Int. J. Miner. Process.

90 (2009) 35–44

Contents lists available at ScienceDirect

Int. J. Miner. Process.

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j m i n p r o

The effect of hydrodynamic conditions on true flotation and entrainment in flotation

of a complex sulphide ore
Emin Cafer Cilek ⁎
Department of Mining Engineering, Faculty of Engineering and Architecture, S. Demirel University, TR 32260, Isparta, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: The separation efficiency and selectivity of flotation are directly proportional to recoveries of the mineral
Received 26 November 2007 species in the feed due to true flotation and entrainment. In this study, effects of the hydrodynamic
Received in revised form 26 September 2008 conditions on true flotation and entrainment were investigated by using a fractional factorial experimental
Accepted 19 October 2008
design. A method previously described in the literature was applied to determine the contributions of true
Available online 5 November 2008
flotation and entrainment in flotation of a complex sulphide ore. In order to apply the method, the kinetic
Keywords:
flotation tests were conducted under various hydrodynamic conditions defined by some physical variables.
Froth flotation Some of these tests were conducted in the presence and absence of a collector to evaluate the self-induced
Flotation kinetics floatability. The selectivity index of the mineral species for entrainment was seen to be suitable evaluation of
Sulphide ores the non-selectivity and efficiency of the entrainment. Furthermore, the results of the size-by-size analysis of
the froth products indicated that the presence of the self-induced hydrophobic particles in the feed is as
important as the presence of very fine particles for accurate estimation of true flotation and entrainment in
flotation of a complex sulphide ore. In addition, the estimated results for entrainment in flotation of the
complex sulphide ore can be misleading. Therefore, a new approach would be necessary to determine the
contributions of true flotation and entrainment in flotation of a complex sulphide ore.
© 2008 Elsevier B.V. All rights reserved.

1. Introduction Several methods were proposed to determine entrainment and

true flotation in the literature (Trahar and Warren, 1976; Warren,
Flotation, as a selective separation process, is applied for separating 1985; Ross and Van Deventer, 1988; Subrahmanyam and Forssberg,
different minerals from each other, and utilizes the difference in 1988a; Smith and Warren, 1989; Ross, 1990; George et al., 2004), and
physico-chemical surface properties of various minerals within an ore. the results of some experimental studies were also published
Hydrophobic mineral particles (naturally or rendered hydrophobic by comparing these methods by Ross (1991) and George et al. (2004).
using a suitable surfactant) generally transfer to the froth product Ross (1991) used an artificial mixture of some minerals in his
primarily by true flotation, but they can also be transferred by experimental study to compare the results given by each of the
entrainment in the inter-bubble water, along with hydrophilic mineral methods proposed in the literature. On the other hand, George et al.
particles. The true flotation and entrainment mechanisms may (2004) studied the assessment of true flotation and entrainment in the
operate simultaneously for the same mineral depending on the size flotation of submicron particles by using silica particles ranging in
fractions present in the pulp. The fast and slow floating hydrophobic diameter from 40 to 150 nm and alumina particles with a d50 of
minerals generally reach the froth product by true flotation, but the 150 nm. They have compared the methods described in the literature
slow floating hydrophobic minerals can also be recovered by and have found reveal highly consistent results. It must be noted that
entrainment like hydrophilic minerals. Lynch et al. (1981) stated the artificial materials (a mixture of minerals or a single mineral) were
that several origins of the slow floating valuable minerals could be used in the studies mentioned above. However, Subrahmanyam and
recognized: presence of the valuable minerals in the composite Forssberg (1988b) used a copper ore having some locked (or
particles, unsuccessful preparation of particles surfaces and the fine composite) particles and natural hydrophobic particles to study the
size fraction in the feed. The recovery of hydrophobic particles by true effects of different frothers on the entrainment by using the method of
flotation does not steadily rise with decrease in particles size, but has a Trahar and Warren (1976). In this method (Trahar and Warren, 1976),
maximum at an intermediate size. The recovery by entrainment rises it assumed that all the particles captured in the froth with no collector
steadily as the particle size decreases (Lynch et al., 1981; Smith and added, will be by entrainment only and therefore, results of the two
Warren, 1989; Nguyen and Schulze, 2004; Nguyen et al., 2006). different flotation tests are compared. The first test involves a
suspension with both a collector and a frother added, while the
⁎ Tel.: +90 2462370428; fax: +90 2462370859. second involves just a frother. The difference between the solids
E-mail address: [email protected]. recovered versus the water recovered, for the collector and non-

