Jiří
Tauchman
a,
Kateřina
Hladíková
a,
Filip
Uhlík
b,
Ivana
Císařová
a and
Petr
Štěpnička
*a
aDepartment of Inorganic Chemistry, Faculty of Science, Charles University in Prague, Hlavova 2030, 12840 Prague, Czech Republic. E-mail: stepnic@natur.cuni.cz
bDepartment of Physical and Macromolecular Chemistry, Faculty of Science, Charles University in Prague, Hlavova 2030, 12840 Prague, Czech Republic
First published on 26th March 2013
2-Ferrocenyl-4(3H)-quinazolinone (2) was obtained by acylation of 2-aminobenzoic amide with ferrocenecarbonyl chloride (FcCOCl) and subsequent base-catalysed cyclisation of the intermediate 2-(ferrocenecarboxamido)benzamide (1). The related benzoxazin-4-one (4) was obtained analogously (though in one step) from FcCOCl and 2-aminobenzoic acid. Compounds 2 and 4 were thionated with P2S5 to give the corresponding ferrocenylated 4(3H)-quinazoline thione (3), and a mixture of benzothiazine-4-thione (5) and benzothiazine-4-one (6, major), respectively. Attempts to prepare 3-amino-2-ferrocenyl-4(3H)-quinazolinone from 4 and hydrazine hydrate failed. The reaction gave only 2-(ferrocenecarboxamido)benzohydrazide (7), which reacted with 2-chloro-1,3-dimethylimidazolidinium chloride to afford a mixture of cyclic and open (major) condensation products, both bearing the 1,3-dimethylimidazolidin-2-ylidene terminal group (8 and 9). All compounds were characterised by spectroscopic methods and the molecular structures of 1–7 and 9 were determined by X-ray diffraction analysis. Cyclic voltammetric measurements revealed that the compounds undergo a single one-electron oxidation, localised presumably at the ferrocene unit. DFT computations for 2 as a representative example performed in the gas phase and crystal state, reproduced well the experimental geometry and further supported the assignment of the redox processes.
Although the structurally related ferrocenyl quinolines,7 quinolinones,8 flavones9 and quinoxalines10 have been already described, there is, up to the best of our knowledge, no report available in the literature concerning the structurally related, ferrocenyl-substituted quinazolinones and similar compounds, which all represent the basis of various biologically active substances.11 This prompted us to prepare and structurally characterise the directly ferrocenyl-substituted quinazolinone, viz. 2-ferrocenyl-4(3H)-quinazolinone (2), its corresponding 3,1-benzoxazin-4-one and their thionated analogues. Details of our investigations are presented in this contribution.
Scheme 1 Synthesis of quinazolinone 2 and the corresponding thione 3. |
The structurally related benzoxazin-4-one 4 (Scheme 2) was obtained in one step by the reaction of ferrocenecarbonyl chloride with anthranilic amide in pyridine. Thionation of 4 with P2S5 provided a chromatographically separable mixture containing unreacted 4 (recovered: 17%) and two new compounds: the fully thionated benzothiazine-4-thione (5, 5%) and benzothiazine-4-one (6, 65%), in which only the in-ring oxygen atom was replaced with sulphur. It is noteworthy that the reactivity observed for 4 contrasts with the literature reports13 suggesting that the reaction of 2-substituted 3,1-benzoxazin-4-ones with P2S5/NEt3 proceeds under the replacement of both oxygen atoms to afford exclusively the corresponding benzothiazin-4-thiones.
Scheme 2 Preparation and thionation of benzoxazinone 4. |
Attempts to prepare 3-amino-2-ferrocenyl-4-quinazolinone by hydrazinolysis14 of 4 led only to an opened amide-hydrazide 7 as a product of formal hydrolysis. Likewise, attempted cyclisation of 7 in the presence of 2-chloro-1,3-dimethylimidazolidinium chloride failed.15 Instead of the anticipated cyclic product, the reaction afforded a mixture of two rather unexpected condensation products, viz. the cyclic 1,3-dimethylimidazolidin-2-ylidene hydrazide 8 and its opened zwitterionic analogue 9 in 14% and 75% isolated yields, respectively (Scheme 3).16
Scheme 3 Opening of benzoxazinone 4 with hydrazine and the reaction of the resulting amido-hydrazide 7 with 2-chloro-1,3-dimethylimidazolidinium chloride. |
All compounds were characterised by 1H and 13C{1H} NMR, and IR spectroscopy, conventional and high-resolution electrospray ionisation (ESI) mass spectrometry, and by elemental analysis. Besides, the molecular structures of 1–7 and 9 were determined by single-crystal X-ray diffraction analysis.
