Copper(ii) self-assembled clusters of bis((pyridin-2-yl)-1,2,4-triazol-3-yl)alkanes. Unusual rearrangement of ligands under reaction conditions.

The reaction of two structurally related bridging ligands bis[5-(2-pyridyl)-1,2,4-triazole-3-yl]methane (H2L1) and bis[5-(2-pyridyl)-1,2,4-triazole-3-yl]ethane (H2L2) with copper(ii) salts resulted in a surprising wide variety of complex structures [Cu2(H2L1)Cl2]Cl2·4CH3OH (1), [Cu4(L1)4]·4H2O (2), [Cu(H2L2)(ClO4)2] (3) and [Cu3(OH)Na2(L')6](ClO4)·5H2O·C3H6O (4), where HL' is 3,5-bis-(pyridin-2-yl)-1,2,4-triazole, which were structurally characterized by the X-ray diffraction method. Complexes 1 and 2 were prepared on the H2L1 basis and have binuclear and tetranuclear structures, respectively, demonstrating strong impact of the type of counter anion on the coordination mode of the ligand. In contrast, the reaction between Cu(ClO4)2 6H2O and H2L2 led to the preparation of mononuclear complex 3. The reaction of H2L2 with Cu(ClO4)2 under alkaline conditions led to oxidative rearrangement of the ligand and the homoleptic pentanuclear complex 4 with anionic ligand L' was prepared. Magnetic properties were studied for compounds 1, 2 and 4 and for all of them the antiferromagnetic interactions between the Cu atoms were confirmed and analyzed by the spin Hamiltonian formalism. Furthermore, the occurrence of the antisymmetric exchange was confirmed in 4. The magnetic data analysis was supported by the X-band EPR measurements performed for complexes 1, 2 and 4.


Introduction
Current significant interest within supramolecular chemistry involves the design and synthesis of new ligand molecules to control the outcome of metal-assisted self-assembly processes. 1 For this purpose, flexible ditopic ligands have attracted recent attention. Fine tuning of these ligand systems has resulted in the isolation of many structural types including grids, boxes, cylinders, various types of molecular polyhedra and helicates. 2 Significant progress has been achieved in the investigation of polynuclear complexes based on relatively simple bis( pyrazolyl-pyridine) bridging ligands and transition metal dications. 3 M. Ward and coworkers have reported an extensive series of high-nuclearity cages based on self-assembly of the abovementioned ligands including an impressive number of polyhedral shapes of polynuclear complexes, e.g. M 4 L 6 tetrahedra, M 6 L 9 trigonal prisms, M 8 L 12 cubes and 'cuneane', M 12 L 18 truncated tetrahedra, and M 16 L 24 tetra-capped truncated tetrahedra. 4 Surprisingly, despite the attractive features of the nitrogenrich triazole unit (e.g., multiple interaction modes, limited steric hindrance, moderate ligand field), the self-assembled structures have been much less studied for compounds with pyridyl-1,2,4-triazole ligands. It is noteworthy that the bis[5-(2pyridyl)-1,2,4-triazole-3-yl]alkanes are excellent multidentate flexible ligands to construct supramolecular coordination complexes with various structures. They exhibit a strong chelate effect arising from the presence of the adjacent heterocyclic rings and furthermore, they can adopt various conformations due to large rotational flexibility of four aromatic rings, which may contribute to the preparation of complexes with interesting structures. Recent studies demonstrated that use of flexible alkyl spacers in linking pyridyltriazolyl chelating arms leads to the formation of novel polynuclear clusters with an exciting topology and structure. 5 In this work, we present four novel Cu(II) complexes based on recently described 5a bis(( pyridin-2-yl)-1,2,4-triazol-3-yl) methane (H 2 L1) and 1,2-bis(( pyridin-2-yl)-1,2,4-triazol-3-yl) ethane (H 2 L2) ligands. Both ligands have two interrelated characteristic features, a flexible backbone and the possibility to fulfill variable coordination modes, so that, depending on the coordination requirements. Varying the reaction conditions (counter-ions and pH value) allowed us to successfully isolate new complexes for the previously described ligands.

