Exploring Supramolecular Assembly Space of Cationic 1,2,4-Selenodiazoles: Effect of the Substituent at the Carbon Atom and Anions

Chalcogenodiazoles have been intensively studied in recent years in the context of their supramolecular chemistry. In contrast, the newly discovered cationic 1,2,4-selenodiazole supramolecular building blocks, which can be obtained via coupling between 2-pyridylselenyl halides and nitriles, are virtually unexplored. A significant advantage of the latter is their facile structural tunability via the variation of nitriles, which could allow a fine tuning of their self-assembly in the solid state. Here, we explore the influence of the substituent (which derives from the nitrile) and counterions on the supramolecular assembly of cationic 1,2,4-selenodiazoles via chalcogen bonding.


Introduction
Recent decades have seen significant progress in the creation of artificial supramolecular systems with increasing complexity. Hydrogen bonding (HB) and coordination at the metal center has been employed often in the creation of complex structures from simpler building blocks [1][2][3][4]. Although the construction of larger and more complex structures is an important task within this field, the other major challenges include the utilization of other weak interactions (apart of HB and metal coordination) for obtaining more complex aggregates, determining recognition units, as well as designing and synthesizing novel recognition units. The attractive non-covalent interactions between polarizable main group elements and Lewis basic species, which include chalcogen bonding (ChB) and halogen bonding (XB), have recently attracted considerable attention and emerged as powerful alternatives to complexation and HB in supramolecular chemistry [5][6][7][8][9][10][11][12].
We recently introduced novel cationic 1,2,4-selenodiazole building blocks, which can be easily prepared via the coupling between 2-pyridylselenyl halides and nitriles in high yields [22,23]. In contrast to the exhaustively studied earlier chalcogenodiazoles, the supramolecular chemistry of these novel 1,2,4-selenodiazole is virtually unexplored, but could be a fruitful research field in the near future. A significant advantage of novel 1,2,4-selenodiazole building blocks is their facile structural tunability via the variation of nitriles, which could allow the fine tuning of their self-assembly in the solid state. Here, we explore the effect of the substituent (which derives from the nitrile) on the supramolecular assembly of cationic 1,2,4-selenodiazoles via ChB non-covalent interactions.
Earlier, we established that the adduct derived from propionitrile formed [Se-Cl] 2 squares in the solid state. Interestingly, when we switched higher homologs (i.e., n-butyronitrile and even n-hexanenitrile), the corresponding adducts 3 and 4 did not form dimers in the solid state, but self-assembled into 1-D supramolecular polymers via Se···Cl ChB and H···Cl HB ( Figure 2). In these cases, other weak interactions outcompeted [Se-Cl] 2 and [Se-N] 2 square formations in the solid state. Expectedly, the transoid position towards the N atom around the selenium center was occupied by the chloride due to the presence of the supporting H···Cl HB, but the cisoid position remained unoccupied in the crystals of 3 and 4.  When the solution of 3 was slowly crystallized from dichloromethane over a period of ca. one month, aside from the major product 3, a minor amount of crystals of 6 also precipitated ( Figure 3). Compound 6 represented an adduct of 1,2,4-selenadiazole 3 with 2-pyridylselenylchloride hydrochloride. The heterodimer 6 featured Se-N-Se-Cl squares, which we have not observed previously for the adducts of nitriles with 2-pyridylselenylhalides. Compound 6 could potentially form via partial hydrolysis of 3, which results in the formation of HCl, and could react with another molecule of 3 to generate 2-PySeCl HCl. The latter could coprecipitate with an additional molecule of 3 to form 6. However, further investigations are required for detailed insight of unusual heterodimers containing Se-N-Se-Cl squares.
Further we switched to a bulkier diphenylacetonitrile, which formed an expected adduct 5 with 2-pyridylselenylchloride and self-assembled into 1-D chains via Se···Cl and H···Cl interactions ( Figure 4) in the solid state, in the same fashion as 3 and 4. However, in contrast to 3 and 4, adduct 5 contained a cocrystallized CH 2 Cl 2 molecule which featured a ChB interaction with a selenium center and occupied a cisoid position at the selenium atom ( Figure 4). Arguably, this was allowed by the pocket formed by the bulky diphenylmethyl substituent. . Ball-and-stick representation of the crystal structure of 5, demonstrating its self-assembly into 1-D supramolecular polymer via Se···Cl and H···Cl interactions and ChB interactions with a cocrystallized CH 2 Cl 2 molecule. Green, purple, blue, grey, and light-grey spheres represent chlorine, selenium, nitrogen, carbon, and hydrogen, respectively.
