In this task, TUDelft investigates the principal feasibility of ligninases, specifically peroxygenases to oxidatively cleave the backbone of common thermoset composites. The chemical philosophy of this approach is based on the higher reactivity of C-H bonds next to heteroatoms such as O or N. The acetal-like structure of the resulting hydroxylation products is expected to lead to spontaneous breakage of the original C-N- or C-O-bond and concommittant depolymerization (Scheme 1). Amongst the model compounds tested methyl derivates of phenol and aniline were converted smoothly (albeit at significantly different rates) by the Pada-I variant of the peroxygenase from Agrocybe aegerita yielding the desired products (Figure 1).

With more complex model compounds, however, also more complex product distributions were obtained indicating that next to the desired a-heteroatom hydroxylation also further oxyfunctionalisation reactions such a ring-hydroxylation occurred (Figure 2). Overall, these results demonstrate that even with monomeric, structurally relatively simple model compounds the existing peroxygenases are poorly regioselective. Apparently, next to electronic effects (C-H activation energies) also steric effects (binding of the the starting material to the enzyme active site) play an important role for the peroxygenase-catalysed depolymerisation of polymers.

This information puts the basis for further enzyme engineering of more selective peroxygenases.