Zn(II) Heteroleptic Halide Complexes with 2-Halopyridines: Features of Halogen Bonding in Solid State

Reactions between Zn(II) dihalides and 2-halogen-substituted pyridines 2-XPy result in a series of heteroleptic molecular complexes [(2-XPy)2ZnY2] (Y = Cl, X = Cl (1), Br (2), I (3); Y = Br, X = Cl (4), Br (5), I (6), Y = I, X = Cl (7), Br (8), and I (9)). Moreover, 1–7 are isostructural (triclinic), while 8 and 9 are monoclinic. In all cases, halogen bonding plays an important role in formation of crystal packing. Moreover, 1–9 demonstrate luminescence in asolid state; for the best emitting complexes, quantum yield (QY) exceeds 21%.


Introduction
Halogen bonding (XB) is a specific kind of non-covalent interaction (according to IUPAC, it "occurs when there is evidence of a net 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" [1]), which was intensively investigated in recent years. Attention on this phenomenon is driven by both fundamental interest and its consideration as an additional tool for directed design of functional supramolecular systems. Indeed, formation of XB can influence different properties of compounds-in solution and, especially, in solid state. These include, in particular, luminescence, which can be amplified [2][3][4][5][6] by XB, solvatochromism [7,8], catalytic activity [9][10][11], and even odor [12]. As a result, "XB strategy" can be widely applied in materials science (especially in development of various sensors [13][14][15]).
The strategy of our work focused on two points. First, we prepared the whole [(2-XPy) 2 ZnY 2 ] series in order to see how the differences in X or Y affect the features of crystal packing and XB patterns in solid state, additionally examining the XB by theoretical methods. Second, we investigated the luminescent behavior of these complexes in comparison with other [ZnL 2 Y 2 ] compounds, in particular, with [(2-MePy) 2 ZnX 2 ] (X = Cl (10), Br (11), and I (12)). Both tasks were fulfilled; results are presented below.

