Minerals Engineering 24 (2011) 50–57

Contents lists available at ScienceDirect

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

The dependency of the critical contact angle for flotation on particle

size – Modelling the limits of fine particle flotation
Daniel Chipfunhu a, Massimiliano Zanin a,⇑, Stephen Grano b
a
Ian Wark Research Institute, University of South Australia, The ARC Special Research Centre for Particle and Material Interfaces, Mawson Lakes Campus,
Adelaide, South Australia 5095, Australia
b
Institute for Mineral and Energy Resources, The University of Adelaide, South Australia 5005, Australia

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

Article history: The flotation behaviour of fine particles is studied in this work. Fine methylated quartz particles within
Received 13 July 2010 the size range from 0.2 to 50 lm, and with varying contact angles, were floated in a mechanical flotation
Accepted 23 September 2010 cell. Results indicate that particles of a given size need to possess a minimum critical contact angle, which
increases in value as particle size decreases, for flotation to be initiated. As a consequence, a non-floating
component exists within a given size fraction. This is interpreted as a fraction consisting of particles
Keywords: below the critical contact angle for flotation for that size. The critical contact angle for flotation is
Froth flotation
explained in terms of the existence of an energy barrier for bubble–particle attachment. The flotation
Flotation modelling
Contact angle
results are interpreted by means of Scheludko et al. (1976) and Drelich and Miller (1992) models for
Fine particles flotation the floatability of fine particles. The experimental data compared very well with calculations using the
Drelich and Miller equation, allowing extension to the prediction of the critical contact angle for flotation
down to particle sizes well below the previous limits investigated, bridging the gap existing in the
literature.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction and particles. Decreasing particle size results in a decrease in colli-

In flotation, fine particles show lower flotation rate, resulting in
low flotation recovery. The recovery of particles by air bubbles dur-
ing flotation occurs after three consecutive sub-processes of
collision, attachment and stability have taken place to form a bub-
ble–particle aggregate. The whole process, termed collection, may
be expressed as (Dai et al., 1998):
E ¼ Ec  Ea  Es
where Ec, Ea and Es are the collision, attachment and stability effi-
ciencies respectively. Collision is mainly controlled by hydrody-
namics (e.g., bubble rise velocity, bubble size) in the flotation cell,
while attachment is dominated by interfacial behaviour (surface
forces) between bubble and particle. For fine particles, which pos-
sess low inertia, the stability efficiency is essentially unity (Dai
et al., 1998). Particle recovery by flotation is sensitive to both parti-
cle size and contact angle (Crawford and Ralston, 1988), with very
fine particles (<10 lm) and coarse particles (>100 lm) floating
poorly, but for different reasons. Coarse particles have low flotation
rate due to detachment problems associated with disruptive forces
in the flotation cell (Pyke et al., 2003), whereas fine particles’ poor
floatability emanates from low collision efficiency between bubbles
⇑ Corresponding author. Tel.: +61 8 8302 3263; fax: +61 8 8302 3683.
E-mail address: massimiliano.zanin@unisa.edu.au (M. Zanin).
sion efficiency as particles are not able to deviate from fluid stream-
lines to collide with bubbles (Weber and Paddock, 1983; Yoon and
Luttrell, 1989). The bubble–particle collision frequency is controlled
by cell hydrodynamics and kinetics, making the problem of fine par-
ticle flotation partly a kinetic one. This implies that particles should
eventually float if given sufficient residence time in the flotation
environment. However, flotation of fine particles may be hindered
by the existence of an energy barrier that prevents successful bub-
ð1Þ
ble–particle attachment after collision, thus making the problem of
fine particle floatability also a thermodynamic one. Put differently,
fine particles may possess insufficient kinetic energy to displace the
intervening liquid layer between the colliding particle and bubble
(Hewitt et al., 1995).
The energy barrier manifests as a critical contact angle, which
represents the amount of activation energy that must be overcome
before bubble–particle attachment can occur. Theoretical consider-
ations and experimental data published to date support the exis-
tence of a critical contact angle below which flotation does not
occur (Crawford, 1986; Gontijo et al., 2007; Miettinen, 2007), but
validation of the experimental data for fine particles with theoret-
ical predictions has not yielded satisfactory agreement. This dis-
parity is discussed further below.
In their theoretical analysis of the floatability of fine particles,
Scheludko et al. (1976) proposed that the kinetic energy of
particles must be larger than the energy needed to overcome the

