycle. As was observed with As(III), the Sb(III) centers are coordinated to two hydroxylate and two carboxylate oxygen donors, leading again to a disphenoidal coordination sphere. Furthermore, these Sb(III) tartrate complexes are only known to form from (R,R) or (S,5) combinations; a mixed macrocycle containing both tartrate enantiomers does not form. Meso- tartrate complexes containing the (R,5) isomer of tartrate are also unknown. Even more interestingly, when Sb(III) and As(III) are combined and added to the tartrate salt, mixed macrocycles containing one of each metal center form, which reinforces the idea that these macrocycles form through genuine self-assembly processes and concomitant rearrangement of weak metal-ligand bonds.38 The pnictogens are also known to assemble with a variety of organic and inorganic components to form cyclic oligomers. One example (Figure 1.17) involves the formation of an Sb6Na6 alternating ring, supported by the presence of 3- methylcatecholate and coordinated THF molecules.39 As (III) and Sb(III) complexes also both can assemble with Group 13 elements such as Ga to form cyclic oligomers.40,41 Alternating structures of these types may lead to use as interesting precursors for the preparation of semiconductor materials. The ring-like structures 31 are often prepared through reactions that include salt or alkane elimination, as well as dehalosilylation reactions. As Bi(lll) compounds are often challenging to handle due to undesirable properties such as the ease of formation of insoluble oxide salts, its supramolecular chemistry is not nearly as well-developed as is that of As (Ill) and Sb(lll). However, due to its unique position in the periodic table as the heaviest nonradioactive element, it has some important differences in coordination and reactivity from the rest of the Group 15 elements. A large atomic radius and the availability of expanded orbitals leads to a more variable coordination sphere; coordination numbers as high as 9 and 10 have been observed. Because of this fact, for example, tartrate complexes of Bi(lll) (used in medicine to treat syphilis) tend to form polymeric structures with coordination number five (instead of four as observed with the lighter pnictogens) in order to satisfy its greater coordination demands. 32 Figure 1.17. Antimony atoms are interspersed with alternating Na(I) cations and coordinated THF molecules. Sb(IJI) is shown as larger purple spheres, while Na(I) is shown as smaller light blue spheres. A few notable examples that illustrate the intriguing coordination chemistry of Bi(III) are discussed here. Thiosalicylic acid and Bi(III) form an octanuclear complex where a ring of six Bi(III) centers are supported by coordination to six oxygens and one sulfur atom, while two additional bismuth atoms lie above and below the ring's center and are bound by three sulfur and three oxygen atoms. The "empty spaces" of the coordination spheres on the two Bi(III) atoms point into the cavity formed by the ring, possibly forming a secondary Bi...Bi interaction.42 33 Finally, it has been shown that the high coordination number preferences of Bi(I1I) can be sidestepped by the use of a multidentate capping ligand that leaves only a few coordination sites available for the coordination of bridging ligands. This approach has been used to prepare a dinuclear Bi(I1I) macrocycle - triazacyclononanes are used to cap the Bi(I1I) centers through their nitrogen groups, while the carboxylate group bridges to the second Bi(IlI) center.43 Group 16 and 17 - Chalcogens and Halides In general, elements of Group 16 and 17 have not found widespread use as building blocks for supramolecular architectures. The heavier chalcogens in Group 16 such as selenium and tellurium do provide some notable exceptions. These elements are commonly found in the (II) or (IV) oxidation states and their secondary interactions with the halogens have been well-documented in model systems.44 Tellurium(IV) has also been shown to form a trinuclear Te3L3 macrocycle (L =l,2-benzenedicarboxylate) with two p-tolyl supporting ligands (Figure 1.18).45 The Te(IV) cations are arranged in an equilateral triangular fashion with an average Te(IV) ...Te(IV) distance of 6.30 A. The coordination geometry around each Te(IV) center is slightly distorted from an ideal disphenoidal structure: the 0-Te-O bond angle averages 168.3°, while the C-Te-C angle averages 98.5°. While this structure may be categorized as organometallic due to the presence ofTe-C bonds, the assembly ofthe dicarboxylate ligands supports the macrocyclic structure itself. 34 Figure 1.18. Supramolecular triangle based on coordination of dicarboxylates to organotellurane centers. The average distance between Te(IV) centers is 6.30 A. Orange =Te(IV). In a particularly elegant example of synthetic fabrication, Stang and co- workers prepared a molecular square [I4.l4]4+ containing iodinium cations in a cis- square-planar arrangement and a linear biphenyl group as the organic Iinker.46 As direct crystallographic observation was not possible with this assembly, based on ESI-MS and NMR data, an energy-minimized model was prepared (Figure 1.19). While this molecular square should perhaps not be considered a true supramolecular or self-assembled structure due to the stepwise nature of its synthesis, it is nonetheless an excellent and illustrative example of a main group element being used to direct the formation of a nanoscale structure. 35 I +1-( K1~1+ Figur'e 1.19. A molecular squar'e based on coordination of biphenyl bridges to cis-square planar' iodinium cations. A schematic view is shown on the left; a space filling molecular model is on the right. The diagonal distance across the square is approximately 1.5 nm, Purple =iodine. It is clear from the few examples of Group 16 and 17 supramolecular structures that these elements provide interesting motifs for the construction of supramolecular assemblies; the unique manner in which these elements can bind to organic groups will likely expand the variety of structures containing main group elements in the future. Host-Guest Chemistry of Supramolecular Main Group Complexes As research in main group supramoIecular chemistry is less developed than in transition metal-based supramolecular chemistry, likewise the host-guest chemistry of the main group congeners is less explored as well. Many of the 36 structures described have host cavities of insufficient size or inappropriate shape to contain common guest molecules. Additionally, the electronic properties of main group elements often vary substantially from those of the transition metals, which may affect the nature ofthe host structure's cavity in a manner precluding the entry of a guest molecule. A notable example is the AszL3 structure described in "Group 15 - Arsenic, Antimony, and Bismuth": all efforts to bind small guests or metal ions (even protons) in the cluster cavity have failed. Without a more spacious cavity, this assembly appears to be devoid of host-guest chemistry. In contrast, both the Ga4L6 tetrahedraz3 prepared by Raymond and coworkers and the mercuracarborands prepared by Hawthorne1? have exhibited a wide variety of rich host-guest chemistry which has led to a greater understanding of the solution phase behavior of supramolecular assemblies. A combination of NMR studies and computer calculations of tetrahedra containing a variety of metal centers has revealed the mechanism for guest exchange in these structures. Guests escape the cavity through a deformation of the host structure which permits guests to enter and exit through existing apertures, rather than via partial dissociation of one or more ligands from one of the metal center vertices (Figure 1.20). For example, guest exchange rates were comparable when the tetrahedra were assembled from Ge(IV) or Ga(II1). 