1168 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
of significance, depending on ionic strength. Thus, for each
ionic strength, the following effects were considered to
influence the separation:
• For a leachate of 3.2 mol/kg, the steric hindrance
and pore swelling effect were considered. The ionic
strength was too high for the Donnan effect to influ-
ence separation. Indeed, the variation of Donnan
potential is too low to have any effect on the separa-
tion (0.1 mV)
• For a leachate of 1.6 mol/kg, only steric hindrance
was considered, because no pore swelling was experi-
mentally recorded. The variation of Donnan poten-
tial is considered negligeable.
• For a leachate of 0.7 mol/kg, only steric hindrance
was considered. The variation of Donnan potential is
still too low to be considered
• For a leachate of 0.36 mol/kg, steric hindrance, the
Donnan effect and the image force effect were con-
sidered. For this solution, the variation of Donnan
potential has raised to 20 mV, explaining the impor-
tance of this effect. The choice of a low membrane
dielectric constant explains the high rejection rate
obtain while modelling the image force effect, but
it is considered a realistic value for polyamide mem-
brane. (Fridman-Bishop and Freger 2017).
Speciation Effect
The speciation of the starting solution was calculated in the
feed solution, the retentate and the permeate to determine
whether its knowledge improved the simulation results.
Simulations with the reactive transport model (with
speciation) and with the transport model (without specia-
tion) were compared. For an ionic strength of 3.2 mol/kg,
the difference between the theoretical and experimental
rejection for the three cations considered were respectively
5, 17 and 3.5% for nickel, magnesium and iron when feed
speciation was modeled, against 9, 19 and 8% when it was
ignored.
Figure 3 shows the contribution of speciation to the
simulation of ion transport at 0.36 mol/kg. Compared to
the previous example, it is possible to observe a larger gap
between the two models. Indeed, while a reactive transport
model shows a difference between theoretical and exper-
imental rejection of 2, 3 and 3% respectively for nickel,
magnesium and iron, a transport model which doesn’t
take into account speciation presents a simulation much
further from reality, with a theoretical/experimental devia-
tion of 10, 11 and 10% respectively. The interest of a reac-
tive transport model becomes even more apparent when
the ionic strength decreases and the Donnan and Image
force effect are present. This is due to the importance of
the charge of ions through the membrane to calculate their
rejection. Indeed, since the membrane is positively charged
in an acidic medium, highly positively charged ions are
much more effectively rejected than negatively or weakly
charged ones.
CONCLUSION
With the arrival of new membranes resistant to extreme
leachate conditions in the hydrometallurgical industry,
nanofiltration is now a possible process to recycle hyperac-
cumulating plant ash leachate acid (ionic strength between
3.2 and 0.15 mol/kg) while concentrating divalent ele-
ments in a retentate with rejection rates greater than 80%
in some cases. An optimum ionic strength of 0.36 mol/
kg has been determined for this process. Rejection rates
could reach values greater than 80% for divalent elements,
demonstrating the feasibility of a membrane acid recycling
process for leachates with high ionic strength. To identify
the predominant phenomena and propose a simulation of
the nanofiltration of such a solution, the SEDE model has
been selected.
The addition of input solution speciation to the vari-
ous phenomena commonly encountered in transport mod-
els dedicated to membrane separations enabled satisfactory
simulations of the nanofiltrations performed. Several les-
sons can be drawn from these observations, and will be
valid for the analysis of agromining leachates:
Table 2. Rejection rates of nickel, magnesium and iron,
modelled by the reactive transport without speciation
simulation
Element
Ionic
Strength,
mol/kg
Experimental
Rejection Rates,
%
Simulated
Rejection Rates,
%
Ni 3.2 35 44
1.6 63 73
0.7 76 75
0.36 85 95
Mg 3.2 31 50
1.6 59 78
0.7 75 82
0.36 85 96
Fe 3.2 43 51
1.6 65 77
0.7 79 80
0.36 87 96
of significance, depending on ionic strength. Thus, for each
ionic strength, the following effects were considered to
influence the separation:
• For a leachate of 3.2 mol/kg, the steric hindrance
and pore swelling effect were considered. The ionic
strength was too high for the Donnan effect to influ-
ence separation. Indeed, the variation of Donnan
potential is too low to have any effect on the separa-
tion (0.1 mV)
• For a leachate of 1.6 mol/kg, only steric hindrance
was considered, because no pore swelling was experi-
mentally recorded. The variation of Donnan poten-
tial is considered negligeable.
