Photochemical spin trapping experiments

Photochemical spin trapping experiments are the stock in trade, and the most difficult ones to judge with respect to mechanism because of their high complexity. The method became popular at a time when the effect of light upon molecules was believed to result mainly in homolysis of bonds, principally because of its ready use in combination with epr spectroscopy and 

Photochemical ET reactions can be classified in at least three categories (which can co-exist), namely (i) simple homolysis of bonds of neutral molecules to give radicals of low redox reactivity (ii) excitation of a species D to produce an excited state D which initiates a second-order ET reaction involving another component of acceptor type, A, with formation of the radical pair D + A (iii) direct excitation of a charge transfer (CT) complex formed between two reaction components D and A to form the same radical pair D + A -. The first case is obviously an ideal situation if it can be realized, but this is seldom the case. The incursion or predominance of situations (ii) and/or (iii) in almost any system is possible, and precautions must be taken to avoid these complications. Much can be done by controlling the wavelength of the light source, but it is also possible to affect the chemistry in a predictable manner. 

This scheme, set up with reactions in dichloromethane, gave spin adducts from several of the nucleophiles discussed above (F, Cl , AcO, CN, tetramethylsuccinimide anion and triethyl phosphite), provided UV light was employed. With filtered light of A 435 nm, no spin adducts were detected. This is expected, since PBN cannot then be excited. With water as the nucleophile, only benzoyl nitroxide [9] was seen, indicating that any HO-PBN disappears too rapidly to be detectable 10 s in acetonitrile) and/or that its rate of formation from PBN +ConW is too low (see above). The complication that nucleophilic addition-oxidation might compete was ruled out experimentally in dichloromethane, but detected for fluoride ion in chloroform, using dioxygen to oxidize the intermediate hydroxylamine anion. 

DMPO did not produce spin adducts in the ComW scheme, except with tetramethylsuccinimide ion. This was possibly due to the fact that the Amax of DMPO, at 242 nm, lies somewhat outside the lower limit of the wavelength region of the lamp used. 

A second scheme involved the use of a sensitizer, 2,4,6-tris(4-methoxyphenyl)pyrylium ion ( °(A /A ) = 1.8 V, Amax = 422nm), to 

---

This formation was linear as a function of time and cyclohexanol was only detected as a very minor product. Cyclooctane was similarly oxidized with the formation of cyclooctanone as a major product. Other iron(III)porphyrins involving different axial ligands, like Fe(TDCPP)Q and Fe(TPP)Cl, were much less efficient and less selective. Irradiation of Fe(TDCPP)OH in the presence of 02 and cis-stilbene failed to give any formation of stilbene epoxide, and cyclohexene was mainly oxidized on its allylic position with formation of cyclohex-2-enol and cyclohex-2-enone under such conditions. Finally, the intermediate formation of OH radicals was detected by spin trapping experiments during cyclohexane oxidation by 02 with photochemically-activated Fe(TDCPP)OH. 

Photohomolysis reactions of cobalt(III) pseudohalide complexes can be used to effect photoreduction to cobalt(II) complexes. Thus, intramolecular photoelectron transfer in the complexes Co(CN)5N3 and Co(NH3)5N3 leads to oxidation of the azide ion to the azide radical, and reduction of the cobalt(III) center to cobalt(II). Evidence for the initial formation of the azide radical comes from the photolysis of solutions containing Co(CN)5N3 and iodide ion, when the iodine anion radical I2 is observed in the solution. This formation of I2 results from the photochemical generation of the azide radical, which then oxidizes the iodide ion to an iodine atom (Scheme 2.1). Subsequently, spin trapping experiments with phenyl-N-t rf-butyl nitrone has been used to verify the formation of azide radicals from the photolysis of Co(CN)5N and Co(NH3)sN3 / ... 

