1166 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
a 5 L tank, fitted with a heat exchanger in Hastelloy, a cor-
rosion-resistant material, (iii) a gear pump in Hastelloy, up
to 250 L/h, (iv) a centrifugal pump, capable of covering a
pressure range from 5 to 15 bars. This allowed the leachate
to flow from the tank to the flat membrane housing, which
contained two nanofiltration membranes of 1.51 10–2 m2
(Suez (GE) JX (150 Da)) consisting of a thin active layer
(0.5 μm), allowing separation, as well as a support layer to
achieve satisfactory mechanical properties while reducing
the mass transfer resistance of the membrane (Zhu et al.,
2021). These membranes are specially designed to operate
continuously at sulfuric acid mass concentrations between
5 and 20% (between 0.5 and 2 M).
Each nanofiltration was conducted in parallel to test
reproducibility of the results. Equilibrium was reached after
one hour of filtration, temperature was maintained at 25 °C
by cryostat.
All the membranes were cleaned between experiments
by circulating ultrapure water at 25 °C at 15 bars for 1 hour.
Membrane Characterization
The characterization of the membrane helps to obtain some
crucial data needed to simulate the transport in a mecha-
nistic transport model. The pore size of the nanofiltration
membrane was determined by filtration of 3 PEGs (poly-
ethylene glycol, 95%, Sigma-Aldrich) of 200, 400 and 600
Da, at a concentration of 4g/L. Average pore sizes were
0.77, 0.64 and 0.69 nm respectively. To determine whether
average pore size changed during speciation, boric acid
((H3BO3, Sigma-Aldrich, 95%) at 4 g/L was added to the
leachate in preliminary experiments. Under these pH con-
ditions, it was neutral, so any changes in pore size could be
observed by analyzing its rejection rate.
Transport Model
The simulation model chosen for this study was the SEDE
model (Szymczyk and Fievet 2005). This model has two
advantages: (i) hydrometallurgical leachates have character-
istics (pH, concentration, and species diversity) that make
it difficult to determine an ionic permeance, especially if
this depends on the composition of the feed solution. A
mechanistic model, which does not take ion permeances
as input data, therefore offers an advantage (ii) the con-
struction of the SEDE model makes it possible to take into
account a large number of phenomena, but also to exclude
them selectively, in order to study the importance of each
of them in the rejection of an ion.
In order to couple the transport and the speciation
models (PHREEQC 3.0, USGS), the IphreeqcCOM mod-
ule (USGS) was used by imitating the geochemical model
approach (Muniruzzaman and Rolle 2016).
The simulation of the solution speciation is done before
and after the simulation of the transport. The calculation
model used to determine coefficients of activity was the SIT
(Specific Ion Theory) model (Brönsted 1922).
Main equations of the mathematical model are listed in
the supplementary material section and input and output
parameters of the system are resumed in Figure 1.
RESULTS AND DISCUSSION
Effect of Dilution on Nickel Rejection Rates
Nanofiltrations experiments have shown that up to 0.37
mol/kg, the decrease of the ionic strength led to an increase
in rejection rates, from 35% to 85% for the nickel (for
an ionic strength of 3.2 and 0.37 mol/kg respectively),
as shown in Figure 2. A further decrease in ionic strength
resulted in a lower rejection rate. Therefore, the experi-
ments showed optimum rejection optimum for both nickel
and magnesium for a leachate with an ionic strength of
0.37 mol/kg. This trend was followed by all solutions con-
taining divalent cations, which achieved similar rejection
rates (85% for magnesium and 87% for Fe). It is interesting
to note a very low rejection rate for sulfate ions, rejection
increasing nevertheless for ionic strengths from 0.3 to 0.15
mol/kg. This trend was followed by all monovalent ions in
the solution.
No pH variation was recorded between permeate and
retentate in any of the experiments. It is considered that
the H+ ions were not rejected by the membrane. However,
the concentration of all metals in the retentate allows the
acid in the permeate to be recycled. These results define
the range of ionic strength in which this nanofiltration is
further studied: from pure solution of 3.2 mol/kg to the
rejection rate optimum of .0.37 mol/kg.