• For a leachate of 0.7 mol/kg, only steric hindrance
was considered. The variation of Donnan potential is
still too low to be considered
• For a leachate of 0.36 mol/kg, steric hindrance, the
Donnan effect and the image force effect were con-
sidered. For this solution, the variation of Donnan
potential has raised to 20 mV, explaining the impor-
tance of this effect. The choice of a low membrane
dielectric constant explains the high rejection rate
obtain while modelling the image force effect, but
it is considered a realistic value for polyamide mem-
brane. (Fridman-Bishop and Freger 2017).
Speciation Effect
The speciation of the starting solution was calculated in the
feed solution, the retentate and the permeate to determine
whether its knowledge improved the simulation results.
Simulations with the reactive transport model (with
speciation) and with the transport model (without specia-
tion) were compared. For an ionic strength of 3.2 mol/kg,
the difference between the theoretical and experimental
rejection for the three cations considered were respectively
5, 17 and 3.5% for nickel, magnesium and iron when feed
speciation was modeled, against 9, 19 and 8% when it was
ignored.
Figure 3 shows the contribution of speciation to the
simulation of ion transport at 0.36 mol/kg. Compared to
the previous example, it is possible to observe a larger gap
between the two models. Indeed, while a reactive transport
model shows a difference between theoretical and exper-
imental rejection of 2, 3 and 3% respectively for nickel,
magnesium and iron, a transport model which doesn’t
take into account speciation presents a simulation much
further from reality, with a theoretical/experimental devia-
tion of 10, 11 and 10% respectively. The interest of a reac-
tive transport model becomes even more apparent when
the ionic strength decreases and the Donnan and Image
force effect are present. This is due to the importance of
the charge of ions through the membrane to calculate their
rejection. Indeed, since the membrane is positively charged
in an acidic medium, highly positively charged ions are
much more effectively rejected than negatively or weakly
charged ones.
CONCLUSION
With the arrival of new membranes resistant to extreme
leachate conditions in the hydrometallurgical industry,
nanofiltration is now a possible process to recycle hyperac-
cumulating plant ash leachate acid (ionic strength between
3.2 and 0.15 mol/kg) while concentrating divalent ele-
ments in a retentate with rejection rates greater than 80%
in some cases. An optimum ionic strength of 0.36 mol/
kg has been determined for this process. Rejection rates
could reach values greater than 80% for divalent elements,
demonstrating the feasibility of a membrane acid recycling
process for leachates with high ionic strength. To identify
the predominant phenomena and propose a simulation of
the nanofiltration of such a solution, the SEDE model has
been selected.
The addition of input solution speciation to the vari-
ous phenomena commonly encountered in transport mod-
els dedicated to membrane separations enabled satisfactory
simulations of the nanofiltrations performed. Several les-
sons can be drawn from these observations, and will be
valid for the analysis of agromining leachates:
Table 2. Rejection rates of nickel, magnesium and iron,
modelled by the reactive transport without speciation
simulation
Element
Ionic
Strength,
mol/kg
Experimental
Rejection Rates,
%
Simulated
Rejection Rates,
%
Ni 3.2 35 44
1.6 63 73
0.7 76 75
0.36 85 95
Mg 3.2 31 50
1.6 59 78
0.7 75 82
0.36 85 96
Fe 3.2 43 51
1.6 65 77
0.7 79 80
0.36 87 96