The dependence of alkali release in fluidised bed combustion on fuel compos
ition was studied using excimer laser fragmentation fluorescence (ELIF), an
in-situ method that determines gas-phase alkali species with sub-ppb sensi
tivity. The PFB-reactor at DMT was operated under combustion conditions and
ELIF-signals were detected in the flue gas after the cyclone. Within each
experiment, the coal composition was varied systematically by doping the co
al with known amounts of additives. The chlorine and sodium contents were i
ncreased separately, by respectively doping methylene chloride into the air
feed of the fluidised bed or sodium acetate into the coal. Sodium chloride
was used as dopant in some runs, but this raised both components simultane
ously. The addition of chlorine alone leads to over-proportional increases
in the alkali concentration whereas that of sodium acetate lead to relative
ly moderate increases. Thus the amount of chlorine in the coal and in the F
B reactor is decisive for the level of alkali emitted. In contrast, the pre
sence of clay minerals e.g. kaolin in the coal or the fluidised bed is foun
d to strongly suppress alkali release, especially that of potassium species
. (C) 2001 Elsevier Science Ltd. All rights reserved.