Int. J. Miner. Process.

76 (2005) 225 – 233

www.elsevier.com/locate/ijminpro

Effect of frothers on bubble size

Rodrigo A. Graua,T, Janusz S. Laskowskib, Kari Heiskanena
a
Helsinki University of Technology, Helsinki, Finland
b
British Columbia University, Vancouver, B.C., Canada
Received 21 May 2004; received in revised form 18 January 2005; accepted 20 January 2005

Abstract

The size of bubbles in flotation cells was measured using the UCT bubble size meter and the HUT bubble size analyser. While
both methods provided similar critical coalescence concentration (CCC) values for the three tested frothers, they also revealed
important differences. Using the HUT bubble size analyser, which does not impose any lower bubble size limit, it was possible to
find differences in bubble size measured for three frothers at concentrations exceeding the corresponding CCC values.
Frothers affect both the break-up phenomena in the impeller/stator zone in flotation cell, and bubble coalescence. While
longer chain polyglycols (e.g. DF-1012) are able to produce larger stable bubbles than the short chain DF-250 and DF-200
polyglycols, the CCC values for DF-1012 are lower than those for DF-250 and DF-200.
D 2005 Elsevier B.V. All rights reserved.

Keywords: flotation; flotation frothers; bubble size; critical coalescence concentration; surface tension

1. Introduction first-order flotation rate constant is linearly related to

Flotation frothers are used to facilitate air dis-
persion into fine bubbles, and to stabilize the froth.
According to Leja–Schulman’s penetration theory
(Leja and Schulman, 1954; Leja, 1956/57), frothers
accumulate preferentially at the water/gas interface
and interact with collector molecules adsorbed onto
solid particles in the particle-to-bubble collision and
attachment. The most dramatic effect of the frothers
on flotation come from their effect on bubble size. The
T Corresponding author. Tel.: +358 9 451 3829; fax: +358 9 451
2795.
E-mail address: Rodrigo.Grau@hut.fi (R.A. Grau).
0301-7516/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.minpro.2005.01.004
S b, the bubble surface area flux, which in turn is
inversely related to the bubble size.
The effect of frothers on bubble size has recently
been extensively studied (Cho and Laskowski,
2002a,b; Laskowski et al., 2003; Laskowski, 2003).
It was found that the effect of frothers on bubble size
results from their ability to prevent bubble coales-
cence. As shown schematically in Fig. 1, with
increasing frother concentration, the degree of bubble
coalescence decreases and at a particular concentra-
tion (critical coalescence concentration, CCC), the
coalescence of the bubbles is completely prevented.
As Fig. 1 explains, the CCC values are obtained by
finding the intersection of the horizontal asymptote to
226 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233

Coalescence determines Table 1

bubble size
The breakage process in the List of tested frothers
impeller neighbourhood
Bubble Size

controls the size of the
bubbles
dbccc
Zone I Zone II
CCC
Frother concentration
Fig. 1. Schematic diagram of the effect of the frother concentration
upon bubble size in a flotation cell.
the bubble size–concentration curves at higher con-
centrations with the sloped line approximating the
curve at lower concentrations. Further increase of the
frother concentration above the CCC value does not
affect the bubble size.
In the experiments discussed in this paper, the
bubble size was determined with the use of the UCT
(University of Cape Town) bubble size meter and the
HUT (Helsinki University of Technology) bubble size
analyser. In the UCT technique, bubbles are suctioned
into a capillary tube, where the length and velocity of
movement of each bubble are measured. The size of
the capillary tube imposes a lower bubble size limit. In
addition, the suction rate and the inner diameter of the
capillary tube seem to impose a maximum bubble size
limit. Grau and Heiskanen (2002) observed that
bubbles greater than 3 mm in diameter broke when
entering a capillary tube of 1 mm in inner diameter.
Since large number of the experimental data have been
generated using the UCT bubble size meter, and since
the critical coalescence concentration for many froth-
ers has been given based on such measurements, in this
Common name Chemical formula Molecular weight
DF-200 CH3(OC3H6)3OH 206.29
DF-250 CH3(OC3H6)4OH 264.37
DF-1012 CH3(OC3H6)6.3OH 397.95
Two series of tests were carried out: one at the
Helsinki University of Technology using Outokumpu
flotation cell, and the other one at the University of
British Columbia using an Open-Top Leeds lab
flotation cell. In the former the HUT bubble size
analyser was utilized, while the UCT bubble meter
was used in the latter.
The HUT bubble size analyser (HUT BSA)
consists basically of a sampler tube, a viewing
chamber and image acquisition system as shown in
Fig. 2. The sampler tube is partially immersed in the
liquid phase, below the froth–liquid interface. The
sampler tube is used to collect rising air bubbles,
which are drawn into the viewing chamber. The
function of the chamber is to expose the bubbles to a
monochrome CCD camera which is connected to a
personal computer, where the visual information is
stored and processed. The images captured during
sampling are automatically processed using commer-
cial image analysis software and in-house visual basic
user-interface. The operation of the HUT bubble size
analyser has been described in detail by Grau and
Heiskanen (2002, 2003). Bubbles captured by the
HUT BSA are shown in Fig. 3.
Viewing
Chamber
Progressive
Scan Camera

