Kinross Gold’s Paracatu Mine in Minas Gerais, Brazil, had
budgeted for 80 per cent gold recovery in its second
concentration plant, but one year after commissioning
in 2008, the operation had topped out at 77 to 79 per
cent. In 2009, the mine called in Blue Coast Metallurgy to audit
the plant. “They approached us to go down there and really
help them understand why the gold recovery was lower than
the projected numbers,” says Chris Martin,
principal metallurgist at Blue Coast.
Over-grinding leads to losses
Originally, the plant’s flowsheet began
with a semi-autogenous grinding mill, which
then fed into two parallel ball mill circuits.
Before material entered each ball mill, it
passed through a set of cyclone density separators
that diverted the finer particles to be
recovered through conventional flotation
and cyanide leaching. The remaining coarser
material was ground in the ball mill, and
sent to a jig for recovery. Anything not recovered
by the jig was returned to the cyclone
and ball mill circuit.
To make their assessment, Blue Coast
personnel and Paracatu’s metallurgical team
collected samples from three different days
and 24 different points in the plant. “With
samples from all the streams, we can kind of
paint a picture of how the plant is working at a given moment
in time,” explains Martin. The Blue Coast team assayed the
material by mass and found high gold grades at midsize fractions
of +38 and +75 microns. However, these composed less
than 10 per cent of the concentrate produced by the jig. Of the
weight in the jig concentrate, 70 per cent was in the +300
micron fraction, containing only five to 10 per cent of the gold.
While building the plant originally, Kinross
had assumed that the gold would be
associated with coarse arsenopyrite and had
chosen radial Pan-American jigs designed to
recover coarser particles. But the majority of
gold, as it turned out, liberated at finer fractions.
“We found that there’s very little supercoarse
gold,” explains Martin. “It really only
starts to liberate at 150 microns.”
The jigs had been dismissing these
smaller particles back to the plant’s milling
circuit again and again, where repeated
grinding damaged their recoverability. “It
was getting very over-ground,” says Martin,
“and there’s quite a lot of evidence to suggest
that the gold surfaces were being degraded
in some way that rendered them less susceptible
to the downstream flotation process.”
The samples taken from the flotation circuit
showed that flotation had been working,
as long as liberated gold was coarse enough.
All of the gold coarser than 12 microns had made it into the concentrate. But only 15 per cent of finer gold
floated, suggesting that over-grinding had serious consequences
As a temporary fix to over-grinding, the Paracatu team
decided to increase the cutoff size for material sorted into the ball
mill by the cyclones from about 90 microns to about 110. Coarsening
the grind led to increased throughputs without reducing
gold recovery. Recoveries actually increased slightly, presumably
because fewer gold particles were being over-ground.
Flash flotation prevents over-grinding
Blue Coast’s plant survey showed that the main gravity
recovery circuit was failing to catch gold that was liberated at
medium particle sizes. The company suggested that if coarser
particles could be exposed to flotation much earlier, then recovery
rates would improve.
Flash flotation cells, used for base and precious metals since
the 1980s, are designed to do this. So named because they
work more rapidly than conventional flotation, flash methods
recover liberated particles that respond easily to flotation but
are too coarse to enter mainstream flotation tanks. In contrast
to typical flotation cells, flash cells sit within the grinding circuit.
By using flotation early in the process, gold that is liberated
at that stage can be recovered to avoid over-grinding, while
coarser particles return to the mill.