Different coordination patterns for two related unsymmetrical compartmental ligands: crystal structures and IR analysis of [Cu(C21H21O2N3)(OH2)(ClO4)]ClO4·2H2O and [Zn2(C22H21O3N2)(C22H20O3N2)]ClO4

The crystal structures of the new complexes [Cu(HL1)(OH2)(ClO4)]ClO4·2H2O (1) and [Zn2(HL2)(L2)]ClO4 (2), derived from two related, phenol-based compartmental ligands, are described. Compound 2 constitutes the first report of a complex obtained from H2L2. The metal compounds are structurally different; 2 is a dimer in which all the heteroatoms of the ligand take part in coordination, while 1 is mononuclear containing a pair of cis-oriented ligands that complete an “open” coordination sphere, in which the aldehyde group of HL1 is not involved. The protonation status of the central phenol groups of HL1 and H2L2 are also dissimilar between the complexes. Infrared vibrational analyses of both complexes, as well as their respective ligands, were performed to connect the observed spectral features with the structural properties of the solids. While some distinctive bands shifted upon complexation, it was not possible to confirm involvement of the aromatic aldehyde group in coordination by this technique. 1H NMR experiments involving 2 suggest that its particular protonation status is maintained upon dissolution in d6-DMSO. Graphical Abstract


Introduction
The study of model dinuclear metal complexes has become an important tool for gaining insight into the function of biologically occurring bimetallic cores. Ligands termed compartmental are defined as the class of polydentate, chelating ligands able to bind simultaneously to two metal ions in the presence of two adjacent coordination sites [1]. The recent recognition of the asymmetric nature of a number of polymetallic biosites has greatly increased the interest in unsymmetrical bioinspired ligands [2,3]. Furthermore, polynuclear complexes derived from such ligands can display catalytic activities [4,5]. Thus, the design of binucleating ligands capable of generating asymmetric dinuclear complexes is of interest. The compartmental ligands presently studied are HL1, 3-{[bis(pyridin-2-ylmethyl)amino]-methyl}-2-hydroxy-5-methylbenzaldehyde, and the associated compound H 2 L2, 2-hydroxy-3-{[(2-hydroxybenzyl)(pyridin-2-ylmethyl)amino]-methyl}-5-methylbenzaldehyde (chart 1), which are classified as phenol-based compartmental ligands, since the donor that acts as an endogenous bridge between the metal centers in the dinuclear complexes is a phenolic oxygen. Moreover, these ligands have different coordination possibilities, and the phenol group (s) can either remain protonated or be deprotonated in their metal complexes. Both HL1 and H 2 L2 contain a central para-cresol ring, with a tridentate coordinating side arm and an aldehyde carbonyl occupying ortho-positions relative to the hydroxyl substituent of this central unit. The coordinating side arm is constituted by a tertiary aminomethyl nitrogen with a pair of 2-methylpyridine substituents (HL1), or by a tertiary aminomethyl nitrogen exhibiting a 2-methylpyridine and a 2-methylphenol as substituents (H 2 L2).
The literature reports some complexes derived from these compartmental ligands. de Oliveira et al. [6], for example, reported the X-ray structure of two related mononuclear copper(II) complexes containing HL1 and L1 -, that, in aqueous solutions, are able to hydrolyze the peptide bonds of bovine serum albumin. Compounds involving other transition metals have also been synthesized, such as the dinuclear cobalt(II) and manganese(II) complexes [Co 2 (L1) 2 ] 2+ and [Mn 2 (L1) 2 ] 2+ [7] and a mononuclear dioxovanadium(V) complex, [VO 2 (L1)] [8]. Regarding H 2 L2, to the best of our knowledge, there are no reports in literature concerning the synthesis of complexes from this ligand. However, this compound was used as a precursor in some studies published by Prof. Ademir Neves' group, at the Federal University of Santa Catarina, Brazil, aimed at preparing higher complexity binucleating ligands [9,10].
