A new copper(II) chelate complex with polyamines as fire retardant and epoxy hardener: Synthesis, crystal and electronic structure, and thermal behavior of (ethylenediamine-N,N′)-(diethylenetriamine-N,N′,N″)-copper(II) hexafluoridosilicate

2.1 Materials and methods

In order to obtain the [Cu(eda)(deta)]SiF₆ chelate complex, the following chemicals were used: copper(II) hexafluoridosilicate hexahydrate – CuSiF₆·6H₂O (deep blue crystals, M = 313.5 g·mol⁻¹, d²⁰ = 2.21 g·cm⁻¹) (Ray et al., 1973) and polyethylenepolyamine (pepa) – H₂N[—C₂H₄NH—]_nH, where n = 1 (eda) and 2 (deta) (yellowish transparent viscous liquid, $\overline{M}$ = 81.5 g·mol⁻¹, d²⁰ = 0.927 g·cm⁻¹, t_flash = 94 °C, t_ignition = 97 °C, t_{self-ignition} = 390 °C) (Baratov and Korolchenko, 1990), IR (KBr cuvette, cm⁻¹): 3372, 3254 and 3210 (ν_NH); 2934, 2908 and 2790 (ν_CH); 1600 (δ_NH); 1460 (δ_CH). All reagents were purchased through commercial sources and used as-received without further purification.

IR absorption spectra were recorded in the spectral range of 4000–500 cm⁻¹ with a resolution of 2 cm⁻¹ on a Perkin Elmer Spectrum Two FTIR spectrometer. The IR measurements were performed with a solid sample of 1 pressed in spectroscopically pure KBr pellet, or a liquid sample of pepa using a KBr cuvette.

Thermal gravimetric analysis for crystalline phase of 1 was performed on a Q-1500D thermal analyzer (F. Paulik, J. Paulik, L. Erdey system) in the temperature range from 25 to 700 °C with a heating rate of 5 °C per minute. The experiments were carried out in dynamical regime in air atmosphere. The weight of each solid-phase sample amounted to 100 mg. Aluminum oxide was used as a standard.

2.2 Synthesis

[Cu(eda)(deta)]SiF₆ has been synthesized by direct interaction of anhydrous copper(II) hexafluoridosilicate with pepa. Polycrystalline CuSiF₆ (2.05 g, 0.01 mol) was placed into a porcelain mortar, and 1.75 cm³ (0.02 mol) of pepa was added. The reaction mixture was triturated till the appearance of a homogeneous dark-blue liquid suspension. It was left for a few days at room temperature, and the solid phase appeared according to the following reaction: CuSiF₆ + NH₂C₂H₄NH₂ + NH₂C₂H₄NHC₂H₄NH₂ = [Cu(NH₂C₂H₄NH₂)(NH₂C₂H₄NHC₂H₄NH₂]SiF₆

Yield: 100% (with respect to CuSiF₆), IR (KBr pellet, cm⁻¹): 3416, 3238 and 3142 (ν_NH); 2952, 2930 and 2886 (ν_CH); 1584 (δ_NH); 1454 (δ_CH). Next the solid phase was dissolved in water. Dark-blue single crystals of 1 suitable for X-ray crystal structure determination were obtained out of the aqueous solution by recrystallization.

2.3 Crystal structure determination

A single crystal of 1 was mounted on the Xcalibur diffractometer (Mo K_α-radiation, λ = 0.71073 Å, graphite monochromator) equipped with an Oxford Cryosystem cooler and a CCD detector. 180 ω oscillation images with a frame width of 0.5° and an exposure time of 20 s were collected at 100.0(1) K using crystal-to-detector distance of 60 mm. After integration the data was corrected for Lorentz and polarization effects (Oxford diffraction, 2002). Unit cell parameters were obtained by the least-squares refinement based on reflection angles in the range 6.20° < 2θ < 59.06°. Structure 1 was solved by direct methods applying SHELX software package (Sheldrick, 2008). Other non-hydrogen atoms were located from the Fourier difference synthesis and refined by the least-squares method in the full-matrix anisotropic approximation. The hydrogen atoms of —NH— and —NH₂ groups also were located from the Fourier difference synthesis, while the hydrogen atoms of —CH₂— groups were placed geometrically. All of hydrogen atoms were refined as riding atoms with the C—H and N—H preset distances, 0.99 Å and 0.92 Å, respectively as well as with isotropic temperature factors of 1.2 times the U_eq values for the parent carbon or nitrogen atoms. The crystal data, collected reflections and parameters of the final refinement data for 1 are reported in Table 1. The positional and thermal displacement parameters are given in Table S1 (Supp. info). All images of the crystal structure of 1 were prepared using DIAMOND program (Brandenburg, 1998). The atom numbering system for the asymmetric part of 1 is represented in Fig. 1.