Structure of complexes
In our previous studies, 5 the products of the reaction between H 2 L1 and copper perchlorate were investigated. It was shown that by varying the molar ratio between the reagents we were able to prepare the tetranuclear complexes with different degrees of ligand protonation: [Cu 4 (H 2 L1) 4 ] 8+ and [Cu 4 (HL1) 4 ] 4+ , respectively. In all cases, the perchlorate anions are non-coordinating and occupy the cavities of the crystalline lattice. In this paper, we investigated the interaction of a bis(5-( pyridin-2-yl)-1,2,4-triazol-3-yl)methane (H 2 L1) with two different copper salts containing potentially coordinating anions -CuCl 2 ·2H 2 O and Cu(CH 3 COO) 2 ·H 2 O.
The complex 1 was formed upon reaction between H 2 L1 and CuCl 2 ·2H 2 O in a 1 : 1 molar ratio. In contrast to the previous studies the H 2 L1 ligand now acts as a bis-bidentate bridging ligand in the dinuclear double helicate [Cu 2 (H 2 L1)Cl 2 ]Cl 2 , in which two ligands are wrapped helically around a pair of metal ions. The crystal structure of 1 is shown in Fig. 1.
The Cu⋯Cu separation is 6.083(2) Å and this reflects the length and flexibility of the methylene spacer group of the H 2 L1 ligand. The Cu(II) ions are five coordinated by two bidentate N-donor ligand fragments and one chlorido ligand. According to Addison classification, the coordination polyhedra of the Cu(II) atoms adopt distorted square pyramidal geometry (SPY, the parameter τ is 0.36). 6 The heavy distortion from the ideal SPY coordination geometry arises from the very non-equivalent Cu-N bond lengths: d(Cu-N Py ) = 1.989 and 2.091 Å, d(Cu-N Tz ) = 1.970 and 2.187 Å (N Py stands for pyridine nitrogen atoms and N Tz stands for triazole nitrogen atoms). Both near-planar pyridyltriazolyl fragments are twisted relative to each other by the angle of 69.88(4)°. The angle C(5)-C(1)-C(6), which represents the flexure of the methylene 'belt' within the dimer, was found to be 112.64(3)°on each strand, suggesting that minimal distortion of the ligand is required in order to conform to the helical topology for similar ligands. 7 Replacement of the chloride ligand by acetate-anions led to the preparation of tetranuclear complex 2 in which the H 2 L1 ligand is fully deprotonated (L1 2− ) in contrast to previous studies. 5,8 The molecular structure of 2 is shown in Fig. 2   N2-triazolyl units. All the copper atoms are pentacoordinated with the N 5 donor set. Each copper(II) atom displays slightly distorted square-pyramidal geometry with the nitrogen atom from the pyridine ring in the axial position (τ = 0.13). Four copper(II) centres are arranged into an unusual Cu 4 N 8 trigonal pyramidal core (Fig. S1). The [2 × 2] array of Cu(II) ions has the Cu⋯Cu separations of 4.031-4.471 Å, which are slightly longer than in the case of the complexes involving protonated forms of the ligand (3.93-3.98 Å for [Cu 4 (H 2 L1)] 8+ and 3.97-4.06 Å for [Cu 4 (HL1)] 4+ ) 5 and these values are typical of the μ-triazolyl coordination mode. 8 The ligand strands are oriented in a "head-to-tail" arrangement at the Cu(II) sites; the "head" and "tail" terms refer to the tridentate, and bidentate donor pockets, respectively. The L1 2− ligand has a substantially planar structure with the average deviation of the atoms from their best-fit mean plane of 0.13 Å.
