Synthesis of “Acetylene-Expanded” Tridentate Ligands

Synthetic routes to four new tridentate ligands with large cavities have been developed. Each ligand features two halides at the termini of the molecules that could be used for further elaboration of the system. Such compounds are ideal for encapsulating organoiodide guests using charge-transfer interactions.


Introduction
Aromatic heterocycles have found considerable utility in supramolecular synthesis [1,2]. Their rigid, planar structure and rich substitution chemistry have allowed the deliberate construction of multidentate host molecules with precisely engineered geometries. In most cases, the guests in these systems have been metal atoms, ions or complex fragments. In addition to acting as metal ligands, however, aromatic heterocycles can form charge-transfer complexes with highly polarizable molecular species such as I 2 or organoiodides [3][4][5][6]. This interaction is highly directional and has a strength on the order of a hydrogen bond. As part of a crystal engineering effort to design nanoporous solids with electron-rich cavities, we have found it necessary to develop synthetic routes to "acetylene-expanded" multi-ring hosts. In these systems, the heterocycles are separated by acetylene or a multi-acetylene linkage. Related systems have recently been prepared for use in molecular electronic [7,8] and nonlinear optical devices [9] and as precursors for topopolymerization [10][11][12]. The cavity sizes and shapes in the new hosts reported here are designed to accomodate large organoiodide guests and to assemble in a predictable manner. These compounds are also potential precursors to large macrocycles [13,14].

Results and Discussion
Our synthetic strategy involves the coupling of a symmetrical "base" heterocycle with an asymmetrical "arm" heterocycle (Scheme 1). The diethynyl heterocycles 1 and 2 are readily prepared by palladium-catalysed coupling of the corresponding 2,6-dibromopyridine or 2,5-dibromothiophene with trimethyl-silylacetylene, followed by removal of the TMS protecting group with KOH [13,14]. Both of these base units are best stored cold and in the absence of light to prevent decomposition. Arm unit 4 could be prepared in better than 90% yield by heating 2-bromothiophene in the presence of iodine and nitric acid [15], but 2-bromo-6-iodopyridine (3) proved to be more difficult to synthesize. A simple synthetic route to 3 was recently reported [16], though no spectroscopic data was provided. In this procedure, 2,6-dibromopyridine was treated with isopropylmagnesium chloride and iodine to give the desired product in a reported 90% isolated yield. Despite numerous attempts however, in our hands these reaction conditions inevitably resulted in a mixture containing substantial amounts of the starting dibromide and diiodopyridine as well as the desired product. The compounds proved to be impossible to separate, forcing us to develop an alternative route.
Using the method of Johnson and co-workers [17,18], 2-amino-6-bromopyridine was prepared from epichlorohydrin in 48% yield (this compound is now available from Aldrich). Conversion of this species to 3 using Sandmeyer conditions was found to be problematic, probably due to the basicity of the pyridine ring. Treatment with isoamyl nitrite in diiodomethane however, gave 3 in moderate yield after column chromatography [19].
Compounds 5-8 could be prepared by Sonogashira-Hagihara coupling [20] of the appropriate base and arm groups in a 1:2 ratio. The reactions were run in diethylamine at room temperature for 12 hours, giving the desired products in moderate yields (6-63%). In the cases of 5-7, chemoselective reaction at the iodo position was observed. All new compounds were characterized by IR, NMR and mass spectrometry, and each gave satisfactory elemental analysis.
Compounds 5-8 are each expected to have an arm heteroatom to arm heteroatom distance of more than 9 Å. This is large enough to serve as a guest for organoiodides such as tetraiodoethylene or odiiodobenzenes [3][4][5][6]. Single crystals of 7, grown from methylene chloride, were analyzed by X-ray crystallography. The heteroatoms in the arm groups are rotated to the interior of the molecular cavity ( Figure 1), indicating little communication between the lone pairs. The molecule is nearly planar in the solid state, with the thiophene arms rotated out of the molecular plane by 19.8° and 29.9° (Fig 2).  Compound 5 shows extremely limited solubility in organic solvents, diminishing its usefulness for further synthetic elaboration. We are currently developing routes to derivatives of 5 in which the bromine atoms have been replaced by other, solublizing substituents.
The incorporation of thiophene rings into the system improves solubility. Unfortunately, we have been unable to effect coupling of the terminal thiophene rings in compound 7 to acetylenes. In an attempt to increase the reactivity of the halide, we reacted 1 with diiodothiophene to produce 8. Interestingly, the yield of this reaction was much worse than in the synthesis of 7. Compound 8 also failed to couple under Sonogashira-Hagihara or related coupling conditions. We do not yet understand the reason for this lack of reactivity, but it may be due to chelation of the palladium catalyst by the tricyclic systems or due to halogen bonding between heteroatoms and the halide.
Compound 6 is both soluble enough to work with and is sufficiently reactive to undergo palladiumcatalyzed coupling to add additional acetylene groups onto the complex (9, Scheme 2). This deprotects cleanly to give 10. We will describe the use of 10 and related complexes as precursors to macrocycles and nanoporous solids elsewhere.

Scheme 2 Conclusions
Palladium-catalyzed coupling of symmetrical bis(acetylene) heterocycles with asymmetric dihaloheterocycles results in new tridentate ligands with large cavities. Thiophene-containing systems are more soluble than the trispyridine complex, but terminal thiophenes are unreactive towards further coupling. Halogens on terminal pyridines will couple with acetylenes in modest yields to form viable precursors for macrocycles.

Acknowledgements
This work was supported by the Petroleum Research Fund.

Experimental
General 1 H-and 13 C-NMR spectra were obtained using a Bruker WM-300 spectrometer. Mass spectrometry data were obtained using a Kratos MALDI-TOF MS. All reagents were obtained from Aldrich Chemical Company (USA) and were used as received. Solvents were obtained from commercial sources and were dried and/or purified by standard techniques and stored over activated sieves when necessary. Carbon, hydrogen, and nitrogen analyses were performed by Atlantic Microlabs, Norcross GA.