956 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
configuration. The difference in chemical affinity is usually
quantified as separation factors. As a result, the extraction
and purification process for REEs typically involves numer-
ous stages with intricate circuitry. In a conventional REE
SX plant, up to 1,000 mixer settlers may be needed for full
separation of all REEs. These large and complex plants are
a notable barrier to entry, and as such, many separation
facilities opt to only perform partial separation of products
for a feedstock tuned to an REE distribution for a specific
ore deposit.
Owing to the large number of equipment and com-
plexity of the SX circuits, these operations are notori-
ously difficult to design and optimize. Traditional SX
design procedures use empirical analysis of batch labora-
tory tests, engineers experience, and piloting to estimate
the number of units needed to obtain a desired degree
of separation. Textbooks often cite the McCabe-Theile
approach, which can be used for the analogous process of
distillation. However, in the context of solvent extraction
for REE, it has been recognized that the McCabe-Thiele
method, traditionally employed for binary distillation,
is unsuitable. This is attributed to several critical factors.
Non-ideal behavior and complex chemical interactions are
exhibited by REE systems, diverging significantly from the
ideal assumptions upon which McCabe-Thiele’s method is
predicated. Furthermore, these separations involve multi-
component mixtures rather than simple binary systems,
a complexity not accommodated by McCabe-Thiele dia-
grams. Additionally, the intricate chemical equilibria, varia-
tions in phase behavior, and kinetic factors crucial in these
processes are not addressed by the method. Therefore, the
necessity for more sophisticated modeling approaches and
experimental validation for the design and optimization of
REE separation circuits has been established. Continuous
pilot trials can be long, tedious, and expensive, while failing
to yield an optimal solution. Chemical engineering prin-
ciples suggest that three to four retention times per reac-
tor are needed to obtain steady state, and this time scales
significantly when given the large number of reactors and
the dilute, but accumulating, recycle streams that are com-
mon in SX circuits. One study by SGS showed that after 21
days of operation, a pilot REE extraction circuit had not yet
reached steady state, thus calling into question the stability
and viability of the results (Bourricaudy et al., 2016).
One means to obviate this challenge is through the
use of advanced model-based process design and optimiza-
tion. Over the past several decades, many researchers have
developed empirical and phenomenological solvent extrac-
tion models. Koermer (2022) has provided a brief review
of the various modeling approaches, delineating them by
their method and theoretical basis. Many are fundamen-
tally derived from the Nernst Distribution Law but assume
constant behavior accross a wide variety of concentra-
tions. Moreover, some of the older techniques, such as the
McCabe-Thiele approach are appropriate for simple, single
or few element systems, but break down, either function-
ally or practically when expanded to the large, multi-ele-
ment problem associated with REE separations. Others
assume constant distribution ratio or constant separation
factor across a range of input conditions, and thus become
error prone as the solution concentrations and other factors
change in as solvent extraction battery.
A more recent modeling technique, proposed by
Turgeon, describes a fundamental approach based on
chemical equilibrium (Turgeon et al., 2023). Since the
model is based on fundamental reaction chemistry, it is
both robust and potentially scalable for large-scale REE
separation problems. In prior work, Turgeon has shown
both the utility of the method for circuit and plant-scale
simulation as well as methods for producing the requisite
data from laboratory analyses (Turgeon et al., 2016).
Building from the Turgeon model, we have integrated
the fundamental modeling approach into a linear circuit
analysis framework to provide a robust framework for com-
plex circuit design and optimization. As a solution meth-
odology, linear circuit analysis provides a consistent and
scaleable approach to define the superstructure of the pro-
cess circuitry in a manner that allows arbitrary complexity.
The combined Turgeon-LCA approach thus rapidly pro-
duces circuit solutions and provides a rigorous basis for cir-
cuit improvement and optimization. The objective of this
paper is, thus to describe the combined methodology and
discuss its application.
METHODOLOGY
Equilibrium Calculations
Details on the Turgeon modeling approach can be found in
Turgeon et al. (2016), and Turgeon et al. (2023) however,
salient points are summarized here for the benefit of the
reader.
