Introduction

The chemistry of aminothiazoles and their derivatives has attracted the attention of chemists, since they exhibit important biological activity in medicinal chemistry [1], such as antibiotic, anti-inflammatory, anthelmintic or fungicidal properties [2–4]. 2-Aminothiazoles are known mainly as biologically active compounds with a broad range of activities and as intermediates in the synthesis of antibiotics, well known sulfa drugs, and some dyes [5, 6]. In addition, it has been shown that 3-substituted cyclobutane carboxylic acid derivatives exhibit anti-inflammatory and antidepressant activities [7], and also liquid crystal properties [8].

The gauge-including atomic orbital (GIAO) [9, 10] method is one of the most common approaches for calculating nuclear magnetic shielding tensors. It has been shown to provide results that are often more accurate than those calculated with other approaches, at the same basis set size [11].

In this study, we present results of a detailed investigation of the synthesis and structural characterization of 5-(4-aminophenyl)-4-(3-methyl-3-phenylcyclobutyl)thiazol-2-amine using single crystal X-ray, IR, NMR, and quantum chemical methods. GIAO 1H- and 13C-NMR chemical shifts of the title compound in the ground state have been calculated by using the Hartree–Fock (HF) and DFT (B3LYP) methods with 6–31G(d) basis set. These calculations are valuable for providing insight into molecular parameters and the vibration and NMR spectra.

Experimental

Synthesis

All chemicals were of reagent grade and used as commercially purchased without further purification. Melting point was determined by Gallenkamp melting point apparatus. IR spectra of the compound were recorded in the range of 4000−450 cm−1 region with a Mattson 1000 FT-infrared spectrometer using KBr pellets. The 1H-, and 13C-nuclear magnetic resonance spectra were recorded on a Varian-Mercury 400 MHz spectrometer. Synthesis of the compound was performed with the minor modification of literature [12] method as follows (see Scheme 1). A mixture of 2.2271 g (10 mmol) 1-phenyl-1-methyl-3-(2-chloro-1-oxoethyl)cyclobutane and 4-phenylthiosemicarbazide (1.6723 g, 10 mmol) in 50 ml absolute alcohol was refluxed with continuous stirring. The course of the reaction was monitored by IR spectroscopy. The product was formed over about half an hour period. After cooling to room temperature, the solution was then made alkaline with an aqueous solution of NH3 (5%) and light brown solid separated out. The brown colored shiny crystals which are suitable for X-ray analysis was obtained by the crystallization from ethanol (yield 81%, m.p. 528 K). Characteristic 1H NMR (DMSO-d6, δ ppm): 1.40 (s, 3H, CH3), 2.27 (dd, j 1  = 8.4, j 2  = 10.3, 2H, –CH2–), 2.60 (m, 2H, –CH2–), 3.58 (q, j = 9.1 Hz, 1H, >CH–), 5.20 (s, 2H, –NH2), 6.60 (d, j = 8.05 Hz, 2H, aromatics), 6.79 (s, 2H, –NH2), 6.95 (d, j = 8.05, 2H, aromatics), 7.12-7.14 (m, 3H, aromatics), 7.28 (t, j = 7.68 Hz, 2H, aromatics); Characteristic 13C NMR (DMSO-d6, δ ppm): 165.78 (C1), 114.67 (C2), 153.30 (C3), 28.80 (C4), 41.07 (C5), 38.87 (C6), 30.80 (C8), 148.42 (C9), 120.34 (C10), 130.19 (C11), 125.32 (C12), 125.80 (C15), 128.81 (C16), 119.60 (C17), 148.06 (C18).

Scheme 1
scheme 1

Synthesis scheme of the title compound

Full size image

Crystal data for the title compound

CCDC 676830, C20H21N3S, M w = 335.46, monoclinic, space group P21/c; Z = 4, a = 9.4350(5), b = 11.2796(6), c = 18.4170(8) Ǻ, α = 90, β = 113.378(3), γ = 90°; V = 1799.09(16) Å3, F(000) = 712, D x = 1.238 g/cm3. Full crystallographic data are available as supplementary material.

