Although formally forbidden, it is demonstrated that 1,3 migrations can occur and can be facile in heterocumulenes. An investigation of this migration in α-oxoketenes showed that the 1,3 migration of the chlorine substituent in 2- phenyl-3-oxoprop-2-enoic acid chloride is detectable below room temperature, using DNMR. The migration barrier derived from these DNMR experiments is 42 kJ mol-1 This is close to the reported B3LYP/6-311+G(3df,2p)//B3LYP /6-31G* calculated value of 36 kJ mol-1. Moreover, using the B3LYP/6-311 + G(3df,2p)//B3LYP/6-31G* method we confirmed reported G2(SVP, MP2) migration aptitudes. The B3LYP/6-311+G(3df,2p)//B3LYP/6-31G* migration barriers are in the range 37 to 202 kJ mol-1 and differ from the G2(SVP, MP2) values by ±10 kJ mol-1. The substituents with lower migration barriers have lone-pairs that may interact with the heterocumulene LUMO, and thus promote the migration.
A computational (B3LYP/6-311+G(3df,2p)//B3LYP/6-31G*) investigation of the 1,3 migration in the α -imidoylketene— α -oxoketenimine system demonstrated that this system has a higher migration barrier than the α -oxoketenes (113 vs. 69 kJ mol-1 for NH2 and 79 vs. 43 kJ mol-1 for N(CH3)2). A calculation that followed the reaction path confirmed that the 1,3 migration transition states are on a pathway between the respective α -imidoylketenes and α -oxoketenimines. In addition, it is demonstrated that the other possible conformation changes are unimportant to the 1,3 migration. The barrier to rotation of the amide or amidine group is much lower than the 1,3 migration barriers, being 26 kJ mol-1 for rotation of either group in the NH2 case. The effect of the adjacent heterocumulene is to increase this rotation barrier by 6-15 kJ mol-1 compared to an adjacent carbon-carbon double bond. The hindered amine rotations are reduced by the adjacent heterocumulene to a level comparable to that found in the presence of an adjacent carbon-carbon double bond. For the amides the hindered amine rotation barrier changes from 83 kJ mol-1 for formamide to about 70 kJ mol-1, and in the amidine cases changes from 58 and 49 for the E and Z imine conformers of formamidine to 30-52 kJ mol-1.
The α -imidoylketenes may also undergo a ring closure of the s-Z conformer to form an azetinone. This is found to have a lower activation barrier (60 kJ mol-1) than the 1,3 migration in the NH2 case, but the azetinone is less stable than the α -imidoylketene (by 42 kJ mol-1) and the most stable α -oxoketenimine (by 61 kJ mol-1). This means that when the 1,3 migration can take place any azetinone in the system will be able to ring open, and then undergo the 1,3 migration. The importance of this ring closure in the laboratory then depends on whether the azetinone can undergo a side reaction that removes it from the migration pre-equilibrium.
Recently our laboratory reported a novel linear ketenimine. This led to an investigation of the effect of electron-withdrawing groups on the nitrogen inversion barrier of ketenimines. Some ketenimines with electron withdrawing groups on the C-terminus were prepared. Additionally, inversion barriers for these ketenimines and for ketenimines with electron-withdrawing groups on the N-terminus were calculated using B3LYP/6-311+G(3df,2p)//B3LYP/6-31G*. It was found that the preparation of bismethylsulfonyl ketenimines (beyond the reported N-methyl case) was not achievable here. Instead biscyano- and cyano-methylsulfonyl ketenimines (N-methyl, N-ethyl and N-phenyl) were prepared. These were usually prepared from suitable methylthioenamines using flash vacuum pyrolysis (FVP). The synthesis of biscyano-N-phenylketenimine was also achieved by photolysis of 1-azido-2,2-dicyano-1-phenylethene. On warming a neat film of cyano- N-phenylketenimine from -140°C, the ketenimine recaptured the codeposited methanethiol to reform our FVP precursor at around -80°C, demonstrating the reaction is reversible.
The calculations (B3LYP/6-311+G(3df,2p)//B3LYP/6-31G*) showed that the effect of a cyano group on the ketenimine inversion barrier is very similar to the effect of the methylsulfonyl group. For the N-H ketenimines, a CONH2 C-substituent reduces the inversion barrier by 10-20 kJ mol-1, while two cyano or methylsulfonyl groups reduce the barrier by about 28 kJ mol-1 (from 51 for ketenimine to 23 kJ mol-1 for biscyanoketenimine). For the N-phenyl ketenimines we find one cyano group reduces the inversion barrier by ~12 kJ mol-1 (to 15 kJ mol-1 from 27 kJ mol-1). Moreover two cyano groups, one cyano and one methylsulfonyl group or two methylsulfonyl groups reduce the barrier by ~18 kJ mol-1 (to 9 kJ mol-1 from 27 kJ mol"^). Interestingly, replacing the N-terminal substituent with electronegative elements dramatically increases the inversion barrier (from 51 kJ mol-1 up to 194 kJ mol-1 for N-Fluoro). By contrast, replacing the N-substituent with a -electron withdrawing acyl or cyano group dramatically lowers the inversion barrier (from 51 kJ mol-1 to 13 kJ mol-1 for N-formyl). This decrease is cumulative with biscyano C-substitution where the barrier reduces by a further 6 kJ mol-1 (to 7 kJ mol-1 for biscyano-N-formyl). In a comparison between the values calculated using the B3LYP/6-311+G(3df, 2p)//B3LYP/6-31G* method and reported inversion barriers, the calculations underestimate the inversion barrier by 3-17 kJ mol-1 for all the ketenimines studied.
The experimental investigation of the ketenimines revealed that their infrared spectra are sensitive to the media in which they are observed. The N-methyl-bismethlysulfonylketenimine had the ketenimine infrared band vary from 2280 (solid) to 2150 (Ar matrix) cm-1. The calculated ground state infrared spectrum of this compound agreed with the lower value, while the higher value was close to the band from the calculated linear transition state. Further work to include the effect of the medium in the calculations is ongoing in the laboratory.