project, the bubble size was also measured using a

visual technique developed at the Helsinki University LED Backlight
of Technology (HUT method, Grau and Heiskanen,
Sampling Analog signal to
2002) to verify how reliable are the CCC values tube frame grabber
obtained with the use of the UCT bubble sizer.

2. Experimental

The tests were carried out with three commercial

Dow Frothers (Table 1). The acronym DF stands for
the trade name Dowfroth. Fig. 2. Schematic diagram of the HUT bubble size analyser.
 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233
(a)
Fig. 3. Images captured using the HUT BSA method: (a) frother DF-200 9.7  10
The experimental work was conducted in a 50 dm3
Outokumpu laboratory flotation cell. Bubble size
distributions were measured at only one location in
the cell, 110 mm above the plane of the stator, as
shown in Fig. 4. In a non-coalescing environment, it
can be assumed that the bubble size distribution is
entirely determined in the impeller/stator zone and it is
preserved throughout the rest of the cell. Thus, it
should be possible to collect a representative sample
irrespective of the sampling location in the cell. Even
though this assumption does not consider a likely
spatial rearrangement of the bubbles, the validity of
this hypothesis is backed up by the experiments
conducted by Jameson and Allum (1984) in a
medium-sized flotation cell and by the results
published by Parthasarathy et al. (1991).
The cell was filled with aqueous solutions of the
tested frothers (Table 1), which were prepared using
municipal tap water. The HUT sampler system was
also filled with the same aqueous solution. All the
experiments were conducted at room temperature (it
fluctuated between 19 and 22 8C). The impeller speed
was set at 1050 rpm (tip speed: 6.9 m/s) and the air
350 mm
225 mm
Fig. 4. Measurement location in the cell.
227
(b)
5 5
M (20 ppm) and (b) DF-1012 2.510 M (10 ppm).
flow rate was set at 77 dm3/min (superficial gas
velocity J g = 1.3 cm/s). On average about 4800
bubbles were sized in each run. Each measurement
was at least duplicated The number of measured
bubbles, at low frother concentrations (2–4 ppm), was
relatively small in comparison with the number of
bubbles present at high frother concentrations. The
bubble sizes measured with the use of the HUT
technique are reported at standard temperature (298
K) and pressure (101325 Pa) conditions.
The UCT bubble-sizing system comprises a glass
capillary tube and two pairs of photo-transistor-LED
detectors. Bubbles are drawn up through the capillary
tube under vacuum and collected in a modified
burette. The suction rate is controlled by a peristaltic
pump. The bubbles are transformed into cylinders
inside of the glass capillary as shown in Fig. 5, and
their velocities and length measured using the optical
detectors. The total volume of gas collected in the
burette is utilized in order to estimate the absolute size
Fig. 5. Bubble entering the glass capillary of the UCT bubble-sizing
system.
228 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233