0892-6875/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2010.09.020
 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57
resistance to the formation and expansion of the three phase con-
tact line for successful particle–bubble attachment to occur. This
energy must be supplied by the kinetic energy of the particle. At
equilibrium, the kinetic energy balances the energy needed to form
and expand the three phase line of contact, from which a minimum
particle diameter, Dmin,K, for flotation maybe estimated (Scheludko
et al., 1976):
" #13
3c2slv
Dmin;K ¼ 2
V 2t Dqclv ð1  cos hÞ
where h is the receding contact angle, cslv is the line tension (which
is the reversible work which must be done to isothermally expand
the unit length of the three phase contact line), clv is the liquid–va-
pour surface tension, Dq is the difference between the particle and
fluid densities and Vt is the bubble–particle approach velocity. To
date, this theoretical model has not been satisfactorily validated
with experimental data due to the uncertainty in the value of the
line tension and the lack of data for very fine particles. There is
no consensus on the line tension value used in Eq. (2) (Mingins
and Scheludko, 1979; Drelich, 1996; Amirfazli and Neumann,
2004; Lin et al., 1993; Li et al., 1992), since the experimental values
of line tension in the literature are quite varied, ranging from 1012
to 105 N, while theoretically determined values range from 1012
to 1010 N (Drelich, 1996). A recent review of the theoretical and
experimental data on line tension concluded that sometimes the
values of line tension are inappropriately compared for dissimilar
systems which may not be reasonably expected to have the same
value even from a theoretical perspective (Amirfazli and Neumann,
2004). The values of line tension quoted by Scheludko et al. (1976)
were estimated for particles of a given size (about 2 lm) and con-
tact angle (between 20° and 40°), and thus cannot be expected to
fit experimental data for substantially larger particles (about
10 lm or larger).
Drelich and Miller (1992) revised Scheludko et al. (1976) ap-
proach and proposed that a pseudo-line tension, which takes into
account surface heterogeneities and roughness, will give a better
correlation with experimental data compared to the line tension.
Using the same theoretical basis used to derive Eq. (2), Drelich
and Miller (1992) recognised that:
cslv ¼ clv  ð1  cos hÞ  rc
where rc is the critical radius of wetting, i.e. the bubble drop radius
below which there is no effective attachment between the solid and
dispersed phases. Substituting Eq. (3) into Eq. (2), an alternative
expression for the minimum particle size, Dmin,K, was proposed:
" #13
3r 2c clv ð1  cos hÞ
Dmin;K ¼ 2
V 2t Dq
Drelich and Miller (1992) determined experimental values for
the critical radius of wetting, rc, as a function of contact angle, h.
The rc values were estimated by extrapolating a plot of h versus
r to h = 0, for methylated quartz plates in distilled water at pH
5.6.
In this paper, the flotation behaviour of fine methylated quartz
of different particle size and contact angle was studied. The exis-
tence of a critical contact angle for flotation, which is particle size
dependent, is shown. Further, to investigate contact angle hetero-
geneity within individual classes of particles, the contact angle of
feed and tails from selected tests was measured by means of film
flotation, which uses the critical surface tension of wetting
(Zisman, 1964). Other methods, such as Washburn technique
(Washburn, 1921), were considered not reliable for fine particles
(<10 lm), because fine particles tend to form aggregates when
dry, biasing the measurement. The critical surface tension of wet-
51
ting was determined in film flotation at 0% floating (Yarar and Kao-
ma, 1984), as opposed to 50% floating which is also suggested in
the literature (Fuerstenau et al., 1991). This is because for very fine
quartz particles the first process to occur when particles are placed
on the liquid surface is capillary penetration, resulting in particles
sinking as aggregates. At 0% floating (or 100% sinking) it is ensured
that both aggregates (which sink due to gravity at higher surface
tension values compared to individual particles) and individual
particles sink, eliminating any ambiguity in the value of the surface
ð2Þ
tension of wetting. In essence, this approach gives the minimum
value of the surface tension of wetting from a range of possible val-
ues. Film flotation experiments were carried out on samples of feed
and tailings, and differences in contact angle interpreted to be a re-
sult of segregation in flotation of particles having different size and
contact angle values.
Finally, the flotation results are validated against the Scheludko
model (Scheludko et al., 1976) and its variation (Drelich and Miller,
1992), based on theoretical considerations in regards to the limits
of fine particle flotation.
2. Materials and methods
All glassware was cleaned by immersion in concentrated so-
dium hydroxide of pH above 13 and sonication for at least
30 min to remove any organic contamination. The glassware was
then thoroughly rinsed with de-ionised water to remove all the so-
dium hydroxide and finally dried overnight in a clean oven at
110 °C. All chemicals used in this work were of analytical reagent
quality and were used without further purification.
2.1. Materials preparation
Microcrystalline quartz of particle size less that 10 lm was ob-
tained from Sigma Aldrich. These particles are designated as Sigma
quartz. Other quartz size fractions were obtained by wet grinding
lumpy crystalline quartz in a laboratory stainless steel ball mill
and wet sieving at 38 lm. The 38 lm size fraction was passed
through a 1” cyclone (precyclone) with the overflow discarded
and the underflow classified in a Warman cyclosizer that split
the particles into six different size fractions (C1, C2, C3, C4, C5,
ð3Þ and C6). Selected size fractions (C3, C5, C6 and Sigma) were subse-
quently cleaned separately to remove any organic and inorganic
contaminants (such as iron from the grinding media) before subse-
quent manipulation to alter the contact angle values.
2.2. Cleaning quartz particles
ð4Þ Hot aqua regia and piranha solution were used to clean each of
the selected quartz size fractions of any surface impurities. The
reaction time was varied from 6 to10 h depending on the particle
size and the reagent used. The acid was washed out by rinsing
the particles with de-ionised water until the pH became neutral.
A centrifuge was used to accelerate particle settling for the finest
size fractions (Sigma and C6). The particles were dried overnight
in a clean oven at 110 °C in a covered clean beaker and subse-
quently stored in clean and sealable glass containers.
2.3. Methylation of silica particles
The contact angle of clean silica particles was manipulated to
different degrees of hydrophobicity by using trimethylchlorosilane
(TMCS) (Blake and Ralston, 1985). Because TMCS reacts readily
with water in the atmosphere, the methylation procedure was car-
ried out in a glove box under dry nitrogen. A known mass of
cleaned quartz for each size fraction was placed in a 1 l laboratory
52 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57