37 ... a b Figure 1.20. Mechanism for guest exchange from M4 L6 host structures showing host distortion. (0) Cartoon representation. (b) Molecular model representation. The host is shown as a stick representation, while the guest is shown as a dark red space filling model. Ge(IV) is more inert to ligand substitution than Ga(III); therefore, it could be inferred that partial ligand dissociation was not a significant factor in guest exchange. Molecular modeling corroborated this. It was shown that an aperture of sufficient size for ingress and egress of guest molecules could be formed through only deformation of the host structure. Furthermore, a cleverly designed amphiphilic guest containing a Ru(III) sandwich complex and an alkylsulfonate tail can enter through one host aperture without disrupting the structure, leaving the "tail" of the guest outside the cavity, as illustrated in Figure 1.21. This may have 38 a b Figure 1.21. Partial encapsulation of an amphilic guest molecule. Ru(lIl) sandwich complex (a) with alkylsulfonate tail is encapsulated at the hydrophobic end; the hydrophilic sulfonate head group is exposed (b). significant implications for applications in catalysis or synthesis of linear polymers in a stepwise fashion inside a properly designed nanoscale reaction vessel. The mercuracarborands described previously also show rich anion-binding properties, especially with the halides,17 For example, the formation of [12]mercuracarborand-4 requires the presence of a halide template; in its absence, the smaller and less strained [9]mercuracarborand-3 forms instead (Figure 1.6). Furthermore, it has been shown that the iodide guest may be removed from the tetranuclear host with silver acetate; the resulting empty macrocycle loses its planar structure and adopts a folded "butterfly" conformation. This now-empty host has been shown to complex both nitrate anions as well as Bl0H102-. 39 The trinuclear macrocyclic host has been involved in the formation of several interesting sandwich-type structures, such as a 2:1 carborand:benzene structure observed in the solid state. Hexamethyl-[9]mercuracarborand-3 also forms 2:1 complexes with halide ions (Figure 1.22). The halide is coordinated to six Hg(lI) centers, forming what is thought to be three equivalent three-center, two-electron bonds between the anion and each macrocycle. This is supported by the decreasing distance between the trimer planes with decreasing guest size and the equidistant mercury-halide atoms that were less than the sum of the van der Waals radii. Figure 1.22. Halide sandwich complexes ofhexamethyl-[9]mercuracarborand-3. View from top is shown at left; side view at right. The halide shown is iodide - similar complexes form from both chloride and bromide. The distance between Hg-macrocycle planes is J- (4.90 A) > Br- (4.764 A) > Cl- (4.672 A). Pink polyhedra =parent carborand and purple spheres =halogen guest. 40 Conclusions The main group elements have been used to prepare a wide variety of supramolecular, self-assembled structures. Many of these structures contain geometric elements which are either inaccessible or require use of blocking ligands when using transition metals as directing groups. While the main group elements have certainly not seen the widespread use in supramolecular chemistry enjoyed by the transition metals, it is clear that there is much utility to be found in this section of the periodic table, particularly in situations where the properties of a transition metal might preclude the use of a desired experimental technique. This has been particularly important in studying the solution state behavior of catecholamide- based supramolecular structures. The paramagnetism of iron (III) prevents study of these clusters by NMR. Gallium(III), on the other hand, served as an extremely effective model for these systems, allowing for more rigorous characterization of the supramolecular structures and dynamics. As more is understood about the supramolecular coordination chemistry of the main group elements, more and larger structures will continue to be prepared. We anticipate that these structures will see novel uses in the design and synthesis of functional materials and perhaps as synthons for materials for optics, electronics, and other emerging fields. One must also note that there are many challenges associated with work in the main group - coordination geometries are often unpredictable and can lead to unexpected difficulties with solubility or characterization. Crystalline samples of 41 these large structures are also at times difficult to prepare and one must rely on combinations of solution techniques to characterize products. Nevertheless, the stability imparted by self-assembly can lead to robust main group complexes. Given the interest in Bi-containing radiopharmaceuticals and other applications of main group chemistry, supramolecular structures such as those describes herein may find application in a broad range of fields. 42 CHAPTER II LIGAND DESIGN FOR MAIN GROUP SUPRAMOLECULAR ASSEMBLIES Supramolecular chemistry is frequently defined as "chemistry beyond the molecule" - that is to say very broadly, it is the study of noncovalent interactions between molecules. More specifically, the supramolecular chemist's goal is to use information stored in individual molecules to understand the myriad interactions that govern chemical and biological processes. Through careful selection of components, the supramolecular chemist is thus able to impose order in chemical systems and build successively larger assemblies and arrays that exhibit a wide range of functional properties through self-assembly processes that can sidestep lengthy covalent syntheses. Health and Environmental Concerns Associated with the Main Group Elements The main group elements are widely used in many applications. They are frequently a component of electronics, byproduct of mining activity, ETC. Their widespread usage has led to concerns over their toxicity - the pnictogen elements, especially As (III), are well-known to have deleterious effects on both human and environmental health, and lead poisoning is considered to be the most common disease of environment origin in children.47 In fact, only recently was lead banned as 43 a component in children's toYS.48 It is therefore extremely important to gain a better understanding of the coordination chemistry of these elements, both for treatment of acute and chronic poisoning, as well as prevention through more effective sequestration of these contaminants. As shown in the previous chapter, the pnictogens have a strong preference for trigonal pyramidal coordination geometries, especially in the presence of thiolate ligands. In addition, these elements frequently interact with nearby arene rings, which can stabilize rather unusual structures.33 The nature of this interaction is unclear and will be discussed in greater detail in Chapter V. Lead(II) is another significant and dangerous environmental contaminant that has been used in many applications,47,49,sO ranging from cosmetics to a gasoline additive; it is this widespread use that has made its presence the danger that it is today. There are many pathways by which lead contamination can enter the human body, ranging from skin contact in the case of organolead compounds to ingestion and inhalation in the case of lead carbonate based paints. The toxic effects of lead poisoning are also quite variable, from mild developmental impairment to acute psychotic episodes in the unfortunate case of several lead plant workers in the 1920's.sO Lead is known to compete for calcium and zinc protein-binding sites, in particular ALAD heme biosynthetic pathway. Pb(H) has been found to bind preferentially over Zn(H) to models of ALAD; lead in fact binds over zinc by 500:1.51 44 In addition to the structural distortions found when Pb(II) substitutes for Zn(II)52 (much of the structural change is related to the much larger volume occupied by lead)53 the Lewis acidity of the metal center is reduced, contributing to the inhibition of the synthesis of porphobilinogen and the buildup of aminolevulinic acid. Pb(II) also binds to certain DNA sequences with unexpectedly high affinity - one striking example is that of the thrombin binding aptamer, which folds into a unimolecular structure based on formation of a guanine quartet around the Pb(I!) center.54 The structural implications ofthis unimolecular DNA folding suggest a pathway for genotoxic effects of Pb(II); in contrast, lead induces the hydrolysis of RNA,55,56 Supramolecular Desien Principles and Application to Main Group Chemistry Many of the supramolecular assemblies detailed in Chapter I suffer a significant drawback: their preparation frequently relies on serendipity. Introducing functionality into larger assemblies requires a well-defined framework in order to obtain predictable results; in application, this is frequently dubbed rational design, in which chemical building blocks contain information that guides their assembly into larger structures. In particular, this design strategy has been exemplified in supramolecular coordination chemistry by the elegant two- dimensional structures and polyhedra prepared by Raymond,22,57 Fujita, and Stang.2 45 The "information" that is supplied