Flash vacuum pyrolysis of 2-acetyl-3-azido[1]benzothiophene

Flash vacuum pyrolysis (FVP) of 2-acetyl-3-azido[1]benzothiophene at 300 °C provides 3-methyl [1]benzothieno[3,2-c]isoxazole (72%). At higher temperatures, the heteroindoxyl 1,2-dihydro[1]benzo-thieno[3,2-b]pyrrol-3-one was obtained in low yield (ca. 10%). The heteroindoxyl exists as a mixture of keto and enol forms in DMSO solution. Because of the easy oxidative dimerisation of these products to indigotin (and its heteroanalogues), such reactions are excellent examples of the synthetic advantages of FVP with the monomeric products conveniently generated under vacuum in a solvent-free, air-free environment


Introduction
We have recently employed a gas-phase nitrene insertion process to generate indoxyl (1) 1 and its heterocyclic analogues (3) 2 and (4) 3 under flash vacuum pyrolysis (FVP) conditions (Scheme 1).2][3] Because of the easy oxidative dimerisation of these products to indigotin (5) (and its heterocyclic analogues) such reactions are excellent examples of the synthetic advantages of FVP with the monomeric products conveniently generated under vacuum in a solventfree, air-free environment.

Scheme 1
Unfortunately, application of the method to the quinoline analogue (6) gave the dinitrile (7) as the major product.We believe this product is formed via a known nitrene insertion in the hetero ring 4 , and subsequent rearrangements in which the elimination of ketene is the final step.Although the heteroindoxyl (8) could be reliably detected in the complex pyrolysate, it was never the major product and proved too unstable to isolate (Scheme 2). 3 In this paper, we complete our current studies of heteroindoxyls by using the nitrene strategy to achieve the successful synthesis of the benzothiophene analogue of 3, viz.1,2-dihydro [1]benzothieno [3,2b]pyrrol-3-one (9).This example demonstrates that the presence of a fused benzene ring in the precursor does not necessarily preclude the formation of an indoxyl by the nitrene strategy.In addition, a detailed comparison is now possible between the behaviour of the azide precursors to 1, 3 and 9 and the tautomeric properties of the products.Scheme 2. Reagents and conditions: (i) FVP 600 °C.

Results and Discussion
A one-step route to 2-acetyl-3-amino [1]benzothiophene (10) is available 5 ; this product was obtained in 80% yield after a 16 h reaction time (see Supplementary Material file, p. S2).Since the structure of the corresponding 2-acetyl-3-amino [1]benzofuran has been reported, 6 the X-ray crystal structure of 10 was obtained for comparison (Fig. 1); there are two molecules in the asymmetric unit (data for the second molecule and the benzofuran equivalent are shown in Table 1).The heavy atom skeleton of the molecule, as a whole, is planar {mean deviation from best plane 0.095 Å (0.034 Å).Maximum deviation from best plane is 0.0245 Å at C (11)   10) adopts an s-E configuration.With the exception of the region around the heteroatom, there are no significant differences in bond lengths between the benzothiophene (10) and the corresponding benzofuran. 6Similarly, the expected push-pull conjugation between N(3) and the carbonyl group of 10 has little effect on the C(2)-C(3) bond length, by comparison with a 2-aroylbenzothiophene previously reported.Due to the low solubility of 10 in dilute HCl, standard diazotization conditions gave only recovered starting material.When phosphoric acid was used instead, however, the increased solubility of 10 (and its phosphate salt) in the medium allowed diazotization and reaction with sodium azide to provide the azidobenzothiophene (11) in 47% yield, initially as a foam which could be purified by recrystallisation (Scheme 3).15) Å]}, another general feature of covalent azide geometry. 8In the absence of hydrogen bonding, the acetyl group itself adopts the opposite (s-Z) configuration to that of the acetyl group of the amine (10) (s-E).
FVP of the azide (11) was complete at 300 °C, producing 3-methyl [1]benzothieno[3,2-c]isoxazole (12) (72%).The parent benzothienoisoxazole is the only previously known example of this ring system and it was prepared from 3-azido [1]benzothiophene-2-carboxaldehyde by a solution-phase nitrene insertion strategy. 9VP of 11 at higher temperatures (>500 °C) resulted in rearrangement to the poorly soluble heteroindoxyl (9) via the nitrene (13), accompanied by a range of more soluble byproducts which contributed to the low isolated yield of (11%).FVP of the isoxazole (12) at 600 °C gave a similar, complex, pyrolysate.The product (9) was characterised by the CH 2 resonance of the keto form 9K [δ H (9K) 4.32 -c.f.δ H (3K) 2 4.22] and the CH resonance of the enol form [δ H (9E) 6.62 -c.f.δ H (3E) 2 6.51]; both tautomeric forms of 9 were present in DMSO solution (see below).The temperature profile for the decomposition of the azide (11) is shown in Fig. 3, and key parameters are listed in Table 2 with those for the sequences 14  15  1 and 16  17  3 listed for comparison (Scheme 5).Data for the benzothiophene sequence (11  12  9) are rather similar (within 25 o C) to those for the transformation of 2-acetylphenyl azide (14) to indoxyl (1) via anthranil (15) 1 , but both differ from the corresponding parameters for the thiophene (16  17  3).In particular, the azide (16) requires a higher temperature than 11 or 14 for complete decomposition, and the onset of formation of the thienopyrrolone (3) occurs at a much lower temperature (≥125 o C) than for 1 or 9, so there is no temperature at which the thieno[3,2-c]isoxazole 17 is the sole product.