The reaction of H 2 L2 with Cu(ClO 4 ) 2 ·6H 2 O in MeOH afforded a blue solution from which the [Cu(H 2 L2)(ClO 4 ) 2 ] (3) compound was obtained. X-ray crystallography (Fig. 3) showed that in 3 the Cu ion has the CuN 4 O 2 coordination environment with the H 2 L2 ligand acting as an equatorial tetradentate chelator (d(Cu-N Pyr ) = 2.0631(17) Å, d(Cu-N Tz ) = 1.9690(17) Å) and with two perchlorate ligands in axial positions. The coordination geometry is an axially elongated octahedron due to the Jahn-Teller effect (d(Cu-O) = 2.521(2) Å). The two bidentate pyridyltriazolyl arms are not quite coplanar, which can be described by the angle 9.77°between the two mean planes involving each bidentate fragment. It appears that the elongation of the bridging aliphatic part between the two coordinating fragments makes the tetradentate coordination mode more favorable, when compared with the related ones. 5 A very interesting result was obtained by studying the reaction of H 2 L2 with copper salts in an alkaline medium in an attempt to determine the coordination modes of the deproto-nated form of 1,2-bis(5-( pyridin-2-yl)-1,2,4-triazol-3-yl)ethane. In order to achieve this goal we tried various ratios of reagents and different types of copper salts (CuCl 2 ·2H 2 O, Cu(CH 3 COO) 2 ·H 2 O and Cu(ClO 4 ) 2 ·6H 2 O) and bases (NaOH, MeONa, Et 3 N). In all cases, unidentifiable mixtures of compounds were obtained. However, prolonged reflux (6 h) of H 2 L2 in combination with two equivalents of Cu(ClO 4 ) 2 ·6H 2 O and in the presence of an excess of NaOH allowed us to isolate the crystals of the unusual complex, which, according to elemental analysis and electrospray ionization (ESI) data, might have an approximate formula [Cu 3 (OH)Na 2 (L′) 6 ](ClO 4 ) where HL′ is 3,5-bis-( pyridin-2-yl)-1,2,4-triazole. The ESI spectra of 4 contain several predominant peaks corresponding to the molecular ion and its decay products. A characteristic signal for a single charged ion at m/z 1586.9 is assigned to the species [Cu 3 (OH)Na 2 (L′) 6 ] + , based on a reasonable agreement of the isotopic distribution pattern and this structural motif was further confirmed by the single-crystal XRD measurements. The presence of water and acetone molecules in the complex has been confirmed by thermogravimetric analysis. The TGA plot recorded for 4 shows two consecutive and distinct weight losses of 2.8% and 5.2% at about 55-60°C and at 95-110°C, respectively, and associated with the presence of one acetone molecule (calculated m m −1 loss = 2.94%) and five water molecules (calculated m m −1 loss = 4.91%) in the crystal packing of 4 ( Fig. S2 †).
Complex 4 crystallizes in the triclinic space group P1 as a solvate with 11 water molecules (some solvate molecules that were disordered were removed by the SQUEEZE procedure to achieve reasonable refinements. 10 ). Differences in the solvate composition for a polycrystalline sample and a single crystal are associated with partial desolvation of the sample upon drying. The crystalline lattice of 4 consists of trinuclear [Cu 3 (μ 3 -OH)Na 2 (L′) 6 ] + complex molecules and perchlorate counter anions. The structure of 4 is shown in Fig. 4.
Three Cu and two Na cations define a trigonal-bipyramidal polyhedron, in which the Cu⋯Na, Cu⋯Cu, and Na⋯Na average distances are 4.564, 3.510, and 8.177 Å, respectively.
Six 3,5-bis-( pyridin-2-yl)-1,2,4-triazolate mono-deprotonated ligands wrap around the complex pentaheteronuclear cluster. Two homochiral [Na(L′) 3 ] units are placed in the apical positions of the molecule and they simultaneously coordinate the {(μ 3 -OH)Cu 3 } 5+ central core. In this manner, the ligands coordinate to the Cu 3 Na 2 unit with helical arrangement, while the whole complex cluster has no intrinsic D 3 symmetry. Note that enantiomers (ΔΔ, ΛΛ) of the cation unit induced by the helical coordination arrangement of the homochiral unit coexist as a racemic form in the crystal.
Three equatorial Cu(II) atoms possess the distorted N 4 O trigonal-pyramidal geometry (τ = 0.76, 0.74, 0.68), in which four nitrogen atoms come from the remaining coordination sites of two L′ − ligands (Fig. 4b). The structure of the trinuclear core is noticeably asymmetric, which is reflected in various Cu-O bond lengths and Cu-Cu distances as well as Cu-O-Cu angles Fig. S3. † The oxygen atom from the hydroxido ligand is slightly raised above the plane (by 0.157 and 0.353 Å for two positions of the disordered OH-group) formed by three copper atoms. The phase purity of the bulk samples was confirmed by XRD analyses as shown in Fig. S4 In summary, the geometry of a complex cation is similar to those described in previously reported studies, 11 in which two terminal triple-stranded [Na(L′) 3 ] units wrapped a planar [Cu 3 (μ 3 -OH)] 5+ core to form a bis(triple-helical) complex with trigonal-bipyramidal topology. The rigid bisbidentate L′ − ligand links one apical Na ion and one equatorial copper(II) ion in a cis-bridging mode.

