Some Factors Influencing Gold Recovery by Gravity Concentration

Some factors influencing gold recovery by gravity
concentration
by M. D. BATH,. M.A.Se. (Met.) British Columbia (Visitor)
A. J. DUNCAN,. B.Sc. (Hons.) Rhodes, M.S.A.C.I. (Visitor)
E. R. RUDOLPH,* M.Se. (Eng.) Rand (Member)
SYNOPSIS
Recent studies have shown that the overall extraction of gold from Witwatersrand ores is increased as the pro-
portion of gold reporting to the amalgamation circuit is increased. Testwork has therefore been carried out with a
view to improving recovery by gravity concentration.
In this paper, the history of gold recovery by gravity concentration is reviewed, and methods and machines at
present in use are briefly described. The results of plant and laboratory tests are presented, and the limitation of
existing concentrators is discussed.
A pilot plant is being built for the testing of improvements to the existing concentrators and of several new
devices, particularly the Reichert-cone concentrator.
SINOPSIS
Onlangse studies het getoon dOltdie totale ekstraksie van goud uit die Witwatersrandse ertse toeneem namate die
hoeveelheid goud in die amalgamasiekring toeneem. Daar is dus toetse uitgevoer met die oog op die verbetering
van die herwinning deur swaartekragkonsentrasie.
Hierdie verhandeling gee 'n oorsig oor die geskiedenis van goudherwinning deur swaartekragkonsentrasie en
beskryf kortliks die metodes en masjiene wat tans in gebruik is. Die resultate van aanleg- en laboratoriumtoetse
word uiteengesit en die beperkings van die bestaande konsentreerders bespreek.
Daar is 'n proefaanleg in aanbou om die verbeterings aan die bestaande konsentreerders en verskeie nuwe toe-
stelle, veral die Reichert-keelkonsentreerder, te toets.
HISTORICAL concentration on the Rand mines (2) Johnson concentrators,

Gravity concentration of gold has goes back to the corduroy blankets, (3) plane tables,
been used since antiquityl but its which were highly labour intensive. (4) endless-belt concentrators, and
importance has decreased since the As a result, the endless-belt con- (5) jigs.
turn of the century, when the cy- centrator, the Johnson drum con-
Corduroy-blanket strakes
anide process was developed. In the centrator4, the plane table, and
These, or rubber-blanket strakes,
early days of the Witwatersrand2, jigs came into use. Jigs have not
were apparently still in use on 24
before the introduction of cyanid- gained wide acceptance, but the
plants in 1961. They were placed at
ation, the ores were stamp-milled two concentrators are still in current
the tube-mill outlets, and were
and the gold recovered by amal- use on Anglo American Corporation
manually washed at intervals to
gamation. This practice became ob- mines. An additional bonus in the
recover the concentrates. The
solete as the oxidized ores became form of osmiridium is realized on
method was highly labour-intensive
exhausted. The introduction of tube plants employing gravity-concen-
and constituted a security risk and,
mills with corduroy-blanket strakes tration circuits, and this often out-
as far as is known, has now been
on the discharge to collect the coarse weighs the cost of the treatment.
entirely replaced by mechanical con-
gold, which was recovered by amal- Douglas and Moir5 reviewed South
centrators. This change has been
gamation later, gave rise to the African gold-recovery practice in
accompanied by circuit modifications
cyanide plant that we know today. 1961, pointing out that two-thirds
to provide selective grinding and by
The vastly more efficient cyanide of the major plants incorporated
the installation of as few machines
process has caused many plants to gravity concentration followed by
as possible.
dispense with gravity concentration, cyanidation of the bulk ore, up to
although they tacitly use longer 73 per cent of the gold being re- J ohnson concentrator
residence times in the cyanide solu- covered by gravity concentration This mechanical concentrator6 was
tion to compensate for the lack of when the gold was relatively coarse. developed in 1925 and has since
concentration. Doff and Bosqui3 In this paper we shall attempt to gained wide acceptance. It consists
emphasize the importance of the put the argument for gravity con- of a continuously rotating cylindrical
early removal of gold from the centration in a more balanced light, drum, having its axis inclined at
milling circuit and advocate gravity especially since computer correlation between 2,5 and 5 degrees to the
concentration, especially for those studies have shown it to have a horizontal, and lined internally with
ores in which a significant propor- significant effect on the overall riffled rubber. Pulp flows down the
tion of the gold is associated with plant recovery. length of the drum, and the heavy
base metal sulphides. minerals settle in the riffles and are
The history of modern gravity GRAVITY CONCENTRATORS carried out of the pulp stream to the
In 1961, the following methods of top, where they are washed into a
...Anglo American Research Laboratories,
Anglo American Corporation of South gravity concentration were in use: launder by sprays. In 1961, mines
Africa Limited (1) corduroy-blanket strakes, using these machines in conjunction
JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY JUNE 1873 383
with endless. belt concentrators for has saw-tooth riffles, 5 mm deep with practise gravity concentration. Gold
upgrading the drum concentrate a pitch of lOmm, running across the recoveries by amalgamation range
reported an average gold recovery belt. The belt moves continuously from 21 to 52 per cent.
of 38 per cent by concentration and against the pulp stream at a speed of Only two types of concentrators
amalgamation, the range being from 30 cm per minute. A spray water- are in use, in all cases the Johnson
30 to 52 per cent. pipe washes concentrate from the drums acting as roughers and the
Plane table riffles after the belt has passed over endless belts as cleaners. Flowsheets
This device, developed at Rand the head pulley. The unit has been are similar in all plants, differences
Leases, consists of riffled rubber used as a primary concentrator and, being in the number of concentra.
covering a series of inclined smooth more usually, as a cleaner concen. tors, pulp feed rates, and the ratio
surfaces, with steps between them. trator for the retreatment of the of Johnson drums to endless belts
The pulp flows down the length of product from J ohnson drums. (Figs. la and 1b). Table I gives
the riffles, and the riffles collect the Jigs mechanical details and operational
gold concentrate, which is drawn off These machines have never gained data for the concentrators in the
through slots at the steps. The faster. wide acceptance in the South African various plants.
flowing pulp is carried across the gold industry, probably owing to Typically, secondary tube.mill
slots and is collected at the end of the large bulk of concentrate. At product is pumped to tertiary cy.
the table. At Rand Leases, the plane- Durban Roodepoort Deep and West. clones whose underflows are distri-
table concentrates were reground ern Areas, Yuba.Richards jigs are buted among several Johnson drum
and re concentrated on secondary used. At Western Areas, the jigs concentrators. The tailings are re-
plane tables, followed by a final treat secondary mill discharge, pro. turned to the mill circuit, while the
concentration stage using James ducing a concentrate that is de. concentrates are retreated on end.
tables. In this circuit, 51 per cent watered and reground before re. less- belt concentrators, there usually
of the gold was recovered by amal. concentration in secondary jigs. being one belt for every 4 to 8
gamation. Tailings from these jigs are returned Johnson drums. The belt concentrate
Endless-belt concentrator to the regrind-mill cyclones, and constitutes the feed to amalgamation
The belt concentrator was de- concentrates are cleaned on endless. while the belt tailings are returned
veloped at Brakpan in 1949 to re- belt concentrators and a James to the mill circuit.
place corduroy blankets. The unit table. The overall plant recovery is As part of a major project aimed
consists of a flat, endless rubber belt of the order of 30 per cent by gravity at increasing gold recovery in Anglo
150 cm wide, with 300 cm between concentration. American plants, improvements in
head and tail pulleys, and inclined gravity concentration are being in.
at a slope of 3 degrees to the hori- ANGLO AMERICAN PRACTICE vestigated both in the laboratory and
zontal. The upper surface of the belt Of the Group's 13. plants, 8 on the plants. The project is divided
,'Hue t from
primary lIIi!lint
c"nc to
cyoniclation belt
amalgamation
.1
oncentrator
tail to
high grode .econdary
.econdory
cyclone .teady head
mill.
pump
Fig. la-Concentration with secondary milling
364 JUNE 1973 JOURNAL. OF THE SOUTH AFRICAN INSTITUTE OF MINING AND MET AL.L.URGY
primary product
cyanidatian amalgamation
secondary tertiary
mills mills
Fig. Ib-Concentration with tertiary milling
into two phases: the first is short viously be no single liberation size- are more likely to be lost to the
term, with its object the optimi- the larger particles of gold are tailings. The fact that the same
zation and improvement of existing liberated at a much coarser grind effect was not observed with the
machines and circuits, and the than the finer particles, for example Vaal Reefs material is probably due
second is longer term-the develop- those enclosed within pyrite grains. to the much coarser gold in this ore-
ment of alternative machines and/or Tests were carried out to deter- even when flattened, the grains are
circuits. The motivation for this mine what percentage of the gold still effectively caught in the riffles.
investigation was the finding, during could be recovered at various degrees The results of these tests suggest
a computer correlation study of of grinding. Ore samples were milled that gravity concentration could be
plant operating data, that the overall for various periods and the products effectively incorporated much earlier
gold recovery is increased as the passed over a small vibrating strake in the circuit than it is at present,
proportion of gold recovered by covered with riffled rubber. Curves and that high recoveries can be
concentration and amalgamation is for recover'y versus grinding are obtained with simple equipment
increased. Without going into details shown in Fig. 2 for three different very similar in principle to that al-
this is obviously due to the fact that samples of ore. Recoveries in the ready in use. Early recovery of
recovery by. amalgamation is more range 75-80 per cent were achieved liberated gold has long been ad-
efficient than that by cyanidation. at relatively coarse grinds. The Vaal vocated but somehow seems to have
Reefs ore, in particular, yielded a been forgotten over the years. The
GRINDING AND STRAKING gold recovery of 75 per cent without deleterious effects of overgrinding
Grinding tests were carried out to any milling at all (crushed to pass -not only flattening but also com-
determine liberation grinds for ores 1,5 mm). As milling was continued, minution-of the gold, would thus
from several of the Group's mines. recovery increased slightly to 79 be avoided.
The term "liberation" as used in per cent at 93 per cent minus 74 /km.
this paper is best defined as that The curves for Western Deep Levels TESTWORK ON THE JOHNSON
particle size which maximizes profit high-grade and low-grade ores show DRUM CONCENTRATORS
for the subsequent process. Hence, increasing recovery up to about 75 A series of tests was carried out
for cyanidation or any leaching per cent minus 74 /km, after which on one of the low-grade Johnson
operation, the mineral grain need there is a marked decrease in re- drum concentrators at Western Deep
be only partly exposed, whereas, covery with finer grinding. This Levels Mine. This drum had been
for gravity concentration, it must be effect is thought to be due to a modified to facilitate sampling, and
almost completely separated from flattening of the liberated gold a variable-speed drive was installed.
the gangue. particles as the result of overgrind- Composite samples of the feed,
In an ore as unhomogeneous as the ing. The flattened particles settle far concentrate, and tailing were taken
gold-bearing reefs, there can ob- less easily than spherical grains, and under various conditions of solids
feedrate, feed pulp density, and
drum speed. The least varied was the 100
pulp density, which was adjusted
only to ensure that the pulp did not
sand out in the drum, the highest
possible pulp density and hence the
longest residence time possible being
thus achieved. The gold content of
the feed was an interfering variable, VRWICS
ranging from 91,5 to 197,6 g/t, but ...
Z
no correlation was found between ..,
V
this and the concentrate grade, Ill:
..,
which varied from 920 to 5468 g/t. A.
The mass of concentrate was 0,44
t/h at the lower figure and 0,07 t/h
at the higher. The solids feedrate
>-
Ill:
was varied from 5,1 to 26,4 t/h. ..,
The percentage recovery was found
to be inversely proportional to the
~
V
WDLIHG
..,
solids feedrate (Fig. 3), while the Ill:
:t
~.
100
,,
80 PERCENT MINUS 74 ~m
.~ Fig. 2-Strake tests-recovery versus grind