by individual chemical units has many sources. It arises from the coordination preferences of metals. It can be found in the strong directionality of hydrogen bonds and even in the weak interactions between two or more arene rings in close proximity. Central to supramolecular chemistry, however, is the idea of "strength in numbers." Multiple weak noncovalent interactions can provide the driving force needed to cause the spontaneous self-assembly of multiple components to form assemblies otherwise completely inaccessible through traditional means. This include assemblies held together by "seams" of hydrogen bonds, the many structures that utilize the well- understood coordination preferences of the transition metals, and more recently structures that utilize the more unusual and less predictable coordination geometries accessible to the main group elements that has been reviewed in Chapter I. While an overview of main group coordination preferences has been presented in Chapter I, certain elements within this set deserve a deeper examination. In particular, the "lower" pnictogens (As, Sb, and Bi) exhibit coordination preferences that simultaneously mirror the bonding preferences of the "higher" pnictogens (nitrogen and phosphorus), while also retaining a great deal of metallic character. Particularly, there is a marked preference for a trigonal pyramidal bonding pattern, and a prominent stereochemically active lone pair of 46 electrons. In contrast to the strong covalent character involved in nitrogen and phosphorus bonding, As - Bi exhibit bonding patterns that are closer to coordinate bonds in character. The kinetically labile nature of bonds to these lower pnictogens presents a unique opportunity to prepare analogs of existing organic macrocyclic and macrobicyclic cage compounds through a coordinative self-assembly pathway rather than a covalent synthesis. This approach provides the benefits of supramolecular assembly processes and allows access to an increased variety of structures. In addition, some ofthe heavier elements such as lead(II) and mercury(lI) share many of the same coordination preferences despite their lower oxidation state and tendency toward the formation of oligomers in the solid state. Secondary Bonding Interactions Secondary or noncovalent bonding interactions comprise a huge segment of supramolecular chemistry and indeed macromolecular chemistry as a whole;58 their key common feature is their relative weakness (compared to covalent bonds) and reversibility. The strongest ofthese interactions are those involving electrostatic attractions between oppositely charged ions - the crystalline lattice of salts exemplifies this type of interaction. These stronger interactions are also the most difficult to control, as they lack directional character beyond that gained through the preferred packing array in the lattice. Ion...dipole interactions are also quite strong, but can be effectively used to guide the formation of specific assemblies. 47 Among the most widely used interaction in supramolecular chemistry is the hydrogen bond - a vast array of host assemblies have been prepared due to the strength, reversibility, and predictable geometries of hydrogen bonds. Cation."n and n."n interactions also play major structural roles in biological systems and supramolecular assemblies. Although not as strong as primary metal-ligand interactions, secondary bonding interactions play an important role in the overall structure of metal-ligand assemblies. The main group elements in particular exhibit a wide variety ofthese interactions - the frequency at which they present stereoactive lone pairs and hemidirected primary coordination spheres provides many opportunities for additional, weaker interactions. The pnictogens display these secondary interactions with chalcogens (especially oxygen and sulfur),59 halogens, hydrogen bond donors, and arenes in particular. Ligand Design for Pb(II) and Hg(II) In a nearly exhaustive review of the known coordination chemistry and biochemistry oflead(II), Claudio et al. have shown apparent trends in the crystal structures of lead complexes pertaining to coordination environment, coordination number, and chelate ring size.49 The most common coordination numbers found in the Cambridge Structural Database are 4 and 6.27 (See Chart 2.1 for the distribution of coordination numbers found in this study.) In the case of six-coordinate lead, an 48 octahedral, "holodirected' geometry is typically observed. In the case of the four- coordinate lead, a stereoactive lone pair is usually observed, with distinct geometric consequences. In hemidirected complexes, the 6s2 pair of electrons hybridizes with p orbitals, resulting in what appears to be a "hole" in the crystal structure. Four- coordinate Pb(lI) therefore assumes a distorted trigonal bipyramidal or "sawhorse" coordination sphere. The existence of the 6s2 lone pair in Pb(lI) is of interest on its own. It is generally quite resistant to oxidation due to the stabilization afforded by relativistic contraction of the s orbitals; calculations show that this contraction can be as much as 25% in the 1s orbita1.6o-62 In addition, Pyykk6 showed that 2s orbitals and beyond contract by nearly the same magnitude. A combined crystallographic and computational study of the coordination preferences of Pb(II)27,49 found several distinct trends in the coordination preferences of lead (II): • Coordination number: A bimodal distribution (Chart 2.1) was observed that found a preference for a coordination number of four, with a secondary preference for coordination number six. • Coordination sphere: A hemidirected coordination sphere was preferred, in which ligands are unevenly distributed about the metal center that represents the prominent stereochemically active lone pair of electrons. • Donor atoms: The most commonly observed donors to the lead (II) center were oxygen and sulfur.63,64 49 • Chelate ring size: Five and six membered chelate rings were frequently observed, with a preference for five membered rings. • Steric bulk a/ligands: Bulkier ligands tend to drive formation of holodirected complexes, while hemidirected complexes are more common with less sterically demanding ligands. The computational study complementing this crystallographic survey found an interesting parallel between the idealized structure of four-coordinate complexes oflead(II) and the structures observed in the supramolecular complexes of mercury(II) as described in Chapter I. Indeed, two ligands tend to form a nearly linear L-Pb-L triad, while the second L-Pb-L triad forms a much sharper angle (near 90°). Several examples of four-coordinate oxygen and sulfur donor Pb(II) complexes have been reported. 63,64 Figure 2.1 shows two particular examples in which the orientation of both the bite angle between chelate groups and the angle between arene ring scaffolds varies based on the connectivity of the ligand framework. Furthermore, the computational studies suggest a strong preference for hemidirected coordination; in some cases, the hemidirected geometry is more stable than the holodirected by as much as 29 kcaljmol. 50 70 60 50 ~ 40 c ~ 30 <:T ~ 20 U- 10 o r- r- r- - -.... ,., n n I 2 3 4 5 6 7 8 Coordination Number 9 10 Chart 2.1. Frequency of coordination numbers found in Pb(I1) complexes. Note the bimodal distribution centered at coordination numbers four and six. Complexes with coordination number five and below are generally hemidirected; complexes with coordination number six and above are generally holodirected. a Figure 2.1. Examples oflow coordination number Pb(U) complexes with oxygen and sulfur donors. (a) Bis(thiomaltolato)lead(I!)65 orients the chelate groups at 81°, while the aryl rings themselves are oriented at 133° to one another. (b) A bis(thiohydroxamato)lead(I1) structure63,64 orients the chelate groups at 84°, while the aryl substituents are much closer to parallel at 22°. Design and Synthesis of 2-Mercaptoacetamide Ligands With these coordination preferences for Pb(II) in mind, we sought to design twofold symmetric ligands that would assemble to form either two- or three- stranded coordination assemblies. The preference for mixed oxygen/sulfur 51 coordination spheres and five-membered chelate rings led us to work with the 2- mercaptoacetamide ligand class. This ligand class was expected to provide the desired coordination number (4) and maintain the presence ofthe stereochemically active lone pair, which had been shown to be a definite preference of Pb(II) centers. Based on molecular models (CAChe, MM3) that suggested that a PbzLz macrocycle was a feasible supramolecular structure, a bisfunctional 2-mercaptoacetamide ligand with a phenylene linkage was prepared (Figure 2.2 and Scheme 2.1). Direct condensation ofthioglycolic acid with aniline and p-phenylenediamine yielded ligands 1 and 2 in a single step. The monofunctional ligand 1 was prepared as a model compound for comparison with complexes prepared from the bifunctional ligands. When 2 was treated with a variety of Pb(II) sources under a broad series of conditions, only insoluble yellow solids were obtained. An alternate synthesis was developed to prepare ligand 3, based on a durene spacer - the methyl groups were intended to increase the solubility ofthe ligand and any complexes. The insolubility problem persisted, which was attributed to the formation of a hydrogen bond network between the ligands. Ligand 4 with a tertiary amide instead of a secondary amide was thus prepared, eliminating the hydrogen bond donors. Indeed, this ligand was much more soluble than 3; no discrete product could be formed when this was treated with Pb(II) sources. ,,/(" ",~~" > 20(/)); 0.0678 and 0.0954 (all), GOF =1.013 for all 9082 reflections, max/min residual electron density +0.483/- 0.340 e A-3. CeDe 685997. 