Bed material agglomeration was studied experimentally in a fluidized-bed biomass combustor. Four biomass fuels, four bed materials, three bed additives, and three NO x additives were tested in a temperature range of 670−870 °C and at two pressure levels of 1.0 and 1.5 MPa. Two types of agglomeration were observed, a homogeneous and a heterogeneous type. The first occurred at low temperature and could be partly compensated for by erosion of the bed. The second took place at high temperatures and often involved interaction between the fuel ash and the bed material. The immobility of the bed particles made the heterogeneous agglomeration a self-accelerated process. The occurrence of hot spots in the bed was the precondition for heterogeneous agglomeration being induced. When silicon was present, alkali metals were the main contributors to heterogeneous agglomeration. Aluminum and iron compounds were able to suppress agglomeration through the high melting point of the eutectics that were formed.

Reduction in the amount of ammonia in fuel gas from biomass gasification was studied. Experiments were carried out in a fixed-bed reactor at 200-1000°C, 21 atm. A kinetic model for ammonia decomposition was developed. The partial pressure of hydrogen in the fuel gas was a key factor to model ammonia decomposition. Activation energies in the empty reactor, on carbon, and in a sand bed were similar, 130-140 kJ/mol. The frequency factors for carbon and sand were 10 times as large as for the empty reactor. The activation energy for a Ni-based catalyst was 111-113 kJ/mol. Carbon deposit deactivated the Ni-based catalyst. High temperature was found to be essential for avoiding carbon fouling and for achieving high ammonia removal efficiency. Estimation of the ammonia reduction for fuel gas showed that a moderate amount of ammonia could be removed by use of the Ni-based pellets at 800°C.

A packed column has been used to study the absorption of nitrogen oxide in an alkaline solution of sodium chlorite. The reactions taking place during the absorption have been examined and a lumped reaction model has been used to estimate rate constants from experimental data. Several parallel and consecutive reactions were found to take place during the absorption. NO was found to be oxidized to NO 2 and/or to NO 2 ± , and ClO 2 ± was reduced to Cl ± and/or to ClO ± . The pH value of the absorbing liquid was found to have a great impact both on the absorption rate and on the extent of the different redox reactions within the liquid. Experimental results indicate that sodium chlorite mainly works as an agent to oxidize NO to NO 2 and that the major part of the nitrogen oxides are absorbed via the hydrolysis of N 2 O 3 and N 2 O 4 .

No abstract

A packed column has been used to study the absorption of nitrogen oxide in an alkaline solution of potassium permanganate. The reactions taking place during the absorption have been examined and the rate constants have been estimated from experimental data. The experiments show that potassium permanganate is an excellent absorbent for nitrogen oxide. However, to avoid formation of Mn02, the hydroxide concentration has to be very high, i.e. > 3 mol/l. It was found that the reaction could be expressed as first-order with respect to NO and with respect to KMn04. The rate constant may be expressed in terms of the hydroxide concentration as follows: k, , , , = 6114.9CNaoH m3 mol-' s-'.