2.1 Cycloaddition of diaryldiazomethanes Ar₂CN₂ across Cl₃C–CHN–CO₂Et

In analogy to the protocol using diphenyldiazomethane as a dipole, the cyloaddition of di(p-chlorophenyl)diazomethane, (p-chlorophenyl)phenyldiazomethane and di(p-methoxyphenyl)-diazomethane produces at 20 °C in toluene as solvent, the expected Δ³-1,2,4-triazolines 2b–d as colourless powders (see Scheme 1). Subsequent thermolysis of these heterocyclic compounds in toluene at 60° leads to (most probably via the transient azomethine ylides 3 [21]) the corresponding 1,1-diaryl-4,4-dichloro-2-azabutadienes 4b–d as pale-yellow stable solids. However, it is not necessary to isolate first the heterocyclic intermediates 2. If the imine 1 is treated with Ar₂CN₂ in warm toluene, in situ transformation affords straightforwardly the 2-azadienes 4. The crude residue of 4c is contaminated by 10–15% of tetrakis(p-chlorophenyl)diazine (p-ClC₆H₄)₂CN–NC(p-ClC₆H₄)₂ as a minor product, which can be separated by fractional crystallisation from hot ethanol [22].

As observed in the proton NMR spectrum of 4a, the vinylic hydrogen resonances of 4c and 4d are quite low-field shifted and appear as singlets at δ 7.07. Furthermore, two distinct singlets due to the methoxy groups are found at δ 3.83 and δ 3.87 in the spectrum of 4d. In the case of 4b, two singlets in a 2:1 ratio are detected at δ 7.03 and δ 7.01 in CDCl₃ solution, respectively. Two singlets, attributed to the CN resonance, are also observed in the ¹³C{¹H} spectrum at δ 167.0 and 166.9 in an approximate 2:1 ratio. Initially, we rationalized this finding by formation of two stereoisomers resulting from different approaches of the dissymmetric diaryldiazoalkane across the CN bond of imine 1 in the transition state. A careful examination under an optical microscope confirmed that the morphology of all crystals of 4b within this batch was quite homogenous, the melting point being sharp between 64–66 °C. Surprisingly, dissolving a single crystal in CDCl₃ again gave rise to two sets of signals in the 2:1 ratio. A survey of the literature revealed that this phenomenon is most probably due to a hindered inversion at the imine nitrogen. For example, Jennings et al. investigated in detail a series of Schiff bases of type p-XC₆H₄(C₆H₅)CNR reminiscent to 4b, illustrated in Scheme 2 [23].

The activation barrier for this imine inversion has been estimated by means of dynamic NMR spectroscopy to be of some 30 kcal/mol, the value varying in function of substituent X. We suppose, that in the case of 4b, a similar inversion process may operate giving rise to a co-existence of E and Z invertomers in solution. Since this phenomenon is beyond the topic of this article, we did not undertake NMR studies at variable temperature.

2.2 Reactivity of 4a and 4c towards NaSAr, NaSC₆H₁₁, NaSMe and NaSⁿBu

Three strategies have been developed in the past to synthesise 2-azabuta-1,3-dienes with thioether substituents: (E)-1-methylthio-2-azabuta-1-3-diene-4-carbo-nitriles like A and B were prepared by the addition of thioamides to methoxymethylene compounds or ketene dithioacetals and subsequent methylation, [24] whilst 3-aza-2-(dimethylamino-4-(methylthio)-2,4-hexadiene (C) was obtained from treatment of N-(thiopropionyl)acetamidine with CH₃I and subsequent deprotonation of the resulting N-ylidene amidinium salt [25]. The crystallographically characterized dithioether D and related compounds were obtained by alkylation of the dithiolates Na₂[S₂CC(R)–NCPh₂] (R = CO₂Me, CN) with CH₃I or Br(CH₂)_nBr (n = 2, 3) [26].

In precedent papers we have demonstrated that 4a reacts under mild conditions in dry DMF with an excess of NaSR leading exclusively to the dithioether derivatives of type [Ph₂CN–C(H)C(SR)₂] (R = Ph, p-Tolyl, iPr), consistent with formal substitution of the two chloro-substituents on the vinylic C3 atom by SR⁻. Their crystal structures have been already published [14,15,19a]. An addition-elimination mechanism, initiated by nucleophilic addition to the C3 atom, elimination of chloride followed by addition of a second thiolate on the C3 atom and termination by dissociation of the remaining chloride has been suggested [14]. A mono-substituted thioether intermediate, [Ph₂CN–C(H)C(Cl)(SPh)], in which the SR group is cis-situated relative to the imine nitrogen, has been structurally characterized [19b]. Note, that for vinylic substitutions also other mechanisms such as S_RN1-type single-electron transfer may operate [27–29].