2-Dimensional NMR Spectra 98 4 3 6 1 6 1 ___-./l\". JL 10 20 30 40 3 4 QI Q e~ 130 5 1) "~ 2 ©; ~ 140 150 -r---"--r--.--.~.,--.....-,.,......,.....---,-../J. I • j • I i I • iii i j ppm 7.2 7,1 7.0 6.9 6.8 6.7 4.1 4.0 3.9 3.8 3.7 3.6 3.5 Figure 3.7. gHMBC of AszL23 (CDzCh, 500 MHz, -20 0e). Correlations H3-C5 and H4-C2 establish connectivity on the aromatic scaffold. 24 3 6 11 ,I J~ JINI Jt i 1<1' i V'\ ,'!i '1!""---- ""'....~'!.............; _'.....__L_ll ....'Ii ..._ ...' ....n...._.Jt__U....J;\............,..... l~....·· '\0.... __............."~ 3S 40 4S 125 130 135 140 99 i 3,> 3.4 3.3 Figure 3.8. gHMQC of AszLZ3 (CDzClz, 500 MHz, -20 QC). H3-C3 and H4-C4 correlations establish connectivity on aromatic scaffold. 100 Variable Temperature NMR Spectra .1J.... .I A __ A. A.-- ~ 11-- 1 A_~ 1 "---- A A. - ~ fI......-.. 1 A....- 1 iii i I I I 4.1 4.0 3.9 3.8 3.7 3.6 3.5 Figure 3.9. Variable temperature IH NMR of AszLZ3 in CDCh. /I.... fj ) l I'. '\ '\..//\ --~~--------------~--~ ~------------~~~~----------------- -----~----------- ---~ ~~ --~ ------ ~~------------~ ...-/'\;,---------- --- /"-;;----------- --_/'.._--- ---~ ----~ ""------/\. -- ~----------' -~-----------' [\------- ____A -' 1'\ _ ~~ -------/1,------/'------ ~A'-------~ J'C---- ~---_.- ~---- ---- ~~ -~-- -/5;----- -----~---------~ '- 3~S Figure 3.10. Variable temperature IH NMR of AszLZ3 in CDCIzCDCIz. 101 Thermodynamic Parameters ] -o,Ot>·cc t------'~_'-- __~_.r'----'r_---"r~'·-'~------~------r--- ~' '" 6585- 1 ..g L.91-S it aI ] &,'10$ eo .5 tdiE5 1B.I) iG'e 120 300 Figure 3.11. Chemical shift ofH3 in AszL23 as a function oftemperature in CDzClz. Marked data points represent experimental data, while the solid line represents calculated chemical shifts. Residuals [8-8_hat) are plotted above. 102 6.995 6.990 6.985 ...... 6.980 E Cl. Cl. rti 6.975 :t ..... it:c 6.970 III 'iU .!:! 6.965 E GI .c u 6.960 6.955 6.950 320310300280 290 Temperature (K) 270260 6.945 +-----,------,-------,------,--------,---------.--------, 250 Figure 3.12. Observed and calculated chemical shifts for 83 of ASZL23 in CDCh. Experimental values are shown as solid circles and the calculated values are shown as the solid line. Residuals are plotted above. 103 7.000 6.990 6.980 E6.970 Q. Q. l'lC 6.960 =: :.c l/l iii 6.950 u 'E III ti 6.940 6.930 6.920 350330310290 Temperature (K) 270250 6.910 +---~------,,-------,---------,.-------.----------r-------, 230 Figure 3.13. Observed and calculated chemical shifts for "3 of AszLZ3 in CDCIzCDCIz. Experimental values are shown as solid circles and the calculated values are shown as the solid line. Residuals are plotted above. ATable 3.2. Observed chemical shift of 83 and derived values for Pc, ll, and K for AS2L23 in CD2Ch. 104 Temperature (K) 193 203 213 223 233 243 253 263 273 283 293 303 bobserved 6.859 6.870 6.881 6.893 6.905 6.917 6.928 6.938 6.948 6.958 6.969 6.977 0.854 0.817 0.777 0.736 0.694 0.653 0.612 0.574 0.537 0.502 0.470 0.440 6.859 6.870 6.881 6.893 6.905 6.917 6.928 6.939 6.949 6.959 6.968 6.977 K 0.17 0.22 0.29 0.36 0.44 0.53 0.63 0.74 0.86 0.99 1.13 1.27 A Table 3.3. Observed chemical shift of 83 and derived values for Pc, ll, and K for ASZL23 in CDCb. Temperature (K) 253 263 273 283 293 298 303 313 bobserved 6.950 6.958 6.965 6.972 6.978 6.981 6.984 6.989 Pc 0.670 0.623 0.577 0.533 0.491 0.471 0.452 0.416 A II 6.950 6.958 6.965 6.971 6.978 6.981 6.984 6.989 K 0.49 0.61 0.73 0.88 1.04 1.12 1.21 1.40 "- Table 3.4. Observed chemical shift of 83 and derived values for Pc. D. and Kfor AszLz3 in CDChCDCh. Temperature bobserved Pc "- KD(K) 248 6.921 0.520 6.920 0.92 253 6.927 0.486 6.927 1.06 258 6.933 0.453 6.933 1.21 263 6.940 0.421 6.940 1.37 268 6.944 0.392 6.945 1.55 273 6.950 0.364 6.951 1.75 278 6.957 0.338 6.956 1.96 283 6.961 0.314 6.961 2.19 288 6.966 0.291 6.965 2.44 293 6.970 0.270 6.969 2.70 298 6.974 0.251 6.973 2.98 303 6.977 0.234 6.977 3.28 313 6.983 0.202 6.983 3.94 323 6.988 0.176 6.988 4.68 333 6.992 0.154 6.992 5.50 343 6.996 0.135 6.996 6.40 105 106 DFT Calculations DFT calculations were performed with the Gaussian 2003 package using the 6-31 +G* basis set for all atoms and the B3LYP functional. Table 3.5. Atomic Coordinates for (d,d)-AszL23 DFT Geometry. Atom X Y Z Atom X Y Z C 3.021 2.053 -2.682 H 0.003 -4.8 -3.017 C 1.714 1.731 -3.076 S 5.459 0.858 -1.745 C 0.649 2.586 -2.797 S 5.592 -1.442 0.836 C 0.846 3.776 -2.082 S 4.943 1.841 1.376 C 2.152 4.098 -1.697 C -1.366 4.008 -0.813 C 3.227 3.256 -1.995 C -0.933 -2.974 -2.401 H 1.525 0.802 -3.611 C -1.638 -0.996 3.736 H -0.354 2.311 -3.117 C -2.996 -1.221 4.021 H 2.337 5.016 -1.141 C -3.822 -1.9 3.127 H 4.227 3.524 -1.667 C -3.315 -2.382 1.909 C 4.147 1.104 -3.032 C -1.966 -2.147 1.62 H 3.737 0.129 -3.309 C -1.141 -1.458 2.513 H 4.717 1.472 -3.895 H -3.409 -0.868 4.965 C -0.33 4.663 -1.721 H -4.871 -2.049 3.372 H -0.835 4.997 -2.638 C -4.184 -3.106 0.915 H 0.046 5.575 -1.238 H -1.554 -2.49 0.674 C 0.628 0.09 4.309 H -0.101 -1.289 2.254 C 1.685 -0.826 4.38 C -2.721 4.366 -0.908 C 2.969 -0.479 3.953 C -3.678 3.81 -0.06 C 3.237 0.801 3.45 C -3.312 2.88 0.925 C 2.176 1.714 3.361 C -1.962 2.513 1.011 C 0.896 1.366 3.791 C -1.004 3.067 0.158 H 1.506 -1.822 4.781 H -3.033 5.088 -1.661 H 3.774 -1.205 4.029 H -4.723 4.091 -0.172 H 2.358 2.711 2.965 C -4.325 2.273 1.863 H 0.095 2.099 3.73 H -1.654 1.777 1.75 C -0.763 -0.29 4.77 H 0.033 2.758 0.252 H -0.682 -0.944 5.65 C -2.136 -3.159 -3.106 H -1.287 0.61 5.118 C -2.988 -2.089 -3.374 C 4.65 1.188 3.088 C -2.662 -0.789 -2.957 H 5.333 0.353 3.262 C -1.468 -0.606 -2.25 H 4.992 2.021 3.715 C -0.618 -1.681 -1.971 C 4.002 -3.1 -0.88 H -2.404 -4.154 -3.454 C 3.611 -3.022 -2.225 H -3.919 -2.265 -3.909 C 2.323 -3.385 -2.622 C -3.562 0.386 -3.239 C 1.379 -3.83 -1.687 H -1.197 0.386 -1.899 C 1.772 -3.907 -0.343 H 0.303 -1.498 -1.428 Table 3.5 (continued) Atom X Y Z Atom X Y Z C 3.06 -3.548 0.056 H -3.824 1.678 2.632 H 4.328 -2.687 -2.972 H -4.909 3.05 2.37 H 2.046 -3.316 -3.672 S -5.636 1.208 1.074 H 1.058 -4.252 0.403 H -3.567 -3.583 0.15 H 3.335 -3.607 1.107 H -4.784 -3.882 1.403 C 5.413 -2.733 -0.486 S -5.457 -2.055 0.036 H 5.944 -3.597 -0.068 H -3.021 1.324 -3.093 H 5.974 -2.391 -1.36 H -3.929 0.363 -4.271 C -0.032 -4.174 -2.116 S -5.13 0.473 -2.225 H -0.5 -4.797 -1.34 As -4.277 -0.116 -0.194 As 4.178 0.167 0.017 107 Table 3.6. Atomic Coordinates for (Ll,A)-AszLZ3 DFT Geometry Atom X y Z Atom X y Z C -0.49 2.751 -1.501 C 1.197 -1.544 -1.855 C -0.801 4.081 -1.201 C 2.204 -1.937 -2.747 C -2.123 4.509 -1.414 C 2.094 -3.198 -3.35 C -3.095 3.641 -1.901 C 1.025 -4.043 -3.052 C -2.778 2.309 -2.217 C 3.368 -1.035 -3.059 C -1.463 1.881 -2.004 S 4.93 -1.416 -2.105 C 0.226 5.074 -0.661 H -1.247 -5.305 -2.648 C 1.61 4.514 -0.411 H 3.108 0.017 -2.908 C 2.56 4.467 -1.444 H -6.662 -1.665 -1.328 C 3.827 3.929 -1.231 H -0.778 -5.259 -0.961 C 4.196 3.432 0.029 H -6.936 -2.85 -0.042 C 3.248 3.474 1.059 H 3.675 -1.141 -4.104 C 1.975 4.011 0.843 H -3.483 3.356 2.955 C 5.582 2.89 0.278 H 0.213 -1.598 5.316 S 6.115 1.525 -0.867 H 5.028 -3.301 0.676 As 4.63 -0.105 -0.272 H -0.561 -2.701 4.197 S 6.137 -1.17 1.061 H -2.774 3.114 1.354 C 5.171 -2.675 1.561 H 5.864 -3.203 2.228 C 3.852 -2.416 2.25 H 5.695 2.56 1.315 C 3.766 -1.622 3.404 H 6.341 3.663 0.105 C 2.545 -1.429 4.048 H -4.17 1.703 -3.756 C 1.368 -2.019 3.563 H -3.377 0.365 -2.922 C 1.455 -2.807 2.409 H -4.115 3.995 -2.033 C 2.677 -2.999 1.759 H -2.395 5.539 -1.186 C 0.04 -1.782 4.247 H 0.522 2.388 -1.351 C -0.795 -0.631 3.689 H -1.189 0.853 -2.232 C -0.388 0.147 2.602 H 0.299 5.912 -1.369 C -1.189 1.19 2.127 H -0.163 5.506 0.271 C -2.421 1.486 2.72 H 2.3 4.855 -2.427 C -2.832 0.703 3.812 H 4.542 3.897 -2.05 C -2.029 -0.327 4.292 H 3.507 3.093 2.044 C -3.276 2.606 2.182 H 1.259 4.034 1.662 S -4.982 2.107 1.626 H -0.852 1.772 1.273 As -4.422 0.447 0.17 H 0.561 -0.057 2.116 S -5.371 1.198 -1.776 H -2.368 -0.915 5.143 C -3.811 1.359 -2.778 H -3.793 0.899 4.282 S -6.164 -0.824 0.899 H 2.503 -0.811 4.943 C -6.222 -2.116 -0.435 H 4.664 -1.152 3.799 C -4.9 -2.763 -0.766 H 2.714 -3.614 0.862 C -4.37 -2.658 -2.062 H 0.557 -3.273 2.009 C -3.155 -3.258 -2.393 H -4.923 -2.106 -2.819 C -2.423 -3.976 -1.438 H -2.768 -3.164 -3.406 C -2.951 -4.077 -0.143 H -2.404 -4.635 0.615 C -4.166 -3.476 0.191 H -4.546 -3.562 1.207 C -1.101 -4.624 -1.799 H 0.963 -5.016 -3.537 C 0.026 -3.658 -2.147 H 2.856 -3.525 -4.055 108 Table 3.6 (continued) Atom X Y Z C 0.127 -2.39 -1.557 Atom H H x y z 1.253 -0.566 -1.38 -0.636 -2.059 -0.857 109 110 CHAPTER IV ARSENIC-TI INTERACTIONS INFLUENCE MECHANICAL COUPLING IN A SERIES OF SUPRAMOLECULAR ASSEMBLIES The material in this chapter was prepared in collaboration with Sean A. Fontenot, Aaron C. Sather, Lev N. Zakharov, and Darren W. Johnson for submission to Angewandte Chemie, International Edition in English. AszL43was prepared by S.A.F.. AszL53was prepared by A.C.S. Crystallography was performed by L.N.Z., and editorial assistance with manuscript preparation was provided by D.W.J. Introduction: Mechanical Coupling in SupramoIecular Chemistry Fabrication of complex structures from simple components has been a research topic of great interest in recent years. ZZ,Z,93-95 Supramolecular self- assembly processes and dynamic covalent chemistry96,97 present a powerful set of tools for the bottom-up synthesis of complex structures with new properties and emergent functionality,98 A defining feature of these synthetic strategies is that information contained within relatively simple components determines the formation of much more complex structures when combined under appropriate conditions. While the self-assembly of metal-ligand complexes incorporating transition metals is well established, with an enormous library of components, main 111 Figure 4.1. The new dome ofthe German Reichstag building in Berlin displays a prominent pair of right-handed helices: the helical twist created by the spiraling walkway echoes that of the corresponding negative space. group supramolecular chemistry is still developing the tools for the predictable formation of well-defined structures.1 These structures are simultaneously chemically elegant and aesthetically pleasing, often echoing the structural and geometric features common found in art and architecture (Figure 4.1). The main group elements possess properties that diverge from those of the transition metals, presenting an attractive target for the supramolecular chemist seeking novel properties and alternative functionalities. As helicates have been described as "the drosophila of supramolecular chemistry,"99 probing the contributions of secondary bonding interactions to the overall structure of main group helicates will provide vital insights toward the construction of larger assemblies. In this communication, we present three new structures based on a self- assembly framework driven by arsenic-n interactions. Mechanical coupling across 112 the ligand scaffold influences the local stereochemistry at the metal centers, revealing that weak secondary bonding interactions may communicate stereochemical information in the same manner as more traditional systems bearing chelate rings. The secondary bonding nature of this interaction, however, introduces additional freedom in the geometric possibilities for complex structures: the helical twist in the ligand spacer is decoupled from the stereochemistry at the metal centers. The vast majority of metallosupramolecular helicates rely on rigid chelating groups; for example, 2,2'-bipyridine, catecholate, ~-diketonate, or 8- hydroxyquinolinate.83 The use of secondary bonding interactions to complete a chelate ring suggests future possibilities combining symmetry based self-assembly with the weak-link approach.100,101 Lewis acid/base adducts formed by interactions between pnictogens and arene rings have been described extensively by Schmidbaur and co-workers,10Z-108 but only recently have they been used in a supramolecular context as a structural motif in the design of larger, higher order assemblies.33-35.109,73 We have previously reported the synthesis and crystal structures oftwo AszL3 mesocates based on trigonal pyramidal coordination of bis(mercaptomethyl)arene ligands to As (III). Primary As-S bonds are providede by the mercaptomethyl groups, while a five- membered chelate-type ring is completed by the secondary intramolecular As-rr interactions. In this context, arsenic-rr interactions are defined as contacts between 113 the trivalent metal center and arene carbon atoms less than the sum oftheir respective van der Waals radii; the interaction is measured between the As(III) center and the centroid of these close contacts.1]z-coordination is typically observed, though in some cases hapticity can be as high as 1]6 when sterically demanding ring substituents center the metal over the ring.103,104,110 Preliminary DFT studies estimate that this mode contributes >7.4 kcaljmol of stabilization per interaction.33 Stereochemical Descriptions of AszL3 and Other Supramolecular Assemblies The first reported supramolecular assembly based around As-n interactions and the bis(mercaptomethyl)arene scaffold, (Li,A)-AszLl3 (HzLl =1,4- bis(mercaptomethyl)benzene), bridged two As(III) centers with a phenylene spacer.33 In this mesocate, the interaction between the As (II!) centers and the arene rings ofthe ligands pulled the lone pair of the metal center toward the interior of the assembly; the twisted disposition of ligands around the [3 axis generated a stereocenter at each metal in the crystalline state. The bidentate nature of the primaryjsecondary coordination spheres creates in effect a distorted octahedral coordination environment around the As (III) center which can then be described with established stereochemical conventions: the absolute configuration at each metal is defined by the direction of torsional twist observed down the As-As axis. A 114 clockwise twist is designated as b. and a counter clockwise twist is designated as A (Scheme 4.1). Helicity in metallosupramolecular architectures has two other origins dictated by the dictated by the ligand spacer: helical induction by (i) structural or (ii) axial chirality.111,l1Z The asymmetric conformation adopted by the diphenylmethane spacer in (b.,A,Mjb.,A,P)-AszLZ3 (HzLZ = 4,4'- SH S\AS , ,, ,, , 1. Base ' r· '~ ,, , ,·ZAS " . . ,.. As/\2. AsCI3 ' __J___ ___ f___ 'lAS SH S 3 a 1\ o II Scheme 4.1. General self-assembly reaction forming AszL3 helicates and mesocates from bis(mercaptomethyl)arene ligands and illustration of local stereochemistry of arsenic(lll) centers. Dashed lines represent the edges of arene rings; sulfur atoms lie at the ends of wedged bonds. Ligand spacers are defined below. 115 bis(mercaptomethyl)diphenylmethane)73 causes the assembly to crystallize as a racemic pair of helices. This complex is an example of a chiral structure originating from a coordination assembly lacking homoconfigurational metal centers. Descriptions of the overall chirality of a given system can be challenging due to the variety of nomenclatures,11z however, based on these examples, we will describe the chiral elements of AszL3 assemblies first by absolute configuration of the metal centers (/1/1\), then by the direction of helical pitch along the ligand spacer (M/?). The choice of ligand spacer in the design of supramolecular helicates directly affects final product, as the local stereochemistry of metal centers in supramolecular helicates and higher order structures is communicated between the metal centers by the ligands' structure.l13 This mechanical coupling is especially apparent in homoconfigurational catecholamide M4L6 tetrahedra in which the overall chirality is enforced by communication along the six ligands that form the edges of the cage.114,115 The general method for preparation of AsZL3 assemblies based on this design strategy is shown in Scheme 4.1: a Cz symmetric bis-thiolate ligand with a rigid or semi-rigid aromatic spacer is treated with a base, followed by the addition of AsCh. Slow evaporation of solvent from a solution of the assemblies yielded crystals of AszL33, AszL43, and AszL53suitable for analysis by X-ray crystallography (Figure 4.2). 