Deactivation of palladium and platinum catalysts due to coke formation was studied during hydrogenation of methyl esters of sunflower oil. The supported metal catalysts were prepared by impregnating γ-alumina with either palladium or platinum salts, and by impregnating α-alumina with palladium salt. The catalysts were reused for several batch experiments. The Pd/γ-Al 2 O 3 catalyst lost more than 50% of its initial activity after four batch experiments, while the other catalysts did not deactivate. Samples of used catalysts were cleaned from remaining oil by repeated extractions with methanol, and the amount of coke formed on the catalysts was studied by temperature-programmed oxidation. The deactivation of the catalyst is a function of both the metal and the support. The amount of coke increased on the Pd/γ-Al 2 O 3 catalyst with repeated use, but the amount of coke remained approximately constant for the Pt/γ-Al 2 O 3 catalyst. Virtually no coke was detected on the Pd/α-Al 2 O 3 catalyst. The formation of coke on Pd/α-Al 2 O 3 may be slower than on the Pd/γ-Al 2 O 3 owing to the carrier's smaller surface area and less acidic character. The absence of deactivation for the Pt/γ-Al 2 O 3 catalyst may be explained by slower formation of coke precursors on platinum compared to palladium. , temperature-programmed oxidation (TPO). FIG. 2. (A) Measured coke content on samples contacted with oil for various times and temperatures. (B) Coke formation on the catalysts vs. time on-stream. Each symbol indicates one batch. The coke concentration was not analyzed for the Pd-α-Al 2 O 3 catalyst after batches 2 and 3. The other two curves represent catalysts on γ-Al 2 O 3 support.

The influence of various poisoning compounds on the hydrogenation of fat is reviewed. Sulfur‐containing compounds, present in both vegetable and marine oils, create serious catalyst deactivation problems. In this paper, the poisoning efffect of sulfur and its role on the reactant adsorption in different hydrogenation reactions are discussed. Several factors affecting the sulfur resistance of the catalyst metal are summarized. Other inhibitors influencing the fat hydrogenation, such as phosphorus compounds, free fatty acids, sodium soaps, chlorophyll, halogen compounds and products of lipid oxidation, are also considered.

This is a preprint of an article whose final and definitive form has been published in Applied Catalysis B-Environmental.

The diesel oxidation catalyst (DOC, Pt/γ-Al 2 O 3) was used in a synthetic-gas catalyst test bench to study internal mass transfer limitations during NO oxidation. A simple and fast experimental methodology, by varying the washcoat thickness in monolithic DOCs was developed and the results were evaluated using various experimental time scales. The ratio between the reaction time constant and the washcoat diffusion time constant was useful in identifying temperatures where the DOCs tested transitioned between a kinetically controlled region and an internal mass transfer controlled region. The NO conversion was shown to be significantly limited by internal mass transfer already at 175 °C for an average washcoat thickness of 110 µm.

There are various methodologies to account for mass transfer within non-uniformly distributed washcoats in monolith reactors in 1D models (axially). However, 1+1D models (axially/radially) fail to capture local variations in mass transfer from different coating thicknesses or cracks. In this paper, we present a novel way to account for local material properties in a washcoated monolith reactor. The suggested method uses an existing 1+1D modelling framework and sectionalizes the washcoat into multiple tangential segments which are solved independently. Intelligent gravimetric analysis and scanning electron microscopy are used in combination to calculate local effective diffusivity as an input for each simulation. The new model is compared to the original 1+1D model using NO light-off simulations. The new model predicted increased conversion at elevated temperatures, where mass transfer limitations are present, due to the higher porosity in the corners. The simulation time for each model was similar due to the parallelizable nature of the new model.