D
D
...
actual yield of gold (kg/h) was
... roughly constant (Fig. 4).
:0
III Fig. 3 is a least-squares fit of
. percentage gold versus feedrate
~
c.
~
60
and shows gold recovery to increase
hyperbolically with decreasing ton-
~
... nage. This is to be expected since
c:
D the residence time of the pulp in the
.! drum is increased when the tonnage
E is decreased. What is not so ex-
l pected is that the actual mass of
i
.
... ~
0 gold recovered per hour remains
] constant, as can be gauged from
~ 0 Fig. 4. Here the masses of gold
8 recovered per unit time are plotted
-. on normal probability paper in
....
Cb
order of increasing magnitude
~0
:: against the cumulative observed
20 0 00 frequency. The fact that a reasonably
'" 0
~« straight line is obtained indicates
0 that the differences observed are
due to experimental error and the
values are not distinguishable from
the mean. The explanation for the
constant mass recovery is to be
found in the fact that, although the
10 20 30
residence time is increased with de-
creased tonnage, the mass of heavy
Feedrate tonnes/hour minerals, including gold, is reduced
Fig. 3-Johnson drum concentrator-recovery versus feed rate proportionately.
366 JUNE 1973 JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY
latter for a small-scale belt con-
centrator specially made for the
purpose.
2 .
Plant machines
0
5
The two plant machines are each
fed by the combined concentrates
10
of three high-grade and three low-
grade Johnson drum concentrators.
20 The results represent day-shift
~ 30
samples taken over a period of six
months and are hence inherently
>- 40 more reliable than those obtained
~
e
~ SO in the laboratory. The average gold
1 60 recovery was 63 per cent, iron re-
e covery (including tramp iron) was
..; 70
0 29 per cent, and uranium recovery
1
u
60 was 9 per cent. The concentrate
mass was 3,8 per cent.
The uranium recovery is con-
sidered to be of interest since some
85
of this uranium comes from the
0
low -grade circuit via the J ohnson
drum concentrators, and an im-
118
provement in this could be of con-
88
400 600 600 1000 siderable economic benefit. Doubling
Au Reco.ery gromsthour the speed of one of the belts resulted
in an overall increase in total gold
Fig. 4-Johnson drum concentrators-Gold recovery versus cumulative frequency, recovery by amalgamation of 2,25
showing the recovery to be normally distributed per cent, but at the expense of in-
The constant mass recovery of
gold per unit time is also a feature
of the high-grade drums, where
about 1 kg of gold was recovered per
hour as compared with the 0,55 kg
of gold per hour on the low-grade
drums. This is attributed to the
much higher heavy-mineral content
of the high-grade circuit, the mass of 1
gold being assumed negligible in .!
parts per million.
The Johnson drum concentrator is
a cheap, easily run machine. Al-
.cc
'0
!
4 x 1.1t no.1
0
I
1.lt no.2
though it is not very efficient as a
unit, it nonetheless, because of a
large re circulating load, recovers up
to 70 per cent of the gold entering ..
the plant. Screen analyses showed
the recovery of gold to be poorest
.
.J:I
in the minus 43 /Lm fraction and
best for the coarse sizes. This is in "'V
common with all gravity-concen-
tration machines and is due to the
.0:.
!
lower settling rates of the fine ! 1

particles. ::t
oC
TESTWORK ON THE ENDLESS-

BELT CONCENTRATORS
Both plant7 and laboratory results
have been obtained for these con- 2 3 4 5 6
centrators, the former for two ma- F..d to b.1t Au kglt
chines in routine service and the Fig. 5-Endless-belt concentrators-recovery as a function of feed grade
creasing the concentrate mass by If:>
l-~~";< ~0 ..... 0 ..;<
nearly 50 per cent.
E-<
Z A
<N
L-
'-'
~...:I C!h-
...:1....;' ..;<
'"
<t5..;:re:...::
~~<N
... <t5
lQ <t5 0..;< re:
..;<
lQ
,=:2;
~-+O
The effect of doubling the speed <0
~~S 0x
<N '"
lQ ~lQlQ '0-+0
000
of both belts would be to increase <0 ~g
~<:::
<0 0~ "1 ..;:
Po<
~;:::L- <:<I<t5 re: ..... <t5 00 <:<I ~8
gamation to 41,5 per cent (an in- ... 0 ~"" <N """ 00 L- ..;< L-
.....
"" "" <0
crease of 4,55 per cent), but such a lQ
"'" ""
move would double the concentrate
mass, which would overload the 0
..... ,=:2;
0'1
existing amalgamation facilities. E-< A <0 Q)-+O
~oj
-< 0"
When the feed grade to the belts ~p
>'1
x <N lQ
<N.....
....;'L- lQ ""<O<Noo 0 0 ~0 0 ~.t:
A ..... L- 00...";:'; <:::
~C "" '"
<0 "" <N ~8
recovered per tonne of feed, a most ... <N
"'
interesting result is obtained. The
recovery Increases in direct pro- L-
L- ..;<
L- lQ
portion to the feed grade (Fig. 5), and <N 0 <NOO<O";< 00 ~<N
<0 ...:I 0~ ..;: 00....;'.;<t5 re: ... <:::
>'1A ""'"
L-"" <N <0 ~1$
the situation on the Johnson drum, 0'1 >'1
~0
00
lQ 0
L-
<0 0<0<000 lQ
L-
'" ~00 .~ >::
where a high feed grade automati- c 00" <t50;<t5..: 0; .Q' ... 0;
lQ ~~'"<t5
lQ
<N";<"'" <N 00
""
<N
~8
cally follows from a decreased ton- I~ .....
""
""
nage (but a constant mass of gold), lQ
and the residence time on the belt en <N 00""
, -+000 ,=:2;
increases. Z ~~<0
L- lQ oj<N ~-+O
00 '"
~c 0 ~oj
~>'1 x ~~<N <N" ";< <O";<""L- <0 lQ
~.t: Q)
Laboratory small-scale belt A ..... 0 0; ~'"
<0 L-"00... 0 lQ
", ..... re:
"'" ~g
<"' ~~0'1
>'1 00 -+0 ..;< "" 0 <0 ..... .....
~0'1 '"
<N"""" <N ..... 00 .~ ~'" '003
This machine has a riffled-rubber ~~lQ lQ >:: ...
~<N .Q'"" ~8
Z~ ... -+0";<
03";< 0
belt 12,5 cm in width, with 60 cm ~p
00
~of exposed surface, and was driven UO'1
Zz -...:I
0
up a plane inclined at 15 degrees to oS z I I

