Está en la página 1de 14

Studies on Differential Flotation Characteristics of

Arsenopyrite/Pyrite Concentrate
Investigation of the differential flotation characteristics between
arsenopyrite and pyrite in a ore concentrate have been carried out on a
laboratory Leeds flotation cell. Experimental results indicate that some
limited separation between arsenopyrite and pyrite is possible by
differential flotation under the conditions of high pH and low pulp redox
potential. The best perfonnance in this testwonc was achieved by using an
oxidizing agent at pH 10.7. Of the oxidizing agents tested, NaOCl gave
better oxidizing effect than KMn04. It has also been observed that the
pulp redox potential depends on the pulp pH value. At high pH values a
hydroxide layer is fonned on the arsenopyrite surface, which has the
effect of a depressant on the arsenopyrite.
Arsenopyrite is an arsenic mineral. It is usually associated with
precious metal ores, and the minerals galena, sphalerite and
pyrite. Its presence in an ore deposit can be of vital economic
significance. Arsenopyrite may carry significant fraction of the
gold present in certain ores. Such gold may be present as separate
grains between arsenopyrite crystals and may be extracted by
direct cynidation (Heinen et al, 1980). Gold can also be found in
solid solution or small inclusions in arsenopyrite (Clark, 1960)
which necessitates the use of more unusual treatment techniques
(Addison, 1980). It may be useful to separate arsenopyrite and
pyrite so that they can be subsequently processed by different
methods to recover gold.
When arsenopyrite in an ore is not associated with gold values,
it is considered to be a nuisance impurity and its selective
depression is beneficial. The presence of arsenopyrite in sulphide
concentrate can cause severe health hazards and generate arsine
during pyrometallurgical and hydrometallurgical processing and
refining (Habashi and Ismail, 1975).
Rotation is the only cost effective method for separation of
arsenic bearing concentrate from pyrite. In the past few decades
researchers have made attempts to separate arsenopyrite and
pyrite by flotation. In the early-1960s, several Russians scientists
studied the separation of arsenopyrite and pyrite in the following
way (Glembotski, Klassen and Plaskin, 1963):
1. flotation of pyrite by reducing the dissolved oxygen;
2. depression of arsenopyrite by using lime;
3. depression of both minerals followed by activation of
arsenopyrite using copper sulfate;
4. depression of both minerals by using lime, followed by
activation of pyrite by ammonium chloride; and
5. depression of both minerals by using sodium sulfide
which was removed subsequently by dewatering,
followed by oxidation with oxidising agents (pyrolusite).
Beattie and Poling (1988) conducted laboratory flotation tests
on several ores and bulk concentrates to evaluate the
1. Western Australian School of Mines, PO Box 597, Kalgoorlie
2. Kalgoorlie Metallurgical Laboratory, Chemistry Centre (WA),
Department of Minerals and Energy, PO Box 881, Kalgoorlie
effectiveness of chemical oxidising agents as selective
depressants for arsenopyrite. They reported that the maximum
flotation recovery of arsenopyrite occured at pH values less than
approximately 7.0. Increasing pH resulted in decreasing
flotability of arsenopyrite only under oxidising condition. The
selective depression of arsenopyrite from bulk pyrite-arsenopyrite
concentrate was achieved through the use of an appropriate
oxidising agents such as hydrogen peroxide or sodium
Li and Zhan (1989) showed that arsenopyrite could be
depressed heavily in alkaline media. However, the presence of
heavy metal ion, such as Cu
+, made separation difficult
O'Conner and Bradshaw (1990) recovered 74.8 per cent
arsenopyrite and only 8.4 per cent pyrite by two-stage differential
flotation. They used dithiophosphate in the first stage at pH 11
and copper sulfate and dithiocarbonate in the second stage of
flotation. Iwasaki et al (1989) found that flotability of
arsenopyrite was improved in a nitrogen atmosphere and
decreased markedly by increasing pH above 7.
In this investigation an attempt was made to separate
arsenopyrite and pyrite by differential flotation using a laboratory
Leeds flotation cell. Rotation tests were carried out to observe
the control that could be exerted over the flotation of arsenopyrite
through the use of oxidising agent. All flotation tests were
performed with an arsenopyrite-pyrite concentrate obtained from
a Western Australian Gold Mine.
Materials used were as follows:
MIRC (frother), PAX (collector), NaOH (pH modifier), NaOCI
(oxidant), Ca(OClh (oxidant) and KMn04 (oxidant). The ore
sample studied was flotation concentrate 80 per cent passing 200
mesh (74 microns) produced from a flotation circuit. The major
elements were determined by chemical analysis. The average
contents are as follows: Au = 60 g/t, Fe = 20 per cent, S = 20 per
cent, and As = ten per cent. Mineralogical examination using
XRD, light microscopy and scanning electron microscopy
showed that the concentrate was mainly composed of pyrite and
arsenopyrite, with talc and minor quartz. The grains of these
minerals ranged between 100 Ilm and sub-micron sizes. The
concentrate also contained minor amounts of iron-nickel
sulpharsenide, slightly manganoan magnetite, galena and silver
bearing gold with silver content estimated to be three to five per
Experimental procedure
The flotation tests were conducted in a three litre Leeds flotation
cell. The ore sample, unless otherwise stated, was tap water
washed using a pressure ftlter; this was then used as the flotation
feed. The pulp density was adjusted by addition oftap water to 25
per cent (by weight). The pH level was adjusted with NaOH.
Unless otherwise stated, the dosage of PAX was 50 g/t and MIRC
38 g/t of the flotation feed. The pulp redox potential Eh was
adjusted either by NaOCI or by KMn04 and measured using a
redox potential meter. The redox potential readings obtained from
the meter were values relative to a Ag/AgCIIKCI (1.0 M)
reference at 25C. The feed was conditioned for ten minutes.
Extractive Metallurgy of Gold and Base Metals Kalgoorlie. 2628 October 1992 217
Flotation results on the Leeds cell.
Test No. Test Conditions Gr:lde (%) Recovery (%) Con. (%) Note
pH Eh Oxidant FeAsS FeS2 FeAsS FeS2
1. 5.44 121
16.0 28.9 14.9 22.7 21.1
2. 8.44 100
13.5 35.4 15.1 32.5 24.9 pH
3. 10.68 19
11.3 47.8 8.6 26.8 16.3
4. 8.08 140 NaOC1 16.5 23.4 22.2 28.1 30.7
5. 8.24 175 NaOCI 17.8 12.2 23.0 14.8 29.0
6. 8.10 250 NaOCI 17.4 10.0 18.1 8.5 22.7 NaOCI
7. 4.80 '220 Ca(OClh 18.7 20.3 21.3 18.9 26.1
8. 8.36 -90
16.5 28.6 21.0 29.5 27.0
9. 4.30 100
21.7 37.4 61.6 793 59.6
10. 5.60 160 KMn04 17.0 35.9 30.9 53.6 41.4
11. 530 360 KMn04 17.0 12.4 21.6 27.8 30.8
In the present investigation, the following methods were tested:
1. Flotation of pyrite by depressing arsenopyrite at high pH
2. Flotation of pyrite by depressing arsenopyrite using
oxidants; and
3. The combination of the above two.
The recoveries and grades of arsenopyrite and pyrite minerals
under various test conditions are listed in Table 1.

