USF Libraries
USF Digital Collections

Metal-organic networks based upon dicarboxylato ligands

MISSING IMAGE

Material Information

Title:
Metal-organic networks based upon dicarboxylato ligands
Physical Description:
Book
Language:
English
Creator:
Wang, Zhenqiang
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Crystal engineering
Coordination polymers
Flexibility
Secondary building units
Supramolecular isomerism
Topology
Dissertations, Academic -- Chemistry -- Masters -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Network structures based upon metal-organic backbones represent a new class of functional materials that can be rationally constructed by employing the concepts of supramolecular chemistry and crystal engineering. The modularity of design strategies, the diversity of prototypal structures, and the dynamic features of networks have afforded great advantages over traditional materials syntheses. The research presented in this thesis is primarily concerned with developing an in-depth understanding of the basic principles that govern the supramolecular behaviors of metal-organic networks and gaining an experimental control over the structure and function of these new classes of hybrid materials.The use of rigid and angular organic ligands along with transition metal clusters gives rise to a wide variety of novel metal-organic architectures ranging from zero-dimensional nanostructures to three-dimensional frameworks. Conformational analysis of these structural models suggests the geometric foundations for the existence of superstructural diversity. Controlled crystallization experiments further reveal the synthetic factors that might determine the formation of supramolecular isomers.Careful selection of more labile organic components, on the other hand, leads to flexible metal-organic networks exhibiting dynamic characteristics that have not been observed in their rigid counterparts. The guest-dependent closing/opening of cavities and the ease of fine-tuning their chemical environments demonstrate the effectiveness of such a strategy in the context of generating tailored functional materials.
Thesis:
Thesis (M.A.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Zhenqiang Wang.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 87 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001909584
oclc - 173147522
usfldc doi - E14-SFE0001669
usfldc handle - e14.1669
System ID:
SFS0025987:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001909584
003 fts
005 20070919133553.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 070919s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001669
040
FHM
c FHM
035
(OCoLC)173147522
049
FHMM
090
QD31.2 (ONLINE)
1 100
Wang, Zhenqiang.
0 245
Metal-organic networks based upon dicarboxylato ligands
h [electronic resource] /
by Zhenqiang Wang.
260
[Tampa, Fla] :
b University of South Florida,
2006.
3 520
ABSTRACT: Network structures based upon metal-organic backbones represent a new class of functional materials that can be rationally constructed by employing the concepts of supramolecular chemistry and crystal engineering. The modularity of design strategies, the diversity of prototypal structures, and the dynamic features of networks have afforded great advantages over traditional materials syntheses. The research presented in this thesis is primarily concerned with developing an in-depth understanding of the basic principles that govern the supramolecular behaviors of metal-organic networks and gaining an experimental control over the structure and function of these new classes of hybrid materials.The use of rigid and angular organic ligands along with transition metal clusters gives rise to a wide variety of novel metal-organic architectures ranging from zero-dimensional nanostructures to three-dimensional frameworks. Conformational analysis of these structural models suggests the geometric foundations for the existence of superstructural diversity. Controlled crystallization experiments further reveal the synthetic factors that might determine the formation of supramolecular isomers.Careful selection of more labile organic components, on the other hand, leads to flexible metal-organic networks exhibiting dynamic characteristics that have not been observed in their rigid counterparts. The guest-dependent closing/opening of cavities and the ease of fine-tuning their chemical environments demonstrate the effectiveness of such a strategy in the context of generating tailored functional materials.
502
Thesis (M.A.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 87 pages.
590
Adviser: Mike Zaworotko, Ph.D.
653
Crystal engineering.
Coordination polymers.
Flexibility.
Secondary building units.
Supramolecular isomerism.
Topology.
690
Dissertations, Academic
z USF
x Chemistry
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1669



PAGE 1

Metal-Organic Networks Based Upon Dicarboxylato Ligands by Zhenqiang Wang A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Chemistry College of Arts and Sciences University of South Florida Major Professor: Michael J. Zaworotko, Ph.D. Mohamed Eddaoudi, Ph.D. Julie P. Harmon, Ph.D. Date of Approval: June 9, 2006 Keywords: crystal engineering, coordination polymers, flexibility, secondary building units, supramolecular isomerism, topology Copyright 2006 Zhenqiang Wang

PAGE 2

Acknowledgments First and foremost, I would like to express my sincere appreciation to my advisor Dr. Michael Zaworotko for his guidance and suggestions throughout the course of this work. In addition, I would like to thank Dr. Mohamed Eddaoudi and Dr. Julie Harmon, members of my supervisory committee, for their generous supports. Finally, I would like to thank Dr. Victor Kravtsov, Dr. Rosa Walsh, Greg McManus and Jason Perman for their help on crystallography. I would also like to thank John Perry, Tanise Shattock, Joanna Bis, Miranda Cheney, Daniel Warrensford, David Weyna, Sheshanka Kesani, and other group members for their helpful comments and encouragements. Thank you again for your support!

PAGE 3

i Table of Contents List of Tables iv List of Figures v Abstract viii Chapter 1 Introduction 1 1.1 Preamble: Crystals 1 1.1.1 Crystals and th e Science of Crystals 1 1.1.2 The Crystal as Molecular Entity 2 1.1.3 Solid State Chemistry 3 1.2 Supramolecular Chemistry 5 1.2.1 History and Scope 5 1.2.2 Supramolecular Chemistry in Solution 7 1.3 Crystal Engineering: a Supramolecular Perspective 10 1.3.1 History and Scope 10 1.3.2 Intermolecular Interactions 12 1.4 Metal-Organic Networks 13 1.4.1 History and Scope 13 1.4.2 Design Principles 14 1.4.3 Structural Analysis of Metal-Organic Nets 17 Chapter 2 Metal-Organic Networks Ba sed upon Rigid Angular Dicarboxylates 19 2.1 Introduction 19

PAGE 4

ii 2.1.1 Secondary Building Units (SBU’s) 19 2.1.2 Supramolecular Isomerism 22 2.2 Metal-Organic Networks from SBU I and BDC or Its Derivatives 24 2.2.1 Nanoballs 25 2.2.2 Two-Dimensional Tetragonal Sheets and Kagom Lattices 28 2.2.3 Three-Dimensional Structures and Some Predicted Structures 31 2.3 Metal-Organic Networks from SBU I and 1, 3-Adamantanedicarboxylate 36 2.4 Metal-Organic Networks from SBU II and BDC or Its Derivatives 38 2.5 Experimental 43 2.5.1 Syntheses 43 2.5.2 Characterizations 45 Chapter 3 Metal-Organic Networks Based upon A More Flexible Dicarboxylates 49 3.1 Introduction 49 3.1.1 Rigidity vs. Flexibility 49 3.1.2 Conformational Analysis of Organic Ligands: A CSD Survey 51 3.2 Tetragonal Sheets from Tetrafl uoro-1, 3-benzenedic arboxylate (TFBDC) 54 3.2.1 1D Structures 54 3.2.2 Guest-Dependent Opening/Closin g of Two Types of Cavities in 2D Structures 56 3.2.3 Functionalization of Inter-layer Cavities in 2D Structures 63 3.3 Experimental 67 3.3.1 Syntheses 67 3.3.2 Characterizations 68

PAGE 5

iii Chapter 4 Conclusions and Future Directions 75 4.1 Summary and Conclusions 75 4.2 Future Directions 78 References 80

PAGE 6

iv List of Tables Table 1.1 A comparison of intermolecular forces. 12 Table 1.2 Comparison of coordination sequences of diamond and lonsdaleite nets. 18 Table 2.1 Comparison of chemical and structural information for compound 2 3 and their parent compound. 31 Table 2.2 Crystallographic data for the two three-dimensional hypothetical structures H1 and H2 34 Table 2.3 Crystallographic data for compounds 1 ~ 8 46 Table 3.1 Crystallographic data for compounds 9 10a ~ c 11 12 69

PAGE 7

v List of Figures Figure 1.1 Rebek’s molecular capsules: the “sof tball” (left) and the cylinder (right). 8 Figure 1.2 Fujita’s octahedral M6L4 cage (left) and Raymond’s tetrahedral M4L6 cage (right) 9 Figure 1.3 Number of citations containing th e key word “coordination polymers” in titles or abstracts in the past 16 years. 14 Figure 1.4 Representative examples of organic ligands used in metal-organic networks. 15 Figure 1.5 Node-and-spacer” representations of metal-organic networks. 16 Figure 1.6 Vertex-linked Polygons or Polyhedra” (VLPP) representations of metal-organic networks. 16 Figure 2.1 Metal-organic networks based upon 4, 4’-bipyridine and mono-metal centers. 19 Figure 2.2 Four commonly encountered sec ondary building units (SBUs) in metalorganic networks. 20 Figure 2.3 Distribution of the paddle-wheel SBUs I deposited in the Cambridge Structural Database (CSD) among various transition metal ions. 21 Figure 2.4 Interpretations of SBU I and II from both node-and-spacer and VLPP perspectives. 21 Figure 2.5 Schematic illustrations of fi ve supramolecular isomers based upon SBU I and BDC. 23 Figure 2.6 Four possible configuratio ns associated with BDC-linked SBU I 24 Figure 2.7 Ball-and-stick and sc hematic representations of nanoballs assembled from SBU I and BDC. 25 Figure 2.8 Crystal structure and crystal packing of 1 27 Figure 2.9 Ball-and-stick representation s of prototypal tetragonal sheet ( a and b ) and Kagom lattice ( c and d ). 28

PAGE 8

vi Figure 2.10 Crystal packing of compound 2 ( a ) and 3 ( b ). 30 Figure 2.11 Crystal structures of USF-1 D ( a ) and CdSO4 net E ( b ). 32 Figure 2.12 Ball-and-stick and schemati c representations of hypothetical structure H1 33 Figure 2.13 Ball-and-stick and schemati c representations of hypothetical structure H2 34 Figure 2.14 Perspective and side views of hypothetical structure H3 in a ball-and-stick mode. 35 Figure 2.15 Perspective and side views of hypothetical structure H4 in a ball-and-stick mode. 35 Figure 2.16 Crystal structures of compound 4 36 Figure 2.17 A predicted cylindrical structure ( H5 ) based upon ADC and SBU I 37 Figure 2.18 1D ladder structure ( a ) and its packing ( b ) in compound 5 39 Figure 2.19 2D layer structure ( a ) and its packing ( b ) in compound 6 40 Figure 2.20 Crystal structures of 7 41 Figure 2.21 2D layer structure ( a ) and the crystal packing ( b ) of compound 8 42 Figure 2.22 Experimental and simulated XPD pattern of 1 48 Figure 3.1 Guest-dependent deformation of a metal-organic network that leads to spin crossover. 50 Figure 3.2 Planes that define the to rsion angles of 4, 4’-bipyridine ( a ) and benzoates/benzoic acids ( b ). 51 Figure 3.3 Histograms showing the distributions of torsion angles for both noncoordinated ( a ) and coordinated ( b ) 4, 4’-bipyridine. 52 Figure 3.4 Histograms showing the distributi ons of torsion angles for noncoordinated ( a ), coordinated ( b ), fluoro-substituted ( c ) and other halogen-substituted ( d ) benzoates/benzoic acids. 53 Figure 3.5 1D zigzag chain structures of the ligand H2TFBDC ( a ) and compound 9 ( b ). 55

PAGE 9

vii Figure 3.6 Crystal structure ( a ) and crystal packing ( b ) of compound 10a 56 Figure 3.7 Crystal structure ( a ) and crystal packing ( b ) of compound 10b 57 Figure 3.8 Crystal structure ( a ) and crystal packing ( b ) of compound 10c 59 Figure 3.9 The open ( a ) and closed ( b ) inter-layer cavities in 10c 60 Figure 3.10 Three axial ligands of SBU I used for the functionalization of inter-layer cavities. 63 Figure 3.11 Crystal structure ( a ) and crystal packing ( b ) of compound 11 64 Figure 3.12 Two hydrogen-bonding motifs occu rred between ethanol guests and the frameworks in 11 65 Figure 3.13 Crystal structure ( a ) and packing ( b ) of compound 12 66 Figure 3.14 TGA trace of compound 10a 71 Figure 3.15 TGA trace of compound 10b 72 Figure 3.16 TGA trace of compound 10c 72 Figure 3.17 TGA trace of compound 11 73 Figure 3.18 Experimental and simulated XPD of compound 10a 73 Figure 3.19 Experimental and simulated XPD of compound 10b compared with simulated XPD of 10a 74 Figure 3.20 Experimental and simulated XPD of compound 10c 74

PAGE 10

viii Metal-Organic Networks Base d upon Dicarboxylato Ligands Zhenqiang Wang ABSTRACT Network structures based upon metal-orga nic backbones represent a new class of functional materials that can be rationall y constructed by employing the concepts of supramolecular chemistry and crystal engineer ing. The modularity of design strategies, the diversity of prototypal structures, and the dynamic features of networks have afforded great advantages over traditional materials sy ntheses. The research presented in this thesis is primarily concerned with developi ng an in-depth understanding of the basic principles that govern the supramolecular behaviors of metal-organic networks and gaining an experimental control over the stru cture and function of these new classes of hybrid materials. The use of rigid and angular organic ligands along with transition metal clusters gives rise to a wide variety of novel me tal-organic architectures ranging from zerodimensional nanostructures to three-dimensi onal frameworks. Conformational analysis of these structural models suggests the geometric foundations for the existence of superstructural diversity. Controlled crys tallization experiment s further reveal the synthetic factors that might determine the formation of supramolecular isomers. Careful selection of more labile organic components, on the other hand, leads to

PAGE 11

ix flexible metal-organic networks exhibiting dy namic characteristics that have not been observed in their rigid counterparts. The gues t-dependent closing/opening of cavities and the ease of fine-tuning their chemical environm ents demonstrate the effectiveness of such a strategy in the context of genera ting tailored functional materials.

PAGE 12

1 Chapter 1 Introduction 1.1 Preamble: Crystals 1.1.1 Crystals and the Science of Crystals “ These were little plates of ice, very flat, very polished, very transparent, about the thickness of a sheet of rather thick paper...but so perfectly formed in hexagons, and of which the six sides were so straight, and the six angles so equal, that it is impossible for men to make anything so exact. ” Ren Descartes, 16351 For centuries, the extraordinary beauty of crystals2 has captivated people’s fondness and curiosities. Snowflakes, diamonds and common salt are familiar examples of crystals and their distinctive and beautiful patterns have sparked the interest of writers, poets, photographers, philosophers, mathematicians, and scientists throughout history. Although it is almost impossible to determine at what point in the history did mankind begin their fascination with crystals, it has been known that as early as 135 B.C., ancient Chinese had recorded their observations of snow as “always six-pointed”. The first attempt to fundamentally understand the nature of a crystal, i.e., to relate the external form or shape of a crystal to its underlying structure, was made in 1611 by Johannes Kepler, who speculated that the hexagonal close-packing of spheres may have something to do with the morphology of snow crystals.3 Robert Hooke went on to ex tend this idea to other crystals and show how different shapes of crystals--rhombs, trap ezia, hexagons, etc.-could arise from the packing together of spheres and globules. Ren Just Hay (also known as Abb Hay, 1743-1822) discovered that crystals of the same composition possessed the same internal nucleus, even though their external forms differed. The now

PAGE 13

2 banished molcules intgrantes that he persistently used in his original treatise4 eventually transformed into “unit cells”, the contemporary term to describe the smallest building block of a crystal, and for this reason, he is arguably regarded by some as the father of modern crystallography. The modern development of the science of crystals, however, began after the discovery of X-ray by W. C. Rntgen in 1895 and, in particular, when Max von Laue demonstrated in 1912 that passage of a narrow beam of X-ray through a crystal of copper sulfate resulted in a pattern of spots on a phot ographic plate due to the diffraction of very short waves by the crystal. Shortly thereafter, W. H. Bragg (1862-1942) and his son, W. L. Bragg (1890-1971) utilized and extended this diffraction method to determine the arrangement of the atoms within such simple crystalline materials as NaCl, pyrite, fluorite, and calcite. By examining the pattern of X-rays diffracted by various crystals, the Braggs were able to establish the fundam ental mathematical relationship between an atomic crystal structure and its diffraction pa ttern--the Bragg’s Law. Since that time, the improvement of the techniques of X-ray crystallography has resulted in an enormous increase in the store of scientif ic knowledge of matter in the solid state, with consequent impact on the development of the sciences of physics, chemistry, biology, and geology. Today, hundreds of thousands of crystal structures have been determined for a wide spectrum of molecules ranging from simple inorganic and organic compounds to complex multi-chained proteins and nucleic acids.5 1.1.2 The Crystal as Molecular Entity