0301-7516/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.minpro.2008.10.002
36 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44

stator system) in a mechanical cell could appreciably affect the gangue

recovery and the water recovery. The impellers combined with a
stator consisting of radial and relatively high blades give rise to a less
turbulent pulp flow in the upper cell volume and, therefore, to a more
stable froth phase then other impeller–stator systems. Such a system
provides the best flotation results because it reduces the entrainment
of gangue particles (Nguyen and Schulze, 2004).
Airflow rate is related to the bubble Reynolds number, and
characterizes how fast the air bubbles exit the cell. If airflow rate (or
air capacity number) is too high, the froth surface can be unstable. Air
capacity number is proportional to the power distribution between
the liquid and air/liquid phases, and characterizes the intensity of
aeration in the system (Nelson and Lelinski, 2000; Nguyen and
Schulze, 2004).
The entrainment of the fine particles can directly be related to the
water recovery. The mechanisms proposed to explain the water
recovery into the froth include the hydrodynamic boundary layer, the
wake and swarm effect of rising bubbles. The formation of a wake is
Fig. 1. Determination of true flotation and entrainment by method of Trahar and Warren
limited by the hydrodynamic conditions. The wake volume is a
(1976).
function of velocity of the bubble rise, the bubble size and the liquid
viscosity. The wake volume can be described as a function of the
bubble Reynolds number. The velocity of the bubble and the bubble
collector tests, gives an estimate of the mass recovered by true size are dependent to amount and type of frother used in flotation, as
flotation. A schematic representation of this method is shown in Fig. 1 well as airflow rate and impeller speed. Some of these operational
(Subrahmanyam and Forssberg, 1988c; Ross, 1991; George et al., factors can also be used to define the cell hydrodynamics. In addition,
2004). the froth removal rate, froth residence time and the stability of the
Subrahmanyam and Forssberg (1988b) stated that the contribution pulp–froth interface directly affect to level of entrainment and the
of entrainment to overall recovery can be individually determined in water recovery. It is known that the probability of bubble particle
certain cases but it is not possible to estimate the recovery by true collision is determined by the hydrodynamics of the system, which is
flotation without entrainment. The authors also pointed out the strongly affected by the particle size, bubble size and the turbulence of
difficulty of determining the contribution of entrainment for naturally the system. In addition to these, turbulence is the main mechanism
hydrophobic minerals. Furthermore, work by Ross (1991) indicated causing the particle detachment for mechanical flotation cells (Nelson
that the method of Trahar and Warren (1976) would be applicable and Lelinski, 2000; Yoon, 2000; Rodrigues et al., 2001; Pan et al., 2003;
when the frother has no collecting properties and when the feed Nguyen and Schulze, 2004). Therefore, it is apparent that the hydro-
material is coarse. For these reasons, it is necessary to adopt a suitable dynamic conditions of the system play a major role in entrainment and
method for determination of the contributions of true flotation and true flotation. The main objective of the present study is to determine the
entrainment in flotation of a complex sulphide ore. In addition, it must contributions of entrainment and true flotation in flotation of a complex
be mentioned that the possibility of naturally induced floatability is sulphide ore and to reveal the importance of the hydrodynamic condi-
widely recognized in the flotation of sulphide ores. For these reasons, tions in the flotation selectivity by using a statistical experimental design.
the method of Ross and Van Deventer (1988) was seen to be suitable
for estimating true flotation and entrainment. In this method, it is 2. Experimental
assumed that the concentration of entrained particles in the froth and
concentrate is identical to the concentration of particles in the pulp. 2.1. Material
Hence, the recovery by entrainment can be estimated by multiplying
the mass of water recovered in the froth by the concentration of the A composite sulphide ore sample was received from the fine ore
pulp, at that time. This method can estimate the rate of flotation with a bin of the Küre Copper concentrator (Turkey). The sample contains
single flotation test. The detailed descriptions of the methods average 2.1% Cu and 24.5% Fe. The main copper mineral is chalcopyrite
mentioned above can be found in the studies cited (Trahar and
Warren, 1976; Warren, 1985; Ross and Van Deventer, 1988; Ross, 1990,
1991; George et al., 2004).
Entrainment is the micro-process by which slurry enters the
flotation froth, moves upward with the froth, and finally leaves the
flotation cell with the concentrate. In agitation flotation cells, the
entrainment is determined considerably by the turbulent flow and,
consequently, by the state of suspension in the pulp phase, in
particular just below the pulp–froth interface. The stability of the
interface, which can be considered as a wall preventing the froth from
contamination by entrained particles, also depends on the turbulent
flow. In mechanical flotation cells, turbulence should be held at an
optimum level for achieving optimum true flotation, as well as
minimum entrainment. In relation to the specific airflow rate,
residence time and height of froth phase, dimensions of Plateau
channels, the water recovery and many other important variables, the
influence of turbulence at the pulp–froth interface has not been yet
understood. In addition to these, Schubert (1999) showed that the Fig. 2. Particle size and mineral species percent distributions of the ground sample used
design of the turbulence generating system (the impeller–baffles– in the experiments.
 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44 37

These dimensionless hydrodynamic parameters were calculated by

following equations:
 
Re = ρp ND2 μ −1
p ð1Þ

CA = Q a N −1 D−3 ð2Þ
3
where ρp is the pulp density (g/cm ) and D is impeller diameter (cm)
(Nelson and Lelinski, 2000; Rodrigues et al., 2001). The following
equation was used to calculate the pulp viscosity (μp):