116 Ligand Scaffolds and Effects on Helicity The diphenylethane spacer in AszL33 (HzL3 =4,4'- bis(mercaptomethyl)diphenylethane) generates the (tJ.,tJ.,MjA,A,P)-AszL33 racemic pair ofhelicates of idealized D3 symmetry, in contrastto (tJ.,A)-AszLl3 and (tJ.,A,Mj tJ.,A,P)-AszLZ3in which the spacers encourage opposing configurations at the metal centers. The conformation of the ethylene spacer generates a helical domain of its own, directing the mercaptomethyl groups at the ends of the ligand strand in opposite directions and encouraging the ligand to "wrap" around the threefold axis rather than coordinate the metal center edge-on.l13 The homoconfigurational (tJ.,tJ.,MjA,A,P)-AszL33 is estimated to be more stable than the meso (tJ.,A,MjtJ.,A,P)- ASZL33 by -1.5 kcaljmol based purely on steric factors,t indicating that the mechanical coupling between the metal centers is relatively weak, an expected result based on the conformational freedom of the spacer.:J: Likewise, the AszL3 assemblies formed on diphenylacetylene and anthracene scaffolds crystallize as D3 helices. Instead of communicating stereochemical information between metal centers through the conformation of the ligand spacer, t Computations were performed with Fujitsu CAChe 5.0; the geometry at the metal centers was held fixed while the ligand spacer's orientation was varied in order to compare, for example, 8,8,M/A,A,P configurations in a given assembly. *Crystallographic data obtained between -100° Cand 23° Cshow only slight structural changes as the temperature is increased. See experimental section for changes to cell parameters with temperature. 117 the information is transmitted directly across the ligand spacer. For AszL43, the 4,4' substitution pattern of the mercaptomethyl groups generates a slight helical pitch around the primary C3 axis and yields (~,~,M/A,A,P)-AszL43 in the solid state. The calculated difference in energy between (~,~,M/A,A,P)-AszL43and (~,A,M/~,A,P)­ AszL43is small- the chiral assembly is estimated to be more stable by only 0.8 kcaljmol. The anthracene spacer in (~,~,P/A,A,M)-AszLs3 (HzLS =2,6- bis(mercaptomethylanthracene)) has a more dramatic effect due to the offset supplied by the 2,6 substitution pattern. Surprisingly, the helical twist ofthe ligand domain is not the same as that of the metal centers' configurations: most AszL3 assemblies prepared thus far bear ligands twisted in the same helicity as that found at the metal centers, and the AS-TI contact is observed to occupy the "leading edge" of the ligand arene rings. The anthracene scaffold deviates from this trend, as the contact occupies the "trailing edge" of the ligand. This is no doubt the result of the offset in the mercaptomethyl substituents located in the 2,6 positions rather than the 4,4' positions occupied in other ligands. Reversing the observed P helicity to M helicity would have the effect of destabilizing the assembly: (~,~,P)-AszLS3 is preferred over (~,~,M)-AszLs3 by -8 kcaljmol. (~,A,M)-AszLs3and (~,A,P)-AszLS3, while not observed in the solid state, are calculated to lie between the homochiral isomers at approximately the same energy. 118 Figure 4.2. Crystal structures of new AsZL3 assemblies shown as the AA isomer. Ligands are represented as sticks, while As(lIl) centers are shown as spheres. Curved arrows indicate direction of helicity in the ligand domain. Structural Effects of Pnictogen...Arene Interactions The relatively consistent length (and likely strength) of the As-n interaction across the series of ligand spacers (Table 4.1) points to its importance in the overall structure of these assemblies. While the specific ligand spacer employed in designed AszL3 assemblies clearly impacts the chirality of the product, the overall strength of the metal-arene interaction that stabilizes the assembly has an effect on the chirality as well. The helical diastereomer is frequently more stable than the corresponding meso diastereomer in dinuclear helicates,116 possibly due to the additional interior space created when a helicate "unwinds" into a mesocate with an increased metal- metal distance. External factors such as guest inclusion can influence this 119 Table 4.1. Stereochemical outcomes for AszL3 assemblies. Rigid spacers and stronger secondary bonding interactions lead to helicate formation over mesocate formation. Complex As/Sb-:rt Oi)la] Configuration1b] AszL13 3.30 f'..,A AszL23 3.48 f'..,A,M+P[c] AszL33 3.29 f'..,f'..,M AszL43 3.33 f'..,f'..,M AszLs3 3.33 f'..,f'..,P SbzLl3 3.37[d] f'..,f'..,P [a] Average distance between metal center and centroid of close contacts with neighboring arene rings. [b] See Figure 2 for description of helical domain designations; for homochiral assemblies, the 11,11 isomer is shown. [c] (11,i\,M)- and (11,i\,P)-ASZL23 form as a racemic mixture. [d] See text for method of comparing strength of secondary bonding interactions. phenomenon, inducing a transformation between a mesocate and the helicate when the metal-metal distance is decreased due to interactions with the guest.117 A similar effect may be observed in supramolecular assemblies incorporating As-rr or Sb-rr interactions - the increased strength ofthe secondary bonding interaction involving Sb(III) relative to As(III) decreases the metal-metal distance. lOB Although the pnictogen-arene interaction distance increases slightly (0.07 A) between AszLl3 and SbzLl3 (Table 4.1), the interaction is in fact comparatively stronger: the As-S bond length is 2.24 A, while the Sb-S bond length is 2.43 A,35 The ratio ofpnictogen."rr interaction distance to pnictogen-sulfur bond length thus decreases from 1.47 to 1.38 between As(III) and Sb(III). Antimony therefore interacts more strongly with arenes than with arsenic;lOB the overall structural effect is that the pnictogen-arene interaction plays a similar structural role as a water guest in the induction of helicity in an otherwise meso supramolecular assembly. Stronger secondary bonding 120 interactions improve the ability of ligands to transfer stereochemical information throughout an assembly. Conclusions: The Arene-CHzSH Group as a Chelating Moiety We have thus shown that bis-bidentate ligands whose chelate ring consists of a pair of a pair of strong (thiolate coordination) and weak (rr-basic) donors may be used to prepare supramolecular AszL3 assemblies in which the local stereochemistry ofthe metal centers is influenced by mechanical coupling across the ligand spacer in three new supramolecular assemblies. This coupling follows the general pattern established by more rigid chelate systems - information is transferred through both ligand structure and conformation, although the helicity of the ligand domain is decoupled from the stereochemistry of the metal centers when the donor groups are offset from one another as shown in 2,6-substituted anthracene assembly. The differing strength of pnictogen-Jt interactions can serve the same structural role as guest inclusion in determining the formation ofhelicates or mesocates with a given ligand. The fact there is nearly an order of magnitude difference in the strength of the As-S bond and the As-rr interaction suggests that this structural motif could treated as a hemilabile ligand under the weak link self- assembly framework1oo -ligand spacers with the ability to transmit stereochemical information are a key component in expanding this design strategy to larger structures. 121 Experimental Details General Information Starting materials and reagents were obtained from commercial sources and used without further purification unless otherwise noted. NMR spectra were obtained on Varian 300 MHz, 500 MHz, and 600 MHz spectrometers as noted; chemical shifts are reported in parts per million (ppm) downfield from tetramethylsilane and referenced to residual solvent as an internal standard. X-ray diffraction data were collected at 173 K on a Bruker Apex diffractometer using MOKa-radiation (1..