Vous êtes sur la page 1sur 8

Minerals Engineering 19 (2006) 726–733

This article is also available online at:


Frother-related research at McGill University

J.A. Finch *, S. Gelinas, P. Moyo
Department of Mining Metals and Materials Engineering, McGill University, The Wong Building, 3610 University Ave.,
Montreal, Canada H3A 2B2

Received 15 July 2005; accepted 7 September 2005

Available online 8 November 2005


Over the past ten years the Mineral Processing group at McGill University has developed techniques to determine gas dispersion
properties (gas superficial velocity, gas holdup, bubble size and bubble surface area flux) in flotation machines. This work is finding appli-
cation in metallurgical diagnostics and cell characterization. The picture, however, will remain incomplete until the impact of chemistry
on bubble production, and hence on gas dispersion, is understood. This has prompted investigations into frothers.
There are two areas addressed in this communication: frother analysis and frother characterization.
Coincident with the centenary, for 100 years there was no convenient frother analysis procedure. A colorimetric technique originally
developed for alcohols had been applied to MIBC (Parkhomovski, V.L., Petrunyak, D.G., Paas, L., 1976. Determination of methyl-
isobutylcarbinol in waste waters of concentration plants. Obogashchenie Rud 21 (2), 44–45). Using this as a starting point, the technique
was successfully extended to a wide range of commercial frothers and shown to be robust against most common ÔcontaminantsÕ. The
technique is readily used on-site and some observations from plant surveys are described.
Characterization of frothers has taken two routes, determining water carrying rate and investigating properties of thin bubble films.
Second only to transporting particles the recovery of water by bubbles has the most influence on metallurgy. The question posed was
whether this Ôwater carryingÕ property could be related to frother type. In a specially designed column the volume rate of water to the
overflow per unit cross-sectional area (Ôcarrying rateÕ, Jwo) and gas holdup (eg) at controlled froth depths were measured. The Jwo–eg
relationship proved approximately linear and dependent on frother type, with four frother ÔfamiliesÕ being identified.
Bubble thin films have been studied for soaps and the techniques were adapted for frothers. From infrared analysis it became apparent
that the frother molecule, while itself not seen, had an impact on organizing water molecules, apparently forming a film of bound water
on the bubble surface. Exploiting the interference pattern generated in UV/Vis the film thickness (d) was determined; for MIBC d was less
than 160 nm while for DF250 d was 600 nm. Taking a representative frother from the four families identified above, the water carrying
rate at a given gas holdup increased with film thickness.
Possible implications of the findings on the role of frother in bubble production are explored.
Ó 2005 Elsevier Ltd. All rights reserved.

Keywords: Flotation machines; Flotation frothers; Flotation bubbles

1. Introduction For over ten years the McGill Mineral Processing group
(and others) has been developing tools (sensors) and tech-
A rather uninspiring title but two topics which should be niques to measure gas dispersion properties (the collective
of interest describing novel methods of frother analysis and term for gas superficial velocity, bubble size, gas holdup
frother characterization will be introduced shortly. First, and bubble surface area flux) (Gomez and Finch, 2002).
however, let us consider the motivation for research in this These measurements are finding increasing application in
area. improving flotation performance (e.g., Cooper et al.,
2004; Gorain, 2005) and characterizing cells. An example
Corresponding author. Fax: +1 514 398 4492. of the latter is Fig. 1, taken from Nesset et al. (2005).
E-mail address: jim.finch@mcgill.ca (J.A. Finch). It shows bubble size (Sauter mean diameter, D32) as a

0892-6875/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733 727

3 taneous analysis de-coupled the two (and gave additional

Outokumpu TC 130 LDI information on both). The presence of butanol was as a
Bubble Size, D32 (mm)