In the present study, the synthesis, crystal structures, and IR vibrational spectra of a new mononuclear copper(II) and a dinuclear zinc(II) complex derived from the phenol-based compartmental ligands HL1 and H 2 L2, respectively, are reported.

Materials and methods
Starting materials were commercially available and were used as purchased. Ligands HL1 [10] and H 2 L2 [11] were prepared according to experimental procedures described previously. EA analyses were performed with a Thermo CHNS-O Analyzer Flash EA 1112 Series. Infrared spectra of the compounds were recorded on a Perkin Elmer FT-IR 2000 apparatus. Samples were measured from 4000 to 450 cm −1 as KBr pellets. . This complex was obtained by adding a methanolic solution of Cu(ClO 4 ) 2 ·6H 2 O (0.5 mM, 0.18 g) dropwise to HL1 (0.5 mM, 0.17 g), dissolved in methanol. The addition of the copper salt immediately resulted in a clear, intense blue color, with no signs of precipitation. The mixture was stirred continuously at 50°C for 20 min and then left to stand at room temperature for slow evaporation of the solvent. One week later, greenish-blue crystals (0.15 g, 45% yield) suitable for X-ray diffraction (XRD) were observed in the beaker. These were filtered off, washed with cold methanol and diethyl ether, and dried at room temperature (ca. 23°C). Elem. Anal. Calcd for CuC 21 (2). This complex was synthesized by mixing a methanolic solution of H 2 L2 (0.5 mM, 0.18 g) with an excess of Zn(ClO 4 ) 2 ·6H 2 O (1.0 mM, 0.37 g). The reaction was stirred at 50°C for 20 min. A yellow powder was then observed in the reaction flask, which was filtered, washed with cold methanol, dried, and analyzed. However, it was shown to be a by-product and was discarded. The mother liquor was allowed to evaporate at room temperature. After one week, yellow crystals (0.08 g, 34% yield) suitable for X-ray structure determination were obtained. Elem. Anal. Calcd for Zn 2  Caution! Perchlorate salts of metal complexes containing organic ligands are potentially explosive. Only small amounts of material should be prepared, and they should be handled with care.

XRD and powder XRD
Single-crystal XRD data were collected on an Oxford GEMINI A-ultra diffractometer. For 1, the measurements were made at 120 K using Cu-K α (λ = 1.5418 Å) radiation, whereas for 2, data were obtained at 150 K with Mo-K α (λ = 0.71073 Å) radiation. Data collection, data reduction, and cell refinement were performed with CrysAlis RED (Oxford Diffraction Ltd, version 1.171.32.38) [12]. Structures were solved and refined using SHELXL-97 [13]. A multiscan absorption correction was applied [14]. All the non-hydrogen atoms were refined anisotropically. Hydrogens bound to carbon were placed at calculated positions and hydrogens bound to oxygen were located from the Fourier difference maps. The coordinates of the phenolic hydrogen in 2 were refined freely. All hydrogens were refined using a riding model. In 2, the perchlorate was disordered and could be refined over two sets of positions in a 0.85/0.15 ratio. Crystal, data collection, and refinement parameters for 1 and 2 are presented in table 1. Crystallographic figures were drawn employing the ORTEP-3 for Windows [15] and Mercury [16] programs.