I 30


--0-- __".&2
10 11
Effect of pH level
From tests 1 to 3 as listed in Table 1 and plotted in Figure I, it
can be seen that pH level has a small favourable effect on
differential flotation between arsenopyrite and pyrite. The
recovery of arsenopyrite decreases and recovery of pyrite
increases, with increase in pH level. The highest recovery of
pyrite (32.5 per cent) oeeured at pH 8.4 and the lowest recovery
of arsenopyrite (8.6 per cent) oeeured at pH 10.7.
The effect of change in pH on the grades of arsenopyrite and
pyrite was similar to that of their recovery. The observation that
the higher pH results in the poorer flotability of arsenopyrite
agrees with that reported by Huang and Wang (1985). Further,
according to Beattie and Poling (1987) the oxidation of
arsenopyrite above pH 7.0 resulted in a ferric hydroxide layer
forming on the surface, which inhibited the oxidation of xanthate
to dixanthogen. Although minor differential flotation between
pyrite and arsenopyrite takes place, their grades in tailings do not
show any significant change at different pH values.
The pulp redox potential Eh was found to be significantly
affected by changes in pH. As can be seen from Table I, the
redox potential drops from 121 to 19 mv as the pH increases from
5.44 to 10.68.
Therefore, it can be concluded that by controlling the pulp pH,
a small amount of differential flotation between arsenopyrite and
pyrite is possible.
FIG I - Effect of pH level.
Effect of oxidising agents
Sodium hypochlorite and calcium hypochlorite as
oxidising agents
As shown in Figure 2 (corresponding to tests 4, 5 and 6) with
increase in redox potential, the recoveries and grades of pyrite
decrease more than those of arsenopyrite. Therefore it is the
pyrite that is being depressed at higher redox potentials. This was
also observed in test 7 where calcium hypochlorite was applied as
an oxidising agent at pH 4.8.
Potassium permanganate as an oxidising agent
In test 8, where the original sample as received from the plant
was used, much NaOH was added to raise the pH to 8.36.
Consequently, in spite of the small addition of KMn04, a
negative redox potential was observed. In this test, both at a
higher pH and in the oxidising environment, better performance
was not achieved when compared to that of test 2.
218 Kalgoorlie. 26 28 October 1992
Extractive Metallurgy of Gold and Base Metals
Fio 2 - Effect of Eh using NaOCl.
. .. ..
110 140 '10 111 200 ..
.. .,.....---------------1
g: tl
t ..

arsenopyrite. Above a redox potential of 150 mv, a lower
recovery of pyrite compared to arsenopyrite resulted. The
arsenopyrite recovery, however, did not change significantly. The
greater depressing effect of KMn04 on arsenopyrite and pyrite is
shown at higher redox potentials.
3. The best conditions for differential flotation between
arsenopyrite and pyrite as determined by this work.
occurs when using an oxidising agent at a high pH.
4. A high pH alone will not usually result in an effective
differential flotation.
Finally, from these series of flotation tests, the best differential
flotation conditions for this type of sample are: maintaining a pH
of 10.7 and addition of a small amount of NaOCI. However, these
conditions alone will not result in an effective pyrite and
arsenopyrite separation.
In tests 9,10 and 11, lower pH values were maintained and
KMn04 was used 10 adjust the redox potential. The best
separation, as shown in Figure 3, occured at a redox potential of



.. --0-- litIloIIII,.11.

21' 10'
,.. ...