>-<u~ :>< 0
the horizontal.
sented
The feed was pre-
about 10 cm down the
~....<
E-< I~. 0 L-
,=:2;
Q)-+O
~oj
...:I
~~~E-<
'"
z x
L-oo
..;: 0lQ 0;.; .;
"'"<NO"" 0;
0 <0
..;: <t5 <:::
~~~E-<0'1
Q 00 "-' .....""""..... """ 00 ~~'0
flowed onto the belt in a sheet Z~ Uj
lQ "" '" "" "" .~ >:: 03
...
~~8 0
immediately behind the feed point >-<E-<
....<.. 0'1 ...
~..c:
to wash gangue out of the concen- 1>-;0'1 Po< .~
u~
~trate. ZP
o£ Po< 0 <O"'<N""
I
Tests ~~were run under varIOUS <N
~...:I 0~ "" .Q'o;o";: <t5 ..... 0
enQ 0'1
L- ...:10; <N ..... ,=:2;
~Q)-+O
Q
0'1
..... "" .....""""..... .....
~conditions of wash-water flowrate ZO'1 0 .....
Z '" x
~~~0'1 lQ
~""'"
""
'"
l-
0 '0-+0
and feed pulp density. A constant 0"' ~;-
<:::
~0'10'1 00 00 ~1$
volume of pulp was delivered so that E-< >'1 lQ 0 00:."" """"oolQ 1:-1 "" "" .~ >::
~~.Q'<:<I.Q'oO ... re:
"" <N ~8
the solids feedrate was a function '"
0'1 ... ~~'" .....
L-
<N""<N ~<N .....
~""
of the pulp density. The concentrate
is expressed as a concentration ratio
.....0.....00 lQ
to suppress the effect of feed grade. <N 0 00" 00.
~....< ...:la;;
.Q' .Q' ",,".Q' 0
<t5 "'" "" ,=:2;
Fig. 6 shows logarithmic plots of '"
"'
0'1
L-
0
'" ....."""".....
"'" .....
L-
~0 ~-+O
~~0'1 ~0 l- "" ""
L- 00 '0-+0
concentration ratio against the per- X "" ~~<N 00 ~>::
centage and >'1 00 0 ""
..... ~1$
mass of concentrate -< lQ 0 "" lQ 0 L- .~ >::
;-
-< 0 lQ"
.Q"';:...oO re:
""L-"""" <t5 ... ~8
percentage recovery, the latter pass- ... ~~~~'"
""'<N""<N 0
lQ
"" <N '"
ing through the Western Deep Levels ""
figure of 63 per cent recovery at a

concentration ratio of 17. A com-
puter analysis of the results (Table
11) showed the concentration ratio S
to be a function of the feedrate and r-:'
Q)
-+0
'" -+0
the wash-water flowrate. 00
~OlQ..;<""
"".....L-L- -+0
~---
b()
:>,
z:2. -+0 bv 0>
Conclusion
~>.15 -+0
~+++1 b()
--- ~§
Q) .; '0" ~ca
x
From the combined results of the 00 88
--- -+003
~--- b()
~00 ~0>
~0>
> .~
~plant and the laboratory, it seems ..c: .8 8 ... >:: 0> ca -+0 ~...
-+0 .15 .8~Q) ~~> 1e -+0
b()
>:: ... 88 '0
03 ca 6 03
... g
that, given a lower feedrate at a b() .t:g
~00 -+0
~..r;; 00 -+0
~~~5b >
;§
~g >::
'"
higher grade, it would be possible, 'S ~~~00'0 '0~03 0 0
~0 ~.0 ~Q) .01 §'" >::
'" g
'" ~"O ~0
by adding wash-water to the belt '"
00 Z 113 en en ;S ... ~~U '" ~0
~368
.JUNE 1973 .JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY
...
... behind the feed point, to Increase
0 ""100""
10
'<I' 0...... '<I' 0> the concentration ratio without
E< 10,
0
......
'<I'
10 00 ~..;~..; ..: ""
0 0'" M
H 0>
""" 0> 00
Z A ""
...... ""
<:'1'<1'<:'1 10 <:'1
'" losing on recovery.
01 ""
Z ...... <:'1
S
rJJ <
X '" <:£
... <:'1 THE PILOT PLANT
~""
~~10 0 00010'" 10 0>, 10
0...... 10 10 10 ...... 0
I'<
M p:;
""
°<:'1 "" 0>
~~";<:'1' 10 10 ~"" The varied operation of gravity-
<:'1 """'<1""" ""
...... '"
0>
" <:'1 concentration plants and their ab-
10 <:'1 <:'1
M
sence on some of the Group's mines
01 '<I'
<:'1 prompted
E< A us to design a portable
~~>'I
~rJJP """
0
<:'1""oco> 0 0 10 pilot plant. This plant consists of
~01A
X 0
...... '"
10
0"" ..:<:£~..: ~<:£ M
"" 0 ......
01 00 "" '"
0> """'<1'<:'1 0>
......
0 ...... 0> 0 10 four modules: the first contains
01 "" 0 00
~" '<I'
0 <:'1 devices for measuring the incoming
~M <:'1 ""
'" feed and presenting it to the other
units; the second holds a sieve bend
......
<:'1 10 00 and facilities for various flow schemes
10 '<I' "'<:'10>""
S...... 10 00 0
'" '"
~M
01 <:'1
""
10, ..:
";
0> """:M~ 00 ... to include or bypass the cyclones on
::;:Z <:'1 """<:'1'<1""'" ... ""
...... <:'1
"""
the third module, which also holds
0""~::;: 0>
x ...... 0 a Johnson drum concentrator; and
>'I
01 A 10
'"
>'I 00 <:'1 <:'1 ... 10
,,
~0 '<I'
0 O...... 0 0
0
'"
"'00<:'1'<1'
<:£
"" <:£
<:'1
""
"" the drum concentrate is fed to a
M p:; 10 ";
0> ..:""..:"" 0 ... ~~""
'<I'
"" <:'1'<1'<:'1 <:'1 """ half-width endless-belt concentrator
"" ""
rJJ
on the fourth module. This belt is of
~0 novel design in as much as it can be
E< '<I' vibrated at 30 Hz in two directions.
~<:'1
Z rJJ ~A flowsheet
E< ~" ...... <:'1 of the plant is shown in
Z <"1'"
~~100 '<1'0>000> ...
0001 rJJ A
X 0
...... '"
"""'<1' 0 '<I' <:'1'0 00...' 10
...... <:'1 00 '"
... ~0
... '"
M Fig. 7. All streams are capable of
"" 0>
~~f'iI>'I ~~00 """"""'<1'<:'1 00
00 '<I' 10
'" "" ~being sampled automatically, and
0 ""
......
E-<° M 2~ '" the major flows can be diverted into
<z
~£ a tonnage box on the second module.
E-<Z
z=~::; The cyclones on the third module
I-IQ::;: E< '<I'
<:'1 will also act as a test-bed for cyclones
~Zo Z ~'<I' of novel design (to be described
HO~ Z 00<:'1 00
00
00
10
.... f'iI 0
~~0 ...'"
1010
0 010010 0> "" '<I'
P=lQ~
rJJ E<
X <:'1 ...... "" ..: M...' M ",' 00 M
"" 0... 0 <:'1' later). This plant is due to be com-
<E-<rJJ ~rJJ 00 "" '<I' '<1"<1' ""
'<I'
00 '<I' 0>
10 '<I'
E-<H~ <:'1 missioned at Vaal Reefs North In
~"'1 I'< 0'<I' 0>
M 00 February 1973.
P=I~ 2~
A~ While the design work was In
~o progress on the first pilot plant, we
HZ
~8 I'< '<I' carried out laboratory gravity-con-
~.. f'iI
f'iI <:'1
1-101 A rJJ
~0> ... centration testwork on a Wilfley
~o ~00 ... 0>
~Z~01 X
0
0
......
0> "" '<1"""""""<1'

..:..;..:....;
<:'1
table. Table III shows the results
~P '"
I> <:'1 ...... ...... 10' 00 0> M M
0 f'iI """ 0> <:'1'<1'<:'1
rJJ 00
~~'"
00
"" 0'" '" "" of Reichert-cone simulations carried
A 0 <:'1
01 f'iI
~M out on the Wilfley table, as well as
01
the tabling of ~some belt-concen-
trator tailings. These results proved
rJJ '<I'
<:'1
~~00 10 so encouraging that we have ordered
f'iI ~...... '<I' 0>
0 <:'1
10
10 00 '<1""'<1'10 0 0> 00
'" ~... a Reichert-cone, which is to be in-
X <:'1 ...... ...... 0" '; ",'......'0""; 00 M ..; ...... 0
>'I 00 0> '<I' '<I'
00 <:'1 stalled in a second portable pilot
< """10"" 00
0~'<I' "" plant that will be commissioned at
M
Vaal Reefs South. Fig. 8 shows the
~flowsheet split into its modules. The
1'1'
0 cone concentrates will be fed to a
'~oj
S triple-deck shaking table, and the
00
oj ~middlings and, possibly, the lower-
a:>
bII
O! grade concentrate will be recycled.
~8
~oj
§ ... CYCLONES AS
t) a:>
.;::
.<::
00
010'<1"<1' "'"
bi5
oj CONCENTRATORS
~""............
~"C
"'" <2..
+++1 "'"
bi5
a:>'
., a:>
Reference to Figs. la and 1b shows
~O! t)I>
'~
a:> 1'1 "'"
 0
X oj $ "'"
bi5 ~a:> that the function of the tertiary
., 's bII
'" -- "'" '" ...
.<::
bII
,S 8 .t! ,S 1'1
;.a .,
~
~oj "C
"'"
cyclones in typical Anglo American
"'" 8 a:> oj
O! ... ...
1'1 ~a:> $ a:>
... a:> "' "0 circuits is that of concentrators
'"
..0 ..a 6D I> bII
'" g
""' I> "'"
1'1 bII
", ;§ a:>
a:> 8.,
a:>
p.. a:>
a:>
"
"'" "C "C
oj
t) 0t) t) O! rather than classifiers. This is par-
0 '0~'" '(;j
a:>
p.. a:> a:> 1'1 a:> 1'10
;n
'" "O ~~p:; E-< Q0 0
~Q "'" ticularly true of the Western Deep
Z U3 00 00 E-<
JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY JUNE 1973c 369
Levels circuit, where there is no
tertiary milling. In other circuits
(Fig. lb), the tertiary cyclones are
20 also concentrators for the feed to
the Johnson drums, but the Johnson
drum tailings are dewatered in
another cyclone before being milled
in the tertiary mills, thus receiving
10 some classification. It is not possible
.!! for a given cyclone to perform both
! functions simultaneously at maxi-
mum efficiency - an increase in
"
.!! classification efficiency automatically
i
ie 5 - results in a decrease in concentration
efficiency. This correlation between
i
v
concentration inefficiency and classi-
fication efficiency is understandable
when one considers that, for high
concentration efficiency, the cyclone
0 weight
is expected to make a separation
in which coarse, light particles
. recovery report to the overflow, and fine,
heavy particles (gold-pyrite) report
to the spigot. This is contrary to
the normal tendency of the cyclone
t. 1 to act basically as a classifier. This
point will be discussed in more
detail later.
perunt
If the above is accepted, a de-
cision must be made on which of the
Fig. 6-Endless-belt concentrator-recovery and mass fraction of concentrate as two functions-classification or con-
a function of concentration ratio centration-is the more important
for the tertiary cyclones. Since the
tertiary overflow is reclassified in the
TABLE II secondary cyclones, low classifying
RESULTS OF TESTWORK ON THE LABORATORY BELT CONCENTRATOR efficiency in the former is not
serious. Any coarse, light gangue,
Solids W as h - F eed- (possibly containing unliberated
Test feed water rate Belt Conc. Mass Rp- Calc. gold) that may overflow the ter-
no. % flowrate relative speed ratio % covery head
I/min units cm/min % Au g/t tiaries will be returned to the
--- --- ~- secondary mills via the secondary
I 23 1,8 4,31 40 8,2 9,1 74,4 365,0
2 23 1,0 4,31 40 2,5 37,7 92,5 513,7 cyclone underflow. It would seem
3 30 1,6 5,91 40 10,3 7,4 76,5 708,8 reasonable therefore, to aim at in-
4 29 1,6 5,67 40 7,9 10,5 82,8 858,9 creased concentration efficiency in
5 35 1,6 7,18 40 9,3 7,.~ 70,0 362,4
6 6 . 0 2,00 40 2,1 45,4 93,5 1311,7 the tertiary cyclones at the expense
7 35 1,1 7,18 40 6,6 1l,7 76,8 365,3 of classification efficiency. The aim
8 25 1,4 4,75 40 6,2 12,7 79,3 533,7
9 15 1,6 2,66 40 5,5 12,2 67,4 204,5 should be maximum recovery of
10 6 0 2,00 40 4,3 18,8 81,1 1149,2 gold in the cyclone underflow at
11 20 1,6 3,67 40 6,1 10,9 66,1 207,1 maximum grade. Obviously, once
12 30 1,4 5,91 80 5,2 14,9 77,9 247,4
13 25 1,8 4,75 80 5,7 14,4 81,8 282,5 this is accepted, it would theoreti-
cally be possible to replace the
cyclones with any other more effi-
concentrate grade cient concentrator. However, the
Concentration ratio =
feed grade cyclones offer so many advantages
from a maintenance and operating
Computer result of regression analysis: viewpoint that it seems logical to
Concentration ratio = 0,555 (feedrate) (wash-water flowrate) +2,578 retain them, and to attempt to
improve their concentrating effi-
Standard error of concentration ratio = 1,489
ciency. As far as is known, no
Standard error of variable = 0,123 serious attempt has been made to
improve the operation of the tertiary
t-test = 4,50 cyclones as concentrators, and the
present investigations were moti- during the initial phase of their vortex-finder before the centrifugal
vated by the considerable scope for acceleration in a radial direction. mass and the centripetal drag forces
improvement in this section of the The former hypothesis has led to the acting on these particles attain
plants. selection of wide-cone angles for equilibrium. Since it is the drag
Recent years have seen the de- concentrators, because particles force that increases the relative
velopment of various eyclones speci- stratify more reliably along the influence of particle size on separ-
fically for concentrating purposes. shallower wall of a wide cone than ation, this force is to be suppressed
Their physical design is such that would be possible along the steep in the operation of cyclones as
classification effects are suppressed wall of a slender cone. The latter concentrators.
and the influence of particle specific hypothesis has led to the preference
gravity is maximized. These are not for relatively short vortex-finder The Compound Water Cyclone
heavy-medium cyclones, which have clearances in concentrators, because The Compound Water Cyclone
been in use in the mineral-processing this design assists in separating (CWC) was invented by J. Visman
industry for some years, but true particles that swirl towards the and developed by him at the Mines
hydrocyclones. They are charact-
erized by blunt cones (i.e., large
included apex angles) in the range Splitt.,
Feed.. plant
60 to 180 degrees and long, large- box
diameter vortex finders. These de-
velopments are largely a result of
work in the coal industry8-12, where
cyclones operating with a water reie"
medium are now in wide use for the
instrument
upgrading of fine coal. Investigations
have also been carried out into the lank
use of cyclones for the beneficiation
of cassiterite13, 14and iron ores, and a
recent Russian paper15 describes the
concentration of gold from milled
conglomerate ores by use of a short- 'ydone
cone hydrocyclone.
As mentioned above, classification
cyclones are of slender design, with
a narrow cone angle-usually 15 to
30 degrees. In these cyclones, the
TO
particles must discharge eithm'
through the vortex finder or the
sp!>fJe, TONNAGE BOX
apex orifice, and the basic require-
ment for separation is the achieve- -ox OR
ment of balance between the centri-