The phase purity of both complexes was confirmed by powder XRD (PXRD). PXRD patterns were collected on a Bruker D8 Advance diffractometer, employing Cu-K α radiation at room temperature. Each sample was scanned between 5°and 40°in 2θ, with 0.02°as step size. A comparison between the experimental PXRD patterns of the isolated materials and those simulated from the single-crystal data by means of the Mercury 3.1 program are in figure S1 (see online supplemental material at http://dx.doi.org/10.1080/00958972.2014. 958080).   [18]. Conversely, deprotonated phenol groups display much shorter Cu-O apical coordination distances as, for instance, reported by Rajendran et al. [19] for chloro{2-[bis(2-pyridylmethyl)aminomethyl]-4-nitrophenolato} copper(II), 2.268(25) Å. The sixth, and also axial, coordination position in 1 belongs to the perchlorate O3 at a distance of 2.511(3) Å from the metal center. As seen in figure 1, N 3 O 3 coordination generates only the mer-isomer for the O and N atoms present in the coordination sphere. Among the HL1 heteroatoms, only the aldehyde O1 is not involved in coordination. A second perchlorate, acting as a counter ion, assures the electrical neutrality. In addition, two crystallization water molecules (not shown in figure 1) per [Cu(HL1)(OH 2 ) (ClO 4 )] + cation were observed.

NMR experiments
As pointed out in the introduction, a very similar six-coordinate mononuclear copper(II) complex derived from HL1, [Cu(HL1)Cl 2 ]·2H 2 O, has already been reported [6]. The main difference between this compound and 1 is the presence, in the former, of a pair of coordinated chlorides instead of the water molecule and perchlorate as observed in 1. As well, Koval and co-workers have prepared the bromide derivative [Cu(HL1)Br 2 ]·½H 2 O, isostructural to [Cu(HL1)Cl 2 ]·2H 2 O [20]. Comparison of the bond lengths between 1 and the other copper complexes of HL1 shows that the bond distances in 1 between Cu and the pyridyl N atoms, N2 and N3, are, on average, 0.041 and 0.036 Å shorter than those in [Cu(HL1) Br 2 ]·½H 2 O and [Cu(HL1)Cl 2 ]·2H 2 O, respectively. It can also be verified that the N(pyridyl) is a better Lewis base than the tertiary amine (N1), since the Cu-N1 bond distance is longer than the other two distances (table 2). In a previous study, we observed the same behavior  (10) for a zinc(II) complex derived from the mycobactericidal drug isoniazid, with the Zn-N (pyridyl) bond distance shorter than the Zn-N(amine) one [21]. In 1, the complex ions are packed, leading to a supramolecular structure stabilized by H-bonds and π-stacking interactions, represented in figures 2 and 3, respectively. The H-bond system forms rings that were ranked by degree (size) and complexity:  classified as moderate, i.e. mostly electrostatic, and associated with energies in the range of 4-15 kcal M −1 , as can be seen by the donor-acceptor distances in table 3 [24]. An intramolecular H-bond interaction between the O1 (aldehyde) and O2 (phenol) is also present [2.599(4) Å]. The π-π interactions involve both pyridine rings and connect the complex cations, generating linear chains (figure 3). The calculated centroid-centroid distance is equal to 3.750(7) Å. Table 3. H-bonding parameters for 1.   Each Zn 2+ cation is coordinated by O3B/O3A (phenol), N1B/N1A (amine), and N2B/ N2A (pyridyl) from the side arms of HL2 − B (bound to Zn1) or L2 2− ligand A (bound to Zn2). Both tridentate arms act as meridional coordination moieties. The fourth coordination sites are occupied by the aldehyde O1A (Zn1) and O1B (Zn2). The coordination pattern is completed by the two phenoxo bridges mentioned above, O2A and O2B. As a result, distorted octahedral geometries are observed around both Zn 2+ cations (table 4). The bridging oxygen of one ligand (O2A or O2B) and the aldehyde oxygen of the other (O1B or O1A, respectively) are in a trans configuration; O2A-Zn2-O1B is 169.28 (8)  Complex 2 exhibits an interesting asymmetry, which is not obvious or even expected considering its dimeric nature. Such a feature is caused by the incomplete deprotonation of HL2 -, in which only one phenol group (O2B) loses its proton. Thus, the only difference in the coordination spheres of the metal centers is related to the protonated (HL2 − ) or deprotonated (L2 2− ) states of O3B and O3A from the terminally coordinated phenol groups. As expected, the distances between the Zn centers and these phenolic oxygens are slightly different; the Zn1-O3B distance (2.198(2) Å) is longer than the Zn2-O3A one (2.092(2) Å). Since both ligands in 2 are anionic, one of them completely deprotonated (L2 2− ) and the other one partially deprotonated (HL2 − ), and two Zn 2+ cations are present in the asymmetric unit (corresponding to the molecular formula), a counter-ion, in this case, a perchlorate anion (not shown in figure 4), is necessary to ensure electrical neutrality. This anion is disordered over two positions. The ligands are not related by symmetry, due to the charge difference between them, which causes a small divergence in some torsion angles. An intramolecular H-bond interaction was observed between O3B-H (phenol) and O3A (phenolate), with a donor-acceptor distance of 2.469(3) Å (table 5), characteristic of a strong, mainly covalent, H-bond [24]. There is no crystallization solvent and neither intermolecular H-bonds nor π-stacking interactions were observed in the structure.