μ p = μ w ð1 − ΦÞ−2:5 ð3Þ

where μw is the water viscosity (cP) and Φ is fraction of the pulp

volume occupied by solids (Rodrigues et al., 2001).
A full factorial study of these three variables at three levels would
have required 27 experiments. Instead, a face-centred cubic experi-
mental design, involving only 15 tests, was used to reduce the number
Fig. 3. The levels of factors used in the designed experiments.
of experiments (Montgomery, 1991; Myers and Montgomery, 1995;
Sheridan et al., 2002). Levels of the factors employed in the experi-
ments are depicted in Fig. 3. Several preliminary experiments were
conducted before the designed experimental program, and repeat-
associated with the siliceous minerals and pyrite. The sample was ability of experiments was measured. The standard deviation for these
crushed to minus 1.7 mm and it was representatively divided to use in tests was 4.89%.
the grinding tests by standard sampling procedure. Prior to flotation, It is assumed that the recoveries of valuable mineral (Cp), gangue
each sample was ground to minus 0.075 mm to obtain a sufficient (G = P + NG) and water (W) can be described by first-order rate equa-
degree of liberation (N90%). The cumulative size, chalcopyrite (Cp), tions. The rate constant for recovery of valuable mineral by entrain-
pyrite (P) and non-sulphide gangue (NG) percent distributions of the ment can be related to the rate constant for recovery of water (Lynch
ground sample are given in Fig. 2. et al., 1981; Smith and Warren, 1989; Polat and chander, 2000; Nguyen
and Schulze, 2004; Angadi and Suresh, 2005). Therefore, true flotation
2.2. Method and entrainment for each mineral estimated by method of Ross and
Van Deventer (1988) were fitted to following the first-order rate
Flotation tests were conducted in Denver laboratory flotation equation:
machine with a two-litre cell. The flotation reagents used in this study
were supplied by Cytec Industries, USA. The pulp was conditioned for Ri = RIi ½1− expð−ki t Þ ð4Þ
10 min of pH (11.5–12) and the collector, Aerophine 3418 A, was added
where Ri is the recovery of a mineral i at time t (min). RIi and ki are the
at a dosage of 50 g/t and conditioned for a further 2 min. After 1 min of
ultimate recovery (%) and the rate constant (min− 1), respectively.
frother (Oreprep F-507) conditioning, the air was introduced. Since
the results of the kinetic flotation tests are required to estimate
3. Results and discussion
entrainment and true flotation for the method, the froth product was
collected at time intervals of 0.5, 1, 2 and 4 min for all flotation tests. As
3.1. Effect of the hydrodynamic conditions on overall flotation
recommended by Ross (1990), the start of the test (t = 0) was taken as
the instant the froth overflowed into the first collecting tray. The froth
Table 1 shows the factor levels and overall recoveries of the pulp
depth (25 ± 2 mm) was kept constant by the addition of make-up
constituents obtained from the tests conducted under various
water to the cell by using a variable speed peristaltic pump. The froth
hydrodynamic conditions or different levels of the factors. As
was scrapped every 5 s. The volume of make-up water added to the cell
mentioned previously, the air capacity number and the Reynolds
was measured and each froth product (water and solids) was weighed,
filtered, dried and weighed again to determine water recovered in the
incremental concentrates. The water recovery is determined from the
ratio of weight of water recovered in an incremental concentrate to the Table 1
Results of the flotation tests conducted under various hydrodynamic conditions
weight of water initially present in the feed pulp (Subrahmanyam and
Forssberg, 1988b; Angadi and Suresh, 2005; Zheng et al., 2006). Even if Hydrodynamic parameters Factors Overall recovery (%)
a high pH was used to depress pyrite that is major sulphide gangue Re CA μp (cP) N (rpm) Q a (L/min) X (%) RCp RP RG RW
mineral of the ore, depressing of pyrite cannot be achieved completely. 77,026.6 0.875 0.01197 1000 5 15 37.47 8.35 5.6 7.62
Therefore, pyrite content of the products was also determined to 77,026.6 0.875 0.01307 1000 5 25 47.41 15.5 13.92 16.25
evaluate the behaviour of pyrite particles separately. The gangue 77,173.9 1.137 0.01248 1000 6.5 20 52.74 13.23 14.45 18.85
77,204.7 1.399 0.01197 1000 8 15 45.91 14.24 10.58 15.6
content of each product was calculated as 100-percentage of 77,173.9 1.399 0.01307 1000 8 25 67.24 30.14 28.82 38.18
chalcopyrite. 96,283.2 0.91 0.01197 1250 6.5 15 64.04 26.13 21.33 32.45
Since the recovery-time set of data for the solids and water are 96,467.4 0.91 0.01307 1250 6.5 25 76.46 34.93 30.65 37.77
required to estimate entrainment and true flotation according to the 96,505.8 0.7 0.01248 1250 5 20 63.03 16.02 13.42 21.06
96,505.8 0.91 0.01248 1250 6.5 20 66.61 33.43 26.84 38.28
method of Ross and Van Deventer (1988), the kinetic flotation test
96,505.8 1.12 0.01248 1250 8 20 70.74 39.99 36.91 40.68
procedure was used in the flotation tests. In addition, the airflow rate 115,539.9 0.583 0.01197 1500 5 15 60.3 17.7 14.95 20.22
(Q a, L/min), the impeller speed (N, 1/s) and percentage of solids by 115,760.9 0.583 0.01307 1500 5 25 75.88 22.2 20.89 29.74
weight in the pulp (x, %) were used as factors. Reynolds number (Re) 115,539.9 0.758 0.01248 1500 6.5 20 77.38 28.56 26.3 52.03
and air capacity number or airflow number (CA) were used to 115,760.9 0.933 0.01307 1500 8 25 81.94 39.14 45.49 56.79
115,807 0.933 0.01197 1500 8 15 62.52 37.14 34.39 46.59
characterize the hydrodynamic conditions of the flotation tests.
38 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44