Close

Fig. 1 The atom numbering system for the asymmetric part of 1. Thermal ellipsoids are displayed at the 30% probability level for non-hydrogen atoms.

Export to PPT

2.4 Quantum-chemical calculation

The numerical simulation of the complexation processes in eda–deta–CuSiF₆ system was carried out by means of quantum-chemical analysis. The ab initio quantum-chemical calculations (the restricted Hartree-Fock method with a basis set 6-31*G) were performed using the HyperChem program version 8.0.6 (HyperChem, 1995–2009; Frank, 1990). The base crystallographic data of CuSiF₆·6H₂O (Ray et al., 1973) and 1 were used for the design of [Cu(H₂O)₆]SiF₆ and [Cu(eda)(deta)]SiF₆ discrete clusters as well as for the construction of free eda and deta molecules.

Calculation of a charge density distribution on atoms for discrete clusters was performed without geometrical optimization of structural fragments whereas the geometry of free molecules was optimized. All calculations were carried out under the hypothesis that the presented clusters (like the free molecules) are isolated and are in vacuum. The results of the quantum-chemical calculation of energy effects of the pepa–CuSiF₆ interaction are given in Table 2; the exhaustive information about the charge density redistribution on the atoms in above particles is represented in Table S2 (Supp. info).

3.1 X-ray crystal structure description

The interaction of pepa (a curing agent of epoxy resins) with noncombustible copper(II) hexafluoridosilicate (a fire retardant) in many respects predetermines the fire retardant-hardener properties for the [Cu(eda)(deta)]SiF₆ complex whose crystal structure is displayed in Fig. 2. In 1, polyamine molecules – eda and deta – coordinate to inorganic cupric salt, forming the chelate complex where all amine groups of the organic ligands are simultaneously bonded to central atom – Cu(II). As can be seen in Fig. 1, the chelation mode determines the structure of 1 that is composed of discrete complex cations resembling those in the [Cu(deta)(eda)](BF₄)₂ (Valdes-Martinez et al., 2006), [Cu(deta)(eda)](ClO₄)₂ (Duan-Jun 1993; Ullah et al., 1996; Patel et al., 2007), as well as in the [Cu(deta)(2,2́-bipyridine)](BF₄)₂, [Cu(deta)(1,10-phenantroline)](ClO₄)₂ (Patel et al., 2005) and [Cu(deta)(tetramethylethylenediamine)]Br₂·H₂O (Saleemh et al., 2017). Similarly, cations of 1 have the same stoichiometry ([Cu(deta)(eda)]²⁺) and together with SiF₆²⁻ anions are main building blocks its crystals.

Close

Fig. 2 Crystal structure projection of 1 onto xz plane. Hydrogen bonds are depicted by dot line.

Export to PPT

The bond lengths and bond angles for 1 are listed in Table 3. The Cu²⁺ ion is pentacoorditated by N atoms (two N atoms from bidentate eda + three N atoms from tridentate deta) in the titled structure. The coordination sphere around the central atom is an elongated square pyramid in which N4, N5 atoms from eda and N1 and N3 atoms from deta form the bottom of the polyhedron, and another nitrogen atom, N2, from deta occupies the apical position. The Cu–N bond lengths in the pyramid base are close and these range from 2.008(4) to 2.043(4) Å; the axial Cu—N2 bond is more lengthy (2.223(3) Å). It should be noted that the bottom of the square pyramid is not planar. The N3 and N5 atoms are above of the N1–N3–N4–N5 plane. Therefore, a τ parameter for coordination environment of Cu(II) (coordination number = 5) was calculated using the procedure proposed by Addison et al. (1984). The obtained value (τ₅ = 0.16) unambiguously fixes the pentahedron shape which has proved considerably closer to square-pyramid than to trigonal-bipyramid.