PAGE 14

3 When cooled sufficiently, the vast majority of substances form one or more crystalline phases, where the atoms, molecule s or ions interplay with each other via different kinds of chemical interactions such as covalent, ionic, and metallic bonds. The resulting entities exhibit a regular repeating arra y of atoms, molecules or ions that extend in three dimensions. Mathematically, these chemical building units can be represented by their centers of gravity and a crystal can be simplified as a threedimensional lattice based upon an infinite number of points orderly arranged in space and entirely related by symmetry. In reality, however, most atoms, molecules and ions are anisotropic and real crystals often feature defects or irregularities in their ideal arrangements. Interestingly, many of the mechanical, electrical and chemical properties of real crystalline materials are critically dependent upon such defects. 1.1.3 Solid State Chemistry Solid state chemistry is concerned with the synthesis, structure, properties and applications of solid materials. Whereas some aspects of glasses and other amorphous solids are also quite relevant to solid state ch emistry, crystalline materials are generally of paramount importance in most cases, and accordingly crystals and crystallography have been often associated with this subject. Solid state compounds represent an important class of materials with high technological relevance and they have been widely used as key devices, such as superconductors, fast ion conductors, magnets, non-linear optics, luminescent materials, laser materials, and hydrogen storage materials, just to name a few. Traditional solid state chemistry usually involves the study of inorganic materials including naturally occurring minerals, and large majority of these compounds are non-

PAGE 15

4 molecular, i.e., their structures are determin ed by the manner in which the atoms and ions are packed together in three dimensions. Although the types of elements involved in these non-molecular solids are relatively limited in the periodic table, the structural diversity exhibited by the materials is nevertheless amazingly striking. For example, of the nearly 25,000 known binary compounds there exist at least 100 structure types, and among the estimated 100,000 possible ternary phases, of which only about 5% have been investigated, already more than 700 structure types have been identified and several thousand more might be expected; not to even mention yet those of quaternary and quinary systems. Historically, the discovery of new solid state compounds, especially those with novel structure types, has largely relied on serendipitous, or at best, empirical processes. The synthesis of extended structure compounds usually takes place at the range of 500oC to 2,500oC and at such high temperatures the control over structure and reactivity is inevitably diminished to a considerable degree. For a long period of time solid state synthesis has been decried as “shake and bake” or “heat and beat”, and there is a widelyheld belief that the preparation of new solid-state compounds based on rational design is not possible. However, this si tuation is gradually being changed and a number of efforts have been devoted to establish a priori synthetic strategies for solid state materials. In particular, two different methods, one of which considers constructing a free energy landscape assisted by computational modeling 6 while the other takes advantage of the concept of molecular building blocks,7 point at the future direction of solid state synthesis: materials by design.

PAGE 16

5 1.2 Supramolecul ar Chemistry 1.2.1 History and Scope “ The relations between toxin and its antitoxin are strictly specific... For this reason it must be assumed that the antipodes enter into a chemical bond which, in view of the strict specificity is most easily explained by the existence of two groups of distinctive configuration of groups which according to the comparison made by Emil Fischer fit each other ‘like lock and key’. ” Paul Ehrlich, 1 9088 Although Nature has established its own supramolecular chemistry through billions of years of evolution, the most elegant examples including enzyme-substrate interactions and DNA double helix formation and replication, that of mankind can be only traced back to the late 19th and early 20th century when Paul Ehrlich, the founder of modern chemotherapy, first introduced the idea of receptor while recognizing that molecules do not act if they do not bind.9 It was Emil Fischer, however, who expressively enunciated the concept of binding selectivity and geometrical complementarity of molecular recognition in his celebrated “lock and key” model.10 In 1948, H. M. Powell described a series of what he called clathrates--inclusi on compounds formed when small molecules, such as methanol, hydrogen sulphide or sulphur dioxide, are completely enclosed in cavities formed by a “host” such as a hydroquinone network.11 In the 1960’s, Charles J. Pedersen showed that some cyclic polyethers, which he termed crown ethers, bind the alkali ions (i.e., Li+, Na+, K+, Rb+, and Ce+) strongly and selectively.12,13 The selectivity is essentially determined by the degree of geometrical match between the cations and the cavities of crown ethers into which the sphe rical metal ions will fit. This discovery represents a breakthrough towards the ambition of many chemists (of then and today!): designing and synthesizing organic molecules that mimic the extraordinary functions of biological systems (e.g., enzymes, DNA, etc). Jean-Marie Lehn and Donald J. Cram

PAGE 17

6 subsequently each developed increasingly sophisticated organic compounds containing holes and clefts that bind cationic as well as anionic and neutral species even more efficiently and selectively.14-18 With this work, Pedersen, Lehn and Cram, who also shared the Nobel Price of Chemistry in 1987, laid the foundations of what is today one of the most active and expanding fields of chemical research--supramolecular chemistry. Thus, supramolecular chemistry, as coined by Lehn, may be defined as “chemistry beyond the molecule”, i.e., it is the chem istry of molecular assemblies and of the intermolecular bond. 19-20 Two main tenets, molecular recognition and supramolecular function lie at the center of understanding the concepts of supramolecular chemistry.21 Whereas mere binding doesn’t necessarily in fer recognition, molecular recognition is generally regarded as a patterned process involving a structurally well-defined set of intermolecular interactions: binding with a purpose.20 It thus implies the storage, at the supramolecular level, of molecular information associated with their electronic properties, size, shape, number, and arrangement. There are generally two partially overlapping areas encountered in supramolecular chemistry: 1) supermolecules well-defined, discrete oligo molecular species that result from the intermolecular association of a few components; 2) supramolecular assemblies poly molecular entities that result from the spontaneous association of a large undefined number of components into a specific phase having more or less well-defined microscopic organization and macroscopic characteristics. More recently, suprasupermolecules a new class of organized entities that bridge the gap between the above two, has been delineated.22, 23 Therefore “supramolecular chemistry” is a broad term that concerns the chemistry of all types of

PAGE 18

7 supramolecular entities including the well-defined supermolecules, the extended, more or less organized, polymolecular associations, and their appropriate combinations. The breadth and especially the unifying power of the perspectives and conceptual framework of supramolecular chemistry developed by Lehn as well as other researchers have become progressively more and more evidenced. In fact, over the past few decades, supramolecular chemistry has fueled numerous developments at the interfaces with biology, physics, and engineering, thus giving rise to the emergence and establishment of supramolecular science and technology.24 Nevertheless, although in principle the molecular recognition events occurring at various levels exhibit similar characteristics, it is perhaps still quite appropriate to note the significantly different aspects of supramolecular chemistry that takes place among different physical states of matters. Notably, the early development of host-guest chemistry was originated from solutions and the fundamental principles governing solution behaviors of molecular aggregates are relatively better understood compared to those in the solid state. 1.2.2 Supramolecular Chemistry in Solution The pioneering examples of synthetic receptors featuring macrocyclic shapes developed by Pedersen, Lehn and Cram have established the field of host-guest chemistry. However, two main drawbacks are inherently associated with this early approach: 1) the construction of host molecules almost exclusiv ely relies upon the tedious and irreversible covalent synthesis of a single structure; 2) the sizes of holes or cavities exhibited by the host molecules are relatively small, thus li miting their recognition capabilities to small guest species such as alkali ions. Accordingly, an alternative synthetic strategy that takes

PAGE 19

8 advantage of multiple building blocks, reversible self-assembly process, and “weak” intermolecular forces, such as hydrogen bonds and metal coordination, is highly desirable. The first examples of self-assembled hydrogen-bonded molecular receptors were described by J. Rebek, Jr. in the 1990’s.25-30 Two self-complementary molecules assemble to form dimers via an array of hydrogen bonds, giving rise to molecular capsules enclosing either spherical/semispherical or cylindrical cavities (Figure 1.1). Depending on the size and shape of the monomeric species, a wide variety of guests can be included inside the capsules and quite often simultaneous encapsulation of more than one guest molecule has been observed. The electronic and geometric restrictions by the confined space result in some unique and interesting behaviors of the guest molecules. For example, the accommodation of p -quinone and 1, 3-cyclohexadiene inside the “softball” capsule dramatically accelerates the Diels-Alder reaction, 31 whereas the unusual associations of pairs of guests within the cylindrical capsule lead to the discovery of “social isomerism”. 32 Nevertheless, since only relatively weak intermolecular interactions, i.e., hydrogen bonds, are involved, the formation and disassociation of the Figure 1.1. Rebek’s molecular capsules: the “softball” (left) and the cylinder (right).

PAGE 20

9 capsules is reversible and the systems reach thermodynamic equilibrium rapidly under mild conditions in solution. Thus they requ ire analytical methods that operate on the same timescale (such as NMR spectroscopy and electrospray mass spectrometry). Furthermore, the inclusion complexes do not survive purification by chromatography and few of these encapsulation complexes have been characterized by X-ray crystallography. Therefore, relati vely stable (that is, longer lived but still reversibly formed) encapsulation complexes should be explored by using the stronger forces of metal–ligand interactions. In this regard, Fujita has taken advantage of pyridine-based monodentate ligands and cis -capped square planar transition metal units and developed a series of cationic supramolecular metal-organic aggregates based upon what he has termed the “molecular paneling” approach.33,34 In particular, a M6L4 type octahedral cage (Figure 1.2) has been shown to possess a cavity large enough to accommodate up to four guest species, which can be used as ideal molecular chambers for mediating chemical reactions such as Diels-Alder reaction, [2+2] cycloadition, and Wacker oxidation.35,36 Most recently, it was Figure 1.2. Fujita’s octahedral M6L4 cage (left) and Raymond’s tetrahedral M4L6 cage (right).

PAGE 21

10 demonstrated that an aqueous solution of the octahedral M6L4 cage induces highly unusual regioselectivity in the Diels-Alder coupling of anthracene and phthalimide guests, promoting reaction at a terminal rather than central anthracene ring.37 Raymond’s group uses an alternative strategy, namely, bidentate chelating ligands and octahedral transition metal units, to direct the assembly of a M4L6 type anionic tetrahedral cage (Figure 1.2) and other related molecular containers.38-40 The most salient feature of this approach is the presence of homochirality as a result of trisbidentate coordination at each metal center that leads to either or configuration. The chiral environment of the cavity turns out to si gnificantly stabilize otherwise short-lived organometallic intermediates and therefore mediate their reactivity toward other substrates.41 1.3 Crystal Engineering: a Supramolecular Perspective 1.3.1 History and Scope Although the roots of crystal engineering can be traced at least as far back as the 1930’s, when Pauling defined the chemical bond in both covalent and non-covalent senses, 42 the term “crystal engineering” was initia lly introduced by Pepinsky in 1955 in an effort to solve the “phase problem” in crystallography.43 However, it was Schmidt who first systematically formulated this idea in the 1970’s in the context of topochemical reactions. He and his co-workers found that the photo-reactivity of dimerizable olefins, such as substituted cinnamic acids, is critically dependent upon the crystal packing of the molecules; in other words, solid state reactivity is a supramolecular property and is characteristic of an entire assembly of molecules. Schmidt therefore proposed an

PAGE 22

11 “empirical” strategy based upon the understanding of intermolecular forces as an approach for the development of organic solid state chemistry, namely, crystal engineering .44 In the late 1980’s and early 19 90’s, Desiraju intensiv ely investigated weak intermolecular forces such as C-H•••X and C-H••• and the roles they play in the design of organic solids.45-48 Thanks to his efforts, these interactions are now widely accepted as an important part of the whole spectrum of hydrogen bonds that are crucial for crystal packing of molecules. In his monograph titled “Crystal Engineering: the Design of Organic Solids”, Desiraju has defined crystal engineering as “the understanding of intermolecular interactions in the context of designing new solids with desired physical and chemical properties”.49 The elucidation of the concept supramolecular synthon ,50 a structural unit within a supermolecule which can be formed and/or assembled by known or conceivable synthetic operations involving intermolecular interactions, has afforded reliable strategies for designing and exploiting crystal structures. Indeed, when crystals are conceived as supermolecules par excellence 51, 52 it is perhaps conceptually instructive to consider crystal engineering as synonymous with supramolecular synthesis in solid state. Interestingly, almost coincident with the establishment of design principles for organic solids, the development of metal-organic compounds and coordination polymers was mainly pushed forward by Robson using a modular “node-and-spacer” approach in the late 1980’s and early 1990’s.53-55 However, these two seemingly isolated areas were not unified under the same context until 2001 when Zaworotko expl icitly delineated their conceptual similarities.56 Today crystal engineering has become a paradigm not only for

PAGE 23

12 constructing organic and metal-organic solids, but also for the design of organometallic and inorganic structures. 1.3.2 Intermolecular Interactions Just as molecular synthesis (organic synthesis in particular) is concerned with the breaking and construction of intramolecular covalent bonds, supramolecular synthesis (crystal engineering in particular) is dictated by the re-organization of intermolecular non-covalent interactions. The existence of attractive and repulsive intermolecular forces and their dynamic balance in crystalline so lids are responsible for holding individual molecules in an ordered array and maintaining particular crystallographic symmetries. Depending upon their distance-dependence and their directionality, intermolecular interactions can be classified as London dispersion, dipole-dipole interaction, stacking, hydrogen bond, and coordination bond, with some overlap between them (Table 1.1) Table 1.1 A Comparison of Intermolecular Forces Force Strength (kJ/mol) Characteristics Examples Coordination bond 50-200 Occurs between metal ions and molecules with lone pairs cis -platin hemoglobin Hydrogen bond 1-160 Occurs between molecules with O-H, N-H, F-H and C-H bonds carboxylic dimers DNA stacking <50 Occurs between electrondelocalized systems graphite Dipole-dipole 3-4 Occurs between polar molecules acetone London dispersion 1-10 Occurs between all molecules; strength depends on size, polarizability CO2, He In classical or Werner type coordination compounds, ligands bind to metal ions almost exclusively via donating their lone pair of electrons, resulting in relatively strong

PAGE 24

13 metal-ligand binding. One would argue such an interaction should be regarded as a type of covalent linkage because of the strength criterion; however, if considering their donoracceptor pattern as well as liable and revers ible nature, coordination bonds exhibit more intermolecular characteristics and therefore have been enormously exploited in the context of crystal engineering of functional solids. 1.4 Metal-Organic Networks 1.4.1 History and Scope Metal-organic networks, also known as metal-organic frameworks, represent a new class of compounds consisting metal ions linked by organic bridging ligands. The structures resulting from metal-ligand linkages can be discrete zero-dimensional (0D) molecular complexes or infinite one-dimensional (1D), two-dimensional (2D) or threedimensional (3D) architectures. Whereas the term “coordination polymers” is more commonly referred to the latter, “metal-organic networks” and “metal-organic frameworks” are applicable in a broader context and are interchangeable in most cases. One of the very first examples of metal-organic networks that have been structurally characterized appeared in 1943, 57 although similar studies can be traced back to the 1930’s. The area of coordination polymers was initially reviewed in 1964 with an emphasis on the preparations. 58 In the early investigations, Prussian Blue based on FeCN-Fe linkages and its analogues were perhaps among the most systematically studied. Surprisingly, however, the field of metal-organic networks was not prospering until the late 1980’s when Robson initiated the now famous “node-and-spacer” approach55 to incorporate both transition metal ions of well-defined coordination geometries and rod-

PAGE 25

14 like organic ligands in the design of framework materials. Subsequently, the work by Zaworotko, 56, 59-61 Yaghi, 62-65 and others66-70 substantially contributed to the field and it is now so rapidly developing that the number of coordination polymeric compounds has witnessed an exponential growth in the past few years (Figure 1.3). In Robson’s original node-and-spacer approach, the nets were usually constructed from organic-based linear spacers and me tal-cation nodes, which could be square, tetrahedral, octahedral, etc. This strategy, however, can be conveniently extended to a much broader context where both metal centers and organic ligands can appropriately function as either nodes or spacers.71 Figure 1.4 illustrates some representative examples of organic ligands with linear/angular, trigonal, and tetrahedral shapes. 1.4.2 Design Principles Figure 1.3 Number of citations containing the key word “coordination polymers” in titles or abstracts in the past 16 years (source: SciFinder Scholar, 07/15/2006).