contaminant in the xanthate, apparently a common situa-
2 tion. (What impact this alcohol had on gas dispersion (bub-
Bateman TC 50 Impala
1.5 Metso RCS 30 Impala
ble size) is a matter of speculation.)
Outokumpu 28U Raglan
Three plant surveys have been conduced to date which
have proofed the technique. After calibrating with the
0.5 plant frother and water, sampling, decantation and dilu-
tion are generally all that was necessary. (Some notion of
0 0.5 1 1.5 2 2.5 3 the dilution required is obtained from the dosage, usually
Superficial Gas Velocity, Jg (cm/s) given in g/ton, and the per cent solids.) For MIBC it was
Fig. 1. Cell characterization: Sauter mean bubble size as a function of gas
necessary to make the determination promptly as there
rate for four cell types. (Adapted from Nesset et al. (2005).) was evidence of loss over time (sealed samples brought
back to McGill showed no detectable frother) probably
due to evaporation. Current evidence is that Dowfroth
function of gas rate (superficial gas velocity, Jg) for a vari- 250 does travel.
ety of cells (all forced air). As Nesset et al. note, the rela- Two examples will serve to illustrate the technique in
tionship can only be properly interpreted when action. In Fig. 2 frother (MIBC) distribution down a bank
ÔchemistryÕ is included. This largely means frother type of 11 cells and between pulp and froth is shown. The bulk
and dosage but could extend to some collectors and high of the frother is introduced in cell 1 (>90%) and clearly dis-
concentration of some salts. The importance of frother in tributes down the bank, the concentration in the pulp
determining bubble size is well documented: Harris liquor remaining 5–7 ppm. The frother did concentrate
(1976) at the A. M. Gaudin Memorial Symposium noted: in the froth, notably towards the rear of the bank.
‘‘ . . . major factor determining bubble size . . . is neither In Fig. 3 a distribution of frother down the 16-cell bank
impeller speed nor air rate but frother concentration’’, was employed, with 50% in the first cell, 25% in cell 5, 15%
and recently Cho and Laskowski (2002) have used the in cell 9 and 10% in cell 13. Process water (PW and QW)
effect on bubble size to characterize frothers. showed residual MIBC was returning to the circuit. The
As steps towards including frother function in our gas distribution scheme showed as step increases in concentra-
dispersion work two topics are discussed in this communi- tion at each addition point (except the last) rising from
cation, frother analysis and frother characterization. 3.5 ppm to 5.5 ppm. The tails still showed 4 ppm.
The concentration was determined at two positions in some
2. Frother analysis: procedure and plant experiences cells, just below the froth (i.e., top, T) and near the impeller
(bottom, B) to check mixing in the cell. Clearly the frother
For the one hundred years of flotation there has been no is well distributed in the cell.
convenient, on-site method of determining frother concen- Although mass balancing was not performed, the evi-
tration. This has arguably restricted engineering the use of dence is that the bulk of frother remains in the pulp water.
frother to achieve metallurgical targets. Approaches have Tsatouhas et al. (2005) have made the same observation. It
included: gas chromatography (Tsatouhas et al., 2005), corresponds to the common experience that frother obvi-
determining total organic carbon (Hadler et al., 2005), ously works its way through a circuit from a limited num-
and using calibrations of bubble size and gas holdup vs. ber of addition points (sometimes only one, at the head
frother concentration (Weber et al., 2003). A literature end). The observation is re-visited below.
search identified a colorimetric method used for MIBC One application of this new tool is to integrate with gas
(methyl-iso-butyl-carbinol) in wastewaters (Parkhomovski dispersion measurements. The metallurgical gain applying
et al., 1976). The technique was based on the Komarowsky
reaction employed in colorimetric analysis of aliphatic
alcohols (Coles and Tournay, 1942).
The procedure, described in detail elsewhere (Gelinas 20
and Finch, 2005), involves dehydration with sulphuric acid froth
MIBC, ppm

followed by reaction with an aldehyde to form a coloured 15

compound. The technique proved amenable to a wide 10

range of frothers. Linear calibrations are given over a range
(e.g., up to 5 ppm Dowfroth 250) and there was no interfer- 5
ence from a range of inorganic and organic reagents and 0
contaminants typically encountered in sulphide flotation. 1 2 3 4 5 6 7 8 9 10 11
Site work encountered two compounds that did react, fuel Cell Number

oil and butanol. However, the absorbance spectra were suf- Fig. 2. Example of frother analysis: Distribution of frother down a bank
ficiently different from the frother that conventional simul- and between pulp and froth.
728 J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733

Fig. 3. Example of frother analysis: Distribution down a bank and within cells.