The particular protonation status presented by 2, involving a terminal phenol coordinated to zinc in its protonated form, has no parallels in the literature concerning dimeric zinc(II) complexes of phenol-based compartmental ligands. Coelho and co-workers [28] reported the structure of the cationic dimer di-μ-chlorido-bis{[2-({[2-(2-pyridyl)-ethyl](2-pyridylmethyl)amino}methyl)-phenol]zinc(II)}, resulting from a tetradentate ligand containing a There are many structural characteristics that make 2 a curious example of a zinc(II) dimeric complex derived from a phenol-based compartmental ligand. Its asymmetric protonation, involving only one phenol group, prevents the ligands L2 2− and HL2 − to be related by symmetry, which results in a non-centrosymmetric species. Six coordination also is not a very common feature in homobimetallic zinc compounds, as well as the presence of a phenol coordinating in its protonated form. Moreover, the inter-metallic distance found in 2  [26]) and 78.40(6)°and 78.67(12)°(Chakraborty [27]).
Finally, this work constitutes, to the best of our knowledge, the first description of any H 2 L2 complex in the literature.

IR vibrational spectra analysis
The infrared spectra of 1 and 2 are very similar, with the main difference related to the perchlorate region (figures 5 and 6, respectively, and table 6). Both complexes show broad bands centered at 3450 cm −1 , which are due to the O-H stretches of protonated phenol (1 and 2) and the coordinated/crystallization water molecules (1). Due to these contributions, the O-H stretching band is stronger in 1. Other bands related to the phenols are the in-plane δ(C-OH) and ν(C-O) vibrations. In free HL1, the in-plane δ(C-OH) mode provides a single absorption at 1379 cm −1 , whereas the H 2 L2 spectrum shows two well-defined frequencies, at 1382 and 1352 cm −1 , for this vibration. We attribute this to the presence of two non-equivalent phenol groups in H 2 L2. Upon complexation, the IR spectra of 1 and 2 show only one band related to this mode, according to their protonation status, which is shifted to lower wavenumber. The ν(C-O) phenol band is also shifted to lower wavenumbers upon coordination, but to a lesser extent (ca. 15-20 cm −1 ).  The band associated to the aldehyde ν(C=O) stretch, present at 1680 and 1682 cm −1 in the HL1 and H 2 L2 spectra, respectively, is shifted to 1650 (1) and 1642 (2) cm −1 . This change is caused by different reasons. In 1, it is related to the moderate H-bond interaction between this group and the phenol hydrogen. In contrast, in 2, this slightly more pronounced shift is due to the coordination of the aldehyde to Zn 2+ cations. The fact that these shifts are very close (30-40 cm -1 ) leads us to conclude that this ν(C=O) absorption could not be used as a diagnostic band for the inference of the involvement of the aldehyde in coordination. As stated in the introduction, Koval and co-workers prepared dinuclear cobalt (II) and manganese(II) complexes of L1 - [7]. In both compounds, the aldehyde oxygens are coordinated to the metals, and the authors assigned the band present at 1640 cm −1 to ν(C=O), with which our proposition agrees. On the other hand, the mononuclear copper(II) complex of HL1 synthesized by de Oliveira et al. displays an intense band related to the stretching of its uncoordinated aldehyde group at 1650 cm −1 [6]. As in our case, this shift is probably also due to H-bond formation. The dioxovanadium(V) compound of L1prepared by Silva et al. in which the phenol is uncoordinated and is not involved in intramolecular H bonding, shows its ν(C=O) mode at 1660 cm −1 [8]. Concerning the phenol and pyridine aromatic rings, the ν(C=C) and ν(C=N) modes of both HL1 and H 2 L2 ligands and 1 and 2 give rise to a series of four bands in the range of 1615-1450 cm −1 .
