Thermal rearrangements of aromatic hydrocarbons are considered to be unimolecular reactions that directly involve the atoms of an aromatic ring structure and require no other reagent than heat. These reactions can be categorized in two major types: one that involves a complete and permanent skeletal reorganization (isomerization), and one in which the atoms are scrambled but no net change in the aromatic ring occurs (automerization). The general reaction schemes of the two types are illustrated in Figure 1.
This class of reactions was uncovered through studies on the automerization of naphthalene as well as the isomerization of unsubstituted azulene, to naphthalene. Research on thermal rearrangements of aromatic hydrocarbons has since been expanded to isomerizations and automerizations of benzene and polycyclic aromatic hydrocarbons.
Mechanisms
Automerizations
The first proposed mechanism for a thermal rearrangement of an aromatic compound was for the automerization of naphthalene. It was suggested that the rearrangement of naphthalene occurred due to reversibility of the isomerization of azulene to naphthalene. This mechanism would therefore involve an azulene intermediate and is depicted below:
Subsequent work showed that the isomerization of azulene to naphthalene is not readily reversible ( the free energy of a naphthalene to azulene isomerization was too high - approximately 90 kcal/mol). A new reaction mechanism was suggested that involved a carbene intermediate and consecutive 1,2-hydrogen and 1,2-carbon shifts across the same C-C bond but in opposite directions. This is currently the preferred mechanism and is as follows:
Isomerizations
The isomerization of unsubstituted azulene to naphthalene was the first reported thermal transformation of an aromatic hydrocarbon, and has consequently been the most widely studied rearrangement. However, the following mechanisms are generalized to all thermal isomerizations of aromatic hydrocarbons. Many mechanisms have been suggested for this isomerization, yet none have been unequivocally determined as the only correct mechanism. Five mechanisms were originally considered: a reversible ring-closure mechanism, which is shown above, a norcaradiene-vinylidene mechanism, a diradical mechanism, a methylene walk mechanism, and a spiran mechanism. It was quickly determined that the reversible ring-closure mechanism was inaccurate, and it was later decided that there must be multiple reaction pathways occurring simultaneously. This was widely accepted, as at such high temperatures, one mechanism would have to be substantially energetically favored over the others to be occurring alone. Energetic studies displayed similar activation energies for all possible mechanisms.
Four mechanisms for thermal isomerizations have been proposed: a dyotropic mechanism, a diradical mechanism, and two benzene ring contraction mechanisms; a 1,2-carbon shift to a carbene preceding a 1,2-hydrogen shift, and a 1-2-hydrogen shift to a carbene followed by a 1,2-carbon shift. The dyotropic mechanism involves concerted 1,2-shifts as displayed below. Electronic studies show this mechanism to be unlikely, but it must still be considered a viable mechanism as it has not yet been disproven.
The diradical mechanism has been supported by kinetic studies performed on the reaction, which have revealed that the reaction is not truly unimolecular, as it is most likely initiated by hydrogen addition from another gas-phase species. However, the reaction still obeys first-order kinetics, which is a classical characteristic of radical chain reactions. A mechanistic rational for the thermal rearrangement of azulene to naphthalene is included below. Homolysis of the weakest bond in azulene occurs, followed by a hydrogen shift and ring closure so as to retain the aromaticity of the molecule.
Benzene ring contractions are the last two mechanisms that have been suggested, and they are currently the preferred mechanisms. These reaction mechanisms proceed through the lowest free energy transition states compared to the diradical and dyotropic mechanisms. The difference between the two ring contractions is minute however, so it has not been determined which is favored over the other. Both mechanisms are shown as follows for the ring contraction of biphenylene:
The first involves a 1,2-hydrogen shift to a carbene followed by a 1,2-carbon shift on the same C-C bond but in opposite directions. The second differs from the first only by the order of the 1,2-shifts, with the 1,2-carbon shift preceding the 1,2-hydrogen shift.