Fig. 4. Effects of the hydrodynamic conditions on cumulative recoveries of chalcopyrite (a), gangue (b), water (d) and cumulative grade of concentrate (c).

number used to characterize the hydrodynamic conditions of the tests and/or the air capacity number. However, the pronounced effects of the
and, their values were determined by using the three factors (N, Q a impeller speed and the pulp density on recovery of each pulp consti-
and x). According to level of the Reynolds number, the hydrodynamic tuent were obtained from the tests conducted under low and medium
conditions of the experiments can be classified in three groups such as levels of the airflow rate, i.e., under low aeration intensities. The effects
quiescent flow conditions (Re = 77026–77173), semi-turbulent flow of the Reynolds number and the air capacity number on recoveries of the
conditions (Re = 96283–96505) and turbulent flow conditions mineral species and cumulative grade of concentrate as a function of the
(Re = 115539–115807). However, the air capacity number exhibited a flotation time are illustrated in Fig. 4(a–d). Similar conclusions can also
pronounced effect on the overall recoveries of the pulp constituents in be drawn from these results, which are consistent with grade of
same value of the Reynolds number as can be followed from Table 1. concentrate-overall recovery trade-off commonly observed in flotation.
Even if the Reynolds number has a greater effect on the responses than It is seen in Fig. 4(a) that change in the hydrodynamic conditions from
the air capacity number, the effect of the air capacity number on the the quiescent to the turbulent flow increased the expected value of the
responses could not be neglected to evaluate the effect of hydro- cumulative chalcopyrite recovery but the recoveries of gangue (b) and
dynamic conditions on the responses. water (d) are also increased. The turbulent conditions and the high froth
From works on entrainment in the literature, one can generally overflow rate increased the amount of the recovered mineral species is
conclude that the gangue recovery (for fine particles) will linearly in- obtained. In addition, the higher water recovery was obtained under
crease with an increase in the water recovery for fixed flotation these flow conditions. These results indicate that the selectivity of
conditions. It is apparent from the results given in Table 1 that recoveries flotation is impaired under turbulent conditions, as seen from Fig. 4(c).
of water and gangue are complicated function of the hydrodynamic
conditions and they cannot be correlated each other linearly (the 3.2. Effect of the hydrodynamic conditions on entrainment and true
coefficient of determination, R2, for the linear correlation is 0.793). It is flotation
apparent that the gangue recovery is not only proportional to the water
recovery but also to other physical variables that are related to the Three contour plots of constant density of the pulp (or x) are given in
hydrodynamic conditions. It is probably due to several reasons: firstly, Figs. 5 and 6 in order to illustrate what combination of the factors
non-selective drainage of the particles from the froth to the pulp is produces high-and-low recoveries of the mineral species by true
decreased by the turbulent flow and the intensive aeration conditions. flotation and entrainment, respectively. It is apparent that the high
Secondly, it should be appreciated that amount of water transferred chalcopyrite recovery by true flotation is obtained with high values of
from the froth into the concentrate is as important as the fine-sized solid the impeller speed and the airflow rate, whereas low entrainment
content of the interstitial liquid, which decreased with a decrease in the recovery of gangue is potentially available when one uses low values of
pulp density/viscosity. Since the air capacity number characterizes the density of the pulp. In fact, extremes of low impeller speeds, low airflow
intensity of the aeration in the system, both rate of the froth removal and rates, and high density of the pulp clearly produce low recoveries of
agitation in the cell increased with increase in the air capacity number. chalcopyrite and gangue, whereas low density of the pulp, low impeller
It should be noted from Table 1 that the effect of the Reynolds speed and high airflow rate result in cleaner concentrate.
number on the water recovery was negligible at the high level of The recovery-time set of data estimated by the method was fitted
the airflow rate (8 L/min) indicating the significance of the airflow rate to Eq. (4), and the values of ki and RIi for entrainment and true
 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44 39

Fig. 5. Effect of the factors on true flotation recovery of chalcopyrite (a–c) and gangue (d–f).