fugal forces accelerating the particles REJECT
radially outward, and the centripetal
drag forces. For this purpose, rc-
1/3
latively large vortex-finder cleal- . ....
ances are common in classification
cyclones, because this assists in
providing the time for particles in
the central flow region to reach
their terminal free-settling velocity
relative to the entraining water.
In contrast,
wide cone angle-in

to 180 degrees-and
concentrating
clones are of squat design, with a
the range 60
are operated
cy-
0 pump
~
to suppress classification phenomena
in favour of gravity-concentration
effects. Briefly, they are based on
two hypotheseslO; (a) a particle bed
0 .a!"pl'~
stratifies according to density along

the conical wall of the lower cyclone RlJEC T
section, and (b) particles entering
the central spiral flow separate Fig. 7-Pilot plant no. I
- 25'-
FEED FROM PLANT
to lont
",een
U/S
0 lompl;ng
TO
MOOULE 2
cyclone feed
pump
10 plonl
FROM
MODULE
middl;ng relurn
from
Re;chorl cono
pulpSG TO
Ro;cherl
<one
& MODULE 3
moss flow mete
feed pump
FROM
TO
MODULE 2
MODULE 4
con 1
REICHERT con 2
It
m;ddling return to
CONE
lurgo tonk ofter
lompl ;ng
tailing
10 plont
I'O~:;' .. reject
TAIL
MID
CON
Fig. 8-Pilot plant no. 2.
TABLE HI charged through the vortex finder
REICHERT-CONE SIMULATION AND OTHER TESTWORK
by a process of elutriation. The
heavy particles, fine as well as
Ana]ysis Distribution coarse, are discharged through the
I
% apex. The separation thus takes place

Ore/Test Fraction Mass 1-- in three steps, as shown diagram-
% Au, g/t ' S, % Au S
- matically in Fig. 10.
Vaa] Reefs West-2° mill Conc. 10,2 467,6 37,4 88,4 92,6
discharge-high-grade Tai]ing 77,7 6,8 0,18 9,8 3,4 TESTWORK AND RESULTS
Reichert-cone simulation S]imes 12,1 8,25 1,36 1,8 4,0
Ca]c. head 100,0 54,0 4,]2 100,0 100,0 Tests were carried out using both
- the CWC 2M (50 mm) unit and the
Vaa] Reefs West-2° mill Conc. 10,0 442,0 39,1 88,7 89,4 CWC4 (100 mm) unit. The test
discharge-low -grade Tai]ing 68,4 5,9 0,25 8,1 3,9
Reichert-cone simulation Slimes 21,6 7,45 1,35 3,2 6,7 cyclone was fed from a centrifugal
Ca]c. head 100,0 49,8 4,37 100,0 100,0 pump, and the overflow and under-
- flow were returned direct to the
Western Deep Leve]s-2° mill Conc. 4,7 581,3 20,9 66,9 88,8
discharge-high-grade Tai]ing 87,2 14,35 0,1 30,6 7,9 pump feed tank. Sampling was
Reichert-cone simulation S]imes 8,1 12,65 0,45 2,5 3,3 carried out by cutting both product
Ca]c. head 100,0 41,00 1,10 100,0 100,0
streams simultaneously, generally
-
Western Deep Leve]s-2° mill Conc. 3,8 409,3 23,5 64,9 89,6 for a period of 5 seconds. The cy-
discharge-low -grade Tai]ing 89,4 9,0 0,09 33,4 8,1 clone was operated in a vertical
Reichert-cone simulation S]imes 6,8 7,1 0,34 2,0 2,3
Ca]c. head 100,0 24,1 1,00 100,0 100,0 position for all tests.
- All tests were performed on a
Western Deep Leve]s-be]t- Conc. 0,9 30198,6 66,2
Midd]ings 7,6 1419,25 26,3
slurry of Western Deep Levels
concentrator tailing-one
pass On table Tai]ing 91,5 33,8 7,5 secondary mill discharge, and the
Ca]c. head 100,0 410,6 100,0 operating conditions included vari-
Conc. & 8,5 4469,9 92,5
mid. ations in feed pulp density, feed
I pressure, cone type, and vortex
finder clearance.
Branch, Department of Mines and around the air core. The results are presented in Table
Technical Surveys, Canada1°, 16. It The remainder of the bed is forced IV, which shows the experimental
consists of a rather short cylinder, into the second conical section (II) conditions, gold assays of overflow
fitted with one of three alternative by new feed entering the cyclone, and underflow, and the distribution
compound cones (Fig. 9). Each substantially without losing its stra- of gold in these products. In addition,
compound cone has three sections, tified character. Here, the central the table includes calculated values
with included angles of 120, 75, and current is much stronger and erodes of the Gaudin "Selectivity Index"
20 degrees. In each case, the widest the top of the bed, where the (S.1.)17. This is defined as "the geo-
angle is nearest the cylinder and middlings are now exposed. The metric mean of the relative re-
the smallest at the spigot, but the light middlings are swept up and coveries and relative rejections of
three types differ in the relative discharged through the vortex finder. two minerals, metals, or groups of
areas of their inner cone surfaces. The heavy middlings that spiral minerals or metals"
The inlet is conventional, but the upward in the central current may RaJb
vortex finder is rather larger and bypass the orifice of the lower i.e., S.l. =
jRbJa
.
longer than that of an ordinary vortex finder owing to their higher
classifying cyclone. specific gravity. Consequently, the In the present case,
Visman16 explains the concentrat- coarse heavy -middlings fraction Ra = recovery of gold in underflow,
ing action taking place in his cyclone tends to recirculate to the stratified Rb=recovery of gangue in under-
as follows (see Fig. 10). bed and finally enters the third flow,
Particles of different sizes and conical section. In this last section J a=rejection of gangue to over-
specific gravity form a hindered (Ill), the bed is finally destroyed as flow,
settling bed in Section I of the coarse particles fan out along the J b=rejection of gold to overflow,
compound cone. Light, coarse par- cyclone wall in a single layer, ex- all values being percentages.
ticles are prevented from penetrating posing the small particles that so The index is used here as a con-
the lower strata of this bed by the far have been protected from being venient way of comparing the results
coarse, heavy fractions and by the washed out. The central current in of tests under various conditions.
fine particles filling the interstices of Section III is relatively weak, as it In Fig. 11, the value of S.L is
the bed. Consequently, the water has nearly spent itself in the pre- plotted against inlet pressure for the
passing from the periphery of the vious sections. The upward current three different cone types CWC 2
cyclone chamber towards its main that remains separates the small unit. The curves all pass through a
outlet (the vortex finder) erodes the particles from the remainder of the maximum at about 0,6 kgfcm2, and
top of the stratified bed and sub- material, with preference for those the value of the S.L increases with
stantially removes the light, coarse of low specific gravity. Thus the cone type in the order M <L<8.
particles via the "central current" fine, light particles are finally dis- The value 4,0-the highest observed
JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY JUNE1973 373
in any of the tests-is mainly due pressures employed were not high inlet pressure. An increase in vortex-
to the very low underflow mass of enough to show the maximum ob- finder clearance also increases the
3, () per cent. served in tests on the ewe 2M value of S.l.
In Fig. 12, similar data have been cyclone, and which presumably Fig. 13 shows that, in the ewe
plotted for the ewe 4 (100 mm) exists for the larger unit. However, 2M (50 mm) cyclone, the value of
cyclone, with the extra variable the L-type cone again gives higher S.l. decreases as the vortex-finder
"vortex finder clearance" included values of the S.l. than does the M clearance is increased, at pressures
(see Fig. 10). It appears that the type. This is true regardless of the above 0,5 kgfcm2. Below this pres-
sure, the value of S.l. at first in-
"I creases and then decreases as the
vortex-finder clearance is decreased.
Screen analyses and gold assays
of the fractions were carried out on
the products from selected tests.
The results are shown in Table V,
together with the gold recovery in
each size range. It is apparent that
recovery is lowest in the finest
(minus 44 /Lm) fraction, and, since
the major proportion of the gold
exists in this size range, it is this
recovery that determines the overall
gold recovery to the spigot product.
TYPr: s Discussion
Although the Selectivity Index
is a useful criterion for comparison
of the concentrating efficiencies of
various machines and experimental
procedures, it is necessary to look
at the results of the present tests
rather more critically, bearing in
mind the potential applications to
existing milling and concentrating
circuits. A high S.I. value can result
from a low mass of underflow at
high grade and low recovery, or,
alternatively, from high recovery in a
large underflow mass at low grade.
Neither of these conditions is ideal
TY,. L from the viewpoint of the present
investigation, the highest values of
both grade and recovery being
desirable.
Examination of the results shows
that gold recovery to the underflow
of 80 per cent is easily achieved with
either cyclone and under a variety of
operating conditions. If this is taken
as the minimum acceptable re-
covery, the "best" result is that
which yields this recovery in the
lowest underflow mass. In three tests
with the ewe 2M unit, this recovery
was achieved in underflow masses of
37 per cent. With the ewe 4 unit,
recoveries in excess of 80 per cent
TYPE M
were obtained in underflow masses
of 45 per cent. The L- type cone was
used in the tests on the ewe 2 unit,
and the M type in the tests on the
Fig. 9- The types of cone used in the Compound Water Cyclone ewe 4 unit. However, a test on the
ewe 4 unit with the L-type cone
gave a recovery of 78 per cent in an
underflow mass of 34 per cent.
I
The recoveries achieved appear to