With the exception of 5a, 5b and 5d, which were isolated as viscous oils, all other derivatives (Scheme 3) formed yellow solids. An X-ray structure analysis has been carried out on a crystal of 5i grown from EtOH (Fig. 1). The s-trans conformation of the azabutadiene chain found in precursor 4a and in 5f, 5g and 5j is also observed in the solid-state structure of 5i. Metric parameters and crystallographic data are gathered in Tables 1 and 2. All in all, they resemble much to those of [Ph₂CN–C(H)C(SPh)₂] 5j [14] and [Ph₂CN–C(H)C(p-STol)₂] 5g [15] and deserve no further comment.

In contrast to the facile preparation of compounds [Ph₂CN–C(H)C(SR)₂] 5 bearing a Ph₂CN imine function, the synthesis of derivatives [(p-ClC₆H₄)₂CN–C(H)C(SR)₂] 5e and 5k was less straightforward. [(p-ClC₆H₄)₂CN–C(H)C(SMe)₂] 5e could be isolated only in a quite modest yield of 20% as a yellow solid. Furthermore, chromatography of the crude residue of 5e provided a second fraction containing the tris(thioether) [(MeSC₆H₄)(p-ClC₆H₄)CN–C(H)C(SMe)₂], albeit in a low yield of 3%. A high-resolution mass spectrum of this bright-yellow solid displayed a peak at m/z 379.66 (Fig. S1), whose simulation matches well with a composition of C₁₈H₁₈ClNS₃, corroborating the formation of a triply SMe-substituted azabutadiene.

During the synthesis of the compound [(p-ClC₆H₄)₂CN–C(H)C(SⁱPr)₂] 5k, spectroscopic analysis did not allow to elucidate its structure unequivocally. Only in the case of [(p-ClC₆H₄)₂CN–C(H)C(STol-p)₂] 5h, a high-yield synthesis could be performed. Since aromatic halides are known to undergo nucleophilic substitutions of the S_NAr type with highly nucleophilic aliphatic thiolates NaSMe and NaSⁱPr, we reacted 4c with a 10-fold excess of NaSⁱPr in DMF under gentle heating. Indeed, after workup, we succeeded in isolating the azadienic tetrakisthioether [(ⁱPrSC₆H₄)₂CN–C(H)C(SⁱPr)₂] 6 with 48% yield in the form of a yellow solid (Scheme 3) [30,31].

2.3 Reactivity of 4a and 4c towards sodium tert-butylthiolate

We extended the reactivity study of 4a,c towards thiolates also on the reaction with NaS^tBu in DMF as solvent. Surprisingly, the spectroscopic features of the substitution products 7, which were isolated after workup as orange-red stable crystalline solids, diverged much from those of 5. For example, the vinylic H resonance of 7a and 7b were detected at δ 5.98 and 6.01 ppm in the ¹H NMR spectrum, whereas the vinylic H atom of 5 gives rise to a singlet at a much lower field of about 7.1 ppm. The ^tBu groups appear as a set of close, but distinct singlets for 7a at δ 1.39 and 1.40 ppm (δ 1.35 and 1.43 for 7b). In the case of 7a, comparison of the experimental chemical shift (5.98 ppm) with the values of the chemical shifts at 5.94 (trans or E with respect to both ^tBu substituents) and 6.61 (cis or Z) ppm calculated by the incremental method of Pascual, [32] allowed us to assign E-stereochemistry around the olefinic double bond. This E-stereochemistry was furthermore ascertained by X-ray crystallography.