Magnetic properties
Variable-temperature direct current magnetization data for polynuclear compounds 1, 2, and 4 were collected on powdered microcrystalline samples over a temperature range from 2 to 300 K and under an applied dc field of 5000 Oe. The magnetic data of dinuclear complex 1 are shown in Fig. 5. The significant decrease of the effective magnetic moment below 50 K and the presence of the maximum on the M mol vs. T curve at T = 6 K confirm the weak antiferromagnetic exchange. Therefore, the following spin Hamiltonian was used for fitting the magnetic datâ for which a simple analytical formula can be derived as  where x = μ B Bg. The analysis of the experimental data resulted in J = −6.4 cm −1 , g = 2.04 and χ TIP = 5.25 × 10 −9 m 3 mol −1 (Fig. 5), where χ TIP stands for the temperature-independent paramagnetism. The magnetic properties of tetranuclear complex 2 are similar, the effective magnetic moment decreases on lowering the temperature practically to zero and there is also the maximum on the M mol vs. T curve at T = 70 K; all these data suggest the presence of the strong antiferromagnetic exchange. The most probable superexchange pathway leads through the bridging μ-N1,N2-triazolyl units; thus the following spin Hamiltonian was postulated where four Cu(II) atoms with S i = 1/2 result in 64 magnetic states. Here it is convenient to use the coupled basis set |αSM S 〉, where α denote the intermediate quantum numbers representing the coupling path. When all local g-factors are equal, it is possible to treat only zero-field magnetic states |αS〉 and with the help of the irreducible tensor operator's technique, 12 the energies ε i,0 (αS) of these states can be evaluated. Then, the magnetic levels can be calculated as ε j (αSM S ) = ε i,0 (αS) + μ B gM S B. Finally, the formula for the molar magnetization at any temperature and magnetic field can be applied as The best-fit to the experimental magnetic data was obtained with J = −58.7 cm −1 and g = 2.09 and the resulting energy levels are also depicted (Fig. 6).
It is interesting to compare magnetic properties of the complexes [Cu 4 (H 2 L1)(ClO 4 ) 8 ], 5b [Cu 4 (HL1)(ClO 4 ) 4 ], 5d from our previous articles and [Cu 4 (L1) 4 ] from the current work. Despite the general similarity, these complex exchange parameters are different: −70, −53.7 and −58.7 cm −1 , respectively. All complexes are structurally tetranuclear and the main difference between them lies in the degree of protonation of the ligand which causes distortion of the Cu-N-N-Cu′ bridge fragment. From the literature 18 it is known that the absolute J value decreases with the asymmetry of the Cu-N-N-Cu′ framework. Asymmetry of the bridge could be evaluated as the Cu-N-N and N-N-Cu′ angle differencehere noted as Δ. Notable that the highest value Δ (10.83°) is observed for the complex [Cu 4 (HL1)(ClO 4 ) 4 ] for which the value of the exchange parameter is the smallest in this series and, in contrast, the lowest value of Δ (8.35°) corresponds to the maximum value of the J-parameter in [Cu 4 (H 2 L1)(ClO 4 ) 8 ].