Fig. 1 View of the molecule of amide 1 in the crystal structure of 1·½H2O showing displacement ellipsoids at the 30% probability level and the labelling scheme. Selected distances and angles (in Å and deg): C11–O1 1.234(2), C11–N1 1.361(2), C21–N1 1.408(2), C27–O2 1.248(2), C27–N2 1.328(2), O1–C11–N1 123.9(2), O2–C27–N2 122.3(2), C1–C11–N1–C21 177.6(2); Fe–Cg1 1.6463(9), Fe–Cg2 1.654(1), ∠Cp1, Cp2 2.6(1). Note: Cp1 and Cp2 are cyclopentadienyl rings C(1–5) and C(6–10), respectively; Cg1 and Cg2 denote their respective centroids. |
In fact, the individual molecules of amide 1 fully exploit their H-bonding ability. In the crystal, they assemble into dimers lying around the inversion centres via N–H⋯O interactions (A in Fig. 2) and further into infinite columnar stacks via the carboxyl-like double hydrogen bridges (B). An additional intramolecular N–H⋯O contact though with a rather acute angle at the H atom (C) is detected between the FcCONH proton and the CONH2 oxygen. The solvating water molecules (not shown in Fig. 2) reside on the crystallographic two-fold axes and between the columnar stacks, which they interconnect into a three-dimensional array via O1W–H1W⋯O1 hydrogen bonds towards the FcCO oxygens (O1W⋯O1 = 2.823(2), angle at H1W = 152°).
Fig. 2 View of a single columnar stack built up from molecules of 1 in the crystals of 1·½H2O. For clarity, the bulky ferrocenyl groups were replaced with filled black squares and the CH protons were omitted. H-bond parameters are as follows. A: N2–H3N⋯O1, N2⋯O1 = 2.975(2) Å, angle at H3N = 176°; B: N2–H2N⋯O2, N2⋯O2 = 2.883(2) Å, angle at H2N = 176°; C: N1–H1N⋯O2, N1⋯O2 = 2.690(2) Å, angle at H1N = 136°. Symmetry codes: A: ½ − x, ½ − y, −z; B: −x, 1 − y, −z; C: intramolecular. |
The overall molecular structure of hydrazide 7 (Fig. 3) compares well with that of the analogous amide 1. Similarly to 1, the ‘amide’ planes in 7 are rotated with respect to their parent aromatic rings [cf. the dihedral angles 9.0(2)° for planes (C11, O1, N1) and C(1–5), and 30.1(1)° for planes (C27, O2, N2) and C(21–26)]. The molecules of 7 assemble into ‘dimers’ via N–H⋯O hydrogen bonds (Fig. 4). However, the cross-linking of these dimers is achieved through N–H⋯O contacts involving the hydrazide NH2 (terminal) moieties, which thus take over the role played by the solvent molecules in the structure of 1·½H2O. As the result, the individual molecules of 7 associate into wave-like layers, which further assemble into a complicated three-dimensional network.
Fig. 3 View of the molecular structure of 7 showing displacement ellipsoids at the 30% probability level and the atomic labelling scheme. Selected distances and angles (in Å and deg): C11–O1 1.235(2), C11–N1 1.358(2), C21–N1 1.405(2), C27–O2 1.245(2), C27–N2 1.327(2), N2–N3 1.415(2), O1–C11–N1 124.1(1), O2–C27–N2 121.9(1), C27–N2–N3 122.6(1), C1–C11–N1–C21–172.6(1), O2–C27–N2–N3 3.7(2); Fe–Cg1 1.6408(6), Fe–Cg2 1.6463(7), ∠Cp1, Cp2 2.15(8). The ring planes are defined as for the analogous bis-amide 1 (cf.Fig. 1). |
Fig. 4 Different N–H⋯O hydrogen bonds (denoted as A–F) formed by a single molecule of 7 in the crystal. For clarity, the bulky ferrocenyl groups were replaced with filled black squares and all CH protons were omitted. H-bond parameters are as follows. A: N2–H2N⋯O1, N2⋯O1 = 3.073(2) Å, angle at H2N = 153°; B: N1–H1N⋯O2, N1⋯O2 = 2.696(2) Å, angle at H1N = 142°; C: N3–H3N⋯O2, N3⋯O2 = 2.751(1) Å, angle at H3N = 109°; D: N3–H3N⋯O2, N3⋯O2 = 3.040(2) Å, angle at H3N = 145°; E ≡ F: N3–H4N⋯O1, N3⋯O1 = 3.066(2) Å, angle at H4N = 151°. Symmetry codes: A: 1 − x, −y, 1 − z; B and C: intramolecular; D: ½ − x, ½ − y, 1 − z; E: x, −y, ½ + z, F: x, −y, z – ½. |
The molecular structures of compounds 2–6 are presented in Fig. 5 and 6, respectively. Structural data for these ferrocenyl-substituted heterocycles, which formally differ only by the type of the heteroatoms and their positions within the six-membered ring, are summarised in Table 1.