a pattern in gas rate to cells in a bank (ÔJg profilingÕ) (Coo- depth, Hf (increasing as Hf decreases) and, we can hypoth-
per et al., 2004; Gorain, 2005) may have a counterpoint in esize, frother type. To isolate frother type it is necessary to
profiling frother. A combination of gas rate and frother (to control froth depth and account for gas rate and especially
control bubble size, Db) may effectively mean the bubble bubble size effects as frother can modify Db and thus alter
surface area flux (Sb) is profiled, perhaps fundamentally water recovery.
the more appropriate metric than Jg (or Db). Care needs The experimental setup is shown in Fig. 4. The column
to be exercised however, as frother is added to control is equipped with an array of conductivity electrodes and
not only bubble size but also froth properties. The question facility to control air and feed water rate and measure over-
goes beyond just distributing frother to distributing frother flow water rate. The water carrying rate was defined as the
function, control of bubble size and froth properties. We overflow rate divided by the column cross-section, Jwo. The
need tools to measure frother functions to augment frother conductivity signals were used both to control froth depth
analysis. These are steps to what may be termed Ôgas dis- (by indicating the froth/water interface and manipulating
persion managementÕ designed to improve flotation circuit feed rate) and to determine gas holdup (eg) employing
performance by controlling the physical variables. MaxwellÕs model. The gas holdup reflects both gas rate
(increasing as Jg increases) and bubble size (increasing as
3. Frother characterization

This topic is attracting growing attention, mainly in

accord with trying to understand the froth zone to devise
more robust flotation models. While fully recognizing the
range of frother (and froth) characterization work there Hf
is no attempt to review it here. Rather two novel Water Overflow
approaches are described based on water transport and Rate = Qwo
properties of bubble thin films. The observations have
prompted speculation on how frothers may function to Conductivity
Jwo= Qwo/ A
gas holdup, eg
control bubble size, which ends the paper.
Feed (water)
3.1. Water carrying rate
Bubbles transport particles and water, which together
dictate the metallurgy (recovery and grade). Restricting
Air, Jg
attention to air–water only, the factors that influence water
recovery include: gas rate (recovery increasing as Jg Fig. 4. Apparatus to measure water carrying rate as a function of gas
increases), bubble size (increasing as Db decreases), froth holdup.
J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733 729

Hf = 7 cm
Free Water
Bound Water
Jwo, cm/s

Hept, Hex, MIBC, DF200

0.2 Octanol, Ethoxy, DF250 Air Bubble



10 15 20 25 30 35 Bulk Water
εg, % (a) (b)

Fig. 5. Water carrying rate as a function of gas holdup showing four Fig. 7. An interpretation of the physical picture based on IR results.
classes of frother. (a) Bubble in air, (b) bubble in water.

Db decreases). It was argued that this metric could serve as depiction of the bubble surface. The spectrum is interpreted
the independent variable to isolate the frother type effect. in terms of water structure. Some peaks are identified with
This seems to be the case. Fig. 5 shows the summary ÔfreeÕ water others with ÔboundÕ water indicative of H-bond-
result for a range of frothers. The plot Jwo vs. eg was ing induced by the polar group of the frother (e.g., OH in
approximately linear (with an intercept, a minimum gas the case of alcohols). The picture this suggests, Fig. 7a, is a
holdup below which no overflow at the given froth depth surface with an inner bound water layer surrounded by an
was achieved) and divided the frothers into 4 classes (or outer free water layer. Fig. 7b translates this to the situa-
ÔfamiliesÕ). As representatives, in order of increasing water tion of interest, a bubble blown in water; the bound later
carrying rate we have: pentanol < MIBC < Dowfroth remains but now the free water is the bulk water. While
250 < F150 (similar to Dowfroth 425). We return to this the interpretation of water structure from infrared spectra
in the next section. is problematic (Nancer et al., 2000), the picture is not unex-
pected since it is known that the polar group will induce
3.2. Bubble thin films local re-organization of the water structure. In the case of
alcohols, this effect, and the consequences, can be quite
The procedure described for soap bubbles (Rödel, 1981; extensive (Wakisaka et al., 2001). The question is how large
Chattopadhyay, 2000) were adapted to study bubbles is this layer?
blown in frother solutions (Gelinas et al., 2005). The bub- This can be addressed from the interference pattern gen-
ble was probed by infrared (for structure) and UV/Vis (for erated in the UV/Vis region. Results for MIBC and Dow-
film thickness using interferometry). froth 250 are given in Fig. 8. The measured (left) is the
The general infrared (FT-IR) spectrum is shown in combined bound and free water layers, which are de-cou-
Fig. 6. The first observation is that no fingerprint of the pled (right). The layer thickness, d, is estimated from the
frother molecule is seen, implying a low concentration following for a beam passing through the center of the bub-
(adsorption density) which contrasts with the common ble (i.e., at right angles to the surface):

Fig. 6. The general FT-IR spectrum for frother bubble film.