Computational procedure

The molecular structure of the title compound in the ground state (in vacuo) is optimized by Hartree–Fock (HF) and DFT(B3LYP) [13, 14] with the 6-31G(d) [15] basis set. For modeling, the initial guess of the title compound was first obtained from the X-ray coordinates. Then, vibrational frequencies for the optimized molecular structures of the title compound are calculated with these methods and then scaled by 0.8929 and 0.9613 [16], respectively. The geometry of the title compound, together with that of tetramethylsilane (TMS) is fully optimized. 1H- and 13C-NMR chemical shifts are calculated within GIAO approach [9, 10] applying the same methods and the basis set as that used for geometry optimization. The 1H- and 13C-NMR chemical shifts are converted to the TMS scale by subtracting the calculated absolute chemical shielding of TMS whose values are 32.90 and 202.14 ppm for HF/6-31G(d), and 32.17 and 190.11 ppm for B3LYP/6-31G(d), respectively. Molecular geometry is restricted, and all the calculations are performed without specifying any symmetry for the title molecule by using GaussView Molecular Visualization Program [17, 18] and Gaussian 03 Program package [19]. The effect of solvent on the theoretical NMR parameters was included using the default model provided by Gaussian 03. Dimethylsulfoxide (DMSO) was used as solvent. To identify low energy conformations, two selected degrees of torsional freedom, T(S1–C2–C15–C16) and T(C5–C6–C9–C10), were varied from −180° to +180° in steps of 10°, and the molecular energy profiles were obtained at the semi-empirical AM1 level.

Results and discussion

IR spectroscopy

FT-IR spectra are obtained in KBr discs using a Mattson 1000 FT-IR spectrometer, and shown in Fig. 1. The band assigned to ν(C=S) vibration of thiosemicarbazide observed at 966 cm−1 disappeared in IR spectrum of thiazol compound because of the formation of thioether (C−S−C). In addition, the absence of any band ν(C=O) in the 1800-1700 cm−1 region of the IR spectrum of the compound signify that carbonyl group of 1-phenyl-1-methyl-3-(2-chloro-1-oxoethyl)cyclobutane reacted with amino group of thiosemicarbazide. The bands at 3443, 3362 and 1635 cm−1 observed in the IR spectrum are assigned to asymmetric, symmetric stretchings and N−H intra-planar bending of NH2 group.

Fig. 1
figure 1

FT-IR spectrum of the title compound

Full size image

It is well known, that the calculated HF and DFT ‘raw’ or ‘non-scale’ harmonic frequencies could significantly overestimate experimental values due to lack of electron correlation, insufficient basis sets and anharmonicity. Much effort has been devoted to accurately reproducing experimental frequencies in theoretical calculations. The Hartree–Fock calculated results are usually more overestimated than the corresponding DFT ones [20]. To compare these, we have calculated the theoretical vibrational spectra of the title compound using both HF and B3LYP method with 6-31G(d) basis set. Frequency calculations at the same levels of theory revealed no imaginary frequencies, indicating that an optimal geometry at these levels of approximation was found for the title compound. We have compared our calculation of the title compound with their experimental results. Theoretical and experimental results of the title compound are shown in Table 1. The vibrational bands assignments have been made by using Gauss-View Molecular Visualization program [17]. To make comparison with experiment, we present correlation graphics in Fig. 2 based on the calculations. As we can seen from correlation graphic in Fig. 2 experimental fundamentals are in better agreement with the scaled fundamentals and are found to have a better correlation for HF than B3LYP.

Table 1 Comparison of the observed and calculated vibrational spectra of the title compound
Full size table
Fig. 2
figure 2

Correlation graphics of calculated and experimental frequencies of the title compound

Full size image

As can be seen from Table 1, the NH2 asymmetric and symmetric stretch bands have been calculated at 3474–3386 cm−1 for HF and 3508–3409 cm−1 for B3LYP with using 6-31G(d) basis set. The experimental C=C and C=N stretch bands were observed at 1559 and 1531 cm−1, that have been calculated with HF and B3LYP at 1602–1547 and 1542–1522 cm−1, respectively. As can be seen from Table 1, there is also good agreement between experimental and theoretical vibration data for the others.

