1222 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
Quantitative XRD analysis of primary WHIMS mag-
netics was undertaken to enable an understanding of rela-
tive recoveries of individual minerals by size fraction.
Mineralogical evaluation of secondary WHIMS mag-
netics by ALS used QEMSCAN analysis on three supplied
fractions of cleaner WHIMS magnetics: +75, 38 to 75 and
minus 38 microns, the objectives being to assess the extent
of liberation of key minerals and their relative abundance.
The head analysis of the composite used for testing is pre-
sented as Table 1.
David Tube Recovery testing at 3000 gauss produced
very low magnetics yields, ranging from 0.8% at a P80 size
of 604 µm to 0.1% at 20 µm, demonstrating progressive
release of magnetics from gangue with increasing grinding
effort, but also indicating that a LIMS precursor stage was
not needed. If significant ferromagnetics content is present
it is prudent to use WHIMS before WHIMS processing to
avoid buildup in the WHIMS matrix which will ultimately
impair efficiency.
Relative mineral abundance for the test sample is
presented as Figure 1. The primary minerals are feldspars
(orthoclase and plagioclase predominantly), quartz, amphi-
bole, garnets and biotite. Allanite makes up 1.3% of the
total feed mass, with significant bias to the +212 micron
fraction, indicating coarse crystal structure. Trace amounts
of fluorocarbonate minerals bastnaesite and synchysite were
also detected.
Locking analysis determined that 87.5% of all allanite
exists as free, pure, or liberated forms (due to grinding), as
depicted in Figure 2. The remaining 12.5% of allanite is
associated with matrix minerals (intergrowths with silicate
gangue). SGS predicted a cerium grade of 9.3% for 94%
recovery, a lanthanum grade of 4.6% for 94% recovery and
a neodymium grade of 3.8% for 95% recovery. In prac-
tice, achieving such high upgrades would be difficult due to
inevitable operational losses as well as challenges associated
with minerals of similar properties, but the data indicated
good potential for upgrade through physical beneficiation.
This analysis however did not anticipate difficulties in sepa-
ration of allanite from iron-bearing gangue minerals.
Optical mineralogical work was undertaken which
identified the amphibole mineral present as hastingsite,
a member of the hornblende family. It was subsequently
found that hastingsite enriched along with allanite with
WHIMS, gravity separation and flotation. Chemical for-
mulae and physical properties for each mineral is presented
as follows (www.mindat.org):
• Allanite(Y):
(Y,Ce,Ca)2(Al,Fe3+)3(SiO4)3(OH)
• Hastingsite: N a C a
2 (F e 2 +4 F e 3 +)
Si6Al2O22(OH)2
Fe2O3 makes up the second highest mineral abundance in
allanite at 19.7%, after silica. This is unusually high as web
database mindat.org indicates a typical content of 10.5%.
Hastingsite typically contains 8.1% Fe2O3 but 29.0%
FeO, the latter being a reduced form of Fe. The mixed Fe(II)/
Fe(III) oxidation state of hastingsite is expected to have
ferromagnetic properties, akin to magnetite. Establishing
magnetic susceptibility of hastingsite is planned for future
work to compare against allanite. The high Fe content is
important to note when evaluating separation efficiency
from other Fe gangue minerals such as hastingsite since
total Fe is reported, not by mineral type.
Similarly, both allanite and hastingsite contain high
levels of silica (41.1% and 36.4% respectively) so measur-
ing success of gangue rejection based on silica content is
also made more complicated.
WHIMS Testing
Primary Separation
Four stage sighter WHIMS testing using a batch GZRINM
WHGMS145 test unit at grind P80 sizes of 500, 250 and
106 µm was undertaken to allow an optimum primary sep-
aration grind size to be established. Test conditions adopted
are summarised in Table 2.
Figure 3 provides graphical representations of recovery-
grade responses for the three grind sizes, plotting Fe and
TREO+Y separately.
TREO+Y recovery at 3000 gauss was surprisingly
high (50 to 61%) given that this is typically the realms of
magnetite and pyrrhotite, possibly maghemite (transition
between magnetite and hematite). The initial thought was
Table 1. Halleck Creek composite head analysis
Rare Earth
Oxide, ppm Value Gangue, %Value
Y2O3 221 SiO2 61.8
La2O3 751 Fetot 5.11
CeO
2 1583 FeO 5.20
Pr
6 O
11 189 Al
2 O
3 15.9
Nd2O3 644 P2O5 0.072
SEGs* 187 CaO 2.87
HREOs† 105 K
2 O 6.03
CREOs‡ 887 Na
2 O 4.24
TREO+Y 3668 TiO2 0.50
*Sum of samarium, europium and gadolinium oxides
† Oxides of terbium, dysprosium, holmium, erbium, thulium,
ytterbium and lutetium
‡ Critical rare earth oxides of Nd, Pr, Tb and Dy used in
supermagnet applications
Quantitative XRD analysis of primary WHIMS mag-
netics was undertaken to enable an understanding of rela-
tive recoveries of individual minerals by size fraction.
