Minerals Engineering 74 (2015) 156162

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Inuence of liberation on bubbleparticle attachment time in otation

Boris Albijanic a,, G.K. Nimal Subasinghe a, Dee J. Bradshaw c, Anh V. Nguyen b
a
Western Australia School of Mine, Curtin University, Kalgoorlie 6430, Australia
b
School of Chemical Engineering, The University of Queensland, Brisbane, QLD 4072, Australia
c
Julius Kruttschnitt Mineral Research Centre, The University of Queensland, Brisbane, QLD 4072, Australia

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

Article history: Flotation is controlled by the bubbleparticle attachment mechanism which depends on the particle sur-
Received 18 January 2014 face properties i.e., the particle composition, the surface liberation of valuable minerals and collector
Revised 3 August 2014 adsorption. This paper focuses on using the bubbleparticle attachment method to understand the fac-
Accepted 11 August 2014
tors affecting attachment time. The attachment time measurements were performed with sized concen-
Available online 16 September 2014
trates obtained by otation of a copper sulphide ore (Northparkes Mine, Australia) in a mechanically
agitated batch otation cell. Quantitative mineral liberation analysis was used to determine the miner-
Keywords:
alogy of otation concentrates. The results showed that the higher the amount of highly and moderately
Froth otation
Attachment time
liberated copper minerals in otation concentrates, the lower the attachment time. By using attachment
Mineral liberation time and collector dosage, we dened a non-linear empirical correlation to estimate Cu grade. The pro-
Empirical correlation posed empirical correlation has shown a satisfactory agreement between the calculated and the experi-
mental Cu grade. These results showed that attachment time measurements are related to the Cu grade.
This relationship may be used in the future to develop a practical method (without assays) to monitor
changing grade for a specied system (otation plant). It also may be possible to infer potential grade
if mineralogy samples are available, but not enough samples are available for conventional otation tests.
However this requires a signicant amount of further work.
Crown Copyright 2014 Published by Elsevier Ltd. All rights reserved.

1. Introduction bubbleparticle attachment time of real ore particles is very

The key mechanism for successful otation is bubbleparticle
attachment which is mainly investigated by contact angle mea-
surements (Leja, 1982; Nguyen and Schulze, 2004). However, these
measurements cannot always predict the otation response with
Halimond tube (Ye et al., 1989). Namely, even though the minerals
are oatable, the contact angle of the minerals can be very small.
For that reason, an alternative measure for predicting the suscepti-
bility of mineral to oat is bubbleparticle attachment time, which
is dened as the time required for an attachment of particles to an
air bubble when they are in close proximity (Albijanic et al.,
2010b). Bubbleparticle attachment time involves the three
stages: the draining of liquid lm between a bubble and a particle,
rupture of liquid lm and the formation of a three phase contact
between air bubble, mineral and liquid (Nguyen et al., 1997).
The measurement and the prediction of bubbleparticle attach-
ment time are very important since the information could poten-
tially be used in the future optimisation, modelling and
simulation of otation circuits. However, the prediction of
Corresponding author. Tel.: +61 8 9088 6117; fax: +61 8 9088 6181.
E-mail address: boris.albijanic@curtin.edu.au (B. Albijanic).
difcult since the surface chemistry of bubbles and real composite
particles cannot be described from rst principles and hence, the
real value lies in ascertaining whether it is possible to establish
an empirical model that links bubbleparticle attachment time
with otation data. However, the literature does not provide an
empirical model describing the relationship between bubble
particle attachment time and otation data. For that reason, the
main objective in this work is to investigate possible relationships
between the bubbleparticle attachment time and the degree of
liberation of a copper-sulphide ore as well as to develop an
empirical model which can successfully predict bubbleparticle
attachment time for an industrial system.
2. Previous work
Numerous researchers have used a device developed by
Glembotsky (1953) to measure bubbleparticle attachment time,
and have shown that otation recovery is inversely proportional
to bubbleparticle attachment time (Ye et al., 1989; Yoon and
Yordan, 1991; Albijanic et al., 2010a). The Glembotsky device is
based on keeping a bubble in contact with an upper surface of