glass bottle and heated in an oven overnight at 110 °C to remove
any physisorbed water. While still hot, the quartz was transferred
to the glove box and allowed to cool, a measured amount of cyclo-
hexane added, and the glass bottle sealed. The sealed bottle was
sonicated for 30 min in a bath in order to disperse the quartz par-
ticles before methylation. A known amount of TMCS was then
added from a precision microsyringe under constant stirring with
a magnetic stirrer. To obtain different contact angles for the quartz,
different concentrations of TMCS and reaction times were used. A
small amount of particles was taken for contact angle measure-
ment using film flotation, while the remainder of the particles were
stored in a clean, sealable container and placed in a desiccator.
2.4. Bubble and particle size distribution measurement
Bubble size distribution in the flotation experiments (Fig. 1a)
was measured using a bubble viewer connected to digital photo-
graphic equipment (Hernandez-Aguilar et al., 2004). The Sauter
mean bubble diameter for 50 ppm MIBC in the mechanical flota-
tion cell was measured to be 0.42 mm. The particle size distribu-
tion for the clean quartz particles, flotation feed (Fig. 1b), tailings
and concentrates was measured using a Mastersizer 2000 instru-
ment from Malvern Particle Sizing Systems. The system measures
particle size by applying a focussed laser beam through the parti-
surface tension. The critical surface tension of wetting was deter-
mined at 0% floating. Once the critical surface tension of wetting
was measured, the corresponding contact angle value was deter-
mined using a calibration curve developed by Miettinen (2007)
(Fig. 3). The calibration curve was developed by methylating five
sets of glass slides to different advancing water contact angles
and then measuring the contact angle on the glass plates as a func-
tion of surface tension of methanol–water mixtures (Miettinen,
2007). The contact angle value is the maximum contact angle value
from a range of possible values as the surface tension of wetting is
the minimum value for the population of particles.
2.6. Flotation procedure
Flotation of the different size fractions was carried out in a
300 cm3 top driven mechanical flotation cell, using demineralised
water, at pH 6 and with potassium nitrate electrolyte concentra-
tion of 1  103 M. The pH was controlled with hydrochloric acid
and potassium hydroxide. The air flow rate was maintained at
150 cm3/min, corresponding to a superficial gas velocity (Jg) of
90

Contact Angle (o)

80
cles, and Mie scattering model is then used to determine the parti-
cle size distribution as equivalent sphere diameters. Scanning 70
Electron Microscope (SEM) images for the particles used in this
study (Fig. 2) show that particles are angular and of irregular 60
shape, and not spherical.
50
2.5. Measurement of particle contact angle 40
20 25 30 35
A small quantity of particles (30–50 mg) was placed on the sur-
Critical surface tension, mN/m
face of methanol–water solutions of different concentration and
controlled surface tension, and the amount of particles that either Fig. 3. Calibration curve for advancing water contact angle of silica particles in
sink or remain on the surface determined as a function of the liquid water versus their critical surface tension used for film flotation, Miettinen (2007).
Number distribution (%)

4 20
(a) (b)
3 15
Frequency

2 10

1 5

0 0
0 0.2 0.4 0.6 0.8 1 0.1 1 10 100
Bubble Size, µm Particle diameter, µm
Fig. 1. (a) Bubble size (Sauter mean bubble diameter = 0.42 mm) in the mechanical flotation cell and (b) particle size distribution for Sigma quartz (), d50 = 3 lm; C6 quartz
(d), d50 = 10 lm; C5 quartz (N), d50 = 16 lm and C3 quartz, (j), d50 = 33 lm.

Fig. 2. SEM images of particles used in this study: Sigma quartz (left) and C6 quartz (right). Other size fractions had similar particle shape to the C6 sample.
 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57 53

