Mononuclear Transition Metal Complexes of 7-Nitro-1,3,5-Triazaadamantane

Complexes of the type [MCl2(7-nitro-1,3,5-triaza-adamantane)2] (M = Zn(II), Pd(II), Pt(II)) and [MCl2(H2O)2(7-nitro-1,3,5-triazaadamantane)2] (M = Mn (II), Co(II), Ni(II)) have been prepared and their structures have been analysed by X-ray crystallography, elemental analysis, IR and solid state C and N NMR spectroscopy, supported by density functional theory/gauge independent atomic orbital (DFT/GIAO) calculations. In each case, 7-nitro-1,3,5-triazaadamantane acts as a mono-dentate ligand and binds to one metal centre only, in spite of the presence of three equivalent amino nitrogens. In the Co(II) and Ni(II) complexes, a two-dimensional intermolecular hydrogen bonding network between the aquaand the chloro ligands is established. The uncoordinated amines of the 7-nitro-1,3,5-triaza-adamantane are not involved in any Hbonding, as a result of the exceptionally low basicity of this compound. Introduction Organic compounds with nitrogen atoms in 1,3 position to each other show interesting properties due to nonbonding interactions based on homoconjugation between the nonbonding orbitals at N. This so called 1,3,n,n interaction is particularly pronounced in azaadamantanes where the nitrogens occupy bridgehead positions in a rigid cage, as in the case of 1,3,5,7-tetraaza-adamantane (better known as hexamethylenetetramine, hmta). The communication between the nitrogen atoms is evident from the typical chemistry of these compounds. The pKa value of 4.89 of the protonated species hmta-H+ is much lower than that of organic R3N-H (pKa 10 – 11) and NH4 (pKa 9.2), indicative of a fairly low basicity of the parent amine hmta. In aqueous systems hmta is mono-protonated only, whereas diprotonation, with a pKa of -2, requires addition of strong acids and leads to decomposition of the hmta framework. Likewise, the metal coordination is compromised, in a sense that coordination to one of the nitrogens alters the coordination behaviour of the others. The coordination pattern is most diverse with Ag(I), where hmta acts as a tetradentate, tridentate or bidentate ligand leading to three dimensional structures, ribbons, sheets or chains. With other metals, tetradentate and tridentate coordination modes have occasionally been observed as well, but bidentate and monodentate binding is far more typical. There are also many examples where hmta is found in the crystal packing without coordination to a metal centre. In these cases, hydrogen bonding between hmta and the metal coordinated ligands is observed. Among the adamantanes bearing three bridgehead nitrogens, 7-phospha-1,3,5-triazaadamantane and the corresponding phosphinoxide and phosphinsulfide have been studied in quite some detail, in particular with respect to bimetallic complexes where one metal coordinates to the P atom and the other is attached to one of the nitrogens, or where the P=O/S moiety coordinates through the O and S atoms. N-coordinated mononuclear complexes have also been characterised. 13] These form with hard metal centres whose preference for P is low. The monodentate coordination pattern is also seen with a related 7-methyl-1,3,5-triazaadamantane derivative, which coordinates Fe(II) to one of the N atoms. 1,3,5-triazaadamantanes bearing heterocyclic substituents such as pyridine or imidazole have been found to undergo aminal imine rearrangement in the presence of Fe(II), and hexadentate tri-imine complexes are formed. 7-Nitro-1,3,5-triazaadamantane (NO2-TAA) tends to monoand bidentate behaviour, as far as one can conclude from the few examples studied so far. These consist of bidentate Ag(I) and Hg(II) complexes, which were characterised in the solid state by single crystal X-ray diffraction, and Hg(II), Zn(II), Cd(II) and Pb(II) complexes whose ligand to metal stoichiometry was determined in CH3CN solution. The H NMR spectra of these complexes point to dynamic behaviour and rapid ligand exchange because the ligand signals shift but do no split in consequence of the reduced symmetry in the coordinated state. The lability is also manifested in the different composition of the Hg(II) complexes, which contain two ligands per metal in solution but only one in the solid state. Pt(II) forms kinetically stable complexes, as shown by multinuclear NMR spectroscopy in CHCl3 solution. Stoichiometry control allows for the selective synthesis of mononuclear compounds where NO2-TAA binds in a monodentate manner. However, the corresponding dinuclear species were obtained as by-products or when two equivalents of the Pt(II) precursor were used, suggesting that the bidentate binding mode might be preferred. In this aspect, NO2-TAA closely resembles the behaviour of hmta as a ligand. In the present work, we describe the synthesis and solid state analysis of a series of new mononuclear transition metal complexes with NO2-TAA. These are the first examples where this ligand binds in a monodentate manner without the need for stoichiometric control, and where no aggregation to polynuclear species takes place upon crystallisation. Results and Discussion (a) Synthesis and characterisation of the ligand: The ligand 7-nitro-1,3,5-triazaadamantane 1 was synthesised from nitro-methane, paraformaldehyde and ammonium acetate as shown in Scheme 1, using a modified procedure from the literature. Alternatively, 1 can be prepared from hexa-methylenetetramine, formic acid and nitromethane, or from tris(hydroxymethyl)nitro-methane, para-formaldehyde and ammonia in similar yields. Scheme 1. Synthesis of 7-nitro-1,3,5-triazaadamantane. Elemental analysis, mass spectrometry, IR and NMR spectroscopic data confirm the structure of the