http://dx.doi.org/10.1016/j.mineng.2014.08.004
0892-6875/Crown Copyright 2014 Published by Elsevier Ltd. All rights reserved.
 B. Albijanic et al. / Minerals Engineering 74 (2015) 156162
particle bed at different controlled contact times, from which the
attachment time is determined at a pre-selected percentage (e.g.,
100%) of particle attachment.
Although the Glembotsky device has been typically used to
determine attachment time, other techniques might be also used
such as the wetting lm stability measurements (Letocart et al.,
1999) as well as using models to back calculate bubbleparticle
attachment time (Danoucaras et al., 2013; Min and Nguyen, 2013).
Various researchers have used mainly pure minerals such as
quartz (Yoon and Yordan, 1991; Gu et al., 2003; Albijanic et al.,
2010a; Subasinghe and Albijanic, 2014; Albijanic et al., 2014) or
coal (Ye et al., 1989) to conrm that bubbleparticle attachment
time is sensitive enough to show changes in physical properties
of minerals (particle size and shape), solution chemistry (pH, dis-
solved ions and surfactant concentration), bubble size and temper-
ature of solution. A comprehensive review of these ndings is
recently summarized by Albijanic et al. (2010b).
Apart from determining attachment time of pure minerals,
attachment time has been also measured in the case of real com-
posite particles (Albijanic et al., 2011, 2012). The most relevant
conclusion of these studies is that attachment time is strongly
affected by surface exposure of valuable minerals, grade of valu-
able mineral and collector dosage. For particles with high grade
of valuable metal, a dramatic reduction of attachment time was
obtained with a small increase of collector dosage. However, in
the case of particles with lower mineral grade, attachment time
was not inuenced by collector addition.
Regarding the prediction of attachment time of real ore parti-
cles from the otation data, Danoucaras et al. (2013) used the P9
otation model (Savassi, 1998) which shows the relationship
between otation mineral recovery and the following variables:
oatability, entrainment, froth and water recovery, the bubble sur-
face area ux and residence time. In the P9 otation model, the
relationship between oatability and the attachment time is
described using fundamental models for particle collection by bub-
ble (Dobby and Finch, 1987). By processing the measurements
obtained by Vianna (2004) for galena size-liberation classes,
Danoucaras et al. (2013) calculated the attachment time.
Albijanic et al. (2012) demonstrated that for high grade particles
(22.240.4%) measured attachment time is close to the value
obtained by Danoucaras et al. (2013) while for lower grade parti-
cles (<12.3%) the measured attachment time was up to 100 times
higher than the attachment time calculated by Danoucaras et al.
(2013). In other words, the prediction of bubbleparticle attach-
ment time from otation data using rst principles remains a
challenge.
3. Materials and methods
3.1. Ore sample
The low grade coppergold sulphide ore (1% Cu) was obtained
from Northparkes mine (New South Wales, Australia). Major valu-
able minerals, found in this ore, are chalcopyrite (1.1%) and bornite
(1.6%), while non-valuable minerals are mainly different type of
silicates such as plagioclase (31.8%), quartz (14.7%), orthoclase
(12.0%) and muscovite (11.4%). It should be noted that non-
valuable sulphide mineral is pyrite (0.2%).
3.2. Flotation experiments
The feed ore was prepared to a P80 of 90 microns by wet milling
at 60% w/w solids in a laboratory rod mill. The ground sample was
transferred to a 5 L bottom driven batch otation cell, and the
required amount of tap water was added to the system (25% w/
157
w solid ratio). The measured pH of the otation pulp was between
8.4 and 8.8. Brisbane tap water was used in the otation
experiments.
The impeller speed was set at 800 rpm. Flotation experiments
were performed as a function of collector dosage of sodium isobu-
tyl xanthate (SIBX) supplied by Senmin Pty Ltd., South Africa. Addi-
tionally, 30 g/t of the Aerooat 208 promoter (Cytec, USA) were
added to the system. After 3 min of conditioning, 14 mg/L of the
Interfroth 6880 alcohol based frother (Chemical & Mining Services
Pty Ltd, Australia) was added to the otation pulp. It should be
noted that all otation chemicals were used as supplied. After
2 min of mixing otation pulp, 15 L/min of air was introduced
through diffusers placed at the bottom of the otation cell, and
the otation took place. An automatic scraper was used to collect
the froth into two launders every 8 s. The concentrates were col-
lected after 0.5 min, 4 min and 10 min and were wet screened to
obtain the sized fraction (53106 lm) for the bubbleparticle
attachment measurements. It is important to note that attachment
time depends on particle size (Yoon and Yordan, 1991; Gu et al.,
2003). However, it was not possible to use the narrower sized frac-
tion for bubbleparticle attachment measurements because the
amount of collected fraction (53106 lm) was less than 1 g. Addi-
tionally, the selected size fraction is in the range of optimum oat-
ability (Muganda et al., 2012) and thus was considered suitable for
the bubbleparticle attachment measurements.
3.3. Bubbleparticle attachment measurements
The bubbleparticle attachment measurements were con-
ducted with the Induction timer from the University of Alberta,
Canada (Gu et al., 2003) as shown in Fig. 1. The sample and a super-
natant solution were transferred to a small cell under the bubble
holder. The solution pH was 8.58.8. A small bubble (1.5 mm in
diameter) was generated with a microsyringe, and the distance
between a bubble and an upper surface of particle bed was
adjusted to be constant in all experiments. The bubble was kept
in contact with the upper surface of particle bed for the controlled
contact time (103000 ms). Then the CCD camera was used to
visually observe whether particles were attached to the bubble.
For each controlled contact time, ten measurements were repeated
to determine attachment efciency (expressed as a number of suc-
cessful attachments in 10 observations). The contact time at which
attachment efciency is 100% represents attachment time
(Albijanic et al., 2012).
3.4. Mineral liberation analysis
Mineral liberation analysis of the sized samples of (53106 lm)
concentrates was performed with a Mineral Liberation Analyser.
For more details about this technique, the readers are referred to
the relevant literature (Gu, 2003; Fandrich et al., 2007). Mineral
liberation measurements based on crosssectional area fraction
of particles are generally given as cumulative liberation yield y
(%) as proposed by Miller et al. (1982), which is the cumulative
fraction of particles having a mineral composition of at least C.
The 95% condence interval of y is generally calculated using the
standard deviation derived by Leigh et al. (1993), and given as
follows:
s
1 1
ry 1:12y1  y 1
N0 N1
where N0 is the number of particles with composition at least C and
N1 is number of particles with composition higher than C. The
number of particles measured on one 2D areal section was about
158 B. Albijanic et al. / Minerals Engineering 74 (2015) 156162