X-ray quality crystals of 7a,b were grown from EtOH and the molecular structures are depicted in Fig. 2. The outstanding information got from concerns the position of the S^tBu substituents. One S^tBu is found, as in all our thioether compounds 5, at the C3 atom and in the cis-position relative to the imine nitrogen. However, and for the first time in this series, we observe the second SR substituent on the C2 atom. The torsion C1–N–C2–C3 angles are far from 180° (7a 111.6(2)°, 7b 116.5(2)°, Table 2), approaching a limit of 90° for which both formal double bonds are perpendicular and their conjugation is almost completely broken. This is nicely demonstrated by the N–C2 bonds (7a 1.411(3) Å, 7b 1.424(2) Å), which are longer than those observed in the s-trans structures where the C2 atom bears the hydrogen atom (see other compounds in Table 1). Thus the s-trans conformation of the azabutadiene chain seen for 5 is strongly perturbed. Such a large distortion from ideal s-trans conformation has been already observed in other 2-aza-1,3-butadiene structures substituted at the C2 atom [33–40] The dihedral angles between “nearly parallel” phenyl ring C4–C9 and the three atom C1–N–C2 plane of the azadienic chain are equal to 15.0(3)° (7a) and 18.6(2)° (7b) with the corresponding N/C1/C4/C5 torsion angles of 12.2° and 16.2°, respectively. These experimental values fall in the range observed for compound 5i. The N/C1/C4 and N/C1/C10 angles in the imine part of the molecules also have values close to those found in 5.

Of course the question arises, why 4 exhibits a different substitution pattern in the case of NaS^tBu compared to the other sodium thiolates, including even steric demanding thiolates such NaSⁱPr and NaSC₆H₁₁. At the first glance, one may suspect a strong sterical repulsion between two geminal -S^tBu groups residing on the terminal carbon atom of a hypothetical isomer [Ar₂CN–C(H)C(S^tBu)₂](7′). However, a literature survey reveals that polythioethers with two gem-S^tBu groups attached on the same carbon atoms exist, [41] for instance the allene [(^tBuS)₂CCC(Ph)(S^tBu)] [42]. Another example is the structurally characterized butatriene-type tetrakisthioether [(^tBuS)₂CCCC(S^tBu)₂] [43]. But in all these examples, the adjacent carbon atom is an allenic one of the C type. But there is also a report on the diene 7,8,8-tris[(^tert-butyl)thio]-6-vinylpentafulvene [(^tBuS)₂CC(^tBuS)–C(H)C₅H₄] [44].

Let's have a look on the simplest model compound, e.g. dichlorovinylidene. Truce et al. have studied the reactivity of Cl₂CCH₂ toward NaSTol-p and NaS^tBu [45,46]. In the case of the more nucleophilic NaS^tBu reagent, they established an addition-elimination pathway via a transient alkyne species rationalizing the formation of cis-1,2-bis(^tert-butylmercapto)ethene as the sole product [45,46]. Some intermediates such as 1,1-dichloro-2-(^tert-butylmercapto)ethane and trans-1-chloro-2-(^tert-butylmercapto)ethene could be isolated. Later work of Tanimoto et al. confirmed that the reaction of dichlorovinylidene with sodium phenylthiolate in DMF affords exclusively cis-1,2-bis(phenylmercapto)ethane [47–49]. We propose a similar addition-elimination mechanism explaining the formation of 7 (Scheme 4).

More recently our group demonstrated that upon the reaction of dibromovinylferrocene [Br₂CC(H)-Fc] with NaS^tBu and NaSAr (Ar = Ph, p-tol) the structurally characterized rearrangement products Z-[(^tBuS)(H)CC(H)-Fc] and Z-[(ArS)(H)CC(SAr)-Fc] were formed [50a]. In a similar manner, treatment of dibromovinyl[2.2]paracyclophane [Br₂CC(H)-PCP] with excess of NaSAr in DMF gives rise to the vicinal dithioethers Z-[(ArS)(H)CC(SAr)-PCP] [50b]. Therefore one may conclude from the above discussion that the formation of the azabutadienes 5 bearing two geminal SR functions could be considered as an exceptional case, whereas the rearrangement encountered in the case of NaS^tBu leading to 7 corresponds rather to the “classical” pathway. We suppose that a combination of both steric and electronic factors is decisive for the formation of 7 using the bulky and highly nucleophilic NaS^tBu. The NaSⁿBu reagent leading to 5a should have a comparable basicity and nucleophilicity with NaS^tBu, but is doubtlessly sterically less demanding.

2.4 Reactivity of 4a towards sodium ethylthiolate

In analogy to the synthesis of compounds 5, the reaction of 4a with NaSEt in DMF at room temperature afforded exclusively [Ph₂CN–C(H)C(SEt)₂] 5d, as confirmed by the presence of the vinylic hydrogen as a singlet at δ 7.10 ppm and two sets of SEt groups in the ¹H NMR spectrum. However, when the reaction of 4a (Scheme 5) was conducted with an excess of NaSEt under reflux in neat ethanethiol for 4 days, imine [Ph₂CN–C(H)(SEt)–CCl₂H] 8 was isolated as a yellow solid. The constitution of 8 has been elucidated from mass spectroscopy and NMR data. Both the ¹H (Fig. S2) and ¹³C NMR spectra of 8 reveal the presence of only one thioethyl group. This unexpected product results probably from formal EtSH addition across the CC bond of 4a, similar to the first step of the reaction sequence shown in Scheme 4.