=0.71073 A). Mass spectra were recorded using an Agilent 1100 Series LCjMSD. Absorption corrections were applied by SADABS. All structures were determined by direct methods and refined on FZ by a full-matrix least-squares procedure. All non-H atoms were refined with anisotropic thermal parameters. H atoms were refined in calculated positions in a model of a rigid group. All calculations were performed by the Bruker SHELXTL 6.10 package (SHELXTL, version 6.10, Bruker AXS, Inc., Madison, WI, 2000). General Method for Synthesis ofAszL3 Assemblies To a solution of HzL in THF under Nz atmosphere was added 2 equivalents of KOH in methanol. This solution was warmed to 50°C and AsCh (0.67 equivalents) was carefully added. After continued heating for 2 hours, the solution was filtered through glass wool and concentrated in vacuo to yield an off-white solid from which 122 crystals suitable for analysis by X-ray diffraction were prepared by slow evaporation. Crystallographic Data AszL33: IH NMR (300 MHz, CDCP): (j = 6.98 (24H, m), 3.78 (12H, s), 2.80 (12H, s). 13C NMR (126 MHz, CDCb, -20° C): (j = 140.14, 136.33, 128.98, 128.17,38.31,35.03; MS [M+H+] C4sH4SAszS6 requires 966.1, found 967.0. Crystal data: C4sH4SAszS6, Mr = 967.06,0.37 x 0.14 x 0.04 mm, monoclinic, C2/c (N 15), a = 30.468(3) A., b = 13.6420(12) A., c = 11.0855 (10) A., f3 = 106.397(lr, V = 4420.2(7) A3, Z = 4, Z=O.5, pealed = 1.453 g cm-I, f1. = 1.829 mm-I, 28max = 54.00°, T = 173(2) K, 24133 measured reflections, 4837 independent reflections [Rint=0.0417], 253 independent refined parameters, Rl = 0.0468, wR2 = 0.1210 (with I> 20-(/)), Rl = 0.0670, wR2 = 0.1416 (all data), GOF = 1.088, max/min residual electron density +1.064/-0.452 e A-3. ASzL43: C4sH36AszS6, Mr = 954.97,0.18 x 0.16 x 0.12 mm, monoclinic, C2/c (N 15), a = 31.093(2) A, b = 13.5902(10) A, c = 10.6951(8) A, f3 = 107.090(lr, V = 4319.7(5) A3, Z = 4, pealed = 1.468 g cm-I, f1. = 1.871 mm-I, 28max = 54.00°, T = 173 (2) K, 19714 measured reflections, 4719 independent reflections [Rint=0.0416], 253 independent refined parameters, Rl = 0.0420, wR2 = 0.0897 (with I> 20-(/)), Rl = 0.0619, wR2 = 0.1015 (all data), GOF = 1.094, max/min residual electron density +0.607/-0.238 e A-3. 123 AsZLS3: CS4H4ZASzS6, Mr = 1033.08,0.07 x 0.06 x 0.05 mm, monoclinic, C2/c (N 15), a =26.660(7) Ab =10.399(3) Ac =20.779(6) A(J =125.192(7t, V =4708(2) A.3, Z = 4, pealed = 1.458 g cm-i,,u =1.723 mm-i, 28max =54.00°, T =173(2) K, 9193 measured reflections, 4993 independent reflections [Rint=0.0856], 280 independent refined parameters, Rl =0.0756, wR2 =0.1014 (with I> 2a(I)), Rl =0.1318, wR2 =0.1626 (all data), GOF =1.004, max/min residual electron density +0.752/-0.453 e A-3. 124 CHAPTER V DESIGN CONSIDERATIONS FOR THE GROUP 15 ELEMENTS: THE PNICTOGEN."rr INTERACTION AS A COMPLEMENTARY COMPONENT IN SUPRAMOLECULAR ASSEMBLY DESIGN We study the origins of the pnictogen."rr effect and relate this to its utility as a design element for the preparation of MzL3 and MzLzClz supramolecular assemblies. Crystallographic, computational, and synthetic strategies have been employed to probe the contributions of electrostatics, charge-transfer interactions, and steric factors in determining the strength and directionality of this interaction in a variety of contexts. The crystal structure database search, structure and data analysis, and writing of this chapter was done by myself with input from D. W. Johnson. Virginia M. Cangelosi and Corrine A. Allen contributed new crystal structures (solved by Dr. Lev N. Zakharov) to include in the discussion of supramolecular assemblies. Introduction The great success of supramolecular chemistry in the preparation of rationally designed supermolecules from simple components118,lO,Z,119,lZO has led to a concomitant surge of interest in study of the main group elements.l Main group 125 metals occupy an interesting position in the periodic table: the nature of their bonding is generally coordinative, while simultaneously mimicking the nonmetals in regards to their geometric preferences.1 As the main group metals are used more in supramolecular chemistry, their coordination preferences have been incorporated into predictive design strategies that have led to the formation of rather spectacular self-assembled structures. The development of these design strategies has, however, been hampered by the unpredictable coordination preferences of these elements.3 Secondary bonding interactions also introduce new challenges to rational design strategies. Weak interactions with arene rings,86 secondary coordination to Lewis basic elements,59 and steric strain all playa role in determining the overall structure of supramolecular assemblies; the main-group elements and pnictogens in particular appear to have a unique susceptibility to unusual coordination behavior. Pnictogen ...Arene Interactions Observed in Mononuclear Complexes Interactions between arene rings and pnictogen metal centers are of particular interest in the development of supramolecular design strategies with main-group elements. While this type of interaction has been known for quite some time,102-106,110,107,108 only recently has it been used as a specific design element in supramolecular assemblies.33-35,78,73 These interactions were first discovered through the observation that combinations of benzene or naphthalene with 126 antimony trichloride produced highly crystalline solids; the unusually high solubility of pnictogen trihalides in neutral organic solvents has also been attributed to noncovalent interactions between the metal and the solvent. This interaction appears to be quite strong, but its geometry is also variable. Examples of the closest pnictogen".arene interactions for As (III), Sb(III) and Bi(III) found in a survey of the Cambridge Structural Database are shown in Figure 5.1. Of note is the fact that for As(III) and Bi(III), the closest interactions with arene rings are found when the space occupied by the metal's lone pair is oriented perpendicular to the face of the ring, while for Sb(III), the closest interaction places the metal in a tilted orientation with respect to the arene. A number of computational studies on these so-called ((Menshutkin complexes" describe the interaction as a ligand to metal charge transfer interaction and suggest that ligand-based donor orbitals interact with 6p acceptor orbitals on the metal center.108 While these calculations do provide insight into the fact that there is likely some degree of charge transfer character to the interaction, there is still disagreement regarding the bonding strength and are considered unsatisfactory to completely explain the bonding observed between pnictogens and neutral arene rings. In particular, antimony complexes bear off-center interactions much more frequently than do arsenic or bismuth - these calculations provide no explanation for this phenomenon. 127 Figure 5.1. Examples of structures from the CSO with particularly short pnictogen···arene distances. Top: Hexaethylbenzene...AsCb adduct with perpendicular orientation. Lower left: Cyclophane...SbCb adduct with tilted geometry. Lower right: Hexamethylbenzene"·Si cluster adduct with perpendicular geometry. Pnictogen ...Arene Interactions in a SupramoIecular Context We have recently reported the development of a supramolecular design strategy for the preparation of helicates, mesocates, and macrocycles from bis(functionaJ) benzylic thiolates bridged by rigid and semirigid aromatic spacers coordinated to trigonal pyramidal pnictogen centers. 33 The primary thiolate coordination sphere is supplemented by intramolecular secondary bonding interactions between the pnictogen and the ligand scaffold. This stands in contrast to previous examples of pnictogen".arene interactions, which tend to be 128 intermolecular adducts. Several examples of supramolecular M2L3 and M2L2Ch structures prepared through this strategy are shown in Figure 5.2. The assembly shown in (aJ is the first example ofpnictogen."arene interactions so intimately involved in the overall structure of a supra molecular coordination complex. This design strategy was shown to be general to ligands bearing benzylic thiolate donor groups; one example of an extended structure is shown in (hp3 Several examples of macro cycles based on this design strategy have been reported as well, and some of these are capable of interacting further with aromatic solvent molecules in the solid state to form inclusion structures,?8 B b c Figure 5.2. Examples of supramolecular assemblies featuring pnictogen...arene interactions. (a) Basic AsZL3 assembly based on phenylene spacer. (b) Expanded AsZL3 assembly based on an extended ligand. (c) AszLzCIz macrocycles capable of dimerizing around a solvent guest in the solid state. 