Driver head
Attached
particles
Capillary
Bubble

Lens
Light

PC Particle bed
CCD camera

Micro-translation stage

Bubble Attached
particles on
bubble surface
Particle bed

Step Step Step

1 2 3

Fig. 1. Schematic representation of bubbleparticle attachment time method (Albijanic et al., 2012).

10,500, and the standard deviation of the cumulative liberation
yield did not exceed 0.03.
4. Results and discussion
4.1. Flotation performance
The typical results of the otation experiments are shown in
Fig. 2. As seen in Fig. 2, for each liberation class a signicant
amount of material was recovered in the rst minute while after
4 min, the rate of recovery was dramatically slower. Additionally,
the maximum otation recovery was accomplished for fully liber-
ated copper sulphide particles, while the lowest otation recovery
was obtained for particles containing less than 25% of copper sul-
phide. It should be noted that otation recovery for each liberation
class was calculated using liberation data given in Section 4.3.
In order to describe the relationship between recovery of the
liberation class and otation time (see Fig. 2), the Klimpel model
(a rst order kinetic model) was applied, giving:
R R1 1  expkt 2

where k is the otation rate constant of the liberation class and R1

100 is the ultimate recovery of the liberation class. The values for the
otation rate constant as a function of mean sulphide liberation
and SIBX dosage are shown in Fig. 3. As seen in Fig. 3, the otation
80 rate constant increases with the increase of mean sulphide libera-
tion, demonstrating that the particles containing higher percentage
Recovery (%)

of copper sulphide minerals were better recovered than the parti-

60 cles with lower percentage of copper sulphide minerals. Addition-
ally, the highest otation rate constant was obtained when
otation pulp was conditioned in the presence of 50 g/t of SIBX.
40

