I would also like to thank all the members of the Dougherty group for their friendship and support

I would also like to thank all members of the Dougherty group for their friendship and support. Analysis of the hyperfine cleavage pattern observed in the Llm5 = 2 transition of 324-Ph reveals that the four-membered ring is planar. The kinetic analysis showed that the variation of decay rates with temperature follows Arrhenius' law, yielding activation parameters of log A = 7.8, Ea0 = 2.29 kcal/mol.

Using the activation parameters for the bridge-flip reaction together with other data results in the development of a detailed model of the kinetic and thermodynamic relationships between 120,320 and 22. Three of these configurations comprise the three components of the triplet state, while the remaining three configurations give giving rise to three singlet states. However, in biradicals where S Lr is very small or zero, ct>s2 gives a poor description of the singlet biradical ground state.

A proper description of the covalent singlet biradical can be constructed by mixing cp82 with ct>a2 so that the ionic terms cancel. In this ground state structure, the cyclobutane ring is flat, but the hydrogens attached to the radical centers have moved out of the plane of the ring.

Figure 1-1. Nonbonding molecular orbitals of a homosymmetric biradical.

In addition, the matrix effect observed for the decay of 320-ηi4 is significantly greater than that of 320. This weak binding is reflected in the very low barrier to the degenerate bridge-flip process that proceeds through the singlet biradical 120. This research remained the only successful attempt of its kind until 1984, when employees of these laboratories reported the observation of the triplet state of 1,3-dimethylcyclobutanediyl (24-Me) when diazene 11 was photolyzed at 4 K.4 We subsequently found that cyclobutanediyls a general class of directly observable, localized 1,3-biradicals and that their spectroscopy and reactivity can be directly investigated as a function of structure.

Earlier work in these laboratories6 showed that addition of N-methyltriazolinedione (MTAD) across the strained central bond of a bicyclobutane provides a convenient method for the synthesis of such urasols. 7 One key to the synthesis of the diazenes is therefore to obtain the appropriately functionalized bicyclobutanes for the MTAD addition. The synthesis of the diphenylurazole 25 then began with a-truxylic acid8 which was converted in four steps to the known l,3-dibromo-1,3-diphenylcyclobutane9 as a 1:1 mixture of diastereomers (Scheme 2-1) ).

Capping of the dibromides to 1,3-diphenylbicyclobutane (19) is efficiently accomplished by treatment with lithium amalgam. One reasonable mechanism for the formation of 26 involves the reaction of 19 to give the endoperoxide 27 followed by rearrangement to the epoxy ketone.

Scheme 2-1

Analogous rearrangements have been observed for a number of 1,2-dioxolanes.10 For example,1 2,3-dioxabicyclo[2.2.l]heptane 28 undergoes rearrangement to epsoyaldehyde 29 with simultaneous homolytic cleavages of the 0-0 and J3 C-C bonds. I0a The existence of the endoperoxide 27 should be only transient at best because the strain energy of the bicyclo[2.1.l]hexane skeleton (about 35 kcal/mol)ll is almost the same as the binding strength of the 0-0 peroxides. The syntheses of methyl-phenyl and monophenyl urazoles (30 and 31) proceed by parallel routes starting with 3-phenylcyclobutanone 12 (Scheme 2-2). Conversion of the hydroxyl groups to bromides using triphenylphosphine-carbon tetrabromide followed by benzylic bromination gave dibromides 36 and 37 in very good yields.

The standard method for converting urazoles to 1,2-diazenes involves basic hydrolysis followed by an oxidation via copper complexes.13 However, several attempts to convert diphenylurazole 25 to diazene 14 using this method gave only very low yields of an obscure copper complex. . We then turned our attention to an alternative hydrolysis-oxidation procedure recently developed in these laboratories.6c,14 This procedure involves partial hydrolysis of the urazoles and isolation of the product semicarbazides. A valuable advantage of this procedure over the previous one is that the oxidation can be carried out at temperatures as low as

J: NCH3 __ K_O_H_-~H

Scheme 3-2

The first experiment involves introducing oxygen into a cold CD2Cl2 solution (-78°C) of bicyclopentane and then monitoring its disappearance as the solution is allowed to warm. If the reaction rate is governed by ring opening (eq 3-5), we estimate (using the activation parameters determined for the bridge rotation reaction) that the half-life of bicyclopentane would be about 30 min at -60°C . . Figure 3-22 shows the extinction graph of 22 according to the first-order norm law.

The pronounced curvature of the plot indicates that the rate of the oxidation reaction does not depend on the (22). Simultaneously with our study of 22, Adam and Wirz undertook a detailed study of the oxidation of 22 to 23.20. Their study was carried out under pseudo-first-order conditions where 22 was present in excess. The lifetime of triplet tetracene, i:tet, was used as an indicator of the oxygen concentration.

The slope of the plot of ln(i:tet) versus time gave the pseudo first-order rate constant, lkobs, where lkobs. Comparing this free energy difference with the free energy of activation of the bridge reversal process ~Gi297 = 16. Note that the activation barriers for singlet biradical cyclization should be considered lower bounds because the rate constant for oxygen scavenging is 22, k2, representing an upper diffusion control limit .20 For example, if the trapping rate constant is an order of magnitude slower than diffusion, ~G297 drops from 13.7 to 12.4 kcal/mol, increasing the pit depth to 4.3 kcal/mol.

It may at first seem surprising that we invoke the capture of a biradical singlet with oxygen, because virtually all examples of biradical capture with oxygen involve the triplet state of the biradical.64 This is not a consequence of a " spin-allowed" triplet-triplet interaction, but it is rather due to the long lifetime of triplet biradicals compared to singlets. The unusual feature of the oxygen scavenging reaction of 120 is that, although it does not have an exceptionally long singlet lifetime (we estimate that the upper limit of the lifetime is 0.3 ns),67 the reaction is nearly quantitative. Therefore, the constant regeneration of 120 ensures that the reaction will eventually complete, even though only a small fraction of the available biradical is captured at any given time.

Although the experimental details of the oxidation reaction of 22 to 23 are consistent with the biradical mechanism discussed above, another mechanism involving the direct addition of oxygen across the strained central bond of 22 is a matter of debate. In a related study, Adam and Wirz have investigated the solution phase lifetimes of 120 and 320 using time-resolved laser flash. However, upon benzophenone-sensitized photolysis in CH3CN, they observed a transient absorption with a lifetime of 25 µs, which was attributed to 320.69 In agreement with the results we have obtained, they find that the only product of the direct and sensitized photolysis is bicyclopentane 22 .