be higher than those obtained in the
I
tertiary cyclones in Anglo American
plants. However, data on the latter
I
are rather scanty, lacking tonnage
measurements in particular. In ad-
I w dition, the test cyclones are small
Cl:
0 units, operated under ideal con-
U
ditions at low feed pulp densities.
I
Cl:
:cc Further laboratory tests that could
be carried out on these units include
~I
- -et.-- investigation of the effect of feed
pulp density and the determination
~I
et. oftheir dso pointIS for gold, which was
VORTEX FINDER-
precluded in the present investi-
CLEARANCE ~I
l 1-.
-~I-
gation because of lack of size data
below 44 /Lm. Furthermore,
cyclones are at a disadvantage under
these
~.
typical gold-plant conditions, be-
I cause the small size dictates

unrealistic level of operator atten-
tion. Therefore, despite the in-
an
I complete nature of the laboratory

investigation, it is felt that further
I tests should be carried out on ewe
i
,][
,u.
LEGEND
8 (200 mm) and/or ewe 12 (300 mm)
cyclones under plant or pilot-plant
conditions. The gravity-concen-
I . GANGUE
tration pilot plant at present being
commissioned provides ideal con-
I fJ
0
LIGHT MIDDLlNGS
HEAVY MIDDLINGS
.
ditions for these tests.
A blunt-cone cyclone is already
operating at Western Holdings Limi-
I C HEAVY CONCENTRATE
ted with apparent success. The unit
was designed by the Anglo American
Research Laboratories and is shown
Fig. ID-Separating process in a Compound Water Cyclone (after Visman)
in Fig. 14. It has a cone included
TABLE IV.
SELECTED RESULTS FROM CONCENTRATION TESTWORK USING COMPOUND WATER CYCLONES
Vortex FE ED UNDERFLOW
Test Cyclone Cone finder
no. diam. type clearance Pressure Solids Solids Au SoJids Solids Mass Cone. Re-
mm kg/cm" % g/I pulp Calc. % g/I pulp % Ratio covery S.l.
g/t Au %
---
1 50 M 13 0,60 25 297 108,1 62,0 1017 17,4 3,39 59,0 2,6
2 50 M 13 0,35 25 297 108,6 60,4 975 23,4 2,68 62,8 2,3
3 50 M 38 0,15 25 297 81,3 44,6 620 54,8 1,63 83,1 2,0
4 50 S 13 0,40 20 230 99,2 54,5 831 3,6 10,20 36,7 4,0
5 50 S 13 0,20 20 230 98,1 44,7 622 17,4 2,95 51,3 2,2
6 50 S 38 0,25 30 370 133,5 63,1 1048 25,5 2,53 64,6 2,3
7 50 L 13 0,05 20 230 71,1 26,9 324 57,1 1,37 78,4 1,6
8 50 L 13 0,60 30 370 90,9 62,1 1018 36,6 2,26 82,6 2,9
9 50 L 38 0,20 30 370 133,0 45,8 644 57,0 1,51 86,3 2,2
10 100 M 35 1,30 12 130 90,5 43,5 599 26,2 1,95 51,1 1,7
11 100 M 55 1,30 12 130 87,3 61,9 1015 37,5 2,05 76,4 2,4
12 100 M 90 1,30 12 130 98,8 64,9 1098 45,7 1,85 85,3 2,6
13 100 L 35 1,05 12 130 91,1 61,7 1011 27,4 2,31 63,3 2,1
14 100 L 55 1,05 12 130 105,6 61,7 1011 33,5 2,33 78,0 2,6
underflow grade
Concentration ratio =
calc. feed grade
plant as a whole. On those plants
having both gravity concentration
and cyanide circuits, the gravity-
concentration circuit must comple-
ment the cyanide circuit in that the
gold entering the cyanide circuit
must be of such a nature as to
respond with maximum efficiency
4
to cyanidation. Feather and Koen19
point out that the majority of the
gold in the residues is free but is
coated with iron oxides and has
gangue minerals pressed into the
surface. These grains are those
which have bypassed the gravity-
concentration circuit for one reason
3
or another, presumably "floating"
into the secondary-cyclone over-
flows after being flattened by re-
)( peated passes through the mill.
CD
The essence, then, of efficient
"C gravity concentration lies in selec-
tive classification of the gold to the
->- gravity-concentration circuit, accom-
:~ 2
- panied by a reduction of the re-
..!
"CD circulating load of gold in the mills.
VI It is felt that blunt-cone cyclones
CWC2M will prove to be of considerable
efficacy here, provided dilution of
13mm VF clearance
the mill circuit is not experienced.
At the same time, considerably
. cone type M improved recovery of the gold by the
concentrators is called for. This can
0 L be achieved in two ways: firstly, an
"
increase in the concentrator capacity
@ "
s to relieve the overloaded conditions
that exist (this is currently being
done on most Anglo American
plants); and, secondly, the install-
ation of novel machines that are
inherently more efficient than the
0.5 1.0
existing concentrators, which are
Pressure kg /cm2 possibly obsolete in terms of modern
Fig. I I-Selectivity Index versus inlet pressure for the Compound Water Cyclone practice. The work currently being
50 mm unit undertaken, coupled with the vastly
superior analytical tools available
angle of 120 degrees and is fed with and achieving similar results to our today (such as the electron micro-
tertiary cyclone underflow, thus re- own.
probe and high-speed computers),
placing a Johnson drum concentrator.
should elucidate the problems in-
The inlet pressure is of the order of
CONCLUSIONS volved and indicate the most prac-
0,2 kgfcm2. Recoveries of 70 per cent
Any improvement in the overall tical method of attaining optimum
in some 30 per cent by mass are
currently being obtained. recovery of gold by gravity con- and efficient operations. It is hoped
Close liaison is being kept with a centration and subsequent amal- that the two pilot plants will be
firm in Germany that is experi- gamation relies on the efficiency of very useful in optimizing the existing
menting in the field of cyclones used each concentration stage in the concentrators and improving the
as concentrators. This firm makes removal of gold from the pulp. concentrator capacity. This applies
cyclones from cast polyurethane, Since no machine can be 100 per especially to the Reichert-cone con-
which has definite maintenance ad- cent efficient, one must aim for the centrator plant, which uses the
vantages. They are currently build- highest unit efficiency in terms of principle of the pinched flume with-
ing concentrating cyclones with the presented pulp, as well as the out introducing wall effects and
wide-angle cones for test purposes highest unit efficiency in terms of the overcomes the problem of choked
riffles by removing the concentrates the probability that gold will become the various mines whose operations
through a slot. coated while in the mill circuit. we investigated, and of those com-
The ultimate result of this work parties who lent us equipment and
will be to reduce the inventory of ACKNOWLEDGEMENTS did testwork on our behalf.
gold held in the mill circuit and to We much appreciate the invalu-
reduce cyanide losses by lessening able assistance and co-operation of References
1. ADAMSON, R. J., ed. Gold metallurgy
" in South Africa. Johannesburg, Cham-
CWC4 ber of Mines of South Africa, 1972.
cone type VF clearance 2. WARTENWEILER, F. Development of
milling and cyanidation on the
mm Witwatersrand. Am. Inst. Min. Met.
0 M Engrs Trans, vol. 112 Milling Methods.
M 1935. pp. 760-784.
3 cD l 3. DoRR, J. V. N., and BOSQUI, F. L.
Cyanidation and concentration of
(i l gold and silver ores, 2nd edition.
X M New York, McGraw Hill, 1950.
.
."
le 4. ROSE, T. K., and NEwMAN, W. A. C.
The metallurgy of gold. 7th edition.
.!: London, Griffin, 1937.
~ 5. DoUGLAs, J. K. E., and MOIR, A. T.
:! 2 A review of South African gold re-
covery practice. Trans. 7th Common-
¥ wealth Min. and Met. Congress, 1961.
J 0 Johannesburg, South African In-
---0- stitute of Mining and Metallurgy.
~ Vol. Ill. pp. 171-1003.
6. JOHNSON, E. H. Concentration and
selective regrinding, J.Chem Met.
and Min. Soc. S. Afr., Apr. 1927.
pp. 215-220.
7. Plant report from Western Deep
Levels Limited, 1972.
8. WEYHER, L., and LOVELL, H. L. The
response of parameter variation in
the hydrocyclone processing of fine
coa!. AIME Trans., vo!. 235. Dec.
1966. pp. 333-340.
0.5 1.0 9. FONTEIN, F. J. Separation by cyclone
according to specific gravity. Cy-
Prenure kg/cm2 clones in industry, K. Rietenia and
C. G. Verver, eds. Amsterdam,
Fig. 12-Selectlvlty Index versus inlet pressure for the Compound Water Cyclone Elsevier Publishing Company, 1961.
100 mm unit pp. 118-134.
10. VISMAN, J. The cleaning of highly
TABLE V friable coals by water cyclones.
SCREEN ANALYSES AND GOLD DISTRIBUTION OF PRODUCTS FROM SELECTED TESTS Fourth International Coal Preparation
Congress, Harrogate, 1962. Paper C2,
pp. 155-162.
11. WEYHER, L. H. E., and LOVELL, H. L.
Overflow Underflow Au Hydrocyclone washing of fine coal.
Test Size recovery AIME Trans, vol. 244. 1969. pp.
no. fraction Au Au .to l!/f 191-203.
JLm Mass Au distrn. Mass Au distrn. III sIze
g/t g/t 12. FALcoNER, R. A., and LOVELL, H. L.
% % % % % The response of varying hydrocyclone
-- -- cone angles in fine coal cleaning.
297 3,7 120,60 3,5
2,9 92 AIME Trans, Dec. 1967. pp. 346-354.
210 22,64 0,5 4,1 67,60 2,1 13. SHEAHAN, P. M. Hydrocyclones. Fed-
149 10,4 7,84 0,6 6,5 52,64 2,7 82 eration of Malaya, Department of
105 15,6 9,52 1,2 4,4 98,92 3,4 74 Mines Research Division, Bulletin
6 74 12,0 11,88 1,1 2,9 178,20 4,0 78 No. 6. 1961.
53 12,0 18,06 1,7 1,9 325,56 4,8 74 14. SHEAHAN, P. M. A proposed dual
44 5,6 41,52 1,8 0,9 834,80 5,8 76 cyclone system for Malayan dredges.
-44 16,0 254,86 31,6 1,1 4 138,60 35,2 53 Min. .T., Lond., Feb. 10, 1961. pp.
146-147.
Original 74,5 66,95 38,5 25,5 316,36 61,5 15. LOPATIN, V. R., and DYESHCHITS,
V. S. Gravity beneficiation of gold-
containing conglomerates in short-
cone hydrocyclones. Tr. Tsent. Nauch.
105 3,5 14,05 0,6 8,6 29,3 2,9 83 Issled. Gornorazved. Inst. Tsvet. Redk.
74 9,8 5,85 0,7 8,3 32,30 3,0 81 Blagorod. Metall., no. 97. 1971.
13 53 14,4 6,30 1,1 5,8 64,80 4,4 80 pp. 97-103. (Translation available
44 8,3 11,30 1,1 2,6 186,10 5,7 84 from AARL).
-44 36,6 78,30 33,5 2,1 1 914,50 47,0 58
16. VISMAN, J. Bulk processing of fine
Original 72,6 42,47 37,0 27,4 197,57 63,0 materials by Compound Water Cy-
clones. Can. Min. Met. Bull. Mar.
1966. pp. 333-346.
17, GAUDIN, A. M. Selectivity Index: A
yardstick of the segregation accom-
plished by concentrating ope"ations
Trans. Am. Inst. Min. Met. Engrs,
vol. 87. 1930. pp. 483-487.
18. DAHLSTROM, D. A. Fundamentals
and applications of the liquid cyclone.
Chem. Eng. Progr., Symp. Series
No. 15, vol. 50. 1954. pp. 41-61.
19. FEATHER, C. E., and KOEN, G. M.
r-
I
The significance of the mineralogical A
and surface characteristics of gold A
grains in the recovery process. .T.
I
I
S.Afr. Inst. Min. Metall., vol. 73,
no. 7 Feb. 1973. pp. 223-234. L..-
DISCUSSION
1. S. K. DE KOK*, B.Sc. (Eng.)
Rand (Fellow)
cone apex angle
I wish to comment on the subject
of the recovery of gold before 1200
cyanidation, the beneficial effect of
which, to my mind, has emerged as
the most significant single finding
of the Anglo American Research
team. Their conclusion supports the
*Assistant Consulting Metallurgist, Anglo-
Transvaal Consolidated Investment Com-
r
II
,~
pany Limited I
4.0
@ section A A
3.0
scale 1:100
)(
. Fig. I4-Blunt-cone cyclone
"c
long-held convictions of the Anglo-
..>- vaal Group, which have found ex-
'>
.... 2.0 pression in the inclusion of gravity
. concentration within the grinding
. circuits and in the employment of
'" flotation, either in the grinding
circuit or in the milled-pulp stream
on its way to the cyanide plant.
CWC2M This contribution is made in the
cone type M hope that our experience in the field
1 of concentration may be of some
V F clearance mm value to the Anglo American in-
0 6-4 vestigators in their further efforts.
. 13
A
On reading between the lines of
the paper by Messrs Bath, Duncan,
X 38 and Rudolph and also those of the
related papers by other members of
the team, I get the impression that
they are possibly on the verge of
0.5' 1.0 repeating the error of omission that
Pre..ure kg/cm2 was made some twenty years ago
when we failed to realize the full
Fig. 13-Compound Water Cyclone 50 mm unit-effect of variation in the vortex-
finder clearance potential of the plane table as a
concentrator of refractory, as well shaking tables and in the barrel Eastern Transvaal Consolidated
as of free, gold. To substantiate this tailings; the net recovery of gold in Group, the plane table concentrates
view, I should first like to elaborate amalgam form was virtually un- after regrinding and removal of the
on the application of this device on changed from that achieved by the free gold and siliceous gangue, are
a few mines within our group and corduroy cloths. Of the gold con- added to the flotation concentrates
then attempt to draw a parallel tained in the shaking-table middlings for roasting and further grinding
with the authors' present work. some 30 per cent proved to be amal- (in cyanide solution) prior to cyanid-
The plane table, functioning gamable. In other words, it was ation.
essentially on the flowing-film prin- shown that the plane tables were Subsequent experience at Harte-
ciple, is an exceptionally high- more effective than the corduroy in beestfontein has confirmed the bene-