PAGE 26

15 Metal-organic networks exemplify how crystal engineering has become a paradigm for the design of new supramolecular materials. Since the structures are composed of at least two components (i.e., metal ions and organic ligands), it appears clear that such components can be pre-selected for their ability to self-assemble. The network structures can therefore be regarded as examples of blueprints for the construction of networks that, in principle, can be generated from a diverse range of chemical components, i.e., they are prototypal examples of modular frameworks. There exist two different strategies that have been widely used to direct the syntheses of metal-organic networks. The first is the above mentioned node-and-spacer approach in which the building blocks are simplified as topological points and lines and the nets are represented in their appropriate combinations. Wells was regarded as the pioneer of this approach thanks to his systematic investigations on the geometric basis of Linear Angular N CO2 4,4’-Bipyridine 1,4-Benzenedicarboxylate Nicotinate 1,3-Benzenedicarboxylate Trigonal Tetrahedral Tri(4-pyridyl)triazine 1,3,5-Benzenetricarboxylate HMTA 1,3,5,7-Adamantanetetracarboxylate Figure 1.4 Representative examples of organic ligands used in metal-organic networks.

PAGE 27

16 crystal chemistry.72-74 Although Wells initial work was primarily focused upon inorganic crystalline compounds, Robson extrapolated this method into the realm of metal-organic compounds and coordination polymers.55 As revealed by Figure 1.5, the node-and-spacer approach has afforded a diverse array of metal-organic architectures ranging from 0D discrete nanostructures to 3D infinite networks, some of which have no inorganic analogues. a ) b ) c ) d ) e ) f ) g ) h ) i ) j ) k ) Figure 1.5 “Node-and-spacer” representations of metal-organic networks: a ) 0D nanoball; b ) 1D zigzag chain; c ) 1D helix; d ) 1D ladder; e ) 2D bilayers; f ) 2D square grid; g ) 2D honeycomb; h ) 3D (10,3)-a net; i ) 3D diamondoid net; j ) 3D primitive cubic net; k ) 3D NbO net. a ) b ) c ) d ) e ) Figure 1.6 “Vertex-linked Polygons or Polyhedra” (VLPP) representations of metal-organic networks: a ) 0D nanoball; b ) 3D ( 10,3 ) -a net; c ) 3D diamondoid net; d ) 3D p rimitive cubic net; e ) 3D NbO net.

PAGE 28

17 Another approach, also based upon geometric principles, takes into account the specific shapes of the building blocks and represents nets as being sustained by vertexlinked polygons or polyhedra (VLPP).75-78 Notably nets shown in Figure 1.5 can be equally well represented in the VLPP fashion (Figure 1.6). Whereas the node-and-spacer approach appears more straightforward in cases involving linear spacers, VLPP perspective more fundamentally reveals the geometrical relationship between various building units. 1.4.3 Structural Analysis of Metal-Organic Nets The employment of geometrical pr inciples not only facilitates the development of reliable design strategies for the synthesis of metal-organic compounds, but also affords an indispensable tool for the recognition and interpretation of some perplexing nets and prediction of novel nets. In this contex t, Wells introduced a simple notation ( n p ) to describe nets, where n is the number of edges of polygons present in the net and p the connectedness of the vertices.72 For example, the planar square grid (Figure 1.5 f ) can be represented as (4, 4) and the symbol (10, 3) implies a 3-connected net based upon 10membered rings (Figure 1.5 h ). Although Wells notation is still widely accepted in the literature, it also has some limitations because of its over-simplification. For example, the above mentioned symbol (10, 3) in fact represents at least seven different 3D nets that are topologically related but distinct. Therefore a more informative system based upon Schlfli symbols, namely, vertex symbols, has been proposed by O’Keeffe.79 In his terminologies, O’Keeffe defined rings as shortest closed circuits without any shortcuts for each angle at a vertex and used

PAGE 29

18 Aa• Bb•[…]• Mm to depict the connectivity of nets, where A B …, M are numbers that represent the ring size and a b …, m are numbers of the respective rings meeting at that angle (subscript “1” is omitted). Thus 105•105•105 indicates there are five 10-rings at each of the three angles, and 102•104•104 suggests two 10-rings at the first angle and four 10rings at each of the other two, whereas in Wells notations, these two nets are designated as (10, 3)a and (10, 3)b respectively. Note that sometimes the subscripts are omitted and the short vertex symbols in these two examples can then bo th be written as 103. However, it should be pointed out even O’Keeffe’s vertex symbol is not entirely satisfying, as exemplified by the case of 4-connected diamond and lonsdaleite (hexagonal diamond) nets. Although belonging to two distinct nets that exhibit significantly different connectivities, these two nets display identical vertex symbols (62•62•62•62•62•62 for both). It thus follows that a more rigorous way of describing detail topological information of nets is necessary and a practical solution is to take into account the concept of topological neighbors--a k th neighbor of a vertex is the one for which the shortest path to that vertex consists of k edges.79 Each different kind of vertex in a net has then associated with it a coordination sequence which is the sequence of n1, n2, …, nk, … where nk is the number of k th topological neighbors. Only by considering coordination sequences, for example, it is possible to distinguish between diamond and lonsdaleite nets (Table 1.2). Table 1.2 Comparison of coordination sequences of diamond and lonsdaleite nets k 1 2 3 4 5 6 7 8 9 10 Diamond 4 12 24 42 64 92 124 162 204 252 Lonsdaleite 4 12 25 44 67 96 130 170 214 264 Difference 0 0 1 2 3 4 6 8 10 12

PAGE 30

19 Chapter 2 Metal-Organic Networks Based Upon Rigid Angular Dicarboxylates 2.1 Introduction 2.1.1 Secondary Building Units (SBUs) Crystal engineering, and in particular, design strategy based upon geometric principles, provides a successful approach to the synthesis of metal-organic networks. Enormous progress has been made in the past decades, giving rise to a large number of aesthetically pleasing and potentially functional coordination polymers.56,63,65,68-70,80 For example, the self-assembly of 4, 4’-bipyridine, a linear spacer, and single-metal ions has afforded, depending upon the coordination geometry of metal ions, a wide variety of superstructures (Figure 2.1).56 a ) c ) b ) d ) Figure 2.1 Metal-organic networks based upon 4, 4’-bipyridine and mono-metal centers: a ) 1D chain; b ) 1D ladder; c ) 2D square grid; d ) 3D diamondoid net.

PAGE 31

20 Although this “M—N” (M being referred to single metal ion and N pyridyl nitrogen) based approach proves to be extremely successful, it is nevertheless inherently associated with a number of issues that could potentially be of weakness, especially in the context of porous materials. For instance, the single “M —N” interactions are less rigid and in most cases, the pyridyl rings are subject to free rotation around the metal centers, thus limiting the degree of control over the final structures; the presence of anionic species due to the cationic nature of the frameworks significantly reduces available free space of the structure; attempts to evacuate/exchange guests within the pores often result in collapse of the host framework. In this context, a so-called “secondary building units (SBUs)” strategy has been employed to overcome the above problems. 65 The concept was originally from zeolite chemistry where SUBs are referred to the comm on structural motifs occurring in various tetrahedral frameworks.81 Yaghi and Eddaoudi extended this idea to metal-organic chemistry and re-defined SBUs as molecular complexes or metal clusters that have welldefined and highly symmetric coordination geometries. Of particular interest are the carboxylate-based metal clusters since the me tal ions are locked into positions by the carboxylates (Figure 2.2). Expansion of SBUs by multifunctional ligands, such as 1, 4benzenedicarboxylate and 1, 3, 5-benzenetri carboxylate, allows for the construction of neutral open frameworks of high structural stability.65 I II III IV Figure 2.2 Four commonly encountered secondary building units (SBUs) in metal-organic networks.

PAGE 32

21 In this thesis, we focus upon exploiting SBU I and II both of which have a general formula of M2(RCOO)4 (axial ligands omitted). SBU I, a paddle-wheel dimetal tetracarboxylate, has been well known for decades because of its ubiquity and easy accessibility. It is perhaps the most frequently used SBU and is present in over 1,300 crystal structures deposited in the Cambridge Structural Database (CSD).82 As revealed by Figure 2.3, the paddle-wheel pattern is most commonly seen among metals such as Cu, Rh, Ru, and Mo, etc. SBU II on the other hand, is far less common than I and remains largely unexploited in the crystal engineering of metal-organic networks. Nevertheless, I and II are related in that both can be simplified as 4-connected nodes according to nodeand-spacer approach while they are characterized by their distinct shapes from VLPP perspective (Figure 2.4). Figure 2.3 Distribution of the paddle-wheel SBUs I deposited in the Cambridge Structural Database (CSD) among various transition metal inos. Figure 2.4 Interpretations of SBU I and II from both node-and-spacer and VLPP perspectives.

PAGE 33

22 2.1.2 Supramolecular Isomerism In molecular chemistry, it is a well known phenomenon that some elements and compounds exist in more than one form and the presence of various molecular isomers is due to different arrangements of atoms, which can be exemplified by the four different forms of carbon, i.e., diamond, graphite, C60, and carbon nanotube. A direct analogy can be drawn in supramolecular chemistry where some molecules are capable of interacting with their partners in different ways, giving rise to a diverse range of superstructures. Zaworotko first recognized superstructural diversity in metal-organic networks in 1997 where he observed three supramolecular isomers (two of which are schematically shown in Figure 1.5d and 1.5e) resulting from T-shaped metal centers linked by a conformationally labile bidentate ligand in a 1:1.5 stoichiometry. 60 He subsequently defined supramolecular isomerism as “the existence of more than one type of network superstructure for the same molecular building blocks”.56 Indeed, as illustrated by Figure 1.5, other pairs of nets can also exhibit similar supramolecular isomerism: zigzag chain vs. helix and honeycomb vs. (10, 3)-a net, for example. The existence of supramolecular isomerism might be seen as a problem from a design perspective since it necessarily implies the difficulty of control over final structures. In this regard, a detail understanding of the factors that could potentially affect the outcome of crystallization, including solv ent polarity, templates, and temperatures, is necessary in order to facilitate the selec tive formation of one isomer over the others. Ironically, it is also possible to view supramolecular isomerism as an opportunity because gaining a better and more fundamental understanding of the factors that influence crystal nucleation and growth will undoubtedly improv e the ability to engineer crystalline solids.

PAGE 34

23 In addition, if one considers that the bulk properties of crystalline solids are as critically dependent upon the distribution of molecular components within the crystal lattice as the properties of its individual molecular components, it is perhaps quite appropriate, from a material perspective, to regard the occurrence of supramolecular isomers as a huge bonus. In fact, each of the four carbon polymorphs represents an extremely important class of materials in both academic and industrial areas. Previous work from our group, which focuses upon Cu(II)/Zn(II)-based paddlewheel SBUs I and angular spacer 1,3-benzenedicarboxylate (BDC), has afforded an ideal system for the investigation of supramolecular isomerism.75-76, 83-84 Depending upon various crystallization conditions, such as solvents, templates (molecules that might or might not be directly involved in the final structures but participate in some way during the crystallizations), and axial ligands, a total of five supramolecular isomers have been A B C D E Figure 2.5 Schematic illustrations of five supramolecular isomers based upon SBU I and BDC: A) nanoball; B) tetragonal sheet; C) Kagom lattice; D) USF-1; E) CdSO4 net.

PAGE 35

24 isolated: 0D nanoballs ( A ), 2D tetragonal sheets ( B ) and Kagom lattices ( C ), and 3D USF-1 net ( D ) and CdSO4 net ( E ) (more supramolecular isomers are expected: see the discussions in section 2.2.3). 2.2 Metal-Organic Networks from SBU I and BDC or Its Derivatives: The fact that 1, 3-benzenedicarboxylate (BDC) is a rigid and angular bridging ligand subtending an angle of 120o has made it an extremely versatile building block. In particular, the presence of two carboxyl groups at the metapositions affords a unique opportunity for the investigation of supramol ecular isomerism. For example, if one considers each individual paddle-wheel SBU I along with the four BDC moieties that are attached to it, it should be noted that, in prin ciple, there exist four possible configurations in which one of the following situat ions is applicable: 1) all four metacarboxyl groups are facing down (or up); 2) two adjacent metacarboxyl groups are facing down; 3) two opposite metacarboxyl groups are facing down; and 4) three of the four metacarboxyl groups are facing down (Figure 2.6). For the sake of simplification, we will designate these as “4D”, “1, 2-D”, “1, 3-D”, and “3D” respectively. As will become apparent “4D” “1, 2-D” “1, 3-D” “ 3D” Figure 2.6 Four possible configurations associated with BDC-linked SBU I : four downs (“4D”), two adjacent downs (“1, 2-D”), two opposite downs (“1, 3-D”), and three downs (“3D”).

PAGE 36

25 below, the very presence of multiple possible arrangements of the molecular building blocks accounts for the occurrence of some supramolecular isomers that are assembled from BDC and SBU I It should be pointed out that a CSD survey reveals that while “1, 2-D” is the predominant conformation and a few other examples exist for “4D”, either “1, 3-D” or “3D” has been hardly observed. 2.2.1 Nanoballs Nanoscale small rhombihexahedra A (cubic phase) are spontaneously formed by the self-assembly of Cu(NO3)2 and H2BDC under appropriate conditions.75 As revealed by Figure 2.7a, 12 SBU I ’s are convergently bridged by 24 BDC moieties, generating 8 triangular windows and 6 square windows. Note that each of the 12 SBUs adopts the same “4D” conformation described above. Surprisingly, a closely related form of the nanoballs, i.e., that of hexagonal symmetry, arises from the identical building blocks under slightly different conditions. This supramolecular isomer of A has an equal number of triangular and square windows and, most importantly, the same “4D” arrangement of SBUs also accounts for its discrete architecture. Degradation of the symmetry of SBU I a ) b ) Figure 2.7 Ball-and-stick and schematic representations of nanoballs assembled from SBU I and BDC: a ) cubic phase; b ) hexagonal phase.

PAGE 37

26 (i.e., the D4h symmetry no long holds) as well as a small distortion of the bridging ligand BDC has been attributed to the formation of this second form. Although visually quite similar, these two compounds can nevertheless be easily distinguished by their connectivities: contrary to the cubic phase which only has one type of node (vertex symbol: (3•3•4•4)12), the hexagonal phase is binodal (vertex symbol: (3•3•4•4)6(3•4•3•4)6). While structure A and its hexagonal isomer are spectacular molecular complexes on their own right thanks to their discrete architectures and nanoscale cavities, it occurs to us that these nanoballs can serve as the building blocks for constructing architectures of higher hierarchy, i.e., they can act as the nodes of much larger infinite networks. For instance, functionalization on the outer surfa ce of nanoballs, which can be realized on either BDC site or SBU site, allows for the cross-linking of adjacent nanoballs. Specifically, several design strategies can be applied: if each nanoball is only linked to two adjacent neighbors, a 1D chain is possible to form; when it is tetrahedrally associated with four neighbors, then a super-diamondoid net is readily accessible; similarly, a primitive cubic or body-centered cubic net can be expected by arranging each nanoball to six or eight adjacent nanoballs, respectively. Indeed, crystals of methoxylated, neutral nanoballs of formula [Cu2(5-MeOBDC)2(MeOH)x(H2O)1.83x]12, 1 result from the modular self-assembly in MeOH under ambient conditions of 70 molecular components: 24 5-MeO-bdc moieties, 24 Cu(II) cations (from copper (II) nitrate), and 22 coordinated solvent (MeOH or H2O) molecules.23 The molecular mass of each molecule is ca 6.9 kDa and their molecular volume is ca 11.5 nm3. It should be noted the nanoballs in 1 exists in the hexagonal form.

PAGE 38

27 The 24 methoxy moieties are disposed at the exterior of the nanoball, and they are capable of coordinating to metal centers th rough their ether oxygen atoms. In fact, two methoxy moieties on each nanoball coordinate to axial sites on adjacent nanoballs in such a way that double cross-linking occurs. As revealed in Figure 2.8, this cross-linking also occurs at the opposite face of each nanoball, thereby generating an infinite 1D chain of nanoballs. The Cu-O distances, averaging 2.26 , are consistent with expected values and the separation between centers of adjacent nanoballs is 2.15 nm. The manner in which the 1D chains pack can be described as hexagonal packing of parallel cylinders (rods) .79 In effect, compound 1 has exemplified the principles of suprasupermolecular chemistry.22-23 a ) c ) b ) Figure 2.8 Crystal structure and crystal packing of 1 : a ) illustration of the methoxy moieties that bridge adjacent nanoballs in blue; b ) 1D chain of nanoballs sustained by double cross-links; c ) hexagonal packing of nanoball chains represented as green cylinders (rods).