NMR spectroscopy

GIAO 1H and 13C chemical shift calculations have been carried out using the HF and B3LYP methods with 6–31G(d) basis set for the optimized geometry. The results of these calculations are tabulated in Table 2. Since experimental 1H chemical shift values were not available for individual hydrogen, we have presented the average values for NH2, CH2 and CH3 hydrogen atoms. The signals assigned hydrazinic NH of the thiosemicarbazide in the 9-12 ppm range disappear in the 1H-NMR spectrum of the thiazol compound [21, 22]. The singlets observed at 5.20 and 6.79 ppm are assigned to N(3)H2 and N(1)H2 groups, respectively, that have been calculated with HF and B3LYP at 3.73–4.74 and 3.31–4.15 ppm. The -CH2- signals of the cyclobutane are observed at 2.27 and 2.60 ppm. The C-H signals belonging to p-aminophenyl group are shielded at 6.60 and 6.95 ppm. However, the C-H signals of phenyl adjacent to the cyclobutane are deshielded at 7.12-7.14 (3H) and 7.28 (2H) ppm. The singlet signal of the methyl group is observed at 1.40 ppm.

Table 2 Theoretical and experimental 1H and 13C isotropic chemical shifts (with respect to TMS, all values in ppm) for the title compound
Full size table

13C-NMR spectra of the thiazol compound show the signal at 165.78 ppm due to C atom next to N(1)H2. This signal has been calculated as 182.01 ppm for HF and 161.99 ppm for B3LYP. The signals at 114.67 and 153.30 ppm are assigned to C atoms next to sulfur and nitrogen atom of thiazol ring, respectively. While the C atoms of methylene group belonging to the cyclobutane ring are observed at 41.07 ppm, methine C atom appeared at 28.80 ppm. The signal at 38.87 ppm is related to the last C atom of the cyclobutan ring.

Comparing calculational and the experimental data, we studied the relativity between the calculation and the experiments, and obtained that the linear function formula is y = 1.02406x – 0.4658 for HF; where R2 is 0.99547, and y = 0.95535x + 0.67306 for B3LYP; where R2 is 0.99493. According to these results, it is seen that, the results of HF method have shown better fit to experimental ones than B3LYP in evaluating 1H and 13C chemical shifts.

Description of the structure

The title compound, a DIAMOND [23] view of which is shown in Fig. 3, crystallizes in the monoclinic space group P21/c with four molecules in the unit cell. The asymmetric unit in the crystal structure contains only one molecule.

Fig. 3
figure 3

The molecular structure of the title molecule, showing the atom-numbering scheme. Displacement ellipsoids are drawn at the 50% probability level and H atoms are shown as small spheres of arbitrary radii

Full size image

The title molecule is composed of a central thiazole ring, with an amino group connected to the 2-position of the ring, a (3-methyl-3-phenyl)cyclobutyl group in the 4-position, and a p-aminophenyl group in the 5-position. The thiazole ring is planar with a maximum deviation of -0.0055(10) Å for atom N2. In the crystal structure, the benzene ring and 5-(4-aminophenyl)thiazol-2-amine group are in cis positions with respect to the cyclobutane ring. The dihedral angles between the thiazole plane A (S1/N2/C1−C3), the benzene plane B (C9−C14), the cyclobutane plane C (C4−C7) and the other benzene plane D (C15−C20) are 88.68(6)° (A/B), 83.60(8)° (A/C), 43.75(4)° (A/D), 45.42(10)° (B/C), 47.99(8)° (B/D) and 67.89(8)° (C/D).

Although close to being planar, the cyclobutane ring is puckered. The C7/C4/C5 plane forms a dihedral angle of 18.92(15)° with the C5/C6/C7 plane. This value is smaller than those in the literatures; 23.5 [24], 25.74(6) [25], and 19.26(17)° [26]. However, when the bond lengths and angles of the cylobutane ring in the title compound are compared with these, it is seen that there are no significant differences.

There are two obviously different C−N bond distances in the thiazole ring, viz. N2−C1 and N2−C3. The C2−C3 bond distance is 1.350(2) Å, characterizing a C=C double bond. The N1−C1 bond distance [1.348(2) Å] is shorter than the single bond N2−C3 [1.3972(19) Å], but longer than that of the double bond N2−C1 [1.300(2) Å], which can be attributed to the conjugation of the electrons of atom N1 with atoms C1 and N2. In addition, the S1−C1 bond distance [1.7338(17) Å], which is shorter than the S1−C2 bond [1.7503(16) Å], being both much shorter than the lower-quartile value [1.809 Å;[27]] for single bonds between three-connected C and two-connected S atoms, arises from the conjugation of the electrons of atom S1 with atoms C1 and N2.

