Silver Alumina Catalysts for Lean NOx Reduction - Influence of Preparation Methods and Type of Reductant, with Transient Kinetic Modelling

Licentiate thesis, 2008

The objective of this work was to compare three different preparation methods for silver alumina catalysts for lean NOx reduction. Characteristics such as catalytic activity with propene and n-octane as reductants, chemical state of the silver and silver dispersion were analyzed. Furthermore, a global kinetic model for hydrogen assisted selective catalytic reduction of NOx with n-octane was developed for one of the prepared catalysts. Sample preparation was performed by the impregnation, thermally dried sol-gel and freeze-dried sol-gel methods. The latter is a new preparation method developed in this work. Characterization by XPS shows the presence of silver clusters on the impregnated and thermally dried sol-gel samples; however no such clusters are found on the freeze-dried sol-gel samples. Together with XRD and TEM analyses, the XPS results indicate that the silver is very finely dispersed throughout the alumina matrix in the freeze-dried sol-gel samples. This work has further shown that XPS enables detection of small silver clusters on the surface of silver alumina catalysts. The catalytic evaluation in flow-reactor shows high NOx reduction using propene as reductant for the freeze-dried sol-gel samples, however at elevated temperatures. The maximum NOx reduction, and also the corresponding temperature, is generally lower using n-octane as reductant. Moreover, the temperature window for NOx reduction is broader with n-octane compared to propene. The best catalytic performance with n-octane is found for a thermally dried sol-gel sample. The results from the flow reactor study imply that the ratio between silver clusters and oxidized silver species is crucial to achieve high NOx reduction. The kinetic model for H2 assisted n-octane-SCR of NOx was developed for the sample with the highest NOx reduction activity. The model was constructed from transient experimental data, with and without hydrogen feed, and it generally follows the changes due to these feed variations well. It is based on the removal of surface nitrates by hydrogen, also accounting for large temperature increases due to hydrogen combustion.