XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 457
as indicated by the DE-XRT sensor response, were quite
marginal. This was particularly evident in the dynamic
test using the –4.0/+2.0 size fraction feed material, which
showed essentially no DE-XRT relative density variations
between ejected and unejected rocks, in contrast to the
more pronounced differences observed with the –2.0/+1.0
and –1.0/+0.5 size fractions. Elemental and mineralogi-
cal analyses of the sorting products from these tests (not
shown in the figure) further corroborated this observation,
displaying more distinct compositional variations in the
sorting products from tests E7 and E5b, which used the
–2.0/+1.0 and –1.0/+0.5 feed materials, respectively.
Based on these findings, the decision was made
to focus the production tests exclusively on the
–2.0/+1.0 and –1.0/+0.5 size fractions of the scandium ore
sample.
Production Tests with –2.0/+1.0 and –1.0/+0.5 Size
Fractions of the Scandium Ore
Table 5 details the ore sorter conditions and parameters
used in the production tests with –2.0/+1.0 and –1.0/+0.5
size fractions of the scandium ore conducted using the
processing schemes shown in Figure 7. Metallurgical bal-
ances for the production tests are provided in Table 6, while
Figure 8 presents the cumulative recovery vs. mass pull
curves obtained for selected elements (Si, Al, Fe, Mg, Ca,
Na, Sc) in comparison to the main mineral groups found
within the ore, i.e., iron-bearing minerals (Fe-MIN) and
feldspars (FSP). In Table 6, the total concentration of REEs
other than Sc was quantified as REE-britholite (REB) with
formula (REE)5(SiO4)3(OH).
Table 6 indicates that the concentrations of scandium
(Sc) and rare earth elements (REEs) expressed as REE-
britholite in the sorting products show no significant varia-
tions, suggesting that the sorting process did not effectively
upgrade these elements. Figure 8 supports this, revealing
only a marginal preconcentration of iron-bearing miner-
als (Fe-MIN) and a slight prerejection of feldspars (FSP)
in both size fractions during the production tests. In light
of these outcomes, conducting further particle sorting pro-
duction tests at larger scales is not recommended. However,
to better understand the limited sorting efficiency observed
in these tests, it is proposed to conduct heterogeneity tests
on batches of 50–100 rocks from one or more of the inves-
tigated size fractions. The aim of these tests would be to
assess the compositional variability among individual rocks
within a specific size fraction of the scandium ore. Such an
analysis could help determine whether the marginal sorting
success was indeed a result of minimal compositional het-
erogeneity among the rocks in each size fraction.
REE Carbonatite Ore Sample
Scanning of Rocks from Distinct Size Fractions with the
Various Sensors
Results of the scanning of rocks from the –4.0/+3.0,
–3.0/+1.5, and –1.5/+0.5 size fractions of the REE ore
sample with the DE-XRT, color, VIS-NIR, and SWIR sen-
sors are presented in Figure 9.
The DE-XRT analysis (Figure 9a-c) revealed moderate
potential for sorting, as some rock-to-rock differences were
noticeable between particles within a given size fraction.
However, it was often challenging to determine if these dif-
ferences were due to changes in the rocks’ relative density or
variations in particle thickness.
In contrast, the analysis of raw images from the color
sensor (Figure 9d-f) proved unsuccessful, making it diffi-
cult to draw any meaningful conclusions from these images.
Similarly, while image analysis was feasible for the hyper-
spectral images captured by the VIS-NIR (Figure 9g-i)
and SWIR (Figure 9j-l) sensors, the results indicated that
Table 5. Conditions and parameters of production tests with size fractions of the scandium ore
Production Test with –2.0/+1.0 Size Fraction Production Test with –1.0/+0.5 Size Fraction
Separation
Run No.
DE-XRT Separation
Run No.
DE-XRT
RD* %Area† RD* %Area†
1 105−230 55−100 1 115−230 55−100
2 105−230 50−100 2 105−230 50−100
3 100−230 50−100 3 105−230 40−100
4 95−230 50−100 4 95−230 50−100
5 95−230 45−100 5 95−230 40−100
6 85−230 45−100 6 85−230 50−100
7 75−230 50−100
8 55−230 40−100
*RD parameter as defined in Table 1.