FIO 3 - Effect of Eh using KMn04.
1. In the differential flotation of arsenopyrite and pyrite pH
is an important factor. The optimum pH suitable for
depressing the arsenopyrite is 10.7.
2. The pulp redox potential was found to be dependent on
the pH. Even if the redox potential is negative, flotation
still occurs as long as the pH is suitably maintained.
In addition, in terms of the surface oxidation, oxidant NaOCl
shows much higher depressing behaviour than KMn04. The
greater depressing effect of NaOCl on arsenopyrite at lower
redox potentials resulted in higher recovery of pyrite compared to
The authors wish to thank Or John Hosking, Director, Chemistry
Centre (W.A.) for permission 10 present this paper and are also
indebted 10 Or Tony Bagshaw and Professor David Spottiswood
for their comments. This investigation was funded by a MERIWA
grant, for which the authors are grateful.
Addison, R, 1980. Gold and silver extraction from sulphide ores. Min
Congr J, Oct, pp 47-54.
Beauie, M J V and Poling, G W, 1987. A study of the surface oxidation of
arsenopyrite using cyclic voltarnetry, InJernational J Mineral
Proassing, pp 87-108.
Beauie, M J V and Poling, G W, 1988. Flotation depression of
arsenopyrite through use of oxidising agents, Trans IMM, 97(C), pp
15-20 (The Institution of Mining and Metallurgy: London).
Clark, L A, 1960. The Fe-As-S system: phase relations and applications,
Econ Geol, 55, pp 1631- 1652.
Glembolski, V A, Klassen, V I and Plaksin, I N, 1963. Flotation,
MonumenJ press, New- York, USA, p 540.
Habashi, F and !smail, M I, 1975. Health hazards and pollution in the
metallurgical industry due to phosphine and arsine, CIM Bull,
August, pp 99-104.
Heinen, H J, McClel1and, G E and Lindstrom, R E, 1980. Recovery of
gold from tusenopyrite concentrates by cyanidation-carbon
adsorption, USBM, RI 8458, pp 1-40.
Huang, K and Wong, D, 1985. A study of selective flotation of antimonite
and arsenopyrite, Nonfe"ous Metals, 37:(2), pp 22-29.
Iwasaki, I, Malicsi, AS, Li, X and Weiblen, P W, 1989. Insights into
beneficiation losses of platinum group metals from gabboric rocks,
Challenges in Mineral Processing, Society of Mining Enginurs (Ed:
P Somasundran), pp 433-447.
Li, G and Zhang, H, 1989. Effect of alkaline oxidants and cupric ions on
arsenopyrite flOlation, Nonfe"ous Met Chin Sac Met, 41:(4), pp 27-
O'Conner, C T, Bradshaw, D J and Upton, A E, 1990. The use of
dithiophosphates and dithiocarbonates for flotation of arsenopyrite,
Miner Eng, 3:(5), pp 447-459.
Extractive Metallurgy of Gold and Base Metals Kalgoorlie, 26 - 28 October 1992 219
Kalgoorlie, 2628 OCtober 1992 Extractive Metallurgy of Gold and Base Metals
Factors Affecting the Recovery and Grade of
Complex Lead-Zinc Ores by Flotation
The size-by-size batch flotation behaviour and flotation rate constant of
galena and sphalerite from three lead-zinc ores were determined. The
experimental data were fitted to the Klimpel model of first order flotation
It was found that the flotation behaviour of the coarse (+63 J.IlI1) and
fine (-10 lUD) size ranges of galena was significantly poorer than that of
the intermediate (-45+10 J.IlI1) size range, and the fine size range of
sphalerite was poorer than the other size ranges. The flotation rate
constant was found to be a maximum at some intermediate particle size.
The contact angle of galena and sphalerite was measured under the
same reagent condition as the flotation tests by using a simple method of
bubble-particle attachmenL It was found that the over all flotation
behaviour varied according to the trends in the contact angle.
Bubble size was measured for the different types of frother used in the
flotation tests. It was found that stronger frothers produced the smallest
bubbles and gave high recovery and high flotation rates, while weaker
frothers produced larger bubbles and gave higher grades.
It was also found that the more complex mi;neral assemblages resulted in
poorer flotation behaviour than the ore containing relatively simple
mineral assemblages.
Differentiation between true flotation and the entrainment of mineral
pa:ticles during the flotation process was determined. The results show
that fine galena was entrained in the froths at short flotation times, and the
true flotation rate constants were higher than the overall flotation rate
constants for all size fractions of sphalerite and for all fractions greater
than 10 J.IlI1 for galena.
Production of lead and zinc concentrates from complex lead-zinc
ores by using froth flotation is an important part of the production
of lead and zinc metals. In the past, ores were high grade, there
was only one metal of interest, and it was relatively simple to
extract. However, this is no longer the case. Nearly every
mineralisation has some problem either in the mining or the
extraction of the metal.
Most of the complex lea4-zinc ores contain lead-zinc minerals
in finely disseminated form. Although flotation is by far the most
important unit operation of mineral concentration, the recovery
achieved by using flotation for these fine ores is often poor. This
is because of the relationship among the various physical and
chemical properties of fine particles, and their behaviour in
Surface and electrochemical propertie& of fine particles tend to
be different from coarse particles of the SaI1)e material. Due to
the small mass and momentum of fine particles, they are carried
into the froth by entrainment, which is different from the
mechanism of particle-bubble attachment in flotation. If gangue
minerals are included in such entrained particles, the result is a
reduction in the grade of the concentrate.
Finer mineral particles have higher specific surface energies
and this may influence flotation in a number of ways. It may
introduce undesirable impurities into solution, affecting
1. Metallurgical Engineer, No 1 Mining Company, Kanbe Road,
Yankin PO Rangoon, Myanmar.
2. Senior Lecturer and Acting Head, Department of Minerals
Engineering and Extractive Metallurgy, WA School of Mines,
PO Box 597, Kalgoorlie, WA 6430.
collector/mineral interactions. Rapid oxidation may also render
some minerals non-floatable under the conditions used for their
flotation. The high surface energy of fine particles also increases
the tendency of collectors to adsorb non-specifically. Fine
particles have low collision probabilities because of their small
mass, which results in a low flotation rate and low recovery. Fine
particles at the liquid/vapour interface may also stabilise the
froth, causing concentrate handling problems.
Because of the extremely fine dissemination and interlocking
of minerals in complex sulphide ores, the treatment of these ores
represents one of the most complicated problems in base metal
flotation. The difficulties arise in producing high grade or high
recovery or both in flotation circuits. This problem comes from
incomplete liberation, poor flotation response at fine particle size
and/or interaction with some components of the complex ores.
However, due to losses of mineral and metal values in the fine
size range, considerable interest is growing in developing new
processes and improving old processes for the recovery of fine
particles. The objective of this work is to examine the flotation
behaviour of fine particles in terms of the flotation kinetics of
different size classes in lead and zinc concentration from bench
scale flotation test work. The variables studied were type of
frother, type of collector, collector concentration, and ore type.
The principle recovery mechanisms are presumed to be
genuine flotation (bubble attachment and levitation) and
entrainment (carry-over with water which enters the concentrate
via the froth). The other possible recovery mechanisms ,
including entrapment in the froth, carrier flotation and the
influence of slime coatings, froth modification by fines, or
possible size effects associated with the return of particles from
froth to pulp, are not considered.
Particle size is recognised as being a very important flotation
variable, and major problems in flotation arise in many instances
from the relatively poor response of coarse and very fine
particles. Recovery falls sharply above 100 lJ.Il1 but only
gradually below 10 lJ.Il1.
Not all minerals show a maximum recovery in exactly the
same size range, but there is no doubt that recovery is best for
particles of an intermediate size.
The presence of gangue minerals in the pulp also effect
different particle sizes differently. of glmgue durIng the
tlotation of lead and zinc at Broken Hill was found to increase
!with decreasing particle size, particularly below 10 lJ.Il1 (Kelsall et
al, 1974; Lynch and Thome, 1974). Granite gangue was
recovered much better as the pulp became finer when synthetic
mixtures of galena and granite were floated in laboratory batch
cells (Gaudin et al, 1931). It was attributed to 'mechanical
carry-over' of fme gangue. The entrained fine gangue causes a
decrease in both concentrate grade and the flotation rate of the
valuable mineral, with decreasing size.