Fig. 5 View of the molecular structure of 2 showing the displacement ellipsoids at the 30% probability level. |
Fig. 6 Views of the molecular structures of compounds 3–6 showing the atom labelling scheme and displacement ellipsoids at the 30% probability level. |
Parametera | 2 | 3 | 4 | 5 | 6 |
---|---|---|---|---|---|
a Definitions: ψ is the dihedral angle subtended by the Cp1 plane and the least-squares plane of the entire heterocyclic moiety (N11, C12, Y13, C14, and C15–C20); cyclopentadienyl rings: Cp1 = C(1–5), Cp2 = C(6–10), Cg1/Cg2 are the respective ring centroids. | |||||
Y13/Z | N13/O | N13/S | O13/O | S13/S | S13/O |
N11–C12 | 1.298(2) | 1.302(2) | 1.279(2) | 1.289(3) | 1.277(3) |
C12–Y13 | 1.382(2) | 1.382(2) | 1.379(2) | 1.762(2) | 1.767(3) |
Y13–C14 | 1.371(2) | 1.361(2) | 1.390(2) | 1.737(2) | 1.780(3) |
C14-Z | 1.238(2) | 1.672(1) | 1.198(2) | 1.655(2) | 1.218(4) |
C14–C15 | 1.452(2) | 1.447(2) | 1.460(2) | 1.455(3) | 1.455(4) |
C15–C16 | 1.406(2) | 1.412(2) | 1.401(2) | 1.410(3) | 1.416(4) |
C16–N11 | 1.388(2) | 1.382(2) | 1.395(2) | 1.393(2) | 1.393(4) |
C16–N11–C12 | 117.1(1) | 117.3(1) | 117.5(1) | 121.9(2) | 122.6(2) |
N11–C12–Y13 | 123.3(1) | 122.4(1) | 125.5(1) | 125.5(2) | 126.7(2) |
C12–Y13–C14 | 123.5(1) | 124.3(1) | 121.1(1) | 104.9(1) | 103.4(1) |
Y13–C14–C15 | 114.7(1) | 114.7(1) | 115.2(1) | 126.5(2) | 118.4(2) |
ψ | 12.40(7) | 8.95(6) | 7.06(7) | 14.7(1) | 4.0(1) |
Fe–Cg1 | 1.6391(7) | 1.6457(6) | 1.6451(7) | 1.644(1) | 1.640(1) |
Fe–Cg2 | 1.6506(9) | 1.6597(7) | 1.6497(8) | 1.652(1) | 1.649(2) |
∠Cp1, Cp2 | 0.3(1) | 2.91(9) | 0.34(9) | 3.3(1) | 0.6(2) |
The crystal structures reveal practically planar arrangements for the entire heterocyclic units (10 atoms) and their attached CO/CS bonds. The maximum observed displacement from the mean planes of the bicyclic units of 0.042(1) and 0.045(1) Å are seen for N13 atoms in 2 and 3, respectively. These compounds also show the largest departure of the exo-cyclic O/S atoms from the plane of their parent heterocyclic moiety, the perpendicular distances from the mean ring plane being 0.082(1) Å for O in 2, and 0.071(1) Å from S in 3. The in-ring distances and angles (Table 1) clearly illustrate a distortion of the six-membered heterocyclic ring upon replacement of the oxygen and nitrogen atoms by the relatively larger sulphur.
The heterocyclic units and their bonding cyclopentadienyl ring in 2–6 are rotated from a mutually coplanar arrangement. The rotation varies both in the magnitude (see ψ in Table 1) and also in the sense. Whereas the CO bond in 4 is directed inwards the ferrocene moiety (i.e., towards the Fe atom), the CO and CS bonds in all other structurally characterised compounds point away from the ferrocene unit. These rotations not exceeding ca. 15° do not seem to destabilise the structures much by, e.g., a decreased conjugation (see DFT computations below), and can be accounted for by the steric factors at molecular level and also the crystal packing effects.
Finally, the ferrocene units in 2–6 are regular, showing only negligible tilting. The observed shortening of the Fe–Cg1 distances as compared to the matched Fe–Cg2 distances, which is relatively small (0.3–0.9%) but systematic and statistically significant, can be explained by a strengthening of the Fe ← Cp1 donation due to an electron density transfer from the ferrocene unit to the electron-withdrawing heterocyclic unit.
Only compounds 2 and 3 possess both NH protons as the typical H-bond donors and suitable H-bond acceptors. Not surprisingly, therefore, the molecules in their crystal structures associate via N–H⋯Z (Z = O, S) interactions giving rise to dimers around the crystallographic inversion centres (Fig. 7a). Probably due to its spatial proximity, the adjacent CH proton (H5) becomes involved in a similar interaction (C5–H5⋯O), too. The molecules further associate via offset π⋯π stacking interactions of the ferrocene C5H5 ring and the C6H4 ring within the heterocyclic moiety (Fig. 7b). The distances of the ring centroids and the interplanar angles are 3.839(1) Å and 11.08(9)° for 2, and 3.8559(9) Å and 5.93(8)° for 3. Since these π-interactions involve molecules related by elemental translation along the b-axis, they interconnect the H-bonded dimers into infinite columnar stacks.
