Relatively few coordination Cu^II complexes with coordinated fluoride ion [1] are described in the literature, compared to the solid-state materials bearing a Cu^II–F bond, which draw attention due to their particular magnetic properties [2,3]. Several qualitatively different approaches have been described for obtaining them. While the most direct method is the use of sparingly soluble CuF₂ as starting material [4,5,6], in most cases, the fluoride ion has come from an anion such as BF₄⁻ or PF₆⁻. The decomposition of BF₄⁻ to yield complexes, described by Reedijk et al. as early as 1974 [7], was later exploited in several groups for the preparation of a variety of coordination complexes with terminal or bridged fluoride ligands, with different transition metal ions [8,9,10,11,12,13,14]. Several authors have described obtaining fluorido complexes “adventitiously” [15] or in purpose [16] when utilizing BF₄⁻ salts as metal sources, initially with Cu [9], and more recently with Cd [17], Zn [17], Cr [18] and Fe [19].

However, very little has been said about the factors promoting the incorporation of a F⁻ ligand from the dissociation of BF₄⁻ or PF₆⁻ anions, and apart from Reedijk’s early report [20] the corresponding mechanism remains virtually unexplored. Some authors claimed that bulky, very basic ligands [21], as well as the use of alcohols as solvent [16] promote dissociation of BF₄⁻ and PF₆⁻. In several instances, complexes bearing a Cu^II–F bond have resulted from exposure to oxygen of Cu^I complexes with BF₄⁻[22] or PF₆⁻ [23,24,25] as counterions. It also seems that presence of sterically crowded and electron rich ligands, pointed out by Straub as essential for the formation of low-coordinate Cu^I species [23], could promote the dissociation of the fluoroanions and/or the stabilization of the coordinated fluoride.

In this context, we would like to contribute with the X-ray study of a couple of closely related dinuclear Cu^II complexes, in which a BF₄⁻ ion is the F⁻ source for the formation of Cu^II–F bonds. The present study is of limited interest from a synthetic point of view, because of reproducibility and separation issues we were unable to address. However, this work clearly shows that X-ray diffraction allows the accurate assignment of F⁻ sites in the coordination sphere, and that this ligand is distinguishable from other potential ligands having almost similar scattering power, OH⁻ and H₂O, providing that each ligand affords a specific hydrogen bond pattern in the resulting crystal structure.

Both complexes are based on the octahydro[4,5]diimidazo[1,6]diazecine cyclic framework, abbreviated dimeim hereafter. This ligand has proven excellent for yielding di-copper complexes with a variety of counterions [26,27,28]. By reacting 2:1 equivalents of Cu(BF₄)₂·6H₂O and dimeim in methanol, we obtained the expected complex, [Cu₂(H₂O)₄(dimeim)](BF₄)₄ (1, see Figure 1), a deep-blue compound initially intended for a study about the influence of the counterion on the magnetic properties of the cation. However, the very first crop of crystalline material contained light-blue small crystals, for which analytical data and X-ray crystallography gave a different formula, [Cu₂(H₂O)₂(F)₂(dimeim)](BF₄)₂·2H₂O (2, see Figure 1). The relative yields for 1 and 2 varied from synthesis attempts, and a satisfactory separation could not be achieved. However, on the basis of their colors, suitable single crystals for 1 and 2 were picked from the raw material, and their crystal structures determined.

Numerous studies [26,27,28,29,30] carried out during the last ten years showed that the central 10-membered diazecine heterocycle of these compounds is systematically stabilized as a C_2h skewed boat-chair-boat conformer. Such a rigid, highly symmetrical conformation strongly favors the formation of centrosymmetric binuclear coordination compounds, using diazecine and imidazole N atoms as coordination sites [26,27,28,29,30]. This rule is again observed for 1 and 2, two binuclear complexes with five-coordinate Cu^II centers.

Synthesis of the complexes. The ligand dimeim was prepared as previously described [26]. To 50 mL of a methanolic solution containing two mmol of Cu(BF₄)₂·6H₂O, one mmol of solid dimeim was added in several portions, which dissolved upon coordination. Small light blue crystals of 2 were collected after one day, calculated analysis for [Cu₂(H₂O)₂(F)₂(dimeim)](BF₄)₂·2H₂O: C, 30.05; H, 5.55; N, 14.02%. Found: C, 30.46; H, 5.85; N, 14.27%. Deep blue crystals of 1 were collected after a few days. Calculated analysis for [Cu₂(H₂O)₄(dimeim)](BF₄)₄: C, 25.69; H, 4.74; N, 11.99%. Found: C, 25.38; H, 4.74; N, 11.93%.

X-ray diffraction. Data for complexes 1 and 2 were collected at room temperature on a Bruker P4 diffractometer [34] using the Mo-Kα radiation (Table 2). Raw data were corrected for absorption effects [35] and the structures refined with the SHELX programs [36]. Although sample for 2 was a rather long crystal not uniformly irradiated during data collection, a reduction of the sample size was not intended, because very few single crystals were available. However, we assume that absorption correction for 2 is still reliable (R_int = 0.028). Accurate determination of sites occupied by H atoms was critical for water molecules, in order to discriminate F⁻ and water molecules (and possibly OH⁻ groups, assuming that hydroxyl groups could be present in the reaction media). All water H atoms in both complexes were thus assigned from difference maps, and refined with free coordinates, although soft restraints were applied to O–H bond lengths and H···H separations, 0.85(1) and 1.34(1) Å, respectively. Sensible orientations for water molecules are obtained in the final models, since all O–H groups are engaged in efficient hydrogen bonds. On the other hand, in the case of 2, final difference map is featureless in the vicinity of the coordinated fluoride site, corroborating that Cu–F bonds have been formed, rather than Cu–OH or Cu–OH₂ bonds. CIF data, including structure factors, are deposited as supplementary material.

In conclusion, substitution of basal water by F⁻ ligand in complex 1, and concomitant release of water to form the hydrate 2, allows the stabilization of a crystal structure featuring strong intermolecular hydrogen bonds between coordinated and free water molecules and the fluoride ligands, which also act as acceptors for hydrogen bonding.