compound. The H NMR data in CDCl3 agree well with literature data in CD3CN and in CD3COOD. In CDCl3 solution, the spectrum displays a 6 proton singlet at 3.83 ppm for the C-CH2-N protons, and two doublets at 4.12 and 4.48 ppm accounting for 3 protons of the axial and equatorial protons of the N-CH2-N groups. The signal of the axial protons is at higher chemical shift than that of the equatorial ones, as confirmed from H NOE experiments. The two types of protons couple to each other with a geminal coupling constant of 13.2 Hz, and the lines of each signal are slightly broadened due to an unresolved long range coupling to the C-CH2-N protons. The C NMR spectrum shows three signals at 59.7 (C-CH2-N), 72.6 (C-NO2), 73.4 (N-CH2-N) ppm. The C MAS-CP NMR in the solid state suggests that the rotation of the nitro group is blocked, and the compound is less symmetric than in solution. The C-CH2-N groups appear as two signals at 56.6 and 59.4 ppm, in a ratio 2:1, whereas the N-CH2-N resonance at 71.8 ppm accounts for three carbons. The quaternary carbon at 74.1 ppm was identified by its disappearance at short (50 μs) CP contact time. The CH2 groups are more shielded and the quaternary C is less shielded than in CDCl3 solution. N MAS-CP NMR spectra, run at natural abundance against external glycine = +12.5 ppm as a reference (NH4NO3 = 0 ppm scale), show the amine nitrogens at 28.4 ppm but the NO2 group was not detected. On the basis of the nuclear properties of N this is a plausible result. The typical N relaxation times T1 of amines can vary widely but the nuclear Overhauser enhancement factors are usually fairly favourable (e.g. BuNH2 70 s/NOE -3.9, quinuclidine approx. 1300 s, indole NH 3.5 s, heterocyclic tertiary N 25 s, NOE -5.0 ), whereas nitro groups relax very slowly (e.g. T1 of nitrobenzene: 400 s [26] or 180 s ) and display unfavourable NOE factors near -1 (e.g. (HOCH2)2CMeNO2 -1.0 ). The chemical shift of the NO2-TAA amines is quite different from that of simple aliphatic amines (typically 2 ppm ) but comparable to that of hmta (26.5 ppm ) where the nitrogens are also subject to 1,3,n,n interactions. From the slightly higher chemical shift in NO2-TAA one might anticipate a lower charge density at N than in hmta. This is in good agreement with results obtained from DFT calculations [18] and also compatible with the extremely low pKa value of the protonated NO2TAA (3.42, as compared to 4.89 in protonated hmta ). There is also evidence from IR and electronic spectroscopy for a charge transfer from the N atoms to an antibonding orbital of the NO2 group, which of course is absent in hmta. (b) synthesis and characterisation of the complexes: The synthesis of the transition metal complexes, outlined in Scheme 2, was guided by the relatively poor solubility of the ligand in most solvents. Near saturated solutions of NO2-TAA in CH2Cl2 (solubility 1.7 ± 0.8·10 mol dm) were carefully covered with a ethanol solution of the metal salt (ZnCl2, MnCl2·4H2O, CoCl2·6H2O, NiCl2·6H2O) and crystals of the complex were grown by slow diffusion. Independent of the ligand to metal ratio, which was varied from 4:1 to 1:1, complexes with a 2:1 stoichiometry formed as the sole products, except for Co(II) where small amounts of oligomeric side products with a 3:2 and 4:3 stoichiometry were detected. In the case of Zn(II) this is in contrast to the previously described complexes where a 1:1 stoichiometry was found in CH3CN solution. The Pd(II) and Pt(II) complexes were obtained from [MCl2(PhCN)2] by ligand exchange in dichloromethane solution, or starting from PdCl2 or K2[PtCl4] in DMF or aqueous solution. The former reaction occurs via the intermediacy of the mixed ligand [MCl2(PhCN)(NO2-TAA)] complex, which, in the case of Pt(II), could be isolated and characterised. Scheme 2. Synthesis of the transition metal complexes. Elemental analyses confirmed the molecular formula attributed to the complexes. In the IR spectrum, the NO2 asymmetric stretch is shifted to higher wavenumbers and shows as two lines, since two ligands are present in the molecule. Also the CH2 scissoring is seen at slightly higher wavenumbers, and the signals associated with the fundamentals of the azaadamantane cage in the fingerprint region split into several lines due to the lower symmetry of the coordinated ligand. From the general IR pattern one can group the complexes into three classes: Those of Mn(II), Co(II) and Ni((II) contain coordinated H2O molecules which are identified by the characteristic strong symmetric and asymmetric O-H stretching vibrations in the range of 3390 to 3220 cm and a H-O-H bending vibration at 1613-1615 cm. The spectra are nearly super-imposable, indicating that these complexes are structurally similar. Among the water-free compounds, the IR of the Pd(II) and Pt(II) complexes are almost identical, but quite different from that of the Zn(II) complex. This reflects the different coordination geometry which is square planar in the former and tetrahedral in the latter. DFT calculations (B3LYP/6-31G* and Lan2DZ for Pd and Pt) were used to assist the spectra interpretation. A good agreement between experimental and calculated vibrational frequencies was obtained, as shown in Figure 1 for ligand 1 and Zn complex 2. Figure 1. IR spectra of the free ligand 1 (red) and the Zn complex 2 (grey), in comparison with the calculated (B3LYP/6-31G*) IR vibrational frequencies (top: experimental transmission spectra; bottom: calculated vibrational frequencies and their respective IR intensities). Scheme 3. C NMR assignment of the coordinated ligand. Table 1. C MAS-CP coordination shifts  = (complex) (free ligand) (ppm). Group Assignment Complex 2 (Zn) 6 (Pd) 7 (Pt)