Fig. 7 (a) View of the hydrogen-bonded dimer in the structure of 2 showing the hydrogen bonds as dashed lines. The H-bond parameters are as follows: N13–H1N⋯O, N13⋯O = 2.836(2) Å, angle at H1N = 173°; C5–H5⋯O, C5⋯O = 3.225(2) Å, angle at H5 = 136°. Compound 3 forms a similar dimer, the respective parameters being: N13–H1N⋯S, N13⋯S = 3.361(1) Å, angle at H1N = 161°; C5–H5⋯S, C5⋯S = 3.525(2) Å, angle at H5 = 132°. (b) π⋯π Stacking interactions of molecules of 2 leading to propagation of the molecular array along the crystallographic b-axis. |
In contrast, the molecules of 4–6, lacking the NH groups, associate predominantly via π⋯π interactions. In the case of 4 and 6, these interactions are formed between the C5H5 rings and the heterocyclic moieties in molecules related by elemental translation along the b-axis. A similar interaction dominates in the crystal assembly of 5, involving molecules related by the crystallographic inversion.
The molecular structure of compound 9 is presented in Fig. 8 together with selected geometric data. According to a search in the Cambridge Structural Database,21 compound 9 represents the only structurally characterised compound possessing an extended (acyclic) (C3N2)⋯N⋯N–C(O) linkage, where ⋯ stands for any bond. An additional search in Chemical Abstracts for a similar fragment bearing the 1,3-dimethyl-2-imidazolidinylidene moiety resulted in only a few compounds (<30)22 that were all formulated conservatively as the tautomeric 2-(1,3-dimethyl-2-imidazolidinylidene)hydrazides, (1,3-Me2C3N2)N–NHC(O)R.
Fig. 8 View of the molecular structure of 9 showing the hydrogen bonds as dotted lines. The displacement ellipsoids correspond to 30% probability. Selected distances and angles (in Å and deg): C1–C11 1.481(2), C11–O1 1.229(2), C11–N1 1.357(2), C22–C27 1.504(2), C27–O2 1.280(2), C27–N2 1.314(2), N2–N3 1.402(2), N3–C28 1.317(2), C28–N4 1.330(2), C28–N5 1.344(2), Fe–Cg1 1.6457(7), Fe–Cg2 1.6491(8), ∠Cp1, Cp2 1.6(1); O1–C11–N1 124.7(1), C22–C27–O2 121.4(1), C22–C27–N2 114.4(1), O2–C27–N2 124.1(1), C27–N2–N3 108.7(1), N2–N3–C28 123.5(1), N4–C28–N5 110.9(1). Note: The ring planes are defined as for 1. H-bond parameters: N1–H1N⋯O2, N1⋯O2 = 2.582(2) Å, angle at H1N = 147°; N3–H3N⋯O2, N3⋯O2 = 2.496(2) Å, angle at H3N = 110°. |
In our case, the good-quality diffraction data allowed to locate unambiguously the NH protons and thus to formulate compound 9 as a zwitterion (1,3-Me2C3N2)NH(+)–N(−)–C(O)RFc (RFc = 2-(ferrocenylcarboxamido)phenyl). Parameters describing the geometry of the C11(O1)N11 unit in 9 are very similar to those found in 1 and the precursor 7. On the other hand, the other amide moiety, N2(−)–C27O2, gains a partial N2(δ−)C27–O2(δ−) character, very likely due to an electron density transfer from the negatively charged N2, which in turn results in an elongation of the C27O2 bond and shortening of the C27–N2 bond. The C28N3(+) and N2–N3 bond lengths in 9 are similar to the respective distances in 4-chloro-N-(imidazolidin-2-ylidene)-1H-indazol-1-amine (N–N = 1.402(2) Å, NNC = 1.307(2) Å),23 whereas the N2–N3 separation in 9 is significantly longer than the N–N distances reported for 1,2-dibenzoylhydrazine (1.385(3) Å)24 and benzaldehyde benzoylhydrazone (1.368(7) Å).25
The extensive electronic delocalisation in the molecule of 9 corresponds with the unique all-planar geometry this compound assumes in the solid state. Taking the amide linker {C11, O1, N1} as the reference plane, the bonding cyclopentadienyl ring is found rotated at the dihedral angle of 8.2(2)°, while the entire extended pendant comprising as much as 18 atoms [N(1–5), O2 and C(21–32)], which are coplanar within ca. 0.07 Å, is rotated by 6.9(2)°. In addition to conjugation, the planar geometry and perhaps also the particular tautomeric form seem to be stabilised by intramolecular N–H⋯O contacts that can be detected in the structure albeit with unfavourably acute angles at the hydrogen atoms (see Fig. 8).