730 J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733

Bound water

<160 nm

A 200 400 600 800 1000

0.10 Free water


0.08 200 400 600 800 1000

200 400 600 800 1000

Wavelength (nm)

Bound water

~600 nm
500 600 700 800 900 1000 1100

Free water


500 600 700 800 900 1000 1100

500 600 700 800 900 1000 1100
Wavelength (nm)

Fig. 8. The interference pattern generated in UV/V is region by bubble film. Left: measured pattern; right, de-coupled ÔboundÕ and ÔfreeÕ water layers with
estimate of bound water thickness.

DN f k1 k2 modifies this much like changing from a hydrophobic to a

d¼ ; ð1Þ
2nðk1  k2 Þ hydrophilic solid will alter the adjacent water structure.
Table 1 gives results for the frothers representative of
where Nf is the number of cycles between wavelength k1
the four families identified in Fig. 5 along with the water
and k2 and n is the refractive index of water (assumed
carrying rate (at equal gas holdup). The macroscopic mea-
1.33 and independent of frother addition and of wave-
sure of water transport appears to be related to the micro-
length). The high frequency pattern is thus associated with
scopic (fundamental) measure of bound water film
the thicker film, attributed to the free water. The thickness
is typically 6–7 lm, common to both frothers and soaps
(Chattopadhyay, 2000). In contrast, the bound water thick-
4. Frother and bubble production
ness (low frequency pattern) is different for the two fro-
thers, for MIBC d < 160 nm (ÔdÕ is below the k range of
Let us try to put the observations together to propose a
the instrument) compared to Dowfroth 250, d  600 nm.
mechanism of frother action in bubble production. A com-
The magnitude seems quite surprising, implying the influ-
mon theory is that frothers prevent coalescence, i.e., there
ence on water structure extends tens even hundreds of
is a primary process that generates fine bubbles and frother
interrupts a secondary process, coalescence.1 Coalescence
An attempt to depict the physical (molecular) picture
requires the out flow of liquid between adjacent bubbles.
suggested by the findings is given in Fig. 9. A lone MIBC
molecule assembles water molecules into an array with
some order extending about 160 nm. As a start to interpret- Coalescence itself may have two sub-categories. Coalescence of the
ing the larger film thickness of Dowfroth 250 it is depicted type described by Tse et al. (2003) occurs away from the point of
anchored via the polar end group with the chain extending generation and continues as long as bubbles encounter one another. This
is a relatively minor contributor to the overall bubble size compared to the
into the water, an orientation perhaps supported by a sec- coalescence of interest here, namely that associated with the moment of
ond molecule. An air surface, like a solid surface, perturbs bubble production. The former might be called secondary coalescence, the
the local water structure; the suggestion here is that frother latter, primary coalescence.
J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733 731

Fig. 10. Artists impression of gas plume downstream of rotating impeller

blade and formation of bound water patches due to frother which may
stabilize bubbles against coalescence at the moment of production (by
shear in this case). (Adapted from Schubert and Bischofberger, 1998.)
Reprinted from International Journal of Mineral Processing, Vol. 52, with
permission from Elsevier.