PAGE 39

28 2.2.2 Two-Dimensional Tetragonal Sheets and Kagom Lattices Tetragonal sheet B76 and Kagom lattice C83 represent two of the prototypal twodimensional structures that can be assembled from angular ligand BDC and square SBU I While B is based upon the linking of square cavities that are consisted of four SBUs I C is characterized by the presence of triangular windows composed of three SBUs I (Figure 2.9). Both B and C exhibit the undulating nature as a result of the 120o angle subtended by BDC and the presence of such a curvature is critical for the formation of Kagom lattices, whereas topologically related tetragonal sheets have been generated from linear spacers such as 1, 4-benzenedicarboxylate. In contrast to the “4D” configurations that are a ) b ) c ) d ) Figure 2.9 Ball-and-stick representations of prototypal tetragonal sheet ( a and b ) and Kagom lattice ( c and d ). b ) and d ) highlight the structural reason for the existence of both isomers. DD UU DD UU

PAGE 40

29 observed in the discrete nanoball structures, both B and C exhibit the “1, 2-D” dispositions, which presumably account for their two dimensionalities. However, a fundamental question still needs to be raised and answered: what exactly causes the existence of these two isomers since they are built from the same building blocks that adopt similar configurations? Close examination of the two structures suggests that the answer lies in the combined effects of the angular nature of ligand BDC and the reducing symmetry of SBU I Molecular modeling study indicates that SBUs I in the most symmetric forms of B and C possess D2h symmetry, which is lower than its ideal D4h symmetry. In fact, the dihedral angles between the adjacent two planes defined by the carboxyl groups are not identical. If we designate “D” for the plane that contains a metacarboxyl group facing downward, and “U” otherwise (Figure 2.9b, d), then the dihedral angles can be written as either DD (same as UU!) or DU. Notice that in structure B DD is slightly larger than DU, whereas in structure C it is the just opposite. Although such a difference might not seem obvious, it nevertheless dramatically influences the connectivity of the networks and ultimately leads to the generation of two completely different architec tures (see the blue motifs shown in Figure 2.9b and 2.9d for an appreciation of this argument). Whereas the principles of crystal engineering provide reliable blueprints for the construction of prototypal structures, as illustrated by the tetragonal sheets B and Kagom lattices C they also afford a great opportunity to chemically functionalize these model compounds, which might be crucial in terms of improving the material’s performances. As chapter 3 will focus upon a series of tetragonal sheets that are derivatives of B we discuss two examples of functionalized Kagom lattices C herein.

PAGE 41

30 Compound 2 of formula {[Cu2(5-MeO-BDC)2(4-MeO-Pyridine)2](guest)x} was obtained as crystalline materials from an ethanol solution of Cu(NO3)2•2.5H2O and 5MeO-H2BDC using 4-MeO-Pyridine as the base and nitrobenzene as the template. In a similar fashion, crystals of compound 3 {[Cu2(5-MeO-BDC)2(MeOH)2](guest)x}, was obtained from a methanol solution of Cu(NO3)2•2.5H2O and 5-Br-H2BDC using pyridine as the base and nitrobenzene as the template. Both compounds manifest 2D Kagom topology, i.e., they contain triangular cav ities as well as hexagonal cavities that result from the linking of triangular units. The size of the triangular and hexagonal cavities in both structures is comparable to 1 nm and 2 nm respectively, which is consistent with their parent compound C However, the crystal structures of compound 2 and 3 significantly differ in the mann er in which the networks stack with respect to each other. The 2D Kagom sheets in 2 eclipse right on top of each other, giving rise to an “AAA” packing, as is also the case in the parent compound; those in 3 are, on the other hand, a ) b ) Figure 2.10 Crystal packing of compound 2 ( a ) and 3 ( b ). Atoms highlighted in purple are methoxy (in 2 ) or bromo (in 3 ) groups.

PAGE 42

31 slipped in the c direction by ca. 33.3%, i.e., every fourth layer repeats, thus resulting in an “ABCABC” sequence (Figure 2.10). The interlayer distances for 2 3 and C are 11.6, 10.4 , and 9.9 , respectively (Table 2.1), underlining the different sizes of the substituted groups at the 5position of BDCs. Table 2.1 Comparison of chemical and structural information for compound 2 3 and their parent compound. Compound R L (axial ligand Space Packing Interlayer (5-R-BDC) of SBU) Group Sequence Distance ( ) 2 MeO 4-MeO-Pyridine P-3 AAA 11.6 3 Br MeOH R-3 ABCABC 10.4 parent H Pyridine P-3C1 AAA 9.9 Kagom lattices are an extremely important class of compounds for a number of reasons: 1) Kagom lattice C is one of the most famous examples of geometrically frustrated topologies, which have been highly pursued by both physicists and chemists;85 2) They are inherently suitable for the generation of multifunctional materials since they are magnetically active and they contain nanoscale cavities and channels; 3) They are modular in nature and they contain multiple sites for steric and/or electronic modification. Compound 2 and 3 ideally illustrate these features and therefore represent a step forward toward tailored functional materials. 2.2.3 Three-Dimensional Structures and Some Predicted Structures In addition to the zero-dimensional nanoballs and two-dimensional tetragonal sheets and Kagom lattices, the self-assembly of SBU I and BDC and its derivatives has also resulted in a number of three-dimensional structures, two of which are shown in Figure 2.11, namely, USF-1 net D and CdSO4 net E respectively. Similar to those in the two-

PAGE 43

32 dimensional structures B and C the SBUs in D and E also display “1, 2-D” predispositions. However, in both cases, the configurations of SBUs I are significantly twisted and the ligands BDC are considerably out-of-plane, which explains the higher dimensionality of the resulting structures, as compared to the cases of structures B and C The differences between D and E on the other hand, can be rationalized on the basis of their different torsion angles. It should be pointed out that D and E represent two examples of 4-connected nets that are both based upon square nodes (Figure 2.5). The vertex symbols can be written as 62•62•62•1250•63•63 and 6•6•6•6•62•*, for D and E respectively. While CdSO4 net represents a common topology for a diverse range of metal-organic networks, 86-90 the connectivity of USF-1 net is truly unprecedented and compound D is thus far the only example that has been observed.91 a ) b ) Figure 2.11 Crystal structures of USF-1 D ( a ) and CdSO4 net E ( b ). Motifs shown in the blue boxes illustrate the distorted “1, 2-D” conformations of SBUs.

PAGE 44

33 We have so far experimentally produced at least 6 different supramolecular isomers (i.e., two forms of nanoballs A tetragonal sheet B Kagom lattice C USF-1 net D and CdSO4 net E ) that are assembled from SBU I and ligand BDC. A conformational consideration has been invoked to facilitate the rationalization of this remarkable supramolecular isomerism. In summary, SBUs in nanoballs A (including both cubic and hexagonal phases) take up a “4D” configuration, and those in structures B ~ E belong to a “1, 2-D” conformation. Such a conformational analysis further suggests the possibility of other supramolecular isomers that might be isolated from this system and we will briefly describe below four of these hypothetical structures, which are based upon “1, 3-D” (structure H1 ), a combination of “1, 2-D” and “1, 3-D” (structure H2 ), a combination of “4D” and “1, 2-D” (structure H3 ), and a combination of “3D” and “1, 2-D” (structure H4 ), respectively (Figure 2.12~2.15). Structures H1 and H2 are both three-dimensional architectures. The inherent topology of H1 is related to that of the sodalite net seen in zeolites.81,92 Note that the 1, 3alternative configuration of SBUs has in effect rendered each node a pseudo S4 symmetry Figure 2.12 Ball-and-stick and schematic representations of hypothetical structure H1 Blue box illustrates the “1, 3-D”configuration of SBUs in the structure.

PAGE 45

34 (since the SBU only has D2h symmetry), resulting in a tetrahedral framework (Figure 2.12). H2 is based upon a 1:1 mixture of “1, 2-D” and “1, 3-D” nodes and its topology is associated with that of the PtS nets (Figure 2.13).93 Table 2.2 shows a short summary of the crystallographic data for H1 and H2 Table 2.2 Crystallographic data for the two three-dimensional hypothetical structures H1 and H2 Compound Space Group a/ b/ c/ / o / o / o V/ 3 H1 Pn-3m 26.343 26.343 26.343 90 90 90 18280.8 H2 P42/nnm 18.6273 18.6273 26.3430 90 90 90 9140.4 Structures H3 and H4 illustrates two examples of two-dimensional hypothetical structures that can be derived using the same principles of conformational consideration. Interestingly, H3 bears the same connectivity as structure C i.e., that of Kagom lattices. Nevertheless, it differs from C by the following aspects: 1) H3 is based upon a 1:2 mixture of “4D” and “1, 2-D” nodes, whereas C is purely from “1, 2-D” type nodes; 2) The lattice symmetry of H3 has been reduced to orthorhombic from trigonal seen in C ; 3) Figure 2.13 Ball-and-stick and schematic representations of hypothetical structure H2 Blue box illustrates a combination of “1, 2-D” and “1, 3-D” configurations of SBUs in the structure.

PAGE 46

35 The two-dimensional sheets of H3 exhibit a much more undula ting nature than those of C (Figure 2.14). H4 is quite an unusual two-dimensional lattice in that it is composed of triangular, square and hexagonal windows due to the presence of its mixed “1, 2-D” and “3D” nodes (Figure 2.15). It is perhaps worth pointing out th at there likely exist other possible structures from the same SBU I -BDC system. However, it should also be kept in mind that although these hypothetical structures are of reasonable geometric plausibility, the chemical feasibility of their formations remains unclear. Figure 2.14 Perspective and side views of hypothetical structure H3 in a ball-and-stick mode. Blue box illustrates a combination of “4D” and “1, 2-D”configurations of SBUs in the structure. Figure 2.15 Perspective and side views of hypothetical structure H4 in a ball-and-stick mode. Blue box illustrates a combination of “1, 2-D” and “3D” configurations of SBUs in the structure.

PAGE 47

36 2.3 Metal-Organic Network from SBU I and 1, 3-Adamantanedicarboxylate Similar to BDC, a ligand that subtends an angle of 120o, 1, 3-adamantanedicarboxylate (ADC) represents another rigid and angular dicarboxylato ligand that can be employed in the construction of novel metal-organic networks. In contrast to BDC, however, ADC has a relatively smaller angle which is close to 109o, and the two planes defined by the COOgroups are instead not parallel to each other (Figure 2.16 a ). Accordingly one would expect different types of structures can be assembled from ADC and SBU I Indeed, single crystals of {[Zn2(ADC)2(Pyridine)2](MeOH)2}, 4 were attained by layering a methanolic solution of H2ADC and pyridine onto a methanolic solution of a ) b ) c ) d ) Figure 2.16 Crystal structures of compound 4 : a ) ligand ADC; b ) the 1D ladder; c ) interdigitation of 1D ladders, leading to a 2D sheet; d ) packing of 2D sheets (guest molecules MeOH in CPK mode).

PAGE 48

37 Zn(NO3)2•6H2O that contains nitrobenzene as the temp late. As illustrated in Figure 2.16, the square SBUs I are double-linked by ADC motifs resulting in 1D architecture of molecular ladder topology (Figure 2.16 b ). These 1D ladders are running along (100) with two different orientations alternatively such that two neighboring ladders have an angle of ca. 107o. Interdigitation occurs between adjacent ladders through face-to-face ••• interactions ( dcentroid-centroid = 4.3 ). As result, an undulating 2D sheet whose mean plane parallels (110) plane is generated by virtue of combining relative strong metal-ligand coordination bonding and weak ••• interaction (Figure 2.16 c ). These 2D sheets are further packed into three dimensions in an “ABAB” fashion, therefore producing 1D channels of ca. 4.9 5.0 . Two methanol molecules per SBU occupy this free space and are hydrogen bonding to the carboxylato oxygens of ADC, which presumably further stabilizes the overall structure (Figure 2.16 d ). The features of compound 4 are salient from a design perspective: a) The ladder topology exemplifies another pattern in which square building units can be linked to each other; b) The fact that the angular ligand ADC is geometrically compatible with square SBUs I suggests other rigid angular organic linkers as reasonable candidates for the Figure 2.17 A predicted cylindrical structure ( H5 ) based upon ADC and SBU I

PAGE 49

38 design of novel metal-organic networks; c) In principle, other supramolecular isomers of 4 might as well be possible. In fact, a cylindrical structure H5 1D analogue of 2D tetragonal sheets B has been proposed (Figure 2.17). 2.4 Metal-organic Networks from SBU II and BDC or Its Derivatives The supramolecular isomerism demonstrated above by the SBU IBDC system has been remarkably impressive, and our conformational analysis reveals the fundamental geometric relationships among the various isomers. As metal-organic networks continue to be intensively exploited in the context of functional materials, an enhanced understanding on the formation of multiple forms of metal-organic compositions becomes especially critical not only from a de sign perspective, but perhaps even more importantly, from a synthetic perspective. In this context, we have explored the use of another type of dimetal tetracarboxylate, SBU II (Figure 2.2) along with BDC ligands, in order to determine the experimental parameters that might potentially determine supramolecular isomerism. As a result, we have found both templates and axial ligands play an important role in this regard. Whereas SBU I exemplifies a versatile square building block in terms of generating various metal-organic networks, SBU II can potentially serve as a pseudo square building block with an ideal symmetry of C2h (Figure 2.4; although the highest possible symmetry for SBU II is D2h, it is usually not achievable due to its less rigidity). A CSD analysis indicates the motif of SBU II exists for a wide array of transition metals, although its occurrence is much less often than that of SBU I

PAGE 50

39 Compound 5 {[Zn2(BDC)2(4-PhPy)4](Benzene)} (4-PhPy = 4-Phenylpyridine), was isolated as single-crystalline materials from a methanolic solution of BDC and Zn(NO3)2•6H2O using 4-Phenylpyridine as axial ligand and benzene as template. X-ray single crystal diffraction reveals a 1D ladder structure in which SBUs II are doubly bridged by BDC in a convergent fashion along a single direction, resembling the structure of compound 4 (Figure 2.18 a ). Each of the Zn(II) ions manifests an octahedral coordination geometry which is surrounded by two oxygens from one chelating carboxyl group, two oxygens from two bridging carboxyl groups, and two nitrogens from two 4phenylpyridine ligands. The Zn-O distances fall in the range of 1.991~2.292, and Zn-N distances average 2.188. The elongated aromatic systems of the axial ligands 4phenylpyridine engage in multiple ••• interactions in such a way that interdigitation occurs between neighboring ladders, thus generating cavities in which benzene molecules inhabit (Figure 2.18 b ). b ) Figure 2.18 1D ladder structure ( a ) and its packing ( b ) in compound 5 Benzene guests are shown in a space-filling mode. Blue box illustrates the convergent fashion in which SBUs II are linked by BDC. a )

PAGE 51

40 When replacing benzene with toluene and leaving everything else in the reaction mixture unchanged, colorless crystals of a new form, compound 6 of formula {[Zn2(BDC)2(4-PhPy)4](Toluene)2} precipitate out. In contrast to the ladder structure of 5 BDC moieties in 6 connect SBUs II in an alternative manner, i.e., divergently, therefore giving rise to a 2D layer structure (Figure 2.19 a ). The Zn(II) ions maintain an octahedral geometry and the Zn-O distances range from 2.014 to 2.505, somewhat larger than those observed in 5 The Zn-N distances (an average of 2.164), on the other hand, are close to or even shorter than those of 5 The interdigitation again occurs between 4-phenylpyridine moieties from adjacent layers with toluene occupying in the interlayer cavities (Figure 2.19 b ). 5 and 6 might be distinguished from a number of ways, among which is their packing efficiency. Apparently the lower dimensionality of 5 has facilitated a better staking of the bulky 4-phenylpyridyl groups, th us generating cavities of smaller size that can only fit benzene (but not toluene), while the higher dimensionality of 6 seems to prevent the same bulky groups from coming as close. Retrospectively, therefore, benzene b ) Figure 2.19 2D layer structure ( a ) and its packing ( b ) in compound 6 Toluene guests are shown in a space-filling mode. Blue box illustrates the divergent fashion in which BDCs link SBUs II a )

PAGE 52

41 preferentially induces the formation of 5 whereas toluene is probably a better template for 6 based upon a size-matching principle. That a small variation on the size of templates has such a dramatic effect on the outcome of superstructures underscores the importance of a careful control over cr ystallization conditions. Similar tuning effects exerted by axial ligands on supramolecular isomers can also be demonstrated by compound 7 and 8 In this context, we use a substituted BDC, namely, 5-hydroxy-1, 3-benzene-dicarboxylate (5-OH-BDC), to bridge SBUs II Note that hydroxyl groups are ideally suited for engaging in complementary supramolecular interactions since they are both hydrogen-bond donors and acceptors. Two different pyridine-type bases, namely, 3, 5-lutidine and isoquinoline, are employed as axial ligands in an effort to direct individual crystallization processes while benzene is used as the template in both cases. a ) b ) c ) Figure 2.20 Crystal structures of 7 : a ) 1D ladder; b ) 2D sheet sustained by complementary hydrogen bonds; and c ) the packing of the 2D sheets. Benzene molecules are shown in space-filing mode.