approximately 0.5 cm/s. Methyl isobutyl carbinol (MIBC) was used
as a frother, at a concentration of 50 ppm. Methylated quartz
(about 1.0 g) of known size fraction and contact angle, was placed
in a 250 cm3 beaker, and about 150 cm3 of electrolyte solution
added. The pulp was stirred for 5 min to properly disperse the par-
ticles, and transferred to the flotation cell. The pulp in the cell was
made up to about 300 ml, the pH was adjusted and frother added
(2 min conditioning time). Air was turned on and flotation carried
out, at an impeller speed of 3000 rpm, for a total time of 8 min,
with concentrate collected at 1, 3, 5 and 8 min intervals. For the
finest size fractions (Sigma and C6) the preparation of the flotation
feed also included sonication for 10 min, to avoid the formation of
aggregates.
Since keeping methylated quartz in aqueous conditions for
more than approximately 25 min results in loss of hydrophobicity
and a decrease in flotation recovery (Blake and Ralston, 1985), the
conditioning time was kept as short as possible. The concentrates
and tailings were dried in an oven at 110 °C and weighed to deter-
mine the flotation recovery. Selected samples of feed, concentrate
and tailings, prior to oven drying, were also analysed to determine
particle size distribution and contact angle.
2.7. Analysis of error and reproducibility
The error on the flotation recovery results was estimated by car-
rying out each flotation test in duplicate. Where the two repeated
experiments gave recoveries differing by more than 5%, a third
experiment was carried out. For contact angle measurements, the
critical surface tension was determined using methanol-water
solutions with an approximate liquid/vapour surface tension dif-
ference of 2 mN/m, giving an error of approximately 1 mN/m sur-
face tension and an equivalent error of 4° in the contact angle
measurement.
2.8. Subtraction of entrainment
The recovery of fine particles is both a function of true recovery
and mechanical carryover, or entrainment. In this work the ap-
proach by Ross (1991) is used to estimate recovery by entrainment.
The rate of entrainment in the ith size fraction, Ei(t), at any given
time is given by Ross (1991,1990):
W i ðtÞ  C mi ðtÞ
Ei ðtÞ ¼ X i ðtÞ  ð5Þ
C W ðtÞ
where Wi(t) is the rate of recovery of water (g/min), Cmi(t) and Cw(t)
are the concentrations of the particular mineral species and water
(g/l of pulp) at time t, respectively. Xi(t) is a factor which character-
ises the degree of entrainment and is particle size dependent. In this
work, the entrainment factor is the gradient of the curves (Fig. 4)
obtained by floating hydrophilic quartz particles and plotting parti-
cle recovery against water recovery.

3. Results
45
Recovery, %

3.1. Flotation recovery as a function of particle size and contact angle

30

Flotation recovery was calculated as a function of time for the

15 four different particle size fractions. For each size fraction the flo-
tation recovery was also determined as a function of contact angle.
0 Flotation recovery was directly proportional to both nominal parti-
0 20 40 60 cle size and contact angle, across all size fractions (Fig. 5).
Water Recovery, % In Fig. 6 the fraction of particles remaining in the flotation cell is
reported versus time (log scale) for all size fractions and for contact
Fig. 4. Solids recovery versus water recovery for hydrophilic Sigma (), C6 (d), C5
angles of 60° and 88°. Ideally, for equisized and surface homoge-
(N) and C3 (j) quartz (300 ml mechanical cell pH 6, Jg = 0.5 cm/s, 0.001 M KNO3,
8 min flotation time, impeller speed 3000 rpm, MIBC frother concentration neous particles, the flotation rate constant would be the same for
50 ppm). all particles, and a plot of log R (where R is recovery) against time

100
(a) (c)
80

60

40
Recovery, %

20

0
(b) (d)
80

60

40

20

0
0 2 4 6 8 0 2 4 6 8 10
Time, Minutes Time, Minutes
Fig. 5. Flotation recovery versus time for quartz (300 ml mechanical cell pH 6, Jg = 0.5 cm/s, 103 M KNO3 flotation time 8 min, impeller speed 3000 rpm, MIBC frother
concentration 50 ppm) (a) Sigma quartz: CA 51° (N), CA 58° (D), CA 71° (d), CA 79° (), and CA 88° (x), (b) C6 Quartz: CA 35° (e), CA 42° (j), CA 48° (N), CA 60° (D), CA 73° (d)
and CA 88° (x). (c) C5 Quartz: CA 33° (e), CA 38° (s), CA 42° (j), CA 52° (N), CA 60° (D), CA 71° (d), CA 77° () and CA 88° (x). (d) C3 Quartz: CA 21° (+), CA 26° (h), CA 30° (;),
CA 42° (j), CA 60° (D), CA 71° (d) and CA 88° (x). The points are experimental values and the lines are the model fit for R = Rmax(1  ekt).
54 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57

Time, min.
0 2 4 6 8 0 2 4 6 8 10
1
(a) CA = 88 o (b) CA = 60 o

Fraction Remaining
Sigma

Sigma
C6 C6
0.1
C5
C5 C3

C3

0.01

Fig. 6. Fraction remaining, log scale, versus time for Sigma (j), C6 (), C5 (N) and C3 (d) quartz for contact angle 88° (a) and contact angle 60° (b). The points are
experimental values and lines are model fit for R = Rmax(1  ekt).

