1224 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
Iron recovery also increased to over 90% indicating free
amphibole and hematite were being recovered.
Increasing to 10 000 gauss and beyond, the stage grade
and recovery fell away, which indicates co-recovery of
partially locked minerals and less magnetic iron minerals
such as goethite and the Fe-feldspar mineral clinochlore.
TREO+Y cumulative recovery tapered off due to falling
grades and stage mass yields. Rare earth minerals in these
stages were most likely partially locked with silica/silicates.
From the analysis, there did not appear to be any ben-
efit in grinding finer 500 µm P80 for the primary WHIMS
stage because the difference in mass yields was small (31.9%
for 500 µm vs. 29.1% for 250 µm). Gangue rejection was
65% of feed mass, which is unusually high for such a coarse
feed size, just inside the limit for efficient WHIMS opera-
tion, whilst still achieving nearly 94% TREO+Y recovery.
This offers a large saving in grinding energy for subsequent
regrind stages.
On conclusion of the sighter testing with the batch
WHGMS145 unit, 300 kg of ore was ground to a P80 of
500 µm and subjected to continuous WHIMS operation
using a Longi pulsating LGS500 continuous WHIMS
unit. Initially only rougher and single scavenger stages
were adopted, with field strengths of 3000 and 6000
gauss respectively due to the batch test results indicated
that most of the allanite is recovered by this ultimate field
intensity. However, it was found that with only two stages
of WHIMS, REO recovery was poor, and a decision was
made to include two additional scavenging stages to boot
yield and recovery. Even with these additional stages, over-
all TREO+Y recovery did not reach the levels achieved in
batch testing.
A plot of cumulative yield vs. cumulative TREO+Y
recovery is included as Figure 4 to illustrate that recovery
is largely linked to yield as the gradients of the two data
plots (batch and continuous) are similar. The grade of
Scavenger 3 magnetics was 9146 ppm TREO+Y, dropping
from Scavenger 2s 11693 ppm, but potentially could ben-
efit from further WHIMS processing. However, five stages
of WHIMS are impractical and expensive to implement,
pointing towards identifying a more efficient means of
recovering remaining liberated allanite.
Analysis of recovery of TREO+Y and gangue elements
by particle size grouping, presented as size class mid-points,
provides insight into where losses were occurring, depicted
in Figure 5.
Figure 2. Liberation of rare earth minerals by size fraction
Iron recovery also increased to over 90% indicating free
amphibole and hematite were being recovered.
Increasing to 10 000 gauss and beyond, the stage grade
and recovery fell away, which indicates co-recovery of
partially locked minerals and less magnetic iron minerals
such as goethite and the Fe-feldspar mineral clinochlore.
TREO+Y cumulative recovery tapered off due to falling
grades and stage mass yields. Rare earth minerals in these
stages were most likely partially locked with silica/silicates.
From the analysis, there did not appear to be any ben-
efit in grinding finer 500 µm P80 for the primary WHIMS
stage because the difference in mass yields was small (31.9%
for 500 µm vs. 29.1% for 250 µm). Gangue rejection was
65% of feed mass, which is unusually high for such a coarse
feed size, just inside the limit for efficient WHIMS opera-
tion, whilst still achieving nearly 94% TREO+Y recovery.
This offers a large saving in grinding energy for subsequent
regrind stages.
On conclusion of the sighter testing with the batch
WHGMS145 unit, 300 kg of ore was ground to a P80 of
500 µm and subjected to continuous WHIMS operation
using a Longi pulsating LGS500 continuous WHIMS
unit. Initially only rougher and single scavenger stages
were adopted, with field strengths of 3000 and 6000
gauss respectively due to the batch test results indicated
that most of the allanite is recovered by this ultimate field
intensity. However, it was found that with only two stages
of WHIMS, REO recovery was poor, and a decision was
made to include two additional scavenging stages to boot
yield and recovery. Even with these additional stages, over-
all TREO+Y recovery did not reach the levels achieved in
batch testing.
A plot of cumulative yield vs. cumulative TREO+Y
recovery is included as Figure 4 to illustrate that recovery
is largely linked to yield as the gradients of the two data
plots (batch and continuous) are similar. The grade of
Scavenger 3 magnetics was 9146 ppm TREO+Y, dropping
from Scavenger 2s 11693 ppm, but potentially could ben-
efit from further WHIMS processing. However, five stages
of WHIMS are impractical and expensive to implement,
pointing towards identifying a more efficient means of
recovering remaining liberated allanite.
Analysis of recovery of TREO+Y and gangue elements
by particle size grouping, presented as size class mid-points,
provides insight into where losses were occurring, depicted
in Figure 5.
Figure 2. Liberation of rare earth minerals by size fraction