Flotation tests
Lead-zinc ores from three deposits, Cadjebut (WA), Woodlawn
(NSW) and Bawdwin (Myanmar) were used for the flotation
Extractive Metallurgy of Gold and Base Metals Kalgoorlie. 26 - 28 October 1992 221


40 u
FIG 1 - Cumulative recovery of PbS at different times as a function of
different size fractions, (0.023 kg/t NaCN, 0.05 kg/t NaEX, H407 0.022
kg/t, pH 8.5 in PbS flotation of Cadjebut ore).
--0-- 20 - 40 60
--120 ---<>- 240 s _ Rate
20 30 40 50 60
Mean Size (microns)
Flotation tests were conducted in a modified Leeds cell.
Recrystallised sodium ethyl xanthate (NaEX), sodium amyl
xanthate (NaAX), liquid CMS 41 (secondary butyl
dithiophosphate) and CMS 42 (hexyl dithiophosphate) were used
individually or in combination with one another as collectors.
Liquid Nalflote series frothers (polyoxypropylene glycol ethers)
and Dowfroth frothers (polypropylene glycol ether) were used
individually as frothers. Sodium cyanide (NaCN), or a
combination of sodium cyanide and zinc sulphate (NaCN +
ZnS04), were used as depressants for sphalerite in the lead
sulphide flotation. Copper sulphate (CUS04) was used as
activator for sphalerite in the zinc sulphide flotation.
Flotation concentrates and tailing were wet screened to
produce six size fractions, -75+63 llm, -53+45 l!ffi, -45+38 llm,
-38+20 l!ffi, -20+10 l!ffi and -10 llm. These size fractions were
studied, in terms of flotation kinetics.
Flotation response is a function of the three factors; chemical,
equipment and operation factors. In this study, the equipment
factors and operation factors (per cent solid, pulp density,
temperature, air flow rate, ete ) were held constant.. Different
types of kinetic models (First-order model, Gamma, Kelsall,
Modified Kelsall and Klimpel model) were used to fit the
experimental data, but the Klimpel model gave the best fit of the
FIG 2 - Cumulative recovery of Zns at different times as a function of
different size fractions, (0.182 kg/t CUS04, 0.08 kg/t NaEX, H407 0.022
kg/t, pH 10.5 in Zns fltation of Cadjebut ore.
--0-- 240. _ Rate
20 30 40 50
Mean Size (mlcrona)
80 ::tl