Concentration (mmol/dm3)
0 0.04 0.08 0.12 0.16 0.2 0.24 0.28
4

Sauter mean diameter (mm)

3.5 UCT HUT

2.5

1.5

0.5

0
0 0.04 0.08 0.12 0.16 0.2 0.24 0.28
3
Concentration (mmol/dm )

Fig. 6. Graphical determination of the CCC values for Dowfroth 200.

of the bubbles. A detailed description of the instru-
ment has been published by Randall et al. (1989).
The experimental work was carried out in a 1 dm3
Open-Top Leeds flotation cell. Aqueous solutions of
the tested frothers were prepared using distillated
water. The impeller speed was set at 1000 rpm and the
air flow rate at 5 dm3/min (superficial gas velocity
J g =0.43 cm/s). Bubble size was determined utilizing
the UCT Bubble size analyser. A glass capillary
sampler of 0.8 mm in inner diameter was selected for

3
Concentration (mmol/dm )
0 0.04 0.08 0.12 0.16 0.2
4

3.5 UCT HUT

3
Sauter mean diameter (mm)

2.5

1.5

0.5

0
0 0.04 0.08 0.12 0.16 0.2
3
Concentration (mmol/dm )

Fig. 7. Graphical determination of the CCC values for Dowfroth 250.

 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233 229

the experimental work. The bubble sampler of the Table 2

UCT bubble size meter was positioned 50 mm above CCC values for the investigated frothers
the stator. All the tests were conducted at room Frother CCC, HUT technique CCC, UCT technique
temperature (21 F1 8C), and approximately 3000 mmol/dm3 ppm mmol/dm3 ppm
bubbles were sampled for each run. DF-200 0.067 13.8 0.084 17.3
DF-250 0.035 9.1 0.033 8.7
2.1. Surface tension DF-1012 0.017 6.6 0.015 6.0

Du Nouy Ring Tensiometer was used to measure
the surface tension of aqueous solutions at varying
frother concentrations.
3. Results
The results of the measurements are shown in the
form of bubble size vs. frother concentration curves in
Figs. 6–8. As these figures reveal, each curve is
clearly divided into two concentration ranges, the low
concentration range which can be approximated by a
steep linear fit and the flat horizontal portion. The
intersection of the horizontal asymptote to the Sauter
bubble size–concentration curve with the steep linear
portion can easily be determined. This is so-called
critical coalescence concentration (Cho and Laskow-
ski, 2002a,b). Only the curve for DF-200 frother
obtained using the HUT method (Fig. 6) is not linear
over the whole concentration range from Co (system
without frother) to CCC. In this case, two linear
portions could be identified over the Co to CCC
concentration range, and in line with the adopted
principle, the intersection of that linear portion which
meets the horizontal asymptote is used to determine
CCC. The experimentally determined CCC values for
the tested frothers are given in Table 2.
4. Discussion
Perhaps the most noteworthy are patterns clearly
visible in Figs. 6–8. All the experimental bubble size
vs. frother concentration curves have the same shape:
a very steep part at low frother concentrations and a
horizontal flat part at higher frother concentrations.

3
Concentration (mmol/dm )
0 0.02 0.04 0.06 0.08
4

3.5 UCT HUT

3
Sauter mean diameter (mm)

2.5

1.5

0.5

0
0 0.02 0.04 0.06 0.08
3
Concentration (mmol/dm )

Fig. 8. Graphical determination of the CCC values for Dowfroth 1012.

