Weak intermolecular interactions of various types are of great interest for many years [1-4]. Special attention is paid to interactions called halogen bonds [5, 6]. It was demonstrated that halogen atoms in inorganic and organic halides can form complexes with molecules containing heteroatoms with lone pairs, for example, oxygen and nitrogen [7, 8]. At first, such interactions were found in the crystal phase [9], but evidence of their existence in a solution [10, 11] and even in a gas phase [12, 13] was also presented. The further study of halogen bonds has shown that these interactions are important for the formation of supramolecular complexes [14-16], protein-ligand interactions [17-22], crystal structures [23, 24].
As recommended by IUPAC, a halogen bond can be defined as “…an attractive interaction between an electrophilic region associated with a halogen atom in a molecular entity and a nucleophilic region in another or the same, molecular entity” [25]. Halogen bonds are described in the same way as hydrogen bonds. Exactly, a halogen bond is denoted by the three dots in R–X…Y, where R–X (X = Cl, Br, I) is the halogen bond donor with an electron-poor region and Y is the halogen bond acceptor containing an electron-rich region. Such a definition is based on the theoretical analysis of halogen bonding. A positive electrostatic potential region has to be located on the outermost portion of the halogen’s surface centered on the R–X axis and was named “σ-hole” [26-28]. The strong directionality and electrostatic character of halogen bonding were expected [25, 29, 30], but a more detailed study of the nature of this interaction resulted in the conclusion that halogen bonds depend more strongly on the contributions from dispersion forces [31, 32].
The next question concerns the interaction strength of halogen bonding. The strength of a classical hydrogen bond can be evaluated from the characteristics of a (3,-1) bond critical point determined within Bader’s theory “Atoms in molecules” between a hydrogen atom and a proton acceptor atom [33]. According to Espinosa’s correlated equation, the potential energy in a (3,-1) bond critical point correlates with the hydrogen bond energy [34]. However, the application of this equation to weak interactions like halogen bonds is very questionable. Another way to estimate the strength of intermolecular interactions is by studying their role in crystal packing formation [35]. Such a method was applied for the analysis of intermolecular interactions in crystals of the simplest halomethanes [36-38] where halogen bonds were expected to be the strongest interactions. However, the role of halogen bonds in crystal structure formation proved to be smaller than the role of non-specific interactions indicating the weak character of these interactions.
Despite their low strength, weak intermolecular interactions are quite effective in the formation of host-guest complexes [39-41]. The search for new efficient host molecules for molecular complexes led us to the synthesis (Scheme 1) of the fluorenonophane 3 representing a new group of macrocyclic receptors for organic molecules. Earlier, it was demonstrated that such types of hosts have a hollow intramolecular cavity limited by two fluorenone and two phenylene fragments [42]. Geometric parameters of the cavity allow different organic molecules, especially aromatic and heteroaromatic molecules, to settle easily within this cavity forming stable molecular complexes. The main intermolecular interaction that stabilizes complexes of such a type is the stacking between electron-rich fragments.
In this paper, we present the complexes of fluorenonophane 3 with the chloroform (3Cl) and bromoform (3Br). The halogen….π interactions are expected to be the strongest in these complexes providing their existence.
Experimental part.
Synthesis
Dibromide 1 was prepared according to the literature procedure [43], hydroquinone monobenzoate is commercially available and was used without further purification. Anhydrous acetonitrile and DMF were distilled from calcium hydride, other solvents were used as received. Column chromatography purification was performed with Silica gel 60 (Merck, 0.063–0.100 mm). Thin-layer chromatography (TLC) was performed on Merck pre-coated plates (silica gel 60 F254). The plates were inspected by fluorescence quenching under UV light or, if required, developed in I2 vapor. 1H NMR spectra were recorded with a Varian VXR-300 spectrometer at 300 MHz with DMSO-d6 as a solvent. Chemical shifts were reported in ppm downfield from internal Me4Si. Electron impact mass spectra were obtained from an MX-1321 mass spectrometer with direct sample admission into the ion source operating at 70 eV.