(t), should give a straight line (Imaizumi and Inoue, 1965) to low
fraction remaining values (<0.05). However, the curves obtained
for the different classes of particles deviate significantly from
straight lines at high fraction remaining values, suggesting the sig-
nificant presence of at least two floating components. Therefore, a
floating component (Rmax) and a non-floating component (100 
Rmax) have been assumed, and the experimental data fitted to the
first order rate equation:
separately on feed and tailings of selected samples. Fig. 8 reports
the cumulative percentage of floating quartz particles versus the
surface tension of wetting.
Consistently, for all samples the maximum contact angle of the
tailings was found to be lower than the maximum contact angle of
the feed (Table 2), suggesting contact angle heterogeneity, as dis-
cussed further below.

R ¼ Rmax  ð1  ekt Þ ð6Þ 3.3. Particle size heterogeneity

The described procedure gave very close agreement between
the experimental data points and the fitted recovery value
(Fig. 6). The fitting parameters (Rmax, k) are reported in Table 1. It
is apparent that there is a larger proportion of unrecovered parti-
cles, non-floating component (100  Rmax), at contact angle 60°
compared with contact angle 88° and that this non-floating com-
ponent increases with decreasing particle size and contact angle.
Even at very high contact angles (e.g., 88°), where recovery is ex-
pected to be virtually 100%, there are some particles that are not
recovered, meaning that either they required greater residence
time in the flotation environment or they were below the critical
contact angle for flotation.
Fig. 7 shows the maximum recovery, Rmax, as a function of par-
ticle contact angle and size. The results point to the existence of a
critical contact angle (hcrit) below which flotation does not occur, in
agreement with previous studies (Crawford and Ralston, 1988).
This critical contact angle is obtained by extrapolating the tangent
at the inflexion point to 0% recovery (Gontijo et al., 2007). The error
in determining hcrit is estimated to be equal to the error in deter-
mining the contact angle (±4°). The critical contact angle for the
range of particle sizes investigated increases with decreasing par-
ticle size (Table 1).
The particle size distributions of feed, concentrates and tailings
of some of the size fractions are shown in Fig. 9. A difference was
found in particle size distribution, with the concentrates having
on overall greater particle size than the tailings. For example, for
a sample of Sigma particles having contact angle 51°, the average
particle size of the feed was 3 lm, while the average particle size
of concentrates and tailings was about 5 lm and 1 lm, respec-
tively (Fig. 9). Similar trends were also observed for other size
fractions.
4. Modelling the limits of fine particles flotation
The system studied by Drelich and Miller (1992) to determine
the values for rc is similar to the flotation system in this study
(i.e., quartz-TMCS), allowing direct comparisons and the use of rc
100
80

60
Rmax, %

3.2. Film flotation and heterogeneity of contact angle values

Heterogeneity of contact angle in each class of particles was 40

investigated by means of film flotation, which was carried out
20
Table 1
Experimental critical contact angles derived from Fig. 7.
0
Designation Critical contact Particle size, CA = 88° CA = 60° 0 20 40 60 80 100
angle, hcrit (°) d50 (lm)
Rmax k Rmax k Contact Angle (o)
(%) (%)
Sigma 55 3 81 0.38 65 0.30 Fig. 7. Rmax versus contact angle for Sigma (), C6 (d), C5 (N) and C3 (j) quartz
C6 37 10 86 0.56 82 0.50 (300 ml mechanical cell pH 6, Jg = 0.5 cm/s, 1 mM KNO3 flotation time 8 min,
C5 30 16 97 1.42 93 1.51 impeller speed 3000 rpm, MIBC frother concentration 50 ppm). The critical contact
C3 20 33 98 2.34 95 2.06 angle for flotation, hcrit, is determined by extrapolation of the curves to 0% recovery
from the tangent at the inflexion point.
 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57 55

100
(a) (b)
80

% Floating
60

40

20

0
20 40 60 0 20 40 60 80
Surface Tension, mN/m
Fig. 8. The cumulative % of floating quartz particles versus wetting surface tension as obtained from film flotation with aqueous methanol solutions (a) Sigma quartz CA 51°
() and CA 88° (j); (b) C5 quartz CA 58° () and CA 77° (j).

Table 2
Contact angles (CA) of feed and tails after flotation of selected samples. Film flotation Table 3
was used to ascertain the contact angle. Critical bubble drop radius, rc, versus contact angle (CA): A comparison between
model values (Eq. (7)) and measured values from Drelich and Miller (1992).
Designation Particle size, Critical CA, Max. CA Max. CA
d50 (lm) hcrit (°) feed (°) tailings (°) CA of quartz plates (°) Measured rc (lm) Model rc (lm)
C3 33 20 32 27 36 0.85 1.05
C5 16 30 58 36 46 0.30 0.32
C6 10 37 41 32 48 0.29 0.25
C6 10 37 88 48 60 0.09 0.06
Sigma 3 55 88 48 72 0.01 0.01