230
The curves can easily be fitted with two linear
segments with the intersection giving the CCC value.
Despite the differences in the size of bubbles as
measured using the HUT method and the UCT
method, very similar CCC values were obtained in
these two series of the measurements (Table 2).
While the trends are clear, Figs. 6, 7 and 8 also
show characteristic differences. The bubble sizes
obtained from extrapolation of the horizontal asymp-
totes for the UCT measurements give very similar
bubble size values for all three tested frothers (see also
Fig. 9). When the HUT method was utilized, for all
three tested frothers the larger bubbles were measured.
There may be many reasons for these differences.
First, it is to be pointed out that these two series of the
tests were carried out in different flotation machines.
Also, whereas tap water was utilized in the HUT
measurements, distilled water was utilized in the UCT
measurements. In spite of these differences almost
identical CCC values were obtained. This implies that
the CCC values for a given frother are independent of
the type of machine or mechanism used for dispersing
the air and independent of the operating conditions
(air flow rate and impeller speed) in the cell. The same
conclusion was earlier drawn by Cho and Laskowski
(2002a), as they compared CCC values measured in a
lab flotation cell and in the system comprising a three-
0 0.1
4
3.5
Sauter mean diameter (mm)
3
2.5
1.5
0.5
0
0 0.1
Fig. 9. Comparison of the bubble size vs. concentration plots for DF-frothers obtained using UCT and HUT methods.
R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233
hole sparger. That indicates that the CCC values can
be treated as material constants for frothers.
The primary factors that determine bubble size
include the following: bubble generating system (its
geometry and hydrodynamics) and superficial gas
flow rate. The secondary factors, which practically
may be more important that the primary ones, first of
all include bubble coalescence that is determined by
frother concentration (Cho and Laskowski, 2002a,b).
In mechanical cells, like the Outokumpu and the
Open-Top Leeds flotation cells, air is dispersed into
fine bubbles in the impeller/stator zone. Experimental
evidence indicates that as air is introduced into the
flotation cell, the air collects behind the blades of the
impeller forming air cavities (Grainer-Allen, 1970).
Bruijn et al. (1974) observed that the shape and the
size of the gas cavities change with the air flow rate.
Further, the air cavities enlarge as the gas flow rate
increases and eventually the air cavities can cover
entirely the rear face of the blade. Primary bubbles are
generated from the air cavities; according to Rigby et
al. (1997), gas bubbles break from the cavity due to
instabilities on the upper and lower interfaces of
cavity. The primary bubbles are further broken up
under turbulent conditions, mainly in the neighbour-
hood of the impeller and dispersed throughout the
volume of the cell. The two utilized flotation cells, the
3
Concentration (mmol/dm )
0.2
DF-200 HUT
DF-250 HUT
DF-1012 HUT
DF-200 UCT
DF-250 UCT
DF-1012 UCT
0.2
3
Concentration (mmol/dm )
 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233 231

Outokumpu cell and the Open-Top Leeds cell, have
different impeller/stator mechanisms. In addition, they
were operated under different conditions with a
superficial gas velocity J g = 0.43 cm/s in the Open-
Top Leeds cell when the UCT method was used to
follow bubble size, and J g = 1.3 cm/s in the Out-
okumpu cell when HUT method was utilized to
follow the bubble size. The influence of the gas flow
rate (superficial gas velocity J g) upon bubble size in
mechanical cells is illustrated in Fig. 10 (Grau and
Heiskanen, 2003). Because of these differences, it is
difficult to identify all the reasons that may be
responsible for dissimilarities in the sizes of bubbles.
The following is obvious, in the case of the UCT
bubble meter the size of the bubbles at the frother
concentration exceeding CCC seems to be determined
to a large extent by the measuring method. Therefore,
the differences in bubble sizes over the C N CCC
concentration range is very small for all the three
tested frothers. These differences are quite substantial
for the measurements carried out using the HUT
bubble-sizing equipment. It is surprising that the least
surface active frother (see Figs. 9 and 11) produced
the finest bubbles, and that the largest bubbles were
measured when the most surface active frother, DF-
1012, was used.
Fig. 11 shows surface tension isotherms for DF-
200, DF-250 and DF-1012 frothers. Comparison of
the surface tension data with the CCC values given
in Table 2 indicates that while for the DF-200
frother the surface tension at the CCC concentration
for this frother (8.4  10 5 M) is almost that of pure
water, the surface tension for DF-250 at its CCC
value (3.3  10 5 M) is about 66 mN/m and for the
DF-1012 (CCC =1.5  10 5 M) it is about 61 mN/
m. Since surface tension values are interrelated with
adsorption via Gibbs adsorption isotherm, this
means that the adsorption of DF-1012 at its CCC
is much larger than the adsorption of DF-250 at the
corresponding CCC values for DF-250; in turn, the
adsorption of DF-250 is much higher than the
adsorption of DF-200 at their respective CCC
values. Yet, the finest bubbles were measured with
the DF-200 frother and the coarsest with the DF-
1012 frother. It follows then that the adsorption for
larger (and more surface active) molecules of
propylene polyglycols must be larger to prevent
bubbles from coalescence. This conclusion may as
well indicate that the bstaticQ surface tension
measurements cannot be directly utilized in the
analysis of the properties of dynamic systems, as
foams. These foams are very unstable, exist only
during bubbling gas and collapse when the foam
formation process is stopped. It is likely that the
stability of such systems is determined by surface
elasticity forces. A deformation (compression or