PAGE 53

42 7 {[Zn2(5-OH-BDC)2(3,5-lutidine)2](Benzene)2}, is structurally related to 5 in that it also exhibits a 1D ladder topology (Figure 2.20 a ) and both Zn-O and Zn-N distances are within the expected ranges and comparable to those observed in both 5 and 6 As is contrary to those seen in the previous two compounds, however, each of the Zn(II) ions in 7 displays a tetrahedral geometry which is completed by two oxygens from two bridging carboxyl groups, one oxygen from one mono-dentate carboxyl group and one nitrogen from 3,5-lutidine. As a result, the ladders in 7 are inevitably prone to be undulating and more significantly, such a wavy disposition allows the hydroxyl group (hydrogen-bond donor) and uncoordinated carboxyl oxygen (hydrogen-bond acceptor) on each 5-OH-BDC moiety in close contact with their partners from adjacent ladder in such a way that 2-fold hydrogen bonding occurs between neighboring ladders (Figure 2.20 b ). These complementary hydrogen bonds thus assemb le 1D ladders into 2D sheets, which in turn pack into 3D architecture and generate both cavities and channels that are occupied by benzene molecules (Figure 2.20 c ). 8 {[Zn2(5-OH-BDC)2(isoquinoline)3](Benzene)1.5}, was isolated when replacing 3,5-lutidine with isoquinolin e and the resulting compound bears a close resemblance to 6 a ) b ) Figure 2.21 2D layer structure ( a ) and the crystal packing ( b ) of compound 8

PAGE 54

43 i.e., a 2D planar sheet. Nevertheless, the coordination of Zn(II) ions in 8 demonstrates somewhat surprising diversity and within each SBU II one of the two zinc centers assumes a trigonal bipyramidal shape while the other the trigonal pyramidal. Similar to the situations observed in 7 distortion away from an octahedral geometry results in an uncoordinated carboxyl oxygen on each 5-OH-BDC moiety, which further engages in hydrogen bonding with nearby hydroxyl group within the same 2D sheet (Figure2.21 a ). In short, we have investigated two different approaches that involve careful selection of either templates or axial ligands and that aim to gain a better control on the formation of desired supramolecular isomers. Although more efforts need to be accomplished and still more data need to be co llected, our systems clearly sugge st a well-founded direction, i.e., supramolecular isomerism is experimentally controllable. 2.5 Experimental 2.5.1 Syntheses The materials in the synthesis were used as received from reliable commercial sources (Sigma-Aldrich or Fischer Scientific); solvent methanol was purified and dried according to standard methods. Synthesis of [Cu2(5-MeO-BDC)2(MeOH)x(H2O1.83-x]12, 1 Green plate crystals of compound 1 were formed by layering 3mL of a methanol solution containing 5-methoxyisophthalic acid (20 mg, 0.10 mmol) and 2,6-lutidine (0.035 mL, 0.30 mmol) onto 3mL of a methanol/nitrobenzene solution (2:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.099 mmol). Typical yield of the reaction is ca 18mg for each vial. Synthesis of {[Cu2(5-MeO-BDC)2(4-MeO-Pyridine)2](guest)x}, 2

PAGE 55

44 Compound 2 were obtained by layering 3 mL of an ethanol solution containing 5methoxyisophthalic acid (20 mg, 0.10 mmol) and 4-methoxypyridine (0.031 mL, 0.30 mmol) onto 3 mL of an ethanol/nitrobenzene solution (2:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.099 mmol). Some green-blue precipitates appeared immediately and green hexagonal crystals fo rmed at the interlayer boundary within 3 days. Typical yield of the reaction is ca 14 mg for each vial. Synthesis of {[Cu2(5-Br-BDC)2(MeOH)2](guest)x}, 3 Compound 3 were obtained by layering 3 mL of a methanol solution containing 5bromoisophthalic acid (11 mg, 0.050 mmol) and pyridine (0.012 mL, 0.15 mmol) onto 3 mL of a methanol/nitrobenzene solu tion (2:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.10 mmol). Green-blue crystals (mostly twinned) appeared at the interlayer boundary within 3 days. Typical yield of the reaction is ca 12 mg for each vial. Synthesis of {[Zn2(ADC)2(Pyridine)2](MeOH)2}, 4 Compound 4 were obtained by layering 4 mL of a methanol solution containing 1,3-adamantanedicacarboxylic acid (112 mg, 0.500 mmol) and pyridine (0.24 mL, 3.0 mmol) onto 5 mL of a methanol/nitrobenzene solution (3:2, v/v) containging Zu(NO3)2•6H2O (149 mg, 0.500 mmol). Colorless crystals appeared at the interlayer boundary after 7 days. Synthesis of {[Zn2(BDC)2(4-PhPy)4](Benzene)}, 5 Compound 5 were obtained by layering 6 mL of a methanol solution containing isophthalic acid (33 mg, 0.20 mmol) and 4-phenylpyridine (93 mg, 0.60 mmol) onto 6 mL of a methanol/benzene solution (2:1, v/v) containging Zn(NO3)2•6H2O (60 mg, 0.20 mmol). Colorless crystals appeared after 7 days.

PAGE 56

45 Synthesis of {[Zn2(BDC)2(4-PhPy)4](Toluene)2}, 6 Compound 6 were obtained by layering 6 mL of a methanol solution containing isophthalic acid (33 mg, 0.20 mmol) and 4-phenylpyridine (93 mg, 0.60 mmol) onto 6 mL of a methanol/toluene solution (2:1, v/v) containging Zn(NO3)2•6H2O (60 mg, 0.20 mmol). Colorless prism crystals appeared within 3 days. Synthesis of {[Zn2(5-OH-BDC)2(3,5-lutidine)2](Benzene)2}, 7 Compound 7 were obtained by layering 20 mL of a methanol solution containing 5hydroxyisophthalic acid (182 mg, 1.00 mmol) and 3, 5-lutidine (0.342 mL, 3.00 mmol) onto 20 mL of a methanol/benzene solution (3:1, v/v) containging Zn(NO3)2•6H2O (297 mg, 1.00 mmol). Colorless needle crystals appeared after 24 hours. Synthesis of {[Zn2(5-OH-BDC)2(isoquinoline)3](Benzene)1.5}, 8 Compound 8 were obtained by layering 20 mL of a methanol solution containing 5hydroxyisophthalic acid (182 mg, 1.00 mmol) and isoquinoline (0.354 mL, 3.00 mmol) onto 20 mL of a methanol/benzene solution (3:1, v/v) containging Zn(NO3)2•6H2O (297 mg, 1.00 mmol). Colorless block crystals appeared after 24 hours. 2.5.2 Characterizations Crystal Structure Determination Single crystals suitable for X-ray cr ystallographic analysis were selected following examination under a microscope. Intensity da ta were collected on a Bruker-AXS SMART APEX/CCD diffractometer using Moka radiation ( = 0.7107 ). The data were corrected for Lorentz and polarization effects and for absorption using the SADABS program. The structures were solved using direct methods and refined by full-matrix least-squares on

PAGE 57

46 |F|2. All non-hydrogen atoms were refined an isotropically and hydrogen atoms were placed in geometrically calculated positions and refined with temperature factors 1.2 times those of their bonded atoms. All crystall ographic calculations were conducted with the SHELXTL 5.1 program package. Table 2.3 Crystallographic data for compounds 1 ~ 8 Compound 1 2* 3 4* Chemical formula C251H140Cu24 N3O162 C24H20Cu2 N2O12 C18H12.22Br2 Cu2O10.67 C120H80N8 O40Zn8 Formula weight 7314.62 655.51 686.07 2796.88 Temperature, K 100(2) 100(2) 100(2) 100(2) Crystal system Triclinic Trigonal Trigonal Orthorhombic Space group P-1 P-3 R-3 P212121 a, 24.172(8) 18.800(3) 18.203(4) 8.5977 b, 24.212(8) 18.800(3) 18.203(4)) 18.0229 c, 33.226(11) 11.600(5) 31.268(13)) 22.2479 deg 91.724(6) 90 90 90 deg 91.854(6) 90 90 90 deg 107.513(6) 120 120 90 V, 3 18518(10) 3550.62 8972(4) 3447.43 Z 2 9 4 calcd, gcm-3 1.312 1.143 mm-1 1.432 3.102 1.44 F(000) 7318 3002 1416 Crystal size, mm 0.11 x 0.09 x 0.03 0.10 x 0.10 x 0.02 range for data collection, deg 1.04 to 20.15 1.45 to 20.85 Limiting indices -16<=h<=23 -23<=k<=23 -32<=l<=32 -11<=h<=18 -18<=k<=6 -29<=l<=31 -9 =2sigma(I)] R1 = 0.1660 wR2 = 0.3744 R1 = 0.1219 wR2 = 0.3333 R indices (all data) R1 = 0.3285 wR2 = 0.4806 R1 = 0.1986 wR2 = 0.3611 Large diff. peak and hole, e-3 1.321 and -1.132 1.593 and -1.273 The poor quality of X-ray diffraction data for 2 and 4 and their structural refinements only result in reliable structural models and respective cell parameters.

PAGE 58

47 (Continued) Compound 5 6 7 8 Chemical formula C66H50N4O8Zn2 C45H34N2O8Zn2 C42H34N2O10Zn C52H38N3O10Zn2 Formula weight 1157.84 861.48 792.08 995.59 Temperature, K 100(2) 100(2) 298(2) 298(2) Crystal system Triclinic Orthorhombic Triclinic Monoclinic Space group P-1 Pccn P-1 P21/c a, 13.1612(10) 22.805(3) 9.1895(12) 19.9198(18) b, 13.7613(10) 15.961(2) 14.0959(18) 11.5997(10) c, 16.8083(12) 16.557(3) 15.864(2) 21.407(2) deg 83.7480 90 77.608(2) 90 deg 67.5420 90 88.590(2) 113.253(2) deg 69.4630 90 82.073(2) 90 V, 3 2633.5(3) 6026.4(16) 1987.9(4) 4544.6(7) Z 2 6 2 4 calcd, gcm-3 1.460 1.424 1.323 1.455 mm-1 0.976 1.250 0.677 1.120 F(000) 1196 2652 820 2044 Crystal size, mm 0.50 x 0.40 x 0.20 0.25 x 0.20 x 0.15 0.20 x 0.05 x 0.05 0.30 x 0.10 x 0.10 range for data collection, deg 1.78 to 25.04 1.56 to 21.50 1.31 to 28.31 1.11 to 28.34 Limiting indices -13<=h<=15 -16<=k<=15 -19<=l<=20 -14<=h<=23 -16<=k<=16 -16<=l<=17 -12<=h<=12 -18<=k<=18 -21<=l<=20 -22<=h<=25 -15<=k<=15 -28<=l<=22 Reflections collected 14042 21750 17426 28359 Unique reflections 9162 3468 9024 10625 R(int) 0.0168 0.3049 0.0547 0.0366 Completeness to 98.4 % 99.9 % 91.3 % 93.4 % Absorption correction None SADABS None None Max. and min. transmission 1.00000 and 0.329032 1.000 and 0.527 1.000 and 0.790 1.000 and 0.846 Data/ restraints/ parameters 9162 / 0 / 721 3468 / 3 / 337 9024 / 0 / 496 10625 / 0 / 606 Goodness-of-fit on F2 1.051 1.150 1.032 1.028 Final R indices [I>2sigma(I)] R1 = 0.0333 wR2 = 0. 0880 R1 = 0. 1440 wR2 = 0. 4073 R1 = 0. 0673 wR2 = 0. 1728 R1 = 0. 0417 wR2 = 0. 0932 R indices (all data) R1 = 0. 0382 wR2 = 0. 0912 R1 = 0. 1727 wR2 = 0. 4181 R1 = 0. 1149 wR2 = 0. 2004 R1 = 0. 0570 wR2 = 0. 1002 Large diff. peak and hole, e-3 0.519 and -0.385 2.188 and -1.100 0.821 and -0.446 0.436 and -0.347 Other Characterizations Low resolution X-ray Powder Diffr action (XPD) data were recorded on a Rigaku RU15 diffractometer at 30kV, 15mA for Cu K ( = 1.5418 ), with a scan speed of 1/min and a step size of 0.05 in 2 at room temperature. The simulated XRPD patterns

PAGE 59

48 were produced using and Powder Cell for Windows Version 2.4 (programmed by W. Kraus and G. Nolze, BAM Berlin, 2000). Figure 2.22 Experimental and simulated XPD pattern of 1

PAGE 60

49 Chapter 3 Metal-Organic Networks Based upon a More Flexible Dicarboxylate 3.1 Introduction 3.1.1 Rigidity vs. Flexibility Metal-organic networks, or coordination polymers, have been intensively investigated in the last decade as new classes of functional materials, in part due to the unique characteristics of meta l-ligand interactions, namely, they are relatively strong and highly directional, but also kinetically labi le. In addition, the well-established molecular synthetic chemistry has afforded, in the context of constructing hybrid network structures, a wide variety of organic ligands ranging from robust rod-like spacers to conformationally versatile linkers. The modular assembly of these building blocks can therefore be easily fine-tuned by judicious selection of either components56, 65 and it is perhaps not surprising to encounter the accomm odation of both rigidity and flexibility in the same class of compounds. Metal-organic frameworks that are able to remain intact under intense conditions (such as high temperatures, removal of guest species, etc.) are of high technical importance because of their potential applications in separation, storage, and heterogeneous catalysis.68-70 One of the most representative examples, MOF-5, is a highly porous cubic open frame work with remarkable thermal stability, which is assembled from SBU III (see Figure 2.2) and 1, 4-benzenedicarboxylate, a rigid and linear building block.64 In the previous chapter, we focus upon incorporating rigid but angular dicarboxylato ligands into the frameworks, wh ich has been proved to be of success in terms of generating a wide array of supramolecular isomers from simple building blocks.

PAGE 61

50 Recently, attentions have been paid to a novel type of metal-organic networks that are integrated with more flexible structural elements.94-101 This new class of compounds are characterized by the dynamic features of their porous architectures and the ability to undergo structural deformations upon external stimuli while maintaining crystallinity of the materials, i.e., they are capable of guest-induced shape-responsive fitting and resemble the degree of induced-fit behavior of bioenzymes such as metalloproteins.102 An elegant example of dynamic metal-organic networks, in which reversible release and uptake of guest molecules cause substantial changes in the local geometry of metal centers (Fe(II)) and lead to interesting spin crossover properties, has been recently reported (Figure 3.1).103 In principle, the resilience of metal-organic networks can be mainly attributed to the flexibility on the molecular level (i.e., flexib ility of both metal coordination geometries and ligand conformations) as well as on the supramolecular level (i.e., low energy barriers among multiple arrangements of molecular building blocks). Although it is not unfeasible to exploit the dynamic aspect s of metal-organic networks from both Figure 3.1 Guest-dependent deformation of a metal-organic network that leads to spin crossover.

PAGE 62

51 perspectives, we will herein primarily highlig ht the influence of ligand conformation on the supramolecular structures. 3.1.2 Conformational Analysis of Organic Ligands: A CSD Survey As delineated above, the manner in which different parts of a molecular entity are disposed with respect to each other will have critical impact on the resulting superstructures; in other words, the intrinsic rigidity or flexibility of metal-organic frameworks will be in part dictated by the co nfigurations of organic ligands. Therefore a detailed investigation on three-dimensional structures of organic functional groups is reasonably justified. In this respect, CSD, a database that houses more than 360,000 organic and metal-organic crystal structures in total and over 330,000 with 3D coordinates determined,82 provides an ideal platform because a systematic analysis of structural parameters can be conveniently realized with the aid of appropriate softwares.104 In particular, we are concerned with two prototypal ligands, namely, 4, 4’-bipyridine (4, 4’-bipy) and benzoates/benzoic acids (molecules that contain at least one carboxyl group attached to a benzene ring), since they represent two of the most widely used ligand systems.56 We define torsion angle of 4, 4’-bipy as the dihedral angle between the a ) b ) Figure 3.2 Planes that define the torsion angles of 4, 4’-bipyridine ( a ) and benzoates/benzoic acids ( b ).