Results and Discussion
According to XRD data, 1-7 are isostructural to each other (triclinic, see the experimental section), but not to 8 and 9. This fact was not obvious at the initial stage of our work, considering overall situation for [Py 2 M II X 2 ] complexes. For example, all four compounds in the [(3-XPy) 2 CuY 2 ] series (X = Cl, Br; Y = Cl, Br) [25,26,38] are monoclinic and isostructural to each other while the corresponding [(3-IPy) 2 CuX 2 ] (X = Cl, Br) [27,29] crystallize in the least symmetric group (triclinic); a similar feature can be noted for [(3-XPy) 2 PdCl 2 ] (those are isostructural for X = Cl and Br, but not I) [18]. From this point of view, the family of 1-7 is especially interesting since it allows the direct comparison of XB energies for different pairs of halogens (see below).
In all cases, Zn(II) retains the tetrahedral coordination environment ( Figure 1). Zn-Y and Zn-N bond lengths are given in Table 1; it can be seen that their differences are negligible.  A noteworthy observation can be made while analyzing the X···Y distances (d XY ) in 1-7. The identity of crystal packing results in the identity of hypothetical X···Y interaction patterns: suggesting their existence, one-dimensional chains must form via the contacts of halide ligands and halogen atoms of 2-XPy units ( Figure 2) in all cases. However, comparison of d XY with the sums of the corresponding Bondi's van der Waals radii (S XY ) [39,40] ( Table 1) indicate that the situation is actually rather different. In the complexes with 2-chloropyridine, d XY slightly (by < 0.1 Å) exceeds S XY with only two exceptions: in 4 and 7, the S XY -d XY values are +0.018 and −0.027 Å, respectively. The highest (S XY -d XY ) were observed for the complexes with 2-IPy (0.316 and 0.280 for 3 and 6, respectively). On the one hand, these facts confirm that 2-IPy is a better XB donor than 2-BrPy, and especially 2-ClPy (as it was noted in related Co(II) complexes [41]). On the other hand, this allows us to draw the hypothesis that X···Y interactions can also be present in the structures where S XY < d XY (such situations were described earlier [42,43]); to verify this, we performed DFT calculations (see below). Moreover, 8 and 9 represent the isostructural pair. Surprisingly, preparation of their single crystals of sufficient quality became a non-trivial task: after numerous XRD experiments, we succeeded in isolation of 9 to give R int = 0.037 (see Table S1 in Supplementary Materials), though revealing strong residual density peaks. For 8, the best result was 0.080; this experiment confirmed that: 1) the crystal contains only [(2-BrPy) 2 ZnI 2 ] units and 2) its cell parameters (8.7876, 14.8613, 11.7322 with β = 93.308 • ) are very similar to those in 9 (Table S1, Supplementary Materials), allowing judging on phase purity of 8. However, the quality of SCXRD data does not allow us to consider that interatomic distances can be estimated reliably in this case (this explains the "missing line" in Table 1). Interestingly, despite significant differences in symmetry and cell parameters, the patterns of X···Y interactions ( Figure 3) in 9 and, very likely, 8, are very similar to those in 1-7 (it can be seen comparing C-X-Y and Zn-Y-X angles, Table 1). The interatomic distances found in the structures reported by C. Hu (no. 1984259-1984263 in CSD, see above; all determined at room temperature) match well with those found in corresponding compounds of the 1-9 series.
Comparison of crystallographic data for 1-9 and other structurally related compounds (neutral [(2-XL) 2 MY 2 ] with tetrahedral M, L = o-halogen-substituted 6-membered unit and Y = halide) allows detecting that all those are isostructural to either 1-7 or, in one case, to 8 and 9. Interestingly, the same situation was observed for o-methyl-substituted derivatives ( Table 2)    Zn(II) 2-MePy I - [46] Cd(II) 2-MePy Br 1-7 [46] Cd(II) 2-MePy I - [46] Mg(II) 2-MePy Br 1-7 [47] For estimation of the energies of hypothetic XB in 1-9, we applied an approach that was successfully used by us [48][49][50][51] and other researchers [52][53][54][55][56][57] for relevant supramolecular systems: atomic coordinates for model clusters were obtained by XRD and used for DFT calculations and computation of the properties of electron density in the bond critical points (3, −1) within the Quantum Theory of Atoms in Molecules (QTAIM) method, "as is", without optimization (we did not use fully relaxed geometries because we were interested in evaluating the interactions as they stood in the solid state instead of finding the most global minimum energy of the complex, see Computational Details section for details). Results are summarized in Table 3; their graphical visualizations are presented in Figure 4 and Figure S23, Supplementary Materials. As follows from these data, bond critical points (3, −1) can be found even in the case of 7 where the halogen···halogen distance exceeds the sum of van der Waals radii by over 0.1 Å. This observation provides an additional argument in favor of the point of view that the "straightforward" approach towards description of non-covalent interactions in the crystalline state, based exclusively on van der Waals radii, may be misleading in certain cases. Even though such "non-conventional" contacts seem rather weak (≈1 kcal/mol), their presence can affect the packing. Moreover, results of calculations confirm the conclusions made by us earlier: [41]; the ability of coordinated 2-XPy to serve as XB donors increase in Cl < Br < I row, so that the highest energies were found for I···Cl interactions in 3 (3.8 kcal/mol). Table 3. Values of the density of all electrons-ρ(r), Laplacian of electron density-∇ 2 ρ(r) and appropriate λ 2 eigenvalues, energy density-H b , potential energy density-V(r), and Lagrangian kinetic energy-G(r) (a.u.) at the bond critical points (3, −1), corresponding to the non-covalent interactions X···Y (X, Y = Cl, Br, I) in 1-7 and 9, as well as energies for these contacts E int (kcal/mol), defined by different approaches.  As follows from PXRD and element analysis data (see Supplementary Materials), 1-9 can be prepared as pure phases, making investigation of their luminescent properties possible. All complexes reveal emission in the blue-green range when irradiated with UV (spectra for 1 are presented on Figure 5, for all other complexes-in Supplementary Materials). A short summary is given in Table 4. The most intensive peaks are between 360 and 520 nm. Emission observed in 7 and 9 is found to be red-shifted with respect to those of bromo-and chloro complexes, following the trend I > Br > Cl. This effect in emission maxima is associated to an increasing electrondonating nature (I − < Br − < Cl − ) of the halide ligands. For iodide complexes, emission is fundamentally different from those for Cl-and Br-containing species.   All complexes reveal fluorescence behavior with lifetime decay of few nanoseconds, except the complexes for which it was not possible to measure lifetime decay due to very poor emissive response when excited with the pulsed source. The compounds show a mono exponential fit of each components of the decay curve, with lifetimes in the range 0.2-8.2 ns. The values reported are the average of three independent determinations for each sample. A decrease of emission quantum yield from chloride to iodide in the series of complexes is likely a consequence of the increase in the atomic number of the halogen atom bound to the zinc center; this causes a spin-orbit enhancement, which in turn favors intersystem crossing.

Materials and Methods
All reagents were obtained from commercial sources and used as purchased. Solvents were purified according to the standard procedures. Complexes 10-12 were prepared similarly to the procedure described earlier [46], (corresponding Zn(II) halide and 2-MePy (1:2) in ethanol) and identified by PXRD and element analysis data. All experiments were performed at room temperature.