Bisphenol 2. Dibromide 1 (3.66 g, 10 mmol) and hydroquinone monobenzoate (5.14 g, 24 mmol) were added to a suspension of K2CO3 (9.96 g, 72 mmol) in dry MeCN (200 mL) under an argon atmosphere. The suspension was stirred vigorously and heated under reflux for 15 h. After cooling down to room temperature, the suspension was filtered and the solvent was removed in vacuo. The residue was washed with diluted HCl (1:10) (50 mL), H2O (50 mL), and hot MeOH (30 mL). The crude bisbenzoate was added to a solution of KOH (2.72 g, 49 mmol) in H2O (27 mL) and EtOH (523 mL). The reaction mixture was refluxed then for 6.5 h. After cooling down to room temperature, the solvent was removed in vacuo, diluted HCl (1:10) was added to the residue to adjust the pH of the aqueous layer to 5 and the crude bisphenol was extracted with CHCl3 (4´100 mL). The organic phase was washed with saturated aqueous NaHCO3 (2´200 mL), H2O (200 mL), and saturated aqueous NaCl (200 mL) and then dried (MgSO4). The solvent was removed in vacuo to afford 2 as a yellow solid (2.33 g, 55%): mp > 250°C dec. EIMS 424 (M+); 1H NMR δ 5.00 (s, 4H), 6.65 (d, J = 8.7 Hz, 4H), 6.79 (d, J = 9.0 Hz, 4H), 7.57–7.64 (m, 4H), 7.71 (d, J = 8.4 Hz, 2H), 8.80 (s, 2H). Anal. Calcd. for C27H20O5: C, 76.40; H, 4.75. Found: C, 76.27; H, 4.99.
Cyclophane 3. A solution of 1 (732 mg, 2 mmol) and 2 (848 mg, 2 mmol) in dry degassed DMF (60 mL) was added dropwise over 10 h to a stirred suspension of K2CO3 (1.656 g, 12 mmol) in DMF (140 mL) under argon at 80°C and the reaction mixture was then heated and stirred for further 45 h. After being cooled down to room temperature, the suspension was filtered off, the solid was washed with DMF (2´20 mL), and the solvent was removed in vacuo. The residue was partitioned between CHCl3 (300 mL) and aqueous 5% NaOH (70 mL). The organic phase was washed with H2O (100 mL) and saturated aqueous NaCl (100 mL) and then dried (MgSO4). Evaporation of the solvent afforded a residue which was subjected to column chromatography (CHCl3/MeOH, 100:1) to yield 3 as a light-yellow solid (188 mg, 15%): mp > 250°C dec. EIMS 628 (M+); 1H NMR δ5.15 (s, 8H), 6.71 (s, 8H), 7.44 (s, 4H), 7.55 (s, 8H). Anal. Calcd. for C42H28O6: C, 80.24; H, 4.49. Found: C, 80.38; H, 4.63.
X-ray diffraction study.
The crystals of 3Cl (C42H28O6∙2CCl3) are triclinic, space group P. At 100 K a = 9.163(2) Å, b = 10.802(3) Å, c = 11.757(4) Å, a = 113.59(3)o, b = 95.10(3)o, g = 112.77(2)o, V = 941.5(5) Å3, dcalc = 1.530 g/cm, Z = 1, m = 0.509 mm-1. Intensities of 23762 reflections (5472 unique reflections, Rint = 0.025) were measured using an “Xcalibur-3” diffractometer (graphite-monochromated MoKa radiation, w-scan, a CCD detector, 2qmax = 60o). Absorption corrections were performed using the multi-scans method (Tmin = 0.898, Tmax = 0.971).
The crystals of 3Br (C42H28O6∙2CBr3) are monoclinic, space group P21/n. At 100 K a = 10.9417(2) Å, b = 12.4492(2) Å, c = 14.4471(3) Å, b = 90.933(2)o, V = 1967.66(6) Å3, dcalc = 1.914 g/cm, Z = 2, m = 6.172 mm-1. Intensities of 8506 reflections (4401 unique reflections, Rint = 0.029) were measured using an “Xcalibur-3” diffractometer (graphite-monochromated MoKa radiation, w-scan, a CCD detector, 2qmax = 55o). Absorption corrections were performed using the multi-scans method (Tmin = 0.468, Tmax = 0.732).