PAGE 63

52 two pyridyl rings and that of benzoates/benz oic acids as the inclination between carboxyl group and benzene ring (Figure 3.2). In the case of 4, 4’-bipy, while it is obvious that the two pyridyl rings are allowed to have certain degree of free rotations, there exists a clear-cut between the planar and torsional conformations, as indicated by the sharp peaks representing near-zero torsion angles and a much smoother distributions among higher torsion angle regions (Figure 3.3). It is worth noting that far less hits are seen in the range of large torsion angles, although coordinating to metal ions does slightly push such a limit to a higher extent. Similar trends can also be observed in the case of benzoates/benzoic acids, i.e., a large number of hits are narrowed within a small range of relatively low torsion angles and the metal-ligand interactions somehow contribute to increase the distortions. However, the distributions of torsion angles tend to be more continuous than those of 4, 4’-bipy, indicating a generally higher flexib ility for the aromatic carboxylates/carboxylic a ) b ) Figure 3.3 Histograms showing the distributions of torsion angles for both noncoordinated ( a ) and coordinated ( b ) 4, 4’-bipyridine.

PAGE 64

53 acids. In particular, we found an even higher degree of distortion for the fluorosubstituted ligands within this family, as co mpared to aromatic carboxylates/carboxylic acids in general. Interestingly, other haloge n-substituted carboxylates/carboxylic acids do not share this same pattern, suggesting an electronic rather than steric reason for the high flexibility of fluorinated ligands (Figure 3.4). a ) b ) c ) d ) Figure 3.4 Histograms showing the distributions of torsion angles for noncoordinated ( a ), coordinated ( b ), fluoro-substituted ( c ) and other halogen-substituted ( d ) benzoates/benzoic acids.

PAGE 65

54 3.2 Metal-Organic Networks from Tetrafluoro-1, 3-benzenedicarboxylate (TFBDC) 3.2.1 1D Structures In the previous chapter, we conc entrate upon the use of angular dicarboxylato ligand BDC, which prefers to adopt planar or near-planar conformations. The combination of angularity and rigidity of the ligand has thus far led to a diverse range of metal-organic network structures that are of particular interest from both scientific and technical perspectives. It hence intrigued us as what could be expected if higher flexibility is integrated along with angularity. Our CSD analysis above indicates that fluorinated carboxylates/carboxylic acids are ideal candida tes in this regard since the presence of fluorine atoms significantly increases the flex ibility of molecules. In this context, we have systematically investigated a particular compound, namely, tetrafluoro-1, 3benzenedicarboxylate (TFBDC), a fluorinated version of BDC, to explore its use in the context of metal-organic networks. The solid state structure of H2TFBDC reveals that of 1D zigzag chain motifs which are sustained by an array of carboxylic dimers (Figure 3.5 a ).50 The O•••O distances in each dimer are ca 2.6, well within the anticipated range for such interactions. As expected, the torsion angles of carboxyl planes with respect to the aromatic rings have the values of 39.00 and 41.34o, which are considerably higher than those observed in BDC. Interestingly, the zigzag chain pattern exhibited in the crystal structure of the free ligand has been literally retained by compound 9 [Cu2(TFBDC)2(Py)4], which was obtained from an ethanol solution of Cu(NO3)2•2.5H2O and TFBDC in the presence of pyridine and nitrobenzene. The analogy can be further drawn by comparing the dimeric units seen in 9 which are composed of two Cu(II) centers, two bridging bifurcated

PAGE 66

55 carboxyl oxygens, two mono-dentate carboxyl oxygens and four pyridines, to the carboxylic dimers in the free ligand (Figure 3.5). Each Cu(II) displays a trigonal bipyramidal geometry and the Cu•••Cu distance is as far as 3.44 , also within the reasonable range expected for this type of chromophore although significantly larger than those seen in SBU I These dimeric units are doubly bridged by TFBDC moieties to give rise to 1D polymeric chains, which in turn close-pack into three dimensions, excluding nitrobenzene from entering the crystal structure. The centroid-centroid distances of each pair of TFBDCs and pyridines are 4.53 and 3.98, respectively, indicating fairly weak ••• stacking for the former and moderate one for the latter. It should be noted that similar 1D coordination polymers have also been isolated using BDC and Cu(II) as building blocks; however, they are mostly based upon mono-copper centers and no such dimeric units are identified in those structures. a ) b ) Figure 3.5 1D zigzag chain structures of the ligand H2TFBDC ( a ) and compound 9 ( b ).

PAGE 67

56 3.2.2 Guest-Dependent Opening/Closing of Two Types of Cavities in 2D Structures The flexibility of the ligand TFBD C and its impacts on supramolecular structures not only can be exemplified by the above 1D structures, but more remarkably, as will be specified below, they are also well demonstrated in a series of 2D architectures that are built upon paddle-wheel SBU I and that are closely related to the tetragonal sheets B discussed in chapter 2. Compound 10a {Cu2(TFBDC)2(quinoline)2}, was acquired as green singlecrystalline materials from an ethanol solution of Cu(NO3)2•2.5H2O and TFBDC using quinoline as the base and relatively large aromatic molecules (such as toluene and xylenes) as the template. X-ray diffraction study discloses a contracted 2D tetragonal sheet topology for 10a thanks to a pronounced distorted effect of TFBDC in which the torsion angles of two carboxyl planes are 57.92o and 75.29o, respectively. The fluorinated rings of two opposite TFBDC ligands are facing toward each other ( dcentroid-centroid = 3.665) and they therefore engage in fairly strong interactions. Such a short contact, a ) b ) Figure 3.6 Crystal structure ( a ) and crystal packing ( b ) of compound 10a

PAGE 68

57 however, effectively closes off the potential cavities that would otherwise be available to guest species (Figure 3.6 a ). Notably, the axial ligand quinolines also participate in, among themselves, considerably strong C-H••• interactions with the D (distance from C to the aromatic ring) being 3.683 within each layer and 3.757 between adjacent layers. As a result of such efficient close-packing, no inter-layer space exists either and thus 10a can be described as an “apohost” framework (a host framework without guest molecules). Such an apohost framework, however, exhibits quite intriguing dynamic characteristics. Indeed, by careful selection of other aromatic templates of appropriate sizes, as compared to those larger ones used in the synthesis of 10a we are able to open up the potential cavities and introduce guest species into the fra mework. Even more significantly, crystallographic study demonstrates it is possible to selectively open either intra or inter -layer free space by means of co ntrolling molecular recognitions. When employing p -dichlorobenzene instead of toluene or xylenes as the template, we obtained compound 10b {[Cu2(TFBDC)2(quinoline)2]( p -dichlorobenzene)0.5}, as the major product. Single-crystal X-ray diffraction suggests that 10b retains a very similar a) b ) Figure 3.7 Crystal structure ( a ) and crystal packing ( b ) of compound 10b The axial ligand (quinoline) is omitted in a ) for the purpose of clarity.

PAGE 69

58 2D architecture as 10a i.e., a distorted tetragonal sheet. In contrast to 10a however, guest molecules p -dichlorobenzene enter into the crystal structure of 10b and force to open the intra-layer cavities that are each defined by four SBUs I and four TFDBC moieties (Figure 3.7 a ). Surprisingly, p -dichlorobenzene occupies only half of these cavities, leaving the other half remain clos ed. Such a dissymmetric occupancy leads to two remarkably distinct dimensions for the open and closed cavities and their centroidcentroid distances between opposite TFBDC rings vary by more than 2.4 (6.793 vs. 4.390)! Within the open cavities, each of the crystallographically disordered p dichlorobenzene molecules is sandwiched by two TFBDCs and the centroid-centroid distance from p -dichlorobenzene to each of the TFBDC rings is 3.397, exactly half of the value 6.793, indicating perfectly parallel interactions between these aromatic systems. It is perhaps of interest to compare the centroid-centroid distances of the closed cavities in 10b (4.390) with those in 10a (3.665) and presumably such a discrepancy can be attributed to the structural distortion of 10b which is induced by the presence of p dichlorobenzene. The structural deformations caused by p -dichlorobenzene can be further exemplified by the subtle changes of intra-layer and inter-layer interactions among quinolines. Whereas quinolines within each layer still participate in C-H••• interactions (D = 3.787), only half amount of such interactions prevails because of a much larger separation for the other half (D = 7.106) due to the expansion of the open cavities. The inter-layer interactions between quinolines, on the other hand, manifest an accommodation of both ••• ( dcentroid-centroid = 3.355 and 3.341) and C-H••• bonding (D = 3.765), in contrast to the solo appearance of C-H••• interactions in 10a

PAGE 70

59 Surprisingly, although 10b significantly differs from 10a from a supramolecular perspective, X-ray powder diffraction (XPD ) and Thermal Gravimetric Analysis (TGA) experiments indicate that if removed from mother liquor under ambient conditions, 10b quickly undergoes a phase trans ition, most likely back to 10a suggesting the thermodynamic instability of the former. Another form of 10 was isolated as single-crystalline product when using chlorobenzene as the crystallization template. This new compound, namely, 10c, with a formula of {[Cu2(TFBDC)2(quinoline)2](chlorobenzene)0.5}, also exhibits a 2D distorted tetragonal sheet topology with an identical network composition as in 10a and 10b As illustrated in Figure 3.8 a the 2D framework displays a closed mode and the two types of short contacts between opposite TFBDC rings ( dcentroid-centroid = 4.149 and 4.652; see below for an explanation of such a difference) clearly suggest an efficient ••• stacking. Quinolines again play an important role in st abilizing each of the 2D layers by engaging in an array of C-H••• interactions (D = 3.977). What makes this structure so unique, a ) b ) Figure 3.8 Crystal structure ( a ) and crystal packing ( b ) of compound 10c

PAGE 71

60 however, is the position of ch lorobenzene molecules within the metal-organic framework. Instead of going into the intra-layer space as seen in 10b chlorobenzene is found to be hosted by the inter-layer ca vities that are enclosed by quinolines on the sides and TFBDCs from the top and bottom (Figure 3.9 a ). While these cavities are mainly constructed from quinolines which engage in alternative ••• stacking ( dcentroid-centroid = 3.927) and C-H••• bonding (D = 3.576 and 3.977), the entrapped chlorobenzene molecules are sandwiched by TFBDC rings from adjacent layers through two-fold ••• interactions ( dcentroid-centroid = 3.998). Nevertheless, only half of these inter-layer cavities are occupied by chlorobenzene molecules and the other half remain guest-free (Figure 3.9 b ). Calculations105 further suggest a volume of ca 1303 for the first type of cavities, in good accordance with the molecular volume of chlorobenzene (98.53), 106 and a nearzero volume for the second type. One would probably be amazed by the extremely high local molar concentration ( ca 12.8M!) of the enclosed guest species. The alternative occupancy of the inter-layer cavities by chlorobenzene also accounts for the aforementioned two different centroid-centroid distances observed within each layer in a ) b ) Figure 3.9 The open ( a ) and closed ( b ) inter-layer cavities in 10c

PAGE 72

61 10c (4.652 vs. 4.149; the former belongs to the ones involved with chlorobenzene) since the interactions between the TFBDC rings and the sandwiched chlorobenzene molecules are driving TFBDC rings slightly away from their opposite partners from the same layer with which they are simultaneously interacting. It is worth noting that both XPD and TGA experiments suggest that complex 10c is much more stable than 10b and the guest species stay in the structure even after removed from mother liquor at room temperature. Compounds 10a ~ c therefore represent a prototypal example of metal-organic networks that are robust and flexible enough to adjust the frameworks under different environments. It appears clear to us that fluorination on the dicarboxylato ligand plays a critical role in this regard, since the re markable flexibility of the functionalized frameworks hasn’t been observed in the original compounds that are based upon the ligand BDC. In contrast to other highly rigid compounds, these new classes of dynamic structures are capable of responding to various host-guest recognition events and accommodating a wide array of guest species, which is especially important in the applications of separation, molecular sensing and storage. Until now, nevertheless, the follo wing questions concer ning the host-guest relationships and the diversity of molecular recognitions remain unanswered: 1) why would dichlorobenzene only reside in the intra-layer cavity whereas chlorobenzene exclusively stays within the inter-layer cav ity, even though these two molecules are electronically and chemically quite similar? 2) Which factors (e.g., energetic or steric effects) determine that only half of the intr a-layer or inter-layer cavities are occupied by guest species? 3) Does the presence of guest molecules in the final structure indicate their

PAGE 73

62 pre-organization and subsequent template effects on the formation of the ordered arrays of metal-organic compositions, or is it simp ly a result of post-synthetic molecular recognitions? Although further theoretical and experimental investigations are undoubtedly necessary, and a thorough understanding of these questions will la rgely facilitate the design of future generations of functional mate rials, we speculate tentative answers to the above as such: 1) Whereas the dimensions of the intra-layer cavities are suitable for both chlorobenzene and p -dichlorobenzene, the limited space enclosed by each of the interlayer cavities has eliminated th e inclusion of slightly larger p -dichlorobenzene; and since structures with guests sitting in the intra-layer cavities have been shown to be less stable, the thermodynamic forces are probably driving chlorobenzene into the inter-layer cavities. In short, steric effects play a central role in the case of dichlorobenzene while thermodynamic factors are the key for the case of chlorobenzene; 2) both the size of guests and the degree of deformation the framework can sustain, among others, decide that only half of the intraor inter-layer cavities of 10 can be fulfilled by chlorobenzene and dichlorobenzene, respectively. One could imagine complexes of 10 with full occupancy of either type of cavities; however, they are most likely over-distorted and therefore become thermodynamically unstable. In fact, as will be demonstrated below, benzene, a guest of smaller size, is able to fully occupy the intra-layer cavities of a related tetragonal sheet; 3) the existence of apohost 10a implies that the presence of aromatic guests is not indispensable for the formation of the metal-organic network; yet the welltrapped scenario of chlorobenzene as sugges ted by the fairly high thermal stability of

PAGE 74

63 complex 10c might indirectly indicate the possible template effects of host-guest interplay at the early stag es of crystallization. 3.2.3 Functionalization of Inter-layer Cavities in 2D Structures Thus far we have demonstrated an effective approach, namely, fluorination of organic ligands, for the modification of prototypal metal-organic networks. The introduction of highly electron-negative fluo rine atoms on the BDC rings dramatically alters the electronic properties of the ligand and results in a much higher level of framework flexibility. In fact, compounds 10a ~ c exemplify a new family of compounds with functionalized intra-layer cavities as the dynamic features of these structures are not observed in their un-substituted counterparts. Since both intra-la yer and inter-layer cavities are amenable to investigation in these structures, it is perhaps appropriate to further evaluate the feasibility of using a similar strategy to transform the nature of inter-l ayer cavities. Quinoline, a relatively large Figure 3.10 Three axial ligands of SBU I used for the functionalization of inter-layer cavities.

PAGE 75

64 hydrophobic aromatic system, has been shown to serve as the axial ligand of SBU I and play an important role in the construction of inter-layer cavities in 10a ~ c It therefore occurs to us that other types of axial ligands, such as 2-picoline (a hydrophobic but smaller aromatic molecule) and ethanol (a sm aller but less hydrophobic molecule), might as well be suited to direct the formation of various inter-layer cavities (Figure 3.10). Indeed, green crystals of 11 {[Cu2(TFBDC)2(EtOH)2](EtOH)2(benzene)}, precipitate from an ethanol solution containing Cu(NO3)2•6H2O, TFBDC, benzene and 2,6-lutidine. Structural analysis reveals a familiar 2D distorted tetragonal sheet and contrary to 10b where only half of the intra-layer cav ities are filled with guest species, each of the holes in 11 is inhabited by one benzene mol ecule that interacts with TFBDC rings through ••• stacking (Figure 3.11 a ; dcentroid-centroid = 3.481 and 3.542). Due to the weak coordination ability of 2, 6-lu tidine, solvent molecu les ethanol instead coordinate at the axial positions of SBUs I therefore modifying both steric and electronic environments on the surfaces of the 2D network. As a result, the inter-layer cavities become less hydrophobic and two ethanol molecules (instead of benzene!) are sitting as a ) b ) Figure 3.11 Crystal structure ( a ) and crystal packing ( b ) of compound 11 Guest molecules (EtOH and b enzene ) are re p resented in a CPK mode.