Fig. 9 Cyclic voltammograms of compounds 2 and 4 as recorded on a Pt disc electrode in 1,2-dichloroethane containing 0.1 M Bu4N[PF6] as the supporting electrolyte. Scan rate: 200 mV s−1. |
Compound | E°′ [V] | Compound | E°′ [V] |
---|---|---|---|
a The potentials were determined as an average of the peak potentials in cyclic voltammetry and are given relative to ferrocene/ferrocenium reference. Conditions: Pt-disc electrode, ca. 0.5 mM solutions in dry 1,2-dichloroethane containing 0.1 M Bu4N[PF6] as the supporting electrolyte. Scan rate: 200 mV s−1. | |||
1 | 0.21 | 5 | 0.27 |
2 | 0.22 | 6 | 0.26 |
3 | 0.26 | 7 | 0.21 |
4 | 0.25 |
All compounds display one electrochemically reversible redox transition in the anodic region, which is controlled by diffusion as indicated by the peak potentials increasing linearly with the square root of the scan rate (ip ∝ ν½). Only in the case of thione 3 the primary redox change is complicated by some follow-up processes affording other redox-active species and is associated with a pre-wave (see Fig. S1, ESI†). The observed redox changes, which can be attributed to the ferrocene/ferrocenium couple, occur in a rather narrow range of ca. 60 mV for the whole series and at potentials more positive than the ferrocene/ferrocenium reference (Table 2). The anodic shift corresponds with the overall electron-withdrawing character of the amide and heterocyclic moieties that can be expected to make any electron removal more difficult (cf. the Hammett’s σp constants for CONH2: 0.36, and 2-pyrimidinyl: 0.53).26
The discrepancy between φ calculated for an isolated molecule and the value determined from X-ray diffraction data was tentatively attributed to crystal packing effects. Indeed, the geometry obtained from DFT calculations (PBE/6-31G*) performed for a periodic structure with four molecules in the unit cell gave φ = 15°, which is in a perfect agreement with the experimental value. In this case, however, a slightly worse average relative error for the bond length and angles was obtained (ca. 1% for the bond lengths and 0.9% for the angles), which is in line with the well known, slightly poorer performance of the PBE functional (that was used to make the calculations of the crystal structure more economical) as compared with B3LYP.
Next, starting from the geometry optimised for an isolated molecule, we carried out a fully relaxed potential energy surface scan (DFT/B3LYP) for this torsion angle and found a second minimum at φ = –172° (Fig. 10). The potential energy difference between the minima is negligible (about 0.02 kJ mol−1) with respect to thermal energy kT at ambient conditions (ca. 2.5 kJ mol−1). The energy barriers separating these minima are about 16 and 22 kJ mol−1 high. Similar calculations using PBE functional gave practically identical results. The energy minima were found at φ ca. −4° and –174°, and the calculated energy barriers were ca. 17 and 24 kJ mol−1 (for a comparison of the two energy profiles, see Fig. S2, ESI†). Considering the height of the energy barriers, one can expect that the ferrocenyl and quinazolinone moieties undergo a fast rotation at room temperature in solution. This is in accordance with the 1H and 13C{1H} NMR spectra showing degenerate signals for the CH groups in α and β positions of the substituted cyclopentadienyl ring. It is also noteworthy that structures corresponding to the minima are very nearly mirror images of each other except for a rotation of the unsubstituted cyclopentadienyl ring.
Fig. 10 The dependence of the relative DFT (B3LYP/6-31G*) potential energy on the torsion angle N11–C12–C1–C2 (φ) for an isolated molecule. |
Furthermore, we investigated the electronic structure of 2. The highest occupied molecular orbital (HOMO; Fig. 11) was found to be contributed mainly from d-orbitals centred at the iron atom (68%). Since this contribution represents the only major entry (>10%) from one particular atom to HOMO, one may expect that the initial oxidation will occur there, which is in accordance with the observed electrochemical behaviour (vide supra). On the other hand, the lowest unoccupied molecular orbital (LUMO; Fig. 11) is more delocalized with major contributions from p-orbitals of the carbon atoms forming the quinazolinone ring (23% for the pivotal atom C18 and 11% for C12; for numbering, see Fig. 5), and further from d-orbitals of the iron atom (15%).
Fig. 11 B3LYP/6-31G* HOMO (top) and LUMO (bottom) orbitals of quinazolinone 2 shown at the ±0.05 a.u. level. |
The extensive conjugation in the molecule of 2 is clearly manifested in UV-vis spectra (Fig. 12). For instance, the low-energy absorption (d–d transitions) of 2 appears shifted to lower energies and is more intense (2: λmax = 458 nm, ε ≈ 220 M−1 cm−1) than for unsubstituted ferrocene (λmax ≈ 440 nm)27 and even ferrocenecarboxamide (FcCONH2; λmax ≈ 445 nm).