Transport Model Through the Nanofiltration
Membrane
Mechanistic Study
Ion transfer across the membrane involves several effects:
steric hindrance, the pore swelling effect, the Donnan effect,
Table 1. Elemental composition of the ash leachate
Element Ni Mg Ca Al Fe Zn Mn K S
Concentration (mg/L) 8,790 4,760 660 170 500 30 20 500 144,000
a 5 L tank, fitted with a heat exchanger in Hastelloy, a cor-
rosion-resistant material, (iii) a gear pump in Hastelloy, up
to 250 L/h, (iv) a centrifugal pump, capable of covering a
pressure range from 5 to 15 bars. This allowed the leachate
to flow from the tank to the flat membrane housing, which
contained two nanofiltration membranes of 1.51 10–2 m2
(Suez (GE) JX (150 Da)) consisting of a thin active layer
(0.5 μm), allowing separation, as well as a support layer to
achieve satisfactory mechanical properties while reducing
the mass transfer resistance of the membrane (Zhu et al.,
2021). These membranes are specially designed to operate
continuously at sulfuric acid mass concentrations between
5 and 20% (between 0.5 and 2 M).
Each nanofiltration was conducted in parallel to test
reproducibility of the results. Equilibrium was reached after
one hour of filtration, temperature was maintained at 25 °C
by cryostat.
All the membranes were cleaned between experiments
by circulating ultrapure water at 25 °C at 15 bars for 1 hour.
Membrane Characterization
The characterization of the membrane helps to obtain some
crucial data needed to simulate the transport in a mecha-
nistic transport model. The pore size of the nanofiltration
membrane was determined by filtration of 3 PEGs (poly-
ethylene glycol, 95%, Sigma-Aldrich) of 200, 400 and 600
Da, at a concentration of 4g/L. Average pore sizes were
0.77, 0.64 and 0.69 nm respectively. To determine whether
average pore size changed during speciation, boric acid
((H3BO3, Sigma-Aldrich, 95%) at 4 g/L was added to the
leachate in preliminary experiments. Under these pH con-
ditions, it was neutral, so any changes in pore size could be
observed by analyzing its rejection rate.
Transport Model
The simulation model chosen for this study was the SEDE
model (Szymczyk and Fievet 2005). This model has two
advantages: (i) hydrometallurgical leachates have character-
istics (pH, concentration, and species diversity) that make
it difficult to determine an ionic permeance, especially if
this depends on the composition of the feed solution. A
mechanistic model, which does not take ion permeances
as input data, therefore offers an advantage (ii) the con-
struction of the SEDE model makes it possible to take into
account a large number of phenomena, but also to exclude
them selectively, in order to study the importance of each
of them in the rejection of an ion.
In order to couple the transport and the speciation
models (PHREEQC 3.0, USGS), the IphreeqcCOM mod-
ule (USGS) was used by imitating the geochemical model
approach (Muniruzzaman and Rolle 2016).
The simulation of the solution speciation is done before
and after the simulation of the transport. The calculation
model used to determine coefficients of activity was the SIT
(Specific Ion Theory) model (Brönsted 1922).
Main equations of the mathematical model are listed in
the supplementary material section and input and output
parameters of the system are resumed in Figure 1.
RESULTS AND DISCUSSION
Effect of Dilution on Nickel Rejection Rates
Nanofiltrations experiments have shown that up to 0.37
mol/kg, the decrease of the ionic strength led to an increase
in rejection rates, from 35% to 85% for the nickel (for
an ionic strength of 3.2 and 0.37 mol/kg respectively),
as shown in Figure 2. A further decrease in ionic strength
resulted in a lower rejection rate. Therefore, the experi-
ments showed optimum rejection optimum for both nickel
and magnesium for a leachate with an ionic strength of
0.37 mol/kg. This trend was followed by all solutions con-
taining divalent cations, which achieved similar rejection
rates (85% for magnesium and 87% for Fe). It is interesting
to note a very low rejection rate for sulfate ions, rejection
increasing nevertheless for ionic strengths from 0.3 to 0.15
mol/kg. This trend was followed by all monovalent ions in
the solution.
No pH variation was recorded between permeate and
retentate in any of the experiments. It is considered that
the H+ ions were not rejected by the membrane. However,
the concentration of all metals in the retentate allows the
acid in the permeate to be recycled. These results define
the range of ionic strength in which this nanofiltration is
further studied: from pure solution of 3.2 mol/kg to the
rejection rate optimum of .0.37 mol/kg.
Transport Model Through the Nanofiltration
Membrane
Mechanistic Study
Ion transfer across the membrane involves several effects:
steric hindrance, the pore swelling effect, the Donnan effect,
Table 1. Elemental composition of the ash leachate
Element Ni Mg Ca Al Fe Zn Mn K S
Concentration (mg/L) 8,790 4,760 660 170 500 30 20 500 144,000