over. This stabilized bubble concept is depicted in the

Ôexploded viewÕ in Fig. 10.
This interpretation accommodates several features: that
Fig. 9. Interpretation of formation of bound water layer resulting from
presence of frother.
apparently little frother is needed at the surface, indicated
by the plant frother analysis results and the FT-IR bubble
film spectra, and that the action is virtually instantaneous,
coincident with the moment of bubble production. The
Table 1
Water carrying rate and bubble film thickness frother may be considered, by altering local bulk water
properties, to have ÔconditionedÕ the water to become an
Frother Water carrying rate, Bound water
Jwo (cm/s) (at eg = 25%, Fig. 5) film thickness (nm)
anti-coalescence medium. The frotherÕs modest surface
activity then helps it stay at the air surface, which is a plus.2
n-Pentanol 0.07 <160
MIBC 0.18 <160
This argues against the periodic attempts to introduce
DF 250 0.25 600 frother via the air (e.g., as an aerosol) to get it more rapidly
F 150 0.35 1100 to the air–water interface; it is actually needed in the water.
On a practical side it seems as well that frothers are not
strongly surface active otherwise, given the large surface
area of gas injected into flotation cells, they would be
The mechanisms invoked that may retard (even reverse the depleted rapidly, making it difficult to keep sufficient in
flow) include Gibbs elasticity and the Marangoni effect dri- the system to perform the functions. (Frother remaining
ven by surface tension gradients and increased surface vis- in the water may be undesirable from environmental and
cosity due to structural effects deriving from the nature of water recycling perspectives but those are other matters.)
the surfactant. Mostly these are discussed in terms of froth The magnitude of anti-coalescence will be a function of
stability, a relatively quiescent region compared to the concentration, preserving the concept of a critical coales-
dynamic, turbulent region where bubbles are generated. cence concentration (CCC) as a useful aid to thought
A speculation is offered on the latter. (Cho and Laskowski, 2002).
An artistÕs impression of the plume of air downstream of While an anti-coalescence property of frothers can be
a rotating impeller blade is depicted in Fig. 10. This is demonstrated there are some questions applying it to bub-
injected into a solution of frother, whose molecules, by ble production. One is the numbers involved. In a typical
analogy with the picture in Fig. 9, are surrounded by a experiment it is common for bubble size to reduce from
region of ÔorganizedÕ water. By chance a nascent bubble 2.5 to 0.5 mm as frother is added up to the CCC. In
shearing from the air surface encounters a frother mole- the coalescence model this means 0.5 mm is the primary
cule, or at least its Ôsphere of influenceÕ on the water struc- generation (i.e., starting) size. Since volume is preserved
ture. The bubble surface now has a ÔpatchÕ of bound water this calculates to one hundred 0.5 mm bubbles ((2.5/0.5)3)
which may be sufficient to hinder the drainage of the water
film between adjacent bubbles and prevent coalescence. 2
This modest surface activity may be the factor enabling frothers to
(This might be classed as a structural effect.) Once the bub- retard coalescence at much lower concentration that inorganic ions (salts)
bles start to move apart the opportunity for coalescence is some of which likewise can modify water structure and retard coalescence.
732 J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733

Fig. 11. Evidence of small bubble generated as a consequence of coalescence. Reprinted from Chemical Engineering Science, Vol. 58, Tse et al. (2003) with
permission from Elsevier.

coalescing to form the 2.5 mm bubble in the absence of materially to flotation? Like most research often more
frother. While not impossible it goes give pause for questions than answers arise.
The alternative mechanism is break-up of the air stream. 5. Conclusions
From just observing the end result, fine bubbles, it is not
possible to distinguish an anti-coalescence action from Bubble production in flotation machines results from
one of break-up. Perhaps we should entertain that frother the interaction of the generating mechanism and chemistry,
may condition the water as an air break-up medium. most importantly that of frother. A practical contribution
Opening that door leads to the possibility of both anti- is made introducing a frother analysis technique, for the
coalescence and break-up mechanisms being at play. Such first time suited to on-site determinations. Combined with
an observation is reported by Tse et al. (2003). The act of gas dispersion measurements, this promises better manage-
two bubbles coalescing was captured (Fig. 11) which ment of flotation systems. A fundamental contribution is
momentarily produced a sausage-like object (third image made measuring water carrying rate and investigating bub-
in sequence) that recoiled to recover a more spherical shape ble thin film properties as a function of frother type. A
(to minimize interfacial energy) and so doing shed a small bound water layer is identified with a thickness that corre-
bubble, 0.1 mm. This invites some questions: what is the lates with water carrying rate. An anti-coalescence mecha-
role of frother (apparently small bubbles persist (Machon nism based on bound water interrupting drainage between
et al., 1997)); will the sampling followed by imaging tech- adjacent bubbles is suggested. This accommodates the fact
nique capture these small bubbles; do they contribute that the bulk of frother remains in the water.
J.A. Finch et al. / Minerals Engineering 19 (2006) 726–733 733

