A special case of negative ion photoelectron spectroscopy occurs when the negative ion has geometry corresponding to a neutral transition state (Figure 2).

Figure 2

In this case, it is possible to carry out transition state spectroscopy. Transition state species have very short lifetimes, existing for only one vibrational period, and are therefore extremely difficult to study using conventional techniques. Negative ion photoelectron spectroscopy is one of the few experimental methods that can be used to obtain detailed information about species as short-lived as these.

A system that is expected to amenable to transition state spectroscopy in this way is the dicyano-Cope system:

The negative ion should have the same geometry as the transition state, so that the transition state will be formed upon vertical photodetachment.

One should not be mislead by the simple pictures drawn for the transition state of the Cope rearrangement. The actual electronic structure of the transition state is very complicated, but can be understood using a biradical description. The molecular orbitals for the transition state can be constructed by considering the union of two allyl fragments. The important MOs for the system are shown below:

The transition state has two electrons that occupy these two orbitals. Therefore, it is convenient to use a two-configuration wave function:

The actual description of the transition state depends on the relative values of c₁ and c₂ in the wave function. If c₁ >> c₂, the electrons are completely delocalized over the entire molecule. This is the case when the b_u orbital is much more stable than the a_g orbital. If the a_g orbital is much more stable than the b_u orbital, then c₂ >> c₁. Again, the electrons are delocalized over the entire molecule, although there is some preference for the 1- and 4-positions, which have the large coefficients. If the energies of the orbitals are such that c₁ = c₂, the wave function is a mixture of the two configurations. The physical picture that corresponds to this is a molecule that has its electron density localized on the 1- and 4- positions, which is a localized singlet biradical.

From this description, it is clear that an important consideration for the transition state of the Cope rearrangement is the relative energy of the b_u and a_g orbitals, because that is what dictates the CI coefficients in the wave function. Not surprisingly, the orbital energy difference is sensitive to substituents on the system. For the hydrocarbon, the b_u orbital is more stable, and c₁ >> c_2. If we add an oxygen substituent, the a_g orbital is much more stable and c₂>>c₁. In fact, the a_g orbital is so stable that the "Cope intermediate" is not even a transient, but is a stable molecule:

 The interesting part is what happens in-between. Substituents that allow for delocalization, such as CN and Ph, will stabilize the a_g orbital with respect to the b_u, but not to the same extent as oxygen. The question, then, is will substituents stabilize the orbital enough to give the intermediate a biradical-like structure? This has been the key issue for the Cope rearrangement for 25 years.

Previous studies have generally involved kinetics measurements to try to determine if the intermediate is delocalized or biradical-like. The experiment that would actually answer the question directly would be to measure the energy difference between the b_u and a_g orbitals in the transition state. This requires transition state spectroscopy, like that described above.

Although the photoelectron measurements will provide structural information about the transition state, they will not give the orbital energy splitting in the transition state. However, our group has devised an experimental method for doing so. Consider the electronic structure of the negative ion. In the ion, two electrons occupy the lower energy orbital, while the higher energy orbital is singly occupied:

If we were to shine light on the ion, it would absorb at a wavelength that corresponds to the energy difference between the b_u and a_g orbitals, which is exactly the energy splitting that we are looking for. Therefore, we can get the energy difference between the b_u and a_g orbitals by measuring the absorption spectrum of the negative ion. The only requirement is that the lowest energy geometry of the negative ion is similar to that of the transition state. This is undoubtedly the case for p withdrawing substituents, like CN, NO₂, and CHO, probably true for phenyl substituents, possibly true for vinyl substituents, but not true for the hydrocarbon or in the case of p donors (including halogens).

We are examining the substituent effects on the orbital energy separation in the Cope intermediates in order to determine the nature of the CI coefficients in the two configuration wave function for both the chair and the boat pathways of the Cope rearrangement.