capacity roughing concentrator, recovering fine free gold, and that fit of a preliminary acid treatment.
which is a decided advantage in the shaking tables were not as On this plant, plane tables were
view of the trend towards large effective as the plane tables in initially installed in the secondary-
grinding units and high circulating retaining the flaky grains. The regrinding circuits, as at Rand Leases,
loads. At the usual pulp dilution, it mainder of the gold returned to the but in this case only 60 to 70 per
has a throughput of some 1500 ton- circuit was either encased in sul- cent of the gold in the mill feed was
nes of solids per day per metre width, phides or tarnished by iron oxide removed, the lower recovery almost
which means that it does not lend stains. The circuit finally adopted certainly being due to the assoc-
itself easily to laboratory or even was a stratagem devised to reduce iation of much of the gold with the
pilot-plant investigation, although the load on the shaking tables and light mineral, thucholite, and to the
I understand that NIM has had to subject the concentrate, grading extreme fineness of the free gold.
some success in this regard. It is 36 per cent minus 200 mesh, to a After re-dressing of the plane-table
extremely simple to operate, but separate fine grind. The concentrates concentrates on an endless-belt con-
requires careful initial adjustments from ten tables installed in the centrator and a shaking table, the
to get it functioning; once working, secondary grinding circuits were fed net gold recovery by amalgamation
it never stops. It was invented at to a pebble mill-2 m by 6 m (6 ft was reduced to 11 per cent. Con-
Rand Leases in 1949 by D. McLean 6 in by 20 ft)-and re concentrated centration was consequently aban-
for the sole purpose of replacing on a single plane table treating the doned. The cyanide residue value
corduroy strakes that were installed mill effluent. The concentrate from was far higher than desirable, even
in the secondary grinding circuits1. this table, approximately 15 at an extremely fine grind, but it
Now the corduroy cloth is probably t/d, was delivered to the shaking was subsequently reduced by 50
the most selective concentrator of table and thence to the amalgam per cent on the introduction of the
free-milling gold that has ever been barrel, while the classifier overflow, reverse-leach process of uranium-
used because it produces a small amounting to 3,6 per cent of the gold recovery, which subjected the
bulk of very high-grade concentrate total plant feed and grading 96 total pulp to acid treatment before
that is easily dressed on a shaking per cent minus 200 mesh, joined the cyanidation2. It is, of course, not
table before being charged into an secondary-mill products for delivery claimed that the reverse.leach pro-
amalgam barrel. The plane table, to cyanidation. Following the re- cedure is the economic solution to
on the other hand, produces a very duction in feed tonnage to the the problem in all cases, but it does
much greater bulk of lower-grade shaking table and the release of more point the way to the more effective
pyritic concentrate, which, to a far gold by the finer grind, total recovery treatment of concentrates containing
greater extent requires reconcen- by amalgamation increased by 5 gold that either is filmed with
tration so that the quantity is re- per cent to 50 per cent. The net coatings, for which expensive scrubb-
duced to that which can be con- effect of the circuit was that 3,6 ing or prolonged cyanidation may
veniently handled in the barrel. But per cent by mass of the ore contain- be the only alternative answer, or is
the removal of total gold from the ing 40 per cent of the total gold, encased in acid-soluble minerals,
grinding circuit is far higher than including the major portion of the which would not otherwise permit
that achieved by corduroys. At refractory gold, was allowed to cyanide attack. After all, uraninite
Rand Leases, the plane tables re- proceed to the cyanide plant, to- was first detected in our ores from
covered some 90 per cent of the gold gether with the remaining 96,4 per an examination of corduroy con-
being fed to the plant, which was cent of the ore containing only 10 centrates.
double that previously recovered in per cent of the gold, without any I turn now to the authors' work on
amalgam form by the cloth. How- preliminary treatment other than the more effective application of con-
ever, because of the constraint im- finer grinding. The mistake I have centration in an attempt to improve
posed on the system by the small referred to lay in not subjecting overall recovery. The Anglo Amer-
capacity of the amalgam barrels this small tonnage of high-grade ican plants employ the Johnson
and because not all the gold re- material to separate treatment such concentrator as the primary device;
covered by this primary or "rough- as more-intensive cyanidation or this is a fairly low-capacity, mech-
ing" device was amalgamable, some pretreatment with acid. It is not anized form of the corduroy cloth,
50 per cent of the value removed was irrelevant to note that, in the although it produces a greater bulk
returned to the mill in the middlings treatment of the far more refractory of concentrate. For cleaning of the
and tailings products from the ores of the three mines of the Johnson concentrate, use is made of
endless-belt concentrators, in which for the recovery of the finer sizes of gold, possibly in the form of a
the belt movement is against the the heavy minerals is, I feel, one of "clean-cut" for direct smelting or by
direction of pulp flow. For this the fields of research in which our in- flotation. The reduction in cyanide
purpose, I think consideration should dustry should be actively engaged. residue value resulting from im-
be given to the use of the Gallagher Arithmetical averages of the fig- proved amalgamation, noted by the
table, which is extensively used in ures presented by the authors, which authors, is not, in my opinion, due
Rhodesia but is hardly known in are sufficiently accurate for illus- to an increase in amalgam recovery
this country; in this device, the tration purposes, indicate that the per se, but rather to a reduction in
belt moves transversely at right endless- belt concentrators recover the quantity ofrefractory gold in the
angles to the pulp flow. I fully 70 to 75 per cent of the gold de- tailings from the amalgam barrel
agree with the authors' statement livered to them by the Johnsolls, that proceeds to cyanidation. There
that early removal of the gold is and that the amalgamation of the is undoubtedly far more of this
most desirable to avoid the deleter- concentrate so produced results in a material present in the endless-belt
ious effects of overgrinding, but net recovery of 35 per cent of the tailings than was ever present in the
the Anglo American grinding and gold fed to the plant. If, as stated, barrel tailings before changes were
concentration circuits as at present the Johnson concentrators recover made in the techniques of amalgam-
constituted are not conducive to some 70 per cent of the total gold, ation. Contrary to the authors'
efficiency in this respect. Concen- it is obvious that the endless-belt assertion that recovery by amal-
tration is confined to the tertiary tailings contain some 15 per cent gamation is inherently more efficient
stage of classification/milling, at of the gold and, in the circuit as at than that by cyanidation, I believe
which stage much of the liberated present composed, this is merely re- that amalgamable gold is the gold
but maltreated gold has already turned to the grinding circuit for that is also most easily cyanided. If
escaped to the cyanide circuit via further defacement and delivery to special treatment is required to
the secondary cyclones. For an in- the cyanide circuit; at least, at render the gold amalgamable, that
creased concentration recovery, re- Rand Leases this material was given same treatment will render it more
liance must therefore be placed on a separate finer grind. After em- amenable to cyanide. It is not im-
the concentrating effect of the cy- phasizing the importance of a high portant that the gold recovery as
clones to deliver as much of the gold recovery of gold before cyanidation, amalgam be 30 per cent or 50 per
as possible to the Johnson con- the authors have quite rightly cent or any other figure; what is
centrators. As a consequence, the applied themselves to improving the important is that, if concentration
authors are having to investigate primary recovery, but so far have has been considered necessary to
wider-angled cyclones, which tend neglected to investigate the correct relieve the cyanide plant, advantage
to suppress the classifying and treatment of the gold that is lost in be taken of the opportunity to
accentuate the concentrating effects the process of reducing the bulk to correctly prepare for cyanidation the
of settling. Earlier removal of the suit the amalgamation process. They gold in the primary concentrate that
gold would seem to be the logical admit that doubling the speed of has not reached the barrel, as well
method, but this would require the the endless belts increased total as the gold that amalgamation has
concentrators to be placed within the gold recovery by amalgamation by failed to recover. In failing to make
circulating loads ofthe primary- and 2,5 per cent but at the expense of full use of this opportunity, the
secondary-milling circuits. The low increasing concentrate mass by 50 authors are perpetuating the myth
capacity of the standard-sized John- per cent. I submit that there is a far that concentration is synonomous
sons is not a factor in their favour greater potential pay-off in investi- with amalgamation, just as we did
for this purpose, and the authors are gating the non-amalgamable gold, until we saw the light. It is not
investigating devices such as the which the primary concentrators surprising that, in the days when
blunt-cone cyclone and the Reichert have handed to them on a plate, concentration was regarded merely
cone, which is but a version of the than in an improvement in the as a means to amalgamation, the
plane table without its side-wall amalgamation process itself. Indeed, proponents of all-cyanidation were
effects. Experiments have been the complete elimination of amal- not convinced of the value of con-
carried out at the Anglovaal Lab- gamation, besides removing the re- centration.
oratory with laboratory-sized vessels strictions imposed by the limitations The advantages of separate treat-
lined with riffled rubber. These are of the plant, would obviate the very ment of the high-grade, more-re-
somewhat similar to the Johnson real danger of mercury escaping with fractory portion of the ore are
concentrator but are conical in shape the pulp and inhibiting cyanidation. highlighted at three of our mines
and run at super critical speeds; This prompts me to ask whether on which flotation plants were in-
the centrifugal forces developed give, the improvement in total recovery stalled for the recovery of pyrite
in open circuit, 90 per cent recovery following an improvement in amal- required for acid manufacture for
of the gold at high ratios of concen- gamation is not in part due to a the uranium industry3. Anglovaal
tration, but, as with all centrifugal reduction in mercury loss. Alter- broke with the traditional practice
machines of this type, continuous native methods of treatment of the of recovering this pyrite from the
discharge of the concentrate presents total primary concentrate could be uranium-plant tailings by inter-
difficulties. The development of a con- devised that would provide for the posing the flotation plants between
centrator utilizing centrifugal forces recovery of the osmiridium and free grinding and cyanidation. At one
of the Hartebeestfontein plants, the 2. J. LEVI N*, B.Sc. (Eng) Rand tised, the quantity of material fed
flotation cells are installed within (Member) to the concentrate is generally
the grinding circuits, whereas, at The National Institute for Metal- several times that of the new feed
Loraine and the other Hartebeest- lurgy is engaged on tests on the entering the circuit, and the pos-
fontein plant (formerly Zandpan), concentration of uranium, and, be- sibility of applying concentration
flotation is carried out on the final cause gold and uranium are closely in an open circuit is worth con-
mill pulp. Gravity concentration in associated in the ores of the Wit- sidering. To ascertain if recoveries
the grinding circuit has been re- watersrand, some of the observations from an open circuit would be