The magnetic data of trinuclear complex 4 displayed in Fig. 7 exhibit very strong antiferromagnetic exchange, as it is evident from the room temperature value of the effective magnetic moment (2.3μ B ), which is significantly lower than the theoretical value of 3.0μ B for three non-interacting spin centers (S i = 1/2 with g = 2.0). The effective magnetic moment is constantly lowering on cooling down to ≈100 K, and then the plateau with a value of 1.8 μ B is reached in the temperature range from 80 to 40 K, which agrees well with the S = 1/2 ground state. However, further cooling results in a drop of μ eff down to 1.1μ B at T = 2 K. This drop can be explained by the antisymmetric exchange among copper atoms within the triangle arrangement. The concept of this interaction was developed by Dzyaloshinsky and Moriya, 13 recently reviewed, 14 and is typical of antiferromagnetically coupled trinuclear copper complexes. 11,15 Then, the suitable spin Hamiltonian is of the form fied to d = (0, 0, d z ) according to Moriya rules. Thus, we are left with three free parameters, J, g and d z . Here, the local basis set |S 1 M 1 〉|S 2 M 2 〉|S 3 M 3 〉 was utilized and the molar magnetization was calculated as where Z is the partition function and the direction of the magnetic field is defined as B a = B(sin θcos ϕ,sin θsin ϕ,cos θ). Then, the averaged molar magnetization of the powder sample was calculated as the integral (orientational) average First, the experimental data were analyzed without introducing the antisymmetric exchange (ASE), which resulted in J = −213 cm −1 and g = 2.02; however, this model cannot account for low temperature data as it is evident from Fig. 7. Therefore, the ASE was included into the fitting procedure. The best-fits were acquired with J 12 = −236 cm −1 , J 13 = J 23 = −223 cm −1 , g = 2.05 and |d z | = 16.8 cm −1 (Fig. 7) or alternatively with J 12 = −218 cm −1 , J 13 = J 23 = −231 cm −1 , g = 2.05 and |d z | = 16.7 cm −1 (Fig. S5 †). It must be noted that the equilateral model with J 12 = J 23 = J 13 = J was unable to concurrently describe temperature and field dependent magnetic data for 4 and therefore the isosceles model was applied. The respective energy levels are shown in Fig. 7 and Fig. S5, † where the energy pattern corresponds to two S = 1/2 and one S = 3/2 spin states split due to the antiferromagnetic coupling. Also, it is apparent that the ASE generates large magnetic anisotropy of two S = 1/2 levels, which can be quantified by calculating the g-factors for the lowest Kramers doublets using effective spin S eff = 1/2, which results in g xy = 0.83 and g z = 2.05.

EPR studies
The X-band EPR spectra of complexes 1 and 2 were recorded on powder samples at room and 50 K temperatures. The magnetic exchange properties of the polynuclear Cu(II) species resulting from the antiferromagnetic interactions very often cause no EPR signals to be observed at room temperature (apparently, because relaxation phenomena hamper the observation) and only badly resolved spectra at low temperatures. In the present case, the room temperature EPR spectra of title complexes exhibit anisotropic signals ( Fig. 8 and Fig. S6 †) typical of square pyramidal/trigonal-bipyramidal Cu(II) centres and show a poorly resolved rhombic feature both at room temperature and at 50 K.
The EPR spectra of polycrystalline samples 1 and 2 contain anisotropic resonances in the half-field region at 1600 G, assignable to a forbidden transition (ΔM S = ±2, g ≈ 4.2), as well as anisotropic signals around the typical 3200 G region (ΔM S = ± 1).
The spectra are described by a rhombically distorted spin Hamiltonian with a fine structure Ĥ ¼ βðg x S x H x þ g y S y H y þ g z H z S z Þ þ DðS z 2 À SðS þ 1Þ=3Þ þ EðS x 2 À S y 2 Þ ð 8Þ where S = 1, 2 is the total spin for the tetramer 2 and S = 1 for the dimer 1, S x , S y , and S z are the projections of the total spin  on the x, y, z axis, respectively, D and E are the components of the fine-interaction tensor, g x , g y , and g z are the components of the g-tensor and H is the applied magnetic field. The bestfitted values are collected in Table 1.
The EPR spectrum of the tetramer was simulated via the Belford (eigenfield) method 16 as the sum of the spectra of four complexes, three of which have spin 1 (but the parameters of the complexes with E 2 (1) = E 3 (1) = 0 were assumed to be identical) and one is spin 2. The complex concentrations in different spin states were calculated from Boltzmann's populations of the corresponding levels at given temperatures (Fig. 8). The best fit parameters D and E and the components of the g-tensor are given in Table 1.
