1314 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
and that it should not be interpreted as representing a con-
tinuity in the dataset.
From the literature review, L3 determined that digly-
colamides should be further investigated. L3 initially
selected and evaluated dimethyl dioctyl diglycolamide
(DMDODGA) based on an article by Stamberga et al.,
2020. During the bench scale process development effort, it
became apparent that the DMDODGA very strong extrac-
tion potential coupled with its tendency for third phase
formation would create operational issues. The L3 team in
discussion with the Oak Ridge National Laboratory scien-
tists, pivoted to investigating dimethyloctyl dihexyl diglyco-
lamide (DMODHDGA, DG6) in place of DMDODGA.
The behavior of DG6 in the HCl-PLS circuit was
observed in a series of bench scale experiments. In one
such experiment, a series of three successive extractions
using the same organic phase with fresh HCl-PLS was per-
formed. The organic phase is composed of equal parts (25
v% each) of DG6 and tridecyl-alcohol, diluted in kerosene
(Exxsol D80). The contact time was set to 30 minutes to
ensure equilibrium and the experiments were performed
at ambient temperature. Extents of extraction, shown as
Figure 5, reveals a significant selectivity for REEs over iron,
even when the iron concentration in the HCl-PLS is two
to three orders of magnitude higher. This high selectivity
can also be observed as the relative reduction of extraction
extent is more significant for iron than for the REE.
Furthermore, the DG6 circuit can be selectively
scrubbed of its iron and a portion of its light REEs using
strip solutions that control acidity and activity. The effect
of acidity and activity on the stripping behavior of the
loaded DG6 from the experiments of Figure 5 is presented
as Figure 6.
While technically a DG6 circuit has proven to be suf-
ficient as a standalone selective REE extraction circuit,
bench scale evaluations, piloting and economic model-
ing of the resulting commercial process led the L3 team
to add an iron removal circuit upstream of the DG6 cir-
cuit. Reducing the amount of iron reporting to the DG6
circuit had two significant benefits: it reduced the overall
capital and operating cost while allowing for a more effi-
cient hydrochloric acid recovery circuit. TBP was selected
because is commonly used for similar applications and its
extraction behavior is well documented for Fe (Haggag et
al., 1976) and Sc/Th (Peppard, Mason and Maier, 1956).
In the initial design built on bench scale test work, the
TBP circuit was composed of an extraction battery where
all iron, scandium and some thorium where extracted, a
scrubbing battery where the scandium, terbium and tho-
rium were scrubbed off and a stripping battery to recover
the iron. The distribution of relevant elements is presented
in Figure 7. In the presented data, the scrubbing solution
was a 2.25 M hydrochloric acid solution while the stripping
solution was a 0.3M hydrochloric acid solution.
Extended operation of the circuit revealed that while
it operated as expected with regards to iron, the rare earth
elements and thorium, the scandium balance did not close
as it appeared to be accumulating in the system. A scan-
dium rich phosphate precipitate was observed during rou-
tine maintenance of the circuit and was recovered. Scoping
Figure 5. Sequential extraction extent of selected element by DG6
and that it should not be interpreted as representing a con-
tinuity in the dataset.
From the literature review, L3 determined that digly-
colamides should be further investigated. L3 initially
selected and evaluated dimethyl dioctyl diglycolamide
(DMDODGA) based on an article by Stamberga et al.,
2020. During the bench scale process development effort, it
became apparent that the DMDODGA very strong extrac-
tion potential coupled with its tendency for third phase
formation would create operational issues. The L3 team in
discussion with the Oak Ridge National Laboratory scien-
tists, pivoted to investigating dimethyloctyl dihexyl diglyco-
lamide (DMODHDGA, DG6) in place of DMDODGA.
The behavior of DG6 in the HCl-PLS circuit was
observed in a series of bench scale experiments. In one
such experiment, a series of three successive extractions
using the same organic phase with fresh HCl-PLS was per-
formed. The organic phase is composed of equal parts (25
v% each) of DG6 and tridecyl-alcohol, diluted in kerosene
(Exxsol D80). The contact time was set to 30 minutes to
ensure equilibrium and the experiments were performed
at ambient temperature. Extents of extraction, shown as
Figure 5, reveals a significant selectivity for REEs over iron,
even when the iron concentration in the HCl-PLS is two
to three orders of magnitude higher. This high selectivity
can also be observed as the relative reduction of extraction
extent is more significant for iron than for the REE.
Furthermore, the DG6 circuit can be selectively
scrubbed of its iron and a portion of its light REEs using
strip solutions that control acidity and activity. The effect
of acidity and activity on the stripping behavior of the
loaded DG6 from the experiments of Figure 5 is presented
as Figure 6.
While technically a DG6 circuit has proven to be suf-
ficient as a standalone selective REE extraction circuit,
bench scale evaluations, piloting and economic model-
ing of the resulting commercial process led the L3 team
to add an iron removal circuit upstream of the DG6 cir-
cuit. Reducing the amount of iron reporting to the DG6
circuit had two significant benefits: it reduced the overall
capital and operating cost while allowing for a more effi-
cient hydrochloric acid recovery circuit. TBP was selected
because is commonly used for similar applications and its
extraction behavior is well documented for Fe (Haggag et
al., 1976) and Sc/Th (Peppard, Mason and Maier, 1956).
In the initial design built on bench scale test work, the
TBP circuit was composed of an extraction battery where
all iron, scandium and some thorium where extracted, a
scrubbing battery where the scandium, terbium and tho-
rium were scrubbed off and a stripping battery to recover
the iron. The distribution of relevant elements is presented
in Figure 7. In the presented data, the scrubbing solution
was a 2.25 M hydrochloric acid solution while the stripping
solution was a 0.3M hydrochloric acid solution.
Extended operation of the circuit revealed that while
it operated as expected with regards to iron, the rare earth
elements and thorium, the scandium balance did not close
as it appeared to be accumulating in the system. A scan-
dium rich phosphate precipitate was observed during rou-
tine maintenance of the circuit and was recovered. Scoping
Figure 5. Sequential extraction extent of selected element by DG6