The crystal structure does not exhibit intramolecular or π−π stacking (face-to-face) interactions. There are, however, one N−H···N and two X−H···Cg(π-ring) (edge-to-face) intermolecular interactions, details of which are given in Table 3. Amino atom N1 in the molecule at (x, y, z) acts as hydrogen-bond donor, via atom H1A, to ring atom N2 in the molecule at (1−x, 1−y, 1−z), so generating by inversion a centrosymmetric dimer, centered at (1/2, 1/2, 1/2) and characterized by an R 22 (8) motif [28] (Fig. 4). The same amino atom N1 at (x, y, z) forms a N—H···Cg(π-ring) contact, this time via atom H1B, with the centroid of the C9−C14 ring [fractional centroid coordinates: 0.8017(2), 0.8569(3), 0.6145(2)] of the molecule at (1 − x, 1 − y, 1 − z). In addition, atom C14 at (x, y, z) forms a C—H···Cg(π-ring) contact, via atom H14, with the centroid of the C15−C20 ring [fractional centroid coordinates: -0.2799(3), 0.3161(2), 0.3606(2)] of the molecule at (−x, −0.5 + y, +0.5 − z), which links the R 22 (8) dimers into a molecular chain.

Table 3 Hydrogen bonding geometry (Å, °) for the title compound
Full size table
Fig. 4
figure 4

Part of the crystal structure of the title molecule, showing the formation of a chain of centrosymmetric R 22 (8) dimers. For clarity, only H atoms involved in hydrogen bonding have been included

Full size image

Theoretical structures

Some selected geometric parameters experimentally obtained and theoretically calculated by HF and B3LYP with 6-31G(d) as the basis set are listed in Table 4. When the X-ray structure of the title compound is compared with its optimized counterparts (see Fig. 5), conformational discrepancies are observed between them. The dihedral angles between A, B, C and D planes are calculated at 87.45° (A/B), 86.56° (A/C), 62.69° (A/D), 35.89° (B/C), 38.02° (B/D) and 67.25° (C/D) for HF, and at 86.28° (A/B), 83.50° (A/C), 45.89° (A/D), 36.38° (B/C), 47.55° (B/D) and 67.59° (C/D) for B3LYP. According to X-ray study, dihedral angle between the C7/C4/C5 and C5/C6/C7 planes is 18.92(15)°, whereas the dihedral angle has been calculated at 25.53° for HF and at 24.82° for B3LYP.

Table 4 Optimized and experimental geometries of the title compound in the ground state
Full size table
Fig. 5
figure 5

Atom-by-atom superimposition of the structures calculated (red) [A = HF; B = B3LYP] over the X-ray structure (black) for the title compound. Hydrogen atoms omitted for clarity. The RMS overlay error of 0.308 and 0.244 Å does not include hydrogen atoms

Full size image

For the optimized geometric parameters, various methods including HF method estimates some bond lengths well to some extent [29–31]. We noted that the experimental results belong to solid phase and theoretical calculations belong to gaseous phase. In the solid state, the existence of the crystal field along with the intermolecular interactions have connected the molecules together, which result in the differences of bond parameters between the calculated and experimental values. It is well known that DFT optimized bond lengths are usually longer and more accurate than HF, due to inclusion of electron correlation. However, according to our calculations, HF method correlates well for the bond length compared with the other method (Table 4). The largest difference between experimental and calculated HF bond length is about 0.026 Å. The B3LYP method leads to geometric parameters (angles), which are much closer to experimental data. This pattern was not found for bond length, as can be seen from Table 4, whereas in the case of B3LYP method, the biggest difference between calculated and experimental values of bond lengths was 0.033 Å. The bond angles provided by B3LYP method is the closest to the experimental values (Table 4). The largest difference is about 1.57° in the case of B3LYP method, while this difference is 1.90° for HF. The same trend was also observed in torsion angles. The largest differences are 20.9° and 6.1° for HF and B3LYP, respectively. As a result, the optimized bond lengths obtained by HF method, and bond angles and torsion angles by DFT (B3LYP) method show the best agreement with the experimental values.