PAGE 76

65 guests inside each of them. Interestingly, these EtOH guests are hydrogen-bonding with the frameworks in two distinct motifs, one of which only involves the coordinated EtOH while the other takes advantage of both coordinated EtOH and the carboxyl oxygens (Figure 3.12). Four different hydrogen-bond distances ( do•••o = 2.609, 2.817; 2.633, and 2.980) are also well within the anticipated range for this type of interactions. When replacing uncoordinating 2, 6-lutidine with coordinating 2-picoline and using hexafluorobenzene (HFB) as the template, we obtained another new compound, 12 of formula {[Cu2(TFBDC)2(2-picoline)2](EtOH)1.3(HFB)}. 12 again manifests an alike 2D tetragonal sheet which has found no guests w ithin the intra-layer cavities (Figure 3.13). The centroid-centroid distance between opposite TFBDC rings is 4.281, in good consistence with those of its analog 10c ( dcentroid-centroid = 4.149 and 4.652) but slightly larger than those in 10a ( dcentroid-centroid = 3.665 ) The use of a smaller axial ligand 2picoline, as compared to the more bulky quinoline, has resulted in the following salient a ) b ) Figure 3.12 Two hydrogen-bonding motifs occurred between ethanol guests and the frameworks in 11

PAGE 77

66 features concerning the inter-layer cavities: 1) the cavities are not as well-defined as in the case of quinoline and the free space is in fact continuous along one direction, thus in effect transformed into 1D channels; 2) the aromatic molecules (HFB) and solvent species (EtOH) are co-existing as guests within the channles; 3) the average number of guest molecules per SBU I (1 HFB and 1.3 EtOH) is larger than other cases; 4) instead of associating with TFBDC rings, HFB molecu les orientate themselves toward 2-picoline moieties in such as a way that the pairs participate in face-to-face ••• stacking ( dcentroidcentroid = 3.632). In short, in addition to the use of fl uorinated ligands as flexible structural ingredients, we have illustrated another compelling strategy that can be employed to manipulate supramolecular structures and functions, i.e., systematically fine-tuning the chemical nature of the surfaces of 2D metal-organic networks. Since both approaches are based upon well-established supramolecular chemistry and crystal engineering principles, we a ) b ) Figure 3.13 Crystal structure ( a ) and packing ( b ) of 12 Half of 2-picoline ligands and all EtOH guests are crystallographically disordered. Guest molecules (EtOH and HFB) are represented in a CPK mode.

PAGE 78

67 anticipate them to be of general implications for the design of other useful metal-organic materials. 3.3 Experimental 3.3.1 Syntheses The materials in the synthesis were used as received from reliable commercial sources (Sigma-Aldrich or Fischer Scientific); solvent methanol was purified and dried according to standard methods. Synthesis of [Cu2(TFBDC)2(Pyridine)4], 9 Compound 9 were obtained by layering 4 mL of an ethanol solution containing 2, 4, 5, 6-tetrafluoroisophthalic acid (169 mg, 0.500 mmol) and pyridine (0.12 mL, 1.50 mmol) onto 4.5 mL of an ethanol/nitrobenzene solution (2.5:2, v/v) containging Cu(NO3)2•2.5H2O (116 mg, 0.500 mmol). Blue rod-like crystals formed at the interlayer boundary within 24 hours. Synthesis of [Cu2(TFBDC)2(Quinoline)2], 10a Compound 10a were obtained by layering 3.5 mL of an ethanol solution containing 2, 4, 5, 6-tetrafluoroisophthalic acid (23 mg, 0.10 mmol) and quinoline (0.059 mL, 0.50 mmol) onto 3.5 mL of an ethanol/toluene solution (2.5:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.10 mmol). Green block crystals formed at the interlayer boundary within 24 hours. Synthesis of {[Cu2(TFBDC)2(quinoline)2](p-dichlorobenzene)0.5}, 10b Compound 10b were obtained by layering 3.5 mL of an ethanol solution containing 2, 4, 5, 6-tetrafluoroisophthalic acid (23 mg, 0.10 mmol) and quinoline (0.059 mL, 0.50

PAGE 79

68 mmol) onto 3.5 mL of an ethanol/ p -dichlorobenzene solution (2.5:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.10 mmol). Green block crystals formed at the interlayer boundary within 24 hours. Synthesis of {[Cu2(TFBDC)2(quinoline)2](chlorobenzene)0.5}, 10c Compound 10c were obtained by layering 3.5 mL of an ethanol solution containing 2, 4, 5, 6-tetrafluoroisophthalic acid (23 mg, 0.10 mmol) and quinoline (0.059 mL, 0.50 mmol) onto 3.5 mL of an ethanol/chlorobenzene solution (2.5:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.10 mmol). Green block crystals formed at the interlayer boundary within 24 hours. Synthesis of {[Cu2(TFBDC)2(EtOH)2](EtOH)2(benzene)}, 11 Compound 11 were obtained by layering 2.5 mL of an ethanol solution containing 2, 4, 5, 6-tetrafluoroisophthalic acid (23 mg, 0.10 mmol) and 2, 6-lutidine (0.034 mL, 0.30 mmol) onto 2.5 mL of an ethanol/benzene solution (1.5:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.10 mmol). Green block crystals formed at the interlayer boundary within 24 hours. Synthesis of {[Cu2(TFBDC)2(2-picoline)2](EtOH)1.3(HFB)}, 12 Compound 12 were obtained by layering 3 mL of an ethanol solution containing 2, 4, 5, 6-tetrafluoroisophthalic acid (23 mg, 0.10 mmol) and 2-picoline (0.030 mL, 0.30 mmol) onto 3 mL of an ethanol/hexafluorobenzene (HFB) solution (5:1, v/v) containging Cu(NO3)2•2.5H2O (23 mg, 0.10 mmol). Green block crystals formed at the interlayer boundary within 24 hours. 3.3.2 Characterizations

PAGE 80

69 Crystal Structure Determination Single crystals suitable for X-ray cr ystallographic analysis were selected following examination under a microscope. Intensity da ta were collected on a Bruker-AXS SMART APEX/CCD diffractometer using Moka radiation ( = 0.7107 ). The data were corrected for Lorentz and polarization effects and for absorption using the SADABS program. The structures were solved using direct methods and refined by full-matrix least-squares on |F|2. All non-hydrogen atoms were refined an isotropically and hydrogen atoms were placed in geometrically calculated positions and refined with temperature factors 1.2 times those of their bonded atoms. All crystall ographic calculations were conducted with the SHELXTL 5.1 program package. Table 3.1 Crystallographic data for compounds 9, 10a ~ c 11 12 Compound 9 10a 10b Chemical formula C18H10Cu F4N2O4 C34H14Cu2F8N2O8 C37H16ClCu2F8N2O8 Formula weight 457.82 857.55 931.05 Temperature, K 100(2) 100(2) 100(2) Crystal system Monoclinic Orthorhombic Triclinic Space group C2/c Pbca P-1 a, 19.275(3) 12.2599(9) 11.2399(9) b, 11.4617(16) 12.1377(9) 12.9837(11) c, 17.162(2) 21.1716(16) 13.4073(11) deg 90 90 89.5790(10) deg 115.903(2) 90 67.1740(10) deg 90 90 79.6010(10) V, 3 3410.5(8) 3150.5(4) 1769.5(3) Z 8 4 2 calcd, gcm-3 1.783 1.808 1.747 mm-1 1.353 1.456 1.377 F(000) 1832 1704 926 Crystal size, mm 0.20 x 0.05 x 0.05 0.10 x 0.10 x 0.05 0.30 x 0.05 x 0.05 range for data collection, deg 2.13 to 28.27 1.92 to 28.30 1.60 to 28.26 Limiting indices -24<=h<=20 -11<=k<=15 -22<=l<=22 -15<=h<=9 -14<=k<=15 -27<=l<=28 -14<=h<=14 -16<=k<=17 -17<=l<=17 Reflections collected 10053 18652 15388 Unique reflections 3941 3743 7977 R(int) 0.0543 0.0559 0.0327 Completeness to 93.4 % 95.5 % 91.1 % Absorption correction None None None Max. and min. transmission 1.000 and 0.857 1.000 and 0.808 1.000 and 0.920 Data/ restraints/ parameters 3941 / 0 / 262 3743 / 0 / 244 7977 / 0 / 550 Goodness-of-fit on F2 1.061 1.089 1.026

PAGE 81

70Final R indices [I>2sigma(I)] R1 = 0. 0508 wR2 = 0. 1020 R1 = 0. 0476 wR2 = 0. 0981 R1 = 0. 0442 wR2 = 0. 0998 R indices (all data) R1 = 0. 0733 wR2 = 0. 1094 R1 = 0. 0641 wR2 = 0. 1047 R1 = 0. 0593 wR2 = 0. 1075 Large diff. peak and hole, e-3 0.542 and -0.696 0.629 and -0.531 0.596 and -0.339 (Continued) Compound 10c 11 12 Chemical formula C37H16Cl0.50Cu2F8N2O8 C30H30Cu2F8O12 C31H21.50Cu2F8.50N2O9.25 Formula weight 913.32 861.62 858.58 Temperature, K 100(2) 100(2) 100(2) Crystal system Triclinic Monoclinic Monoclinic Space group P-1 P21/n P21/n a, 11.3895(11) 13.2869(10) 13.0006(14) b, 12.7032(12) 13.5884(11) 12.9131(13) c, 13.2957(13) 19.2207(15) 20.042(2) deg 89.200(2) 90 90 deg 69.464(2) 103.6920(10) 103.889(2) deg 78.878(2) 90 90 V, 3 1764.6(3) 3371.6(5) 3266.2(6) Z 2 4 4 calcd, gcm-3 1.719 1.697 1.746 mm-1 1.343 1.367 1.409 F(000) 909 1744 1720 Crystal size, mm 0.50 x 0.40 x 0.20 0.10 x 0.10 x 0.02 0.20 x 0.10 x 0.04 range for data collection, deg 1.64 to 28.27 1.69 to 28.26 1.70 to 27.50 Limiting indices -14<=h<=14 -16<=k<=16 -17<=l<=17 -15<=h<=17 -11<=k<=17 -25<=l<=25 -16<=h<=16 -16<=k<=16 -25<=l<=15 Reflections collected 15312 20876 19720 Unique reflections 7939 7877 7389 R(int) 0.0416 0.0581 0.0940 Completeness to 90.7 % 94.3 % 98.6 % Absorption correction None None None Max. and min. transmission 1.00 and 0.824 1.000 and 0.842 ? Data/ restraints/ parameters 7939 / 0 / 523 7877 / 0 / 483 7389 / 1 / 451 Goodness-of-fit on F2 1.036 1.024 0.923 Final R indices [I>2sigma(I)] R1 = 0. 0539 wR2 = 0. 1258 R1 = 0. 0532 wR2 = 0. 1148 R1 = 0. 0565 wR2 = 0. 1152 R indices (all data) R1 = 0. 0740 wR2 = 0. 1370 R1 = 0. 0865 wR2 = 0. 1286 R1 = 0. 1124 wR2 = 0. 1271 Large diff. peak and hole, e-3 0.997 and -0.494 0.980 and -0.826 0.669 and -0.661

PAGE 82

71 Other Characterizations Thermogravimetric analysis was performed under nitrogen at a scan speed of 4C/min on a TA Instrument TGA 2950 Hi-Res. Low resolution XRPD data were recorded on a Rigaku RU15 diffractometer at 30kV, 15mA for Cu K ( = 1.5418 ), with a scan speed of 1 or 2/min and a step size of 0.05 in 2 at room temperature. The simulated XRPD patterns were produced using and Powder Cell for Windows Version 2.4 (programmed by W. Kraus and G. Nolze, BAM Berlin, 2000). Figure 3.14 TGA trace of compound 10a

PAGE 83

72 Figure 3.15 TGA trace of compound 10b Figure 3.16 TGA trace of compound 10c

PAGE 84

73 Figure 3.17 TGA trace of compound 11 Figure 3.18 Experimental and simulated XPD of compound 10a

PAGE 85

74 Figure 3.19 Experimental and simulated XPD of compound 10b compared with simulated XPD of 10a Figure 3.20 Experimental and simulated XPD of compound 10c

PAGE 86

75Chapter 4 Conclusions and Future Directions 4.1 Summary and Conclusions The research presented in this thesis is primarily concerned with developing an indepth understanding of the basic principles that govern the supramolecular behaviors of metal-organic networks and gaining an experimental control over the structure and function of these new classes of hybrid materials. In particular, this work has contributed to the rationalization of supramolecular isomerism, a phenomenon referred to the existence of more than one type of superstructure from the same set of molecular building blocks, and the functionalization of prototypal metal-organic materials. To summarize, we have illustrated the following aspects: i) Under various conditions, the self-assembly of rigid and angular ligand 1, 3benzenedicarboxylate (BDC) and Cu(II)/Zn(II )-based paddle-wheel secondary building units (SBUs), a dimetal tetracarboxylate, whic h can be viewed as a molecular square, generates a wide array of metal-organic networks ranging from 0D nanoballs, 2D tetragonal sheets and Kago m lattices, to 3D CdSO4 net and an unprecedented “USF-1” net. The remarkable diversity of the resulting superstructures from such simple structural ingredients can be rationalized on the bases of angularity and distortion of the molecular building blocks. A detail conformation and configuration analysis not only reveals the fundamental geometric relationships among the existing supramolecular isomers, but also predicts a number of other interesting structures that are in principle possible to be isolated.

PAGE 87

76 In addition, the employment of two geometric principles, namely, node-and-spacer and VLPP approaches, and the use of rigorous topological descriptions, such as vertex symbols and coordination sequences, have significantly facilitated the recognition, interpretation, and prediction of complicated metal-organic network structures. ii) Other angular ligands, such as 1, 3-adamantanedicarboxylate (ADC), also selfassemble with paddle-wheel square SBUs to give rise to some novel structures including 1D ladder topology. These results, along with those already mentioned in i), highlight the myriad possibilities of lin king square building units.107It is quite obvious that structures based upon square building units can always be simplified as 4-connected nets, which are probably among the best-understood classes of topologies.72-73,79,108 From a topological point of view, square and tetrahedral nodes ar e in fact interchangeable in the sense that each square-based net can be equally represented as a tetrahedron-based net by adjusting the shape of linkers (see Figure 2.12 for an example of tetrahedral frameworks illustrated in a square fashion), and vice versa If taking into account the numerous examples of tetrahedron-based zeolite nets, 81 it is perhaps appropriate to regard the design principles we delineate in this work concerning the use of square SBUs as a potential alternative to zeolite-like metal-organic fram eworks (ZMOFs), a recently developed area pioneered by Eddaoudi.109 We believe the key to the success re lies upon the rational selection of suitable spacers that can link square building blocks in a desired manner. iii) The assembly of BDC and its hydroxyl derivative with another dimetal tetracarboxylate, a pseudo -square SBU, also results in a series of supramolecular isomers such as 1D ladders and 2D sheets. Our controlled experiments demonstrate the subtle influences of both templates and axial ligands of SBUs on the resulting superstructures.

PAGE 88

77 These results unambiguously indicate that it is in principle possible to experimentally control supramolecular isomerism. From an added perspective, if one considers supramolecular isomerism in a broader sense, the 1D ladders and 2D sheets based upon pseudo -square SBUs can also be regarded as supramolecular isomers of those structures mentioned in i) since they all contain the same framework compositions. It thus further complicates the situation of superstructural diversity because not only the contributions from organic ligands (e.g., their angularity and conformation) but also those from metal ions (e.g., various factors that determine the formation of a certain chromophore) need to be taken into consideration. Nevertheless, the presence of an overwhelming amount of superstructures from a limited number of easily accessible building blocks might as well be considered as an opportunity from a materials point of view, as is exactly the case for the four different forms of carbons. iv) The introduction of fluorine atom s to BDC moieties has been shown to lead to a dramatic increase of flexibility of the molecu le and the incorporati on of tetraf luoro-1, 3benzenedicarboxylate (TFBDC) with paddle-wh eel square SBUs results in a wide array of 2D metal-organic networks that are based upon a distorted tetragonal sheet topology. The flexibility on the molecular level is thus translated into the supramolecular level as these 2D networks manifest guest-dependent closing/opening of intraand inter-layer cavities, a unique aspect that hasn’t been observed in the original un-fluorinated compounds. It therefore represents an effective approach toward functionalized metalorganic networks.