Scheme 5. Comparison of ring systems.
The thermal reactivity of the azide (11), and stability of the isoxazole (12), relative to their monocyclic analogues ( 16) and (17), may be qualitatively rationalised by the maintenance of aromaticity in the benzene ring of 12, whereas the aromaticity of the thiophene ring is lost in the formation of 17.Despite the apparent oquinonoid character of anthranil (15), it is known that such structures maintain significant aromatic character 10 which may account for the similarity in behaviour of 14  15  1 and 11  12  9.Because of the low yield of 9 and its poor solubility, it was not possible to study its chemistry in detail.A complete NMR characterisation of both keto (9K) and enol (9E) forms, however, was carried out in DMSO solution.When first dissolved, the solution contained essentially 100% keto tautomer, suggesting that 9 exists in the keto form (9K) in the solid-state.After several hours in solution, the enol form (9E) predominated (ca.80%).By comparison, indoxyl itself is present as the enol form (1E) almost exclusively (>95%) in DMSO, whereas the thiophene analogue (3) shows an unexpected preference for the keto form (3K) (80%) in the same solvent. 2

Flash vacuum pyrolysis reactions
The precursor was volatilized under vacuum through an empty, electrically heated silica tube (35  2.5 cm) and the products were collected in a U-tube cooled with liquid nitrogen, situated at the exit point of the furnace.CAUTION: aryl azides are potentially explosive when heated.Although we experienced no problems with the reactions reported here, the following precautions were always taken: 1.Each individual pyrolysis was carried out on a scale no greater than 250 mg., 2. A metal inlet heater was always used, and 3.The apparatus was protected by a blast shield while in use.Upon completion of the pyrolysis, the trap was allowed to warm to room temperature under a nitrogen atmosphere.The entire pyrolysate was dissolved in solvent and removed from the trap.The precursor, pyrolysis conditions [quantity of precursor, furnace temperature (T f ), inlet temperature (T i ), pressure (P) and pyrolysis time (t)] and products are quoted.

Figure 1 .
Figure 1.Plot of one of the molecules of 10 showing the crystallographic numbering scheme.

Figure 2 .
Figure 2. Plot of one of the molecules of 11 showing crystallographic numbering scheme.

Table 1 .
Bond lengths of benzothiophene molecules of 10 found in asymmetric units of crystal structure (crystallographic numbering)

Table 2 .
Data for the temperature profile of 11, forming 12 and 9 Diffraction data for 10 and 11 were collected with Mo-K radiation on a Bruker Smart Apex diffractometer equipped with an Oxford cryosystems low-temperature device operating at 150 K.