129 Factors Contributing to Pnictogenn.Arene Interactions The nature of the pnictogennarene interaction clearly has some bearing on its utility in supramolecular assembly design. The possibility of covalent character (ifthe interaction were caused, for example, by interactions between 1t electrons and the As-S 0* orbital) would imply that very specific structural arrangements are required for this interaction to be used as a design element; conversely, a primarily electrostatic interaction would allow considerably more latitude in how the interaction could be incorporated into assembly design. Charge Transfer Model for the Pnictogen.nArene Interaction A closer examination of assembly AszLl3 in Figure 5.2 provides a visual depiction of how the pnictogen. n 1t interaction might be caused by overlap between the As-L antibonding orbital and the arene's 1t electrons to form a charge transfer complex. In Figure 5.3, the "stick" representation of AszLl3 is shown in (a). The primary ligand coordination sphere about the As (III) center is shown in more detail in (b); in particular, the a* orbitals associated with the As-S bonds are drawn in their approximate locations, though not to scale. Figure 5.3(c) is a view down the As-As axis of the supramolecular assembly, showing how the 1t systems ofthe ligand spacer arene rings might interact with these a* orbitals to stabilize the overall assembly. 130 ®e / As·", I fII L L ~L a b c Figure 5.3. A charge transfer model for the pnictogen...arene interaction in supramolecular assemblies. (0) Assembly As2L13 shown as a stick model. (b) Locations of 0* orbitals in the primary coordination sphere about an As(III) center; the generic ligand L replaces the sulfur ligand in this view. (c) View down the As-As axis showing the possibility of interactions (red arrows) between the IT systems of the arene spacers with the 0* orbitals. Structural Survey A systematic survey of the Cambridge Structural Database was undertaken in order to examine how pnictogen size and other factors contribute to the strength of geometry of their interactions with arene rings.§ Structures included in this survey are those containing As, Sb, or Bi in close proximity (i.e., less than the sum of their van der Waals radii) to aromatic rings. As shown in Figure 5.4, the distance for the interaction d and the interaction angle ()were then measured, tabulated, and § In most cases, the pnictogen center interacts in an 1]2 or 1]3 fashion with neighboring arene rings. On occasion the pnictogen may have 1]1,1]4,1]5, or 1]6 contacts with arenes. The interaction is measured using the centroids of the close contacts. loa 131 compared with the interactions observed in supramolecular pnictogen."arene assemblies. Figure 5.4. Schematic for measurement ofpnictogen...arene interaction distances and interaction angles. The interaction distance d refers to the distance between the centroid of observed close contacts and the pnictogen center, while the interaction angle erefers to the angle between this centroid, the pnictogen center, and the ligand (L) opposite the interaction. The As••• rr interaction is depicted here as an rJ2 contact. Trends Observed in Pnictogen •••Arene Interactions Examination of the distances found in the CSO search broken down by pnictogen type shows two important facts about the pnictogen."n interaction. The first is that the average distance decreases with increasing pnictogen size, as shown in Figure 5.5. Arsenic· ..n interactions average 3.45 'A, while interactions with antimony and bismuth are 3.40 and 3.32 A. respectively. The second notable feature is that as the pnictogen size increases, the distribution of the interaction distance increases considerably. The standard deviation for arsenic."n interactions is 0.09 'A, or 2.5%, which increases to 8.62% for bismuth. This data, while not particularly 132 conclusive, does suggest that increasing pnictogen size results in a wider range of possible interactions. The angular preferences associated with this interaction provide a slightly different picture of the pnictogen."rr interaction. The mean angle associated with the pnictogen."arene interaction (Figure 5.6) decrease only slightly as the metal size is increased, and the standard deviations associated with these values are also significant. The actual deviation of the angle, however, does not increase or decrease significantly with increasing pnictogen size. > 6 .; " I I J.S i ! I 3.4 ~ "i 3,3 X 3.2 3' csn As %CSO-' Sb -/"s2L3 C~p$uk:~ ClAs,ll2C:2 f.'!;;n::racydgr; 133 Chart 5.1. Mean pnictogen".n distances observed in the CSD. See Table 5.1 for data values. 134 lao 1/0 16~ C .!< 11&0 "~ :;; 15$ lS0 Chart 5.2. Mean angles for pnictogen....rr interactions observed in the CSD. See Table 5.1 for data values. Table 5.1. Mean interaction distances and angles for pnictogen."n interactions. Pnictogen Meand Deviation Deviation Mean (J Deviation Deviation Type (A) (A) (%) CO) CO) (%) As - CSD 3.45 0.09 2.49 163.38 13.71 8.39 Sb - CSD 3.40 0.19 5.49 162.35 10.40 6.40 Bi - CSD 3.32 0.29 8.62 160.53 11.36 7.08 AszL3 3.34 0.12 3.54 155.22 9.76 6.29 AszLzCIz 3.33 0.08 2.36 153.51 5.19 3.38 o CSD -As CSD - Sb • CSD - Bi • As2L3 Capsules o As2L2CI2 Macrocycles CD c '00 ......... c: liillii, .. I 135 Arene· .. pnictogen-L angle (0) Chart 5.3, Angular preferences for pnictogen."arene interactions as observed in the CSD. The angular distribution is weighted as 1/sin(8),12l A second way to examine the angular data's distribution is shown in Figure 5.7. As reported by Steiner,122 a noncovalent interaction's angle can be associated with the frequency at which it occurs in order to generate a distribution. The frequency is weighted by sin Bin order to properly observe angular preferences.121 The angular distribution shows that there is a preference for this angle to lie between 150° and 170°, A more linear angle is preferred for antimony and bismuth in particular, which suggests that the interaction may have some increased degree of covalent character for these elements. Conversely, a more bent angle is more 136 frequently observed in the supramolecular assemblies reported by our laboratory. This is most likely due to the geometric constraints applied by the relatively short mercaptomethylene tether. Are Interaction Distances and Angles Correlated? While there does appear to be some preference for the angle between the pnictogen-ligand bond and the noncovalent pnictogen."TI interaction to approach a linear orientation (160° to 170° are the most frequently observed angles), comparison between distance/angle pairs is necessary to determine whether there is a relationship between the strength ofthe interaction and its approach to linearity. Ifthe orbital overlap model shown in Figure 5.3 is correct, there should be a strong correlation between the pnictogen".TI interaction distance d and the interaction angle: as d decreases, the angle should approach 180°. In fact, a plot of the distance/angle pairs (Figure 5.5) reveals an essentially random distribution of distances and angles. The lack of correlation between d and () suggests that the pnictogen".arene interaction is primarily electrostatic in character, corroborating the studies that indicated that a charge transfer model is insufficient to describe this type of interaction. The slight decrease in d as the pnictogen's size is increased is likely due to decreased repulsions between the TI system of the arene ring and the lone pair of 137 electrons on the pnictogen. This is likely the result of increased delocalization of these electrons.10B The fact that this interaction is primarily electrostatic in character is actually quite useful for the design of supramolecular assemblies, as the interaction's geometric preferences are less likely to dominate the final structure as much as the primary coordination sphere and the structure of the chosen ligands. o " Q Q " .. 160 , ~. J;: 't! « e gU 'ii !! ! 12U Gil, "Sb O£:h lOU l.S 1.7 1.1U lnt"ra.:3JI($f'$:Z~.'(\.fJJ):i:"!l .. ~~=t1C1t:::$I:$~*c:~±$I:$~~(~~c:~) ..:... ' . 3~__ ~ ~_~L ~~__~ ~: : _ ~---~-----------_.------~--~-,~,._-----_._-_._---_._-_ .. _j--_.,---- -- -~----- 5 Error Analysis • Parameter Chanqe 6 Parameter Modification: 1.00E-06: : ~ ·.·.·······················'·····::~~t~·; ••·•• ,·.·.·.·.·.·.·.·.·.·.·.·••'~~dify 4HO ······.Modify··.I1So····· ··.··~·:~·;;~···~~~···············.··~Odify··~~C······· .....2...-.~.t:t.~.. 2.125E+03· ?j?$g±g~. 2:i2sE+03 2.i2$g+g).2i24.7242s8 -.1Q..~.~.~.;?:~?qE.±2QL.,?:~?qE..±22?:~?qE.±9Q;...7.470E+00. 7.469759 ~ ~=()~f:l.'!__~...__ ?-:~9?~E-±Qg_L_7-:}Q2~±.QQ.L 7.10 2E+0Q~~?J,O? ~±Q.0~:~_~?1018?~ 12 ~ closed 6.818E+00 6.818E+00-6~8i8E+00 6.818E+00 6.818313 13 : ' 14 ' • ~tv10difif:l