10 n

Contact angle measurement
The contact angle of galena and sphalerite were estimated using
the bubble-particle attachment method (Hanning and Rutter,
1989). This involves determining the diameter of the largest
particle from a population of particles immersed in water that can
be raised against gravity by a captive air bubble.. The contact
angle can be calculated by using the equation of Scheludko et al
(1976), as given below:
=2 (3ywv( 1{2 sin 8(2. (1)
where d
maximum particle size captured by bubble
Ywv surface tension of liquid-vapour
p density difference between the solid and water
g gravitational acceleration, and
8 equilibrium contact angle
Bubble size measurement
The bubbles generated from the Leeds flotation cell were
captured by a capillary tube at the centre of the cell. The bubbles
are sucked up the capillary tube where the number of bubbles, as
well as their diameter, were calculated using a Randall bubble
size measurement unit. The unit detects the beginning and end of
a bubble in the capillary as it passes a photo diode. For a
capillary of known diameter, the volume of each bubble is
measured and hence the bubble diameter is calculated.
The experimental data were fitted to the Klimpel model of first
order flotation kinetics, as represented by equation 2.
R = R_ [l-(l/(kt(1-exp(-kt] (2)
where, R the cumulative recovery for time 1.
R_ recovery at infinite time
k flotation rate constant
The effect of particle size on recovery and flotation rate
Different particle sizes of galena and sphalerite exhibited
different flotation rates.
The results of the flotation tests using NaEx and Dowfroth 400
are shown in Figures 1 and 2. Galena recoveries for all tests had
similar characteristic curves. Sphalerile recoveries in other tests
also had characteristic curves similar to Figure 2.
Significant differences in recoveries and flotation rate constants
were obtained for the different size fractions. The highest
recovery and highest flotation rate were exhibited by the
intermediate size fraction (-38+20 l!ffi) for PbS. With sphalerite
flotation, there is no clear maximum in recovery and flotation rate
over the size range measured, with xanthate as collector. The
minimum flotation rate of sphalerite in Zns flotation is obtained
in the fine size range, and the rate slightly increased with size.
In the discussion of flotation kinetics here, the flotation rate
constants of sphalerite in PbS flotation and the flotation rate
constants of galena in Zns flotation have not been considered.
The content of each mineral in the other mineral's concentrate
may be due to incomplete liberation and the relation between the
flotation rate constants of one mineral in the other mineral's
concentrate and particle size was random. Hence their behaviour
is uncertain.
Kalgoorlie, 26 - 28 October 1992 Extractive Metallurgy of Gold and Base Metals
A 5IJm I
-----+-- CMS41
--<>-- CMS42
20 30 40 50
Mean Size (microns)
FIG 4 - Flotation rate constants for different size fractions of sphalerite
(ZnS flotation) for diffemt types of collector (CUS04 0.182 kg/t, collector
0.08 kg/t, H407 0.22 kg/t and pH 10.5).
FIG 3 - Flotation rate constants for different size fractions of galena (PbS
flotation) for different types of collector (NaCN 0.023 kg/t, collector 0.05
kg/t, H407 0.022 kg/t and pH 8.5).
~ 6
~ 4
~ 3
.S! 2
0 10
The flotation kinetics of galena and sphalerite in the fine size
region was universally poor. This behaviour appears to be in
agreement with the literature. As the particle size decreases. two
characteristics begin to dominate. The specific surface becomes
large. and the mass of the particle becomes very small. The
interfacial energy per unit surface area and surface and
electrochernical properties of different particle sizes are likely to
vary (rrahar and Warren (1976). Fuerstenau (1980). Klassen and
Mokrousov (1963). Leja (1982. For example. fme particles are
more reactive than coarse particles. due to an increase in high
energy sites with decreasing size and hence could oxidise easily.
Fines that are heavily oxidised are not floated with conventional
Fine particles have low inertia because of their small mass.
When approaching a bubble, they tend to follow the path of the
liquid and deviate around the bubble, whereas a large particle
would have continued along its rectilinear path due to its higher
momentum and collided with the bubble.
At the other end of the scale, although very fme particles are
less likely to establish contact by collision with bubbles. when
contact is established. they are more difficult to detach from
bubbles than coarser particles. If the large and small particles are
equally hydrophobic. large particles are more likely to collide
effectively than small ones. But large particles require a longer
time to establish contact than small particles; that is. the larger the
particle. the smaller the probability of adhesion. According to the
proposal of Klassen and Mokrousov (1963). the lower the
momentum of the fine particle. the harder it is to thin and rupture
the disjoining fllm between the particle and bubble. If the
collection mechanisms were considered alone. the larger particles
would be more easily captured and the recovery would be
expected to remain higher. The decline in recovery then. must be
due to a detachment mechanism which begins to dominate as the
particle size increases. The larger particles are easier to separate
from the bubbles than fine particles owing to their larger mass.
For any given bubble. the smaller the attached particle the greater
is the probability of the bubble-particle system being able to
withstand the disruptive forces present in the flotation cell.
Trahar (1981) has discussed possible explanations for coarse
particle flotation behaviour. The degree of hydrophobicity
required to promote a high level of floatability increases with an
increase in particle size. The flotation rate constant thus exhibits
an optimum value at an intermediate particle size.
Although the results of the flotation rate constants of sphalerite
do not agree with this expression up to 63 ~ it may be that the
optimum size is coarser than this size and the flotation rate
constants will decrease at some coarser size.
No. of Carbon Atoms In Collector molecule
FIG 5 - Flotation rate constants for diffemt size fractions of galena (pbS
flotation) as a function of collector chain length (CUS04 0.182 kg/t,
H407 0.022 kg/t and pH 8.5).
The effect of collector on the flotation rate
The influence of some collector types on Cadjebut ore, with
constant depressant, activator. frother dosage and pH is shown in
Figures 3 - 10. All of the collectors used in these tests were thiol
The flotation rate constants at different size fractions of galena
in PbS flotation and sphalerite in Zns flotation show a significant
difference. Results of the flotation rate constants, recovery, and
grade in galena flotation, suggest that the strong collectors NaAX
and CMS 42 give higher flotation rate constants with poorer
grades than the other collectors, and better selectivity is found
with the weaker collector CMS 41. NaAX and CMS 42 are
considered stronger collectors because their hydrocarbon chain
length is longer than for NaEX and CMS 41 (see Figures 5 and
6). Moreover. larger contact angles were aChieved by collectors
NaAX and CMS 42 than with NaEX and CMS 41.
The combined collector. CMS 41 and NaEX, gave a better
flotation rate constant and grade than either reagent on its own.
These results seem to be in agreement with the literature
(Dowling et al 1986). Strong collectors not only increased the
4 s
Extractive Metallurgy of Gold and Base Metals Kalgoorhe. 26 - 28 October 1992 223
1---0.025 kg/! --0-- 0.050 kg/! --0.10 kg/! I
70 60 50 40 30 20 10
0 . . . ~ 4 4 4
No. of carbon Atoms in Collector molecule
Mean Size (microns)
FIG 6 - Flotation rate constants for diffemt size fractions of sphalerite
(lnS flotation) as a function of collector chain length
(CUS04 0.182 kg/t. H407 0.022 kg/t and pH 10.5).
FIG 9 - Flotation rate constants for galena (PbS flotation) as a function of
NaEx concentration and particle size
(NaCN 0.023 kg/t. H407 0.22 kg/t and pH 8.5).
________ 69 ---0-- 49 --42
I--- 0.04 kg/! --0-- 0.08 kg/! --0.16 kg/! I
----Q-- 29 --15 ---0-- 5
~ m
70 60 50 30 40 20 10
.5 5
.! 4
C 3
J! 2
Mean Size (microns)
0.02 0.04 0.06 0.08
Collector Concentration (kg/t NaEx)
~ 7
C 15
~ 5
~ 4
~ 3
! 2
-------- 69 --0-- 49 ---- 42
FIG 7 - Flotation rate constnats for diIfemt size fractions of galena (PbS
flotation) as a function of NaEx concentration
(NaCN 0.023 kg/t, H407 0.022 kg/t and pH 8.5).
----Q-- 29 --15 ---0-- 5
~ m
0.05 0.1 0.15 0.2
Collector Concentration (kg/t NaEll)
FIG 8 - Flotation rate constants for diffemt size fractions of sphalerite
(ZnS flotation) as a function of NaEx concentration
(CUS04 0.182 kg/t, H407 0.022 kg/t and pH 10.5).
FIG 10 =Flotation rate constants for sphalerite (lnS flotation) as a
function of NaEx concentration and particle size
(CUS04 0.182 kg/t, H407 0.022 kg/t and pH 10.5).
flotation rate constant but also improved the recovery, but with
poor selectivity. These strong collectors also affected the gangue
minerals which were recovered in the froth phase, and hence the
grade decreased.
However, whether the behaviour of the collector is strong or
weak, fme size and coarse size particles had lower flotation rate
constants and recoveries compared to the intermediate size range.
Stronger collectors slightly improve the flotation rate constants in
the fine size range.
In sphalerite flotation, the use of strong collectors, NaAX and
CMS 42, or weak collectors NaEX and CMS 41, does not result
in any difference in grade. Strong collectors can improve the
flotation rate constant and recovery, especially with the CMS 42
collector. The combined collector CMS 41 and NaEX also gave
higher flotation rates when compare to using either CMS 41 or
NaEX alone. In this case, the strong collectors did not show the
effect on gangue particles. Collector activity on zinc is lower
than for lead, and collector consumption may be higher in
sphalerite flotation. Therefore, strong collectors affect zinc in
ZnS flotation, not only by improving the flotation rate constants,
but also recoveries, without affecting selectivity. Consequently,
224 Kalgoorlie, 26 - 28 October 1992 Extractive Metallurgy of Gold and Base Metals
I---+-- 69 ---0-- 29 - 5 IIm I
FIG 12 - Flotation rate constants for sphalerite (2'nS flotation) as a
fWlction of contact angle and particle size.
Contact Angle
26 25 24 23 22
.5: 5
! 4

