XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 2651
to either falcon centrifugal separation or shaking table con-
centration. The overall tail was then merged to produce the
feed sample for flotation.
Collector used for flotation is a 97%-purity 3-(iso-
decyloxy)propylamine supplied by Ceca (France). This
etheramine was solubilised using analytical grade acid ace-
tic (Sigma Aldrich, USA) at 1:1 molar ratio. pH modifiers
used are H2O4 (97%, Sigma-Aldrich, USA) and NaOH
(99%, VWR, USA). All reagents were prepared using milli-
Q water, resistivity of 18.2 MΩ/cm (Aquadem, Veolia,
France).
Flotation tests were carried out on 500g of ore in a
LA-500 rotor-stator apparatus (Agitair) with a 1.5 L cell.
Conditioning and flotation were both carried out in the
flotation cell with solid ratio of 66% and 25%, respectively,
with a rotor speed of 800 rotation per minute at room tem-
perature (20–25°C). Ore samples were conditioned with
tap water and and the desired doasage of etheramine for 5
min. The pH was set at the selected pH value (± 0.1) and
monitored using a 3310 pH meter equipped with a sentix
41 pH probe (WTW, Germany). The air flow rate during
flotation was set at 0.27 m3/h. Froth was manually recov-
ered every 1–2s during 3 minutes. Then all products were
dried in an oven overnight and recovered for both chemical
analysis and mineral characterisation.
Particle Size Distribution
Particle size distribution (PSD) was acquired using laser
light scattering method with a Helium-Neon Laser Optical
System Mastersizer 3000 (Malvern Panalytical, Malvern,
United-Kingdom). One analysis consist of the average value
of 5 measurements. For this study, all samples (crushed core
samples, underflow and overflow of the hydrocycloning
step and flotation products) were analysed to get their PSD.
Energy Dispersive X-Ray Fluorescence
Samples were assayed by X-ray fluorescence using a Niton
XL3t GOLDD+ spectrometer (Thermo Scientific, USA).
An Ag-anode X-ray tube with a voltage of 50 kV is used as
X-rays source and X-rays are detected using a silicon drift
detector (SDD). An assay lasts 120s divided in four filters
(i.e., Main, Low, High and Light). The spectra are inter-
preted by the Mining Hf/Ta mode of the device.
RESULTS
Comminution
Aliquots were sampled during the pilot test and showed
promising results, as both the amount of material and the
solid flow rate imposed a total operating time that was too
short for good material equilibrium in the pilot. This can
be seen by looking at the flow rate of the ball mill (Table 4).
Desliming
Partition curves were built to see the effective cut-off of the
desliming stage. They are displayed in Figure 2. From these
plots, it appears that the slope of the partition curves for the
sieved samples is greater than for the sample submitted to
hydrocyclone (Figure 2). This implies that the cut-off of the
sample is narrower for sieved samples.
From these plots, the effective cut-off of each sample is
about 7, 45 and about 80µm. The difference between the
observed and expected cut-off size can be linked to (i) the
shape factor of lepdiolite which can biais the sieving step
and (ii) the measurement of the PSD. It was observed that
for coarse samples, a biais can be introduced with laser light
scattering method producing coarser PSDs than the one
observed by standard sieving tests.
Flotation tests
Physico-Chemical Conditions
The design of experiments was used to model both the
recovery of lithium and the enrichment ratio of lithium in
the floated products. The statistical analysis of the DOE is
given in Table 5. From this table, it appears that the both
the recovery and the enrichment ratio can be efficiently
modelled from experimental observations. Both responses
have acceptable R2 values for both low RMSE and p-values
(Table 5). Therefore, such models can be used to predict the
recovery and enrichment ratio during flotation of lepidolite
and can also be used to build surface responses along the
studied ranges.
Table 4. Streams sampled during the pilot test along with their flow rate and solid percent
Stream Feed In Ball Mill Out Ball Mill Thickener
Flow (kg/h) 48.8 52 63 49.7
%Solid 100 60 60 8
%Li
2 O 1.06 1.36 1.32 1.04
Sn (ppm) 825 759 720 803
Nb (ppm) 96 79 91 96
Ta (ppm) 172 132 128 158
to either falcon centrifugal separation or shaking table con-
centration. The overall tail was then merged to produce the
feed sample for flotation.