Perchlorate anions display characteristic ν 3 and ν 4 infrared-active modes. As discussed in the crystallographic description of the structures, the number of perchlorate groups present in 1 and 2 is not the same. The perchlorate region of the infrared spectra, as expected, reflects this difference. Complex 2 has a clean pattern, with bands at 1092 (ν 3 ) and 939 (ν 4 ) cm −1 , typical of a free, uncoordinated tetrahedral ClO 4 - [30]. On the other hand, the ν 3 mode of 1 shows a very complex multi-component envelope, with main frequencies at 1116 and 1090 cm −1 , and a series of shoulders. The presence of two ClO 4 − anions and the fact that one of them is coordinated to copper contribute to the splitting of the ν 3 mode. Attempts to relate the energy of the absorptions to the coordination status of ClO 4 − were fruitless. In previous reports published by Rey et al. [31,32], two dinuclear copper(II) complexes containing perchlorate, i.e. [Cu 2 (μ-OH)(L asym )(ClO 4 )]ClO 4 (L asym = 2-[N,N-di (pyridin-2-ylmethyl)aminomethyl]-4-methyl-6-[(6-methyl- [1,4]diazepan-6-yl)imino-methyl] phenolate) [31], and [Cu 2 (μ-OH)(L sym )](ClO 4 ) 2 ·H 2 O (L sym = 4-methyl-2,6-bis[(6-methyl- [1,4]diazepan-6-yl)imino-methyl]phenolate) [32] were synthesized and fully characterized by XRD. Nevertheless, in those studies, the reported infrared bands at 1144, 1110, and 1091 cm −1 , and 1147, 1117, and 1082 cm −1 , respectively, could not be connected to the structural roles of the ClO 4 − anions. In the IR spectrum of 1, the ν 4 mode was observed as a single band at 930 cm −1 . However, as Lewis and co-workers showed several years ago, the presence of a pair of bands between 700 and 600 cm −1 , assigned to the A1 and E modes of a monodentate ClO 4 − anion (C 3v symmetry), provides a clear indication of the involvement of this anion in coordination [33]. In the IR spectrum of 1, an asymmetric band of medium intensity with maximum at 625 cm −1 is observed. The ligand HL1 does not show any bands in this region.

