1232 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
zircon ahead of direct flotation of minerals such as colum-
bite or pyrochlore (Kanyika, Malawi), and for barite flo-
tation ahead of bastnaesite flotation (Ngualla, Tanzania).
Testing on feed material produced poor selectivity due to
the high proportion of feldspars and silica and was aban-
doned early in the ore testing part of the program.
Cleaner WHIMS Magnetics Flotation
Earlier promising results with direct flotation of ore were
not repeated with cleaner WHIMS magnetics because the
majority of non-Fe bearing minerals had been rejected by
WHIMS, leaving primarily a hastingsite-rich concentrate
with allanite and minor gangue impurities. Direct flota-
tion of WHIMS magnetics demonstrated that floatabil-
ity of hastingsite exceeded that of allanite. The best result
achieved was a 38% stage yield of 28% mass at a grade of
1.58% TREO for 46.4% stage recovery, using sodium ole-
ate as the rare earth collector. These poor results prompted
a revisit of reverse silica flotation to reject hastingsite.
For reverse silica flotation, the best result achieved
was a “sinks” (value stream) of grade 2.46% TREO+Y for
40.7% recovery in 25.9% mass yield. The target conditions
for cleaner WHIMS magnetics flotation were 72% recovery
into 31% stage mass yield, neither of which were achieved
due to competition with hastingsite.
The following conclusions are summarised from the
flotation testwork program:
• Whilst direct flotation of whole ore feed was suc-
cessful in rejecting substantial feed mass and still
achieve high REO+Y recoveries, this appears to be
less successful on WHIMS magnetics. Most of the
non-ferruginous free silica and felspar minerals were
rejected, leaving iron silicate minerals, predomi-
nantly hastingsite.
• Though of different silicate groups (hastingsite is an
inosilicate of the hastingsite family and allanite is an
orthosilicate of the epidote family), they are similar
enough to be extremely challenging to separate with
available direct flotation collectors and modifiers. As
a result of their similarity, determining the extent
of rejection of hastingsite is difficult from chemical
analyses alone as they both contain, Fe, Al, Ca and
Si, x-ray diffraction (XRD) or QEMSCAN are the
only means of determining effectiveness in flotation.
Figure 11. WHIMS magnetics flotation product yield vs. TREO+Y recovery
zircon ahead of direct flotation of minerals such as colum-
bite or pyrochlore (Kanyika, Malawi), and for barite flo-
tation ahead of bastnaesite flotation (Ngualla, Tanzania).
Testing on feed material produced poor selectivity due to
the high proportion of feldspars and silica and was aban-
doned early in the ore testing part of the program.
Cleaner WHIMS Magnetics Flotation
Earlier promising results with direct flotation of ore were
not repeated with cleaner WHIMS magnetics because the
majority of non-Fe bearing minerals had been rejected by
WHIMS, leaving primarily a hastingsite-rich concentrate
with allanite and minor gangue impurities. Direct flota-
tion of WHIMS magnetics demonstrated that floatabil-
ity of hastingsite exceeded that of allanite. The best result
achieved was a 38% stage yield of 28% mass at a grade of
1.58% TREO for 46.4% stage recovery, using sodium ole-
ate as the rare earth collector. These poor results prompted
a revisit of reverse silica flotation to reject hastingsite.
For reverse silica flotation, the best result achieved
was a “sinks” (value stream) of grade 2.46% TREO+Y for
40.7% recovery in 25.9% mass yield. The target conditions
for cleaner WHIMS magnetics flotation were 72% recovery
into 31% stage mass yield, neither of which were achieved
due to competition with hastingsite.
The following conclusions are summarised from the
flotation testwork program:
• Whilst direct flotation of whole ore feed was suc-
cessful in rejecting substantial feed mass and still
achieve high REO+Y recoveries, this appears to be
less successful on WHIMS magnetics. Most of the
non-ferruginous free silica and felspar minerals were
rejected, leaving iron silicate minerals, predomi-
nantly hastingsite.
• Though of different silicate groups (hastingsite is an
inosilicate of the hastingsite family and allanite is an
orthosilicate of the epidote family), they are similar
enough to be extremely challenging to separate with
available direct flotation collectors and modifiers. As
a result of their similarity, determining the extent
of rejection of hastingsite is difficult from chemical
analyses alone as they both contain, Fe, Al, Ca and
Si, x-ray diffraction (XRD) or QEMSCAN are the
only means of determining effectiveness in flotation.
Figure 11. WHIMS magnetics flotation product yield vs. TREO+Y recovery