values from the former study. In the current study, an exponential
expression is used to determine the values of rc. An equation that
best fits the experimental relationship between rc and contact an-
gle was determined by the least squares method (Eq. (7)). Experi-
mental rc values (Drelich and Miller, 1992) and those predicted by
Eq. (7) are shown in Table 3, giving a value of R2 = 0.96 for the
goodness of fit.
r c ¼ 79:68  e0:12h
Ultimately, Eq. (7) estimates and extrapolates the value of rc be-
low the contact angles for which it was determined by Drelich and
Miller (1992).
Eq. (7) was used to calculate the values of rc for the critical con-
tact angles in Table 1. The minimum particle size for flotation,
Dmin,K, was determined by substituting the calculated rc values into
Eq. (4). The other parameters used to calculate the minimum par-
ticle size in Eq. (4) were the liquid vapour surface tension,
clv = 71.6 mN/ml, the liquid density, ql = 1000 kg/m3, quartz den-
sity, qs = 2650 kg/m3, and the bubble–particle approach velocity,
Vt = 0.3 m/s. It is noted that the value of the line tension, or more
precisely the pseudo-line tension, in the approach by Drelich and
Miller (1992) is calculated as a function of particle contact angle
ð7Þ
(Eq. (3)), and shows a decreasing profile as contact angle increases
(Table 4).
The model fit to the experimental data is also shown in Fig. 10,
as well as the experimental data by Crawford (1986). On the plot,
the region of critical contact angle/particle size values calculated
from Scheludko et al. (1976) is also reported.

100
Sigma C5 CA 58
o
C6 CA42
o
o
80 CA 51

60
40
20
Cumulative Volume %

0 o
Sigma CA 88 o o
C5 CA 88 C6 CA88
80

60

40

20

0
0 3 6 9 0 10 20 30 40 0 10 20 30
Particle size, µm
Fig. 9. Cumulative particle size distribution of feed (N), concentrates (j) and tailings (h) for Sigma, C6 and C5 quartz.
56 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57

Table 4
Theoretical and experimentally determined Dmin,K values for the fine particles in this
study. The model to determine Dmin,K is based on the Drelich and Miller (1992) 100

Particle size, µm
approach.
Particles
hcrit Exp. Dmin,K Model rc Pseudo L.T., cslv Model Dmin,K float
10
(°) (lm) (lm) (N) (lm)
(b)
20 33 3.98 3.2E08 31
30 16 1.52 2.1E08 18 1
37 10 0.70 1.4E08 12 (a)
Particles
55 3 0.14 3.4E09 4
don’t float
0.1
0 15 30 45 60 75 90
5. Discussion
Contact Angle, (o )
5.1. Flotation recovery as a function of particle size and contact angle Fig. 10. Comparison of the model based on Drelich and Miller (1992) (solid line)
with the experimental values obtained in this study (N), Crawford (1986) (D) and
The experimental results show that the flotation recovery of Miettinen (2007) (x). Upper and lower limits of the Scheludko et al. (1976) model
(dotted lines) are also reported where line (a) corresponds to cslv = 2.8  1010 N and
fine particles is a function of both particle contact angle (hydro-
line (b) corresponds to cslv = 5.6  1010 N. For both lines, Vt = 0.3 ms1,
phobicity) and size. More precisely, the recovery increases with clv = 71.6 mN/m and Dq = 1650 kg/m3.
an increase in contact angle and decreases with a decrease in
particle size, in agreement with previous studies (Crawford and fine Sigma particles, aggregation may also have contributed to the
Ralston, 1988). sinking of particles due to gravity, at higher surface tension values.
The increasing recovery with increasing contact angle, or parti- This may have magnified the apparent contact angle heterogeneity.
cle hydrophobicity (Fig. 5), can be explained by the fact that the With respect to particle size heterogeneity, the greater average
attachment efficiency is increased. Kinetically, for successful bub- particle size of the concentrate with respect to the feed and tailings
ble–particle attachment to occur, the induction time (total time (Fig. 9) is because larger particles have a lower critical contact an-
needed for the thinning and rupturing of the thin liquid film sepa- gle of flotation (and possibly higher contact angle) than smaller
rating particle from bubble) must be shorter than the time it takes particles in the same feed sample.
a particle to slide on the bubble surface, also called the sliding The existence of a non-floating component, even in a narrow
time. Increasing surface hydrophobicity results in a more rapid size/contact angle class of particles, can be explained in terms of
rupture of the intervening liquid film between bubble and particle heterogeneity of both contact angle and particle size as well as
(Hewitt et al., 1993), i.e. shorter induction time. From a thermody- the existence of a critical contact angle. A qualitative description
namic perspective, as particle hydrophobicity increases, the inter- of the phenomenon can be given assuming a statistical distribution
action potential energy between the particle and the bubble of contact angles in a class of particles around a nominal value. In
becomes more attractive thus enhancing the attachment efficiency this hypothesis (Fig. 11), the proportion of particles with contact
and ultimately flotation recovery. angle above the critical value, hcrit, is equal to the maximum recov-
Further, fine particles in this work floated at a lower rate than ery Rmax in the flotation test (it is assumed that the total flotation
coarser particles having similar contact angle values (Fig. 5 and time is long enough to ensure complete recovery of the floatable
Table 1), pointing towards the existence of a critical contact angle fraction). Inversely, 100  Rmax is the fraction remaining in the flo-
(hcrit), which is particle size dependent (Fig. 7). This becomes espe- tation cell at long flotation times. For the same particle size range,
cially important in sulphide mineral systems where fine particles and therefore the same critical contact angle, hcrit, the proportion of
are believed not to float well because they may have lower contact recovered particles (Rmax) increases with the increase in contact
angle due to surface oxidation compared to relatively coarser par- angle. This is described in Fig. 11, in which the different maximum
ticles, even though hydrophobised under the same conditions. recovery for the 77° and 58° contact angle samples is ascribed to
the different mass of particles below the critical contact angle for
flotation in the two samples.
5.2. Heterogeneity of contact angle and particle size