In this approach, the solvent extraction system can be
described as a series of reversible chemical reactions involv-
ing the exchange of cations between phases of varying mis-
cibility. Generally:
Ln RH LnR 3 3H 3+
3 E +++
where Ln refers to any general REE in an aqueous phase
and RH refers to an extractant in an organic phase. The
aqueous is mixed with the organic where the REE will
configuration. The difference in chemical affinity is usually
quantified as separation factors. As a result, the extraction
and purification process for REEs typically involves numer-
ous stages with intricate circuitry. In a conventional REE
SX plant, up to 1,000 mixer settlers may be needed for full
separation of all REEs. These large and complex plants are
a notable barrier to entry, and as such, many separation
facilities opt to only perform partial separation of products
for a feedstock tuned to an REE distribution for a specific
ore deposit.
Owing to the large number of equipment and com-
plexity of the SX circuits, these operations are notori-
ously difficult to design and optimize. Traditional SX
design procedures use empirical analysis of batch labora-
tory tests, engineers experience, and piloting to estimate
the number of units needed to obtain a desired degree
of separation. Textbooks often cite the McCabe-Theile
approach, which can be used for the analogous process of
distillation. However, in the context of solvent extraction
for REE, it has been recognized that the McCabe-Thiele
method, traditionally employed for binary distillation,
is unsuitable. This is attributed to several critical factors.
Non-ideal behavior and complex chemical interactions are
exhibited by REE systems, diverging significantly from the
ideal assumptions upon which McCabe-Thiele’s method is
predicated. Furthermore, these separations involve multi-
component mixtures rather than simple binary systems,
a complexity not accommodated by McCabe-Thiele dia-
grams. Additionally, the intricate chemical equilibria, varia-
tions in phase behavior, and kinetic factors crucial in these
processes are not addressed by the method. Therefore, the
necessity for more sophisticated modeling approaches and
experimental validation for the design and optimization of
REE separation circuits has been established. Continuous
pilot trials can be long, tedious, and expensive, while failing
to yield an optimal solution. Chemical engineering prin-
ciples suggest that three to four retention times per reac-
tor are needed to obtain steady state, and this time scales
significantly when given the large number of reactors and
the dilute, but accumulating, recycle streams that are com-
mon in SX circuits. One study by SGS showed that after 21
days of operation, a pilot REE extraction circuit had not yet
reached steady state, thus calling into question the stability
and viability of the results (Bourricaudy et al., 2016).
One means to obviate this challenge is through the
use of advanced model-based process design and optimiza-
tion. Over the past several decades, many researchers have
developed empirical and phenomenological solvent extrac-
tion models. Koermer (2022) has provided a brief review
of the various modeling approaches, delineating them by
their method and theoretical basis. Many are fundamen-
tally derived from the Nernst Distribution Law but assume
constant behavior accross a wide variety of concentra-
tions. Moreover, some of the older techniques, such as the
McCabe-Thiele approach are appropriate for simple, single
or few element systems, but break down, either function-
ally or practically when expanded to the large, multi-ele-
ment problem associated with REE separations. Others
assume constant distribution ratio or constant separation
factor across a range of input conditions, and thus become
error prone as the solution concentrations and other factors
change in as solvent extraction battery.
A more recent modeling technique, proposed by
Turgeon, describes a fundamental approach based on
chemical equilibrium (Turgeon et al., 2023). Since the
model is based on fundamental reaction chemistry, it is
both robust and potentially scalable for large-scale REE
separation problems. In prior work, Turgeon has shown
both the utility of the method for circuit and plant-scale
simulation as well as methods for producing the requisite
data from laboratory analyses (Turgeon et al., 2016).
Building from the Turgeon model, we have integrated
the fundamental modeling approach into a linear circuit
analysis framework to provide a robust framework for com-
plex circuit design and optimization. As a solution meth-
odology, linear circuit analysis provides a consistent and
scaleable approach to define the superstructure of the pro-
cess circuitry in a manner that allows arbitrary complexity.
The combined Turgeon-LCA approach thus rapidly pro-
duces circuit solutions and provides a rigorous basis for cir-
cuit improvement and optimization. The objective of this
paper is, thus to describe the combined methodology and
discuss its application.
METHODOLOGY
Equilibrium Calculations
Details on the Turgeon modeling approach can be found in
Turgeon et al. (2016), and Turgeon et al. (2023) however,
salient points are summarized here for the benefit of the
reader.
In this approach, the solvent extraction system can be
described as a series of reversible chemical reactions involv-
ing the exchange of cations between phases of varying mis-
cibility. Generally:
Ln RH LnR 3 3H 3+
3 E +++
where Ln refers to any general REE in an aqueous phase
and RH refers to an extractant in an organic phase. The
aqueous is mixed with the organic where the REE will