X-ray Diffractometry
Data sets for single crystals of 1-7 were obtained at 130 (1 and 2), 140 (3)(4)(5), 150 (6), or 143 (7) K on Agilent Xcalibur diffractometer equipped with an area AtlasS2 detector (graphite monochromator, λ(MoKα) = 0.71073 Å, ω-scans). Integration, absorption correction, and determination of unit cell parameters were performed using the CrysAlisPro program package (CrysAlisPro 1.171.38.41. Rigaku Oxford Diffraction: the Woodlands, TX, USA, 2015). For the crystal of 9, the data were obtained on Bruker D8 Venture diffractometer with a CMOS PHOTON III detector and IµS 3.0 source (Mo Kα radiation, λ = 0.71073 Å). All measurements were performed at 150 K, the ϕand ω-scan techniques were employed. Absorption correction was performed using the SADABS program (Bruker Apex3 software suite: Apex3, SADABS-2016/2 and SAINT, version 2018.7-2; Bruker AXS Inc.: Madison, WI, USA, 2017). The structures were solved by a dual space algorithm (SHELXT) and refined by the full-matrix least squares technique (SHELXL) [62] in the anisotropic approximation (except hydrogen atoms). Positions of hydrogen atoms of organic ligands were calculated geometrically and refined in the riding model. The crystallographic data and details of the structure refinements are summarized in Table S2 (Supplementary Information). CCDC 2059579-2059586 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Center at http://www.ccdc.cam.ac.uk/data_request/cif.

Computational Details
The single point calculations based on the experimental X-ray geometries of 1-7 and 9 were carried out at the DFT level of theory using the dispersion-corrected hybrid functional ωB97XD [63] with the help of the Gaussian-09 program package. The second-order scalar relativistic Douglas-Kroll-Hess calculations, requested relativistic core Hamiltonian, were carried out using the DZP-DKH basis sets [64][65][66][67] for all atoms. The topological analysis of the electron density distribution, with the help of the atoms in molecules (QTAIM) method developed by Bader [68], was performed by using the Multiwfn program (version 3.7) [69]. The Cartesian atomic coordinates for model supramolecular trimeric associates are presented in Table S3. Currently, two general theoretical approaches for studies of non-covalent interactions in the solid state are widespread. The first is the "molecular" approach that is typically applied for molecular crystals, and it includes modeling of a separate isolated supramolecular adduct without considering the neighboring molecular environment, and periodicity of the real crystal (for reviews see [70] and [71]). This is a rather rough approximation, but it is useful when fine effects are not under study and high accuracy is not needed. The second approach typically includes time-demanding "true" periodic conditions calculations. It is perfectly correct for highly symmetrical ionic crystals (for review see [72]), but has certain limitations for less symmetric, disordered molecular crystals, and crystallosolvates. In our recent works [73,74], we showed that a single point "molecular" approach for calculation of supramolecular associates, particularly involving transition metal complexes, agree well with Kohn-Sham calculations, with periodic boundary conditions. Moreover, other researchers widely used such single point "molecular" approaches in studies on halogen bonding and other non-covalent interactions in similar chemical systems [52,55,56].

Powder X-ray Diffractometry (PXRD)
Details are provided in Supplementary Material.

Luminescence Spectra
Registration of emission and excitation spectra at room temperature, as well as determination of absolute emission quantum yields, were performed with Fluorolog-3 (Horiba Jobin Yvon), equipped with a 450 W Xe lamp, an integration sphere, double grating excitation, and emission monochromators. Emission and excitation spectra were corrected for source intensity (lamp and grating) and emission spectral response (detector and grating) by standard correction curves. Lifetime measurements were performed on a Horiba Fluo-roCube lifetime instrument by a time-correlated single-photon counting method using a 350 nm LED excitation source.

Conclusions
Despite overall similarity of XB patterns in the crystal structures of [(2-XL) 2 ZnY 2 ] complexes, not all members of this group are isostructural. The presence of halogen···halogen non-covalent interactions was detected by means of QTAIM analysis, even in cases where corresponding distances exceeded the sum of the van der Waals radii. Considering that similar situations were reported in previous studies, it can be concluded that judging on presence (or absence) of non-covalent interactions, based only on analysis of distances, can be misleading: in some "borderline" cases, DFT calculations give more reliable answers. Although most [(2-XL) 2 ZnY 2 ] reveal moderate luminescence, there are two compounds demonstrating quantum yields exceeding 20%. Taking into account that XB can affect photophysical characteristics [2], examination of luminescent behavior of 1-9 in presence of other XB-forming building blocks (i.e., perfluoroiodoarenes) can be an interesting research task; corresponding experiments are underway in our group.