PAGE 89

78 By systematically modifying the ch emical nature of ligand s on the axial site of paddle-wheel SBUs, we have further shown our capability of adjusting the environments on the outer surfaces of the 2D frameworks, which in turn results in a better control over the size, shape and hydrophobicity of the inter-layer cavities. In particular, a small variation on the size of the axial ligands (i.e., from quinoline to 2-picoline) has transformed the inter-layer free space from discrete cavities to 1D continuous channels that can be utilized by a much higher amount of guest species. In short, the main effort of this work has been devoted to illuminating basic principles of supramolecular chemistry and crystal engineering in the context of designing metal-organic networks, which are applicable to a much broader range of functional supramolecular materials. 4.2 Future Directions Although it was deliberately intended that a focus on design and structural aspects would be placed upon the main body of this thesis, as is limited by the scope of this document, it is the function of solids that should be driving the field of crystal engineering and metal-organic networks into its next level of advancements. Specifically, interfacing with other cutting-edge areas, such as materials science, bio-sciences and nanotechnology, is rapidly becoming and will continue to be a main theme in the coming decades. It is the author’s belief that the ultimate goal of this field is to “make molecules at will”. Even though our understanding on the supramolecular and suprasupermolecular level remains relatively limited, as comp ared to that on the molecular level,110 it is only a

PAGE 90

79 matter of time that such a dream will be rea lized, especially with the view of increasing progress that have been made in gaining better controls on metal-organic systems, as unambiguously demonstrated by this thesis. As part of efforts that are aimed at this ambition, we propose the following initiatives, among others, to highlight the future direction of our research: 1) Stronger tools for structural determina tion of molecules, including effective techniques for routine elucidation of stru ctures of polycrystalline and amorphous solids; 2) A thorough understanding of hierarchies of weak intermolecular forces and the roles they play in the supramolecular entities; 3) Controlling supramolecular structure by manipulating molecular structures 4) A direct correlation of structure and function of molecules.

PAGE 91

80 References 1. Frank, F. C. "Descartes' Observations on the Amsterdam Snowfalls of 4, 5, 6 and 9 February 1634," J. Glaciology 1974 13 535. 2. A Oxford English Dictionary (online version, http://www.oed.com/ ) definition of “crystal”: A form in which the molecules of many simple elements and their natural compounds regularly aggregate by the operation of molecular affinity: it has a definite internal structure, with the external form of a solid enclosed by a number of symmetrically arranged plane fa ces, and varying in simplicity from a cube to much more complex geometrical bodies. 3. Kepler, J. The six-cornered snowflake 1611; reprinted with English translation and commentary by the Clarendon Press, Oxford, 1966 4. Hay, R. J. Essai d'une thorie sur la structure des cristaux ; Paris, 1784 5. Numbers of entries contained in the major primary crystallographic databases: i) 348,481 Cambridge Structural Database of organic and metal-organic structures (CSD, v5.27, including a Jan 2006 update); ii) 89,366 Inorganic Crystal Structure Database (ICSD, as of Dec 2005); iii) 35,701 Protein Data Bank (PDB, as of Mar 21, 2006); iv) 3,647 Biological Macromolecule Crystallization Database (BMCD, v2.00); v) 3,000 Nucleic Acid Database (NDB, as of Mar 14, 2006). 6. Jansen, M. Angew. Chem. Int. Ed. 2002 41 3746-3766. 7. Yaghi, O. M.; O'Keeffe, M.; Kanatzidis, M. J. Solid State Chem. 2000 152 1-2.

PAGE 92

81 8. Paul Ehrlich, Nobel Lecture, December 11, 1908 ; from Nobel Lectures, Physiology or Medicine 1901-1921 Elsevier Publishing Company, Amsterdam, 1967 9. Paul Ehrlich, Studies on Immunity Wiley, New York, 1906 10. Emil Fischer, Ber. Deutsch Chem. Ges. 1894 27 2985. 11. Powell, H. M. J. Chem. Soc. 1973 61–73. 12. Pedersen, C. J. J. Am. Chem. Soc. 1967 89 2495-2496. 13. Pedersen, C.J. J. Am. Chem. Soc. 1967 89 7017-7036. 14. Cram, D. J.; Kaneda, T.; Helgeson, R. C.; Lein, G. M. J. Am. Chem. Soc. 1979 101 6752-6754. 15. Trueblood, K. N.; Knobler, C. B.; Maveri ck, E.; Helgeson, R. C.; Brown, S. B.; Cram, D. J. J. Am. Chem. Soc. 1981 103 5594-5596. 16. Dietrich, B.; Lehn, J. M.; Sauvage, J. P. Tetrahedron Lett. 1969 2885 and 2889. 17. Dietrich, B.; Lehn, J. M.; Sa uvage, J. P.; Blanzat, J. Tetrahedron 1973 29 1629. 18. Dietrich, B.; Lehn, J. M.; Sauvage, J. P. Tetrahedron 1973, 29 1647. 19. Lehn, J. M. Angew. Chem. Int. Ed. 1988 27 89. 20. Lehn, J. M. Supramolecular Chemistry—Concepts and Perspectives VCH, Weinheim, 1995 21. Desiraju, G. R. Nature 2001 412 397-400. 22. Newkome, G. R.; He, E. F.; Moorefiled, C. N. Chem. Rev. 1999 99 1689-1746. 23. McManus, G. J.; Wang, Z.; Zaworotko, M. J. Cryst. Growth Des. 2004 4 11 – 13. 24. Ungaro, R.; Dalcanale, E. (Editors) Supramolecular Science: Where It Is and Where It Is Going Kluwer, Dordrecht, The Netherlands, 1999

PAGE 93

82 25. Wyler, R.; de Mendoza, J.; Rebek Jr., J. Angew. Chem. Int. Ed. 1993 32 16991701. 26. Rebek Jr., J. Chem. Soc. Rev. 1996 255-264. 27. Conn, M. M.; Rebek Jr., J. Chem. Rev. 1997 97 1647-1668. 28. Rebek Jr., J. Acc. Chem. Res. 1999 32 278 – 286. 29. Rebek Jr., J. Chem. Commun. 2000 637 – 643. 30. Rebek Jr., J. Angew. Chem. Int. Ed. 2005 44 2068-2078. 31. Kang, J.; Rebek Jr., J. Nature 1997 385 50 – 52. 32. Shivanyuk, A.; Rebek Jr., J. J. Am. Chem. Soc. 2002 124 12074-12075. 33. Fujita, M. Chem. Soc. Rev. 1998 27 417-425. 34. Fujita, M.; Umemoto, K.; Yoshizawa, M.; Fujita, N.; Kusukawa, T.; Biradha, K. Chem. Commun. 2001 509–518. 35. Fujita, M.; Tominaga, M.; Hori, A.; Therrien, B. Acc. Chem. Res. 2005 38 371380. 36. Fujita, M.; Oguro, D.; Miyazawa, M.; Oka, H.; Yamaguchi, K.; Ogura, K. Nature 1995 378 469-471. 37. Yoshizawa, M.; Tamu ra, M.; Fujita, M. Science 2006 312 251-254. 38. Kersting, B.; Meyer, M.; Powers, R. E.; Raymond, K. N. J. Am. Chem. Soc. 1996 118 7221. 39. Caulder, D. L.; Raymond, K. N. Acc. Chem. Res. 1999 32 975-982. 40. Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N. Acc. Chem. Res. 2005 38 351-360.

PAGE 94

83 41. Fiedler, D.; Bergman, R. G.; Raymond, K. N. Angew. Chem., Int. Ed. 2006 45 745-748. 42. Pauling, L. The Nature of the Chemical Bond ; Cornell University Press: Ithaca, New York, 1948 43. Pepinsky, R. Phys. Rev 1955 100 971. 44. Schmidt, G. M. J. Pure Appl. Chem. 1971, 27 647-678. 45. Sarma, J. A. R. P.; Desiraju, G. R. Acc. Chem. Res. 1986 19 222-228. 46. Desiraju, G. R. Acc. Chem. Res. 1991 24 270-276. 47. Desiraju, G. R. Acc. Chem. Res. 1996 29 441-449. 48. Desiraju, G. R. Acc. Chem. Res. 2002 35 565-573. 49. Desiraju, G. R. Crystal Engineering: the Design of Organic Solids ; Elsevier: Amsterdam, 1989 50. Desiraju, G. R. Angew. Chem., Int. Ed. 1995 34 2311-2327. 51. Dunitz, J. D. Pure Appl. Chem. 1991 63 177-185. 52. Dunitz, J. D. Perspectives in Supramolecular Chemistry ; Desiraju, G. R., Ed.; Wiley: New York, 1996 ; Vol. 2. 53. Hoskins, B. F.; Robson, R. J. Am. Chem. Soc. 1989 111 5964-5965. 54. Batten, S. R.; Robson, R. Angew. Chem., Int. Ed. 1998 37 1461-1494. 55. Robson, R. J. Chem. Soc., Dalton Trans. 2000 3735-3744. 56. Moulton, B.; Zaworotko, M.J. Chem. Rev. 2001 101 1629-1658. 57. Griffith, R. L. J. Chem. Phys. 1943 11 499-505. 58. Bailar Jr., J. C. Prep. Inorg. React. 1964 1 59. Subramanian, S.; Zaworotko, M. J. Angew. Chem., Int. Ed. 1995 34 2127-2129.

PAGE 95

84 60. Hennigar, T. L.; MacQuarrie, D. C.; Losier P.; Rogers, R. D.; Zaworotko, M. J. Angew. Chem., Int. Ed. 1997 36 972-973. 61. Biradha, K.; Seward, C.; Zaworotko, M. J. Angew. Chem., Int. Ed. 1999 38 492495. 62. Yaghi, O. M.; Li, G.; Li, H. Nature 1995 378 703-706. 63. Yaghi, O. M.; Li, H.; Davis, C.; Richardson, D.; Groy, T. L. Acc. Chem. Res. 1998 31 ,474-484. 64. Li, H.; Eddaoudi, M.; O’Keeffe, M.; Yaghi, O. M. Nature 1999 402 276-279. 65. Eddaoudi, M.; Moler, D. B.; Li, H.; Chen, B. L.; Reineke, T. M.; O’Ke effe, M.; Yaghi, O. M. Acc. Chem. Res. 2001 34 319-330. 66. Chui, S. S. Y.; Lo, S. M. F.; Charmant, J. P. H.; Orpen, A. G.; Williams, I. D. Science 1999 283 1148-1150. 67. Seo, J. S.; Whang, D.; Lee, H.; Jun, S. I.; Oh, J.; Jeon, Y. J.; Kim, K. Nature 2000 404 982-986. 68. Janiak, C. J. Chem. Soc., Dalton Trans. 2003 2781-2804. 69. James, S.L. Chem. Soc. Rev 2003 32 276-288. 70. Kitagawa, S.; Kitaura, R.; Noro, S. Angew. Chem., Int. Ed. 2004 43 2334-2375. 71. Papaefstathiou, G. S.; MacGillivray, L. R. Angew. Chem., Int. Ed. 2002 41 20702073. 72. Wells, A. F. Three-Dimensional Nets and Polyhedra ; Wiley: New York, 1977 73. Wells, A. F. Further Studies of Three-dimensional Nets ; ACA Monograph: Washington, DC, 1979

PAGE 96

85 74. Wells, A. F. Structural Inorganic Chemistry, 5th Edition ; Clarendon Press, Oxford, 1984 75. Moulton, B.; Lu, J.; Mondal A.; Zaworotko, M. J. Chem. Commun. 2001 863864. 76. Bourne, S. A.; Lu, J. J.; Mondal, A.; Moulton, B.; Zaworotko, M. J. Angew. Chem., Int. Ed. 2001 40 2111-2113. 77. Lu, J. J.; Mondal, A.; Moulton, B.; Zaworotko, M. J. Angew. Chem., Int. Ed 2001 40 2113-2116. 78. Wang, Z.; Kravtsov, V. Ch.; Zaworotko, M. J. Angew. Chem., Int. Ed. 2005 44 2877-2880. 79. O’Keeffe, M.; Hyde, B. G. Crystal Structures I. Patterns and Symmetry ; Mineralogical Society of America: Washington, DC, 1996 80. Zaworotko, M. J. Chem. Commun. 2001 1-9. 81. Baerlocher, Ch.; McCusker, L. B. Database of Zeolite Structures : http://www.iza-structure.org/databases/ 82. Allen, F. H. Acta Crystallogr. Sect. B 2002 58 380-388. 83. Moulton, B.; Lu, J. J.; Hajndl, R.; Hariharan, S.; Zaworotko, M. J. Angew. Chem. Int. Ed. 2002 41 2821-2824. 84. Moulton, B.; Abourahma, H.; Bradner, M. W.; Lu, J. J.; McManus, G. J.; Zaworotko, M. J. Chem. Commun. 2003 1342-1343. 85. Greedan, J. E. J. Mater. Chem. 2001 11 37-53. 86. Abrahams, B. F.; Hoskin s, B.F.; Robson, R. J. Am. Chem. Soc. 1991 113 3606.

PAGE 97

86 87. Carlucci, L.; Ciani, G.; Macchi, P.; Proserpio, D.M. Chem. Commun. 1998 18371838. 88. Power, K. N.; Hennigar, T. L.; Zaworotko, M. J. Chem. Commun. 1998 595-596. 89. Shin, D. M.; Lee, I. S.; Chung, Y. K.; Lah, M. S. Chem. Commun. 2003 10361037. 90. Niel, V.; Thompson, A.L.; Munoz, M.C.; Galat, A.; Goeta, A.E.; Real, J.A. Angew. Chem., Int. Ed. 2003 42 3760-3763. 91. O’Keeffe, M.; Yaghi, O. M. Reticular Chemistry Structure Resource : http://okeeffe-ws1.la.asu.edu/RCSR/home.htm 92. Vertex symbol for the sodalite net: 4•4•6•6•6•6. 93. Vertex symbol for the PtS net: (4•4•82•82•88•88)( 4•4•87•87•87•87). 94. Gardner, G. B.; Venkataraman, D.; Moore, J. S.; Lee, S. Nature 1995 374 792. 95. Choi, H. J.; Suh, M. P. J. Am. Chem. Soc. 1998 120 10622. 96. Kitaura, R.; Fujimoto, K.; Noro, S.-i.; Kondo, M.; Kitagawa, S. Angew. Chem., Int. Ed. 2002 41 133. 97. Li, D.; Kaneko, K. Chem. Phys. Lett. 2001 335 50. 98. Kitaura, R.; Seki, K.; Akiyama, G.; Kitagawa, S. Angew. Chem., Int. Ed. 2003 42 428. 99. Kitagawa, S.; Kondo, M. Bull. Chem. Soc. Jpn. 1998 71 1739. 100. Kitagawa, S.; Uemura, K. Chem. Soc. Rev. 2005 34 109-119. 101. Uemuraa, K.; Matsudab, R.; Kitagawa, S. J. Solid State Chem. 2005 178 2420–2429.

PAGE 98

87 102. Matsuda, R.; Kitaura, R.; Kitagawa, S.; Kubota, Y. ; Kobayashi, T. C.; Horike, S.; Takata, M. J. Am. Chem. Soc. 2004 126 14063–14070. 103. Halder, G. J.; Kepert, C. J.; Moubaraki, B.; Murray, K. S.; Cashion, J. D. Science 2002 298 1762-1765. 104. ConQuest Version 1.8, Copyright CCDC, 2006. 105. Barbour, L. MCAVITY Molecular Cavity Calculator Copyright, 2003. 106. Webster, C. E.; Drago, R. S.; Zerner, M. C. J. Am. Chem. Soc. 1998 120 5509-5516. 107. Eddaoudi, M.; Kim, J.; Vodak, D.; Sudik, A.; Wachter, J.; O’Keeffe, M.; Yaghi, O. M. Proc. Natl. Acad. Sci. U.S.A. 2002 99 4900-4904. 108. O'Keeffe, M.; Eddaoudi, M.; Li, H.L.; Reineke, T.; Yaghi, O.M. J. Solid State Chem. 2000 152 3-20. 109. Liu, Y.; Kravtsov, V. Ch.; Larsen, R.; Eddaoudi, M. Chem. Commun. 2006 1488-1490. 110. Nicolaou, K. C.; Vourloumis, D.; Winssinger, N.; Baran, P. S. Angew. Chem. Int. Ed. 2000 39 44-122.