† %area parameter of the DE-XRT sensor as defined in Table 1
as indicated by the DE-XRT sensor response, were quite
marginal. This was particularly evident in the dynamic
test using the –4.0/+2.0 size fraction feed material, which
showed essentially no DE-XRT relative density variations
between ejected and unejected rocks, in contrast to the
more pronounced differences observed with the –2.0/+1.0
and –1.0/+0.5 size fractions. Elemental and mineralogi-
cal analyses of the sorting products from these tests (not
shown in the figure) further corroborated this observation,
displaying more distinct compositional variations in the
sorting products from tests E7 and E5b, which used the
–2.0/+1.0 and –1.0/+0.5 feed materials, respectively.
Based on these findings, the decision was made
to focus the production tests exclusively on the
–2.0/+1.0 and –1.0/+0.5 size fractions of the scandium ore
sample.
Production Tests with –2.0/+1.0 and –1.0/+0.5 Size
Fractions of the Scandium Ore
Table 5 details the ore sorter conditions and parameters
used in the production tests with –2.0/+1.0 and –1.0/+0.5
size fractions of the scandium ore conducted using the
processing schemes shown in Figure 7. Metallurgical bal-
ances for the production tests are provided in Table 6, while
Figure 8 presents the cumulative recovery vs. mass pull
curves obtained for selected elements (Si, Al, Fe, Mg, Ca,
Na, Sc) in comparison to the main mineral groups found
within the ore, i.e., iron-bearing minerals (Fe-MIN) and
feldspars (FSP). In Table 6, the total concentration of REEs
other than Sc was quantified as REE-britholite (REB) with
formula (REE)5(SiO4)3(OH).
Table 6 indicates that the concentrations of scandium
(Sc) and rare earth elements (REEs) expressed as REE-
britholite in the sorting products show no significant varia-
tions, suggesting that the sorting process did not effectively
upgrade these elements. Figure 8 supports this, revealing
only a marginal preconcentration of iron-bearing miner-
als (Fe-MIN) and a slight prerejection of feldspars (FSP)
in both size fractions during the production tests. In light
of these outcomes, conducting further particle sorting pro-
duction tests at larger scales is not recommended. However,
to better understand the limited sorting efficiency observed
in these tests, it is proposed to conduct heterogeneity tests
on batches of 50–100 rocks from one or more of the inves-
tigated size fractions. The aim of these tests would be to
assess the compositional variability among individual rocks
within a specific size fraction of the scandium ore. Such an
analysis could help determine whether the marginal sorting
success was indeed a result of minimal compositional het-
erogeneity among the rocks in each size fraction.
REE Carbonatite Ore Sample
Scanning of Rocks from Distinct Size Fractions with the
Various Sensors
Results of the scanning of rocks from the –4.0/+3.0,
–3.0/+1.5, and –1.5/+0.5 size fractions of the REE ore
sample with the DE-XRT, color, VIS-NIR, and SWIR sen-
sors are presented in Figure 9.
The DE-XRT analysis (Figure 9a-c) revealed moderate
potential for sorting, as some rock-to-rock differences were
noticeable between particles within a given size fraction.
However, it was often challenging to determine if these dif-
ferences were due to changes in the rocks’ relative density or
variations in particle thickness.
In contrast, the analysis of raw images from the color
sensor (Figure 9d-f) proved unsuccessful, making it diffi-
cult to draw any meaningful conclusions from these images.
Similarly, while image analysis was feasible for the hyper-
spectral images captured by the VIS-NIR (Figure 9g-i)
and SWIR (Figure 9j-l) sensors, the results indicated that
Table 5. Conditions and parameters of production tests with size fractions of the scandium ore
Production Test with –2.0/+1.0 Size Fraction Production Test with –1.0/+0.5 Size Fraction
Separation
Run No.
DE-XRT Separation
Run No.
DE-XRT
RD* %Area† RD* %Area†
1 105−230 55−100 1 115−230 55−100
2 105−230 50−100 2 105−230 50−100
3 100−230 50−100 3 105−230 40−100
4 95−230 50−100 4 95−230 50−100
5 95−230 45−100 5 95−230 40−100
6 85−230 45−100 6 85−230 50−100
7 75−230 50−100
8 55−230 40−100
*RD parameter as defined in Table 1.
† %area parameter of the DE-XRT sensor as defined in Table 1