C 3
J! 2
18 19 20 21
sphalerite particles required higher collector dosages in these
tests. Particles in the fine size range do not response well to
either strong or weak collectors with respect to flotation rate
constants and recovery.
The effect of an increased addition of NaEX collector, with
constant depressant, activator, frother dosage and pH, is shown in
Figures 7 - 10. Even though the recovery did not improve in the
coarse size fraction (+63 increasing the collector dosage has
a markedly positive effect on the flotation rate constants of the
larger size fractions. While the recOvery and flotation rate of
galena in PbS flotation and sphalerite in Zns flotation were
higher, the increase was most pronounced in the intermediate size
fractions. For sulphide minerals, high collector dosages are really
needed to float the coarser particles.
In galena flotation, selectivity was poor in all size fractions as a
result of increasing the collector dosage. This means that the
collector can adsorb to a greater extent, not only on galena
particles but also on the gangue minerals. In sphalerite flotation,
increasing the collector dosage improved the selectivity as well as
the flotation rate constants and recoveries. This means that
sphalerite particles still require more collector to float In fact,
the influence of collectOr dosage can dominate almost any
industrial circuit response.
The flotation rate constants were expected to show an increase
from low contact angles to high contact angles, for all size
fractions. Figures 11 and 12 indicate that at low contact angles,
the rate of flotation is relatively constant, only increasing for
contact angles exceeding 48 for galena.
The effect of an increased addition of collector, NaEX, on the
behaviour of galena and sphalerite flotation is given in Figure 13.
These results illustrate that the concentration of NaEX has litde
effect on the contact angles of galena and sphalerite. The greater
the collector concentration, the higher the contact angle will be,
up to a limit governed by the optimum surface coverage of the
sulphide surface by the collector. These results would suggest
that the optimum surface coverage has been achieved. It is
interesting to note the relatively small values of contact angles;
50 for galena and 22 for sphalerite were found compared with
the contact angles found in the literature (Wark and Cox, 1936).
Especially sphalerite did float very well in these tests with a low
contact angle. On the other hand, the difference in contact angles
between this investigation and the literature may be due to the
different methods of measurement for contact angle, or the galena
and sphalerite in either investigation may have included other
Contact angles
The contact angles of hand picked galena and sphalerite particles
were measured.
Contact angle measurement is a significant method of
investigation available for the study of the attraction between air
bubbles and minerals.
Figure 11 and 12 shows the comparison of flotation rate with
contact angles for different types and dosages of frothers and
collectors. The function of frothers is to permit the transfer of
minerals from pulp to froth and stabilise the collected minerals
already in the froth prior to froth removal. Significant differences
exist in the contact angles obtained for different types of frothers,
in galena and sphalerite flotation. Strong frothers H407 and
H405 gave higher contact angles for galena and sphalerite
compared to weak frothers,H402 and H41O.
Strong collectors show higher contact angles than weak
collectors. Longer hydrocarbon chain amyl xanthate gave the
highest contact angle, anq the weakest collector, secondary butyl
dithiophosphate, gave the lowest contact angles on galena and
sphalerite. These results agree with the literature.
I---+-- 69 ---0-- 29 ---+-- 5 IIm I
I---Galena ----0- Sphalerite I

.. 70
Cl 40
0( 30
8 10
0+--.....--4_--+--+_......_ ......_ .....

