Abstract:
Carbonyl sulfide (COS) gas is emitted from aluminium smelters at concentrations of 1 – 8
ppmv or 0.2 – 1.6 kgt-1 Al produced. While COS emissions are relatively small compared to
emissions of HF or SO2, previous measurements suggest that COS is not efficiently captured
in the scrubbing systems used by aluminium smelters and it has been suggested that the
industry is a significant anthropogenic emitter of COS. COS is a unique sulfur gas in the
atmosphere, due to its stability, lifetime of several years and implication in ozone depletion.
The source of COS and other sulfur gases from aluminium smelters, is the 1 – 4 mass% S
content of anodes, consumed during aluminium electrolysis. These anodes are primarily
made from petroleum coke, a low value and therefore high sulfur product, of crude oil
refining. This thesis examines aspects of the formation and fate of COS in aluminium
reduction cells.
One thesis objective was to confirm the nature of sulfur species present in petroleum cokes
and anodes. A XANES (X-ray Absorption Near Edge Structure) spectroscopy study of cokes
from the major suppliers and anode core samples showed that organic sulfur containing ring
structures were the dominant sulfur species, before and after anode baking and even after
significant desulfurisation at 1500ºC.
Thermodynamic calculations performed using the HSC Chemistry© computer package
agreed with previous smelter measurements that COS gas was the dominant sulfur gas to
form at the anode face and SO2 became the dominant sulfur gas after the anode gases left the
cell and mixed with an induced air draft in the cell hooding.
Previous smelter measurements showed that some COS (about 5 ppmv) did survive through
the cell ducting into the dry scrubber and that there were only negligible reductions in COS
concentration after dry scrubbing. The dry scrubber contains smelter-grade alumina, which
adsorbs HF and other smelter gases, before the alumina is recycled to the cells. Another objective of this thesis was to study the effect dry scrubbing had on COS emissions
in greater detail in the laboratory to better understand any reactions that may be occurring. A
thermodynamic analysis using HSC Chemistry© determined that alumina would reduce the
stability of COS at a typical dry scrubbing temperature of 80ºC. The dry scrubber was
modelled in the laboratory using an alumina containing fluidised bed reactor, with the inlet
and outlet gas composition monitored using mass spectrometry. Adsorbed species resulting
from COS adsorption on alumina were studied separately using DRIFTS (Diffuse
Reflectance Infrared Fourier Transform Spectroscopy).
Experiments showed that COS was stable in the fluidised bed apparatus in isolation. The
presence of smelter-grade alumina led to the adsorption of COS in its molecular form and as
adsorbed hydrogen thiocarbonate. The latter species is believed to be an intermediate in the
hydrolysis of COS into H2S on alumina. The effect of conditions which may vary in the dry
scrubbing system, such as humidity, SO2 and HF concentrations, were studied. The presence
of optimum humidity levels ensured sustained hydrolysis of COS over several hours, but
adsorbed SO2 and HF interfered with COS conversion. This suggests that COS hydrolysis
would be unlikely to occur to a great extent in commercial dry scrubbers used in the
aluminium smelter industry.