Ozonolysis

Ozone, an allotrope of oxygen, is a 1,3-dipole that undergoes [4s + 2s] cycloaddition to alkenes. The structure of ozone may be written as a resonance hybrid of zwitterionic structures, as shown here. Although the energetically favored structures on the left are 1,2-dipoles, ozone can only react as a terminal 1,3-dipole, since the central oxygen has a filled valence shell.

The initial product of ozone cycloaddition to an alkene is called a molozonide. Molozonides are very unstable and rapidly decompose; nevertheless, spectroscopic evidence for the transient existence of a molozonide has been obtained at -100 ºC. The manner in which molozonides decomposes has been the subject of many investigations. As shown in the following diagram, it is usually depicted (the Criegee mechanism) as a concerted cycloreversion leading to a carbonyl fragment and a zwitterionic species, sometimes referred to as a carbonyl oxide. Calculations indicate an activation energy of 20 kcal/mol or less for such a transformation. An alternative cleavage of one of the two very weak O–O bonds in the molozonide, followed by immediate fragmentation of the resulting diradical would lead to equivalent intermediates. The zwitterionic intermediate is a 1,3-dipole of the same type as ozone; however, it prefers to cycloadd to carbonyl functions rather than alkene double bonds. In this way an isomeric ozonide, known as a Staudinger ozonide, is formed. Such ozonides are much more stable than their molozonide precursors, and in some cases may be isolated as pure crystalline compounds. In general, ozonides must be handled with care, since they may decompose explosively.
If ozonolysis is carried out in an alcohol solvent, instead of the customary methylene chloride solvent, the dipolar carbonyl oxide may be trapped as the α-hydroperoxide of an ether. The reaction of styrene, shown below the green line, is an example. In the case of unsymmetrical alkenes such as styrene, the carbonyl oxide zwitterion is stabilized by alkyl substituents on the positively charged carbon, and the fragmentation of the molozonide favors this intermediate. Likewise, nearby electron withdrawing substituents destabilize the carbonyl oxide species.

In addition to the alcohol trapping results, there is a growing body of evidence supporting key elements of the mechanism shown here. Some examples will be displayed above by clicking on the mechanism diagram. Thus, the carbonyl oxide zwitterion undergoes cycloaddition reactions with extraneous carbonyl compounds, as shown by ozonide C in equation 1. Although two zwittterions could combine to give a dimeric bis-peroxide, the extreme reactivity of these intermediates makes such an encounter improbable, compared with other reactive interactions. Dialkyl substitution serves to improve the stability of the carbonyl oxide, and in the case of reaction 1, the bis-peroxide B is obtained in small amounts together with the ozonides A and C.
Intramolecular recombination of the carbonyl and carbonyl oxide fragments is usually favorable, as the ozonolysis of cyclopentene (equation 2) demonstrates. The bicyclic [3.2.1] ozonide is obtained pure in yields up to 80%. Curiously, cyclohexene does not yield a similar monomeric ozonide, but instead, an assortment of oligomeric ozonides and peroxides. An explanation for this behavior may reside in the configuration of the intermediate carbonyl oxide zwitterion. Just as oximes and other imine derivatives may exist in syn and anti stereoisomeric configurations, the RHC=O–O grouping may adopt similar structures. The resonance description of a carbonyl oxide is presented in the gray shaded box above. Of the four Lewis structures shown, the most favorable is clearly the one on the left. The C=O–O unit is planar and bent, with a bond angle ca. 120 º. If the barrier to rotation (or inversion) about the C=O bond is sufficiently high, the carbonyl oxide will have a distinct configuration relative to the substituents on carbon. The example shown here is syn. If a syn configuration of the carbonyl oxide is required for intramolecular cycloaddition of short chains, and the cyclohexene molozonide fragments to an anti isomer, its failure to form a monomeric ozonide is understandable. A beautiful demonstration supporting this explanation will be displayed above by clicking on the mechanism diagram a second time. Cyclopentene and cyclohexene derivatives, each carrying an appropriately sized, deuterium labeled, aldehyde substituent, were ozonized in methylene chloride at -78 º.C. Molozonide fragmentation in each case produces a carbonyl oxide having two equal length aldehyde chain substituents. Unlike the unsubstituted examples noted above, both compounds lead to monomeric ozonides in good yield. If the carbonyl oxide intermediate is formed in a stereo-random fashion, or if the isomeric forms are rapidly interconverted, then the deuterium label will be scrambled between the bridgehead location and the remaining aldehyde side chain. The data presented in the diagram clearly demonstrates stereoselective fragmentation of each molozonide and a strong preference for syn-cycloaddition.

Equation 1 in the following diagram illustrates formation of a typical carbonyl oxide intermediate by oxygen addition to a carbene, generated by photochemical elimination of nitrogen from a diazo compound. This carbonyl oxide exhibits the same reactivity as those formed by ozonolysis, including cycloaddition to an aldehyde carbonyl function. In the same manner, a sufficiently stable carbonyl oxide species, permitting spectroscopic characterization, was prepared recently, as shown in equation 2. Although still highly reactive, this intermediate could be examined in solution at temperatures below -80 ºC. Further irradiation isomerized the zwitterion to its neutral dioxetane isomer, shown on the right. Despite the apparent ring strain of this compound, it is stable up to 20 ºC and could be crystallized. Spectroscopic and X-ray diffraction data confirm the structure shown here.

In most synthetic applications of ozonolysis, oxidative or reductive decomposition of the Staudinger ozonide to carbonyl products or their acetal derivatives is the final stage of the reaction. Two common methods employed in such work-up were described in an earlier section of this text, and many others have proven useful. For example, quenching the ozonolysis reaction mixture in a THF solution of lithium aluminum hydride results in reduction of both carbonyl moieties to alcohols. A particularly useful set of conditions that permit two symmetrically equivalent aldehyde functions to be released in different forms or oxidation states is shown in the following diagram. These procedures all begin with a low temperature ozonolysis in the presence of methanol. Once the double bond is completely converted to the initial ozonide product, as evidenced by the characteristic blue color of unreacted ozone, the excess ozone is removed by a stream of nitrogen. At this point one of the aldehydes is free and the other exists in the form of an α-hydroperoxide methyl ether. In procedure A, addition of p-toluenesulfonic acid converts the free aldehyde to a dimethyl acetal. The acid catalyst is then neutralized with sodium bicarbonate, and the hydroperoxide is reduced to a hemiacetal by treatment with dimethyl sulfide, a generally useful reductant for ozonides or peroxides. This reduction is shown by the upper equation in the blue shaded box.

Following the initial ozonolysis, procedures B and C proceed by first removing excess methanol as a benzene azeotrope. The key reaction in both cases is an eliminative oxidation of the α-hydroperoxide methyl ether, as shown by the bottom reaction in the shaded box. This reaction is effected either immediately (conditions B) or following acetal formation as in procedure A (conditions C).