Based on HF/6-31G(d) and B3LYP/6-31G(d) optimized geometry, the total energy of the title compound has been calculated by these two methods, which are −1330.434223 and −1337.204818 a.u., respectively. In order to define the preferential position of the aminobenzene fragment with respect to thiazole ring, and the preferential position of the benzene ring with respect to the cyclobutane ring, respectively, a preliminary search of low energy structures was performed using AM1 computations as a function of the selected degrees of torsional freedom T(S1–C2–C15–C16) and T(C5–C6–C9–C10). The respective values of the selected degrees of torsional freedom, T(S1–C2–C15–C16) and T(C5–C6–C9–C10), are -42.30(19) and -38.8(2)° in X-ray structure, whereas the corresponding values in optimized geometries are -62.8068 and -40.4267° for HF, and -46.0387 and -39.7911° for B3LYP. Molecular energy profiles with respect to rotations about the selected torsion angles are presented in Fig. 6. According to the results, the low energy domains for T(S1–C2–C15–C16) are located at -50 and 130° having energy of 92.165 and 92.142 kcal mol−1, respectively, while they are located at -40 and 140° having energy of 92.156 and 92.158 kcal mol−1, respectively, for T(C5–C6–C9–C10). Energy difference between the most favorable and unfavorable conformer, which arises from rotational potential barrier calculated with respect to the two selected torsion angles, is calculated as 2.926 kcal mol−1 when both selected degrees of torsional freedom are considered.

Fig. 6
figure 6

Molecular energy profiles of the optimized counterpart of the title compound against the selected degrees of torsional freedom

Full size image

The molecular energy can be divided into bonded and non-bonded contributions. The bonded energy is considered to be independent of torsional angle changes and therefore vanished when relative conformer energies are calculated. The non-bonded energy is further separated into torsional steric and electrostatic terms [32]. Since the title compound contains no intramolecular hydrogen bond, it can be deduced from the computational results that the most stable conformer of the title compound is principally determined by the non-bonded torsional energy term affected by packing of the molecules.

The Mulliken atomic charges and natural population analysis (NPA) atomic charges for the non-H atoms of the title compound calculated at HF/6-31G(d) and B3LYP/6-31G(d) levels are presented in Table 5. The calculated results show that the two amino N atoms have bigger negative charges along with their suitable spatial configuration, which result in that they are the potential sites to react with the metallic cores. Namely, the title compound can act as multidentate ligand to bind one or two metal centers, so resulting in interesting metal complexes with different coordination geometries.

Table 5 Atomic charges (e) of the title compound at HF/6-31G(d) and B3LYP/6-31G(d) levels
Full size table

Figure 7 shows the distributions and energy levels of the HOMO − 1, HOMO, LUMO and LUMO + 1 orbitals computed at the B3LYP/6-31G(d) level for the title compound. Both the highest occupied molecular orbitals (HOMOs) and the lowest-lying unoccupied molecular orbitals (LUMOs) are mainly localized on the aminobenzene and aminothiazole fragments indicating that the HOMO-LUMO are mostly the π-antibonding type orbitals, and are perpendicularly oriented to each other. The value of the energy separation between the HOMO and LUMO is 4.626 eV and this large energy gap indicates that the title structure is very stable.

Fig. 7
figure 7

Molecular orbital surfaces and energy levels given in parentheses for the HOMO – 1, HOMO, LUMO and LUMO + 1 of the title compound computed at B3LYP/6-31G(d) level

Full size image

Conclusions

As a result, X-ray structure is slightly different from its optimized counterparts, and the crystal structure is stabilized by N−H···N type hydrogen bonds and edge-to-face interactions. Crystal packing of the title compound is dominated only by intermolecular interactions formed during preparation or crystallization. These hydrogen bonds supply leading contribution to the stability and to the order of the crystal structure, and are presumably responsible for the discrepancies between the X-ray and optimized structures of the title compound. For the geometric parameters, the results of B3LYP method has shown a better fit to experimental ones than HF in evaluating geometrical parameters. However, the HF method seems to be more appropriate than B3LYP method for the calculation of vibrational frequencies and chemical shifts.