Ph.d. forsvar af Mads Lützen

Principal supervisor:

Associate Professor Christian Danvad Damsgaard

Co-supervisor:

Associate Professor Jakob Munkholt Christensen, DTU Kemiteknik

Examiners:

Professor Jakob Kibsgaard, DTU Fysik
Principal Research Scientist Ramchandra Tiruvalam, Topsoe A/S
Prof. Dr. Silke Behrens (IKFT) Karlsruhe Institute of Technology

Chairperson at defence:

Associate Professor Paul Kempen
 

Abstract:

This thesis uses atmospheric pressure transmission electron microscopy (TEM) to
investigate thermal heterogeneous catalysis. Additionally, the ContourFilter is presented. This allows for automated segmentation of supported nanoparticles in scanning electron microscopy (STEM) images.
Catalysis allows for the conversion of chemicals with the use of a catalyst. This includes synthesizing potential future chemical energy carriers, e.g., methanol. These carriers are a way to store energy produced from renewable energy sources, like solar and wind, for the days when the sun does not shine, and the wind does not blow. Such energy carriers are vital for the transition towards renewable energy sources and to improve self-reported life satisfaction globally by easing access to energy. Atmospheric pressure TEM provides images with an atomic resolution in 1 bar of pressure at catalytically relevant temperatures. This was used to investigate an industrial Cu/ZnO/Al2O3 catalyst for methanol synthesis. Here, atomic resolution
of a Cu nanoparticle was achieved with a dose rate of only 10 e/(Å2s). The dose rate limits for the sample were found for different gas compositions. Further production of CO and H2O, along with consumption of CO2 were confirmed with mass spectroscopy. This indicated a measurement of the reverse water gas shift reaction. Size-selected Pt nanoparticles were also investigated using atmospheric pressure TEM. CO oxidation was measured from 3.8nm nanoparticles with a projected coverage of 1%. The ContourFilter automates the segmentation of supported nanoparticles in STEM images. The algorithm’s concepts are detailed and tested against manual measurements of CoCu2Ga nanoparticles supported on SiO2. These tests showed that the ContourFilter results match the size distribution obtained by manual measurements, with a -2.7% relative median difference.