It seems that the basicity of NaSEt does not permit the elimination of HCl. However, the reaction of compound 8 in toluene with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) or an equivalent of NaOMe in methanol yields quantitatively the dehydrochlorinated compound 9. We identified a mixture of two diastereomers (Z)-9/(E)-9 in a 9:1 ratio. The proton NMR spectrum allows distinguishing two vinylic protons at 5.22 and 5.52 ppm. By comparison with the chemical shifts at 5.23 and 5.51 ppm respectively, calculated by the incremental method of Pascual, [32] we can assign the major compound (singlet at 5.52 ppm) to the stereoisomer (Z)-9. The synthesis of the azabutadiene 7c, which is structurally related to 7a,b, was achieved by reacting 9 with an excess of NaSEt in ethanol. The proton NMR spectrum revealed again the presence of a mixture of two stereoisomers (Z)-7/(E)-7 in an approximate 8:2 ratio. Two vinylic protons are observed at 5.20 and 5.47 ppm. Again, the incremental method of Pascual allowed a stereochemical assignment of the resonances. For the dominating stereoisomer (Z)-7c, a chemical shift of 5.23 ppm was calculated. For the vinylic hydrogen of the minor isomer (E)-7c, a chemical shift of 5.47 ppm was calculated, demonstrating the excellent match with the experimental data.

2.5 Reactivity of 4a towards the sodium salt of piperidine

With the aim to prepare amine-substituted azabutadienes related to 5 and 7, we have reacted 4a with secondary amides X₂N⁻Na⁺ (X = Ph, Et) as well as with NaNH₂. But despite varying the reaction conditions, we failed to isolate any product. Fortunately, treatment of 4a with cyclic amides was more successful. We have communicated that the reaction of 4a with the sodium salt of pyrrole leads, through an addition-elimination process, to (E)-4-chloro-1,1-diphenyl-3-(1-pyrrolyl)-2-azabuta-1,3-diene, in which the chloro and the N-pyrrole substituents are in a trans-configuration [16]. For comparison, we treated 4a with the in situ prepared sodium salt of piperidine using neat piperidine as the reaction medium. After work-up, a yellow solid [Ph₂CN–C(H)C(NC₅H₁₀)₂] (10) was isolated in 45% yield (Scheme 6).

Since the spectroscopic data, notably the chemical shift of the vinylic H-proton resonating at δ 5.40 ppm did not allow an unambiguous assignment of the substitution pattern of the azadienic chain, we have grown single-crystals of 10 from ethanol and determined the molecular structure of this piperidino-functionalized 2-azabutadiene. An ORTEP plot depicted in Fig. 3 confirms that the molecular geometry and substitution pattern correspond to that of [Ph₂CN–C(H)C(OPh)₂] [14] and [Ar₂CN–C(H)C(SR)₂] (5). Again, one aryl group is almost coplanar with the transoid azabutadiene array (N1/C1/C4/C5 torsion angle −14.5(2)°, Table 1). The torsion angle of the azadiene array (C1/N1/C2/C3) reaches 178.4(1)°. Both piperidine substituents on the C3 atom adopt a chair conformation. The mean C–N distance of the CC(NC₅H₁₀)₂ moiety bearing the two bulky piperidyl groups amounts to 1.395(2) Å and compares well with that reported for methyl-3,3-dipiperidinodithioacrylate [MeSC(S)–C(H)C(NC₅H₁₀)₂] (1.35(4) Å) [51] as well as with other structures reported in this article.