Nevertheless, the coordination sphere of Cu(II) can be complemented by one more F atom of neighboring hexafluoridosilicate anion (see Fig. 2). Thus, a distorted square bipyramid arises in unit cell of 1 in consequence of supplementary coordination of the central atom by atom of F2 that occupies another apical position on the opposite side of a square pyramid (Cu—F2 3.086(3) Å). Observed loosening of Cu—N2 and Cu—F2 bonds is mainly caused by electrostatic repulsion of electron lone pair of nitrogen atom or fluorine atom from electron pair located on $d_{z^2}$ atomic orbital (AO) of Cu²⁺ ion (Jahn-Teller effect (Tanaka et al., 1979)).

The SiF₆²⁻ anions have an almost ideal octahedral shape (see Table 3 and Fig. 2). The bulky SiF₆²⁻ units are incorporated into crystalline framework of 1 where these crosslink the cation stacks by N—H…F hydrogen bonds (the branched system of the hydrogen bonds for 1 is presented in Table 4). These hydrogen bonds impart a directed character to the ionic interaction Cat²⁺…An²⁻. The examined hydrogen bonds are a striking example of especial influence of the directed ionic interaction on the crystal structure forming (Desiraju, 2000, 2002).

3.2 IR spectroscopy

Besides X-ray structure analysis, 1 was also identified by IR spectroscopy (Fig. 3). Due to the fact that cupric salt is able to interlink the polyamine molecules into chelate complex, one can observe the shifts of N—H absorption bands after complexing. Therefore, the absorption bands caused by the stretching and bending of NH₂ and NH groups of the coordinated and free molecules of polyamine will interest us primarily (Nakamoto, 2009). So, for free pepa (Fig. 3, A) the high-frequency region 3372 and 3254 cm⁻¹ is related to the stretching of —NH₂ whereas the band observed at 3210 cm⁻¹ is ascribed to the stretching of —NH—; the observed oscillation frequency at 1600 cm⁻¹ is attributed to the bending of N—H bonds. After bonding of the pepa (eda + deta) with CuSiF₆ (Fig. 3, B), one of the N—H absorption bands is shifted to the high-frequency region observed at 3416 cm⁻¹ whereas two other bands are shifted to the low frequency region (3238 and 3142 cm⁻¹). The observed wavenumber at 1584 cm⁻¹ related to the bending of N—H bonds is also shifted to low frequency region.

Close

Fig. 3 IR spectra of pepa (A) and 1 (B).

Export to PPT

3.3 Electron-molecular structure

The complexing in pepa – CuSiF₆ system is accompanied by some changes in the electronic parameters of coordinated eda and deta molecules in comparison with ones in uncoordinated state. The carried out quantum-chemical calculations have revealed that the electron density of nitrogen atoms within coordination core efficiently shifts to the Cu(II) atom owing to chelate effect. So, the values of charge density (δ) on amine nitrogen atoms of uncoordinated eda and deta molecules are −0.266, −0.381, −0.393 $\overline{e}$ for N1, N2, N3 atoms and −0.366, −0.360 $\overline{e}$ for N4, N5 atoms, respectively, and the δ value on Cu atom in copper(II) hexafluoridosilicate hexahydrate equals to −0.575  $\overline{e}$ (Fig. 4, A–C). However, the electron density on —NH₂ and —NH— groups of eda and deta molecules is decreased (the δ values are −0.026, −0.148, −0.114 $\overline{e}$ for N1, N2, N3 atoms and −0.094, −0.087 $\overline{e}$ for N4, N5 atoms, respectively) after the [Cu(eda)(deta)]SiF₆ formation while the electron density on copper(II) atom, conversely, is increased (the δ value is −0.672 $\overline{e}$ for Cu atom) (Fig. 4, D).

Close

Fig. 4 Charge density (±δ, e ¯ ) distribution on the atoms in eta (A), deta (B), [Cu(H₂O)₆]SiF₆ (C), [Cu(eda)(deta)]SiF₆ (D).

Export to PPT

The above-mentioned distribution of electron density within the coordination core results in the degeneracy elimination of the 3d-AOs in the Cu²⁺ ion. According to theory of crystal field advanced in (Van Vleck, 1932), the degenerated 3d-AOs are split influenced by the square-pyramidal crystal field. The constructed diagram of molecular orbitals (MOs) for [Cu(eda)(deta)]SiF₆ and 3d-AOs splitting by the square-pyramidal crystal field for chelated Cu²⁺ ion are represented in Fig. 5. Taking into account the chelate effect, the split parameters of the 3d-AOs by the square-pyramidal crystal field acquire next sequence in 1; the energy levels are disposed in the following way (in ascending order): d_xy < d_yz < d_xz <  $d_{z^2}$  <<  $d_{x^2-y^2}$ (see Fig. 5).

