Supramolecular Sandwiches: Halogen-Bonded Coformers Direct [2+2] Photoreactivity in Two-Component Cocrystals

The halogen-bond (X-bond) donors 1,3- and 1,4-diiodotetrafluorobenzene (1,3-di-I-tFb and 1,4-di-I-tFb, respectively) form cocrystals with trans-1,2-bis(2-pyridyl)ethylene (2,2′-bpe) assembled by N···I X-bonds. In each cocrystal, 2(1,3-di-I-tFb)·2(2,2′-bpe) and (1,4-di-I-tFb)·(2,2′-bpe), the donor molecules support the C=C bonds of 2,2′-bpe to undergo an intermolecular [2+2] photodimerization. UV irradiation of each cocrystal resulted in stereospecific and quantitative conversion of 2,2′-bpe to rctt-tetrakis(2-pyridyl)cyclobutane (2,2′-tpcb). In each case, the reactivity occurs via face-to-face π-stacked columns wherein nearest-neighbor pairs of 2,2′-bpe molecules lie sandwiched between X-bond donor molecules. Nearest-neighbor C=C bonds are stacked criss-crossed in both cocrystals. The reactivity was ascribed to the olefins undergoing pedal-like motion in the solid state. The stereochemistry of 2,2′-tpcb is confirmed in cocrystals 2(1,3-di-I-tFb)·(2,2′-tpcb) and (1,4-di-I-tFb)·(2,2′-tpcb).


Introduction
Halogen bonding (X-bonding) is a well-established [1] supramolecular synthon, as this directional, noncovalent force can reliably direct the self-assembly of molecules into supramolecular architectures [2,3]. Electrophilic regions associated with halogen (X) atoms that serve as X-bond donors recognize complementary, nucleophilic (:Nu) functionalities (e.g., N-atoms) on acceptor molecules such that an X-bond, a special class of Lewis acid-base adduct, X···Nu, results [4].
Due to the highly directional character of X-bonds, X-bond donor molecules with rigid geometries have attracted much interest in the field of crystal engineering. A necessary, albeit admittedly ambitious, aim is to achieve a functional level of predictive power in the design of desired geometries and/or topologies based principally on geometries of the X-bond donor and acceptor molecules [2,3]. Successes would be invaluable in the design of functional materials and would constitute a major accomplishment for the fields of crystal engineering and supramolecular chemistry.
In this context, an application of X-bonding of interest to our group is that of exploiting X-bonds as supramolecular synthons to design photoreactive solids. We have devoted efforts to investigate whether multicomponent systems can be assembled into predictable geometries based primarily on the directional character of X-bonds and molecular structures of the components. Ideally, judicious choice of coformers would enable access to solids of functional significance, such as reactivity [5][6][7][8][9].
Perfluorinated iodoarenes are well-established X-bond donors and are currently among the most frequently employed building blocks in the crystal engineering of X-bonded materials. In the context of Scheme 1. Components for cocrystals.

Cocrystal
Primary Assembly Secondary Assembly Photoreactivity infinite 1D chain based on N···I and face-to-face π-stacks --

General Experimental
All reagents and solvents (synthesis grade) were purchased from commercial sources and used as received unless otherwise stated. 1,3-diiodotetrafluorobenzene (1,3-di-I-tFb) was purchased from Apollo Scientific; 1,4-diiodotetrafluorobenzene (1,4-di-I-tFb) was purchased from Aldrich; trans-1,2bis(2-pyridyl)ethylene (2,2′-bpe) was purchased from TCI America; and chloroform (CHCl3; certified ACS grade, ≥99.8%, approximately 0.75% EtOH as preservative) was purchased from Fisher Chemical. All cocrystal syntheses were conducted in screw-cap glass scintillation vials. For cocrystal syntheses, "thermal dissolution" refers to the process of combining both solid cocrystal components in a vial, adding solvent portion-wise while maintaining a saturated mixture at room temperature (rt), and tightly capping the vial and heating the mixture on a hot-plate until all solids dissolve to afford a homogeneous solution with the minimum necessary volume of solvent. Compositions of all single crystals were shown to be representative of the bulk material by matching experimental powder X-ray diffraction (pXRD) patterns with those simulated from single-crystal X-ray diffraction (scXRD) data. Yields refer to isolated yields of analytically-pure compounds unless otherwise stated. Melting points (mp) were recorded on samples in open capillary tubes using a manual MEL-TEMP apparatus (Electrothermal Corporation, Staffordshire, UK) and were uncorrected. Photoreactions were conducted in either a NuLink 36 W UV lamp apparatus (λ = 400 nm) or an ACE photocabinet

Powder X-ray Diffraction (pXRD)
Powder X-ray diffraction data were collected at room temperature on a Bruker D8 Advance X-ray diffractometer on samples mounted on glass slides. Each sample was finely ground using an agate mortar and pestle prior to mounting. Instrument parameters: radiation wavelength, CuKα (λ = 1.5418 Å); scan type, coupled TwoTheta/Theta; scan mode, continuous PSD fast; scan range, 5-40 • two-theta; step size, 0.02 • ; voltage, 40 kV; current, 30 mA. Background subtractions were applied to all experimentally collected data within the Bruker DIFFRAC.EVA v3.1 software suite. All data were plotted in Microsoft Excel 2016. Simulated pXRD patterns were calculated from scXRD data within the Cambridge Crystallographic Data Centre (CCDC) Mercury [27] software suite.

Single-Crystal X-ray Diffraction (scXRD)
Single-crystal X-ray diffraction data were collected on either a Bruker Nonius-Kappa APEX II CCD or a Bruker Nonius-Kappa CCD diffractometer, each equipped with an Oxford Cryosystems 700 series cold N 2 gas stream cooling system. Data were collected at either room temperature (298.15 K) or low temperature (150.15 K) using graphite-monochromated MoKα radiation (λ = 0.71073 Å). Crystals were mounted in Paratone oil on a MiTeGen magnetic mount. Data collection strategies for ensuring maximum data redundancy and completeness were calculated using the Bruker Apex II TM software suite. Data collection, initial indexing, frame integration, Lorentz-polarization corrections, and final cell parameter calculations were likewise accomplished using the Apex II software suite. Multi-scan absorption corrections were performed using SADABS [28]. Structure solution and refinement were accomplished using SHELXT [29] and SHELXL [30], respectively, within the Olex2 [31] graphical user interface. Space groups were unambiguously verified using the PLATON [32] executable. All non-hydrogen atoms were refined anisotropically. All hydrogen atoms were attached via a riding model at calculated positions using suitable HFIX commands. The occupancies of the major and minor positions for the disordered C=C cores within 2(1,3-di-I-tFb)·2(2,2 -bpe) and (1,4-di-I-tFb)·(2,2 -bpe) and for the disordered cyclobutane C-C cores within 2(1,3-di-I-tFb)·(2,2 -tpcb) and (1,4-di-I-tFb)·(2,2 -tpcb) converged to their respective ratios after each was identified in the difference map and freely refined. Figures of all structures were rendered in the CCDC Mercury [27] software suite.

Conclusions
We demonstrated that X-bonds support intermolecular [2+2] photodimerizations of 2,2'-bpe in the solid state. The reaction occurred stereospecifically and quantitatively to generate 2,2 -tpcb. The cocrystals 2(1,3-di-I-tFb)·2(2,2 -bpe) and (1,4-di-I-tFb)·(2,2 -bpe) provide mounting evidence that X-bonds can be utilized in a more general way to sustain the formation of C-C bonds in solids. Our efforts are now focused on rational approaches to utilize X-bonds to direct stacking and control reactivity in the solid state.