flotation were determined. In addition to this, the calculated rate data hydrodynamic conditions is consistent with the change in the froth
for entrainment and true flotation were statistically analysed to overflow rate, collapse of the bubbles and agitation in the cell. Based
determine the effects of the factors on true flotation and entrainment. on the results of the statistical analysis for the estimated data, the
The analysis of variance for the factors (N, Q a and x) on the responses relative significance of the factors in a decreasing order can be
(Ri, RIi and ki for entrainment and true flotation) is summarised summarised as the impeller speed, the pulp density and the airflow
Table 2. The change in the expected values of RIi with change in the rate. From Table 2, it can be noted that changing the levels of factors
40 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44

Fig. 6. Effect of the factors on entrainment recovery of chalcopyrite (a–c) and gangue (d–f).

leads to change in the value of RICp but not kCp. It is known that this Recovery by true flotation of gangue is large as can be seen from
means that the chalcopyrite recovery can be improved without any Fig. 4. It is probably due to the possible presence of the self-induced
impact on the flotation time required (Xu, 1998). According to the floatable pyrite particles and not liberated particles (composite) in the
statistical results, the amount of chalcopyrite recovered by entrain- feed, because the gangue content of each concentrate was defined as a
ment was not affected by change in the levels of factors, whereas summation of pyrite and non-sulphide gangue. As it is followed from
interaction effects of the factors on the entrainment rate of Table 2 and Fig. 4, similar conclusions can be drawn for pyrite. The
chalcopyrite are significant. The reliability and validity of the effect of the hydrodynamic conditions on the true flotation recovery of
estimations will be discussed in detail in the following section. pyrite and gangue is more important than that of the entrainment of
Table 2
Analysis of variance for true flotation and entrainment (theoretical F value [F0.05, 1, 6] is 3.78 for all factors)

Responses Source of variance Sum of squares Dfa Mean square FCalculated

Recovery (Ri) Chalcopyrite True flotation N (rpm) 1113.121 1 1113.121 157.698
Qa (L/min) 181.435 1 181.435 25.704
x (%) 582.536 1 582.536 82.529
NQa 51.040 1 51.040 7.231
Entrainment N (rpm) 0.300 1 0.300 23.101
Qa (L/min) 0.280 1 0.280 21.587
x (%) 0.562 1 0.562 43.264
Pyrite True flotation N (rpm) 364.926 1 364.926 17.043
Qa (L/min) 328.472 1 328.472 15.341
Entrainment N (rpm) 16.582 1 16.582 28.650
Qa (L/min) 18.436 1 18.436 31.853
x (%) 31.325 1 31.325 54.122
Gangue True flotation N (rpm) 208.464 1 208.464 25.562
Qa (L/min) 401.224 1 401.224 49.198

E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44

NQa 62.625 1 62.625 7.679
Entrainment N (rpm) 59.890 1 59.890 30.230
Qa (L/min) 65.182 1 65.182 32.902
x (%) 110.538 1 110.538 55.796
Ultimate recovery (RIi) Chalcopyrite True flotation N (rpm) 1073.124 1 1073.124 270.370
Qa (L/min) 163.965 1 163.965 41.310
x (%) 555.340 1 555.340 139.916
NQa 44.617 1 44.617 11.241
Entrainment N (rpm) 0.276 1 0.276 10.915
Qa (L/min) 0.390 1 0.390 15.415
x (%) 0.730 1 0.730 28.879
Pyrite True flotation N (rpm) 353.281 1 353.281 16.384
Qa (L/min) 337.998 1 337.998 15.675
Entrainment N (rpm) 15.314 1 15.314 13.295
Qa (L/min) 26.856 1 26.856 23.316
x (%) 41.254 1 41.254 35.816
Gangue True flotation N (rpm) 204.306 1 204.306 26.613
Qa (L/min) 384.229 1 384.229 50.050
NQa 68.556 1 68.556 8.930
Entrainment N (rpm) 56.638 1 56.638 14.230
Qa (L/min) 94.994 1 94.994 23.867
x (%) 145.726 1 145.726 36.613
Rate constant (ki) Chalcopyrite True flotation – – – – –
Entrainment NQa 0.104 1 0.104 11.031
Qax 0.067 1 0.067 7.094
Pyrite True flotation x (%) 0.927 1 0.927 6.179
Nx 1.322 1 1.322 8.812
Qax 0.930 1 0.930 6.200
Entrainment NQa 0.093 1 0.093 14.157
Qax 0.063 1 0.063 9.573
Gangue True flotation N (rpm) 0.596 1 0.596 11.816
Entrainment NQa 0.089 1 0.089 12.569
Qax 0.056 1 0.056 7.816
a
Degrees of freedom.