Contact Angle
30 35 40 45 50 55
o 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.18
Collector Concentration (kglt, NaEx)
FIO 11 Flotation rate constants for galena (pbS flotation) as a function
of contact angle and particle size.
FIG 13 - Contact angle of galena and sphalerite (Cadjebut ore) for
different concentrations of NaEx (0.22 kg/t H407 and pH 8.5).
Extractive Metallurgy of Gold and Base Metals Kalgoorlie, 26 - 28 October 1992 225
70 60 50 40 30
Mean Size. (microns)
20 10 o
Flotation behaviour is related to the size and surface
composition of the particles. Valuable minerals are present in the
flotation feed as liberated particles and as locked or composite
particles, with a wide variety of compositions, and these particles
consequently have a wide range of flotation behaviour.
The earth's minerals are not distributed evenly and no two
mineral deposits are the same. The Cadjebut deposit is a
stratabound Mississippi Valley type deposit. Woodlawn and
Bawdwin deposits were formed under similar origins of volcanic
environments although the mineralogical constitution of the ores
are different. These sulphide deposits are characteristically
Due to their different mineralogy, the liberation characteristics
of these ores are also different. Table 1 shows the X-ray
diffraction results for the mineral composition of these ores.
According to the simple mineral assemblages characteristic of
Mississippi Valley type deposits, liberated galena and sphalerite
could be found in the Cadjebut sample. The galena in the
volcanogenic base metal ores occurs as intergrowths with
sphalerite, chalcopyrite and pyrite and as inclusions in pyrite
grains. The intergrowth galena is well bonded to the sphalerite,
chalcopyrite and pyrite. The minute galena grains in pyrite and
the galena grains with chalcopyrite/sphalerite inclusions were
difficult to liberate during primary grinding or regrinding because
they are small and are enclosed within each other. The sphalerite
layer in the volcanogenic base metal ores are coarse grained, but
generally contain inclusions of pyrite, galena, chalcopyrite and
other minor minerals. The massive sphalerite occurs as
interstitial filling between the pyrite grains. For most
volcanogenic ores, much of the massive sphalerite can be
liberated but some pyrite, chalcopyrite or galena remains attached
to grains that were inclusions in the massive sulphide. The
attachment of pyrite and chalcopyrite can changed the flotation
behaviour of the sphalerite.
The mineralogy agreed with the better flotation behaviour of
Cadjebut ore. The difference in floatability of the three ore types
could not be related to any difference in contact angle since the
method used for contact angle measurement required coarse
single particles of each mineral. With the fine grain size of
Woodlawn and Bawdwin ores, this was not possible.
FIG 15 - Flotation rate constants for diffemt size fractions of sphalerite
(ZnS flotation) for diffemt ore types (0.182 kg/t CUS04. 0.08 kg/t NaEX,
0.22 kg/t H407 and pH 10.5).
I --Cadjebut --0-- Woodlawn ---+- Bawdwin I
~ 3
;: 2
70 60 50 40 30 20 10
o .I-_-+- ~ . . _ _ _ _ _ _ t
Effect of ore mineralogy on flotation rate
Flotation of three different run-of-mine ores were studied with
constant collector, depressant, activator, frother dosage and pH.
The results for flotation rates are summarised in Figures 14 and
15. There were significant differences in recovery and flotation
rate constants at different size fractions of galena and sphalerite.
Better recovery for all size fractions of galena and sphalerite was
achieved with Cadjebut ore than with the other ores. Maximum
flotation rate constants for Cadjebut ore occurred at intermediate
size fractions for galena, and coarse size fractions for sphalerite.
Although the flotation rate constants of galena and sphalerite, at
all size fractions, for Bawdwin ore are not significantly different,
the highest flotation rate was obtained in the intermediate size
range, 30 - 50 Il-m, for bot!'! galena and sphalerite. The flotation
rate constants for all size ranges are not significantly different for
the Woodlawn ore.
minerals, or the surface of the galena and sphalerite may not have
been clean. Every effort was made to minimise these factors in
this investigation.
Under laboratory conditions, the attainment of a maximum
contact angle, even on a clean homogeneous surface, required a
certain collector concentration. As well as this, it took time for
adsorption of xanthate to take place, and in certain systems, the
maximum contact angle may not have been achieved for an hour
or more. If the particle remained for a long time in the pulp, the
greater would be the contact angle and presumably the higher
would be the rate constant.
For freshly liberated minerals, surface properties differ from
place to place on the surface. The exposed surface may be those
of different crystallographic planes, or impurities may be present.
The adsorption of xanthate on a freshly liberated sulphide mineral
surface would hence vary from site to site. Therefore the contact
angle may also vary from site to site, and this can vary the overall
apparent contact angle. Thus for the same overall contact angle,
the flotation behaviour of different sized particles may be
dissimilar.. Therefore in the relationship between contact angle
and flotation rate constant at different size fractions, it is difficult
to show a reasonable connection between the two factors.
Although the actual value of contact angle is not generally
applicable -to flotation tests, the trends demonstrated will almost
certainly be followed in the flotation tests. In general, the higher
the contact angle for galena and sphalerite, the higher the
flotation rate.
FIG 14 - Rotation rate constants for different size fractions of galena
(pbS flotation) for diffemt ore types (0.023 kg/t NaCN, 0.22 kg/t H407,
0.05 kg/t NaEX and pH 8.5).
__Cadjebut -0-- Woodlawn ---- Bamwin
Mean Size (microns)
226 Kalgoorlie. 26 - 28 October 1992 Extractive Metallurgy of Gold and Base Metals
X-ray diffraction analysis ofthe three ore types; flotation feed sample.
ORE Calc Dol Chlor Musc Qtz Gal Pyr Sph Talc OCC OF OXI OXX
TYPE (%) (%) (%) (%) (%) (%). . (%) (%) (%) (%) (%) (%) (%)
Cadjebut 2 29
. .
5 9 3' 51
. .
. .
- - -
3' 57 18
. .
5 2
8 20 19 43 2 1
. .
Cadjebut Bawd Bawdwin
calcite Pyr pyrite
dol-ankerite Sph sphalerite
chlorite OCC Calcite and/or chalcopyrite
muscovite OF Microcline/rutile/sphene
quartz OXl Cubanite CuFe2S3
galena OXX Unidentified phase
unsure of mineral identification and/or abundance
1-Entrained --0-- Total Zn I
The effect of entrainment on the flotation rate
Determination of true flotation by the method proposed by Ross
(1990) involves only one batch flotation test, in which the
recoveries of solids and water. and the consequent changes in the
composition of the pulp, and the characteristics of the froth with
time, are used in the estimation of the contribution of entrained
particles that belong to species i.
For example, the results for one test are shown in Figures 16 to
It was found that there were entrained particles at all conditions
of flotation, and the degree of particle entrainment varied. There
were significant differences in entrainment between all of these
tests, especially for different types of frothers. Galena
entrainment in PbS flotation, at all size fractions was higher for a
strong frother than for a weak frother. Sphalerite entrainment in
Zns flotation, at most of the size fractions, was also higher for a
strong frother. Thus frother strength has a great affect upon
8 c
0 10 20 30 40 50 60 70
1_Enlrelned --0-- Total Pb I
S 2.5

0 10 20 30 40 50 SO 70
Mean Size (microns)
FIG 16 - Mass of lead recovered by entrainment in the flotation of galena.
Mean Size (microns)
FIG 17 - Mass of zinc recovered by entrainment
in the flotation of sphalerite.
entrainment. Strong froths rely upon slow rates of drainage of
liquid from the bubble walls or borders within the froth. As a
result of this slow drainage, strong frothers that increase recovery
will also contain more water. Therefore, frothers that produce
strong froths also tends to have higher entrainment and less
selectivity. Although there were significant differences in
entrainment of particles between the different types of frothers,
each showed a similar behaviour in that more entrainment of
particles were found in the ftne size ranges, for both lead and
The particle size distribution of the ore presented to the
flotation test will also, in part, have an influence upon the
entrainment during the flotation process. If the ore contains a
higher percentage of 10 slimes, the froth will not drain as
efficiently as it would in the absence of slimes. The froth will
tend to be more stable, and carry more water and entrained
Extractive Metallurgy of Gold and Base Metals Kalgoorlie, 2628 October 1992
Extractive Metallurgy of Gold and Base Metals
------- 69
---0-- 49 -42
---<>-- 29 --15 ----<>-- 5

C 25

,g 20

0 50 100 150 200 250
Time (seconds)
FIG 20 - Per cent of lead entrained in the galena concentrate with time.
FIG 21 - Per cent of zinc entrained in the sphalerite ooncentrate with time.
true flotation. The exception to this rule is the 10\ in
galena flotation.
Particle size has a pronounced effect on the flotation galena
and sphalerite as determined by rate of flotation, recovery and
selectivity. In terms of recoveries, the overall recovery of galena
and sphalerite was limited by the relatiyely poor response of
coarse and fine size fractions, and especially for the fine sIZe
fraction when compared with the intermediate size fraction. It
was experimentally shown that the optimum size for the best
recovery of galena is the intermediate size range from 20 to 45
Jlffi, and for sphalerite the best recovery is in the +63 Jlffi or +45
size range.
Flotation recovery and rate are often critically dependent on
frother, collector, and collector concentration.
The laboratory flotation studies reported here generally
confirmed that:
Kalgoorlie, 2626 October 1992
20 30 40 50
Mean Size (microns)
FIG 19 - Per cent of zinc in the sphalerite concentrate due to entrainment.