2.6 Coordination of 6 on CuBr yielding a 16-membered macrocyclic complex

As stated in the introduction, one of the aims of this work was the preparation of a wide range of potentially bi- or poly-dentate unsaturated ligands for coordination chemistry. In a preliminary study on the coordination properties of tetrathioether 6, we reacted this ligand with CuBr in a 1:1 metal-to-ligand ratio. A single-crystal diffraction study (Fig. 4) indicates that from four available sulphur coordination sites in the ligand, only one aryl-bound SiPr group and one vinylic SiPr group are involved in the complexation of 6 on CuBr leading to the original dinuclear 16-membered macrocyclic complex [{Cu₂Br₂{p-iPrSC₆H₄)₂CN–C(H)C(SiPr)₂}₂] (11). The cycle is built over the crystallographic symmetry centre in the P1¯ space group. Consequently, the central phenyl rings (C4 through C9 between the Cu/Cu* atoms) are parallel but slipped with respect to each other. The distance between these two planes is close to 3.4 Å and may so suggest a kind of intra-molecular π stacking (Fig. S3) as encountered in cyclophanes.

4.1 General methods

The NMR spectra were recorded on a Bruker Avance 300 apparatus (300.13 and 75.1 MHz for ¹H and ¹³C) in CDCl₃, unless otherwise stated. FT-IR analysis was performed on a Shimadzu IR Affinity-1 apparatus (KBr, ν in cm⁻¹). Melting points were obtained on a Büchi (SMP10) capillary melting point apparatus. UV–vis spectra were measured with a VARIAN-Cary 100 spectrophotometer in CH₂Cl₂ at room temperature, λ_max in nm (ɛ in M⁻¹ cm⁻¹). The hydrazones Ar₂CN–NH₂ were prepared according to the literature and converted to Ar₂CN₂ using yellow HgO [53–55]. The thiols were purchased commercially from Aldrich and Alfa Aesar.

4.2 General procedure for the synthesis of the Δ³-1,2,4-triazolines

An equimolar mixture (10 mmol) of the diaryldiazomethanes and N-ethoxycarbonyl-N-(2,2,2-trichloroethylidene)amine in 10 mL of toluene or dichloromethane, was stirred at room temperature for 2–3 days. The solvent was then evaporated, and the residue was filtered and washed with ethanol.

4.3 General procedure for the synthesis of azadienes 4

An equimolar mixture (10 mmol) of the diazo compound and of N-ethoxycarbonyl-N-(2,2,2-trichloroethylidene)amine in 10 mL of toluene is stirred at 65 °C for 3 days. Then, the toluene is evaporated and the oily residue is chromatographed on silica or on neutral alumina, using a CH₂Cl₂-petroleum ether mixture (50:50) as the eluant. Recrystallization from ethanol provided 4 in the form of pale yellow crystals.

4.4 General procedure for the synthesis of 2-azabuta-1,3-dienes 5, 6, and 7.

1,1-Diaryl-4,4-dichloro-2-azabuta-1,3-diene (1.1 mmol) was stirred with an excess of thiolate (10 mmol) in dry DMF (10 mL). The reaction mixture was stored at room temperature for 8 h, then poured into water (100 mL) and extracted with diethyl ether (150 mL). The organic solution was washed three times with water, dried over anhydrous sodium sulfate and evaporated. The crude residue was recrystallized from ethanol or hexane.

4.5 Crystal structure determinations

Single crystals of 5i, 7a, 7b, 10 and 11 were mounted on a Nonius Kappa Apex-II CCD diffractometer equipped with a nitrogen jet stream low-temperature system (Oxford Cryosystems). The X-ray source was graphite-monochromated Mo-K_α radiation (λ = 0.71073 Å) from a sealed tube. The lattice parameters were obtained by least-squares fit to the optimized setting angles of the entire set of collected reflections. Intensity data were recorded as ϕ and ω scans with κ offsets. No significant intensity decay or temperature drift was observed during data collections. Data were reduced by using DENZO software [56]; the missing absorption corrections were partially compensated by the data scaling procedure in the data reduction. Absorption correction was further applied by using MULTISCAN [57]. The structures were solved by direct methods with SHELXS97 [58a] or SIR [58b] programs. Refinements were carried out by full-matrix least-squares on F² using the SHELXL97 program on the complete set of reflections [58a]. All non–hydrogen atoms were refined with anisotropic thermal parameters. The hydrogen atoms were placed in calculated positions and included in final refinement in a riding model with the isotropic temperature parameters set to U_iso(H) = 1.2U_eq(CH), U_iso(H) = 1.3U_eq(C_aromatic) and U_iso(H) = 1.5U_eq(CH₃).

Crystallographic data (cif files) have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication numbers 835275 (5i), 835276 (7a), 835277 (7b), 889646 (10) and 941426 (11). Copy of these data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge CB2 IEZ, UK, fax: 144-(0)1223-336033 or e-mail: deposit@ccdc.cam.ac.uk.