Close

Fig. 5 The MOs diagram for [Cu(eda)(deta)]SiF₆ (on the left) and splitting of 3d-AOs by square-pyramidal crystal field for chelated Cu²⁺ ion (on the right).

Export to PPT

3.4 Thermal behavior

The bonding efficiency of a combustible organic polyamine with an incombustible inorganic salt largely predetermines the thermal stability of the [Cu(eda)(deta)]SiF₆ chelate complex – fire retardant-hardener of epoxy resins. Therefore, the thermal behavior of 1 was studied by thermal gravimetric analysis. The curves of thermogravimetry (TG), differential thermogravimetry (DTG) and differential thermal analysis (DTA) are displayed in Fig. 6.

Close

Fig. 6 TG, DTG and DTA curves of polycrystalline sample of 1.

Export to PPT

The TG curve of the thermal decomposition of 1 is represented by four sections. The first involves the temperature range from 20 to 160 °C. It is accompanied by a loss of weight of 13.4 wt% and by the appearance of a minimum on the respective section of the DTA curve. At this temperature interval (DTA curve) an endothermic process is observed at 112 °C.

The second stage of the thermal decomposition lies in the temperature range from 160 to 253 °C (Δm = 14.3 wt%) with an endothermic minimum at 212 °C. The stepwise break-down of the inorganic part of complex takes place at these two temperature intervals: $$[\text{Cu}(eda)(deta)\text{Si}{\text{F}}_6\stackrel{20-253{}^{\text{O}}\text{C}}{\to }[\text{Cu}(eda)(deta)]{\text{F}}_2+\text{Si}{\text{F}}_4\uparrow$$

The third stage of the thermal decomposition occurs in the temperature range from 253 to 368 °C and is accompanied by the appearance of exothermic and endothermic effects on the DTA curve. The loss of weight amounts to 39.9 wt% at this stage. In this temperature interval, the partial thermal-oxidative destruction of the amine part of the complex (i.e., a dehydrogenization of the coordinated eda and deta with a concurrent combustion of H₂) takes place.

The fourth stage of a thermal destruction takes place in the temperature range from 368 to 544 °C and corresponds to a complete combustion of the pyrolytic residue of organic constituents of the complex. It is accompanied by the appearance of a pronounced exothermic effect on the DTA curve with a maximum at 488 °C. The weight loss amounts to 15.3 wt% at this final stage.

At the very end, we would like to consider the chemical influence of the pepa–CuSiF₆ interaction and the complex 1 formation on the combustibility of the coordinated polyamine. So, if the ignition point for free eda and deta is 45 °C and 97 °C, respectively then the coordinated eda and deta in 1 do not inflame at all. In other words, a combustible curing agent of epoxy resins (pepa), which is made up of the highly inflammable eda and inflammable deta after bonding to copper(II) hexafluoridosilicate (a fire retardant) turns into a practically incombustible substance. Undoubtedly, the reason for this is the additional chemical bonds that arise between the combustible organic molecules (eda and deta) and the incombustible inorganic salt (CuSiF₆). To break down the Cu(II) ← N bonds, a considerable part of the heat energy coming from the ignition source must be expended. Moreover, in order to get a flame, it is necessary that eda and deta form a mixture of gaseous polyamines with air in which concentration of the saturated vapor eda and deta was greater than the bottom concentration limit of the flame spread (Hurley, 2016). However, even at temperature exceeding 450 °C, ignition of eda and deta was not observed above the surface of melted down complex 1. It is not less interesting to compare the thermal behavior of complex 1 with such behavior for pepa. Unlike pepa, whose total weight loss is observed in the temperature range from 20 to 170 °C, the complex 1 fully degrades at a lot more temperature (544 °C). Consequently, the polyamine molecules through Cu(II)—N bonds firmly hold in the complex, therefore their instantaneous vaporization and further ignition becomes difficult. Obviously, the complexation is the main cause of a fire retarding effect at burning of nitrogen-containing hydrocarbons.