The analyses for the four samples A1–A4 were found to give
reproducible results, as was the case for the analyses for B1–B2.
Very small aluminum particles could not be identified with the
proposed elemental analysis method, which is why the 2mm
and sludge fractions were little taken into account in this study.
In order to recover aluminum scraps from 2–40 mm fractions,
remaining non-ferrous metals were firstly separated applying a
dry and a wet eddy current method. Subsequently the metal scraps
were separated using a metal detecting sensor separator after
which the aluminum scraps could be separated from the heavy
metals using a density separation method (Muchová et al., 2006).
The aluminum scraps in the +40 mm could be retrieved directly
using hand sorting. It is noted that the Al particles that were too
heavily corroded were manually discarded from further analysis.
This is justified since in an industrial separation plant (e.g. based
on eddy currents or density) the heavily corroded particles with
relatively little metallic aluminum can also not effectively be
recovered. For a standard aluminum beverage can in Europe, the deep-drawn body is commonly composed of aluminum alloy 3004
or 3104 while the top lid is made from alloy 5182. Typically the
body makes up 75% of the beverage can weight and the lid 25%.
The identification of these alloys in the different size ranges was
conducted using a combination of chemical detection methods
(Hu and Rem, 2009) and XRF analysis. The chemical method is
applied as a first separation criterion because it is faster and cheaper
than XRF analysis. For the chemical test a part of the particle surface
is polished and cleaned to remove corrosion, after which a few
droplets of a chemical liquid are applied. The presence of the targeted
alloying metal is detected visually by changes in color. For
each alloying metal a different chemical liquid was used. The
experimental procedure depicted in Fig. 1 is used for identifying
cans alloy in the retrieved aluminum scraps from the samples
A1–A4 and B1–B2. The used separation procedures for non-ferrous
metals and aluminum density recovery were verified by adding
given amounts of aluminum and copper to a subsample of bottom
ash from which all the metals were first removed. It was found that
the separation procedures have efficiencies themselves and are
determined to be better than 95%. To check reproducibility the separation
procedure was repeated for each sample. Meanwhile, the
effectiveness of the chemical test for determining Cu, Fe, and Mn
in cans alloy was verified by testing given samples of virgin beverage
cans body and pure aluminum. It proved the chemical detection
methods can reliably pick out the targeted elements from
the cans body. The chemical method was applied to a sample
(30–50% of total) of the density-separated aluminum scraps to
select the scraps that most likely originated from cans.