3.5

2.5
Bubble size, mm

1.5

1 Sauter diameter d32, OK Cell, 50 dm3

Average diameter d10, OK Cell, 50 dm3

0.5
Laskowski et al. (2003), 1 dm3 Open-
Top Leeds Cell
0
0 0.5 1 1.5 2
Superficial gas velocity Jg, cm/s (STP)

Fig. 10. Effect of gas velocity upon bubble size. Measurements conducted with DF-250 at concentrations exceeding its CCC value (Grau and
Heiskanen, 2003). OK stands for Outokumpu.
232 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233

1E-005 0.0001 0.001 0.01

75 75

70 70

65 65

60 60
Surface Tension (mN/m)

55 55

50 50

45 45
DF-200

DF-250
40 40

DF-1012
35 35

30 30

1E-005 0.0001 0.001 0.01

3
Concentration (mol/dm )

Fig. 11. Surface tension isotherms for DF-200, DF-250 and DF-1012 frothers.

dilation) of the liquid–bubble interface produces 5. Conclusions

variations in surface tension, which provide a
restoring force that tends to counteract the disturb-
ance (Malysa, 1992; Harris, 1982; Leja, 1982). This
is known as the Marangoni effect. Under dynamic
conditions the coalescence must then be related not
so much to the frother adsorption as to the rate with
which it can be adsorbed (Comley et al., 2002).
Furthermore, surface elasticity forces might also
have a strong effect on bubble break-up, therefore, the
frother adsorption rate may also be controlling bubble
generation in the cell. Walter and Blanch (1986)
concluded that long hydrocarbon chain surfactants
produce larger stable bubbles than short chain
surfactants in solutions with similar surface tensions.
A similar effect is observed in the measurements
carried out with the use of the HUT bubble size
analyser (see Fig. 9); shorter chain molecules of
propylene glycols (DF-200) were found to produce
finer bubbles than the longer chain molecules of DF-
1012 propylene glycol at concentrations exceeding the
CCC values.
The following conclusions can be drawn from this
project
1. The bubble size measurements carried out using
the HUT bubble size analyser in a 50 dm3
Outokumpu lab flotation cell and the UCT bubble
size meter in a 1 dm3 Open-Top Leeds flotation cell
gave practically the same CCC values for three
Dow Frothers (DF-200, DF-250 and DF-1012).
The former tests were carried out in tap water while
the latter were conducted in distilled water.
2. Almost identical CCC values obtained when using
different equipment confirms that the CCC values
can be treated as material constants for frothers.
3. The experimental bubble size–frother concentra-
tion curves are clearly divided into two concen-
tration ranges: the low concentration range which
can be approximated by a steep linear fit and the
flat horizontal portion. Since the CCC values are
determined graphically as the intersection of the
 R.A. Grau et al. / Int. J. Miner. Process. 76 (2005) 225–233
horizontal asymptote to the Sauter bubble size–
concentration curves with the steep low concen-
tration range linear portion the precision of the
bubble size measurements apparently are not
critical for obtaining the CCC values.
4. While the obtained experimental bubble size–
frother concentration curves show the same trends
irrespective of the bubble-size measuring method
the bubble sizes measured at frother concentrations
exceeding the CCC values were much larger when
the HUT method was utilized for all the three