In related work, the reactions of hydrogen peroxide with iron(II) complexes, including Feu(edta), were examined.3 Some experiments were carried out with added 5.5"-dimethyl-1-pyrroline-N-oxide (DMPO) as a trapping reagent fa so-called spin trap) for HO. These experiments were done to learn whether HO was truly as free as it is when generated photochemically. The hydroxyl radical adduct was indeed detected. but for some (not all) iron complexes evidence was obtained for an additional oxidizing intermediate, presumably an oxo-iron complex. 

From the above, it is evident that every photochemical system must be carefully analysed in order to establish the nature of the process of spin adduct formation. Not all systems have the inbuilt diagnostic features of the fluoride or carboxylate nucleophiles, and it must therefore be accepted that mechanistic certainty will be difficult to attain. It also must be remembered that many studies in the past were designed without regard to the inverted spin trapping mechanism and are difficult to judge owing to lack of critical experiments to test this particular aspect. 

The unexpected formation of the blue crystalline radical cation (97) from the reaction of triazinium salt (98) with tetracyanoethylene has been reported and the product identified by its EPR spectrum and by X-ray crystallography (Scheme 42).199 Carboxylic acids react with the photochemically produced excited state of N-t-a-phenynitrone (PBN) to furnish acyloxy spin adducts RCOOPBN. The reaction was assumed to proceed via ET oxidation of PBN to give the PBN radical cation followed by reaction with RCO2H.200 The mechanism of the protodiazoniation of 4-nitrobenzenediazonium fluoroborate to nitrobenzene in DMF has been studied.201 Trapping experiments were consistent with kinetic isotope effects calculated for the DMF radical cation. The effect of the coupling of radicals with different sulfur radical cations in diazadithiafulvalenes has been investigated.202... 

Singlet oxygen ( O2) may also be another source of the oxidized spin traps, since it reacts with spin traps to give the same products as OH (73-75). The potential for the production of singlet oxygen arises because both Cr(VI) and Cr(V) complexes are photochemically active (22) and it is conceivable that they, or Cr(lV) intermediates, could produce O2. Evidence for the presence of O2 has been gained using the spin trap 2,2,6,6-tetramethyl-4-piperidone (TMPD), which is believed to be specific for O2. Experiments with this spin trap indicated the presence of O2 in the blood of Cr(Vl) treated mice (76). The possible involvement of O2 needs to be explored further in more simple in vitro systems. 

Photolysis of 3-buten-l-ol nitrite affords no cyclized products (Cy5/Cy4) neither does 5-hexen-l-ol nitrite (Cy6/Cy7). The same result is obtained on peroxydisulphate oxidation of 5-hexen-l-ol. In the Cy6/Cy7 case an important competitive pathway is probably 1,5-intramolecular ally lie hydrogen abstraction and, indeed, esr spin trapping by nitrosodurene " provides evidence of this. Cyclization in the Cy6/Cy7 case was considered to explain the reaction products of tetrahalogeno-o-benzoquinones with 2,3-dimethylbut-2-ene but was discarded in favor of a direct cycloaddition process on the basis of spin trapping and deuteration experiments. As discussed before, cyclization in the Cy3/Cy4 case must be difficult to observe because of the high j5-scission rate of oxyranylalkyl radicals. Nevertheless, this pathway has been used recently to explain the formation of diepoxides in the thermal-, photochemical-, or ferrous-salt-induced decomposition of unsaturated cyclic peroxides. In view of the multistep scheme involved this conclusion must await further confirmation. 

Esr studies wth organic spin traps have shown that azide radicals (N3) are formed in the reaction of hydroxyl radicals with azide ions, and that the radicals N 3 and OCN are formed by persulfate oxidations of azide and cyanide. However, azide radicals could not be detected in the photochemical reductive cis elimination of cw-diazidobis(triphenylphosphine)platinum(II) to give Pt(PPh3)2, even though they have been observed in the esr spectra of the photolysis products of other azido complexes. It is suggested, therefore, that the azide ligands are cleaved off as Ne (hexaazabenzene). Calculations have shown that Ne is slightly stabilized and could therefore be stable at the low temperatures of these experiments. 

---