Journal article
Deactivation of SCR catalysts by potassium: A study of potential alkali barrier materials
Department of Chemical and Biochemical Engineering, Technical University of Denmark1
CHEC Research Centre, Department of Chemical and Biochemical Engineering, Technical University of Denmark2
Freiberg University of Mining and Technology3
Haldor Topsoe AS4
Department of Chemistry, Technical University of Denmark5
Centre for Catalysis and Sustainable Chemistry, Department of Chemistry, Technical University of Denmark6
Organic Chemistry, Department of Chemistry, Technical University of Denmark7
The use of coatings in order to protect vanadia based SCR catalysts against potassium poisoning has been studied by lab- and pilot-scale experiments. Three-layer pellets, consisting of a layer ofa potential coating material situated between layers of fresh and potassium poisoned SCR catalyst, were used to test the ability of the barrier layer to block the diffusion of potassium across the pellet.
Of MgO, sepiolite and Hollandite manganese oxide, MgO was the most effective potassium barrier, and no potassium was detected in the MgO layer upon exposure to SCR conditions for 7 days. Two monoliths have been exposed to KCl aerosols at 350 °C in a pilot-scale setup for about 1000 hours. A 3 wt.% V2O5-7 wt.% WO3/TiO2 reference catalyst deactivated with an average rate of 0.91 %/day, and SEM-EDS analysis showed complete potassium penetration of the catalyst wall.
A similar monolith coated with 8.06 wt.% MgO deactivated with a rate of only 0.24 %/day, relative to the fresh activity of the uncoated reference. The initial observed activity of the coated catalyst was, however, only 58 % of that of the reference, likely due to increased transport limitations and loss of active material during the coating process.
Potassium had to some extent penetrated the MgO coat, and SEM analysis revealed it to be rather thick and fragile. Despite these observations, the coating did protect the SCR catalyst against potassium poisoning to some degree, leaving promise of further optimization.
Language: | English |
---|---|
Year: | 2017 |
Pages: | 56-64 |
ISSN: | 14353199 |
Types: | Journal article |
ORCIDs: | Olsen, Brian Kjærgaard , Schill, Leonhard and Jensen, Anker Degn |