Fig. 12 UV-vis spectra of 2 (red) and ferrocenecarboxamide (blue) as recorded for ca. 5 × 10−4 M solutions in methanol (optical path: 1 cm). |
The calculated electrostatic potential (ESP) is shown in Fig. 13 with values coded by colours on a contour of total electron density drawn at the 0.002 a.u. level that roughly corresponds to the Van der Waals envelope.28 The domain with the most negative ESP (most attractive for a positive test charge) is located on the NH group of quinazolinone, whilst the domain with the most positive ESP is localized at the CO moiety. This corresponds with the solid-state packing based on an assembly of dimers formed via the N–H⋯OC hydrogen bonds. A supportive C–H⋯O interaction also detected in the crystal structure (C5–H5⋯O see Fig. 5) probably reflects a polarisation of this CH group by the negatively charged CO oxygen from the other molecule, which is indicated by an increased ESP at this H atom as compared to α-CH close to the heterocyclic N atom.
Fig. 13 Electrostatic potential (with magnitude <0.06 a.u.) mapped on a B3LYP/6-31G* electron density contour corresponding to 0.002 a.u. |
NMR spectra were recorded with a Varian UNITY Inova 400 MHz spectrometer (1H, 399.95 MHz; 13C, 100.58 MHz) at 298 K. Chemical shifts (δ/ppm) are given relative to internal tetramethylsilane. Infrared spectra were recorded in Nujol mulls with a Nicolet 6700 FTIR spectrometer in the range 400–4000 cm−1. UV-vis spectra were recorded with Unicam UV300 spectrometer using methanolic solution and 1 cm quartz cells. Electrospray ionization (ESI) mass spectra were obtained with an Esquire 3000 (Bruker; low resolution) or an LTQ Orbitrap XL instrument (Thermo Fisher Scientific; high resolution).
Electrochemical measurements were performed with a computer-controlled potentiostat μAUTOLAB III (Eco Chemie, Netherlands) at room temperature (ca. 23 °C) using a standard Metrohm three-electrode cell equipped by a platinum disc working electrode (2 mm diameter), platinum sheet auxiliary electrode, and a double-junction Ag/AgCl (3 M KCl) reference electrode. The compounds were dissolved in dry 1,2-dichloroethane (Sigma-Aldrich, absolute) to give solutions containing ca. 0.5 mM of the analyte and 0.1 M Bu4N[PF6] (Fluka, puriss. for electrochemistry) as the supporting electrolyte. The solutions were deaerated with argon before the measurement and then kept under an argon blanket. The redox potentials (accuracy ca. 5 mV) are given relative to a ferrocene/ferrocenium reference (E°′ = 0.46 V under the experiment conditions).
M.p. 161–163 °C (toluene). 1H NMR (CDCl3): δ 4.25 (s, 5H, C5H5), 4.43 (virtual t, J = 2.0 Hz, 2H, C5H4), 4.43 (virtual t, J = 1.9 Hz, 2H, C5H4), 5.70 (br s, 1H, NH2), 6.32 (br s, 1H, NH2), 7.07 (ddd, J = 1.2, 7.4, 7.9 Hz, 1H, C6H4), 7.53 (ddd, J = 0.4, 1.6, 7.4 Hz, 1H, C6H4), 7.57 (ddd, J = 0.3, 1.5, 7.9 Hz, 1H, C6H4), 8.78 (dd, J = 1.1, 8.6 Hz, 1H, C6H4), 11.72 (br s, 1H, NH). 13C{1H} NMR (CDCl3): δ 68.63 (CH C5H4), 69.97 (CH C5H5), 71.06 (CH C5H4), 76.48 (Cipso C5H4), 117.72 (CCipso C6H4), 121.23 (CH C6H4), 122.07 (CH C6H4), 127.36 (CH C6H4), 133.60 (CH C6H4), 140.95 (NCipso C6H4), 169.75 (CONH2), 171.50 (CONH). IR (Nujol): νmax 3380 m, 3320 m, 3169 m, amide I 1655 vs, 1624 s, 1595 s, 1579 s, amide II 1514 vs, 1448 vs, 1400 vs, 1308 vs, 1270 s, 1217 w, 1171 w, 1142 m, 1124 w, 1107 m, 1083 w, 1043 w, 1029 m, 1002 w, 950 w, 886 w, 848 m, 825 s, 760 s, 682 w, 641 m, 527 m, 496 m, 482 m, 464 w cm−1. MS (ESI+): m/z 371 ([M+Na]+). HR MS (ESI+) calc. For C18H16O2N2FeNa ([M+Na]+) 371.0453, found 371.0453. Anal. Calcd for C18H16N2O2Fe·0.15 CH2Cl2: C 60.40, H 4.55, N 7.76. Found: C 60.37, H 4.63, N 7.54.