Previously, we showed that 2-pyridylselenylchloride eagerly reacts with cyanogen bromide forming a corresponding cationic 1,2,4-selenadiazole in an almost quantitative yield [23]. Within this work, we explored the reaction of 2-pyridylselenylbromide with cyanogen bromide, which also resulted in the formation of the adduct 8, which selfassembled in the solid state in the same manner as its chloride analog, which featured [Se-Cl] 2 square dimers in the solid state ( Figure 5). Interestingly, switching from a chloride to a bromide resulted in an intriguing peculiarity: the Br(C)-Br distance (3.678 Å) in the crystal of 8 was shorter than the Br(C)-Cl separation (3.774 Å) in the analogous adduct of 2-pyridylselenylchloride with cyanogen bromide [23]. This indirectly indicates that the Br···Br interactions in 8 are attractive. Overall, molecules of 8 formed 1-D polymer via Se···Br ChB and Br···Br XB in the crystal. Further, we were interested in how variation of the anion would affect the selfassembly of the adduct 2-pyridylselenylchloride with cyanogen bromide. For this reason, we performed an addition of a saturated solution of NaBPh 4 in MeOH to a solution of 8 in the same solvent, which resulted in an immediate formation of microcrystalline precipitate of 9 ( Figure 6). Introduction of the bulky anion resulted in the destruction of the polymeric chain. The solid-state structure of 9 featured "isolated" 1,2,4-selenadiazole cations which were involved in two chalcogen-π interactions. It should be noted that chalcogen-π interactions are a bonding motif found in biological systems such as proteins [34]. Thus, compound 9 could be an interesting potential model structure for the exploration of chalcogen-π ChB.
Further, we explored the influence of charged (cationic) substituents by the nitrile group on the cyclization with 2-pyridylselenylchloride. p-Cyano-N-methylpyridinium methanesulphonate and o-cyanodiazonium tetrafluoroborate readily reacted with 2-pyridylselenylchloride, which resulted in the formation of adducts 7 and 10 in high yields. Interestingly, the presence of the diazonium group did not affect the cyclization, demonstrating a remarkable functional group tolerance of the coupling with nitriles.
The structural analysis revealed that dicationic 1,2,4-selenadiazoles in 7 formed [Se-Cl] 2 squares (Figure 7), which we observed earlier [22]. It should be noted that the second chloride did not participate in ChB (Figure 7). The adduct 10 featuring diazonium moiety and two BF 4 anions did not form supramolecular dimers or polymers in the solid state ( Figure 8). The Se center was involved in two Se···F ChBs, while the N 2 + group-in three N···F pnictogen bonding interactions ( Figure 8). Inspection of the crystallographic data revealed the presence of various non-trivial non-covalent interactions in the crystal structures of 3-10. To understand the nature and quantify strength of these non-covalent interactions, DFT calculations followed by a topological analysis of the electron-density distribution within the QTAIM approach [35] were carried out at the ωB97X-D3/Sapporo-DZP-2012 level of theory for model supramolecular associates (see Computational Details in the Materials and Methods, and the attached xyzfiles in the Supplementary Materials). The results of the QTAIM analysis are summarized in Table S2 (the Poincaré-Hopf relationship was satisfied in all cases). The contour line diagrams of the Laplacian of electron density distribution ∇ 2 ρ(r), bond paths, selected zero-flux surfaces, visualization of electron localization function (ELF), and reduced density gradient (RDG) analyses for some of these non-covalent interactions are shown in Figure 9 and Figures S1-S6. The QTAIM analysis of model supramolecular associates 3-10 demonstrated the presence of bond critical points (3, -1) for the non-covalent interactions listed in Table S2 and shown in Figure 9 and Figures S1-S6. In all cases, with the exception of Se25···Cl27 and Se25···Cl26 contacts in 6, the low magnitude of the electron density (0.007-0.028 a.u.), positive values of the Laplacian of electron density (0.020-0.066 a.u.), zero or very close to zero energy density (0.000-0.002 a.u.) in these bond critical points (3, -1), and estimated strength for appropriate short contacts (0.9-5.0 kcal/mol) are typical for weak hydrogen bonds [3,4,36] and non-covalent interactions involving halogen [8][9][10][11][12]32,37,38] and chalcogen [22,23,[39][40][41][42] atoms in similar chemical systems. The strongest non-covalent interactions in the studied model supramolecular associates 3-10 were chalcogen bonds: Se1···Cl24 in 3 (4.7 kcal/mol), Se1···Cl30 in 4 (4.4 kcal/mol), Se1···Cl38 in 5 (5.0 kcal/mol), Se1···Cl24 in 6 (3.5 kcal/mol), Se29···Cl56 in 7 (5.0 kcal/mol), Se1···Br15 in 8 (3.5 kcal/mol), Se2···C31 in 9 (1.3 kcal/mol), and Se1···F29 (4.7 kcal/mol) in 10. In cases of Se25···Cl27 and Se25···Cl26 contacts in 6, the significant magnitude of the electron density (0.064 and 0.069 a.u.), clear negative energy density (−0.015 and −0.018), and relatively high estimated strength (14.7 and 16.6 kcal/mol) allow us to consider these closed-shell interactions more like coordination bonds rather than non-covalent interactions [43,44]. Note that the energies of non-covalent interactions can also be potentially estimated on the basis of the experimental electron-density distribution functions [45,46]. The balance between the Lagrangian kinetic energy G(r) and potential energy density V(r) at the bond critical points (3, −1) reveals the nature of these interactions: if the ratio −G(r)/V(r) > 1 is satisfied, then the nature of appropriate interaction is purely non-covalent, whereas in the case of -G(r)/V(r) < 1, some covalent component takes place [47]. On the basis of this criterion, one can state that a covalent contribution in all interactions listed in Table S2, except Se25···Cl27 and Se25···Cl26 contacts in 6, was absent, whereas Se25···Cl27 and Se25···Cl26 contacts in 6 had significant degrees of covalency. The Laplacian of electron density is typically decomposed into the sum of contributions along the three principal axes of maximal variation, giving the three eigenvalues of the Hessian matrix (λ 1 , λ 2 and λ 3 ), and the sign of λ 2 can be utilized to distinguish bonding (attractive, λ 2 < 0) weak interactions from nonbonding ones (repulsive, λ 2 > 0) [48,49]. Thus, all contacts listed in Table S2 are attractive.