In flotation tests on narrow size/contact angle classes of parti- 77o

cles, the maximum contact angle of the tailings was found to be
lower than the maximum contact angle of the feed (Table 2), sug-
gesting heterogeneity in the feed sample. Further, the contact angle
100 -Rmax
of tailings was very close to the previously determined critical va-
lue, giving some confidence to the latter. The fact that the contact
angles of the tailings are slightly above the critical contact angle
suggest the presence of a small amount of weakly hydrophobic par- 58o
ticles in the tailings possibly for reasons related to rate processes.
100 -Rmax
Further indication of heterogeneity of contact angle was appar-
ent in the trends observed for the film flotation experiments, which
also suggest a spread of contact angle values within a sample
(Fig. 8). For very high contact angle particles there is a very narrow
surface tension range over which particles change from zero wet-
ting to complete wetting. As the hydrophobicity of the particles de-
creases, i.e. at low contact angle (e.g. Sigma quartz with contact Contact Angle, ( o )
angle of 51°), the transition gradient becomes shallower and the
Fig. 11. Relationship between critical contact angle for flotation and Rmax for
majority of particles are wet even with water (surface tension particles of contact angle 58° and 77°, assuming statistical distribution of contact
72.8 mN/m). This indicates that the particles are not homogeneous angles in the sample. It is assumed that particles with contact angle below hcrit do
with respect to contact angle (Fuerstenau et al., 1991). For the very not float.
 D. Chipfunhu et al. / Minerals Engineering 24 (2011) 50–57
5.3. Limits for fine particles flotation
This study also provides insights on the limits for fine particles
flotation. The results presented support the hypothesis (Scheludko
et al., 1976) that fine particle floatability is limited by the existence
of an energy barrier which is linked to the line tension. The mod-
elling procedure presented in this work is based on the pseudo-line
tension values proposed by Drelich and Miller (1992) and linking
the line tension to particle contact angle and critical radius of
wetting.
There is good agreement (Fig. 10) between the experimental data
and predictions obtained using the model. On the contrary, calcula-
tions from Scheludko et al. (1976) apparently underestimate the
minimum particle size for flotation. It is the line tension values
which are very different in the two approaches. It has to be stressed
that values of cslv were calculated by Scheludko et al. (1976) under
somewhat arbitrary assumptions. In particular, a minimum particle
size for flotation, Dmin,K = 2 lm was assumed for particles between
20° and 40° contact angle, with a bubble approach velocity of
0.2 m/s. Scheludko et al. (1976) recognised the uncertainty of line
tension values available in the literature at the time. Further, there
is good agreement between the predictions of this model, the exper-
imental data in this study and the experimental data of Crawford
(1986) in the fine particle ranges (<38 lm) (Fig. 10).
The procedure seems to provide accurate prediction of the mini-
mum particle size for flotation over a wide range of particle contact
angle values. The variation in the line tension values with particle
size brings about the concept of the pseudo-line tension, which takes
into account surface imperfections and roughness and is better able
to fit experimental data (Drelich and Miller, 1992). The experimental
data presented in this work, and model calculations, allow extension
to the prediction of the critical contact angle for flotation to finer par-
ticle size ranges from the previous limits in Crawford (1986), bridg-
ing the existing gap in the literature.
6. Conclusions
The flotation of fine particles is limited not only by collision effi-
ciency (rate process), but also by the lack of kinetic energy to rup-
ture the intervening liquid film between particle and bubble, and
subsequent expansion of the three phase line of contact. A non-
floatable fraction of a flotation feed has been established and ema-
nates from particles that are below the critical contact angle re-
quired for flotation. The critical contact angle for flotation
depends on particle size, and increases as the particle size de-
creases. Recoveries obtained in this work suggest that even though
particle size is limiting the collision efficiency, if fine particles have
adequate hydrophobicity they may attach to bubbles. In other
words, flotation of fine particles is possible if a critical contact an-
gle value is attained.
The theoretical calculations for limits of flotation for fine parti-
cles developed by Scheludko et al. (1976) do not give good agree-
ment with the experimental data in this work. On the contrary,
there is very good agreement between the experimental data and
a model which has been developed following Drelich and Miller
(1992) approach for the calculation of the line tension. Drelich
