Selective and controllable purification of monomeric lignin model compounds via aqueous phase reforming

Depolymerization of lignin into its monomeric constituents is a promising way to produce aromatic bulk chemicals from lignocellulosic biomass and lignin waste streams. In order to obtain an industrial product further downstream, processing of the monomeric mixture will be needed. Therefore, we selectively removed methoxy (MeO) and hydroxy (OH) groups from mixtures of monomeric lignin model compounds on a Pt-based catalyst in an aqueous environment and achieved a narrow range of products, which is controllable via limiting the hydrogen supply. This well-balanced supply of hydrogen is crucial to push the reaction towards phenol while simultaneously preventing ring hydration. We could show that at temperatures of approx. 250 degree C with a Pt-based catalyst the MeO group was converted into an OH group while reacting with water to methanol (MeOH). This MeOH is then instantly reformed on the Pt catalyst to 3 H sub(2) and CO sub(2) providing an in situ supply of hydrogen directly at the active sites of the catalyst, which facilitates MeO and OH group removal. Therefore, the amount of MeO groups limits the supply of hydrogen. Ring hydration does not occur because the hydrogen is produced in situ and consumed immediately on the catalyst. Adding small amounts of MeOH in the beginning accelerates the reaction as expected; nevertheless the hydrogenation of phenol seems to be the slowest reaction in the reaction route, making it a promising product of the process. Pt catalysts with gamma -Al sub(2)O sub(3), ZrO sub(2), TiO sub(2), and activated carbon as a support were investigated, while a Ni/C catalyst was also tested as an alternative. Pt/ZrO sub(2) showed the best results with regard to conversion, followed by Pt/C and Pt/ gamma -Al sub(2)O sub(3). Pt/TiO sub(2) and the Ni/C catalysts showed no significant conversion. The support comparison reactions were done for ten hours taking liquid samples every hour. A simple reaction network is proposed and reaction rates are estimated and then fitted onto measured data.