41
42 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44

Fig. 7. The selectivity indexes of the mineral species for true flotation (a) and entrainment (b) as a function of the Reynolds number.

these mineral species. This is, in fact, inconsistent with what is indexes were calculated to be very similar in magnitude for the
expected in true flotation and entrainment. Although no clear conclu- mineral species. It means that the hydrodynamic conditions did not
sion can be drawn, this effect is believed to be due to some locked affect the entrainment rate of the mineral species. These results are
chalcopyrite and/or pyrite particles transferred by their hydrophobic not in accordance with the observed results given in Fig. 3, as is
parts. generally observed in lab or plant flotation. As described in the
previous section, the hydrodynamic conditions applied in these tests
3.3. The efficiency of entrainment were varied markedly and it is known that an increase in the Reynolds
number and/or the air capacity number could cause an increase in the
As it is known entrainment is a non-selective sub-process. The entrainment, resulting in the low concentrate grade values. As can be
amount of the fine sized mineral particles (valuable or gangue) to be followed from Fig. 4(c), the lowest grade of concentrate was obtained
recovered by entrainment is proportional to their specific gravity, shape from the flotation test conducted under turbulent conditions. It is
and amount in the feed. It must be kept in mind that the ratio of gangue evident that this can directly be related to entrainment.
to valuable could be very high up to 20 in a complex sulphide ore and, On the other hand, the selectivity index values were less than one for
therefore, it is expected that the most of the entrained particles consist all cases, as seen in Fig. 7(b). However, the SICp/P values were higher than
of the gangue. In our case, the amount of chalcopyrite recovered by that of Cp/G; i.e., the entrainment rate of the gangue particles is faster
flotation or entrainment in terms of mass for a chalcopyrite recovery of than that of pyrite for all cases. It is well known that the particle density
23.74% equals to the amount of the gangue recovered by flotation or is one of the important factors affecting entrainment. Since the non-
entrainment in terms of mass for a gangue recovery of 1%. Even though sulphide gangue minerals have low density and their amount is higher
the gangue recovery (in percentage) is much less than the chalcopyrite than that of the sulphide minerals existing in the feed, the high
recovery (in percentage), the amount of the gangue in the concentrate entrainment rate of gangue (P + NG) is obtained.
could be more than the amount of chalcopyrite, resulting in the From the results in Fig. 7, one can conclude that the main mechanism
impairment of the concentrate grade and the separation efficiency. In of the gangue transfer into the froth is true flotation rather than
addition to these, it is known that the separation selectivity of flotation entrainment. This conclusion drawn from the predictions cannot be
can be measured by the selectivity index. Therefore, the mineral species realistic because it is the well known that mechanisms of the gangue
selectivity index of the entrainment estimated by the method was seen transfer into the froth are entrainment, true flotation of the locked
to be suitable for evaluation of the effects of the factors on the particles, slime coating etc.
entrainment. As described in details by Xu (1998), the selectivity index is
defined as the ratio of the modified rate constant (kmodi = RIiki) of mineral 3.4. Effect of the self-induced floatability
I to the modified rate constant of mineral II.
Because of these findings and discussions above, the effect of the self-
−1
SII=II = RII kI ðRIII kII Þ : ð5Þ induced floatability and particle size on the levels of entrainment and
true flotation must experimentally be determined to prove the above
The rate parameters were used to calculate the selectivity indices of conclusions drawn. The levels of true flotation and entrainment were
chalcopyrite/pyrite (SICp/P) and the chalcopyrite/gangue (SICp/G). The determined by the method of Trahar and Warren (1976). To do this, test 8
values of the SICp/P and SICp/G generated from the kinetic model was repeated in the absence and in the presence of the collector, and a
parameters of entrainment and true flotation are given in Fig. 7. As size-by-size analysis was conducted for products of each test. For the
expected, the true flotation rate of chalcopyrite particles was faster than sake of brevity, only two fractions (−0.075 + 0.02 mm, −0.02 mm) were
that of pyrite and non-sulphide gangue particles in all cases. However, used in the size-by-size analysis and the results are illustrated in Fig. 8.
the values of SICp/G were higher than the values of SICp/P and, this is As expected, there were significant differences between the levels of
consistent with what is generally recognized in flotation of the sulphide entrainment and true flotation obtained in the absence and the presence
ores having pyrite. of the collector.
Fig. 7(b) shows the selectivity indexes of the mineral species for Approximately 30% of pyrite particles (both two fractions) in the feed
entrainment in all cases, proving once again the non-selectivity of were recovered by true flotation. Even if it was defined as a sulphide
entrainment. Since the mineral were transferred into the froth gangue mineral in this study, the high recovery due to true flotation is
according to their specific gravities and percentages in the feed, the attributable to the self-induced floatability of pyrite rather than the
entrainment rate of chalcopyrite was slower than those of the other presence of the locked particles in the feed. Subrahmanyam and
minerals. As can be seen from Fig. 7(b), the entrainment selectivity Forssberg (1988b) reported similar conclusions. However, a very low
 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44 43

Fig. 8. Size-by-size recoveries of chalcopyrite (a–b), pyrite (c–d) and gangue (e–f) due to overall flotation, true flotation and entrainment estimated from the method of Trahar and
Warren (1976) (Run #8; N = 1500 rpm, Q a = 8 L/min; x = 25%).