M.p. 190 °C dec. (chloroform). 1H NMR (CDCl3): δ 4.25 (s, 5H, C5H5), 4.58 (virtual t, J = 1.9 Hz, 2H, C5H4), 5.24 (virtual t, J = 1.8 Hz, 2H, C5H4), 7.50 (ddd, J = 1.5, 6.8, 8.1 Hz, 1H, C6H4), 7.73 (ddd, J = 0.6, 1.5, 8.2 Hz, 1H, C6H4), 7.78 (ddd, J = 1.5, 6.7, 8.2 Hz, 1H, C6H4), 8.36 (ddd, J = 0.5, 1.5, 7.9 Hz, 1H, C6H4), 11.20 (br s, 1H, CONH). 13C{1H} NMR (CDCl3): δ 68.13 (CH C5H4), 70.03 (CH C5H5), 71.45 (CH C5H4), 76.06 (Cipso C5H4), 120.53 (CCipso C6H4), 125.92 (CH C6H4), 126.38 (CH C6H4), 127.35 (CH C6H4), 134.81 (CH C6H4), 149.92 (NCipso C6H4), 155.14 (NCN), 163.49 (CO). IR (Nujol): νmax 3178 w, νNH 3122 m, 3053 w, amide I 1660 vs, 1596 vs, amide II 1564 s, 1507 m, 1440 s, 1413 w, 1339 m, 1291 s, 1265 w, 1246 w, 1219 w, 1153 m, 1107 m, 1099 w, 1063 w, 1041 w, 1030 w, 1020 w, 1003 w, 957 m, 892 m, 872 w, 836 w, 818 s, 782 s, 729 w, 690 w, 623 w, 609 w, 551 m, 494 s, 484 s cm−1. MS (ESI+): m/z 331 ([M+H]+), 353 ([M+Na]+). Anal. Calcd for C18H14N2OFe·0.04 CHCl3: C 64.69, H 4.23, N 8.37. Found: C 64.71, H 4.14, N 8.22.
M.p. 170–172 °C (toluene). 1H NMR (CDCl3): δ 4.24 (s, 5H, C5H5), 4.58 (virtual t, J = 2.0 Hz, 2H, C5H4), 5.00 (virtual t, J = 1.9 Hz, 2H, C5H4), 7.48 (ddd, J = 1.3, 7.0, 8.2 Hz, 1H, C6H4), 7.69 (ddd, J = 0.5, 1.3, 8.2 Hz, 1H, C6H4), 7.78 (ddd, J = 1.5, 7.0, 8.3 Hz, 1H, C6H4), 8.68 (ddd, J = 0.5, 1.5, 8.1 Hz, 1H, C6H4), 10.38 (br s, 1H, CONH). 13C{1H} NMR (CDCl3): δ 67.82 (CH C5H4), 70.20 (CH C5H5), 71.76 (CH C5H4), 74.96 (Cipso C5H4), 127.36 (CH C6H4), 127.97 (CH C6H4), 127.97 (CCipso C6H4), 130.49 (CH C6H4), 135.72 (CH C6H4), 145.41 (NCipso C6H4), 153.28 (NCN), 186.29 (CS). IR (Nujol): νmax 3200 w, 3121 w, 1582 vs, 1566 vs, 1499 vs, 1481 vs, 1432 s, 1410 w, 1346 m, 1318 w, 1293 w, 1256 w, 1241 w, 1225 vs, 1205 vs, 1153 m, 1138 w, 1107 m, 1056 w, 1023 m, 1006 w, 925 m, 866 m, 859 w, 842 w, 824 m, 806 w, 769 w, 774 s, 767 s, 742 w, 724 w, 706 w, 592 w, 559 w, 483 s, 438 m cm−1. MS (ESI+): m/z 369 ([M+Na]+). Anal. Calcd for C18H14N2SFe: C 62.44, H 4.08, N 8.09. Found: C 62.28, H 4.08, N 8.04.
