Die Kinetik der Piancatelli-Umlagerung

• Kinetics of the Piancatelli rearrangement

Vömel, Lukas Theodor; Liauw, Marcel (Thesis advisor); Leitner, Walter (Thesis advisor)

Aachen (2019)
Dissertation / PhD Thesis

Dissertation, RWTH Aachen University, 2019

Abstract

The Brønsted acid catalyzed Piancatelli rearrangement of the substrate furfuryl alcohol (FA) to 4-hydroxycyclopent-2-enone (4HCP) with polymeric by products was experimentally investigated in three different reactor types (stirred-tank batch, tubular reactor and continuously stirred-tank reactor (CSTR)). Initially, the influences of the pressure p, proton concentration c(H+), volume fraction toluene ϕ(toluene) and cosolvent N-methyl-2-pyrrolidone c(NMP) on the selectivity in the tubular reactor were excluded. Subsequently, the reaction parameters temperature T, residence time τ and concentration of the substrate c$_{0}$(FA) were empirically optimized in a self-optimizing tubular reactor. High selectivities were achieved at high temperatures and low initial substrate concentrations. High temperatures and residence times lead to higher conversions. The optimal reaction parameters of the tubular reactor are thus low initial substrate concentrations at high temperatures and a corresponding residence time. In order to generate a kinetic model for the reaction, concentration-time curves (c t) of the reaction from FA to 4HCP were recorded in the three reactors at 180 ≤ T/°C ≤ 240 and 15 ≤ p/bar ≤ 70. The c-t curves were investigated in the tubular reactor with eight models and combined with a convection model, in the batch reactor with four different kinetic models. The models of both reactors were combined to 16 complete models and validated on the c-t curves of the CSTR. It was shown that the concentration curves in both the tubular reactor and the CSTR are best described by the same kinetic model. This confirms the validity and robustness of this kinetic model. The other kinetic models could be excluded by the Akaike information criterion. Three components were considered in the best model: FA, 4HCP and the polymeric by-product. The model shows that the activation energy of the chain growth of the polymers is significantly smaller than that of the product formation to 4HCP. This leads to increased product selectivity at high temperatures. However, since the activation energy of the nucleation of the polymers is higher than that of the product formation to 4HCP, the selectivity reaches a plateau at very high temperatures. In addition, the speed constant of the chain growth of the polymers is smaller than that of the nucleation of the polymers. This relationship leads to higher yields in the tubular reactor than in the CSTR. Therefore, based on the data available in this thesis, a Plug-flow reactor (PFR) is best suited for the Piancatelli rearrangement of furfuryl alcohol due to its low back-mixing.