Combustion and thermal processes produce severe toxic components, like particulate matter, polychlorinated dibenzo-p-dioxins, polycyclic and aromatic hydrocarbons, environmentally persistent free radicals (EPFR). These radicals are thought to be generated by interaction of aromatic compounds with the environmental transition metal oxides (iron, copper, zinc, nickel oxides) [1, 2] and are expected to produce reactive oxygen species (ROS) like OH●, O●2−, HO2● HO2●−, H2O2 [3]. ROS are able to induce biological damages, going from inflammation to serious DNA damage, hence the relevance of acknowledging the mechanism of radical formation and in particular the catalytic role of the transition metal ions.

In the last 10 years the mechanism proposed by Dellinger et al. [1, 2] was dominating all the investigations. It has been suggested that firstly the molecular aromatic precursor chemisorbs on the oxide surface, with elimination of water, then an electron transfer from the aromatic molecule to the transition metal center takes place and causes both reduction of the metal and EPFR generation (scheme 1a). The above mechanism [4] has been assessed detecting the radical species by the electron spin resonance (ESR); however the same ESR technique did not give sure evidence of the reduction of the transition metal center.

Scheme 1
scheme 1

a Dellinger mechanism [4] for the formation of phenoxyradical. b Cosentino mechanism [5] for the formation of phenoxyradical. In the reaction scheme also the relative electron energies of intermediates with respect to separate reagents are reported (in round parentheses kcal/mol) as well as the activation energies (in square parentheses kcal/mol). For the A5 intermediate, which is very stable with respect to separate reagents, a biradical singlet may be hypothesized featuring the antiferromagnetic coupling that quenches the ESR lines. The mechanism is independent from the presence of the OH group in the aromatic molecule

Full size image

On the other hand the spectroscopic and computational model proposed by Cosentino, Morazzoni et al. [5], gives evidence of a mechanism triggered by O2 interaction, where the decrease in intensity of the Cu(II) ESR lines was attributed to antiferromagnetic interaction between the phenoxyl radical and the Cu(II) center (Scheme 1b, species A5), instead of the Cu(II) reduction proposed by Dillinger.

More recently different authors investigated the formation of EPFR radicals on several oxides; the abilities of the metal oxides to promote EPFR formation were in the order Al2O3 > ZnO > CuO > NiO [6], in particular the radicals are most stable on ZnO. This is a highly ubiquitous oxide where the radical longevity may present an ongoing environmental hazard. Zn2+centers cannot be significantly reduced to Zn+ under contact by aromatics, thus the Dellinger’s mechanism becomes even more questionable in this case. Very recently an investigation of the electronic state of ZnO after chemisorption of the aromatic molecule demonstrated, through photoelectron spectroscopy (XPS) and DFT calculations, that an electron transfer is active from the oxide to the aromatic molecules [7], opposed to that proposed by Dellinger. Accordingly the trend of the oxides in promoting the EPFR formation is parallel to their ability to stabilize the superoxide anion.

By collecting all these suggestions and mainly following the Cosentino mechanism [5], it seems to me that like in the case of CuO, a trigger action was played by O2 also for ZnO. It is well known that the superoxide species Zn2+-O2− stabilizes at the oxide surface [8] Thus Zn2+-O2− can be thought the product of the electron transfer from ZnO to O2 (probably via conduction electrons) and should allow the formation of the phenoxy radical on ZnO as well as on CuO, without reduction of Zn2+(A4 and A5 in Scheme 1b).

A similar mechanism was reported by Orlandi, Morazzoni et al. [9] and is concerning the catalytic formation of the phenoxy radical by interaction between E-methyl ferulate (ROH) and [N,N’bis(salicylidene)ethane-1,2-diaminato] cobalt(II), [Co(salen)]. The ESR investigation demonstrated the following pathway.

$$\left[ {{\text{Co}}^{{{\text{II}}}} \left( {{\text{salen}}} \right)} \right] + {\text{ROH + O}}_{2} \rightleftharpoons \left[ {{\text{Co}}^{{{\text{III}}}} \left( {{\text{salen}}} \right)\left( {{\text{ROH}}} \right)\left( {{\text{O}}_{2} } \right)} \right]^{ - }$$
(1)
$$\left[ {{\text{Co}}^{{{\text{III}}}} \left( {{\text{salen}}} \right)\left( {{\text{ROH}}} \right)\left( {{\text{O}}_{2} } \right)} \right]^{ - } + {\text{ROH}} \rightleftharpoons \left[ {{\text{Co}}^{{{\text{III}}}} \left( {{\text{salen}}} \right)\left( {{\text{ROH}}} \right)\left( {{\text{RO}}^{ \cdot } } \right)} \right] + {\text{HO}}_{2}^{ - }$$
(2)
$$\left[ {{\text{Co}}^{{{\text{III}}}} \left( {{\text{salen}}} \right)\left( {{\text{ROH}}} \right)\left( {{\text{RO}}^{ \cdot } } \right)} \right]^{ - } + {\text{HO}}_{2}^{ - } \rightleftharpoons \left[ {{\text{Co}}^{{{\text{III}}}} \left( {{\text{salen}}} \right)\left( {{\text{RO}}^{ - } } \right)\left( {{\text{RO}}^{ \cdot } } \right)} \right] + {\text{H}}_{2} {\text{O}}_{2}$$
(3)

where the O2 interaction allows the formation of the superoxide derivative [Co(salen)(ROH)(O2−)] (step 1). The coordination of a second E-methyl ferulate, (step 2), removes O2− and gives rise to the phenoxy radical derivative. The above paper [9] also details that in the presence of a stronger electron donor molecule (like methanol of pyridine), the superoxide derivative remains stable and the phenoxy radical does not form.

It seems to us that the role of Zn 2+ can be thought as parallel to that of Cobalt(II) in the organometallic process, that is the formation of the phenoxy radical on ZnO goes through that of a superoxide species. If this has high stability, the formation of the phenoxy radical is hindered. In agreement with this opinion, the recent results reported by D’Arienzo et al. [10] outline that the phenoxy radical becomes stable on ZnO when the superoxide anion does not form. If, like in the case of defective ZnO, reactive oxygen vacancies are present and stabilize Zn2+-O2−, the formation of the phenoxy radical is strongly inhibited.

The way of avoiding the formation of phenoxy radicals is to not diffuse oxide particles with regular surfaces.