M.p. 162–164 °C (heptane). 1H NMR (CDCl3): δ 4.23 (s, 5H, C5H5), 4.55 (virtual t, J = 1.9 Hz, 2H, C5H4), 5.08 (virtual t, J = 1.9 Hz, 2H, C5H4), 7.46 (ddd, J = 1.2, 7.3, 7.8 Hz, 1H, C6H4), 7.57 (ddd, J = 0.6, 1.2, 8.2 Hz, 1H, C6H4), 7.77 (ddd, J = 1.6, 7.3, 8.1 Hz, 1H, C6H4), 8.20 (ddd, J = 0.6, 1.6, 7.9 Hz, 1H, C6H4). 13C{1H} NMR (CDCl3): δ 69.26 (CH C5H4), 70.06 (CH C5H5), 71.86 (CH C5H4), 73.12 (Cipso C5H4), 116.47 (CCipso C6H4), 126.39 (CH C6H4), 127.20 (CH C6H4), 128.64 (CH C6H4), 136.52 (CH C6H4), 147.60 (NCipso C6H4), 159.97 (NCO), 162.46 (OCO). IR (Nujol): νmax 1760 vs, 1708 w, 1620 vs, 1599 vs, 1570 s, 1474 vs, 1447 vs, 1409 m, 1395 w, 1351 w, 1329 w, 1318 m, 1307 m, 1262 vs, 1224 w, 1205 w, 1162 w, 1108 w, 1095 s, 1056 w, 1045 w, 1036 m, 1024 m, 1008 vs, 968 w, 936 m, 892 m, 842 w, 816 m, 794 w, 779 vs, 735 w, 726 w, 690 m, 629 m, 500 s, 485 s cm−1. MS (ESI+): m/z 332 ([M+H]+), 354 ([M+Na]+). Anal. Calcd for C18H13NO2Fe: C 65.28, H 3.96, N 4.23. Found: C 65.11, H 4.00, N 4.11.
M.p. 173–175 °C (chloroform). 1H NMR (CDCl3): δ 4.26 (s, 5H, C5H5), 4.44 (virtual t, J = 1.8 Hz, 2H, C5H4), 4.90 (virtual t, J = 1.9 Hz, 2H, C5H4), 7.07 (ddd, J = 1.2, 7.4, 7.9 Hz, 1H, C6H4), 7.47 (ddd, J = 0.4, 1.6, 7.9 Hz, 1H, C6H4), 7.52 (ddd, J = 0.4, 1.6, 7.4 Hz, 1H, C6H4), 7.56 (br s, 1H, NH), 8.74 (dd, J = 1.2, 8.4 Hz, 1H, C6H4); signals due to hydrazide protons were not found. 13C{1H} NMR (CDCl3): δ 68.61 (CH C5H4), 69.98 (CH C5H5), 71.09 (CH C5H4), 76.41 (Cipso C5H4), 117.70 (CCipso C6H4), 121.33 (CH C6H4), 122.36 (CH C6H4), 126.23 (CH C6H4), 133.28 (CH C6H4), 140.20 (NCipso C6H4), 169.61 (CONH), 170.04 (CONH). IR (Nujol): νmax 3393 m, 3319 s, 3206 w, 3111 w, 3075 w, amide I 1637 vs, 1615 m, 1594 vs, amide II 1522 vs, 1445 vs, 1339 m, 1315 s, 1299 s, 1268 m, 1240 w, 1220 w, 1163 w, 1148 m, 1106 w, 1078 w, 1045 w, 1033 w, 1006 w, 945 s, 830 s, 756 s, 736 w, 669 w, 558 w, 534 m, 500 s, 486 m cm−1. MS (ESI+): m/z 386 ([M+Na]+). Anal. Calc. For C18H17N3O2Fe: C 59.52, H 4.72, N 11.57. Found: C 59.23, H 4.73, N 11.32.
The diffraction data (±h ±k ±l; θmax range 27.0–27.5°, data completeness ≥99.8%) were collected with an Apex II CCD diffractometer (Bruker) equipped with a Cryostream Cooler (Oxford Cryosystems) using graphite monochromated MoKα radiation (λ = 0.71073 Å). The data were corrected for absorption using the methods incorporated in the diffractometer software.
The structures were solved by the direct methods (SHELXS9731) and refined by full-matrix least squares based on F2 (SHELXL9731). All non-hydrogen atoms were refined with anisotropic displacement parameters. The NH and OH hydrogens were identified on the difference density maps and refined as riding atoms with Uiso assigned to 1.2 Ueq of their bonding atoms. Other hydrogen atoms were included in their calculated positions and refined similarly (riding model). Relevant crystallographic data and structure refinement parameters are available as (Table S2, ESI†).
The geometric data and structural drawings were obtained with a recent version of the PLATON program.32 The numerical values were rounded with respect to their estimated deviations (ESDs) given to one decimal place. Parameters relating to atoms located in constrained positions are given without ESDs.
Similar calculations were done also for the crystal structure using periodic boundary conditions (PBC). As the unit cell contains four molecules, a computationally less intensive method was used, namely a functional developed by Perdew, Burke and Ernzerhof (PBE)36 with standard 6-31G* basis set. The calculated geometric data (distances and angles of the non-hydrogen atoms) are available as (Table S1, ESI†).
Footnote |
† Electronic supplementary information (ESI) available: Cyclic voltammogram of 3 (Fig. S1), a comparison of calculated and experimentally determined bond lengths and angles for 2 (Table S1), comparison of energy profiles calculated for 2 at two different levels of theory (Fig. S2), and summary of the crystallographic data (Table S2). CCDC 922133–922140. For crystallographic data in CIF or other electronic format see DOI: 10.1039/c3nj00182b |
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2013 |