tained at Loraine, both for osmirid- that have been made have a bearing substantially lower than from a
ium recovery and because of our on the problems of gold extraction. closed circuit, laboratory tabling
experience, verified on the Eastern Currently, an ore from the Vaal Reef tests were done on material ground
Transvaal Consolidated Group of is being examined, and the infor- to various extents, and the results
mines, that flotation is least success- mation obtained so far points to the were compared with those obtained
ful in recovering the very finest and following tentative conclusions, which from a test in which closed-circuit
the very coarsest fractions of the are based on the experimental results grinding and concentration were
gold particles. On all the mines, the shown in the accompanying tables. simulated. Table I summarizes the
recovery of gold into the flotation (1) Gravity concentration for the results obtained. The figures actually
concentrates exceeds 80 per cent. It recovery of gold and uranium suggest that, if the optimum degree
is of interest to note that well over can be carried out advantage- of grinding for open-circuit con-
50 per cent of the total gold can be ously in an open circuit, instead centration is used, the recoveries in
floated with the addition of frother of in the usual closed grinding- open circuit will be higher than from
only and without the addition of classification circuit. a closed circuit, but it is preferred
xanthate or any other pyrite pro- (2) The proportion of concentrate merely to draw the conclusion that
moter, the main dilutent in the made by gravity or flotation the recoveries of gold and uranium
high-grade concentrate being fine concentration should be fairly from an open circuit should not be
siliceous gangue. The separate re- high-between 5 and 10 per markedly lower than from a closed
grinding and intensive cyanidation cent of the ore-so that the circuit. Table I also shows that
of the pyrite concentrates result in maximum amount of gold and substantially increased recoveries are
99 per cent recovery of the contained uranium can be recovered. effected when the mass of concen-
gold. The net results have been (3) Extraction of gold by cyanid- trate is increased from 5 to 10 per
either increased overall recoveries or ation is increased significantly cent of the feed.
unchanged recoveries with increased when cyanidation follows acid- Gold extraction from products of
tonnage throughputs, the natural .leaching for the dissolution of gravity concentration
consequence of decreased depend- uranium; a significant increase A batch of coarsely ground ore
ence on fine overall grinding. At is observed even when the was concentrated by use of a Hum-
Hartebeestfontein, the combination tailings are subjected to prior phreys spiral and a shaking table
of reverse-leach treatment of flo- acid leaching. to provide three concentrates and a
tation tailings, reground concen- (4) Satisfactory extraction of gold tailing. The four products obtained
trates, and calcines has resulted in from gravity concentrates can were ground to various degrees, acid
an increase in recovery of some 2t be obtained by cyanidation after leached, and cyanided. Fig. 1 shows
per cent at a grind coarser by 18 fine grinding and acid leaching, how the concentration was effected
per cent minus 200 mesh. without resort to amalgamation. and the results obtained. Table II
(Amalgamation would not be shows the results of cyanidation of
In conclusion, I wish to congratu-
applicable to the relatively large the individual products of con-
late the authors on their stimulating
proportion of concentrate being centration. The very high extractions
paper. Of their overall philosophy
recommended. ) of gold obtained from the first con-
and approach there can be no (5) Flotation effects a higher re-
criticism. They are tackling the centrate are noteworthy, and it
covery of gold than does appears that this concentrate does
problem in a methodical sophisti-
gravity concentration. not include gold in a form that needs
cated manner; I venture to predict
(6) When gravity concentration is intensive treatment for its extraction
that the evidence they are still to
used, the tailing is the largest by cyanidation. However, all the
gather from their pilot-plant will
source of gold not extracted by other products need to be finely
soon result in a revision of their
cyanidation. When flotation is ground and acid leached to enable
priorities.
used, the first xanthate con- satisfactory extractions of the gold
centrate is the largest source of to be achieved. The high gold con-
References
gold not extracted by cyani- tent of the tailings should be noted.
1. ZADKIN, T. The Rand leases plane dation.
table. .f. Chem. metall. min. Soc. S.Afr., Concentration by flotation
vol. 54. Feb. 1954. pp. 292-298. Gravity concentration Fig. 2 illustrates the procedure
2. BRITTEN, H., and DE KOK, S. K. When concentration in a closed used in concentrating the ore by
Uranium in South Africa, vol. 1, p. 462. grinding-classification circuit is prac-
3. BUSHELL, L. A. The flotation plant flotation, and Table III gives the
of the Anglovaal Group. J. S.Afr. Inst. *Head, Ore-dressing Division, National results of the pretreatment and
Min. Metall., Jan. 1970. Institute for Metallurgy cyanidation tests done on the flo-
JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY
JUNE 1973 381
tation products. It can be seen that TABLE t
COMPARISON OF GRAVITY CONCENTRATION IN OPEN AND CLOSED CIRCUITS
57,6 per cent of the gold was re-
covered by flotation with only a
Gold recovery, % at
frother. Cyanidation of the frother
concentrate was very effective, al- Circuit Grind, %-75fLm 5 % mass of conc. 10% mass of conc.
though acid leaching is essential if simulated
maximum extraction is to be Open 40 43,0 63,8
achieved, and acid leaching in this Open 50 51,0 62,0
Open 61 48,2 65,4
instance appears to be more impor- Closed 3 stages; 45,5 58,0
tant than fine grinding. Cyanidation 1st stage
of the xanthate concentrate was far 40%-75fLm
less satisfactory, and it is evident
that these concentrates contain gold
in a form that requires intensive treat-
ment for its extraction. The residues
TABLE 11
from the flotation tailing are mar- CYANIDATION OF PRODUCTS OF GRAVITY CONCENTRATION
kedly lower than those obtained (See Fig. 1).
from gravity concentration, but
even it benefits from acid leaching Distbn. of gold, % of gold in:
Gold in cyanide
prior to cyanidation, the gold assay residue g/t Prod. of conc. Original ore
value being reduced by approxi- Concentra- Grind
mately 0,14 g/t for all degrees of tion Product fLm Dir.* Rev.t Dir.* Rev.t Dir.* Rev.t
grinding tested. 1st conc. -150fLm 7,0 3,1 0,30 0,20 0,09 0,06
Mineralogical examinations -100fLm 4,5 4,0 0,21 0,14 0,06 0,06
- 53 fLm 10,0 4,9 0,54 0,20 0,16 0,06
Approximately half the gold ob-
served in the first two gravity con- 2nd conc. -150 fLm 8,4 3,6 9,0 3,8 2,49 1,05
-100fLm 5,0 2,1 5,4 2,3 1,49 0,64
centrates was in the form of free 53 fLm 3,7 2,4 3,5 2,5 0,98 0,69
grains, the remainder being assoc- -
iated in roughly equal amounts with 3rd conc. -150 fLm 10,5 3,9 9,1 3,4 I,ll 0,42
-100 fLm 7,5 2,6 6,5 2,2 0,79 0,27
sulphide or uraninite. In the third 5,2 2,7 4,2 2,4 0,51
- 53 fLm 0,29
gravity concentrate, free gold is less
abundant, and gold associated with Tailing 53%-75fLm 0,84 0,52 13,5 8,4 4,04 2,50
65%-75fLm 0,74 0,37 11,9 6,0 3,56 1,78
uraninite is more abundant; this 76%-75fLm 0,47 0,25 7,6 4,0 2,26 1,20
concentrate also contains a marked 87%-75fLm 0,39 0,26 6,3 4,2 1,87 1,25
proportion of grains in which thu-
cholite is attached to uraninite.
Although most of the gold grains *Direct cyanidation
tCyanidation after acid leaching
appeared to be discoloured or stained
to various extents, the considerable
proportion of grains in which gold
TABLE III
is associated with uraninite suggests CYANIDATION OF PRODUCTS OF FLOTATION
that much of the benefit obtained (See Fig. 2).
from acid leaching prior to cyani-
dation can be ascribed to the Distbn. of gold, % of gold in:
association of gold and uranium. Gold in cyanide
residue g/t Prod. of flotn. Original ore
However, there is no definite evi- Concentra- Grind
dence of the magnitude of the tion Product Dir.* Rev.t Dir.* Rev.t Dir.* Rev.t
-- ~-
benefit that might have resulted Frother conc. -150 fLm 17,8 4,7 1,20 0,30 0,69 0,17
from the removal of coatings from -100 fLm 13,7 2,1 0,90 0,14 0,52 0,08
the gold grains by the acid leaching - 53 fLm 11,1 3,1 0,72 0,20 0,42 0,11
~-
step. 1st xanthate -150 fLm 15,7 10,8 10,70 7,40 3,30 2,78
The frother concentrate obtained conc. -100 fLm 10,3 6,9 7,10 4,70 2,19 1,45
by flotation contained numerous - 53 fLm
~-
free grains of gold and numerous 2nd xanthate -150 fLm 11,4 8,5 26,8 20,0 1,34 1,00
grains in which the gold was asso- conc. -100 fLm 7,0 4,4 16,5 10,4 0,82 0,52
ciated with pyrite or uranium min- - 53 fLm
erals; the latter were predominantly Tailing 57%-75fLm 0,40 0,27 29,0 19,7 1,91 1,30
thucholite, but there were also 67%-75fLm 0,30 0,22 21,4 16,2 1,42 1,05
75%-75fLm 0,31 0,17 22,6 12,2 1,49 0,80
numerous grains in which uraninite 85%-75fLm 0,24 0,10 17,39 7,0 1,15 0,46
was associated with the thucholite,
and some grains that appeared to *Direct cyanidation
be wholly uraninite. tCyanidation after acid leaching
xamination of the two xanthate [rother concentrate and. that more 3. E. A. D. RUBIOGE- (i=ellow)
concentrates indicates that they gold was associated with sulphides
The paper reported many factors
contained less free gold than the than with uraninite or thucholite.
relating to the efficiencies attainable
in amalgamation.
At one stage the paper defines the
word "liberation" when applied to
gold particles and states that the
Batch grind to 45 % - 75 /Lm
mineral grain "for gravity concen-
tration ... must be almost com-
I pletely separated from the gangue."
HumPh'!" "p"'" In his address relating to this para-
meter, Mr Rudolph remarked that
hydrated iron oxides often covered
the surfaces of the gold particles and
I
prevented or delayed the process of
Bulk concentrate amalgamation.
It is on this point that I wish to
I
report on some experiments carried
Bhok;.. tab,!, out in the Rand Leases Assay
Department in 1954. They dealt
with concentrates from the Virginia
Mine. The metallurgist in charge
I I
found that barrel tailings repeatedly
I I
Concentrate 1 Concentrate 2 Concentrate 3 Tailing contained "floured" mercury, which,
Mass, % I
0,29
I
5,71
I
2,01
T 91,99
of course, carried much gold. He
asked that tests be carried out to
indicate the cause of the trouble
and, if possible, provide information
Gold, gjt 2005 92,5 116,0 6,2
that would lead to effective amal-
Gold distribution, % 30,4 27,7 12,2 29,7 gamation.
The material contained many par-
U30., gjt 4126 1258 1976 230 ticles that were described as "rusty
gold", which was, of course, caused
U3O. distribution, % 3,6 21,4 11,9 63,1
by the "hydrated iron oxide" re-
Fig. I-Procedure and results for gravity concentration test
ferred to in the paper under dis-
cussion.
The line of investigation under-
taken on that occasion was mainly
directed to conditioning of the con-
centrate by chemical means before
proceeding with the amalgamation
Batch grind to 52%-75/Lm
tests.
Only five of the many significant
I tests will be reported here, with the
Float in stages
purpose of high-lighting the part
played by the use of cyanide to clean
I the gold particles before the addition
of mercury and rolling for a very
I I I I
brief period without further grinding.
Frother 1st xanthate 2nd xanthate Tailing Weighed portions of concentrate,
concentrate concentrate concentrate
with water added, were dealt with
I
as follows: (see table on page 384)
Mass, % I
0,74
I
4,12
[
2,29
I
!!2,85
Note: In all these tests, no grind-
ing balls were added. It was assumed
Gold, gjt 1520 146 42,6 1,38 that the grinding that usually accom-
panies amalgamation work adds to
Gold distribution, % 57,6 30,8 5,0 6,6 the build-up of hydrated iron oxide
U.O., gjt
film. This would not, of course,
5284 1748 1945 212
occur to any great extent in the
U3O. distribution, % 11,1 20,4 12,6 55,9
*Formerly Group Assayer, Anglo- Trans-
vaal Consolidated Investment Company
Fig. 2-Procedure and results for concentration by flotation Limited
Test
No. Treatment given Observations
---- --- ----.
Washed, CaO and KMnO. added, and rolled on lab- Much "floured"
oratory rolls mercury
la Washed out all CaO and KMnO. from No. 1 above, 98 % of gold found in
added CaO and KCN, and rolled for 30 minutes amalgam
with Hg
2 Washed three times with water, then added CaO and 68,3 % of gold
Hg, and rolled for 15 minutes recovered
3 Treated as in 2 above, added aaa, KCN, and Hg, 92,6 % of gold in