true flotation recovery of chalcopyrite particles in 0.02 mm was obtained entrainment and true flotation in flotation of this ore. A different test
surprisingly. These estimates by method of Trahar and Warren (1976) procedure, which resulted from the following assumptions, was adopted
indicate that the overall recovery of the self-induced floatable minerals and used to overcome this difficulty. It is assumed that the recovery by
(chalcopyrite and pyrite) comprises one part dependent on entrainment entrainment in the collectorless flotation of the ore can be reduced to a
and another part dependent on true flotation. Therefore, the presence of negligible level by using a flotation test conducted under very quiescent
self-induced hydrophobic particles in the feed is as important as the flow condition, deep froth and addition of wash water. In this way, it is
presence of fine particles for accurate estimation of true flotation and expected that most of the solids recovered by the collectorles flotation
entrainment in flotation of a complex sulphide ore. Although these two will be composed of the naturally hydrophobic particles, which collide
recoveries must be distinguished to accurately determine/estimate the the rising bubbles and can be attached to their surfaces. In other words,
contributions of entrainment and true flotation, this cannot be achieved the only naturally hydrophobic particles in the ore can completely be
by using the two methods used. Opposite to this, George et al. (2004) recovered by using the collectorless flotation conducted under these
have compared the methods described in the literature and have found flotation conditions. In order to test this, firstly a collectorless flotation
highly consistent results by using an artificial mixture of the natural test was performed to recover the self-induced floatable particles. Then,
hydrophilic minerals (silica and alumina). the pulp that remained in the cell was further conditioned with the
It is clear that the self-induced floatability of the sulphide minerals is collector and, a kinetic flotation test was performed in the second stage.
a main difficulty for accurate determination of the contributions of In order to compare the results of these tests with the results of test 15
44 E.C. Cilek / Int. J. Miner. Process. 90 (2009) 35–44

Table 3 effect of the self-induced floatability on the estimation of the

Results of additional flotation tests conducted in sequence entrainment could be overcome. Since the self-induced minerals
Collectorless flotation Flotation (chalcopyrite and pyrite) in the feed were partly removed/recovered
Products Mineral (%) Recovery, Ri (%) conditions in this stage, this attempt failed. It is apparent that a new approach
CuFeS2 FeS2 Gangue CuFeS2 FeS2 Gangue would be necessary to determine the contributions of true flotation
Concentrate 48.83 44.67 51.17 9.07 1.05 0.08 Qa = 3.5 L/min
and entrainment in flotation of the complex sulphide ores.
N = 1000 rpm
Tailings 5.84 50.16 94.16 90.93 98.95 99.92 x = 15% Acknowledgements
Q ww = 0.5 L/min
Tf = 25 mm
The help and support of the staff at the Kure Copper Concentrator,
Kinetic flotation
Concentrate-1 17.54 57.63 82.46 37.72 14.42 11.01 Q a = 6.5 L/min Etibakir A.S., is gratefully acknowledged. The author would also like to
Concentrate-2 15.68 59.82 84.32 49.97 22.19 16.68 N = 1250 rpm acknowledge the anonymous referees for their helpful comments on
Concentrate-3 13.59 61.68 86.41 54.62 28.86 21.56 x = 15% an earlier version of this paper.
Concentrate-4 12.44 63.24 87.56 59.26 35.07 25.89 Tf = 20 mm