FIG 1g - Per cent of lead in the galena concentrate due to entrairunent.

particles. If the ore contains ":Iainly coarse a
very small proportions of slimes, the froth will dram qUIte
According to Klimpel's model and the experimental results
shown in Figures 20 and 21, it can be postulated that most of the
-10 Jlffi galena was entrained in the froths at very short times,
that the amounts of other size fractions of galena and sphalente,
entrained in the froth, gradually increases with time.
The grade of the flotation concentrate is affected by the amount
of entrained particles, which include not only the valuable
minerals but also the gangue minerals that are contained within
the liquid associated with the froth. The entrainment of particles
carmot be eliminated, but the effect on concentrate grade and
flotation kinetics can be significantly reduced with attention to
the entrainment mechanism.
Determination of the flotation kinetics for true flotation were
carried out for a number of test conditions. The results are
plotted against the apparent flotation rate (total flotation) in
Figures 22 and 23. The majority of the data points lie below
median line indicating that the flotation rate constants for all Stze
fractions, in both galena and sphalerite flotation, are higher for

------- 69
---0-- 49 -+-42
---0-- 40 -00
--240_ --<>--- 29 --.-15
50 50
45 45

C 25

15 N

5 5
0 0
0 10 20 30 40 50 70
0 50 100 150 200 250
Mean Size (microns)
Time (seconds)
... I
C 10


a: .. I
, .:

.. I
~ .

~ .
0 2 4 6 8 10 12 14 16 18 20
True Flotation Rate (min-1)
FIG 22 - True flotation rate of lead particles versus total flotation rate of
lead in the galena concentrate.
Strong frothers gave the highest contact angles and weak
frothers gave low contact angles. The flotation rate
constant increased with increasing contact angle; and
strong collectors gave high contact angles on both galena
and sphalerite, and the flotation rate increased with contact
angle. .
The flotation behaviour of the simple mineral assemblage,
characteristic of the Cadjebut deposit, is superior to the complex
mineral assemblage of the Woodlawn and Bawdwin deposits.
Particle size and particle compositions are both important for the
behaviour of flotation.
There are significant differences in entrainment for the
different types of frothers. Stronger frothers have a greater effect
upon entrainment. Fine size ranges of both galena and sphalerite
are more entrained than other size ranges.
Finally, even though the factors affecting grade and recovery,
in the flotation of complex lead - zinc ore, can improve the
flotation behaviour of the minerals, fine size and coarse size
particles have lower flotation rate constants and recoveries
compared to the intermediate size particles.
FIG 23 - True flotation rate of zinc particles versus total flotation rate of
zinc in the sphalerite concentrate.
Dowling, E C, Klimpel, R R and Aplan, F F, 1986. Use of kinetic models
to evaluate new flotation collecting reagents, in Schuhmarm
Symposium, pp. 129-144.
Fuerstenau, D W, 1980. Fine particle flotation in Fine Particlt
Proctssing, Chap 35, pp 669-705.
Gaudin, A M, Groh, J 0 and Henderson, H B, 1931. Effect of particle
size on flotation in Amer Insl Min Mttall Engrs, Ttch Publ No 44.
Hanning, R N and Rutter, PR, 1989. A simple method of determining
contact angle on particles and their relevence to flotation in
Inttrnalional Journal ofMintral Proctssing, 27, pp 133-146.
Kelsall, D F, Stewart, P S B and Trahar, W J, 1974. Diagnostic
Metallurgy, a systematic method of plant optimisation, Symposium on
Optimization and Control of Mintral Proctssing Plants, Brisbane,
AMIRA, pp 53-65.
Klassen, VI and Mokrousov, V A, 1963. An introduction to tht thtory of
flotation, pp 493 (Butterworths: London).
Klimpel, R R, 1989. The industrial sulfide mineral flotation system, in
Advanct in Coal and Mintral Proctssing Using Flotation, Chap 30,
pp 273-285.
Leja, J, 1982. Surfact chtmistry offrolhflotation. New York and London.
Lynch, A J and Thome, G C, 1974. Research on flotation simulation and
control at the Julius Kruttschnitt Mineral Research Centre, in
Symposium on Optimisalion and Control of Mintral Proctssing
Plants, Brisbane, pp 67-80 (Min Ind Res Assoc: Melbourne).
Ross, V E, 1990. Flotation and entrainment of particles during batch
flotation tests, in MintraIs Enginttring 3:(314}, pp 245-256.
Scheludko, A, Toshev, B V and Bojadjiev, D T, 1976. Attachment of
particles to liquid surface (capillary theory of flotation), in J Chem
Soc Farad Trans 1, vol72, pp 2815-2828.
Trahar, W J and Warren, L J, 1976. The flotability of very fme particles,
in Inttrnational Journal ofMintral Proctssing 3, pp 101- 131.
Trahar, W J, 1981. A rational interpretation of the role of particle size in
flotation, in Inttrnalional Journal of Mintral Proctssing, 8, pp
Wark, 1 Wand Cox, A B, 1936. An experimental study of influence of
sodium-sulphide, Alkalies and cOPPer sulphate on effect of xanthate
at mineral surface, in Principlt of mintral flotation.
The authors wish to thank the Australian International
Development Assistance Bureau for sponsoring this work and the
companies operating the Cadjebut, Woodlawn and Bawdwin
mines for the supply of samples for this project.
12 10

6 4
True Flotation Rabl (min-1)
~ i
.... I
.... I
.....~ .-
........ .".
...// ../ ..../ .r;
strong collectors gave higher flotation rate constants for
galena and sphalerite in all size fractions, compared with
weaker collectors;
the weaker collectors not only gave low flotation rate
constants but also gave poor recoveries for galena and
sphalerite in all size fractions;
the combined collectors gave better flotation rates for
galena and sphalerite than either reagent on its own;
the weaker collector gave the best grade for galena in PbS
flotation, at all size fractions;
all of the collectors in these tests gave approximately the
same grades in ZnS flotation;
collector concentration has a significant effect on the
flotation rates and recoveries of galena and sphalerite, at all
size fractions. The increase was most pronounced in the
intermediate size fraction for both galena and sphalerite;
significant differences in the contact angles of galena and
sphalerite were obtained for different types of frothers.
4 0
Extractive Metallurgy of Gold and Base Metals Kalgoorlie. 26 - 28 October 1992
230 Kalgoorlie. 26 28 October 1992 Extractive Metallurgy of Gold and Base Metals