Acknowledgments Hadler, K., Aktas, Z., Cilliers, J.J., 2005. The effects of frother and collector
distribution on flotation performance. Miner. Eng. 18, 171–177.
Harris, C.C., 1976. Flotation machines. In: Fuerstenau, M.C. (Ed.), A.M.
This work is funded by ÔChair RenewalÕ grant, under the Gaudin Memorial Volume, vol. 2. SME of AIME, pp. 753–815.
Collaborative Research and Development program of Machon, V., Pacek, A.W., Nienow, A.W., 1997. Bubble sizes in electrolyte
NSERC (Natural Sciences and Engineering Research and alcohol solutions in a turbulent stirred verssel. Trans. Inst. Chem.
Council of Canada) sponsored by Inco, Noranda, Falcon- Eng. 75 (Part a), 339–348.
bridge, Teck Cominco, COREM, and SGS Lakefield Re- Nancer, M., Sperline, R.P., Miller, J.D., 2000. Anomolous dispersion
effects in the IR-ATR spectroscopy of water. Appl. Spectrosc. 54 (1),
search. The ideas have germinated from discussions with 138–143.
a series of talented students and colleagues. Points raised Nesset, J.E., Hernandez-Aguilar, J.R., Acuna, C., Gomez, C.O., Finch,
by Prof. Jan Miller after the Keynote presentation are also J.A., 2005. Some gas dispersion characteristics of mechanical flotation
gratefully acknowledged. The engineers and operators who machines. In: Jameson, G., Yoon, R.-H. (Eds.), Centenary of
enthusiastically supported the frother analysis site work are Flotation. Proceedings of International Conference. SME-AusIMM,
Brisbane, pp. 243–250.
also recognized. Parkhomovski, V.L., Petrunyak, D.G., Paas, L., 1976. Determination of
methylisobutylcarbinol in waste waters of concentration plants.
References Obogashchenie Rud 21 (2), 44–45.
Rödel, K., 1981. Proof of boundary surface water and its effect on
Chattopadhyay, A., 2000. Time-dependent changes in a shampoo bubble. colloidal-chemical properties: IR spectroscopy of foam films. Tenside
J. Chem. Educ. 77 (10), 1339–1343. Det. 18 (3), 141–148.
Cho, Y.S., Laskowski, J.S., 2002. Effect of flotation frothers on bubble Schubert, H., Bischofberger, C., 1998. On the microprocesses air disper-
size and foam stability. Int. J. Miner. Process. 64, 69–80. sion and particle-bubble attachment in flotation machines as well as
Cooper, M., Scott, D., Dahlke, R., Gomez, C.O., Finch, J.A. 2004. Impact consequences for the scale-up of macroprocessses. Int. J. Miner.
of air distribution profile on banks in Zn cleaning circuit, CIM Bull. Process. 52, 245–259.
Vol. 97 (1083), extended abstract p. 73, full paper. Available from: Tsatouhas, G., Grano, S.R., Vera, M., 2005. Case studies on the
<www.cim.org>. performance and characterization of the froth phase in industrial
Coles, H.W., Tournay, W.E., 1942. The Komarowsky color reaction for flotation circuits. In: Jameson, G., Yoon, R.-H. (Eds.), Centenary of
aliphatic alcohols. Ind. Eng. Chem. Anal. Ed. 14 (1), 20–22. Flotation. Proceedings of International Conference. SME-AusIMM,
Gelinas, S., Finch, J.A., 2005. Colorimetric determination of common Brisbane, pp. 377–384.
industrial frothers. Miner. Eng. 18 (2), 263–266. Tse, K.L., Martin, T., McFarlane, C.M., Nienow, A.W., 2003. Small
Gelinas, S., Finch, J.A., Gouet-Kaplan, M., 2005. Comparative real-time bubble formation via a coalescence dependent break-up mechanism.
characterization of frother bubble thin films. J. Coll. Interface Sci. 291 Chem. Eng. Sci. 58, 275–286.
(1), 187–191. Wakisaka, A., Komatsu, S., Usui, Y., 2001. Solute–solvent and solvent–
Gomez, C.O., Finch, J.A., 2002. Gas dispersion measurements in flotation solvent interactions evaluated through clusters isolated solutions;
machines. CIM Bull. 95 (1066), 73–78. preferential salvation in water–alcohol mixtures. J. Mol. Liq. 90, 175–
Gorain, B., 2005. Optimisation of flotation circuits with large flotation 184.
cells. In: Jameson, G., Yoon, R.-H. (Eds.), Centenary of Flotation. Weber, T., Gomez, C.O., Finch, J.A., 2003. A frother concentration
Proceedings of International Conference. SME-AusIMM, Brisbane, meter, In: Proceedings of the 35th Annual Meeting of the Canadian
pp. 843–852. Mineral Processors (division of CIM), Jan 21–23, pp. 639–652.