References
Selectivity Chalcopyrite/ Chalcopyrite/
index Pyrite Gangue Angadi, S.I., Suresh, N., 2005. A Kinetic Model for the Prediction of Water Reporting to the
True flotation 4.21 6.48 Froth Products in Batch Flotation. Trans. of the Inst. of Min. and Metall. (Section C:
Entrainment 0.125 0.073 Miner. Process. and Extr. Metall.), vol. 114, pp. 225–232.
George, P., Nguyen, A.V., Jameson, G.J., 2004. Assessment of true flotation and
entrainment in the flotation of submicron particles by fine bubbles. Miner. Eng. 17,
847–853.
Lynch, A.J., Johnson, N.W., Manlapig, E.V., Thorne, G.C., 1981. Mineral and Coal Flotation
(see Fig. 3), the medium levels of factors were applied in the second Circuits, Their Simulation and Control. Elsevier, Amsterdam, pp. 56–96.
stage. Table 3 shows the results of these tests performed in sequence. Montgomery, D.C., 1991. Design and Analysis of Experiments, Third Edition. Wiley &
A clean concentrate was produced from the collectorless flotation Sons, Inc., New York, pp. 521–548.
Myers, R.H., Montgomery, D.C., 1995. Response Surface Methodology: Process and
and, the results given in Table 3 indicate that the entrainment was product optimization using designed experiments. J. Wiley & Sons, Inc., New York,
minimized to a negligible level. However, recoveries of chalcopyrite pp. 297–341.
and pyrite were very low too. As can be followed from Table 3 and Nelson, M.G., Lelinski, D., 2000. Hydrodynamic design of self-aerating flotation
machines. Miner. Eng. 13 (10–11), 991–998.
Fig. 6, there are no significant differences between the values of SICp/P
Nguyen, A.V., Schulze, H.J., 2004. Colloidal Science of Flotation. Surfactant Science
and SICp/G obtained from the second stage test and from test 15. This is Series, vol. 118. Marcel Dekker, Inc., New York, pp. 709–775.
clearly indicated that only a part of the self-induced minerals in the Nguyen, A.V., George, P., Jameson, G.J., 2006. Demonstration of a minimum in the
recovery of nanoparticles by flotation: theory and experiment. Chem. Eng. Sci. 61,
feed was recovered by means of true flotation in the first stage
2494–2509.
(collectorless flotation). Therefore, it was observed that the recoveries Pan, C.M., Nguyen, A.V., Miller, J.D., Evans, G.M., Jameson, G.J., 2003. Investigation of
of the mineral species obtained from the second stage were very close bubble–particle interactions. Int. J. Miner. Process. 72, 239–254.
to those obtained from test 15. These results clearly indicate that the Polat, M., chander, S., 2000. First-order flotation kinetics models and methods for
estimation of the true distribution of flotation rate constants. Int. J. Miner. Process. 58,
self-induced minerals in a complex sulphide ore can partly be 145–166.
recovered in the absence of a collector. Further investigations would Rodrigues, W.J., Leal Filho, L.S., Masini, E.A., 2001. Hydrodynamic dimensionless
be necessary to use this approach for the determination of the parameters and their influence on flotation performance of coarse particles.
Miner. Eng. 14 (9), 1047–1054.
contributions of true flotation and entrainment in flotation of a Ross, V.E., 1990. Flotation and entrainment of particles during batch flotation tests.
complex sulphide ore. Miner. Eng. 3 (3/4), 245–256.
Ross, V.E., 1991. Comparison of Methods for Evaluation of True Flotation and
Entrainment. Trans. of the Inst. of Min. and Metall. (Section C: Miner. Process.
4. Conclusions and Extr. Metall.), vol. 100, pp. 121–126.
Ross, V.E., Van Deventer, J.S.J., 1988. In: Sastry, K. (Ed.), Column Flotation'88 —
The kinetic flotation tests were conducted to estimate true Proceedings of an International Symposium. SME Inc., Littleton, pp. 383–386.
Schubert, H., 1999. On the turbulence-controlled microprocesses in flotation machines.
flotation and entrainment. In order to evaluate the self-induced
Int. J. Miner. Process. 56 (1–4), 257–276.
floatability of the sulphide minerals in the ore, some of these tests Sheridan, M.S., Nagaraj, D.R., Fornasiero, D., Ralston, J., 2002. The use of a factorial
were conducted in the presence and absence of a collector. The rate experimental design to study collector properties of N-allyl-O-alkyl thionocarba-
mate collector in the flotation of a copper ore. Miner. Eng. 15, 333–340.
data generated from the experiments were used to calculate the
Smith, P.G., Warren, L.J., 1989. Entrainment of particles into flotation froths. Miner.
selectivity index of each mineral for entrainment and true flotation. Process. Extr. Metall. Rev. 5, 123–145.
The results show: Subrahmanyam, T.V., Forssberg, E., 1988a. Froth stability, particle entrainment and
drainage in flotation—a review. Int. J. Miner. Process. 23, 33–53.
1- The selectivity indexes were calculated to be very similar in Subrahmanyam, T.V., Forssberg, E., 1988b. A study of particle entrainment in flotation
magnitude for entrainment of the mineral species. It means that with different frothers—the case of copper ore. In: Forssberg, E. (Ed.), XVI
International Mineral Processing Congress. Elsevier, pp. 785–796.
there is no effect of the hydrodynamic conditions on the Subrahmanyam, T.V., Forssberg, E., 1988c. Froth characteristics and grade-recovery
entrainment of the mineral species. relationships in flotation of lead–zinc and copper ores. Miner. Eng. 1 (1), 41–52.
2- The main mechanism of the gangue transfer into the froth is true Trahar, W.J., Warren, L.J., 1976. The floatability of very fine particles—a review. Int. J.
Miner. Process. 3, 103–131.
flotation rather than entrainment. Warren, L.J., 1985. Determination of the contributions of true flotation and entrainment
in batch flotation tests. Int. J. Miner. Process. 14, 33–44.
It is clear that these findings cannot be realistic and the estimated Xu, M., 1998. Modified flotation rate constant and selectivity index. Miner. Eng. 11 (3),
results for true flotation and entrainment can be misleading. In 271–278.
addition, the results of the size-by-size analysis of the froth products Yoon, R.-H., 2000. The role of hydrodynamic and surface forces in bubble–particle
interaction. Int. J. Miner. Process. 58, 129–143.
indicated that the presence of self-induced hydrophobic particles in
Zheng, X., Franzidis, J.P., Johnson, N.W., 2006. An evaluation of different models of water
the feed is as important as the presence of fine particles for accurate recovery in flotation. Miner. Eng. 19, 871–882.
estimation of true flotation and entrainment in flotation of a complex
sulphide ore. It was hoped that if all the self-induced minerals could
be removed/recovered by a suitable test procedure, the detrimental