Electron paramagnetic resonance spectra of the ground crystalline sample of 4 have been recorded at room temperature and at 2 K at X-band frequencies. At room temperature, the EPR spectrum of the trimer 4 contains one weak unresolved peak (Fig. S7 †). Upon cooling to 2 K, however in addition to the main signal at g = 2.04 the EPR spectrum exhibits a second band at lower fields with g = 1.01 (Fig. 9). These features are associated with the presence of the antisymmetric exchange resulting in large anisotropy of the g-tensor of the ground state with the effective spin 1/2. 11 Moreover, the experimental value of g xy = 1.01 is in good agreement with the value of g xy = 0.83 obtained from simulation of the magnetic data.

General
The reagents and solvents employed were commercially available and used as received without further purification. The C, H, and N microanalyses were carried out with a PerkinElmer 240 elemental analyser. Electrospray mass spectra of complexes were recorded on a Finnagan TSQ 700 mass spectrometer in the positive ion mode. Samples were prepared at a concentration of ∼2 mg ml −1 MeOH. Spectra were acquired over an m/z range of 50-2000; several scans were averaged to provide the final spectrum. Magnetic susceptibility measurements were performed with the use of a Quantum Design magnetometer/susceptometer PPMS-9 under an external magnetic field of 0.5 T in the temperature range of 2-300 K. The diamagnetic contributions of the samples were estimated from Pascal's constants. The X-band EPR spectra were recorded on an Elexsys E-680X radiofrequency spectrometer (Bruker). The commercially available, CuCl 2 ·2H 2 O, Cu(CH 3 COO) 2 ·H 2 O and Cu(ClO 4 ) 2 ·6H 2 O were used as reactants. The synthesis of H 2 L1 and H 2 L2 was described previously. 5
Data Preparation of complex 2. The same synthetic procedure as for 1 was used except that the CuCl 2 ·2H 2 O was replaced by Cu (CH 3 COO) 2 ·H 2 O giving sea-green X-ray-quality single crystals of 2. Preparation of complex 3. 1,2-Bis(5-( pyridin-2-yl)-1,2,4triazol-3-yl)propane (0.318 g, 1 mmol) was suspended in a MeOH-water mixture (1 : 1 v/v) (10 cm 3 ) and a solution of Cu(ClO 4 ) 2 ·6H 2 O (0.378 g, 1 mmol) in MeOH (10 cm 3 ) was added. The resulting blue solution was vigorously stirred for 3 h at room temperature. Then, the solution was left to stand for a few days at room temperature and blue single crystals of X-ray quality precipitated.
Crystallography. Single crystal X-ray diffraction data were collected using a Bruker SMART APEX II diffractometer with a CCD detector (1, 2 and 4) or an Xcalibur™2 diffractometer (Oxford Diffraction Ltd) with the Sapphire2 CCD detector (3), both the devices contained a monochromatic radiation source (MoKα radiation, λ = 0.71073 Å). The structures of complexes were solved by direct methods and refined in the full-matrix anisotropic approximation for all non-hydrogen atoms. Some hydrogen atoms were found in differential Fourier maps and their parameters were refined using the riding model. All the calculations were performed by direct methods and using the SHELX-97, SHELXL-2014/7 and SHELXL-2018/3 program packages. 17 The crystallographic parameters and refinements are given in Table S1. † More details can be found in the ESI and in CCDC no. 1860806-1860809. † All of the structures suffered from disorder to some extent. In all cases, crystals exhibited the usual problems of this type of structure, namely, weak scattering due to a combination of poor crystallinity, extensive solvation, and disorder of anions/solvent molecules. Some solvent molecules in 4, which could not be localized, were removed by the SQUEEZE procedure. 10 The structure 2 was solved with disordering of the O atom of MeOH molecules in two positions (0.2/0.5), restraints on O-C bond lengths (DFIX). The structure 4 was solved with O1 M atom disordering in two positions (0.55/0.45), restraints on bond lengths and angles as well as on displacement parameters (DFIX, FLAT, RIGU and SIMU). In 1 EADP constrain and DFIX restraints were used for the refinement of partly occupied positions of solvate molecules. In each case, the basic structure and connectivity of the complex cation could be unambiguously determined, which is all that is required for the purposes of this work.

Conflicts of interest
There are no conflicts to declare.