and rolled for 15 minutes amalgam
4 Treated with CaO & KNOs and, without washing, 74,1 % of gold in
added Hg and rolled for 15 minutes amalgam
5 Treated as in 4 above but added KCN before rolling 95,4 % of gold in

for 15 minutes X amalgam
short time of rolling reported in the It will be seen that cyanide would or assay proved the need for such a
above tests. None-the-Iess, it is dissolve the very fine or flattened move, but returning the tailings to
of importance to note that, with particles of gold (probably a desir- circuit (or other further treatment)
the gold surfaces cleaned (and thus able feature!), and the transfer of after more than 90 per cent of the
"almost completely separated") by this gold-bearing solution would coarse gold has been removed would
cyanide, the amalgamation appears have to be attended to in any not, I am sure, raise any anxieties.
to take place almost immediately. plant designed to follow the sug- Perhaps the authors of the paper
Many other tests were carried out, gested procedure. would find it useful to apply their
and, wherever cyanide was omitted, The need for the further grinding very thorough skills to proving the
the extraction was much poorer in of the resulting barrel tailing might, value or:. ,otherwise of the above
these short-time runs. of course, be necessary if inspection findings.
Company affiliates
The following members have been Hewitt-Robins-Denver (Pty) Limited. Rustenburg Platinum Mines Limited
admitted to the Institute as Com- Highveld Steel and Vanadium Corpo- (Rustenburg Section).
pany affiliates. ration Limited. St. Helena Gold Mines Limited.
Impala Platinum Limited. Shaft Sinkers (Pty) Limited.
AE & Cl Limited.
AfroxjDowson and Dobson Limited. James Sydney & Company (Pty) S.A. Land Exploration Co. Limited.
Amalgamated Collieries of S.A. Limi- Limited. Stilfontein G.M. Co. Limited.
ted. Kinross Mines Limited. The Griqualand Exploration and Fi-
Apex Mines Limited. Kloof Gold Mining Co. Limited. nance Co. Limited.
Associated Manganese Mines of S.A. Lennings Holdings Limited. The Messina (Transvaal) Develop-
Limited. Leslie G.M. Limited. ment Co. Limited.
Blackwood Hodge (SA) Limited. Libanon G.M. Co. Limited. Trans-Natal Coal Corporation Limit-
Blyvooruitzicht G.M. Co. Ltd. Lonrho S.A. Limited. ed.
Boart & Hard Metal Products S.A. Loraine Gold Mines Limited. Tv!' .Cons. Land & Exploration Co.
Limited. Marievale Consolidated Mines Limit- Tsumeb Corp. Limited.
Bracken Mines Limited. ed. Union Corporation Limited.
Buffelsfontein G.M. Co. Limited. Matte Smelters (Pty) Limited. Vaal Reefs Exploration & Mining Co.
Consolidated Murchison (Tvl) Gold- Northern Lime Co. Limited.
Limited.
fields & Development Co. Limited. O'Okiep Copper Company Limited.
Venters post G.M. Co. Limited.
Doornfontein G.M. Co. Limited. Palabora Mining Co. Limited
Vergenoeg Mining Co. (Pty) Limited.
Durban Roodepoort Deep Limited. President Stern G.M. Co. Limited.
Virginia O.F.S. Gold Mining Co.
East Driefontein G.M. Co. Limited. Pretoria Portland Cement Co. Limit-
Limited.
East Rand Prop. Mines Limited. ed.
Free State SaaiplaasG.M. Co. Limited. Prieska Copper Mines (Pty) Limited. Vlakfontein G.M. Co. Limited.
Fraser & Chalmers SA (Pty) Limited. Rand Mines Limited. Welkom Gold Mining Co. Ltd.
Goldfields of SA Limited. Rooiberg Minerals Development Co. West Driefontein G.M. Co. Limited.
The Grootvlei (Pty) Mines Limited. Limited. Western Deep Levels Limited.
Harmony Gold Mining Co. Limited. Rustenburg Platinum Mines Limited Western Holdings Limited.
Hartebeesfontein G.M. Co. Limited. (Union Section). Winkelhaak Mines Limited.

Some Factors Influencing Gold Recovery by Gravity Concentration

Uploaded by

Copyright:

Available Formats

Some Factors Influencing Gold Recovery by Gravity Concentration

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Some Factors Influencing Gold Recovery by Gravity Concentration

Uploaded by

Copyright:

Available Formats

Some factors influencing gold recovery by gravity

HISTORICAL concentration on the Rand mines (2) Johnson concentrators,

Fig. la-Concentration with secondary milling

Fig. Ib-Concentration with tertiary milling

.~ Fig. 2-Strake tests-recovery versus grind

lower settling rates of the fine ! 1

TESTWORK ON THE ENDLESS-

up a plane inclined at 15 degrees to oS z I I

figure of 63 per cent recovery at a

0> "" '<1"""""""<1'

ment of balance between the centri-

wide cone angle-in

stratifies according to density along

Fig. 8-Pilot plant no. 2.

% apex. The separation thus takes place

The recoveries achieved appear to

I cause the small size dictates

I complete nature of the laboratory

3 Treated as in 2 above, added aaa, KCN, and Hg, 92,6 % of gold in

5 Treated as in 4 above but added KCN before rolling 95,4 % of gold in

You might also like