tested frothers.
Acknowledgments
The authors would like to thank Ms. Maaria
Kinnunen for the bubble size measurements at
Helsinki University of Technology, and Mr. Francisco
Melo for the surface tension measurements at the
University of British Columbia. The authors are
grateful to Outokumpu Technology for their contri-
bution to the research project carried out at Helsinki
University of Technology.
References
Bruijn, W., Van’t Riet, K., Smith, J.M., 1974. Power consumption
with aerated Rushton turbines. Trans. Inst. Chem. Eng. 52,
88 – 104.
Cho, Y.S., Laskowski, J.S., 2002a. Effect of flotation frothers on
bubble size and foam stability. Int. J. Miner. Process. 64, 69 – 80.
Cho, Y.S., Laskowski, J.S., 2002b. Bubble coalescence and its effect
on dynamic foam stability. Can. J. Chem. Eng. 80, 299 – 305.
Comley, B.A., Harris, P.J., Bradshaw, D.J., Harris, M.C., 2002.
Frother characterization using dynamic surface tension measure-
ments. Int. J. Miner. Process. 64, 81 – 100.
233
Grainer-Allen, T.J., 1970. Bubble generation in froth flotation
machines. Trans. Inst. Min. Metall., Sect. C 79, C15 – C22.
Grau, R., Heiskanen, K., 2002. Visual technique for measuring
bubble size in flotation machines. Miner. Eng. 15 (7), 507 – 513.
Grau, R., Heiskanen, K., 2003. Gas dispersion measurements in a
flotation cell. Miner. Eng. 16 (11), 1081 – 1089.
Harris, P.J., 1982. Frothing phenomena and frothers. In: King,
R.P. (Ed.), Principles of Flotation. South African Institute of
Mining and Metallurgy Monograph Series, pp. 237 – 263.
Johannesburg.
Jameson, G., Allum, P., 1984. A survey of bubble sizes in industrial
flotation cells, report prepared for AMIRA limited.
Laskowski, J.S., 2003. Fundamental properties of flotation frothers.
In: Lorenzen, L., Bradshaw, D.J. (Eds.), Proc. 22nd Int. Mineral
Processing Congress, vol. 2. The South African Institute of
Mining and Metallurgy, Cape Town, pp. 788 – 797.
Laskowski, J.S., Tlhone, T., Williams, P., Ding, K., 2003.
Fundamental properties of the polyoxypropylene alkyl ether
flotation frothers. Int. J. Miner. Process. 72, 289 – 299.
Leja, J., 1956/57. Mechanism of collector adsorption and dynamic
attachment of particles to air bubbles as derived from surface-
chemical studies. Trans. IMM 66, 425 – 437.
Leja, J., 1982. Surface Chemistry of Froth Flotation. Plenum
Press.
Leja, J., Schulman, J.H., 1954. Flotation theory: molecular
interactions between frothers and collectors at solid–liquid–air
interfaces. Trans. AIME 199, 221 – 228.
Malysa, K., 1992. Wet foams: formation, properties and mechanism
of stability. Adv. Colloid Interface Sci. 40, 37 – 83.
Parthasarathy, R., Jameson, G.J., Ahmed, N., 1991. Bubble breakup
in stirred vessels—predicting the Sauter mean diameter. Trans.
IChemE, Part A 69, 295 – 301.
Randall, E.W., Goodall, C.M., Fairlamb, P.M., Dold, P.L., O’Con-
nor, C.T, 1989. A method for measuring the sizes of bubbles in
two- and three-phase systems. J. Phys., Sect. E: Sci. Instrum. 22,
827 – 833.
Rigby, G.D., Evans, G.M., Jameson, G.J., 1997. Bubble breakup
from ventilated cavities in multiphase reactors. Chem. Eng. Sci.
52, 3677 – 3684.
Walter, J.F., Blanch, H.W., 1986. Bubble break-up in gas-liquid
bioreactors: Break-up in turbulent flows. Chem. Eng. J.
Biochem. Eng. J. 32, B7 – B17.