Collector used for flotation is a 97%-purity 3-(iso-
decyloxy)propylamine supplied by Ceca (France). This
etheramine was solubilised using analytical grade acid ace-
tic (Sigma Aldrich, USA) at 1:1 molar ratio. pH modifiers
used are H2O4 (97%, Sigma-Aldrich, USA) and NaOH
(99%, VWR, USA). All reagents were prepared using milli-
Q water, resistivity of 18.2 MΩ/cm (Aquadem, Veolia,
France).
Flotation tests were carried out on 500g of ore in a
LA-500 rotor-stator apparatus (Agitair) with a 1.5 L cell.
Conditioning and flotation were both carried out in the
flotation cell with solid ratio of 66% and 25%, respectively,
with a rotor speed of 800 rotation per minute at room tem-
perature (20–25°C). Ore samples were conditioned with
tap water and and the desired doasage of etheramine for 5
min. The pH was set at the selected pH value (± 0.1) and
monitored using a 3310 pH meter equipped with a sentix
41 pH probe (WTW, Germany). The air flow rate during
flotation was set at 0.27 m3/h. Froth was manually recov-
ered every 1–2s during 3 minutes. Then all products were
dried in an oven overnight and recovered for both chemical
analysis and mineral characterisation.
Particle Size Distribution
Particle size distribution (PSD) was acquired using laser
light scattering method with a Helium-Neon Laser Optical
System Mastersizer 3000 (Malvern Panalytical, Malvern,
United-Kingdom). One analysis consist of the average value
of 5 measurements. For this study, all samples (crushed core
samples, underflow and overflow of the hydrocycloning
step and flotation products) were analysed to get their PSD.
Energy Dispersive X-Ray Fluorescence
Samples were assayed by X-ray fluorescence using a Niton
XL3t GOLDD+ spectrometer (Thermo Scientific, USA).
An Ag-anode X-ray tube with a voltage of 50 kV is used as
X-rays source and X-rays are detected using a silicon drift
detector (SDD). An assay lasts 120s divided in four filters
(i.e., Main, Low, High and Light). The spectra are inter-
preted by the Mining Hf/Ta mode of the device.
RESULTS
Comminution
Aliquots were sampled during the pilot test and showed
promising results, as both the amount of material and the
solid flow rate imposed a total operating time that was too
short for good material equilibrium in the pilot. This can
be seen by looking at the flow rate of the ball mill (Table 4).
Desliming
Partition curves were built to see the effective cut-off of the
desliming stage. They are displayed in Figure 2. From these
plots, it appears that the slope of the partition curves for the
sieved samples is greater than for the sample submitted to
hydrocyclone (Figure 2). This implies that the cut-off of the
sample is narrower for sieved samples.
From these plots, the effective cut-off of each sample is
about 7, 45 and about 80µm. The difference between the
observed and expected cut-off size can be linked to (i) the
shape factor of lepdiolite which can biais the sieving step
and (ii) the measurement of the PSD. It was observed that
for coarse samples, a biais can be introduced with laser light
scattering method producing coarser PSDs than the one
observed by standard sieving tests.
Flotation tests
Physico-Chemical Conditions
The design of experiments was used to model both the
recovery of lithium and the enrichment ratio of lithium in
the floated products. The statistical analysis of the DOE is
given in Table 5. From this table, it appears that the both
the recovery and the enrichment ratio can be efficiently
modelled from experimental observations. Both responses
have acceptable R2 values for both low RMSE and p-values
(Table 5). Therefore, such models can be used to predict the
recovery and enrichment ratio during flotation of lepidolite
and can also be used to build surface responses along the
studied ranges.
Table 4. Streams sampled during the pilot test along with their flow rate and solid percent
Stream Feed In Ball Mill Out Ball Mill Thickener
Flow (kg/h) 48.8 52 63 49.7
%Solid 100 60 60 8
%Li
2 O 1.06 1.36 1.32 1.04
Sn (ppm) 825 759 720 803
Nb (ppm) 96 79 91 96
Ta (ppm) 172 132 128 158