NMR studies of complex 2 in DMSO solution
A complete assignment of the hydrogens of free H 2 L2 was carried out with a 1 H-1 H correlation map from a COSY NMR experiment (figure S2, Supplementary material). Figure 7 (bottom) shows the 1 H NMR spectrum of H 2 L2 in the region ranging from 6.5 to 10.5 ppm, with the respective assignments. The aldehyde (-HC=O) hydrogen, as expected, appears as a sharp singlet at the largest ppm value, 10.22 ppm. Among the aromatic hydrogens, those belonging to the pyridine ring are at larger ppm values, especially H19 (8.56 ppm) and H21 (7.79 ppm). Between 7.40 and 7.31 ppm, there is an intricate multiplet corresponding to approximately four hydrogens. This was attributed to the remaining pyridine hydrogens, H22 and H20, and to H3 and H5, the aromatic hydrogens alpha to the methyl group. In general, the carbon-bound hydrogens from the phenolic rings are at lower chemical shifts; H16, H14, H13, and H15, which belong to the coordinating side arm, are all observed below 7.20 ppm. The methylene protons could not be seen, since the water band associated with DMSO overlapped. Finally, the methyl hydrogens give a singlet at 2.21 ppm (not shown in figure 7). The 1 H NMR spectrum of 2 is also presented in figure 7 (top) for comparison. The presence of the Zn 2+ cations causes a significant broadening in some of the signals, which could be attributed to the site-exchange processes involving coordination to the metal center [34] and the asymmetry of the HL2 − and L2 − ligands. Because of this, a detailed study on the 1 H NMR spectrum of 2 could not be performed. Nevertheless, some differences when compared to the 1 H NMR spectrum of the ligand H 2 L2 are easily noticeable. The aldehyde hydrogens appear at even larger ppm values upon coordination, at 10.49 ppm. In general, the aromatic protons are at lower ppm values in the spectrum of 2, but this trend is not valid for all the signals. The presence of a singlet, corresponding to one hydrogen, at 10.19 ppm is assigned to the phenolic hydrogen H3B. This assignment was confirmed by the addition of some D 2 O drops to the sample. As expected, the signal disappeared indicating H/D isotopic exchange.
Not many studies involving zinc(II) complexes with ligands containing protonated phenol groups are available in the literature, and even less so with regard to 1 H NMR spectroscopy experiments. Among these, to the best of our knowledge, there are no examples in which the protonated phenol group is involved in coordination. Thus, there is no literature value available for comparison of the chemical shifts. Seena and Prathapachandra Kurup [35] studied a series of salicylaldehyde N(4)-phenylthiosemicarbazone (H 2 L) coordination complexes, in which one of them, [Zn(HL) 2 ]·EtOH, retained a protonated phenol group after complexation. The authors assigned the signal at 11.38 ppm (in CDCl 3 ) to this uncoordinated phenolic proton. Khalil and co-workers prepared a zinc complex of a salicylaldehyde thiosemicarbazone with the formula [Zn(Tsc)(HTsc)]·H 2 O [36], with protonated and deprotonated phenol groups, although the former group was not coordinated to zinc. The signal at 10.9 ppm (in d 6 -DMSO) was attributed to the phenolic hydrogen.

Conclusion
Although similar experimental conditions were employed in the synthesis of 1 and 2, the resulting metal compounds are quite different. While both the copper and zinc sites are octahedral, 2 is a dimer in which all the heteroatoms of the ligand take part in coordination. On the other hand, 1 is a mononuclear complex in which the aldehyde group of HL1 is not involved in coordination, containing a pair of cis-oriented (solvent-derived and perchlorate) ligands that complete an "open" coordination sphere. The protonation status of the central phenol group of HL1 and H 2 L2 is also dissimilar between the complexes. Where this group is protonated in 1, both H 2 L2 molecules lose their para-cresol phenol protons during the reaction to form 2. The only phenol proton remaining in 2 belongs to the side arm of one of the H 2 L2-derived ligands, and this proton is shared between the two terminally coordinated phenol oxygens. 1 H NMR experiments indicated that this particular protonation status is maintained upon dissolution in d 6 -DMSO. While IR spectroscopy showed some distinctive band shifts upon complexation, the technique did not exhibit good performance for determining whether the aldehyde carbonyl is coordinated. There are many structural characteristics that make 2 a curious example of a homodinuclear, dimeric zinc(II) complex derived from a phenol-based compartmental ligand. Its asymmetric protonation prevents HL − and L − to be related by symmetry, which results in a non-centrosymmetric species. Also, the presence of a phenol coordinating in its protonated form is definitively not a common feature in homobimetallic zinc compounds. Also, the intermetallic distance in 2 is one of the shortest Zn⋯Zn distances reported to date for compounds of this type. To the best of our knowledge, 2 constitutes the first report of a complex with H 2 L2 in the literature.