Liberation -1
k(min ) Rinf 4.2. Bubbleparticle attachment measurements
100% 5.75 0.98
20 100%-75% 5.27 0.98
75%-25% 4.60 0.90 The bubbleparticle attachment time measurements were
less than 25% 2.77 0.52 performed on the sized fraction (53106 lm) obtained for the
otation concentrates (00.5 min (Con-1), 0.54 min (Con-2) and
0
0 2 4 6 8 10 410 min (Con-3)) from a batch otation test in which otation
pulp was conditioned at different SIBX dosages.
Flotation time (min)
Fig. 4 shows the typical results for attachment efciency (num-
Fig. 2. Cumulative sulphide recovery for different liberation classes when otation ber of successful attachments in 10 observations) as a function of
pulp was conditioned in the absence of SIBX. contact time. The attachment time is dened as the contact time
 B. Albijanic et al. / Minerals Engineering 74 (2015) 156162 159

7 60
SIBX(g/t) SIBX(g/t)
0 50 0
Flotation rate constant (min -1)

6 25

Cu mineral grade (%)

25
50
40 50

5
30

4 20

10
3
0
Con-1 Con-2 Con-3
2 Fig. 5. Copper mineral grade of the 53106 lm fraction in concentrates.
0 20 40 60 80 100
Mean sulfide liberation (%)
Fig. 5 shows that the Cu mineral grade of oated products
Fig. 3. Flotation rate constant as a function of mean sulphide liberation. decreased with the increase of otation time. For example, when
otation pulp was conditioned in the presence of 50 g/t of SIBX,
Cu mineral grade decreased from 54% (Con-1) to 14% (Con-3).
100 Additionally, the increase of collector dosage caused more adsorp-
Con-1 tion of this molecule on the surface of Cu minerals, and thus at
Con-2 higher collector dosage more Cu minerals were recovered.
Con-3 The mineral liberation measurements also provided liberation
Attachment efficiency (%)