The structures were solved by the direct method using the SHELXTL program package [44, 45] implemented in the OLEX2 program [46]. Positions of the hydrogen atoms were located from electron density difference maps and refined by the “riding” model with Uiso = 1.2Ueq of the carrier atom. Full-matrix least-squares refinement of the structures against F2 in anisotropic approximation for non-hydrogen atoms using 5472 (3Cl), 4401 (3Br) reflections was converged to: wR2 = 0.049 (R1 = 0.035 for 4141 reflections with F>4σ(F), S = 1.092) for structure 3Cl and wR2 = 0.073 (R1 = 0.036 for 2684 reflections with F>4σ(F), S = 0.917) for structure 3Br. The final atomic coordinates and crystallographic data for 3Cl and 3Br have been deposited with the Cambridge Crystallographic Data Centre, 12 Union Road, CB2 1EZ, UK (fax: +44-1223-336033; e-mail: [email protected]) and are available on request quoting the deposition numbers CCDC 647971 for 3Cl and CCDC 2098245 for 3Br.
Analysis of crystal structures from the energetic viewpoint.
Crystal structure analysis was performed within the approach based on quantum chemical calculations of pairwise interaction energies between molecules in a crystal [47]. Any molecule in a crystal may be considered as a basic unit of a crystal packing (BU0), and its first coordination sphere can be constructed using the standard procedure within the Mercury program [48]. This option allows us to determine all molecules for which the distance between atoms of the basic BU0 and its symmetrical equivalents is shorter than van der Waals radii sum plus 1 Å at least for one pair of atoms. In the case of Z′ > 1, this procedure should be applied to each of the molecules found in the asymmetric part of the unit cell. The selected fragment of the crystal packing can be divided into dimers where one molecule is basic and the other one belongs to its first coordination sphere. The molecular geometries of these dimers were not optimized. Taking into account the well-known effect of X−H bonds shortening in the X-ray diffraction study [49], the positions of hydrogen atoms were normalized to 1.089 Å for C−H and 0.993 Å for O−H bonds, according to neutron diffraction data [50]. The pairwise interaction energies were calculated using the B97D3 density functional method [51] with the def2-TZVP basis set [52, 53] and corrected for a basis set superposition error by the counterpoise method [54]. All single-point calculations were performed within the Gaussian03 software [55].
The energy-vector diagrams (EVD) were proposed for the graphic representation of the obtained data [56]. The calculated interaction energy between two molecules takes on vector properties if it originates from the geometrical center of the basic BU0 and is directed toward the geometrical center of a symmetrically equivalent BUi. The use of such an assumption makes it possible to visualize the interaction energies in a crystal as a set of such vectors (Li) originating from the geometrical center of the basic BU. The length of each energy vector Li is calculated using the following equation:
where Ri is the distance between the geometrical centers of the interacted building units BU0-BUi, Ei is their interaction energies, and Estr is the strongest pairwise interaction energy in the crystal.
The energy-vector diagram represents the image of a molecule in terms of the strength and directionality of its intermolecular interactions in a crystal. Replacing the basic molecule with such a vector image and applying all symmetry operations to it result in the visualization of a crystal packing in terms of interaction energies between molecules. This method allows us to define the most strongly bound fragments of a crystal packing such as a primary basic structural motif (BSM1) or a secondary basic structural motif (BSM2) [57].
Analysis of the interaction energy
The estimation of the true energy of halogen bonding
To evaluate the true energy of halogen bonding the approach proposed earlier [58] was used. The difference in intermolecular interaction energies for halogen bonded dimers and model dimers where the interacting halogen is replaced by the hydrogen atom may be used as the approximate (within 0.1÷0.2 kcal/mol) true energy of a halogen bond.
Decomposition of the Interaction Energy
The contribution of all types of interactions to the total interaction energy was studied using the modified method of Morokuma and Kitaura [59] namely Localized Molecular Orbital Energy Decomposition Analysis (LMOEDA) [60] implemented in the GAMESS-US software package [61]. For these calculations, we used the geometry of molecule pairs extracted from the experimental data. The calculations of interaction energies were carried out using the m06-2x method [62] and the 6-311G(d,p) basis set [63] and corrected for a basis set superposition error by the counterpoise method. The accuracy of the DFT grid [64] was increased to ultrafine (99 radial shells with 590 Lebedev points in each). Pulay's direct inversion of the iterative subspace (DIIS) interpolation [65, 66] was used in these calculations to increase the convergence speed.