and Miller (1992) model takes into account the correlation
between surface hydrophobicity (contact angle) and surface rough-
ness of the particles by using a pseudo-line tension which better
represents flotation conditions of the particles used in the current
study.
In this work, it was possible to extend the prediction of the crit-
ical contact angle for flotation to finer particle size ranges from the
57
previous limits in Crawford (1986), down to values below approx-
imately 2 lm, bridging the existing gap in the literature.
Acknowledgements
The authors would like to acknowledge the sponsors of the
AMIRA International P260E Project and the Australian Research
Council (ARC) for the financial support to this study. D.C. thanks
the Ian Wark Research Institute and AMIRA for his scholarship
support.
References
Amirfazli, A., Neumann, A.W., 2004. Status of the three-phase line tension: a review.
Advances in Colloid and Interface Science 110, 121–141.
Blake, P., Ralston, J., 1985. Controlled methylation of quartz particles. Colloids and
Surfaces 15, 101–118.
Crawford, R.J., 1986. Particle Size, HYDROPHOBICITY and Flotation Response.
MAppSci Thesis, Department of Applied Chemistry. Swinburne Institute of
Technology.
Crawford, R., Ralston, J., 1988. The influence of particle size and contact angle in
mineral flotation. International Journal of Mineral Processing 23.
Dai, Z., Dukhin, S., Fornasiero, D., Ralston, J., 1998. The inertial hydrodynamic
interaction of particles and rising bubbles with mobile surfaces. Journal of
Colloid and Interface Science 197, 275–292.
Drelich, J., 1996. The significance and magnitude of the line tension in three phase
(solid–liquid-fluid) systems. Colloids and Surfaces A: Physicochemical and
Engineering Aspects 116, 43–54.
Drelich, J., Miller, J.D., 1992. The effect of surface heterogeneity on pseudo-line
tension and the flotation limit of fine particles. Colloids and Surfaces 69, 35–43.
Fuerstenau, D.W., Diao, J., Williams, M.C., 1991. Characterization of the wettability
of solid particles by film flotation 1. Experimental investigation. Colloids and
Surfaces 60, 127–144.
Gontijo, C.D.F., Fornasiero, D., Ralston, J., 2007. The limits of fine and coarse particle
flotation. Canadian Journal of Chemical Engineering 85, 739–747.
Hernandez-Aguilar, J.R., Coleman, R.G., Gomez, C.O., Finch, J.A., 2004. A comparison
between capillary and imaging techniques for sizing bubbles in flotation
systems. Minerals Engineering 17, 53–61.
Hewitt, D., Fornasiero, D., Ralston, J., 1993. Aqueous film drainage at quartz/water/
air interface. Journal of Chemical Society of the Faraday Transactions 89, 817–
822.
Hewitt, D., Fornasiero, D., Ralston, J., 1995. Bubble–particle attachment. Chemical
Society Faraday Transactions 91, 1997–2001.
Imaizumi, T., Inoue, T., 1965. Kinetic consideration of froth flotation. In: Roberts, A.
(Ed.), Sixth International Mineral Processing Congress. Cannes.
Li, D., Cheng, P., Neumann, A.W., 1992. Contact angle measurement by
axisymmetric drop shape analysis (ADSA). Advances in Colloid and Interface
Science 39, 347–382.
Lin, F.Y.H., Li, D., Neumann, A.W., 1993. Effect of surface roughness on the
dependence of contact angles on drop size. Journal of Colloid and Interface
Science 159, 86–95.
Miettinen, T., 2007. Fine Particles Flotation. PhD Thesis. Mawson Lakes, University
of South Australia.
Mingins, J., Scheludko, A.D., 1979. Attachment of spherical particles to the surface of
a pendant drop and the tension of the wetting perimeter. Journal of Chemical
Society, Faraday Transactions.
Pyke, B., Fornasiero, D., Ralston, J., 2003. Bubble particle heterocoagulation under
turbulent conditions. Journal of Colloid and Interface Science 265, 141–151.
Ross, V.E., 1990. Flotation and entrainment of particles during batch flotation tests.
Minerals Engineering 3, 245–256.
Ross, V.E., 1991. Comparison of methods for evaluation of true flotation and
entrainment. Transactions of the Institution of Mining and Metallurgy Section C
– Mineral Processing and Extractive Metallurgy 100.
Scheludko, A., Toshev, B.V., Bojadjiev, D.T., 1976. Attachment of particles to a liquid
surface (capillary theory of flotation). Journal of the Chemical Society, Faraday
Transactions 1: Physical Chemistry in Condensed Phases 72, 2815–2828.
Washburn, E.W., 1921. Dynamics of capillary flow. Physical Review 17, 374–375.
Weber, M.E., Paddock, D., 1983. Interceptional and gravitational collision
efficiencies for single collectors at intermediate Reynolds numbers. Journal of
Colloid and Interface Science 94, 328–335.
Yarar, B., Kaoma, J., 1984. Estimation of the critical surface tension of wetting of
hydrophobic solids by flotation. Colloids and Surfaces 11, 429–436.
Yoon, R.H., Luttrell, H., 1989. The effect of bubble size on fine particle flotation.
Mineral Processing and Extractive Metallurgy Review 5, 101–122.
Zisman, W.A., 1964. Relation of the equilibrium contact angle to liquid and solid
constitution. In: GOULD, R.F. (Ed.), Contact Angle Wettability and Adhesion.
American Chemical Society, Los Angeles, California.