80
60
40
20
0 concentrates.
1 10
Contact time (ms)
Fig. 4. Attachment efciency for the concentrates as a function of contact time
obtained when otation pulp was conditioned with 50 g/t of SIBX.
at 100% attachment efciency. As seen in Fig. 4, there is the critical
contact time above which attachment efciency increases dramat-
ically. The reason for this behaviour is because above the critical
contact time there is more time for drainage of liquid lm between
a bubble and a particle, resulting in stronger bubbleparticle
attaching force. Determination of this force as a function of contact
time was beyond the scope of this work, and has not yet been
quantied; however, it would be an interesting topic for future
research.
4.3. Mineral liberation measurements
The copper mineral grade of particles in otation products
(Con-1, Con-2 and Con-3) collected when otation pulp was condi-
tioned at different SIBX dosages is given in Fig. 5. It should be noted
that in Fig. 5, combined Cu mineral grade was shown instead of
grade of bornite and chalcopyrite, because these two minerals
are equally oatable in the presence of collector (Ackerman et al.,
1989).
data for the Cu minerals. Fig. 6 shows the Cu grade and liberation
data for the three concentrates in three classes: highly liberated
(Cu mineral liberation > 75%), moderately liberated (75% > Cu min-
eral liberation > 25%), and poorly liberated (25% > Cu mineral liber-
ation). As seen in Fig. 5, the amount of highly-liberated particles
recovered decreased with the increase of otation time. A greater
amount of this liberation class was recovered when the otation
pulp was conditioned in the presence of SIBX. For example, when
otation pulp was treated with 25 g/t of SIBX, the Con-2 had 10%
more highly liberated particles than in the absence of SIBX. The
increase of SIBX dosage also improved recovery of particles having
less than 25% of copper sulphide, particularly in the later
100 1000
4.4. Effect of liberation and collector dosage on attachment time
Fig. 7 shows the effect of the liberation classes (see Fig. 6) on
attachment time. As seen in Fig. 7ac, at constant collector dosage,
the higher the amount of highly or moderately liberated Cu miner-
als, the lower the attachment time. For example, when otation
pulp was conditioned in the presence of 25 g/t of SIBX, the
increased amount of fully liberated Cu minerals in otation con-
centrates resulted in decreasing attachment time over 20-fold from
400 ms (36%) to 20 ms (55%). On the other hand, it can be also
observed from Fig. 7d, that the relationship between the amount
of the poorly liberated Cu minerals and the attachment time is
not clear, but this may be dependent on the increased error of mea-
surement for this class. Further repeat tests are recommended to
conrm this and evaluate whether a similar relationship to that
obtained with the other classes develops or not.
Regarding the effect of collector dosage on the attachment time,
it was found that the increase of collector dosage resulted in signif-
icant decrease in attachment time and a slight increase in the o-
tation rate constant. The reason for this is because the increase of
SIBX dosage enhanced the surface coverage of SIBX molecules on
Cu minerals, and thus surface hydrophobicity of these minerals
was improved. However, quantitative determination of the surface
coverage was beyond the scope of this work.
160 B. Albijanic et al. / Minerals Engineering 74 (2015) 156162

35 16 7
SIBX(g/t)
Con-1 Con-2 Con-3
0
% Cu sulfide minerals

30 14 6
25
25 50 12
5
10
20 4
8
15 3
6
10 2
4
5 2 1

0 0 0
100%Lib 100%>Lib>75% 75%>Lib>25% Lib>25% 100%Lib 100%>Lib>75% 75%>Lib>25% Lib>25% 100%Lib 100%>Lib>75% 75%>Lib>25% Lib>25%

Fig. 6. Distribution of copper sulphide minerals for different liberation classes in the concentrates (Con-1, Con-2 and Con-3) as a function of SIBX collector dosage.

10000 10000

k(0 g/t)=5.75 min-1 k(0 g/t)=5.27 min-1

Attachment time (ms)

1000 Attachment time (ms) 1000

100 100 k(25 g/t)=5.1 min-1

k(25 g/t)=6.1 min-1

k(50 g/t)=5.4 min-1

10 10
k(50 g/t)=6.1 min-1

1 1
0 5 10 15 20 25 30 0 2 4 6 8 10 12
Fully liberated Cu minerals (%) 75%<Liberation of Cu minerals<100%

10000 k(0 g/t)=4.6 min-1 10000

Attachment time (ms)

Attachment time (ms)

k(0 g/t)=2.77 min-1

1000 1000

k(50 g/t)=3.4 min-1

100 k(25 g/t)=4.4 min-1 100

10 k(25 g/t)=3.5 min-1

10
k(50 g/t)=5 min-1

1 1
2 4 6 8 10 12 2 4 6 8 10 12
25%<Liberation of Cu minerals<75% Liberation of Cu minerals<25%

Fig. 7. Attachment time as a function of mass percentage of Cu sulphide minerals in the concentrates (h Con-1, 4 Con-2 and s Con-3) at different SIBX dosages.

