XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 2345
and high carbon samples (Figure 4). In the low carbon
samples, the spectra were deconvoluted into the D band
(1,348 cm–1), G band (1,577 cm–1), and S band (2,702
cm–1). The averaged low carbon samples are characterized
by a low intensity D band and a high intensity, narrow G
band (Figure 4). The spectra also display a broad S band.
These characteristics suggest that the low carbon samples
contain ordered, crystalline CM. In the high carbon sam-
ples, the spectra were deconvoluted into the D band (1,330
cm–1) and the G band (1,606 cm–1). The averaged high
carbon samples are characterized by a broad, relatively-high
intensity D band and a high intensity, relatively narrow G
band (Figure 4). The spectra do not indicate an S band.
These characteristics suggest that all the high carbon sam-
ples contain disordered, amorphous CM.
Samples with varying CM concentrations and gold-
robbing capacities were analyzed with the expectation that
a high gold-robbing capacity was due to the presence of
amorphous carbon. Results show that all four of the highly
gold-robbing samples analyzed contain purely amorphous
carbon, while the four non gold-robbing samples contain
crystalline carbon (Figure 4). This implies that carbon
structure can be an indicator for gold-robbing in carbon
containing polymetallic ores to a certain extent. It is impor-
tant to note that gold deportment may be complicating
the analysis further, as it is anticipated that there may be
a varying concentration of gold in solid solution in pyrite,
thereby influencing how much truly leachable gold is avail-
able for recovery.
These different “types” of carbon are also anticipated
to impact flotation response, but well controlled flotation
laboratory studies are hampered by high froth recoveries
and better control that can be applied to a laboratory test
versus the challenges in managing froth at a plant scale,
where froth recoveries are typically low, and the impact of
any instability on the froth phase can be very challenging to
manage. It is anticipated that the more disordered the car-
bon is, the more reagent will be consumed in flotation, and
the greater disruptions it will cause in the froth phase, lead-
ing to challenges controlling the flotation separation. By
the same token, more ordered carbon is anticipated to be
more easily controllable in flotation, creating less upheaval
in subsequent flotation separations. An additional compli-
cation in operation is that even measuring organic carbon
in practice is a challenge and layering on the need to under-
stand the “type” of carbon at scale, is a solution that does
not yet exist. It is possible, however, that a proxy indica-
tor, whether through geology or geochemistry, automated
measurements of reagent “demand” (consumptions), could
potentially be leveraged, but it will take significant effort to
discern a reliable proxy within the bounds of what is typi-
cal “noisy” plant performance information. Using a gold
robbing index could be one such proxy indicator. Further
benchmarking work, leveraging the recently developed
Figure 4. Averaged Raman spectra for CM in the low and high carbon samples
and high carbon samples (Figure 4). In the low carbon
samples, the spectra were deconvoluted into the D band
(1,348 cm–1), G band (1,577 cm–1), and S band (2,702
cm–1). The averaged low carbon samples are characterized
by a low intensity D band and a high intensity, narrow G
band (Figure 4). The spectra also display a broad S band.
These characteristics suggest that the low carbon samples
contain ordered, crystalline CM. In the high carbon sam-
ples, the spectra were deconvoluted into the D band (1,330
cm–1) and the G band (1,606 cm–1). The averaged high
carbon samples are characterized by a broad, relatively-high
intensity D band and a high intensity, relatively narrow G
band (Figure 4). The spectra do not indicate an S band.
These characteristics suggest that all the high carbon sam-
ples contain disordered, amorphous CM.
Samples with varying CM concentrations and gold-
robbing capacities were analyzed with the expectation that
a high gold-robbing capacity was due to the presence of
amorphous carbon. Results show that all four of the highly
gold-robbing samples analyzed contain purely amorphous
carbon, while the four non gold-robbing samples contain
crystalline carbon (Figure 4). This implies that carbon
structure can be an indicator for gold-robbing in carbon
containing polymetallic ores to a certain extent. It is impor-
tant to note that gold deportment may be complicating
the analysis further, as it is anticipated that there may be
a varying concentration of gold in solid solution in pyrite,
thereby influencing how much truly leachable gold is avail-
able for recovery.
These different “types” of carbon are also anticipated
to impact flotation response, but well controlled flotation
laboratory studies are hampered by high froth recoveries
and better control that can be applied to a laboratory test
versus the challenges in managing froth at a plant scale,
where froth recoveries are typically low, and the impact of
any instability on the froth phase can be very challenging to
manage. It is anticipated that the more disordered the car-
bon is, the more reagent will be consumed in flotation, and
the greater disruptions it will cause in the froth phase, lead-
ing to challenges controlling the flotation separation. By
the same token, more ordered carbon is anticipated to be
more easily controllable in flotation, creating less upheaval
in subsequent flotation separations. An additional compli-
cation in operation is that even measuring organic carbon
in practice is a challenge and layering on the need to under-
stand the “type” of carbon at scale, is a solution that does
not yet exist. It is possible, however, that a proxy indica-
tor, whether through geology or geochemistry, automated
measurements of reagent “demand” (consumptions), could
potentially be leveraged, but it will take significant effort to
discern a reliable proxy within the bounds of what is typi-
cal “noisy” plant performance information. Using a gold
robbing index could be one such proxy indicator. Further
benchmarking work, leveraging the recently developed
Figure 4. Averaged Raman spectra for CM in the low and high carbon samples