Mineralogical evaluation of secondary WHIMS mag-
netics by ALS used QEMSCAN analysis on three supplied
fractions of cleaner WHIMS magnetics: +75, 38 to 75 and
minus 38 microns, the objectives being to assess the extent
of liberation of key minerals and their relative abundance.
The head analysis of the composite used for testing is pre-
sented as Table 1.
David Tube Recovery testing at 3000 gauss produced
very low magnetics yields, ranging from 0.8% at a P80 size
of 604 µm to 0.1% at 20 µm, demonstrating progressive
release of magnetics from gangue with increasing grinding
effort, but also indicating that a LIMS precursor stage was
not needed. If significant ferromagnetics content is present
it is prudent to use WHIMS before WHIMS processing to
avoid buildup in the WHIMS matrix which will ultimately
impair efficiency.
Relative mineral abundance for the test sample is
presented as Figure 1. The primary minerals are feldspars
(orthoclase and plagioclase predominantly), quartz, amphi-
bole, garnets and biotite. Allanite makes up 1.3% of the
total feed mass, with significant bias to the +212 micron
fraction, indicating coarse crystal structure. Trace amounts
of fluorocarbonate minerals bastnaesite and synchysite were
also detected.
Locking analysis determined that 87.5% of all allanite
exists as free, pure, or liberated forms (due to grinding), as
depicted in Figure 2. The remaining 12.5% of allanite is
associated with matrix minerals (intergrowths with silicate
gangue). SGS predicted a cerium grade of 9.3% for 94%
recovery, a lanthanum grade of 4.6% for 94% recovery and
a neodymium grade of 3.8% for 95% recovery. In prac-
tice, achieving such high upgrades would be difficult due to
inevitable operational losses as well as challenges associated
with minerals of similar properties, but the data indicated
good potential for upgrade through physical beneficiation.
This analysis however did not anticipate difficulties in sepa-
ration of allanite from iron-bearing gangue minerals.
Optical mineralogical work was undertaken which
identified the amphibole mineral present as hastingsite,
a member of the hornblende family. It was subsequently
found that hastingsite enriched along with allanite with
WHIMS, gravity separation and flotation. Chemical for-
mulae and physical properties for each mineral is presented
as follows (www.mindat.org):
• Allanite(Y):
(Y,Ce,Ca)2(Al,Fe3+)3(SiO4)3(OH)
• Hastingsite: N a C a
2 (F e 2 +4 F e 3 +)
Si6Al2O22(OH)2
Fe2O3 makes up the second highest mineral abundance in
allanite at 19.7%, after silica. This is unusually high as web
database mindat.org indicates a typical content of 10.5%.
Hastingsite typically contains 8.1% Fe2O3 but 29.0%
FeO, the latter being a reduced form of Fe. The mixed Fe(II)/
Fe(III) oxidation state of hastingsite is expected to have
ferromagnetic properties, akin to magnetite. Establishing
magnetic susceptibility of hastingsite is planned for future
work to compare against allanite. The high Fe content is
important to note when evaluating separation efficiency
from other Fe gangue minerals such as hastingsite since
total Fe is reported, not by mineral type.
Similarly, both allanite and hastingsite contain high
levels of silica (41.1% and 36.4% respectively) so measur-
ing success of gangue rejection based on silica content is
also made more complicated.
WHIMS Testing
Primary Separation
Four stage sighter WHIMS testing using a batch GZRINM
WHGMS145 test unit at grind P80 sizes of 500, 250 and
106 µm was undertaken to allow an optimum primary sep-
aration grind size to be established. Test conditions adopted
are summarised in Table 2.
Figure 3 provides graphical representations of recovery-
grade responses for the three grind sizes, plotting Fe and
TREO+Y separately.
TREO+Y recovery at 3000 gauss was surprisingly
high (50 to 61%) given that this is typically the realms of
magnetite and pyrrhotite, possibly maghemite (transition
between magnetite and hematite). The initial thought was
Table 1. Halleck Creek composite head analysis
Rare Earth
Oxide, ppm Value Gangue, %Value
Y2O3 221 SiO2 61.8
La2O3 751 Fetot 5.11
CeO
2 1583 FeO 5.20
Pr
6 O
11 189 Al
2 O
3 15.9
Nd2O3 644 P2O5 0.072
SEGs* 187 CaO 2.87
HREOs† 105 K
2 O 6.03
CREOs‡ 887 Na
2 O 4.24
TREO+Y 3668 TiO2 0.50
*Sum of samarium, europium and gadolinium oxides
† Oxides of terbium, dysprosium, holmium, erbium, thulium,
ytterbium and lutetium
‡ Critical rare earth oxides of Nd, Pr, Tb and Dy used in
supermagnet applications