4.5. Effect of Cu grade and collector dosage on attachment time collector dosage, the higher the Cu grade, the lower the attachment
time. The reason for the drop in the attachment time is not only
The results of attachment time measured on concentrates were due to the decrease of Cu grade but also because of the decrease
correlated with their Cu grade and collector dosage added during of Cu liberation with the increase of otation time. In other words,
conditioning of otation pulp. As seen in Fig. 8, at constant both Cu grade and Cu liberation affect the attachment time.
 B. Albijanic et al. / Minerals Engineering 74 (2015) 156162 161

30 A satisfactory agreement between the experimental and the cal-

Con-1 culated Cu grade was 13.3%, as seen in Fig. 9. It should be noted
Con-2 that the proposed model is limited to the ore tested which com-
Con-3 prised of predominantly chalcopyrite and bornite. However, it is
0 g/t
shown that the relationship exists between attachment time and
20 the ore grade of the concentrate and this could be used in the
Cu grade (%)

future.
It is important to note that the effect of otation time is indi-
25 g/t
rectly incorporated into the attachment time i.e., the longer the o-
tation time the more poorly liberated particles would report to the
concentrates, resulting in longer attachment time.
10
The empirical correlation described by Eq. (2) can be used to
50 g/t estimate Cu grade in otation concentrates. However, attachment
time method is an off line laboratory technique which is manually
performed, and thus is time consuming. This method requires a
skilled operator who changes contact time and visually observes
0 whether bubbleparticle aggregate has occurred. Additionally,
1 10 100 1000 10000 100000
another drawback of this method is that all measurements should
Attachment time (ms) be performed with sized samples having simple mineralogy.

Fig. 8. The relationship between the attachment time, Cu grade and collector
dosage. The line represents calculated Cu grade with Eq. (2).
5. Conclusions

In this study, the bubbleparticle attachment and mineral liber-

However, these effects have not yet been decoupled and it is worth
further investigations.
Fig. 8 also showed that at constant Cu grade, the increase of col-
lector dosage resulted in the decrease of the attachment time since
more molecules of collectors are adsorbed on Cu minerals, making
these minerals more hydrophobic.
By applying regression analysis to the experimental data,
shown in Fig. 8, Cu grade (g) was correlated with the attachment
time (tat, ms). For this purpose, a power-law relationship showed
the lowest deviation between experimental and calculated Cu
grade, and given by:
g a  tat 0:3
where a is a constant which depends on collector dosage, C (g/t).
The empirical correlation between a and C was found with LabFit
statistical software as follows:
 
a 96:4  0:4 e7C=100
30
ation measurements were used to relate the degree of liberation of
Cu minerals with attachment time. The results demonstrated that
the presence of highly and moderately liberated Cu minerals
resulted in a decreased attachment time. This work showed that
a non-linear correlation between the attachment time and Cu
grade exists which may be used to predict the concentrate Cu
grade. This work should contribute to a better understanding of
bubbleparticle attachment mechanism occurring during otation
of copper-sulphide ores. This relationship may be used in the
future to develop a practical method (without assays) to monitor
changing grade for a specied system (otation plant). It also
may be possible to develop a method to infer potential grade if
3 mineralogy samples are available, but not enough samples are
available for conventional otation tests. This requires a signicant
amount of further work.
Acknowledgement
4
Financial support from the AMIRA International P9O Project and
the member companies is gratefully acknowledged. We thank the
anonymous reviewers for their great feedback.
References

Ackerman, P.K., Harris, G.H., Klimpel, R.R., Aplan, F.F., 1989. Evaluation of otation
+15% collectors for copper suldes and pyrite, III. Effect of xanthate chain length and
branching. Int. J. Miner. Process. 21 (12), 141156.
Albijanic, B., Hampton, M., Nguyen, P., Ozdemir, O., Bradshaw, D., Nguyen A., 2010a.
20
(Cu grade) calc