General Remarks
All manipulations were carried out in air unless specified otherwise. All reagents used in this study were obtained from commercial sources (Aldrich, TCI-Europe, Strem, ABCR). Commercially available solvents were purified by conventional methods and distilled immediately prior to use. NMR spectra were recorded on a Bruker Avance III (Karlsruhe, Germany). Chemical shifts (δ) are given in ppm, coupling constants (J) are given in Hz. C, H, S, and N elemental analyses were carried out on a Euro EA 3028HT CHNS/O analyzer (Pavia, Italy). 2-Pyridylselenyl chloride, 2-pyridylselenyl bromide, and di-(2-pyridyl)-diselenide were synthesized as reported earlier [50,51].

X-ray Crystal Structure Determination
The single-crystal X-ray diffraction data for 3-10 were collected on three-circle Bruker D8 Venture or Bruker Smart Apex II diffractometers (Centre of Joint Equipment of Kurnakov Institute of General and Inorganic Chemistry, Russian Academy of Sciences) using ϕ and ω scan modes. The data were indexed and integrated using the SAINT program [52]. Absorption corrections based on measurements of equivalent reflections (SADABS) were applied [53]. The structures were determined by direct methods and refined by a full-matrix least squares technique on F2 with anisotropic displacement parameters for non-hydrogen atoms. The hydrogen atoms in all compounds were placed in calculated positions and refined within riding models with fixed isotropic displacement parameters (Uiso(H) = 1.5Ueq(C) for the CH3-groups and 1.2Ueq(C) for the other groups). All calculations were carried out using the SHELXTL program [54] and OLEX2 program package [55]. For details, see

Computational Details
The DFT calculations based on the experimental X-ray geometries of 3-10 were carried out at the ωB97X-D3/Sapporo-DZP-2012 level of theory [56][57][58][59] with the help of the ORCA 4.2.1 program package [60]. The RIJCOSX approximation [61] was utilized. The topological analysis of the electron density distribution, with the help of the QTAIM approach [35], was performed using the Multiwfn program (version 3.7) [62]. The Cartesian atomic coordinates for model supramolecular associates are presented in the attached xyz-files in the Supplementary Materials.

Conclusions
In summary, eight novel cationic 1,2,4-selenadiazoles were structurally characterized. Solid state structures exhibited multiple ChB interactions, which were studied by DFT calculations and topological analysis of the electron-density distribution within the framework of Bader's theory (QTAIM method). Formation of [Se-N] 2 squares, which we observed earlier, did not occur for selenadiazoles studied within this work. Although substituentdependent self-assembly in the solid state was observed, a priori prediction of the packing preference is still challenging due to the presence of multiple competing weak interactions. Overall, we demonstrated several new types of structural organization of novel cationic 1,2,4-selenadiazole building blocks in the solid state, which involved attractive ChB interactions. Further studies into the chemistry of cationic 1,2,4-selenadiazoles and their applications from our laboratory are underway, and will be reported in due course.
Supplementary Materials: The following supporting information can be downloaded: Table S1 with crystal data and structure refinements for 3-10; Table S2 with details of QTAIM analysis [63,64]; Figures S1-S6 with contour line diagrams of the Laplacian of electron density distribution ∇ 2 ρ(r), bond paths, selected zero-flux surfaces, visualization of electron localization function (ELF), and reduced density gradient (RDG) analyses for some of non-covalent interactions in 5, 6, 8, 9, and 10; xyz-files with Cartesian atomic coordinates for model supramolecular associates. Funding: This work was performed under the support of the Russian Science Foundation (award no. 20-73-00094). X-ray crystal structure determination was performed under the support of the RUDN University Strategic Academic Leadership Program.