An integrated study of bubbleparticle attachment mechanisms. In: XXV

International Mineral Processing Congress, pp. 17031709.
-15% Albijanic, B., Ozdemir, O., Nguyen, A.V., Bradshaw, D., 2010b. A review of induction
and attachment times of wetting thin lms between air bubbles and particles
and its relevance in otation separation of particles. Adv. Colloid Interface Sci.
159 (1), 121.
Albijanic, B., Amini, E., Wightman, E., Ozdemir, O., Nguyen, A.V., Bradshaw, D.J.,
10 2011. A relationship between the bubbleparticle attachment time and the
mineralogy of a copper-sulphide ore. Miner. Eng. 24 (12), 13351339.
Albijanic, B., Bradshaw, D.J., Nguyen, A.V., 2012. The relationships between the
bubbleparticle attachment time, collector dosage and the mineralogy of a
=13.3% copper sulphide ore. Miner. Eng. 3638, 309313.
Albijanic, B., Ozdemir, O., Hampton, M., Nguyen, P., Nguyen, A., Bradshaw, D., 2014.
Fundamental aspects of bubbleparticle attachment mechanism in otation
0 separation. Miner. Eng. 65, 187195.
Danoucaras, A.N., Vianna, S.M., Nguyen, A.V., 2013. A modelling approach using
0 10 20 30 back-calculated induction times to predict recoveries in otation. Int. J. Miner.
(Cu grade)exp Process. 124, 102108.
Dobby, G.S., Finch, J.A., 1987. Particle size dependence in otation derived from a
fundamental model of the capture process. Int. J. Miner. Process. 21 (34), 241
Fig. 9. Comparison between experimental and calculated Cu grade. 260.
162 B. Albijanic et al. / Minerals Engineering 74 (2015) 156162
Fandrich, R., Gu, Y., Burrows, D., Moeller, K., 2007. Modern SEM-based mineral
liberation analysis. Int. J. Miner. Process. 84 (14), 310320.
Glembotsky, V.A., 1953. Izv. Akad. Nauk SSSR (OTN) 11, 15241531 (In Russian).
Gu, Y., 2003. Automated scanning electron microscope based mineral liberation
analysis. J. Miner. Mater. Charact. Eng. 2, 3341.
Gu, G., Xu, Z., Nandakumar, K., Masliyah, J., 2003. Effects of physical environment on
induction time of air-bitumen attachment. Int. J. Miner. Process. 69, 235250.
Leigh, G.M., Sutherland, D.N., Gottlieb, P., 1993. Condence-limits for liberation
measurements. Miner. Eng. 6 (2), 155161.
Leja, J., 1982. Surface Chemistry of Froth Flotation. Plenum Press, New York, NY.
Letocart, P., Radoev, B., Schulze, H.J., Tsekov, R., 1999. Experiments on surface waves
in thin wetting lms. Colloids Surf., A 149, 151159.
Miller, P.R., Reid, A.F., Zuiderwyk, M.A., 1982. QEM SEM image analysis in the
determination of model assays, mineral association and mineral liberation. In:
XIV International Mineral Processing Congress, 813.R., Reid.
Min, M.A., Nguyen, A.V., 2013. An exponential decay relationship between micro-
otation rate and back-calculated induction time for potential ow and mobile
bubble surface. Miner. Eng. 40, 6780.
Muganda, S., Zanin, M., Grano, S.R., 2012. Benchmarking the otation performance
of ores. Miner. Eng. 26, 7079.
Nguyen, A.V., Schulze, H.J., 2004. Colloidal Science of Flotation. Marcel Dekker, New
York.
Nguyen, A.V., Schulze, H.J., Ralston, J., 1997. Elementary steps in particlebubble
attachment. Int. J. Miner. Process. 51 (14), 183195.
Savassi, O.N., 1998. Direct estimation of the degree of entrainment and the froth
recovery of attached particles in industrial otation cells. PhD thesis. University
of Queensland, Brisbane.
Subasinghe, N., Albijanic, B., 2014. Inuence of the propagation of three phase
contact line on otation recovery. Miner. Eng. 57, 4349.
Vianna, S., 2004. The effect of particle size, collector coverage and liberation on the
oatability of galena particles in an ore. PhD thesis. The University of
Queensland.
Ye, Y., Khandrika, S.M., Miller, J.D., 1989. Induction-time measurements at a particle
bed. Int. J. Miner. Process. 25 (34), 221240.
Yoon, R.H., Yordan, J.L., 1991. Induction time measurements for the quartzamine
otation system. J. Colloid Interface Sci. 141 (2), 374383.